Relating Events in Narrative, Volume 2: Typological and Contextual Perspectives

RELATING EVENTS IN NARRATIVE VOLUME 2: TYPOLOGICAL AND CONTEXTUAL PERSPECTIVES This page intentionally left blank R...

Author: Sven Strmqvist | Ludo Verhoeven | Sven Stromqvist

39 downloads 906 Views 27MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

RELATING EVENTS IN NARRATIVE VOLUME 2: TYPOLOGICAL AND CONTEXTUAL PERSPECTIVES

This page intentionally left blank

RELATING EVENTS IN NARRATIVE VOLUME 2: TYPOLOGICAL AND CONTEXTUAL PERSPECTIVES

SVEN STROMQVIST Department of Linguistics Lund University, Sweden

and

LUDO VERHOVEN Department of Social Sciences University ofNijmegen, The Netherlands

In collaboration with Ayhan Aksu-Ko^, Edith Bavin, Ruth Berman, Frederikke Blytmann Trondhjem, Penelope Brown, Herbert Clark, Elisabeth Engberg-Pedersen, Dennis Galvan, Maya Hickmann, Iraide Ibarretxe-Antunano, Aylin Kiintay, Judy Kupersmitt, Kei Nakamura, Asa Nordqvist, Hrafnhildur Ragnarsdottir, Dan Slobin, Sarah Taub, Goklem Tekdemir, Asa Wengelin, David Wilkins, Peerapat Yangklang, and Jordan Zlatev

2004

LAWRENCE ERLBAUM ASSOCIATES, PUBLISHERS Mahwah, New Jersey London

The cover illustration is based on a "frog-crossing" road sign in the Austrian Alps, first noted by Nini Hoiting, and photographed by Dan Slobin.

Copyright © 2004 by Lawrence Erlbaum Associates. Inc. All rights reserved. No part of this book may be reproduced in any form, by photostat, microform, retrieval system, or any other means, without prior written permission of the publisher. Lawrence Erlbaum Associates. Inc., Publishers 10 Industrial Avenue Mahwah, New Jersey 07430 Cover design by Kathryn Houghtaling Lacey Layout and Editorial Assistance for this book was provided by Marcus Uneson. Library of Congress Cataloging-in-Publication Data Stromqvist, Sven. Relating events in narrative : typological and contextual perspectives : volume 2 of the frog story studies / Sven Stromqvist, Ludo Verhoeven. p. cm. Continuation of a study published in 1994 under the title: Relating events in narrative : a crosslinguistic developmental study. Includes bibliographical references and index. ISBN 0-8058-4672-7 (alk. paper) 1. Discourse analysis, Narrative. 2. Typology (Linguistics). 3. Context (Linguistics). I. Verhoeven, Ludo Th. II. Title. P302.7.S77 2004 401'.41—dc22 2003058420 CIP Books published by Lawrence Erlbaum Associates are printed on acid-free paper, and their bindings are chosen for strength and durability. Printed in the United States of America 10 9 8 7 6 5 4 3 21

Contents Contributors

iii

Preface and Acknowledgements

vii

Prologue Typological and Contextual Perspectives on Narrative Development Sven Stromqvist & Ludo Verhoeven

Part I: Typological Perspectives

1 3

15

Focusing on 'where' An Analysis of Warlpiri Frog Stories Edith L. Bavin

17

Position and Motion in Tzeltal Frog Stories The Acquisition of Narrative Style Penelope Brown

37

Focus on Action in Motion Descriptions The Case of West-Greenlandic Elisabeth Engberg-Pedersen & Frederikke Blytmann Trondhjem

59

Motion Events in Basque Narratives Iraide Ibarretxe-Antunano

89

Time, Space, and Manner in Swedish and Icelandic Narrative Construction in Two Closely Related Languages Hrafnhildur Ragnarsdottir & Sven Stromqvist

113

The Verbalization of Motion Events in Arrernte David P. Wilkins

143

A Third Way to Travel The Place of Thai in Motion-Event Typology Jordan Zlatev & Peerapat Yangklang

159

The Encoding of Motion Information in American Sign Language Dennis Galvan & Sarah Taub

191

The Many Ways to Search for a Frog Linguistic Typology and the Expression of Motion Events Dan I. Slobin

219

Part II: Contextual Perspectives

259

The Role of Context in Developing Narrative Abilities Ruth A. Berman

261

Coherence, Cohesion, and Context Some Comparative Perspectives in Narrative Development Maya Hickmann

281

Interplay between Narrativity and Mindreading A Comparison between Turkish and English Ayhan Aksu-Koc & G ok I em Tekdemir

307

Linguistic Strategies Serving Evaluative Functions A Comparison between Japanese and Turkish Narratives Aylin C. Kilntay & Keiko Nakamura

329

Writing the Frog Story Developmental and Cross-Modal Perspectives Sven Stromqvist, Asa Nordqvist & Asa Wengelin

359

Form-Function Relations in Spanish Narratives A Comparative Study of Bilingual and Monolingual Speakers Judy Kupersmitt

395

Bilingualism and Narrative Construction Ludo Verhoeven

435

Epilogue

455

Variations on a Ranarian Theme Herbert H. Clark

Appendices Appendix Appendix Appendix Appendix Appendix

457

477 I: Frog, where are you? Ha: Frog-Story Research as per 1994 lib: Frog-Story Research after 1994 He: Typological Overview of Frog-Story Research III: Format and Abbreviations for Glosses

479 487 501 515 519

References

523

Subject Index

575

Person Index

595

Contributors Ayhan Aksu-Ko9 Dept. of Psychology Bogagici University P.K. 2 Bebek 80815 Istanbul Turkey email: [email protected]

Herbert Clark Dept. of Psychology Jordan Hall, building 420 450 Serra Mall Stanford University Stanford, CA 94305 USA email: [email protected]

Edith Bavin

Elisabeth Engberg-Pedersen Dept. of General and Applied Linguistics University of Copenhagen Njalsgade 80 DK-2300 Copenhagen S. Denmark email: [email protected]

School of Psychological Science La Trobe University Bundoora, Victoria 3083 Australia email: [email protected]

Ruth Berman Dept. of Linguistics Tel Aviv University Ramat Aviv Israel 69978 email: [email protected]

Dennis Galvan Dept. of Psychology Gallaudet University 800 Florida Avenue, NE Washington, DC 20002-3695 USA email: [email protected]

Frederikke Blytmann Trondhjem Dept. of General and Applied Linguistics University of Copenhagen Njalsgade 80 DK-2300 Copenhagen S. Denmark

Maya Hickmann

Penelope Brown

Iraide Ibarretxe-Antunano Facultad de Filosofia y Letras. Filologia Inglesa Universidad de Deusto/Deustuko Unibertsitatea Apartado 1 E-48080 Bilbao Spain email: [email protected]

Max Planck Institute for Psycholinguistics PB310 NL-6500 AH Nijmegen The Netherlands email: [email protected]

Lab. Cognition et Developpement, CNRS Institut de Psychologie University Rene Descartes - Paris V 71, Avenue Edouard Vaillant 92774 Boulogne Billancourt cedex France email: [email protected]

Contributors

Aylin Kiintay Dept. of Psychology Ko? University Rumeli Feneri Yolu, Sariyer 80910 Istanbul Turkey email: [email protected]

Dan Slobin Dept. of Psychology University of California Berkeley, CA 94720 USA email: [email protected]

Judy Kupersmitt POB 1292 Givat Ada 37808 Israel email: [email protected]

Sven Stromqvist (editor) Dept. of Linguistics University of Lund Helgonabacken 12 SE-223 62 Lund Sweden email: [email protected]

Kei Nakamura

Sarah Taub

Institute of Cognitive & Linguistic Studies Keio University 2-15-45 Mita, Minato-ku Tokyo 108-0073 Japan email: [email protected]

Dept. of Linguistics and Interpretation Gallaudet University 800 Florida Avenue. NE Washington. DC 20002-3695 USA email: [email protected]

Asa Nordqvist Dept. of Linguistics Goteborg University Box 200 SE-405 30 Goteborg Sweden email: [email protected]

Goklem Tekdemir Dept. of Psychology Bogagici University P.K. 2 Bebek 80815 Istanbul Turkey email: [email protected]

Hrafnhildur Ragnarsdottir Iceland University of Education StakkahliD IS-105 Reykjavik Iceland email: [email protected]

Ludo Verhoeven (editor) University of Nijmegen Montessorilaan 5 P.O. Box 9104 65000 HE The Netherlands email: [email protected]

Contributors

Asa Wengelin Dept. of Linguistics University of Lund Helgonabacken 12 SE-223 62 Lund Sweden email: [email protected]

David Wilkins Center for Aphasia and Related Disorders V.A. Northern California Health Care System I S O M u i r R o a d 126(s) Martinez, Ca 94553-4695 USA email: [email protected]

Peerapat Yangklang Dept. of Linguistics, Faculty of Arts Chulalongkorn University, Phyathai Rd. Bangkok 10330 Thailand email: [email protected]

Jordan Zlatev Dept. of Linguistics University of Lund Helgonabacken 12 SE-223 62 Lund Sweden email: [email protected]



VARIATIONS ON A RANARIAN THEME There have been rumors for the past several years that the worldwide frog population is declining, perhaps due to effects of global warming. But this book shows that the worldwide frog-story population is growing and thriving. This will be of little comfort to lovers of amphibians, but a cause for pleasure to students of cognitive linguistics, developmental psychology, and narratology. Two decades ago the two of us started our search for the frog, at first using Mercer Mayer's Frog, where are you? as a lure to speakers of Hebrew and English to reveal to us something of their understanding of temporality. That project expanded to include German, Spanish, and Turkish. And it soon became evident that the frog story was a remarkable device for exploring a range of cognitive, developmental-psycholinguistic, and narrative topics. We put that all together a decade ago, in Relating events in narrative: A crosslinguistic developmental study (Berman & Slobin, 1994). Since then, a lot of little tadpoles have grown into frogs - in such unexpected cold places as Greenland and Iceland, as well as more likely tropical regions like Thailand and Yucatan. These new frogs jump about energetically and provide fresh insights in the present volume, which we like to think of as "Frog II." As proud foster-parents of this amphibious exuberance, we welcome the frogs that have been gathered together by our friends and colleagues, Sven Stromqvist and Ludo Verhoeven. Ruth Berman & Dan Slobin

* The OED (2 nd Edn.) entry for ranarian is: ra'narian, a. nonce-wd. 'Froggy'. [Mod.L., f. rdna frog] 1814 T. L. PEACOCK. Sir Proteus i. Note, Ranarian minstrels of all ages and nations have entertained a high opinion of their own melody.


TRANSFORMATIONS ON A RANARIAN THEME In one of his Metamorphoses, Ovid tells the story of a group of malevolent peasants who, by the force of a spell cast by the goddess Latona, were turned into frogs. Living their transformed life in a pool, they are described by the poet in the famous line quamvis sint sub aqua, sub aqua maledicere temptant ('though they may be under water, even under the water they try to utter maledictions')*, where the chiastic reduplication sub aqua, sub aqua ('under water, under water') serves to mimetically amplify their new mode of communication. Working with the new set of frog-story studies has been a truly enchanting journey, felicitously among benevolent colleagues, with lots of charms as well as a few spelling mistakes evoking one or two maledictions. How we, the editors, will ultimately be perceived after numerous transformations is entirely for the readership to decide. We want to acknowledge our source of inspiration, Ruth Berman and Dan Slobin, for their constant parental support, and our many colleagues and peer tadpoles (or, indeed, toad pals) who have made this second collection of frogstory studies possible; Cathleen Petree and Art Lizza of Erlbaum, whose good advice and patient rescheduling have helped us keep our editorial noses just above the surface of the water; Johan Segerback for proof-reading; and our editorial assistant Marcus Uneson for his magic touch, making the manuscript visually appetizing.

Sven Stromqvist & Ludo Verhoeven Editors

*Ovid, Metamorphoses, Book VI, line 376. English translation by Frank Justus Miller. Cambridge, Mass.: Harvard University Press, MCMLVI.

Prologue


Typological and Contextual Perspectives on Narrative Development

Sven Stromqvist & Ludo Verhoeven

1. INTRODUCTION This book is a collection of articles converging on a common theme: the comparative study of elicited narrative production - across languages, cultures, modalities (speech, sign, writing), and settings. It is a follow-up volume to Berman and Slobin's Relating events in narrative: A cross-linguistic developmental study (1994). Working closely with Ruth Berman and Dan Slobin, we (the editors) have brought together a wide range of scholars who, inspired by the 1994 book, have all used Mercer Mayer's wordless picture booklet Frog, where are you? (Mayer 1969) as a basis for their research. Narratives can be seen as an important device for human communication. People use narratives to recapitulate past experience by matching a verbal sequence of clauses to the sequence of events which actually occurred. Thus, the reproduction of narratives can be seen as a major vehicle for cultural transmission. In order to understand the process of narrative (re)production, the question of how narratives can be represented becomes highly relevant. An important question is how children acquire the ability to (re)produce narratives. Studies of language acquisition have made clear that children typically command the grammatical principles and rules governing their native language by the age of four years (see, for example, Slobin 1985; Radford 1995). Several studies have also shown language development to continue into the school years at the levels of both grammar and language-use preferences. The later language development of children can be characterized by a growing command of discourse (Berman 1998a; Nippold 1998; Berman & Verhoeven 2002a). According to KarmiloffSmith (1995), narrative can be seen as one of the most significant domains of later language acquisition. Around the age of five years, critical developmental shifts take place from basic structures to additional functions and from


clause juxtaposition to clause-combining devices. In the process of narrative development, children must leam that linguistic forms are multifunctional and that the use of any particular form is multiply defined. Berman (1997a) has explored what preschool children know about language structure and language use. By the age of five, children are adept at combining clauses and have mastered a great deal of complex syntax; they can also construct sequentially well-organized narratives and express different perspectives on events. In other research reviewing a large number of crosslinguistic studies on discourse organization in the domains of person, space, and time, Hickmann (1995) concluded that the progression in narrative organization involving inter-relations between the sentence and discourse levels is relatively late. The shift from exophora to endophora appears to be the main process acquired by school-aged children to displace reference and, more generally, to use narrative as a new kind of context in which to anchor utterances when necessary. In their ground-breaking volume, Relating events in narrative (1994). Berman and Slobin presented a window on language, thought, and development. Studying children in the age range from three to nine and adults, they unvealed developmental differences in the linguistic construction of events in narrative. They also discovered differences with respect to event-encoding and narrative emplotment in five different languages: English, German, Hebrew, Spanish, and Turkish. The basis for this research is the re-telling of the so-called frog story (Mayer 1969), a series of 24 pictures representing a number of dynamic interactions between animate beings over time and in different physical settings, by narrators of different ages and different languages. In the present volume, typological and contextual perspectives on narrative construction and its development will be reviewed. The narrative, as displayed by the frog story, is taken as the main point of reference. With respect to the typological perspective, the analyses in the present volume are focused on the organizing processes used by children and adults to produce narratives in a typologically broad variety of languages. With regard to the contextual perspective, the focus on narrative construction by means of oral retellings of the frog story in monolingual, middle-class subjects is counterbalanced with variation across genres, modalities, and populations.

2. HOW TO TELL THE FROG STORY The recipe for frog-story research is simple. Ask your subjects to browse through the picture booklet (the 24 component pictures of Frog, where are you? are presented in Appendix I). Then ask them to tell the story, in their own words, as they page through the booklet at their own pace. Repeat the data-collection procedure with subjects of a different language or age, or with subjects with impaired language development. Or repeat the procedure with the same subjects


in a different condition (for example speaking versus writing or first versus second language). Then compare the results. The method is simple, yet powerful. The subjects' ways of telling the frog story is revealing of their age, the informationstructuring qualities characteristic of their language, their skills or problems, etc. It is therefore not surprising that the method has gained a great deal of popularity. An overview of research based on the frog story is presented in Appendices Ila (which is a reprint of the research overview given in the 1994 volume) and lib (which lists later additions). These appendices contain descriptions of more than 160 projects covering a range of 72 different languages and 13 different phyla (language families). Appendix He gives a typological overview of these languages in terms of a genetic-geographic classification. A notational system for coding the considerable grammatical and morphological diversity represented by these languages is given in Appendix III. 1 The method has several advantages. First, you will find that, with few exceptions, your subjects enjoy the story-telling task - whether they are young or old, from Western urban or Austronesian rural culture. Probably, the type of plot represented by the frog story - a journey and a search - is a crossculturally pertinent theme. It may be that telling stories in relation to pictures - from cave paintings to holiday photographs - is a deeply ingrained practice in many cultures. Whatever the explanation, the method works well. Second, the data elicited through the frog story represents connected meaningful discourse. This allows the analyst not only to perform lexical and grammatical analyses on the data, but also to identify discourse functions served by the narrators' lexical and grammatical choices (for example, introducing versus maintaining reference to story characters, or foregrounding versus backgrounding story situations). While simple in some respects, the construction of a spoken, signed, or written narrative based on the picture story Frog, where are you? is a complex and demanding task which is sensitive to a host of cognitive and cultural factors. Unlike a personal narrative, where specific real-world experience is communicated, the frog story is fictional. Any act of communication concerning the story characters and their doings presupposes that relevant aspects of the A world in question are shared between sender and receiver. Further, the frog-story task is a retelling task. Mercer Mayer has already told the story by means of a wordless picture booklet and the subject's task is to retell it in words. Let us consider, in greater detail, some of the things involved in the construction of a frog-story narrative. /.

Perceive and interpret the individual pictures: a}

perceive and interpret lines, curves, shades as depicting objects (such as ajar, a floor, etc.), animate beings (e.g., a boy, a frog) and relations (e.g., jar on floor, frog in jar);


/'/.

//'/'.

iv.

b)

interpret the animate beings (from i-d) as having sensations, thoughts, intentions and attitudes, and other invisible properties such as, for example, proper names;

c)

interpret the objects and animate beings as involved in situations (such as seeing something, moving somewhere, scolding someone);

d)

impose a temporal profile on the situations depicted in the individual pictures (thereby interpreting, for example, one situation as an event occurring within the temporal extension of another situation, a process);

e)

foreground and background information in the individual pictures.

Interpret the individual pictures as inter-related: a)

interpret certain objects, animate beings, or places in the different pictures as identical;

b)

interpret the situations depicted in the individual pictures as temporally ordered or spatially inter-related;

c)

interpret the situations depicted in the individual pictures as logically inter-related (among many other things, organize situations in relation to the point of the story).

Take emotional and emotive stances on the situations constructed in / and //: a)

explain, justify, doubt, evaluate situations;

b)

become emotionally involved in situations (e.g., identify with or side with story characters).

Find the adequate expressive means to communicate the complex construct assembled through /-///.

In practice, the "functional" schema above is not followed, step by step, by the story-telling subject. Browsing through the pictures in advance of wording the


story, the subject gets a chance to construct a first overall idea about the story: its characters and plot. As the subject then revisits the pictures, while actually telling the story, he will probably reinterpret some of their details or elaborate certain aspects further. Also, the schema should not be taken to imply a strict bottom-up process in story interpretation or story construction. Already Step i-a is guided by top-down control processes based on knowledge about drawings and knowledge about story structure. For example, the degree of naturalistic representation in drawings varies with cultures and traditions, and the viewer is always left with a job of filling in gaps - often according to culturally determined principles (Gombrich 1960). And a story schema guides the subject's attention in that it makes him try to find, for example, a candidate main character, etc. Now assume that certain aspects of the construction of a frog story, according to the schema above, require a greater degree of make-believe than others. Thus, the objects and states constructed in i-a require relatively little make-believe - we tend to "see" them as a jar, a boy, etc. In contrast, the constructs in i-b, for example, the boy's thoughts or his name, cannot be "seen" in the pictures; they exist only within the game of make-believe. When the story is to be told, that is, communicated, the degree of make-believe has implications for the selection of expressive means. For example, the objects and states in i-a can be pointed at, but that is not true of the boy's thoughts or his name. His thoughts have to be described linguistically, and the name, itself a linguistic device, has to be predicated of the boy before it can be used to refer to him. Make-believe is, as it were, parasitic on the real world. For example, once you have established the drawing of the boy as a boy in the make-believe world (i-a in the schema above), everything you know about boys in the real world potentially applies to the boy in the make-believe world (Evans 1982). This renders the make-believe constructs sensitive to cultural experience as well. In this way, a potentially very rich information array is derived. The array represents a huge resource but also a problem. It needs structuring and control. The structuring key is the purpose of the activity - to tell a spoken, signed, or written story. Both the story schema and the particular language in which the story is going to be told impose selectional restrictions and a perspective on the information array. The story schema leads the subject, among other things, to select information which is relevant to the plot and to structure information from the perspective of the main character, the one who is doing things and to whom things happen. Taking the linguistic restrictions into account means selecting and structuring information in ways which allow the mapping of this information onto accessible linguistic means. This conceptual structuring process, driven by the purpose of linguistic information encoding, is what Slobin has proposed to call thinking-for-speaking (Slobin 1987, 1996b, 2002a, this volume), and it is the leitmotif of the present volume.


The frog story consists of several events and episodes of different types, presenting different cognitive and linguistic challenges to the narrator. For example, Pictures 11 and 12 (see Appendix I) show several temporally overlapping events, something which puts a special demand on the temporal organization of the story (see i-dand i-e in the functional schema above) and on the use of tense, aspect, and means for temporal clause linkage in the linguistic construction of the narrative. Pictures 13-15 show an episode where the main character discovers that what he originally thought was a pair of branches is in fact the antlers of a deer. This episode puts a special demand on epistemic operations and presupposes an advanced theory of mind in the narrator (see i-b and iii-a in the functional schema above). And Pictures 16-18 show an episode which is rich in motion events (i-c in the functional schema above). Crosslinguistic comparisons of this particular episode offers a fortuitous testing ground for hypotheses related to Talmy's distinction between verb-framed and satellite-framed languages and the extent to which languages detail path and manner of motion (Talmy 1985, 1991,2000a,2000b). These selected episodes and events provide, as it were, more controlled windows on thinking-for-speaking, in that they allow us to see in greater detail how different languages promote different perspectives on, for example, motion events. As Berman and Slobin put it in the 1994 volume: "In addition to the fact that narrators can take expressive options in the selection and arrangement of events, and the highlighting or downgrading of event components, each language requires or facilitates particular choices." (Berman & Slobin 1994:11).

3. RELATING FORMS AND FUNCTIONS IN NARRATIVES The 1994 volume ("Volume I"), Relating events in narrative: A cross-linguistic developmental study, brought together three important and productive trends in research on languages, cognition, and development: the functionalist approach to language acquisition and language development (see, e.g., Karmiloff-Smith 1979), the crosslinguistic study of language acquisition (see, e.g., Slobin 1985-1997), and the renewed interest in linguistic relativity (see, e.g., Lee 1996; Gumperz & Levinson 1996). For reviews of the 1994 volume, see Bamberg (1994b), Barbieri (1995), Chafe (1998), Gee (1994), Gerrig (1995), Hickmann (1998), Koopmans (1995), Lieven (1996), and Minami (1995). Volume I elaborated a "functionalist-conceptual" approach to language development, claiming, among other things, "that the development of grammar cannot be profitably considered without attention to the psycholinguistic and communicative demands of the production of connected discourse" (Berman & Slobin 1994:2). This approach was applied, among other things, to the use of


tense and aspect for the construction of an internal and external temporal profile of narrative events and for the foregrounding and backgrounding of events (see i-d, i-e, and ii-b in the functional schema above). Three complex tenets provided "guiding themes" for the volume: /.

Filtering Theme: The world does not present "events" to be encoded in language. Rather, experiences are filtered - through choice of perspective, and through the set of options provided by the particular language - into verbalized events.

//.

Packaging Theme: A skilful narrative does not simply consist of a linear chain of successive events located in time and space. Rather, events must be packaged into hierarchical constructions.

//'/'.

Development Theme: Younger children take fewer expressive options, because a) cognitively, they cannot conceive of the full range of encodable perspectives; b) communicatively, they cannot fully assess the listener's viewpoint; and c) linguistically, they do not command the full range of formal devices.

Within a typological crosslinguistic framework, the developmental relations between form and function were explored with respect to the following functional categories: temporality, event conflation, perspective, connectivity, and narrative style. Across the different functional categories, the grammatical forms that are accessible and obligatory within a language were primarily found to channel the attention of the learner towards particular functions. That is, frequently used forms direct learner attention to the functions of those forms. Finally, the acquisition of a form to express a particular function was also found to provide an opening wedge for the acquisition of the more advanced functions associated with that form. Berman and Slobin's volume also made clear how narrative competence develops. It was shown that children, as they grow older, learn to take more perspectives on events, to use a larger range of linguistic devices to express narrative functions, and to arrive at a better understanding of the listener's point of view. Given the crosslinguistic design of the study, an attempt was also made to show what is crosslinguistically valid and what is language-specific in narrative development. Crosslinguistically valid developmental characteristics concerned the understanding of the plot of the story and the mastery of linguistic cohesion and thematic coherence. The younger children tended to treat scenes as isolated

10


events with only minimal anchoring in time (that is, they did not go beyond / in the functional schema above). Older children, on the other hand, turned out to relate events more by means of temporal links, such as the use of tense to mark simultaneity or sequencing (// in the functional schema). By the age of nine, children are able to create a more or less explicit global narrative structure, to divide the narrative into foregrounded and backgrounded information, to express temporal and causal connections, and to make evaluative reference to states of mind of the characters in the narrative. In these respects, their narrative production approaches adult-like competence. However, language-specific characteristics in narrative development were also manifest. There was clear evidence that the formal and expressive options of the five languages under consideration affect the nature of narrative production. Four main typological dimensions emerged. First of all, it was found that the expression of tense and aspect as markers of the time slot of an event and its internal temporal constituency differed across languages. The grammatical marking of the progressive in English, Spanish, and Turkish, but not in German and Hebrew, forms a case in point. Second, the allocation of information between the main verb and its supporting satellites was found to be different: in satelliteframed languages, such as English and German, verbs tend to conflate motion and manner while additional elements encode direction, whereas in verb-framed languages, such as Hebrew, Spanish, and Turkish, verbs tend to conflate motion and direction, while manner is elaborated in separate elements. Third, languages differ in perspective, or in their means of expressing agent and patient, topic and focus, and foreground and background. Word order can be seen as one of the main devices to express perspective. In some languages, such as English, word order tends to be strict, whereas in a language like Turkish it is highly flexible. The fourth and final typological dimension concerns connectivity, as expressed by means of clause-linking devices such as syntactic conjunction, subordination, relative clauses, non-finite verb forms, and elliptical constructions. Different languages use different formal constructions to express connections in the text. From a comparative point of view, two types of complex sentence structure can be distinguished: co-ranking structures and chaining structures (Longacre 1985). Co-ranking structures, such as those found in most Indo-European languages, may consist of several verbs having the same rank. Traditionally, a distinction is made between co-ordinate and subordinate linkage relations. Co-ordinate clauses are autonomous, and are joined together with or without co-ordinating conjunctions. In subordinated constructions, one of the clauses (the main clause) is modified by one or more subordinate clauses. Subordinate clauses are grammatically dependent on the main clause and usually introduced by a subordinating conjunction. A chaining structure, in contrast, typically ends with a dominating verb of a fuller structure than any of the preceding verbs. For


instance, Turkish uses converbs in which the verb of the first clause receives a suffix indicating the relation to the verb of the second clause, which is marked for tense, aspect, person, and number. These four typological dimensions in the 1994 study show that the developmental path towards narrative competence is highly influenced by the subject's language. Different languages were found to be responsible for differences in thinking-for-speaking (Slobin 1987, 1991, 1996b). The claim made is that thoughts formulated while speaking in a language may be particular to the linguistic features of that language. Berman and Slobin brought forward a modified version of the Sapir-Whorf hypothesis, maintaining that one's native language directs one's attention to particular ways of filtering and packaging information. Berman and Slobin restrict their hypothesis to the mental processes that are accessed in the course of constructing linguistic messages, without any claims about possible effects of language on other mental processes. Berman and Slobin's work can be seen as a major attempt to arrive at a theoretical account of narrative development. The crosslinguistic differences show that cognitive activity is reflected in narrative, but at the same time shaped by the available linguistic forms. This evidence calls for further crosslinguistic studies with language varying across the typological scale.

4. THE PRESENT VOLUME This new book is organized in two main parts, and features a broad linguistic and cultural diversity. Contributions focusing on0crosslinguisticperspectives make up the first part. This part is concluded by Dan Slobin with an analysis and overview discussion of factors of linguistic typology in frog-story research, "The many ways to search for a frog". The second part offers a variety of theoretical and methodological perspectives, all dealing with contextual variation of narrative construction in a wide sense: variation across medium/modality (speech, writing, signing), genre variation (the specific frog-story narrative compared with other genres), frog-story narrations from the perspective of theory of mind, and from the perspective of bilingualism and second language acquisition. This second part is initiated by Ruth Berman with an analysis of the frog-story findings in the light of different genres, elicitation settings, and age groups, "The role of context in developing narrative abilities". The new book represents a rich overview and illustration of recent advances in theoretical and methodological approaches to the crosslinguistic study of narrative discourse. Several of the contributions also deal with developmental perspectives, but, unlike in the 1994 book, that is not the only focused issue. A recurrent perspective throughout the book is Dan Slobin's theory of thinking-

12


for-speaking (Slobin 1987, 1996b, 2002a, this volume), an approach to cognitive consequences of linguistic diversity. The book ends with an epilogue by Herbert Clark, "Variations on a ranarian theme". Unlike the 1994 book, the current volume is a collection of articles, rather than a set of logically interrelated chapters. One advantage of this arrangement is flexibility of order: the reader is free to follow his own reading preferences. A disadvantage is that there is an overlap of theoretical and methodological background information between the articles. All readers are advised to read carefully the three key contributions mentioned above (by Slobin, Berman, and Clark) and to consult the appendices (Appendix I: the picture series Frog, where are you?; Appendix II: an overview of research projects based on the frog story: and Appendix III: a list of abbreviations and grammatical codes recurring in the volume). The three themes focused on in Volume 1 provided guidelines for comparative analyses of frog-story data from different age groups (three to nine years and adults) and five languages (English, German, Hebrew, Spanish, and Turkish). The focused theme of Volume I was Development, but its discovery of filtering effects meant an important contribution to the exploration of linguistic relativity, a contribution which is extended and deepened in the present volume (Volume II). The possibility of detecting filtering effects was probably enhanced by the frog-story methodology. The older tradition of comparative philology as well as much of modern crosslinguistic research has relied heavily on translation exercises, and that creates a bias towards the information structure represented by the source-language data. By the same token, analyses in this tradition are often directed towards the detection of those lexical items, phrases, and constructions which present the closest equivalents, both grammatically and semantically, across two different languages. In contrast, the frog-story methodology is perfectly neutral with respect to source and target languages, and this paves the way for the detection of filtering effects. The subject is not biased by any linguistically encoded information structure, but can freely exercise the linguistic and rhetorical habits and preferences characteristic of his/her linguistic community. Under these circumstances, the semantic differences between the narratives produced can be expected to be greater. Importantly, this semantic variation has been found to be partly determined by typological factors. The ground-breaking work by Leonard Talmy on differences between verb-framed and satellite-framed languages (Talmy 1985, 1991, 2000a, 2000b) guided the exploration of crosslinguistic differences between frog-story data already in Volume I, and it is at the core of the typological analyses presented in Part I of the present volume. Volume II extends and deepens the comparative explorations of Volume I in several ways. First, the range of languages analysed is widened. The present volume deals with American Sign Language, Arrernte, Basque, English. Hebrew.


13

Icelandic, Japanese, Spanish, Swedish, Thai, Turkish, Tzeltal, Warlpiri, and WestGreenlandic. That makes fourteen languages and six phyla, as compared with five languages and three phyla in Volume I (see Appendix He for a typological overview). This rich and typologically varied set of data suggests, among other things, that Talmy's dichotomy between verb-framed and satellite-framed languages is over-simplified and should be replaced with a scale or cline, where languages can show a combination of properties of verb- and satellite-framed languages or be oriented towards either pole. Further, the data suggest that future typological comparisons can profit from taking several other dimensions into account, such as intonation, tone of voice, speech rhythm, and gestures (Slobin, this volume). Similarly, Clark (this volume) argues that the Platonic distinction between diegesis (describing) and mimesis (showing - e.g., through quotations, mimetics, and iconic gestures) is important to capture the subtleties involved in producing a good story. The frog-story data explored in Part I of the present volume were produced under different language conditions, with the discourse task kept constant. The frog story as a window on languages, thought, and development has proved successful in many ways, but it is a very narrow window. As Hickmann (1998) pointed out in her review article of Volume I: "the entire book is devoted to a particular discourse genre: the production of narratives based on a picture book and addressed to an adult." Narratives which, furthermore, are spoken and monolingual, we might add. Berman (this volume) puts the frog story into perspective, by comparing it with other genres and elicitation settings. Berman's discussion sets the scene for the second part of Volume II. In that part, a wider range of communicative conditions and processing constraints are explored as determinants of how the frog story is told: genre variation, speaking versus writing, monolingual versus bilingual frog stories. The analyses presented suggest that these determinants have semantic effects on, for example, the organization of the temporal and logical inter-relationships of story events, and the construction of rational and emotive stances (// and //'/, respectively, in the functional schema). Returning to our leitmotif, the analyses suggest that thinking-for-speaking differs from thinking-for-writing, and that thinking-for-speaking in a bilingual condition differs in certain ways from thinking-for-speaking in a monolingual condition (cf. Verhoeven & Stromqvist 2001). Similarly, thinking-for-narrating is partly different from thinking-forargumentation, and so on. In addition, Part II provides cognitive and linguistic perspectives on the construction of mental states in story characters. As already mentioned, a case in point is the episode shown in Pictures 13-15, where the main character discovers that what he originally thought was a pair of branches is in fact the antlers of a deer. Among other things, this situation requires the ability to contrast the narrator's omniscient perspective with the boy's temporary lack of knowledge;

14


that is, it demands a theory of mind that can separate the boy's state of knowledge from the narrator's. We end this introduction with a brief detour beyond the facts of the fiction. What would the frog-story characters - the boy, the dog, and the frog - say if they were aware of our frog-story quest, the observations and analyses made by the research community? Would they, with a gentle touch of irony, suggest to us that the branches of the tree of grammar we have held on to so steadfastly should rather be thought of as something similar to the antlers of a deer? We don't know. We are convinced, however, that the quest still has a host of discoveries in store, and we hope that this volume will provide a source of inspiration for its further pursuit.

NOTES 1.

Throughout this book all grammatical and morphological codings and glosses of foreign-language examples are made in accordance with the notational conventions used by Bernard Comrie (1981). Furthermore, some authors have chosen to represent semantic concepts by giving the initial letter in upper case (e.g.. Path or Manner).

Part I: Typological Perspectives


Focusing on 'where' An Analysis of Warlpiri Frog Stories

Edith L. Bavin

1. INTRODUCTION Australian Aboriginal people have a close affinity with their traditional lands, something that non-Aboriginal people find hard to understand. Warlpiri people typically talk about where people are and where they are going and where things happened. This chapter examines the use of locative markers in Warlpiri children's stories, and the age at which they sound like adults in terms of expressing locative concepts, such as goal, path, and source1. Based on the cultural values associated with ngurra 'country', it was predicted that locations and directions would be frequently used even in the stories from four- to fiveyear-olds. Reporting on crosslinguistic narrative data elicited with Mayer's (1969) picture book Frog, where are you?, Berman and Slobin (1994) illustrate differences across languages in the way people talk about events. Both crosslinguistic and developmental differences are discussed in relation to a number of issues, in particular temporality. When comparing narratives across cultural groups, crosscultural differences might also be considered. As well as differences in content, cultural differences may involve procedural issues. For example, what is regarded culturally as good performance (Bauman 1977) will be heavily influenced by the culture. According to Hymes (1975), a speaker presents a narrative in a way that will elicit evaluation by the listener, but not all cultures value the same type of behaviour. Cultural values will influence what a speaker determines as important when telling a story; for example, the deer in the frog story may be viewed as helping the boy search for his frog, and thus the fall from the cliff could be an accident rather than a deliberate throw by the deer. Nor do cultures agree on the "right" way to approach tasks. People may pay attention to features of a task that other cultures might view as irrelevant.

18

Edith L. Bavin

Brewer (1985) questions whether there is a notion of a universal story schema, in which an initiating event leads to a series of attempts (e.g., Labov & Waletzky 1967; Labov 1972). In addition, we might question whether speakers across cultural groups use a story for similar functions; this relates to who in a culture tells stories, and to which listeners. We might also question how important rhetorical style is in storytelling.2 For example, does the narrator hold the floor for the length of the narrative, or is the audience expected to join in, as reported by Tsitsipis (1983) for the Albanian spoken by the older people in a minority community in Greece. When people draw upon the resources available in their language to construct narratives, the particular resources selected must be influenced by such cultural factors. My main objective in this chapter is to present data showing developmental patterns in stories from the Warlpiri, a group of central Australia. The data are from the stories I collected using Frog, \vhere are you? (Mayer 1969; Appendix I). 1 focus on the use of locative/spatial markers. Other aspects of Warlpiri children's language are discussed in other papers (e.g., Bavin 1992, 1995, 1998. 2000). The analyses show the number of locative markers used at different age levels and the number of users, the types of locatives included in the stories at different age levels, and the verbs with which the locatives are typically associated. In the next section. I discuss the strong affiliation Warlpiri people have with their land (ngwra 'country") as well as other cultural factors associated with storytelling. The language-specific ways for encoding locations are then discussed, followed by a summary of the use of these markers at different age levels in the frog stories collected.

2. FEATURES OF WARLPIRI CULTURE In Aboriginal Australia, land is the basis of the social, spiritual, and economic organisation of a community (see, e.g., Nash, manuscript; Sutton et al. 1988). The Warlpiri in central Australia, like other Australian indigenous groups, have no desire to leave their traditional lands, which means leaving to go away for educational, vocational, or economic purposes is not valued. Nor do the people own land in the Western sense; after all, the land is there long after people die. The land is sacred and the people perceive of themselves as its custodians. In Aboriginal mythology, ancestral beings created the physical features of the land, and ancestral narratives are embodied in its topographical features. Many of the stories told are accounts of the travels of these ancestral beings from site to site (Sutton et al. 1988). Aborigines from desert areas have been reported to have almost total recall of topographical features of their country, and it has been suggested that they develop a topographical schema, a mental map (see, e.g., Klich 1988). A mental map indicates a detailed awareness of the important landmarks and ancestral sites

Focusing on'where'

19

of their country, as well as a familiarity with the narratives which interlink the sites. As suggested by Lewis (1976a), desert Aboriginal people have an ability to continually update their mental maps, in terms of time, distance, and orientation; realignments to any changes in direction are also made. Skills in route finding, geographic orientation, and a memory for locations are valued abilities for the Warlpiri as well as other Aboriginal groups. On the basis of experimental work, Kearins (1981) reports that Aboriginal people from the western desert have good spatial memory. She found that western-desert children, whether semi-traditional or more westernized, were far superior to non-Aboriginal children in recalling and replacing arrays of objects, arranged in sets of twelve or twenty, some of which were natural objects and some manufactured. Good orientation skills are not just associated with western-desert people. In discussing Aboriginal storytelling from the north-east of Australia, Haviland (1993) argues that the Guuguur Yimidhir people can take the perspective of the people in the stories and point out directions from the perspective of where those people would be located at any point in the story, rather than where the storyteller actually is. Traditionally, drawings in the sand were associated with Warlpiri storytelling (e.g., Munn 1973); the drawings used symbols for the people, where they travelled, and the tracks of animals. Some Warlpiri still use sand drawings when telling stories but paintings are also now put onto canvas and the painter can tell the story of his/her painting. Thus a visual representation of a narrative is culturally relevant, although the nature of the representation may be different from that of the Western-style representation of a story in a picture book. In a culture that has only recently introduced literacy skills, and where most of the people have no functional need of books and no books in their camps, the use of a picture book for eliciting stories could be criticized. However, the children have access to a limited number of illustrated books written in Warlpiri in the school, and as stated above, stories from mature, often non-literate, speakers are typically based on a painting or drawing. The paintings are from an areal perspective, as if the viewer is looking down on the land; dot patterns used in various colours on canvas represent important land areas, as well as the people and the paths along which they travelled. In telling the "story", it is of the greatest importance to locate the event or site. A number of different themes emerge in stories told by the Warlpiri. The people are traditionally hunter-gatherers, and one theme is the search for food. There are also stories about the mythical beings who were transformed into topographical features of the land, for example, into rock holes and hills. These stories provide children with knowledge of their cultural history. Other stories feature the cleverness or skill of someone in outwitting another. The best storytellers are mature, older speakers who have knowledge of cultural history to impart to the young. People tell stories as they sit around their camp fires.

20

Edith L. Bavin

The young mothers I have worked with remember the stories they heard around the camp fires as children. Today, in some families, storytelling is not as prolific as it once was. In the past fifteen or so years, satellites have allowed television reception into the community, and television programmes and videos now for some people occupy the hours that would have been spent listening to stories in former times. As children grow up and start to walk around the community with their peers, they gain information which they can contribute when talking with others. They construct personal narratives which are more typical from young children than are stories about cultural history. Storytelling for the Warlpiri is not a collaborative affair; the teller holds the floor. Other activities are collaborative. A rhetorical feature of the narrative style is repetition, which is reported to be a common feature of narratives in oral tradition (Brewer 1985). Information is repeated, maybe in a different form or word order. A "build-up" style is often used: information is repeated with some new added. There can be a gradual build-up of the information; for example, in telling the frog story a speaker might give the information that someone fell, then someone fell to the water, then someone fell down to the water, and then specify that it was the child and dog who fell. So not all is revealed at once, and perhaps this is a way of holding the attention of the listeners. Repetition is noted to some extent even in narratives from five-year-olds. There are several orientations to narrative analysis reported in the literature (Berman 1995). Nelson's (1986) event-representation approach and Schank and Abelson's (1977) script approach fall into the category of content-focus (Berman 1995). As Berman argues, these reflect two complementary points of departure, one focusing on children's emergent notion of scripts, as generic, prototypical sequences of events anchored in mental schemata (Berman 1995:5). Based on the Warlpiri people's ties to their country and their frequent journeys out of Yuendumu to their traditional country, particularly at weekends w hen the preferred activity is to collect favored bush food (e.g., \vardapi 'goanna' and other kuyu 'meat', as well as berries and sugarbag), we would predict that location and path would be essential features of the mental schemata associated with journeys. Thus, although the boy and his dog in the frog story are. presumably, not perceived to be travelling in Warlpiri land, we would predict that the frog stories told by Warlpiri people would include details about where the boy and his dog are and where they travel.

3. MARKING LOCATIONS IN WARLPIRI Languages differ in the way they lexicalize components of meaning. For example, as discussed extensively by Talmy (1983, 1985). Germanic languages typically

Focusing on 'where'

21

incorporate manner into verbs of motion, unlike Romance languages. These typically incorporate path with motion, while Germanic languages typically add a separate lexical item to mark path (for example, stroll along, climb up, continue on}. In Warlpiri, case markers, directional affixes, verb morphology, locative nouns, and preverbs plus combinations of these all provide information about location (see Hale 1982 for details of Warlpiri syntax). Path may be marked with a locative noun, a case marker, a verbal affix, or a preverb - a word which modifies the meaning and sometimes argument structure of the verb. A person going hunting will return to the starting point (ngurra 'camp'). For this reason, as Laughren (1978) suggests, the locative noun kamparru 'ahead' may also be interpreted as 'behind' since the ultimate goal of the journey, the starting point, is behind the traveller in some stages of the journey. However, there is also a locative \\o\mpirdangirli 'behind'. Thus in choosing which marker to use, a speaker takes different orientations and perspectives. Directions are not given in terms of streets and buildings, nor in relation to sides of the body (right, left). Locative nouns for the cardinal points (e.g., yatijarra 'north') are used to indicate locations and directions. Directional affixes can be used on verbs and locative nouns. The forms rra 'thither', rni 'hither', and mpa 'across' typically appear on verbs of motion. Note that the verb yani is both 'go' and 'come'; a directional affix may be used to distinguish 'toward' or 'away' from the speaker or some other point of reference, as in yanurra 'someone went away' versus yanirni 'someone came towards'. However, the speaker has the choice of adding the affixes or not. Case markers marking location include the general locative ngkalrla. This is not specific; its meaning includes 'in the area of and 'at'. The form ngka is used with two-syllable words and rla with longer words. Other case markers are hurra (allative) 'to/towards', ngarlu or janka (elative) 'from', and wana (perlative) 'along', which a speaker may choose to use if the reference object has a dimension of length, such as a creek, a line of hills, or the edge of a grassy area. A combination of case marker and locative noun gives a more specific location (as in [5]). The locative nouns include hulhurru 'between, middle', kankarla 'up, on', kanunju 'down, below', and wurnturu 'far'. To these, directional affixes may be added, as with kankarlarni. Other affixes include purda '-wards'. Note that the posture verbs ngunami 'lie', nyinami 'sit', and karrimi 'stand' are used to show where someone or something is, the particular verb selected depending on the shape of the object being located. For example karrimi 'stand' is used for a tree, ngunami for a branch on the ground, and nyinami for a small stone. Example [1] focuses on where the man is, rather than that he is in a sitting posture, although in his camp he probably is sitting. In [2] an added noun with a locative case would be needed to specify where the dog is entering.

22

[1]

Edith L. Bavin

Wati man

ka IPFV

nyina-mi sit-NONPAST

ngurra-ngka. camp-Loc

yuka-mi enter-NONPAsr

kaninja-rra. down-oiREC

'The man is in his camp.'

[2]

Maliki dog

ka IPFV

'The dog is entering down/going down into (something).'

[3]

Ya-nu-rra

watiya-ngurlu.

go-PAST-DIREC

tree-ELAT

'He/she/it went away from the tree.'

[4]

Ya-nu-rna go-pAST-lso:suB

pina back

ngurra-kurra. camp-ALLAT

'I came/went back home.'

[5]

Parnka-ja-lpa-rna run-PAST-iPFv-lsG:suB

pirli-ngka hill-ioc

kangkarla-rni. up-oiREC

'I was running on the hill this way.'

[6]

Marna grass

wana along

parnka-ja -rna. run-PAST-lsG:suB

'I ran by/along the grass.' Laughren (1978) illustrates that there can be a number of interpretations given to complex locatives in Warlpiri. For example, kankarla + rni 'up' + 'this way' used with the posture verb ngunami 'lie' could mean something was actually below the speaker but facing upwards and towards the speaker. So the interpretation of kankarla 'up' and kanunju 'down' depends on what the speaker uses as a reference point. A complex locative might include information about the orientation of the located object as well as source, path, direction, and location.

Focusing on 'where'

23

Preverbs also provide locative information. These words modify the meaning of a verb by adding different components of meaning, including path, manner, and quantity (see Nash 1982). Some preverbs are restricted in use. Others are more productive, combining with a number of lexical verbs; one example is warm 'around'. A preverb with a lexical verb may give the equivalent of a lexicalized verb in English, as with 'chase' from the preverb wajili 'running' used with the verb pinyi 'attack'; however, wajili can also be used with other lexical verbs. Preverbs do not always appear next to the verb. Some preverbs change a verb's argument structure. One example isjurnta 'away', which adds a dative argument, the person or thing from whom someone moves. Preverbs include: warm ' around \jaala 'back and forth', yaarl 'over top of, kulpari 'returning', pina 'in return, back\julyurl 'into water, fire', andjurnta 'away'. Nominals may also be added to a lexical verb to derive new meanings, as with the noun yakarra 'awake' with the verb pardija 'arise, start off, grow' which, combined, has the meaning 'wake up'. There are comparatively few lexical verbs in Warlpiri, probably less than 130 (see Nash 1986). The motion verbs include yam 'go\parnkami 'run', wapami 'move about', wantimi 'fall, drop', warrkarni 'climb, mount', yukami, 'enter, arrive', jitimi 'descend, dismount', purami 'follow'. Transitive verbs can be created by adding the causative mani to a nominal. The inchoative jarrimi can be added to a nominal to form an intransitive verb, as with wurdungu-jarrimi 'to become silent' from wurdungu 'silent, dumb'. A derivation with -njarni changes the meaning of the verb to 'go along x-ing', as with pankanjarni 'go along running' and kanjarni 'go along carrying'. While this combination indicates that an event was protracted, it also adds a path-like element, as is clear with the example wankanjarni 'go along talking'.

4. DATA With the exception of the stories told by ten teenagers, the stories were collected in Yuendumu in the centre of Australia. Stories from all groups were collected in culturally appropriate settings, that is, outside in view of others but not close enough for others to hear. The teenagers' stories were collected in a large park in Melbourne when they were on a visit away from the community. The speakers were allowed as much time as they needed, and they were only encouraged to tell the story if they showed initial willingness and interest. The children were first shown the book and the sequence of pictures; the researcher turned the pages for the younger children (four to five years) but the older children turned the pages themselves. To check that they knew the words for the objects/creatures in the story, as they were looking through the book the youngest children were asked

24

Edith L. Bavin

the words for 'jar', 'deer', and 'gopher'. An appropriate word was supplied for children who did not remember one. All testing was in Warlpiri. The task was to tell a good story (yimi ngurrju). All stories were audio-taped. Narratives from fifty Warlpiri people, aged 4;6-14, were transcribed and analysed. In order to investigate developmental patterns, the fifty stories were divided into five equal groups often, based on the age of the storyteller: four/fiveyear-olds (range 4:6-6;0); six/seven-year-olds (range 6; 1-7; 10); eight-year-olds (range 8:0-8;11); nine/ten-year-olds (range 9;9-10;7); teenagers (range 13;014; 11). In addition, stories from seven Warlpiri adults were taped and analysed in order to determine the extent to which the adults would impose a story schema onto the set of pictures. The total number of locative elements used in each group was calculated, as was the number of different locative types used in each group. The length of each story was determined on the basis of number of verbal sentences used. For each age group, the number of locative elements as a proportion of the story length was calculated, as was the number of different locative types as a proportion of the number of locative tokens. Tokens of each verb were counted. In addition, the types of locative elements used with each verb type were investigated.

5. RESULTS AND DISCUSSION 5.1 Analysis 1 The mean number of locative tokens, the mean number of locative types, and the mean length of stories by age are given in Table 1. As can be seen from these figures, although the differences between six/seven-year-olds and eight-year-olds are minimal, there is an increase overall in the number of locatives used by age as well as the number of locative types used. The mean story length was lower for the eight-year-olds than for the six/seven-year-olds and nine/ten-year-olds, but, overall, older speakers told longer stories. The lower rate for the eight-year-olds may be related to the fact that the group had the smallest age range; all were eight years. Variation was evident within all groups. Three separate one-way ANOVAs (see, for example, Iverson & Norpoth 1987) were used to determine if the differences across ages were significant for the number of locative elements, the number of different types, and the story length. The number of locative elements was found to be significant at the .001 level, F(4,45) = 5.61. The number of different locative types was also significant across groups at the .001 level, F(4,45) = 5.70. Length was not significant, F(4,45) = 2.16, p = .089. Post-hoc Scheffe tests revealed significant differences for number of locative tokens between the four/five-year-olds and the teenagers.

Focusing on 'where'

25

Table 1. Use of locative in children Age

4-5

6-7

8

9-10

13-14

Mean number of locative elements used

3.6

8.4

9.9

15.6

20.9

Mean number of locative types used

2.3

4.3

4.6

6.4

7.1

21.9

31.2

27.0

33.4

36.3

Number of locatives used in proportion to story length (as a percentage)

15.6%

26.4%

34.4%

39.8%

57.9%

Number of types in proportion to number of locative elements used (as a percentage)

73.9%

57.3%

52.0%

55.0%

35.7%

Mean number of verbal sentences used

Post-hoc Scheffe tests also revealed significant differences for the number of locative types between the four/five-year-olds and the nine/ten-year-olds, and between the four/five-year-olds and the teenagers. Speakers older than eight used significantly more types than the four/five-year-olds. 5.2 Analysis 2 Because longer stories allow more opportunity for locatives to be used, the number of tokens in relation to story length is a more revealing measure of developmental trends. Table 1 also shows the number of locative tokens in relation to story length. Note that the largest increases are between the four/five-year-olds and the six/seven-year-olds, and between the nine/ten-year-olds and the teenagers. The last row of Table 1 shows the mean number of types as a percentage of the number of tokens. The higher the figure, the fewer types were used in relation to the number of tokens. Thus the four/five-year-olds, the youngest group, used about as many types as tokens, while for the older groups more types were used, and also more instances of each type. Two separate one-way ANOVAs were used to determine any significant differences across age groups, based on the means in the last two rows of Table 1. The number of locatives used in relation to story length was significant at the .000 level, F(4,45)= 11.08. Post-hoc Scheffe tests revealed that the teenagers used significantly more locatives than the three youngest groups, but not more than the nine/ten-year-olds. In addition, the nine/ten-year-olds used significantly more than the four/five-year-olds. The number of different types in relation to the tokens was also significant, at the .001 level, F(4,45) = 3.67. Post-hoc Scheffe tests revealed a significant difference only between the four/five-year-olds and the teenagers.

Edith L. Bavin

26

Table 2. Use of locatives in teenagers and adults Age

13-14

Adults

20.9

47.4

7.1

11.0

36.3

74.3

Number of locatives used in proportion to story length (as a percentage)

57.9%

63.8%

Number of types in proportion to number of locative elements used (as a percentage)

35.7%

25.9%

Mean number of locative elements used Mean number of locative types used Mean number of verbal sentences used

In summary, the oldest group (the teenagers) used significantly more locatives in proportion to story length than the children under nine, and the nine/ ten-year-olds used significantly more than the four/five-year-olds. The teenagers used significantly more types as a proportion of the number they used than the youngest children, the four/five-year-olds. It is clear from these findings that the youngest children focus less on locative information than might be predicted given the cultural focus on land and location and the emphasis on where people are and where they are going. However, the four/five-year-olds included a few locatives. The stories from the teenagers were compared with those from the adults, in order to determine how similar they were in terms of locative use. If differences between teenagers and adults are not significant, it would indicate that the teenagers had developed a mental schema of journeys and searches and an associated script upon which they and the adults draw. Comparisons of locative use by the adults and the teenagers are given in Table 2. These figures illustrate that a greater number of tokens and types were used by the adults than by the teenagers, and that the adults' stories were longer. A series of five independent t-tests for the five variables showed four significant differences: number of locatives used, number of different types, length of stories, and percentage of type to total. The number of locatives used as a percentage of story length was not significantly different. 3 5.3 Analysis 3 A further question is which locative types the youngest children chose to include, and which locative types were more likely to be included by the older children

Focusing on 'where'

27

and the adults. The data show that the general locative-case marker was the one most likely to be used by children in the four- to five-year age range. For the children aged from six up, a goal (with allative case) and directions were more likely to be included in the narratives. A source (elative case) as well as locative nouns to give more specific information about path and direction were more likely to be selected by the older participants. Preverbs were not frequent except from the older teenagers and adults; those used by the children were typically with one verb only, to form verbal concepts that are not lexicalized in Warlpiri, e.g.,yakarra withpardimi 'wake up\julyurl with wantimi 'swim' and wajili with pinyi 'chase'. The preverb pina was more flexible, used with yani or kanyi 'go back' or 'carry back'. In addition, warru 'around' was more flexible and used by a few six/seven-year-olds but more of the older speakers. Details of how many children in each group used particular locative elements is revealing. These are discussed below. The figures represent the number of children in each group who used a particular locative element at least once. This analysis is not based on frequency of tokens but number of users. Altogether six different types of locatives were used by the four/five-yearolds. These were: a) locative case (used most frequently with the verbs wantimi 'fall' and nyinami 'sit'); b) directionals, rni 'hither', rra 'thither', wana 'along'; c) njarni 'go along doing'; djjunga 'straight'; e) hurra (allative); and/) kankarla 'up'. However, other than the locative case, these locative elements were infrequent, usually one token from one speaker. The locative case was used by seven of the group. Recall that the mean number of locatives used by this group was only 3.6, whereas the mean number of verbal sentences was 21.9. Thus, on average, only one in six sentences included a locative element. Of thirteen types used by the six/seven-year-olds, only eight were used by more than one child in the group. The others were used idiosyncratically. The locative case was used by nine children, and the allative (kurrd) by eight. So, by age six to seven years, the children are more likely to specify a goal than are the younger children. Five people used the directional affix, and kanunju 'down' was used by four, as was warru 'around'. The forms njarni 'go along doing', wana 'along', and ngurlu 'from' were each used by two people. Sentences with the verb for 'fall', 'sit', and 'climb' were the ones in which most speakers added a general locative-case marker, the locative form used by most speakers. Of the eleven types of locative elements used by the eight-year-olds, only eight types were used by more than one speaker. The locative case was used by eight children and the allative (goal) by five, while the elative case (source) had six users, as did the directional affixes. From these data it seems that children aged from eight years up typically specify the starting point, goal, and a direction. The sentences in which most children used a locative-case marker contained the verbs for 'climb' and 'sit'. More speakers used the allative in sentences with the verbs for 'go' and 'throw', which is not surprising given the inherent goal

28

Edith L. Bavin

orientation for these verbs. More speakers specified a source with the verbs for 'jump' and 'arise'. For the nine/ten-year-olds, thirteen different locative types were used; of these, three types were used by more than 50% of the group: locative case (nine children), allative (nine children), and directionals (seven children). Six other types were used by three or more speakers: kanunju 'down' and kankarla 'up' were used by four speakers, while ngurlu 'from', wana 'along', warm 'around'. pina 'back' were used by three speakers. Sentences with the verbs for 'fall', 'climb', 'go', 'arise', and 'see' were the most likely to include a variety of locative types. More speakers used the general locative case in sentences with verbs for 'climb' and 'sit'. In sentences with yani 'go', more directionals were used than in other sentences. Overall, the average use of locatives was one for about every two verbs, which contrasts with the youngest group with one for about six verbs. Fifteen locative types were used by the teenagers, with eleven types used by more than one person. All ten teenagers used the locative case, the allative case, and wana 'along', while seven used the elative case (source), six used a directional affix, and six used kanunju 'down'. Five used the preverb warru 'around' and four speakers used pina 'back'. Based on the stories collected, by age thirteen, Warlpiri speakers are likely to include a locative element of some type for every 1.75 verbs, and they draw upon a variety of locative elements. All ten speakers added a locative case in sentences with the verbs for 'see' and nine speakers did so for sentences with 'climb'. The locative case was also used in sentences with the verbs for 'throw' (seven people) and 'sit' (six people). Speakers were more likely to choose to use an allative in a sentence with the verbs for 'fall' and 'carry' than in other sentences, while more speakers specified a source with sentences containing the verb for 'arise' than for sentences with other verbs. Sentences with the verbs for 'go', 'see', 'call', and 'fall' were more likely to include a range of locative types from more speakers in the group than were sentences with other verbs. In summary, while a few of the youngest children who told stories used a locative marker other than the locative case, the data show that the locative case was the one most likely to be used. This finding can be compared with earlier research findings. In a study testing comprehension and production of locative terms, Warlpiri children were asked to place objects in specific locations in relation to a set of reference objects and, in a follow-up experiment, to identify locations where a number of objects were placed (Bavin 1990, Study 3). It was found that while some of the four/five-year-olds produced kankarla 'up' and kanunju 'down', some relied on the general locative-case marker on the reference object, although they understood the locative nouns. Children do not necessarily draw upon familiar linguistic forms when representing events in narratives. As they become more familiar with the schema for events and the scripts used

Focusing on 'where'

29

by older speakers to report them, they adopt these as the scripts for their own stories. For the seven adults, the number of different locative types totalled seventeen, with nine used by 50% or more of the group, that is, by four or more speakers. Locative-case and directional affixes were used by all seven adults, as was the 'going along doing' construction, and the allative case and wana 'along' were used by six adults. The nominals kanunju 'down' and kankarla 'up', and elative-case ngurlu orjangka (or both) and the preverbpma 'back' were each used by five of the adults. Four used nyampu 'here' and three included wurnturu 'far'. More adults were likely to add a locative case to sentences with the verbs for 'see' (6 adults), 'sit' (6), and 'climb' (5) as well as 'enter' (4). These speakers were likely to choose an allative case for sentences with the verbs for 'go' (5) and 'fall' (4). In addition, more adults chose to include directionals for the verb for 'go' (5) and the verb for 'climb' (5) than for other verbs, and more speakers chose to add a source with the verb for 'arise' (4). The form warm 'around' was used with the verb for 'see' ('look around') by three adults. Speakers chose kankarla 'up' (typically with a directional affix) for the verb for 'climb', and two speakers also used the locative noun for 'up' with the verb wantimi 'fall'. In such uses, the noun does not represent a path but an orientation, with the boy facing up. Locations and goals as well as orientations were prolific in the adult stories. The speakers' perceptions of scenes were implicit in the selection of wana over a general locative case, and mpa 'across', as used with kanyi 'carry', yani 'go', and pardami 'arise' . Adults were likely to include a source with pardami 'arise'. But some included a source with wantimi 'fall' and nyanyi 'see'. With verbs of motion, path components were also likely to be included. Overall, the frog stories as told by the older participants were longer. However, while more locative tokens were used as well as more types, these were not in direct proportion to length of story. Developments that change a child speaker to an adult speaker in terms of the locative elements used are complex: the children need to have acquired the concepts which underlie the use of specific forms, such as for ngurlu 'source'. They also need to have acquired the meanings of the lexical verbs, and know which locative meaning components can be assumed in the meaning of the verb so they are not added as separate elements. As concepts such as causality and causal agents are included in the stories, more opportunities may arise for elaborating with associated spatial concepts. For example, with the older groups there was more elaboration in the scene of the boy falling from the deer (although not for the dog's fall from the window). As shown in Tables 3a and 3b, many of the speakers from the older groups used two or more sentences to capture the event. While the younger speakers focused on

30

Edith L. Bavin

Table 3a. Verb and locative use in the "Dog falling from window" scene for each age group N Age Verb Number of V + Locative constructions 6 4-5 fall 2 jump 5 6-7 fall 1 + allative 2 jump 1 + source 5 8 fall 1 + down & allative 4 1 + source jump 4 1 + down 9-10 fall 1 + source 1 + down & allative 1 jump 9 fall 13-14 1 + allative 1 + source 1 + source & allative 1 + down & allative 1 1 + wana jump 2 1 + here & down, and 2nd sentence + 'fall* + allative fall Adults 1 + down & rra 1 1 + source & allative jump The 'adults' group comprised 7 subjects, all others 10; N in the third column refers to the number of subjects who used the verb in question.

falling, a typical pattern for most of the speakers from the older groups was to use kijirni 'throw' and to follow this with a sentence with wantimi 'fall'. Sometimes a third sentence or more followed, often with 'throw' or 'fall'. Typically, one or all had accompanying locative elements. In addition, not shown in the table, a follow-on sentence often included ngapa 'water' and a locative case. So in the elaboration of this event, the speakers focused on the cause of the change of location, the path and goal of the journey, where the participants ended up, and who they were. The notion of causality was more likely to be encoded by the older speakers, who used kijirni more frequently than the younger speakers, who focused more on the fall. Speakers sometimes chose to specify kanunju 'down' with the verb wantija 'fell'. Some used combinations of kanunju with a goal, as in kanunju ngappakurra 'down to the water'. Other combinations were direction, path, and goal, as in wantijarra kanunju ngapa-kurra 'fell away down to the water'. Repetition of information, using the build-up style discussed above, was notable in this scene. Speakers built up information as they elaborated on the path, direction, and goal. In contrast, the "dog falling from the window" scene elicited few directions or path elements.

31

Focusing on 'where'

Table 3b. Verb and locative use in the "Boy falling to water from deer" scene for each age group

Age 4-5

Verb

fall

N 6

throw 4

6-7

fall

7

throw 3

8

fall

5

throw 5

9-10

fall

2

throw 8

13-14

fall 2 throw 8

Adults fall

2

throw 5

Number of V + Locative constructions 1 + directional 2 + allative in 2nd sentence 2 + allative 1 + locative in 2nd sentence 2 + allative 1 + locative 1 + down & intensifier in 2nd sentence 2 + allative 2 + down & allative 2 + allative + rra & locative + allative in 2nd sentence + 2 more sentences with 'throw' and 'fall' + allative + allative + 2nd sentence with 'fall' + allative + allative in 2nd sentence + 2nd sentence with 'fall' + PV + allative + 'throw' in 2nd sentence + 'fall' + rra + locative in 3rd sentence 3 + allative + locative + allative + 'throw' + allative in 2nd sentence + allative + 2nd sentence with 'fall' + allative + 2nd sentence with 'fall' + allative 2 + allative 1 + allative 1 + allative +2"d sentence with 'fall' + allative + 3rd sentence with 'fall' + down + rra 1 + allative + 2nd sentence with 'fall' + locative 1 + allative + 2nd sentence + 'fall' 2 + allative + 2nd sentence with 'fall' + allative 1 + down + rra + 2nd sentence with 'fall'+ allative 1 + down + rra + allative + 2nd sentence with 'fall' + allative 1 + source, locative, up, allative & ahead; l-2nd sentence + 'fall'; 3rd sentence with 'fall' + preverb & allative, 4th sentence with 'throw' + down, rra; 5lh sentence with 'fell' + allative 1 + down & allative; 2nd sentence with 'fall' + locative 1 + down & rra; 2nd sentence with 'throw' + allative; 3rd sentence with 'fall' 1 + allative; 2nd sentence with 'fall' + allative, 3rd sentence with 'throw' 1 + there, 2nd sentence with 'fall' + rra, source & up

The 'adults' group comprised 7 subjects, all others 10; N in the third column refers to the number of subjects who used the verb in question.

32

Edith L. Bavin

6. VERBS MOST FREQUENTLY USED: ANALYSIS OF TOKENS For the total group of participants, the nine verbs most frequently used were (most frequent first): nyanyi 'see', \vantimi 'fall',yani 'go\purlami 'call', nyinami 'sit'. warrkarni 'climb', pardimi 'arise', kijirni 'throw', and kanyi 'carry'. For each verb there were at least 70 tokens (range 71-380 tokens), and no less than 50% of the speakers in each age group used the verb at least once (except for nyinami and kijirni, which were used by four users in the youngest group). Because the frog story involved a search, it is appropriate that the verb nyanyi would be frequent. The verb typically appears in an ergative-absolutive frame (ergative subject and absolutive object). However, with a change of case frame to ergative-dative. the meaning of the verb changes to attempted action ('look for'). Most of the children used this derived frame to encode the boy's search, rather than the specific verb warrirni 'search for'. Table 4 shows the number of tokens by group for each of these verbs and the percentage of tokens which appeared with at least one locative element. It also shows the number of people who used the particular verb at least once in each group. Note that, overall, while the verb nyanyi 'see' had more tokens, sentences containing this verb did not have the highest percentage of locatives. Sentences with warkarni 'climb', yani 'go', and \vantimi 'fall' had more. Sentences with purlami 'call' did not. The use of nyanyi 'see' and purlami 'call' show that the children were aware of the search for the frog, but the perception verb was more typically accompanied by locative elements. That is, stating where the boy and dog were looking was more likely than stating where the boy was calling for the frog. For nyinami 'sit', about half the verbs appeared with locative elements for all groups. Recall that this verb has a locative function in Warlpiri. Adult usage can be considered to be representative for the language, and thus it is valuable to consider the adult data separately. About half the number of locative tokens from the adults were for the four verbs nyanyi 'see', \\-antimi 'fall', \varrkarni 'climb', and yani 'go'. Three of the verbs are motion verbs. A further 20% were for the verbs purlami 'call', pardimi 'arise', parnkami 'run". ngunami 'lie', and kanyi 'carry', with about an equal number for each. Two of these are motion verbs. The third group of verbs, responsible for a further 15% of the locative tokens from the adults, were kijirni 'throw', karrimi 'stand', \vajilipinyi 'chase', yukami 'enter', and mani 'take'. Two of these are non-motion verbs. The remaining locative markers used were distributed across other verbs, some used by only one adult. Elaboration with locative elements involves more verbs as the children get older. For the younger speakers, a lower percentage of sentences contained locative elements. Considering the most frequently used verbs, half or more of the sentences for the four/five-year-olds with the verbs for 'go', 'see', and 'sit' contained a locative element. For the six/seven-year-olds, half or more of the

Focusing on 'where'

33

Table 4. Verbs with most tokens: number with locative elements over total; percentages; and number of users for each age group Verb 1. see

4-5

39/50

10

3/43

9

9/25

78%

2. fall

7/25

3. go

4. call

5/9

4/9

4

0/3

5

3/10

2

1/6

4

2/6 33%

2/9

12/20

14/24

0/8

6/11

6

2/11

3/8 38%

6/53

9

10/24

6

12/18

2/22

6/12

14/15

7

5/10

7

6/11

7

4/10

6

3/10

4/17

4/7

10/14

4/11

6

7/18

5/14

7

30%

1/5 20%

10

21/40

9

19/31

6/14

5

13/24

16/21

9/19

8

10/10

8

8/12

9

5/12 42%

18/38

7

18/24

7

75% 7

15/32

6

47% 8

5/14

7

36% 7

8/18

7

44%

67% 4

7

47%

100% 7

11/38 38%

47% 10

7

62%

76% 6

7

53%

54% 8

7

47%

43%

36% 4

33/72

70%

39% 7

10

67%

36%

40% 3

16/23

42%

55% 5

7

Adults

31%

57%

50% 8

22/33

24%

93% 4

8

67%

50% 8

31/100

42%

9% 10

10

11%

79% 5

18% 5

9

32%

55%

17%

10. stand

11/14

-

30%

9. carry

6

58%

-

8. throw

3/16

13-14

9-10

19%

60%

44%

7. arise

10/31

22%

56%

6. climb

7

19% 7

5. sit

12/62

36% 6

64%

0/12

8

7%

28%

9/14

8

6-7

6

8/11

2

73%

The 'adults' group comprised 7 subjects, all others 10.

sentences with verbs for 'climb', 'lie', 'sit', and 'throw' contained locatives. For the eight-year-olds, half or more of the sentences with verbs for 'arise', 'climb', 'enter', 'go', 'sit', and 'throw' contained locatives. For the nine/ten-year-olds, half or more of the sentences with verbs for 'enter', 'go', and 'sit' contained locatives. For the teenagers, half or more of the sentences with verbs for 'carry', 'climb', 'enter', 'fall', 'go', 'run', 'sit', or 'throw' contained locatives. One similarity across the groups is the use of locatives with the verb nyinami 'sit'. Even some of the youngest storytellers included a locative element in sentences with nyinami. With the exception of the six/seven-year-olds, sentences with the motion verb yani 'go' were also likely to be produced with a locative element, across all age groups.

34

Edith L. Bavin

7. SUMMARY AND CONCLUSION In conclusion, the main differences found across the age groups were the frequency with which locative elements were included, and the types used. More specific locations were indicated in the stories from the older groups. Complex locatives specified where people were and where they were moving; these were combinations of elements: location, orientation, source, goal, path, and direction, or a combination of these. In comparison with other languages, reported in Berman and Slobin (1994), the use of locatives in Warlpiri stories patterns more like that in stories from the Germanic languages than from Spanish, Turkish, or Hebrew. Slobin (1990:245) reports that by age nine, the speakers of English and German "tend to put together directionality, source, and/or goal in one complex clause". In comparison, speakers of Hebrew, Turkish, and Spanish tend to use a strategy in which the scene is set first and then a change of location is shown with a motion verb. For example, the translation of one Spanish example cited is The deer took him to a place where below there was a river. Then the deer threw the dog and the boy to the river (Slobin 1990:245). The lack of elaboration of locative elements in the three non-Germanic languages is attributed in part to the availability in these languages of verbs with inherent directionality. A question raised in the introduction was to what the extent do cultural factors influence the stories produced by children and adults. In the frog stories told by the Warlpiri, locatives appeared with posture verbs and nyanyi, indicating that the speakers assumed it was important to say where the boy and his dog were looking and where they were located, not just where they were going. Overall, the adults included more episodes in their stories than did the children, and with these, details about location were included. Knowledge is gained through experience in Warlpiri culture; older people are respected for the knowledge they have and for the details they can remember. Clearly, it takes a number of years for Warlpiri children to gain knowledge of an appropriate framework for a story about a journey and a search. While even the smallest children have many experiences of journeys and food gathering, the script which the adults follow in their stories can only be gained by repeated exposure to journeys and to the stories which draw upon the script, and by becoming familiar with the language used. People have options in how they construct dialogue and narratives, but they are heavily influenced by the lexicalization patterns of their language (Berman & Slobin 1994). The stories from the older Warlpiri children about the boy. his dog, and the frog show that they are well on their way to becoming adult storytellers.

Focusing on 'where'

35

NOTES 1.

My sincere thanks to the late Napaljarri who gave such devoted assistance over the years and who made sure my transcriptions did justice to the language and the storytellers. Nancy Napanangka, Cecily Napanangka, and Barbara Napaljarri have also assisted greatly. Mary Laughren's materials on directions in Warlpiri have been of enormous help.

2.

Not all cultures follow the same rhetorical style. For example, the frog stories 1 collected from speakers of Western Nilotic languages revealed that the use of proverbs marks a good story in western Kenya and northern Uganda. Such an indirect means of conveying a moral is more highly valued than the direct style typically adopted in our culture.

3.

The differences between the two groups for number of locatives used, types, and length were significant at the .05 level. Equal variance was not assumed; for total, t(7.5) = 2.7; for types, t(13.2) = 2.9; for length, t(7) - 2.9. The differences between the type/total ratios were also significant at the .05 level, t(14.99) — 2.26. The differences for total as a proportion of length was not significant, t(13.55) =.98, p = .34.


Position and Motion in Tzeltal Frog Stories' The Acquisition of Narrative Style

Penelope Brown

1. INTRODUCTION In Mercer Mayer's wordless picture-book Frog, where are you? (see Appendix I), the following story is portrayed: a boy has a dog and a pet frog; the frog escapes from its bottle and runs away; the boy and dog look for it across hill and dale, through woods, and over a cliff, till they find it, and return home with a baby frog child of the original pet frog. As a visual prompt for eliciting what is essentially a 'journey' story, the picture-book provides a rich resource for examining the temporal and spatial organization of events in narratives from different languages and cultures. It has therefore been the focus of an extensive cross-linguistic investigation into narrative style and its development in children (Berman & Slobin 1994; see also Slobin 1991, 1994; Ozsaliskan & Slobin 1998; Ozyurek & Ozcahskan 1998; Wilkins 1997). Major points of the Berman and Slobin research include the following: /.

There are clear differences between languages with respect to the event components that speakers encode, due to the 'channeling of attention' involved in thinking-for-speaking. In particular, the semantic typology of a language as 'verb-framed' or 'satellite-framed' (Talmy 1985) is reflected in the linguistic expression of locative trajectories. In verbframed languages such as Spanish, the preferred locus for Path is in the verb; furthermore, these tend to be concise, bare verbs of motion, often without further locative specification (e.g., 'he ascended'). Satelliteframed languages such as English prefer to encode Path in a satellite (particles such as 'up', 'down', 'in', etc.), and complex trajectories are often encoded in one clause (e.g., 'he climbed up the rock to the tree branch'). Slobin (1996b) further argues that what is asserted versus

38

Penelope Brown

what is implied varies in accordance with this typology; for example, characteristically, English asserts trajectory and implies endstate, while Spanish asserts endstate and implies trajectory. This language-specific propensity for channeling attention ensures that the foreground and background are constructed by the narrator, not given by the pictures. Frog stories have a language-specific flavour or style. /;'.

Another important discovery in the Berman and Slobin research is that, in children's narrative development, there are language-specific interactions between narrative organization and linguistic expression. Narrative functions that are expressed by obligatory devices all appear early (by about age three), but even by age nine children have not fully acquired the narrative style of their language.

;'/'/.

A final point, one particularly relevant to this chapter, is that linguistic devices do not have to be obligatory to influence narrative development: children can learn early linguistic forms that are late-learned by children learning other languages (e.g., relative clauses, passives) if they are either a) grammaticized, or b) 'accessible' (frequent, in contrast sets with other frequent items, prosodically emphasized), i.e.. if they are easily expressed in the child's own language.

The present study of Tzeltal frog stories contributes to this large comparative project additional data with two new features. The first is that the data come from a predominantely non-literate, monolingual peasant community. The language of this study is Tzeltal (Mayan), as spoken in the remote rural community of Tenejapa, Chiapas, Mexico; the cultural differences from Western societies, and especially the absence of books, makes for a quite different interpretation of the events portrayed in the story, although it is still treated as a journey story. Secondly, this study focuses on the integration of position (stasis) and motion in event descriptions. It is argued that specification of position as well as motion in journey descriptions is just such an 'accessible' characteristic of the language and influences Tzeltal children's narrative style from age three. In frog-story research to date, a great deal of attention has been devoted to linguistic and cultural differences in the packaging of event descriptions and the deployment of spatial language (locative trajectories, the lexicalization of Manner and Path, etc.) (Berman & Slobin 1994; Wilkins 1997). No attention has been paid to Position (as location, as the end result of motion, or positionwhile-moving), which from an Indo-European perspective seems irrelevant to locative description.2 But Tzeltal elaborates position as an important element in descriptions of location (Brown 1994). Here I'll argue that it can also be important in descriptions of motion, and, more specifically, that these positionals

Position and Motion in Tzeltal Frog Stories

39

can bring Manner (e.g., how the Figure is positioned while moving) into the clause.

2. TZELTAL LINGUISTIC RESOURCES There are four relevant Tzeltal resources used in journey descriptions; none of these is grammatically obligatory.3 The first three encode Path; the fourth, Position. /.

A core set of motion verbs encoding Motion + Path (e.g., 'go'/'come', 'enter'/'exit', 'ascend'/'descend', 'pass by', 'return'. ('Fall' also encodes Motion + Path, though it is not a member of this core Tzeltal set.) In this respect, Tzeltal is a verb-framed language.

//.

A set of directionals, adverbials that directly follow the verb and are grammaticized from the core motion verbs. In this respect, Tzeltal is satellite-framed (or outside of Talmy's typology). Directionals convey Path (Direction of motion or Orientation of a path or of a static array for example, 'he exited coming' or 'he is falling coming downwards').

/'//.

A set of auxiliaries, grammaticized from the same core set of Pathencoding motion verbs, which precede the verb and convey a 'motioncum-purpose' meaning (for example, 'he ascends in order to view the countryside').

iv.

Several hundred positional, a class of verbal roots which convey Position of animate or inanimate things (in stasis, or concurrent-with, or as-a-result-of motion). In a clause, these roots can take any one of three predicative forms. With a -V,l suffix4 they become a stative predicate ('be-in-position'), which is by far the predominant form they take in the frog stories. But they may also be derivationally causativized (to indicate transitive 'put-into-position'), as in 'he [the dog] stood himself up on his hind legs'. And they may be intransitivized (conveying 'come-tobe-in-position' or 'be-normally-in-position')5. The 'positional' I count in these analyses are mostly in the -V,l predicate form (including bodyposition words such as 'sitting', 'standing', 'kneeling', 'lying face-up', 'perched', but also words for positions or configurations of inanimate objects, e.g., 'tightly-inserted', 'bowl-shaped inverted', 'hanging').

Tzeltal is in part a verb-framed language with a core set of Path-encoding motion verbs, so that the bare bones of the frog story can consist of verbs translating as

40

Penelope Brown

'go', 'come', 'pass by', 'ascend', 'descend', 'enter', 'exit', 'arrive', 'return', etc. But Tzeltal narrators also make heavy use of the Path-encoding adverbials which provide the direction of motion or the orientation of static arrays. Furthermore, motion is not only encoded barebones, but vivid detail may be provided by positional verbs which can describe the position of the Figure as an outcome of a motion event; motion and stasis are thereby combined in a single event description, packed into one clause.6 For example (throughout the chapter, positionals and their glosses are given in italics, while motion/path expressions and their glosses are underlined): [1]

xojol moel sjol ta ala plastiko. 'His head [the dog's] is inserted-tightly upwards into the little plastic thing.'

[2]

chawal ya x'anix te kereme. 'Lyingface-up the boy has fled [from the bird].'

[3] jipotjawal ta lum. 'He [the boy] has been thrown lying-face-upwards-spread-eagled'to

the ground.'

[4] pakal\\\e\ tz'i' ta yanil ton. 'The dog is lying-belly-down remaining-behind at the bottom of the stone.'

Why should Tzeltal-speakers be inclined to combine Position and Motion/Path in the same clause? One reason is that positionals can bring manner-like information into the verbal clause, although this is not manner-of-motion but 'manner' of the state of affairs described by the whole clause, picking out the Figure. As we shall see, Tzeltal resources for expressing manner of motion are much less elaborated: there are a few Motion + Manner verbs (e.g., 'swim', 'flee', 'chase', 'walk', 'run', 'jump/fly'), and some specialized devices (e.g., reduplication, affect verb affixes). Positionals fill the gap.

3. COLLECTING TZELTAL FROG STORIES 3.1 Hypotheses Given the importance of positionals as an integral (though non-obligatory) part of locative and motion descriptions, Tzeltal-speakers might be more prone to attend to details of positions when describing a motion journey than speakers of other languages. Two possibilities are compatible with what is already known about Tzeltal: Positionals might be used for backgrounded events and picturedescription (e.g., 'shoes standing at the foot of the bed'), with motion verbs used


41

for foregrounded events that carry the journey story forward (e.g., 'he fled from the bees'). Alternatively, positionals might be integrated with motion in the same clause; in this case both would be foregrounded (e.g., 'he fell down lying-faceupwards'). 1 hypothesized that: /.

A relatively large proportion of Tzeltal frog-story narrative clauses will specify the position of the Figure.

/'/.

Tzeltal children will attend to position (their attention will be channeled to it) by age three.

//'/'.

Rather than simply alternating background/scene-setting positiondescriptions with descriptions of pure motion that move the story along through time, speakers will express position in the same clause as motion, to convey not just pure motion but a kind of'manner' associated with the scene. Position and motion will both be foregrounded in such clauses.

iv.

Although cross-linguistically it appears that children even by age nine have not fully mastered adult narrative style (Berman and Slobin 1994), relatively early Tzeltal children will show this pattern of motion-eventplus-position description in the same clause.

3.2 Data For the largely non-literate population of Tenejapa, books are not culturally salient items, nor are they a part of Tzeltal children's lives until they go to school. In addition, the traditional narratives which used to be an important part of socialization practice are nowadays very much in decline. However, storytelling in the form of gossip and everyday narratives (what happened today) is a well-developed conversational genre, and visual representations in the form of photographs have been part of the culture for over twenty years. The story as portrayed in the pictures has several cultural oddities from a Tzeltal point of view. Chief among these is the fact that frogs would never be considered pets, and therefore none of my frog-story tellers took the main focus of the journey to be a search for a lost pet frog; nor do they usually identify the frogs at the end with the frog at the beginning. Nonetheless, for all the adults and many of the children it was clear that a journey was at issue (perhaps going out for a walk, or hunting), motivating the narrative style appropriate for journey stones, although in many cases this was taken to be a relatively goalless journey.

42

Penelope Brown

Table 1. Tzeltal frog-story texts Age group

Number of texts

Number of tellers

Total number of clauses

3-5

11

8

800

6-7

13

10

784

8-13

8

7

658

Adults

10

10

1699

3.3 Method Frog stories were collected from ten adults and nineteen children, in videotaped sessions mostly in the subjects' homes. A pure frog-story approach (Berman & Slobin 1994:22-25) was unworkable in this community, since even some adults tended to construe the task as a picture-labelling exercise. Therefore I or another caregiver first went through the storybook with the subject, identifying the protagonists and the objects portrayed. Then the narration event was conducted in a culturally natural interactive way, with other interlocutors (siblings, caregivers, the investigator) interjecting conversational responses and prompting questions. Several of the children provided multiple retellings on different occasions; all of their stories are included in the analyses. For this analysis 1 have edited out irrelevant picture-descriptive detail (about lamps, shoes, moon, etc.), looking only at event descriptions which move the narrative forward through time and focusing on those expressing motion and/or position. The children are grouped into three age groups: 3-5 (prior to schooling), 6-7 (the beginning years for schooling), and 8-12 (school children). Table 1 gives the details for each group.

4. ADULT TZELTAL FROG STORIES The narrative style of adult storytellers in this task involves introducing protagonists and scene-setting elements with the predicate ay 'exist', often specifying their position as well as existence, using positionals for static setting description, and using motion or action verbs along with directionals and positionals to move events and actions through time. Here is a relatively literal translation of one adult frog story, w ith irrelevant detail edited out, to give the narrative flavor:


There's one little boy, with his little dog. But there's a frog there. The little dog wants to get it coming out. He sees that it will come out if he grabs it. There's a little girl. She's asleep on her bed with her dog. ... Two little shoes and a tortillamaking table, and a bottle. But the frog is jumping coming outwards. On the ground there's a piece of cloth dumped-in-a-heap. Maybe it's the sleeping girl's shirt. It's nighttime. ... There, lying-face-down on the bed is the little boy. There's something he's looking down at. Maybe he's looking down at the little shoes standing there. And the little shirt, it's still spread out there on the floor. But the little frog isn't there. It's gone. I don't know where it went. There's a little boy again, he's playing with his shoe. There's something he's dumping coming out of it. And the little dog, it looks like there's something he's trying to get, but the little dog doesn't catch anything. There's nothing visible there. ...And the little tortilla-making table, the dog tipped it over. And here the little shirt is hung up, with the little boy's pants. Here, there's the boy again standing at the window, with the dog, who looks like he wants to catch something. But he didn't get anything from inside the little bottle. For no reason is his nose entered in there. Here again, the boy is looking out of the window. But there's the little dog again there, he thinks that the little frog is there, he inserted his head in there again. But he threw himself coming down, the little dog fell coming down. He arrives at the ground. The little boy has 'lost his soul' (been scared). He raised up his eyes because of it. He's afraid his dog will die. But now he descended coming from his window, he came and picked up his little dog. He's sad. He gets licked, he gets kissed, he gets loved (by the dog), because he got looked for by his master. Now here where there's a big tree, there the little boy is standing, with his little dog. Maybe he wants to grab the little birds, lots of them are flying around. But he doesn't catch any; dogs cannot fly. The boy is sad because he wanted to catch the birds. "Why don't I catch any?" maybe he asks himself. But the dog doesn't catch any, because they've flown.... Here's his little house, and they've come to the woods, there are bees there. And a hollowed-place in the tree. There's nothing there, it's just where water puddles. The boy and dog aren't there. Now here again is the little boy under the tree. They arrived coming out there where the big tree is standing; there's the little dog. They arrived to see where a bee's nest is hanging. The dog wants to get it, but the bees came out. There they are flying you see. The little dog wants to get them down, but he can't reach them, they are hung up. The little boy is kneeling on the ground, he's waiting for the honey to fall coming down. But the boy sees there's a hole in the ground, he looks. But a 'groundhog' came out there. But the little dog, he wants to get coming down the bees, he doesn't see his master nearly got bitten by the 'groundhog'. But the little boy was really scared, so his soul went. The dog scratches on the tree, the bees fell coming down because of it. The bees came out (of the nest). Lots came out, lots of bees came out. He's scratching it coming down toward himself. But the little dog doesn't get bitten; he doesn't get bitten. And his master has fled, he ascended the tree. He's afraid he'll be bitten by the bees. And he's afraid of the 'groundhog'. That's why he climbed the tree. Because he won't be caught up there. He looks into the hole in the tree, I don't know what he'll find there now. Then he saw one bird come out there, where the belly of the tree has a hole in it. It's maybe what

43

44

Penelope Brown

we call an 'owl'. The boy was really scared, he fell coming down face-up to the ground. The little dog was chased coming by the bees. There he is running away. He's scared away. I don't know where he ran away to, to his land perhaps. And his master, he has fallen coming down. The little bird flew away, the owl we call it. Then the boy feels better. "It's gone" he says. "They've all gone, nothing got me" he says. "But I've lost my little dog, I guess" he says. The dog is nowhere to be seen. He goes to yell to the top of a stone. He calls the dog. But he (dog) has hidden himself beneath the stone. Look where the bird that they are afraid of is looking towards them. I don't know what this is, it looks like a deer. He didn't see it come, he got perched onto the deer. He's mounted on its head. The little dog looks for his master, he smells to find where his master went. He's been perched on the deer; the deer has run away. There the little boy is mounted awaywards. "Ijj, my master has gone!" says the little dog attached there awaywards. Let's see where they go now. He's been thrown by the little deer. The little boy has fallen down. The little dog arrives and makes a bed of his master. Upside-down faceupwards to the ground, the little dog. He fell down from high up. He arrived at a lake. This is the lake, and woods, where they are playing. The boy and the dog are immersed in the lake. Sitting they are sad. "We nearly died" they are thinking. The water is low-down (shallow), you can see the boy is sitting. "Here's there's a big tree stump," says the little boy. He went and grabbed it. there he got out. But the stump has a hole in its belly again. I don't know what he'll be scared by next. He's just getting scared. So the little boy puts himself tummy-down there, he comes out. And the little dog, he comes out. The little dog says. "There's something here!" The little dog saw it first. "There's something here. There's something squatting here" he says. "Oh". It's the little frogs again. The little boy was really scared again. The whites of his eyes raised up. he was really scared. He was sad, with the little dog too. "Shall we grab them, or not?" "Maybe, there are only two", he says. But only then does he see that there are lots. Maybe this holey place is their house here. So they are scared. "Maybe they won't bite." he says. The little boy looks like he's going to laugh. So the little boy grabbed one. He grasps it in his hand. The lake is shallow where they are. It's shallow, the little dog is going across the middle of the lake. But look where other frogs are squatting on top of the log. But there's one who wants to ascend too. because he's only a little one, he can't get up. But his companions have gone up. But it's high up where they are. There it finishes." This story illustrates the uses of positionals and motion/path descriptors for the journey as construed by this teller. Clearly, many clauses construe events in terms of motion alone, and many others encode position of protagonists or aspects of the setting alone in a clause. But sometimes motion and position occur together in one event construal, as when the little boy is described as being 'mounted awaywards' - i.e., mounted on the deer while the deer moves awaywards - and when 'he fell downface-up to the ground' - i.e., while he fell he was face upwards. Looking, across all the adult stories, only at clauses which describe the activities/positions of the protagonists (boy, dog, frogs, bees. bird, and deer).


45

and ignoring for current purposes clauses describing just the setting (fallen-down logs, tipped-over tables, etc.), we find many clauses that combine stasis (adjectival position) with motion. Some examples of the different possibilities follow: /'.

Scene-setting descriptions (ay 'exist' + positionals):

[5]

ay ala kerem, nakal, sok jkojt ala tz'i'. There's a little boy, he's sitting, with one little dog.'

[6]

ay ala kerem, pakal ta swaeb. 'There's a little boy lying-face-down on his bed.'

//'.

Event-motion descriptions (single action or motion verbs, or motion directional):

[7]

sok ayix pokok ochem koel ch'ix yakan ta ala plastiko 'And there's now a frog whose one leg has entered downwards in the little plastic thing.'

[8]

pero ya xwil tal lok'el ala xpokok tey a. 'But the little frog there jumps out towards us.'

[9]

ya xben, spisil ya xbenik ta te'tikil. 'He's walking, both of them are walking in the woods.'

[ 10]

ch'ay taj. koel i kerem i. 'The boy fell coming downwards.'

///'.

Static scene descriptions (position only, position and direction, or verbinto-position):

[11]

tik 'il pokok tey a. 'A frog is inserted-in there.'

[12]

in te tz'i'e, ay bijipil ta te', ya syaan talya stejk'an sba yu'un. 'As for the dog, there's something hung in the tree, he's stretching towards it, he stands-himself-on-hind-legs because of it.'

[13]

koel tz'imil sni' me ye tz'i stz'i' i. 'His dog's nose is hanging-extended downwards'

46

Penelope Brown

[14]

te xpokoke jukul ya xk' abuot. The frog [= groundhog] is being looked at squatting.'

[15]

ay jukul j\\e\ jkojt. 'There's one (frog) squatting left behind/

[ 16]

ay jipil tal koel xux. There are bees hung coming downwards.'

;v.

Non-motion and direction mixed:

[17]

ya sk'abu ochel. 'He looks into it (bottle).' [lit.: 'looks entering']

[18]

ya sjutz'ita ochel xpokok ta yutil ala limete. 'He smells inwards to the frog at the inside of the little bottle.

[19]

la sk'ok tal koel xux tzT i. The dog broke off the bees down towards him.'

[20]

li' to k'alal ta meru ya' moelix te step. 'All the way upwards to his thigh are the shoes.'

v.

Position and motion/direction mixed (see also Examples

[21]

te ala tz'i'e kotol ochem sjol ta banti ala xalu. The little dog, standing his head has entered the little bottle.'

[22]

ch'ay koeljawal niwan ek. 'He fell downwards face-up perhaps too.'

[23]

xpejkunaj xben yilel. 'He [dog] looks like he's low-crouching walking [= he's limping].

[24]

kejel ya xch'oj bel. "Kneeling he throws it (frog) awaywards.'

[25]

kojkoltza' ochem. 'Upside-down entered in [dog's head, into bottle].'


vi.

47

Position and non-motion verb mixed:

[26]

kerem, tek'el spetoj stz'i'i. 'The boy, standing he has picked up his dog.'

[27]

kerem tek'el ya stik xchoj band lekbe te tz'i'. 'Boy standing he's feeling his cheek where the dog licked him.'

[28]

tz'i\ tek'el ya sk'abu moel xux. 'Dog, standing he's looking up. at the bees.'

The interpretation of such combinations depends of course on the component semantics. With a non-motion action verb and a directional, what is specified is the direction or orientation of the action (as in Examples [17-19]). With an existence verb or no verb, you simply get direction of a static array, as in 'upwards to his thigh' in Example [20]. With motion and direction it may be the same path reiterated in both motion verb and directional (as in Example [10], 'he fell coming downwards'), or it may be Motion + Path with the directional supplying the direction away from Source or towards Goal, as in Examples [7-8]. With a positional, what is conveyed is the Figure's position during or as a result of the action. And with Position + Path, the position of the Figure combines with direction of motion, as in luchul bel 'mounted awaywards', i.e., he was mounted on the deer and moving awaywards.

5. ADULT AND CHILD NARRATIVES COMPARED Tzeltal adults mix Position with Motion; both are important aspects of the frog story for them. But do children? And at what age? I took a central sequence where an elaborate trajectory is portrayed (the 'cliff scene'; see Pictures 14-19 in Appendix I): the scenes illustrated show the boy climbing onto a rock, being carried off by a deer, thrown over a cliff into the water below, and sitting up with the dog on his head. In descriptions of this cliff scene, all clauses which express motion or path or position, or some combination of these, were analyzed to see what proportions are used of different semantic elements in different forms. The proportion of such clauses which express Position alone (with positionals), Motion/Path alone (with motion verbs or directionals), or Position + Motion/Path, are given in Table 2, for adults and for the three groups of children. As Table 2 shows, Position alone or Motion/Path alone are preferred by the youngest children aged 3-5; for example (Position alone in [29-30], Motion/Path alone in [31]):

48

Penelope Brown

Table 2. Cliff scene - position and motion/path expressed in clause Age group

[29]

Position alone

Motion/Path alone

Position and Motion/Path

Total clauses Motion and/or Position

3-5

24 (52%)

6-7

20 (23%)

21 (46%)

1 (2%)

46

62(71%)

5 (6%)

87

8-13

20 (25%)

41 (52%)

18(23%)

79

Adults

74 (39%)

99 (52%)

17(9%)

190

t'umul sok yala tz'i'. 'He [boy] is immersed-in-water with his little dog.' (Pet, 3;5)

[30]

chawal ch'in kerem. 'The little boy is lying-face-upwards.' (Sil, 5;0)

[31]

moemtate'. 'He [boy] has ascended (into a) tree.' (Mai, 3;0 )

There is one example of Path combined with Position in the three- to five-year-old children's stories: [32]

jok 'ol tal koel yakanik. 'His feet are hanging coming downwards.' (Lus, 4;6)

But the very high proportion of clauses encoding Position alone (52%) reflects the fact that at this age, the task was to a high degree treated as a picture-description task, not a narration. By age 6-7 there is a big increase in Motion/Path descriptions with a motion verb or a directional, or both; e.g.: [33] ch'ayix tal koel i ala kerem i. 'The little boy has fallen coming downwards.' (X'an, 6; 1)

Clearly, by age 6-7 the children display considerable narrative attention to Motion/Path, more in fact than do the older groups, suggesting the possibility of a U-shaped developmental curve reflecting their emphasis on Motion and Path, although at this age they still are not using much Motion/Path + Position combined. By this age they have the adult-like pattern of combining motion verbs with directionals routinely to specify the direction of motion. This is virtually obligatory (stylistically but not grammatically) with verbs of falling, even though


49

semantically the directional is redundant with the verb. But the children of this age group also use directionals with many other verbs - both motion and non-motion verbs ([34-38] from Mat (6;0); [39^12] from X'anton (6;!)): [34] ya xlok' tal. 'It [the frog] comes out towards (us).' [35]

lok'em bel. 'It [the frog] has exited awaywards.'

[36] och tal te limete ta sjol tz'i' i. 'The bottle entered onto [lit: 'entered coming'] the head of the dog.' [37]

xux i ya sti' moel te tz'i'. 'The dog barks ug at the bees.'

[38] l i ' a y tal koel tz'i'. 'Here's the dog towards (us) downwards'. [39] ya yil koel ala xpokok i tz'i' i. 'The little dog sees downwards the little frog.' [40] ya x'ochix bel ta van ton i yala tz'i' i. 'His little dog enters awaywards at the bottom of the stone.' [41] xk'echlayej julel yakan. 'He [the boy] has carried high arriving his legs.' [42]

ya k'an mook bel sk'ok tal i ala tz'i' i. 'The little dog wants to go up awaywards to pick it [beehive] towards him.'

[43] ay bi xi' tal koel. 'There's something he [the boy] has been scared coming down by.' This attention to directional meaning that adds to, rather than is redundant with, the verb's meaning contrasts with the finding of a similar U-shaped curve for Spanish by Sebastian and Slobin (reported in Berman & Slobin 1994:263^1). They found, in the younger group of age 4-5, a heightened use of directional adverbs (when compared with older groups) that were generally redundant with the verb's meaning, suggesting that "these children feel a need to 'reinforce' the directional meaning inherent in the verb of motion" (Berman & Slobin 1994:264). This cannot be the explanation for the heightened use of Motion/Path expressions with many different verbs by the Tzeltal six/seven-year-olds.

50

Penelope Brown

The expression of Position + Path in a single clause increases with age, for the three groups of children. For example: [44] kajal moel ta sjol te'tikil chij. 'He [boy] is mounted-on upwards the head of the deer.' (Pon, 6:3)

[45] xjawawet taj koel 'He [boy] is face-up-mouth-open coming descending.' (Alu, 6:5)

[46] chawalik k'oel. 'They [boy and dog] are lying-face-up-arms-outstretched

arriving-there.' (Sil. 6;1)

[47] pakal k'oel i ala tz'i' i. The little dog is lying-face-down arriving-there.' (X'anton, 6:1)

The relative scarcity of these clauses at age 6-7 (6% in the cliff scene) suggests that the children do not yet fully have the ability (or the stylistic inclination) to combine Motion and Position in one clause. In the group aged 8-13 we find a very large spurt in the expression of Position + Motion/Path. These are expressed concurrently in 23% of the cliff-scene motion and/or position-encoding clauses, a notably higher proportion than that for the adults (9%). Adults of course told much longer and more elaborate stories than the children, more than twice as long on average, resulting in many more motion and/or position clauses than among the children. (See Table 2.) This contrast between the behavior of the group aged 8-13 and the adults warrants another analysis, eliminating the clauses that have only position or only manner-of-motion verbs, and comparing just those clauses that encode Path with those that encode Motion + Path (columns 2 and 3 of Table 2). This is presented in Table 3. On the face of it, this pattern by the 8-13 group also reflects a U-shaped curve, possibly attributable to a heightened sensitivity to the language-specific pattern which adults manifest to a lesser degree. Another possibility is that these children are using positionals much more for irrelevant picture-description detail than the adults, who are using them for narrative coherence. However, a scan of their stories compared with those of the adults suggests that this is not the whole explanation for the differences; both adults and children mention many picturedescribing details that are not plot-advancing, and many stories lack a clear-cut plot. Finally, it must be noted that this 8-13-year-old group of children has more schooling and consequently more exposure to Spanish than most of the adults: a potential link with Spanish narrative style is worth exploring/ We turn now to consider what semantic elements can be combined in one clause in Tzeltal, and how these are manifested in the frog stories as a whole.


51

Table 3. Cliff scene - motion/path alone and combined position + motion/path as a proportion of path-encoding clauses Age group

Motion/Path alone

Position and Motion/ Path

Total Path clauses

3-5

17(94%)

1 (6%)

18

6-7

45 (90%)

5 (10%)

50

8-13

34 (65%)

18(35%)

52

Adults

67 (80%)

17(20%)

84

across different ages. Motion descriptions in Tzeltal can only specify Source or Goal in one clause, but not both. There is only one general-purpose preposition in the language, which can be glossed as 'at', 'to', 'from', 'in', 'on', etc., depending on the verb. Tzeltal cannot encode complex paths in one clause in the way that English allows: you cannot say the equivalent of 'He went from A to B via C', but only something like 'He left A, he went to B, he passed by C, he arrived there at B.' 8 So in Tzeltal we find trajectories from a Source, or towards a Goal, but not both in the same clause; they are not coded as the same event (taking the clause as representing an event). For example (from Source in [48-50], towards Goal in [51-53]): [48]

ay xpokok, ya xwil tal lok'el ta ti' limete." 'There's a frog, he's jumping coming out of the edge of the bottle.'

[49]

ch'ay tal koel ta sjol men chije. 'He fell down from the head of the deer.'

[50]

jipot koel ta witz 'He was thrown down from the mountain.'

[51 ]

echajikix k'oel ta ja' a. 'They are spread-eagledarriving-there at the water.' [spread-eagled upon arrival]

[52]

smeyoj wan k'oel ta sjol ek i tz'i'. 'He hugged arriving on his head the dog.' [hugged as he arrived]

[53]

spetoj koel ta lum. 'He carried it [dog] downwards to the ground.' [carried it while descending]

52

Penelope Brown

Table 4. Proportion of clauses with complex predicates Age group

Motion + Path

3-5

40 (20%)

6-7

73(19%)

8-13

76 (22%)

Adults

129(16%)

Other verb + Path

Motion/ Path + Position

Other verb + Position

Total complex predicates

Total clauses Mot./Path/ Position

30(15%)

6 (3%)

112(29%)

19(5%)

2(1%)

78 (38%)

204

3(1%)

207 (54%)

381

95 (27%)

25 (7%)

3(1%)

199(58%)

346

139(17%)

37 (5%)

15(2%)

320 (40%)

796

Nonetheless, even though complex trajectories cannot be conveyed with multiple prepositional phrases in one clause, other complex semantic packages can be expressed by a combination of motion verb plus directional or positional, as we have seen in the examples above. A crude index of how frequent such a strategy may be in Tzeltal narratives can be obtained by looking at all the complex clauses - with a verb plus a directional, or a verb plus a positional, or a positional plus a directional - in the frog stories. For this analysis, the complete frog story was used; the analysis summarized in Table 4 shows, for the adults and children, the percentage of complex predicates with different semantic elements encoded in one clause as a proportion of all Motion and/or Positionencoding clauses: /'.

Motion + Path (with a motion verb and directional; e.g., 'he is walking awaywards');

//.

other verb + Path (a non-motion, non-positional verb plus directional: e.g., 'it [dog] is looking upwards at him');

iii.

Motion/Path + Position (with a motion verb or directional and a positional; e.g., 'he is falling down face-up');

iv.

other verb + Position (a non-motion verb and positional, e.g., 'he was thrown face-up spread-eagled').

As we can see in Table 4, the practice of combining these semantic elements in one clause is frequent indeed - of the total clauses conveying either motion or Position or some combination of these, from 38% for the 3-5-year-olds to 58% for the 8-13-year-olds have a combination. For all the age groups, verb + Path (a verb plus a directional) is the favorite combination, but all age groups, even the youngest, have some positional + motion verb or directional combinations, and


53

Table 5. Percentage of simple versus complex Motion/Path/Position in position/motion/path clauses Age group

non-complex

complex path

complex position

Motion only

Position only

Motion + Path

other verb + Path

Position + Path

Position + other verb

3-5

31%

30%

20%

15%

3%

1%

6-7

26%

19%

19%

29%

5%

1%

27%

7%

1%

17%

5%

2%

8-13

18%

25%

22%

Adults

23%

37%

16%

Table 6. Percentages of Position versus Motion/Path in position/motion/path clauses Age group

Motion/Path (columns 1, 3, 4, 5)

3-5 6-7

Position (columns 2, 5, 6)

Total clauses Motion/ Path/Position

69%

34%

204

79%

25%

381

8-13

74%

33%

346

Adults

61%

44%

796

again, the 8-13-year-olds outstrip the adults in the percentage of clauses with both Position + Path encoded. Still, clearly Motion/Path is more crucial to the frog-journey texts than is Position, as measured by proportion of clauses encoding these semantic elements. For all age groups, the percentage of clauses expressing Motion/Path outstrips that expressing Position. Looking at the same data in a different way, Table 5 compares these same semantic notions when expressed alone in a clause with those occurring in combinations. Comparing expression of Motion/Path versus expression of Position - regardless of whether or not they occur in combinations - it is clear again that Motion/Path is more central than Position in these stories, at all ages. Table 6 extracts the percentages from Table 5 for Motion (columns 1, 3, 4, and 5) versus Position (columns 2, 5, and 6).

6. POSITION AND MANNER IN TZELTAL FROG STORIES How can we explain the heavy use of positionals in Tzeltal frog stories? Are they predominantly stereotypical combinations? On the whole, they are not; although this is indeed the case for some motion verb + directional combinations

54

Penelope Brown

(especially with verbs of falling), it is not the case for positional which very freely combine with many different verbs and directionals. Can we then account for them by suggesting that Tzeltal tellers are treating this task not primarily as a narration, but as a picture-describing device, and hence there is heavy use of positionals to describe static scenes? Although this may be part of the reason, especially for the younger narrators, it cannot explain the frequent use of positionals in combination with motion verbs in descriptions of scenes which move the narration forward. Except in the 3-5-year-old group, positionals are not being used just for scene-setting background details. Could positionals contribute in some way to narrative coherence? This is difficult for a non-native speaker to assess. For the most part, positionals as used here do not seem to be essential for discourse needs, for the addressee's understanding of the narrative, or for the temporal/spatial organization of the story. They are not on the whole being used to portray aspects of the story which, for these storytellers, are essential to the plot. Rather, I suggest, they provide color and interest to what appears to be (for many tellers) a relatively motiveless journey. What really seems to motivate the use of positionals is a language- and culture-specific property: speakers of Tzeltal tend to display a special interest in the configurational details of Figure and Ground in locative and motion scenes, and the language provides rich resources for this, allowing speakers to compress properties of the Figure (such as how it is positioned) with those of motion so that they are equally foregrounded in the clause. Sometimes this combination can express manner ("he fell lying-face-up") or resultant state ("he arrived sitting"). This suggests that we might gain some insight into the role of positionals by taking a look at how Manner is expressed in these stories. Tzeltal, as a verb-framed language, would be expected not to have an elaborate repertoire of manner-of-motion verbs, and even to express Manner only when this is foregrounded in the narrative (Oz^ahskan & Slobin, in press-a). It is indeed the case that manner verbs are very few in these stories. There are only a handful of types of manner-of-motion verbs ('walk', 'run', 'swim', 'jump/fly'), and of Manner + Path verbs ('flee', 'chase', 'fall'); for example: [54] te yakal ta animal te tz'i'e. 'There the dog is in the process of running.'

[55] ja' a snutz te yajwale, te luchul bel. 'He chases his master, who is perched awaywards there.' There are not many tokens of these manner verbs in the frog stories; Table 7 summarizes these for the adults and children.


55

Table 7. Manner clauses in Tzeltal frog stories Age group

Manner alone

Manner + Path

Total clauses Motion/ Path/Position

3-5

4 (2%)

22(11%)

204

6-7

17(4%)

15(4%)

381

8-13

7 (2%)

12(3%)

346

Adults

32 (4%)

49 (6%)

796

It seems reasonable to conclude that one reason for the proliferation of Position in frog stories is to provide some of the graphic visual detail for which, in satelliteframed languages, manner-of-motion verbs are used.9

7. CONCLUSIONS The 'Tzeltal type' of frog narration integrates Motion + Path (expressed with motion verbs and directionals) and 'Manner'-like stasis (expressed with positionals), to provide vivid snapshots of events in the frog journey. Tzeltal, like other Mayan languages, has rich resources for discriminating paths and even richer resources for nuances of position, configuration, and orientation. This characteristic of journey descriptions is compatible with the observation (Brown 1994) that, in Tzeltal, position is integral to descriptions of where things are and where they are going. This is a property of thinking-for-speaking which Tzeltal children have to learn. Even 3-5-year-olds are attuned to this propensity, as seen in the roughly equal incorporation of position and motion in their stories (Table 2), although for the most part they lack the syntactic ability to combine both in one clause. Yet the fact that position is already heavily attended to even at this age reinforces the conclusion from the Berman and Slobin research that lexicalization patterns shape narrative style, via the shaping of attention produced by habitual language use. The expression of Position and Motion/Path in one clause means that both can be foregrounded. This usage increases with age for children. However, the adults produce less as a percentage of total clauses than do the 8-13-year-olds, and, in any case, this is not a highly frequent linguistic strategy. This suggests that expressing Position + Motion/Path in one clause is a stylistic resource that children even at 3-5 are beginning to acquire; by the 8-13 age range they are outstripping adults, showing 'over-sensitivity' (Berman & Slobin 1994) to this language-specific feature of Tzeltal narrative style.

56

Penelope Brown

Though Tzeltal is a verb-framed language - since the preferred locus for Path is in the verb or the adverbial directional - complex predicates play an important role in motion description, and while most of these involve Motion + Direction, a significant portion of these (up to 7% in the frog stories taken as a whole; up to 23% in the cliff scene) convey Path and Position in the same clause. 10 The results of this exploratory study are broadly compatible with the Berman and Slobin findings on narrative development, despite the unfamiliarity of the task for these subjects and the consequent departures from the Berman and Slobin elicitation procedure. The prolific use of positional in combination with motion verbs in the Tzeltal frog stories provides evidence for a language-based sensitization to particular characteristics of the events described. Children of 3-5 do seem to be already attuned to features of narrative style in their language, even when non-obligatory, if these are accessible and easily expressed in the language. It remains to consider the effect of cultural framing of the task on these results. One effect of non-literacy, and of the cultural oddity of the task, on frog stories by adults is that some adults' stories appear to be non-'proficienf in comparison with their naturally produced (non-elicited) narratives; they do not always produce a cohesive, coherent frog-story text. However, their stories are certainly 'native'. Another important question is the extent to which having a visual stimulus (in this case, a picture-book) increases the amount of descriptive detail in narratives - would an alternative procedure (for example, having them first look at the book and then tell the story from memory) produce many fewer positionals? One presumes so. And finally, the difference between the 8-13-yearolds and the adults remains to be explained. One obvious difference between these two groups is schooling: few of the adults received more than a third-grade education, while most of the 8-13-year-olds had exceeded that and also had some familiarity with Spanish. The next step in this research is to compare the Tzeltal frog stories with naturally-occurring Tzeltal narratives, and to investigate the effect of schooling on the narratives of children and young adults.

NOTES 1.

A first version of this chapter was presented at the 30th Stanford Child Language Research Forum and published in the proceedings (Brown 2000). 1 am very grateful to Dan Slobin for detailed comments and suggestions for improvement, and to Seyda Ozcahskan for providing me with the latest version of the motion-verb coding scheme used in the Berman and Slobin project. I am also grateful to Jiirgen Bohnemeyer for on-going discussions on the nature of positionals in Mayan languages.

2.

There are occasional exceptions; Slobin (personal communication) cites a handful of English examples: 'lying', 'knocked over', 'hanging down', 'get stuck' (between


57

horns), 'landed on his back', 'lean over' (log), 'sitting up'. And Slobin (1996b) cites some Spanish examples with a positional endstate, e.g., el nino estd subido en el drbol. 3.

See Brown (in press) for the semantic and grammatical details of these Tzeltal forms.

4.

-V ( l is a suffix where the vowel matches the vowel of the root; for example tek' 'stand' becomes tek '-el 'standing', but nak 'sit' becomes nak-al 'sitting'.

5.

What I am here calling 'positional' includes predicates formed from the Tzeltal positional roots and also from 'bivalent' roots, which share some characteristics of the positional class but others of the transitive class (elsewhere I have called this superordinate class of roots 'dispositionals'). There is actually a cline - based on morphological proclivities - between these two large verb classes. See Brown (1994, in press); and Bohnemeyer & Brown (to appear); see also Haviland (1994), for more details of these verb classes in the two closely-related Mayan languages Tzeltal and Tzotzil.

6.

A clause is taken to be "a linguistic unit containing a unified predicate in the form of a verb or an adjective" (Berman & Slobin 1994), or in the case of Tzeltal, potentially both. Since both adjectives and verbs take absolutive endings to form a predicate, a sentence like ch'ay koel jawal niwan ek 'he fell down lying-face-up too perhaps' could be argued to have two clauses (ch'ay-0 koel 'he fell down') and jawa 1-0 niwan ek 'he was lying-face-upwards too perhaps'. But in the examples given in this chapter of such constructions, there is no prosodic or grammatical marking of clause boundary; they are treated as one communicative unit.

7.

Melissa Bowerman (personal communication) suggests a third possibility which remains to be explored - could the 8-13-year-olds actually be using positional + verb or directional creatively in ways that adults would not consider acceptable Tzeltal, over-extending the potential of these resources to combine freely in a clause? This is difficult for a non-native speaker to assess, but running all the 8-13-year-olds' examples past Tzeltal adult assessors would be a valuable first step.

8.

See Bohnemeyer (in press-b) for detailed evidence of this point for another Mayan language, Yukatek Maya.

9.

The way in which Tzeltal spatial language is used to convey motion events integrated with stasis is reminiscent of a different language-specific way of doing this in another verb-framed language - the 'associated-motion construction' in Arrernte (Wilkins 1997).

10. This compares with 24% of motion-event clauses in another verb-framed language - Spanish - having complex predicates (Slobin 1996b); for example: solid corriendo, partieron zumbando, etc.


Focus on Action in Motion Descriptions The Case of West-Greenlandic

Elisabeth Engberg-Pedersen & Frederikke Blytmann Trondhjem

1. INTRODUCTION The studies of narrations of Frog, where are you? (Mayer 1969) in Berman and Slobin (1994) focused both on age-dependent but language-independent differences and on differences in style due to typological variation. Irrespective of language, children were found to go through the same development in plot narration. To the extent that the youngest children of three years old did not talk about other things or engage the experimenter in conversation, they tended to name items in the pictures or describe events in each picture independently of the other pictures; they did not tell a coherent story. Some five-year-olds, but not all, told stories with a global structure. These children made the individual parts of the story hang together: the boy discovers that the frog has disappeared, he looks for it, and he finds a frog or a substitute for the missing frog. Almost all nine-yearolds recounted both that the boy discovered that the frog was missing and that he looked for it, but only about two-thirds connected the frogs at the end with the frog in the beginning of the story. By contrast, differences in narrative style can be interpreted in the light of theories of linguistic relativity and linguistic determinism, or rather linguistic typological influence on information processing at the moment of language production. As an alternative to the theory that linguistic structure determines world view or habitual thought, Slobin (1991, 1996b) developed the concept of thinking-for-speaking, which implies "picking those characteristics of objects and events that (a) fit some conceptualization of the event, and (b) are readily encodable in the language" (Slobin 1996b:76). Through this double identification, Slobin leaves open the possibility that individuals conceive events independently of their language, but that in the process of describing the events, they focus on those aspects that can be "readily encoded", either lexically or grammatically.

60


When language users communicate their experiences to others, they compromise between their experiences and the expressive means provided by their language: "In the evanescent time frame of constructing utterances in discourse one fits one's thoughts into available linguistic frames" (Slobin 1996b:76). When children acquire their first language, they acquire specific ways of thinking-for-speaking. It is thus likely that at the local level - and not so much at the global level of plot development - retellings of the same picture story in different languages will emphasize different aspects of the story because the languages offer different "linguistic frames" or differ with respect to what is "readily encodable". In this chapter1, we will present a study of narrations of Frog, where are you? (Mayer 1969; see Appendix I) by five-year-olds, ten-year-olds, and adults in WestGreenlandic. We will focus on the development in global plot narration as well as the development of a local feature, namely descriptions of motion events. The West-Greenlandic children's score on the presence of the core components of the story's plotline is somewhat low compared with that of the children in Herman and Slobin's (1994) study. This result is interpreted as suggesting that differences in plot narration may vary not only with age, but also with language. West-Greenlandic is a polysynthetic language with very complex verbal and nominal morphology (for a brief introduction to its structure, see the Appendix). The verb is the obligatory part of the clause, and children acquiring West-Greenlandic seem to concentrate particularly on verbs and consequently have a fairly low score on core components of the plot: they tend not to use nominals that might identify what the boy is looking for (the second core component of the plot; see Section 3), nor to relate the frogs at the end of the story to the frog in the beginning (the third core component). West-Greenlandic belongs to the type of languages that Talmy (1991) calls "verb-framed". In descriptions of motion events, such languages predominantly code the path or trajectory taken by the moving figure in the verb stem, as in the English verb of Romance origin (he) entered ('went into') (the room). By contrast, "satellite-framed" languages predominantly express the path outside the verb stem, e.g., by means of verb particles or pre- or postpositions, as indicated by the "translation" in parentheses ('went into') in the English example. Aske (1989) further shows that Spanish does not always code the path in the verb even though it is a verb-framed language. What Aske calls "mere locative paths" may be expressed in satellites, as in la botellafloto hacia la cueva 'the bottle floated towards the cave' (after Aske 1989:3), while telic descriptions of an end-state of the figure must be expressed by a verb, as in la hotel la salioflotando 'the bottle exited floating', where it is asserted that the bottle ended up outside the cave. Satellite-framed languages such as English accept satellites that make the event telic, as in The bottle floated out of the cave. Slobin and Hoiting (1994) reinterpret the difference between the two types of path expressions as a difference between

Focus on Action in Motion Descriptions: The Case of West-Greenlandic

61

path focus and boundary focus. The types of path that must be expressed in the verb in Spanish focus on the figure's crossing a boundary (cf. the Spanish equivalents of enter, exit, and cross). Whenever the path crosses a boundary, a verb-framed language demands a new verb, as in "... elld [sic!] bajo la escalera y salio de la cascf (Slobin 1997a:441). A satellite-framed language may just add another satellite: ... she went downstairs and out of the house (translation of the Spanish sentence; after Slobin 1997a:441). The different treatment of boundary-crossing paths has immediate consequences for texts in the two types of language. Since verb-framed languages use a new verb whenever a boundary is crossed, Slobin proposes that "[t]exts in Vlanguages [verb-framed languages - EEP & FBT] will generally be characterized by the occurrence of fewer ground elements per verb, in comparison with texts in Slanguages [satellite-framed languages - EEP & FBT]" (1997a:442). This hypothesis is supported by surveys of texts in different languages (Slobin 1996a, 1997a). The typological difference cannot explain, however, why texts in verb-framed languages never have verbs with more than two ground-denoting expressions even when a protagonist moves with respect to several grounds without crossing a boundary. Slobin suggests that a general narrative style emerges in a language, based on systematic linguistic constraints, but going beyond them. The combined effect of a lexicon of pathverbs along with a boundary-crossing constraint is a style in which most path segments are encoded by separate verbs with a limited number of ground nominals per verb. The result is a sort of "rhythm" that becomes a habit or a norm, setting a framework that goes beyond its core linguistic determinants. (1997a:443; author's emphasis)

Slobin (1996a) finds that Spanish frog stories have more bare motion verbs, i.e., motion verbs without any satellite, than English stories; that English stories include more elaborate descriptions of complex paths than their Spanish counterparts; and that stories in Spanish have more static setting descriptions. The narrative style of the language as a norm or habit is thus a further determinant of thinking-for-speaking. The analysis of West-Greenlandic frog stories confirms the hypothesis that stories in a verb-framed language have more bare motion verbs and fewer ground-denoting expressions with motion verbs than do stories in a satelliteframed language, but the West-Greenlandic stories have remarkably more bare motion verbs and fewer ground-denoting expressions than even the Spanish frog stories. Moreover, the West-Greenlandic stories are considerably closer to the English style when it comes to the elaboration of path descriptions distributed over several clauses. The contrast between satellite-framing and verb-framing is thus not sufficient to explain the differences in style.

62


After a presentation of the narrators and the data in Section 2, we describe the development in plot narration in the West-Greenlandic stories in Section 3. Section 4 is devoted to the results of the analysis of the four aspects of motion description first presented in relation to English and Spanish frog stories in Slobin (1996a). In Sections we suggest a typological explanation for the apparently conflicting results of the analysis, including the possibility of influence from cultural factors as suggested by Wilkins (1997).

2. THE WEST-GREENLANDIC NARRATORS The West-Greenlandic frog stories were collected in a kindergarten and a school in the capital of Greenland, Nuuk, and among adult Greenlanders in Copenhagen. Denmark.2 The number of stories collected and the narrators' ages are shown in Table 1. As a region of the Kingdom of Denmark with home-rule status, Greenland is a bilingual society. Most children live in homes where West-, East-, or NorthGreenlandic is the only or the predominant language. All schools offer courses in Danish, and many schools at least in Nuuk also have classes where Danish is the medium of instruction. Because of a shortage of qualified Greenlandic-speaking teachers, a number of teachers speak Danish only. All the parents of the children from the kindergarten in this study indicated that the language spoken in the home was West-Greenlandic or - in one case - West-Greenlandic and Danish. All the ten-year-olds attended a class where West-Greenlandic was the medium of instruction, and the parents of four of these children indicated that the language used in the home was West-Greenlandic or predominantly West-Greenlandic. AlO's and FlO's parents indicated that the language used at home was Danish and predominantly Danish, respectively. Most of the parents have graduated from high school and have some further education (up to four years). The adult narrators all reported using West-Greenlandic at home. Their educational level is equivalent to the educational level of the children's parents. The children's stories were video- and audio-recorded in the kindergarten and the school. During the recording sessions, two adult native speakers of WestGreenlandic were present, one concentrating on the recording equipment and the other instructing the children. The only language used was West-Greenlandic. The adult narrators were video- and audio-recorded in their homes or in Frederikke Blytmann Trondhjem's home; here again, only West-Greenlandic was used. The instructions were a translation into West-Greenlandic of the instructions found in Berman and Slobin (1994:22). The narrators were told that the book is about a boy, his dog, and his frog. They were asked to look first at all the pictures and then tell the story. They were encouraged to go through the book again while telling the story.


63

Table 1. West-Greenlandic narrators in the study

5y jars

Adu Its*

l O y ears

ID

age

ID

age

ID

age

A5

4;10

A10

10;0

AV

17

B5

4;10

BIO

10;2

BV

42

C5

5;5

CIO

10;1

CV

41

D5

5;0

DIG

10;1

DV

33

E5

4;11

E10

9;4

EV

40

F10

9;8

*'V for voksen 'adult'.

3. DEVELOPMENT IN PLOT NARRATION As Frog, where are you? (Mayer 1969) is "clearly a product of Western culture" (Berman & Slobin 1994:21), it could be expected that Greenlandic children would find it difficult to relate to it. Furthermore, the story includes animals and vegetation not found in Greenland. In one picture, the boy holds on to some twigs which turn out to be a deer's antlers. In the area around Nuuk there are no trees, only moss, lichen, and very low shrubs. Nevertheless, the children did not seem to have more problems with the storyline or the animals and the vegetation than, for instance, Danish children (Engberg-Pedersen 2000). Children in Nuuk are used to watching children's TV programmes from Denmark, including many series from other Western countries. The Greenlandic children and adults had problems naming the gopher, some describing it as a lemming. Some called the deer a reindeer, and some the owl an eagle, but without any signs of hesitation. Frog, where are you? (Mayer 1969) has a plot which is typical of Western narrations. There is a problem which motivates the action, the action develops as the protagonist attempts to solve the problem, and at last the problem is solved. Berman and Slobin describe the core components of the plot in the frog story thus: [...] we stipulated three elements of the story as criterial of the ability to relate the contents of the picturebook as an integrated whole. These core components are: I - the onset of the plot: the boy's realizing that his frog has disappeared; II - unfolding of the plot: the boy's search for his missing frog; and III resolution of the plot: the boy's finding the frog he has lost (or one to take its place). (Berman & Slobin 1994:46; authors' emphasis)

64


The first component is present in the story only if the narrator explicitly states that the boy realizes that the frog is gone. It is not enough for the narrator to say only that the frog is gone or that the jar is empty, since only the boy's realization that it is gone leads up to the following search. The second core component is present only if the narrator says that the boy is searching for the frog; just mentioning that the boy is searching or shouting is not enough. Moreover, the narrator must say that the boy is searching for the frog in the forest, not just when he is inside the house. Otherwise, the central part of the story, the boy's and the dog's adventures in the forest, do not contribute to the plot. The presence of the third core component depends on whether the narrator says that the frog which the boy finds or the one he takes home at the end is either the same frog as the one that disappears in the beginning or one to take its place. Without this explicit link between the beginning of the book and the end, the plot does not appear as a global whole. The study presented in Berman & Slobin (1994) demonstrated that the older the narrators, the greater the chance that they included one or more of the three core components, no matter which language they spoke. Table 2 shows the figures from narrations in five languages. Berman and Slobin's criteria for presence of the core components may seem very strict. The five-year-old child D5 says, for instance (see the Appendix for abbreviations used in the examples): [1]

aana uku-a here

iter-mata

naraseq pe-qan-ngil-aq

DEM-RELIPL wake.up-CAUs:3pL frog:ABS

be-have-NEG-iNDic:3sG

'Here [in the picture] when those there wake up, the frog is not there.' (D5)

Does D5 understand that the boy's realization that the frog is gone is what motivates the action, i.e., the search in the forest? It seems so, since at the transition from the section about the search inside the house to the search in the forest, she mentions that the boy is looking for the frog. Strictly speaking, she does not fulfil the criterion for the first core component, however. Another child, B5, does not mention that the boy sees that the frog is gone, but he does use quoted speech saying "Where are you?" repeatedly and "Are you in there?" in relation to the picture where the boy is searching the hole in the ground, thus demonstrating that he knows that the boy is looking for someone - but does he know that the boy is looking for the frog? We cannot know. In Table 3, yes under Component I means that the narrators fulfil Berman and Slobin's strict criterion. No/yes means that the narrators do not fulfil the strict criterion but do mention that the frog is gone and demonstrate, by immediately afterwards saying that he is looking for "it" (a transitive verb) or by subsequently saying that the boy is searching for the frog (with an explicit nominal object), their knowledge that the boy has realized that the frog is gone.


65

Table 2. Percentage of narrators across five languages Element

3 years

5 years

9 years

Adults

1

17%

78%

94%

100%

II

15%

52%

98%

100%

III

10%

41%

62%

92%

The figures refer to the percentage of narrators fulfilling the criteria for each of the three core components in Berman and Slobin's study (1994:48). Each age group comprises 58 narrators.

As regards the criteria for the second core component, the narrator must say that the boy is searching for the frog, and that he is doing so not just when he is inside the house but also when he is in the forest. Problems with these criteria arise, for instance, in relation to C5's story. In relation to Picture 14, where the boy is seen standing on a stone, cupping his hands in front of his mouth, she says: taava-lu then-and

suaar-amiuk

suaar-amiuk

shout-CAUS:4SG:3SG

shout-CAUS:4SG:3SG

pakkaluaq butterfly:ABS

eeh ehm

sanioqqutip-poq3 PASS-INDIC:3SG

'And then when he calls for it/him/her, a butterfly passes by.' (C5)

The first verb in [2] (suaaramiuk) is transitive, its suffix indicating both a subject and an object. That is, C5 clearly indicates that the boy is shouting for something definite, but not that he is shouting for the frog - and that is criterial for this component of the plot. E5 uses the same verb intransitively with quoted speech in relation to the same picture: [3]

taava

sujaartor-ami4

then

shout-CAUS:4SG

anaanaa mummy

'Then he shouts: "Mummy!" ' (E5)

E5 clearly does not fulfil the criterion for the second core component and demonstrates how uncertain the interpretation of C5's sentence in [2] is. C5 too may be thinking that the boy is shouting for his mother. C10 demonstrates yet another variant. She uses an intransitive form of the verb itjaasi- 'search', which means that she does not link the search with the

66


Table 3. Presence or absence of the core components in the West-Greenlandic stories (cf. Berman &Slobin 1994:46) Subject

Component I

Component II

Component III

A5

yes

no

no

B5

no

no/yes

no

C5

no

no/yes

no

D5

no/yes

no/yes

no

E5

yes

no/yes

no

A10

yes

no/yes

yes

BIO

no/yes

no/yes

no/yes

CIO

yes

no

no/yes

D10

yes

yes

no

E10

yes

no/yes

no/yes

F10

yes

yes

no/yes

AV

yes

no/yes

yes

BV

yes

yes

yes

CV

yes

yes

yes

DV

yes

yes

yes

EV

yes

yes

no

Component I no/yes: : the narrator mentions that the frog is gone and later that the boy is searching for 'the frog' or for 'it', but not that he realizes that it is gone. Component II no/yes: the narrator uses a transitive verb ('is searching/shouting for it") or quoted speech ('Where are you?') but does not mention the frog. Component III no/yes: the narrator uses a transitive verb meaning, for instance, 'see', or an intransitive verb meaning, for instance, 'be there', without an explicit nominal that implies that the object or subject, respectively, is known from earlier in the story.


67

overall plot. Several children use transitive forms of verbs meaning 'shout' or 'search' in descriptions of Picture 8, where the boy is seen between the house and the forest, but in their description of his search in the forest they use intransitive forms of the same verbs. In Table 3, yes under Component II indicates that the narrator uses the noun naraseq 'frog' as the object of a verb meaning 'look for', 'search for', etc. in the middle section of the story, while no/yes indicates that the narrators say only the equivalent of 'is searching/shouting for it', i.e., use a transitive verb with a suffix that indicates an object without specifying its nature, or that they, like B5, use a quotation relevant to the search for the frog ('Where are you?'), but without mentioning the frog. None of the five-year-olds fulfils the criterion for core component III, i.e., none links the frogs that the boy sees or the frog in his hand at the end with the frog that disappears in the beginning. In describing the boy's encounter with the frogs at the end of the story, two ten-year-olds, D10 and E10, use transitive forms of verbs meaning 'hear' and 'find' respectively: [4]

taav(a)

tusaa-vaa

narase-qar-toq

then

hear-iNDic:3SG:3SG

frog-be-PART.Mooo:3SG

'Then he hears there is frog.' (D10)

[5]

taava then

nassaa-ri-vaat find-have.as-iNDic:3PL:3sG

naraseq frog:ABS

Then they find a/the frog.' (E10)

In [4], D10 uses a verb in the participle mood derived from the noun naraseq 'frog', thereby making the meaning 'frog' less individualized. DIG thus cannot be said to make it explicit that the frog at the end has anything to do with the frog in the beginning of the story. West-Greenlandic does not indicate definiteness in the nominal. An object in the absolutive case with a transitive verb showing both subject and object agreement signals high discourse value of the object as contrasted with one of the less transitive verb forms, but the referent of the object need not be given. The most neutral word order is SOV, while the order VO puts emphasis or focus on the object since the position after the verb is most strongly felt to relate to new information (Fortescue 1984a: 181), but the referent of a postverbal object need not be new. In [5], the word order may simply indicate that after much searching, they did indeed find the frog. That is, a transitive verb form is most often used when the referent of the object is given, but it may be new, and the word order VO is most often used when the referent of the object is new, but

68


it may be given. Strictly speaking, then, [5] codes the referent of the object as neither given nor new. The other four ten-year-olds use transitive or intransitive verbs without any nominal argument. Their verb forms imply rather than make it explicit that the object or subject refers to the frog from the beginning of the story or to whatever was missing. F10 says that the boy asks the dog to be quiet and continues: [6] shh shh

imaar(qa) maybe

tassani-i-ssa-pput there-be-FUT-iNDic:3pL

tulli-a-n(i)

taava then

taku-gamikkut see-CAis:3R:pL:3SG

taku-aa(t)

next-3so:POSS:SG-LOC

see-INDIC:3PL:3SC

' "Shh! Maybe they/some are there." Then they see it. In the next [picture] they see it.'(FlO)

The listener has to infer that the frog seen by the boy and the dog is identical to the one in the beginning of the story. A10 also uses a transitive verb without an explicit object nominal, but this time the lexical meaning of the verb makes a certain connection with an earlier part of the story clearer: [7] piara(q)

ataase(q) tigu-sima-gamiu(k)

young.one:ABS onerABS

take-PERF-CAUS:4SG:3sd

eqqoriar-lugu guess-CONTEMP:(4SG):3SG

'He took one young one assuming that it is the one.' (A10)

The meaning of the verb eqqoriar-lugu (guess-coNTEMP.3Rsg.3sg) implies that the referent is known from earlier. With respect to core component III, no/yes in Table 3 indicates cases where the child uses a transitive verb form without an explicit nominal object that refers to the frog, or uses a form ofnaraseq 'frog' in the absolutive case with a transitive verb. In these cases, it is implied rather than stated explicitly that the boy finds the frog that he lost in the beginning. All the adults fulfil the criterion for the first component by saying that the boy sees that the frog is gone or that he cannot find 'the frog' or looks for 'the frog' in vain. As for the second core component, the youngest of the adult narrators (17 years old) uses only a transitive verb without an explicit nominal argument, like some of the children (no/yes in Table 3). The adults otherwise use various strategies to indicate definiteness. One uses a - definite - proper name, Takorna. meaning 'stranger', to refer to the frog both in the description of the boy's search in the forest and in the description of his encounter with the frogs at the end. Some use anaphoric demonstratives in the nominals that refer to the frog (underlined


in [8]), or possessive constructions (underlined in [9]), thereby making the expressions definite: [8]

...

taku-aa

taan(na)

see-INDIC:3SG:3SG DEM:ABS

narase(q)

narasi-m(ik)

alla-m(ik)

frog:ABS

frog-INSTR

other-INSTR

ila-qar-toq companion-have-PART.MOOD:3sG '... he sees that frog with another frog as its companion.' (AV)

[9]

Piita-li

suaartar-tavar-poq

Peter-and

stand.and. shout-continuously-IN D : 3 SG

narasi-uti-ni frog-alienable.possession-4so:POSS:SG

suaartar-lingu stand.and.shout-CONTEMP:(4SG):3sG

'And Peter keeps shouting as he is shouting for his frog.' (BV)

The strategies used by the adults to make the expressions denoting the frog definite are thus the following: a proper name, nominals with an anaphoric demonstrative, and nouns in possessive forms linking the frog with the boy. Berman and Slobin claim that the development in plot narration is independent of the language spoken. Compared with the children in their study, the West-Greenlandic children include fewer core components - unless we use more lenient criteria, especially accepting transitive verb forms without an explicit argument referring to the frog. In relation to component II, an explicit nominal argument naraseq 'frog' as the object of a verb meaning 'shout for' or 'search for' would make it clear that the boy was looking for a/the frog, thereby fulfilling the criterion for the second component. In relation to component III, a transitive verb without an explicit object would make it clear that the boy found something given, but not what he found. A transitive verb with an explicit object in the absolutive would indicate that he found 'frog', but not that the frog was given. The adults who do make it clear that the frog at the end is identical to the frog at the beginning use a number of optional strategies, some of which are structurally simple while one requires quite complex morphological marking of possession on the noun. The number of narrators in this study is too small for it to be conclusive, but it seems that the language spoken may influence the number of core components present in children's stories at different ages. Verbs in West-Greenlandic include information about the subject and, for transitive verbs, the object, and the only obligatory part of the clause is the verb. The children's stories generally have few

70


nominals compared with verbs, which suggests that their style may be influenced by the requirement that they acquire the verbal morphology, something which in its turn influences the extent to which they fulfil the criteria for core components II and III. They leave out the nominal necessary for component II and, along with the nominal, also the means necessary for fulfilling the criterion for component III, i.e.. for marking definiteness or .co-reference one way or another. The adult stories amply demonstrate that the core components can be expressed in West-Greenlandic.

4. MOTION IN WEST-GREENLANDIC VS. ENGLISH AND SPANISH 4.1 Path denotation and ground denotation in West-Greenlandic Talmy's (1991) typological distinction between verb-framed and satellite-framed languages combined with Aske's (1989) suggestion that verb-framed languages always need a new verb to indicate telic - or in Slobin and Hoiting's (1994) analysis: boundary-crossing - motion led Slobin to predict that frog stories in verb-framed languages would include more bare motion verbs and fewer grounddenoting expressions with motion verbs than frog stories in satellite-framed languages. West-Greenlandic is a verb-framed language in which the path can be expressed in the verb root, as in the following examples (verb roots underlined): [l0a] taava then

naraseq

ani-voq

frog:ABS

move.out-INDIC:3SG

Then the frog crept out.' (BIO)

[l0b] taava then

orpim-mut tree-ALLAT

qaqi-sima-voq reach.the.top-PERF-INDIC:3SG

'Then he climbed the tree.' (A5)

[l0c] nakka-qa-aq fall.down-intensity-iND:3SG

igalasserfim-minngaanniit

nuna-mut

window. frame-ABL

ground-ALLAT

'It falls from the window frame onto the ground.' (CV)

As in other verb-framed languages, we also find verbs that express manner and not path in the root, such as pissi- 'jump' and arpa- 'run'. Some verbs denote motion and direction in relation to an implicit ground which may, but need not (cf. [10a]), be made explicit by means of a ground-denoting noun in one of the locative cases:


71

locative, allative, ablative, and prosecutive (Fortescue 1984a:226-228). Examples of such verb roots are qaqi- 'reach the top', iser- 'move into', pula- 'poke into', pi- 'arrive at', and tup- 'move into contact', which all imply a goal, and ani- 'move out', which implies a source. These verbs have the telic or boundary-crossing character described by Aske (1989), Slobin and Hoiting (1994), and Slobin (1997a). Roots that do not imply a ground, such as nakkar- 'fall down', may also occur with explicit adjuncts denoting the source or the goal, as in [lOc], In the West-Greenlandic frog stories, the path and the ground in the motion descriptions are expressed in the following five ways outside the verb (cf. Fortescue 1984a:226ff, 259ff): /'.

Noun-derived directional adverbials. Adverbials denoting direction which are derived from case-inflected nouns with a locative meaning (underlined in [11 ]). They may occur with or without a ground-denoting expression: pulamil-luni fall.forward-coNTEMp:4sG

nakkar-toq fall.down-pART.MOOD:3sc,

"... as it falls forward and down.' (DV) Ammut 'downwards' is derived from a noun, but is no longer morphologically transparent. Nouns with a locative meaning (e.g., tunu- 'back') in a locative case do not by themselves denote the ground, but rather the path, as they can combine with a pronominal affix that denotes the ground. When tunu- 'back' combines with a pronominal affix, the whole denotes the backside of something (see iv below). The pronominal affix can then be further specified by a noun in the relative case which specifies the item at issue, i.e., the ground (see v below). The nine tokens of type / are all found in the adult stories. Two occur in combination with other types of path-denoting expressions. //'.

[12a]

Deictic and anaphoric adverbial demonstratives. Underlined in [12a] is an anaphoric adverbial demonstrative, and in [12b] a deictic demonstrative used as the root of a verb. Such demonstratives denote path only: ilanngul-luni Jensi aamma tamaanga nakkar-poq be.included-CONTEMP:4SG Jens:ABS also there:ALLAT fall.down-IND:3SG 'Jens [the dog] also fell down there.' (DV)

72


[ 12b] ataaseq one:ABS

uunngar-mat come.there-CAUS:3SG

'One goes there.' (A5)

In tamaanga 'to there' in [12a], ta- is a prefix making the form anaphoric (Fortescue 1984a:143). Uunngar- 'come there' in [12b] is deictic, i.e., it is used to direct the listener's attention to the book. The anaphoric adverbial demonstratives in the data, tamaanga 'to there', tappika 'up there, up above', tassunga 'to there', and takanunga 'down there', are used only by the adults. There are altogether eleven tokens, four of them in combination with other path-denoting expressions. The deictic adverbial demonstratives in the data are uunga 'over there'. ugguuna 'this way', uani 'here, there', and kanunga 'down there'. They are used only by the five-year-olds, altogether twelve times and once in combination with another path-denoting expression. Demonstrative pronouns denote ground elements, but they are rare in the data. As they all occur with another ground-denoting expression, they are not listed separately here. Two adults use the anaphoric demonstrative pronouns in the relative case taassuma 'that one, that one's' and taakkua 'those, them, theirs' altogether three times and in all cases in combination with another ground-denoting expression. One ten-year-old uses a deictic demonstrative pronoun in combination with another ground-denoting expression and one adult uses a deictic demonstrative pronoun as a determiner with a noun. None of the five- or ten-year-olds uses anaphoric adverbial or pronominal demonstratives, and by far the majority of the deictic expressions in the material (12 tokens out of 14) are produced by the five-year-olds.

///.

Case-inflected nouns. The noun refers to the ground and the case denotes the path, as in [10b]. This form is found in the stories from all age groups and is by far the most frequent way of referring to the ground. The two tokens of ground-denoting expressions in the stories from the five-yearolds are of this type, 24 out of 28 ground-denoting expressions in the stories from the ten-year-olds are also of this type, and the adults use this type 46 times out of 98 ground-denoting expressions.

iv.

Noun-derived directional adverbials with a possessive affix. An adverbial derived from a case-inflected noun with a locative meaning, as in Type /. but with a possessive pronominal affix (underlined in the example) that denotes the ground:


[13] tunga-a-nut

73

aarsaa-qattaar-ami

direction-3SC:POSS:SG-ALLAT

fly.down-again.and.again-CAUS:4SG

'It flies down in his direction again and again.' (EV)

This type is not found in the stories from the five-year-olds, the tenyear-olds have only one token, and the adults have ten tokens of this type. v.

Noun-derived directional adverbials with a possessive affix and a noun denoting the possessor. An adverbial derived from a case-inflected noun with a locative meaning as in Type /, but with a possessive pronominal affix that denotes the ground as in iv and this time with a noun in the relative (possessive) case which specifies the possessive affix and thus refers to the ground. An example is seen in [14], where EV extends her clause in [13] (the ground-denoting expressions are underlined):

[14] nukappiaqqa-p tunga-a-nut boy-RBL

direction-3so:poss:sG-ALLAT

aarsaa-qattaar-ami fly.down-again.and.again-CAUs:4so

"It flies down in the boy's direction again and again.' (EV)

This type is the second-most frequent form in the data. It is not used by the five-year-olds, however, and only once by the ten-year-olds. One ten-year-old uses a related form - a locative noun with a possessive affix as the stem of a verb accompanied by a demonstrative in the relative case (uuma qaav-a-niit-toq [DEM:REL surface-3so.POSS.SG-be.inpART.MOOD:3sG] 'as he was on top of that one there [= the deer]'). The adults have 38 tokens of type v. Besides the five main types, there is one example, in the story from a ten-yearold, of a ground-denoting noun as a verb root (ujara-ssuar-mukar-niar-uttor-luni [stone-big-go.to-want-just.as-coNTEMp:4so] 'just as he wanted to climb a big stone', with the root ujara- 'stone'). Moreover, the adults combine the different types in one nominal several times. 4.2 Bare verbs of motion Motion verbs in Spanish may occur with prepositional phrases that indicate the source or the goal of the path in descriptions of what Aske (1989) calls "mere locative paths". Nevertheless, Slobin (1996a) finds a considerable difference in the number of bare verbs in descriptions of downward motion in English and

74


Table 4. Percentage of downward-motion descriptions with bare verbs Language

3-5 years

5 years

9 years

1 0 years

Adults

English

16%

13%

15%

Spanish

56%

54%

36%

West-Greenlandic

97% (29/30)

71% (19/27)

51% (21/41)

The percentages from English and Spanish are from Slobin (1996a:200).

Spanish frog stories, in contrast to verbs with some locative addition such as a particle, a prepositional phrase, or an adverbial expression indicating downward direction, source, or goal of motion (e.g., jump down in English). Verbs with no such elaboration of the path denote directionality only inherently. The results from Slobin's analysis and from an equivalent analysis of descriptions of downward motion in the West-Greenlandic frog stories are shown in Table 4.5 The English narrators do not show any age-related development with respect to the number of bare verbs compared with verbs with some locative addition. From an early age, English-speaking children prefer verbs with some extra locative material to the same extent as adults. By contrast, narrators speaking the two verb-framed languages, Spanish and West-Greenlandic, show an age-related development and overall have a higher percentage of bare motion verbs than English narrators. But in addition to that, there is a remarkable difference between the Spanish and the West-Greenlandic stories. West-Greenlandic pre-schoolers use almost exclusively bare verbs, and even in the adult stories on an average every second motion verb is a bare verb. 4.3 Minus-ground and plus-ground clauses In the next step in his comparison of English and Spanish motion descriptions. Slobin (1996a) distinguishes minus-ground clauses and plus-ground clauses. English motion descriptions may be made by means of bare verbs only (e.g.. jump), verbs with a satellite without any ground-denoting expression (jump down), and verbs with a satellite combined with a ground-denoting expression (jump down onto the ground). Only the latter type includes a ground-denoting expression, and it is termed a plus-ground clause. Table 5 shows the percentages of plus-ground clauses from three age groups of English and Spanish narrators (from Slobin 1996a:201) and the results from the West-Greenlandic stories.6 Again we see a clear difference between English and Spanish narrators. The West-Greenlandic stories demonstrate that the percentage of plus-ground


75

Table 5. Percentage of plus-ground clauses English

Spanish

West-Greenlandic

5 years

60%

50%

2%

9 years

62%

61%

Age group

33%

1 0 years

82%

Adults

63%

40%

The results from English and Spanish are from Slobin (1996a:201). The absolute numbers for the West-Greenlandic data can be found in Note 8.

clauses increases with age and stays below the level of the English stories. The difference between verb-framing and satellite-framing cannot, however, explain the considerable difference between the stories in the two verb-framed languages, Spanish and West-Greenlandic. Again we see a much greater reluctance to elaborate on the path and the ground in relation to motion verbs in WestGreenlandic than in Spanish. In one respect, West-Greenlandic may be as elaborate as Spanish, however. In 216 Spanish stories, Slobin (1996a) finds only three examples of one clause with both a source-denoting expression and a goal-denoting expression in relation to a single verb. There is one such example in the 16 West-Greenlandic stories, namely [10c]. Another example is [15], with two ground-denoting expressions (underlined) with a single verb, both denoting goals. [ 15] aalateri-ffige-riar-lugit wave.goodbye-TRANS-first-

tako-qqis-sa-nera-riar-lugit see-again-FUT-say.that-firstCONTEMP:3PL

orpip-passui(t) tunga-an-nut tree-many :REL direction-3PL:POSS: SG-ALLAT

ingerla-qqip-put illu-mi-nnut go-again-INDIC:3PL house-3R:PL:POSS: SG-ALLAT

'First they waved goodbye to them and then said "See you again" to them and went back towards the forest to their house.' (DV)

The West-Greenlandic stories show a remarkable increase in the variation and complexity of the path-denoting and ground-denoting expressions between the age groups. One five-year-old uses case-inflected non-locative nouns (Type //'/), but otherwise the five-year-olds use only deictic adverbial demonstratives (Type ii). As mentioned, the five-year-olds use twelve out of the fourteen deictic expressions in the data. The ten-year-olds use more case-inflected non-

76


locative nouns (Type //'/) and extend their inventory with two locative nouns with possessive affixes (Type /v), one of them also with a specifying noun in the relative case (Type v), but they do not use any anaphoric adverbial demonstratives (Type //) and only one anaphoric pronominal demonstrative. Only the adults use the anaphoric forms of the adverbial demonstratives, and they use more complex locative forms than the children; they combine several types and they extend the individual types. 4.4 Complex paths Slobin (1996a, 1997a) further extends the investigation of the English and Spanish motion descriptions beyond the clause limit, examining whether Spanish narrators elaborate descriptions of complex paths over more than one clause. Based on the descriptions in the frog stories, he analyses the scene of the fall from the cliff into six narrative segments (after Slobin 1996a:203; see Pictures 16-18 in Appendix I):7 1. deer starts to run; 2. deer runs, carrying boy; 3. deer stops at cliff; 4. deer throws boy (off antlers/down); 5. boy and dog fall; 6. boy and dog land in water. As the pre-schoolers provided few details of the journey, Slobin considers only stories from the nine-year-olds and the adults and finds that [o]nly 40 per cent of Spanish 9-year-olds provided three segments, and none of them provided more than three. By contrast, 92 per cent of American 9year-olds provided three or more segments, and, of these, almost half provided more than three segments. Of the adult narrators. 100 per cent of the Americans and only 75 per cent of the Spaniards provided three or more segments of the journey. (Slobin 1996a:203)

He concludes that the Spanish narrators do not compensate for their minimal use of clauses with ground-denoting expressions by means of a series of separate action clauses. However, Slobin (1996a) does not make it clear to what extent all aspects of a segment must be included for the segment to be counted as present in a story (e.g., both that the deer runs and that it carries the boy for Segment 2). None of the West-Greenlandic ten-year-olds mentions the cliff (Segment 3: "deer stops at cliff'), but two mention that the deer - or rather reindeer - suddenly stops, w ithout


77

specifying where. Three mention the water in relation to Segment 4 ("deer throws boy [off antlers/down]") and not in relation to Segment 6 ("boy and dog land in water"), and one in relation to Segment 6. Only one mentions that the boy is on the deer's antlers when it is running, while one says the equivalent of 'then they go' with a plural subject affix, and another says 'the reindeer is running, is running'. One child and several adults mention that the dog follows, a segment that Slobin does not include, probably because it cannot be said to specify part of the path, but rather one more actant. Two adults mention both that the deer arrives at a cliff and that it suddenly stops at the border of the cliff, i.e., two segments of the journey where only one is mentioned in Slobin's list (see also Wilkins 1997). If we focus on the part of Slobin's characterization of the parts of the journey that specifies the action or event related to the path (i.e., rather that the deer runs than that it carries the boy), it is possible to analyse the West-Greenlandic descriptions of the journey into the six segments. Three of the ten-year-olds, i.e., 50%, provide three segments, while three provide only two. None mentions Segment 1 ("deer starts to run"), while five mention Segment 5 ("boy and dog fall"), or only that the boy or the dog falls. The West-Greenlandic ten-yearolds thus resemble the Spanish nine-year-olds in that about half mention three segments and none mentions more than three segments. By contrast, the adults elaborate the path, or rather the actions or events taking place along the path. One adult mentions all six segments, two mention five (and one of them mentions the extra segment "deer arrives at cliff), one mentions four plus the extra segment, and one - the youngest one, aged 17 - mentions only one segment. In other words, 80% mention three or more segments. Here, the adult West-Greenlandic narrators clearly deviate from the pattern with the West-Greenlandic stories differing from the English stories in the same way as, but to a larger extent than, the Spanish stories. West-Greenlandic adult narrators do indeed elaborate the different stages of a complex series of actions and events over an extended path, although they do not elaborate the path as such. 4.5 Static setting descriptions Although Spanish narrators use a smaller set of motion verbs, mention fewer ground elements in individual clauses, and describe fewer segments of a journey than the English narrators, Slobin finds that "their narratives, overall, seem to 'tell the same story' as English accounts" (1996a:204). The reason seems to be that the Spanish stories focus more on setting than on motion: English, with its rich means for path description, can often leave setting to be inferred; Spanish, with its sparser path possibilities, often elaborates descriptions of settings, leaving paths to be inferred. (Slobin 1996a:204).

78


One Spanish-speaking seven-year-old says, for instance, Los tiro a un precipicio donde habia harta agua. Entonces se cayeron. '[The deer] threw them at a cliff where there was lots of water. Then they fell.' (Slobin 1996a:204, Example 7a). The listener can infer that the boy and the dog fell into the water which is presented in a static description. Looking at how the setting is expressed in the West-Greenlandic stories in the scene of the fall from the cliff, we focus on how the narrators introduce the cliff and the water. Only one out of five West-Greenlandic five-year-olds mentions the water and does indeed introduce it in a static clause: [16] uani there

taava

aakanna

imaq

then

that.down.there

sea:.-\BS

'There then down there is the sea.' (D5)

None of the five-year-olds mentions the cliff. Five out of six ten-year-olds mention the water, and all introduce it by means of a noun in the allative case with an active verb, as in [17]: [17] taava then

nakkar-tip-paa

imer-mut

fall.down-make-iNDic:3sG:3sG

sea-ALLAT

'Then it makes him fall into the sea.' ( B I O )

None of the ten-year-olds mentions the cliff. Of the five adults, three mention the cliff, all five mention the water, and one talks about 'the border of a river' (kuup sinaanut) rather than a cliff. In seven out of these nine cases, the setting is introduced by means of a noun in the allative case or as the object in the absolutive of a dynamic verb meaning 'be on the way', 'arrive at/bring to', 'throw', 'hurl', or 'make fall'. There are only two examples of static setting description in the adult stories: [18] taatsuma-li this:REL-and

ata-a-niip-poq

taserar-savaq

down-4sG;poss:sG-be.at-iND:3sG

puddle-big:ABS

'Below it [the cliff] there is a lake.' (BV) [19] una DEIC:DEM:ABS

kuu-ju-ner-poq river-be-possibly-iNDic:3sG

imaaluunniit or

tasera-a-ner-poq orpippassu-it eqqa-a-niit-toq puddle-be-possibly-iNDic:3sG forest-PLiREL next.to-3pi.:poss:sG-be.onPART.M(X)D:3SG

79


Table 6a. Number of narrators providing static setting descriptions in the scene of the fall from the cliff 5 years

9 years

English

1(12)

1(12)

0(12)

Spanish

1(12)

5(12)

3(12)

Language

West-Greenlandic

1 0 years

0(6)

1(5)

Adults

2(5)

The numbers for English and Spanish are converted from the percentages given in Slobin (1996a:205). The numbers in parentheses indicate the total number of narrators. Table 6b. Number of narrators introducing the cliff or the lake with either a dynamic or a static verb in the West-Greenlandic stories First mention of Age group

Cliff dynamic

Lake static

dynamic

5 years

1

10 years Adults

static

5 4

3

2

'Next to the forest is what may be a river or a lake.' (CV)

Table 6a summarizes the analysis of the number of narrators who provide a static setting description in relation to the scene of the fall from the cliff in the three languages, and Table 6b shows the number of static versus dynamic introductions of the cliff and the lake in the West-Greenlandic stories. It is possible that the numbers for the West-Greenlandic stories in Table 6 would look more like the numbers for the Spanish stories if there were as many West-Greenlandic as Spanish narrators, but it is also important to note that out of fifteen introductions of the cliff and the lake in the West-Greenlandic stories from all age groups, only three are static descriptions. In all other cases, the setting is introduced with a dynamic verb. Moreover, there is very little setting description in other parts of the story. None of the children mentions that the beehive is in a tree even when they include the episode of its fall. All the adults mention the beehive, but only one (BV) says that it is hanging from a tree; two (BV, EV) mention the tree when they describe how the dog is shaking it (i.e., they mention the tree as the object of a transitive verb), but none mentions the tree in the description of the fall of the beehive.

80


Summing up, the frog stories in West-Greenlandic seem at first sight to live up to the expectations raised by Slobin's comparison of stories in English and Spanish. Like the stories in Spanish, the other verb-framed language, the West-Greenlandic stories have more bare verbs for downward motion, i.e.. verbs without any locative satellite, than the English stories, and they have fewer ground-denoting expressions per motion verb than the English stories. Hovvev er. West-Greenlandic is much farther from English than Spanish is, the differences between Spanish and West-Greenlandic being about as great as the differences between Spanish and English. Moreover, the West-Greenlandic adult stories are closer to the English type with respect to the number of segments used in the description of the complex path - or rather complex series of actions - in the scene of the fall from the cliff. What is more, both the ten-year-olds and the adults introduce the few ground elements that they do mention by means of adjuncts with motion verbs, like the English narrators, rather than introducing them in static setting descriptions, as do some Spanish narrators, but the absolute numbers here are small. The overall conclusion is that the typological difference between verb-framing and satellite-framing languages is not sufficient to explain the results of the comparisons.

5. DISCUSSION The results of the comparison between English, Spanish, and West-Greenlandic descriptions of motion events raise doubt about the typological distinction between satellite-framing and verb-framing languages as an exhaustive explanatory factor. Even though Spanish and West-Greenlandic are both verbframed, they differ radically in their morphological type. West-Greenlandic is a polysynthetic language with extensive use of derivational and inflectional morphology. The verbal derivational affixes code many semantic distinctions, e.g., the "embedding" functions of judging, saying, wishing, causing, and intending; modifications of degree, manner, completion, frequency, and duration; tense; epistemic modality; negation; subjective coloration; "conjunctional" meaning of, for instance, sequencing ('after'), simultaneity ('while still'), and condition ('if'), and many more meanings (Fortescue 1980). To give the reader an impression of the distribution of information in a polysynthetic language among verbs in superordinate (VSUP) and subordinate (VSUB) moods, nominals (NOML). and adverbials (ADV), we will present BV's description of the scene of the fall from the cliff in a tagged close "translation": 1.

The reindeer takes off

NOML VSLP


in the forward direction, him being on its head, as it is bothered by Peter and wants to try and get rid of him.

ADV

And his dog follows it slowly, trying to defend, without success, Peter.

NOML

Then they arrive at a cliff.

ADV

NOML

4.

And this one's down below is a big puddle.

NOML VSUP NOML

5.

As the reindeer suddenly stops suddenly at the border of the cliff, it causes to fall Peter into the big puddle.

NOML ADV VSUB NOML NOML VSUP NOML NOML

6.

And as the dog comes along, running along, not noticing undeliberately him, it also falls too itself via the cliff.

NOML VSUB VSUB VSUB ADV VSUP NOML NOML

Yet the cliff is not so very high, and the lake is filled with many bushes.

nominalized participial VSUP VSUB NOML VSUB

VSUP VSUB NOML

VSUP

NOML VSUP NOML VSUP NOML

81

82


8.

Thus as Peter and his dog hit, they are not hurt at all.

ADV NOML NOML VSLB VSLP ADV

Altogether there are 17 verbs + 1 nominalized verb, 20 nominals, and 6 adverbials in the extract, this list exhausting the number of words (44). Moreover, including the participial nominalized verb, 9 out of the 18 verbs are in one of the subordinate moods. In the ten-year-olds' descriptions of the complex journey we find approximately the same ratio of verbs to nominals as in the adult stories, but only one case of two verbs in subordinate moods with one verb in a superordinate mood, and no sentences as complex as in BV's description. The five-year-olds, on an average, use more than two verbs per nominal, including locative adverbials. in the description of the complex journey. In a recent study of oral and written frog stories from different age groups, Trondhjem (2000) found that even children as old as ten combine clauses with two subordinate moods without any superordinate mood, thereby demonstrating the problems children encounter in acquiring the obligatory parts of the verb. In the five-year-olds' stories, there are many cases of semi-repetitions of verbs with the same root, but with different obligatory mood and number marking or with different derivational affixes, again showing how the children struggle with the obligatory and optional parts of the verb: [20] una

imer-nia-rami

nakkar-tit-tuus-sa-vai

DEM:3SG:ABS drink-intended-CAUS:4SG fall-make-happen.to.one-FUT-INDIC:3SG:3PL

taava nakkar-tik-kamiuk

tako-riar-paa

then

see-surprisingly-INDIC:3SG:3SG

fall.down-make-4sc:3sG

nakkar-toq fall.downPART.MOOD:3SG

'As that one intends to drink, it happens that it makes them fall. Then it makes him fall. It sees him falling.' (E5)

[21 ] nakkar-ti-ta-a-

nakkar-tik-kiartor-ti-ta-a-

fall.down-make-

fall, down-make-go, to-make-

PASSIPART.MOOD-be-

PASS:PART.MOOD-be-

uja-raa

taava

uani

nakkar-tik-kiartor-paa

look.for-PART.MOOD: 3G:3SG

then

here

fall.down-make-go.to-INDIC:3SG:3SG


kanunga down.there

nakka-rami fall.down-INDIC:3SG

83

qimmer-lu dog.3SG:ABS-and

'(He) is thrown down - (he) is taken over to be thrown down. He looks for him. Then he goes to make him fall here, he falls there, the dog too.'8 (D5) [22] taava then

taava then

nakkar-til-lugu fall.down-make-coNTEMp:(4SG):3SG

nakkar-ti-le-ramigit fall.down-make-about.toCAUS:4SG:3PL

nakkar-tik-kamigit sikorner-luni fall.down-make-CAUs: close.eyes4SG:3PL CONTEMP:4SG

taava then

nakkar-tor-lu fall.downPART.MOOD:3SG-and

'Then it makes him fall. When it is about to make them fall, then it makes them fall as it closes its eyes, and then he falls.' (A5) The semi-repetitions also demonstrate how the children struggle to acquire the adult narrative style with repetitions of the same verb root in different moods. A recurrent function of such semi-repetitions in the adult frog stories is to contribute to cohesion: the first form in the indicative mood is used to denote an event, and the second form in a subordinate mood leads on to a new verb in the indicative mood denoting a new event:9 [23] Piita-p

qimmi-a dog-3SG:sc

ingitsa-at kamaal-lungu bee-PL:REL be.angry.with-coNTEMp:(4sG):3sG

malersi-ler-sima-vaat,

malersi-ler-uttor-aal-li

pursue-begin-pERF-iNDic:3pL:3sc

pursue-begin-just.as-pART.MOOD:3pt:3sG-and

uppi-ip owl-REi,

Piita PetenABs

putu-a-ninngaanii(k) hole-3sG:poss:sG-ABL

nui-ffmga-a come.forward-TRANS-iNDic:3sG:3sG

orpi-ip tree-REL

qaarusu-usa-minngannii(k) cave-similar.to-ABL

'As the bees are angry with Peter's dog, they have started pursuing it, and while they are pursuing it, an owl comes forward towards Peter from the hole in the tree similar to a cave.' (BV) The second verb form (malersi-ler-uttor-aal-li [pursue-begin-just.as-PART.MOOD: 3pL:3so-and]; note the affix -uttor [just.as]) is used to relate the different events to each other in time, a function that in other languages might be fulfilled by

84


a temporal conjunction or adverb. As the children generally have problems acquiring the functions of the different verb moods (Trondhjem 2000), their repetitions and semi-repetitions, as evidenced by Examples [20-22], may be an attempt at acquiring the adult way of indicating narrative cohesion. What the adult stories in West-Greenlandic reveal is a style using an overwhelming number of verbs with much affixal derivation as well as highly complex sentence patterns with many verbs in subordinate moods. The children seem to focus in particular on the acquisition of verbal forms, the derivational and inflectional affixes of verbs, and the relative distribution of the different moods. The morphological form of the verb is far from transparent. Fortescue and Lennert Olsen talk about "the total lack of either phonemic or suprasegmental clues as to the internal structure of the 'word'" (1992:215-216). Children can thus be expected to concentrate on acquiring the partially obligatory and semantically very rich verb morphology. This concentration on verbs immediately explains the large number of bare verbs and small number of ground-denoting expressions, even compared with the stories in the other verb-framed language, Spanish. The verb-rich style may also contribute to explaining why the West-Greenlandic children have a low score on the second core component of the plotline compared with the children in Herman and Slobin's study (1994). To fulfil the second core component of the plotline strictly, the narrator must mention the frog in the section about the boy's adventures in the forest. But in accordance with their emphasis on verbs, many West-Greenlandic children leave out the nominal and use only transitive forms of a verb meaning 'search for it' or 'shout for it'. Two more results call for explanations: the low score on the third core component by the strict criterion and the unexpectedly detailed descriptions of the complex journey in the adult stories. In itself, the focus on verbs cannot explain why the children do not use one of the optional means to denote definiteness in order to link the frogs at the end with the frog in the beginning of the story. We can only speculate that the lack of obligatory definiteness marking in the nominal and the focus on verbs make this feature of plot narration less urgent. As regards the detailed description of the complex path in the scene of the fall from the cliff, Slobin's six components include both the ground as sections of the path and actions that take place in relation to these different sections. The only two ground elements mentioned in the West-Greenlandic stories in descriptions of the journey are the cliff and the lake. The focus in the descriptions is thus rather on what happens in the different sections of the journey than on where it happens - again a consequence of the focus on verbs. An alternative explanation of especially the children's small number of ground-denoting expressions and low score on components II and III of the plotline might be that the children are not trained to tell a story that can be understood independently of the context. After all, the experimenter can see


85

the pictures, too. However, while the five-year-olds do indeed use deictic expressions, the ten-year-olds hardly use any. They have learnt to tell a story independently of the immediate context, and the adult stories are almost totally context-independent. In the conclusion to their acquisitional study of West-Greenlandic, Fortescue and Lennert Olsen (1992; see also Fortescue 1984b) emphasize the morphology as a complicating factor in the acquisition of the language, but they also point out that the children may be helped by "the clear-cut and pervasive distinction between referring expressions (with distinctive nominal endings) and action/ state expressions (with distinctive verbal endings), irrespective of the internal complexity of the expressions themselves" (Fortescue & Lennert Olsen 1992: 216). They further hypothesize that in a hunting-gathering society in rather specific physical environments such as the Greenlandic society was and to some extent still is, there is less need for referential specificity, less "emphasis on 'things' rather than on actions and states (for the description of which polysynthetic 'holophrasis' can be extremely efficient)" (Fortescue & Lennert Olsen 1992:215). This explanation for the small number of nominals combines the language type with the language users' communicative needs, given the type of society and the physical context. Wilkins (1997) goes even further in giving priority to cultural influences on narrative style. In a study of descriptions of the frog's escape from the jar and the scene of the fall from the cliff in the Central Australian language Arrernte, he finds that Arrernte-speakers on an average code more segments of the paths than English-speakers, even though Arrernte is a verb-framed language. He ascribes this stylistic feature to the culture-specific concern for motion and orientation among nomads in Central Australia, mentioning, however, also the fact that Arrernte verbs have a distinct morphological slot for "an elaborate category of inflections called the category of associated motion, which is used to indicate that the verb stem action happens against the background of a motion event with a specific orientation in space" (1997:298-299; author's emphasis). Australian Aboriginal communities are known to put much weight on locations and directions, but Eskimo languages have elaborate systems of locative terms (Denny 1982; Fortescue 1984a, 1988), which might raise expectations of elaborate path and ground descriptions in these languages too. Furthermore, locative information can be included in the verb stem, since demonstratives can occur as stems of verbs. In contrast to the Arrernte verb, however, the WestGreenlandic verb does not have what Wilkins describes as "a distinct slot for an elaborate category of inflections" (1997:298) or any equivalent verbal inflection expressing locative information. It is doubtful whether we shall ever be able to sort out the different influences of language and culture. But besides culture, a strong factor in determining rhetorical style may be what is coded in the verb in polysynthetic languages.

86


In these languages, the verb is the only obligatory part of the clause, which means that children acquiring the language first have to come to terms with verb morphology and the semantic distinctions coded more or less obligatorily in the verb. In his description of thinking-for-speaking, Slobin talks about what is "readily encodable" (1996b:76) in a language. It cannot be claimed that the ground is difficult to code in West-Greenlandic. As shown, the ground-denoting and path-denoting expressions may become very complex, but the ten-year-olds choose the fairly simple expressions also offered by their language, i.e., nouns inflected for one of the four locative cases. This solution also accounts for almost 50% of the adult ground-denoting expressions. Still, both children and adults use many bare verbs and few ground-denoting expressions. The rhetorical style of the West-Greenlandic frog stories seems to be influenced not so much by what is "readily encodable" as by what is coded semantically richly in the verb - the only obligatory part of the clause.

APPENDIX West-Greenlandic is a polysynthetic language with very complex nominal and verbal morphology. The morphology includes affixes that may nominalize or verbalize roots from other word classes. Verbs can take derivational affixes from a rich inventory with many more or less lexicalized combinations (Fortescue 1980) and they must take inflectional suffixes of mood and person and number agreement for the subject and the object, including a switch reference system, i.e.. the subject affix on the verb indicates whether the current subject is co-referential or not with the subject of the superordinate verb. Word derivation and inflection are subject to complex morpho-phonological rules which make the morphological structure of the word quite opaque. There are two kinds of mood: superordinate moods comprising indicative, imperative, optative, and interrogative, and subordinate moods comprising contemporative, causative, conditional, and participial (Fortescue 1984a:287ff). West-Greenlandic has ergative morphology. Ergative case is traditionally called relative, and this case is also used to mark the possessor. (See also Fortescue 1980, 1984a; Lennert Olsen & Hertling 1988.) In the glosses, a hyphen is used to separate morphemes and a period indicates that more than one English word is used for one West-Greenlandic morpheme or that the West-Greenlandic unit contains more than one morpheme. The abbreviations used to gloss obligatory morphemes, as well as a few derivational affixes that occur in small sets of contrastive elements (Fortescue, personal communication), are all according to the notational conventions presented in Appendix III of this volume, with the exception of the following four categories:


87

CONTEMP

contemporative mood. Subjects of verbs in the contemporative mood are always co-referential with the subject of their superordinate verb, and the inflectional affix of transitive verbs in the contemporative mood indicates only the object (Fortescue 1984a:297ff). To distinguish transitive and intransitive verbs in the contemporative mood, we indicate the subject of transitive verbs within parentheses, e.g., "coNTEMp:(3RSG):3sG". PROS prosecutive (a case meaning 'through' or 'via') 4sG the subject-agreement marker of subordinate verbs must indicate whether the subject is co-referential with the subject of the superordinate verb (Fortescue 1984a:146ff). "4so" indicates coreferentiality. PART.MOOD participle mood The English translations of the examples emphasize relevant West-Greenlandic structure rather than idiomaticalness.

NOTES 1.

We would like to thank Michael Fortescue and Birgitte Jacobsen for comments on an earlier version of this chapter. The earlier version, Fra barn til voksen: Udviklingen i genfortcelling af en billedbog pa gronlandsk ['From child to adult: The development in retelling a picture book in Greenlandic'], was published in Gronlandsk Kultur- og Samfundsforskning 98/99, Nuuk: Ilisimatusarfik/Atuagkat 2000.

2.

The children's stories were collected and transcribed in March 1998 by Maren Louise Brandt, Annga Lynge, and Kistara V. Motzfeldt, students at the Department of Greenlandic Language and Culture, Ilisimatusarfik, University of Greenland, as part of a course on language acquisition held by Elisabeth Engberg-Pedersen. The adult stories were collected and transcribed by Frederikke Blytmann Trondhjem later in 1998 in Copenhagen among Greenlanders who had all been in Denmark for no more than a year. Frederikke Blytmann Trondhjem glossed all the stories.

3.

The form sanioqqutippoq is not possible, but may be an attempt at saying sanioqquppoq (pAss-iND:3so) with an overgeneralization of -t from the stem and insertion of an otherwise latent vowel.

4.

The -j- in sujaartorami is a speech error.

5.

Included among the West-Greenlandic verbs are verbs meaning 'hit' (especially in the episode about the boy and the dog falling into the lake), 'jump' (about the dog jumping out of the window), and 'turn over' (about the boy falling over underneath the tree as it can be seen in Picture 12), all used in the fall episodes.

6.

The West-Greenlandic-speaking adults use altogether 246 motion verbs, they have 88 ground-denoting expressions with a noun (i.e., Types in and v plus 4 combined


expressions) and 10 such expressions where the ground is expressed only pronominally in an affix (Type ;'v). The ten-year-olds use 86 motion verbs and altogether 26 grounddenoting expressions with a noun and 2 with only a pronominal affixal expression of the ground. The five-year-olds use 82 motion verbs and altogether 2 ground-denoting expressions, both with a noun of Type Hi. We have here included only those grounddenoting expressions that occur with motion verbs. To a limited extent, the ten-yearolds and the adults also use nouns in one of the locative cases with verbs meaning 'see' or 'shout' (e.g., 'he shouts from the window'), but neither the verbs nor the nominals have been included here. 7.

Slobin (1997a:448) lists "four potential event components", while Wilkins (1997) analyses the "cliff journey" into 14 distinct path segments based on 36 frog stories in Arrernte (Central Australia). As Slobin (1997a) does not include data from children, we use Slobin (1996a) here.

8.

The first two verb forms do not include the obligatory marker of mood and agreement.

9.

In the West-Greenlandic frog stories, repetition and semi-repetition of verb forms are also used to indicate duration or iteration, as in Examples [2] from a five-year-old and [9] from an adult, and to indicate a wish or possibility followed by a statement about the realization of the wish or possibility.

Motion Events in Basque Narratives1

Iraide Ibarretxe-Antunano

\. INTRODUCTION Motion events are situations "containing movement or maintenance of a stationary location" which are analysable into a set of semantic components (Talmy 1985: 85). Depending on the different lexicalization patterns that languages use in order to package these components into linguistic forms, Talmy has typologically divided languages into two main broad groups: satellite-framed and verb-framed languages (Talmy 1985, 1991, 2000a, 2000b). The former usually provide speakers with a set of locative particles, so-called satellites, which encode the core schema, i.e., the Path (change of location); the latter supply speakers with a set of different verbs for each change of location. Slobin (1987, 1996a, 1996b, 1997a, 2000) has taken this typology one step further and has proposed a modified form of linguistic relativity in the thinking-for-speaking hypothesis - the fact that specific language effects can be demonstrated in the on-line use of language. Following this line of research, this chapter gives a detailed account of the linguistic and gestural expression of motion events in Basque. I argue that Basque is a verb-framed language, and as such it shows a tendency to encode Path in the verb and to express Manner in a satellite. However, there are also specific characteristics particular to Basque and, therefore, crucial to thinking-forspeaking in this language. These are discussed in the second part of this chapter. 1 focus on four areas: a) the description of Path, and the complete-path hypothesis; b) the description of physical setting; c) the use of onomatopoeic words and sound-symbolic expressions; and d) the utilization of different iconic gestures for encoding both Manner and Path. The data come from fifteen video-taped narratives by adult native speakers of Basque from different dialectal areas. Following Berman and Slobin's (1994) procedure, the data were elicited from a wordless picture-book (Mayer 1969; see Appendix I).

90


2. THINKING-FOR-SPEAKING ABOUT MOTION EVENTS As already mentioned, motion events are situations containing movement or maintenance of a stationary location, which are analysable into a set of basic semantic elements (Talmy 1985:85). These elements are of two types: internal components and external co-event components. Among the former, Talmy (1985: 61) distinguishes four: Figure, the moving object; Ground, the entity or entities that the Figure is moving in relation to; Path, the course followed by (and the trajectory of) the Figure; and Motion, the actual predication of a motion act. Among the latter, Talmy includes Manner, the way in which motion is performed: and Cause, what originates the motion itself. The concept of motion seems to be present in all the languages of the world. However, neither the way in which motion is expressed in these languages - the lexicalization patterns - nor the aspects of motion (motion components) that each of these languages focuses on - thinking-for-speaking - are exactly the same. A great deal often Talmy's (1985, 1991, 2000a, 2000b) work in linguistics is devoted precisely to the study of the variation that languages show in mapping lexical resources onto semantic domains. According to Talmy, the systematic relations between semantic elements (meaning) and surface elements (linguistic forms) do not show a one-to-one correspondence across language types. In fact, this relationship may take different forms, with several semantic elements being expressed by one surface element, or a single semantic element being expressed by several surface elements. Following this line of research, Talmy (1991) proposes that the world's languages can be divided into a two-category typology on the basis of how the core schema of a specific semantic domain is mapped onto syntactic and lexical structures. Verb-framed languages are those where the core schema is mapped onto the verb, and satellite-framed languages are those where the core schema is mapped onto some other element. In the case of motion events, the core schema is Motion. As we can see from Talmy's (1985:487) classic example of the description of a bottle floating out of a cave in Spanish [1] and English [2], languages contrast with each other in their preferences when it comes to encoding the semantic components of Motion. [1]

La botella salio flotando. The bottle exited floating/

[2]

The bottle floated out.

The first difference between these two sentences relates to the encoding of the Path of motion. In Spanish, the verb, salir 'exit', expresses the core information

Motion Events in Basque Narratives

about the Path of movement, whereas in English, it is a particle or - in Talmy's terms - a satellite to the verb, out, that encodes this information. The second difference relates to the expression of the Manner of movement. Since English does not encode Path in the main verb, this slot is available for Manner verbs such as float. In Spanish, on the other hand, this slot is already occupied by Path verbs, and Manner of motion tends to be conveyed in a separate expression such as the gerundflotando 'floating'. Languages such as Spanish are referred to as verb-framed, while languages such as English are satellite-framed. Slobin (1987, 1991, 1996a, 1996b, 1997a, 2000) has taken this typology one step further and has proposed a modified version of the classic Sapir-Whorf arguments on linguistic relativity which were much debated in the first half of the twentieth century (Sapir 1924; Whorf 1940): the thinking-for-speaking hypothesis. Slobin's hypothesis differs from Sapir's and Whorf's view in that its main goal is not to prove the effects of grammar on world view or non-linguistic behaviour, but to show how speakers of a language organize their thinking in accordance with the linguistic tools offered by that language. In other words, Slobin claims that specific language effects can be demonstrated in the on-line use of language. The main idea of the thinking-for-speaking hypothesis is that an experience cannot be verbalized unless the verbalizer takes a specific perspective, which is influenced, if not determined, by the typological characteristics and lexicalization patterns of a given language. For Slobin (1997a), a motion event can be described as situated in two different cognitive frames: a discourse frame - which refers to a journey that a moving figure undertakes - and a typological frame - which refers to the tools provided to and constraints imposed on speakers in expressing Motion in a particular language. The discourse frame is common to all languages, but the typological frame is particular to each language. That is to say, the typological frame imposes a subjective tendency upon the schematization of experience because it "provides and limits the means of expressing components of a motion event" (1996a:195). Each language has a different set of grammatical and lexical choices, and as Slobin and his collaborators have shown, these choices seem to lead speakers of different languages into dealing with the same experience in different ways. Let us draw upon some examples to illustrate this point. As I have said before, the main contrast between satellite-framed and verb-framed languages relates to the encoding and conflation of Manner and Path of motion in the verb root. This typological difference has several consequences for the schematization of motion events in these languages:

92


/'.

The lexical items for Manner description available in the lexicon of satellite-framed languages are not only much more numerous, but also more expressive2, than those of verb-framed languages (Slobin 2000).

/'/'.

Manner of motion is mentioned more frequently in satellite-framed languages than in verb-framed languages (Slobin 2000).

iii.

The rhetorical style of these languages also varies a great deal. Whereas satellite-framed languages devote more narrative attention to the description of Manner, verb-framed languages focus more on the scene setting (Slobin 1991, 1996a).

iv.

The number of Ground elements attached to the verb, as well as the number of Path elements per trajectory, is larger in satellite-framed languages than in verb-framed languages (Slobin 1996a, 1997a).

Each of these consequences is observed not only in on-line use of language, but also in different modes of expression such as writing, reading, talking, translating, and so on (Slobin 2000). In this chapter, I give a detailed account of the typological frame in the description of motion events in Basque, a language isolate spoken on both sides of the western end of the Pyrenees. In the first half of the following section, I present some data that lead us to classify Basque in the verb-framed type of languages. In the second half, I focus on those specific characteristics that are particular to Basque and, therefore, crucial to thinking-for-speaking in this language.

3. MOTION EVENTS IN BASQUE 3.1 Data-elicitation procedure The data that I use in this chapter come from fifteen elicited narratives by adult native speakers of Basque from different dialectal areas3. The procedure I followed in gathering these data is that of Berman and Slobin (1994). I used a wordless picture-book, Frog, where are you? (Mayer 1969; see Appendix I), to elicit these narratives. In 24 different pictures, this little book tells the story of a boy who, accompanied by his dog, goes in search of his pet frog. The story starts when the frog escapes from its jar while both the boy and the dog are sound asleep. The following morning, when they discover that the frog is gone, they start searching for it through the woods. This involves the description of different types of motion events such as falling, running, and climbing: The


93

dog falls from a window. The boy climbs and falls from a tree. The dog runs away from a swarm of bees. The boy climbs a rock and gets entangled in the antlers of a deer. The deer throws both the boy and the dog over a cliff into a pond. Finally, both of them get out of the water and climb over a log to find their missing frog. The method is very simple. The informant is told that he has to tell a story about a boy, a dog, and a frog. He is shown the book and allowed to look through the pictures. Then he is asked to tell the story to the researcher in front of a video camera.4 The informant follows the book picture by picture. 3.2 Basque as a verb-framed language The main distinction between verb-framed and satellite-framed languages relates to the encoding and conflation of Manner and Path in the verb root. According to our data, the "characteristic expression of motion"5 in Basque can be considered to be that typical of a verb-framed language. That is, Path and Motion are conflated in the verb and Manner of motion is usually conveyed in a separate adverbial expression. Let us illustrate these characteristics with some examples. [3]

Eta

ontzitikan

atera

intzan

and

jar:ABL:Loc

exit:ppv

do.Aux

'And [the frog] went out of the jar.' (B20b)

In [3], the verb atera 'exit' conflates the self-motion component and the directionality or Path of that motion: 'out'. In the fifteen frog stories that form our corpus, there are a total of 63 different types of Motion verbs. As expected in this type of language, a large number of these verbs - 29 types (49%) - are verbs of conflated Motion and Path. [4]

Eta

erlauntzatik

erle

guztiak

and

beehive:ABL

bee

all:ABs:ARi:pL exit:ppv

irten

ziren

hegaka

AUX

flying

'And all the bees flew out from the beehive.' (B20i)

In [4], the verb irten 'exit', another verb with Path and Motion conflated, is complemented with the adverb hegaka 'flying', an adverb which expresses the Manner of motion in this particular motion event. This type of adverbial expressions for the description of motor patterns of the Figure are usually constructed with the iterative suffix -ka in Basque. Other examples from our corpus are: arinka 'running', biriboraka 'rolling', ihesika 'escaping', korrika 'running', and saltoka 'jumping'.

94


There are also several other adverbial expressions that denote the "rate" or speed of motion such as bapatean 'suddenly', berehala 'fast', and ziztu bizian 'fast', as illustrated in [5]: [5]

Eta and

ziztu flash

bizian joan quicknoc §O:PFV

ziren AUX

Zikinen ZikiniGEN

atzetik behindiABL

'And [the bees] rushed after Zikin [the dog].' (B201) Finally, there is also a great deal of adverbial complementation denoting what Slobin (1997a:460) has called the "protagonist's inner state" (of mind), as in [6]: [6]

Oreina deenABS

ere also

ikaratuta scared

korrika running

hasi startiPFv

zen AUX

'The deer started to run, very scared.' (B20i) Other such expressions are aztoratuta 'bewildered', beldurtuta 'scared', haserre 'angry', ikaratuta 'scared', isil-isilik 'quietly, silently', izututa 'scared', keztatuta 'worried', larrituta 'worried', lasai-lasai 'calmly',poz-pozik 'happily', andzainzain 'watchful'. The number of Manner verbs is also very small. There are only fifteen different types (23%), of which six convey Manner as well as Path, e.g.. ihes egin 'escape'). Slobin (1997a:459) distinguishes a two-tiered lexicon of manner verbs: the neutral, everyday verbs, such as walk and fly and climb: and the more expressive or exceptional verbs, such as dash and swoop and scramble. As expected in a verb-framed language, the second tier is not very elaborate in Basque; most of the verb tokens found belong to the first tier: airatu 'fly', astindu 'shake', ibili 'walk', irriste egin 'slide', korrika egin 'run', makurtu 'bend'. paseatu 'stroll', saltatu 'jump', salto egin 'jump'.

4. PECULIARITIES OF BASQUE In the previous section, I presented some data confirming that Basque can be classified among the verb-framed languages. However, Basque is not a prototypical example of this category. In what follows, I argue that Basque has several peculiarities which, in some cases, may place it closer to the satelliteframed languages than to a prototypical verb-framed language such as Spanish or French.


95

4.1 Path elaborations: The complete-path hypothesis One of the most striking differences between Basque and a prototypical verbframed language such as Spanish is the elaboration of Path descriptions. In a contrastive study of English and Spanish Path descriptions of scenes depicting falls 6 in the frog story, Slobin (1996a:200) argues that the general tendency is to use bare verbs, i.e., verbs with no elaboration of Path beyond the inherent directionality of the verb itself, much more often in Spanish than in English (Sp. 36% vs. Eng. 15% in adult data). English tends to add some locative information to the Motion verb. Even when the comparison between these two languages is made with regard to the use of minus-Ground clauses, i.e., bare verbs and verbs with directional satellites, on the one hand, and plus-Ground clauses, i.e., verbs with the source and/or the goal conveyed in prepositional phrases, on the other, the tendency remains the same. English narratives tend to use more Ground adjuncts than Spanish ones (in adult data, minus-Ground: Eng. 18% vs. Sp. 37%; plus-Ground: Eng. 82% vs. Sp. 63%). In Basque, the tendency is to provide an extensive and detailed elaboration of Path. In the same falling scenes, the use of bare verbs is only 11% and the use of Ground descriptions is 88%, as illustrated in Table 1. As we can see from Table 1, the elaboration of Path in Basque occurs regardless of the type of verb motion. In other words, it does not matter whether the verb of Motion is 'neutral' asjoan 'go' in [7], has conflated Path asjausi 'fall' in [8], or has conflated Manner as salto egin 'jump' in [9]. [7] Azkenean,

habixe,

erleabixe,

lurrera

jun

zan

finally

nestiABs

beehive:ABS

groundiALLAT

§O:PFV

AUX

'In the end, the nest, the beehive, went down to the ground.' (B20d)

[8]

Bapatean,

erlauntza

lurrera

jausi...

suddenly

beehiveiAes

groundiALLAT

fall:ppv

'Suddenly, the beehive fell down to the ground ...' (B20i)

[9]

Zakurrak ere

zuhaitzen gainera

salto egin

dog-.ERG

tree:GEN

jump

also

top:ALLAT

duela

dirudi

make:ppv AUX:CMPLR seems

'It seems that the dog has also jumped onto the tree.' (B20f)

The extensive elaboration of Path descriptions might be a consequence of the very nature of the Basque language and its rich lexical resources for spatial expression. There are five locational cases and over thirty postpositions, also inflected for these cases, that allow finer and more detailed descriptions of space and of motion

96


Table 1. Comparison between bare verbs and Ground descriptions Verb

Scene Dog (12)

fall Bare verb

2 16% 2 16% 8 66%

Total

fall throw

2 16%

go Ground description

Beehive (12) 4 33% 4 33% 5 41% 2 16% 1 8%

Total Boy (13) 1 7% 1 7% 9 69% 2 15%

jump (no verb) Total

10 83%

8 66%

1 7% 12 92%

Deer (22)

(59)

13 59% 6 27% 2 9 1 4%

7 11% 7 11% 35 59% 10 16% 5 8% 1 1%

22 100%

52 88%

First row = number of tokens, second row = percentages. Numbers in parentheses in the Scene columns denote number of motion-event descriptions for each scene.

in space. In fact, Turkish - a language of similar characteristics in this respect - is also reported to have more elaborations of Path than other verb-framed languages (Aksu-Koc 1994). Although Turkish and Basque share this frequent and extensive elaboration of Path, it seems that each of these languages may have a different motivation for doing so. Aksu-Ko? (1994) suggests that the motivation in Turkish is due to discourse factors: bare verbs occur when the Ground information can be presupposed from the context, and more extensive locative modification occurs when there is a need for more detailed information about the source and goal. The motivation in Basque is somewhat more complex. Path elaboration occurs not only in these discourse environments, but also in cases such as [10] and [11] below (with the same motion verbs joan 'go' andjausi 'fall' as before): [10] Bapatean suddenly

Txuri txakurra

leihotik

txuri

window:ABL below:ALi_AT

dog:ABS

behera

'Suddenly, Txuri the dog went down from the window.' (B20i)

joan

zan

goippv

AIX


[11] Eta and

97

zuhaitzatik behera

jausi

tree:ABI.

falhppv

below:ALLAT

zen AUX

'And (the boy) falls down from the tree.' (B20i)

In these two sentences, there is very specific information about the path followed by the dog in [10] and the boy in [11]. These locative phrases specify both the source - leiho 'window' and zuhaitz 'tree' - and the goal - behe 'below' - of the motion. But if we look closely at the lexical items that fill these roles, we notice that they are different. Whereas the first element, the source, is a noun, the second element is a postposition. That is to say, the second element is much more general in meaning than the first one. What is more, in [11] the second element is somewhat redundant, because the semantic information provided in behera 'below:ALL' is already conveyed by the verb jausi 'fall'. These are examples of what I call the complete-path hypothesis', the tendency to linguistically express in the same clause both the source and the goal of a translational motion, even in cases where one of the source- or goal-denoting components is pleonastic. The use of complete-path clauses (henceforth, "CP") in the data is very high. CPs are found in all fifteen frog stories. Most of the CP examples (18 tokens, 30% of the Ground descriptions) are used to describe the "falling" scenes; therefore, the second element in these examples is always the postposition behera. However, CP clauses are not constrained to downward movements.7 [12] Eta kristalezko and

ontzitik kanpora irtetea

glass:iNSTR:ADN 8 jar:ABL

out:ALLAT exit:GBR:ART

lortu

zuen

obtain:ppv AUX

'And (the frog) managed to come out of the jar.' (B20h)

In [12], the direction of the CP is not downwards, but outwards. Again, as in the other two examples, the semantic content of the postposition kanpora 'out: ALL' is already conveyed by the verb irten 'exit'. CP clauses are not a grammatical requirement of Basque. They are not compulsory. Sometimes speakers use them, sometimes they do not. Although further research into this matter is needed, I hypothesize that CP clauses are used in those cases where the path is delimited, i.e., the locations of both the source and the goal are conceptualized as fixed points in space, and as a result the distance between them (Path) is viewed as a restricted, delimited trajectory between these two fixed points. This hypothesis seems to be further supported by the gestural information that sometimes accompanies these clauses. As will be explained in more detail in Section 4.4, CP gestures depict very clear beginnings (sources) and ends (goals). The gesture never indicates that the goal

98


of motion is, say, "somewhere towards the right", but always, say, a specific point to the right. 4.2 Path elaborations: Descriptions of physical setting Another piece of evidence that shows how Basque devotes special attention to the description of Path comes from the analysis of the physical setting where the movement takes place. According to Slobin (1996b), the amount of narrative attention devoted to the description of the physical setting in satellite-framed and verb-framed languages is different. Satellite-framed languages such as English seem to leave the setting to be inferred, focusing more on movement and, thus, elaborating the Path description more. Verb-framed languages such as Spanish seem to follow the opposite strategy: they leave the Path to be inferred and allocate more narrative attention to the description of the physical setting. As Slobin argues, the information about the Ground provided by the narrators is the same in these two languages. What seems to be different is, first, the means by which this information is conveyed: in Path descriptions or in setting descriptions; and, second, as a consequence, the focus of narrative attention: dynamics of movement or static scene-setting. Basque, in this respect too, is more similar to English than to Spanish. If we analyse the same description of the fall from the cliff in the deer scene that Slobin uses for the illustration of these differences, the results for Basque are close to those for English. Let us look at some examples: [13] [DJanak amildegitikan behera alliABS

cliffiABLiLoc

erori

below:ALi_AT falliPFv

zian

ibai

batera

AUX

river

one:AU.AT

'All of them fell from the cliff down into the river/ (B20b)

[14] [A]mildegitik behera cliffiABL

gure Zikina eta gure Mikeltxo erori

belowiALLAT our

zikina and

our

mikel

Erori

ziren

halako

osin

batera

fallrppv

AUX

that:ADN

pond

oneiALLAT

ziren.

falliPFV ALX

'And our Zikina [dog] and our Mikel [boy] fell down from the cliff. They fell into some pond.' (B201)

Descriptions such as those in [13] and [14] are the norm in our data (80%). The Ground information (the river's location below the cliff) is to be inferred from the description of the trajectory ("from the cliff'; "falling down": "into a river/pond")

99


adjoined to the verb of Motion. The only difference between these descriptions and the English data is the Basque-speakers' tendency to separate into different clauses each of the segments of the falling trajectory. In Example [14], instead of the falling scene being described in a single clause "fell from the cliff into the river", the first clause describes a complete-path trajectory from the cliff down, and the second clause complements the first one by giving us information about the place where the boy fell: the pond. A static description of the same Ground is found in only three cases (Examples [15-17]): [15]

[E]ta and

gure Andoni our andoni

erori falliPFv

eta and

amildegitik behera cliff:ABL below:ALLAT

txakurra dogiABS

ziren, baina AUX but

amildegiaren cliff:GEN

azpian

erreka

below:Loc

river:ABS

zegoen zorionez

eta

ez

zitzaien

ezer

gertatu

was

and

NEC

AUX

nothing

happen

fortunately

'And our Andoni [boy] and the dog fell down from the cliff, but fortunately, there was a river below the cliff, and nothing happened to them.' (B201)

[16]

Nora erori? where:ALLAT falliPFV bat one

zegoen was

eta and

Ba well

uretara,

han

water:ALLAT

there:Loc below:Loc river-oiM

erdian putzu middle :LOC pond

putzua

zela

pond:ABS

was:cMPLR otherwise well

bestela,

ba,

baina putzu erdira

erori

but

fallippv

pond

middle:ALLAT

behean

handi big

bat, one

erreka-txo

eta eskerrak and thanks

ba,

han

well

there:LOC remainiFUT AUX

geratuko

ziren;

ziren AUX

lehenengo Aitor eta gero first aitor and then

atzetik Txuri back:ABL txuri 'Where did they fall? Well, to the water, down there, there was a little river, and in the middle a pond, and thank God it was a pond, otherwise, well, well, they would have died there; but they fell into the middle of the pond, first Aitor [boy] and then Txuri [dog].' (B20d)

100


[17] [AJmildegitik behera jausi ziren; baina zorionez, amildegiaren cliff:ABL below:ALLAt fall:ppv AUX falliPERF fortunately cliff:GE\ azpian below:toc

erreka river

bat one

zegoen was

eta and

uretara jausi wateriALLAT fallippv

ziren ALX

They fell down from the cliff, but fortunately, there was a river below the cliff and they fell into the water.' (B20j)

However, it is only in one of these three cases - Example [15] - that the static description is the only information we have about the Ground: "there is a river below the cliff'. In the other two - [16] and [17] - the static description co-occurs with a dynamic description of movement. In sum, although Basque is a verb-framed language, both the elaboration of Path and the focus of narrative attention in this language are more similar to those of satellite-framed languages. The use of Ground adjuncts with verbs of Motion is frequent, and the number of static descriptions of the physical setting is small. In other words, Basque tends to devote more narrative attention to the dynamics of movement; therefore, its rhetorical style in this respect seems to be more akin to that of satellite-framed languages. Slobin (1996b:205) suggests that, in the case of English, this tendency is probably due to the "availability of verbs of motion (often conflated with manner) that can readily be associated with satellites and locative prepositional phrases to trace out detailed paths in relation to ground elements." This may also be a plausible explanation for this tendency in Basque. Although its lexicon of Motion verbs is more similar to those of typical verbframed languages, Basque has rich lexical and grammatical resources, five locational cases, and over thirty postpositions which allow the description of detailed Paths with regard to Ground elements. However, rich lexical and grammatical resources cannot be the sole reason for this tendency. In Turkish, another agglutinative verb-framed language with similar resources (cases and locative nouns), there are also more frequent elaborations of Path (see the discussion above), and yet narrative attention seems to focus on static scene-setting (Slobin 1996b, 1997a). Perhaps another decisive factor for the explanation of this tendency is the frequent omission of the verb (mostly periphrastic) in Basque. As I will explain in Section 4.3, it is very common to find utterances with detailed descriptions of the Path of motion (source of motion, goal of motion, etc.) but no mention of the verb auxiliary or the main verb. The possibility of leaving the main (Motion) verb unmentioned, while still producing a grammatically correct utterance, may be another reason for the focus on the dynamics of movement through the elaboration of Path in Basque9.


101

One difference between Basque and satellite-framed languages is, nevertheless, its inclination to represent in separate clauses each of the segments in a given complex Motion event. This is very typical of verb-framed languages, but instead of "retarding] the fluent depiction of a journey, [and] perhaps leading to a preference to limit the specification of path details unless absolutely necessary", as Slobin (1996b:217) argues concerning Spanish, in Basque this practice produces a vivid, detailed, and dynamic description of the Path taken in each of these segments10, with no clear preference to limit the use of Path elaboration. 4.3 Sound symbolism Hinton, Nichols, and Ohala (1994:1) define sound symbolism as the "direct linkage between sound and meaning". Although in traditional linguistic theory (Saussure 1916), the relation between these levels of language is assumed to be arbitrary, several studies of sound symbolism (Abelin 1999; Hamano 1998; Hinton, Nichols & Ohala 1994) have been successful in proving this assumption, if not completely wrong, then not completely right, either. There is a non-arbitrary element in the way some meanings are linked to some sounds, and in the way some sounds are linked to some meanings. Furthermore, it seems that sound symbolism occurs in many languages around the world, perhaps in all of them, even though some languages, such as Basque, appear to be more prone to sound-symbolic formations than others. The soundsymbolic system, though largely under-studied, is very rich and exceptionally important in some areas of the Basque lexicon, especially in the semantic fields related to small creatures (amphibians, reptiles, insects, crustaceans, birds, and fish), types of activity, weather phenomena, noise-making instruments, physical characteristics, and sexual terms (Trask 1997:259). It is therefore no wonder that we find some sound-symbolic words in our frog-story data. Hinton, Nichols and Ohala (1994:2-5) distinguish four different categories of sound-symbolic words: /.

Corporeal sound symbolism, that is "the use of certain sounds or intonation patterns to express the internal state of the speaker, emotional or physical";

//.

Imitative sound symbolism, that is "onomatopoeic words and phrases representing environmental sounds";

in.

Synesthetic sound symbolism, that is "acoustic symbolization of nonacoustic phenomena"; and

102

iv.


Conventional sound symbolism, that is "analogical association of certain phonemes and clusters with certain meanings".

In our data, there are examples of each of these categories. There are corporeal sounds such as zas! and ras!, which could be said to function as adverbs in sentences [18] and [19] since they indicate a rapid movement. [18] [T]a and

ras!

Salto

egina

bat

enbor

ganera

ras

jump

made

one

log

abover.ALLAT

'And "quickly" jump onto the log.'

[19] [E]ta and

zas!

Sagu

kanpora

zas

mouse

outsideiALLAi

'And suddenly, a mouse [comes] out.' (B20d)

Note that rapid movements and rapid audible motion through air are usually associated with fricative phones, as in these two examples (Hinton, Nichols & Ohala 1994;Ohala 1984). There are also several imitative sound-symbolic words, such as those referring to the sounds of animals: txau txau 'dog's barking' and korroka (korroko 'belch' + iterative suffix -ka) or kroak kroak 'frog's croaking'. Further, there are conventional sound-symbolic sequences such as sibilants" for expressing silence in isil-isilik 'quietly' or rapid movement in irrist egin 'slide' and ziztu bizian 'very fast'. Finally, but perhaps more interestingly for the purpose of this chapter, there are a few examples of sound-symbolic expressions for describing different types of motion. Hinton, Nichols, and Ohala (1994:3) call these cases "movement imitatives" and define them as "representation^] of movement w ith the same kinds of sound symbolic forms that are used for the representation of sounds". Although I will not discuss the appropriateness of this term for these cases in Basque any further, I nevertheless feel inclined to disagree with the inclusion of these cases in the second category of imitative sound symbolism. It is true that to some extent they do imitate the sound produced while carrying out these activities, but as I will show later, they also follow some precise phonaesthetic rules. Therefore, I would rather consider these cases, together with those described above, not as belonging to one category or another, but as belonging to what Hamano (1998:9) refers to as an iconic resemblance scale; that is, a continuum ranging from very iconic cases to cases with less or almost lost iconic resemblance.


103

There are two examples in the data12: plisti-plasta 'to wade' and taka taka 'to walk with small and short steps, to creep, to crawl'. Let us analyse these cases in some detail. With respect to their denotation, both of them mean 'to walk'. Plisti-plasta specifically refers to walking through water; that is, it contains information about the Ground, or rather about the medium (Slobin 1996a, 1997a). It is used to describe the scene where the boy and the dog walk out of the river and climb over the log. Taka taka denotes a special kind of walking, in small and short steps, or even toddling. It is used to describe the boy's and the dog's motion from the river, over the log, to the place where the frogs are. With regard to their form, both expressions are uttered alone, with no other lexical items such as a Motion verb. This is actually quite a common feature of both oral and written narratives in Basque (Ibarretxe-Antunano 2001). Basque verbal morphology is largely periphrastic, apart from a few very common synthetic verbs. The periphrastic expression is composed of a non-finite form marked for aspect and a finite auxiliary marked for tense and mood, plus all the agreement markers. There is a strong tendency to leave unmentioned either the finite auxiliary or, as in these cases, both the finite and the non-finite forms. This latter type of double ellipsis is frequently found with Manner-of-motion adverbs such as these sound-symbolic expressions, or with goal complements such as kanpora 'outside: ALL' in [19] above. As I have argued elsewhere (Ibarretxe-Antunano 2001), these patterns seem to be discourse-constrained. They are allowed not only when the information can be inferred from context (subject, tense, etc.), but also when the focus of attention is placed on those pieces of information conveyed by the items mentioned. That is to say, the most relevant pieces of information in these cases are the Manner of motion in the sound-symbolic expressions, and the goal of motion in [19]. This is precisely what is coded linguistically; the rest is not important, it is, in Talmy's terms, backgrounded and, therefore, left out. Interestingly enough, both of these sound-symbolic expressions are complemented with gestural information, as I will explain in the following section. Both expressions follow typical sound-symbolic strategies. Based on the studies included in the book edited by them, Hinton, Nichols, and Ohala (1994: 9) point out that there is an overall tendency, across languages, to use infrequent segments and supra-segmentals. This tendency holds true for both of the above cases. Initial voiceless plosives, such as /t/, and initial clusters, such as /pi/, are considered infrequent segments in Basque phonology (Trask 1997:258). Although they are common in present-day Basque, they were most probably not present in historically early Basque phonology. In early-Basque phonotactics, only the consonants corresponding to the graphemes , , , <s>, , and seem to have been allowed word-initially, and there were no word-initial clusters at all (cf. Hualde 1991; Hualde, Lakarra & Trask 1995; Michelena 1990; Trask 1997).

104


Another strategy in sound symbolism is reduplication. Hinton, Nichols, and Ohala argue that "European languages in general utilize reduplication less than the rest of the world" (1994:9); Basque, however, is an exception to their generalization. Reduplication, as in taka-taka, and "partial reduplication", as in plisti-plasta, are two of the most salient phonaesthetic characteristics of Basque sound-symbolic words. In fact, as we have seen in some of the previous examples13, reduplication is not constrained to this type of words; it is also used as an "expressive" or "emphatic" discourse strategy. 4.4 Gestural expression of Manner and Path The importance of gesture in human communication is not a new topic in linguistic discussion (Kendon 1982). However, in recent years it has become a more important field of research for those who are interested in understanding the conceptual structure of a language, and the speaker's conceptual representation. Gestures are not just signals produced by body movements. Gestures are meaningful. They represent thought in action, and are an integral part of language totally intertwined with speech in time, meaning, and function (McNeill 1992). This is why, as McNeill (2000) has shown in a recent article, it is also possible to see differences in the association and elaboration of gestures for the encoding of visuospatial cognition between typologically different languages such as Spanish (verb-framed) and English (satellite-framed). These differences support Slobin's thinking-for-speaking hypothesis. In this section, I will briefly analyse the relationship between gesture. Motion events, and thinking-for-speaking in the Basque frog stories. The results I will present, although robust in relation to these data, are to be treated as indications or suggestions of what a more detailed analysis of gestures might reveal about the gestural system in this language. The gestures I discuss are divided into three areas: a} those that describe Manner, b) those that describe Path, and c) those that describe Path and Manner. Let us start with the first group. As we have already seen, Manner is usually expressed in a separate adverbial expression in Basque: korrika 'running', saltoka 'jumping', biriboraka 'rolling'. hegaka 'flying', and so on. Another way of expressing Manner is by means of a gesture, as in [20] :14 [20] (E)ta / [ / brst ], holako hegazti bat atera zion, and

brst

that:ADN bird

one exit

zera. mozolo bat

ALx:3so:3so what owl

one

'And, suddenly, a bird like this, ehm, an owl came [flew] out to him.' (B20d) (1) BH, B spread shape, palms down, RH in right P, LH in left P: Motion: PS. 2SM.


105

In this sentence, the speaker uses a particular gesture to encode the Manner of motion: both hands, situated on either side of the head, flutter. They symbolize the bird's wings and their movement. It is only thanks to the gesture15 that we know how the owl came out, since Manner is not verbally coded in [20]. Cases of this type are examples of "Manner fogs", i.e., scenes described by Manner gestures in the total absence of Manner information in speech. (McNeill 1992, 2000; McNeill & Duncan 2000). In the data, there are several other examples of Manner fogs. A recurrent case is the description of the scene where the dog is shaking the tree with the beehive. The verb of Motion used in these cases is eragin 'cause to move'. Eragin does not specify how the entity - the tree in our case - is made to move; this information is given in these narratives by the accompanying gesture: both hands, S-shape - as if gripping something -, shaking. It is important to note that the use of gesture to encode Manner is not necessarily a compensation for the absence of Manner in verbal expressions. There are examples such as [21 ] in our data where Manner is encoded both in the gesture (both hands rotating) and in the linguistic expression (korrika 'running'). [21] (T)xakurra arinka / arinka, hor doa, [ korrika ]| / [ korrika / korrika ]] dog quickly quickly there goes running running running 'And there goes the dog, very fast, running and running.' (B20e) (1) BH, S shape, RH in right C, LH in left C; Motion: AB, 2SM.

In a recent paper, McNeill (2000) argues that the use of gesture in two typologically different languages - English (satellite-framed) and Spanish (verb-framed) - is related not to the tendency to incorporate Manner verbally or not, but to the mechanisms by which Manner is incorporated. McNeill states that "Spanish appears to have a 'gesture strategy' whereby manner is an integral part of conversational speech but is not (always) spoken. English has a 'speech strategy' in which gesture performs a different function, that of modulating spoken manner" (2000:53). In other words, Manner in Spanish (verb-framed languages) is optional; therefore, if it is expressed in any way (gestural or verbal) in an utterance, it immediately carries communicative weight. In English, on the other hand, Manner is not optional; therefore, the only way for the speaker to indicate the communicative weight of Manner information is by choosing to include a Manner gesture (enhancing) or by choosing not to (minimizing). Basque seems to behave like a verb-framed language in this respect. The second type of gestures in the data are those which depict Path. As we have seen previously, Basque, despite being a verb-framed language, tends to give a detailed elaboration of Path in most cases. This is also reflected in the gesture data.

106


There is a tendency to accompany the verbal expression of a Path description with a gesture. Let us look at some examples. [22]

[ (L)ehen esan before say:ppv

dugu hori ], AUX that

/erle / bee

zera / what

ba

lurrera

erori

zela

well

ground:ALLAT

falliPFv

AUX:CMPLR

erlekabia beehive:ABS

That, what we said before, the bee... uhmmm, that the beehive fell to the ground.' (B20d) (1) BH, palm against palm, BH upper P; Motion: PF down, 2SM.

In this sentence, the gesture starts with both hands together situated above the head of the informant, depicting the situation of the beehive hanging from the tree. They are kept in that position for a second and then they move downwards, representing the fall of the beehive. Interestingly enough, the gesture precedes the verbal expression of "falling". In sentence [23], the gesture is more complicated: it has two parts. In the first one, the right hand with the palm down moves from the left to the right, depicting the distance covered by the deer before it arrives at the cliff and stops. In the second part, the hand in the same position moves upwards a little bit just at the same time as the emphatic locative case -an is uttered, and then moves downwards, coinciding with the postposition in the allative case behera 'below: ALL'. In other words, this gesture depicts the complete path conveyed in the verbal expression. [23] (E)ta azkenian [ amildegi batetikan and

finally

cliff

one:ABL:Loc

bai umia

eta bai

txakurra

yes boy:ABS

and yes

dog:ABS

behera

bota

below:ALLAT

throw:pp\

zun ], ALX

'And finally, [the deer] threw both the boy and the dog down from the cliff.' (B20b) (1) RH, B spread shape, palm down, RH in C; Motion: PF from left to right. PF downwards.

According to McNeill (2000), one of the main differences between English and Spanish in relation to Path gestures is that English tends to segment curvilinear paths into straight-line segments. Although we would probably need a more complex and extended Path scene to be sure, it could be argued that this sentence


107

might be considered an example of the kind of segmentation of Path that seems to occur in satellite-framed languages. Different segments of the deer scene are, in fact, embodied in the gestures. First, the segment of the deer running with the boy. This is depicted by the flat palm moving from the left to the right. Then, the segment when the deer stops at the edge of the cliff and throws both the dog and the boy. This is represented by the brief upward movement. Finally, the segment of the boy and the dog falling down. This is embodied in the downward-motion gesture. We should not forget either that this scene is not verbally expressed. This fact supports McNeill's (2000:48) claim that "gestural segmentation is not a direct mapping of linguistic path content but is its own source of path segmentation". In the following example, the motion of the hands depicts a horizontal path that represents the translational motion of the dog running away from the swarm of bees. [24] Txuri, txuri

esan sayiPFv

dugu, AUX

[ korrika batean running one:LOC

pasatu passiPFv

zela^ AUXICMPLR

'We said that Txuri [the dog] passed by running.' (B20d) (1) BH, palm against palm AB, BH in C; Motion: PF from left to right, 2SM.

In [24], the informant establishes very clearly what the source and the goal of the motion are. He sets up the space where the motion starts at the very left and keeps moving his hands up to the very right. This information is not so clearly expressed in the text. The verb pasatu means 'go by, pass', but this verb does not necessarily convey a translational motion along the path. The use of this verb in [24] simply tells us that the dog passes in front of the boy. The gesture is what really describes the path taken by the dog. Sentence [25] is a better example. Here, there is no verbal expression of Motion at all - there is neither a Motion verb nor a source or goal phrase, only an adverbial expression for the rate of Manner of motion: rostean 'very quickly'. [25] Eta and hor there

hemen here bai ?, yes

berriz, zierbain lepoan, buruan, again deer back:Loc head:Loc

gainean, top:LOC

mutila, boy:ABS

[ rostean ] | quickly

'And here again, on the deer's back, on its head, on top of it, the boy, there yes?, goes by quickly.' (B20h) (1) RH, RH in C; Motion: PF from centre to right.

108


The path is expressed only in the gesture - the right hand moving from the centre towards the right -, which is executed at the same time as the Manner adverbial expression. These sentences are very similar to those "Manner fog" examples where Manner of motion is conveyed only by the gesture. The difference is that in these cases, what is expressed by the gesture is the Path of motion. Therefore, I will call these "Path fogs", i.e., scenes described by Path gestures in the total absence of Path information in speech. The third type of gestures are those that give us information about both Manner and Path of motion. This information can be supplied in two different ways: either with Manner gestured first and Path afterwards, or with both Manner and Path gestured at the same time. In [26], Manner and Path are gestured separately. First, both hands are rotating while the informant says ibaitik, which can be considered another example of Manner fog since the rolling is encoded only in the gesture. At the end of the last rotation, the right hand stops on top of the left hand, and from that point, which represents the source of the motion, it moves downwards. [26] [T]a umia and

ta

boyrABS

zakurra batera

and dog:ABS

erori

zian

together fallippv

ALX

[ ibaitik ][ [ an

amildegira ]2 biak

batera

rivenABL.'Loc

cliff:ALL

together

IWCXABSIPL

'And the boy and the dog fell down [rolling] from the river to the cliff, the two together. " 6 (B20b) (1) BH, BH in C; Motion: rotation, PF, 2SM. (2) RH, RH in C, on top of LF; Motion: downwards, PF.

In the following example, on the other hand, both Manner and Path are encoded in the gesture at the same time. [27] [E]ta han and

isil

isilik

[ arrapaladan hasi

there silence silence:PARTiT fast

dira lurretik ]t

start:ppv ALX ground:ABL

'And there they go, they start to [walk] very quietly along the ground.' (B20e) (1) BH, B spread shape, palms down, BH in C; Motion: AB, LH after RH.

The gesture in [27] embodies both the Manner of motion - walking - and the Path - going away - at the same time. Both hands symbolize the feet. The speaker moves one hand after the other ("walking"), at the same time as she moves them


109

away from her body (Path). It is also important to point out that this sentence is another case of both Manner fog and Path fog. The same gesture is chosen for the two sound-symbolic expressions analysed in the previous section: taka taka andplistiplasta. In these cases, each movement of the hand is synchronized with the verbal expression, so that the first taka or plisti coincides with the left-hand movement and the second taka or plasta with the right-hand movement. Each of these sound-symbolic expressions is repeated twice, so there are a total of four "steps" forward.

5. CONCLUSIONS One of the main ideas in Slobin's thinking-for-speaking hypothesis is that each language has a different set of grammatical and lexical choices that lead its speakers into dealing with experiences in different ways. Each event is anchored in two different cognitive frames, a discourse frame common to all languages and a typological frame particular to each language. In this chapter, I have analysed the discourse frame of Motion events in the typological frame provided by the Basque language. In terms of Talmy's binary classification of languages, Basque is a verbframed language. Its characteristic expression of Motion is that typical of these languages: Path and Motion are conflated in the verb, and Manner of motion is expressed in separate adverbial expressions. Basque behaves like a prototypical verb-framed language in relation to the expression of Manner. The number of Manner verbs is very small, and all are typically first-tier verbs. The linguistic, as well as the gestural, use of Manner adjuncts is optional; as a consequence, whenever Manner is explicitly expressed it carries great communicative weight. There are different types of Manner adjuncts: those that describe motor patterns (hegaka 'flying'), those that describe the rate of motion (ziztu bizian 'fast'), and those that describe the protagonist's inner state of mind (ikaratuta 'scared'). A characteristic particular to Basque is the use of movement imitatives, i.e., sound-symbolic words for the description of Manner (taka taka 'walk in small steps'). With regard to Path elaboration, Basque seems to describe and use Path quite differently from other, prototypical, verb-framed languages. The use of Ground adjuncts with verbs of Motion is high, and the number of static descriptions of the physical setting is small. Consequently, Basque tends to devote more narrative attention to the dynamics of movement; therefore, its "rhetorical style" in this respect seems to be more akin to that of satellite-framed languages. The extensive and detailed elaboration of Path descriptions in Basque is further supported by two other specific characteristics of this language. First, there is a tendency to express

110


linguistically, in the same clause, both the source and the goal of a translational motion, even in cases where one of the source- or goal-describing components is pleonastic (the complete-path hypothesis). Second, there is a tendency to accompany verbal descriptions with gestural information. In many of these cases, the gesture not only depicts the Manner of motion, but also the Path. The analysis presented in this chapter has drawn upon elicited oral narratives. Therefore, it can only be taken as a possible indication of how the typological frame in Basque imposes a subjective tendency upon the schematization of experience in Motion events. Further research and analysis of Motion events in other types of narratives (novels, translations, etc.) will tell whether the ideas put fonvard in this chapter are a true reflection of thinking-for-speaking in this language.

NOTES 1.

This chapter is dedicated to Dan Slobin for launching me into the world of Motion research. I would also like to acknowledge the support and help I received from the International Computer Science Institute and the Institute of Human Development at UC Berkeley. This research is supported by Grant BFI99.53.DK from the Basque Country Government's Department of Education, Universities, and Research.

2.

By "expressive" Slobin refers to the number of distinctions that Manner verbs make in terms of motor pattern, rate, affect, and evaluation of movement. Although both types of languages do have "expressive" means for referring to Manner of motion, satelliteframed languages have more expressive Manner verbs than verb-framed languages do. Slobin attributes this fact to the "availability of the combined slot for Motion and Manner, [which poses] no additional cost to adding richer manner expressions, since the slot must be filled by some verb or other in order for a syntactically complete sentence to be produced" (2000:113).

3.

The informants were speakers of the Bizkaian, Gipuzkoan. and Low Navarrese dialects of Basque. I would like to take this opportunity to thank them for their help and collaboration in this project. Special thanks to Joselu Fdez. Bolado who left us all so unexpectedly.

4.

This is the major innovation compared with Berman and Slobin's (1994) procedure, since they only audio-taped their narratives.

5.

Talmy (2000a:27) defines as characteristic those expressions which are "colloquial in style, [...], frequent in occurrence in speech, [...] and pervasive".

6.

There are four of these: the dog falling out of the window, the beehive falling from the tree, the boy falling from the tree, and the deer throwing the boy into the water.

7.

In fact, CP clauses can also be used metaphorically in expressions such as goitik behera (above:ABL below:ALLAi) 'everywhere, thoroughly", hitzetik hortzera (word:ABL tooth:ALLAi) 'chattering', and berlatik bertara (right.there:ABi. right.there: ALLAT) 'at once, immediately, then and there'.


8.

ADN stands for adnominal.

9.

Slobin (personal communication) informs me that omission of the verb is not common in Turkish. This fact seems to support my hypothesis.

10. This is especially interesting if we also take into account that Basque informants, for the same deer scene, tend to mention an average of 4.4 segments per person (Ibarretxe-Antunano, manuscript). Slobin (1996b:203) points out that in a complex Motion-event scene, Spanish-speakers do not tend to " 'compensate' for the minimal use of source-goal clauses by means of a series of separate action clauses that analyse a journey into its components." In fact, while 100% of the American English narrators mention three or more segments, only 75% of the Spanish narrators do so. In Basque, too, 100% of the narrators mention more than three segments. 11. There are three main sibilants in the Basque phonological system: voiced laminoalveolar, voiceless apico-alveolar, and voiceless palato-alveolar, corresponding to the graphemes , <s>, and <x>, respectively. The latter is also used for "expressive palatalisation" - for example, isil-isilik can be produced as ixil-ixilik without change of meaning but with an additional expressive charge (see Hamano (1998) for similar processes in Japanese). 12. There is a third type in the data that could have been included in this group. This is hiriboraka 'rolling'. I do not discuss it, though, because its structure is slightly different: it contains the iterative -ka. Nevertheless, the phonaesthetic rules that I explain in this section are applicable to this adverb as well. The number of movement imitatives is quite large in the Basque lexicon; see Ibarretxe-Antunano (in preparation) for a detailed study. 13. Recall lasai-lasai 'calmly', poz-pozik 'happily', zain-zain 'watchful'. Some other examples from the frog-story data are: buzti-buztita 'soaked', kontu-kontari 'chatting away', haser-haserra 'very angry', txiki-txikia 'very small\freska-freska 'refreshed', and hazter-bazterrean 'in the corner' (the last element takes the case marking). 14. I follow some of McNeill's (1992) coding of gestures. [ ] = onset and end of motion, underlining = a gesture hold in which the hands stop in mid-air, / = silent in speech. BH = both hands, RH = right hand, LH = left hand, P = periphery, C = centre. For motion, TB = towards body, AB = away from body, PF = parallel to front of body, PS = parallel to side of body, 2SM = both hands same motion. The index of the right square bracket corresponds to the appropriate specification of the gesture following the example. 15. It could be argued that the onomatopoeic word brst might also be a clue for flying. However, I would rather argue that it refers more to the rate of motion, i.e., a rapid, quick, audible motion, than to its motor pattern, i.e., flying. 16. Note that even though there is a production error in this sentence - the informant substitutes the river for the cliff, and vice versa -, the speaker's message is clear, not only from the sentence before this one, where she correctly says amildegitikan behera ... ibai batea 'from the cliff down ... to the river', but also from the second part of the gesture, where she depicts a downward Path.


Time, Space, and Manner in Swedish and Icelandic1 Narrative Construction in Two Closely Related Languages

Hrafnhildur Ragnarsdottir & Sven Stromqvist

1. INTRODUCTION The present study is concerned with the linguistic encoding of time, space, and manner for narrative-event construction in the two closely related Scandinavian languages Icelandic and Swedish. We will be concerned with the employment of tense and aspect categories in general, and with distinctions of space and manner for the construction of motion events in particular. The study focuses on how differences in detail between the two languages influence event construction in narrators of different age groups. Both Icelandic and Swedish have simple present and past tenses, whereas neither of them has a full-fledged aspectual system. Each language, however, offers means for encoding aspectual distinctions along a cline of grammaticalization. On this cline, Icelandic provides a larger set of relatively highly grammaticalized distinctions of on-going versus completed action. For more detailed descriptions of the two languages, see Thorell (1973), Haugen (1987), Fridjonsson (1989), Sigurdsson (1989), Irainsson (1990), and Ragnarsdottir & Stromqvist (1999). Icelandic and Swedish are both satellite-framed languages and they provide rich morpho-syntactic means especially for detailing direction. Further, like many other satellite-framed languages, they offer a rich variety of lexical options but more restricted syntactic means for encoding manner of motion. Despite their typological proximity, Icelandic and Swedish differ in several ways. For example, while the inflectional system with four cases characterizing Old Norse, a common Scandinavian language still spoken around the year 1200, was abandonded by Swedish in the late Medieval ages, it is preserved in present-day Icelandic. What are the implications of these differences between Icelandic and Swedish for narrative event construction? Do the differences influence language development and narrative development? In which ways can the approach taken

114


in the present study contribute to language typology? In order to explore these questions, we used narrative data from matched groups of Icelandic and Swedish children, adolescents, and adults. The next section (2) describes in greater detail cross-linguistic similarities and differences between Icelandic and Swedish in the domains of time, space, and manner. Section 3 describes the methodology and Sections 4 and 5 the results. The chapter is concluded with a discussion (Section 6).

2. TIME, SPACE, AND MANNER IN ICELANDIC AND SWEDISH 2.1 Completed versus progressive As already mentioned, both Icelandic and Swedish have two simple tenses, present and past, but they differ in terms of the means available for constructing an action as on-going or completed. To refer explicitly to an action as completed, Icelandic presents three types of perfect constructions where Swedish has one. As illustrated in Examples [1-3] below, all three Icelandic perfects are translated into English by have + past participle, and into Swedish by the corresponding construction ha + perfect participle. [1]

Icelandic: Snoopy Snoopy

haf-di have-PAST:3:sc

Swedish: Snoopy Snoopy

ha-de ak-t med honom have-PAST go-PERF.PART with him

far-id go-PAST.PART

med honum with him

English: Snoopy had gone with him

[2]

Icelandic: Strakur-inn boy-DEF.-NOM

var buinn be:PAST:3sc finished

Swedish: pojk-en ha-de boy-DEF:sc have-PAST

fang-at catch-PERF.pART

ad to

veida catch

en ARTMNDEF

frosk frog

groda frog

Time, Space, and Manner in Icelandic and Swedish

115

English: The boy had (already) caught a frog

[3]

Icelandic: Froskur-inn frog-DBFiNOM

er horf-inn be:pREs:3sc disappear-PAST.PART

Swedish: Grod-an ha-r frog-DEF:sG have-PKES

forsvunn-it disappear-PAST.PART

ur krukk-unni from jar-DEFiDAr.FEM

fran burk-en from jar-DF.F:sc

English: The frog has disappeared from the jar

Icelandic and Swedish both use a construction with auxiliary 'have' + past participle (just like English) which we will henceforth refer to as "HAPERF" (Ragnarsdottir, Stromqvist & Nordqvist 2000). The second Icelandic construction is vera buinn ad F-INF 'be done at V-INF', henceforth "BUPERF", and the third is vera F-PAST.PART 'be V-PAST.PART', henceforth "BEPERF". The Icelandic HAPERF is highly grammaticalized and can in principle be applied to all verbs as well as accomplish all functions of the perfect (Jonson 1992), whereas the two other perfects are more constrained. The BUPERF is used primarily with action verbs and preferably with animate, agentive subjects. Further, BUPERF is used to refer to specific, recently completed actions/events, whereas HAPERF is used to refer to more remote, general completions. The BUPERF is preferred to the HA?ERF in resultative contexts, and in informal, spoken language it is the primary candidate for the typical current-relevance function of the perfect (Svavarsdottir & Jonsdottir 1998; Wide, manuscript). The HA?ERF, on the other hand, is the only perfect for expressing evidentiality (FriSjonsson 1989: 106). The relative use of the BUPERF and the HAPERF varies with situations, genres, and texts (Wide, manuscript). The BUPERF is more likely to be encountered in spoken, informal genres than in formal or written ones, whereas the reverse holds for HAPERF.

Unlike the HA?ERF and the BUPERF, the BEPERF has many functions. With transitive verbs it forms the passive. With intransitive verbs of movement and change, on the other hand, it functions as a resultative perfect. This construction is thus more limited in scope than the other two. On the other hand, it applies to very frequent verbs (and highly relevant ones for telling the frog story), such as 'come', 'go', 'disappear', and it is stylistically neutral. The BEPERF is not included in all descriptions of the perfect in Icelandic (see, however, Wide, manuscript:

116


Kress 1982). Its inclusion in our comparative study of aspectual markers in Icelandic and Swedish is motivated by the fact that its historical equivalent in Swedish has been replaced by the HA?ERF. Icelandic also has a relatively highly grammaticalized means for encoding the notion of on-going action, the construction vera ad F-INF 'be at V-INF', henceforth called "BEPROG", which is very similar in meaning and use to the English progressive 'be V-ing'. In contrast, Swedish resorts to means of a considerably lower degree of grammaticalization. Typological descriptions of Swedish (e.g.. Dahl 1985) state that the construction halla pa att V-INF 'keep on V-ing' is the main periphrastic strategy for marking an action as on-going. Another strong candidate is the V och V construction, henceforth "V&V" (e.g., sitter och laser 'sits and reads', star och ropar 'stands and shouts'), where the first verb is a verb of posture (or, possibly, locomotion), and the second an intentional verb. Further, the two verbs in V&V show tense agreement, and the second V receives phrasal stress. Example [4] provides illustrations. [4]

Icelandic: Stelpa-n er ad les-a girl-DEF be:pREs:3so at read-iNF Swedish: Flicka-n sitt-er girl-DEF:so sit-FRES

/ /

Stelpa-n situ-r og les girl-DEF SJI-PRES and read:PRES

och la's-er and read-pRES

English: The girl is reading V&V is a permissible structure in Icelandic, too. In Icelandic, however, the construction does not seem to have the phrasal-stress property. Instead, both Vs tend to be stressed. These differences between the V& V-constructions in Icelandic and Swedish are probably a reflection of their different status in the respective language system. In Swedish, V&V and halla pa att V-INF 'keep on V-ing' are the prime candidates for the encoding of progressivity and they show prosodic traits of being promoted to a more grammaticalized status, which is not the case with the Icelandic V&V. Again, in Icelandic, BEPROG constitutes the dominant means for encoding progressivity.2 2.2 Grounds, particles, and prepositions There are both similarities and differences between the languages of the world with respect to how they encode spatial aspects of a motion scenario. A crosslinguistically valid tendency is to encode the moving object, the "trajector"


117

(Langacker 1987) or "figure" (Talmy 1985), as grammatical subject in intransitive constructions or as object in transitive constructions. In an intransitive construction, such as he fell, the subject of the intransitive verb fell encodes the trajector. In a transitive construction, such as she made him fall or she threw him off, the trajector is encoded as grammatical object and the subject position is used for encoding the causal agent of the motion event. The spatial profile of the linguistic description of a motion event further encompasses various "landmarks" (Langacker's term) or "grounds" (Talmy's term) that the trajector departs from (source), travels along (path), or travels towards (goal). In a typology proposed by Talmy (see, e.g., Talmy 1985, 1991), a distinction between satellite-framed and verb-framed languages is made. Satellite-framed languages, such as the Germanic ones, tend to incorporate or "frame" notions of direction into "satellites" of the verb, such as particles. In contrast, verb-framed languages, such as the Romance ones, tend to incorporate or "frame" notions of direction into the verb. An example of a satellite-framed motion concept is the English go out (for further examples, see [5-8] below). In Spanish, a typical verb-framed language, the same concept is lexicalized as salir. The two closely related languages Icelandic and Swedish belong to the Scandinavian subgroup of the Germanic language group. They are both typical satellite-framed languages, and yet they differ from each other in certain details, spelled out below, starting with Examples [5-8]. In the present study, we will consider a wider range of linguistic markers of spatial relations and analyse not only satellites (in Talmy's sense), but also other sorts of spatial morphemes, such as adverbs and prepositions. Given this wider range of phenomena under study, some prosodic and syntactic differences between Icelandic and Swedish should be noted; these are illustrated in [5-8] where italics indicates that a word is stressed. [5]

Icelandic: taka frosk-inn take frog-DEFiACC

[6]

Swedish: ta

upp

take up

grod-an

frog-DEF:sG

[7] Icelandic: taka frosk-inn take frog-DEF:Acc [8]

uppur up-from

krukk-u-nni jar-DEFiDAi

Swedish: ta upp grod-an ur take up frog-DEF:so of

kruk-an jar-DEFisc

118

Hraftihildur Ragnarsdottir & Sven Stromqvist

Icelandic has separate stress on the verb and on the particle/adverb (Example [5]).4 Swedish has phrasal stress on the particle/adverb (Example [6]). Icelandic places syntactic material between the verb and the particle/adverb (Example [5]). In Swedish, there is syntactic contiguity between the verb and the particle/adverb (Example [6]). When a verb phrase is further elaborated with a prepositional phrase (henceforth "PP") in Icelandic, the particle/adverb (henceforth "prt/adv") is attracted to the PP and enters into a compound preposition, thus shifting the elaboration of spatial semantics from the verb to the PP (Example [7]). In Swedish, by contrast, the prt/adv remains faithful to the verb (Example [8]). Note, further, that there are more lexical constraints in Icelandic on using the compound items as satellites (see, further, Ragnarsdottir & Stromqvist 1997). Also, the shifting of the elaboration of spatial semantics from the verb to the PP is reinforced by the fact that Icelandic, unlike Swedish, is a case language. The dative and accusative case endings on the NP in [7] illustrate this point. Stromqvist et al. (1995) proposed that the case marking and the more severe constraint on particles in Icelandic are pulling for more ground expression (for a similar argument concerning case marking, see IbarretxeAntunano, this volume). Examples [5-8] also illustrate the semantic force of the prt/adv when it participates in the linguistic construction of motion. The verb taka (\cd)/ta (Sw) 'take', which has an inherent possessive rather than spatial meaning, is shifted into a causative spatial verb when it is combined with upp/op 'out', etc. Similarly, a verb of perception such as nosa (Sw) 'sniff' can participate in the description of motion, when combined with a prt/adv as in medan hunden nosade runt i ndrheten (Sw) 'while the dog was sniffing around in the vicinity' (Stromqvist et al. 2000). In an earlier cross-linguistic study (Ragnarsdottir & Stromqvist 1997). we analysed the distribution of spatial particles/adverbs and prepositional phrases in verb phrases across longitudinal samples of early Icelandic and Swedish child language. Clear language-specific effects could be seen: •

there were more prepositional phrases than particles/adverbs in Icelandic;

•

there were more particles/adverbs than prepositional phrases in Swedish;

•

this distributional asymmetry was found both in the input to the child and in the child's own production.

A similar distributional asymmetry has been found by Behrens (1997) between English child-language data (more PPs) on the one hand and German (more pit/


119

adv) on the other. This observation weakens the claim that case marking is pulling for more ground expressions, since German is a case language whereas English is not. 2.3 Manner and the verb lexicon As described above, satellite-framed languages tend to incorporate notions of direction into "satellites" of the verb. Notions of manner, in contrast, are often incorporated into the verb, and satellite-framed languages tend to have a rich menu of lexical options describing manner of motion. Verb-framed languages, on the other hand, tend to incorporate notions of direction into the verb, whereas notions of manner of motion are often encoded syntactically in the form of a subordinated participle. For example, where a satellite-framed language such as English says / ran out, a verb-framed language such as Spanish would say sail corriendo lit. 'I exited running'. Icelandic and Swedish -just like English - make some use of this latter option, too. The by far most productive usage of the construction [V, + V2 PRES.PART] for describing manner of motion in Icelandic and Swedish is constrained to motion verbs where V, is koma (\ce\)/komma (Sw) 'come' as in, for example, hun kom hjolandi (\CQ\)/hon kom cyklande (Sw) 'she came biking'.

3. METHOD What effects do the differences detailed in Section 2 between Icelandic and Swedish have on narrative construction and the content of linguistic descriptions of motion events? To the extent that we find differences in content, these can be more safely attributed to the differences in linguistic details between our two closely related languages than what would be the case if two less typologically related languages were compared. Are the hypothetical cross-linguistic effects visible already in childhood, and how do children compare with adults? What are the implications of the answers to these questions for language typology? In order to explore these research questions in a controlled way, we decided to keep the discourse task constant across languages and subjects. Narrations were elicited by means of Mercer Mayer's wordless picture booklet Frog, where are you? (Mayer 1969) from Icelandic and Swedish subjects in four age groups: five-year-olds, nine-year-olds, fifteen-year-olds, and adults. The design thus allows us to explore similarities and differences in the description of the same picture story and the same individual motion events between the two minimally different languages Icelandic and Swedish, age group by age group. Table 1 gives a summary overview of the amount of data elicited. As can be seen, the mean number of finite clauses per narrative is very similar between the Icelandic and the Swedish peers.

120


Table 1. Size of the Icelandic and the Swedish corpora

Age

n subjects

Mean clause s per subject

n finite clauses

Icel

Sw

Icel

Sw

15

497

459

33

31

15

817

827

54

55

Icel

Sw

5

15

9

15

15

15

15

935

951

62

63

Adults

15

15

1568

1491

114

99

Several previous studies using picture-story-elicited discourse converge on the finding that there is a qualitative change in the discourse elicited between the ages of five and nine (see, e.g., Karmiloff-Smith 1981; Ragnarsdottir 1992; Berman & Slobin 1994; Nordqvist 2001). Children's productions shift from being predominantly descriptions of individual pictures without an overall narrative structure in the youngest age group to coherent stories with a very clear narrative structure by the age of nine. In an earlier study of the episodic structure of a smaller group of Icelandic subjects (10 per age group). Ragnarsdottir (1992) found that out of a total of seven story components defined for the frog story, the mean number produced by the five-year-olds included in that study was 3.4. as compared with a mean of 6.7 for the nine-year-olds, and 7 for the adults. The difference between the five- and nine-year-olds was statistically significant. As shown in Table 2, these findings were replicated exactly in the present sample of 15 subjects per age group. Accompanying the shift in story structure between ages 5 and 9, the anchor tense shifted to being predominantly the past by age 9 (Table 3). The slight decrease for the preterite in the adult group stems from the fact that four adults use the historic present as their anchoring tense.

4. RESULTS 4.1 Time The shift from predominantly individual-picture descriptions without an overall narrative structure in the five-year-olds to coherent stories with a clear narrative structure in the nine-year-olds can also be clearly seen in the present study. All the nine-year-olds tell a proper story with a setting, an initiating event, a plot, and an ending, whereas the five-year-olds build a discourse which is much more fragmented from the point of view of narrative structure. Both tense and aspect categories participate in the encoding of that structure, and the structural change

121


Table 2. Mean number of story components in Icelandic frog stories Age

n

Mean number of story components (max 7)

5

15

3.4

9

15

6.7

Adults

15

7

Table 3. Anchor tense in Icelandic frog stories. Age

n

Present

Preterite

Mixed

5

15

6 (40%)

6 (40%)

3 (20%)

9

15

2(13%)

12(80%)

1 (7%)

Adults

15

4 (27%)

9 (60%)

2(13%)

Total

45

12 (27%)

27 (60%)

6(13%)

is reflected by an overall change in the proportions of use of the different tense/ aspect categories (see Tables 4 and 5). In both the Icelandic and the Swedish data, we see a sharp increase for the past tense as the children's productions shift from being predominantly picture descriptions at age 5 to narratives with an episodic structure at age 9 (Table 3). A slightly higher proportion of Icelandic subjects in all age groups tell the story in the past tense. This is reflected in the distribution of clauses in present and past tense, respectively, shown in Tables 4 and 5. The perfect (HAPERF) has a relatively stable usage throughout development in Swedish, and it does not become really frequent until adulthood (see Table 5). In Icelandic, on the other hand, we see an increase with age in the use of the perfect constructions. A more detailed analysis of the three different perfects, BuPERF, HAPERF, and BEPERF (see Table 6), reveals a differentiated development between the three types of perfect. The use of BUPERF (recent, specific) decreases with age and is very infrequent in the adult stories (0.7%). HA?ERF (remote, unspecific), in contrast, is extremely rare in the productions of the youngest children (0.4%) but takes off at age 9 with the qualitative change in narrative stucture and continues to increase in the adult stories. The use of the BEPERF also increases slightly between ages 5 and 9 but remains stable in the adult productions. This differentiated pattern of development can be related to several factors. Let us start by considering HAPERF. Firstly, whereas descriptions of settings are very rare in the stories of five-year-olds, the nine-year-olds are notably more

122


Table 4. Distribution of T/A-categories (% of finite clauses) in the Icelandic frog-story material by age groups Age

n finite clauses

Present

Preterite

Perfect

BEPROG

V&V

5

497

32.8%

53.7%

4.8%

8.7%

0

9

817

20.8%

68.5%

7.3%

3.3%

0

Adults

1568

31.8%

57%

9%

2.2%

0

Table 5. Distribution of T/A-categories (% of finite clauses) in the Swedish frog-story material by age groups HAPERF

V&V

4.1%

1 .3%

Age

n finite clauses

Present

Preterite

5

459

75.2%

19.4%

827

46.0%

48.2%

3.6%

2.2%

39.6%

50.2%

8.2%

2.0%

9

1491

Adults

Table 6. Distribution of BEPERF, BUPERF, and HA?ERF in the Icelandic frog stories Age

BEPERF

BUPERF

HAPERF

5

2.6%

1.8%

0.4%

9

3.3%

1.6%

2.4%

Adults

3.4%

0.7%

4.8%

likely to construct a background to the first scene(s) of the story, and that type of background is typically provided by means of HAPERF. [9]

Tense/aspect use in the beginning of a story by an Icelandic nine-year-old: Oli var ad fara ad hatta (PAST, INCH.ASP).

Oli was about to get ready for bed.

Hann hafdi farid og vein (PAST HAPERF) frosk i dag.

He had gone and caught a frog today.

Hann hafdi verid mjog lengi a6 na (PAST HA?ERF, BEPROG) froskinum.

He had been [= it had taken him] very long to catch the frog.

123


Snati hafdi farid (PAST HAPERF) me6 honum.

Snati had gone with him.

Hann var ordinn (PAST BEPERF) ofboSslega breyttur.

He had become terribly tired.

Hann sofnar (PRES) en hann gleymdi (PAST) ad loka burinu

He falls asleep but he forgot to close the cage

thar sem Omar froskur var (PAST).

where Omar the frog was.

Um morguninn vaknar (PRES) hann.

In the morning he wakes up.

Hvar hefur Omar fariS (PRES HAPERF)? hugsar hann.

Where has Omar gone? he thinks.

Snati hefur ekki getad teki6 (PRES HAPERF) hann

Snati has not been able to take him

pvi hann svaf (PAST) vid lappirnar a mer...

because he slept at my feet...

Second, the use of the HA?ERF is clearly related to the increasingly cohesive story structure. The increased usage of the HA?ERF reflects, among other things, more long-distance backtracking - something which serves to increase the global cohesion of the discourse. Interestingly, our data suggest that backtracking to an event already mentioned is relatively more frequent in the Icelandic than in the Swedish narrations. Example [10], taken from a Swedish adult subject, provides an analysis for the purpose of illustration. [ 10] Backtracking with reference to old and new events in a Swedish adult: Torsten fortsatter letandet

Torsten continues the search

new event =7

och nu har han tappat bade sin kompis Lasse och Grodan Boll

and now he's lost both his buddy Lasse and Frog Ball

new event =2

sa han eh ropar fortvivlat

so he ehr calls out in desperation

new event =3

och en nagot skamsen Lasse kommer framtrippandes

and a somewhat ashamed L comes tripping forth

new event =4

och nar Torsten da har ropat sa mycket pa Lasse och pa Grodan Boll

and when Torsten then has called so much for Lasse and Frog Ball

old event =3

BT

BT

124


sa nar ban bar start uppe pa den bar jattestora stenen

so when he's been standing up on this huge stone

new event =5

BT

The above story fragment is analysed into event descriptions, and each event is given a numerical index. Co-indexed descriptions refer to the same event. Each event is further categorized in terms of its newness and in terms of whether it represents an instance of backtracking (BT), that is, of moving back in terms of story time. Example [ 10] contains two instances where new events (not previously described in the discourse) are introduced through BT, and one instance of BT to an old event (an event already mentioned). BT is a rare type of narrative strategy in the youngest age groups in both languages but starts gaining ground among the nine-year-olds. Also, the strategy of backtracking to an old event (old event = 3 in [10]) is much more frequent in the Icelandic data, where 50 instances were found in the frog stories (see Table 7) compared with only 2 in the Swedish data. The five-year-old child, on the other hand, is typically concerned with describing the activites in each picture, and connects them in a sequential, resultative manner. Example [11] illustrates the relevance of the BLPERF (and of the present tense and the progressive) for these purposes. [11]

Tense/aspect use by an Icelandic five-year-old: Og sidan litur (PRES) strakurinn ofani einhverja holu og kallar (PRES)

And then the boy looks down-into some hole and calls

og hundurinn er ad reyna a6 na and the boy is trying to get down (PRES BEPROG) nidur holfi flugunnar the fly's cage her er - begar hann er biiinn ad kalla (PRES BUPERF)

here is - when he is done calling

ba kemur (PRES) ein moldvarpa uppiir

then one mole comes up-out of

og svo begar (ad) hundurinn er buinn ad na (PRES BUPERF) bvi nidur,

and then when the dog is done getting it down

ba kemur (PRES) fullt af byfiugum ur

then a lot af bees come out

og strakurinn er enn ad kalla og kalla (PRES BEPROG)

and the boy is still calling and calling

125


Table 7. Backtracking to old events in the Icelandic frog stories Age

n

n of BT to old events

5

10

3

9

10

12

Adults

10

35

Finally, the reversal with age in the relative importance of euPERF and HA?ERF participates in a stylistic change from casual to more formal style in Icelandic, whereas the corresponding HAPERF is stylistically neutral in Swedish. Let us now turn to the encoding of progressivity of events. In Icelandic (see Table 4), BEPROG shows a peak in the youngest age group: 8.7% of all finite clauses produced by the Icelandic five-year-olds contain BEPROG. The rate then decreases in a linear way to 2.2% in the adult group. This decrease in use reflects a change in the narrators' communicative intent and the corresponding change in function of BEPROG. The five-year-olds often employ BEPROG to describe what is going on in an individual picture, whereas the older subjects use BEPROG for constructing the relationship between events (foregrounding/backgrounding). The construction V&V is non-existent in the Icelandic data (see Table 4). In Swedish (Table 5), the usage of V&V increases with age to reach 2.0% in the adult group - a figure which perfectly matches the frequency of usage of BEPROG in the Icelandic adults. The Swedish construction halla pa att V-INF 'keep on V-INF', which is considered by Dahl (1985) to be a main periphrastic strategy for marking an action as on-going, is used only marginally (altogether 9 instances across all age groups, or 0.2% of the utterances) in our narrative corpus, whereas V&V is abundant. Since the first verb in V&V must be a verb of posture, and since the story events are rich in body postures, the frog story presents a discourse that is particularly conducive to the usage of V&V for marking ongoing action. Conversely, the V&V-constructions of the Swedish frog stories are multifunctional: they mark progressivity of actions at the same time as they add information about body posture. 4.2 Space In Section 2.2, we observed that the elaboration of spatial semantics in relation to a verb tends to drift more towards the prepositional phrase in Icelandic (V + PP), whereas it is more evenly distributed between the VP and the PP in Swedish. Recall, in particular, that Ragnarsdottir and Stromqvist (1997), in a crosslinguistic study based on longitudinal case studies, found an asymmetry between

126

Hrafnhildur Ragnarsdottir & Sven Stromqv ist

early Icelandic and Swedish child language: Icelandic child language was richer in spatial verbs + (prt/adv) + PPs than in spatial verbs + prt/adv, whereas Swedish was richer in spatial verbs + prt/adv than in spatial verbs + (prt/adv) + PP. Does the same tendency hold true for the frog-story data? An exhaustive analysis of the syntactic contexts of all motion verbs in the five-year-old subjects did indeed reveal the same tendency. Tables 8 (Icelandic) and 9 (Swedish) summarize the constructions used by Icelandic and Swedish fiveyear-olds in clauses describing motion events in the frog story. In both the Icelandic and Swedish data, the development can be described as a drift towards a greater proportion of constructions including a prepositional phrase. The magnitude of this increase is similar for the two languages, in the order of 30%, but the points of departure at five years of age are very different. Whereas 41% of the constructions used already by the Icelandic five-year-olds include a PP, the corresponding figure for the Swedish five-year-olds is 13%. The Icelandic adults "land" at a proportion of 71% constructions containing a PP. and the Swedish adults at a proportion of 42%. The difference in preferred constructions between Icelandic and Swedish subjects does not merely represent a difference in grammatical ways of encoding the same or very similar information. It also represents a tendency towards a difference in semantic content - a difference in "filtering" in Berman and Slobin's (1994) sense. More precisely, referential information about grounds is filtered out when V + prt/adv is used (as is more often the case in Swedish), whereas it filters through when the VP is further elaborated with a prepositional phrase (as is more often the case in Icelandic). Whereas the NP in a PP like ofan-i sjo (Icel) 'above-into sea:ACc' explicitly refers to a ground (the goal), the ground(s) are only implied, and not explicitly referred to, when the VP is elaborated with a prt/adv only, as in, for example, ramla ner (Sw) 'fell down' (see Examples [12] and [13]). We thus have very clear evidence not only of crosslinguistic differences in preferred (or dispreferred) grammatical constructions in descriptions of motion events, but also of concomitant differences in semantic filtering effects: 71% of the adult Icelandic narrators make explicit reference to grounds, as compared with only 42% of the adult Swedish narrators. Again, the semantic difference (filtering effect) is there already in the youngest age group: 41% of the Icelandic five-year-olds make explicit reference to grounds as compared with 13% of their Swedish peers. Example [12] shows a typical description of the fall from the cliff by an Icelandic five-year-old, and Example [13] a typical description by a Swedish five-year-old. [12] Icelandic: og svo datt hundurinn og strakurinn 'and then fell the dog and the boy

ofan-i sjo above-into sea:Acx '

127


Table 8. Constructions used by Icelandic subjects in descriptions of locomotion Age

n

V only

V + prt/adv

V + (prt/adv) + PP

n clauses

5

15

36%

23%

41%

182

Adults

15

14%

15%

71%

416

Table 9. Constructions used by Swedish subjects in descriptions of locomotion Age

n

V only

V + prt/adv

V + (prt/adv) + PP

n clauses

5

15

24%

63%

13%

164

Adults

15

12%

46%

42%

321

[13] Swedish: pojken ramla ner "the boy fell down' 4.3 Manner and the verb lexicon Table 10 gives an overview of the motion verbs occurring in the Icelandic frogstory corpus and Table 11 shows the verbs occurring in the corresponding Swedish corpus. Only verbs with inherent motion-related semantics have been included in the tables. Thus, for example, a verb such as taka (Icel)/ta (Sw) 'take' has been excluded, although it is often combined with a particle, e.g., tit (\cQ\}lut (Sw) 'out', to describe a motion event. Further, causative verbs of posture or position, for example setja (\ce\)/sdtta (Sw), leggja (lcel)/lagga (Sw), stalla (Sw) 'put' (or, more exactly, 'cAusE:sit', 'CAUSE:lie', 'cAUSE:stand', respectively), are excluded, although the (re)positioning of the object can be seen as a motion event. Motion verbs in the present participle, taking part in the construction [V, + V2 PRES.PART] for the description of a motion event, are included (e.g., kom traskande (Sw) 'came trotting'). In Tables 10 and 11, the items are ordered in descending frequency (the total number of instances in all age groups is shown in the rightmost column of the tables). In order to shed further light on later phases of lexical development, a group of Icelandic fifteen-year-olds and a group of Swedish fifteen-year-olds were added. The number of subjects in each age group for each language is 15. Tables 10 (Icelandic) and 11 (Swedish) reveal similar overall tendencies. First, there is a large range of motion verbs: 88 different verbs (lemmas) in the Icelandic and 70 different ones in the Swedish data. Second, the first few of the most frequent verbs account for a very large proportion of all tokens. Among

128

Hrafnhildur Ragnarsdoftir & Sven Stromqvist

Table 10. Motion verbs in the Icelandic frog-story corpus: frequency distributions Icel. motion verb fara delta koma hlaupa elta klifra henda Icedast stokkva fljiiga skrida hrista(st) hoppa lenda labba steypa(st) komast

flyja Jyigj" prila sleppa ganga bera(st) falla veita eftirjbr veifa hrinda svnda rddast a rekast kasta laumast stiga fella strjuka lyfta snarstoppa beygja sig risa troda(st) hamast hrapa sveima peytast

English translation go fall come run chase, follow climb toss creep/sneak jump/leap fly crawl shake hop land walk (col.) fall head first (from a height) reach (a place) flee follow, accompany clamber release/escape walk carry fall persecute/ follow wave shove swim attack collide with throw sneak secretly stride, step topple, cut down escape lift stop dead bend over, bow rise, ascend stuff, push forward go at it hammer and tongs fall from a great height swarm fling off

Age 5 n= 15 41 39 33 7 17 4 4 4 3 3 1 5 i

4 1 0 2 0 0 0 0 0 0 0 0 0 4 2 0 0 0 0 2 0 0 0 0 2 0 0 0 0 0 0

ARC 9 Age 1 5 Adults Sum n= 15 n= 15 n= 15 \ = 60 69 84 48 242 57 186 37 53 55 44 170 38 37 20 82 18 7 11 16 51 39 9 16 10 6 26 7 9 25 9 4 8 24 7 11 3 12 23 0 8 11 2 16 3 1 3 12 21 4 18 9 3 5 14 3 2 4 4 12 3 1 0 11 12 1 1 7 11 10 2 5 3 5 10 0 5 9 9 0 0 2 8 3 3 1 7 8 0 4 7 0 3 4 7 0 3 0 6 0 6 0 0 6 6 5 1 0 0 1 1 1 5 2 2 5 1 i 2 1 5 5 1 4 0 5 1 0 4 4 1 1 0 2 4 0 2 i 2 4 0 1 4 0 3 1 4 0 3 1 0 3 0 1 3 1 1 1 3 0 2 3 3 0 0 0 0 3 3 3 0 0 ^ 3 3 0 0


Table 10. (cont.) fleygja stinga hrasa hysja (=peysa) Jeer a sig hrifa missa rasa hlammast h rev fast klongra.it kiitveltast skokka spretta sveifla(st) seek/a rdfa .itanga stddva a last Jitekja forda (ser) hrekja hrokklast sloppa snarstansa dilla hada (in) endasendast hvolfa krafia sig pommsa olmast skella slitna smevgja ser sni'ia sturta toga rega up pa vera a ferd ogflugi velta vingsast vfirgefa

Total types Total tokens

throw away dive stumble go very fast move out of the way sweep away drop involuntarily raise flop down heavily move clamber (over difficult terrain) roll head over heels

jog take off abruptly swing fetch wander about butt, gore stop (formal style) hurry tangle up take to flight chase/shock off be chased off stop stop abruptly wag spread fall head over heels turn over struggle through plop down move fast slam to be torn apart thread one's way turn pour down pull balance move all over roll over wingle abandon

1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

1 1 1 1 1 1 1 1

22 182

31 243

49 338

67 416

1 1 1 1

1 0 2 2 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0

1 1 1 1

1 1

0 0 0 0

1 1 1 1 2 2 2

2 2 2 2 0 0 0 0 0 0 0 0 0 0 0

1

1 1 1

1 1 1 1 1 1

130


Table 11. Motion verbs in the Swedish frog-story corpus: frequency distributions Swedish motion verb

English translation

Age 5 Age 9 Age 15 Adults Sum n= 15 n= 15 n= 15 n= 15 N = 60

komma

come

32

64

28

49

173

ga ramla

go fall

13

42

84

30

169

35

34

81

13

163

springs

run

7

26

75

35

143

hoppa

jump

10

19

14

72

klattra

climb

15

11

29 22

23

71

trilla

fall

11

1

43

58

flyga

fly

6

3 2

15

11

34

falla

fall

4

18

nmma

run away

1 t

4

10

9 5

21

32

lyfla

lift

3

3

20

throw

4

6 2

8

kasta

8

18

9

6

18 14

smita

slip away

1

4 2

stanna

stop

4

2

7

smyga

sneak

0

7

3 5

16

landa

land

0

2 ->

13

throw away

4

0

8 5

3

sldnga

3

12

jolja

follow

0

4

4

3

11

krvpa

crawl

1

5

•>

1

9

resa sig

rise

1

1

1 i

9 8

vdnda

turn

0

sticka

pop off

1 2

6 5

0

1

bege sig

go to

0

0

5

5 •>

fara

travel

1

0

4

0

7 5

8

kliva

stride

0

1

4

tvdrbromsa

brake suddenly

0

3

0 •>

5

hamna

end up

0

0 2

0

2

4

rusa

rush

0

1

3

4

stoppa

stop

0 2

1

0

0

3

puna

push

1

2

0

0

0 i

3

0

3

5

3

tvdrstanna

stop dead

1

hdmta

fetch

0

1

0 2

plumsa

splash

0

0

0

3

3

ttimla

tumble

0

0

0

3

slappa

let go of

1

1

0

0

3 i


131

Table ll.(cont) valtu

overturn

1

1

0

0

2

plaska

splash

1

0

1

0

dm

pull

1

0

0

1

7 7

simma

swim

0

2

0

0

2

knuffa

shove

0

1

1

0

7

storma

charge

0

1

1

0

tvcirnita

jam on the brakes

0

1

1

0

2 7

svarma

swarm

0

0

1

2

tJ/ngla

dangle

0

0

1

1 1

stota

thrust

0

0

1

1

2

Irippa

trip along

0

0

0

2

7

traska

trot

0

0

0

2

2

rasa

collapse

0

0

0

2

ruska

jolt

0

0

0

2 2

0

2

7

2

strcicka

migrate

0

0

7

">

fly

flee

0

0

0

2

linka

limp

0

0

0

1

hissa

elevate

1 1

0

0

0

1

lunka

j°g ride

0

0

1

0

1 1

0

ricla

0

0

1

aka

go by vehicle

0

0

0

1

0

1

0

1

0

1

0

1

tap pa

drop accidentally

0

0

tindla

go sprawling

0

0

Icimtia

leave

0

0

hraka

crash down:sound

0

0

1 1 1 1 1

slunga

sling

0

0

1

0

1

stanga

butt, gore

0

0

1

0

1

dumpa

dump

0

0

0

1

jiylla

move

0

0

0

j1a.\a

flap wings

0

0

0

skicka

send

0

0

0

skutla

scamper away

0

0

0

slappla

totter

0

0

0

storta

crash, dash

0

0

0

svciva

float, hover

0

0

0

1 1 1 1 1 1 1 1

1 1 1 1 1 1 1

Total types

29

32

42

47

70

Total tokens

164

250

458

321

1193

132


these top verbs are fara (Icel) 'go'/ga (Sw) 'walk/go' and koma (lce\)/komma (Sw) 'come'. From a semantic point of view, 'go' and 'come' belong to the cross-linguistically attested set of nuclear verbs of motion (Viberg 1993). and they are highly polysemous. As we move down Tables 10 and 11 in descending order of frequency, we find verbs with a more specific meaning and, importantly, verbs encoding more fine-grained distinctions of manner of motion. From a developmental point of view, it can be noted that the the top ten verbs are found in all age groups, whereas the diversity of the verbs encountered increases with age. especially in later phases of development (beyond nine years). As an illustration, consider Tables 12 and 13, which present the distribution of motion verbs used to describe one particular scene in the frog story: the boy's fall from a cliff down into a pond (Picture 17 in Appendix I). Further, from a developmental point of view, the proportion of tokens accounted for by the first few verbs is greater for the younger age groups and it is smallest for the adults. Thus, the younger subjects seem to rely more on the most frequent verbs. This effect can be clearly seen for the top three verbs, whereas it is less obvious for the top ten. These figures are summarized in Tables Hand 15. In both Icelandic and Swedish, there is an increase of motion verbs (lemmas) with age. The increase is greater for Icelandic at all age intervals. The Icelandic five-year-olds have 22 lemmas, the nine-year-olds 31, the fifteenyear-olds 50, and the adults 67. The Swedish five-year-olds have 29 lemmas, the nine-year-olds 32, the fifteen-year-olds 42, and the adults 47. Further, the proportion of verb tokens made up by the top three verbs is consistently greater for Icelandic. A possible explanation for this latter difference is that the general predicate go has a lexicalized form in Icelandic, the verb fara. but not in Swedish. In fact, fara 'go' is the most frequent motion verb in the Icelandic data and it has a broader range of usage than its closest Swedish equivalents, ga 'walk' and aka 'go (by means of a vehicle)'. No such difference in proportion can be seen for the top ten verbs (see Tables 14 and 15). Richness in verbs differentiating manner of motion is a feature distinguishing satellite-framed from verb-framed languages. In a contrastive study of motion verbs in English and Spanish frog-story narratives (12 subjects in each of five age groups: 3, 4, 5, 9, and adult - that is, 60 subjects per language). Slobin (1998) found, in total, 47 different motion verbs in the English data, and only 27 different ones in the Spanish data. The all-in-all 70 different motion verbs in our Swedish data and the 88 verbs in our Icelandic data may be interpreted to suggest that the notion of satellite-framed languages is a scalar notion and that the two Scandinavian languages are located closer to the satellite-framing end point of that scale than English. It should be noted, however, that the different age composition of the 60 subjects in the Scandinavian groups (5. 9. 15. and


133

Table 12. Verbs used to describe the boy's fall from the cliff by Icelandic subjects Icelandic verb

English translation

detta henda

Age 5 (n=15)

Age 9 (n=15)

Adults (n=15)

fall

9

9

toss

4

4

9 2

fara a bolakaf

plunge deep

2

1

2

steypa

throw head first

1

1

1

lenda

land

1

1

4

hrinda

shove

3

0

0

flevg/a

throw

1

0

0

pommsa

splash

0

1

1

sleppa

release, let go of

0

2

0

kasta

throw

0

1

0

stanga ofani

butt down into

0

1

kutveltasl

roll head over heals

0

0

0 2

peytast

whirl off

0

0

(Idta) falla

cause to fall

0

0

3 2

velta

roll

0

0

1

hlamrnast

flop down heavily

0

0

1

hendast

dash, rush

0

0

1

steypast a bolakaf

plunge head first

0

0

1

hrista af ser

shake off

0

0

1

Total types

19

7

9

14

adults) compared with the English/Spanish ones (3, 4, 5, 9, and adults) is very likely to contribute to the relatively higher number of verbs in the former, although it hardly explains all the difference.

5. SPATIAL MORPHEMES REVISITED In the previous section, we explored the increasing lexical variety of motion verbs, which contributes, above all, to distinctions of manner of motion. Across development, there is also an increase in the variety of morphemes (particles, adverbs, prepositions) which contribute to distinctions and details of space in


134

Table 13. Verbs used to describe the boy's fall from the cliff by the Swedish subjects Swedish verb

English translation

ramla

fall

Age 5 (n=15)

Age 9 (n=15)

Age 15 (n=15)

Adults (n=15)

12

8

7 8

kasta

throw

4

4

10 2

sldnga

throw

4

0

5

3

trilla

fall

2

1

6

putta

push

1

1 2

0

0

falla

fall

1

1

5

3

flyga

fly

1

0

1

1

kmtffa

push

0

1

1

0

landa

land

0

0

4

2

rasa

collapse

0

0

1

2

stota

thrust

0

0

1

1

plaska

splash

0

0

1

0

braka

crash (sound)

0

0

1

0

tumla

tumble

0

0

0

1

plumsa

splash

0

0

0

1

storta

crash, dash

0

0

0

1

dumpa

dump

0

0

0

1

svava

float, hover

0

0

0

1

12

14

Total types

18

7

6

descriptions of motion events. Consider Table 16, which summarizes the various constellations of spatial morphemes occurring with the top three motion verbs in the frog stories by the Swedish five-year-olds. For example, this table shows that the verb ramla 'fall' is combined with the following five constellations of morphemes: / 'into', ner 'down', ner i 'down into', ner fran 'down from', in 'out'. These constellations contain four different function words: / 'into', ner 'down', fran 'from', ut 'out'. If we turn to the top ten verbs, we observe, among other things, that the verb hoppa 'jump' attracts six different morphemes (function words) and the verb rymma 'run away' none, and so on. The average number of different morphemes per verb (top ten verbs) for the Swedish five-year-olds is 2.9. Table 17 provides the corresponding summary overview for the Swedish adults, whose average variety of morphemes across the top ten verbs is 8.2.

135


Table 14. Lemmas and some proportions of motion verbs in the Icelandic frog-story material Age

n verbs (lemmas)

n tokens

Top 3 (% of tokens)

Top 10 (% of tokens)

5

22

182

62%

87%

9

31

243

62%

85%

15

50

338

52%

77%

Adults

67

415

39%

65%

All (n = 60)

88

1178

51%

74%

Table 15. Lemmas and some proportions of motion verbs in the Swedish frog-story material Age

n verbs (lemmas)

n tokens

Top 3 (% of tokens)

Top 10 (% of tokens)

5

29

164

49%

81%

9

32

250

56%

84%

15

42

458

42%

79%

Adults

47

321

29%

73%

All (n = 60)

70

1193

42%

78%

Table 16. Top three motion verbs and their variety of constellations of morphemes: Swedish fiveyear-olds Verb

Spatial morpheme

komma 'come'

upp 'up', ut 'out'

ga 'go'

hem 'home', ivdg 'away', under 'under', upp 'up', ut 'out'

ramla 'fall'

/ 'into', ner 'down', ner i 'down into', nerfran 'down from', ut 'out'

136


Table 17. Top three motion verbs and their variety of constellations of morphemes: Swedish adults Verb

Spatial morpheme

komma 'come'

from ' forward \fram till 'all the way lo\fran 'from', hem 'home', tillbaka 'back', tillbaka till 'back to', upp 'up', upp ur 'up from out of, ut 'out', ut ur 'out of

ga 'go'

bort 'away', bort ifran 'away from', bort till 'away to', bort at till 'away in the direction of, fram 'forward', fram till 'up to', hem 'home', in 'into', in i 'into', in mot 'in towards', ivag over 'away across', mot 'towards', ner in i 'down into', ner mot 'down towards', ner till 'down to', upp pa 'up onto', ut 'out', ut i 'out into', ut pa 'out onto'

ram la 'fall'

/ 'into', omkull 'over', ner 'down', nerfran 'down from', ner for 'down along', ut 'out', ut genom 'out through', ut ur 'out of

Table 18. Number of morphemes occurring with the top ten motion verbs: Icelandic Age Mean number Range Verb with most morphemes of morphemes

5 9 Adults

3.3 5.9 7.7

0-14 0-21 0-19

far a 'go' far a 'go' koma 'come'

Table 19. Number of morphemes occurring with the top ten motion verbs: Swedish Verb with most Age Mean number Range morphemes of morphemes

5 9 Adults

2.9 5.1 8.2

0-6 0-11 2-15

hoppa 'jump' hoppa jump' ga 'go/walk'


137

From a contrastive analysis of the variety of morphemes between the Icelandic and Swedish frog-story data, we obtain similar patterns. These are summarized in Tables 18 (Icelandic) and 19 (Swedish). Thus, for both languages, the mean number of spatial morphemes (for the top ten verbs) increases from around three in the five-year-olds to around eight in the adults, although the range is greater for the Icelandic than for the Swedish subjects. Further, among the verbs attracting the greatest lexical variety in terms of spatial morphemes are the two very frequent nuclear verbs come (Icel koma, Sw komma ) and go (Icel/ora, Sw ga ). See Tables 18 and 19.

6. DISCUSSION Our observations of the linguistic encoding of time, space, and manner in Icelandic and Swedish frog-story narrations lead us to a few concluding remarks about development, cross-linguistic differences, and the interaction between the two. The narrators' communicative intentions, including what genre or type of narrative to construct, change with age, and the linguistic means used change accordingly. We propose that this is the main explanation for the change in the distribution of tense and aspect marking we have observed: the change from present to past as anchoring tense, the increase in usage of the perfect with age in both languages, and the decrease in usage of progressive in Icelandic. The greater differentiation of the Icelandic tense/aspect system allowed us to pursue the developmental analysis beyond these general observations. Thus, where Swedish has only one type of perfect tense, Icelandic has three, and these three do not all show the same developmental profile. Rather, their relative frequency of usage is related to the type of discourse being produced (picture description versus coherent story). In effect, since the type of discourse changes with age, this means that the use of the different perfects has different relevance to narrators in different age groups. With increasingly complex narrative structure, the Icelandic subjects increase their use of the remote, generalized, formal HAPERF in their narratives, whereas for the purposes of the younger children, the recent, specific, colloquial euPERF is more relevant and hence more frequently used. In contrast, BEPERF remains stable across age groups. BEPERF is the only type of perfect which is insensitive to genre and modality. However, it is lexically constrained to intransitive verbs of motion and change, and many of these (such as come, go, disappear) are among the most frequent verbs in the frog-story narrations in all age groups. From a cross-linguistic point of view, our Icelandic and Swedish frog-story data strengthen the claim that these two languages are prototypical satellite-

138


framed languages. Our broader investigation of spatial morphosyntax beyond satellites (in Talmy's sense) and of verb vocabulary allows us to describe differences between Icelandic and Swedish in a more detailed fashion. In terms of the use of morphemes to specify direction of motion and in terms of the richness of the verb vocabulary used to describe motion, especially distinctions of manner of motion, Icelandic outperforms Swedish (Icelandic > Swedish). On the basis of similar frog-story data reported for English (Slobin 1998) we could tentatively position English on the same scale: Icelandic > Swedish > English More research is needed, however, before this hierarchy can be said to be properly consolidated. The difference between Icelandic and Swedish with respect to richness of spatial morphemes in constructions with motion verbs is manifest in all age groups of our study. In contrast, the difference in richness of motion-verb vocabulary emerges in late phases of language development. It is manifest in the fifteen-year-olds and in the adults, but not in the younger age groups. This observation supports the notion that grammaticalized categories tend to be acquired early whereas an important amount of lexical growth takes place beyond childhood. However, since all very frequently used morphemes are grammatical morphemes (but not all grammatical morphemes are very frequently used), the deeper, underlying dimension might have to do with frequency. And, again, frequency also helps to explain why the top ten motion verbs are common to all age groups, whereas less frequent verbs tend to be manifested only in the older age groups. The observed differences in language usage between Icelandic and Swedish narrators of the frog story represent not only differences in form, but also in content. Differences in very frequent categories affect language-specific profiles already at an early age, whereas less frequent categories and items contribute to cross-linguistic differences in later phases of development. And we further observe that content differences which are present at an early stage often level out with age. How should we describe these differences and developments? We propose to talk about them in terms of "privilege of access". This notion is based on the assumption that frequent categories and items are more easily accessible than less frequent ones (both in acquisition and in production). Thus, a user of Icelandic is privileged compared with his Swedish peer when it comes to the linguistic construction of an action as on-going. Icelandic makes the marking of progressivity easily accessible through a highly grammaticalized and very frequent form, BEPROG, whereas the Swedish-language user has to work harder to construct the same temporal profile from partly lexical means (V&V


139

or halla pa att 'keep on V-ing'). And in the domain of talking about space, the Icelandic peers are further privileged in terms of making reference to grounds. Similarly, a child acquiring Icelandic is privileged in that he will be more frequently and systematically exposed to highly grammaticalized linguistic forms associated with the notion of progressivity of actions than will his "disadvantaged" Swedish peer. And contrasting Icelandic and Swedish child language in the frogstory task, we discover a disequilibrium: the Icelandic five-year-olds much more often mark an action as progressive (8.7% of all finite verbs are marked for BEPROG) than their Swedish peers (1.3% of all finite verbs are V&V). Long-term development, however, involves not only the acquisition of new linguistic resources but also the expansion of functions. In the frog-story task, five-year-olds and adult language users do not employ the linguistic marking of on-going action for exactly the same purposes. Interestingly, adult Icelandic subjects end up at 2.2% BEPROG out of all finite verbs and adult Swedish subjects at 2.0% V&V, a perfect equilibrium between the two languages. We interpret this distribution as functionally motivated. The Icelandic and Swedish adults have developed a similar conceptual story representation, including the distinction between backgrounded and foregrounded events, and they seek to map this distinction onto adequate linguistic means. For Icelandic narrators, BEPROG is an easily accessible and adequate means, and its new usage for backgrounding events, together with the withdrawal of its usage for simple picture description, results in a decrease in overall usage compared with the younger age groups. For Swedish narrators, V&V is not as accessible a means, but it is very adequate and so the functional motivation pushes the narrator to employ it. More generally, we conclude that the advanced learner is not necessarily he who merely has acquired the grammatical or lexical forms, but rather he who has an advanced functional design of the narrative and is in command of the linguistic resources necessary to implement that design. The need to model long-term development in terms of an expansion and change of functions is also illustrated by the Icelandic categories of BEPROG and HAPERF. Both categories are encountered in the youngest age group of frog-story narrators, but only in the function of describing on-going versus completed events in individual pictures. Again, it is only later in development that we observe BEPROG in the function of backgrounding events and HAPERF in the function of summing up completed events of relevance to the forwarding of the plot. Linguistic items and constructions in discourse are, as a rule, multifunctional, and the eventual functional equilibrium observed between the Icelandic BEPROG and the Swedish V&V concerned only the function of constructing an action as on-going and backgrounded. The V&V-construction crucially participates in the attainment of that functional equilibrium, but, what is more, it also participates in creating a new functional disequilibrium in terms of specification of body

140


posture. Since the first verb in the V&V-construction is a verb of posture, the Swedish narrators of the frog story incidentally predicate the posture of story characters to a greater extent than their Icelandic peers. The use of V&V to mark an action as on-going carries with it, as it were, a postural "bonus effect". In the domain of lexical development, the opposite tendency is detected, that is, from greater to less cross-linguistic similarity. The same small set of around ten verbs make up the most frequent verbs in all age groups, and the older the age group, the larger the extension to that set. The Icelandic data indicate that the adult group reaches a variety of 67 different verbs, whereas the Swedish adults reach only a modest 47. From Roman Jakobson's visionary ideas, relating children's language acquisition to assumed language universals and particulars (Jakobson 1941), one would predict that maximal cross-linguistic similarity should be found across early child languages and that samples from more advanced learners should exhibit more particular traits as children come to conquer languageparticular details only in later phases of acquisition. Jakobson's tenets were related to observations in the domain of phonology. The basic idea, however, can be extended to grammar, and a number of works on early child language development following the generativist/nativist approach provide an illustration (e.g., Hyams 1986). In contrast, the need to study the cross-linguistic diversity of early language acquisition has been pointed out by many scholars (see. e.g., Slobin (ed.) 1985-1997). We would like to add our voices to that call for research by arguing that children may be more at risk of "falling victim" to languagespecific effects in an early phase of their language development, and that longterm development serves to overcome these initial restrictions as the learner widens his repertoire of linguistic means to expand his expressive capacity and to implement partly new communicative activities and discourse types.

NOTES 1.

This work was supported by grants from the Nordic Research Council (Nordiska samarbetsnamnden for humanistisk forskning, NOS-H) to both authors, and from the Icelandic Research Council to Hrafnhildur Ragnarsdottir. The analyses presented here concerning time and space build on studies previously published in the Proceedings from the Stanford Child Language Research Forum 1996, 1999 (see Ragnarsdottir & Stromqvist 1997; Ragnarsdottir, Stromqvist & Nordqvist 2000). We extend our gratitude to the editor of the Proceedings, Eve V. Clark.

2.

Our description of the prosodic properties of V&V rests on auditory judgment only. Further analyses in our lab will help determine the phonetic detail of the perceived differences.


141

3.

The diacritic signs in the Icelandic examples are not related to stress.

4.

In Icelandic, when an item such as ut occurs in a verb phrase such as hljop ut 'ran out' it is classified as an adverb according to Icelandic grammatical descriptions (see, e.g., trainsson 1990). The construction V + prt with phrasal stress on the particle is reserved for lexicalized (non-compositional) meanings, for example, lita ut 'look/look like'.


The Verbalization of Motion Events in Arrernte1

David P. Wilkins

1. INTRODUCTION At the time of first colonization, there were approximately 250 distinct and healthy languages indigenous to Australia. Now, only about 20 Australian languages are still being acquired by children as a first language. Arrernte2, spoken in the deserts of Central Australia, is one of those 20 viable languages, as is its better known neighbor to the northwest, Warlpiri. While the acquisition of Warlpiri has been studied extensively, primarily through the work of Bavin (1993, 1992, 1991, 1990), there has been very little consolidated work on the acquisition of the other healthy Australian languages, and the study to be reported here represents the first investigation of any kind into the acquisition of Arrernte. In particular, this chapter will examine the construction of complex motion events in Arrernte narratives from both a developmental and a cross-linguistic perspective, and in so doing it attempts to bring together two distinct domains of research - ethnographic and ethnolinguistic research among Aboriginal communities in Central Australia, on the one hand, and crosslinguistic developmental research on the acquisition of narrative rhetorical style, on the other.

2. BACKGROUND AND HYPOTHESES The Desert area of Australia, particularly the Central Desert area where the Arrernte and Warlpiri live, has long been recognized as a cultural area within which a number of different Aboriginal communities share certain cultural and linguistic traits. For instance, it has been common for primarily English-speaking anthropologists and linguists to give emphasis in their descriptions to the observation that Desert Aborigines appear to have a unique and special concern for

144

David P. Wilkins

orientation in space and for giving detailed attention to motion paths and journeys (e.g., Lewis 1976b; Laughren 1978; Glass 1980; Koch 1984; Myers 1986: Munn 1986; Wilkins 1991). The following quotes illustrate this observation: Australian Aborigines are well known as being a nomadic people. As such, travel was an intrinsic part of their daily life. [...] In the desert, in particular, they travelled long distances. It is not surprising therefore that the languages of the Australian Aborigines should have a variety of lexical, morphological and syntactic devices to deal especially with travel. (Glass 1980:123) Warlpiri children of 3, 4 and 5 handle directional terminology (up. down. on. under, hither, thither, here, there, etc.) including the points of the compass and positions and directions relative to the points of the compass with ease and competence. (Laughren 1978) Orientation in space is a prime concern for the Pintupi. Even their dreams are cast in the framework of spatial co-ordinates. (Myers 1986) The concept of locale and journey provides the framework for men's songs and narratives about ancestral events. (Munn 1986:132)

The research into this special concern for travel and the linguistic attention given to motion events has tended to be, at best, only implicitly comparative, and has had no developmental examination. This brings us to the second domain of research, which is explicitly comparative and developmentally oriented, namely the "frog story" work of Slobin and his colleagues (Slobin 1996a, 1991; Berman & Slobin 1994; Slobin & Bocaz 1988). Three points suggested by that research are of particular relevance to this chapter: /.

even nine-year-olds have not fully acquired the rhetorical style of their speech community (Berman & Slobin 1994:593, 599);

//'.

from early in acquisition, learners stick closely to the set of semantic distinctions provided by their language and tend not to improvise distinctions (Berman & Slobin 1994:641);

//'/'.

there appears to be a correlation between the semantic type of a language and its rhetorical structuring of motion events; in particular, it is suggested that Talmy's (1985) distinction between verb-framed and satellite-framed languages correlates with the degree of narrative attention devoted to the dynamics of movement in the description of a journey3 (with satellite-framed languages devoting greater attention to

The Verbalization of Motion Events in Arrernte

145

the dynamic structuring of motion events than verb-framed languages (Slobin 1996a)). Against the background of these two research domains, comparative and developmental hypotheses can be formulated and investigated. On the basis of the ethnographic and ethnolinguistic research, one could try to find support for the following comparative hypothesis (Hypothesis la): The special Arrernte concern for motion and orientation will manifest itself even in elicited narratives and, as a consequence of this, the structuring of motion events in Arrernte will be both qualitatively and quantitatively different from that in English. For example, Arrernte-speakers should deploy spatial language to construct more elaborated paths and journeys than English-speakers.

However, since Arrernte is a verb-framed language (like Spanish) rather than a satellite-framed language (like English), Slobin's (1996a) typological observations suggest a counter-hypothesis (Hypothesis Ib): Since Arrernte is typologically a verb-framed language, we would expect it to have a rhetorical style which devotes less narrative attention to the dynamics of movement than English, which is a satellite-framed language. Arrernte-speakers should deploy spatial language to construct less elaborated paths and journeys than English-speakers.

In short, if there are any differences between the narrative treatment of motion events in Arrernte and English, will they be in line with areal cultural predictions or predictions from linguistic typology? Along with these comparative hypotheses, the chapter will also examine the following developmental hypothesis (Hypothesis 2): Although children around the age of nine are not expected to have fully adult-like narrative event structure, Arrernte children's productions will, nonetheless, look more like those of Arrernte adults than those of English-speaking children.

3. THE FROG-STORY TASK AND SUBJECTS To investigate the above hypotheses, elicited narratives were collected following the procedures outlined by Berman and Slobin (1994:20-28): /.

subjects are shown a copy of the wordless picture storybook entitled Frog, where are you? by Mercer Mayer (1969; see Appendix I);

146

David P. Wilkins

//.

subjects are told in advance that they are going to be asked to tell a story (the story that is shown in the book);

/'/'/'.

subjects first go through looking at the book, picture-by-picture;

iv.

after looking through the book once (to see what the story is about), the subject is asked to tell the story to the investigator, while again going through picture by picture from beginning to end;

v.

the investigator works with each subject individually, and makes an audio-recording of the story; and

v/.

the investigator keeps track of which picture(s) the narrator is attending to when relating a certain portion of text.

This task was undertaken with six children and six adults residing in Alice Springs. The six Arrernte children were all students in the Central Arrernte primary class of the Yipirinya School (an Aboriginal-controlled bilingual and bicultural school). There were two seven-year-olds, two eight-year-olds, one nine-year-old, and one ten-year-old. While Arrernte is the first language of these children, all are also speakers of some variety of English. The Arrernte children's data were collected in September 1993 by a Central Arrernte person. Carmel Ryan, the children's teacher. She was instructed in the procedure by the author, and she herself identified the best means for conveying the instructions in Arrernte. The data for the six Arandic adults (aged between 25 and 55) were collected by the author in the period 1992-1995. There are four distinct (mutually intelligible) Arandic varieties represented in the adult group: three Central Arrernte speakers, one Eastern Arrernte speaker, one Western Arrernte speaker, and one Alyawarr speaker. Like the children, all the adults also speak some variety of English, and most also speak another Central Australian language. In what follows, the elicited Arrernte narratives will be compared with the American English narratives that were used for analysis in Berman and Slobin (1994) and Slobin (1996a). Dan Slobin kindly provided the transcripts of these American English frog stories, and the comparison here will involve the twelve nine-year-olds and the twelve adults. The investigation will focus on two narratively significant 'journeys' from the frog story: 'the journey from the jar' and 'the journey off the cliff (Pictures 1-3 and 15-18, respectively, in Appendix I).


147

4. SOME FACTS ABOUT ARRERNTE GRAMMAR Since this study is concerned with identifying how spatial language is deployed to render complex motion events (journeys) within narratives, there are three aspects of Arrernte grammar that are relevant to the discussion. First, Arrernte has an extensive system of case marking involving fourteen distinct cases which are marked on the final element of the NP (Wilkins 1989). For current purposes, the three most relevant spatial cases are: -nge ablative (ABL), -werne allative (ALLAT), and -ke dative (DAT). With a motion verb such as tnye- 'fall', the allative case would only indicate motion towards the goal, without entailing that the goal is reached, while the dative case would indicate motion to the goal, entailing that the goal is reached. Second, in Talmy's (1985) terms, Arrernte is verb-framed (like Spanish) rather than satellite-framed. That is, it is the motion verb that tends to encode path information, rather than a satellite to the motion verb. Thus, for example, it typically uses monomorphemic path-incorporating verbs which translate 'enter' and 'exit' rather than complex phrases like 'go in' and 'go out'. English, by contrast, is satellite-framed. Third, a particularly interesting feature of the verb morphology is a distinct slot for an elaborate category of inflections called the category of associated motion, which is used to indicate that the verbstem action happens against the background of a motion event with a specific orientation in space. These inflections have meanings like 'do verb stem act while going upwards', 'do verb stem act while moving past', 'do verb stem act and then start to go off, etc. (Wilkins 1991, 1989). It must be stressed that such inflections conflate motion and path, that they can attach to any verb stem (e.g., 'sit', 'eat', "know', 'sing', 'fall', etc.), and that they fall outside Talmy's 'verb-framed' versus 'satellite-framed' typology.

5. THE 'JOURNEY FROM THE JAR' In this section, the motion descriptions which accompany picture 2 (see Appendix I) of the storybook are discussed.4 The preceding picture had shown a boy's bedroom in which a boy and his dog were looking at a frog in ajar. In the picture we are concerned with, the boy and dog are asleep, and the frog is depicted in the midst of raising himself out of the jar (just one leg is out of the jar). The following picture shows the boy and dog awake looking at the empty jar. In all three pictures, the bedroom window is partly opened. Examples of the motion description are given in [1-4]: [1 ]

... and the frog is getting out of the jar. (English child 9D)

148

[2]

[3]

David P. Wilkins

The frog gets out of the bottle and escapes through the open window. (English adult 20L)

..., kenhe

nhenhe

frog-frog

nhenge

mameye

..., CONJ

here

frog

DEM

mother

ikwere-nhe

mape-werne

alpe-rlenge

3SG-GEN

mob-ALLAT

gO.back-DIFF.SLBJ 5

'..., while here, that frog is going back to his mother and family' (Arremte child AR, 8 yrs)

[4]

..., antyetyerre re

kenhe

..., frog

CONJ

BSGISL'BJ

arrate-rle.lhe-me-le exit-DO&GO 6 -PRES-SAME.SUBJ

alwirre-me-le

alhe-rlenge

run.away-pRES-SAME.suBJ

go.away-oiFF.suBj

'..., while the frog got out (of the jar) and started going off, fleeing away, leaving (the scene)' [lit. 'go off having fled, having exited and started moving off.'] (Arrernte adult MM)

As the above examples show, it is possible to identify two distinct path segments in the description of the motion scene. The first segment can be labelled frogleaves-jar and the second can be \abe\\edfrog-departs-scene. Examples [2] and [4] show an English adult and an Arrernte adult who encode both path segments in their telling (i.e., 'frog leaves jar and then departs scene'). As a contrast, the English child in [1] only encodes the first path segment frog-leaves-jar, and the Arrernte child in [3] only encodes the second path segment frog-departs-scene. Considering all the storytellers, it is possible to compare the different age and language groups with respect to how common it is to code one or other or both of these two path segments in describing the motion event associated with the second picture of the storybook. The results of this comparison are given in Table 1. As the table shows, all but one of the English children overtly mention that 'the frog leaves the jar', while only a third of them mention that 'the frog departs the scene'. This is in stark contrast to the Arrernte children, all of whom overtly encode that 'the frog departs the scene', while only one child also encodes that 'the frog leaves the jar'. This suggests that, in Berman and Slobin's

149


Table 1. Group comparison of segment coding for the 'journey from the jar' Path segment

Example

1 .frog-leaves-jar t— \

'exit', 'get out of, 'emerge from', 'escape from'

11 92%

11 92%

1 17%

6 100%

'rush away', 'run away to', 'get away', 'escape', 'leave', 'return to'

4 33%

6 50%

6 100%

6 100%

ffl 2 . frog-departs-scene

English children n=12

English Arrernte Arrernte adults children adults. n=12 n=6 n=6

terms (1994:613), the Arrernte children's narratives show a different channeling of attention in comparison to those of the English children. Looking at the adults, Table 1 shows that English adults were very similar to the English children - once again, all but one of the twelve encode the first path segment in their telling, and the number of storytellers that encode the second path segment increases from 33% to 50%. The Arrernte adults are the only group who, without exception, overtly encode both path segments in their narrative. One thing that Table 1 does not reveal is the fact that Arrernte adults do indeed show a similar channeling of attention to that of Arrernte children. Whenever an English-speaker encoded both path segments for this journey, it was always through a sequence of simple conjoined clauses. However, 83% (5/6) of the Arrernte adults encode the description of the second path segment (frogdeparts-scene) as the head verb (primary clause), upon which the description of the first path segment (frog-leaves-jar} is dependent. In other words, the Arrernte children all encode the one segment which the Arrernte adults treat as the head, and they overwhelmingly neglect the segment which Arrernte adults treat as the dependent clause. As Slobin and Berman (1994:153) observe, one developmental trend in the English encoding of this motion event is the fact that no English child selects the lexical item 'escape' in his or her description, whereas 58% of the adults do. It is worth pointing out, however, that 25% use 'escape' to describe the first path segment ('escaped from the jar') and 33% use it to describe the second path

David P. Wilkins

150

segment ('escaped through the window'). Similarly, while 66% of the Arrernte adults use the associated-motion inflection -He. I he 'do verb stem act and then start to go off' (e.g., 'exit and start to go off' as in Example [4]), none of the Arrernte children uses this or any other associated-motion inflection.

6. THE * JOURNEY OFF THE CLIFF' In this section, the narrative descriptions given for Pictures 15-18 (see Appendix I) of Frog, where are you? are considered. These pictures depict a 'journey' which is given close scrutiny in Slobin's (1996a) investigation of the rhetorical structuring of motion events in Spanish and English, and which is called by him 'the scene of the fall from the cliff'. Examples [5] and [6], from an English and Arrernte child, respectively, give some sense of this journey. [5]

Then there was a deer living behind the rock and he got caught, (pet. 15) and then ran and ran and Ralph was right in front of them. (pet. 16) Then the deer bucked him off (pet. 1 7) and he fell into the pond with the rocks, (pet. 1 8) (English child 9J)

[6] Then then Kele O.K.

thing thing

ine-ke,

alhe-ke

3SG-ACC

get-PAST

gO-PAST

re thing-ke SSGISUBJ thing-DAT

antye-ke ride-PAST

re 3sc:suBj

gO-PAST

itere-werne, side-ALLAT

kngwelye dog

Re 3sc:suBj

re-nhe

Ihe-ke

nhenge

tnye-ke fall-PAST

Re first, 3so:stJBJ first,

anteme now anteme now

kweke little anteme now

and then and then

kwele. QIOT

kngwelye dog

too too

(pet. 15)

kwele QLOT

(pet. 16)

kwele

ante

QUOT

and

(pet. 17)

Kngwelye

anteme.

dog

now.

tnye-me fall-pRES

itere-nge. side-ABL

Kele O.K.

kngwelye-0 ikwere kertne, re-therre kwatye-ke tnye-tyalpe-ke. (pet. 18) dog-s 3so:DAT top, 3Du:suBJ water-DAT fall-return-pAST

The Verbalization of Motion Events in zArrernte

Schematization of the 14 path segments of the

151

'cliffjourney'

'Then something got him [the boy], and went. (pet. 15) So he rode the thing now towards the edge, he went now, and so did the dog. (pet. 16) He fell now, apparently. The dog now. Him first, and then the dog are falling from the edge. (pet. 17) So the dog is on top of him, and the two of them fell back down in the water, (pet. 18)' (Arrernte child JR, 9 yrs)

In such descriptions, the continuous path trajectory of the boy is the focus of interest - from where the boy is picked up by the deer to where the boy lands in the water. In examples such as [5] and [6], it is possible to identify and abstract out the linguistically coded path segments. In [5], the English-speaking child builds the journey using four distinct path segments - the deer goes off with the boy ("ran and ran"); the boy moves from the deer ("the deer bucked him off); the boy moves downwards ("he fell"); and the boy moves to be in water ("fell into the pond"). In [6], the Arrernte child encodes three of the same path segments as the English child, but instead of encoding that 'the boy moves from the deer', she uses another distinct path segment 'the boy moves from the cliff' ("Him first, and the dog are falling from the edge"). Taking all 36 narratives together, it is possible to identify 14 distinct path segments each of which receives overt linguistic encoding in at least one of the narratives. These are presented in Figure 1 (and Table 3). There is no claim here that this is the logical number of segments out of which the path is constructed, only that these segments emerge from the Arrernte and English data examined. In this, the segmental comparison is different from that of Slobin (1996a). It is easy to imagine how other languages (or other narrators) could break the path into further segments. Moreover, a similar two-language comparison may identify fewer (or different) segments.

152

David P. Wilkins

Figure 2. Comparison of mean number of segments in journey off the cliff.

If English-speakers and Arrernte-speakers are different with respect to the degree of narrative attention they devote to the dynamics of motion events, then this should be reflected in differences in the number of path segments the speakers of each language string together in building their account of the 'journey off the cliff'. Figure 2 shows that Arrernte adults average 7.8 path segments in their description of this journey. This is roughly three path segments more than English-speaking adults, who show a mean of 4.7 segments. English adults average approximately one segment more than English-speaking children, who show a mean of 3.8. The Arrernte children's mean of 4.5 segments is down around the English-speakers, but is much closer to the English adult average than it is to the English-speaking children's average. Thus, Arrernte adults appear to build much more elaborated 'journeys' than any of the other groups, and Arrernte children are much further from the adult target than English-speaking children are. However, the Arrernte children do appear to build slightly more complex journey descriptions than the English children. These differences may be easier to detect in Table 2, which reveals how many segments are encoded by individual narrators. While both of the child groups show a range of three to five segments, most of the English children build paths of four segments, while most of the Arrernte children build paths of five segments. English adults show a range of four to six segments, but half of the adults build paths of five segments (much like the Arrernte children). By contrast, the Arrernte adults have a range which begins where the English adult range ends (i.e., six segments) and which goes all the way up to nine distinct path segments. In fact, half the Arrernte adult group encode nine path segments in their narration of the 'cliffjourney'.


153

Table 2. Group comparison of the number of segments used for the 'cliff journey' by different narrators, English children and adults versus Arrernte children and adults Number of segments

English children n=12

English adults Arrernte children n= 12 n=6

3

3 25%

1 17%

4

8 67%

5 42%

1 17%

5

1 8%

6 50%

4 66%

6

1 8%

Arrernte adults n=6

2 33%

7 8

1 17%

9

3 50%

A further question is whether the groups differ in terms of which of the 14 path segments they are most likely to encode in their narratives. Table 3 gives this comparison. Note that, while all 14 path segments are needed to account for the Arrernte data, only 10 segments would have been needed to account for the English data on their own. That is, the number of path segments emerging from the English narratives is a proper subset of the number that emerges from the Arrernte narratives. In fact, there are three path segments which are only found in adult Arrernte narratives - number 3, 'move past viewer'; number 9, 'boy move toward the ground'; and number 12, 'boy/dog move past cliff. Interestingly, each of these three path segments is instantiated through the use of'associated motion' inflections, and all six speakers encode one or other of these three segments. Further, it should be pointed out that all Arrernte adults mention the path segment 'start to go', while none of the Arrernte children overtly codes this, and only one English child and three English adults include this segment in their telling. Of the six Arrernte adults, five render this particular segment using the 'associated motion' inflection meaning 'do act and then start to go off (e.g., ine-rle.lhe-ke get-DO&oo-PAST 'got and then started to go off). The fact that all Arrernte adults encode a 'start to go' component seems to correspond to another regularity that is revealed in Table 4. The majority of

154

David P. Wilkins

Table 3. Group comparison of journey segments that receive overt coding, English children and adults versus Arrernte children and adults Journey segment

Example

1 . start to go

deer gets him and starts to go; the deer takes off with him

Eng. ch. n= 12

Eng. ad. n= 12

Arr. ch. n=6

1 8%

3 25%

0

8 67%

8 67%

6

6

100%

100%

Arr. ad. n=6

6 100%

2. go (towards the cliff)

the deer carries the boy; the boy rides the deer; the boy sits while going

3. move past viewer

he lies while moving past

0

0

0

•> 33%

4. approach cliff

they're approaching a cliff

0

1 8%

1

0

1 7%

3 25%

4 33%

the deer just stopped; the deer put the brakes on

1 8%

3 25%

1 7%

1 17% •> 33%

7. boy move from deer deer leaves boy behind; boy comes off the deer; the deer bucks the boy off

6 50%

7 58%

•) 33%

100%

8. boy move downwards

the boy falls; the deer drops the boy; thrown downwards

8 67%

10 83%

6

6

100%

100%

9. boy move toward the ground

deer brings him down to ground

0

0

0

10. boy/dog move from cliff

the deer has thrown the boy over the cliff; the boy and dog fall off the ledge

1 1 . boy /dog move waterwards

the two of them fell down towards the water; they're heading for a pond

5. go be at cliff

the deer ran to a cliff; the 'demon' arrived at the cliff

6. deer stop (at cliff)

1

6

i 33%

12. boy /dog move past so the child fell past the (body cliff of) the cliff towards the water 13. boy /dog move to be in/at water

they both fall into some water; the boy and dog land in the water

14. boy/dog be in water

the boy was lying in the water

Mean number of journey segments

0

7 58%

7 58%

3 50%

4 66%

0

1 8%

1 17%

4 66%

0

0

0

i 33% 5 83%

12

12

6

100%

100%

100%

0

0

1

1

1 7%

1 7%

3.8

4.7

4.5

7.8

(46/12)

(56/12)

(276)

(476)


155

Table 4. Group comparison of the point of departure for the 'journey', English children and adults versus Arrernte children and adults Picture No..

English children n= 12

15

0

16 17

English adults Arrernte children n= 12 n=6

Arrernte adults n=6

0

3 50%

5 83%

12 100%

10 83%

3 50%

1 17%

0

2 17%

0

0

Arrernte adults (83%) start their narration of the actual journey (i.e., give their first translocational-motion encoding) when their attention is on Picture 15 in the storybook. No English-speaker, child or adult, starts his or her motion account when his or her attention is on this picture; in fact, the vast majority give their first motion description when their attention is on Picture 16. In Picture 15, the deer is standing behind the rock with the boy on its head. The Arrernte adults overwhelmingly choose to see this as the logical prelude to the translocationalmotion event that will follow and indicate that the deer (and boy) 'start to go' at this stage. Importantly, even though none of the Arrernte children encodes a 'start to go' path segment, 50% of them do use their first translocational-motion verb in the account of this journey when their attention is on Picture 15 (see Example [6] above). In short, the Arrernte children once again show a similar channeling of attention to the Arrernte adults. In this case, they identify the rhetorically 'correct' place to begin the account of the 'journey' even if they do not use the same linguistic means.

7. DISCUSSION Regarding the comparative hypotheses outlined in Section 2, we do see qualitative and quantitative differences between speakers of Arrernte and English with respect to the construction of motion paths. Against the typological prediction (i.e., Hypothesis Ib), speakers of Arrernte, a verb-framed language, build more complex motion paths than speakers of English, a satellite-framed language. Thus, it is the areal ethnographic observations that informed Hypothesis la which here appear to be more predictive of the findings. Of course, the fact that Arrernte adults build more complex motion descriptions is likely to be closely connected

156

David P. Wilkins

to the availability of another 'motion-dedicated' linguistic resource which lies outside the satellite-framed versus verb-framed typology, namely the associatedmotion inflections described in Section 4. With respect to Hypothesis 2, the developmental hypothesis, we can say that, although Arrernte children may not be closer to Arrernte adults than they are to English-speakers in terms of the complexity of the paths they build, they do show some evidence of greater motion-event complexity than the English-speaking children (see Section 6). More significant, however, is the fact that the Arrernte children show unique adult-like channeling of attention which is very different from their English-speaking counterparts. Berman and Slobin (1994:641) correctly observe that each language shapes its own world of expression, and children are sensitive to that, even if they are not masters of it. However, exactly how they become sensitized to the rhetorical predispositions of their speech community remains an open question. For instance, the Arrernte data show that it is not a simple matter of verb-framed versus satellite-framed typology which predicts whether speakers rhetorically code journeys with more or fewer path segments. In fact, I would predict that speakers of Central Australian desert languages would behave in essentially the same way, independently of language type, owing to areal cultural factors. Moreover, while Arrernte children do not always use the same spatial language and patterns of event construction as adults, they still demonstrate certain adult patterns of'thinking-for-speaking', and it seems reasonable to question whether language is the only contributor to this.6 Indeed, Berman and Slobin (1994:594) suggest an interesting possibility when they write: ... in the course of development, the child learns to use the expressive options of a particular native language to carry out general discourse functions. Further, those general functions are, to some extent, shaped and channeled by features of individual languages (and- though not considered here - by cultures as well}. (emphasis mine - DPW)

We are left to ponder, then, what is the relative contribution of linguistic typology versus culture in sensitizing children to the narrative interests and rhetorical style preferred by the adult members of their speech community. This chapter does not pretend to address that question, only to raise it.

NOTES 1.

Data collection was supervised and supported by the Yipirinya School, for which I am grateful. Carmel Ryan, in particular, is to be acknowledged for her part in the collection and the analysis of the Arrernte children's data. Margaret Heffernan's help and support in the field was invaluable, as was that of Robert Hoogenraad and Penny


157

Evans. I would also like to thank Shanley Allen, Melissa Bowerman, Eve Danziger, Ken Drozd, Dan Slobin, and Barbara Villanova for helpful advice with respect to the research, analysis, and/or data presentation. 1 would like to acknowledge the MPI for Psycholinguistics for funding my field research during the period of this research. This chapter initially appeared as a paper in Proceedings of the twenty-eighth annual child language research forum (1997), edited by Eve Clark, and is little changed from that version. It was included in the current volume at the request of the editors and I am grateful to them for the improvements they have made. 2.

Until recently, Arrernte was most commonly rendered as 'Aranda'.

3.

This chapter uses Slobin's (1996a) notion of journey, which refers to a complex path that is built up from a series of linked paths, or is an extended path with subgoals.

4.

For people who do not have access to Mercer Mayer's Frog, where are you?, the pictures are reproduced in Appendix I.

5.

Arrernte has switch-reference: SAME.SUBJ = same subject (as matrix clause); DIFF.SUBJ = different subject (from matrix clause).

6.

DO&GO reads 'do verb action and then go away'.

7.

In Central Australia, narratives of all sorts tend to be supported and augmented by two other semiotic systems, sand drawings and absolutely-oriented gestures. Either system, or both together, could draw a child's attention to certain aspects of the rhetorical structuring of narrative and lead him or her to hypotheses concerning the critical features of adult-like 'thinking-for-narrating'.


A Third Way to Travel The Place of Thai in Motion-Event Typology

Jordan Zlatev & Peerapat Yangklang

1. INTRODUCTION It has become almost impossible to consider motion-event descriptions, especially from a typological perspective, without starting from the seminal work of Talmy (1985, 1991, 2000), and its application to narrative and to the issue of linguistic relativity by Slobin (1996a, 1997a, 1997b, 2000), as also testified by many of the contributions to this volume (cf. Galvan & Taub, Ibarretxe-Antunano, Brown, Ragnarsdottir & Stromqvist, Bavin, and Wilkins). To recapitulate very briefly, in order to avoid unnecessary repetition: holding the expression pole of the linguistic sign (cf. Saussure 1916) constant, languages can be divided into path-type, manner-type, and figure-type, depending on which semantic category is preferably lexicalized in the root of their motion verbs (Talmy 1985; Slobin & Hoiting 1994; Zlatev 1997).' On the other hand, starting from the semantic pole, and the highly abstract semantic-conceptual notion of core schema ("the association function that sets the figural entity into a particular relationship with the ground entity", Talmy 2000b:218), which in the case of motion events corresponds above all to the category of Path, languages fall into one of only two categories depending on which form class preferably expresses this notion: the verb root, or the "satellite" - a verb prefix, verb particle, or some other kind of (subordinate) element in close association ("in a sister relation") with the verb root. This distinction is supposed to cut a line through the world's languages, as stated by Talmy: Languages that characteristically map the core schema into the verb will be said to have a framing verb and to be verb-framed. Included among such languages are Romance, Semitic, Japanese, Tamil, Polynesian, most Bantu, most Mayan, Nez Perce, and Caddo. On the other hand, languages that characteristically map the core schema onto the satellite will be said to have a. framing satellite and to

160


be satellite-framed languages, and included among them are most Indo-European minus Romance, Finno-Ugric, Chinese, Ojibwa and Warlpiri. (Talmy 2000b:222: author's emphasis)

While remaining very influential, Talmy's "two-category typology" (cf. IbarretxeAntunano, this volume) of verb-framed versus satellite-framed languages has also proved to be problematic. For example, as Brown (this volume) parenthetically notes, Tzeltal, a Mayan language, has both Path-expressing verbs and directionals. where the latter could be classed as "satellites", especially given the rather vague definition of that concept. Thus, Mayan would appear to belong to both sides of the two-category typology. Particularly problematic are languages with verb serialization, where two or more verbs, with or without arguments, co-occur in the same clause, apparently expressing the same event, as illustrated by the Thai single-clause sentence [1]. which includes three different kinds of motion verbs: the Manner verb desn 'walk', the (non-deictic) Path verbs khaam 'cross' and khaw 'enter', and the Deictic verb paj 'go'. [1]

chan desn khaam thanon khaw paj naj suan I walk cross road enter go in park 'I walked across the road and into the park.'

Such so-called serial-verb languages are widespread in a variety of language families, such as Niger-Congo, Hmong-Mien, Sino-Tibetan, Tai-Kadai, MonKhmer, and Austronesian, and have attracted considerable attention from typologically oriented linguists (e.g., Stahlke 1970; Thepkanjana 1986: Chuwicha 1993; Bisang 1995; Durie 1997; Senft 2001). With regard to motion-event typology, Talmy (1991, 2000b) classifies at least two serial-verb languages, Chinese and Lahu, as satellite-framed. On the other hand, Slobin and Hoiting (1994) suggest that "serial-verb languages be reclassified as complex verb-framed types" (ibid:502), and in a recent publication Slobin (2000) remarks that "serial-verb languages like Chinese may represent a third type of lexicalization pattern, lying between S-languages and V-languages" (ibid: 134). In a recent paper, Ameka and Essegbey (in press) pose the crucial question already in the title, Serializing languages: Satellite-framed, verb-framed or neither? Based on an analysis of motion events in Ewe and Akan (NigerCongo), they conclude that these languages share some features of both types and therefore cannot be said properly to belong to either. In this chapter, we will address Ameka and Essegbey's question from the standpoint of (Standard) Thai. In Section 2, we will describe the Thai translocative serial-verb construction, exemplified in [1], and draw preliminary

A Third Way to Travel

161

conclusions from such "competence" data, i.e., from analysis based on the grammatical intuitions of native speakers. At first glance, it appears that Thai should belong to the verb-framed category since the "core schema" is expressed in Path verbs such as khaw 'enter' and ?ook 'exit'. On closer inspection, however, this conclusion turns out to be doubtful since these verbs can be accompanied by preceding Manner verbs. There is no syntactic or semantic evidence that either type is subordinate to the other, and therefore they need to be given equal status. Nevertheless, as increasingly practised in linguistic typology (e.g., Slobin 1996a: Pederson et al. 1998), it is not enough to recognize the presence of a particular structure in a language - it should be investigated to what extent this structure is actually used in discourse and how it correlates and interacts with other discourse features in the formation of a rhetorical style (Slobin 1996a). To satisfy this requirement, we used the frog-story framework (Berman & Slobin 1994), with certain adaptations for Thai, called for by its status as a serial-verb language, as we will describe in Section 3. Thus, properly armed with data from linguistic analysis and narrative production, we will turn in Section 4 to the two versions of Talmy's typology - the one concerning lexicalization patterns (Talmy 1985, 2000a) and the one concerning "event integration" (Talmy 1991, 2000b) - and consider the place of Thai in a general typology of motion events. In Section 5, we will look more specifically at some of the features shown by Slobin to distinguish typical satellite-framed languages ("S-languages"), such as English, from typical verbframed languages ("V-languages"), such as Spanish, on the discourse level, namely: /. //. iii. iv.

higher level of Manner-verb use; higher level of specification of Ground2; higher granularity of event encoding; fewer static descriptions and less "scene setting".

Finally, in Section 6 we will summarize and conclude that in accordance with the proposals of Slobin (2000) and Ameka and Essegbey (in press), Thai - and by extension other serial-verb languages - cannot be properly categorized as either verb-framed or satellite-framed. Furthermore, we will argue that these languages do not take a position that is intermediate between the two, as suggested by Slobin and Hoiting (1994), but rather instantiate a "third type", which in turn calls for a radical extension and/or revision of the currently dominant motion-event typology.

162


2. TRANSLOCATIVE MOTION IN THAI: VERBAL EXPRESSION 2.1 Conceptual foundations Work in motion-event typology, which inevitably involves comparisons between languages, as well as different authors' descriptive frameworks, has been hampered by a lack of clear definitions. In particular, there seems to be some confusion as to what actually is a "motion event" and how: to decide in a systematic way what semantic category is being "conflated" by different expressions. Talmy (1985) is somewhat to blame for this, since in this highly influential paper these terminological issues were left unclear, but the recent revised version (Talmy 2000a) is more explicit with respect to the basic concepts. A Motion event (with a capital M) is defined as "a situation containing motion and the continuation of a stationary location" (ibid:25). This very general notion is further divided into translational motion, where "an object's basic location shifts from one point to another in space", and self-contained motion, where "an object keeps its same basic, or 'average' location" (ibid:35). What makes distinguishing between these two categories problematic is that many Mannerof-motion verbs such as float and kick can be used in both a "translational" [2a] and a "self-contained" context [2b]. [2a] He kicked the ball over the wall. [2b] He kicked the wall. Talmy tries to account for this variation in terms of polysemy. For example, in [2] there would be two different senses of kick involved, where the first is translocative' and the second locative. But this is clearly problematic, and even Talmy admits that a "constructional approach" (Goldberg 1995) may be preferable. We will assume precisely such an approach and regard the meaning of verbs such as kick as semantically general, rather than polysemous. even though the distinction between polysemy and generality may sometimes be difficult to draw (cf. Geeraerts 1993; Zlatev, in press). We will therefore treat as potentially translocative not only clear Path verbs such as khaw 'enter' but also verbs expressing an activity that can lead to a change in location: both intransitive verbs such as kradoot 'jump' and transitive ones such as plak 'push'. On the other hand, when working with production data - as we do with the Thai frog stories in later sections - the decision on whether to code a particular use as translocative or locative can (and will) be made based on the actual context. In dealing with the rather ill-defined category of Manner, we will have in mind a rather general notion, corresponding to Talmy's (2000b) "co-event".


163

which can be subcategorized depending on the various "relations" it stands in with respect to the translocative event, where Manner-of-motion in a narrow sense is but one. Other types of Manner in the broad sense involve Cause, Speed, Vehicle, and Means. However, we will refrain from using the term "co-event" since we consider it highly aprioristic to regard Path (or more generally the "core schema") as the "main/framing event", irrespective of whether it is encoded by a (main) verb, as in V-languages, or by a (subordinate) particle or affix, as in S-languages. While it may seem reasonable to apply this terminology to typical V-languages such as Spanish and French, where the "co-event" is expressed by an optional adverbial phrase, it seems to us highly forced to treat the Manner verbs in, e.g., Germanic languages as expressing a "co-event", leaving the "framing event" for the verbal particle. Similarly, we find the distinction "main event" versus "co-event" inapplicable to the Thai translocative serial-verb construction, where, as we will argue, Path and Manner verbs have equal status. 2.2 (Non-deictic) Path verbs At first glance, Thai seems to be a fairly typical V-language, similar to the Romance languages in that it has a considerable number of Path verbs, e.g., khaw 'enter', Took 'exit', laaj 'pass', klap 'return', phaan 'pass', khaam 'cross', khiun 'ascend', lorj 'descend', thooj 'back off', joon 'reverse', trorj 'go straight', won 'circle', thaam 'follow'. Each one of these can appear as the sole verb in a sentence, for example as an answer to the question tham ?araj 'what am/is/are/ was/were I/you/he/she/we/they doing?', as in [3] and [4]. [3]

chan khaw noon, I enter room 'I went into the room.'

[4]

chan ?ook caak hoot] I exit from room 'I went out of the room.'

Several verbs from this group can combine in a serial-verb (henceforth "SVC"), as shown in [5]. [5]

chan won joon klap khaw hoot] I circle reverse return enter room (a) (b) (c) (d) 'I returned circling back into the room.'

construction

164


Thepkanjana (1986) makes the interesting generalization that when several Path verbs combine in an SVC, their order is a) geometric shape of the path, b) direction with respect to the preceding path, c) direction with respect to an object, and d) "direction resulting from the interaction between the path and the outside world" (ibid: 136); each of the these four categories is instantiated by one verb in [5], as indicated by the letters in parentheses. On the other hand. Muansuwan (2000) argues that no definite ordering between Path verbs in the SVC can be established. Our position is intermediate: We believe that there are indeed constraints on ordering, resulting in the ungrammaticality of, e.g., [6]. On the other hand, the ordering is hardly as strict as Thepkanjana (1986) proposes, since, e.g., [7] is possible. Furthermore, Thepkanjana's distinction between c) and d) is not clear. The issue is definitely interesting and in need of further analysis, but as it is not directly relevant to our main argument, we leave it at this for the present. [6]

*chan I

khaw joon noon, enter reverse room (d)

[7]

(b)

chan won klap joon khaw noon, I circle return reverse enter room (a) (c) (b) (d) 'I returned circling back into the room.'

2.3 Deictic-Path verbs On the other hand, it is clear that when one of the two Deictic-Path verbs paj 'go' and maa 'come' appears in a translocative SVC, it always appears in the final position of the verb series, as in [8]. If the deictic verb precedes a motion verb as in [9], that motion verb is not part of the translocative construction proper, and arguably not part of the same clause. The whole construction, of which the translocative construction is but a part, is a purposive one (Thepkanjana 1986). involving two clauses rather than one, with a semantic difference reflected by the different English translations of [8] and [9]. [8]

chan klap

khaw paj/maa naj

I return enter go/come 'I came back into the room.' [9]

inside

[[chan klap pay/maa]trans khaw I return go/come enter 'I came back [in order] to enter the room.'

hoot] room

(naj) (inside)

h33rj]purp room


165

The Path expressed by paj and maa is oriented not only with respect to the Deictic Center4 (DC) - "away from" and "towards", respectively -, but also with respect to the landmark, when this is expressed by a nominal. In this latter case, the Path expressed is "allative": the end point of the motion event is the reaching of the landmark. 2.4 Manner verbs The analysis presented so far reinforces the impression that Thai is a V-language, or at least a path-type language, as actually claimed by, e.g., Weinold (1995). However, a first indication that this conclusion may be problematic comes from the fact that Thai also has a large number of verbs expressing Manner of motion. This includes not only the relatively small set typical of S-languages, including verbs such as daan 'walk', wirj 'run', piin 'climb', khlaan 'crawl', klirj 'roll', and kradoot 'jump', but also verbs making rather fine semantic distinctions, typical of Manner-type languages, and correspondingly of S-languages, e.g., kracoon 'leap', jorj 'tiptoe', kaaw 'stride', khajerj 'limp', liuaj 'creep' (of snake-like animals), luj 'wade', biurj 'rush', nee 'parade', and khajap 'bulge/extend'.5 To these one could easily add other more general Manner verbs such as daenthaarj 'travel'6, haen 'travel by air', boj 'float', and khii 'ride'. It should be noted that all these verbs express intransitive motion in the sense that even when they can take a direct object (as with, e.g., khii 'ride'), it is the subject of the sentence that is the focal moving entity, referred to as the figure (Talmy 1985) or trajector (Langacker 1987). Manner verbs can appear alone as the sole verb of the sentence, or followed by a deictic verb [10]. At least in the latter case they clearly express translocative rather than simply "self-contained" motion. [10] chan daan I walk

(paj) go

'I am walking (away from DC, towards something).'

Even though they can appear alone, Manner verbs and Path verbs often combine with Path verbs in a translocative SVC, always in the order Manner verb > NonDeictic-Path verb > Deictic-Path verb (cf. Thepkanjana 1996; Muwansuan 2000; Zlatev 2003), as in [11]. The sequence of Path verbs can consist of at least five different verbs, as in [12]. [11] chan daan khaw paj I walk enter go 'I am walking in (away from DC, into something).'

166


[12] chan daan won klap joon khaw I walk circle return reverse enter 'I am walking in a circle, returning back inside.'

paj go

It is sometimes claimed, e.g., by Sak-Humphry, Indambraya, and Starosta (1997), that when a Path verb follows a Manner verb in examples such as [11], it is not really a verb but a "deverbal adverb" of some sort, "homophonous" with the corresponding true Path verb. If this is true, then, e.g., khaw, 'enter' should be treated as an adverb, and therefore as a "satellite", in [11], and possibly even in [12], but as a verb in [3]. We find such an analysis unconvincing for three reasons. First, there seems to be no semantic difference between the "truly verbal" (e.g., Example [3]) and the adverbal (e.g., Example [11]) "homophones", as would be expected if the latter were the result of a process of grammaticalization. Second, it is not clear how such analyses apply to cases such as [12] where there are several Path verbs following the Manner verb. It would be rather implausible to claim that these are all deverbal adverbs in this context but not in [5], which is identical save for the presence of a preceding Manner verb. Third, and perhaps most importantly, all verbs in the translocative SVC, individually [13-15] or together [16], are within the scope of negation, which, when preceding verbal phrases, is marked with the negative marker majdaj. [13] khaw majdaj dean khaw paj, khaw wTrj khaw paj he/she NEC walk enter go he/she run enter go '(S)he is not walking in (away from DC) - (s)he is running in.' [14] khaw majdaj daan khaw paj, khaw daan Took he/she NEC walk enter go he/she walk exit '(S)he is not walking in (away from DC) - (s)he is walking out.'

paj go

[15] khaw majdaj daan khaw paj, khaw daan khaw maa he/she NEG walk enter go he/she run enter come '(S)he is not walking in, away from DC - (s)he is walking in, towards DC." [16] khaw majdaj daan khaw paj, khaw wtrj Took maa he/she NEG walk enter go he/she run exit come '(S)he is not walking in, away from DC - (s)he is running out. towards DC." While none of the three reasons provided (i.e., synonymy. Path-verb sequences without Manner verbs, wide negation scope) is conclusive in itself, their combination makes it highly implausible to argue that Path verbs following Manner verbs are actually "satellites".


167

2.5 Manner + Path verbs The perceptive reader may have noticed that among the examples of Manner and Path verbs given above there were no verbs corresponding to fall and sink, which have been treated differently by different authors. Slobin (1996a) regards fall as a Path verb, since its primary meaning is downward direction of motion, while sink is treated differently since it implies motion through a medium different from air. But since there is a minimal paradigmatic contrast between these verbs (in English), it can be argued that fall also expresses Medium; that is, at least in the default interpretation, the downward movement is "through air". Furthermore, both (kinds of) verbs contain the additional information that the motion is caused by gravity and that there is a corresponding lack of control on the part of the trajector. Finally, as is the case with other Manner verbs, the "framing event" is expressed by a satellite, e.g., down in he fell down. We will suggest that such verbs take an intermediate position between "pure" Manner verbs (which could be locative, at least in some contexts, e.g., he walked for hours on the treadmill} and Path verbs proper such as enter and exit, since they seem to conflate Manner (in the broad sense of the term) and Path. Hence, an appropriate label for these verbs would be Manner + Path verbs, or "MP verbs" for short. The class of MP verbs is well represented in Thai, for example by phloo 'pop out', thalu 'pierce', and thim 'puncture', which include a semantic aspect of going through a landmark; by tok 'fall', Ion 'fall' 8 , lorn 'collapse', hoklom 'trip and fair, and com 'sink', which express a downward motion; and by laj 'chase', which like the Path verb taam 'follow' specifies motion in the direction of a moving object, but also conveys Manner-related information (purposive action, high speed, etc.) With a liberal interpretation of "translocation", we could also consider here verbs which express a directional postural change of the human body such as kom 'bend down', rjssj 'straighten up', and nan 'turn around', where the Path-related information is reflected by the English glosses. Intriguingly, the class of MP verbs, thus far characterized semantically, also constitutes a syntactic category in Thai. This becomes clear when MP verbs are used in combination with the other motion-verb types - Manner, Path and Deictic-Path verbs -, as in [17a]. Interestingly, the position of the MP verb is exactly where it would be predicted on semantic grounds: following the Manner verb, but preceding the Path verb(s). Violations of this constraint, e.g., [17b] and [17c], result in ungrammaticality. [17a]chan dean I walk

phloo pop out

M popped out, walking.'

?5ok exit

paj go

168


[17b]*chan I

phloo pop out

daan walk

Took exit

[17c]*chan I

daan walk

?ook exit

phloo pop out

paj go paj go

To summarize, the structure of the intransitive translocative serial-verb construction is characterized by the order specified in [18], where the asterisk represents the possibility for several verbs of a given type (with or without arguments) to follow each other, obeying constraints that are still not completely clear. [18] [Manner-V* > MP-V > Path-V* > Deictic-Path-V] clause The precedence order of the four different types of motion verbs \\-ithin a single clause can be established with a good deal of certainty. This constitutes an important grammatical fact and poses impediments to assigning Thai to only one side of the binary motion-event typology. A case in point is the following: Considering the presence of Path verbs, Thai ought to be regarded as a Vlanguage. But one characteristic feature of V-languages is that they obey the so-called boundary-crossing constraint (Slobin & Hoiting 1994), according to which Manner verbs cannot be used to describe situations in which a boundary is crossed. However, as Example [1], repeated in [19], shows, the Manner verb dssn, 'walk' combines not only with one but with two boundary-crossing verbs, and the sub-events described by them are within its scope. [19] chan dssn khaam thanon khaw paj naj suan I walk cross road enter go in park 'I walked across the road and into the park.' Note also that it is possible to have more than one landmark (Ground) per clause, again unlike V-languages but similarly to S-languages such as English. In these ways Thai behaves quite similarly to other serial-verb languages such as Ewe and Akan, as described by Ameka and Essegbey (in press). 2.6 Cause-of-Motion verbs A characterization of the expression of translocative motion in Thai would not be complete without considering "transitive motion" as well, i.e., utterances where the grammatical subject is not identical to the trajector, but rather constitutes the agent (or cause) of motion, while the trajector is expressed by the direct object in the sentence, as in [20].


[20] khaw he/she

joon cast

169

luukbon ball

?ook exit

paj go

'(S)he cast the ball out.'

Verbs such as joon 'cast', conflating Motion with Cause-of-Motion ("CoM"), are also numerous in Thai, e.g., Taw 'take', jip 'pick', saj 'insert', phaa 'take along', thoot 'take off, jok 'lift', and phlak 'push'. There are considerable similarities between the transitive translocative-motion construction and the intransitive one described in previous sub-sections. The "slots" for the Path verb(s) and the Deictic-Path verb appear to be the same. CoM verbs can also combine with Manner verbs, as in [21]. [21] khaw phaa luuk dasn Took he/she take-along child walk exit '(S)he took his/her child out walking away.'

paj go

However, it is impossible to combine CoM with most MP verbs in a single clause, as shown in [22]. The reason appears to be that most MP verbs imply volitional motion, i.e., agency on the part of the trajector, which is what is lacking when the trajector is the direct object of a transitive CoM verb. [22] *khaw joon he/she cast

luukbon ball

phloo pop out

?ook exit

paj go

'(S)he cast the ball and it popped out.'

This interpretation is supported by the fact that when the MP verb does not imply agency, as is the case with tok 'fall', combination with a CoM verb is indeed possible, as in [23]. [23] khaw

phlak

dek tok lorj

paj

he/she push child fall go-down go '(S)he pushed the child down.'

On the other hand, the Path verbs combine freely with CoM verbs, provided of course that they do not specify conflicting spatial values. Thus, [24a] is possible, whereas [24b] is not. [24a] khaw saj luukbon khaw he/she insert ball enter 'S(h)e put the ball inside the box.'

paj naj klorj go in box

170


[24b]*khaw he/she

saj insert

luukbon ball

Took exit

paj naj klorj go in box

This can be explained assuming that Thai Path verbs lack the element of agency as part of their semantic, "overt" meaning, while still being able to imply it "covertly" (Zlatev 1997), pragmatically (Levinson 2000), in the intransitivemotion constructions exemplified in [3-16]. The alternative would be to opt for a "polysemy" explanation with intransitive Path verbs expressing agency and transitive ones lacking this semantic feature. But since this latter account increases the complexity of the analysis and starts us on the slippery slope of postulating new "polysemes" for new constructions, we find the first account preferable. 2.7 Summary and conclusions A general linguistic analysis of translocative-motion constructions in Thai, based on native-speaker intuitions, shows that at least the following four characteristics make it extremely difficult to classify Thai as either verb-framed (a V-language) or satellite-framed (an S-language). /'.

Thai is not forced to "choose" between Path verbs and Manner verbs (and Cause-of-Motion verbs), since each has a separate slot in the serial-verb translocative-motion construction. Because of this, Thai features the kind of Path verbs typical of V-languages, and at the same time features a large number of Manner verbs making fine semantic distinctions, as in S-languages.

ii.

Furthermore, there is a separate slot for MP verbs, conflating Manner and Path in a single lexical item.

/'//.

Even when occurring together with Manner and MP verbs, the Path verbs exhibit the properties of full lexical verbs rather than those of, e.g., "deverbal adverbs", and hence cannot be treated as "satellites".

iv.

Unlike typical V-languages such as Spanish, Thai does not obey the boundary-crossing constraint (Slobin & Hoiting 1994) and allows more than one Ground expression in a single clause.

In order to see if these conclusions hold also in the face of quantitative, "usage" data, we analysed the data from the Thai frog-story corpus, as described in the following two sections.


171

3. METHODOLOGY: THE THAI FROG-STORY CORPUS The quantitative study of verbal expressions of Motion in Thai is based on a corpus of 50 Thai frog stories from five age groups. In constructing this corpus, we by and large adopted the well-known method for eliciting cross-linguistically comparable narratives originally developed by Berman and Slobin (1994), with some adaptations with respect to the definition of the basic units of analysis word, clause, and utterance - as described in this section. However, for the reader not interested in these methodological issues and eager to continue with the argument from the preceding section, we recommend skipping directly to Section 4. 3.1 Data collection The child data were collected in one pre-school and two primary schools in Bangkok and the adult data were collected from students and faculty members of Chulalongkorn University, in both cases during the first half of the year 2000. The interviewer, always a native Thai-speaker, first let the subject scan through the book for about five minutes. The instructions were approximately as in the following English translation, highlighting the narrative character of the task: "This story is about a boy, his dog and a frog. I'll let little brother/sister9 [= the subject] take a look at the pictures of the story, first. Then, big brother/sister [= the interviewer] will ask little brother/sister to tell big brother/sister the story, picture by picture." 3.2 Transcription and segmentation Ten narratives from each of the age groups 4, 6, 9, 11, and Adult were transcribed in standard Thai orthography. In almost all cases this transcription was performed by the person who carried out the interview, soon after the recording took place. Then the Thai transcription was converted into a phonemic notation via a semiautomatic computer program. The transliteration system shown in Table 1 for consonants and Table 2 for vowels was used. Tones were marked at the end of each syllable according to the following scheme: mid 0; low 1; falling 2; high 3; rising 4. While current versions of the CHAT notation (cf. MacWhinney 2000) support non-ASCII fonts, we found that sticking to ASCII increased the "portability" of our data considerably. We will use this notation in the remaining part of this chapter. Unlike Western orthographic systems, Thai writing does not place spaces between words, but rather uses them to separate semantic units, corresponding sometimes to phrases, sometimes to whole sentences. Hence, the phonemic

172


Table 1. The transliteration system for Thai consonants

Stop + voice - asp.

Labial

Alveolar

b

d

Stop - voice - asp.

P Stop - voice + asp. ph Fricative

f

Semi-vowel

w

Nasal

m

Palatal

Velar

Glottal o

t

c

k

th

ch

kh h

s j N

n

Lateral

1

Trill10

r

Table 2. The transliteration system for Thai vowels Front

Central

Back

Close

i

U

u

Mid

e

q

o

Open

X

a

O

transcription had to be segmented into word-level segments manually. This was straightforward in most cases since the vast majority of Thai words, especially in the colloquial register, are monosyllabic. However, it is not always clear whether certain multi-syllabic expressions should be treated as a) mono-morphemic words, b) multi-morphemic words including lexical compounds, or c) phrases consisting of one or more words. In deciding how to analyse particular cases, we used the following criteria: ;'.

Mono-morphemic word iff at least one of the syllables in the expression does not have a transparent separate meaning, e.g., naa2taaNl 'window". Even though this expression is probably a compound diachronically, the compounding is not transparent to present-day speakers, and therefore it was treated as mono-morphemic.

//.

Multi-morphemic word ("+" between the syllables) iff all the syllables have transparent separate meanings, but the meaning of the whole is not derivable by combining that of the parts, e.g., phuu2+jaj 1 'person' + 'big' = 'adult'. Lexical compounds are one sub-class of this category. Derivations such as khwaamO + sukl '-ness' + 'happy' = 'happiness') are


173

also included in this category, even though their derivation is semantically regular. ///. Phrase (a space between the syllables) iff the syllables have separate meaning, and combine systematically to give the meaning of the whole, e.g., maa4 'dog' + noj4 'little' = 'little dog' and raN4 'nest' + phUN2 'bee' = 'beehive'. In the same group as the second category, and thus marked in the same way (with a "+" connecting the parts), were formulaic expressions, e.g., mayO+penO+rajO 'never mind'. Thus, compounds and other multi-morphemic words, formulaic expressions, and reduplications could easily be treated as single lexical items (at the same time as analysis could easily be performed on their parts if required). The phonemic transcription was checked against the original tape recordings and adaptations were made to bring the transcription closer to the actual speech produced. Deviations from the normative pronunciation were represented using the CHAT convention of placing the citation form in square brackets after the transcription of the sub-standard form, e.g., laNO [: raNO] 'nest'." All pauses were marked in the transcriptions as short (#) or long (##). Repetitions and re-tracings were marked using the CHAT conventions, i.e., the repeated or re-traced material was surrounded by angle brackets "< >" (if consisting of more than one word) and followed by [/], [//], or [///]. The first indicates a repetition, the second a retracing with some change, usually a self-correction, and the third a retracing with substantial change, usually a paraphrase.12 One of the most serious methodological difficulties we encountered was how to define and segment the narratives in "sentence-length" units in a consistent and clear manner. In this respect there is a conflict between the instructions in the manuals of CHILDES, which urge the transcriber to focus on the utterance: "Each main line should code one and only one utterance." (MacWhinney 2000: 16), and those of Berman and Slobin (1994), which use the clause as a major unit of analysis, defining it as follows: Each clause should be transcribed on a new text line. We define a clause as any unit that contains a unified predicate. By unified, we mean a predicate that expresses a single situation (activity, event, state). Predicates include finite and non-finite verbs, as well as predicative adjectives. (ibid:660)

The serial-verb nature of Thai, exemplified in Section 2, made it difficult to decide what to regard as a "unified predicate". While knowing that our decisions will not satisfy all scholars of Thai linguistics, we used the following set of criteria (/-/v) for identifying clause boundaries in Thai. Thus, a clause boundary was marked with the conventional CHAT symbol [c]

174


/.

before the introduction of a new explicit or implicit subject;

//'.

a) before the relative-clause markers thii2 and sUN2 'which'; b) if there is only a noun phrase (i.e., the subject) between the previous [c] and the relative-clause marker, the clause boundary [c] is instead placed at the end of the relative clause; c) in other places where a relative-clause marker may be inserted without change of meaning;

///.

a) where clause boundaries are indicated by the presence of conjunctions such as 1x3, Ixxw3 'and', Ixxw3 kO2 'and then', kO2 'then', thxxl 'but', phrO3 'because', mUa2, phOOO 'when', conO 'until'. mxx3 'though', and phUa2 'for the sake of; b) in other places where a conjunction may be inserted without change of meaning;

iv.

after wa2 'that', if it is both preceded and followed by text segments with main verbs (thus excluding cases where wa2 is a main verb, and where it has nouns and other non-verb expressions as complements).

To illustrate, following these criteria, a popular pop-song line 13 consisting of seven verbs in a row and a pronoun [25] can be segmented into three clauses by the insertion of clause-boundary symbols after pajO (iii-b), after waaO (/v), and after thqqO (/). [25] jaakl dqqnO khaw2 pajO [c] bOOkl waaO wish walk enter go speak say/that

[c] rak3 thqqO [c] love you

'I want to come in to say I love you.'

Given this way of marking clauses, we were faced with the dilemma of how to satisfy the CHAT convention of having only one utterance per line (tier). If we chose only phonetic criteria, i.e., pauses and intonation contours, to define utterance boundaries, we would have to break up clauses - as previously defined - into many lines, and thus decrease readability and analysability. On the other hand, if we neglected prosodic criteria and only segmented the text into clauses, we would miss the fact that some clause boundaries coincided with pauses, and thus seemed to constitute processing units, while others did not. We resolved this dilemma with a compromise, operationally defining a criterium (v) for


175

Table 3. Size of the Thai frog-story corpus in terms of utterances, clauses, word tokens, and word types, excluding repetitions and retracings, by age Age

Utterances

Clauses

Word tokens Word types (excluding reps) (excluding reps)

4

380

433

2273

277

6

508

631

3141

312

9

616

780

4638

467

11

551

700

4227

448

Adults

957

1235

8144

855

Total

3012

3779

22423

1149

"utterance", i.e., the unit to be placed on a single line/tier, through a combination of phonetic and grammatical criteria: v.

An utterance boundary (.) occurs when there is both a phonetic indication of utterance closure - a short pause (#), a long pause (##), or a vowel lengthening (:) - and a. clause boundary, marked with [c].

This means that if there is only a pause but no clause boundary, the pause is marked within the utterance. Likewise, if there is a clause boundary but no pause or vowel lengthening, the utterance is assumed to continue until a clause and an utterance boundary coincide.14 Based on the above definitions, Table 3 shows the size of the corpus in terms of utterances, clauses, word tokens, and word types, the last two excluding repetitions and re-tracings. The Thai frog-story corpus, as thus far described, is available to the linguistic community through the data-exchange system CHILDES (childes.psy.cmu.edu). 3.3 Marking and coding of translocative utterances The forms involved in the expression of motion events in Thai - the verbs described in Section 2, but also prepositions and locative nouns (cf. Zlatev 2003) - are frequently polysemous between spatial and non-spatial meaning. Furthermore, one and the same form, e.g., thUN4 'to', 'reach', can appear as either verb or preposition. Finally, as pointed out in Section 2, it is important to distinguish translocative from self-contained motion, and sometimes this can be decided only by context. Factors such as these make it necessary to find all instances of translocative constructions manually and to "tag" their relevant

176


constituents. This was done on a subsidiary tier (%dir), as shown in [26]. Note that since there are two clauses in this utterance (according to the definitions presented above), each one including a translocative construction, the two are coded separately on two different subsidiary tiers. (The English translation above the utterance is added for expository purposes only; it is not included in the actual coded corpus.) [26] *** File "09j.cha", lines 54-56 he/she walk

until wonder enter

go

in

forest

*CHI: khaw3 dqqnO [c] conO loN4 khaw2 pajO r.ajOpaal '.c\*. %dir: dqqnO:MAN %dir: loN4:MAN khaw2:PATH pajO:DEI najO:RN paal:LM

Table 4 shows the tags used for coding the expressions in the translocative utterances in the corpus, the semantic categories they represent, and their form class as well as one example per tag.

4. THE PLACE OF THAI IN MOTION-EVENT TYPOLOGY As pointed out in Section 1, the general motion-event typology developed by Talmy can be formulated in two complementary ways: /.

in terms of the dominant lexicalization pattern for the verb root: Path + Motion, Manner + Motion, or Figure + Motion; or

//.

in terms of the dominant form of core schema expression: verb-framed (V-language) or satellite-framed (S-language).

Let us consider each one of these in turn, in the light of the Thai frog-story data. 4.1 Lexicalization patterns: Manner-type or path-type? According to Talmy (1985, 2000a), all Indo-European languages but the Romance ones, as well as many others, notably Chinese, predominantly conflate Motion with Manner in their verb roots, i.e., they are manner-type (Slobin & Hoiting 1994). On the other hand, the Romance languages, and quite a few Asian languages, including Thai (Wienold 1995), are supposed to be path-type, owing to a pronounced presence of verbs conflating Motion with Path. 15 Unlike the similar observation made by, e.g., Tesniere (1959). Talmy has consistently argued (e.g., Talmy 1991, 2000a) that these lexicalization types form an exhaustive typology, i.e., that any language will use only one type of verbs in

177


Table 4. Tags, semantic categories expressed ("conflation pattern"), form classes, and examples of translocative-motion expressions in the Thai frog-story corpus

Tag

Conflation pattern

Form class

Example

COM

Motion + Cause-of-Motion

Verb

joonO 'throw'

MAN

Motion + Manner

Verb

dqqnO 'walk'

MP

Motion + Manner + Path

Verb

phlool 'pop out'

PATH

Motion + Path

Verb

khaw2 'enter'

DEI

Motion + Path + Deixis

Verb

payO 'go'

PREP

Path

Preposition

caakl 'from'

RN

Region

Region noun

najO 'in'

CN

Region

Class noun

khaaNO 'side'

LM

Landmark

Noun

paal 'forest'

its most "characteristic expression of motion", defined (Talmy 1985:62, 2000a: 27) as that which /. /'/. ///'.

is colloquial in style, rather than literary, stilted, etc. ... is frequent in occurrence in speech, rather than only occasional. ... is pervasive, rather than limited, that is, a wide range of semantic notions are expressed in this type.

Even though English has a number of Path verbs such as enter, exit, ascend, and descend - notably of Romance origin -, these lack properties /-///. There are a few Path verbs such as pass and cross that do fulfill criteria / and /'/, but they are relatively few in number, i.e., they fail //'/. It has been argued, e.g., by Slobin and Hoiting (1994) and by Zlatev (1997), that the distinction between path-type and manner-type languages is much more of a continuum, and indeed Talmy (2000) allows for a certain degree of mixed types. Still, if one can maintain the claim that in any language, there is only one "most characteristic expression of motion" (cf. above), one may still decide, for instance, that despite Manner verbs such as rouler 'roll', French is a path-type language, while English can be classed as manner-type notwithstanding the few exceptions mentioned above. In both cases, the exceptions will be attributed to a lesser, ftcw-characteristic pattern. However, Thai appears to be problematic for this general claim, owing to its sizeable number of Manner verbs and Path verbs as well as to the structurally distinct classes of Manner + Path (MP) verbs, Deictic-Path verbs, and - in the case of "transitive motion" - Cause-of-Motion (CoM) verbs. The question to be

178


Table 5a. Cause-of-Motion verbs. Manner verbs, Manner + Path verbs (including posture-change verbs (P)), Path verbs, and Deictic-Path verbs in the 50 Thai frog stories Manner verbs

Cause-of-Motion verbs

awO plOjl sajl joonO phaaO phlakl capl jok3 namO thOOtl thee jUUn2 jipl khwitl Iaj2 phuN2 pqqtl soNl thUU4 thiN3

'take' Met go' 'insert' 'cast' 'take along' 'push' 'grab' 'lift' 'carry' 'take off' 'kick' 'pass', 'give' 'pick' 'poke' 'chase' 'throw' 'open' 'send' 'carry' 'throw away'

32 6 6 5 4 3 2 2 2 2 2 1 1 1 1 1 1 1 1 1

'fall' 'escape' 'chase, follow' 'pop out' 'fall, collapse' 'fall off 'bend down' (P) 'crawl under' 'bend down' (P) 'fall' 'turn around' (P) 'rise' (P) 'pierce' 'sink' 'stick in' 'pop up' 'look up' (P) 'stretch out' (P) 'dive' 'trip and fall' 'stand up' (P) 'turn ups. down'

"run" 'climb' 'fly' 'jump' 'walk' 'tiptoe quietly' 'crawl' 'float' 'slip (off)' 'swim' 'travel (far)' 'ride' Toll' 'leap' 'spread out" 'wade' 'trot (in mud)' 'wander" 'move frantically" 'crawl'

122 64 52 49 40 8 5 4 4 3 2 2 2 2 2

'exit' "descend" 'ascend' 'enter" 'follow' 'return' 'cross' 'reach' 'pass' "leave" 'encircle' 'go along' 'pass by"

228 213 146 64 49 46 8 6 4 3 1 1 1

'go" 'come'

523 392

1 1 1 1

Path verbs

Manner + Path verbs

tokl nii4 Iaj2 phlool Iom3 lonl kom2 mut3 chaONook2 ruuaN2 han4 Iuk3 thim2 comO jxxl phutl NqqjO chaONqq3 diNl hokl+lom3 jUUnO khwam2

wiN2 piinO binO kraOdoot 1 dqqnO JON2 khlaanO 1OOJO lutl waaj2+naam3 dqqnO+thaaNO khiil kliN2 kraOcoonO kra 1 cujO++kra 1 caaO lujO jam2 loN4 talkiaakl-H-talkaajO tajl

196 68 36 35 14 14 8 8 7 7 5 5 5 4 3 2 1 1 1 1 1 1

OOkl loNO khUn2 khaw2 taamO klapl khaam2 thUN4 phaan 1 caakl OOm2 1O3 IqqjO Deictic-Path verbs

pajO maaO

179


Table 5b. Total number of types, tokens, and mean number of tokens per type of Cause-of-Motion verbs, Manner verbs, Manner + Path verbs, Path verbs, and Deictic-Path verbs in the 50 Thai frog stories Verb type

Types

Tokens

Tokens/type

Cause-of-Motion verbs

20

75

3.8

Manner verbs

20

367

18.4

Manner + Path verbs

22

423

19.2

Path verbs

13

770

59.2

2

915

457.5

Deictic-Path verbs

answered is whether these classes also fulfil the criteria of being a) colloquial and b) frequent in everyday discourse. Table 5a and 5b show this to be the case for at least three of the classes. Table 5a gives the type-token count for each individual verb of the five classes present within the 50 Thai frog stories, while Table 5b summarizes the counts class-wise. With respect to the requirement of being colloquial, we can observe that all five classes are represented in this undoubtedly colloquial genre. With respect to frequency, however, we can see some clear differences. The Causeof-Motion verbs are least frequent, and have a fairly low token/type ratio. This rather surprising finding can probably be explained by the relatively young age of most of the narrators, as well as the actual content of the story, where selfpropelled "intransitive" motion dominates. The Deictic-Path verbs, on the other hand, are extremely frequent but have only two types, suggesting that it may be proper to regard Deixis as a grammaticalized category in Thai. However, if we focus on the remaining three categories - MAN, MP, and PATH -, we can observe that the differences between them are hardly sufficient to motivate regarding one as overshadowing the others and being the "characteristic" one. Each class is represented by a large number of types, where the first five or six types are highly frequent. In the present analysis, the MP class includes verbs expressing body-posture change such as han4 'turn around', involving the translocation of an implicit or even explicit 16 body-part trajector. If posture verbs - marked (P) in Table 5a - are included, MP is the most numerous category, and MP and MAN have an almost identical token/type ratio. But even if one excludes the posture-change verbs, MP is still very well represented. Furthermore, one can argue that from a semantic standpoint, MAN and MP could be considered as a single (macro-)category of "Manner verbs". Then this class of Manner verbs would have a total frequency that is almost the same as that of the PATH verbs. As regards the higher token/type ratio of the PATH verbs, it is predictable from their more schematic semantics (cf. Section 2).

180


Finally, the fact that MP verbs have a separate slot, positioned between the MAN and PATH categories, both syntactically and semantical ly, reinforces the conclusion that Thai belongs at least to a "mixed-lexicalization type". However, our belief is that its serial-verb characteristics make Thai not just a mixed type, but an altogether different type from the typical satellite-framed and verb-framed languages. We will argue for this conclusion in the remainder of this chapter. 4.2 Path expression: Verb or satellite? Since the "core schema" includes Path in the domain of motion events by definition, and since, as was demonstrated in Sections 2 and 4.1 above. Path is expressed verbally in Thai, it would seem that Thai should be regarded as a verbframed language. However, two quite different problems put this conclusion in doubt. The first is that the somewhat "bleached" semantics and high token/type ratio of Path verbs (see above) suggest that they could possibly be on their way to becoming "satellites" of the preceding Manner/Cause verbs. It is such a tendency, combined with less verb-like morpho-syntactic characteristics, that seems to motivate Talmy's (1991, 2000b) classification of Chinese and Lahu (typologically similar to Thai) as satellite-framed languages. In Section 3, it was argued on the basis of linguistic analysis that Thai Path verbs are indeed "true verbs", but can this conclusion be further supported by production data? For this we may consider the proportion of independent use of the different kinds of verb types in discourse. If it turns out that Path verbs appear predominantly together with Manner verbs, but that the opposite is not the case, i.e., if Manner or MP verbs often appear "alone" as in Example [10] above, then one may still suspect that Path verbs are progressing along a cline of grammaticalization that might lead them into "satellitehood". To investigate this, we checked the proportion of co-occurrence and independent use of the three main candidates for head-verb status in the translocative serial-verb construction: MAN, MP, and PATH. Table 6 presents the results. What Table 6 shows is that all three verb types appear independently of each other to a considerable degree. Far from being in a more dependent position, and therefore possibly developing into a "satellite", Path verbs occur independently even more often than the other two categories. Even if we were to subtract the 45 instances in which the category co-occurs with Cause-of-Motion (not counted in Table 6), it is still the case that in about half of all occurrences the Path verbs appear independently of other motion-verb types, save possibly for Deictic-Path verbs. In all these cases, the Path verb can be nothing else but "head" in the translocative construction. The conclusion is that in a binary- motion-event typology. Thai must be regarded as a verb-framed language - possibly as a "complex verb-framed type".

181


Table 6. Proportion of co-occurrence and independent use of the three verb categories Manner (MAN), Manner + Path (MP), and Path (PATH) in the Thai frog stories Category

Independent use (instances)

Total Co-occurrence (instances)

Percentage of independent use

MAN

367

+ MP=58 + PATH = 1 7 2

367-230= 137

37%

MP

423

MAN + =58 + PATH = 218

423-276= 147

35%

PATH

770

770-390 = 380

49%

MAN+ MP+

=172 =218

as proposed for serial-verb languages by Slobin and Hoiting (1994). However, the implicit premise of this conclusion is the second and more serious issue in doubt - what reasons are there to insist that the typology should be binary, apart from general considerations of simplicity? A good reason would be if it can be supported with empirical evidence that all languages fall into one of the two types with respect to a number of grammatical and discourse characteristics, as argued for a large sample of languages by Slobin in a series of recent publications (e.g., Slobin 1996a, 1997a, 1997b, 2000). However, whether serial-verb languages, and Thai in particular, pattern along with the typical verb-framed languages such as the Romance languages is highly doubtful, as we will show in the following section.

5. THAI AND V-LANGUAGES To answer the question above, let us compare specific data from the Thai frog stories with comparable data from typical verb-framed and satellite-framed languages, collected mainly by Slobin and his associates. Table 7, based on data reported by Slobin (2000), shows the proportion of Manner versus Path verbs used to describe a single event in the frog story: the sudden appearance of the owl from within the tree (Picture 12). As Slobin shows, the satellite-framed Germanic languages use at least a certain proportion of Manner verbs (the "Path verbs" reported seem to be mostly deictics such as come), while Russian, which can pack Path into the verbal prefix and Manner into the root, uses Manner verbs exclusively. The situation is reversed in the verb-framed Romance, Turkish, and Hebrew. When we introduce the Thai data into the comparison, however, the clear typological divide breaks down, or rather the binary distinction seems

182


Table 7. "The Owl's Exit": Percentages of Manner and Path verbs, considering only those narratives where the event was expressed (29 for Thai), in a number of different languages, based on Table 1, Slobin 2000 Language

Manner verb

Path verb

English

32% (16)

68% (34)

German

18% (5)

82% (23)

Dutch

17% (4)

83% (19)

100% (18)

-

French

-

100% (21)

Spanish

-

100% (64)

Turkish

-

100% (53)

Hebrew

3%(1)

97% (78)

MAN: 38% (11) MP: 28% (8)

PATH: 90% (26) DEI: 100% (29)

Satellite-framed

Russian Verb-framed

9

Thai

inapplicable. First of all, as for all serial-verb languages, the percentages need not add up to 100% since there is no requirement that there should be only one verb per clause. This "exiting" event in question is described by a (non-deictic) Path verb in 26 of the 29 instances, and with a Deictic-Path verb in all 29. The most noteworthy result is, however, that the use of Manner verbs and Manner + Path verbs is at least on the level typical of the satellite-framed Germanic languages, if not higher. Another result that typologically distinguishes Thai from the discourse pattern typical of V-languages comes from looking at the proportion of Ground expressions to the total number of motion verbs. In verb-framed languages, each path segment or change of direction (not just "boundary-crossing" subevents) needs to be encoded by a separate verb since there are no satellite-based constructions permitting a complex path with a single verb, such as in he went out of the house, down the hill, through the forest, and into the village. Since in a non-serial-verb language, a new main verb would imply a new clause, and a string of separate clauses describing the same motion event would be stylistically heavy.

183


Table 8.

Percentages of Ground expression per verb (and per clause) in Thai, English, and Spanish, based on data from Slobin (1997b), Table 1.4 Age

English

Spanish

Thai (per motion verbs)

Thai (per translocative utterances)

3-4

54%

48%

71/316(22%)

71/177(40%)

5

60%

50%

92/480(19%)

92/247 (37%)

9

62%

61%

114/564(20%)

114/257(44%)

117/477(25%)

117/229(51%)

147/728(20%)

147/315(47%)

11

Adult

82%

63%

Slobin predicted that speakers of V-languages would express only key segments of a path, leaving the rest to inference from context and background knowledge. This, in turn, would imply a lower proportion of Ground elements to verbs than in satellite-framed languages, where, as the English example shows, it is quite possible to include several Grounds in a single clause. As shown by Slobin (1997b), this prediction is only weakly confirmed for children, but the difference between speakers of V-languages and S-languages becomes more prominent for adults; cf. the first three columns of Table 8. As pointed out in Section 2, Thai does not obey the boundary-crossing constraint, and furthermore allows more than one Ground expression in the same clause - both claims can be illustrated with the example sentence presented in [1] and repeated in [19]. This leads us to expect a level of Ground expression that is higher than that typical of V-languages and possibly similar to that of S-languages. However, the results from the analysis of the frog-story data were contrary to this prediction. Considering the proportion of Ground expressions in Thai 17 , we find a level that is significantly lower than in Spanish, not only when we calculate the Ground/verb ratio (Table 8, column 4), which was indeed expected to be low because of the possibility to have a large number of different motion verbs in a single translocative construction, but also when we calculate the proportion of Ground expressions to translocative utterances (Table 8, column 5). Furthermore, as the breakdown into age groups shows, there seems to be hardly any development in the proportion of Ground expressions after the age of nine. How can this finding be accounted for? Prior to answering this question, let us look closer at four motion events in the frog story involving falling events, analysed previously by, e.g., Slobin (1996a) and Ibarretxe-Antunano (this volume). Comparison between Thai and Basque, another "untypical" V-language, shows how non-homogeneously languages assigned to the same side of the motion-event typology can behave with respect to this parameter; cf. Tables 9-10.

184


Table 9. "Bare verb(s)" or Ground in four different "falling events": comparison between Basque and the first 40 Thai frog stories (ages 4, 6, 9, and 11), based on Ibarretxe-Antunano (this volume), Table 1 Scene

Basque

Thai

Bare verb

Ground

Bare verb(s)

Ground

Dog falls

2

10

21

17

Beehive falls

4

8

20

2

Boy falls from tree

1 -

12

12

15

22

31

48

7(11%)

52 (88%)

84(51%)

82 (49%)

Boy and dog fall Total

Table 10. Comparison of the degree of Ground specification between the "problematic" languages Basque and Thai, the S-language English, and the V-language Spanish Language (category)

Bare verbs

Verb + Ground

English (S)

18%

82%

Spanish (V)

37%

63%

Basque (V?)

11%

88%

Thai (?)

51%

49%

The results in Table 10 are particularly troublesome for the binary typology, since it turns out that Basque, a V-language, has a higher level of Ground specification even than the S-language English, while Thai has a considerably lower level than Spanish. This shows that the presence or absence of Path verbs in a language can at best be a contributing factor to the tendency to specify Ground through nominal expressions. Another and seemingly more important factor is the presence of a developed adnominal spatial semantic system, such as the locative-case system in Basque (as suggested by Ibarretxe-Antunano) or in Icelandic, which is shown by Ragnarsdottir and Stromqvist (this volume) to have a higher proportion of Ground specification than the closely related Swedish, which lacks a case system. As for Thai, the reason for the relatively low level of Ground specification is clearly not a low level of Path expression, as hypothesized for V-languages by Slobin, but rather the opposite: since the Motion event is often so richly specified verbally (cf. Table 7), nominal Ground specification is often omitted since the information can be worked out from context. This can also be seen as part of a general tendency in Thai, and other South-East Asian languages, towards referential implicitness or "zero anaphora" (Clark 1992).


185

Table 11. Average number of event segments mentioned and percentage of narrators mentioning three or more segments, based on Slobin (1997:449). (Ten stories for each language, in Thai ten stories produced by nine-year-olds.) Language type

Language

Satellite-framed languages

Germanic (DU, EN, GE, 1C, SW)

3.0

86%

Slavic (PL, RS, SC)

2.8

76%

Verb-framed languages

Romance (FR, PR, SP)

2.1

30%

Serial-verb languages

Thai

3.0

80%

Mean number of Narrators mentioning > 3 segments event segments

This explanation for the surprisingly low level of Ground specification in Thai finds further confirmation when Thai is compared with the typical S-languages and V-languages with respect to how many events in a single trajectory of motion are explicitly encoded. According to Slobin (1997), one specific dramatic event in the frog story involves "four potential event components", which he specifies as follows (Slobin 1997:448): /. /'/. //'/. /v.

change of location: deer moves, runs, arrives at cliff; negative changes of location: deer stops at cliff; cause of change of location: deer throws boy, makes boy/dog fall; change of location: boy/dog fall into water.

Table 11 shows that in this respect, Thai is much more similar to the S-languages analysed - in both cases approximately 80% of the narrators mentioned at least three (two mentioned four, possibly five) of these "event segments" - than to the V-languages, where the corresponding percentage was 30%, with a significantly lower average number of event segments mentioned. Finally, Slobin hypothesized that this relative "under-specification" of V-languages in the above respect would be compensated for by a higher proportion of static scene descriptions, e.g., He fell. There was a lake below. The Romance languages have indeed been shown to have a higher proportion of such descriptions than the S-languages, in which they are practically non-existent. In this respect, Thai seems to be intermediate, since in two of the ten event descriptions analysed for Table 11, i.e., 20%, there were such "scene settings": more than in S-languages, but less than in V-languages. But clearly, a larger sample is necessary for any clear typological tendencies to be discernible.

186


To summarize, even though Thai clearly expresses Path (or the "core schema") through verbs, and not satellites, it remains problematic to categorize it as a Vlanguage since with respect to several properties, such as the degree of Mannerverb use and the degree of sub-event expression, it resembles the contrasting class of S-languages much more. These properties of Thai, as well as its insensitivity to the boundary-crossing constraint, can be attributed to its typological feature of verb serialization. The low proportion of Ground expressions was somewhat surprising, but this too can be understood as resulting from a high rather than a low - as in the Romance languages - level of trajectory specification, and a general "pragmatic" tendency to favour referential implicitness.

6. SUMMARY AND CONCLUSIONS The aim of the research described in this chapter was to investigate whether Thai, and by extension other languages with verb serialization, belongs to the verbframed (V-languages) or to the satellite-framed (S-languages) category of the well-known Talmian motion-event typology - or possibly to neither of the two. as suggested by Ameka and Essegbey (in press). We began by describing the structural properties of the Thai translocativ e serial-verb construction, establishing that intransitive motion can be encoded by four structurally and semantically different classes of motion verbs: Manner. Manner + Path (MP), Path, and Deictic-Path verbs, occurring with or without complements. Transitive (causative) motion is similarly expressed, though with verbs expressing Cause instead of Manner, and combining only with MP verbs that express non-volitional motion. Since Manner verbs combine with Path verbs in expressing boundary-crossing events, it could be concluded that Thai does not obey the "boundary-crossing constraint", which states that in V-languages. Manner-of-motion verbs can be used to describe only non-boundary-crossing situations (Slobin & Hoiting 1994). Furthermore, since several Path verbs can combine in a single translocative-motion construction, there is no constraint to encode only a single "path segment" per clause, as is the case in V-languages but not in S-languages. To see how these structural characteristics were "translated" into specific discourse properties, we investigated all translocative-motion utterances in a corpus of 50 Thai frog stories. The results showed that both the verbs expressing Cause/Manner (the classes CoM, MAN, and MP) and those expressing Path (MP. PATH, and DEI) were highly represented in terms of both tokens and types, though with explainable differences within the groups. The fact that the Manner + Path (MP) class straddles both sides of this divide reinforces the conclusion that Thai avoids the dilemma of either-Manner-or-Path through a synthesis: not either-or. but both!

187


Table 12. Comparison between structural and discourse characteristics of V-languages (verbframed), S-languages (satellite-framed), and Thai Parameter

V- Language

S-Language

Thai

Core schema (Path) expression

Verb

Satellite

Verb

Co-event (e.g., Manner) expression

Adverbial

Verb

Verb

Boundary-crossing constraint

Yes

No

No

Several path segments per clause

No

Yes

Yes

Manner-verb use

Low

High

High

Ground specification Lower (63% for Spanish)

Higher (82% for English)

Lower (25% per verb; 51% per clause)

Event granularity

Lower "Cliff scene": < 3

Higher "Cliff scene": > 3

Higher "Cliff scene": > 3

Scene setting

Yes

No

Some

Still more quantitatively, we compared the Thai frog-story data with similar data for other languages, collected by, e.g., Slobin and Ibarretxe-Antunano, with respect to the discourse parameters a) Manner-verb use (in the encoding of a particular "exiting" event), b) level of Ground specification, c} event granularity (number of sub-events encoded), and d) presence of static, "scene-setting" expressions. With respect to these factors, Thai appeared to resemble S-languages more than V-languages, especially if one disregards Ground specification, which was surprisingly low - something that could be attributed to a "pragmatic" principle of avoiding explicit nominal reference whenever Ground information can be contextually recovered. Table 12 summarizes these findings and juxtaposes them with the characteristics of (typical) verb-framed and satellite-framed languages. This clearly shows that even though by the first definitional criterion, Thai should be regarded as "verb-framed", it resembles typical V-languages very little, and in most structural and discourse parameters it behaves much more like the Slanguages. This seeming contradiction - Thai is a V-language that mostly behaves like S-languages - can be resolved only by stating that the language belongs to a third class. We hypothesize that this is the class to which all (typical) serial-verb (SV) languages belong. This last claim is supported by the research of Ameka and

188


Essegbey (in press), who come to essentially the same conclusions as we do for Thai in their analysis of the Niger-Congo languages Ewe and Akan. 18 Slobin and Hoiting (1994) acknowledge the special characteristics of serialverb languages, but describe them as being on a "cline" between S-languages and V-languages, rather than belonging to a separate third type: In this paper, we focus our attention in a language's tendency to encode paths in full verbs rather than in other types of elements. Perhaps it would be more appropriate, therefore, to speak of verb-framed and non-verb-framed languages. In this sense, languages like Lahu are at the verb-framed end of the cline. (Slobin & Hoiting 1994:502)

The problem with this proposal is that while on a "cline" based on a parameter such as verbal encoding of Path, serial-verb languages would indeed be "at the verb-framed end of the cline", but based on other parameters, e.g., verbal encoding of Manner, they would be towards the satellite-framed end. As a matter of fact, based on the findings reported in this chapter, summarized in Table 12. we would claim that there is no consistent metric according to which serial-verb languages can be positioned "between " S-languages and V-languages. Hence, we fully agree with Ameka and Essegbey that serial languages belong to a third type, or - to extend Slobin's (1996a) well-found metaphor - that they offer a "third way to travel". There is an even more fundamental reason why serial-verb languages should be considered a distinct type: The motivation behind the original dichotomy is the presence of a single main-verb slot in the clause structure of the languages analysed. Given this basic constraint, languages can fairly naturally be distinguished on the basis of whether they preferentially use this slot to encode Path, leaving Manner to an optional adverbial, or rather to encode Manner, leaving Path to a "satellite". But in the serial-verb languages, this basic constraint is relaxed - the language need not choose between the first and the second strategy since it can easily have it both ways: Path and Manner are expressed in two different verbs, which are structurally and discursively of equal status. The conclusion can only be that the existence of serial-verb languages such as Thai, Ewe, and Akan entails the need to extend the Talmian binary motionevent typology at least into a ternary one. In the process of this extension, the role of structural parameters such as the number of verbs per clause and the presence/absence of a locative-case system will need to be rethought. Theoretical frameworks such as distributed spatial semantics (Sinha & Kuteva 1995) and holistic spatial semantics (Zlatev 1997, 2003), which emphasize the many-tomany mapping between, on the one hand, (lexical and grammatical) form classes and, on the other, semantic categories such as Path, Manner, and Region, are likely to prove useful for this purpose. But this is the topic for another story ...


189

ACKNOWLEDGEMENTS We wish to thank especially Dan Slobin, but also Melissa Bowerman, Sven Stromqvist, Kiyoko Takahashi, Ann Lindvall, and Lars-Ake Henningsson for useful comments on earlier drafts of this chapter. We would also like to gratefully acknowledge the financial support of The Swedish Foundation for International Cooperation in Research and Higher Education (STINT), as well as the hospitality of the Department of Linguistics, Faculty of Arts, and the Centre for Research in Speech and Language Processing (CRSLP) at Chulalongkorn University, Bangkok, and in particular the Director of the CRSLP, Prof. Sudaporn Luksaneeyanawin. Finally, we wish to emphasize that despite our somewhat critical, "revisionary" attitude concerning some of the work of Len Talmy and Dan Slobin, we are very much aware that we are, at best, standing on the shoulders of these two linguistic giants.

NOTES 1.

It is generally unknown to the English-reading world that the currently famous typological distinction between Manner verbs and Path verbs was introduced by Tesniere (1959) under the terms mouvement and deplacement, respectively (cf. Walchli 2001). Tesniere observed that French motion verbs express predominantly deplacement, while in German, Latin, and Russian motion verbs express mostly mouvement.

2.

Talmy's (and Slobin's) notion of ground corresponds closely to that of landmark (Langacker 1987). We will use the two terms interchangeably, except in cases such as He went outside, where a ground is arguably expressed without there being an explicit landmark. To avoid any possible misunderstandings, we will capitalize the term, i.e., Ground.

3.

We will use this term, which we find considerably more transparent than the term translational.

4.

This is usually equated with the Speaker in discussions of Deixis within Thai linguistics, but the two are not identical (cf. Zlatev 2003), even though they appear to coincide more often than in English (cf. Fillmore 1966).

5.

Takahashi (1997) describes 26 such verbs expressing "global locomotory body motion", and this list is by no means exhaustive.

6.

Etymologically, the two syllables of the verb mean 'walk' and 'road', and synchronically, the verb implies traversing a long distance using effort and - by pragmatic implicature - by means of a vehicle. This motivates our inclusion of this verb in the Manner category.

190


7.

For example, English satellites are within the scope of negation too, at least when stressed: He did not walk into the room - he walked out of it.

8.

The difference between these two kinds of "falling" is subtle and difficult to pinpoint, but it seems that tok is more general while Ion implies a preceding event causing the fall. Accordingly, rain can tok but not Ion.

9.

"Big brother/sister" and "little brother/sister" are clumsy English translations of the terms phii and noon,, respectively, which are extremely common in spoken Thai and roughly indicate whether the speaker is of higher or lower age/status than the addressee.

10. The contrast between /!/ and /r/ is often not preserved in colloquial style, but it is a feature of Standard Thai, and we have consistently added the standard form in brackets in the cases where /r/ is pronounced as [1]. 11. Most of the CLAN programs make an automatic substitution of the second form for the first, which gives greater reliability in, e.g., counting word types. This substitution can be easily cancelled and analysis performed on the first forms by using the switch "+r5". 12. All CLAN programs except MLU and MODREP include the repeated/retraced material by default, and in order to exclude it, the switch "+r6" needs to be used. 13. "Somsaan", LOSO (Bangkok: More Music, 1998). 14. An exception from the condition Ul was made when there was only a single word (or short phrase) serving as a "filler" between the clause boundary and the actual pause: in this case, the utterance was terminated after this filler. 15. The "third type" according to this classification: conflation of Motion with Figure. i.e., information about the nature of the moving entity, exemplified by some NorthAmerican languages such as Atsugewi and Navajo, need not concern us here. 16. As in He turned his head around. One could argue that the corresponding motion verbs in Thai should rather be regarded as verbs of Cause-of-Motion. but in the present analysis, both kinds of uses, i.e., with explicit and implicit landmarks, are coded as MP verbs, for the sake of simplicity. 17. These were defined either as concrete nominals or, if those are absent, as "Class nouns" combined with "Region nouns", e.g., khaaNO 'side' + najO 'in', the combination of which corresponds to the English inside. 18. Ameka and Essegbey (in press) summarize their results in a table that is almost identical to our Table 12, except that they generalize directly to "serializing language" and do not include data on Manner-verb use and degree of Ground specification.

The Encoding of Motion Information in American Sign Language

Dennis Galvan & Sarah Taub

1. INTRODUCTION The human mind constructs complex conceptual models of real-world situations. Yet much of this information is not explicitly represented in the linguistic messages people use to communicate about these situations. This chapter, following Talmy (1985), is concerned with the relation between conceptual structures and linguistic elements. We focus on how typologically different languages encode information about motion events: in particular, which conceptual elements are encoded, and by what linguistic means. The present study compares American Sign Language (ASL) with English, drawing data from adults' retellings of the story Frog, where are you? As Talmy suggests, in this process we may look at the conceptual elements that a language might choose to express, the linguistic "surface" forms that are available to express them, and the patterns of how conceptual elements are encoded by particular types of surface forms. His 1985 paper examines several conceptual areas (e.g., motion, causation, temporal distribution), but the area that has inspired the most subsequent research concerns motion events. A motion event - that is, an event where an entity moves from place to place, or is identified as located at a particular place - may be divided into a number of basic conceptual pieces: the Figure or moving/located entity; the Ground or landscape against which it moves; the fact of Motion (or location) itself; and the Path along which it moves (or location where it stays). We may also identify the Manner of its movement; Manner is less conceptually unified but may include the locomotion method of an animate figure (e.g., running, clambering, etc.) and local details of the motion (e.g., sliding, bouncing, tumbling). Finally, for some events a Cause or external event is seen as providing the conditions for the motion event.

192


Talmy used this analysis to identify three different patterns of how languages express these elements; the patterns are exemplified in [1]. [la] la botella

entro

a

la cueva

flotando

the bottle moved-in to the cave 'the bottle entered the cave floating' (Spanish; path-type)

floating

[Ib] the bottle floated into the cave (English; manner-type)

[Ic] /' -w3rd pers. subject

ca-

st'aq'-

ic't

a

from wind blowing on it

for runny matter to move/locate

into liquid

3rd pers. subject

=> [c'wast'aq'ic'ta] 'icky stuff blew into liquid' (Atsugewi; figure-type)

The first grouping (dubbed path-type by Slobin and Hoiting 1994) expresses Path and Motion in the verb root; Manner and Ground information, if present at all, occur in surrounding elements or satellites. This group includes Romance. Semitic, and Polynesian languages; [la] gives an example from Spanish. The second group (manner-type}, consisting of most of the other Indo-European languages and Chinese and exemplified by English in [Ib], conflates Manner and Motion in the main-verb root, with Path given in satellites; these languages can also combine Cause and Motion in verb roots. Finally, the figure-type languages conflate Figure and Motion in the verb root and give Cause, Path, and Ground in satellites; the Hokan group of Native American languages are of this type, as exemplified by Atsugewi in [Ic]. In later work, Talmy (1991) picked out Path as the "core schema" of motion events, and categorized languages based on how they expressed Path. Languages with Path in the verb root are called verb-framed, and those with Path in satellites are satellite-framed. Thus path-type languages are by definition verb-framed, and all other types are satellite-framed. Talmy noted that satellite-framed languages could actually include more Path information per sentence than verb-framed languages, since they could concatenate many satellites specifying complex paths in a single sentence. While Talmy focused on encoding patterns at the sentence level, Slobin (1996a) looked at whole texts or narratives. One might suppose that languages like Spanish, though they encode less Path information in each sentence, might "catch up" over the course of an entire text. This apparently is not the case: Slobin found that Spanish texts included less Path, Manner, and Ground information overall than comparable English texts. It seems that disparities in encoding persist


193

across longer stretches of text; Slobin identified these disparities as an influence of language type on rhetorical style. There is also evidence (McNeill 1992) that speakers use gesture to squeeze in additional Path and Manner information, and that speech and the accompanying gestures form a tightly integrated whole. For example, McNeill (1999) found that Spanish-speakers, when uttering motion descriptions that contain no Manner information, frequently move their hands in iconic depiction of climbing, walking, falling, and other Manners. To our knowledge, however, there has not yet been a thorough comparative study to see whether the amount of encoded conceptual information equalizes across languages when gesture is taken into account. The preceding studies all focused on spoken languages. Now let us survey the resources signed languages have for describing motion events. Just as spoken languages do, signed languages have a vocabulary of lexical verbs (e.g., ASL's verbs glossed as RUN, WALK, THROW) to express motion concepts. All signed languages studied to date (see Brennan 1990 and Engberg-Pedersen 1993 among others), however, also have a complex system for representing the details of specific motion events (and in fact any kind of spatial description). These systems have received a variety of analyses (e.g., McDonald 1982; Supalla 1982; Schick 1987; Liddell 1995) and names (classifiers,polymorphemic verbs, etc.); the basic facts, however, are clear: when describing a motion event, the signer will use a conventional (and often iconic) configuration of articulators to represent the moving entity, and will move or position the configuration in ways that represent the actual movement or location of the entity. Some researchers (e.g., Supalla 1982) claim that these movements consist of discrete morphemes; others (e.g., Liddell 1995) argue that they are gestural and based on a conceptual blending of an envisioned "mental space" with the space surrounding the signer. Even lexical verbs will sometimes be spatially modulated to show specific paths and movements. One class of lexical verbs (often called spatial verbs) nearly always shows such modulation; another class (plain verbs) in general does not show modulation but may do so under special circumstances (cf. EngbergPedersen 1993 on "discourse modulation"). ASL spatial verbs include those glossed as RUN, DRIVE, and MOVE-OBJECT; for these verbs, the starting location represents the initial location of the verb's theme argument, and the final location represents the theme's final location. Another resource available to the signer is what has been called "role shift" or "referential shift" (Emmorey & Reilly 1995). Here the signer's own face and body are taken to represent the face and body of some entity being described; the signer takes on that entity's "persona" and demonstrates the entity's actions or experiences. The start of a referential shift is signaled by a shift in gaze away from the addressee; personas are differentiated by the angle of gaze and (often) by characteristic facial expressions or poses. Skilled storytellers can shift into four or five characters distinguishable by gaze and expression alone. This phenomenon is

194


difficult to analyze in terms of conventional linguistic morphemes, but lends itself to a conceptual blending account (Liddell 1995). ASL's classifiers and referential shifts are iconic in nature: the resources available to a signed language - the signer's body, hands, face, and surrounding space - are assembled via conventional, language-specific means into iconic representations of spatial information (cf. Boyes Braem 1981; Taub 2001). With the high degree of iconicity afforded by classifiers, spatial modulations of verbs, and role shifts, signed languages can easily encode a great deal of motion information. Impressionistically, a well-told signed story has the impact of a cinematic presentation. The purpose of this study is to give a quantitative measure to this qualitative impression: compared with English (a satellite-framed spoken language), how much conceptual information does ASL encode? Additionally, is there a tendency for certain linguistic structures to encode particular types of information?

2. STUDY 1 2.1 Data collection and transcription The children's story book Frog, where are you? (Mayer 1969; see Appendix I) was used as a prompt to elicit narratives from twenty adults. This book contains only pictures, with no accompanying text. The pictures present a set of conceptual elements which narrators may choose to include, emphasize, or ignore when retelling the story. Narrations of the story by users of different languages give insight into how linguistic type influences incorporation of conceptual information. Ten deaf, native users of ASL were videotaped signing the story and ten hearing, native English-speakers' stories were recorded and transcribed.


195

One scene, involving a boy, a dog, and a deer, was coded for this analysis. Pictures 16-18 of the book (Appendix I) depict this scene: the deer runs to a cliff with the boy on its head and the dog alongside barking at the deer, the deer stops, and the boy and dog fall over a cliff and land in a pond. This scene was chosen because elements of Figure, Motion, Path, Manner, and Ground were all present and could all be included in the narration if desired; in addition, the characters' affect was clearly visible from the pictures. The scenes were coded by the authors, a linguist fluent in ASL and a native signing psychologist, both professors at Gallaudet University. We coded the data together using each others' insights. To give a flavor of our data, we have placed two versions of the scene in Appendices 1 and 2. Appendix 1 gives an English-speaker's version; we will assume that our readers are familiar with English, and we give no detailed comments on the narrative. Appendix 2 gives an ASL-signer's version in glossed form; a few of his signs are further illustrated in Figures 1, 2, and 3. (Footnotes in the transcript indicate which glosses correspond to these figures.) Our transcription conventions for ASL are summarized in Appendix 3. To handle the simultaneous presentation of various types of information, we have used four different categories in the transcript, corresponding to columns in Appendix 2. The three leftmost columns show what the signer's hands are doing, the first for signs produced by the right hand alone, the second for signs produced by both hands together, and the third for the left hand alone. The fourth column, labeled "Perspective," gives the persona or character that the signer has assumed at each moment in the narrative. (The exact means used by the signer to mark each persona are not transcribed here, but they include characteristic facial expressions, postures, and eye-gaze directions.) As we can see, during the scene the signer switches from a neutral, "narrator" persona to that of all three of the featured characters; signs uttered while in a character's persona are understood as the utterances and actions of that character (e.g., the dog barking at the boy and telling him to get down off of the deer).

196


The transcript lets us illustrate the use of lexical verbs and classifiers to express spatial information. For example, the lexical verb BARK, delivered in the dog's persona, is directed upward and to one side; that area of signing space represents, from the dog's perspective, the relative location of the boy atop the deer's head. When the signer says GET-DOWN imper', GET-DOWN imper, again in the ° dog's persona, the imperative form of GET-DOWN starts near the boy's actual location and moves downward toward the desired location. Figures 1-3 present a few of the many classifier forms used in the excerpt: these particular forms all describe the boy's and the dog's fall into the pond. (We will return to these forms in our discussion of Study 2.) Let us begin with Figure 2, the simplest example. This classifier uses the V handshape (index and middle fingers extended) on both hands to represent an animate creature with four legs; contextual information lets us deduce that this classifier refers to the dog. The signer moves his hands on a path that arcs forward and down, showing the path of the dog over the edge of the cliff. At the same time, his facial expression shows the carelessness of the dog in not noticing the cliff. For Figure 2, both hands contribute to a single classifier form, but this is not the only possibility. Signers will often show two aspects of a scene via two classifiers, one created by each hand, as we see in Figure 1. Here the signer's right hand forms a V-shaped classifier (cl:v), representing an animate entity, and his left hand is in the flat B-shape (cl:b), a classifier representing a flat object. Context lets the viewer identify the cl:v with the boy, and the cl:b with the deer's head. According to ASL's conventions, since the signer is right-handed, the righthand classifier presents the focused Figure and the left-hand classifier presents a background element. The cl:v starts on top of the cl:b, and then moves off and downward in an arc path, representing the initial location of the boy on the deer's head and his subsequent path over the cliff's edge and down. Simultaneously, the signer's referential shift and the scared expression on his face indicates the boy's distressed emotional state. The signer's body will often contribute further meaning to a classifier construction. In Figure 3, the signer forms cl:v's with both hands, representing two animate beings. At the beginning of the construction, the left-hand cl:v is held at a mid-level position in signing space; by context, we identify this with the dog's position on the ground. The right-hand cl:v is held at the signer's forehead; in this construction, the signer's head is representing the head of yet a third animate being, whom we identify as the deer. Thus, the cl:v at the signer's head represents the boy on the deer's head. From these initial positions, both clrv's move through signing space: first the left cl:v rises in an arc, and then the two cl: v's move downward together, revolving around a common axis, so that the left cl:v ends up on top of the left one. These movements represent the simultaneous falls of the boy and the dog over the cliff, and their final position with the dog on the boy's stomach. J

e


197

For another example, see the first line of Appendix 2: the right hand holds the cl:v animate classifier atop the signer's head, while the left hand produces a Y-shaped shape-and-size classifier to the side of his forehead. This construction means that an animate being is lying on top of another being's head, near a long, curved protrusion. In the context of the story, we identify the first being with the boy, the second with the deer, and the protrusion with one of the deer's antlers. 2.2 Hypotheses, coding, and analysis We now continue with the description of our methodology. The experimental hypothesis for all of our contrasts was that English and ASL would show differences as to what was encoded, how often it was encoded, and what degree of complexity that encoding attained. The null hypothesis therefore was that there would be no differences between English and ASL. It concerned us that our analysis might give an unfair advantage to the ASL narratives, since those stories were videotaped, allowing full access to spatial and non-manual elements, and the English narratives were typed transcripts that did not notate any accompanying gestures or body shifts. David McNeill (1999, personal communication) told us, however, that speakers of English provide very little gesture when telling the frog story with book in hand; for the most part, gestures are limited to pointing to the book. Since for users of both languages, the task involved looking at the book while telling the story, the ASL video and English transcript versions were more comparable than one might expect. Yet, since English-speakers do regularly provide information in their free gestures, the best possible study would involve speakers and signers telling the story from memory, to allow maximum gesturing by the speakers. Three analyses of the selected scene were performed. The first determined whether the languages differed on how many pieces of conceptual information their narrators specified. The second analysis checked for how many times narrators encoded each piece of information. The third catalogued the morphosyntactic forms that were utilized in each language to express the content of the scene. The coding system for these analyses was based on a list of all possible elements presented in the scene's pictures within the categories of Figure, Spatial Relations, Motion, Path, Manner, Ground, and Affect. Each utterance referring to the scene was coded for whether or not a particular element was included. The possible Figures were the boy, the dog, and the deer. (Though each sentence has its own Figure, and in some sentences the deer is a Ground with respect to which the boy is located, the boy, dog, and deer were the main protagonists and functioned as Figures within the scene.) Spatial Relations between the Figures were also listed. The four Spatial Relations presented in the pictures (boy on deer's head among antlers, dog at deer's feet, dog next to boy in the fall, dog on boy's stomach) were

198


analyzed as each consisting of three different relations: simple accompaniment, a specification of direction (e.g., on, beside for English), and specific details of the arrangement (e.g., strewn across the deer's antlers). These relations were listed independently. Motion information included the following movements: the deer moving toward the cliff, the boy being carried, the dog moving toward the cliff, the deer stopping, the boy falling, the dog falling, the boy landing, and the dog landing. Manner information included the deer's running, the dog's running or jumping, the boy's bouncing as he was being carried, the deer's stopping with effort, the boy's falling, the dog's falling or jumping, and the boy's and the dog's splashing in the water. Information on Affect or internal states of the Figures in the pictures included the dog's concern for the boy, the dog's barking at the deer, the deer's anger, the boy's dismay over his predicament on the deer, the dog's carelessness in running off the cliff, and the boy's and the dog's fear while falling or on impact. Ground information in the pictures included the level ground, the cliff, the trees, the water or pond, and other minor landmarks. Path information included path segments and locations leading away from the rock, across ground, to the cliff, at the cliff, from the deer's head, over the cliff, down, towards the water, into the water, and in the water. We used this list to code the conceptual information present in each narrator's version of the scene. Based on this coding, we created two sets of variables: one set enumerated distinct conceptual elements encoded by each narrator, and the other set enumerated presentations of conceptual elements, including redundant ones. Separate variables summarized Figure, Spatial Relations. Motion, Path. Manner, Ground, and Affect information. These summary variables were then analyzed using the Mann-Whitney U test. Analysis of the first set of summary variables focused on whether or not a particular conceptual item was encoded. A significance level of .01 was used for each variable investigated. For this analysis, six variables (Spatial Relations. Motion, Path, Manner, Ground, Affect) were examined for an overall .06 error level. 2.3 Results Since each of the narrators mentioned all of the Figures at least once, no difference between languages was observed for Figure information, and the Figure variable was not included in this analysis. There was clearly no difference between ASL and English users on Ground information included in the narratives (Mann-Whitney U = 49.5, significance = .957, two-tailed). The patterns of what Ground information was encoded were remarkably similar: overwhelmingly, storytellers mentioned the cliff and the water, but no other landmarks. The languages differed, however, on encoding specifics of the Spatial Relations in the scene. This contrast approached significance (Mann-Whitney


199

U = 20.0, significance = .019, two-tailed). A significance level of .019 indicates that there is a 1.9% probability that these results could have occurred by chance alone. While a significance of .019 would be considered significant in many studies, we chose a .010 significance level for each of the six contrasts so that the overall error rate (.06) of the study would be acceptable. Had we used more narrators and had they responded similarly to the narrators we used, this contrast could have been statistically significant. In summary, ASL narrators gave more specific accounts of the Spatial Relations in this scene and this difference narrowly missed statistical significance. The measure of Affect information also approached significance (MannWhitney U = 19.0, significance = .017, two-tailed). ASL users included more information related to internal affective states of the Figures than did English users, but this difference is best described as a trend. The next three contrasts were all significant at the .01 level. The contrast on Motion was significant (Mann-Whitney U = 16.0, significance = .007, twotailed), indicating that ASL narrators specifically mentioned instances of Motion significantly more often than did English narrators. The contrast on Path was also significant (Mann-Whitney U = 12.5, significance = .003, two-tailed). The ASL narrators specified the Path of the Figures significantly more often than the English users. Finally, the contrast for Manner was significant (Mann-Whitney U = 13.0, significance = .005, two-tailed): ASL users specified Manner of motion significantly more often than English users. To repeat, these variables only measured how much distinct information was given, not how often a narrator repeated a given piece of information. Table 1 summarizes these findings. The second analysis in Study 1 included information on repeated or "reinforced" encoding of information. Impressionistically, ASL narrators tended to encode the same information repeatedly while English narrators did not; this analysis was performed to see whether the difference was significant. The results parallel the first analysis but show bigger differences between the ASL and English narratives. Again a .01 level of significance was used for each of the seven contrasts, giving us a .07 total alpha level for this analysis. As with the first analysis, there was clearly no difference in the presentation of Ground information (Mann-Whitney U = 48.5, significance = .906, twotailed). The two groups describe background features of the scene quite similarly, referring primarily to the cliff and water. Our impression was that the ASL narratives repeatedly presented the same (i.e., body-type) information about the three Figures of this scene, the boy, the dog, and the deer, through the use of classifiers and referential shifts. A contrast with English on this variable was not significant at the .01 level but did approach significance (Mann-Whitney U = 23.0, significance = .040, two-tailed). The third contrast dealt with all references to Spatial Relations between the three Figures in this scene. This contrast also approached significance (Mann-

200


Table 1. Results of Mann-Whitney U test on number of encoded pieces of information (excluding repetitions) Mann- Significance Whitney U (two-tailed)

Means ASL (n=10)

Means English (n=10)

Ground (unique)

2.0

1.9

49.5

0.957

Spatial Relationship (unique)

7.4

5.0

20.0

0.019

Affect (unique)

3.5

1.3

19.0

0.017

Motion (unique)

6.3

4.3

16.0

*0.007

Manner (unique)

5.7

3.4

13.0

*0.005

Path (unique) 5.6 Significant at the .01 level of significance.

4.1

12.5

*0.003

Variable

Whitney U = 18.5, significance = .017, two-tailed). The ASL narrators not only gave more specific accounts of the Spatial Relationships, they gave them more often, and again this difference narrowly missed significance at the .01 level. The two contrasts dealing with Spatial Relations together can be interpreted as a clear trend towards the ASL narrators including more specific Spatial Relations more often than the English narrators. The remaining four contrasts all reached a .01 level of significance. These contrasts compared the two groups on how often they specified Affect (Mann-Whitney U=15.5, significance = .008, two-tailed), Motion (MannWhitney U = 7.0, significance = .001, two-tailed), Path (Mann-Whitney U = 4.5, significance = .001, two-tailed), and Manner (Mann-Whitney U = 9.5. significance = .002, two-tailed). Again, ASL users repeated their specifications of Motion events, the Path of the motion, and the Manner of those events more often than English users. In addition, ASL users reported internal affective states of the three Figures more often than English users. Table 2 summarizes these results. The third analysis describes what kinds of syntactic forms were used by each group to convey the conceptual information of the story. We present a rough summary of the following categories: Figure, Spatial Relations. Motion. Path, Manner, Ground, and Affect; these results are also given in Table 3. The English-speakers referred to the three Figures a total of 83 times in the analyzed scene (an average of 8.3 times per narrative). They used exclusively nouns (66%) and pronouns (34%) to refer to the three Figures. The ASL signers referred to these Figures a total of 122 times (an average of 12.2 times per narrative). ASL narrators used nouns (39%) and pronouns (1%) much less frequently than the

201


Table 2. Results of Mann-Whitney U test on total encodings of information (including repetitions) Variable

Means ASL (n=10)

Means English (n=10)

MannWhitney U

Significance (two-tailed)

Ground (total)

2.9

2.8

48.5

0.906

Figure (total)

13.0

8.9

23.0

0.040

Spatial Relationship (total)

12.8

6.7

18.6

0.017

Affect (total)

4.6

1.3

15.5

*0.008

Motion (total)

13.6

5.7

7.0

*0.001

Path (total)

10.0

5.0

4.5

*0.001

Manner (total)

10.0

3.8

9.5

*0.002

* Significant at the .01 level of significance.

English narrators and instead relied on classifiers (60%). In addition, the ASL narrators used referential shift in 14% of their references to these Figures. (Unfortunately, we do not have access to similar information for the Englishspeakers, as the English transcripts do not indicate if the narrator at any time "shifted" into one of the character's roles through the use of intonation or other mechanisms.) We should note that owing to their iconicity, ASL's classifiers and referential shifts provide much more information about the Figures' shape and size than do English's nouns and pronouns. A comparison of what forms the two groups used to express the Spatial Relationships between the three Figures also shows marked differences. The English narrators used a total of 41 forms, relying primarily on prepositional phrases (32%), plural pronouns indicating conjunction (15%), adverbs (15%), and other conjunctions (10%), with a few other forms thrown in. (Approximately 12% of the subjects established the fact that the boy was on the deer in a section of the narrative which was not included in our sample.) The ASL narrators used adverbs (11%), verbs such as FOLLOW (16%), prepositional phrases (7%), plural pronouns (2%), and conjunctions (2%) only occasionally. Instead, ASL narrators relied heavily on classifier constructions (55%), adding referential shift occasionally (11%). Motion patterned similarly to Spatial Relations. The English narrators used a variety of verb forms (52 instances) to indicate Motion, including simple verbs of motion (73%), causative verbs of motion (21%), and other verbal constructions (6%). In contrast, the ASL narrators' 86 instances of Motion information relied heavily on classifiers. While ASL narrators did use some


202

Table 3. Type of syntactic category used to encode conceptual elements

ASL

English

Figures

39%

66%

Pronouns

1%

34%

Classifiers

60%

Nouns

Referential Shift

*14%

Manner Simple verbs

5%

12%

Other verb constructions

8%

Participles Classifiers Referential Shift

74%

5%

Causative verbs

95% *21%

Path Prepositional Phrases

10%

60% 20%

Verbs Spatially modulated verbs

12%

Verb/Particle constructions

3%

Classifiers

74%

Referential Shift

*6%

20%

Ground Nouns

34%

96% **4%

Verbs Classifiers

47%

Fingerspelling

18%


203

Table 3. (cont.) Type of syntactic category used to encode conceptual elements

ASL

English

Prepositional Phrases

7%

32%

Plural Pronouns

1%

15%

Spatial Relations

Adverbs Other Conjunctions

11%

15%

1%

10% 12%

Previously established Verbs

16%

Classifiers

64% 8%

Miscellaneous Referential Shift

7%

*17%

Motion Simple verbs Causative verbs

20%

73%

2%

21% 6%

Other verbs Classifiers Referential Shift

77%

*21%

Affect Verbs Full Clauses Adjectives

14%

25%

7%

58%

3%

17%

Referential Shifts

55%

Mouth Adverbs

21%

* Percentage of classifiers; Referential Shift can occur simultaneously with classifiers. ** The verb splash.

204


simple verbs (20%) and causative verbs (2%) for this purpose, the vast majority of forms were classifier constructions (77%). In addition, out of the 67 classifier constructions used by ASL narrators, 14 (21%) also included a referential shift that indicated motion. English and ASL narrators used completely different syntactic forms to indicate Path of motion. English narrators referred to Path (45 instances) with prepositional phrases (60%), verbs (20%), and verb/particle combinations (20%). ASL narrators referred to Path (68 instances) using prepositional phrases (10%) and spatially-modulated lexical verbs (11%) much less often. Instead. ASL narrators relied heavily on classifier constructions (68%), sometimes accompanied by referential shifts (6%). English narrators also used verb forms most of the time to indicate Manner (39 instances). These included simple verbs of motion (74%), causative verbs of motion (5%), other verb constructions (12%), and participles (8%). When ASL narrators indicated Manner (77 instances), they used classifier constructions 95% of the time. Simple verbs were used the remainder of the time. Out of the 73 classifier constructions, 15 (21%) included a referential shift to indicate manner. Recall that we found no difference in the amount that English and ASL narrators referred to Ground elements. It was striking to us that all narrators in both languages referred to the cliff and the pond almost exclusively. They did not use the same syntactic forms to do so, however. English narrators used only nouns (96%; the only exception was the verb splash] to refer to these Ground elements (28 instances). ASL narrators made 38 references to Ground elements, using nouns (34%), classifier constructions (47%), and fingerspelling (18%). English and ASL narrators used very different forms to express Affect information. English narrators primarily used verbs (23%) and full clauses (54%) to describe the thoughts and emotions of the Figures. ASL narrators used these kinds of forms occasionally (14% and 7%, respectively) and instead relied mostly on referential shifts (55%) and mouth adverbs (21%). To summarize this analysis, English users typically expressed motion information through nouns, pronouns, verbs, and prepositional phrases; ASL users occasionally used these lexical categories for the same information, but more often relied on classifiers and referential shifts. Our impressionistic conjectures were confirmed by the results of Study 1. ASL signers consistently incorporate far more conceptual information into their descriptions of motion events than do English-speakers. In particular, significantly more Motion, Path, and Manner information was included, and there was a strong trend towards greater Affect and Spatial Relations information. Moreover. ASL signers repeat this information more frequently than English-speakers - this contrast was significant for Motion, Path, Manner, and Affect, and approached significance for Spatial Relations. (Interestingly. ASL matched but did not exceed English on amount and frequency of Ground information.) We also noted a strong


205

preference in the ASL narratives for expressing conceptual elements through classifiers and referential shifts. Our findings show the deep influence of iconicity on ASL descriptions of motion events. Even when alternative forms (nouns, lexical verbs, fingerspelling, extended explanations) are available in ASL's resources, signers overwhelmingly use iconic classifiers and referential shifts. Engberg-Pedersen (1999) noted this phenomenon in Danish Sign Language, and referred to it as a signed-language narrative ideal - in a signed story, wherever possible, narrators should "show" the information using iconic forms, rather than "describe" it using lexical forms. The greater amount and repetition of motion information in ASL stories can be attributed to iconicity as well. One might conceive of the repetition of information as linguistic anaphora or agreement, but in fact the cause might not be specific to language. When multiple iconic forms of any sort are constructed, they portray many aspects of the referent event, and each iconic form must agree with others in accurately portraying those aspects; this leads naturally to redundancy across forms. For example, nearly all the ASL forms presenting the boy's Motion and Path through space also contained Figural information; this information consistently (and redundantly) portrayed him as a two-legged animate being. Though we often think of repetition or redundancy as boring or unnecessary, the repetition in these iconic narratives serves an important purpose: it reinforces the viewer's mental imagery and helps the viewer comprehend new information that is mixed in with the repeated information. We will return to this point in the discussion of Study 2. This line of thought leads directly to the question of conflation patterns in ASL's representations of motion events: are there constraints or limits on ASL's iconicity? A priori, since the articulators of ASL are objects (i.e., body parts) moving in space, one might expect that signers would encode all the conceptual information about a motion event in a seamless flow of iconic representation. That is, all of Talmy's pieces of a motion event might be "conflated" into one classifier form. Previous work, however, has made it clear that this does not happen. ASL signers do in fact separate different pieces of the event into different linguistic components. (Note that it might make more sense to discuss languages' separation patterns rather than their conflation patterns.) Supalla (1990) took explicit notice of this issue and pointed out a number of separation patterns in ASL. Most classifier-based verbs of motion follow one of two patterns: they may combine Figure information with Manner of locomotion, or else with Path information. The Manner/Figure/Motion verbs use the signer's body to represent the referent's body, and focus on details of the referent's limbs; examples include the use of the signer's fingers or hands and forearms to represent the referent's legs and feet. The Path/Figure/Motion verbs, in contrast, use only the signer's hand or fingers to give an overall view of the referent; the signer's body is not semantically significant. Examples include the uses of the flat B-hand

206


or extended index finger to show movement of a person or thing through space. The "two-legged" classifier made by extended index and middle fingers (Vshape) is an exception, in that it encodes both Manner and Path as well as Figure and Motion. (Though Supalla does not make this observation, the exception could be explained by the fact that the V-shape alone has a part-whole structure that can show limb movement as well as the relative freedom of movement of an item articulated solely by a single hand.) Supalla noted that these two types of motion verbs often combine in a series, allowing the signer to express both Manner and Motion information in the same sentence. Thus, to describe a person limping up a hill, a signer would first produce a Manner-of-motion verb showing the limping body motions, and then a Path motion verb showing the person's trajectory up the hill. Though it is physiologically possible for the signer to move the Manner verb through the uphill path, signers do not do so; instead, they separate a putatively unitary event into two conceptual chunks articulated in series. Supalla referred to these patterns as arbitrary linguistic restrictions on which aspects of the event could appear together in which forms. One may question Supalla's interpretation, however. Native ASL signers do accept some Manner/Figure/Motion verbs that show a definite path (Samuel Hawk, personal communication). It may be more accurate to interpret Supalla's observations as tendencies, rather than absolutes. Moreover, one might seek a motivation for the tendencies in general cognitive and perceptual functioning rather than in linguistic restrictions. We shall return to this point in our discussion of Study 2. Slobin and Hoiting (1994) studied verbs of motion in Sign Language of the Netherlands (SLN), whose classifier system is like ASL's in nature though not in every detail. They supported Supalla's view that motion events are conceptually separated into different forms, but disagreed on the details. In particular, whereas Supalla claimed that in a series of a Manner and a Path motion verb, the Path verb is secondary and "reduced", Slobin and Hoiting claimed that the Path verb is actually primary. To support this point, they adduced examples from spoken languages that use verb series of Manner and Path verbs; in those cases, the Path verb generally functions as the main verb of the sentence. They also contended that, since (according to Talmy 1991) Path is the "core schema," and since ASL verbs of motion necessarily display Path, ASL is a verb-framed, path-type language. They introduced the notion of complex verb-framed languages; this type would consist of verb-framed languages such as Turkish, ASL. and SLN that use verb series to convey motion information. Slobin and Hoiting noted another characteristic of verb-framed languages: they give special treatment to crossings of boundaries. Satellite-framed languages such as English treat boundary-crossings in the same way as any other Path segment, through the addition of a satellite element specifying that segment

207


(cf. [2a]); these satellites normally follow a verb of Manner. But verb-framed languages differ. As we have seen (cf. [la] and [2b]), for the most part verbframed languages express Path in the main verb, with a satellite expressing Manner. However, in certain contexts (called path-focused by Slobin and Hoiting), Path may appear as a satellite with a Manner main verb (as in [2c]). Yet those Paths may never include a crossing of a boundary. A sentence such as [2d], which attempts to combine a boundary-crossing Path with a Manner verb, is unacceptable. [2a] The man ran out of the surf, across the field, and into the house. [2b] el hombre the man

entro moved-in

[2c] el hombre corrio the man ran [2d] *el hombre the man

corrio ran

corriendo running

hasta up-to de from

a to

la casa the house

la casa the house la playa the beach

en in

la casa the house

These examples show the special treatment of boundary-crossings in Spanish, and Slobin and Hoiting provide analogous examples for French, Turkish, Japanese, and Korean. According to their data, the same is true for SLN - each boundarycrossing is marked by its own verb. The scope of our investigation did not permit an extensive analysis of separation patterns in our data set. We chose to focus on separation patterns in the description of a single complex event in Frog, where are you? We hypothesized that there would be a limit on the amount of conceptual information that could be presented in a single classifier form. Though our focus was too narrow to fully test and compare Supalla (1990) and Slobin and Hoiting's (1994) claims, our observations partially confirm and extend their results. Specifically, we confirm the tendencies that they observed but question whether they involve arbitrary linguistic rules or cognitive heuristics.

3. STUDY 2 We investigated the separation and repetition patterns associated with the classifier V handshape (cl:v) in one event of the story, the boy's and the dog's fall into the water. This "fall" is complex, involving all three Figures and simultaneous Paths from the deer's head (for the boy) and from the top of the cliff (for the dog), over the edge of the cliff, down, and into the pond. Logically, this event could have

208


been encoded in one complex form using both hands and the signer's head: one hand using a cl:v (animate classifier for the boy) could have started from the forehead (representing the deer), moved in an arc downwards, and stopped with a bounce, while the other hand, also using a cl:v (animate classifier for the dog), would have started lower and to the side of the neutral signing space (representing the top of the cliff), and moved in a parallel path also ending with a bounce. Only one of the 37 forms produced by our native signing subjects did this (Figure 3). Most forms were far simpler. Moreover, every narrator produced at least three cl:v forms, each focusing on different aspects of the scene. This suggests the existence of linguistic or perhaps cognitive constraints on simultaneous encoding of information. The patterns of use are summarized in Table 4. Our ASL-signing narrators all repeated the cl:v an average of 3.7 times with a range from three to five instances. These cl:v's were combined and recombined with different additional information. Out of 37 instances of the cl:v in the "fall," only one could be considered an exact repetition of a preceding form (and this person did use three distinct cl:v forms). Some clear separation patterns emerged from the data. All signers' cl:v forms (with one possible exception; see note in Table 4) included Path information. The signers divided the Path into three pieces: the initial arc (denoting a loss of balance and fall), a downward segment, and a final bounce. We saw cl:v forms that incorporated the entire Path (4), the arc, and the downward segment (21). the downward segment and the bounce (2), only the arc (2), and only the downward segment (8). The initial arc was included in 27 forms, while 6 included the bounce; moreover, no forms included only the bounce. This suggests a tendency to focus on the initial part of the event rather than the end; also, awareness of the lexicalized sign meaning/all, derived from a cl:v with an arc path, may have influenced signers to include the arc more frequently. We can divide the cl:v forms into four groups based on what conceptual elements combined with Path: Motion + Figure, showing an animate entity falling (12); Motion + dual Figure, showing the boy and the dog falling together (7); Motion + Figure + Ground (initial landmark), showing the boy falling from the deer's head or the dog falling from the cliff (13); and Motion + Figure + Ground (boundarycrossing), showing an animate entity entering the water (5). Additionally, there was one highly complex form which included separate Paths, Figures, and Motions for the boy and dog, along with the Ground (initial landmark) of the deer's head. If there is an upper limit on the amount of conceptual information in a cl:v form, we might expect that the second, third, and fourth groups and the complex form would accommodate less complex Paths than the first group. The first group, containing the forms with the fewest additional elements (Motion + Figure), could combine with Paths of any degree of complexity. We

209


Table 4. Encoding of conceptual elements with the cl:v form Conceptual Categories

Scene Elements

Form

Encoded Path Information start

Motion + Path + Figure

Motion + Path + Figure (dual)

animate fall (boy or dog)

cl:v

arc

down

*1 X

6

X

2

X

X

cl:v + animate fall (dog; four-legged animal) cl:v

X

X

cl:v + cl:v

X

X

X

X

dual animate fall (boy and dog)

end

X X

X

X

fall from head of deer (boy)

cl:v + forehead

1 •>

T

X

X

Motion + Path + Figure + Ground (initial landmark)

N

1

1

X

2 1

X

X

X

X

X

5

X

X

X

3

X

X

X

fall from cliff (dog) cl:v + cl:b

X

X

X

2

Motion -r Paths (2) + Figures (2) + Ground (initial landmark)

fall from deer (boy) cl:v + cl:v & fall from cliff (dog)

X

X

X

1

Motion t Path + Figure + Ground (boundary-crossing)

animate enters water (boy or dog)

X

5

cl:v + cl:b

cl:v + cl:b

X

The arc in this item may have represented continuous aspect rather than arc Path.

2

210


should note that two of the forms in this category actually used both hands to represent all four legs of the dog, as illustrated in Figure 2. The second group of forms added a second cl:v to represent the second Figure. In all of these cases, the two hands moved symmetrically, either in parallel (for the fall) or in alternation (for a few depictions of the bounce). Because of this, we consider these forms to have a dual Figure (the boy and dog "chunked" together) tracing a single Path. One of these forms is illustrated in Figure 1. The third group of forms incorporated a Ground element representing the initial landmark. In six cases, the signer's head was used to represent the deer's head; in five cases, the non-dominant B handshape (cl:b) represented the deer's head; and in five cases, the cl:b represented the cliff. Table 4 marks the use of this initial landmark in the column labeled "start." In considering the second and third groups of cl:v forms, we found that adding the initial landmark or the complex dual Figure did not affect the allowable complexity of the form's Path. Both the initial-landmark and the dual-figure forms incorporated one, two, or all three Path elements. A possible exception to this generalization is the group of forms in which the signer's head was used for the initial landmark: none of these six forms included the final bounce. A referential shift is needed to define the signer's head as the deer's head, and adding the final bounce to that form might have made the form too complex for most signers. The fourth group of forms is perhaps the most remarkable. In these forms, the non-dominant cl:b represented a Ground element, the surface of the water, and the cl:v was moved downward past the cl:b. All five of these forms incorporated only the downward Path segment. No articulatory constraint limits the Path of the cl:v: presumably it could move in an initial arc down past the water to a final bounce, but this never happened. This is a striking difference from the initial-landmark group, which also incorporated a single Ground element yet could include any Path elements. It suggests that the number of conceptual elements is not the only factor in determining separation patterns, and that boundary-crossings are in some way special in ASL. To further investigate this hypothesis, we examined all other boundarycrossings in our sample often ASL narratives. We found that if the boundary does not impede the Figure's motion, the crossing is always represented either with a lexical verb (e.g., EXIT) or with a straight-path classifier. Boundaries that do impede forward motion (e.g., a windowsill) are coded with arc-path classifiers or even more complex classifier forms. This suggests that the nature of the boundary is crucial to its effect on separation patterns. Finally, the single most complex form is illustrated in Figure 3 and described under Study 1 above. It incorporates arc and downward Path elements for two separate Figures, plus an initial Ground for one Figure, and an ending configuration showing a complex Spatial Relationship between the Figures. We note that before producing this form, the signer had already produced


21

two simpler cl:v forms (Figures 1 and 2); nearly all the information in the complex form had already been presented. This is consistent with a cognitive constraint on the level of complexity: this much complexity might only be comprehensible when the addressee has already had a chance to assimilate it in smaller chunks. To summarize, all signers used at least three different cl:v forms to describe the fall, showing that some degree of conceptual separation is necessary for this complex event. Also, all signers repeated Figure, some Path, and other information in their different forms, showing the importance of repetition in building up a coherent picture of the event. Nearly all the cl:v forms consisted of Motion, Figure, Path, and one possible additional element (Ground element or second Figure chunked with the first). We saw only one form of greater complexity, which followed (and summarized) two simpler forms. This suggests that there is a strong tendency to limit the amount of conceptual information that cl:v forms can express. Yet complexity is clearly not the only factor, as we see from the contrast in allowable Path complexity between forms with initial landmarks and forms with boundary-crossings.

4. DISCUSSION AND CONCLUSIONS Our results support Supalla's claim that ASL signers do in fact separate a unitary event into different conceptual pieces. Though ASL's resources for describing motion events are highly iconic, this is not a simple, directly depictive kind of iconicity. Instead, the event is decomposed into pieces which can be presented separately, and there is a strong tendency to limit the amount of conceptual information that appears in each iconic form. It does not seem, however, that there is an absolute upper bound on the amount of information. This suggests that many of ASL's separation patterns stem from limitations on cognitive processing rather than on linguistic co-occurrence restrictions. We also find evidence supporting Slobin and Hoiting's view that boundarycrossings are special, though the matter may be more complex than they suggest. In our selected scene, complex paths never appeared in simultaneous construction with boundary-crossings, whereas they freely co-occurred with representations of other landmarks. Yet complex paths did co-occur with other boundary-crossings in the narratives, notably those in which the boundary impeded the Figure's motion. We propose that this differential treatment of impeded and non-impeded boundary-crossings is not due to arbitrary co-occurrence rules but instead to cognitive factors. Presumably, signers have mental images or models of the entire motion event which they wish to describe. Let us hypothesize that for boundary-crossings, they focus their attention on a small piece of the image immediately surrounding the

212


crossing (cf. Langacker 1987's notion of scope). For non-impeded crossings (e.g.. the boy's fall into water), the immediate neighborhood of the crossing contains only a straight-line path and no effort on the Figure's part. But for impeded crossings (e.g., the frog's escape from a jar), the neighborhood contains an arc path and potentially complex maneuvering by the Figure. If signers encode what they construe as being in the crossing's neighborhood, they would create forms with straight paths for non-impeded crossings and complex paths for impeded ones. This focus-based theory matches the observed data without requiring arbitrary rules. The finding that ASL signers "focus in" on boundary-crossings gives some support to Slobin and Hoiting's claim that ASL should be considered a verb-framed language, since these languages give special treatment to boundary-crossings. Consider, however, that ASL encodes much more conceptual information than English, a satellite-framed language. If it is true that verb-framed languages (e.g., Spanish) tend to encode significantly less Path and Manner information than satellite-framed languages, we might conclude that if ASL is a verb-framed language, it is highly atypical. Pending future studies, we maintain neutrality on this issue — it may even be the case that we need to expand Talmy's typology to accommodate languages which can conflate Path, Manner, and Figure into a single verb form. At this point we would like to discuss further whether ASL's separation patterns are arbitrary and linguistic, as Supalla maintains. Some of the patterns that Supalla noted are conventional facts about ASL - for example, the fact that signers use extended index fingers rather than index and middle fingers to represent a referent's legs - and thus are arbitrary to some degree. But the overall pattern he observed of a separation between Manner of locomotion and Path may not be arbitrary at all. McNeill (1999) reported on a study of gestures that accompany spokenlanguage descriptions of motion events; his study involved speakers of English. Spanish, and Mandarin, from age 3 up until age 12. His earlier work (McNeill 1992) showed that these gestures are tightly bound to the spoken message they accompany, and that much conceptual information may be expressed through these gestures; in particular, gestures may carry (at least) Path, Manner, and Figure information. The recent study showed that starting at age 3. speakers of all three languages separated Manner and Path information in their gestures. That is, the children never produced gestures that included both Manner and Path information. Adult speakers, however, regularly produced such combinations in their gestures. McNeill suggests that there may be a limit on children's ability to represent motion events; though they may recognize the Manner and the Path of the motion events, they cannot yet combine them into a single gestural portrayal. Given these results, we suggest that the locomotion/path split is a cognitive universal, not a language-specific one, and that ASL's motion verbs' patterns


213

derive from the same cognitive structures as spoken-language gestures. The fact that most of these patterns seem to be tendencies rather than absolute rules also supports a cognitive rather than a linguistic origin. This proposal could reconcile Supalla (1990) and Slobin and Hoiting's (1994) observations with Liddell's (1995) analysis of the use of space in ASL as "gestural." The discretization of the representation of space and movement in ASL would then draw not on "morphemes" in the traditional sense, but on separate cognitive representations. The logical next step is to look for separate cognitive/perceptual structures in the brain that would recognize and represent different aspects of motion events, and other structures that would integrate those representations. Maintaining the locomotion/path separation in gesture and language could serve a number of functions. First, it could be conceptually easier to produce expressions that match one's internal representation of the event. If the process of perception requires this split, then it might be simpler to match that structure than to combine the pieces into a single representation. This would fit with McNeill's observation that combined Manner/Path gestures develop after age 12, when a child is cognitively ready to integrate these elements. Second, for similar reasons, separated representations might be easier for the addressee to understand. Why create a complex fully integrated classifier when the addressee must break it down again in the comprehension process? Presenting a separated, "pre-digested" form instead would allow the addressee to build up a model of the event gradually, from simpler conceptual pieces. In addition, the repetition of "old" information with new information would help him or her reassemble the separated pieces into a coherent whole. The serial and sequential constructions noted for ASL may in fact be better suited to communication than a single mimetic portrayal of all the aspects of a complex event. This hypothesis is supported by the fact that our only example of a highly complex form follows and repeats the information previously given in two simpler forms. Finally, separating out pieces of information allows the expresser to direct the addressee's attention to particular pieces. As Talmy (1985) noted, information conflated into one lexical item is "backgrounded" and not available for focus; compare I flew to Hawaii and / went by plane to Hawaii - in the second, the mode of transport is available for emphasis. Thus, if signers and gesturers can present particular aspects of the scene in isolation, they can emphasize those aspects. To sum up, ASL encodes a great deal of conceptual information about motion events, significantly more than English and presumably more than most other spoken languages. This information is largely presented through referential shift and iconic classifier forms. Yet the iconic forms do not usually give a fully simultaneous portrayal of all aspects of complex motion events. Instead, many aspects (Manner of locomotion, details of Path, different Figures, and Ground elements) are routinely separated into different pieces and presented sequentially;

214


moreover, some aspects are presented repeatedly to provide coherence to the separated presentation. This type of separation occurs both in signed languages and in gestures accompanying spoken languages; it may be a consequence of our cognitive and perceptual resources for understanding motion events rather than of arbitrary language-specific rules.

APPENDICES Appendix 1. English transcription English transcription of one scene from Frog, where are you? Adult male. Pictures 16, 17, and 18 (Appendix I). Longer pauses are represented by line breaks, shorter by dashes. the deer takes off with - the boy strewed across his antlers and the dog runs at his feet yelling at him to - to stop it. um - they're approaching a cliff and the deer - stops abruptly which causes the boy to lose his balance and fall with the dog down into the stream um - or a little puddle.

We would like to thank Virginia Marchman, Tanya Renner, and Dan I. Slobin at the University of California at Berkeley for the use of these data. These data are currently available through the CHILDES database.


215

Appendix 2. ASL transcription ASL transcription of one scene from Frog, where are you? Adult male, Pictures 16, 17, and 18. Right Hand

Both Hands

cl:v (on top of head) animate on head

Left Hand

Perspective

cl:y (on side of forehead) antlers

narrator

Translation

'The boy is on the deer's head.' narrator

cl:5 (to side of head) deer cl:vv fourlegged animate is running

deer

'The deer is running.'

BOY

narrator

cl:v (on head) animate on head STUCK

boy (scared)

cl:vv fourlegged animate is running 'The scared boy is stuck on the running deer's head.'

DOG

narrator BARK (dir: up and to side)

dog narrator

SAYS GETDOWN imper GETDOWN. mper (dir: up & side)

dog (concerned/ worried about boy)

'The concerned dog barks up at the boy and says "Get down, get down." ' narrator

DEER cl:vv fourlegged animate runs and runs TO LAND cl:bb 'flat land to cliff'

deer

narrator

'The deer is running to the cliff.'

216

Right Hand


Both Hands

Left Hand

clrvv (same location as the cliff) fourlegged animate stops at cliff STOP

Perspective

Translation

deer (scared)

'The scared deer screeches to a halt at the edge of the cliff.' narrator boy (scared)

BOY

cl:vb (same location of the deer's stop and cliff) animate falls off*

'The scared boy falls off the deer's head and over the cliff.' WITH

narrator

cl:vv (at cliff location) fourlegged animate falls

dog

DOG

DOESNTNOTICE

narrator clrvv (at cliff space) fourlegged animate falls**

dog

'With the dog. the dog falls. without noticing (the cliff) the dog falls." cl:v (at head) animate falls from head

clrv (at cliff location) animate falls from cliff

boy ( scared )

cl:vv two animates intertumble 'As the boy falls from the head of the deer, the dog falls from the cliff, resulting in the two of them inter-tumbling until they land, with the dog on the boy.'*** ; form. *** Figure 3 represents this entire line.


21'

Appendix 3. ASL transcription conventions The glosses of ASL in Appendix 2 are given in five categories. Normally, these are arranged in lines, with an implicit time axis running from left to right. In this case, however, the time axis has been set to run from top to bottom, in order to reduce line breaks and retain readability. The five categories appear as column headers. The three leftmost columns describe what the signer's hands are doing: the first shows the right hand when it acts alone, the second shows what the hands do together, and the third shows what the left hand does on its own. The fourth column gives the persona assumed by the signer: narrator, boy, deer, or dog. Personas' affect is sometimes described in parentheses. The fifth column gives the English translation. Lexical items are glossed by English words in CAPITAL letters (e.g., DOG). A dash joining two English words indicates a multi-word gloss of a lexical sign (e.g., DOESN'T-NOTICE). The subscript IMPER indicates that the verb has the imperative inflection. Classifiers' glosses consist of the letters "cl", followed by an abbreviation for the classifier handshape, followed by a description of the classifier's meaning. For some classifiers, a parenthetical note gives additional placement information. For example, cl:y (on side of forehead) antlers denotes a Y-shaped classifier placed at the side of the signer's forehead to indicate antlers. Classifier handshapes appearing in this transcript include: v y 5 b

index and middle fingers extended from a closed fist; fingers may be curved or straight; thumb and pinky finger extended from a closed fist; all fingers and thumb extended and spread; all fingers extended and touching; thumb may touch fingers or be extended outward.


The Many Ways to Search for a Frog Linguistic Typology and the Expression of Motion Events

Dan I. Slob in

1. INTRODUCTION The chapters in this volume, along with the extensive list of frog-story studies in Appendix II, provide a rich database for the exploration of particular questions of language use and acquisition. The studies reported in Part I reflect a range of languages of different types, making it possible to focus on the role of linguistic typology in narrative construction.1 A recurrent concern in those studies is the expression of motion, which is one of the dominant themes of Frog, where are you? In one way or another, all of the studies confront Talmy's by now familiar typology of verb-framed and satellite-framed languages (Talmy 1985, 1991, 2000b). Briefly, the typology is concerned with the means of expression of the path of movement. In verb-framed languages ("V-languages") path is expressed by the main verb in a clause ('enter', 'exit', 'ascend', etc.), whereas in satelliteframed languages ("S-languages") path is expressed by an element associated with the verb ('go in/out/up', etc.). This dichotomy has engendered a good deal of research and debate in the literature on motion-event descriptions over the past decade or so.2 In this concluding chapter on typological perspectives I suggest that several different sorts of factors "conspire" to produce a range of frog-story varieties. These varieties result from combined influences of linguistic structure, on-line processing, and cultural practices. Talmy's typology was designed to characterize lexicalization patterns, and it has provided important insights into the overall set of structures that define individual languages. However, the typology alone cannot account for discourse structures, because language use is determined by more than lexicalization patterns. It is striking how much has been learned by application of the V-language/S-language contrast, and it still plays a part in the mix of factors considered here. But a fuller account of narrative organization will require attention to a range of morphosyntactic, psycholinguistic, and pragmatic

220

Dan I. Slobin

factors. Some of these factors are explored in this chapter, with regard to motion events in the frog story. The aim is to come to a fuller explanation of the ways in which languages differ in rhetorical style. In order to arrive at this goal, the chapter examines three major components of motion events across languages: manner of motion, path of motion, and grounds (landmarks). Talmy's typology casts light on crosslinguistic differences in the expression of each of these components in narrative. However, in order to account more fully for these differences, a number of additional factors present themselves. A language provides its speakers with a range of ways of describing motion events - combinations of lexical items and grammatical morphemes in various construction types. In attempting to define patterns of rhetorical style, it is also necessary to consider language-processing capacities (short-term memory, cognitive load), paralinguistic factors (voice quality, gesture), cultural practices, and modality (speech, writing, sign language). In learning a particular language, one develops habitual patterns of organizing all of these factors in guiding attention to dimensions of experience for narrative expression. On the basis of detailed examination of frog-story data, a revision of Talmy's typology is proposed. It is also proposed - with regard to manner of motion - that it is more useful to rank languages on a cline of manner salience than to allocate them to one of several distinct typological categories. This is because of the finding that a number of factors contribute to the degree of salience of manner in a language. In the first volume of frog-story studies (Berman & Slobin 1994) there were three V-languages (Spanish, Hebrew, Turkish) and two S-languages (English. German). Berman and Slobin proposed the following typological contrasts between narratives in these languages: Satellite-framed languages allow for detailed description of paths within a clause, because the syntax makes it possible to accumulate path satellites to a single verb, along with prepositional phrases that add further specification (e.g.. the deer threw them off over a cliff into the water). The satellite-framed languages in our sample also tend towards greater specification of manner, probably because the lexicon provides a large collection of verbs that conflate manner with change of location (crawl, swoop, tumble, etc.), often conflating cause as well (dump, hurl, shove, etc.). In verb-framed languages, such elaboration is more of a "luxury," since path and manner are elaborated in separate expressions, which are generally optional, and which are less compact in form [e.g., 'exit flying (from the hole)' vs. 'fly out (of the hole)']. As a consequence of these differences, it seems at least in our data that English and German narrations are characterized by a great deal of dynamic path and manner description, while Spanish. Hebrew, and Turkish narrations are less elaborated in this regard, but are often more elaborated in description of locations of protagonists and objects and of endstates of motion. (Berman & Slobin 1994:118-119)

The Many Ways to Search for a Frog

221

These patterns are now reflected in a much greater range of languages of the two types; but there is more to the story.3 Consider "the owl episode" in Frog, where are you?, attending only to the boy and the owl as protagonists. There are two pictures (see Appendix I, Pictures 11-12). In the first, the boy is in a tree, straddling a branch, looking into a hole. In the second, an owl with extended wings is located in the hole; the boy is lying on his back on the ground, with arms and legs extended upwards. From these two pictures - building on the previous ten pictures and accessing world knowledge - a complex physical event can be interred: /'. //. ///. iv.

boy boy owl boy

climbs tree; looks into hole; emerges from hole; falls from tree to ground.

Based on plot comprehension and world knowledge - both physical and psychological - a complex causal event can be inferred: a) i and ii are motivated by the boy's search for the runaway frog; b) the boy's action has aroused the owl; c) the emergence of the owl (///) causes the boy to fall from the tree (iv). This much is shared by all mature "readers" of the pictures. However, the accumulated collections of frog stories show tremendous diversity in the ways in which this event is narrativized across ages and languages. Because the goal of this chapter is to focus on crosslinguistic and typological issues, most of the examples will be drawn from adult narratives. That is, we will primarily focus on developmental "endstates" - the language-particular goals of learning. To simplify the task, consider two motion events: the emergence of the owl and the fall of the boy, along with their associated locations with regard to the tree. Let us begin with English. The owl episode can be related in a matter-of-fact way, as in the following example from an American English-speaking adult (leaving out intervening clauses that deal with the behavior of the dog and the bees): [1]

the boy climbs on top of the tree the hole the boy was looking into was the home of an owl and the owl came out of the hole and scared the boy the boy fell off the tree and landed on his back

Every clause in this narrative segment presents inferred information, that is, information that goes beyond the two static pictures. Verbs are used to move the protagonists: climb, come out, fall, land. The ground elements are labeled with

222

Dan I. Slobin

nouns: tree, hole. We have to infer that the hole is in the tree; and we are not told why the owl emerged or why the boy fell, though we can draw an inference about the latter event from the verb scare. This, of course, is not the only way of narrating the owl episode in English. Another American English-speaking adult is rather different, especially in the use of verbs of motion: [2]

after that he continues and climbs up on this tree and looks in this hole for the frog the boy tumbles down from the branch because of an owl who's popped up from the hole

The verbs that move the protagonists provide additional information about manner of motion: tumble, pop up (in addition to climb, shared with the first version). Again, we have to infer that the hole is in the tree; but we're told that the emergence of the owl somehow caused the boy to fall (because of). Turning now to Spanish, we find narratives that are similar to [1]. The following is from a Chilean adult: [3]

el nino sigue buscando en un arbol que tiene un hueco dentro de ese arbol vacio aparece un buho que asusta al nino y el nino se cae

the boy continues searching in a tree that has a hole inside of that empty tree appears an owl that scares the boy and the boy falls

However, there are no Spanish equivalents of [2] in frog stories gathered in Chile, Argentina, and Spain. In Spanish, the boy 'ascends' (subir) the tree; the owl 'appears' (aparecer, asomarse) or 'exits' (salir), and the boy 'falls' (caerse). There is no climbing, clambering, popping out, flying out, tumbling, crashing to the ground, or the like. That is, the languages differ in their attention to manner of movement. By contrast, the Spanish stories often are precise in locating the ground objects, as in the description in [3]: un arbol que tiene un hueco 'a tree that has a hole'. Other examples are: uno de los huecos que hay entre las ramas del arbol 'one of those holes that there are between the branches of the tree'; un arbol que tenia un buho dentro 'a tree that had an owl inside'. English often leaves the locations of ground elements to inference, as in [1] and [2], or makes use of compact phrases to convey such locative information: the hole in the tree, or simply a tree hole. Often it is the owl that is the foregrounded reference point, rather than the tree: there was only an owl in the big old hollow tree.


223

As discussed in "Frog I" (Berman & Slobin 1994) and elsewhere (Slobin 1996a, 1997a, 2000, 2002b), languages differ systematically in rhetorical style - that is, the ways in which events are analyzed and described in discourse.4 And, as I have proposed earlier, these differences are grounded in typological characteristics of morphosyntax and lexicon, often allied with cultural narrative practices. These themes are repeated in many chapters in "Frog II" (the present volume). By now we have enough frog-story data, across a range of languages and cultures, to begin to extend and revise the typological picture presented in Frog I. I want to propose that rhetorical style is determined by the relative accessibility of various means of expression, such as lexical items and construction types. That is, ease of processing is a major factor in giving language-particular shape to narratives. At the same time, cultural practices and preferences reinforce habitual patterns of expression. The picture is complex - as we will see - because various options compete or conspire to provide the overall shape to narrative production. And because the various means of expression are not completely obligatory, the data from each language present a range of possible narrative patterns. This chapter tracks these issues across three major components of motion events: manner of motion, path, and ground elements (landmarks). We begin with the dominant typology presented in Frog I and Frog II: Talmy's verb-framed and satellite-framed dichotomy. As noted at the outset, however, it will become evident that this dichotomy is only part of a complex system of interacting factors.

2. MANNER OF MOTION5 The typology can be approached from two angles: How is manner of motion expressed, and how is path of motion expressed? The expression of path defines a motion event in the typology. In Talmy's definition (2000b:25), "[t]he basic Motion event consists of one object (the Figure) moving or located with respect to another object (the reference object or Ground)."6 Path, then, is obligatory; but manner is optional and can range from rather general manners, such as walk, run, and fly, to quite specific distinctions, such as limp, sprint, and swoop. Taking manner as a starting point, it appears that languages vary considerably with regard to this dimension, with V-languages paying much less attention to manner than S-languages, for a variety of reasons. 2.1 The emergence of the owl across languages We begin with the emergence of the owl in five languages categorized by Talmy as V-languages (Spanish, French, Italian, Hebrew, Turkish) and five categorized as S-languages (English, German, Dutch, Mandarin, Russian).7 In V-languages,

224

Dan I. Slobin

narrators of all ages almost always describe the appearance of the owl with a single path verb, meaning 'exit'. For example: [4]

V-languages Spanish: Sale 'Exits French: D'un 'From a Italian: Da 'From Turkish: Oradan 'From Hebrew: Yaca 'Exits

un buho. an owl.' trou hole

de 1'arbre of the tree

sort exits

quest' that

albero esce tree exits

un gufo. an owl.'

bir there

un hibou. an owl.'

baykus an owl

mitox from inside

haxor the hole

yansuf. owl.'

By contrast, many S-language narrators, at all ages, use some kind of manner verb together with a path satellite to add some sort of dynamic information about the owl's emergence. For example: [5]

S-languages English: An owl popped out. German: ... weil da '... because there Dutch: ... omdat er '... because there Russian: Tam vv-skocila 'There out-jumped Mandarin: Fei chu yl zhT 'Fly out one

eine Eule an owl

plotzlich suddenly

een uil an owl

uit-vliegt. out-flies.'

sova. owl.' maotouyTng. owl.'

raus-flattert. out-flaps.'


225

Figure 1. Percentage of narrators using a manner-of-motion verb for the owl's emergence. Spanish, French, Turkish, Italian, and Hebrew are V-languages; Dutch, German, English, Mandarin, Thai, and Russian are S-languages.

Figure 1 shows the percentage of narrators using a manner verb to describe the appearance of the owl, adding data on Thai (Zlatev & Yangklang, this volume). The differences between the two language types are dramatic. Manner verbs are simply not used in the V-language descriptions (with two exceptions, discussed below). On the S-language side, however, there are obvious differences between the relatively low use of manner verbs by the three Germanic languages, the higher use in Mandarin and Thai, and the much higher use in Russian. Here we have the first hint that rhetorical style is based on more than Talmy's two-way typology. (As suggested later in the chapter, it seems more useful to treat these data as representing a continuum of manner salience.) 2.2 Lexical availability and construction types All of the languages in this sample have frequent manner verbs that are applicable to this scene - in particular, versions of 'fly'. Why isn't 'fly' used in the V-languages? The answer probably lies in what Slobin and Hoiting (1994) have called the "boundary-crossing constraint," building on Aske's (1989) identification of the role of telicity in the use of manner verbs in Spanish. It appears that V-languages only license the use of a manner verb as a main verb in a path expression if no boundary-crossing is predicated. Thus it is possible,

226

Dan I. Slobin

across a range of V-languages, to say the equivalent of 'fly to/from the tree* but not 'fly out of the hole'. Perhaps the most salient characteristic of V-languages is the preference to mark a change of state with a verb, rather than by some other device. With regard to motion events, changes of state are boundary-crossing events: enter, exit, cross. In order to add manner to such events, some sort of subordinate construction is required - equivalents of 'exit flying'. Although this option is available in the five V-languages considered here, it was not taken by any narrator, at any age (from three years to adulthood). There are at least two sorts of explanations for this avoidance: a) the construction unnecessarily foregrounds the owl's manner of movement and b) it is "heavy" in terms of processing (production/comprehension). Typically, in V-languages, a neutral verb of motion is used to designate a creature's normal manner of movement: owls 'go', fish 'go', people 'go', cats 'go', and so forth. Manner verbs are used when manner is foregrounded - and then owls can 'soar' or 'flap' (but apparently not across boundaries). The only exception seems to be verbs that encode particular force dynamics - highenergy motor patterns that are more like punctual acts than activities, such as equivalents of 'throw oneself and 'plunge'. Such verbs occur with boundarycrossing in V-languages. This may be because a sudden boundary-crossing can be conceptualized as a change of state, and what is apparently most characteristic of V-languages is the use of verbs to encode change of state. What seems to be blocked is the conceptualization of manner of motion as an activity that is extended in time/space while crossing a boundary (Kita 1999). That is, boundarycrossing is a change of state and manner verbs are generally activity verbs. The only manner verbs that can occur in boundary-crossing situations are those that are not readily conceived of as activities, but rather as "instantaneous" acts. Thus one can 'throw oneself into a room' but one generally can't 'crawl into a room" in V-languages. With regard to the frog story, all of the V-language narrators focused on the owl's emergence or appearance, with an occasional adverbial indication of suddenness, rather than focus on the activities of flying or flapping out. In the entire corpus, there are only two instances of manner verbs in this scene, and. interestingly, neither of them is a boundary-crossing construction. An Italian five-year-old said: // gufo void, il bambino cased 'the owl flew, the boy fell': and an Israeli adult said: yansufkofec meha'ec 'owl jumps from:the:tree'. Note that these are both simple clauses, in which the manner verb is the main verb and no boundary-crossing is expressed. There is no compact construction that allows for simultaneous attention to the owl's sudden appearance, its emergence across a boundary, and its manner of movement. As a consequence, it seems that V-language speakers opt to encode change of state, i.e.. in/out or non-visible/ visible.


227

S-languages, by contrast, do provide such compact expressions, as shown in the examples in [5]. Why, then, is the manner option used relatively infrequently in the Germanic languages (Dutch 17%, German 18%, English 32%)? I suggest that a focus on the owl's emergence predominates in these languages as well. The most common expression is 'come out', thus taking the viewer's perspective and predicating appearance using a readily available expression. In order to add manner to the perspective, speakers of Dutch, German, and English are in the same position as speakers of V-languages: they would require a heavier construction, such as 'come flying out'. Instead, there is a tendency to pick one of two options: 'fly/pop/jump out' or 'come out'. Note that these two options are equally processible: Each has a main verb plus a path particle, and they are apparently easily accessible. Russian presents a different lexicalization pattern. There is no independent verb that is the equivalent of 'come'; rather, a deictic prefix on a motion verb is needed for the expression of motion towards the speaker's perspective. All path particles (satellites) are also verb prefixes in Slavic languages, and prefixes can't be stacked; so there is no way to combine 'come' and 'out' with one verb, as in Germanic. One has to choose betweenpri-letet' 'come-fly' and vy-letet' 'out-fly'. The deictic option (pri-letet') was taken by 11% of the Russian narrators of the owl scene; the remaining 89% focused on the owl's emergence, using vy- with verbs meaning 'fly' (vy-letet'), 'jump' (vy-skocif), and 'crawl' (vy-lezit'). Again, narrators chose a simple construction with a single verb. Note, however, that both options use a manner verb - hence the 100% of Russian manner verb choices in Figure 1. Thus it is not satellite-framing alone that accounts for the rate of use of manner verbs; morphosyntactic structure also plays a role in determining rhetorical style. 2.3 Serial-verb languages Mandarin Chinese has been categorized as an S-language by Talmy. It is a serialverb language in which each verb in the series is morphologically unmarked and monosyllabic. Talmy considers the manner verb to be the main verb and the path verb to be a satellite, because path verbs often do not function as full verbs and because there is a small, closed set of path verbs. However, the two elements in the Mandarin example of [5] are both verbs: fei chu 'fly exit'. Such constructions are highly frequent and are probably easy to process; it is not surprising that they are used by 40% of Mandarin-speakers to narrate the owl's emergence. However, there is a clear developmental trend: whereas narrators of all ages use a path verb, path-manner combinations are never used by three-year-olds, are used by 22% of children aged four to seven, and by 73% of nine-year-olds and adults. Similarly, in Thai, only at age 11 do children use path-manner combinations at

228

Dan I. Slobin

a comparable rate to adults (83% for eleven-year-olds, 86% for adults). There is apparently a cost to adding manner information, especially for the younger narrators. However, deixis seems to be more closely tied to conceptions of path. This is suggested by the fact that most speakers add a final 'come' verb, filling a standard serial-verb construction type, as in Mandarin fei chu lai 'fly exit toward.speaker'. The Mandarin and Thai data raise a problem for Talmy's treatment of path verbs as satellites in serial-verb languages. In these languages both 'exit' alone and 'exit come' are options used in the owl scene; that is, path can be expressed by verbs. These findings make it difficult to follow Talmy in placing such languages in the S-language category.8 Slobin and Hoiting (1994) proposed to treat such languages as "complex verb-framed languages": verb-framed because path is expressed by an independent verb, and complex because the serialverb construction functions as a sort of compound main verb in a clause, with no division between finite and non-finite forms as in "standard" verb-framed languages, which require constructions such as 'exit flying'. That is. manner is not syntactically subordinated to path in serial-verb languages. Because the path verbs can occur alone, they cannot be regarded as satellites, which are verb particles and affixes that do not occur alone. But because manner verbs occur freely with boundary-crossing events, as in the owl scene, these languages do not fully pattern with V-languages either. At present, it may be most appropriate to treat serial-verb languages as a third typological category with regard to motion events. This is the proposal of Zlatev and Yangklang (this volume): and Ameka and Essegbey make the same proposal for West-African languages, based on a large set of criteria: When the properties are tallied, we find that serialising languages share more properties with S-languages than with ...V-languages ... while still possessing a unique property. What this shows is that they cannot be said to belong to either type. Instead, they appear to belong to a class of their own. (Ameka & Essegbey. in press)

I agree, and propose later in the chapter that a third type be added, equipollentlyframed languages, to include serial-verb languages and other types of languages in which both manner and path are expressed by "equipollent" elements - that is. elements that are equal in formal linguistic terms, and appear to be equal in force or significance. (Furthermore, I will suggest that the facts of manner expression are better dealt with by placing languages on a cline of manner salience, rather than placing them into dichotomized or trichotomized typologies.)


229

2.4 Attention to manner: Narrative style Analysis of the emergence of the owl has proven to be fruitful in revealing a collection of factors that can influence narrative style. The requirement that a path verb be used to express boundary-crossing in V-languages does not block the expression of manner in a subordinate construction with a path verb, such as 'exit flying', but the added processing load apparently militates against the choice of that option unless manner is at issue. That is, the speakers of the five V-languages in this sample could have said such things as 'the owl exited flying', but they chose not to. On the other hand, although S-languages provide speakers with a compact way of saying 'the owl flew out', choice of this option is determined by competing factors. In a scene such as this one, where one can focus on either the owl's exit from the hole or the owl's moving toward the narrator's perspective, speakers of Germanic languages pick one option or the other, apparently because either a deictic verb or a manner verb can be used in the standard construction: come out or fly out. In Slavic languages, by contrast, with no independent 'come' verb, and no way of stacking verb prefixes, both options push for a manner verb: 'come-fly' or 'out-fly'. And in serial-verb languages, where both path and manner are of equal morphosyntactic weight, older speakers tend to use both, whereas younger speakers begin with more attention to path; and speakers of all ages also tend to add an additional verb for deixis. In sum, lexicalization patterns interact with ease of processing and available construction types in determining narrative style - for this simple scene, as well as across the frog story and other types of narrative. As a demonstration of the overall degree of attention to manner in a language, Figure 2 compares English with two different sorts of V-languages, Turkish and Spanish. The figure is based on descriptions of all of the many sorts of motion events in the frog story. The data show, for each age group, the percentage of motion events that were described by a manner verb - either alone or in combination with a path verb, and including transitive and intransitive verbs of motion. Two things are striking in these data: a) Turkish and Spanish are roughly equal in their use of manner verbs whereas English is higher; b) these patterns are true at all ages. That is, the lexical encoding of manner is a stable characteristic of a language, acquired early and maintained through life. (In a larger sample, Hebrew patterns like Turkish and Spanish, and German, Dutch, Russian, and Mandarin pattern like English; see Oz9ahskan and Slobin 1999.) The crosslinguistic patterns are typical of creative fiction as well as pictureelicited narratives. Figure 3 compares frog stories with novels in the same three languages - English, Turkish, and Spanish - choosing adult frog stories and intransitive motion verbs for comparison. It is clear that the narrative styles that are revealed in oral narratives elicited by a picture storybook are the same as those

230

Dan I. Slobin

Figure 2. Percentage of motion events described by manner-of-motion verbs (transitive intransitive tokens) in frog stories.

produced by creative writers. Thus, adult frog stories appear to provide a good sample of the overall narrative style of a language. 2.5 Conflation patterns Most of the discussion in the literature on motion events contrasts pure path verbs, such as 'exit' and 'ascend', with pure manner verbs, such as 'run' and 'crawl'. Using such verbs, it is possible to delineate the sorts of typological contrasts we have been considering so far. Talmy's typology is based on the most characteristic and frequent lexicalization pattern used by a language - that is, V-languages tend to use path verbs with subordinate manner expressions and S-languages tend to use manner verbs with associated path forms. However, languages also have minor patterns that may play a role in narrative style. With regard to the domain of motion, some verbs conflate path and manner. An example of such a verb in the frog story is Turkish tirmanmak 'climb'. Unlike English climb, which is neutral with regard to path (one can "climb down from a tree" or "climb out on a branch," for example), the Turkish verb is used only for upward motion in a grasping manner. (The same is true of equivalent 'climb' verbs in other V-languages.) The prelude to the owl scene in Turkish allows us to compare a simple path verb. fikmak 'ascend', with a manner-path verb, tirmanmak 'climb' (Ozcahskan & Slobin 2000a). If ease of processing is a major factor contributing to narrative style, Turkish-speakers should describe the boy's ascent of the tree with either verb and should avoid the complex subordinated form, tirmanarak qikmak


231

Figure 3. Percentage of motion events described by manner-of-motion verbs (intransitive tokens) in adult novels and frog stories.

'by.climbing ascend'. This is, in fact, what they did. None of the Turkish narrators used the 'climbing ascend' construction; thus they opted for ease of processing, using a construction with a single verb. However, Turkish-speakers also operate with a habitual narrative style in which path verbs serve as main verbs. Whereas 88% of English-speaking narrators used the verb climb in this scene, 45% of the Turks described the ascent of the boy as 'ascend' and 55% used 'climb'. 9 It is significant that Turkish-speakers attended to combined path and manner about half of the time, picking a readily available lexical item that conflates the two dimensions; nevertheless, the habitual narrative style of attending to path still plays a role, since 'ascend' was also a frequent choice. It seems, then, that availability of a manner-path verb for a particular event type can lead V-language speakers to pay more attention to manner than they usually do, yet still less than S-language speakers. Linguistic habits are strong. 2.6 Beyond lexicon and morphology The distinction between verb- and satellite-framed languages is based on lexicalization patterns - in the case of motion events, the lexicalization of path as main verb or satellite. We have seen that the morphological expression of path elements also plays a role, comparing Germanic separable particles, Slavic prefixes, and verbs in serial-verb constructions. And minor conflation patterns, such as manner-path verbs, can blur the overall typological distinction when considering the encoding of specific events. However, more is involved in

232

Dan I. Slobin

discourse organization than lexicon and morphology. Here we consider several additional dimensions: adverbial and phrasal expressions of manner outside of verbs, ideophones, and gesture. Alternative descriptions of manner Perhaps V-language narratives are rich in descriptions of manner of motion, but we have been looking in the wrong place. Speakers of Turkish, Spanish, and Hebrew avoid subordinate constructions such as 'exit flying', but they may use various adverbial and other descriptions to add information about such dimensions as suddenness, rate, force dynamics, inner state, terrain, and so forth - that is, information about factors that suggest manner of movement. Ozcahskan and Slobin (in press-b) examined alternative manner expressions in English and Turkish frog stories, as well as novels. English- and Turkish-speaking narrators use such expressions at comparable rates, but for different purposes. In Turkish. 61% of the additions occur with non-manner verbs, indicating that they are often used to "compensate" for the absence of manner in path constructions. For example, a Turkish adult adds 'silently' in describing how the boy 'approaches' the tree trunk at the end of the frog story: [6]

Bir agac kutugune sessizce yaklasti. '(He) silently approached a tree trunk/

By contrast, in English 73% of additions qualify manner verbs, thereby augmenting attention to manner. For example, an Australian English-speaking adult has already chosen a highly descriptive manner verb, slink, but uses adverbs to further elaborate the manner of motion: [7]

Fido very very quietly slunk out of the water.

Apparently, English-speaking narrators are often not satisfied with their already rich lexicon of manner verbs, and add more richness by other means. In addition to qualifying manner of movement, English frog stories and novels provide relatively more details of protagonists' inner states and conditions of the terrain - that is, information that suggests nuances of manner. In sum, even when considering alternative expressions of manner, S-languages texts still show relatively greater attention to manner, in both quantitative and qualitative terms. Posture and manner Brown (this volume) presents data from Tzeltal, a Mayan language with a very large lexicon of positionals, that is "verbal roots which convey Position of animate or inanimate things (in stasis, or concurrent-with, or as-a-result-of motion)." Positionals are acquired early, and play a role in Tzeltal frog stories


233

which is, in some ways, comparable to that of manner verbs in other languages. For example, xpejkunaj indicates a low-crouching posture. Combined with xben 'walk" it conveys a limping manner of movement: [8]

xpejkunaj xben hilel 'He [dog] looks like he's low-crouching walking.'

In other instances, by describing a postural endstate, a positional can suggest manner; e.g.: [9]

j ipot jawal ta lum 'He [boy] has been thrown lying-face-upwards- to the ground.'

Brown suggests that "one reason for the proliferation of Position in frog stories is to provide some of the graphic visual detail for which, in satellite-framed languages, manner-of-motion verbs are used." Again, it seems necessary to expand the binary typology. Tzeltal seems to be a V-language, with its reliance on path verbs and a small lexicon of manner verbs, but positionals add another expressive dimension, not found in the V-languages we have been considering. (Tzeltal also has directional morphemes, discussed below with regard to path expressions.) Ideophones expressing manner There are occasional examples, in all of the languages under study, of the use of sound symbolism to heighten the drama of the frog story. These are found especially in the culminating fall from the cliff into the water: [ 1 Oa] Splash! They both fall into the water. (English adult) [ 1 Ob] Birlikte cumbadanak bir suya diisuyorlar. 'Together cumbadanak they fall into some water!' (Turkish adult) [lOc] jZas! Se cayeron al suelo. 'Zasl They fell to the ground.' (Spanish adult) Ibarretxe-Antunano (this volume) pays special attention to sound symbolism in Basque, where it plays a role in describing manners of motion. Basque is a Vlanguage with a large lexicon of ideophones that serve as "movement imitatives"; in the frog story, speakers use plisti-plasti 'waddle' and taka-taka 'walk with small and short steps, creep, crawl'. In addition, there are ideophones expressing acoustic dimensions, inner states, and so forth, which can serve to suggest manner

234

Dan I. Slobin

of motion. The following example parallels the use of adverbs in [6] and [7] - but note that there is no verb of motion; rather, the reduplicated ideophone indicates quiet walking. [11] eta han

isil

isilik

and there silence

hara

silence:pART there

nola hasi

dira

lurretik

how

MX

ground:ABL

start:PERF

'and there they go, they start [to walk] very quietly along the ground'

Ideophones are widely attested in the languages of the world (Voeltz & KilianHatz 2001), especially in West Africa, East Asia, and Southeast Asia. Clearly, they provide an important option for expressing manner of motion, though they are not highly frequent in the languages of the frog-story sample considered here. Such forms have been well described for another V-language, Japanese ("mimetics"), where they appear to function as in Basque (Hamano 1998; Kita 1997). For example (Hamano 1998:148): [12] doya-doya noisily

haiQ-te-kuru come.in-GER-come

'a group of people came in noisily' Similar examples are found in some Turkish frog stories, e.g.: [13] suyun waterGEN

icine

gumbiir giimbiir

du§uyor

interiorDAT

splash

come.in-GER-come

'(he) falls splash into the water' (Turkish adult)

Word order may play a role, too, in favoring the use of ideophones. Basque. Japanese, and Turkish are all verb-final languages, and in all three, the ideophone occurs early in the sentence. Processing strategies in such languages require the accumulation of information about participants and locations while waiting for the final predicate to tie the information together. This "rightward slant" in processing might favor the development of ideophones or mimetics. placed early in the sentence. That is, information about manner in the form of ideophones might easily fit into an on-line packaging strategy. Ideophones also appear to be frequent in serial-verb languages, where they may fit in with strategies for accumulating verbs into larger meaning units. Ease of processing may thus play a role in facilitating attention to manner by means of sound symbolism, just as it apparently plays a role in S-languages by means of a main-verb slot for manner verbs.


235

It remains to be determined whether recourse to sound symbolism allows a language to elaborate the domain of manner of motion in ways that parallel the large lexicons of manner verbs in the S-languages we have examined. Examples from various languages suggest that ideophones can provide a rich lexicon of manner distinctions, e.g.: gulukudu 'rush in headlong' (Zulu), minyaminya 'stealthily' (Ewe), kitikiti 'at-a-stomp' (Emai), widawid 'swinging the arms while walking' (Ilocano), badi badi 'waddling' (Turkish), dengdeng 'tramping' (Mandarin), tyoko-maka 'moving around in small steps' (Japanese).10 The availability of ideophones thus adds a further revision to the V-language/Slanguage typology, at least with regard to the expression of manner of motion. Gestural depictions of manner Ibarretxe-Antunano finds that ideophones in Basque are often accompanied by gestures; Kita (personal communication) reports the same for Japanese. In [11], above, there is no motion verb at all, but the ideophone isil isilik, indicating quietness, is followed by gestures that depict both manner ("walking" movements of the hands) and path (moving away from speaker). Manner gestures also occur together with path verbs. For example, in the owl scene, a Basque adult said atera 'exit' while flapping his hands on either side of his head to depict flying. We do not have systematic data on manner gestures in frog-story narrations, but David McNeill and his colleagues (McNeill & Duncan 2000; Ozyiirek & Kita 1999) have studied co-speech gestures in narratives in several languages, elicited by an animated cartoon (Canary Row, with "Tweety the bird" and "Sylvester the cat"). English-speakers have the option of augmenting a manner verb with gesture. When they do so, they either provide a gesture that combines path and manner, thus augmenting the lexical expression of manner, or they gesture path alone. There are almost no instances of a path verb accompanied by a manner gesture. By contrast, in three V-languages - Spanish, Turkish, and Japanese - there are frequent occurrences of gestures that depict only path or only manner, as well as path-manner conflations. Manner gestures accompany path verbs as well as manner verbs. The findings suggest that V-language speakers conceive of manner as a separate element - a sort of activity - that can "augment" directed motion, whereas S-language speakers conceive of manner as an inherent component of directed motion (see Slobin 2000)." When more co-speech data are available, we may find that V-language speakers provide gestural manner information at comparable rates to the lexical manner information provided by S-language speakers.12 If so, yet another factor must be considered in predicting discourse patterns from the typology of lexicalization patterns. However, there are several important ways in which gesture and lexicon provide different tasks for language use and acquisition. The kind of information provided by gesture is analog and depictive, and does not require the speaker to attend to the sorts of categorical distinctions provided by lexical items (e.g.,

236

Dan I. Slobin

roll, bounce, bob, tumble, skid, bump, etc.). The categorical contrasts encoded by lexical items are language-specific and have to be learned; gestural depiction, by contrast, is much less categorical and conventional. Furthermore, if gestures do provide "compensation" for lexical gaps and for heavy constructions, this option is only available for oral narrative - that is, for speaking. The considerable differences between written texts in the two language types still remains (Slobin 1996a, 1997a). There are also prosodic expressions of manner, which have not yet been explored in detail. Written texts are poor in the representation of prosodic marking of force dynamics by such devices as shifts in volume, rate, and pitch. Gesture and prosody are features of oral, rather than written, language. Typologies tend to be developed on the basis of written materials - sample sentences, texts, transcriptions of speech. It may well be that the two types of language production require different sorts of typological accounts.13 2.7 Beyond spoken language Galvan and Taub (this volume; see also Taub & Galvan 2000) compare frog stories in English and in American Sign Language (ASL). Because space is used to represent space in signed languages, motion verbs always move from one place to another - that is, the verb is inherently a path verb. However, owing to the nature of the manual modality, the moving hand can also represent the type of figure as well as manner of motion. For example, a downward V-handshape indicates a two-legged being and wiggling fingers indicate moving legs. In addition, independent manner verbs, using the whole body, can express such notions as running, walking heavily, and so forth, while the face can indicate force dynamics and affect. Also, as in Tzeltal, the configuration and orientation of the hand can convey the posture and final orientation of the moving figure. Slobin and Hoiting (1994) had proposed that signed languages are "complex verb-framed languages": verb-framed because of the use of path verbs; complex because of the simultaneity of path, figure, and manner, as well as the serialization of manner and path verbs. Galvan and Taub provide detailed evidence for a high degree of simultaneity in ASL frog stories, along with discourse-based patterns that are reflected in detailed attention to both path and ground elements. They also document what appears to be a boundary-crossing constraint in ASL. They suggest: [W]e might conclude that if ASL is a verb-framed language, it is highly atypical. Pending future studies, we maintain neutrality on this issue - it may even be the case that we need to expand Talmy's typology to accommodate languages which can conflate Path, Manner, and Figure into a single verb form. (Galvan & Taub. this volume. Section 4)


237

It seems appropriate to me to accept Galvan and Taub's suggestion. The visual modality of hands, face, and body offers possibilities that are quite different from the acoustic modality of the voice or the combined modality of speech and gesture. 2.8 Codability and habitual expression The term codability sums up the various factors of accessibility or ease of processing examined above. Various factors make a concept such as manner of motion relatively more codable: /. //'. ///.

expression by a finite rather than non-finite verb; expression by a high-frequency rather than low-frequency lexical item; expression by a single word rather than a phrase or clause.

There is a synergy between these factors in S-languages which apparently leads speakers of those languages to pay special attention to manner. This is a domain that receives habitual expression - a domain that speakers and listeners expect to play a role in descriptions of motion events. Over time, such languages add more lexical items to the domain. English, for example, has gone on innovating mannerof-motion verbs century by century. For example, the following intransitive verbs of goal-directed human motion were added in Oxford English Dictionary in the nineteenth century: barge, career, clomp, cruise, dawdle, dodder, drag oneself, drift, flop, gambol, goose-step, hike, hustle, leapfrog, lunge, lurch, meander, mosey, pounce, promenade, race, sashay, scurry, skedaddle, skitter, slither, slog, slosh, smash, sprint, stampede, tromp, twist, waltz, wiggle, worm, zip. As a consequence, the semantic space of manner of motion in S-languages becomes more saturated, and more finegrained distinctions come to be made. Learners are faced with the task of attending to the details of motion that are relevant to the semantic space, and they end up with an increasingly differentiated conceptualization of manner. The same issues of habitual use and acquisition are probably found in languages with a rich lexicon of manner ideophones. All of these factors, of course, vary in degree from language to language; therefore patterns of discourse and conceptualization can't be predicted from a binary typology alone. With regard to Talmy's typology, however, I propose that verb-framing does not "suppress" attention to manner: manner of motion is too important for human beings to ignore. That is, speakers of all languages talk about manner when it is especially important. However, people are led to focus on and elaborate manner if they use a language with high codability in this domain, as provided by manner verbs in S-languages and serial-verb languages, and by manner ideophones (apparently in verb-final V-languages and serial-verb languages). As I have proposed elsewhere (Slobin 2000b, 2002b), a

238

Dan I. Slobin

highly codable domain plays a major role in thinking-for-speaking, with possible consequences for mental representation. To quote from Frog I: We have proposed (...) that frequent use of forms directs attention to their functions, perhaps even making those functions (semantic and discursive) especially salient on the conceptual level. That is, by accessing a form frequently, one is also directed to the conceptual content expressed by that form. (Berman & Slobin 1994:640)

3. PATH OF MOTION As noted at the beginning of the section on manner of motion, Talmy's typology can be approached from two angles, focusing on the means for expressing manner or path. It is path, however, that forms the core of the approach: in V-languages, path conflates with motion verbs, and in S-languages, path occurs as a "satellite" to a motion verb. Because path is an obligatory component of motion-event expressions, we can't compare languages in terms of the accessibility of path as a category: without a path verb or satellite or other path element, there is no motion event. However, languages differ with regard to the canonical segmentation of paths as well as the relative ease of building complex-path constructions. They also present an array of path elements going beyond the division into verb versus satellite. These differences are only partially determined by the Talmian typology. 3.1 Analysis and packaging of path components The events in the frog story pose two kinds of issues with regard to the expression of path: a) segmentation of an event into path components and b) distribution of path components in clauses of a narrative episode. The first issue is concerned with the degree of granularity of an event description, that is. how many sub-trajectories combine into an overall trajectory. Compare, for example, two hypothetical versions of the frog's escape at the beginning of the story, a schematic summary in [14] and an elaborated path in [15]: [ 14] the frog escaped [ 15] the frog exited the jar, passed through the window, and entered the w oods The second issue is concerned with tightness of packaging. For example, the version in [15] could be compacted as: [16] the frog crawled out of the jar and through the window into the woods


239

These two issues are related. Note that [15] uses three separate path verbs: exit, pass, enter. This is the required pattern in V-languages, because each of the events is a boundary-crossing event. (In fact, the constraint applies whenever a change of path direction occurs, e.g., approach and ascend the tree versus crawl toward and up the tree.)}4 By contrast, [16] uses a single manner verb, crawl, with three path satellites: out, through, in. This pattern is typical of S-languages (and is another factor contributing to the more frequent use of manner verbs in those languages). The following examples are not unusual in S-languages, though they are exceptionally rare in V-languages (in frog stories as well as novels): [ 17] they decided to walk outside the house down to the back of the garden out into the bit of a forest there (Australian adult) [ 18] plotzlich fallt der Hund aus dem Fenster von dem Fensterbrett herunter 'suddenly falls the dog out of the window down hither from the windowsill' (German adult) [ 19] iz-za kamnja olen' vy-skocil 'from-behind [a] rock [a] deer out-jumped' (Russian adult)

An earlier study (Slobin 1997a) examined the "fall from the cliff episode, which has a number of potential path components: moving to the cliff, stopping at the cliff, throwing the boy and dog down, falling of the boy and dog into the water (see Pictures 16-18, Appendix I). The proposal was "that speakers of S-languages are more likely to break up the event into a larger number of components, based on 'narrative habits' of compacting several path components in a single clause. Speakers of V-languages, by contrast, have developed a narrative style that makes more sparing use of individual motion verbs to encode path components" (Slobin 1997a:448). This proposal was supported across a large number of languages, as shown in Table 1 .' 5 Speakers of S-languages tend to mention more path segments than speakers of V-languages (about three versus two), and a greater proportion of S-language speakers do so (about 80% versus 30%). 3.2 Typological factors Apparently, language typology contributes to a typical level of event granularity. The determining factor seems to be the heavy use of a series of separate clauses in V-languages, as compared with the accumulation of path particles and prepositional phrases with a single verb in S-languages. In Note 9 it was mentioned that some English-speaking preschoolers, instead of using the verb climb in the owl scene, used go and get. What is striking, though, is that these light verbs also occurred

240

Dan I. Slobin

Table 1. Path segmentation in the "cliff scene" (adult narrators) Language type

Language family/branch

Languages

S-Languages

Germanic

Dutch English German Icelandic Swedish Polish Russian Serbo-Croatian French Portuguese Spanish Hebrew

Slavic

V-Languages

Romance

Semitic

Average Proportion number of event of narrators segments mentioning > 3 segments

3.0

86%

2.8

76%

2.1

30%

2.0

30%

with strings of path markers: go up into the tree, get up on the tree, and so forth. That is, English-speaking children, at early ages, are disposed to describe complex paths, and in compact constructions. Across ages, the collection of complex locative elements in English exceeds the possibilities provided by path verbs in V-languages. Table 2 shows the range of combinations that occur (and almost all of them are provided by children). Comparable data can be found in Dutch. German, and Russian; see also the "constellations of spatial morphemes" reported for Icelandic and Swedish by Ragnarsdottir and Stromqvist (this volume). However, beyond these patterns - as in the discussion of manner of motion - it is necessary to consider factors beyond Talmian typology in accounting for patterns of narrative style. 3.3 Language-specific constraints Brown (this volume) reports that path descriptions in Tzeltal cannot include source and goal in the same clause. Bohnemeyer (in press-a) reports the same for Yukatek, another Mayan language. The only way to say, for example, that the boy went from the tree to the rock would be, 'the boy, he was at the tree, and then he left, and then he arrived at the rock'. Brown gives frog-story examples that specify either source (e.g., 'he fell down from the head of the deer') or goal (e.g., 'they are spread-eagled arriving-there at the water'), but not both. Although it is possible to mention source and goal in successive clauses, this is rarely done (Brown, personal communication), perhaps owing to processing constraints, though cultural factors may be more important, as discussed below.


241

Table 2. Combinations of path satellites and prepositional phrases in English frog stories ("N" = noun) Satellite down off out over up

Prepositional phrase from N, in(to) N, on(to) N, off N, out of N, over N, through N, to N after N, into N, onto N, over N, through N, to N at N, from N, in(to) N, onto N, the back of N, through N, to N in(to) N, out of N, to N, toward N from N, in(to) N, out of N, off N, on top of N, on(to) N, over N, toward N, to N, from behind N, from out N, from over N

3.4 Morphosyntactic factors Locative case marking Basque is a V-language, as determined by its lexicon of path verbs. However, as discussed by Ibarretxe-Antunano (this volume), Basque frog stories tend to mention both source and goal in path descriptions. A morphological factor seems to play a role here. The language provides a rich set of morphemes encoding features of location and movement: five locational cases that apply to nouns as well as thirty or more postpositions that also take locative inflections. We can, again, appeal to ease of processing as a determining factor: Basque morphology makes it possible to specify path details in a compact fashion. In addition, a predisposition to omit verbs puts greater weight on the nominal encoding of path segments. These factors lead Ibarretxe-Antunano to propose a "completepath hypothesis" for Basque: "the tendency to linguistically express in the same clause both the source and goal of a translational motion, even in cases where one of the components is pleonastic." Thus, there is a "narrative habit" that favors path elaboration. Basque, then, falls toward the S-language side of Table 1. For example, in the cliff scene, ablative and allative case markers indicate path segments: [20] danak all. AHS

amildegi-tikan

behe-ra

erori

zian

ibai

bat-era

cliff-ABULoc

below-ALL

fall:pFV

AUX

river

one-ALL

'all of them fell from the cliff down to the river'

In the following example, there is no verb at all; path is indicated by the allative case on a noun meaning 'outside': [21] eta

zas!

sagu

kanpo-ra

and

zas!

mouse

outside-ALL

'and suddenly a mouse [comes] out'

242

Dan I. Slobin

Aksu-Koc (1994) has reported similar patterns in Turkish, where narrators make use of locative inflections and directional adverbs to present several path segments in compact fashion; for example: [22] geyik deer

bir tane ucurumun one item clifTiGEN

kenanna side-DAT

geliyor comes

ve and

ucurumun clifT:oE\

ucunda

cocugu

basin-dan

asagi

atiyor

side-toe

boy-Ace

its:head-ABL

downwards

throws

'(the) deer comes to the side of a cliff and at the edge of the cliff throws the boy down from his head.' (Turkish adult)

In sum, easily processed construction types can contribute to path analysis and packaging in both types of languages. Directional morphemes Mayan languages are characterized by the use of directional morphemes that combine with path verbs to provide more specific path information, as well as with manner verbs (reported for Tzotzil by Bowerman et al. 1995). These morphemes are derived from path verbs, but function in a fashion that may be compared to that of path satellites in S-languages. They have meanings such as 'away from', 'toward', 'up', 'down', 'in', 'out', and so forth. Directionals are used by Tzeltal children for path elaboration, as in the following example from Brown's chapter (this volume, Example 42): [23] ya k'an mook bel sk'ok tal i ala tz'i'i 'the little dog wants to go up awaywards to pick it [beehive] towards him'

Tzeltal can be classed as a V-language because path is lexicalized by main verbs, but the availability of directionals provides an "S-language flavor." 3.5 Cultural factors David Wilkins (this volume) adds cultural and ecological dimensions to path elaboration. In his Australian frog stories in Arrernte, like Edith Bavin's (this volume) Warlpiri stories, there is great attention to path details. This is apparently based on the importance of journeys in Australian Aboriginal culture, as embodied in attention to real paths and cardinal points in the environment, traditional narrative structures, and an array of linguistic factors. Although Arremte has path verbs, Wilkins also points out the use of locative cases as well as inflections for associated motion, that is, marking a verb for an associated background action


243

in space, such as 'do X while going upwards/moving past', etc. (Note that encoding associated motion lies outside of the Talmian typology.) The Arrernte narrators analyze the cliff scene into many more fragments than the S-language speakers represented in Table 1, although Arrernte, with its path verbs, would be considered a V-language.16 Thus, whereas English-speaking adults average 4.7 path segments in the cliff scene, Arrernte adults average 7.8 segments, and some provide as many as nine. Wilkins' conclusions add another important dimension to the determinants of narrative style: I would predict that speakers of Central Australian desert languages would behave essentially the same way, independently of language type, owing to areal cultural factors. ... We are left to ponder, then, what is the relative contribution of linguistic typology versus culture in sensitizing children to the narrative interests and rhetorical style preferred by the adult members of their speech community. (Wilkins, this volume, Section 7)

3.6 Modality factors Path information can't be left out in signed languages, because the hand has to move from one location to another. Galvan and Taub (this volume) report that ASL frog stories give more specific information about spatial relations and movement than comparable English stories. However, in so doing, the signers tended to break up a complex path into separate elements, often followed by a composite path representation. Part of the explanation may lie in a possible boundary-crossing constraint, but on-line factors may also play a role. Galvan and Taub suggest that processing demands account for this pattern. However, in this case, the organization of path information may be designed for the receiver, rather than the producer: Presenting a separated, "pre-digested" form ... allow[s] the addressee to build up a model of the event gradually, from simpler conceptual pieces. In addition, the repetition of "old" information with new information would help him or her re-assemble the separated pieces into a coherent whole. The serial and sequential constructions noted for ASL may in fact be better suited to communication than a single mimetic portrayal of all the aspects of a complex event. (Galvan & Taub, this volume, Section 4)

3.7 Typology and the encoding of path information Thus, as in the case of manner, a number of factors interact to determine the expression of directional components of motion events in any given language. Although path is, of necessity, always a part of the narration of motion events, languages differ in the degree to which they analyze paths and present complex

244

Dan I. Slobin

trajectories. Again, lexical, morphosyntactic, and cultural factors must be considered. In addition, choices in the encoding of path information also affect the degree to which manner is encoded, given various constraints on the conjoint expression of path and manner in a language. At the same time, forms of encoding path information affect the degree to which grounds are attended to - sources and goals of paths as well as landmarks and mediums encountered along the way.

4. GROUNDS IN MOTION EVENTS 4.1 Path segmentation When a path is segmented, the narrator has the choice of mentioning the ground associated with each segment. For example, the Basque narrator in [20] mentioned both source and goal ('from the cliff down into the river'), whereas the Turkish narrator in [22] mentioned two source elements, but no goal ('at the edge of the cliff... down from his head'). As discussed above, the availability of locative inflections might predispose speakers of V-languages like Basque and Turkish to explicitly mention source and goal. In S-languages, as discussed in the comparison of [15] and [16], narrators have the option of packaging many path segments into clauses with a single verb. In such instances, each path expression (satellite, preposition) tends to carry a ground element with it ('outside the house down to the back of the garden out into the bit of a forest'). A study of verbs and grounds in novels from the two types of languages suggests a typological difference (Slobin 1997a). S-language novels written in English, German, Dutch, and Russian were compared with V-language novels written in French, Spanish, Turkish, and Japanese. S-language writers tended to mention more than one ground element per verb, whereas V-language writers in the sample almost never used a motion verb with more than one ground. Because the literary traditions of these eight languages are rather different, it would be hard to account for these differences on the basis of national characteristics, as opposed to lexicalization patterns. Note, however, that narrators can also mention grounds in several successive clauses. Nevertheless, in novels and frog stories, there is a tendency for V-language narrators to use fewer clauses with motion verbs and grounds, probably owing to processing constraints that remain to be fully described. For example, it might slow down a narrative to provide a series of clauses, each with a motion verb and ground information. Rather, at least in the sample of languages from Western-style, literate cultures (including Israel. Turkey, and Japan), V-language writers and frog-story narrators prefer to provide ground information in scene-setting descriptions rather than in clauses with motion verbs. That is, descriptions of the terrain and locations of landmarks allow the inference of directed motion, as opposed to explicit descriptions of motion.


245

Clearly, cultural preferences interact with typological factors in this regard. Bavin (this volume), for example, describes an Australian narrative style that builds up information by repetition: A "build-up" style is often used: information is repeated with some new added. There can be a gradual build up of the information; for example, telling the frog story, a speaker might give the information that someone fell, then someone fell to the water, then someone fell down to the water and then specify that the child and dog fell.

Eric Pederson (personal communication) reports a build-up style in Tamil frog stories, in which the end of each clause is repeated as a bridge to the next. And Galvan and Taub (this volume; Taub & Galvan 2000) discuss ASL narrative style that relies on separation and repetition of event components. Explanations based on ease of processing may not be relevant to narrative style at the level of episode construction. Avoidance or appreciation of repetition is a cultural issue, and may be most evident in oral narrative. With regard to manner verbs, I invoked processing factors that may influence the relative accessibility of forms and the codability of concepts. But such factors may not play similar roles in accounting for the rate of mention of grounds. In addition to these considerations, there are linguistic factors that can be briefly mentioned. Again, these lie outside of the Talmian typology. 4.2 Locative case marking Ragnarsdottir and Stromqvist (this volume) compare frog stories in two closely related S-languages, Icelandic and Swedish. They report that Icelandic narrators make explicit reference to grounds far more frequently than Swedish narrators. This is apparently due to a small morphosyntactic difference. Where Swedish narrators mark direction with independent satellites, as in English, Icelandic goalmarking is indicated by accusative case on nouns along with complex directional prepositions. Therefore, Icelandic narrators make much more frequent use of nouns, whereas Swedish allows directional marking alone. Compare: [24] Icelandic: og svo datt hundurinn og strakurinn 'and then fell the dog and the boy

[25] Swedish: pojken ramla ner 'the boy fell down'

ofan-i

sjo

above-into

sea:Acr'

246

Dan I. Slobin

4.3 Dependent-marking versus head-marking languages Engberg-Pedersen and Trondhjem (this volume) present frog-story data from West-Greenlandic, a language in which the verb is richly inflected for information about participants, locations, direction, and temporality. That is, all of the critical information for verb-argument structure is marked on the head (the verb) rather than on dependents (noun phrases) - as in the case of all of the other languages we have considered thus far, with the exception of ASL, which is also head-marking. In West-Greenlandic, a clause typically consists of a verb alone. EngbergPedersen and Trondhjem suggest that this "verb-rich style" results in infrequent mention of grounds. West-Greenlandic is a V-language, but the determinants of narrative style seem to lie in the morphosyntax of head-marking rather than the lexicalization pattern for path expressions. The authors suggest that "[t]he rhetorical style of the West-Greenlandic frog stories seems to be influenced not so much by what is 'readily encodable' as by what is coded semantically richly in the verb, the only obligatory part of the clause." However, they also provide a thoughtful discussion of the problems of attributing narrative patterns to linguistic and/or cultural factors: [Fortescue and Lennert Olsen] hypothesize that in a hunting-gathering society in rather specific physical environments such as the Greenlandic society was and to some extent still is, there is less need for referential specificity, less "emphasis on 'things' rather than on actions and states (for the description of which polysynthetic 'holophrasis' can be extremely efficient)" (Fortescue & Lennert Olsen 1992:215). This explanation for the low number of nominals combines the linguistic type with the language users' communicative needs given the type of society and the physical context. ... It is doubtful whether we shall ever be able to sort out the different influences of language and culture.

4.4 The said and the unsaid Recall that the Basque tendency to omit the verb has as one consequence more frequent mention of source and goal - that is, grounds of movement. In Basque, this occurs within the clause. In other V-languages, where there tend to be fewer motion clauses per episode, and fewer ground elements per verb, there is a tendency to provide static scene setting in order to ground reported motion events. For example, in the fall from the cliff, Berman and Slobin (1994:623) report a considerably higher rate of scene setting in three V-languages - Spanish, Hebrew, and Turkish - in comparison with two S-languages - English and German. In ASL (Galvan and Taub, this volume), as in other sign languages, it is obligatory to "set the stage" with appropriately placed landmarks before motion verbs can move from one landmark to another. In such languages, the verb - as in WestGreenlandic - indicates which participants and referent objects are involved.


247

but these are not lexicalized as part of a clause that reports a motion event. The distribution of information between different sorts of lexical items in connected discourse thus affects both how much will be said about grounds and where that information will occur.

5. TYPOLOGY REVISITED 5.1 Limitations of the binary typology The satellite- versus verb-framed typology has been useful in systematically sorting the world's languages as well as providing a framework for discourse analysis. At the same time - in the course of profitable explorations of the influences of typology on narrative style - the limitations of a binary typology have become evident. Talmy's starting point, with regard to motion events, was to identify the means of expression of path - namely, in the main verb of a clause or in a satellite to the main verb. However, both notions - main verb and satellite • are now in need of greater attention. The following sorts of problems can be listed: 17 •

In serial-verb languages, it is not always evident which verb in a series, if any, is the "main" verb.

•

There are languages with bipartite verbs, such as the Hokan and Penutian languages described by DeLancey (1989, 1996), in which the verb consists of two morphemes of equal status, one expressing manner and the other path. Talmy (2000b:113) provides a similar description ofNez Perce manner prefixes (Aoki 1970). Richard Rhodes (personal communication) reports that such constructions are typical of the American Indian languages Algonquian, Athabaskan, Hokan, and Klamath-Takelman. There are languages such as the Australian language Jaminjung (Schultze-Berndt 2000), with a very small verb lexicon of about 24 "function verbs." For encoding motion events, one of five verbs is used, expressing a deictic or aspectual function: 'go', 'come', 'fall', 'hit', 'do'. These verbs are combined with satellite-like elements, "preverbs," that encode both path and manner in the same fashion. Again, neither path nor manner is unequivocally the "main" element in a clause. In V-languages that make heavy use of path adverbs, such as Mayan directionals, it is possible to express path in elements that, formally, are

248

Dan I. Slobin

neither main verbs nor satellites. The same appears to be true in Italian, and is an incipient tendency in other Romance languages. Thus we lack clear criteria for distinguishing satellites from directional adverbs. •

In V-languages like Korean, in which the main verb is a deictic verb, both path and manner verbs are expressed in subordinate, non-finite forms; yet Korean is categorized as a V-language because path is expressed by a verb (Choi & Bowerman 1991). In some S-languages, the meaning of a satellite depends on the construction type in which it occurs. For example, in Dutch the satellite in 'in' encodes path in a particular syntactic context only. For example:

[26] de the

jonge boy

loopt walks

het the

bos woods

in in

'the boy walks into the woods' [path: boy enters woods]

[27] de the

jonge

loopt

in

het

bos

boy

walks

in

the

woods

'the boy walks in the woods' [non-path: boy located in woods]

As Sinha and Kuteva (1995) have pointed out, this is an example of distributed spatial semantics. That is, the spatial meaning of an element does not reside in a single lexical item; rather, it is distributed over form classes and constructions. 5.2 A revised typology At the very least, given facts such as these, it seems necessary to revise the definitions of verb-framed and satellite-framed, adding a third type (Table 3). Such a proposal is not particularly satisfying for discourse analysis, because what is most interesting is the impact of various additional options on the structure of narrative and the allocation of attention - especially to features of path and manner. Rather than put languages into typological categories, it might be more profitable to lay out the collection of factors that, together, interact to contribute to particular rhetorical styles. The linguistic locus of path expression is only one of such factors, as we have seen in considering narrations of frog-story events in several different sorts of languages. Note also, as pointed out in Frog I: "As a general caveat, it should be remembered that typological characterizations often


249

Table 3. Revised motion-event language typology Language type Preferred means of expression

Typical construction type

Examples

Verbframed languages

path is expressed by a Path verb + verb, with subordinate subordinate Manner expression of manner verb

Romance, Semitic, Turkic, Basque, Japanese, Korean18

Satelliteframed languages

Manner verb + path is expressed by a non-verbal element Path satellite associated with a verb

Germanic, Slavic, Finno-Ugric

Equipollentlyframed languages

path and manner are expressed by equivalent grammatical forms

Manner verb + Path verb (serial-verb languages)

Niger-Congo, HmongMien, Sino-Tibetan, Tai-Kadai, MonKhmer, Austronesian

[Manner + Path]verb (bipartite-verb languages)

Algonquian, Athabaskan, Hokan, Klamath-Takelman

Manner preverb + Path preverb + verb

Jaminjungan languages

reflect tendencies rather than absolute differences between languages" (Berman & Slobin 1994:118). With these perspectives in mind, we can begin to lay out factors that influence the habitual expressions of path and factors that influence the habitual expressions of manner across languages. 5.3 Habitual expressions of path With regard to path expression, we have seen that freestanding path particles, such as English satellites, allow the stacking of path components in a clause or succession of clauses with a single verb. This contrasts, for example, with the impossibility of stacking Slavic satellites, which are verb prefixes. We have seen that some languages require a separate verb for each component of a motion event, while other languages can conflate components in a single verb. At the same time, cultural patterns of narrative style can act to foster or limit repetition and elaboration of path components. At the level of lexical expression of path, languages vary in the extent to which they provide verbs that conflate several components of a motion event, such as verbs for climbing, which conflate manner and path. Other examples are verbs that focus on various temporal and motivational dimensions of motion. For

250

Dan I. Slobin

example, Korean ttuda and ttonada both mean 'leave', but the former simply focuses on the fact that the theme is no longer in an enclosed space, while the latter focuses on the source location and continuing motion beyond. Distributed spatial semantics must also be considered, however, because the focus on source location, with ttonada, is modified by case marking on the source. For example, with ablative case, e.g., Seoul-ABi, the meaning treats Seoul as the point of departure: 'he left from Seoul'. Seoul-Ace, by contrast, implies that the person has spent some time in Seoul (Kyung-ju Oh, personal communication; Wienold & Schwarze 1989; Wienold et al. 1991). Distributed spatial semantics turns out to play a major role in almost every corner of the world of motion-event expression. For example, the Spanish verb entrar 'enter' receives a different deictic perspective with the choice of preposition: entro a la casa '(he) entered to the house' implies that the speaker observed the event from outside, while entro en la casa '(he) entered jn the house' is from the perspective of a speaker located inside the house (Ibarretxe-Antunano, under review-a). Examples like these abound in language descriptions. They make it clear that more is at play than expression of path by a verb or a non-verb. It should be possible to lay out systematic interactions of patterns of lexical conflation and morphosyntax, with the aim of arriving at more precise explanations of the array of path expressions in discourse in particular languages. 5.4 Habitual expressions of manner: Degree of manner salience It is the expression of manner, rather than path, that has attracted the most attention in the array of narrative analyses discussed in this chapter. And with regard to manner, a different kind of typological question comes to the fore: How easy or natural is it to add manner information to path expressions in a particular language? Here, rather than a bipartite typology, it seems more useful to put languages on a dine of manner salience: •

High-manner-salient languages: There is an accessible slot for manner in the language, made available in various ways: ° main verb in S-languages; ° manner verb in serial-verb languages; ° manner morpheme in bipartite verbs; ° manner preverb in Jaminjungan languages; ° ideophone.

•

Low-manner-salient languages: Manner is subordinated to path.


251

From this perspective, a tripartite typology with regard to path expression (verbframed language, equipollently-framed language, satellite-framed language) does not seem to be necessary in accounting for the relative attention to manner in a language. In high-manner-salient languages, speakers regularly and easily provide information about manner when describing motion events, whereas in low-manner-salient languages, manner information is provided only when manner is foregrounded for some reason.19 There are diachronic consequences to the typology. In high-manner-salient languages there is a rich lexicon of manner morphemes. This is probably due to diachronic processes. Work with dictionaries and consultants in the several languages considered here suggests that the Romance languages, Turkish, and Hebrew have no more than about 75 intransitive manner verbs in regular use, whereas the Germanic and Slavic languages, Hungarian, and Mandarin have upwards of 150.20 There are no available counts for manner ideophones, but the papers in Voeltz and Kilian-Hatz (2001) suggest a rich inventory. Westermann's (1930) grammar of Ewe gives examples of 37 ideophones that can be used with the verb zo 'walk', with the additional information that these forms can be reduplicated and can occur with high tone for diminutives and low tone to describe motions of large entities. Synchronically, each individual learner is, of necessity, oriented to the conceptual distinctions that are required in order to acquire and use expressions in the domain of manner. Children learning high-manner-salient languages develop a large and differentiated manner lexicon in the preschool years (Slobin 2000) and pay attention to manner of motion in experimental tasks (e.g., Hohenstein 2001). As suggested in Frog I: "A generalization that emerges from our study is that if a linguistic form is highly accessible, its functional development may be accelerated" (Berman & Slobin 1994:624; authors' emphasis). Languages can also move along the cline over time. For example, Italian, in comparison with other Romance languages, makes wider use of directional adverbs with both path and manner verbs (Schwarze 1985), allowing combinations such as nuotare via 'swim away' as well as boundary-crossing uses of prepositional phrases, such as correre sulla strada 'run onto the street' (Hottenroth 1985). This movement of Italian in the direction of an S-language may be stimulated by contact with German, especially in Northern Italy. (Similar patterns are reported for Brussels French, under the influence of Dutch contact, as well as other Romance-Germanic contact situations; see Kramer 1981.) Italian frog stories seem to be richer in manner verbs and manner-path combinations than French and Spanish ones. Languages can also move in the other direction. For example, Slavic path prefixes often fuse with the verb stem, leaving verbs that are semi-transparent manner-path lexical items. The verb

252

Dan I. Slobin

Figure 4. How a language becomes manner salient.0

vz-myvat' 'soar upward', has a transparent path prefix vz- 'upward', but myvat' is no longer an independent verb. Serbo-Croatian path prefixes have become phonologically fused to a neutral verb of'going', and native speakers apparently consider these forms as monomorphemic path verbs (Luna Filipovic, Jelena Jovanovic, personal communication). This is the direction that Latin took, also fusing path prefixes and verb roots - ultimately arriving at monomorphemic path verbs such as Spanish entrar and French entrer, in which one can discern the erstwhile prefixed satellite. 5.5 A diachronic and synchronic model for the emergence of manner salience The collection of factors proposed above can be put together in a model, as shown in Figure 4. A language provides a set of patterns - lexicalization patterns, morphological forms, syntactic constructions. The use of these patterns on-line is modulated by psycholinguistic factors of processability - ease of access, heaviness of construction, etc. Linguistic patterns and processability act jointly to set the degree of accessibility of manner expressions in a language. A cascade of interrelated effects follows. If manner is easily accessible, it will be encoded more frequently and, over time, speakers will tend to elaborate the domain in terms of semantic specificity. Consequently, learners will construct a more elaborate conceptual space for manner, allowing each new generation to continue the cycle


253

of attention to manner. Habitual patterns of language use and attention arise and become entrenched in the individual lifespan and across generations. 5.6 Low-manner-salient languages These include prototypical V-languages, such as Romance, Semitic, and Turkic languages, as well as languages like Korean, in which both path and manner verbs are subordinated to a final deictic verb. With regard to manner, path-salient languages require additional morphology, such as gerunds, converbs, or adverbial expressions. There are also more factors that may work to downgrade attention to manner in V-languages. Every change of directional vector requires a separate verb, which may lead speakers to produce fewer path expressions (and hence fewer manner expressions). Furthermore, these languages appear to have some version of a boundary-crossing constraint that limits the use of certain types of manner verbs in the description of certain types of paths. Thus, although there is nothing in principle that would keep such a language from also paying detailed attention to manner, several factors seem to conspire against such a tendency - at least with regard to habitual descriptions of motion events. 5.7 Conclusion Typologies are useful in sorting out the multifarious dimensions that characterize human languages. But just because of this diversity of dimensions, each language is an intersection of many interacting patterns. Diversity is not boundless and unconstrained. If it were, the growing field of linguistic typology would have died on the vine. The challenge to the typological linguist is to find sets of variables that co-occur and to try to account for those co-occurrences. However, linguistic patterns don't occur in the abstract. They arise in the course of language in use. The frog-story research has demonstrated that one particular use the construction of oral narrative - requires typological descriptions that are sensitive to on-line processing. That is, habitual patterns of language use are shaped by ease of accessibility of linguistic forms - to producer and receiver, as well as by the dynamics of cultural and aesthetic values and the perspectives and communicative aims of the speaker. It is to Talmy's credit that his typology has brought us thus far. We can build upon his insights in working towards typologies of language use. Obviously, a diversity of languages must be studied. At the same time, a diversity of types of data is also required. The frog-story research makes it clear that we need audio and video data - along with grammars and dictionaries, texts and corpora - in order to carry the work forward.

254

Dan I. Slobin

ACKNOWLEDGMENTS Research reported here has been supported by the Max Planck Institute for Psycholinguistics, Nijmegen, Netherlands; the National Science Foundation: and the University of California, Berkeley (Committee on Research, Institute of Cognitive and Brain Sciences, Institute of Human Development). I am grateful to many students and colleagues who have helped me think through the issues discussed in this chapter, and especially to Melissa Bowerman for meticulous reading and steadfast critique.

NOTES 1.

The languages presented in Part I are American Sign Language (Galvan & Taub). Arrernte (Wilkins), Basque (Ibarretxe-Antunano), Icelandic (Ragnarsdottir & Stromqvist), Swedish (Ragnarsdottir & Stromqvist), Thai (Zlatev & Yangklang). Tzeltal (Brown), Warlpiri (Bavin), and West-Greenlandic (Engberg-Pedersen & Trondhjem).

2.

See, especially, Ameka & Essegbey, in press; Berman & Slobin 1994; Billman. Swilley & Krych 2000; Croft et al. 2002; Filipovic 1999. in preparation: Fong & Poulin 1998; Hohenstein 2001; Gennari et al. 2002; McNeill & Duncan 2000: Mora Gutierrez 1998; Naigles et al. 1998; Naigles & Terrazas 1998: Narasimhan 1998: Oh, in preparation; Ohara 2000; Ozcahskan 2002: Ozcahskan & Slobin 1999. 2000a. 2000b, in press-b; Papafragou, Massey & Gleitman 2001; Slobin 1996a. 1997a. 2000. 2002b.

3.

The data reported here come from a range of published and unpublished studies carried out at the Institute of Human Development, University of California. Berkeley, and in collaboration with other institutions, as well as data reported in this volume. Narratives were gathered from preschoolers (age 3-5), school-age children (age 6-11). and adults, with 10-20 stories per age group. The following collaborators have been involved, along with many Berkeley students: Basque: 1. Ibarretxe-Antunano: Dutch: J. Aarssen, P. Bos, L. Verhoeven; English: V. Marchman. T. Renner. G. Wigglesworth: French: H. Jisa, S. Kern; German: M. Bamberg, M. Carroll, C. von Stutterheim: Hebrew: R. A. Berman, Y. Neeman; Icelandic: H. Ragnarsdottir; Italian: P. Cipriani. M. Orsolini; Mandarin: J. Guo, A. Hsiao; Polish: M. Smoczynska: Portuguese: I. Hub Faria; Russian: Y. Anilovich, N. V. Durova, M. Smoczynska, N. M. Yurieva: SerboCroatian: S. Savic; Spanish: A. Bocaz, J. Covey, E. Sebastian: Swedish: A. Nordqvist. S. Stromqvist; Thai: J. Zlatev, P. Yangklang; Turkish: J. Aarssen. A. A. Aksu-Koc. A. Kiintay, S. Ozcalis.kan, L. Verhoeven.

4.

The patterns of rhetorical style discussed in this chapter are found in narratives, and I use the terms "rhetorical" and "narrative style" interchangeably here. However, similar patterns are found in non-narrative discourse, suggesting general languagespecific patterns of rhetorical style across genres (Slobin 1997a. 2000. 2002b).


255

5.

"Manner" covers an ill-defined set of dimensions that modulate motion, including motor pattern, rate, rhythm, posture, affect, and evaluative factors.

6.

Note that this definition includes static location, translational motion, and selfcontained motion. The events discussed in this chapter are only those of translational motion, defined as events in which "an object's base location shifts from one point to another in space" (Talmy 2000b:35). Path is treated as consisting of three components (Talmy 2000b:53-57): a) the Vector (AT, TO, FROM, ALONG, etc), b) the Conformation (POINT, ENCLOSURE, VOLUME, etc.), and c) the Deictic (±TOWARD SPEAKER).

7.

Later we examine intra-typological differences, comparing Germanic and Slavic satellites and considering whether serial-verb languages such as Mandarin and Thai should be classed as S-languages.

8.

Talmy suggests that path verbs in serial-verb languages often show evidence of grammaticizing into path satellites - that is, losing some features of independent verbs. This is certainly an important diachronic path that can lead from one language type to another. But the frog-story data do not make it possible to unequivocally categorize languages like Mandarin and Thai as either S- or V-languages. A Chinese linguist also points out that path verbs are not strictly comparable to English verb particles: "However, different from English, these satellites in Chinese can also function as independent verbs themselves. When such a verb is connected to another verb, a verbal construction called lidn dong shi 'serial verb construction' is formed" (Gao 2001:62).

9.

A small number of younger English-speaking children used the verbs go and get with path particles (go up/up in/up into the tree, get up on the tree}.

10. Sources: Zulu: Msimang & Poulos 2001; Ewe: Ameka 2001; Emai: Schaefer 2001; Ilocano: Rubino 2001; Turkish: Jendraschek 2002; Mandarin: Ying 1988; Japanese: Hamano 1998. 11. Full data have not been published on the Canary Row studies. From published examples, it would appear that morphosyntactic factors play a role in the gesture patterns of the three V-languages. Turkish and Japanese are SOV languages in which manner precedes path in subordinated constructions, e.g., 'rolling descend' in the Canary^ Row examples. Co-speech gestures tend to accompany the two verbs in this order, resulting in separate gestures for manner and path, sometimes followed by a manner-path conflated gesture. Spanish is an SVO language in which the order for phrasal verbs is the opposite, e.g., 'descend rolling'. In the frog stories, such phrasal verbs almost never occurred, and data on phrasal verbs are not available for Canary Row. The published examples suggest that the Spanish manner and manner-path gestures tend to accompany path verbs alone, e.g., 'descend'. (English gestures are timed to occur with verb-satellite constructions such as roll down) Adequate data are needed in order to determine when co-speech gestures in V-languages accompany semantically corresponding verbs and when they add information that is not lexically specified; that is, statistics are needed in order to compare manner gestures that accompany path verbs with those that accompany manner verbs. (See van Hoof 2000, for data on co-speech gesture in frog stories

256

Dan I. Slobin

in English, Dutch, and Spanish, as well as English as a second language in a small sample of Dutch- and Spanish-speakers. In van Hoofs data, manner gestures are rare but they almost always accompany manner verbs for all of the groups of subjects. That is, gesture does not augment attention to manner for these S- and V-language speakers.) 12. Taub, Pinar, and Galvan (2002) have analyzed one scene from Canan- Ro\\- in English and Spanish, attending to both lexical and gestural encoding of information. There are no statistical differences in a small sample, but the directions suggest somewhat greater attention to manner in English. In terms of number of "conceptual elements." English speakers used slightly more lexical elements to express manner (3.58 versus 3.17) and slightly more gestural elements as well (2.58 versus 2.33). Combining speech and gesture, English-speakers provided 4.5 conceptual elements of manner, compared with 4.0 in Spanish. 13. In addition, as Berman (this volume) and Stromqvist, Nordqvist. and Wengelin (this volume) point out, children face new problems in learning how to produce written narratives. 14. Verb-gapping is required in the English construction in [16], but this is probably simpler, in processing terms, than clauses with two distinct verbs in V-language equivalents. Bohnemeyer (in press-a) has proposed a Unique Vector Constraint that apparently applies to both types of languages: "All direction specifications in a single simple clause referring to a single continuous motion event must denote the same 'unbounded 1 direction vector, i.e. the same direction vector..." Accordingly, crawl toward and up the tree must be expressed by two simple clauses in English as well as in V-languages. Nevertheless, gapping of the verb in the second clause, as well as the possibility of using a single manner verb for both path segments, is a special characteristic of (some) S-languages, and thereby contributes to manner salience in those languages. 15. The data come from ten adult frog stories in each of the languages. The research was carried out at a workshop supported by the National Science Foundation held at the Linguistic Institute of the Linguistic Society of America at the University of New Mexico, Albuquerque, in July 1995. Participants involved in the data analysis were: Jeroen Aarssen, Ayhan Aksu-Koc, Michael Bamberg, Edith Bavin, Ruth Berman. Petra Bos, Nancy Budwig, Harriet Jisa, Catalina Johnson Herrera, Sophie Kern. Asa Nordqvist, Barbara Pearson, Hrafnhildur Ragnarsdottir, Judy Reilly, Svenka Savic. Magdalena Smoczynska, Anat Stavans, Sabine Stoll, Sven Stromqvist. and Ludo Verhoeven. 16. Bavin (personal communication) considers Warlpiri an S-language because path is not encoded by path verbs but rather by elements associated with verbs: locative nouns, case markers, verbal affixes, or preverbs. Wilkins considers Arrernte a sort of Vlanguage because of the use of path verbs. The cultural orientation to paths, however, seems to be more important than "framing"-typology for both languages. 17. Talmy (1985, 2000b) is aware of the existence of various mixed language types and the problems of categorizing a lexical item as a satellite or unequivocally assigning


257

a language to one of the two types. However, he finds the binary typology useful for the larger goal of exploring a "typology of event integration." On this broader level, Talmy suggests that S- and V-languages show distinct patterns across a number of event types: motion, aspect, state change, correlated activities, and realization of goals. These suggestions go beyond the aims of this chapter, where we are concerned with applying the binary typology to the narration of motion events across languages. 18. Korean dictionaries tend to treat combinations of a subordinated path verb and a main deictic verb, such as 'enter-come', as compound verbs, giving such compounds separate dictionary entries. Note that this is a small, closed set of compound path verbs. By contrast, combinations of Manner + Path + Deictic are generally not given separate dictionary entries (Kyung-ju Oh, personal communication). The latter, of course, would constitute a very large set. In this sense, then, Korean remains a Vlanguage. 19. Talmy (1985) has noted that manner is backgrounded in habitual S-language expressions, whereas it is foregrounded in subordinated V-language expressions. Recently he has interpreted the differences in attention to path and manner in English and Spanish, reported by Slobin (1996a): "In English, both Manner and Path are characteristically expressed in backgrounding constituents: the main verb root and the closed-class satellite. It should be expected therefore that both of these semantic categories will be readily included in a sentence - and that is what is found. But characteristically in Spanish, only Path is expressed in a backgrounding constituent, the main verb root, whereas Manner is expressed in a foregrounding constituent, a gerundive or an adverb phrase. It would thus be expected that the expression of Path is readily included in a sentence, while that of Manner is not - and, again, that is what is found" (Talmy 2000b: 131). However, this account does not explain why V-languages make relatively less use of manner verbs in constructions where they are allowed as main verbs (i.e., non-boundary-crossing events, such as 'dash up the stairs' or 'stroll to the woods'). I have suggested that V-language use engenders a habitual rhetorical style in which manner is not highly salient. 20. Talmy notes that these facts lie beyond his typology: "Slobin (1996a) has further observed that verb-framed languages like Spanish not only express Manner less readily than satellite-framed languages like English, but that they also have fewer distinct lexical verbs for expressing distinctions of Manner. The four principles posited here do not account for this phenomenon, so further explanation must be sought" (Talmy 2000b:146). Such further explanation has been the goal of this chapter.


Part II: Contextual Perspectives


The Role of Context in Developing Narrative Abilities1

Ruth A. Berman

I . INTRODUCTION This chapter considers whether and how the findings from the original crosslinguistic "frog story" study published in Berman and Slobin (1994) generalize across other communicative contexts and elicitation procedures in later school age as well as in early language development. As pointed out by Maya Hickmann in her thoughtful review article, "the entire book is devoted to a particular discourse genre: the production of narratives based on a picture book and addressed to an adult." (Hickmann 1998:34). In fact, we explicitly note at the outset of our study that a factor "in evaluating our research methodology is the special nature of picture-description tasks in general, and reliance on this storybook and our elicitation procedures in particular ..." (Berman & Slobin 1994:24). And we go on to say that [d]espite the many caveats we note [mainly in Chapters IB and IIA], we feel confident that our findings are robust, because they are consistent with the experience of researchers in diverse cultures, ... [and] because we find common developmental patterns across languages, [and also since] our findings are well in line with what has been found by researchers using different procedures for eliciting connected narratives from children ...

The present chapter re-evaluates these claims, a decade or so after the original frog-story samples were collected and analyzed, in light of other research on children's discourse-production abilities in different languages and various contexts. This chapter, like our original study, thus remains focused on production of narrative texts, rather than on processes of comprehension or recall of story content and structure.

262

Ruth A. Berman

As background, note that there is some confusion between two interrelated yet distinct dimensions of text production: type of discourse or (sub)genre, on the one hand, and the procedures adopted for elicitation of the texts in question, on the other. For example, several distinct procedures have been applied to elicit narrative texts from children on the basis of graphic representations. These include: /.

pictured storybooks, of which Frog, where are you? is perhaps the best known;

/'/'.

series or sequences of four to eight pictures, as in the work of Hickmann and her associates (e.g., Hickmann 1991; Hickmann & Liang 1990). in Karmiloff-Smith's work on determiners and reference (KarmiloffSmith 1979, 1981), in the studies of Nelson (1986) and her associates on the role of scriptal knowledge in storytelling (Shapiro & Hudson 1991), as well as in Hebrew-language research with which this author is connected (Berman & Katzenberger 1998; Katzenberger 1994; Lotan 2003); and

/'/'/'.

short films without words (e.g., Hickmann 1980; Warden 1981: and Chafe's (1980) study with adult subjects).

These different elicitation procedures for narrative text production each imposes distinct cognitive demands. Yet all belong to a shared narrative subgenre: the verbalization of graphic representations of non-veridical, fictive sequences of events. As such, they differ markedly from the reconstruction of real-life events typified by the personal-experience narratives elicited by Labov's (1972. 1997) classic study of young speakers of Black English, in Peterson and McCabe's (1983) large-scale study of narratives produced by three- to ten-year-old children, and in crosslinguistic research on older children discussed in Section 6 below (Berman & Verhoeven 2002). It thus seems important to examine our results for the frog-story picturebook sample in light of what has been found both for children's fictive stories based on short picture-series or films and for how they recount events that they have experienced personally.

2. A SHARED DEVELOPMENTAL SEQUENCE The research question can now be reworded as follows: Do children manifest the same developmental path in other elicitation settings and in performing different narrative tasks as in the frog-story study? The answer will depend to some

The Role of Context in Developing Narrative Abilities

263

extent on how stringently one interprets the notion of "the same developmental path". I adopt a weaker definition, focused on similarity rather than identity of developmental patterning. This yields two predictions: /'.

There will be a shared development in mastery of overall narrative organization as a function of age: older children's texts will be more coherent and better structured globally than those of younger children. But this does not entail chronological identity. Children might be able to construct well-formed narrative texts quite adequately at the age of three in some contexts but not others. For example, they may be well able to report on something that happened to them in conversation or in a highly familiar or well-rehearsed setting (Nicolopolou 1996), but they may encounter difficulty when asked to write a narrative at even quite late school age (Berman 1999).

//.

Relatedly, children may prove able to express relations between linguistic forms and narrative discourse functions in some domains earlier in some narrative contexts than in others. For example, young preschool children appear to have little difficulty in making clear reference to protagonists in personal-experience accounts (Hudson 1993), but they fail to maintain or switch reference appropriately at much later ages when asked to tell a story based on a series of pictures (Berman & Katzenberger 1998; Hickmann 1995).

Consider, next, what was involved in the original task that we administered to preschool children aged 3-5, gradeschoolers aged 9, and adults. As defined in Hickmann's review of our study, the particular genre investigated in Berman & Slobin (1994) "requires that children decode visual information about complex event sequences (line drawings), construct on this basis a global cognitive representation of the story, and 'transform' this representation into a sequentially and hierarchically organized narrative" (Hickmann 1998:34). A key finding was that, in the literate Western-type cultures included in our study, even the youngest children, aged 3-4, were able to relate to the contents of the drawings and they were also all able to translate static, graphic material into dynamic verbal eventdescriptions; by late preschool, around age 5-6, most children were able to express these events in terms of sequentially organized narratives. However, only older school-age children and adults were able to meet the requirements of constructing an overall story representation and so to produce globally wellorganized narrative texts. These conclusions appear well-motivated and hence generalizable, in light of developmental trends revealed by other studies on narrative development. In

264

Ruth A. Berman

line with prediction /', similar patternings emerged in investigations which used very different research strategies. These include Applebee's (1978) analysis of stories which children were asked to make up themselves and Karmiloff-Smith's (1981, 1985) research on specially designed sequences of pictures. Our results are also consistent with findings from a range of other frog-story research, beyond the original five-language sample (Berman & Slobin 1994, Chapter IIA). These include an extended Hebrew-language sample (Berman 1988a). a different sample of American-English children (Reilly 1992), and more recent studies among monolingual French-speakers (Kern 1997) and bilingual Dutch-Turkish (Aarssen 1996) and French-Turkish (Akinci 1999) speakers. Across these studies, children proceed from describing unrelated isolated events to linear chaining and on to hierarchically organized global structuring of narrative texts. Another strong argument for generalizability is the fact that analysis of the crosslinguistic frog-story findings reveals an interesting interaction between stages in the development of narrative abilities, on the one hand, and levels of organization of information in narrative discourse, on the other (Shen & Berman 1997). The following pattern can be identified in both narrative organization and narrative development: /.

sequential chaining of events at a local level of adjacent clauses, to meet Labov's definition of a minimal narrative as "a sequence of two clauses which are temporally ordered" (1972:360);

//.

causal relating of events at a local level (Shen 1990; Trabasso & Nickels 1992); and eventually,

///.

macro-level hierarchical structuring of events in a globally organized narrative discourse (van Dijk 1980; Shen 1988).

Against this background, the assumption underlying the present chapter is that the basic developmental sequence revealed by our frog-story study is not a mere artifact of either the elicitation methods or the narrative subgenre under consideration. It is a development that has strong psychological reality and motivation in the theory of narrative text-construction, at least within the Western cultural conventions of storytelling considered in our study.

3. TASK-SPECIFICITY As against this assumption of a shared, cognitively determined developmental sequence in gaining mastery of narrative structure, I argue that context will


265

play a role in this as in other domains of language knowledge and language use. Specifically, the general pattern of development will be affected by the communicative and cognitive demands evoked in different settings of narrative text production. This means that children may reveal more knowledge of narrative structure and they will perform the act of storytelling better in some situations than in others. This claim is developmentally well-motivated for two reasons. First, a key conclusion of our earlier frog-story study was that acquisition of form-function relations in gaining mastery of linguistic structure and of language use has a long developmental history. And this finding is strongly confirmed by research in the area of "late acquisitions" (Herman 1997). In learning how to tell a story, as in other domains, acquisition is not an "all-or-nothing" leap from no knowledge to full knowledge. Rather, it involves partial knowledge and reorganization and integration of prior knowledge across different domains. As a result, more advanced developmental phases will be manifested earlier under some circumstances and in some areas than in others.2 Developmental phases are recurrent: they may be attained at different times in different cognitive domains and in different subsystems of language knowledge (Herman 1986, 1993b). In the present case, reaching a given phase in narrative development will depend on the specific task or setting in which the text is produced (and also, presumably, comprehended). A second related reason for taking account of the impact of context on narrative development is that children's abilities are task-sensitive.3 This has been argued by developmental psychologists for different cognitive domains (e.g., Donaldson 1978; Fischer 1980; Gelman 1978; Rose & Blank 1974; Wellman & Somerville 1980); it has been demonstrated for different aspects of language acquisition (e.g., Farrar, Friend & Forbes 1993; Hecht 1985; Karmiloff-Smith 1979; Levy 1987); and it has been shown in several studies of narrative abilities (e.g., Allen et al. 1994; French & Nelson 1982; Hudson & Shapiro 1991; Peled & Blum-Kulka 1992; Seidman, Nelson & Gruendel 1986; Shatz 1985; Wellhousen 1993; Wolf, Moreton & Camp 1994). For example, Hudson (1986:103) finds that preschool children "are able to produce organized narratives about past events (either spontaneously ... or in response to experimenters' queries) [whereas] other narrative genres, such as story production, are not mastered till late." And similar conclusions are reached by Seidman, Nelson, and Gruendel (1986), in comparing scripts, episodic recountings, and make-believe accounts of scriptlike events. These findings can be explained in general developmental terms, deriving from the relative vulnerability of children's partial knowledge of systems which are still en route to mastery. In storytelling as in development generally, when knowledge is not fully consolidated and integrated with other domains, children will cope better with tasks that take up less mental space or impose less of a cognitive load (Berman & Katzenberger 1998; Shatz 1985).

266

Ruth A. Berman

The impact of elicitation setting and the role of context on children's narrative productions were analyzed by comparing findings from different samples of Hebrew-language narrative texts. It turned out that three-year-olds were well able to tell some kinds of stories, whereas five-year-olds had difficulty with others, and even eight-year-olds found it hard to construct coherent narrative texts when relating the contents of a film they had seen just once. The studies surveyed here are outlined in Table 1, listed in ascending order of "task difficulty," defined by the age at which children demonstrated command of the different levels of narrative organization: isolated events, linear chaining, and hierarchical global structure. Detailed analyses, illustrations, and motivations for the summary provided in this chapter are presented in Berman (1995:295-302).4 Comparison of results of these five contexts revealed that even some three-year-olds were able to produce temporally sequenced descriptions of a familiar scriptlike event-complex (Study 1 in Table 1), confirming results of a range of studies with this kind of task (Nelson 1986). Moreover, they did better on describing the script of a visit to a doctor than on telling what happens when people fight or quarrel. In contrast, although three-year-olds were also quite good at recounting a fight or quarrel as a specific personal experience (Study 2). they typically needed much scaffolding, in the form of prompts and other adult input, in order to perform this task. Under the same elicitation procedure, fiveyear-olds all proved able to produce fight-story personal-experience accounts that were well-structured monologic texts, with a beginning, middle, and end. These accounts differed from those of older subjects (schoolchildren aged 7 and 9 and adults) mainly in the simplicity of their episodic structure, relative lack of motivation, and formulaic background-setting elements (Berman 200la). The picture-series based elicitations (Study 3 of Table 1) proved to be a rather more difficult task for expression of developing narrative structure. Preschoolers aged 4, 5, and 6, gradeschoolers aged 10, and adults were asked to tell the story depicted in three series of four picture-sets (Katzenberger 1994). Although all subjects were familiar with the scripts depicted in each series (buying an article of clothing, going fishing, and picking fruit), most of the four-year-olds and several of the five-year-olds failed to recognize that the four pictures "told a story"; rather, they treated each of the pictures in isolation. Moreover, in marked contrast to the fight-story sample, only the ten-year-olds and adults were able to maintain and switch reference unambiguously to the two characters represented in each series of pictures (Berman & Katzenberger 1998). Static descriptions and labeling of the contents of the pictures rather than relating to dynamic events were common across the four-year-olds, the youngest group in this study. This contrasted with the Hebrew frog-story sample, where such descriptions (e.g.. here is a boy, this is a frog, he has a dog) were typically confined to the first one or two pictures in the book.


267

Table 1. Narrative-elicitation studies with Hebrew-speaking subjects Story, Content

Elicitation materials

1. Script

Instructions

Task, Type of knowledge

"Do you know what it's like • Verbal reconstruction of to go to a doctor?" familiar event sequences "What happens when people go to a doctor?" • Generic formulation "Do you know what it's like to quarrel with somebody?" • Temporal sequencing "What happens when people quarrel?"

doctor

fight

2. Personal experience

"Have you ever quarreled with someone, had a quarrel, a fight? Tell me about it."

• Verbal reconstruction of personal experience • Fight/quarrel script • Single-episode structure

3. Picture- Three sets series of four shopping, pictures fishing, fruitpicking

"Arrange the pictures so they tell a story, and then tell me the story."

• From visual, static, spatial to verbal, dynamic, temporal • Relation between pictures • Reference, action-structure

4. Picturebook

Fifteenpage wordless picturebook

"The book tells a story about a boy, a dog, and a frog. Look through the pictures and then you'll tell the story."

• From visual, static, spatial to verbal, dynamic, temporal • Adventure-story, search • Complex episode structure

Sevenminute film without words

"Look at this film, try to remember what it is about, and afterwards you tell the story."

• Recall and verbalization of visual scenes • Several unrelated episodes and no clear resolution

fight story

frog story

5. Film pear-story

Yet the frog-story elicitation task (Study 4 in Table 1) proved more difficult than the picture series in overall plot structure. In recounting the contents of the frog-story picturebook, few preschoolers (ages 3, 4, and 5) achieved a globally organized narrative, whereas around 40% of children aged 4-5 and nearly 60% of

268

Ruth A. Berman

children aged 6 did so for at least one of the picture-series used by Katzenberger (1994, Study 3 in Table 1). Both studies used pictures as a basis for narrative elicitation, but the short series yielded more advanced results in some respects (overall narrative structure) and less advanced in others (reliance on static labeling descriptions as a major strategy for reference to people and events). The reason appears to lie in relative episodic complexity. The frog story depicts numerous sub-episodes which make up a single overarching theme, the search for a missing pet. In contrast, sequences of four to six pictures are typically confined to one main episode or a single attempt to reach the goal. Yet the frog-story elicitation may in fact present an easier task than a short series of separate pictures in on-line scaffolding by means of visual support for story content, because children can relate to the picture(s) they see on a single page each time, before turning to the next page of the booklet. In the last type of procedure used for eliciting Hebrew narratives (Study 5 in Table 1), subjects recounted the contents of a short film which they had just viewed for the first time. This was the film constructed by Chafe (1977. 1980) for studying how adults in different cultures recapitulate experience verbally. In the Hebrew study, none of the five-year-olds and only a few eight-year-olds were able to produce a minimal, temporally well-sequenced chain of events to describe the contents of the film. This contrasts markedly with the more advanced levels of narrative ability demonstrated by preschoolers as well as school-age children in the other settings listed in Table 1 - generic scripts, personal-experience accounts, picture-series sequences, and picturebook story. Two factors seem to account for this difference. The task imposed heavy cognitive demands on memory to recall the contents of a seven-minute film, depicting a complex, atypical event structure. The other tasks listed in the table all involve a canonic narrative action-structure: one or more main protagonists who face a problem or goal which arises from a background motivation and initiating event, their attempts to solve this problem or reach this goal, and a resolution which they finally achieve. The "pearstory" film violates these requirements, since it includes several different characters and events that are unrelated to each other and since there is no clear narrativ e closure. This makes it difficult to identify a global narrative structure, even at an age where a canonically organized storyline, such as the one depicted in Frog, where are you?, is clearly recognizable and can be verbalized as such. In sum, a survey of different contexts of narrative text production confirms the first of the two predictions formulated in the preceding section: Along with a shared developmental path in overall narrative structure, the ability to construct a well-organized narrative text will be manifested earlier and better in some situations than in others. The next section considers the second prediction: that children will recruit linguistic forms to express particular narrative discourse functions earlier or better in some settings than in others.


269

4. DIFFERENT CONTEXTS TAP DIFFERENT ABILITIES The claim here is that particular narrative contexts involve particular kinds of abilities, and as a result they will reveal certain aspects of narrative knowledge and storytelling abilities sooner or differently than others. Three domains are considered in this connection: reliance on deictic expressions as a function of shared knowledge; strategies for distinguishing referents; and expression of background, scene-setting story elements. In picture-based studies, it has been shown that shared knowledge between subject and investigator in the storytelling situation, such as in our frog-story study, might well encourage excessive reliance on deixis and other inappropriate means of referring to characters and situations, especially among younger children. This was so in studies which compared situations where children were asked to tell a given story to a blindfolded or a sighted investigator (Kail & Hickmann 1992; Kail & Sanchez y Lopez 1997). Under conditions of "no mutual knowledge" (where the addressee was blindfolded), children did better on providing unambiguous reference to characters across the story. Relatedly, where someone other than the experimenter is told the story, as tried out by Slobin and Kiintay, children used fewer deictic expressions such as this, here, now for referring to characters and situations than in the original design (Berman & Slobin 1994:24-25). Distinctions in making and maintaining reference to characters were also revealed by different situations in a picture-series task (Katzenberger 1994). Two different picture series presented to preschoolers aged 4, 5, and 6, to gradeschoolers aged 10, and to adults yielded rather different reference strategies. These turned out to be a function of the different narrative structure of each of the two series: One set of four pictures "portrays one clearly dominant character, a woman who performs several different activities with the goal of purchasing a hat and who is served by a secondary character, a woman who enables her to do so ... [in the other], in contrast, both characters perform much the same activity across the four pictures, of casting and pulling out their fishing rods, but one is successful in achieving the goal of catching fish, while the other is not" (Berman & Katzenberger 1998:37). For the "hat-purchasing" series, all the more proficient narrators - defined as those who gave independent evidence of a well-developed narrative schema since they constructed texts with a globally organized actionstructure - shared a preferred reference strategy: what Karmiloff-Smith (1981) termed a "thematic subject strategy". They focused on the woman who did the buying and either backgrounded or failed to mention the secondary character of the woman who did the selling. In contrast, a "parallel presentation" strategy was preferred for the other series, which depicts two children performing the same activity of fishing in a body of water, although one is successful and one is not. Moreover, while some of the older subjects used a "labeling strategy" in

270

Ruth A. Berman

both series, it fulfilled a different function in each case: in the hat-purchasing series, it took the form of "role-labeling" of one woman as customer, the other as salesperson; in the fishing series, the children were differentiated by name (e.g., 'Moshe' and 'Uri') and/or by sex (e.g., 'the boy' and 'the girl', 'Rami' and 'Tami'). These findings underscore a major theme of our original frog-story study: there is no one single nor even one "best" way to tell a story or to depict the same external (in this case, graphically depicted and fictive) reality. But here this theme is revealed from a rather different perspective. First, not only young children but even adult speakers of the same language may differ in the particular strategies they prefer to perform such discursive tasks as making reference (Wigglesworth 1997). These individual preferences are a critical feature of rhetorical style in a range of domains, including reference, temporality marking, and textual connectivity (Berman 1988a). Second, and more pertinently to the topic under discussion in this chapter, not only the type of materials and procedure adopted for eliciting a narrative text (series of four pictures) will affect the end product, but so will their specific content (distinct as compared with parallel roles of characters in the story). As a last example of the effect of context in a particular area of narrative construction, consider how children provide background, scene-setting information in different types of narrative texts. I compared this across the "fight story" personal experience accounts and the picturebook "frog story" texts (Berman 200la). It turns out that the type of task and context of elicitation has an impact on both the amount and nature of setting elements across age groups. The prediction of development with age in amount of scene-setting, defined as the proportion of clauses serving this function across the two types of texts, was not confirmed for the frog-story task. It yielded a consistently low proportion of clauses dedicated to scene-setting (the first picture out of a total 24 pictures in the book), between 4% to 7% of all clauses across age groups, including children aged 3, 5, 7, 9, and adults. In marked contrast, fight-story texts elicited from subjects in the same age groups revealed a clear and consistent rise in mean proportion of setting clauses with age, from less than 10% among preschoolers, around one-quarter at school age, and over a third among adults. These findings are robust, since they tally with findings for similar types of elicitation across other populations. Analyses of frog-story texts in other languages from the Berman and Slobin (1994) study, together with French data elicited by the same methodology by Kern (1997), reveal exactly similar trends. The first picture, providing the background antecedent to the plot-initiating event of the frog escaping from its jar, yields the same low figures for adult narrators in English (mean 5.5% of all clauses in the sample), Spanish (mean 7.6%), and French (mean 7.3%). In marked contrast to these low figures, analysis of the


271

setting element in a range of other fight stories elicited in speech and writing from other Hebrew-speaking schoolchildren and adults reveal a closely parallel trend to the earlier original fight-story sample listed in Table 1 (Berman 1999). The second- and fourth-graders (aged 7-10) produced an average of 20% to 25% background-setting clauses out of the total clauses in their narratives, while highschool students (aged 16-17) and university-educated adults devoted an average of as high as one-third to nearly half of their narratives to background-setting clauses. Further evidence for these general trends for "personal experience" stories is that there was no notable difference between the figures for the narratives produced in writing versus speech. These figures are consistent with other findings for amount of backgroundsetting material compared with overall text length as defined by number of clauses across a range of other materials, in English and in Hebrew. Development of children's ability to start a story by providing relevant, and adequate, background, scene-setting information proves sensitive to the particular narrative situation. Children were able to demarcate setting elements better and earlier in personal-experience accounts than in a picturebook story. They did so as young as age 3 when they were free to tell about anything that had happened to them but only from around age 5 when asked to tell specifically about a fight they had experienced. In general, personal-experience accounts appear to provide more authentic contexts for elaborating on scene-setting than picture-based elicitations. Moreover, a study based on a series of six pictures depicting what happens to two children at a zoo, where a monkey snatches the ice-cream from the younger child, indicates that young children may not bother to provide setting elements at all in picture-series narrations (Berman 1998b). They all clearly recognized the "zoo" script situation in the pictures, but only from school age would they provide any kind of scene-setting as a frame for their texts, and only children aged 6 and up produced stories which were organized around an acceptable actionstructure (similar to the findings for another set of picture series in Katzenberger 1994). In contrast to these picture-series elicitations, non-picture-based fictional accounts, where children are asked to create imaginary stories, appear to provide a particularly rich context for expressing early scene-setting abilities (as can be inferred from what kindergarten children do in a "pretend-reading" task). And indeed, background exposition plays a crucial role in construction of literary narratives. Analysis of "scene-setting elements" thus needs take account of the particular kind of story being told, for example, the script of an adventure-story in the genre of children's storybooks in the frog story, or the script of a conflict situation in the genre of personal-experience accounts in the fight stories. Further, suitably controlled research using comparable materials (e.g., based on the same theme, topic, or script) across different tasks is needed to test the claim that certain

272

Ruth A. Berman

story elements should be analyzed from a developmental perspective in some narrative subgenres rather than in others. The findings reviewed in this section for reliance on deixis, reference strategies, and expression of setting elements all point to the fact that different elicitation procedures and communicative contexts will promote, or at least allow expression to, some types of abilities earlier, better, and/or more than others. As against this, in a developmental perspective, once both narrative competence and storytelling performance are well-established, older and more proficient narrators are less susceptible to effects of task and context in formulating narrative story-settings as in other domains. This issue is addressed in the next section.

5. SHARED YET DIFFERENT DEVELOPMENTAL PATHS The comparisons in Sections 3 and 4 suggest that there is no reason to expect the frog-story findings to be directly mirrored in other elicitation settings and in other narrative subgenres. Two related factors determine how well, and how early, children will be able to carry out a particular narrative task. One is anchored in the interaction between "narrative competence and storytelling performance" (Berman 1995): the ability to make effective on-line use of one's knowledge of language and of principles of narrative discourse in order to produce a wellorganized and coherent narrative text. The second factor is the recognition that this ability (and most likely relevant abilities in other domains as well) need not develop along a uniformly linear curve. This appears to contradict evidence from more carefully controlled studies than those reviewed here (for example. Hudson & Shapiro 1991; Wolf, Moreton & Camp 1994), which imply that one story genre develops straightforwardly before another. This does not seem to be the case in fact. For example, generic scripts may not always be easier than accounts of personal experiences, since it depends critically on how familiar, how complex, and how open-ended the particular script happens to be. Cultural factors will also have an impact on relative task difficulty. For example, a "birthday party" is taken as a typically open-ended script in the United States, whereas Israeli preschoolers recognize a highly conventionalized and stereotyped format for such occasions. Nor is it clear that make-believe stories are invariably more of a challenge than generic scripts or personal-experience accounts. On the contrary, digressing into the realm of wolves, witches, and wicked stepmothers may make it easier for children to spin a tale than the requirement that they accurately recount a sequence of events in which they were personally evolved. Relatedly, it is not necessarily the case that one elicitation setting will always take developmental precedence. For example, picture series may not always be easier than picturebook stories, or film recountings harder than either of these two. It depends on a cluster of factors.


273

including: amount of adult input or other props and the nature of such scaffolding; familiarity of the materials and contents of the situation; and episodic complexity and the structural canonicity of the story itself. A second reason has more general, non-domain-specific developmental underpinnings. Many of the contextual effects noted in Sections 3 and 4, such as reliance on deixis as an immature reference strategy in picture-description tasks, are particularly marked among younger children, and exert less impact with age. Older, more proficient respondents cope with narrative as with other tasks in a more autonomous, less context-dependent fashion than young children. This does not mean that task-sensitivity disappears with age; rather, it takes on a different shape. Among young children, overall success is critically related to relative task complexity. As noted, five-year-olds may be able to produce a well-constructed single-episode account of a personal experience before they are able to handle the cognitively more demanding task of relating the narrative contents of a set of pictures. Older speakers, whose narrative competence is fully established, will produce well-formed narrative texts under any of these conditions, in line with the argument that they are less vulnerable to contextual variation. In fact, however, the impact of context-sensitivity manifests a U-shaped developmental curve (Strauss 1982). Like younger, non-proficient narrators, the narratives of older subjects will differ depending on the contextual situations in which they are produced. But the reason for this, and hence its effect, is different: proficient narrators are able to skillfully adapt the particular strategies they select to suit a given discourse context. In the case in point here, this is reflected by how only more mature and literate narrators adapt the register of their language use to the context of written or spoken modalities (Berman 1999; Berman & Verhoeven 2002). And it is shown in the flexibility of older subjects in selecting either a thematic-subject strategy or alternative referential strategies such as "parallel presentation" and "labeling" as appropriate to the particular type of script embodied in the their picture-based narratives (Berman & Katzenberger 1998). As noted there for the domain of reference, so in general, "the apparent similarity in the overt form of strategies adopted by young children and by older, more proficient narrators, is only a surface behavioral phenomenon, one which reflects quite different underlying competence in their two cases" (ibid:38). In sum, young children's command of narrative text production is essentially vulnerable. As long as this competence is not fully established, their skill in storytelling performance will be unevenly affected by context. They will tend to treat narrative production as an interactive communicative enterprise; they will be influenced by the facilitating effect of factors such as scaffolding and familiarity; and they will be highly susceptible to the cognitive demands of the particular elicitation task at hand. The generalizability of the original frog-story findings thus needs to be evaluated in terms of an overall conception of development, as

274

Ruth A. Berman

argued for in the work of Karmiloff-Smith (1985, 1986b, 1992) for both linguistic and other cognitive domains, and by the present author for a range of subdomains in acquisition of Hebrew morphology, syntax, and lexicon (Berman 1986. 1987, 1988b, 1990, 1993b, 1996). Before any relevant knowledge is acquired, prior to the inception of "task-readiness", no amount of prompting, scaffolding, or other contextual aids will avail. After knowledge has consolidated and matured to an integrated level of narrative competence, storytelling performance will once again be relatively free of constraints, and it will reveal flexible adaptation to the requirements of different narrative genres and varied communicative contexts. It is in the critical in-between or en route phases of partial knowledge and reorganizations of abilities that children will be most affected by variations in elicitation tasks and storytelling settings.

6. LANGUAGE DEVELOPMENT ACROSS ADOLESCENCE This concluding section is meant as a postscript to the original paper, which was written some time ago. It considers insights from a study that emerged out of work with Dan Slobin and his Berkeley-centered group culminating in the 1994 book to which the present volume is a sequel, and a crosslinguistic workshop at the Linguistic Institute of 1995, where the ideas motivating this sequel volume were first articulated. In contrast to the original and most subsequent frog-story studies, the project discussed in this section deals with both narrative and nonnarrative (expository) text construction, in both speech and writing, in seven different languages - including three of those in the original study (English. Hebrew, and Spanish) and four others (Dutch, French, Icelandic, and Swedish). The findings of this study, as summed up in the introductory article to Berman & Verhoeven (2002) and in Berman (in press), reiterate and underscore the key themes of our original project. They also elaborate on the original study in several directions, since the design of the new study was extended along a range of variables: a} in age and schooling the youngest, second-grade subjects were at the same age as the oldest in the original study (9-10), and the study included junior high school (12-13) and high school students (16-17) in each of the seven countries, while the adults were all graduate students doing their M.A. or Ph.D. at an established university; b) in text type and genre the new study elicited personal-experience accounts as a special subgenre of narrative to which were added expository discussions; c) in medium subjects produced a story and a discussion in both speech and writing, so that each subject produced four different texts; d) in elicitation procedures participants were all shown the same short video clip consisting of vignettes depicting situations of conflict (moral, social, and physical) in a school setting and then asked to tell and write a story of an incident where they had been involved in a situation of interpersonal conflict


275

and to give a talk and write a composition discussing the topic of "problems between people"; and e) in thematic content all texts related to social issues of interpersonal conflict, and so differ markedly from the children's adventure story depicted in the frogbook. Below I consider directions in which these different research variables serve to both reinforce and elaborate on the guiding themes of the original frogbook study. 6.1 Later developments: High school and beyond A first important conclusion to emerge from the new study concerns the interrelation between (linguistic) form and (discourse) function that constituted a leitmotif of the original study. The first and perhaps major finding of our study is that, indeed, language acquisition has a long developmental history (touched on in Section 3 of this chapter) and that it in fact continues well beyond adolescence. Specifically, we found across a range of topics dealt with in the studies reported on in the special issues edited by Berman and Verhoeven (2002) and by Berman (in press), that there is a major cut-off point between the type of linguistic expression and discourse construction that characterize the texts produced by gradeschool and junior high school children, on the one hand, and high school adolescents and university student adults, on the other. These distinctions underscore the theme of Berman and Slobin (1994) to the effect that the ability to recruit a rich and varied linguistic repertoire to meet a range of discourse functions imposes marked cognitive and planning burdens on speaker-writers, and that it may take as long as late adolescence to reach fruition. This finding is marked in such varied domains as lexical density and diversity (Stromqvist et al. 2002), the use of modal expressions to denote different prepositional attitudes (Reilly et al. 2002), use of passive constructions and other means of downgrading agency (Jisa et al. 2002; Tolchinsky & Rosado, in press), and use of pronominal and other devices in the expression of discourse stance (Berman, in press). The protracted route followed by schoolchildren of well-educated, typically middle- to upper-middle-class backgrounds who formed the population of this study found expression in three additional domains that were not considered in the original frog-story study: genre (expository versus narrative discourse); modality (writing versus speech); and register (level of language use). The new study confirmed the findings for a range of subgenres of narrative text production noted in Section 2 above: In general, by gradeschool age, children have command of a well-established narrative schema, and they can construct a well-formed narrative text organized around introductory setting elements, a sequence of events constituting narrative episodes or a sequentially ordered plotline, and a concluding resolution. Moreover, the youngest group in our sample, aged 8-10, were able to produce such texts in writing as well as in speech. A further finding was that by this age, children also distinguish clearly in linguistic expression

276

Ruth A. Berman

between narrative and expository texts - for example, in use of verb tense, mood, and aspect (preference for past tense and perfective aspect in narratives and reliance on the timeless present and irrealis modalities in expository texts): in forms of reference (use of personal pronouns and other means for specific reference in narrative texts compared with reliance on generic pronouns and impersonal constructions in expository texts); in voice (more use of passive and middle-voice constructions in expository than in narrative texts); and in lexical selection (more dynamic predicates in narratives, greater reliance on abstract nominals in expository texts). As against this early emergence of inter-genre distinctiveness, we found that children did not manifest command of expository-text construction until high school age (Berman & Nir, in press; Katzenberger, in press). Only from adolescence on were participants across the study able to organize the information in their expository texts hierarchically, around a well-defined introductory opening, then to develop key themes presented earlier in the body of the text, and to reach a final summary or conclusion deriving from the contents of the preceding discourse. Again, this was a robust finding across languages, revealing that highschoolers were more similar to the adults in the study, whereas younger children were unable to cope with the cognitively demanding task of generating original and clearly argued text content while concurrently organizing this information in the form of a well-defined global text structure (Berman & Katzenberger, in press; Tolchinsky et al. 2002). Another development which goes way beyond early school age is mastery of written language as "a special discourse style" (Ravid & Tolchinsky 2002). Here, too, even the youngest children in the study demonstrated some distinction between how they wrote and how they spoke. This was shown, for example, in their tendency to express greater informational density and more complex syntactic packaging (Berman & Ravid, in press; Ravid, in press), in their recourse to such linguistic means as passive voice, in greater lexical density, and in avoidance of interactive discourse markers such as English like, ya know, okay in their written compared with their spoken texts (Jisa et al. 2002; Stromqvist et al. 2002; and see, too, Stromqvist, Nordqvist & Wengelin, this volume). But it is not until high school that children manifest a well-developed level of "linguistic literacy" in the different ways in which they organize both the structure and the content of the texts they produce in writing compared with speech, in both narrative and expository types of discourse. Relatedly, the ability to make consistent and appropriate distinctions in linguistic register and level of language usage is a late development, which depends on extensive exposure to and experience with literacy-related activities of a kind that consolidate mainly in high school. This finding, too, emerged across the sample in our project, although this aspect of language use is particularly vulnerable to the sociolinguistic attitudes and values of the cultures involved, as expressed in different systems of language education


277

in the countries in our project (Aisenman 1999a:291-341) and in different levels and types of diglossia in each of its seven languages (Aisenman 1999b). 6.2 Form/function approaches revisited A second leitmotif of the earlier study which the new project reiterates is both methodological and principled: the need to examine linguistic forms in relation to their discourse functions. In the first place, in complex crosslinguistic undertakings such as the original frog-story study and the present project, definition of linguistic forms is essential as a starting point for analyzing their distribution. We attempted to define comparable or at least compatible structural categories for coding across different languages and age groups - along lines described in the special journal issue edited by Berman and Verhoeven (2000) and analysed there for a range of topics: for lexical distribution in terms of diversity and density (Stromqvist et al.), forms of subject nominals (Ravid et al.), verb tense and aspect (Ragnarsdottir et al.), syntactic packaging (Verhoeven et al.), and passive-voice constructions (Jisa et al.). Such studies constitute an invaluable point of departure for more functionally motivated analyses, although, as pointed out in Berman and Nir (in press), it is extremely difficult to arrive at directly comparable categories of coding with respect to both units of analysis (what constitutes a "word", a "clause," or a "clause package") and categories of analysis. For example, does passive voice include both be and get passives in English and what do these correspond to in, say, French, Spanish, or Hebrew? Is the English grammatical system of modal auxiliaries in fact comparable to its counterparts in the other languages, including both agent-oriented counterparts to English can, must in French, Hebrew, and Spanish (poder, deber}! How do these interact with impersonal constructions like French ilfaut, and Spanish and Hebrew hay que, tsarix, respectively? Are generic pronouns like you, one, we analogous in the different languages, and how do these relate to, say, French on, Dutch men, and Swedish man? The new project has attempted to re-evaluate these issues in more functional terms from several perspectives: in relation to the notion of "competition" between alternative means for expressing the agent-downgrading function of passive constructions to explain their differential distribution in five of the languages in our sample (Jisa et al. 2002); in relation to the notion of lexical diversity, density, and wordlength in different languages (Stromqvist et al. 2002); in a reassessment of the traditional division between open- and closed-class lexical items (Berman 2001b); and in analyzing the use of different types of modal expressions in English, French, and Hebrew as means of realizing deontic compared with epistemic propositional attitudes (Reilly et al. 2002). Three other functionally oriented types of analyses have emerged as valuable points of departure for crosslinguistic comparisons in the new project. First, a start

278

Ruth A. Berman

has been made on relating the use of different types of lexical items, mainly verbs and nouns, analysed in terms of both morphological structure (e.g., derived versus basic) and semantic category (e.g., concrete versus abstract) to developmental differences in expressing genre distinctions and the thematic content of texts (Berman 2000; Ravid & Cahana-Amitay, in press). Second, analysis of opening and closing elements of texts in terms of communicative elicitation setting, discourse function, and linguistic structure and content has proved a valuable means of analyzing topdown, global text structuring and organization across age groups and text types (Berman & Katzenberger, in press; Tolchinsky et al. 2002). Third, an initial attempt at redefining the notion of "discourse stance" in crossgenre, cross-modality, and developmental perspective (Berman, Ragnarsdottir & Stromqvist 2002) has yielded a rich set of functional analyses of the use of linguistic systems such as pronouns, passive voice, and impersonal constructions in each of the languages in our sample (Berman. in press). Each of these analyses concerns a single language (except for one that compares use of generic pronouns and passive voice in the two closely related languages of Icelandic and Swedish) - and in this sense they are not strictly crosslinguistic. However, all alike focus on the shared theme - what we defined as "discourse stance" - so that together they provide highly convergent findings for the nature of form-function interrelations in the expression of a more or less distanced discourse stance, of more general or specific reference, and of a more text-based, sender-based, or receiver-based orientation towards the topic of "interpersonal conflict" in narratives compared with expository texts, in speech compared with writing, across four different age groups. The studies all reveal clearly marked distinctions between narrative and expository text structure and content from the earliest age group, but they also show that only around adolescence and particularly among highly literate adults will speaker-writers deploy a rich repertoire of "distancing devices" of different types and at different levels of language usage in order to mark their written expository texts as expressing a less involved and personalized discourse stance than, say, their oral narratives. These analyses underline two important themes of the new study. First, forms of linguistic expression reveal a complex interaction between genre (personal-experience narrative versus expository discussion) and medium (speech versus writing) - often conflated in analyses of linguistic register and style, as clearly articulated in the introduction to the paper by Tolchinsky and Rosado (in press). Second, the latter's analysis of five different means for agent downgrading in (mainland) Spanish, like the cluster analysis conducted by Reilly et al. (in press) on the use of impersonal pronouns, passive-voice constructions, and modal predicates, underscores an idea that served as a guiding theme for the entire undertaking: the notion of a "confluence of cues" (Berman & Verhoeven 2002, Section 3.3), suggesting that a range of different linguistic devices and subsystems conspire together to express a given discourse function.


279

6.3 The lexicon/syntax interface Our study reiterates yet another theme which emerged out of the original frogstory project, and which has found important echoes in Slobin's subsequent analyses of the expression of motion events in different languages, text types, and age groups (Slobin 1996b, 2002b). The idea here is that in the expression of form-function relations in discourse, there is an intimate interconnection and dependency between lexicon (in morphological form and semantic content) and syntax (in both clause-internal and clause-linking constructions). This is revealed, for example, by the form and function of subject nominals in expository compared with narrative texts and in the use of different classes of modal expressions to denote prepositional attitudes (Ravid & Cahana-Amitay, in press; Reilly et al. 2002); by the differential marking of discourse segments and "clause packages" through use of different types of discourse markers (Katzenberger & CahanaAmitay, in press); by the distinction between different classes of discourse markers in different syntactic positions and different discourse functions in speech and writing, in narrative and expository texts; by the interrelation observed in Spanish (Tolchinsky & Rosado, in press) and English (Reilly et al., in press) between use of grammatical constructions and expression of particular types of thematic content; and by the developmental shift observed by Ravid and Cahana-Amitay (in press) from strictly verbal predicates to more nominal and morphologically derived types of predications in Hebrew narrative texts, reflecting the fact that beyond junior high school, participants relate to more abstract situations and make different kinds of generalizations concerning the conflicts they describe. Like the other themes noted here, this idea, too, warrants further and more detailed examination for different languages across a range of discourse genres, elicitation settings, and age groups. The ideas outlined here, like those of the original frog-story project and its extensions in the present volume, will hopefully serve as a springboard for yet more crosslinguistic research on developing discourse abilities from a variety of perspectives.

NOTES 1.

Thanks to Dan Slobin for continuing to be a source of support and inspiration, and to the editors of this volume for their helpful input.

2.

I use the term phases in preference to the Piagetian notion of age-bound, domainneutral stages for reasons similar to Karmiloff-Smith (1986b, 1992).

3.

The contents of this section are based on research for an earlier study comparing results of the Hebrew frog-story sample with findings from a range of other studies on the narrative productions of Hebrew-speaking children (Berman 1995, Section 3).

280

4.

Ruth A. Berman

Texts were elicited from different groups of subjects in each of the studies listed in Table 1. Yet comparisons are justified, since in all cases the subjects were children of educated, middle-class speakers of Hebrew as a first language (like the adult subjects in each study). Besides, highly consistent findings are reported by a carefully controlled study (Lotan 2003) that examined the narrative texts produced by Hebrewspeaking children of similar backgrounds, but in which the same 60 children (aged 5. 7, and 9) produced personal-experience narratives similar to those in Task 2 in Table 1 and also picture-series based narratives (Task 3 in Table 1). All the preschoolers attended Hebrew nursery school from age 2 and kindergarten from age 5 to 6, and the schoolchildren attended grade school from age 6 to 11 or 12.

Coherence, Cohesion, and Context Some Comparative Perspectives in Narrative Development

Maya Hickmann

1. INTRODUCTION Coherence and cohesion are two essential facets of narrative organization, which both evolve during the course of children's discourse development. Although much current research has focused on one or the other of these aspects of children's narrative competence, little is still known about their relation during the course of development. Furthermore, a growing body of results reveals variations in children's and adults' narrative performance, which have implications for our understanding of the development of cohesion and coherence. In this chapter, I focus on two types of variations: cross-linguistic variations, showing the impact of language-specific factors on narrative development, and contextual variations, showing that children's narrative performance partially depends on some dimensions defining narrative situations and tasks. I illustrate below some properties of coherence and cohesion, as well as cross-linguistic and contextual variations in children's discourse. Findings and examples are drawn from a variety of situations, both naturalistic and experimental, with particular attention to various types of narrative productions, including frog stories (Herman & Slobin 1994; Mayer 1969). On the basis of this material, I first argue that an exclusive focus on coherence or on cohesion can provide only an incomplete picture of narrative development. A discussion of several possible views leads to the hypothesis that these two aspects of discourse organization are intricately related, rather than merely parallel components of children's narrative abilities. Furthermore, cross-linguistic and contextual variations highlight the impact of fundamental factors that cannot be ignored in any account of coherence and cohesion during narrative development. I conclude with the suggestion that the use of language (and of particular languages) imposes some constraints and organizing principles on the processes of constructing and communicating stories in relation to context.

282

Mava Hickmann

2. COHERENCE AND COHESION As a starting point, I first use the terms coherence and cohesion in a general way to refer to two complementary aspects of narrative organization: roughly, coherence includes those aspects of a narrative that are directly linked to the overall structure of its content, while cohesion involves those aspects that pertain directly to the linguistic expression of discourse-internal relations across clauses. As will be shown, the more specific definitions that have been provided for these terms revolve around a number of dimensions that have been invoked as a way of differentiating these two different aspects of narrative organization. I return to these dimensions below in a more detailed discussion of the relation between coherence and cohesion, after a brief summary of different approaches to narrative development. 2.1 Coherence and its development A number of researchers have provided a general account of know ledge about complex event sequences in the form of macro-structures or knowledge structures that actively integrate the information to be narrated. Two main types of macrostructures have been postulated in the developmental literature: scripts and stoiy grammars. Notwithstanding important differences between them (see more details in Hickmann 2003a), both have been claimed to provide children with cognitive schemata underlying their internal representations of complex events as well as their external verbalizations of these events in narrative discourse. Scripts correspond to cognitive structures enabling children to represent common behavioural routines, such as going to the restaurant or to a birthday party (e.g.. French & Nelson 1985; Hudson & Nelson 1983; McCartney & Nelson 1981; Nelson 1986; Nelson & Gruendel 1981; O'Connell & Gerard 1985; Shank & Abelson 1977; van Dijk 1980). Such routines typically involve ordered sets of familiar and predictable events, which have typically been experienced by the child on one or more occasions and the structure of which is determined by a logical and/or conventionally agreed-upon chronological sequence. In contrast, stories cover a wider range of event sequences, which typically include unfamiliar events not previously experienced by the narrator on any particular occasion. As will be shown below, they also involve more variability with respect to how these events are presented, owing to a greater variety of organizing principles being inherent to discourse cohesion in this case. Much research has aimed at determining the general principles or rules that characterize structurally well-formed stories (e.g., Applebee 1978; Berman & Slobin 1994; Fayol 1985; Labov & Waletzky 1967; Mandler 1978; Mandler & DeForest 1979; Mandler & Johnson 1977; Stein 1982; Stein & Albro 1997; Stein & Glenn 1979, 1982; Stein & Nezworski 1978; Stein & Trabasso 1981). Labov

Coherence, Cohesion and Context in Narrative Development

283

and Waletzky (1967) have provided what is perhaps the most general and most often cited account, according to which the basic structure of stories minimally comprises several components: a setting, a complication, and a resolution, as well as an evaluation, which provides the narrator's attitude towards the denoted content. A number of story grammars have also been proposed as a way to further specify the structural properties of stories, such as their basic constitutive units and the rules defining how these units can be hierarchically and sequentially arranged into well-formed stories. For example, the story grammar proposed by Stein (1982) generates stories which contain a Setting, followed by an Episode consisting of the following sequence of units: an Initiating Event triggers an Internal Response on the part of the main protagonist, motivating his/her Attempt to reach a goal, which leads to some Consequence resulting in a Reaction on his/her part. Similarly, Mandler (1978) postulates a story grammar consisting of the following units: Setting, Beginning, Reaction, Attempt, Outcome, Ending. Related proposals (e.g., Stein & Albro 1997; Stein & Trabasso 1981; Trabasso & Rodkin 1994) present a more abstract goal/plan structure, revolving around a main protagonist capable of intentional action. Macro-structures have been claimed to be psychologically real and to guide the processing and production of stories by adults and children. When subjects are asked to recall structurally anomalous stories, their recall of the stories decreases and they modify the texts in such a way as to make them more similar to well-formed texts. For example, when narrative units are deleted or displaced to non-canonical positions, story recall decreases and subjects rely on various narrative strategies whereby they make the story more compatible with its canonical version. These strategies may consist of modifying the order of displaced units, integrating dispersed information into an episodic structure, producing stories containing two episodes rather than one, or generating information corresponding to deleted units. These findings suggest that story schemata provide a strong temporal structure, which serves as a retrieval strategy during story recall, even for young children (from five years on) who may not yet be able to provide judgements about anomalous stories. Indeed, another body of research (see a review in Hickmann 2003a; see also below) shows clear developmental progressions in children's ability to judge several aspects of story content and structure, e.g., to recognize central versus secondary information, to make appropriate inferences, to detect and correct inconsistent or missing pieces of information, and to judge what is a good story. For example, although young children (five to seven years) can differentiate stories from non-texts (unconnected sentences), only older children (nine to eleven years) can also differentiate stories from incomplete stories and from scripts (Esperet 1991). Studies examining children's spontaneous or elicited narrative productions further show the central role of various story schemata from the earliest ages on, as well as developmental progressions in how these schemata are externalized in

284

Maya Hickmann

discourse (e.g., Applebee 1978; Bronckart & Schneuwly 1991; Eisenberg 1985; Esperet 1991; Fivush, Gray & Fromhoff 1987; Gopnik 1989; McCabe & Peterson 1991; Menig-Peterson 1975; Menig-Peterson & McCabe 1978; Miller & Sperry 1988; Peterson 1990; Peterson & McCabe 1983). For example, Berman and Slobin (1994) show a basic understanding of the frog-story schema at all ages, as well as some differences in which plotline components are narrated across age groups. Whereas the adults systematically provide all components (onset, unfolding, resolution), children focus on some of these components, e.g., failing to provide explicit references to problem-solving or to internal states until five years of age, even though they can be induced to fill in missing information if sufficient scaffolding is provided (Trabasso & Rodkin 1994). 2.2 Cohesion and its development Cohesion is directly related to the linguistic expression in discourse of the content to be narrated, and more particularly to the regulation of information across utterances (e.g., Chafe 1976, 1979; Clark & Haviland 1977; Halliday & Hasan 1976). Constructing cohesion involves several inter-related principles governing how speakers use linguistic devices to organize discourse-internal relations. According to the general principle of discourse anchoring, speakers must establish the personal and spatio-temporal parameters of the denoted situation in order to allow the interpretation of subsequent discourse. Furthermore, they must continuously mark the status of information as a function of mutual knowledge. For example, when the existence and identity of referents are not mutually known, speakers must first introduce these referents, then differentiate the degree to which their identity can be presupposed in the universe of discourse. A number of linguistic devices are available for this purpose in any given language. Most IndoEuropean languages, for example, provide some local markings in their referring expressions: an opposition between indefinite and definite nominal determiners that differentiates new from given information, and different forms of reference maintenance (e.g., English definite nominals, explicit pronouns, zero elements) that indicate the degree to which the identity of the referents is accessible and can therefore be presupposed in discourse. Finally, speakers must simultaneously ground information as a function of communicative focus. In particular, they must differentiate the foreground and background of discourse as a function of what information is more central or more secondary in the narrative. The foreground corresponds to the skeleton of the plotline, comprising the chronologically ordered situations that make the story move forward in time. In contrast, the background corresponds to secondary information that surrounds the foreground and that need not be chronologically ordered. The best illustration of how this principle is linguistically operationalized is provided by a number of temporal-aspectual devices (verbal morphology and


285

particles, connectives, adverbials) that greatly contribute in all languages to this type of marking (e.g., Hopper 1979a, 1979b, 1982; Tomlin 1987). For example, in the sequence when Susan came in, Peter was washing the dishes, Susan's arrival is presented in the foreground in relation to Peter's dish-washing, which is presented in the background. In contrast, in the sequence when Susan came in, Peter washed the dishes, the two events are presented as successively occurring in the foreground. The developmental literature presents a growing number of findings which bear on both of these principles. However, some of these findings have led to different conclusions concerning the rate, course, and determinants of cohesion development. With respect to the marking of information status, studies focusing on children's use of referring expressions present diverging results concerning the age at which they master these devices (as early as three years and as late as ten years). A synthesis of these results (see a detailed review in Hickmann 2003a) shows at least two sources of this divergence. First, some methodological problems make it difficult to synthesize the results, because studies vary along a considerable number of dimensions. I return to this point below, since it bears directly on the impact of contextual variations on children's narrative performance. Second, researchers have examined different properties of these devices. Whereas an exclusive focus on syntactic aspects of referring expressions has led to claims of early mastery, studies focusing on the semantic and/or pragmatic properties of these devices are more heterogeneous in this respect (Emslie & Stevenson 1981; Hickmann 1991, 1995; Maratsos 1976; Power & Dal Martello 1986; Sauvaire & Vion 1989; Peterson & Dodsworth 1991; Sachs 1983; Warden 1976, 1981; Wigglesworth 1990). In general, studies claiming relatively late mastery place emphasis on the gradual construction of discourse-internal cohesion, which evolves until at least ten years of age. In this view, although young children use a variety of devices from two or three years on, they do not use them in the same way as older children or adults. Furthermore, it is precisely the discourse-cohesive functions of these devices for anchoring, for the marking of information status, and for the grounding of information that emerge late (from six or seven years on) and that continue to develop thereafter. A review of results based on a variety of naturalistic and experimental situations (Hickmann 1991, 1995, 2003a; Karmiloff-Smith 1979) shows that early uses of referring expressions often denote already-known entities (e.g., deictic uses of definite forms) or salient entities to which the child draws attention by primitive markings of information status (e.g., uses of indefinite forms in deictic labellings), and that they also serve other functions (e.g., uses of indefinite forms for non-specific reference). Other analyses based on narrative situations (Karmiloff-Smith 1981, 1985, 1987, 1992) further show a gradual developmental progression in children's discourse strategies. Children are first driven by bottomup processes, whereby they are guided by the external stimuli and do not relate

286

Maya Hickmann

their utterances within discourse. Later on, they rely on top-down processes, whereby they impose a thematic subject strategy, organizing discourse around a main protagonist, for which they exclusively reserve subject pronominals. During a last phase, their narratives show a reliance on both types of processes, resulting in less rigid types of discourse organization. With respect to temporal-aspectual devices, studies have begun to show that discourse factors are fundamental determinants of acquisition across languages (e.g., Bazzanella & Calleri 1991; Berman & Slobin 1994; Bamberg & Marchman 1994; Hickmann 1996, 2003a). Despite some early uses of temporal-aspectual devices, children gradually learn to use these devices to mark pragmatic distinctions in social interaction and/or in narrative discourse. For example, some discourse uses of these devices appear late in frog stories (Bamberg & Marchman 1994; see also below) as well as in other stories (Hickmann 1996, 2003a)'. such as uses that differentiate the discourse foreground and background and/or that mark relations between overlapping events. Such findings have led researchers to modify or even to reject other developmental hypotheses that do not take into account such discourse-functional properties of temporal-aspectual devices. For example, the frequently cited defective-tense hypothesis puts forth that the emergence and development of temporal-aspectual devices are determined mainly by universal concepts of situations, which are available early, and by children's cognitive development, leading them to use past tenses first exclusively to denote results, then to denote a greater range of situations. Finally, some findings show cross-linguistic differences in how a variety of devices are used for personal, temporal, and spatial reference in discourse, casting further doubts on such universalistic hypotheses. I return below to this cross-linguistic variation and to its implications for our understanding of narrative development. 2.3 Relating coherence and cohesion in development Although much research has been carried out on either coherence or cohesion, the literature is in great need of further empirical evidence concerning the relation between these two facets of discourse organization. Many studies focusing on narrative coherence have postulated the existence of cognitive schemata underlying children's and adults' performance in various tasks (e.g.. comprehension, production, and recall of canonical versus structurally anomalous stories), independently of the cohesive properties of the stories. Inversely, many studies focusing on cohesion have been concerned with children's uses of a variety of linguistic devices in different situations (narrative productions, processing of more or less cohesive stories), independently of the structural properties of the stories. The question of precisely how to relate coherence and cohesion in a general account of discourse development remains open. Although the need to


287

address this question has been stressed by some authors (e.g., discussions in Berman & Slobin 1994; Fayol 1985, 1991), studies have only barely begun to examine it empirically. Several approaches to this question are possible. Some studies merely postulate simultaneous parallel developments in coherence and cohesion, without making any further particular claim about the relations that might exist between them. For example, some show developmental progressions in both aspects of children's narrative organization (e.g., Orsolini 1990; Shapiro & Hudson 1991). Others find linguistic correlates of the ability to construct story structure (Stein & Albro 1997) and/or of the ability to comprehend story content (e.g., Stenning & Mitchell 1985). However, no explicit account is given of the relation between such correlates and the postulated underlying schema from a dynamic point of view that would account for developmental change. There is an obvious need to provide such an account, as shown below, because of an increasing literature on narrative development indicating an interdependence between coherence and cohesion that must be further examined. Second, a number of studies argue or imply that macro-structures (scripts or story grammars) are major determinants of the use and acquisition of cohesive devices (e.g., French & Nelson 1985; Orsolini 1990). Scripts guide children's recall and verbalizations of familiar event sequences, for which they typically display a much better narrative performance than for other types of event sequences, suggesting that familiar routinized event sequences facilitate the emergence of some discourse skills. For example, studies show earlier uses of some linguistic devices when children tell scripted narratives based on familiar and experienced events than when they tell other types of narratives. In these situations, children as young as two or three years old use more connectives and adverbials, as well as more inflections marking temporal-aspectual distinctions, and they show some ability to distinguish between some uses of definite and indefinite nominal determiners (French & Nelson 1985). These findings therefore raise the question of the relation between form use and narrated content and/or discourse type (see also below), challenging the idea that children are unable to use devices appropriately merely because of their cognitive immaturity. Finally, yet another different approach postulates a more complex account of children's narrative competence, according to which there would be important interactions between these two aspects of discourse organization, each influencing the other during the course of development. A particular version of this view, which 1 adopt below, argues that cohesion might also be partially constitutive of coherence. Such a view is implicit in the conclusions of Berman and Slobin (1994), although again no explicit account is provided of how the processes of simultaneously constructing coherence and cohesion might unfold. As further discussed below, new research is necessary to address this question, particularly

288

Maya Hickmann

in order to disentangle the effects of different variables related to coherence and/ or to cohesion that have been confounded in previous studies. Approaching the relation between coherence and cohesion raises a number of more specific questions. One question concerns the type of organization that is best captured by these two aspects of discourse. Most authors oppose cohesion and coherence in terms of the local versus global nature of the discourse-organizing processes involved. An illustration of this view is provided by Berman and Slobin (1994:40), who define linguistic cohesion on the microlevel of individual clauses and adjacent clauses, and thematic coherence on the macro-level of plot organization, further noting that "across languages there is a common developmental pattern towards increasing cohesion and coherence". Such a distinction between local and global organization is directly related to (and sometimes synonymous with) another distinction between sequential and hierarchical organization. Such oppositions capture the general idea that a large component of discourse cohesion pertains directly to the linear organization of successive clauses in unfolding discourse (e.g., marking continuity in reference maintenance), while a large component of narrative coherence pertains directly to the hierarchical organization of narrative units into larger chunks (e.g., organizing narrative units into episodes). Nonetheless, these oppositions must be strongly qualified in two major ways. First, global and hierarchical types of cross-clause organization are also fundamental to the construction of cohesive discourse. As noted above and as further discussed below, several central aspects of children's narrative require overall discourse planning that goes beyond adjacent clauses. For example, the marking of information status partially depends on the construction of a global thematic subject or discourse topic (Karmiloff-Smith 1981. 1985). and the grounding of information involves processes of overall temporal anchoring (Berman & Slobin 1994; Hickmann 1995, 1996, 2003a; Hickmann. Hendriks & Roland 1998). Inversely, local and sequential types of organization are also fundamental aspects of coherence. Thus, coherence to some extent involves adjacency, for example in the ordering of successive units within episodes and in the ordering of successive episodes in the narrative schema, to which children are sensitive from an early age on (Mandler 1978). The second question concerns the narrative units postulated by story grammars and the different discourse factors that may determine their inherent properties. For example, the structural category of the Setting (Stein 1982: Mandler 1978), which provides the initial orientation of the story (Labov & Waletzky 1967), plays a crucial role in discourse cohesion, serving to set the initial parameters of the plot: it provides introductions of the main characters, along with an initial spatio-temporal frame, which are necessary for the interpretation of subsequent discourse. It is therefore no trivial question to ask whether the very existence of this category is largely (albeit perhaps not entirely)


289

determined by discourse-cohesive principles. This question is linked to uses of particular linguistic devices which exist in all languages for the organization of discourse cohesion and which presumably affect our judgements concerning the presence or absence of initial story parameters in children's narratives. For example, in the absence of relevant mutual knowledge, the mere mention of the protagonist (e.g., by means of a proper name, a definite noun, or a pronoun) is an insufficient means of establishing the main personal parameter of the story, since the marking of information status requires a more explicit referent introduction containing some adequate linguistic device serving this purpose. Similar remarks can be made about the spatio-temporal parameters of the story, which are marked by a variety of devices. For example, temporal adverbials and inflections provide temporal anchors, while some explicit introductions of entities and of locations provide spatial anchors. Examining the necessary linguistic marking of coherence is therefore an important aspect of how we might go about relating cohesion and coherence in children's narratives. Third, the set of rules assumed to govern the ordering of narrative units constitutes another aspect of narrative structure that might be dependent on uses of cohesive devices. The general principle of information grounding involves a variety of rules or assumptions on the part of interlocutors, including the default principle of chronological order. According to this principle, speakers assume that the order in which events are successively denoted in discourse follows the order in which these events really occurred, unless the utterances are otherwise marked by means of specific linguistic devices. In fact, such a default principle is largely operational in the interpretation of scripts, which present a series of successive events in the foreground of discourse, whereas non-scripted stories typically involve more complex grounding relations in discourse organization. As noted above, the fact that the Setting is placed in the initial position of the narrative is no coincidence, since this unit provides the initial parameters of the story, which serve to anchor the remainder of it. Similarly, the immediately subsequent event (e.g., the Initiating Event in Stein's story grammar) typically introduces new situations and secondary referents that trigger the rest of the plot. Finally, the remaining units are chronologically ordered, even if they call for a number of assumptions or inferences and for general-world knowledge. In this respect, story grammars best capture causal chains among states or events, such as the fact that Initiating Event leads to psychological states that themselves lead to actions and to various consequences resulting from these actions. However, this ordering of units can be violated to some extent without this necessarily resulting in a structurally anomalous story, provided that the order violations are accompanied by adequate linguistic devices that ensure cohesion and thereby allow the interlocutor to reconstruct the correct sequence of denoted events (e.g., oppositions between verbal inflections or connectives such as before, after.

290

Maya Hickmann

while, because, etc.)- Thus, above and beyond the content of particular narrative units, their ordering into a canonical story structure partially reflects discourse principles and is linked to uses of particular linguistic devices. Finally, these points raise the more general theoretical question concerning the existence of abstract story schemata and their actualization in particular discourse situations. Two positions have been implicitly invoked in studies of narrative development. One approach views children's underlying cognitive representation of story structure as a major determinant of their narrative competence. This position assumes that the development of narrative competence is mainly or exclusively the result of an internal abstract and universal story schema which might be present in some form very early, but which becomes gradually more sophisticated and/or is externalized in gradually more sophisticated ways. Some have further assumed that early non-verbal capacities in temporal ordering, which are independent of the language faculty and already available during the pre-linguistic period, constitute fundamental prerequisites underlying the ability to order event sequences in discourse (e.g., Bauer & Mandler 1989). A second approach views children's representation of stories as (at least partially) a byproduct of principles governing discourse cohesion. This view assumes that some such representation is actively constructed on line as the narrative unfolds, being shaped by the very process of linguistic organization as the narrator communicates a particular story in a particular narrative context. It postulates that the way in which we organize events mentally is not independent of the constraints imposed by discourse (Berman & Slobin 1994; Hickmann 2003a; Slobin 1996b). In comparison with the first view, then, this second view emphasizes the fact that non-verbal prerequisites and abstract language-independent cognitive structures cannot account alone for discourse organization, because language use imposes its own organizing principles onto our cognitive activity. When examining these different questions, the evidence first suggests that coherence and cohesion are constructed simultaneously and that they are intricately related. A first point to be noted in this respect is that the same linguistic devices simultaneously ensure linguistic cohesion and mark the macrostructure of stories. For example, in their analysis of frog stories. Bamberg and Marchman (1994) show that, despite some developmental differences, narrators in all age groups use various devices (lexicalized inceptive aspect, forestalling devices, more explicit references to some characters) to foreshadow or wrap up information, simultaneously chunking this information into units and sub-units of various types (episodes, scenes) and indicating how these chunks relate to previous and subsequent discourse. Some studies (Bamberg 1997a: Bronckart & Schneuwly 1991; de Week 1991) also show that children make differential use of linguistic devices across situations (e.g., first-person versus third-person narratives; see also below) and that they use referring expressions and temporal-


291

aspectual devices in a variety of third-person narratives simultaneously to construct discourse cohesion and to mark the macro-structure of stories. In addition, as discussed elsewhere (Hickmann 2003a; Hickmann & Schneider 2000), the types of structural violations that have been tested in previous studies (e.g., displacement, deletion, scrambling of narrative units) typically result in simultaneous violations of story-cohesion principles. For example, such violations occur in stories where units containing referent introductions are deleted and where the displacement of a unit to a non-canonical position (in the absence of additional linguistic devices) in and of itself violates the default principle of chronological ordering. Furthermore, if a displaced unit contains a reference to an entity, the resulting text may display an anomalous sequence of referring expressions (e.g., a definite form maintaining reference to an entity before an indefinite form introducing this entity). Indeed, the evidence shows that children are sensitive to the violations of cohesion principles that result from structural anomalies. The deletion of the Initiating Event causes a drastic decrease in recall and frequent use of a strategy whereby children spontaneously generate the missing information. Similarly, some particular displacements result not only in a high decrease in recall, but also in a strategy consisting of producing two episodes rather than one (see also further discussion below).2 In both cases, it is likely that children's reactions to the structural anomalies may simultaneously result from their attempts to recover both the canonical structure of the story and its cohesive well-formedness. There is indeed some evidence that children are sensitive to anomalies in discourse cohesion. Some findings (Hickmann & Schneider 1993) show that children can restore discourse cohesion when they process stories that contain anomalous referring expressions, such as anomalous definite forms used for referent introductions or anomalous indefinite forms used for reference maintenance. This ability can be clearly observed in some tasks (story recall, repetition of sentences extracted during story presentation), even though children are unable actively to detect and judge these anomalies. Such a gap has been reported in other studies, showing that very young children can produce selfcorrections of cohesive devices before they can reflect on such uses (KarmiloffSmith 1986a, 1992). In addition, studies show that children use cohesive devices such as pronouns to integrate information during on-line text processing from seven years onwards (e.g., Tyler 1984). Finally, it is precisely the inability to take into account anaphors as indices of discourse cohesion that leads some populations of children to encounter difficulties in story comprehension (e.g., Ehrlich, Remond & Tardieu 1999; Garnham, Oakhill & Johnson-Laird 1982; Schneider, Williams & Hickmann 1997). Attempts at relating cohesion and coherence empirically are rare. One study (Garnham, Oakhill & Johnson-Laird 1982) examines how adults and

292

Maya Hickmann

eight-year-old children (good versus bad comprehenders) recall three types of written stories: well-formed stories; randomized versions, in which sentences occur in random order; and restored randomized versions, in which referring expressions are changed to restore the referential continuity of the randomized versions. Although recall was better for all subjects with the well-formed versions, the adults and the children diagnosed as good comprehenders recalled the well-formed and the restored randomized versions significantly better than the randomized ones. These results suggest that cohesive devices facilitate children's processing of stories by the age of eight years. Some other research (Hickmann & Schneider 2000) has begun systematically to compare the effects of violations in both narrative coherence and cohesion. English-speaking children between five and ten years old were asked to listen to and retell stories containing anomalies of two types: cohesion anomalies, which consisted in inappropriate sequences of referring expressions; and coherence anomalies, which consisted in violations of the canonical story structure (e.g., deletion or displacement of a narrative unit). In the latter case, the resulting non-cohesive sequence of referring expressions was either left intact (simultaneous anomalies in cohesion and coherence) or it was changed to restore the cohesive referential chain (structural anomaly with appropriate cohesion). The results show that age, story structure, and discourse cohesion all have a strong impact on children's uses of appropriate linguistic devices for referent introductions, as well as on their reliance on some of the narrative strategies discussed above. For example, children's retellings show frequent use of the strategy illustrated in [1]. In this example, which was elicited with a story version simultaneously involving structural and cohesion anomalies (displaced unit, unrestored cohesion), the child introduces two target referents, rather than one, using devices (underlined) that clearly indicate this double introduction and the resulting double-episodic structure. Further analyses will determine the extent to which such strategies vary systematically with experimental conditions and whether they are accompanied by a decrease in recall. In general, although children are guided by an internal narrative schema, they also rely heavily on linguistic cues which follow rules governing information status, and these two aspects of discourse organization interact during the course of development. [1]

He went and he's seen a leaf in the water. And he went and he jumped in and tried to get the leaf in the water. And he jumped so hard it sank. And then one more day he was swimming in the water. And he found another leaf. (fourth-grader)


293

3. NARRATIVE DEVELOPMENT: COMPARATIVE PERSPECTIVES A further problem to be faced by any attempt to relate coherence and cohesion pertains to the variability in children's performance as a function of a wide range of different variables. I focus below on two types of variation: crosslinguistic variation, showing that language-specific factors may influence the rate and course of narrative development; and contextual variation, showing that children's narrative performance is variable across situations and tasks. After a summary of some relevant results illustrating both types of variation, I turn to their implications for our accounts of narrative development. 3.1 Universals and cross-linguistic variability A first type of variation has been observed in an increasing number of crosslinguistic studies of narrative development, showing two complementary results: general developmental patterns, which are observed in all languages and are presumed to be determined by universal cognitive determinants; and particular variable patterns, which are observed in both adults' and children's performance from language to language (or across language families) and are presumed to be determined by language-specific factors. I will briefly illustrate such general and specific patterns in narrative development, with particular attention to the development of narrative cohesion, where the impact of language-specific factors has been best demonstrated. Although cross-linguistic variation in the development of narrative coherence has been reported anecdotally, less is known about the factors that might determine this type of variability in children's narrative performance. An example of cross-linguistic variation can be found in how children mark information status when denoting story characters. Although markings of information status are available in all languages, they vary a great deal across languages. These devices might be local (e.g., referring expressions) or more global (utterance structure), and they might be optional or obligatory within and across languages. For example, the indefinite/definite opposition is obligatory in English to differentiate new from given information, but it is optional in Chinese, which requires new information to be placed in post-verbal position. Furthermore, the complexity of these devices for the child might vary depending on other functions they might serve for the organization of the sentence and/or of discourse. In this respect, some types of multi-functionality are more complex than others. For example, indefinite nominal determiners can mark non-specific reference in English, presenting some degree of functional complexity. However, word order is central for the marking of grammatical relations, especially in languages that do not provide much morphology, resulting in the need for

294

Maya Hickmann

a cognitively more complex type of functional differentiation that operates simultaneously at the sentence and discourse levels. Studies comparing referent introductions and reference maintenance across several languages (see Note 1) show a general progression in children's ability to differentiate new and given information from about six or seven years on in all languages, suggesting that some universal determinant linked to cognitive maturity might partially underlie this type of development (Hickmann 1995: Hickmann & Hendriks 1999; Hickmann et al. 1996). Young children also prefer local markings (types of NPs) to global markings (utterance structure), regardless of whether they are acquiring languages where local markings are obligatory (e.g., French, English, and German) or languages where they are merely optional (e.g., Chinese). This finding shows that children do not always use the obligatory devices in their language when organizing discourse from the earliest phases onwards, contrary to previous results based on the frog story in other domains (Berman & Slobin 1994, see also below). In particular, Chinese children at first avoid global markings, which are functionally more complex than local markings because they operate at two levels of linguistic organization (discourse status as well as grammatical relations at the clause level). Finally, some general pragmatic principles (discourse topicality, types of cross-clause co-reference) guide children's uses of reference-maintaining devices in all languages, regardless of important structural differences between them (e.g., parametric variations). However, in addition to these general patterns, a large number of crosslinguistic variations occur. For example, narratives show wide differences in the use of clause structure across languages. Children make frequent use of structural variation specifically for the pragmatic purpose of marking information status in French (from four years on) and in Chinese (after seven years). 5 In contrast, children's structural variations depend both on discourse factors and on grammatical constraints in German (e.g., obligatory subject-verb inversion with frequent utterance-initial connectors such as da 'there' or dann "then'), and they are comparatively quite rare in English, where children reserve word order for the expression of grammatical relations. Other databases (e.g.. Kail & Sanchez y Lopez 1998) show related findings in Spanish and French frog stories. In particular, as in French, one finds a heavy reliance on structural variation for referent introductions in Spanish frog stories, but these variations take a different form (e.g., subject-verb inversions, rather than existentials or dislocations). Another illustration of general and language-specific patterns can be found in the area of spatial reference (Hendriks 1993; Hickmann 2003a: Hickmann. Hendriks & Roland 1998). Analyses of our cross-linguistic corpora (see Note 1) show strikingly similar developmental progressions in children's ability to provide an initial spatial anchoring in their narratives, thereby allowing their addressee to interpret locations and changes of locations in subsequent discourse. In all languages considered, and despite important typological differences that are


295

otherwise observed between them (see below), this ability follows a very gradual development, which does not give rise to adequate spatial anchoring before ten years of age and which continues to evolve until the adult age. This finding is in line with studies based on other tasks, such as room descriptions or route directions (e.g., Ehrich 1982; Weissenborn 1986). Such a developmental curve across tasks and languages seems to result from a general initial inability to plan discourse. Findings further show that spatial anchoring develops even later than the ability to anchor reference to the story characters. This delay can presumably be accounted for in two ways. First, in comparison with the story characters, the entities that serve to locate the story characters are not in the foreground of discourse, except perhaps temporarily, when they are being introduced for the purpose of subsequent spatial anchoring. As a result, young children focus more on the animate characters, ignoring the spatial parameters until some later point in the story where they become relevant to locate (or to re-locate) the story characters. Second, the spatial domain involves much inferencing and reliance on general-world knowledge, which results in somewhat different processes for marking information status (e.g., the utterance I fell down while picking apples relies on the assumption that apples are usually picked up in trees). However, the typological properties of languages also determine, at least partially, the particular aspects of denoted situations on which speakers focus and the ways in which they organize this information. For example, Hendriks (1993) reports that, despite a similar general developmental progression in how children learn to establish spatial anchoring in Dutch and in Chinese stories (see Note 1), Chinese children specify initial spatial parameters in more elaborate ways than Dutch children, while they provide fewer details thereafter, relying on presuppositions established in discourse. This difference can be attributed to the fact that these two languages differ with respect to topic-orientation versus subject-orientation, resulting in different strategies when organizing discourse. Furthermore, on the basis of both naturalistic data and narrative productions, several authors show that, depending on their language, adults and children from two years on emphasize different types of information pertaining to motion (Berman & Slobin 1994; Choi & Bowerman 1991; Hickmann 2003a; Hickmann, Hendriks & Roland 1998). Consider, for example, how different languages express displacements involving changes of locations. Following Talmy's typology (1983, 1985, 2000a, 2000b), satellite-framed languages (e.g., English and German) compactly encode the manner of motion in the verb root and give path information in various satellites such as particles or prepositions. In contrast, verb-framed languages (e.g., Spanish and French) encode the trajectory in the verb root, treating the manner as structurally secondary (e.g., she ran across the street versus elle a traverse la rue en courant 'she crossed the street running'). Correspondingly, speakers of satellite-framed languages focus more on detailed information concerning motion, whereas speakers of verb-framed languages

296

Maya Hickmann

focus more on detailed information concerning the settings in which motion takes place. For example, Berman and Slobin (1994) show contrasting patterns in how speakers (adults and children) narrate frog stories in verb-framed languages (Spanish, Turkish, and Hebrew) and satellite-framed languages (English and German). Analyses of the English corpus of frog stories show that children use a great variety of verbs that conflate (among other things) motion and manner. English-speakers also specify the source and goal of motion, even though such spatial specification increases with age. Thus, young children tend to describe the pictures, rather than narrate them, frequently mentioning the end states of motion events. In contrast, older children and adults can simultaneously provide multiple types of information about motion events, including manner, causation, and locative elaboration concerning both the source and the goal of motion. In comparison, Berman and Slobin (1994) and Sebastian and Slobin (1994) note the following properties of the narratives collected in Spanish. First, Spanish-speakers rarely provide information about the manner of motion. Second, they provide a great deal of stage-setting, while English- and German-speakers leave the setting to be inferred. Third, they rarely mention both the source and the goal of motion at any age. In this respect, the Spanish narratives show a three-phase developmental progression with respect to the details of paths. In a first phase, illustrated in [2] and [3], about half of the pre-schoolers describe changes of location by means of the verbs alone, sometimes adding a prepositional phrase specifying the source of motion. The other pre-schoolers represent a second phase, illustrated in [4]. in which they use locative adverbials together with the verb in a redundant way. Finally, the narratives produced from nine years on represent a third phase, illustrated in [5], which is characterized by increased specification of the source and goal of motion and, more generally, by increased locative description. Other results reported in Berman and Slobin (1994) also show similar patterns with respect to the expression of motion in Turkish and in Hebrew, as well as related developmental progressions with respect to ground specification. [2]

Se ha subido. '(He) has ascended.'

[3]

El nino se sube al arbol. 'The boy ascends the tree/

[4]

Sube por arriba por el tronco. '(He) ascends upwards along the trunk.'

[5]

Los tiro a un precipicio donde habia harta agua. Entonces se cayeron. '(He) threw them at a cliff where there was lots of water. Then (they) fell."


297

Related observations were also made in our cross-linguistic comparisons (see Note 1) examining how adults and children narrate stories in French (verb-framed language) versus English, German, and Chinese (satellite-framed languages) (Hickmann 1995, 2003a; Hickmann, Hendriks & Roland 1998). The narratives differ with respect to three inter-related properties: the range of predicates used to narrate motion events (particularly changes of locations); how compactly the relevant information is encoded; and narrators' relative focus on different types of situations. The narrators in English, German, and Chinese show great concern for providing dynamic information, as reflected by the considerable diversity in their uses of dynamic predicates. Most predicates represent several types of information, which is compactly encoded within utterances by various combinations of verbs and satellites. In contrast, the French corpus contains relatively few types of dynamic predicates. These predicates rarely contain multiple types of information concerning motion, and narrators distribute this information across clauses. French narrators also use more static predicates to provide locations, compared with narrators in the other languages. This pattern can be observed at all ages, as is the case for the frog stories analysed by Berman and Slobin (1994). However, in French, it is quantitatively much more marked among the youngest narrators than among the older children and the adults, partly because of their frequent use of predicating constructions to introduce referents in discourse. Furthermore, some types of information present difficulties to young narrators. For example, consider the two story excerpts in [6] and [7] (five-yearolds), which describe a picture in which a cat is hanging from the branch of a tree, trying to get to a bird's nest, while a dog is pulling on its tail (the preceding picture shows the cat climbing up the tree and the next one shows the dog chasing the cat away). In the English excerpt, the child compactly combines causativity, manner, and direction within the same utterance. In comparison, the information is distributed across clauses in the French excerpt and the causal relation between the dog's action and the cat's downward motion is not expressed (the intransitive verb descendre 'to go down' represents voluntary motion, while caused motion would require a complex construction such as faire descendre 'to make come down'). [6]

The dog was pulling the cat down by his tail. (5 yrs)

[7]

Apres le chien lui mord la queue. Apres le chat descend de 1'arbre et le chien 1'attrape. 'Then the dog bites his [the cat's] tail. Then the cat goes down from the tree and the dog catches him.' (5 yrs)

298

Maya Hickmann

Such cross-linguistic differences can also be observed in a large corpus of data I collected as part of a study (Hickmann 2003b) which aims at systematically examining the expression of voluntary-motion and involuntary-motion events in English and in French. For example, with respect to voluntary motion, part of this study compares how adults and three- to six-year-old children narrate short animated cartoons showing various characters performing displacements, such as arrivals and departures, vertical motion (e.g., running up and down a tree), or displacements involving a boundary-crossing (e.g., crossing a river). English-speaking adults and children (from three years on) package both path and manner information compactly within the clause with all types of displacements, as illustrated in [8] (to crawl across, up, back down, away). In French, adults often provide both types of information, particularly with crossing events (e.g.. traverser en nageant 'to cross swimming' in [9]), but they do so less often than the English-speaking adults, particularly with vertical motion and with arrivals and exits (e.g., avancer 'to move forward', monter 'to go up', redescendre 'to come back down', repartir 'to leave' in [10]). Young French children, for their part, rarely express both types of information with any displacements and. when they do so at six years of age, they distribute the information across clauses, rather than compactly packaging it within the clause, as shown in [11 ]. [8]

The caterpillar crawled across the ground to get to the plant, crawled up the plant, took a bite off the leaf, crawled back down and crawled away, (adult)

[9]

II traverse la riviere en nageant. 'He crosses the river swimming.' (adult)

[10] C'est un ecureuil qui se trouve dans la foret, qui avance, qui monte dans un arbre, passe dans un trou, qui ressort du trou, qui redescend de 1'arbre et qui repart. 'It's a squirrel that's in the forest, that moves forward, that goes up in a tree, passes into a hole, that comes back out of the hole, that goes back down from the tree, and that goes away.' (adult) [11] C'est une dame qui a fait du velo. [...] Elle est passee sur les rails. 'It's a lady who rode a bike. [...] She passed onto the railway tracks.' (6 yrs)

Finally, cross-linguistic differences can also be observed in children's uses of temporal-aspectual markings. For example, Berman and Slobin (1994) show that the language-specific properties of temporal-aspectual systems affect how children from three years on use morphological markings to express events in frog stories. Depending on the relative complexity of these morphological distinctions within a given language, children display a greater focus on the temporal


299

contour of events in some groups (e.g., Spanish) than in others (e.g., German). These results suggest that from early on, children rely on the obligatory devices available in their language to organize information in discourse. Furthermore, our cross-linguistic comparisons based on other stories and languages (see Note 1) show some association between resultative events and past and/or perfective markings in all languages, as predicted by the universal defectivetense hypothesis. However, the strength of this association is quite variable across ages and languages: it is strongest in Chinese (at all ages), as well as in English (but only among the younger children), and it is quite weak among French- and German-speakers (at all ages). These differences can be partially attributed to the relative complexity and transparency of different temporal-aspectual systems, which simultaneously result in different types of temporal anchoring in the IndoEuropean languages (frequent anchoring in the past in English, but in the nonpast in French and German) and in a greater reliance on connectives in Chinese. Despite these differences, children's uses of temporal-aspectual devices follow a similar developmental progression towards increasing cohesion. 3.2 Context, genre, and register Other types of variations, which may be grouped under the general heading of contextual variability, concern the situations and tasks in which narrative performance is elicited. This type of variability is much more heterogeneous, since it depends on a great number of different dimensions, all of which affect the ways in which children go about producing or processing narratives. These dimensions include at least the following: the length, complexity, and structure of the stories to be narrated; whether narratives are spontaneously produced or elicited in different types of situations; the mode in which they are presented in experimental situations; the implicit or explicit discourse genre, register, or language modality; and the nature of the interlocutors' knowledge in the narrative situation. All of these variables can be major determinants of language use. As discussed in more detail elsewhere (e.g., Berman, this volume; Hickmann 1991, 1995, 2003a), the divergence in the results reported across studies of narrative development is partly linked to this methodological heterogeneity, something that often makes generalizations quite difficult. In order to illustrate this point briefly, 1 will merely mention the impact of a few variables pertaining to the nature of the communicative situation on children's narrative performance. First, a number of studies (e.g., Bamberg 1997a; Bronckart & Schneuwly 1991; de Week 1991; de Week & Schneuwly 1994; French & Nelson 1985; Preece 1987; Pellegrini 1984) show variations in children's uses of devices such as referring expressions, temporal-aspectual morphology, and connectives across different types of discourse situations. As

300

Maya Hickmann

discussed above, children display early uses of these devices in scripted narratives that are embedded in conversational settings, as compared with other types of narratives. Furthermore, children are sensitive to discourse types, varying their uses of these devices as a function of various other factors, e.g., across other discourse types (narratives of personal experience versus third-person narratives, stories such as fairy tales versus other types of third-person narratives, narrative versus expository discourse), as well as across different language modalities (written versus oral discourse) and/or across different situations involving different registers (more or less formal and/or planned uses of language). Second, some evidence (Hickmann, Kail & Roland 1995; Kail & Hickmann 1992; Kail & Sanchez y Lopez 1998) also shows the strong impact of mutual knowledge on the cohesion of children's narratives. These studies compare how French and Spanish subjects (adults and children between six and eleven years old) narrate the frog story in two situations characterized, respectively, by mutual knowledge (telling the story while leafing the book with the interlocutor) and by the absence of mutual knowledge (telling the story to a naive blindfolded listener). With respect to referent introductions, children of all ages use more indefinite forms in the absence of mutual knowledge than when knowledge is shared. However, nine-year-olds display a maximal differentiation in this respect, whereas younger children nonetheless frequently use definite forms in both situations. Eleven-year-olds, on the other hand, generalize the use of indefinite forms for referent introductions even to the situation where know ledge is shared, thereby marking the discourse as being of the "narrative" type. With respect to reference maintenance, children in all age groups mark cohesion and coherence in both situations. For example, they are sensitive to principles of referential continuity, using more pronouns when there is immediate co-reference across adjacent clauses. In addition, they are sensitive to story structure, over-marking their references to story characters by means of nominals (despite the existence of a tight co-referential chain) when they cross an episodic boundary (thereby marking the internal story structure) and when they cross a picture boundary (thereby marking the external story structure). However, this over-marking disappears at eleven years of age in a situation where mutual knowledge is absent. In this situation only, the oldest children systematically follow principles of discourse cohesion, which becomes their predominant mode of narrative organization. Third, children's prior knowledge of the stories may affect their narrative productions. For example, divergent results are reported by Karmiloff-Smith (1981, 1985) and by Bamberg (1987) concerning the age at which children display the thematic-subject strategy. On the basis of different types of simple stories, elicited with short picture sequences, Karmiloff-Smith finds this phase to be a rather late development. In contrast, on the basis of the much more complex


301

frog story, Bamberg reports such a strategy rather early (three or four years of age), before a later type of strategy whereby children pay attention to the presence or absence of immediate co-reference across clauses. In this respect, it may be no coincidence that the children were quite familiar with the frog story in Bamberg's study (where they had already seen the picture book, as well as heard and told the story before), whereas they were not familiar with the stimuli in KarmiloffSmith's study (where they discovered the stories as they told them for the first time). It is possible, then, that the children's high level of prior knowledge about the story in Bamberg's study may have influenced the ease with which they were able to rely on a top-down type of organization when constructing their representation of the story. To complicate matters further, a number of contextual variations do not take the same form from language to language. Thus, although all languages provide some linguistic means of differentiating situations and/or registers, some languages present a more striking range of devices to mark such distinctions than do others. French presents a strong differentiation of this type, which is reflected by a great number of lexical, morphological, and syntactic indices (Berman in this volume makes a similar point about Hebrew). In this respect, French children's uses of clause-structure variations for the marking of information status (see N o t e l and Section 3.1 above) frequently involve two kinds of structures - existential-relative and dislocated structures - which typically characterize oral speech produced in informal and/or unplanned discourse, but not written, formal, or planned discourse. Existential-relative structures are used to introduce referents (e.g., il y a un chien qui arrive 'there's a dog who comes') or to "reintroduce" them (e.g., il y a la maman qui revient 'there's the mother who comes back'). Dislocations mark topic switches, although their uses evolve with age. Thus, left-dislocations are frequently used at four-five and seven years of age for all kinds of topic switches, including to mark referent introductions, for which they are inappropriate. The ten-year-olds reserve these structures for two specific purposes in reference maintenance: switching subjects (e.g., et elle s 'en va et le chat il regarde 'and she goes away and the cat he's watching') and promoting topics after referent introductions (// y a le chat qui arrive, le chat il regarde les oiseaux et il monte 'there's the cat who comes, the cat he looks at the birds and he goes up'). The corpora further show a drastic decrease of these structures in the adults' narratives. In particular, French adults practically never produce dislocations in this particular discourse situation, but rather use other means of organizing reference maintenance, suggesting that they are relying on a different discourse mode. Other indices are compatible with this hypothesis, such as adults' uses of particular verbal inflections, of different types of subordination, or of zero anaphora, all of which are most typical of written, formal, and/or planned speech. Thus, the greater discourse planning observed with increasing age not only results

302

Maya Hickmann

in increased cohesion (by ten years of age) - it may also be accompanied by a tendency to rely on another register (adults), at least in some languages. 3.3 Implications of variability for the study of coherence and cohesion All of this evidence, then, shows the effects of language-specific and contextual factors on children's narrative performance. Contextual variations indicate that narrative performance is sensitive to a number of dimensions defining narrative situations, such as discourse type, register, and prior or mutual knowledge of the story, none of which can be ignored in our assessments of children's narrative competence and in the explanations we provide for its development. The evidence concerning language variation shows that language-specific properties have an impact on two aspects of children's narratives: the particular content children primarily select when denoting events and the means they use for discourse cohesion when including these events in narratives. For example, the language to be acquired influences the degree to which children focus on the trajectory or the manner of motion, on location or motion, on states or processes, on the temporal contours of situations, and on some types of causal chains. It also influences the ways in which personal, temporal, and spatial information is distributed through a variety of linguistic means serving to express both prepositional content (who did what to whom, when, and where) and discourse relations (what's new or given, topical, in focus, overlapping versus successive, etc.). One implication of these cross-linguistic differences is that language presumably has a general impact on how coherence is constructed, since it affects the content that is selected and placed at the centre of the narrative. In addition, it is highly likely that such differences are simultaneously related to the processes whereby cohesion is constructed across languages. In particular, languagespecific properties result in a differential focus on what is treated as central versus secondary information, and therefore they result in a differential organization of the information in the background versus foreground of discourse. Furthermore, language affects the compactness of the information represented. Thus, in some languages spatial information of several types is tightly packaged within the clause, whereas in other languages the same information is distributed across clauses. In this case also, then, language-specific properties affect not only sentence-internal structure, but also discourse organization, determining how the flow of information is regulated across clauses. In this respect, the speakers of some languages (e.g., French and Spanish) are particularly concerned with the discourse structure of information, as illustrated above by their frequent use of structural variations as a means of marking information status (at least in a particular register). Thus, although no differences were found in children's basic ability to provide spatial anchors in discourse across languages, presumably


303

because this ability reflects the gradual development of discourse planning, wide variations occur in their focus on various types of information when relating dynamic events. It is worth noting that similar late developmental progressions concerning spatial anchoring can be observed across tasks (room descriptions, route directions, narration), across narrative stimuli (frog stories and other stories), across narrative situations (whether knowledge of the story is mutual or not), and across typologically different languages (verb-framed versus satellite-framed languages, topic-oriented versus subject-oriented languages). Thus, with respect to narrative productions, the available findings show some gradual developmental progressions in Spanish, Turkish, and Hebrew frog stories (Berman & Slobin 1994), as well as in other English, French, German, Dutch, and Chinese stories (Hendriks 1993; Hickmann 2003a; Hickmann, Hendriks & Roland 1998). In fact, the strikingly similar curves that we observed in this respect across typologically different languages were not expected, given our original hypothesis that spatial anchoring may be more frequent and more precocious in verb-framed languages than in satellite-framed languages. This general development, then, suggests that, regardless of their language, children strive with some difficulty to establish spatial anchoring in various cohesive ways. However, it remains to be seen whether more language-specific types of developmental phases can be observed across languages in this respect. Following a similar interest in discourse skills, Aksu-Ko? (1994) suggests in her analysis of Turkish frog stories that the uses and non-uses of various devices to elaborate on locative trajectories are motivated by discourse factors. Comparing the narratives produced at different ages, she notes two reasons for the lack of locative modification to elaborate on locative trajectories. The first reason, she argues, is related to the fact that young children are not yet concerned with establishing discourse-internal relations. Second, young children "tend to describe what is most salient as new information in the pictures, which is typically the movement itself, while the source is given and the goal needs to be inferred" (Aksu-Ko*,: 1994:352). Thus, in contrast to young children, older children and adults view the pictures as related in a sequence, rather than as discrete situations, and they create discourse-internal relations, referring back to previous context. The question here is whether such developmental progressions, expected on the basis of universal general cognitive development, follow the same course in all languages or whether variations can be observed in this respect across languages. As we saw, despite wide cross-linguistic variation in how narrators (adults or children) represent spatial information in their narratives, the evidence shows the existence of universal cognitive factors determining the extent to which young narrators are able to plan discourse in such a way as to provide locative information serving as spatial anchors in discourse.

304

Maya Hickmann

A further point concerns possible interactions between contextual and language factors. As was only briefly illustrated above, the context might have different effects depending on the language, for example as a function of available register markings, although this question requires more thorough empirical study. In addition, note that particular effects of language-specific properties may themselves depend on the contextual conditions in which narratives are produced. For example, the infrequent expression of manner in Spanish- and French-speakers" narration of location changes was related to the fact that the manner of motion is structurally secondary and peripherally encoded in verb-framed languages, whereas the trajectory of motion is specified in the verb root. However, imagine a situation where contrastive uses of manner information would be crucially relevant to discourse, for example if the same motion event is performed by the same or different characters in different manners, such as Example [12] below. In this case, speakers of verb-framed languages might be induced to provide manner information that would be omitted in more neutral discourse contexts. Some of our research in progress indeed suggests that, notwithstanding French-speakers' infrequent expression of manner information with some types of displacements in some situations, this information can be systematically elicited from them, provided that the situations focus their attention on this aspect of motion. Crosslinguistic comparisons, then, must also take into account the impact of contextual variations both within and across languages. [12] Pierre a traverse en marchant, Marie en courant et Paul a quatre pattes. 'Pierre crossed by walking. Marie by running, and Paul on all fours.' Finally, research comparing different types of narratives (e.g., Bamberg 1997a) has shown large differences between third-person and first-person narratives. Given these findings, it is likely that the relation between coherence and cohesion will be most evident in a variety of third-person narrative types, where structural and cohesive aspects of narrative discourse are both maximally required. In comparison, first-person narratives, such as narratives of personal experience embedded in conversational settings, involve a great variety of other factors linked to the nature, goals, and unfolding of the interlocutors' interaction, including socio-cultural factors determining inter-personal relations and additional psychological factors linked to personality and identity. In these cases, coherence and cohesion are less central to the unfolding of the narrative, the organization of which follows other rules and depends on other constraints.


305

4. CONCLUDING REMARKS In summary, one of my aims was to suggest the need for serious examination of the relation between coherence and cohesion in the development of children's narrative competence. I argued that narrative competence involves structural aspects of narrative organization, related to the construction of a story schema, as well as procedural aspects, necessary for the construction of cohesive discourse. Therefore, accounts focusing on only one of these aspects of discourse organization provide an incomplete picture of children's abilities which is insufficient for us to be able to understand the nature of their narrative competence. Furthermore, among the different available approaches to coherence and cohesion, those that view them as merely parallel miss the complex interplay between structural and procedural aspects of discourse organization during development. For example, I pointed out the need to consider the very nature of some narrative units from a discourse-functional point of view and the effects of some anomalies concerning these units not only on coherence, but also on cohesion, both of which might affect children's recall and processing of narratives in different situations. More generally, the different approaches discussed above contrast with respect to whether they view discourse-organizational principles as structuring the very process of constructing story schemata. I suggested that discourse organization imposes its own constraints on these constructive processes, raising some questions about the nature of abstract story structures that would be entirely independent of language use. Surprisingly little empirical research has directly addressed this general theoretical question. In addition, the brief review above pointed out the existence of various types of variations in children's narrative performance as a function of languages and contexts. Cross-linguistic analyses of children's narratives show the impact of general factors, affecting discourse development across languages, as well as the impact of language-specific factors, affecting how children go about selecting and organizing information in discourse. General recurrent patterns show the development of a general ability to anchor reference within discourse, as well as the impact of the relative functional complexity of different markings. However, variable patterns across languages also show differences in how children and adults go about selecting and regulating information across utterances. These results highlight the impact of language-specific constraints on the cognitive processes that underlie discourse organization. The resulting claim has been that children engage in the activity of thinking-for-speaking (Slobin 1996b) as they learn to communicate in the mould provided by their particular language, which partially determines how information is filtered and channelled in discourse. Finally, I pointed out how language-specific factors may influence not only the content that is selected for expression in narratives, but also the organization

306

Maya Hickmann

of this content in the foreground or background of narrative discourse. Given the role of functional factors in the uses of linguistic devices for discourse cohesion, it is therefore likely that at least some aspects of the structural differences that can be observed across languages might be highly sensitive to contextual constraints. Special care is therefore necessary to take all of these variations into account when designing further research on the cohesion and coherence of children's narratives. More generally, future accounts of children's narrative competence must bear in mind the fact that language use implies two types of constraints on discourse organization: general constraints resulting from the universal properties of language as a semiotic medium which imposes its own organizational principles on children's cognitive processes when they communicate in particular contexts, and some variable constraints resulting from the particular systemic properties of the languages with which children are confronted during development.

NOTES 1.

These findings and others discussed throughout this chapter stem from large corpora of narrative productions elicited from adults and children (four to ten years old) in different languages, including English, German, French, Dutch, and Mandarin Chinese (Hendriks 1993; Hickmann 2003a). Two stories were elicited ("cat stories" and "horse stories") by means of short picture sequences without text, which contrasted with respect to the status of the animate referents (only one sequence had a clear main protagonist). Data collection involved controlling background knowledge conditions in such a way as to maximize discourse-internal cohesion: the subjects were asked to tell the story to a naive addressee who did not have access to the pictures and who would have to tell the story back.

2.

Structural anomalies have variable effects on children's story recall and on their narrative strategies, depending on the particular units that are deleted or displaced and, in the case of displacements, on their position (see Hickmann 2003a).

3.

The use of structural variations in French and Chinese has been related to the fact that both languages are topic-oriented (Lambrecht 1981).

Interplay between Narrativity and Mindreading A Comparison between Turkish and English

Ayhan Aksu-Koc. & Goklem Tekdemir

1. INTRODUCTION Narratives typically involve a temporally-causally ordered sequence of events directed towards a goal and experienced by one or more protagonists. With this particular structure, narratives simultaneously effect two functions, identified as referential and evaluative by Labov and Waletzky (1967). The referential function is realized by a sequence of clauses that describe the events and actions constituting the story, and the evaluative function by clauses that refer to the mental states of protagonists, motivating circumstances, or consequences of those events. Bruner, with a more general focus on narrative thinking, has similarly underlined the dual levels of organization that inhere in narratives: a "landscape of action" constituted by a sequence of events and a "landscape of consciousness" constituted by the thoughts, beliefs, and emotions of protagonists (Bruner 1986:18). In short, the referential function corresponds to the plane of action and the evaluative function is particular to that of consciousness. In telling their stories, mature narrators move between the two planes and may adopt different perspectives in realizing these functions: that of the author and those of one or more of the protagonists. Such perspective shifts rest on the capacity to hold a theory of mind, which entails understanding the mental states of others and establishing causal links between those belief states and the world. Research in the area of theory of mind shows that children younger than four years old can attribute perceptual states, desires, and emotions to others (Wellman, Phillips & Rodriguez 2000) but have difficulties in distinguishing between appearance and reality and in understanding that people can have beliefs different from their own, even false ones. Put differently, they fail to understand that one and the same object may simultaneously be represented in two different ways; and in making inferences about beliefs and predictions about behaviour,

308

Ayhan Aksu-Ko? & Goklem Tekdemir

they take their own knowledge about reality as pivotal and tend to misjudge the mental representations of others that result in false beliefs (Astington & Gopnik 1991; Astington & Jenkins 1999; Flavell 1999; Flavell et al. 1987; Nelson 1996: Perner 1993). In short, they have difficulties in representing such representational states. Studies of children's developing narrative competence, on the other hand, show that stories that fit the above narrative definition emerge at around the same time, somewhere around the age of four or five. Narratives where the interplay between the two landscapes is coherently displayed develop gradually as children take into account the intentional and emotional states of the protagonists and relate them temporally-causally to the course of events (Bamberg & DamradFrye 1991; Kiintay & Nakamura, this volume). Furthermore, the use of evaluative versus referential statements by manipulating different perspectives appears to show cross-cultural (Kiintay & Nakamura, this volume) as well as intra-cultural (Aksu-Ko? 1996) differences. In this chapter, we look at how narrators identify a false belief resulting from a discrepancy between the apparent and the real identity of an object, a problem that belongs to the landscape of consciousness, and causally link it to ensuing events, a problem that belongs to the plane of action. For this purpose, we focus on the scene in the storybook Frog, where are you? (Mayer 1969) where the boy mistakes the antlers of a deer for the branches of a tree and finds himself accidentally situated on top of the deer's head between its antlers (Pictures 14-16. Appendix I). We analyse the stories produced by narrators of different ages in Turkish and English - two typologically different languages. These two languages differ in terms of the means for expressing changes in perspective: Turkish marks voice shifts inflectionally whereas English uses analytical constructions. Both languages use periphrastic means such as modals, mental verbs, and adverbs to talk about appearance versus reality, belief states, and representational change. Turkish, in addition, has a specific verb for encoding misrepresentation (san 'think that') and an obligatory evidential particle (-ml$) that marks the asserted information as "new" or "unexpected" to the speaker's mind, thereby signalling representational change (Slobin & Aksu 1982)'. It is, therefore, reasonable to expect Turkish narrators to be more sensitive in attending to the mental states of actors as a basis for their actions, and to insert the boy's misrepresentation of reality into the chain of events as a "causal link" at an earlier age than their English counterparts. In the following sections we first explain the coding scheme used in the analysis. Then we present the results, tracing developments in mindreading and narrativity in both languages. Finally, we discuss the similarities and differences between the two groups in mindreading and their preferred linguistic strategies, considering the implications of these for inter-relations between linguistic and cognitive components of narrative thinking.

Interplay between Narrativity and Mindreading

309

2. METHOD 2.1 Sample The Turkish data come from the A. Aksu-Ko9 (1994) corpus and the A. Kiintay corpus (1997). There are six age groups (ages 3;0, 4;0, 5;0, 7;0, 9;0, and adults) with fifteen subjects per group. The English data come from the T. Renner (1988) corpus and consist of five age groups (ages 3;0, 4;0, 5;0, 9;0, and adults) with twelve subjects per group, except for the adult group with only ten subjects. 2.2 Procedure M. Mayer's picture storybook Frog, where are you? was used to elicit the narratives (see Appendix I). The subjects were studied individually. Each subject was asked first to look at the pictures of the book from the first to the last page, and then to tell the story from the beginning, following the pictures. The narratives were audio-recorded and transcribed. 2.3 Scheme for analysis Each narrative was evaluated and coded in terms of the following three parameters: /.

Plot structure, reflecting the narrator's ability as regards narrative organization at the global level. In line with the previous analyses in Berman and Slobin (1994:39-84), reference to three plot components was taken as criterial. These were: a)

Onset: the boy's realization that his frog has disappeared;

b)

Unfolding: the boy's search for his missing frog; and

c)

Resolution: the boy's finding his lost frog or another frog (Berman & Slobin 1994:45-46).2

Reference to all three components was taken to indicate the presence of an overall hierarchical theme to which the plot-constitutive events were related. //.

Misrepresentation ("MR" from here onwards), reflecting the narrator's ability to identify the protagonist's false belief arising from a mismatch

310

Ayhan Aksu-Ko
kiyor yine baginyor, boyle bir hayvanm boynuzlarmin, iki boynuzunun arasmda kahyor. Hayvan $ok sinirleniyor, sonra onlar bir yerden yere dusiiyorlar, gole dusiiyorlar. '[H]e climbed onto that and he is calling again, he gets stuck between an animal's horns, two horns. The animal gets very mad, then they fall off somewhere into a lake.' (t7;10i, Kiintay corpus)

The boy was rendered intentional by 27% of the subjects, using the BOY-AGENT strategy (subject of a transitive verb such as get on, ride}. Thus, the majority of the seven-year-olds perform like the nine-year-olds in effectively relating events on the plane of action. Five-year-olds The five-year-olds did not refer to the MR but focused on the encounter with the deer. The dominant strategy of the English-speaking five-year-olds was to present the boy as non-intentional, thereby ensuring a coherent flow of events on the plane of action; the dominant strategy of the Turkish-speaking group, on the other hand, was to portray the boy as an intentional agent. This response type entails assigning an agentive role to the boy and presenting the encounter with the deer

318

Ayhan Aksu-Koc & Goklem Tekdemir

as an event under his control, with no attention to what happens at the level of consciousness. Talking about the boy as agent, with the boy in the subject position of a transitive verb as in Example [10] below, is a less mature strategy observed only in the children's data. [ 10] [BJurada da £ikip tasm uzerine baginyor. Geyigin iistiine 9ikiyor '[A]nd here he climbs on the rock and calls. He climbed onto the deer' (t5; 11 f, Aksu-Koc corpus) In the English-speaking group, only one subject portrayed the boy as intentional, and two subjects hinted at the MR. However, both were unsuccessful; one equated appearance and reality but failed to integrate this into the causal chain of events (Example [11]), the other negated the validity of the mental state of the boy but was unsuccessful at linking it to reality - the deer (Example [12]). [11] ... and they start calling. And they think that sticks, but it's a d - um - an antler. And then they run, and then they - go in the water. (e5;6d, Renner corpus) [12]...e's gonna climb the rock, and then - he calls - up on the rock - for the frog, and then he - thought it was sticks and - he got on that and - the deer came and - carried him. (e5;llk, Renner corpus) Thus, together with those who rendered the boy non-intentional, 75% of the American subjects displayed an awareness of the accidental nature of the encounter. None of the Turkish five-year-olds hinted at the MR, except for one subject who alluded to it by attributing "lack of knowledge" instead of "false knowledge" to the boy, as in [13]. [13] [B]i hayvamn biseyini go'rdiiler. ne oldugunu bilmiyorlardi. Hayvan cikti ortaya, 'kurbaga nerede'dedi, sonra dedi ki 'neyse ben bu kayanin iistiine biniyim'. Sonra. sonra hayvamn iistiine bindi... '[T]hey saw an animal's something, they did not know what it was. The animal popped up, and said "where is the frog?" then it said, "anyway. I will mount onto that rock". Then, then he mounted onto the animal...' (t5;4a. Aksu-Koc corpus) Instead, 40% presented the boy as non-intentional by using the BOY-ACTOR or BOY-PATIENT strategies, and 47% presented him as intentional using the BOY-AGENT


319

strategy. In Example [14], the verb takil 'get caught' implies the accidental nature of the boy's move to the deer's head, but because he cannot equate the branches with the antlers, the narrator ends up presenting the boy as intentionally mounting the deer, using the verb bin 'mount'. [14] ... simdi £ocuk kann iistune cikmisti, biryere bakiyordu. Sonra pocuk bir dala takildi orda kaldi, dala takildi kaldi. Ondan sonra bir hayvanm ustiine 9ikmisti, bir hayvanm iistune cikmisti. ... ".. .now the boy had climbed on the snow, looking somewhere. Then the boy got stuck between the branches got stuck there, got stuck between the branches got stuck there. And then he climbed up on an animal, climbed up on an animal ...' (t5;2a, Kuntay corpus)

Whatever the strategies they use may be, as the examples above illustrate, the responses of the five-year-olds differed from those of the older subjects in being constrained to the plane of action. The events narrated were sequenced in accordance with the pictures, rather than related to each other for the purposes of accounting for the boy's fall onto the deer's head. Four-year-olds Not surprisingly, the responses of the four-year-olds were less sophisticated than those of the five-year-olds. A few subjects alluded to the unexpected nature of the encounter by referring either to some aspect of narrative reality (Example [15]) or to their own mistaken belief (Example [16]), but they could not recognize the misrepresentation on the part of the boy: [ 15] [R]ock's there - there's a deer hiding up there (no more deers there) he hops on the deer and he's gonna throw the deer down there. (e4f, Renner corpus)

[ 16] [K]eci var burda. O kadar agirmis ki ben bunun ipinde birsey yok zannettim, demek ki keyi varmis. '[TJhere is a goat here. It is so heavy that, I thought there was nothing in here but evidently there is a goat.' (t4j, Kuntay corpus)

The data did not reveal a dominant response strategy: in describing the encounter, the number of children portraying the boy as intentional by using the BOY-AGENT strategy (Example [15]) was about equal to that presenting him as non-intentional by using the BOY-ACTOR or BOY-PATIENT strategy5, as in Examples [17] and [18], respectively.

320


[ 17] [H]e went on a rock and the dog was not on a rock, he slipped on to a deer. and the deer took boy and the dog was behind the deer. (e4;4b, Renner corpus) [18] [SJonra cocuk bi tane dagin iistiine ^ikmis anyormu§ onu. sonra kopek. kopekle cocuk kafmaya baslarken ko'pegi geyik almis. atiyormu$ ikisini de - de, bu gole.. '[T]hen the boy climbed up on a mountain looking for it, then the dog. when the boy started running away with the dog, the deer took the dog, was throwing both of them to this lake...' (t4;7p, Kiintay corpus)

In short, responses that regard the boy as non-intentional, suggesting an awareness of the accidental nature of the encounter, began to appear at four years of age. but in a more random than systematic pattern. Three-year-olds The typical strategy of the three-year-olds was not to mention either the misrepresentation or the encounter. This is not surprising since understanding the discrepancy between appearance and reality in the present context requires the ability to "read" the connections between events in the different pictures, which three-year-olds did not do. While very few mentioned the branches, more than half referred to the deer - but only as a central figure in a new picture: they did not refer to the boy's encounter with it. Only one-third of the children in each group talked about it, but since their accounts were not temporally guided or presented from a stable-protagonist perspective, we could not identify a role that the boy was systematically assigned to; some presented him as the AGENT of transitive verbs such as bin 'ride' or climb, some as the ACTOR of intransitive verbs such as dti§ 'fall', and some as the PATIENT of transitive verbs such as take away with the deer in the subject position. A typical example is given in ([19]): [19] [A]nd the boy tried to climb up it, - he did and the dog is - going away. A deer trying -jumping over, and the dog is running away with deer, and the boy on top of the deer. (e3;5d. Renner corpus) The above survey of the data has shown that our hypothesis regarding the effect of language on mindreading is not supported. There is no significant difference between the Turkish-speaking and the English-speaking narrators in terms of their tendencies explicitly to refer to the boy's misrepresentation of reality and the causal role of this false belief in the course of subsequent events. However.


321

differences were observed in the strategies used by the two language groups in maintaining a coherent flow of events while constructing the story on the plane of action. In order to capture the nature of these differences, we looked at the use of perspective shifts at the local level of the "deer scene". 3.3 Narrativity at the local level: Discourse perspective As was observed above, when the misrepresentation - a state of affairs pertaining to the level of consciousness - is explicitly recognized, then the "why" and the "how" of the boy's location on the deer - an event pertaining to the level of action - is accounted for. In explicating the MR, the narrator introduces the deer into the discourse, and it can then be moved to the subject/topic position as old information, yielding the boy in PATIENT role. Alternatively, the boy can be maintained as the subject/topic from the previous utterance but still be presented in a non-agentive role as ACTOR, accidentally encountering the deer. When the MR is not explicitly mentioned, a meaningful account of the "how" of the subsequent events on the plane of action can still be given, again by shifting topic to the deer and using the BOY-PATIENT strategy, or by maintaining the boy as topic and presenting him as ACTOR. Both of these strategies reduce the boy to non-intentional status and capture the accidental flavour of events. A third option is to maintain the boy both as topic and in an agentive role, disregarding the accidental nature of the following events. Both the BOY-PATIENT and the BOY-ACTOR strategies call for a perspective shift at the conceptual level, from the boy as the controller of events to the boy as an affected entity, whereas the BOY-AGENT strategy does not. The BOY-PATIENT strategy, however, additionally calls for a topic shift from the boy to the deer as subject, at the discourse level. To see if these two types of shift showed any difference by age or by language, further analyses were carried out. Table 3 presents the percentage of subjects assigning the boy to the AGENT, ACTOR, or PATIENT roles, regardless of whether they talk about the MR or not, for the two language groups. A chi-square analysis of the data revealed a significant effect of age on whether the boy is assigned an AGENT, ACTOR, or PATIENT role (using logit, yj = 68.68, df = \2,p< 0.05) but no significant effect of language. This finding indicates that the two language groups did not differ in terms of perspective shifting at the conceptual level. It is observed that younger children prefer the BOY-AGENT strategy whereas older children and adults prefer the BOY-PATIENT or BOY-ACTOR strategies. This may be because the BOY-AGENT strategy represents the most canonical sentence structure, where all three roles of syntactic subject, semantic agent, and discourse topic coincide in a single protagonist whose activities can provide a consistent

322


Table 3. Percentage of subjects assigning the boy to AGENT, ACTOR, or PATIENT roles, by age and by language English

Turkish

Age

AGENT

ACTOR

PATIENT

AGENT

ACTOR

PATIENT

0

3;0

13%

13%

7%

17%

17%

4;0

27%

27%

13%

33%

25%

8%

5;0

53%

20%

13%

17%

17%

58%

9;0

20%

47%

27%

0

33%

67%

20;0

0

67%

27%

0

60%

40%

thread for plot development. The increase in the proportion of this response type between three and five years of age may therefore reflect a systematic attempt to construct a coherently organized story, and it may provide evidence for increased narrative competence in general. Put differently, younger children appear to be constrained to maintain the boy's perspective for taking the storyline forward despite the fact that this results in a deviation from the course of events depicted in the pictures of this particular scene. Although the language effect was not significant, it is observed that the BOY-AGENT strategy is a more prevalent characteristic of the Turkish children's narrative style than of that of the American children, who shift to the more mature strategies at around five years of age. A significant effect of language, however, was found when a chi-square analysis was conducted on the proportion of responses representing the DEERTOPIC versus BOY-TOPIC discourse strategies (using logit analysis. x: = 4.28. df = 1, p < 0.05). There was no significant effect of age. This finding indicates that Turkish- and English-speaking narrators prefer different discourse strategies in talking about this scene. Table 4 shows the distribution of subjects using different discourse strategies by age, for Turkish and English. As can be observed, speakers of Turkish use BOY-TOPIC, that is, topic maintenance, as a dominant discourse strategy, whereas English-speakers prefer the DEER-TOPIC option, displaying topic shifting as their strategy. Considered together, these findings indicate that the narrative performance of the two language groups differs only in terms of topic shifting, that is, discourse-level perspective shifting. In the present context, this means that Turkish narrators preferred to continue the story from the boy's perspective despite the decreased level of agency in his role during the course of events. In fact, older children and adults expressed this by using linguistic structures of reduced transitivity such as passive, reflexive, or stative constructions. In


323

Table 4. Percentage of Turkish- and English-speaking subjects using different discourse strategies, by age English

Turkish Age

BOY-TOPIC

DEER-TOPIC

BOY-TOPIC

DEER-TOPIC

3;0

27%

7%

33%

0

4;0

53%

13%

58%

8%

5;0

67%

20%

33%

50%

9;0

67%

27%

33%

67%

20;0

67%

27%

50%

40%

contrast, the English-speaking narrators, starting around the age of five, shifted topic to the new character, the deer, assigning it an active plot-advancing role. This difference in strategy observed at the discourse level may arise from cultural differences in narrative style between the two groups or from more language-specific factors. To investigate whether it is due to differences at the grammatical level between the two languages, a chi-square analysis (using logit analysis) was carried out on the proportion of responses displaying different types of grammatical structures (namely, the boy as the subject of a) transitive verbs, b) intransitive verbs, c) stative constructions, d} passive constructions, e) reflexive constructions, and/) the boy as the object of transitive verbs). The results revealed no significant effect of language or age. Tables 5 and 6 show the distribution of the different grammatical structures by age for the two language samples. Although the differences between language or age groups were not significant, a comparison between Tables 5 and 6 reveals some observable trends: both the Turkish- and the English-speaking three- and four-year-olds show a preference for constructions with the boy as the subject of transitive verbs. At five years of age, an even higher proportion of Turkish children use this construction type, whereas English-speaking children display a preference for constructions with the boy as the object of transitive verbs, a preference also observed at nine years. Nine-year-olds and adults of both groups use a greater variety of structures, such as intransitives, ge/-passives, and full passives in English, and intransitives, passives, statives, and reflexives in Turkish. Speakers of the two languages thus use different grammatical structures to effect their preferred discourse strategies to reach the same narrative outcomes. Turkish-speakers prefer topic maintenance and the boy-subject strategy, and English-speaking narrators prefer topic shift and the boy-object strategy.


324

Table 5. Percentage of Turkish-speaking subjects using different grammatical structures, by age Object of

Subject of

Age

Transitive

Stative

Intransitive

Passive

Reflexive

Transitive

20%

0

7%

0

0

7%

4;0

40%

7%

7%

0

0

13%

5;0

40%

7%

20%

0

0

20%

9;0

20%

7%

27%

13%

0

27%

0

13%

13%

27%

13%

27%

3;0

20;0

Table 6. Percentage of English-speaking subjects using different grammatical structures, by age Object of

Subject of

Age

Transitive

Stative

Intransitive

Passive

Prep. Phr.

Transitive

3;0

17%

8%

0

8%

0

0

4;0

33%

0

17%

8%

0

8%

5;0

8%

0

0

25%

8%

42%

9;0

0

0

17%

17%

17%

50%

20;0

0

0

10%

40%

10%

30%

4. DISCUSSION In conclusion, our results show that the ability to integrate a misrepresentation that belongs to the level of consciousness into the landscape of action is a late development. We do not find any reference to the misrepresentation in the descriptions of the three- and four-year-olds, nor in the narratives of most of the five-, seven-, and nine-year-olds. The majority of the older children, in particular the seven- and nine-year-olds, however, underline the unintentionality of the boy's ending up on the deer's head by a variety of linguistic means. Thus, around seven years of age children start attending to the level of consciousness by using expressions that attribute different levels of intentionality to their protagonists even if they do not explicitly refer to their mental states. At nine, children's narratives similarly remain limited to the level of action along with an implicit reference to intentional states, although a few subjects also explicitly refer to the misrepresentation. It is only the adults who explicate the misrepresentation and causally link it to the sequence of plot-advancing events. Similar developmental trends are reported by Bamberg and Damrad-Frye (1991) and Kuntay and


325

Nakamura (this volume), who, respectively, compared English and German frog stories and Turkish and Japanese frog stories, studying various types of evaluative comments. Several factors may account for this late development. The first involves the difficulty of negating the apparent and affirming the real identity of the misrepresented, and then integrating this into a sequence of events as a causal explanation. In Labov and Waletzky's terms, this involves co-ordinating a reference to an event that did not occur or is not the case with a reference to a sequence of events regarded as having taken place. In Bruner's terms, what is required is integrating a logical operation characteristic of paradigmatic thinking into the temporal ly-causally ordered sequences of narrative thought. Furthermore, in the present task, the recognition of the misrepresentation rests on the ability of the subjects to make inferences about the mental state of the protagonist from the representational medium of the pictures. Interpreting pictures and making narrative meaning from them, however, is in itself a task that requires experience, and children of different ages and possibly of different cultures differ in this respect. Bamberg and Damrad-Frye (1991), analysing the same database from English-speaking children for a broader range of references to mental and emotional states, indeed refer to children's dependence on perceptual cues, noting that "the coherency conditions for young children require a close link between precipitating event and facial expression", while older narrators rely more strongly on the hierarchical function of such situations (1991:706). A second difficulty involves the demands posed by the task of constructing a complex narrative, which in the present context means structuring a coherent flow of events, manipulating the perspectives of various protagonists and/or the narrator, as well as using the linguistic strategies necessary for their expression. In the picture descriptions by the youngest narrators, the protagonist who is presented as the subject/topic changes from frame to frame. Four- and five-yearolds single out the boy as the protagonist and adhere to his perspective by keeping him as the subject/topic and in an agentive role, suggesting that they are highly constrained by the demands of constructing a coherent story line. Children's competence in structuring the plot appears to become co-ordinated with their ability to use different linguistic structures around the age of seven to nine. At this age, children display a good command of the different linguistic structures (transitive, intransitive, stative verbs or passive/reflexive constructions) and discourse strategies (topic maintenance versus topic shifting) that adults use to talk about the misrepresentation and the boy's subsequent encounter with the deer. However, they still disregard the misrepresentation and use these linguistic strategies to talk about events on the level of action only, suggesting that what is at issue is not a linguistic but a conceptual difficulty. That is, when there are complex task demands on their narrativity skills, children's use of the capacity for mindreading is undermined. Adults, on the other hand, manipulated the different

326


linguistic strategies to skilfully integrate the landscape of consciousness with the landscape of action. Cross-linguistically, our comparison of the Turkish- and English-speaking narrators in terms of their skill at mindreading shows that the two groups do not differ in this respect. The specific devices for encoding false belief or marking representational change in Turkish do not appear to facilitate mindreading in the context of this task. Indeed, English-speaking children appear to be slightly better than Turkish-speaking children at doing this in the representational medium of the pictures. The two groups differ, instead, in terms of narrativity. in particular in terms of shifting perspectives in discourse and the linguistic means used for this purpose. While English-speakers prefer topic shifting as a discourse strategy. Turkish narrators prefer topic maintenance, continuing the episode from the point of view of the boy. These differences may be particular to the frog story, or it may be a function of the cultural practices of storytelling characteristic of the two language groups. This alternative explanation for the difference in the performance of the two cultural-linguistic groups consists in the theory that reference to internal states may be less normative for the storytelling practices of the Turkish-speaking group than for those of the English-speaking group, as suggested by Kiintay and Nakamura (this volume). These authors found no age-related changes in the total number of evaluative devices used in Japanese and Turkish frog stories; they further found that the percentage of clauses containing causal connectors that occur together with references to frames of mind is much lower in the Japanese and Turkish narratives than in the English ones analysed by Bamberg and Damrad-Frye (1991). These findings indicate that cultural differences in narrative style - in what speakers think may be explicitly stated and what they think may remain implicit - should not be disregarded. A final but related point concerns the fact that Turkish has evidential markers which inform the hearer about the epistemic stance of the speaker, and this takes us back to our initial question of whether the availability of such markers facilitates mindreading. Bamberg and Damrad-Frye (1991), who found a significant number of emotional-state references and truth-value qualifications in their English and German frog stories, propose that these two types of evaluation allow narrators to express their epistemic stance - the uncertain and the inferential nature of what they think is going on in the story (Bamberg & Damrad-Frye 1991:698). Kiintay and Nakamura, comparing their data with the data from English-speaking children analysed by Bamberg and Damrad-Frye, however, report relatively few epistemic-stance markers and causal explanations in Turkish and Japanese stories. They attribute this difference to differences in narrative style, pointing to the more informal performance of English-speaking children compared with the more formal style of their Turkish and Japanese peers. What we would like to


327

suggest, additionally, is the possibility that speakers of Turkish, a language which has grammaticalized evidentials, may not feel the same kind of need for using lexical qualifiers of epistemic stance since their language already provides them with grammatical means that function as hedges or disclaimers by indicating different sources of information and thereby different degrees of certainty. The fact that the evidential mood is a special marker of the narrative genre in Turkish makes this an even stronger possibility, which needs to be investigated in future research.

NOTES 1.

The different tense-aspect-modality markers in Turkish indicate different sources of evidence for belief states, such as direct experience (-DI), inference from physical evidence and hearsay (-mis), and knowledge about habitual and normative states of affairs (-DIr). The (-mis) particle marks change in mental representations due to acquisition of new/unexpected information and, depending on context, may yield meanings ranging from surprise and irony to pretense and fantasy.

2.

For more detail of this level of analysis, see Berman and Slobin 1994, Chapter I A.

3.

In the subject IDs of the examples, the first letter indicates the language: t = Turkish, e - English; the numeral indicates the age of the subject; and the following letter is the subject's code in that group. Further, in renderings of individual morphemes, capital letters - as in -ml$ - are used to represent vowel alternations.

4.

Since data from seven-year-olds are available for Turkish only, this age group was not considered in the statistical analyses.

5.

Also suggested by the use of verbs such as emerge and appear.


Linguistic Strategies Serving Evaluative Functions1 A Comparison between Japanese and Turkish Narratives

Ay I in C. Ktintay & Keiko Nakamura

\. INTRODUCTION The main task for a speaker asked to "tell a story" based on a sequence of pictures in a book is to spell out and inter-relate the events depicted in discrete frames, and to package the contents into a set of successive utterances. In Ruth Berman's (1995:295) terms, picture-based storytelling involves representing a "visual, static, spatial" kind of knowledge in a "verbal, dynamic, temporal" form. As summarized in the introductory chapter of Relating events in narrative, Berman and Slobin (1994), along with their collaborators, set out to identify different linguistic means used by narrators of different ages and speaking different native languages in inter-relating pictorially depicted events in the form of a story. For this purpose, they used a wordless picture-book story called Frog, where are you? (Mayer 1969). The focus of the original research endeavour was to study how narrators put into words the thematic progression represented in the series of pictures, using linguistic devices such as tense/aspect markers, voice alternations, inter-clausal connectives, and relative clauses. The efforts culminated in uncovering patterns of event encoding and sequencing which narrators of different ages and speaking different languages display in their narratives. Although presenting a coherent plotline based on the events depicted in the pictures might constitute the crux of picture-book storytelling, effective narration also calls for the narrator's interpretation of these events. In order to tell the story successfully, the narrator must infer and convey aspects of the story that are not overtly available in the individual pictures. For example, intentions of characters are not directly observable from the pictured scenes and therefore must be reconstructed by the teller. Likewise, causal relationships underlying inter-related components of the story are implicit in the pictures and require the narrator's explication. In this chapter, we look at the linguistic strategies Japanese

330

Aylin C. Kiintay & Keiko Nakamura

and Turkish narrators of different ages employ to signal their interpretive stance with respect to the sequence of events to which they are referring in their elicited narrations of Frog, where are you? (Mayer 1969). That is, among the components in Labov and Waletzky's (1967) classic division between the referential and evaluative functions of narrative, we focus on the latter. With the book's sketchy and ambiguous drawings and complex action sequences, the storytelling task elicited various interpretations and judgements from the narrators, providing us with appropriate data for our exploration of such evaluative language. We will discuss similarities and differences in the usage of evaluative devices across ages in Japanese and Turkish, and also compare our results with those of Bamberg and Damrad-Frye (1991), who conducted a similar study with data from Englishspeaking children. Before presenting our study, we will first examine the previous literature on a) evaluative devices in narrative and b) cultural factors in narrative.

2. EVALUATIVE DEVICES IN NARRATIVE In an influential article written in 1967, Labov and Waletzky specified what they proposed to be fundamental properties of narrative structure, providing grounds for much later research to explore various linguistic realizations of these structural categories. Their study was based on the tellings of near-death experiences drawn from 600 interviews conducted with American adolescents and adults. The main question Labov and Waletzky asked the participants was: "Were you ever in a situation where you thought you were in serious danger of getting killed?" In analysing the oral renderings of such emotionally significant personal experiences, Labov and Waletzky distinguished two main functions of narrative: referential and evaluative. Bruner (1986) proposed a similar distinction between the landscape of action and the landscape of consciousness, which correspond to the referential and the evaluative functions, respectively. At the centre of the definition of the referential function of a personal narrative lie the criteria of sequentiality and temporality, namely, that the function of extended discourse composed of sequential clauses is to refer to temporally ordered events or actions. The evaluative function, on the other hand, holds off or complements the advance of the referential function and includes a variety of ways to reveal the teller's interpretation of and attitude towards some of the narrated events (Bamberg & Damrad-Frye 1991). The narrator removes him/herself from the activity of describing behaviours or actions, and focuses on a particular mental state, or an outcome of behaviours and actions, from an evaluative perspectiv e. In his more recent writing, Labov (1997:403) offers the following comprehensive definition for the evaluative function: the "evaluation of a narrative event is information on the consequences of the event for human needs and desires".

Linguistic Strategies Serving Evaluative Functions

331

In addition to highlighting the importance of the evaluative function of narrative, Labov and Waletzky set out to determine the common linguistic devices which serve this function. Their model for story evaluation distinguishes between two types of evaluation: external evaluations that appear outside the eventive clauses and involve the introduction of the narrator's perspective through general interpretive commentary; and internal evaluations that are incorporated into the frame of eventive narrative clauses, serving to specify the extent and motivation of a particular action. In external evaluations, the whole clause serves to enunciate the point of the story, giving insight into the perspective from which the narrative events are being organized. In internal evaluations, the narrator's stance is expressed as part and parcel of the action structure, and therefore is embedded within referential clauses. Internal-evaluation devices include intensifiers such as modifiers and prosodic emphasis cues that serve to highlight events, comparators such as comparatives and negatives that compare the events that occurred with those that did not take place, correlatives such as progressives and other aspectual forms that describe simultaneous events, and expletives such as subordinate clauses that provide reasons and motivations for the events. In short, the evaluative function, as defined by Labov and Waletzky, includes a wide array of means, ranging from explicit statements by the narrator concerning his/her attitude towards the communicated events to such aspects of the internal clausal structure as the use of negatives. Labov and Waletzky found that any linguistic proposition in a narrative can include an internal evaluative element, even though evaluative clauses tend to form an evaluative section, usually placed around strategically relevant points of the narrative, such as the resolution of the story, where complicating actions are brought to a close. As the majority of narratives collected by Labov and Waletzky were about near-death experiences, the notion of evaluation has been initially developed for first-person narratives about peaked and unusual experiences. Labov (1972:371) proposes that eventfulness is an intrinsic aspect of some personal experiences and can be expressed by evaluative devices in personal narratives: "evaluative devices say to us: this was terrifying, dangerous, weird, wild, crazy, or amusing, hilarious, wonderful; more generally, that it was strange, uncommon, or unusual - that is, worth repeating. It was not ordinary, plain, humdrum, everyday or run-of-themill". Since the stories that Labov and Waletzky were looking at were typically centred on an emotionally charged event, evaluations were found to be an essential component. Elicitation of such peaked personal experiences unsurprisingly produces highly evaluated and highly performed narratives - probably owing to the inherent tellability attributed to the events by the storytellers. Near-death experiences often lead to conversational narratives that tend to get repeated for different audiences, leading to the elaboration of both descriptive and dramatic or humorous elements of stories (Ervin-Tripp & Kiintay 1996).

332


Young children also repeat personal narratives that are about unforgettable events in their lives, and eventually consolidate highly elaborate narrative structures with much evaluation for the same story content (Kiintay & Ervin-Tripp 1997, 1998). Both Umiker-Sebeok (1979) and Kernan (1977) have examined the occurrence and distribution of Labovian components, including evaluation, in young children's conversational narratives. Umiker-Sebeok found devices such as lexical intensifies, references to emotional states, and comparators in pre-school children's intra-conversational narratives. The frequency of use and the variety of these evaluative devices showed an increase from three to five years of age. Kernan also found evidence for the Labovian type of expressive evaluation in the narratives of pre-school children. Miller and Sperry (1988), in a study of even younger children in family settings (two-year-olds in South Baltimore), observed interactions rich in talk about affective states of family members. This, Miller and Sperry argued, can be considered a type of narrative-evaluation strategy, playing an important role in the origins of personal-narrative talk in this community. Although the contents of picture-book stories do not usually give rise to such strong personal involvement, narrators demonstrate a considerable effort in expressing the significance of the sequence of events from their own emotional and subjective point of view in third-person stories as well (Bamberg & Reilly 1996). In extending Labov and Waletzky's distinction between evaluative and referential clauses from first-person narratives of personal experience to thirdperson narratives - in particular, frog stories -, Bamberg and Marchman (1990: 111) describe narrative clauses as "linearly ordering event descriptions along a horizontal (temporal) axis", while evaluative clauses signal the "consequentially and meaningfulness of such events according to the vertical axis of hierarchical order". In other words, while action descriptions present a linearly organized order of events represented in the pictures, evaluated outcomes of these actions or descriptions of mental states serve to highlight the overall hierarchical perspective from which the narrative gains coherence, indicating the significance of the individual events and actions. In analysing lexically expressed narrative evaluation in the frog stories told by five-year-old, nine-year-old, and adult speakers of American English, Bamberg and Damrad-Frye (1991) considered five categories of evaluative devices: references to characters' mental and affective states (frames of mind), character speech, hedges, negative qualifiers, and causal connectors. Overall, they observed developmental trends in the frequency of clauses that contain evaluative devices. The evaluative references to frames of mind were not very frequent at pre-school ages, with increasing use in nineyear-olds' and adults' narratives. Berman and Slobin (1994) observed similar differences between Hebrew-speaking nine-year-olds and pre-school children. In fact, the tendency to attribute inner states and affective responses to the story characters was listed as a salient property characterizing the stories of nine-year-


333

olds. Apart from analysing the density of evaluative devices in English frog stories told by different age groups, Bamberg and Damrad-Frye also observed differences between the discourse functions of reference to characters' inner states in the narratives of pre-school children and older speakers. They reported that most of the expressions of inner states in pre-school narratives were direct read-outs of the characters' facial expressions locally precipitated by immediately preceding events. Older narrators, on the other hand, tended to consider the global plot structure of the narrative in their usage of frames of mind, constructing an intentional framework for communicating aspects of this structure to their listeners. Reilly (1992) took into consideration para-linguistic evaluative elements such as narrators' gestures, facial expressions, and prosodic features, in addition to examining explicitly linguistic devices, in her analysis of evaluative language in the frog stories of American children aged three to eleven years. She found that while such para-linguistic evaluative strategies were predominant in the young children's narratives, with increasing age explicitly linguistic devices gained precedence. Reviewing developmental research on both personal-experience accounts and picture-based stories inspired by Labov's work, Berman (1997b) proposes the consideration of attitudinal-evaluative elements within an integrated tripartite analysis of narrative components, which also includes eventive elements and factual-informative elements. Eventive elements correspond to the referential function in Labov and Waletzky (1967) and describe events in temporally ordered clauses. Factual-informative elements provide background information about the location and temporal properties of events and about the attributes of the characters. Evaluative elements "include the narrators' subjective commentary or their attitude towards the events they are reporting and how they assume the protagonists relate to these events in terms of the motivations, emotions, and mental states that narrators attribute to these protagonists" (Berman 1997b: 241). While conceiving evaluation as an analytically separable component of narratives, both Berman (1997b) and Aisenman and Assayag (1999) point out that most past research on evaluation has ignored the distinction between structural, semantic, and discourse-functional criteria in determining the elements they count in different texts and at different ages. They argue that purely structural criteria such as the presence of subordination or negation in a clause cannot sufficiently distinguish evaluative from non-evaluative clauses. Semantic criteria such as whether a given clause contributes to the thematic progression of the plotline are also deemed to be inadequate given the existence of intermediate clauses that combine an advance in the referential content with performing an evaluative function. This is what Aisenman and Assayag (1999) call "action in evaluation". Discourse-functional criteria refer to the rhetorical functions carried

334


out by linguistic devices, and include global-level devices such as repetition as part of the evaluative structure of narratives. Considering all three criteria. Aisenman and Assayag (1999) advocate an approach that considers local, clausebased evaluative elements and global evaluative strategies simultaneously.

3. CULTURE AND NARRATIVE Previous research has also focused on the effects of systematic linguistic differences on narrators' rhetorical styles regarding the preferences of speakers of different languages in their description of the same events (e.g., Berman & Slobin 1994). Furthermore, cultural factors, in addition to cross-linguistic influences, seem to strongly influence the content of stories as far as there exist cultural styles of narrative construction that are transmitted to children. Representing a sociocultural perspective on the study of narrative performance, Gee (1991:20) points out that "the fundamental function of narrative in human life is not to report a chronological sequence of events, but to signal a perspective on events and create a satisfying pattern of themes one has drawn from various social traditions". As far as there exist cultural styles of narrative construction that are transmitted within communities, cultural factors may be speculated to influence how a story is elaborated content-wise. Research on children's narratives collected from different cultures (e.g., African-American, Hawaiian, and Japanese) has shown that from early childhood, children learn the narrative mode of discourse favoured by their caregivers (e.g., Au 1993; Michaels 1981: Minami 1996). McCabe and Peterson (1991), in studying differences in personal narratives by children of different cultural backgrounds, document that unique narrative styles of various cultures affect the structure of children's narratives and reflect their habitual interactions with their parents. For example, in a study of cultural variation of Labovian narrative components, they found that European North American children stress action sequences and focus less on evaluation than African American, Japanese American, and Latino children (Peterson & McCabe 1997). Similarly, Minami and McCabe (1995), with their comparisons of personal narratives told by Japanese and North American children to their mothers, report that there are culture-specific patterns of social interaction in narrative production, and that in different cultures there are differences in both mothers' and children's narrative styles. Furthermore, it is clear that what constitutes a "good" narrative varies according to the culture and language of the speaker. For example. Heath's (1983) work on children learning to use language at home and at school in two communities only a few miles apart in south-eastern United States clearly illustrates that even within the same country, cultural norms for constructing


335

effective narratives vary drastically. The form, structure, occasions, content, and functions of the narratives in the two communities differed greatly. While people in the White working-class community of Roadville tell straightforward, factual narratives, using stories to reaffirm group membership and behavioural norms, people in the Black working-class community of Trackton tell highly exaggerated stories to assert individual strengths and powers. As Heath writes, hearing different kinds of stories, the children in the respective communities develop competence in telling stories in highly contrasting ways. Likewise, Minami and McCabe (1995) report that Japanese children learn how to construct culturally appropriate succinct narratives relating multiple experiences, while North American children formed lengthier narratives focusing on one personal experience. There is a growing body of literature on Japanese narratives. However, most of this research focuses on adults (e.g., Maynard 1993). Several researchers, such as Minami (1996, 1998), looking at personal narratives, Nakamura (1993, 1999), looking at elicited narratives, and Uchida (1990), examining a wide range of topics pertaining to narrative content, have begun to work on Japanese children's narratives. Unfortunately, as in the case of research on narrative development in other languages, most of these studies focus on narrative structure rather than evaluative content. Previous studies on Turkish narratives have also mostly focused on linguistic forms that contribute to the referential content of discourse or on conversational occasions that lead to naturalistic narratives (Aksu-Ko? 1994; Kiintay & ErvinTripp 1997). Aksu-Ko9 (1996), in a study of film retellings obtained from young adults of low education/low income and high education/high income backgrounds, examined whether speakers insert their own interpretive judgements into their stories. She found that the "author's voice" was most clearly invoked by low education/low income male respondents, and proposed that these differences relating to socio-economic status and gender are accountable for the relatively separate sets of interpretive codes employed in different social environments. Similarly to this last study, focusing on the evaluative content of narratives, rather than excluding this as irrelevant from the analysis of plotline events, will help us to reconsider the issue of cultural factors in the area of narrative research.

4. PURPOSE The purpose of this study is to a) examine what types of evaluative devices Japanese and Turkish children and adults use in their construction of oral narratives; b) analyse the data from a developmental perspective, focusing on similarities and differences in the usage of evaluative devices across ages; and c) discuss similarities and differences in the usage of evaluative devices in the two

336

Aylin C. Kuntay & Keiko Nakamura

languages and also compare our results with those of Bamberg and Damrad-Frye (1991), who conducted a similar study with data from English-speaking children. Research on this topic will add to the growing body of literature on narrative development across languages as well as shed light on how much children know about mental states and feelings, and how readily they can attribute these states to others in different contexts. Furthermore, looking at this topic from a crosscultural perspective will help us understand the effects of systematic linguistic and cultural differences on narrators' rhetorical styles regarding their preferences in their description of the same set of events.

5. METHOD

5.1 Subjects Japanese For Japanese, 69 children and 16 monolingual adults served as subjects. The children were divided into four age groups: four-year-olds (17 subjects), fiveyear-olds (17), seven-year-olds (19), and nine-year-olds (16). All participants were from similar middle-class socio-economic backgrounds in the Tokyo and Kobe metropolitan areas of Japan. Turkish The Turkish narrators were residents of Istanbul, also coming from middle-class socio-economic backgrounds. There were 16 adults and 66 children, who were divided into four age groups: four-year-olds (16), five-year-olds (20), seven-yearolds (15), and nine-year-olds (15). 5.2 Procedure An elicited-narrative task using the 24-page wordless picture book Frog, \\~here are you? (Mayer 1969; see Appendix I) was conducted in order to examine the use of evaluative language. As also seen in the other chapters in this volume, and from Berman and Slobin (1994), this method allowed us to compare narratives across ages and languages. Extensive data using this methodology have been collected in languages as diverse as Mandarin Chinese, Spanish. Russian. English, and Australian Aboriginal languages. The story is about a boy, a dog, and their pet frog. In order to tell the story successfully, the child or adult must infer and convey aspects of the story that are not overtly available from the individual pictures. For example, the narrator must infer relationships between characters (e.g., friendship, animosity), attribute emotions and mental states to characters


337

(e.g., tear, surprise, desire, worry, thinking), as well as provide motivations and causal explanations for the characters' actions. All of the stories were tape-recorded and transcribed. The transcripts were coded into clauses to obtain a measure of story length. 5.3 Coding categories The Turkish and Japanese frog stories were coded for lexical items falling into one or more of the following categories of evaluative language. All of these evaluative devices enrich narratives by providing additional information to the plotline, which makes the story more engaging and vivid. They reflect the narrators' ability to infer relationships between individuals and events as well as to infer affective and mental states and attribute them to the characters in the story. In this section, we present the categories, along with our motivations for including them as narrative evaluative devices, and give some examples. The first five categories were adapted from Bamberg & Damrad-Frye 1991. Frames of mind Frames of mind are expressions that refer to mental and affective states of characters, serving to qualify the nature of the links between subsequent events. These internal states must be inferred by the narrator. They also function to encourage empathy and interest in the audience. This category consists of references to emotional states such as happy, scared, surprised, and worried. Emotion verbs (e.g., Turkish: sevmek 'to like/love', sinirlenmek 'to get scared', kizdirmak 'to make angry'; Japanese: odokasu 'to scare someone', bikkuri sasern 'to surprise someone') were included in this category. Further, references to purely cognitive states pointing to the mental activity or the epistemic status of the narrator (e.g., Turkish: du^tinmek 'to think', zannetmek 'to think falsely', sanmak 'to assume'; Japanese: kangaeru 'to think', kimeru 'to decide') were also counted in this category. Hedges Hedges are devices that inform the audience about the narrator's uncertainty with respect to the truth value of the proposition expressed in a clause. These consist of lexical devices that serve as epistemic-distancing devices, suggesting noncommitment to the truth value of the proposition, such as galiba 'probably' and verb-bilir 'might be' in Turkish, andyoo da 'seems like', mitai na 'looks like', and tabun 'probably' in Japanese. This category also includes expressions that convey uncertainty on the part of the narrator with respect to what is asserted, such as anladigim kadariyla 'as far as I can understand', zannediyorum 'I assume' in Turkish, and omou 'I think' in Japanese.

338


Negative qualifiers Negative qualifiers highlight the underlying expectations and sometimes the surprise of the narrator by marking story circumstances or events that might have taken place, but did not. In other words, they mark the discrepancies that the narrator expresses with respect to some canonical-event knowledge. Examples are oraya du$tu, bogulmadi da 'he fell down there [in the lake], and did not drown' (Turkish) and soto o sagashitemo miatarimasen 'even though (they) searched outside, (they) were unable to find him [the frog]' (Japanese). Any direct negation of a state or an action was included in this category. Character speech Character speech serves to attribute speech and, therefore, particular intentional states to a character, contributing an alternative perspective to that of the narrator, bringing immediacy and vividness to the narrative. Direct statements of characters' utterances are often accompanied by animated voice and intonation, bringing a sense of engagement to the narrative. Examples of direct character speech are: sonra kopege cocuk kizmi$, "kurbagamn neden kavanozumi kirdin? " demis, 'then the boy got mad at the dog and said "why did you break the frog's jar?" ' (Turkish) and otoko no ko \va "shizuka ni!" to itta 'the boy said "be quiet!" to the dog' (Japanese). In addition, statements of indirect speech were also examined. Examples of indirect speech include statements such as cocuk kopege sessiz olmasim soyledi 'the boy told the dog to be quiet' (Turkish) and otoko no ko wa inn ni shizuka ni suniyoo ni itta 'the boy told the dog to be quiet" (Japanese). Causal connectors Causal connectors point to the inferred motivation for an action and provide causal frameworks for the implicit relationship between events in a narrative. This category consisted of inter-clausal connectors such as ctinkti 'because', icin 'in order to', and dive 'so that' in Turkish, and kara 'because', sorede 'because of that, therefore', and no ni 'in order to' in Japanese. Enrichment expressions This category consists of adverbial phrases such as 'again' and 'suddenly' (Turkish: yine and birdenbire, respectively; Japanese: mala and kyuu ni) which reveal the unexpected or inferred nature of an action; intensifiers such as emphatic markers, e.g., cok (Turkish) and totemo (Japanese) 'very'; and repetition, which helps to draw the listener's attention to a specific event. Connectives such as ama (Turkish) and demo or shikashi (Japanese) 'however' that preface unexpected or contrastive occurrences of events were also included in this category.


339

Onomatopoeia and mimesis These devices use sound-related effects to attract the listener's attention and convey information in a more intense and marked way, e.g., pattadak 'with a thud' and vizlamak 'to make a buzzing sound' in Turkish. Japanese is particularly abundant in sound symbolism in the form of onomatopoiea and mimesis. Japanese mimetic expressions can be divided into phenomimes, which are phonetic representations of phenomena perceptible by non-auditory senses which try to describe the manner or looks of a situation (e.g., kushakusha 'all rumpled', guruguru 'round and round'); psychomimes, which are phonetic representations of human psychological states (e.g., mutto 'offended, sullen', gakkari 'disheartened'); and phonomimes, which are attempts to imitate sounds (e.g., wanwan 'bow-wow', bunbun 'buzz-buzz') (Martin 1975). These are an integral part of Japanese adult spoken and written language, and are used especially frequently in child-directed speech and children's own speech. Evaluative remarks Evaluative remarks convey the narrator's reflections about the events of the narrative by communicating a subjective point of view. These remarks are mostly asides that suspend the advance of the referential storyline for the purpose of expressing the narrator's mostly ethical or aesthetic judgments on the related events, and they are often found in Turkish narratives. An example is kurbaga da, tabii ki, tabiatm yaratigi 'the frog, of course, is a creature of nature' (Turkish). Verb-style shifts In Japanese, shifts between teineigo 'formal style' and informal style often occur when shifts in narrative perspective occur. Teineigo, usually marked by the use of desu/masu predicates, is often used when the speaker is speaking in the voice of the narrator, while informal style is frequently used for character speech. Among a variety of functions, formal style has been described as functioning as a distancing device, while informal style is often used to mark intimacy. Verb-style choice has also been related to the narrator's external and internal positioning in relation to the event that is being described (e.g., Maynard 1993; Nakamura 1999).

6. RESULTS 6.1 Japanese data The first analysis examined the distribution of linguistic devices for evaluative purposes across the five Japanese age groups. Since all comparisons of the use of evaluative devices have to be related to story length, the overall story length for

340

Aylin C. Kilntay & Keiko Nakamura

each age group was established, using the number of clauses as an indication of story length. The mean number of clauses was 31.1 for the four-year-olds, 41.4 for the five-year-olds, 51.0 for the seven-year-olds, 37.9 for the nine-year-olds, and 57.4 for the adults. Story length increased steadily with age, the nine-yearolds being an exception. As regards the overall use of evaluative devices, there did not seem to be much of a difference in numbers across the age groups. Surprisingly, the nineyear-olds seemed to use the fewest evaluative devices (relative to story length), while the adults used the most (see Table 1). Analyses of the Japanese narrative data revealed various trends in the use of specific evaluative devices across age groups: Frames of mind In general, with increasing age, the number of references to frames of mind increases (with the exception of the nine-year-olds)2. Examples are as follows: [1]

Sorede otoko no ko ga nanka aru ka naa to omotteru no. That is why the boy is thinking "is there something [in this hole]?" ' (4:4. male)

[2]

Kukkii wa hachi no su wo mitsukete oo-yorokobi. 'Cookie [the dog] found a beehive and was very happy/ (5;5. female)

[3]

Otoko no ko wa bikkuri-shite okkotchatta no. The boy was surprised and fell down.'(7;l, female)

The number of types of mental/affective expressions also increased steadily with age. For example, the four-year-olds used 15 different types of mental-affective expressions, the seven-year-olds 22 different types, and the adults 41 different types. Thus, the older narrators were able to use a larger variety of mental/affective terms. While the mental/affective expressions used by the four-year-olds were limited to a handful of basic mental-state and emotion terms, such as omou 'to think', hoshii 'to want', bikkuri sum 'to be surprised', okoru 'to get angry', and yorokobu 'to be happy', the adults were able to provide a wider range of expressions, such as gokigen o torn 'to try to get on someone's good side', ka\vaigaru 'to treat fondly'. mushi sum 'to ignore', yakusoku o sum 'to promise', hotto swu 'to be relieved'. kanshin sum 'to be impressed', and koishii 'to long for'. Hedges The number of hedges was relatively low in all of the age groups, with the exception of the seven-year-olds. No age-related developmental trend was found, either regarding percentage of hedges or types of hedges. For example.


341

Table 1. Percentage of clauses containing types of evaluative devices in Japanese by age 4 years

5 years

7 years

9 years

Adults

Frames of mind

8.7%

8.1%

10.5%

6.4%

13.2%

Hedges

1.3%

1.9%

4.3%

2.3%

2.3%

Evaluative device

Negative qualifiers

6.1%

6.1%

5.3%

5.3%

6.5%

Character speech

9.8%

10.1%

7.4%

5.3%

5.1%

Causal connectors

2.8%

1.0%

3.9%

2.0%

3.7%

Enrichment expressions

4.5%

4.8%

3.9%

9.7%

10.0%

Onomatopoeia and mimesis

3.6%

3.3%

3.1%

2.2%

1 .0%

Total

36.9%

35.2%

38.5%

33.2%

41.9%

Average no. of clauses

31.1%

41.4%

51.0%

37.9%

57.4%

regarding types of hedges, while the four-year-olds used five types of hedges, including deshoo 'probably', moshikashitara 'maybe', and mitai 'like', the adults also used only five types of hedges, some of which overlapped with the forms selected by the four-year-olds, such as deshoo 'probably', while others differed (e.g., voo da 'it seems that'). The large numbers of hedges in the sevenyear-olds seemed to occur owing to the preference of a small handful of the narrators to rely heavily on hedging; there were strong individual differences in the use of this form. [4]

moshikashitara inu ga koko ni noborenai to omou. '(1) think that maybe the dog cannot climb up here.' (4;4, male)

[5]

kitto kono kaerutachi wa kazoku nan deshoo. These frogs are probably a family.' (adult, female)

The hedges served to mark the narrator's uncertainty with respect to the truth value of the utterance. Negative qualifiers The number of negative qualifiers was at about the same level for all age groups. In general, the most common negative qualifiers were negations of states, as in inai 'to be gone (animate)'. However, even the youngest children were able to use negations of actions/expectations, such as ochitenai 'isn't falling', as well as negations of potentials, such as noborenai 'can't climb up' and torenai 'can't put (something) inside', as seen in [7] and [8]:

342

[6]


Sorede kaeru ga inai no. 'And then, the frog isn't there.' (5;4, male)

[7]

Moshikashitara inu ga koko ni noborenai to omou. '[I] think that maybe the dog can't climb up here.' (4;4, male)

[8]

Sagashiterun janakute asonderu. '[The dog] isn't looking [for the frog], he's playing.' (7;8, female)

Overall, with age, children become more skilful at encoding concepts related to negation of expectation and potential. Character speech Analyses of narratives show that character speech decreases with age. Initially, Japanese children seem to rely heavily on character speech with animated tone of voice and gestures, while adults tend to use other forms of evaluative devices. Character speech was the evaluative device used most frequently by the fourand five-year-olds. Direct character speech decreases with age: most of the cases of character speech by the older narrators involved indirect speech. Several examples are given below. [9]

Baibai-tte itta. '[He] said "bye-bye".' (4;4, female)

[ 10] Mogura ga dete-kite sorede kusai-tte itta no. 'A mole came out [of the hole], therefore [the boy] said "[he] stinks!" ' (5:4. female)

[11] Otoko no ko wa inu ni shizuka ni suru yoo ni iikikasete ki no uragawa wo nozoki-komimashita. 'The boy told the dog to be quiet over and over again, and peered around the back of the tree.' (adult, female)

Causal connectors The number of causal connectors did not show any developmental trend. The number of causal connectors was relatively high for the seven-year-olds and adults, but was much lower for the five-year-olds and nine-year-olds. There were strong individual differences in the use of causal connectors, especially among the younger children. For example, only 4 of the 17 four-year-old subjects and 6 of the 17 five-year-old subjects used causal connectors. Furthermore, their choice of causal connectors was limited to kara 'because' and voo ni 'in order to'. However.


343

these narrators tended to use causal connectors relatively often. A larger proportion of the older subjects used causal connectors, and as one might expect, they were also able to use a larger variety of linguistic devices to mark causal relations. Examples of causal connectors from the Japanese narratives are as follows: [ 12] Risu ga deta kara naiteru. "A squirrel came out so [the boy] is crying.' (4;6, male) [ 13] Kaeru ga nigechawanai yoo ni inu ga mihatte iru. 'In order that the frog doesn't run away, the dog is watching him.' (7;2, female) [ 14] Kaeru o yonde mita kara hachi ga dete kimashita. 'Because [he] tried calling for the frog, bees came out.' (9;3, male) [ 15] Inu wa bin no naka o sagashita tame ni bin kara hanarenaku natte shimaimashita. "Because the dog searched the inside of the jar, [he] became unable to separate from the jar.' (adult, female) Almost all of the causal connectors were either purposive causals, providing motivations or reasons for a character's actions or behaviours, or causals marking physical causality relations. Enrichment expressions The use of enrichment expressions also increases with age. While the four-yearolds relied on a small number of adverbial expressions (e.g., mata 'again', mada 'still', itsumo 'always', shizuka ni 'quietly'), the older children and the adults were able to use a wider range of adverbial expressions to heighten the sense of drama, such as ikinari 'suddenly', totemo 'very', dake 'only', and hisshi ni 'with all his might'. The older narrators were also able to use more complicated sentence constructions including connectives such as tokoro ga/kedo 'however' and -temo 'even though' which express occurrences that are contrary to expectation. The following examples appeared in the narratives: [16] Mata ana o miteru. '[The boy] is looking at the hole again.' (4;8, female) [ 17] Sorede isshookenmei sagashitan dakedo inakatta. "And then [the boy] looked as hard as he could, but [the frog] wasn't there.' (4; 11, female)

344


[ 18] Otoko no ko ga ana o nozoki nagara uta o utatte-itara ikinari risu ga tobidashite kita. 'When the boy was singing as he was looking into a hole, all of a sudden a chipmunk came flying out.' (7;7, female)

[19] Dakedo inu wa genki desu. 'However, the dog is fine.' (9;7, male)

Onomatopoeia and mimesis The use of onomatopoeia and mimesis decreases with age. Children tended to use a larger proportion, as well as a larger variety, of mimetic expressions. Adults tended to rely on other linguistic strategies in their narratives. Examples are as follows: [20] Hachi ga bunbun tonda. 'The bees flew buzzing.'(4:0, male)

[21 ] Koomori ga deta kara boten-to koketa. 'A bat flew out, so [the boy] fell down with a thud.' (4:6, male)

[22] Kaeru ga pyon-to nuke-dashimashita. 'The frog sneaked out with a hop.' (5;5, female)

In general, the expressions used by the children were phonomimes (attempts to imitate sounds), with a few phenomimes (phonetic representations of phenomena perceptible by non-auditory senses which describe the looks or manner of a situation). Young Japanese children seem to rely heavily on mimetic devices, which is not surprising as these are used frequently in child-directed speech and in children's own speech. In addition, such devices are often used in children's storybooks and comic books. The decrease in the use of mimetic devices w ith age seems to be accompanied by an increase in the use of adverbial enrichment expressions. Verb-style shifts in Japanese Verb-style choice is now considered more than just an issue of politeness indicating the ingroup—outgroup distinction or inter-personal distance (e.g.. Shibatani 1990), and it has been shown to be a marker for different indices of the self (e.g., an external, "social" self as opposed to an internal, "private" self) as well as a marker for the narrator's external and internal positioning in relation to the event that is being described (Cook 1997; Maynard 1993; Minami 1998: Nakamura 1999).


345

Table 2. Percentage of subjects using formal, informal, and both verb styles in Japanese by age Verb style Informal

4 years

5 years

7 years

9 years

Adults

100%

89%

68%

50%

25%

11%

25%

44%

21%

25%

31%

Formal Both

11%

Looking at the data provided in Table 2, we can see that none of the fouryear-olds used formal style in constructing narratives. Of the five-year-olds, only two children told their story using both formal and informal verb styles. The data from the seven- and nine-year-olds and adults show that older speakers tend to choose the formal teineigo style in narrating stories. In other words, speakers tend to use the formal style to mark the narrator's perspective. Switches between verb styles occurred when the narrator switched from narration in the formal style to direct character speech in the informal style. The young children using both styles sometimes switched incorrectly, probably owing to a lapse in attention to style which occurred while they were paying attention to narrative content. Verb-style choice seemed to be closely related to personal style. Preliminary results from a project on language socialization and narrative structure show that children whose mothers frequently read storybooks to them seemed to be particularly keen on switching between the two verb styles. This is not surprising, given the fact that an informal survey of 200 children's picture books showed that the majority of books for pre-schoolers are written in the formal style, with dialogue in the informal style (Nakamura 1999). In addition, preliminary results from another project examining differences in adult narrative style show that storytelling experience (i.e., whether the person frequently tells/reads stories to children) also influences verb-style choice. Parents who frequently read storybooks to their children are more likely to switch between the two verb styles, while adults without children are less likely to do so. However, it is clear that the relationship between narrative ability and exposure to storybooks needs to be studied more carefully. Relative frequencies of evaluative utterances in Japanese The relative frequency of a given evaluative device was calculated as the percentage of (the total number of) clauses which contained the evaluative device in question. The results, which are summarized in Table 1, show that the relative frequencies of the evaluative devices did change with age. Four-year-olds and five-year-olds used character speech more than any other evaluative device, followed by frames of mind and negative qualifiers. Seven-year-olds, however, used frames of mind more than character speech, although here, too, negative

346


qualifiers came third. Nine-year-olds used enrichment devices more than frames of mind and negative qualifiers/character speech, which ranked second and third respectively, while adults used frames of mind most often, followed by adverbial expressions and negative qualifiers. Two of the coding categories showed no agerelated trends: hedges and causal expressions. 6.2 Turkish data The second set of analyses examined the distribution of linguistic devices for evaluative purposes across the five Turkish age groups. Turkish stories were longer than Japanese stories in every age group. The mean number of clauses was 55.8 for four-year-olds, 65.9 for five-year-olds, 58.6 for seven-year-olds. 55.5 for nine-year-olds, and 65.9 for adults. The overall number of evaluatives relative to story length did not seem to differ across the children's age groups, but there was a big burst in the adult group. Table 3 indicates the percentage of clauses that contain each type of evaluative device listed in the previous section. Frames of mind As in the Japanese stories, the number of references to characters' frames of mind increases with increasing age of the narrators. Examples with references to characters' frames of mind are as follows: [23] Vine "niye kavanozu parcaladin?" diye kizdi. 'Again he got angry, saying "why did you smash the jar?" ' (4;6, male)

[24] Kopek de bu arada korkusundan hizlan kosuyordu. 'In the meantime, the dog was running fast because of fear.' (7:0, female)

[25] Ktifuk 9ocuk ve kopegi bir kurbaga bulmuslar ve cok seviyorlarmis. 'The little boy and his dog found a frog, and they like him very much.' (adult, female)

The range of types of lexical elements employed to express frames of mind increases steadily after the pre-school years. For four- and five-year-olds, 11 different kinds of mental and affective expressions can be attested each. The number of types goes up to 15 for seven-year-olds, 20 for nine-year-olds, and 29 for adults. Hedges Hedges seem to occur only in the narratives of the adult Turkish group. The children do not seem to employ devices that indicate their level of uncertainty


347

Table 3. Percentage of clauses containing types of evaluative devices in Turkish by age Evaluative device

4 years

5 years

7 years

9 years

Adults

Frames of mind

4.9%

3.3%

6.0%

7.6%

8.9%

Hedges

0.5%

0.5%

0.5%

0.1%

3.2%

Negative qualifiers

2.8%

3.0%

3.5%

4.3%

3.7%

Character speech

8.3%

2.3%

2.1%

3.6%

1.4%

Causal connectors

0.5%

1.4%

2.5%

3.4%

1.3% 9.3%

1 .3%

Enrichment expressions

5.6%

6.1%

9.8%

0

0

0

0.5%

0.7%

Onomatopoeia and mimesis

1.0%

2.4%

1.6%

1.0%

5.9%

Total

21.3%

22.0%

20.5%

24.6%

36.2%

Average no. of clauses

55.8%

65.9%

58.6%

55.5%

65.9%

Evaluative remarks

about the events narrated, whereas adults often specify their lack of confidence in their inferences from the pictures. An example of an adult hedge is given below: [26] Onbirinci resimde bu arada kopek agaci sallayarak herhalde an kovanmi dusurmiis durumda. 'In the eleventh picture, in the meantime, the dog, it seems by shaking the tree, is in a state of having knocked over the beehive.' (adult, male)

Negative qualifiers Similarly to the Japanese narratives, the relative frequency of negative qualifiers does not show a developmental trend for Turkish narratives. Negative markers were most commonly used to indicate that the frog is missing, such as in the following example: [27] Bir bakmisti kavanoza, yokmustu. '[The boy] looked at the jar, [the frog] was not there.' (4;6, male)

From younger ages on, negative qualifiers were also used to mark discrepancies compared with expected event representations: [28] Ondan sonra koyun iistune dusmustu, ama bagirmadi. 'And then [he] fell on the bay, but did not scream.' (4; 10, male)

348


In the adult narratives, some of the negative qualifiers used were embedded in subjective evaluative remarks that reflect the narrator's disapproving stance towards the paths of action taken by the narrators. The following example is illustrative of such uses of negation: [29] Ama cocugu almamasi gerekiyor bana sorarsan. 'But he must not take the baby [= the baby frog], if you ask me.' (adult, female)

Character speech In Turkish stories, character speech constitutes the most preferred evaluative device for four-year-olds. Its use decreases from five years of age on. In the stories of pre-school children, all of the references to speech of characters are of the direct kind, while it is only adults who often employ indirect speech, which involves nominalization of the embedded verb, as seen in [30] and [31]: [30] Ondan sonra cocuk da "burnum acidi" diyor. 'And then the boy says "my nose hurts".' (4;7, male) [3 1 ] £ocuk kopege sessiz olmasim soyliiyor. 'The boy told the dog to be quiet.' (adult, male)

Causal connectors The use of causal expressions shows a slight developmental trend in the Turkish narratives. Most of the use in Turkish stories involved purposive causals that provide explicit motivations for a character's actions or undertakings, such as the following: [32] Ondan sonra da ondan sonra da kafasma geciriyor kavanozu kopek aramak 'And then and then the dog puts the jar over his head in order to look for [the frog].' (5;8, male)

[33] Sonra bakiyor cizmelerinin icini anyor girdi mi acaba dive. 'Then [he] looks, searches in his boots, wondering/saying if [the frog] got in there' (adult, female)

In the adult stories, the function of some causal clauses is to present reasons offered by the narrator to explain the state of affairs in the story from a plotexternal point of view. [34] is an example:


349

[34] Ve o sirada baslan tabii ki belaya giriyor 9iinku kendi dogal ortamlarimn disina 9ikiyorlar. 'And of course, at that point they get into trouble, because they go out of their natural environment.' (adult, female)

Enrichment expressions Enrichment expressions are used the most by five-year-old and adult Turkish narrators. However, the types of devices used by pre-schoolers and adults show some differences. The young narrators employ adverbial expressions that carry aspectual meaning, such as yine/gene 'again, still', or that mark manner of motion, such as gizlice 'secretly', and intensifies such as qok 'very'. The older narrators' enrichment expressions include devices that reveal their own attitude towards the story events, such as tabii 'naturally' and halbuki 'nevertheless', in addition to the types of enrichment expressions used by pre-school children. Examples of enrichment expressions follow: [35] Kopegi hala baginyor. '[His] dog is still shouting.' (5;7, male)

[36] Bu arada camin a9ik oldugunu goriiyor ve 9ikiyor tabii. 'In the meantime [the frog] sees the window open and naturally goes out.' (adult, male)

Onomatopoeia and mimesis Unlike the Japanese narratives, the Turkish ones did not exhibit much use of sound-symbolism devices at all, especially in the younger age groups. A few onomatopoeic forms, such as the following one, can be observed in the narratives of nine-year-olds and adults: [37] Pattadak yere diisuyor. '[The boy] fell down on the ground with a thud.' (adult, female)

Evaluative remarks Subjective commentary of ethical or aesthetic content is one device that is particularly favoured by Turkish adults, not so much by younger Turkish narrators, and not at all by Japanese subjects. These clauses constitute stand-alone external evaluative devices that reflect the narrator's point of view or opinion with respect to the encoded events. The examples below illustrate the use of this device by a pre-school child and an adult:

350


[38] Ve kurbaga kavanozun icinden kaciyor. Dogrusu bu pek komik bir seydi. 'And the frog escapes from the jar. In fact, this was a very funny thing.' (5;7, female) [39] Dogal ortamindan yaratiklan cikarmamayi 6'greniyorlar. They learn not to take creatures out of their natural environments.' (adult, female) As in the Japanese stories, the preference for various evaluative strategies changed with age. Character speech is the most frequent evaluative used by four-year-olds, followed by frames of mind and enrichment expressions. From five years of age on, character speech quickly loses its preferred status, giving way to enrichment expressions and frames of mind. The most prevalent devices in the older age groups are also frames of mind and enrichment expressions. Among adults, using evaluative remarks is another favoured strategy.

7. DISCUSSION 7.1 Cross-linguistic comparisons Bamberg and Damrad-Frye (1991), in their study on English-speaking children's narratives, found that adults used three times as many evaluations as five-yearolds, and two-and-a-half times as many as nine-year-olds. This differs from the Japanese and Turkish data, which did not show any age-related differences in the total number of evaluative devices used. As for the specific evaluative categories, in general, we found that the English-speaking children exhibited a stronger developmental trend for frames of mind, hedges, and causal connectives. In addition, the English-speakers seemed to use character speech less frequently than the Japanese children. One surprising finding was that despite extensive use of grammatical evidential in Japanese and Turkish, there were relatively few hedges, compared with the English-speaking children. This may be due to differences in narrative style (e.g.. the Englishspeaking children are more informal, while the Japanese and Turkish children are more formal, making fewer hedges), or to the fact that Japanese children paused longer before saying each clause, taking more time to formulate what they were going to say. These findings differ from Minami's (1998) results, which showed frequent use of hedges in personal narratives, pointing to differences in narrative style according to genre. Many researchers have reported that avoidance of direct assertions through the use of hedges and hedge-like expressions is particularly important in Japanese society, where harmony and politeness are strongly valued: however, it is possible that there are contexts in which hedges occur relatively infrequently (e.g., Maynard 1990; Minami 1998).

351


Table 4. Percentage of causal connectors used by age (English data taken from Bamberg and DamradFrye 1991) 7 years Language 4 years 5 years 9 years Adults Japanese Turkish English

2.8% 0.5%

1.0% 1.3% 3.7%

3.9% 1.3%

2.0% 1.4% 4.7%

3.7% 3.0% 6.9%

One feature of the Japanese data which differed from both the English and Turkish data was the extensive use of onomatopoeia/mimesis. This is not surprising, given its frequent use in child-directed speech as well as in Japanese children's books and comics. The Turkish narratives differed from the Japanese narratives in their use of evaluative remarks, which were especially common in the adult narratives. The Japanese narrators did not use evaluative remarks at all. Another major difference between our results and those of Bamberg and Damrad-Frye was in the children's use of causative expressions. The frequency of use and the functions of causal connectives exhibit some interesting differences cross-linguistically. First of all, at all ages where we could carry out comparisons with the Bamberg and Damrad-Frye results, the percentage of clauses containing causal connectors is much higher for American narrators than for Japanese and Turkish narrators. Table 4 presents the percentages of clauses containing causal evaluative devices by age group and language. Looking at data from American narrators, many researchers have identified causal explanations as crucial to the narrative genre. As opposed to more dialogic genres, narratives entail a reflective stance on observed activity and therefore demand an emphasis on reasons (Kyratzis 1993). Narratives call for inferring causal or enabling relationships between story components such as physically observable events and characters' internal states. Inter-clausal connectors such as because, so, and in order to are used to explicitly encode the inferred relationship between successive story components. Causal connectives frame explanations and reasons for the unfolding events, reflecting the narrator's explicit stance regarding the causal antecedents of events or the motivations of narrative actions associated with story characters. Reference to inner experiences and intentions of story characters can be packaged with story actions either as causes, caused outcomes, or both. Bamberg and Damrad-Frye (1991) find a constellation of references to characters' mental and emotive states and causal connectives in the nine-year-olds' and adults' frog stories. The English-speaking children in their study tended to use causal connectors in clusters with "frame of mind" references (40% in nine-year-olds), as in the following example:

352


[40] and the dog was scared cause all the bees were coming to get the dog (five-year-old) [41 ] and then the boy was mad cause he knew that the frog went away (nine-year-old) McCabe and Peterson (1988), in studying American children's spontaneous uses of connectives, found that psychological relations between an intention and its motivated behaviour were marked with explicit connectives (because and so) more often than physical causality relations, which were mostly marked by the non-causal connective and. Bamberg and Damrad-Frye state that this pattern of integrating frames of mind with causal connectives may be part of a communicative convention in Western European languages: whenever you refer to some mental state of another person, you should give additional information in terms of situational knowledge as to why you attribute this mental state, justifying your perspective when inferences about others' mental/emotional attitudes are verbally expressed. The pattern of using clausal connectives in conjunction with frames of mind is relatively rare in the Japanese and Turkish narratives, which also had relatively few causal expressions (see Table 5). Bamberg and Damrad-Frye (1991) state that this might be due to the fact that inferences about others' mental/emotional states are impossible in Japanese, as they do not fall within the "speaker's territory of information" (Kamio 1979; Kuno 1987). However, looking at the examples, it seems that both Japanese and Turkish children and adults do make such inferences - it is just that the mental/emotional attribute is assumed/inferred, and becomes part of the shared knowledge between the speaker and the listener. This fits in with descriptions of Japanese culture and sociolinguistics emphasizing the need for omoiyari, or concern/consideration for the other, often requiring one to anticipate what the other feels or thinks, without an explicit verbal exchange (Clancy 1986; Doi 1971). Studies such as Marcus and Kitayama (1991) have reported that Japanese speakers tend to downplay personal agency, while Anglo-Americans tend to underscore it. Similarly, in a study on maternal discourse in mother-child play contexts, Shapiro and Fernald (1998) found that Anglo-American mothers tended to use more of what these researchers designate "agentic" language, emphasizing inter-personal exchange, whereas Japanese mothers tended to use more affirming language emphasizing shared experience. For example. AngloAmerican mothers were more likely to use discourse highlighting idea exchange, implying that each individual possesses ideas that others do not possess. On the


353

Table 5. Percentage of causal connectors used in conjunction with frames of mind by age Language Japanese Turkish English

4 years

5 years

7 years

9 years

Adults

21%

29%

19%

0%

15%

0

0

27%

25%

20%

40%

other hand, Japanese mothers were more likely to use language that emphasized mutual affirmation and shared understanding. Along such lines, Clancy (1986) described "empathy training" routines in which Japanese mothers encouraged sensitivity to the needs, wishes, and feelings of others by teaching their children how to anticipate what others are thinking and feeling. According to Minami and McCabe (1995:442), "Japanese adults believe that preschool children should be able to read the mind of others and count on others' filling out parts of stories that can readily be inferred". Such studies seem to point to the possibility that Japanese adults and children might view the information expressible by pairing clausal connectives with frames of mind as part of the shared knowledge of the narrator and the listener, and might therefore consider such pairing as unnecessary in narrative construction. Another possibility is that, since joint construction is a common strategy used when forming narratives in Japanese culture, this information would normally be filled in by the listener. As Ochs and Schieffelin (1984) argue, the discourse patterns used by people implicitly communicate values and beliefs that are prevalent within their cultural surroundings, such as individual autonomy and inter-personal relatedness. In a recent study of causal expressions in a narrative-retelling study, AksuKoc and Kuntay (2001) found that Turkish narrators at all ages (five, eight, and eleven plus adults) reformulate the original causal relations in the story read to them by omitting the psychological part of a cause-effect relationship, merely mentioning the action-oriented part. This finding is also in line with the pattern displayed by Turkish and Japanese narrators of the frog story - a pattern that can be characterized as having relatively less causal packaging of frames of mind into action sequences than that of the American narrators studied by Bamberg and Damrad-Frye. Some typical patterns of use of causal connectives in Turkish and Japanese are demonstrated in the following examples: [42] Ondan sonra agacin arkasinda kurbaga var mi diye bakiyorlardi. 'And then they were looking in order to see if the frog was behind the tree.' (Turkish four-year-old)

354


[43] Otoko no ko ga hachi ga kita kara hikkuri kaetchatta no. The boy flipped over because the bees came.' (Japanese seven-year-old)

[44] Sorede hachi ga oikaketekuru no de, hayaku hayaku hashirimashita. 'Because the bees came chasing, [the dog] ran quickly, quickly.' (Japanese sevenyear-old)

In the Japanese narratives, approximately 17% of overt causals have reference to a frame of mind. Causal expressions are often used to mark physical causality, or occur without a direct reference to frames of mind. This is what happens in the previous two examples, in which a causal expression is provided to explain the behaviour of the characters (flipping over, running). In both cases, the frames of mind of the characters are inferred. Causal connectives were used with frames of mind in all of the age groups except the nine-year-olds. However, compared with the rate of 40% of the American stories, the rates at which overt causals occur together with references to frames of mind are relatively low. In the Turkish stories, no causal connectives are used in conjunction with frames of mind at preschool ages. For the older age groups, on average 23% of the overt causals occur together with references to frames of mind, a rate far below that in the American stories. For the purpose of comparison, one may note that Trabasso and Rodkin (1994) find that 80% of the adult English-speaking narrators indicate sadness on the part of the boy upon discovering that the frog has disappeared. Out of fifteen adult Turkish narrators, on the other hand, only four included an emotional reaction on the part of the boy in the same scene. Two adults attributed sadness and one surprise to him, and one adult enacted character speech indicating sadness on the part of the boy. Among younger Turkish narrators as well, there was only marginal reference to a negative reaction upon the boy's discovery of the disappearance of the frog -just one narrator in each of the age groups expressed either surprise or sadness on the part of the boy. Similarly, of the sixteen adult Japanese narrators, only four described an emotional reaction on the part of the boy in this scene, two attributing surprise, one sadness, and one worry. Only one child in the four-year-old and five-year-old groups, and three of the sevenyear-olds, attributed an emotional reaction to the boy in this scene. It seems that Turkish and Japanese narrators do not feel as compelled as American narrators to talk about an intermediate emotional event that links the boy's noticing of the loss of the frog and the inception of the search. It might be possible to explain this difference in terms of cultural differences in shared knowledge and in what needs to be explicitly stated. However, this might also be due to a tendency for Turkishspeakers to apply a different type of evaluative positioning of the characters in relation to one another (Bamberg 1997b). Although 100% of the American nine-


355

year-olds and 90% of the adults tend to mark the relationship between the boy and the frog as some sort of pet ownership (Trabasso & Rodkin 1994), only 30% of the Turkish adults encode that relationship as a possession relationship. The type of conception of the initial relationship between the characters should have a consequence for the entire story structure, especially as regards its influence on the motivational structure for the story. David Wilkins (e-mail communication, September 13, 1999) has also pointed out that Arrernte narrators never see the frog as belonging to the boy.3 The stereotypical relation between small boys and small animals such as lizards and frogs in the Central Australian setting where Arrernte is spoken is that the boys capture the animals to harm them. The construal of such an antagonistic relationship between the boy and the frog will naturally keep the narrators from indicating disappointment on the part of the boy upon his separation from the frog. However, it is also interesting to point out the fact that the Japanese narrators did resemble the American pattern in that ten of the sixteen adult narrators explicitly described a pet-ownership type of relationship between the boy and the frog, while two narrators described a close friendship. In Japan, children frequently catch insects and reptiles and keep them as pets, and the catching and observation of insects and reptiles is a summer science project frequently assigned to elementary-school pupils. Therefore, the lack of causal expressions and emotional reactions in the Japanese narratives needs to be explained in terms of shared knowledge. From a cognitive-developmental point of view, Harris (1989) found that normal three- to four-year-olds understood that emotion can be caused by certain situations (e.g., nice situations make you feel happy, nasty ones make you feel sad) and desires (e.g., fulfilled desires make you feel happy, unfulfilled ones make you feel sad). They also found that by the age of four to six years, normal children understood that beliefs can affect emotion (e.g., if you think you're getting what you want, you'll feel happy). Another study with American pre-schoolers on the early acquisition of causal connectors has illustrated that children as young as three seem to feel an urge to tie references to psychological/mental states closely to what else they know about the surrounding events, especially to precipitating and consequent actions of the participants involved (McCabe & Peterson 1991). There is no reason to suspect that Turkish and Japanese narrators lack this basic cognitive capacity of inferring psychological states from certain situations, but this understanding does not seem to be as robustly expressed as part of their explicit interpretive stance as in the nine-year-old and adult Americans' narratives that Bamberg and Damrad-Frye (1991) studied. A recent study by Eaton, Collis, and Lewis (1999) indicated that specific prompting about story characters led English-speaking pre-school children to demonstrate their ability to offer coherent causal explanations for a story character's feelings. Thus, the cognitive skill to formulate evaluative causal explanations is separable from the

356


tendency to include these in narratives produced with no prompting. For Japanese and Turkish narrators, causal explanations regarding mental states do not seem to be among the dominant forms of narrative-evaluative devices, at least as far as the frog-story task is concerned. 7.2 Cultural considerations for eliciting narratives One thing that needs to be considered is the difficulty of the task for Japanese children, who are not used to creating monologic narratives. As reported in Minami and McCabe (1995), when telling personal narratives to their mothers, Japanese children receive constant verbal acknowledgement and back-channelling. During the data collection for our project, it was discovered that Japanese children (as well as some of the adults) had difficulty creating monologic narratives and frequently looked at the researcher for verbal encouragement and support. When the researcher failed to provide appropriate feedback, the subjects became even more hesitant, as if assuming that their narratives were incorrect or inappropriate, or that the researcher was bored or uninterested. This is understandable, as Japanese children rarely have opportunities to engage in monologues, since much of Japanese discourse (e.g., personal narratives and casual conversations) involves linguistic strategies such as back-channelling (e.g., to display attention and comprehension), which leads to frequent turn-taking and shorter individual turns. Therefore, Japanese children are used to narrative co-construction (i.e., with their mothers and other caregivers) rather than monologic narratives (Minami & McCabe 1995). Even when Japanese children are first introduced to the concept of monologic narratives in kindergarten, the narrator often receives prompting and questions from the audience, making the narrative interactive. Other researchers have also commented on the difficulty of monologic narratives, especially for younger children. For example, Sophie Kern (e-mail communication. September 10, 1999), who collected 140 frog stories from French children and adults, reported that sometimes her youngest narrators (three- to four-year-olds) expected the listener to help them, making the narrative interactive. Although Turkish children did not appear to have problems with delivering monologic narratives, many of them failed to mention some crucial elements of content. For example, conceiving the significance of the relationship between the boy and the pet animals in the initial scene is crucial for setting up a motivating background for the rest of the plot structure. The relative absence of encoding of the nature of the relationship between the boy and the pet animals in the Turkish stories, in comparison with American and Japanese stories, suggests the workings of different cultural frames for interpreting the same picture contents. Similarly, some of the Japanese children had difficulty identifying some of the animal characters, which may have influenced their use of evaluative language. Not


357

being familiar with the characters and their general personality would be expected to influence narrators' willingness to attribute emotions and mental states to the characters. David Wilkins (e-mail communication, September 13, 1999), based on his work with Arrernte narrators, commented that they tend to treat the characters as role types, as opposed to individual characters. For example, the characters would be generic as opposed to individuals with names and personalities. Of the 85 Japanese narrators, only six provide proper names for the boy. Most of the narrators refer to the boy as a "boy", "young man", or "person". However, many of the narrators personalized the animal characters by adding the polite -san suffix to their name, as in kaeru-san 'Mr Frog' orfukuroo-san 'Ms Owl', which is a common feature of child-directed speech that can also often be found in children's storybooks. Treating characters as distinct individuals as opposed to generic characters would lead to differences in the use of evaluative language. Furthermore, there are other cultural factors which may influence the use of evaluative language in narrators' renditions of the frog story. For example, Kazumi Takahashi (e-mail communication, October 7, 1999), in her research on Japanese adult frog stories, discovered that factors such as familiarity (i.e., being a personal acquaintance of the researcher as opposed to a complete stranger) influenced the evaluative content of the narratives. In Takahashi's study, the narrators recorded their stories on a tape recorder with nobody else present, but the narrators who were not personal acquaintances of the researcher did not use evaluative forms. A study currently being conducted by Nakamura has also found that familiarity influences the length and complexity of stories created by Japanese narrators, especially in younger age groups. In general, stories told to a researcher who was familiar with the children resulted in longer, more detailed stories, with richer evaluative content. 7.3 Directions for future research As mentioned previously, a study is currently being conducted to look at the use of evaluative devices from the perspective of language socialization. Children learn how to construct narratives from their caregivers and other adults. Preliminary data from Japanese mother-child pairs show that Japanese children pay close attention to their mothers' narrative styles, and that similarities emerge quite early in their use of evaluative language. One direction for future research involves the use of affective prosody as an evaluative device in the frog stories. Reilly (1992) and Reilly, Klima, and Bellugi (1990) point out that young children make extensive use of para-linguistic devices in their narratives. Future studies should include prosodic marking of affect and attitude in studies of evaluative language both in children and in caregivers.

358


Another direction for future research is to examine the use of evaluative strategies in conversational narratives. The kind of strategies that speakers use in expressing the evaluative function in narratives that have inherent tellability from their own perspective might plausibly be different from those exhibited in picturebook stories. Such cross-genre comparisons within languages will allow us to see a fuller picture of the communicative conventions of linguistic communities. The study of children's usage of evaluative devices is challenging and meaningful, as the construction of effective narratives is not simply a linguistic endeavour, but also requires sophisticated cognitive skills and social understanding. As illustrated in this study, we must also pay close attention to cultural and cross-linguistic issues, as narrators' choices from their repertoire of evaluative strategies are affected not only by the availability and cognitive ease of certain linguistic forms, but also by the prevalent communicative conventions of their culture.

NOTES 1.

An early version of this chapter was presented at the VI th International Congress for the Study of Child Language in Trieste, Italy, in July 1993 (Kiintay & Nakamura 1993). This research was supported by a Mellon Foundation Grant through the Center for Middle Eastern Studies to Aylin Kiintay and a Humanities Graduate Research Grant to Keiko Nakamura, both at the University of California, Berkeley. We thank the respondents of our Info-CHILDES e-mail survey and Judy Reilly for inspiring us to do this study through a workshop she conducted at the Institute of Human Development at the University of California, Berkeley. We are particularly indebted to Dan Slobin for his invaluable guidance and support.

2.

There seem to be two possible explanations for the poor performance of the the nineyear-olds (in terms of story length and number of evaluative devices): first, the nineyear-old sample consists of twelve boys and four girls (boys tend to produce shorter, less elaborate narratives than girls), and second, school-age children may become less engaged in the task, regarding it as a school task (something to be completed) rather than as a game (something to be creative and have fun with).

3.

In September 1999, we conducted a survey addressed to researchers who had previously collected frog stories in various languages around the world through the Info-CHILDES mailing list. The purpose of the survey was to compare experiences that these researchers had in collecting frog stories in different linguistic and cultural settings, and to examine whether the frog-story method elicits the same type of semantic and cultural content in different linguistic settings and cultures.

Writing the Frog Story1 Developmental and Cross-Modal Perspectives

Sven Stromqvist, Asa Nordqvist & Asa Wengelin

1. INTRODUCTION

This chapter presents analyses of written frog stories, both in terms of their linguistic information encoding (on a lexical, grammatical, and discourse level) and in terms of the flow of the discourse (production rate). Both developmental and cross-modal issues are addressed. A first set of analyses concerns the development of writing compared with that of speech. A second set of analyses is concerned with writing in congenitally deaf subjects. The flow of writing was analysed by means of a computer tool, "ScriptLog", designed specifically for research on the on-line process of writing. The chapter starts with a broader review of similarities and differences between spoken and written language, along with some comparative perspectives on sign language. It ends with a synthesis conclusion on similarities and differences between "thinking-for-speaking" and "thinking-for-writing".

2. SPEECH, SIGNING, AND WRITING 2.1 Towards a model Language appears in a variety of forms, including spoken, signed, and written forms. These forms of language are shaped and modified under the constraints and principles of human information processing (e.g., principles of clarity, speed, effort/economy, and expressiveness; see, e.g., Slobin 1977) to serve a variety of social and cognitive functions. Spoken and written language are used in different socio-cultural contexts and the two forms of language tend to be associated with different communicative conditions and processing constraints.

360


Spoken communication typically allows on-line feedback and mutual adaptation between speaker and addressee (see, e.g., Allwood et al. 1992; Clark 1996). In contrast, written language is typically used for communication where the sender and the receiver are located at different points in time. Further, spoken language is processed under strict on-line constraints: spoken discourse must be perceived on line just as it is produced on line (see, e.g., Stromqvist 1994; Gayraud 2000). In written communication, these on-line constraints can be lifted. A written document (such as a book or a letter) is typically produced on the assumption that it is only the final edited version which is to be perceived and understood by the addressee. In Table 1 we summarize what we take to be the most important components in a context of design features and processing constraints typical of spoken and written language, respectively. In Table 1, different types of communication ("communication technologies") in speech and writing are contrasted in terms of a number of analysis dimensions (italicized in the leftmost column of the table). The spatio-temporal configuration of the communicative setting, the duration of the signal, the use of different sensory modalities, and the relation between production and perception are some of these dimensions. Spoken face-to-face interaction and written communication in the form of a book or a letter are communication technologies which show pronounced differences with respect to all of these dimensions. These two types of technologies are located on opposite sides (left and right) in Table 1. In between the two are communication technologies showing more "mixed" characteristics. Sign language and spoken language show substantial similarities in terms of the dimensions in Table 1. Interestingly, use of the text telephone - a "mixed technology" where written language is employed for on-line communication - is very common in the community of sign-language users. Face-to-face interaction is the type of communicative situation in which a spoken or signed language is first learned by an individual, and it is in all likelihood the situation in which communication first evolved in the species (see, further, Hauser 1996). In this kind of situation, both speech and bodily communication can be used, and mutual on-line feedback and adaptation between the interactants are possible. Further, in this kind of communication, everything has to be performed at a high speed, and utterances have to be perceived on line just as they are produced on line. In contrast, these on-line constraints can be lifted in written communication (as illustrated by, e.g., a book or a letter); normally, it is only the final edited version of a text that you intend the addressee to read. On the other hand, written communication is a solitary form of communication in the sense that you do not get any on-line feedback from the addressee on how he or she perceives and understands your words - feedback which, in turn, may serve as a guideline and support for your further planning and wording of the discourse you are producing. Couched in partly different terms, on-line feedback

Writing the Frog Story

361

Table 1. Spatio-temporal configuration of communication setting: Consequences for speaking, signing, and writing (after Stromqvist, Ahlsen & Wengelin 1999) Spatio-temporal same time, configuration of same place communication setting:

same time, different places

different times, same plac;e

different times, different places

Communication e.g., spoken or signed face-totechnologv: face interaction

e.g., telephone; written on-line communication: text telephone, computer talk systems

e.g., writt en notes, e.g., written inscriptions, off-line dictaphor e communication: books, forms, e-mail

Duration of signal:

very short: very high speed of processing; short-distance editing

speed of processing may be relaxed; long-distance editing

Modalities:

predominantly multi-modal

mono-modal

Distribution of expressive features:

both simultaneous and linear

linear

Relation between perception and production:

discourse perceived on-line, just as it is produced on-line

Interactivity:

on-line feedback and mutual adaptation possible

Normative ideals:

\ /

, . . discourse perceived by receiver only after final editing

and mutual adaptation not possible

In a given linguistic community, speech, sign, and writing tend to be associated with partly different normative ideals, towards which the speakers, signers, and writers are orienting in the production process.

362


giving and second pair parts in spoken interaction serve, among other things, as evidence that a joint communicative action is completed, so that the next step in the communicative activity can be taken (cf. Clark 1996:222ff). One of the difficulties with composing a written text is precisely the absence of such evidence during the production process. Table 1 represents a first sketch towards a model of basic design features, processing constraints, and their inter-relationships in spoken, signed, and written communication. Roughly, the features at the top of the table are assumed to represent the most powerful determining dimensions, whereas dimensions further down can be interpreted more as consequences of the ones at the top. The model further serves as a point of departure for the analysis of how language users actually manage communication in speech and writing. In a classic, finalistic model of communication: We speak to be heard in order to be understood (see Ohman 1979; Jakobson et al. 1952). And, similarly: We sign to be seen in order to be understood and We write to be read in order to be understood. This particular way of formulating the functional raison d'etre for language puts special emphasis on perception as a determinant of production. The activities of speaking, signing, and writing involve the language user's planning, execution, and monitoring of linguistic utterances, while he is adapting to and taking into consideration relevant circumstances of uptake and processing on the part of the addressee, so as to secure a reasonable degree of understanding (see, further, e.g., Levelt 1989; Clark 1996). The circumstances of uptake and understanding are, as we have already argued, typically very different between spoken and written communication. In effect, the management of the highly complex, multiple-constraint-satisfaction process of language use typically looks different in speech and writing. For example, under the on-line constraints of spoken interaction, the length of silences within an utterance affects the speaker's ability to maintain the addressee's attention or to hold the turn (Sacks et al. 1974): long silences may cause the listener to lose attention or lead to the speaker's loss of the turn. In the management of spoken communication, speakers therefore learn to minimize the length of silent pauses when their turn is threatened, and they employ a number of other devices such as "hesitation markers" (e.g.. eh, uhm, you know, etc.), lengthening of continuants, repeats, or use of turnkeeping gestures. Also, self-repairs and revisions of something already said are typically associated with these kinds of behaviour. The hesitations, repeats, etc. are, however, not merely indicative of planning or monitoring activities in the speaker. The speaker's choice and timing of these devices will also affect the processability of the utterance for the listener by making it easier or more difficult for the listener to segment and meaningfully apprehend the discourse attended to (cf. Clark & Wasow 1998; Fox Tree & Clark 1999). Similarly, when a speaker interrupts himself in the middle of a word, he never goes straight on to complete


363

the word, but always reiterates the beginning of the word before completing it, thus preserving the word as an uninterrupted unit (lexical contiguity), something which makes it easier to process for the listener. The presence of intonational contours co-occurring with meaningful units in speech is further conducive to this preservation of units in the context of self-interruptions. In the management of written communication, by contrast, the location or duration of pauses is irrelevant to the processability of the final, edited product. Thus, the writer can allocate more resources to planning or monitoring activity and indulge in long pauses without risking any extra cost or trade-off effect in terms of increased processing effort on the part of the addressee. 2.2 The encoding of information and the distribution of expressive features Consider the dimension Distribution of expressive features in Table 1. First, let us briefly consider linear versus simultaneous ways of distributing information. Spoken face-to-face interaction allows communication in more than one sensory modality (spoken communication is usually multimodal', see, e.g., McGurk & MacDonald 1976; Andersson 1992). This makes possible a considerable extent of simultaneously distributed information, for example smiling and using manual gestures while talking (see, e.g., Argyle 1975). Also, there are simultaneously distributed dimensions of information in the vocal/auditory channel. At the same time as words are articulated, intonation gestures and voice quality contribute important information about, for example, information focus and attitudinal stances (see, e.g., Haberland 1986; Clark & Gerrig 1990; Home et al. 1999). Most probably, these ways of packaging information are related to the on-line constraints of speech. The pressure towards high speed in spoken communication makes the simultaneous distribution of information a more attractive solution to the packaging problem than a strictly linear distribution. Similar solutions can also be seen on the phonetic level of speech, where so-called co-articulation is a central principle (see, e.g., Ohman 1979). And if we turn to sign language, the simultaneous distribution of expressive features is even more exploited and conventionalized than in spoken language (see, e.g., Stromqvist 1994:221-223). In contrast, written language is much more confined to a linear distribution of information. Certain ways of distributing information simultaneously, however, such as bold face, italics, underscore, and the distinction between lower and upper case, are frequent in writing, at least in writing on a computer. Henceforth, we will be concerned mainly with patterns of encoding and distributing information beyond the phonetic or phonological level, that is, with ways of encoding information by means of lexicon, morphology, syntax, and discourse-level patterns. Consider Examples [1-2] from English and Russian (see Stromqvist 1997):

364

[1]

[2]


land on the moon (VP(VPP(PNP(DetN)))) npmryHHTbCfl

[pri-lun-it'-sja]

tO-mOOn-INF-REFL

Given that we want to encode information about the activity of landing and about a particular goal of the locomotion (the moon), English (Example [1]) offers a word, land, with no internal morphological structure, and the specification of the goal (the moon) has to be made by means of a prepositional phrase, on the moon. In contrast, Russian (Example [2]) offers a verb with an internal morphological structure, npHjiyHHTbCH [pri-lun-it'-sja] to-moon-iNF-REFL (by analogy with npH3CMJiHTbCfl [pri-zeml'-it'-sja] to-earth-iNF-REFL), which incorporates the goal. Thus, the same (or very similar) information about a motion event is encoded in different fashions in English and in Russian: in the former case it is distributed over a syntactic phrase, in the latter it is encoded into a morphologically complex verb. Now, consider Example [3]. [3]

... they entered an orbit around the moon ... they landed.

In the syntactic vicinity of landed in Example [3], there is no specification of the goal. But this does not mean that the information necessary to construe or interpret the motion event in question (the landing on the moon) is not accessible. It can be inferred from the information given earlier in the discourse (... they entered an orbit around the moon). There is an important difference, however, between encoding the information explicitly, as in Examples [1-2], and encoding it in such a way that it has to be inferred (Example [3]). Among other things, this leads to a difference in the foregrounding versus backgrounding of information. Assume that the linguistic encoding of information about a scenario necessarily means selecting certain aspects of that scenario while leaving other aspects out. What factors govern that selection? Surely relevance is one major factor and mutual understanding (the sender's assumptions about mutual understanding between himself and the addressee) is another (see. e.g., Grice 1975; Allwood 1986). Still another factor is the accessibility of linguistic forms. Thus, some concepts may be easier to encode into linguistic form than others. Also, the languages of the world differ in terms of the accessibility of forms tailored to talk about a given concept or conceptual domain (see, e.g.. Whorf 1956; Choi & Bowerman 1991; Berman & Slobin 1994). Following Berman and Slobin (1994), we will refer to this process as filtering. In particular, we will subject to contrastive analysis spoken and written discourse about very similar content domains and ask: What information filters through into the linguistic descriptions and what information is filtered out?


365

The detection of filtering effects is partly a function of the size of the linguistic units or the width of the discourse "window" observed/analysed. Thus, if we consider linguistic information encoding on the level of individual words, information about the goal in the encoding of the moon-landing above is filtered out in Example [1], whereas it filters through in Example [2]. In contrast, if the verb phrase is taken as the domain of information encoding, information about the goal filters through in Example [1], too. And if a larger fragment of discourse is taken as the domain, information about the goal filters through even in Example [3] - although, admittedly, in a highly backgrounded fashion. 2.3 The flow of discourse Psycholinguistic analyses of the speech-production process have explored the flow of speech as a source of information about planning, execution, and monitoring activity. The flow of speech contains a number of clues to the interplay between sub-processes in the form of, for example, pauses, stutterings and self-repairs (see, e.g., Goldman-Eisler 1968; Butterworth & Beattie 1978; Clark & Clark 1977; Levelt 1983; Levelt 1989). The flow of discourse is also sensitive to the type of discourse or communicative activity. Analysing the "flow of thought and the flow of language" in relation to spoken narrative discourse in English, Chafe (1979) finds the sentence or clause to be a unit which is more seldom broken up internally by pauses. Instead, pauses tend to come at the boundary between sentences or clauses. Chafe argues that the sentence/clause provides an adequate unit for encoding information about an event and that the flow of speech - reflecting the flow of thought - therefore tends to proceed sentence by sentence (or clause by clause) when a narrator is relating an event. And Beattie (1980), measuring the production rate in spoken interaction, found a cyclic arrangement of hesitant and fluent phases corresponding roughly to a turn in the conversation: at the beginning of a turn there tends to be a more hesitant phase, followed by a more fluent phase; then, again, at the beginning of the next turn there is a more hesitant phase, and so on. Further, the flow of discourse also seems to vary between different types of languages. In their pioneering analysis of American Sign Language (ASL), Klima and Bellugi (1979:181 ff) report a study of the production rate in three adult bilingual speakers/signers of American English/ASL. Among other things, the subjects were asked to relate a special episode they had experienced. The storytelling was performed under three conditions: spoken only, signed only, and simultaneously spoken and signed. The results from the spoken-only and signedonly conditions are summarized in the first two columns in Table 2. The third column summarizes results from a similar narrative task in writing with ten adult monolingual, normally hearing Swedish subjects from our own project archive

366


Table 2. Production rate in speech, signing, and writing, measured as words/signs and propositions per second Production rate words/signs per second (mean)

Speech

Signing

Writing

4.70

2.37

0.23

1.40 1.47 propositions per second (mean) 31.48 Data for speech and signing from Klima & Bellugi 1979; for writing from Stromqvist & Ahlsen 1999.

(see Stromqvist & Ahlsen 1999). We will first discuss the figures from Klima & Bellugi and then return to the figures from the writing task. Table 2 shows that, on average, the bilingual subjects produced nearly twice as many words (4.70) per second when speaking as they produced signs (2.37) when signing. But what about the amount of information expressed over a period of time? Is it twice as large in speech as in signing? Klima and Bellugi's study also included measurements of the pausing rate and the proposition rate2 (see Table 2). It turned out that the proposition rate was almost exactly the same for speech and signing (on average, 1.40 and 1.47 seconds per proposition, respectively). The absence of a difference in proposition rate between speech and signing - in spite of the difference found with respect to the production rate in terms of words and signs per second - calls for an explanation. A reasonable hypothesis is that signs and words are not equal from the point of view of linguistic information encoding. In particular, signs depend crucially on facial expressions and the spatial location of manual gestures for the composition of linguistic meaning, and sign languages exploit the simultaneous multi-dimensional encoding of linguistic information to a greater extent than do spoken languages. Therefore, signs per second is probably a measure which severely understates the actual production rate of signing. The difference in production rate between speaking and signing thus tends to level out as the window of analysis is broadened from words and signs to propositions (sentences/clauses). This conclusion parallels our earlier conclusion (in relation to linguistic information structuring) that the detection of filtering effects seems to be partly a function of the size of the linguistic units or the width of the discourse window observed/analysed (see Examples [1-3] and the discussion in Section 2.2). That is, the wider the analysis window, the greater the extent to which it will be found that similar information can be encoded in different languages. Put somewhat differently, local-level effects on linguistic information encoding or on the flow of language can be compensated for. and a more balanced picture emerges, at a higher discourse level. Taken together, these basic observations of linguistic information encoding and the flow of discourse serve as a very strong argument for studying language and linguistic behaviour in a discourse context.


367

In order to extend the comparative analysis of production rate to written language, we decided to subject written narratives from our own archive of computer-logged narratives (see Stromqvist & Ahlsen 1999) to Klima and Bellugi's measures. Our archive contains, among other things, stories from a composition task with a pre-set topic, "I was never so afraid". Narrative data (keyboard, computer) from ten adult control subjects (monolingual, Swedish, normally hearing subjects) performing this particular composition task were selected for the analysis. The results are shown in the third column of Table 2. As can be seen, the figures for writing are different from both speech and signing, and the differences are huge. First, 0.23 words per second means that it takes on average around 5 seconds to produce a word (including time for planning and monitoring), and second, the proposition rate of 31.48 propositions per second means that it takes a little more than half a minute to produce a proposition in writing. What is the explanation for these differences between speaking and signing on the one hand and writing on the other? Writing on a computer (as well as handwriting) is associated with more effort and is more time-consuming than speaking or signing. But again, although effort most probably explains some of the difference observed, it is unlikely that it explains the major part of it. We propose that the constraints of on-line interaction are the main reason for the high rate of production in speech and signing, and that the much lower rate observed in writing can be afforded when these constraints are lifted. 2.4 Thinking-for-speaking and thinking-for-writing In short, then, speaking and writing are complex, multiple-constraintsatisfaction processes, and the sub-tasks you have to carry out in order successfully to manage communication are not accomplished with equal effort or ease in speech and in writing. Some of them are easier in spoken communication, others are easier in written communication. The allocation of cognitive resources to the different things you do when you speak or write is also dependent on your degree of proficiency. For example, it is much easier to free resources for higher-level planning and monitoring in the construction of written texts if you have automatized lower-level actions (such as spelling and the construction of words). Indeed, attaining this kind of automatization and increasing your control over the construction of discourse are important aspects of learning to write. But since the acquisition of written language takes place in a developmental context, where the learner has already attained substantial knowledge and skills in spoken language, the acquisition of writing interacts with the development of spoken language; the early acquisition and further development of written language will be influenced by spoken language, and the acquisition of written language will come to influence the further development of spoken language (cf. Ong 1982).

368


We believe that there is a causal inter-relationship between communicative conditions, information-processing constraints, and linguistic ways of encoding and structuring information. In the same vein, we hypothesize that the differences in communicative conditions and processing constraints between speech and writing are conducive to a differentiation in terms of linguistic information structuring between spoken and written language. The differences are not merely stylistic differences (see, e.g., Enkvist et al. 1964) in the sense of variations in terms of intimacy, formality, and the like, but, importantly, differences in terms of lexical and prepositional content. For example, Ure (1971) found spoken discourse to be characterized by a lower "lexical density" (i.e., a lower percentage of content words per total number of word tokens) than written discourse. Similarly, Allwood (1996) found that the 50 most frequent words (all function words) made up a greater proportion of the total number of words in a corpus of spoken language than in a corpus of written language. And Chafe (1982) describes written language as characterized by a higher degree of "integration", that is "packing [...] more information into an idea unit than the rapid pace of spoken language would normally allow". In a narrative experiment with 30 fifteen-yearolds, Stromqvist (1998) studied how a non-verbally presented motion episode was described in speech and in writing. Each subject was asked to describe the event once in a spoken and once in a written condition, and the order of the two conditions was controlled for. Among other things, it was found that the encoding of causal relationships (by means of conjunctions and causative verbs) was much more frequent in the written than in the spoken condition. These findings suggest that the decision whether to speak or write about a given topic may have an impact on the conceptual structure of the discourse produced. In the past few years, the study of the inter-relatedness of language, culture, and cognition has been revitalized by revived and revised approaches to linguistic relativity (see, e.g., Lee 1996; Gumperz & Levinson 1996). In that context. Slobin (1996b) presents a theory of thinking-for-speaking. Slobin suggests that already when we are planning the conceptual structure of what we are going to say. we are taking into consideration the constraints imposed on information encoding by the particular language we are going to speak. Since the languages of the world differ in terms of grammatical constraints (what must be encoded) and lexical constraints (for example, the accessibility of lexemes tailored to talk about a given concept or conceptual domain), this thinking-for-speaking varies across the languages of the world. Using the conceptual framework suggested by Slobin (1996b). we propose that thinking-for-\\riting may differ from thinking-for-speaking. The contrastive study of speech and writing suffers from both conceptually and methodologically determined biases. In his detailed criticism of the "writtenlanguage bias" in traditional, mainly Chomskyan, linguistics. Linell (1982) points out that concepts and models tailored to the lexical and grammatical elements


369

and structures of ready-made written texts are of limited use for the analysis of the dynamic properties of speech and spoken interaction. Focusing on the linguistic production process, we suggest that, conversely, some of the tools and perspectives tailored to the process of studying the production of spoken discourse may in fact be useful for exploring more carefully the on-line process of producing written texts. The controversies may partly be due to methodological shortcomings. The availability of high-quality devices for the on-line recording of speech has made possible a close examination of the dynamic properties of spoken language. In contrast, the production process as evidenced in writing is still poorly examined on the level of on-line events (see, however, Eklundh & Kollberg 1995; Kollberg 1998; Stromqvist 1996; Stromqvist 1998). The absence of devices for the on-line recording of written language, and the fact that a written text, as a rule, is encoded and edited in its entirety by a sender before it is decoded by an addressee, have conspired to background almost completely the dynamic properties of the genesis of written texts. By implication, we believe that the coupling of on-line studies of linguistic behaviour and the flow of discourse in both speech and writing with analyses of linguistic information encoding will pave the way for a richer and more fruitful scientific investigation of the production, perception/understanding, and acquisition of language. In the following sections of this chapter, we present two sets of studies which apply this approach to aspects of the writing of frog stories. The first set is concerned with developmental issues and with frog stories written by school children and adult subjects. The second one contains analyses of frog stories written by congenitally deaf subjects.

3. DATA AND METHOD The data relate to altogether five groups of subjects: a group of 15 nine-yearolds (writing with paper and pencil), a group of 16 nine- to twelve-year-olds (writing on a computer), a group of 30 fifteen-year-olds (computer), and a group of 20 adults (computer) - all monolingual, Swedish, normally hearing subjects plus a group of 10 adult congenitally deaf writers of Swedish (computer), all having Swedish Sign Language as their first language. All subjects were required to have at least one year's experience with writing on a computer prior to recruitment for the experiments. For the primary-school pupils, we were not able to find a homogeneous group of nine-year-olds meeting that requirement. Our compromise solution was therefore to collect data from two groups, with the computer criterion relaxed in one and the age criterion in the other. The computer-written narratives were logged by means of a computer tool,

370


Figure I. On-line playback frozen after about 25 seconds from the start of the writing activity in a nine-vear old.

"ScriptLog", which we have developed specifically for the purpose of studying the on-line process of writing (Stromqvist & Malmsten 1998; Stromqvist & Karlsson 2002)3. ScriptLog keeps a record of all events on the keyboard, the screen position of these events and their temporal distribution. From a ScriptLog record, then, you can derive not only the finally edited text from a writing session, but also the "linear" text with its temporal patterning, pauses, and editing operations. You can also play back a recording of a writing activity. ScriptLog allows you to design writing experiments with elicitation stimuli: in the studies reported here, the frog story (Mayer 1969) was administered as such an elicitation stimulus (see Appendix I). More precisely, the subject uses the mouse to activate a new picture in a stimulus window in the left half of the screen and proceeds to write about that picture in an editor window in the right half of the screen. Figure 1 provides an illustration from the playback of a recording of a written frog story.


371

Table 3. Task design Task order speech followed by writing writing followed by speech

Studied condition

Code*

speech

sp_wr_SP

writing

sp_wr_WR

speech

wr sp SP

writing wr sp WR * Henceforth, the different conditions are referred to by the codes given in this column.

In order to enable systematic comparisons between speech and writing, each subject was asked to perform the frog-story task twice: once in a spoken and once in a written condition. The two conditions were administered in sequence with a short interval of 10 to 15 minutes in between. Every second subject performed first in the spoken and then in the written condition and every second subject in the reverse order. The design is summarized in Table 3. The second row of Table 3 represents the pre-eminently spoken condition or "control" condition of speaking, in that the subject performs in speech without being influenced by a preceding performance in writing. Similarly, the bottom row represents the pre-eminently written condition or control condition of writing. In contrast, the rows between them correspond to experimental conditions. The setup represented by the code sp_wr_WR allows us to explore the question of how writing might be influenced by speech, and that of the code wr_sp_SP how speech might be influenced by writing. For the rest of this chapter, the different conditions will be referred to by the codes given in the last column of Table 3. All subjects were encouraged to browse through the picture booklet in advance of the first condition, so as to minimize differences in familiarity with the content of the frog story between the two conditions. The subjects all received the instruction to tell "what is happening in the pictures".

4. THE DEVELOPMENT OF WRITING 4.1 Production rate In order to get an overall profile of the distribution of writing activity across the frog story, keystrokes and pauses were measured (Stromqvist 1998). Figure 2 shows how many keystrokes the subjects spent on each picture from 1 to 24. It consists of three diagrams, one for each age group. The upper, dashed graph in each diagram shows the number of keystrokes in the linear text, that is, the total number of keystrokes made. The lower, solid graph shows the number of keystrokes remaining in the edited version of the text. The distance between the

372

Sven Stromqvist. Asa Nordqvist & Asa Wengelin

graphs, then, indicates the amount of editing made in relation to a picture. The values in Figure 2a-c are means across all subjects for each of Pictures 1-24. From the production profiles in Figure 2a-c, it is clear that, on a group level, the youngest age group - here represented by the 16 nine- to twelve-year-olds writing on a computer - write the least, thefifteen-year-oldsa little more, and the adults the most. Further, the profiles indicate differences between the groups in terms of the distribution of writing activity across discourse. The nine- to twelveyear-olds write a great deal in the beginning, then less and less, and they have a very meagre finish. It is reasonable to interpret this profile as indicative of effort and exhaustion. In contrast, the fifteen-year-olds show a much more smooth and balanced profile. Along the same line of interpretation, they seem to be able better to plan their discourse and to control their expenditure of effort. They also spend a little more on the finish. The adults, finally, describe a very dynamic overall profile: they spend a great deal of their writing activity at the beginning and end of the story, and they vary in between according to the richness and ramifications of the individual pictures. In terms of editing activity, the amount of editing tends to decrease towards the last third of the story for all three age groups. The last picture/episode constitutes an exception in the two older age groups: while the fifteen-year-olds and adults perform a relatively large amount of editing in relation to the last picture, the nine- to twelve-year-olds perform very little. In terms of relative amounts, the nine- to twelve-year-olds edit 23.3% of the keystrokes they spend on the initial picture of the story-elicitation instrument, but only 10.1% of those spent on the final picture. The great majority of the editings concern spelling mistakes, and many spelling errors remain in the edited text. For an illustration, see Example [4] below. The fifteen-year-olds have a higher and more balanced proportion of editings: 30.6% for the initial and 26.6% for the final picture. Again, the great majority of the editings concern spelling mistakes, but here very few errors remain in the edited text. The adults show a perfectly balanced editing rate, 15.5% for the initial and 15.6% for the final picture. The great majority of their editings concern content structure, and their texts are almost completely error-free from the point of view of spelling (for an illustration, see Example [5] below). Thus, both the amount and the nature of the editing operations are subject to developmental change. Example [4] provides an illustration of an elaborate story opening by a nine-year-old in relation to the initial picture of the frog story (whitespace added for clarity; no translation given owing to the fragmentation of the output). The example shows a detailed ScriptLog analysis of the story opening, w ith indications of pauses and editing operations. Pauses (according to the two-second pause criterion) are enclosed in angle brackets and their length is given in seconds, for example . Instances of editing operations are enclosed in square brackets.


373

Figure 2a. Production profiles for 9-12-year-olds in terms of keystrokes for the 24 different pictures of the frog story.

Production profiles for 15-year-olds in terms of keystrokes for the 24 different pictures of the frog story.

Figure 2c. Production profiles for adults in terms of keystrokes for the 24 different pictures of the frog storv.

374


The brackets are indexed and the italicized string contained within the brackets is the string deleted in favour of what stands immediately to the right of the brackets; for example, hab[ b]n illustrates how the writer produces hab. then replaces the b with an n (resulting in han 'he'). Example [4] is filled to the brim with pauses, testifying to the effortful writing of the young subject in question. Out of a total of 15 editing operations. 13 are concerned with spelling errors, and only 2 with content structure. The first content editing (indexed 3) occurs after the very first word det 'once' (upon a time ...) - the writer decides that he wants a title for the story, and he writes it in upper-case letters (POJKEN OCH GRODAN 'THE BOY AND THE FROG'). In the second content editing (indexed 14), the writer changes his mind about what to call the frog (from Inetz to Urban}. By contrast, Example [5] provides an illustration of a story closing by an adult subject. The original ScriptLog record in Swedish, consisting of pauses and keystrokes, is reproduced in the two leftmost columns, and an English translation is given in the column to the right. Both [4] and [5] come from subjects who wrote their stories before they told them in speech. The more selective pausing in Example [5] indicates that the adult subject is in command of the low-level actions of writing and can concentrate on more sophisticated lexical choices and on the planning and monitoring of larger chunks of discourse. Out of a total of six editing operations, only one is concerned with a spelling mistake, the remaining five with content structure. Note, in particular, the way the subject rethinks the relationship between the two propositions a] 'Now they would never part again' and b] 'Before they fell asleep this night the boy told the frog that the dreamland is not outside the bedroom window but it comes to all those who are sleeping' (to save space, we here use the English translation): The subject first encodes a as a simple declarative sentence terminated with a period, then takes away the period and continues with 'and' in order to produce a conjoined sentence structure a 'and' b . Having written b conjoined with a. however, the subject changes her mind anew and puts an exclamation mark w here she originally put the period (which she then deleted). And, as a consequential adjustment, she proceeds to change the initial 'a' of 'and' from lower-case to upper-case. The ScriptLog record of the later part of this complex editing activity is rendered with the notation [ . [ ] ] < 3 . 1 8 > ! < 5 . 4 2 > < R I G H T > [. a; A (where indicates a cursor movement). Among those subjects who started in the spoken condition, there were some lexical editings in the spoken storytellings. but no editings even remotely similar to the one illustrated in Example [5]. Figure 3a-c shows the distribution of pauses (longer than 2 seconds) across the 24 story pictures. The upper, dashed graph shows the total amount of pausing, and the lower, solid graph shows the amount of picture-initial pausing (the pause between the activation of a new picture and the first keystroke made in relation

375


[41 ~4.08,

15.31> 13.7> c3.91> -1

det

OCH GRODAN D et <Space> var en P oj ck L 4 ckl k ee[: el s om tykt e <Space><Space> om att <Space> fis ka en dag nar hab[, bln va

r [,

.I

vid d amen <Space 1 drog han me d haven 1 v attnet <Space> och fick P[,PlUP P en groda

P

<Space> hade

ald

~ 2 . 2 8>

rig set [9

1

t en groda <Space><Space> S5

han to 9 med sig qrodan hen i en glas <Space> [ l i l ml m <space> bur k Nar

P 01 ken kom hen Lll nlm tit t ade han noqa dan och a[,&alsa

Du ska het a inet

ur ban L15.

J

P6

gro

Sven Stromqvist, Asa Nordqvist & h a Wengelin

376

Pojken lyfte forsiktigt upp sin efterldngtade

The boy gently lifted .LF his wished-for

[ efterlangtade]saknade van

[

och vinkade at alla and waved to all the skogens djur som pa ett sa animals of the wocd wh.c underligt such a strange

och markligt satt hade h jalpt hom [ . m] nom

wis.?ed-fzr! rr.issed f r i e 7 . S

. : 1

and peculiar way :?ad helped hin [ .?j I

fram till

all the way to

Kvac k .

Croak.

Nu skulle de aldrig skiljas at mer

Now they wsc;ci re.ier ~ s r r again

[,

. I och

innan de

[

.:

and

before they

somnade den har kvallen talade pojken om for

fell asleep this r.igk: boy told

grodan att

the frog that

dromlandet finns inte utanfor

the dreamland is r.c: outside

sovrumsfonstret utan det kommer till alla som sover.

ir the bedroorr. window b:;. comes to ail tr.ose w t c a r e sleeping.

" 1 1

[.

: :

[,

blbarked

;

[ : 01 O

Voffs h

[-

instammande och viftade glatt pa svansen

klvoffade

in consen: ana r r e r r - l y wagged his tall

:?.e


377

to that picture). Figure 3a-c shows that the nine- to twelve-year-olds have the largest and the adults the smallest proportion of pauses in relation to the amount of writing (as measured by number of keystrokes). Further, the nine- to twelveyear-olds make a relatively short planning pause before they start writing in relation to the first picture of the frog story (on average, less than half a minute), the fifteen-year-olds make longer discourse-intitial planning pauses (around one minute), and the adults make very long planning pauses (on average, more than two minutes). 4.2 Information structuring Let us now turn to the linguistic encoding of information in the spoken and written frog stories. First, we will take a closer look at collocations of word forms in the four conditions described in Table 3 (sp_wr_SP; sp_wr_WR; wr_sp_SP; wr_sp_WR). Second, we will compare strategies for projecting speech onto story characters in spoken and written frog stories. Already from calculations of word frequencies, characteristic differences between spoken and written frog stories can be detected. For example, the general conjunction och 'and' is even more frequent in speech than in writing, whereas aft (ambiguous between subordinator 'that' and infinitive marker) is more frequent in writing than in speech (cf. Johansson 1999). Similarly, frequency distributions of collocations of word forms offer a bird's-eye perspective on similarities and differences between spoken and written frog stories (Stromqvist 1996). Table 4a-c shows the top ten pairs of word forms in the four conditions (sp_wr_SP; sp_wr_WR; wr_sp_SP; wr_sp_WR) for the three age groups. The youngest age group is here represented by the 15 nine-year-olds writing with paper and pencil. Table 4a-c shows that, in the nine-year-olds, collocations with the general conjunction och 'and' dominate in all four conditions. A subset of these collocations is made up by composite sentence connectors (och sa lit. 'and so', och se(da)n 'and then', se(da)n sa lit. 'then so' - all three meaning 'and then') characteristic of casual spoken narrations (Stromqvist & Day 1993). A budding differentiation between speech and writing can be traced insofar that these composite connectors are held back in the written narrations. In the fifteen-yearolds, a more powerful differentiation emerges. Collocations with 'and', including the composite connectors, are at the top in the pre-eminently spoken condition sp_wr_SP, as well as in the condition wr_sp_SP. In contrast, the pre-eminently written condition wr_sp_WR is characterized by a causal/final connector, for att 'because/in order to', as its most frequent collocation (rank 1). The same causal/ final connector is encountered in the other writing condition sp_wr_WR with rank 10, whereas it is as low as rank 272 in the pre-eminently spoken condition

378


Figure 3a. Pause profiles for 9-12-year-olds for the 24 different pictures of the frog ston:

Figure 3b. Pause profiles for 15-year-olds for the 24 different pictures of the frog ston:

Figure 3c. Pause profiles for adults for the 24 different pictures of the frog ston:


379

sp_wr SP. In the condition wr_sp_SP, however, for att is considerably more frequent, holding rank 4. On the level of connectors, then, the fifteen-year-olds encode causal relationships much more frequently in the pre-eminently written condition (writing unpreceded by speech) than in the pre-eminently spoken one (speech unpreceded by writing), but when they speak after having written the story, the encoding strategy typical of writing spills over to the spoken narration. In the adult group, finally, for att is at the top of all four conditions, closely followed by collocations with och 'and'. In the adult narrations, the composite connectors typical of casual speech do not show up among the top ten collocations. If we try to integrate the observations made so far under a developmental interpretation, and extrapolate from the three age groups, a scenario of three stages emerges. At an early stage of writing development, evidenced in the nine-yearolds, the subjects are still heavily influenced by their spoken-language practices when they are writing the frog story. Similarly to oral storytelling, they engage in a minimum of planning before starting the story, and when writing it they rely largely on the general conjunction 'and' to connect sentences. The further qualification of relations between events, for example causal relations, is to a considerable extent left implicit. In several respects, the nine-year-olds seem to be thinking-for-speaking, not only when they are indeed speaking, but also when they are writing. The lexical or grammatical modifications typical of their written narrations consist in the avoidance of certain connectors characteristic of speech rather than in the addition of connectors typical of written language. Further, low-level actions are associated with considerable effort: the nine-year-olds are struggling with spelling and with finding the keys on the keyboard. This results in a generally slow production rate and in an exhaustion effect with a pronounced peak of writing activity in an early phase of the story task, rapidly decreasing to a minimum during the latter half of the narration. Thus, effortfulness associated with low-level actions of writing has an effect on overall content structuring. At a later stage, evidenced in the fifteen-year-olds, there are several signs of a more advanced adaptation to written language. First, the general production rate has increased, indicating increased control of the keyboard and of spelling. Second, the writing activity is more evenly distributed across the story task, with peaks both at the beginning and at the end, indicating an ability on the part of the writer to economize on his resources and to spend some of them on bringing the story to a coherent closure. Also, the fifteen-year-olds make greater use of the relaxation of the on-line constraints in the written condition, engaging in longer discourse-initial planning pauses. As to linguistic information encoding, they show a clear preference for inter-sentential och 'and' in speaking, whereas in writing they more often encode causal relationships explicitly. The increased use of causal/final/or att 'because/in order to'to accomplish inter-clausal connectivity in writing indicates that the fifteen-year-olds often operate with larger units

380


Table 4a. Top ten word pairs in spoken and written frog stories. Nine-year-olds (sp_wr = 7; wr_sp = 8) Rank 1 2 3 4 5 6 7 8 9 10

sp wr WR

word pair

tokens

word pair

wr sp SP tokens word pair

och sa och hunden voff voff och en pojken och och sedan och pojken ardet sedan sa sa ar

47 41 28 20 17 17 15 15 12 10

och hunden och pojken vov vov i ett del var pojken och pa grodan ardu upp pa ramlade ner

26 20 17 11 10 9 8 8 7 7

sp_wr_SP

och hunden och sedan sedan sa del var och sa ramlade ner och pojken och de en groda ner och

wr sp WR

tokens word pair 26 21 19 17 16 14 12 11 11 10

och hunden och ropade och pojken en eroda i ett det var pojken och i en ett hal var en

tokens 18 12 11 10 9 9 8 8 8 7

Table 4b. Top ten word pairs in spoken and written frog stories. Fifteen-year-olds (sp_wr = 15: wr_sp = 1 5 ) Rank 1 2 3 4 5 6 7 8 9 10

sp wr_WR

sp_wr_SP


word pair

tokens

word pair

och hunden och sa och han i ett det var i en detar efter grodan pa en ner i

56 53 38 26 23 21 18 17 17 16

efter grodan 25 det var 24 iett 21 pa en 17 i en 15 pa grodan 15 och tittar 14 att grodan 13 en groda 13 for att 13

och sa och hunden och da for att en liten det var sa han efter grodan i ett na'r han

wr sp WR

tokens word pair 57 38 36 32 26 25 24 23 21 20

for att det var att han och hunden efter arodan pa en pa grodan att leta en stor och ropade

tokens 48 31 21 20 19 17 17 16 16 16

Table 4c. Top ten word pairs in spoken and written frog stories. Adults (sp_wr = 10: wr_sp = 1 0 ) Rank sp_wr_SP word pair

sp wr WR tokens word pair

1 2 3 4 5 6 7 8 9 10

30 30 23 23 20 20 20 18 17 17

for att och hunden sa han en liten pojken och i en den lilla och eh ner i efter grodan

far att och hunden pojken och det var efter grodan i en att han upp pa och pojken i marken


wr sp_WR tokens word pair

61 43 35 31 27 21 19 18 16 14

49 42 28 24 21 21 19 19 18 16

for att och eh och hunden han hade pojken och i ett sa han efter grodan som han i en

for att och hunden pojken och efter arodan i ett aenom fonstret det var att han nar han i en

tokens 51 31 29

2~> 17 17 16 16 15 15

381


Table 4a (translation). Top ten word pairs in spoken and written frog stories. Nine-year-olds (sp_wr = 7; wr_sp = 8) Rank sp_wr_SP word pair

1 -> 3 4 5 6 7 8 9 10

and so and the dog bow wow and a the boy and and then and the boy is it then so so is

sp_wr_WR

wr_sp_SP

wr_sp_WR

tokens word pair

tokens

word pair

47 41 28 20 17 17 15 15 12 10

26 20 17 11 10 9 8 8 7 7

and the dog 26 and then 21 then so 19 it was 17 and so 16 fell down 14 and the boy 12 and they 11 a frog 11 down and 10

and the dog and the boy bow wow in a it was the boy and at the frog are you up on fell down

tokens word pair

and the dog and called and the boy a frog in a it was the boy and in a a hole was a

tokens

18 12 11 10 9 9 8 8 8 7

Table 4b (translation). Top ten word pairs in spoken and written frog stories. Fifteen-year-olds (sp_wr : 15; wr_sp= 15)

wr sp WR

Rank sp wr SP word pair

tokens word pair

tokens word pair

tokens word pair

tokens

1 2 3 4 5 6 7 8 9 10

56 53 38 26 23 21 18 17 17 16

25 24 21 17 15 15 14 13 13 13

57 38 36 32 26 25 24 23 21 20

48 31 21 20 19 17 17 16 16 16

and the dog and so and he in a it was in a it is after the frog on a down in

sp_wr_WR

after the frog it was in a on a in a at the frog and is looking that the frog a frog because/since

wr_sp_SP

and so and the dog and then because/since a small it was so he after the frog in a when he

because/since it was that he and the dog after the frog on a on the frog to search a big and called

Table 4c (translation). Top ten word pairs in spoken and written frog stories. Adults (sp_wr = 10; wr_sp = 10) Rank sp_wr_SP word pair

1 2

3 4 5 6 7 8 9 10

because/since and the dog so he a small the boy and in a the small and eh down in after the frog

sp_wr_WR

wr_sp_WR

tokens word pair

wr_sp_SP tokens word pair

tokens word pair

tokens

30 30 23 23 20 20 20 18 17 17

61 43 35 31 27 21 19 18 16 14

49 42 28 24 21 21 19 19 18 16

51 31 29 22 17 17 16 16 15 15

because/since and the dog the boy and it was after the frog in a that he up on and the boy in the ground

because/since and eh and the dog he had the boy and in a so he after the frog which he in a

because/since and the dog the boy and after the frog in a through the window it was that he when he in a

382


of discourse in their writing than in their speaking, where simple additive och 'and' dominates. They thus seem to have begun to adapt to written language also in terms of thinking-for-writing versus thinking-for-speaking. Indeed, certain aspects of thinking-for-writing may even be a cognitive priority at this stage, since items typical of writing carry over as memory traces from a previous written narration to a subsequent spoken one (but not from speech to writing). At a still later stage, evidenced in the adult group, the adaptation to the processing constraints of written language has proceeded even further. The adults have great control of low-level aspects of writing. They engage in very long discourse-initial planning pauses and then distribute their writing across discourse in a very dynamic fashion with main peaks at the beginning and end. In terms of inter-clausal connectors, the causal/final for att 'because/in order to' has now taken precedence in all four conditions and thus no longer distinguishes written narrations from spoken ones. On the level of language acquisition, the situation resembles the one described by Ong (1982) with respect to societal level and historical change: once written language has emerged in a society, it impregnates spoken language so thoroughly that genuine spoken language can no longer be found in that society. Further, on the level of editing activity, the adults engage in editing of content structure to a far greater extent than the younger age groups. The adults, thus, not only take their time to think (plan) before they write, they also take the privilege of rethinking and changing what they have written. Let us now turn to the functional domain of projected speech - that is, spoken utterances and vocalizations projected onto story characters. Nordqvist (1998, 2001) distinguishes three main types of strategies for reporting or projecting speech: 'indirect speech' with the form "X said that "; 'direct speech' with the form " , X said", and 'free direct speech' with the form " !". Nordqvist investigated the distribution of these three types across spoken and written narratives in the three hearing age groups of the Swedish corpus. Prior to her investigation, Nordqvist balanced the groups for number of subjects, picking 15 subjects from each group. The results are summarized in Table 5. Table 5 shows that direct and free direct speech are by far the most frequently used categories among the nine-year-olds (both in terms of the number of subjects/users and in terms of tokens), in speech as well as in writing. Indirect speech is used by only two subjects and by them only once each. Further, the nine-year-olds show a clear preference for letting the main character of the story, the boy, say things, but they also project speech or animal cries onto other story characters. In their spoken performance, their use of free direct speech is accompanied by changes in voice quality, distinguishing the character's voice from the narrator's voice, and shifts to different story characters are accompanied by correspondingly differentiated shifts in voice quality. Similar shifts in gestural activity and facial expression tend to be part and parcel of these role switches in the narrations. These para-linguistic and non-verbal features efficiently indicate


383

Table 5a. Distribution of indirect and direct speech in spoken and written narratives. Nine-yearolds, paper and pencil (15 subjects) Direct speech Indirect speech Free direct speech " I" "X said that " , X said" boy dog other boy dog other boy dog other Speaking N subjects N tokens Speech-act verbs Writing N subj N tokens Speech-act verbs

2 2

0 0

0 0

10 17

1 1

1 1

7 55

2 8

2 11

9 23

3 5

1 2

7 57

3 9

3 5

3 2 2

0 0

0 0

3

Table 5b. Distribution of indirect and direct speech in spoken and written narratives. Fifteen-yearolds, computer (15 subjects) Indirect speech Direct speech Free direct speech " i" "X said that " , X said" boy dog other boy dog other boy dog other Speaking N subjects N tokens Speech-act verbs Writing N subj N tokens Speech-act verbs

8 9

0 0

1 1

4 7

0 0

0 0

1 5

0 0

0 0

5 5

0 0

0 0

0 0

0 0

0 0

3 4 4

0 0

0 0

2

Table 5c. Distribution of indirect and direct speech in spoken and written narratives. Adult controls, computer (15 subjects) Direct speech Indirect speech Free direct speech " i" " , X said" "X said that " dog other boy dog other boy dog other boy Speaking N subjects N tokens Speech-act verbs Writing N subj N tokens Speech-act verbs

3 3

0 0

0 0

9 28

0 0

5 6

3 4

0 0

2 2

9 44

4 14

6 23

3 11

1 7

1 1

8 0 0

0 0

1 1

33

384


the identity of the story character in question as well as aspects of the emotional and attitudinal stances of these characters. In the written narrations, by contrast, these cues are totally lost, and the many instances of free direct speech in writing are much more ambiguous in terms of speaker identity and stances. Again, the nine-year-olds have not yet adapted to the expressive constraints of writing: the simultaneous multi-dimensional distribution of expressive features in speech (verbal, voice, gesture) is not transformed into a linear distribution in writing; and some aspects of the rich information structure in the spoken version is lost in the written versions (cf. Michaels & Collins 1984). More advanced lexical choices, for example speech-act verbs such as 'jeer' or 'chuckle', would convey some of the emotional and attitudinal information carried by for example voice quality, but the nine-year-olds do not seem to have access to these lexical devices. There are only three speech-act verbs evidenced in the youngest age group: saga 'say', skrika 'shout', and ropa 'call'. In short: when they are writing, the nineyear-olds seem, to a large extent, to be thinking-for-speaking (cf. also Hildyard &Hidi 1985). In the group of fifteen-year-olds, the use of direct speech has decreased substantially and free direct speech is almost completely absent. The most frequently employed strategy in this group is indirect speech, and no other age group uses it nearly as much as the fifteen-year-olds (9 of the 15 fifteen-yearolds use it, compared with 2 out of 15 nine-year-olds and 3 out of 15 adults). Indirect speech is a strategy where non-verbal and para-linguistic features are marginalized and the speech reported/projected (contained in the embedded clause) must have a prepositional form, and with the exception of a single instance the fifteen-year-olds project speech only onto the main character (the boy). Further, the fifteen-year-olds still use a very restricted range of speech-act verbs. Only one speech-act verb is used in the spoken narrations (saga 'say') and four in the written ones (saga 'say', be 'ask, beg\ fraga 'ask. enquire', and vrala 'howl'). The adults are in many respects similar to the nine-year-olds. They show a clear preference for direct speech and free direct speech over indirect speech, and they project speech onto a a greater variety of story characters. Unlike the nine-year-olds, however, the adults typically embed their instances of free direct speech in discourse sequences designed to facilitate inferences about the identity of the speaker, as illustrated in Examples [6] and [7]. In Example [6], indirect and free direct speech are combined within the discourse construct, and the free direct speech (Tyst nu! 'Be quiet!') is parasitic on the preceding indirect speech in that it provides a specification or continuation of what the already-identified story character is saying. The discourse construct in Example [7] combines direct speech with free direct speech, which is extended into a dialogue. The examples show how the adult subjects are able to control larger chunks or units of discourse


385

Table 6. Types of speech-act verbs in the written narratives of the adult subjects 'ask' pipa 'chirp' be 'start' paminna borja 'remind' fraga 'ask' ropa 'call' frdsa 'hiss' rada 'advise' forebra skratta 'laugh' 'reproach' foresla 'propose' skrika 'shout' forsoka 'try' skrocka 'chuckle' 'reply' skdlla 'bark' genmdla 'stammer' hoa 'hoot' stamma 'shout' 'answer' hojta svara 'hush' hyssja saga 'say' 'jeer' 'wonder' hanskratta undra 'exclaim' horas utbrista 'be heard' 'exclaim' lava 'promise' utropa lata 'sound' viska 'whisper' morra 'growl' voffsa 'yelp' muttra 'mutter'

for the distribution of expressive features, which tend be simultaneously distributed in speech. [6]

[7]

Pelle maste ideligen paminna Plutt att inte vifta sa forfarligt pa svansen

'Peter constantly has to remind Plutt not to wag his tail so furiously.'

"Tyst nu!"

' "Be quiet!" '

"Vad ar det?" fragade Bert forvirrat. ' "What is it?" Bert asked confusedly.' "Tror du inte att ban kan ha gomt sig nagonstans?"

' "Don't you think he might have hidden somewhere?" '

"Varfor skulle han gomma sig?"

' "Why would he hide?" '

Further, the adults show evidence of a much richer vocabulary than the fifteenyear-olds in that they use a very large set of speech-act verbs, especially in writing (33 different types). These verbs are listed in Table 6. Unlike the nine- and fifteen-year-olds, the adults often combine these verbs with manner adverbials, for example utbrista fortjust 'exclaim delightedly', fraga surt 'ask sourly', morra forundrat 'growl puzzledly', skdlla oroligt 'bark

386


anxiously', saga med barsk rost 'say in a harsh voice', muttra forebraende 'mutter reproachfully'. See Nordqvist (2001) for further details. The dramatic increase in the use and diversity of speech-act verbs, particularly in writing, from the younger age groups to the adult group indicates a greater awareness or concern on the part of the adults that some characteristics of speech need to be taken special care of in writing and in written speech reports. Interestingly, in his analysis of the history of reading, David Olsen (Olsen, in press) makes a similar observation. Towards the end of the Middle Ages, Olsen observes, there was a dramatic increase in the use of speech-act verbs in written texts, suggesting an increased concern with rendering aspects and qualities of voices in written narratives. Olsen interprets this as circumstantial evidence in favour of the hypothesis that a new mode of reading - silent reading - began to gain ground at about that time in history. Earlier, the standard mode was to read aloud, but when this was abandoned in favour of silent reading, the need to give more information about voice quality and speech acts became increasingly evident. Abandoning spoken language in favour of written language thus puts new functional demands on written language, both in historical societal change and in individual language development - demands which have to be attended to and met through linguistic adaptation.

5. WRITING IN DEAF LANGUAGE USERS Let us first look at the production profile (Figure 4) and pausing profile (Figure 5) of the ten congenitally deaf adult subjects. Just like the normally hearing adults (see Figure 2), the congenitally deaf adults show a very dynamic overall profile (Figure 4) with pronounced peaks at the beginning and end of the narration. And their pause profile (Figure 5) shows a relatively long discourse-initial planning pause, and thereafter a low profile, indicating a high production rate. This profile is very similar to that of the normally hearing adults (see Figure 3), with the exception of a pausing peak at Picture 11. Possibly, this pausing peak is an effect of the writer's thinking-for-signing. Picture 11 depicts two simultaneous events: bees swarming out of the beehive just torn down by the dog, and the boy searching a hole in a tree. In (Swedish) Sign Language, these two events are linguistically located in two different symbolic spaces in front of the signer, who can then quickly and easily shift between the two with deictic strategies. There is no similar mechanism available in written language and this may make the description task effortful for the subjects. The high production rate of the deaf writers was also evidenced in another study by two of us (Wengelin & Stromqvist 2000), comparing the same ten deaf


387

Figure 4. Production profiles for deaf subjects in terms of key strokes for the 24 different pictures of the frog story.

Figure 5. Pause profiles for deaf subjects for the 24 different pictures of the frog story.

subjects with ten of the same fifteen controls on the basis of four different writing tasks (a narrative task with a pre-set topic ("I was never so afraid", see Stromqvist & Hellstrand 1994), a route-direction task, a job-application task, and a "letter to the editor" type of argumentative-discourse task). The two groups were compared in terms of the time they spent on the transition between different types of units during their text writing. More precisely, for each subject, the median transition times in three types of contexts were derived from the ScriptLog recordings: between letters within a word, at a word boundary, and at a sentence boundary. The group means across these median values were then calculated. The results are summarized in Table 7. As can be seen, the deaf writers are faster than the normally hearing controls on all three measures. The clearly greatest difference in production rate is associated with transitions at sentence boundaries: just above two seconds in the

388


deaf group and almost three and a half seconds among the controls. We suggest that this reduced pausing between sentences among the deaf writers may be a reflection of their style of writing when they are using the text telephone. In fact, an important part of the adaptation to the on-line constraints for writing when communicating on the text telephone is precisely to keep inter-sentential pauses to a minimum, in order not to lose the turn. The high production rate of the deaf writers showed no trade-off effect on the accuracy of spelling. On average, only 1.7% of the words written by the deaf subjects were spelled incorrectly. On the level of syntax and grammar, in contrast, the deaf subjects sometimes structured their discourse in a fashion that resembles Swedish Sign Language, making the discourse deviate from normatively sanctioned written Swedish. Example [8] (from Nordqvist 1999) provides an illustration from the frog-story task. [8]

POJKESA

'BOY SAID'

FAR-INTE

'MUST-NOT'

HOPPA-NER

'JUMP-DOWN'

GLAS-SONDER

'GLASS-BROKEN'

SPILLA PA MARKEN

'SPILL ON GROUND'

Further, unlike the normally hearing subjects (see Table 4), the congenitally deaf subjects do not make frequent use of inter-sentential connectors to create discourse cohesion. An exhaustive analysis of collocations of word forms in the written frog stories by the ten congenitally deaf subjects reveals a frequent use of och 'and', but almost exclusively for conjoining phrases (mostly noun phrases), not for connecting sentences. Also, there are very few instances of the causal/final connector for att 'because/in order to'. The top ten pairs of word forms are given in Table 8 (cf. Table 4 for normally hearing subjects). Much like the normally hearing adults, however, the deaf writers make relatively frequent use of direct speech: nine out often subjects use direct speech at least once (see Table 9) in their written narratives (Nordqvist 1999). In all cases but one, the deaf narrators construe the story characters as hearing and speaking beings, and when they project utterances onto the story characters the utterances are construed as spoken rather than signed. In one single case, the writer describes the boy and dog as being deaf and continues to explain why they still use their voices to call for the frog: [9]

Som jag uppgav att de e dova pojken och hunden

'As I reported that they're deaf the boy and the dog


Table 7. Transition times for the construction of words and sentences Location of transition

Congenitally deaf adults Normally hearing adults tot N words = 1 8 000 tot N words = 1 1 000

between letters within a word (s)

0.210

0.247

word boundary(s)

0.542

0.720

sentence boundary (s)

2.170

3.332

Table 8. Top ten word pairs in a group often congenitally deaf writers Rank Word pair

English translation

Tokens

1 och hunden

'and the dog'

58

2 pojken och

'the boy and'

31

3 efter grodan

'after the frog'

29

4 och pojken

'and the boy'

28

5 tittar pa

'is watching'

17

5 det var

'it was'

17

7 och kalle

'and charlie'

16

7 grodan boll

'froggie ball' (name)

16

9 se pa

'look at'

15

9 pa marken

'on the ground'

15

Table 9. Distribution of indirect and direct speech in spoken and written narratives. Adult congenitally deaf subjects Indirect speech Direct speech Free direct speech 1" "X said that , X said" boy dog other boy dog other boy dog other Writing 2 0 0 9 0 0 2 1 2 N subjects N tokens 2 0 0 14 0 0 10 1 2 5 Verb types

390


men det ar grodan som kan hora

but it's the frog that can hear

darfb'r skrek de och ropade grodaaaaaa

therefore they yelled and shouted froooooog'

The form grodaaaaaa in [9] illustrates the usage of an expressive feature, reduplication of letters, which is simultaneously distributed with the formation of the wordgroda' frog'. This feature, which can be considered to be conventionalized, is used to mark the word as free direct speech and as an exclamation with a certain duration. This is a strategy used by both the deaf and the hearing subjects. The deaf subjects, however, use a greater variety of techniques for accomplishing the same or similar functions. Consider, as an illustration, Example [10], where the free direct speech is marked with hyphens between letters: [10]

POJKE [...] ROPAR

'BOY [...] SHOUTS'

H-A-L-L-A H-J-A-L-P

'H-E-L-L-O H-E-L-P'

H-U-N-D

'D-O-G'

6. TEXT WRITING AND VISUAL ATTENTION In order further to explore how a writer distributes his attention during text writing, we have recently combined ScriptLog recordings with registrations of eye movements.4 Example [11] provides an example of output from ScriptLog analysis which integrates selective eye-tracking data (Stromqvist & Karlsson 2002). The example is taken from the recording of an adult subject who is highly proficient on the keyboard, and it shows the first two sentences produced in relation to the first picture of the frog story. After having activated the first picture of the picture-story elicitation instrument, the subject in Example [11] first makes a pause of 5.896 seconds. During that pause she looks at the computer screen, making six fixations (that is, according to the fixation criterion used, focusing her eyes on a point or small region for at least 120 milliseconds), including three fixations on the picture (for example, "FIXSOOPic" means "fixation on the picture with a duration of 500 milliseconds"). She then writes well over a sentence and makes a pause (2.415 sees) before writing en liten hundvalp 'a little puppet'. During that pause she makes two fixations on the picture. The last fixation in Example [11] occurs in connection with an editing operation: the subject has made a typo, h8, fixates on the text for 340 ms, deletes the 8 by backspacing one step and then writes undvalpen, thereby achieving the word hundvalpen 'the puppet'. Adding eye-movement data to a ScriptLog recording makes it possible, among other things, to test hypotheses about what the writer is doing while he


391

is making a pause during text writing. Is he reading his text, is he looking at the picture, etc.? The limited pilot data we have analysed so far (a handful of adults) are clearly insufficient for making general claims, but they are very consistent. The writers spend the great majority of the pausing time doing something other than looking at the pictures or reading their own text. Thus, the proficient keyboard user in Example [11] spent 26% of the total writing-session time pausing (twosecond criterion), but only 2% looking at the pictures and only 1 % looking at her own text. Fixations on the text occur mainly in connection with corrections and revisions of the text, as illustrated in Example [11]. Further, in contrast to eye movements in normal reading, where fixations seldom exceed a duration of 0.5 seconds (see, e.g., Hyona & Olson 1994), several

392


fixations in our pilot data are much longer (> 1 second). These unusually long fixations suggest that the writer is not really reading his text during the fixations in question. Rather, he has found a point which his eyes can rest upon while he is concentrating on something else, such as planning his next move in the textproduction process. A possible interpretation of these patterns is that the writers we have observed so far are trying to avoid being distracted in the text-production process by, among other things, minimizing their search for external information (reading their text or looking at the pictures). A similar explanation was put forth by Argyle (1967) to account for why speakers tend to avoid gaze in face-to-facinteraction. especially when they talk about complex topics.

7. DISCUSSION Does thinking-for-writing differ from thinking-for-speaking? Our analyses allow us to approach this question with reference to three broad and partly intertw ined dimensions. The first dimension has to do with planning and monitoring activity and breaks down into two principles, which could be phrased as Slobinian operating principles (see, e.g., Slobin 1977): first think, then produce! and produce, then rethink! The on-line constraints of spoken communication allow a minimum of thinking in advance of producing, whereas the relaxation of these constraints in most forms of written communication allows a maximum. Pilot data from our research on eye movements in adult text-writers bolster this claim: the great majority of their pauses are not used for reading, but rather for planning. Also, rethinking what was produced and editing it accordingly is easier in writing than in speech. From a developmental point of view, however, these operating principles are not automatically followed as soon as written language enters the scene. Rather, their application is the outcome of a long-term development where the learner is progressively adapting to and exploring the constraints and potentials of written language. It is only in our group of adult writers that we see really long discourse-initial planning pauses and extensive content editing. But not all forms of written communication are "off line". Adaptation to the use of written language for on-line communication - for example, on the text telephone - will, in some respects, make thinking-for-writing closer to thinking-for-speaking. We interpret the minimization of inter-sentential pauses in our congenitally deaf subjects (all users of the text telephone) as a sign of that kind of adaptation. The second broad dimension has to do with the encoding and distribution of information in a linear versus simultaneous fashion and in a uni- versus multidimensional fashion. The simultaneously distributed multi-dimensional encoding


393

of prepositional content (words and grammar), speaker identity, and attitudinal and emotional stance (intonation and voice quality) in the construction of storycharacter speech is quickly and easily accomplished in speech. In writing, by contrast, the same functions must be realized through a linear distribution of units which encode information along the dimensions of words and grammar only. Failure to adapt to this constraint can lead to a loss of information, as when the nine-year-old subjects indulged in free direct speech in their written narrations without stating the identity or describing the stance of the speaking story character. In this particular context, the adult control subjects and the deaf writers show evidence of more sophisticated strategies. The former use a wide range of speech-act verbs, and the latter make frequent use of simultaneous multi-dimensional information encoding which conforms to the conditions of written language, such as indicating free direct speech by quotation marks and/or hyphens between letters (for example, H-A-L-L-A 'H-E-L-L-O'). The third dimension is closely related to the second one and concerns the size of the domain of distributed features. A predominantly linear distribution of expressive features requires a larger domain of planning (thinking-for-writing) than a simultaneous distribution of expressive features (thinking-for-speaking). On the other hand, the off-line character of writing allows more sophisticated lexical choices, and words can be very rich in meaning, resulting in a high information density of the discourse. Conversely, spoken discourse is typically characterized by lower lexical density and greater use of very frequent words (function words) as a way of gaining time and relaxing the cognitive load under the on-line constraints of speaking. From a developmental point of view, there is increased richness of lexical choices and increased control over inter-sentential connectivity with age - especially in later development, as evidenced in the leap between the fifteen-year-olds and the adults in our data. As regards the lexicon, the nine- and fifteen-year-olds use only three or four different speech-act verbs in their spoken and written narrations, whereas the adults use eight types in speech and thirty-three (!) in writing. And unlike the nine- and fifteen-year-olds, the adults often combine these verbs with manner adverbials, for example morra forundrat 'growl puzzledly'. As regards connectivity, the narratives of the nineyear-olds are largely characterized by additive strategies where sentences are connected by the general conjunction och 'and', in speech as well as in writing. The fifteen-year-olds use the same strategy in their spoken narrations, whereas the written ones are distinguished by the causal and final conjunction for aft 'because/in order to', yielding sentences which are more tightly inter-connected. And the adults use the latter strategy in both speech and writing. Further, the adults show evidence of greater control of larger chunks of discourse also in that they carefully embed free direct speech in contexts from which the identity of the story character is evident.

394


In short, the nine-year-olds are still much influenced by their way of thinkingfor-speaking when they are writing. Marks of written language on their written narrations are there first and foremost in the form of effort with lower-level actions (keystrokes and spelling). In the fifteen-year-olds, by contrast, thinkingfor-writing has started to penetrate their written discourse. The fifteen-year-olds behave more strategically when they compose their written narrations, they spend more time on discourse-initial planning, and they knit their written discourse more tightly together by means of causal and final sentence connectors. The increasing weight of thinking-for-writing in the fifteen-year-olds can also be seen from the fact that their frequent use of causal connectors in pure writing (wr_sp_WR) carried over to the subsequent spoken narration (wr_sp_SP), whereas their written narrations produced subsequently to speaking (sp wr WR) remained unaffected by the previous spoken narration (sp_wr_SP). Further research will help determine to what extent these findings generalize across different samples and different languages.

NOTES 1.

This study was supported by the Swedish Council for Social Research (SFR/FAS).

2.

A proposition is a semantic unit (which can be true or false). In Klima and Bellugi's operationalization, a proposition corresponds roughly to a clause (whether embedded or non-embedded). The proposition rate is a measure designed to specify the amount of information expressed over a period of time.

3.

More information about ScriptLog can be found at the website www.ScriptLog.net

4.

The combined technology of ScriptLog and eye-tracking has been developed in the project "The dynamics of perception and production during text writing", sponsored by the Swedish Research Council and the Centre for Reading Research, University College of Stavanger. Project leaders are Sven Stromqvist. Department of Linguistics, and Kenneth Holmqvist, Department of Cognitive Science, both at Lund University. Eye movements are recorded using SMI iView eye-tracking equipment.

Form-Function Relations in Spanish Narratives' A Comparative Study of Bilingual and Monolingual Speakers

Judy Kupersmitt

1. INTRODUCTION In their cross-linguistic study on the development of form-function relations in narrative, Herman and Slobin (1994) concluded that narrative abilities are influenced by three inter-related factors: cognitive command of a narrative schema, the interaction of linguistic forms and narrative functions, and targetlanguage typology. Specifically, the set of "rhetorical options" available in any given language guides speakers to particular ways of organizing the flow of information in discourse. Recent studies have also addressed the question of the development of form-function relations in narratives produced by bilinguals from a contrastive/typological perspective. These studies focused on issues such as the development of cohesive devices in bilingual development (Aarssen & Bos 1999); the use of linguistic structures to encode narrative components in bilinguals' and monolinguals' narratives (Akinci 1999; Akinci, Jisa & Kern 2001); and the relationship between linguistic proficiency and narrative ability in bilingual children from different socio-economic backgrounds (Pearson & Umbel 1996). The present study examines the effect of increasing exposure to Hebrew as L2 on the development of form-function pairings in the narrative texts produced by Spanish-Hebrew bilingual children and adults. The term form refers in this context to the whole range of inflectional and closed-class lexical items, morphological and syntactic constructions, syntactic operations, and lexical expressions existing in the grammar and the lexicon of a language. Function here refers to the semantic and discourse roles that a variety of forms play in constructing a linguistic text (Berman 1996). For example, the temporal notion of "simultaneity" can be expressed by means of morphological markings on the

396

Judy Kupersmitt

verb and/or by lexical expressions between clauses, depending on the devices available in a given language (Aksu-Koc & von Stutterheim 1994). The present chapter addresses three main research questions: /.

Do bilinguals' narratives differ from those of monolinguals in the development of narrative structure across age in terms of explicit reference to an overall plotline: onset of the problem, goal, and resolution of the problem?

//'.

What is the effect of increasing exposure to Hebrew as L2 on the use of linguistic devices to express various functions in bilinguals' narratives, and what factors (psycholinguistic and typological) make a certain form more vulnerable to loss than other forms in the non-dominant language (Spanish)?

;';'/.

Is there evidence of different acquisitional and developmental paths in monolinguals' and bilinguals' language as reflected in the use of morpho-syntactic devices and in the occurrence of errors?

The study thus intends to describe the differences in linguistic forms used by bilinguals and monolinguals, aiming mainly to define the changes occurring in the Spanish of bilinguals. In this sense, the study relates to the issue of language loss and aims to detect systems which seem "underdeveloped" (Andersen 1989). With respect to narrative organization, I assume that the development of narrative structure will follow common patterns across age in monolinguals' and bilinguals' texts. That is, the mastery of an overall narrative plotline will depend on age rather than on linguistic background. In the domain of form-function relations, two contrasting trends are predicted in the narratives of bilinguals compared with those of monolinguals. On the one hand, both will gain in syntactic complexity with age and will include more cohesively packaged clauses, with maturely and appropriately used rhetorical devices. But the texts produced by bilinguals will show a disadvantage in the choice of language-specific devices for encoding different semantic distinctions, such as the temporal contour of events marked by aspectual forms. Whether a certain form is lost or kept depends on three interacting factors: a) typological differences between LI and L2 (such as non-equivalence of grammatical features); b) basic typological properties in LI (such as the pervasiveness of a system in the grammar)2; and c) the development of form-function relations in the language of an individual, with more forms used to encode a given function and more functions fulfilled by a particular form (Berman 1994). In the development of linguistic forms, bilinguals are expected to show three kinds of deviations from monolingual-adult usage:

Form-Function Relations in Spanish Narratives

397

Developmental errors reflect strategies adopted by children when dealing with structural difficulties in the process of language acquisition (Berman 1984; Ravid 1995). Fossilized errors (see Ravid 1988, for LI; and Selinker 1974, for L2) are analogous to "stagnation" errors (Aarssen 1996; Bos 1997). These are forms which manifest a "developmental deficit" in the sense that they occur early in the acquisition process and do not change with development, to conform to target-language norms. Transfer errors derive from what is sometimes called "interference" (Berman 1985) from another language in situations of languages in contact, where structural patterns or lexical distinctions are transferred from the dominant to the weaker language.

2. METHOD Narratives produced in Spanish by Hebrew-Spanish bilingual children and adults, using the same picture book and the same methodology as in Berman and Slobin (1994), were compared with those of their Spanish monolingual peers performing the same task. 2.1 Population Data were collected from Spanish-Hebrew bilinguals of South American Spanishspeaking background living in Israel. The subjects are LI speakers of Spanish and L2 speakers of Hebrew, where LI precedes L2 in temporal order of acquisition but drops behind in prevalence of usage (Polinsky 1997). All subjects were born in Israel to first-generation immigrant parents or immigrated themselves to Israel no later than at the age of two. All share the same middle-class to upper-class socio-economic background, and both parents of all the children are educated native speakers of Spanish. Subjects with an impoverished knowledge of Spanish were excluded from the sample, on the basis of a vocabulary test and a selfevaluation questionnaire (filled out by the parents of the younger children).3 The control group consisted of monolingual speakers of (South American) Spanish, matched for age and socio-economic background. The narratives from monolingual Spanish-speakers were collected by Aura Bocaz in Argentina and Chile (Bocaz 199Id; Sebastian & Slobin 1994). The subjects were divided into four groups as detailed in Table 1. These sub-groups represent different states of literacy - pre-literacy, early and later school age, and developed literacy.

Judy Kupersmitt

398

Table 1. Breakdown of the population by language, age, and level of schooling or education Schooling

Linguistic background

N

Age range

Mean age

monolinguals

10

4;0-5:0

4:5

bilinguals

10

4:0-5:4

4:7

Grades II-III

monolinguals

11

7:0-8:0

7:5

bilinguals

11

7:0-8:3

7:6

Grades VI-VII

monolinguals

10

11:0-12:0

11:5

bilinguals

10

10:8-12:2

11:5

9

Not spec.

Not spec.

8

20-35

27

Pre-school

Secondary school completed monolinguals bilinguals

2.2 Materials The narratives were elicited by means of the booklet by Mercer Mayer Frog, where are you? (Mayer 1969), a wordless booklet with 24 pictures (see Appendix I) which describes the adventures of a boy and his dog who go looking for their frog which has disappeared. This matched the stimulus and the procedures used for the Spanish monolingual samples that served as the control group for this study, and made it possible to control for the content of the narratives elicited. This picture book has also been used successfully for elicitation of narrative texts from monolingual children at the same ages as those included in my study, in a range of languages: English (Wigglesworth 1992), French (Kail & Hickmann 1992; Kern 1997), German (Bamberg 1987), and, most importantly for my purposes. Hebrew (Berman 1988a; Berman & Neeman 1994). The same booklet has recently been used in a number of studies of bilingual children, including children speaking Turkish and Dutch (Aarssen 1996), Moroccan Arabic and Dutch (Bos 1997). and Turkish and French (Akinci 1999; Akinci, Jisa & Kern 2001), as well as Spanish and English (Pearson & Umbel 1996). 2.3 Procedures The subjects were first instructed (in Argentinian Spanish - the native language of the investigator) to go through the entire booklet and to look carefully at the pictures. They were then asked to tell the whole story with the book open. Since this study is not intended to measure vocabulary proficiency, the subjects were encouraged to ask about any word they could not recall, either before they started


399

or while telling the story. This was mainly intended to reduce the subjects' anxiety about "getting stuck" at some point in the story. The instructions were the same as in the Berman and Slobin (1994) study: "Here's a picturebook about a boy [pointing to the picture on the cover], a dog [pointing], and a frog [pointing]. First, 1 want you to look carefully at all the pictures. Pay attention to each picture and then you tell me the story while you look at the pictures again." Little children often needed scaffolding or some kind of help in order to gain confidence in the task of telling a story. The sessions were held in the subjects' homes, often in the presence of the mother (for the little children), and lasted around half an hour. They were recorded by audio tape-recorder and transcribed immediately after the session following the transcription conventions adopted for Spanish in the Berman and Slobin study. As noted, the texts were analysed along two dimensions: narrative structure and linguistic form. In presenting the results (Sections 3-4), I will interpose elements of interpretative discussion together with a quantitative description of the findings.

3. RESULTS: NARRATIVE STRUCTURE To evaluate the effect of different backgrounds (monolingual versus bilingual) on the development of narrative structure across age, the texts were examined in terms of length and reference to narrative components. 3.1 Text length The length of the texts was analysed in terms of the number of clauses, where a clause refers to "any unit that contains a predicate which expresses a single situation, i.e. an activity, an event or a state" (Berman & Slobin 1994:660). Similarly to what was found in other monolingual studies on the development of narrative abilities in different languages (Berman & Slobin 1994:31; Berman 1988a), I expected that the length of the texts would increase with age, and that the patterns of growth would be similar in the monolingual and bilingual groups. Table 2 shows that, as predicted, there is a significant increase in the mean number of clauses per text with age in both monolingual and bilingual groups (r (4()) = 0.73,/7< 0.02 and r = 0.505,p< 0.02, for monolinguals'andbilinguals' texts, respectively). The second prediction was partially confirmed. There was a marked similarity in the pattern of text-length increase in both groups, but the bilinguals' texts, especially those of school-age children and adults, are generally shorter than those produced by monolinguals.

400

Judy Kupersmitt

3.2 Narrative components Narrative quality was analysed by reference to an overall plotline composed of three crucial components of this picture book: the initial event chain (frog leaves jar; boy and dog find that it disappeared), the search motif (boy and dog go out in search of frog), and the ending (boy finds his lost frog, or gets another one), with each component divided into two sub-components4. I assumed that with age, both monolinguals and bilinguals would produce narratives that would be richer in terms of explicit mention of the narrative components. This turned out to be the case in both groups. There was a significant effect of age on the development of overall narrative structure (r (40) = 0.72, p < 0.01 and r, 9 | = 0.75. p < 0.01 for monolinguals and bilinguals, respectively). The four-year-olds mentioned an average of 2.70/2.90 components while the eleven-year-olds and the adults averaged more than five out of a total of six components. This parallel development towards richer, more elaborately constructed stories in monolinguals' and bilinguals' texts is illustrated by four excerpts relating to the ending of the story ("the boy finds the frog'V'the frog is the missing one or a substitute for it"). Examples [1] and [2] are segments produced by four-year-olds and eleven-year-olds who are monolingual (a) and bilingual (b), respectively: [la] Vieron dos sapitos, y todos sus sapitos. Y se puso contento y estaba en el agua con un sapito en la mano, y dijo "jchau!" '(They) saw two toadies, and all their toadies. And (he) became happy and (he) was in the water with a toadie in the hand, and (he) said "bye!" '

[Ib] Y entonces vio la rana y vio todas las ranitas chiquitas. Y despues dijo "chau" a la ranita. 'And then (he) saw the frog and (he) saw all the little froggies. And then (he) said "bye" to the froggie.' (BS4c-4;9)

[2a] Y jsorpresa! Ve a su ranita preferida, con un sapo. De repente, sale la familia, y Pepe esta contento. Pepe se lleva uno de los renacuajos. y se despide de los padres contentos con su linda familia. 'And surprise! (He) sees his favorite froggie, with a toad. Suddenly exits the family, and Pepe is happy. Pepe takes one of the tadpoles, and says good-bye to the parents [who are] happy with their beautiful family/ (MSI lc-1 1 ;3)

[2b] Y ahi vio a dos ranas, y luego vio junto a ellas muchas ranitas chiquitas. Entonces el nino se llevo una de las ranas y se fue de camino de regreso a su casa. 'And there (he) saw two frogs, and then (he) saw together with them a lot of little froggies. Then the boy took one of the frogs and went the way back to his home.'

401


Table 2. Mean number of clauses per text in monolinguals' and bilinguals" texts, by age

Age

Monolinguals' texts

Bilinguals' texts

Range of clauses

Mean

SD

Range of clauses

Mean

SD

4;5

33-54

42.7

6.4

24-88

44.8

16.9

7;5

46-73

61.4

7.4

28-63

49.9

10.9

11;5

45-100

60.1

15.8

38-79

48.1

12.3

Adults

60-170

95.0

37.3

56-140

78.8

28.6

As shown in other studies of monolinguals' frog-story texts as well as other types of picture-based narratives in a variety of languages (e.g., Katzenberger 1994; Kern 1997), younger pre-school children tend to treat each scene as an isolated event, describing the content of each picture at a local level. They do not provide relevant temporal, spatial, or motivational scene-setting, nor do they make explicit reference to the onset of the problem (the frog leaving the jar and the boy's discovery that it is gone). Also, as illustrated in [1], they fail to provide a resolution that implies a connection with the rest of the story. Later, around the age of six, children begin to link events sequentially and are able to create a temporally organized text. Only after the age of seven will stories be thematically unified, with events connected to one another both temporally and causally at a more global level. The examples in [2] show that older children explicitly mention the finding of the frog and also relate to the fact that the frog is either the same as the one that escaped at the outset of the story or another frog from the family of frogs. There is a clear difference between the stories of younger (pre-school) and older (school-age and adult) subjects, since the latter explicitly mention more narrative components, which are effectively interconnected to achieve a hierarchical, global narrative organization. These findings support the claim that the development of narrative organization and structure is largely language-independent, particularly in populations which belong to cultures with shared narrative conventions and similar socio-cultural characteristics, such as the subjects in this study (Kupersmitt & Berman 2001). The findings clearly demonstrate that command of global-level narrative structure develops as a function of age rather than language in the Spanish monolinguals' and Spanish-Hebrew bilinguals' texts alike. However, a more fine-grained analysis of narrative structure, in terms of how the contents of the narrative are verbally realized (Berman 1988a), shows that bilinguals often fail to encode the events appropriately. For example, some of the children had difficulty in giving specific expression to the experiences the protagonists undergo during their search for the frog, as illustrated in [3] and [4]:

402

[3]

Judy Kupersmitt

Y el perro estaba jau au au au! Y las abejas estaban jbbbbzzzzzzzz! 'And the dog was au au au au au! And the bees were bbbbzzzzzz!' (BS7a-7:9)

[4]

El nene no vio la colmena y las abejas puede ser que tiene olor feo. La colmena se cayo y las abejas salieron, y el nene siguio buscando a la rana. 'The boy did not see the beehive and the bees maybe it has a bad smell. The beehive fell down and the bees exited, and the boy continued looking for the frog." (BSlla-11)

Example [3] shows the use of onomatopoeia to replace two verbs related to animal behaviour - ladrar 'bark' and volar 'fly' - and the more specific term zumbar 'buzz'. In [4], an elliptical utterance breaks the flow of the narration, as the subject avoids an unknown or inaccessible lexical item, probably related to what the bees were doing. This bilingual disadvantage seems related to lexical production and becomes particularly noticeable among school-age subjects. Evidently, while they are gaining proficiency as narrators in general cognitive and communicative terms, there develops an increasing gap between Spanish, their home language, and Hebrew, which by now has become their dominant language - as shown by other studies of Israeli bilingual children from similar socio-economic backgrounds (Kupersmitt & Berman 2001; Rabinowitch 1985).

4. RESULTS: LINGUISTIC FORM To analyse the effect of increasing exposure to Hebrew as L2 on the Spanish of bilinguals as compared to monolinguals, I examined the texts in three domains: interclause connectivity (4.1), form-function relations (4.2), and command of specific linguistic devices (4.8). The studied subtypes of form-function relations are treated under separate subheadings for clarity: grammatical aspect (4.3), nonfinite forms (4.4), null subjects (4.5), word order (4.6), and relative clauses (4.7). 4.1 Inter-clause connectivity Connectivity is defined here in terms of "syntactic packaging" or "clause linkage" serving to create skilful connections between the events related in the stow (Berman & Slobin 1994:538-548). Four means of clause linkage were analysed: /'.

verb gapping: clauses linked by omission of either the predicate or a tense-carrying auxiliary in the second clause;


403

//. finite linking: clauses linked by a subordinating or non-subordinating conjunction (such asporque 'because', cuando 'when',pero 'but') with a finite verb in the conjunct clause; ///'.

non-finite linking: clauses linked by gerund and infinitive constructions;

iv.

relative clauses: restrictive and non-restrictive as well as headless relative clauses.

As expected, syntactic clause combining increased with age in both groups, going from less than 10% in the texts of the youngest children to more than 30% in the adult texts. The results on the development of syntactic packaging are detailed in Table3. Table 3 shows that finite and non-finite linking as well as relative clauses increase significantly with age. For example, finite linking goes up from less than 5% to around 20% in both groups (r = 0.7, p < 0.01 in the monolingual and r = 0.52, p < 0.01 in the bilingual group). Verb gapping showed no correlation with age (r = 0.08,/?> 0.5) and was common in the texts of younger bilinguals, as in [5-6]: [5]

El nene se cayo y tambien el perro. 'The boy fell and also the dog.' (BS7a-7;9)

[6]

A la manana el chico se levanto y el perro tambien. 'In the morning the boy got up, and the dog also.' (BS7d-8;2)

This use resembles that of five-year-old Hebrew-speakers, who were reported to frequently use the word gam 'also' for specifying that the same situation applies to different protagonists in the story (Berman & Neeman 1994:317 - and see also Berman 1996, on the use of 'also' as a precursor to syntactic co-ordination in other languages). Overall, young children in both groups constructed their texts from isolated clauses, using numerous deictic, "here and now" expressions such as acd 'here'. Also, most of their texts showed over-marking by sequential terms such as y 'and', despues 'then, afterwards', y despues 'and then' at the beginning of each clause, as in [7]: [7]

Despues se fue. Y el perro estaba escondido, y el nene no. Y despues el animal levanto al nene, y despues tiro al nene y al perro. 'Then (he) went away. And the dog was hidden, and the boy not. And then the animal picked up the boy, and then (he) threw (down) the boy and the dog.' (BS4e-4;3)

404

Judy Kupersmitt

Older children and adults, in contrast, used a range of devices for syntactic packaging, including finite and non-finite subordination as well as relative clauses. However, as Table 3 shows, these devices were used far less among bilinguals than among monolinguals, particularly among eleven-year-olds and adults. This is illustrated by the following excerpts translated into English from an eleven-year-old monolingual [8a] and an eleven-year-old bilingual [8b], relating to the scene where the dog gets stuck in the jar and then falls out of the window: [8a] (While) looking through the window, the boy called the frog, while the dog continued with the jar inserted [= on its head]. The dog fell and the boy was very surprised. When (he) picked him up, the dog licked him a lot, because (it) was very happy.

[8b] (They) went outside to look for it, and the dog, without purpose, entered its head inside the jar, and (he) could not take it out. Suddenly the dog fell from the window and the jar was broken. And the boy got a bit angry.

In [8a], the clauses are well connected by means of various syntactic devices. including non-finite linking, through the initial adverbial clause mi ran do potla ventana '(while) looking through the window'. The clauses are packaged in combinations of backgrounded and foregrounded clauses, with subordinators such as 'while', 'when', and 'because'. In [8b], by contrast, the clauses are connected mainly by y 'and'; also, although they are coherently ordered and thematically unified, they lack the richly embedded and complex packaging achieved by a combination of finite and non-finite subordination. Thus, monolinguals used a wide range of lexical connectives in the category of finite linking, compared with bilinguals. They also used non-finite connectors such as adverbials to describe attendant circumstances that precede or are simultaneous with the main, plotadvancing events. Another major difference was noticed in the functions that forms served when used as connectors. One example of this is in the use of relative clauses. Monolingual adults, for example, used them as a rhetorical device to shift between foreground and background information or to encapsulate events in the story, as shown in [9aj. Bilinguals, on the other hand, used relative clauses with a more local, descriptive function, as in [9b] [9a] Y Tomasito habia decidido meterse en un agujero que habia en un arbol muy muy grande, en el bosque, tratando siempre de buscar a la ranita que no encontraban por ningun lado.

405


Table 3. Mean percentages of packaged clauses, by age and linguistic background, out of the total number of clauses

Age

Linguistic background

4;5

monolinguals bilinguals

7;5

monolinguals bilinguals

11;5

monolinguals

Finite linking

0

4%

Non-finite linking

Relative clauses

0

2%

Total

6%

2%

4%

0

2%

8%

0

12%

4%

8%

24%

7%

14%

2%

8%

30%

0

24%

6%

8%

38%

2%

12%

2%

4%

22%

monolinguals

0

20%

12%

16%

48%

bilinguals

0

18%

4%

8%

30%

bilinguals Adults

Verb gapping

'And Tomasito had decided to enter a hole that was in a very very big tree, in the woods, trying all the time to look for the frog that (they) did not find anywhere/ (MS20f-20)

[9b] De repente salio una lechuza que se desperto en el medio del dia. 'Suddenly exited an owl that woke up in the middle of the day.' (BS20d-26)

An interesting finding was bilinguals' use of unconventional connectors, not found among monolinguals, as in [10], which is from a bilingual adult: [ 10] Mientras tanto estaba el nino buscando la rana adentro de un arbol, pero parece que las ranas tampoco les gusta los arboles. De repente las avispas empezaron a correr atras del perro, cuando el nino estaba buscando adentro del agujero de un arbol. 'In the meanwhile, the boy was looking for the frog inside a tree, but it seems that frogs does6 not like trees either. Suddenly the wasps started to run after the dog, when the boy was searching inside the hole of the tree.' (BS20d-26)

In [10], the connective cuando 'when' (which corresponds to the Hebrew kshe 'when, while') is used instead ofmientras 'while'. Interestingly, the two temporal adverbs used in the bilingual example, de repente 'suddenly' and mientras tanto 'in the meanwhile', correspond to two frequently used adverbials in the Hebrew monolingual narratives: pit'om and benatayim, respectively (see Berman & Neeman 1994:301).

406

Judy Kupersmitt

In sum, the differences in syntactic clause linkage are less marked among the younger children, since both bilinguals and monolinguals use an equally limited range of devices and rely heavily on linear chaining rather than embedding. Among school-age children and adults, bilinguals seem to fall behind monolinguals in three ways: they use a limited set of forms; some of the forms they use resemble Hebrew rather than Spanish forms of expression; and their forms serve local rather than global narrative purposes. These differences can be explained on developmental and typological grounds. In developmental terms, the bilinguals' use of devices for connectivity resembles earlier phases in development, where fewer forms are used for a narrower spectrum of functions. Thus, although Hebrew and Spanish share typological features for finite coordination and subordination, it seems that in the course of telling a story, bilinguals adhere to early-acquired devices such as v 'and\porque 'because', and cuando 'when'. Typologically, the difference seems related to the accessibility of certain devices in the two languages. For example, non-finite forms such as infinitives and gerunds are highly accessible linguistic means in Spanish, and are used as adverbials for clause combining from the age of seven onwards. In Hebrew, non-finite constructions such as gerunds and nominalizations are acquired late, and they constitute highly literate, quite formal-register forms of expression. Another typological contrast regards the use of conjunctions such as mientras 'while'. This lexical connector is acquired by speakers of Spanish somewhere in the middle of the school years and is commonly used to express simultaneity between events. In Hebrew, corresponding terms such as be-od she 'whereas' or bi-zman she 'at the time that' in the sense 'while' are of a higher register, and are used almost exclusively by adult narrators. These differences may explain why Spanish-Hebrew bilinguals adhere to more Hebrew-like means of clause combining, such as cuando 'when', corresponding to the general temporal subordinator kshe 'when, while' in Hebrew, and the temporal adverbial mientras tanto 'in the meanwhile', corresponding to Hebrew benatayim, to express cooccurring events. This suggests an effect of the dominant language on the choice of forms used to organize the flow of information in discourse. 4.2 Form-function relations Linguistic forms were selected to reflect typologically relevant differences as well as similarities in the morpho-syntax of Spanish and Hebrew: rich in Spanish but almost absent in Hebrew (grammatical aspect); existing in both languages but grammatically realized in different forms (non-finite verbs); existing in both languages but differently constrained (null subjects, word-order alternations): and almost identical in the two languages (relative clauses). These categories were presumed to be sensitive to factors involved in language loss and maintenance, such as the time of acquisition or the centrality of the form in the language.


407

The following subsections consider how linguistic forms function as rhetorical devices, in the sense of "the way in which speakers organize the information which they present in ongoing construction of a text by means of selection of appropriate options out of the grammatical and lexical devices available in the target language" (Berman 1998b:199). This covers a speaker's choice of relevant linguistic devices (prosodic, morphological, syntactic, and lexical) to meet such functions as narrative cohesion, by lexico-syntactic marking of inter-clause and inter-sentence relations; narrative perspective, by alternations of voice, valency, and word order; and background-foreground distinctions, by tense/aspect switches to alternate between the sequentially ordered foregrounded events and backgrounded, collateral material in the narrative. We start by considering grammatical aspect. 4.3 Form-function relations: Grammatical aspect The perhaps most striking grammatical feature differentiating Spanish and Hebrew is the aspectual system. Spanish verbs are inflected for mood (indicative, conditional, and subjunctive), tense (present, past, and future) and aspect (perfective and imperfective in the past tense; progressive and perfect in the present, past, and future tenses). Hebrew verbs, in contrast, are inflected only for infinitives, for imperative mood, and for tense (present - which is also participial -, past, and future). This subsection focuses on the use of aspectual marking on the verb to express the temporal constituency of events (such as durative versus punctual) and temporal relations between events (such as anteriority and simultaneity), and to background and foreground events in the narrative. The temporal constituency of an event is determined by the interaction between the aspectual nature of the verb (Aktionsart or lexical aspect) and the temporal construal of the event, as coded in the verb morphology. For example, stative verbs (which are inherently durative) are generally marked in the imperfective, as in el nene no sabia lo que hacer 'the boy did not know-iPFV what to do' (from an eight-year-old bilingual girl). The temporal relation between activities, events, or states is determined by the degree of overlap between situations on a given time axis. If two situations share a value on the time axis, they are considered as simultaneous (Aksu-Ko? & von Stutterheim 1994). If not, one of the situations is either anterior or posterior to the other. Simultaneity is often expressed in Spanish by combining two forms - one progressive and one non-progressive, as in [11]: [11] Todos los mosquitos vienen atras del perro y el perro se esta escapando. 'All the mosquitoes come-pRES after the dog and the dog is running-pRES.pROG away.'(BS4a-4;0)

408

Judy Kupersmitt

Anteriority is achieved by switching to the past perfect, from either the past or the present tense, as in [12]: [ 12] Pepito y su perro estaban mirando la rana que habian cazado por la manana. 'Pepito and his dog were watching-iPFV.pRoc the frog that (they) had caughtPAST.PERF in the morning.' (MS20a)

Finally, foreground-background distinctions are essential for constructing a narrative text, since they mark off plotline events from their associated circumstances (Berman & Slobin 1994:6-9; Hopper 1979a; Reinhart 1984). The foreground of a narrative is provided by the ordered sequence of events that constitutes the main plotline, typically expressed by verbs of achievement and accomplishment in the perfective aspect (in the past), while the background is provided by collateral, descriptive material, typically expressed by states. Also, two activities can be combined in a foreground-background relation (e.g.. the phone rang while I was washing up). In Spanish, backgrounded situations are generally expressed in the imperfective. On the basis of a previous study on the development of form-function relations in the narratives of Spanish-Hebrew bilinguals (Kupersmitt & Berman 2001), it was assumed that the texts produced by bilinguals would display fewer aspectual contrasts by inflectional marking of progressive, perfect, and perfective/ imperfective7 than the texts produced by monolinguals. Figure 1 shows that this prediction was partly confirmed. The use of the progressive aspect was similar in both groups, i.e.. there was no effect of linguistic background on the use of this form (F ( , _, ( = 1.04. p > 0.1 for the present progressive and F (1 =0.65, p > 0.1 for the imperfective past progressive). Among four-year-olds in both groups, more than 10% of the verbs are in some form of the progressive, either present or imperfective past. This frequent use may be related to the deictic function of the progressive, which is used for picture description (Sebastian & Slobin 1994). School-age children in both groups used the progressive to describe durative actions, in either a presentanchored or a past-anchored narrative, as in [13]: [13] Y el perro estaba oliendo y el nene gritaba. gritaba. pero no la encontraban. 'And the dog was smelling-iPFv.pROG and the boy shouted-ippv. shouted-iPF\. but (they) did not find-iPFV it [the frog].' (BS7g-8;2)

Older children and adults used the progressive exclusively to indicate protraction, as a rhetorical choice in specific scenes that were perceived as extended in time compared with other scenes in the narrative. These are typically the sleeping scene


409

Figure 1. Percentage of aspectual marking by age and linguistic background. ([14a]); the scene where the boy and the dog are looking at the frog ([14b]); the different and repeated search scenes ([14c]); and the scene with the bees ([14d]): [ 14a] Estan durmiendo los dos. "Are sleeping-pRES.pRoc both of them.' (BSadult-26) [ 14b] Y estaba por la noche contemplando una rana. 'And was at night observing-iPFV.pROG a frog.' (MS20h-20) [ 14c] Y el y el chico estaban buscando todavia. 'And it [the dog] and the boy were still searching-iPFV.pROG.' (BSllb-ll;7) [14d] El perro estaba jugando con un panal de abejas. 'The dog was playing-iPFV.pROG with a beehive.' (MSI ld-11;4) The occurrence of past-perfect forms showed a significant effect of linguistic background (F = 34.8, p < 0.01) and of linguistic background and age (F = 6.14,;? < 0.01). Among monolinguals, this form is maturely used for the expression of retrospection in narratives from the age of seven onwards ([15]). The use of this form increases with age, and so does the number of contexts where it appears. Seven-year-olds switch to this form only from the past tense, while older children and adults frequently shift to it from the present ([16]): [15]

Habia un chico con un perro que habian atrapado a una rana. '(There) was a boy with a dog that had captured-PAST.PERF a frog.' (MS7g-7;3)

410

Judy Kupersmitt

[16] Y el nino salta y se enoja porque no lo habia esperado. 'And the boy jumps and gets angry because (it) [the dog] hadn't waited-PAST.pERF forhim'(MSlla-ll;l)

This is a typically "late acquired" form since it requires attention to three temporal points - speech time, reference time, and event time (Weist 1986). This can explain why the past perfect is absent from the repertoire of bilingual children and adults. Apparently, they do not manage to acquire it by the time they enter school, and they do not catch up later on, either. Consequently, they opt for the perfective to express retrospection, e.g., habia un arbol que se cay6 'there was a tree that fell-PFv.' (BSlla-11) (instead of se habia caido 'had fallen'); y ahora entiende por que se escapo 'and now (he) understands why (she) [the frog] escaped-ppv' (BSadult-26) (instead of se habia escapado 'had escaped'). The use of the imperfective aspect was similar in both groups across ages. A 2-by-4 ANOVA (see, for example, Iverson & Norpoth 1987) showed no effect of age interacting with linguistic background (F ,,; = 0.14, p > 0.5). In both groups, the imperfective occurred mainly with stative predicates, such as ser and estar 'be', poder 'be able to, can', querer 'want', and tener 'have', as in [17a]; and. to a lesser extent, with dynamic-atelic verbs, such as gritar 'shout', perseguir 'chase', andjugar 'play', associated with durative, repeated, and non-completed activities, as in [17b]; as well as, in a few cases, with non-dynamic verbs (such as verbs of perception), as in [17c]: [ 17a] El nene vio un raton y tenia olor feo. The boy saw a mouse and (it) had-iprv a bad smell.' (BS4j-4;l 1)

[ 17b] Y el nene gritaba y gritaba pero no la encontraban. 'And the boy shouted-ippv and shouted-iprv but (they) did not find it.' (BS 7g-8:2)

[17c] Entonces se sento con el perrito y veian la rana. 'So (he) sat down-ppv with the dog and (they) looked at-iPF\ the frog.' (BS4h-4:6)

Contrasts between perfective and imperfective aspect are used for an immature kind of backgrounding from an early age, as in [18]: [18] El nene creia que era un arbol y no era, y cuando salio el pajarraco el nene se asusto. 'The boy believed-iPFV that (it) was-iPFV a tree, but (it) was-iPFv not. and when exited-PFV the big bird the boy got frightened-PFV.' (MS4i-4:6)


411

Later, these contrasts are interwoven in larger stretches of text, yielding a twodimensional texture, that is, a mixture of sequentially ordered, plot-advancing foreground events (encoded by what Labov termed "narrative clauses") and their associated background or orienting situations, generally unbounded or suspended in time (Silva-Corvalan 1991). This is illustrated in [19]: [ 19] Mientras que el chico dormia en esta noche, la rana se escapaba. Al amanecer, el chico busco y no la encontro. Se habia escapade. El chico se vistio, busco por todas partes, y se asomo a la ventana para ver si la encontraba. 'While the boy slept-iPFV that night, the frog escaped-iPFV. At dawn, the boy searched-ppv but did not find-ppv it. (It) had escaped-PAST.PERF. The boy dressed; PFV. searched-PFV everywhere, and leaned out-PFV of the window to see if (he) found-iPFV it.' (MS 11 a-11)

Example [19], from an eleven-year-old monolingual subject, shows a mature, skilful use of three aspectual forms: the imperfective, perfective, and past perfect, yielding a three-dimensional texture to the opening scene and the subsequent events. This subject suspends the first two events, advances the plot, retrospects, and moves forward again. This example also illustrates the use of the imperfective to express expectation (Fernandez-Ramirez 1986). The last verb, encontraba 'found', is used in the imperfective to create an expectant perspective regarding the result of the search. On the surface, there seem to be no differences in the use of the imperfective between monolinguals and bilinguals. However, Spanish-Hebrew bilinguals appear to show less flexibility than their monolingual peers in using the imperfective aspect for the purpose of backgrounding as a more global, narrative device. After the age of five and at school age, bilinguals construct texts which are poor in aspectual contrasts, which is evident from the massive use of the past perfective in their texts, as shown in the figure above. This lack of variability and flexibility makes the bilingual children's Spanish-language texts more linearly constructed in a way typical of monolingual Hebrew narrators (Berman & Neeman 1994), since they show less in the way of narrative foregroundbackground distinctions, as illustrated in [20] from an eleven-year-old bilingual, relating to the same scene as [19]: [20] En la noche, cuando el nene se fue a dormir y el perro, se escapo la ranita. Despues a la manana, cuando se despertaron. miraron el frasco y vieron que la ranita no esta. La buscaron por todos lados... Mn the night, when the boy went-PFV to sleep and the dog, escaped-PFV the frog.

412

Judy Kupersmitt

Then in the morning, when (they) woke up-ppv. (they) looked-PFv at the jar and saw-ppv that the frog is not (there). (They) searched-PFv for it everywhere...'

Only adult bilinguals seem to counteract this shallowness in aspectual contrasts. which is typical of bilingual school-age children. But even among them, the use of the imperfective is to a greater extent limited to the expression of habituality and durativity, as illustrated in [21], from a bilingual adult: [21] El chico hablaba con el perrito, jugaba con el perrito. hacia todas las cosas con el perrito, aiin iba a dormir con el perrito. Un dia, cuando estaban caminando, encontraron un nuevo amigo... The boy talked-iPFv to the dog, played-iprv with the dog, did-ippv everything with the dog, even went-iPFv to sleep with the dog. One day, when (they) were walking. (they) found-pry a new friend...' (BSadult-31)

The bilinguals are also behind the monolinguals in use of the imperfective with dynamic-telic verbs (1% versus 3% of all verbs in the imperfective). Spanishmonolingual children as young as four show proficient and well-motivated use of the imperfective with verbs that are basically punctual such as salir 'exit' or meterse 'enter' ([22]): [22] Y el sapo siempre se metia adentro de un frasco que tenia el nene. 'And the toad always entered-iPFv inside ajar that had the boy.' (MS7e-7:5) In sum, Spanish-Hebrew bilinguals used the same aspectual variations as monolinguals except for the past perfect. However, there are differences in the way speakers exploit the spectrum of forms for aspectual contrasts to serve different narrative functions. The massive use of past-perfective forms and the lack of flexible and mature use of other aspectual distinctions often made bilinguals' narratives rather flat and monotonous in temporal texture. 4.4 Form-function relations: Non-finite forms Two non-finite forms are analysed in this context: infinitives and gerunds. Infinitives are mainly used as verbal complements with modal verbs such as querer 'want', poder 'can', tratar 'try'; with aspectual verbs such as empezari comenzar a + infinitive 'start to' (inchoativity); and as prepositional complements, e.g., se puso contento de ver a la rana '(he) was happy to see the frog'. They can also function as adverbials in inter-clause connectivity, generally expressing simultaneity or two concurrent events, as in [23]:


413

[23] Al caer, rompe el recipiente. 'On falling, (it) [the dog] breaks the vase.' (MS20h-20)

Gerunds, though different in form, fulfil similar functions in discourse. They can function as complements after aspectual verbs, e.g., salir/partir + gerund 'exit' (progressiveness), seguir + gerund 'go on' (continuativeness); and with verbs of perception, as in [24]: [24] Y de repente vieron seis ranitas chiquititas saltando al costado. "And suddenly (they) saw six tiny froggies jumping by the side.' Gerunds also function as adverbials for inter-clause connectivity and for the expression of simultaneity or immediate posteriority, as in [25]; and to indicate manner, in a way typical of verb-framed languages (Talmy 1985; Slobin 1996a), as in [26]: [25] Se sube Pedro a una piedra tratando de encontrar a la rana amiga. 'Pedro climbs a rock trying to find the frog friend.' (MS20a-20) [26] Se salio y se fue corriendo. '(It) [the frog] exited and went away running.' (BS4h-4;6) Hebrew non-finite forms differ from those of Spanish in several ways. As in Spanish, infinitives are used as complements of modal and aspectual verbs, e.g., race lesaxek 'wants to play', hitxil 'begin, start' (for inceptive aspect), himshix 'go on, continue' (for continuative aspect); but they are also the only verb form to follow prepositions, e.g., hu pa'al bli laxshov paamayim 'he acted without thinking twice'. Hebrew has nothing comparable to the gerund complements of Spanish. Other Hebrew forms translatable to gerund forms are adverbial gerunds to express the notion of "attendant circumstances", as in hu tarak et hadelet behikanso laxeder 'he slammed the door on entering the room' (Berman 1978: 297), and nominalizations, as in hem mamshixim bi-merucat-am 'they continue in-flight-their [= they continue with their flight]' (Berman & Neeman 1994:323). These are typically restricted to formal registers and are rare in children's usage and in spoken Hebrew (Ravid & Avidor 1998). As predicted, the overall use of non-finite forms (gerunds and infinitives) was significantly higher in monolinguals than in bilinguals (t = 3.2, p < 0.01). The monolingual group used these forms twice as much as the bilingual group: a total mean of 22 occurrences (- 8% of the total mean of clauses) versus 11 occurrences (4.7% of the total mean of clauses), counted in terms of tokens. There was also an increase in the number of non-finite forms with age in both groups, particularly large in the monolingual group (r = 0.67,p < 0.01 for monolinguals

414

Judy Kupersmitt

and r(39) = 0.39, p < 0.02 for bilinguals). Among four-year-olds in both groups, there was little use of non-finite forms (a mean of around one occurrence in each group). Later, this gradually increases from three to twelve occurrences in the monolingual group and from two to six occurrences in the bilingual group, in the seven-year-olds' and the adults' texts, respectively. Gerunds were used by monolingual school-age children and adults with verbs of aspect to mark continuativeness and progressiveness ([27]); and to express simultaneity between events, as in [28]: [27] Y siguieron buscando a la rana. 'And (they) continued looking for the frog.' (MSI la-11;2)

[28] Despues el perro se cae con el frasco en la cabeza y el chico mirando a ver si estaba la rana. 'Then the dog falls with the jar on its head and the boy looking to see if the frog was (there)/ (MS7d-7;2)

This last use means that children recognize that the gerund itself is the form that marks progressivity. The use of gerunds in this context was significantly different in the two groups, as it was almost non-existent among bilinguals (/ = 2.88. p < 0.01). Another large group of gerund forms served for clause linkage to express simultaneity, immediate anteriority, or attendant circumstances. This latter use increased wih age in the monolingual group (r (40) = 0.62. p < 0.01) and was frequent among monolingual adults (1.7 versus 0.6 occurrences in the monolinguals' and bilinguals' texts). Example [29], from a monolingual adult, illustrates this function: [29] Buscando a la rana, se van acercando al pequeno bosque. 'Looking for the frog, (they) go approaching the little wood.' (MS20c-20)

Among bilinguals, gerunds were used almost exclusively with the verb seguir 'go on' to mark continuativeness (3.8 out of a total mean of 4.6 gerunds). Here, the influence of Hebrew syntax was particularly noticed. In Hebrew, the aspectual verb himshix 'continue' is similar to its Spanish equivalent, but it is not followed by a gerund, as in Spanish, but by an infinitive, e.g., himshix lexapes 'continued to-search'. This yielded the following non-native, unconventional combinations in the bilinguals' texts ([30]): [30] Y despues salio afuera con el perro y siguieron buscar a la rana. 'And then (he) went outside with the dog and (they) continued to-search for the frog.'(BS1 la-11)


415

It is important to note that these non-native productions occurred in the same subjects' texts side by side with well-formed constructions such as y el nene siguio buscando a la rana 'and the boy continued searching for the frog'. This may indicate that the influence of Hebrew syntax on Spanish constructions is not absolute but rather determined by the exigencies of on-line production. That is, bilingual subjects certainly know the right Spanish construction, but at some point in on-line discourse production they produce Hebrew-like combinations. These are possibly triggered by the existence of an element common to both languages in a certain linguistic form (in this specific case, the verbs seguir and lehamshix 'continue'). Infinitives were used in both groups with the verb empezar 'start' to express inchoativity, and this use increased with age. However, only monolinguals (particularly adults) used infinitives as adverbials for the expression of attendant circumstances, as in [31]: [31 ] Al despertarse, Juan se queda sorprendido. "On waking up, Juan becomes surprised.' (MS20g-20)

In sum, differences in the use of non-finite forms between monolinguals and bilinguals were particularly marked in the use of gerunds for the marking of progressive and continuative aspect, and in the use of both gerunds and infinitives for more mature, complex functions such as inter-clause connectivity. 4.5 Form-function relations: Null subjects In generative terms, Spanish is considered a "pure" pro-drop language since it allows the omission of surface subject pronouns in all types of clauses (Borer 1984; Hyams 1989; Weissenborn 1992). Hebrew behaves in a more restricted fashion in that it licenses subject ellipsis in independent clauses in the past and future tense, but not in the present, and in the first and second person, but not in the third (Armon-Lotem 1996; Berman 1980, 1990). Between clauses, subject ellipsis is obligatory in Spanish in co-ordinated, subordinated, and juxtaposed clauses with the same subject, while in Hebrew it is optional in co-ordination and restricted in subordination and juxtaposition. Previous analyses of null subjects among Spanish-Hebrew bilinguals (Kupersmitt & Berman 2001) yielded the prediction that bilingual subjects would use null subjects as much as, and in similar contexts as, their monolingual peers. The figures in Table 4 corroborate the expectations: the total proportion of null subjects per number of clauses is similar in monolinguals and bilinguals. Overall, 40% of the clauses are subjectless, across age groups. Also, the distribution of different types of null subjects is similar in the monolinguals' and the bilinguals' texts (t = - 1.56, p > 0.1 for Type I Nulls; t = - 0.43, p > 0.5

Judy Kupersmitt

416

Table 4. Mean percentage of null subjects by type of construction, age. and linguistic background, out of the mean number of clauses Bilingu als. age

VIonolinjmals, age

Type of null subject*

adults

4;5

7;5

11:5

adults

4:5

7:5

Type I Nulls

6.5%

10%

10%

12%

7.8%

10.8%

11:5 12.8%

15.4%

Type II Nulls

10.5%

17%

20.7%

20.3%

7.8%

13.4%

28%

23.6%

Type III Nulls

29%

7.6%

3%

2%

20%

8.6%

4.5%

1 .6%

Type IV Nulls

9%

0

0

0

5.5%

0

0

0

34% 35% 55.5% 34.2% Total % * Type of null subject: Type I Nulls: syntactically obligatory null subjects; Type II Nulls: discourse-licensed null subjects; Type III Nulls: situationally licensed null subjects; Type IV Nulls: deictic null subjects.

41.2%

33%

45%

41%

for Type II Nulls; / (77) = 0.55,;? > 0.5 for Type III Nulls; and / ( , T) = 0.22, p > 0.5 for Type IV Nulls). Null-subject types I and II increased with age in both groups. The first are syntactically required null subjects, found in co-ordinated, juxtaposed, and/or subordinated clauses, as in [32], by a seven-year-old bilingual: [32] Y cuando el nene se desperto, 0 vio que no estaba la ranita. 'And when the boy woke up, (he) saw that was not the frog [= that the frog was not there].'(BS7g-8;2) In [31], the null subject appears in a subordination construction and it refers to the same referent as the subject in the immediately anterior subordinate clause. The results also showed that none of the narrators in either group used overt subject pronouns in co-ordinated or subordinated clauses. This means that both monolingual and bilingual narrators know the grammatical constraints regarding deletion of same-subject pronouns in Spanish. Type II Nulls are discourse-licensed null subjects, where subjects are dropped across non-adjacent clauses with different subjects, for the purpose of connectivity. The following excerpts, Examples [33a] and [33b], from an elevenyear-old monolingual and an eleven-year-old bilingual relating to the same scene, illustrate their use: [33a] El chico le dice que se calle al perro. Entonces 0 vio algo. 0 Miran detras del palo, y 0 ven dos ranitas. 0 Se dan vuelta, y 0 vieron que 0 tuvieron hijitos. Entonces los padres, o sea las ranitas le regalaron una ranita.


417

The boy tells to be silent to the dog. Then (he) saw something. (They) look over the log, and (they) see two froggies. (They) turn around, and (they) saw that (they) [the frogs] had little children. Then the parents, in other words the froggies, gave him a little froggie.' (MSlle-11;5)

[33b] Y el nene penso que 0 sabe donde esta la ranita. Y 0 le dijo al perrito que este en silencio, y 0 fueron a buscar a traves de un arbol. Ahi 0 vieron la ranita, y un sapo y nenes chiquititos, 0 eran muy felices. Ahora 0 le pidieron al sapo y a la ranita que le den un sapito chiquitito, y 0 fueron a la casa. 'And the boy thought that (he) knows where is the frog. And (he) told the little dog to be silent, and (they) went to search over a tree. There (they) saw the frog, and a toad and little kids, (they) [the toads] were very happy. Now (they) asked the toad and the froggie to give them a tiny toadie, and (they) went home.' (BS7h-10;5)

These examples show that subjects skilfully maintain reference to three protagonists - the boy, the dog, and the frog - who are sometimes paired, that is, the boy with the dog, the frog with another frog or frogs. This is done by means of topic elision in independent clauses, to keep track of the characters along a certain piece of text in a clear and unambiguous way. Thus, Type II Nulls serve a discourse function rather than a grammatical function. In this respect, SpanishHebrew bilinguals behave differently from monolingual speakers of Hebrew. In the former group, particularly among older children and adults, topic elision across long stretches of the text is a frequently used, favoured device for topic maintenance. In the latter group, this is an optional device, whose use is generally determined by individual rhetorical style. Example [34], from an eleven-year-old Hebrew-speaker narrating the same scene as in [33], illustrates how, in contexts where Spanish-speakers (both monolingual and bilingual) drop the subject, many Hebrew-speakers opt to leave it in: [34] Hakelev saxa lakivun ve 0 raca linboax, ax Dani amar lo she yihiye besheket. hem tipsu me'axorey geza ha'ec haxalul, ve 0 gilu shney cfarde'im - ba'al ve isha. lefeta hem ra'u gam et kol yaldey'hem. Dani bikesh mehem - cfardea exad ve hem natnu lo, ve hu xazar habayta. The dog swam to the direction and (he) wanted to bark, but Dani told him to be quiet. They climbed behind the hollowed log, and (they) found two frogs husband and wife. Suddenly they saw also all their children. Dani asked from them - one frog and they gave him, and he returned home.' (HI lg-11 ;5)9

Type III Nulls and Type IV Nulls show a different developmental trend than Type I and II Nulls. While types I and II increase with age, types III and IV decrease, as illustrated in Table 4 above. Type III Nulls or "situationally

418

Judy Kupersmitt

licensed null-subjects" (Berman 1990) represent half of the null subjects used by the younger children and less than 5% of those used by adults, among both monolinguals and bilinguals. The subject referents are recoverable not from the preceding linguistic context but from general knowledge of the situation, in this case knowledge about the characters in the narrative and their functions. In fact, their use does not require the ability to organize a well-maintained referential sequence. Rather, pronouns are omitted randomly, leaving to the listener the task of overcoming those "referential gaps". This is illustrated in the following excerpt ([35]) from a four-year-old monolingual, who freely jumps from one referent to the other: from the boy to the owl, via the boy and the wasps, and back to the boy: [35] Y el chico se subio a un arbol y llamo: jsapo! y salio una lechuza. 0 Se asusto. 0 Se cayo. Y salieron todas las avispas corriendo al perro. Y 0 tenia miedo de la lechuza. 0 Se subio a una roca. 'And the boy climbed a tree and called: toad! and exited an owl. (He) got frightened. (He) fell. And exited all the wasps running after the dog. And (he) was afraid of the owl. (He) climbed a rock.' (MS4f-4;5)

Here, too, occurrences of Type III Nulls reflect an incompatibility at the syntax-discourse interface. There is nothing grammatically incorrect in these null subjects, but the inability to connect the clauses in the text into a cohesive, thematically unified piece of discourse yields less-interpretable null subjects "hanging" in a sequence of isolated clauses. Finally, Type IV Nulls appear only among the youngest, in both monolinguals' and bilinguals' texts. These are deictic-pronoun omissions, characteristic of a picture-by-picture mode of narration. Also, in these cases, null subjects refer to characters that have not been introduced, as shown in [36] by a bilingual child narrating the beginning of the story: [36] La rana adentro. 0 Duerme. Y la rana aca sale. Ahora 0 esta despierto y 0 ve que la rana se fue. The frog inside.(He) [the boy] sleeps. And the frog here exits. Now (he) [the boy] is awake and (he) sees that the frog is gone.' (BS4g-4;9)

From a contrastive perspective, the findings for bilinguals' and monolinguals' use of null subjects for discourse connectivity and topic maintenance show a striking similarity. At first, most null subjects have a grammatical rather than a discourse function. They are sometimes used in a deictic manner, in isolated clauses. With time, null subjects are used across larger stretches in the text, for the purpose of topic maintenance in thematically unified pieces of the narrative. Thus, in both


419

monolinguals' and bilinguals' texts, subject pronouns are dropped across the board, except when used deictically by some of the youngest subjects in both groups or, conversely, when mentioned for disambiguation or contrast by older children and adults, as in [37]: [37] El biiho lo asusta al nino y 0 lo ataca, hasta que el [el nino] se sube a una piedra. 'The owl frightens the boy and (it) attacks him, until he [the boy] climbs onto a rock.'(MSadult-20b)

4.6 Form-function relations: Word order As Spanish is characterized as a "free inversion" language (Rizzi 1986), it was assumed that the monolinguals' texts across age groups would present considerable manipulation of word-order variation, as compared with the bilinguals' texts. Spanish-Hebrew bilinguals would adhere to the more conservative SV(O), under the influence of Hebrew (Dromi & Berman 1986; Ravid 1995), and they would diverge to VS word order mainly when syntactically required. The results in Table 5, which show the number of occurrences of VS word order in each age group in monolinguals and bilinguals, partially confirm the predictions. Overall, monolinguals surpass their bilingual counterparts in occurrences of VS word order. This tendency is most marked among younger children (78 and 74 occurrences versus 23 and 27 in the four- and seven-year-old groups of monolinguals and bilinguals, respectively), and less so among older children and adults. Syntactically determined predicate-initial constructions occurred with the verb estar1" for presenting the characters ([38]), and with locations ([39]): [38] Una noche estaba Juan con su perro Tomy... 'One night was Juan with his dog Tommy...' (MS20i)

[39] Y aca estaba el pajaro. 'And here was the bird.' (MS4b-4;2)

This predicate occurred in verb-initial position among monolinguals three times more than among bilinguals (32 versus 10 occurrences). Bilinguals preferred the verb haber 'be' (existential) in the present form hay '(there) is/are' or in the past form habia '(there) was/were' to the copula verbs ser or estar 'be' for the presentation and location of referents and characters. These forms correspond to the Hebrew particle yesh 'there is/are' in the present or haya '(there) was:

Judy Kupersmitt

420

Table 5. Total of occurrences of VS word order according to context of appearance and age group Context of VS occurrence

Monolinguals, age (N)

Bilinguals. age (N)

4;5

7;5

11;5

adults

4;5

7;5

11:5

adults

(10)

(11)

(10)

(9)

(10)

(11)

(10)

(8) 5

With the verb estar 'be'

8

8

6

10

With clitic pronouns

9

14

3

1

1

6

0

4

With changeof-state verbs

27

29

11

16

9

9

12

16

With presentatives

34

21

14

15

6

8

11

10

For literary style

0

2

1

2

3

4

1

11

Total occurrences ofVS

78

74

35

44

23

27

25

46

4

0

1

MASC:SING', hayta '(there) was:FEM:SING', hayu '(there) were' in the past, used in existential and possessive constructions.This over-use of the verb haber was found across all ages but was more notable among the four-year-olds, as in [40] by a four-year-old bilingual: [40] Y aca hay dos sapos. Y aca hay dos, y aca hay una mama y un papi y aca hay muchos. Aca hay uno en el agua y uno en su mano. 'And here (there) are two toads. And here (there) are two, and here (there) is a mother and a father, and here (there) are many. Here (there) is one in the water and oneinhishand.'(BS4b-4;5)

VS word order with object clitics occurred mainly among younger monolinguals, for topic maintenance, as in [41], or for manipulating narrative perspective, as in [42], by a seven-year-old monolingual, where the dog, which is more topicworthy than the bees, is preposed before the object clitic, and the bees are defocused in a post-verbal position: [41] Y se cayo el perro y \o alzo el nene. 'And fell the dog and it-Ace picked up the boy." (MS4d-4;3)


421

[42] El chico ve ahi adentro del arbol que habia un buho....y al perrito ]e estaban persiguiendo todas las abejas. 'The boy sees there inside the tree that (there) was an owl.. .and the dog it-Ace were chasing all the bees.' (MS7b-7)

Bilingual children know the grammar of object-clitic constructions, as they use clitics widely in their texts. However, they do not seem able to "exploit" this device in subject-verb inverted constructions for pragmatic purposes. With age, subject-verb inversion with object clitics becomes a less favoured option for topic maintenance and manipulation of topic perspective. Older children and adults maintain topic by means of more complex rhetorical options, such as passive constructions ([43]) or relative clauses ([44]): [43] Del agujero sale un buho que sorprende al nino y lo lira del arbol mientras el perro es perseguido por las abejas. 'From the hole exits an owl that surprises the boy and throws him from the tree, while the dog is chased by the bees.' (MS20d)

[44] Y al mismo tiempo el perro tenia problemas con un monton de abejas que le iban persiguiendo. 'And at the same time the dog had problems with a lot of bees that were chasing hjm.'(MS20i)

Bilinguals, in contrast, use simpler constructions such as causal subordination to maintain the dog as topic, as in [45]: [45] El nene se cae y el perro esta corriendo porque todas las abejas lo estan corriendo a el. 'And the boy falls and the dog is running because all the bees are running to him.' (BSadult-23)

In both groups across ages, the largest proportion of VS word order occurred with change-of-state predicates (e.g., se cayo el panal de abejas 'fell the beehive') and with presentative predicates. These share semantic and syntactic features of "unaccusative" predicates". The examples in [46] illustrate a pragmatically motivated deviation from standard SV order with presentative predicates to introduce a character, by a monolingual and a bilingual, respectively: [46a] Despues el nene se cayo del tronco del arbol y aparecio una lechuza. Then the boy fell from the branch of the tree and appeared an owl.' (MS7i-7;4)

422

Judy Kupersmitt

[46b] Y las abejas estaban bbbzzzzzz, y salio del arbol un buho. 'And the bees were bbbzzzzz, and exited from the tree an owl.' (BS7a-7;9)

The cases of VS word order with presentative predicates included also predicates referring to the disappearance of a protagonist, as in [47]: [47] Y en la manana el nino vio que se escapo la rana. 'And in the morning the boy saw that escaped the frog/ (BS7c-7;3)

The figures in Table 5 indicate that VS word order with change-of-state verbs and presentative predicates decreases with age among monolinguals but increases with age among bilinguals. This, in turn, yields sharp differences in use between the youngest bilingual and monolingual subjects (the four- and seven-year-olds), but not between older children and between adults. Among four- and sevenyear-old monolinguals, VS word order with these predicates is pervasive, that is, whenever these verbs occur with an overt subject, the subject occupies a post-verbal position. Thus, if the grammar of their language allows it, as is the case for Spanish, VS is preferred to SV in this context. This purely grammatical fact interacts with pragmatic factors. It could be that in the flow of a narration, children are more interested in the result than in the process or, relatedly, in the protagonists rather than in the events. This is illustrated in [48], from a four-yearold bilingual: [48] Y despues vieron por ahi y no vieron nada. Y despues se rompio el frasco y se cayo el perro. 'And then (they) looked over there and (they) saw nothing. And then broke the jar and fell the dog.' (MS4d-4;3)

In [48], the two change-of-state verbs (romper 'break' and caer 'fall') appear in verb-initial position, so the focus is on the subjects (the jar and the dog). Among young bilinguals, there were few occurrences of SV inversion with change-of-state verbs and presentatives. This could be due to a lexical deficiency and the influence of Hebrew. In the first case, bilingual children avoided describing a certain event, since they did not know a specific term such as the names of animals or other objects in the story. In [49], a bilingual girl suspends the presentation of the animal by asking a question, and introduces it (the deer) only after the interviewer tells her the specific term: [49] Y busco y busco, y mira que encontro...un.... ciervo. 'And (he) searched and searched and look what (he) found....a ...deer.' (BS7e-7;l 1)


423

Besides, as stated about Example [40], it is common for bilinguals to replace a presentative predicate by the existential verb haber 'be'. This is further illustrated by the contrasting examples in [50] from a four-year-old monolingual (a) and a four-year-old bilingual (b), relating to the deer scene: [50a] Y despues aparecio un caballo y el nene se estaba subiendo al caballo. 'And then appeared a horse and the boy was climbing onto the horse.' (MS4g-4;5)

[50b] Habia una piedra grande y habia un caballo. '(There) was a big stone and (there) was a horse.' (BS4i-4)

Older children and adults in both groups use SV inversion as a rhetorical device for different purposes in the narrative, alternating VS and SV word order (with the same predicates) for change of perspective, or for focusing on the process/event rather than on the referent of the grammatical subject. Examples [5 la] and [5 Ib], from a monolingual and a bilingual adult narrating the same scene, show this use: [51 a] Se asoman a la ventana, empiezan a llamarla a los gritos, y en un momento el perro se cae y rompe lo que era la casita de la rana. '(They) lean out of the window, (they) start to call her with shouts, and in a certain moment the dog falls and breaks what used to be the house of the frog.' (MS20e)

[51b] Abren la ventana y gritan: ^donde estas? jrana, rana! Y no encuentran. El perro se cae por la ventana y rompe el frasco. '(They) open the window and shout: where are you? frog! frog! And (they) don't find. The dog falls out of the window and breaks the jar.' (BSadult-26)

In [5 la] and [51b], the verbs 'falls' and 'breaks' appear after the subject, el perro 'the dog', for the purpose of topic change (that is, to the dog) and to focus on the event (the falling and subsequent breaking of the jar). Finally, some instances of VS word order were found among bilinguals (particularly adults) to achieve a literary style of narration. These inversions occurred with verbs such as decir 'tell/say', llamar 'call', or ver 'see', e.g., "esto es un ciervo " dijo el chico ' "this is a deer", said the boy' (BSadult-31), and were found mainly in direct-speech constructions, which were common in bilingual texts across all age groups. This may indicate a formally simple device which gives the narrative a literary, formalistic style, in keeping with the style of much Hebrew child literature (Ziv 1988). In sum, bilinguals showed far more conservative behaviour regarding VS word order in all contexts.

424

Judy Kupersmitt

4.7 Form-function relations: Relative clauses Relative clauses are largely similar in Spanish and Hebrew. They are post-nominal, and are marked in both languages by an obligatory, invariant complementizer (que in Spanish and she in Hebrew), which is not marked for categories such as gender, number, or animacy (Dasinger &Toupin 1994; Fernandez Lagunilla & Anula Rebollo 1995). In clauses that relativize oblique NPs, Hebrew requires the use of a resumptive pronoun, e.g., ha-kelev she Dany sixek ito 'the dog that Dany played with-it' (Berman & Neeman 1994), and Spanish uses more specific relative markers such as quien 'who', cuyo/cuya 'of whom', or donde 'where', some of them preceded by a preposition as demanded by the subordinated predicate, e.g.. la persona con quien hable trabaja conmigo 'the person with whom (I) talked works with me'12 As expected, I found that the use of relative clauses increased with age in both groups, from 1% of all clauses in the texts of the youngest monolinguals and bilinguals to 8% and 4% in the texts of monolingual and bilingual adults, respectively. This is partly explained by the fact that little children connect clauses by means of the conjuction v 'and' and not by means of the relativizer que 'that', as in [52a] and [52b] by a four-year-old monolingual and a four-year-old bilingual: [52a] Y le salio un raton y le mordio la nariz. 'And for-him exited a mouse and for-him bit the nose.' (MS4f-4:5) [52b] Y el nene vio un raton y tenia olor feo. 'And the boy saw a mouse and (it) had a bad smell.' (BS4j-4;l 1) Presentative relative clauses were the most common in both groups, used to describe or provide information about a character or some other referent in the story. This type of relative clauses occurred in a similar proportion in both groups across all ages (/\ 7 1 | = 1.2, p > 0.1). In the monolinguals' texts, they evolved from being locally used in isolated clauses (Example [53]. by a four-year-old monolingual) to being globally used in multiple subordination (Example [54]. by an adult monolingual): [53] Habian un chico y un perro que estaban mirando una rana. Y despues se escapo el sapo. '(There) were a boy and a dog that were looking at a frog. And then escaped the toad.'(MS4e-4;5) [54] Tomasito empezo a correr hacia el barranco que habia en ese lugar. y donde abajo estaba el rio que seguramente era el lugar donde este ciervo queria desquitarse.


425

'Tomasito started to run towards the cliff that was in that place and where below was the river that certainly was the place where this deer wanted to take revenge/ (MS20f-20)

Among bilingual adults, this development towards more hierarchically embedded relative clauses is less marked. Most descriptive clauses in this group resemble those used by younger children, since they are more local and less globally oriented, as in [55]: [55] El nene esta buscando en un agujero que hay en la tierra. 'The boy is searching in a hole that is in the ground.' (BSadult-23)

This is true for other types of relative clauses, too, as illustrated by the use of continuative clauses (those that serve to advance the narrative plot) in [56a] and [56b], from a monolingual and a bilingual adult: [56a] Eran los cuernos de un ciervo que, molesto por la nueva compania, lo monto a Pedro arriba de el. '(They) were the horns of a deer that, bothered by the new company, him, mounted Pedro, on it [= mounted Pedro on itself].' (MS20a-20)

[56b] Y despues viene un alee que lo alza al nene. 'And then comes an elk that him | raises the boy r ' (BSadult-26)

In [56a], the continuative relative clause is split up, arranged in a non-linear mode of one clause embedded into the other. In contrast, [56b] illustrates a simply embedded construction, attached linearly to the object NP. A significant effect of both linguistic background and age was found for relative clauses used with a retrospective function (f = 3.31, p < 0.03). This may be attributed to the absence of past-perfect aspect in bilinguals' texts, and is well illustrated by the following two openings of the frog story, by a monolingual ([57a]) and a bilingual ([57b]), respectively: [57a] Habia un chico con un perro que habian atrapado a una rana. '(There) was a boy with a dog that had captured a frog.' (MS7g-7;3)

[57b] Un dia habia un nenito que tenia una ranita y un perro. 'One day (there) was a little boy that had a little frog and a dog.' (BS7g-8;2)

In [57a], the relative clause has a retrospective function: it situates the listener in a time before the reference time of the story. In [57b], however, the relative clause has a descriptive/presentative function, since it states a fact about the character

426

Judy Kupersmitt

presented (the boy) and that fact coincides with the reference time of the story. These examples reflect the meeting point between syntax and discourse, in which the availability of certain linguistic devices allows a different manipulation of the discourse flow. To sum up, with age, relative clauses evolve to perform more functions in more syntactically and discursively complex contexts, leading to a considerable gap between the form - which is certainly known by monolinguals and bilinguals alike in all age groups - and the function - which is extended among monolinguals and narrowed among bilinguals. From the point of view of contrastive rhetoric, both quantitative and qualitative differences in the use of this device could be attributed to the influence of Hebrew rhetoric, which, apparently, makes less use of relative clauses in extended discourse than does Spanish (Dasinger & Toupin 1994:495). 4.8 Command of specific linguistic devices The characterization of language development in monolinguals and bilinguals was addressed by means of an error analysis, which was intended to evaluate the command of specific linguistic devices in both groups. Core-grammar errors included a range of violations at the morpho-syntactic level, such as inappropriate gender, number or case marking in determiners, clitics, and/or pronouns; preposition errors; se particle errors; inappropriate use of the double copula ser and estar 'be'; and omission or insertion of words belonging to certain grammatical categories. Many more grammatical errors occurred in the bilinguals' than the monolinguals' texts. For example, of the seven-year-olds, an overall mean of 0.27 children produced the different types of errors in the monolingual group versus 2.8 in the bilingual one. Developmental errors (see Section 1) occurred among young monolinguals and bilinguals. They included the following: /.

Errors in object-clitic agreement features, e.g., el chico lo_ llamaba a la rana 'the boy it-ACC.MASc called-siNG to the frog' (MS4e-4:5). where masculine lo is used instead of la to refer to rana 'frog' (feminine);

//.

Errors in the use of the double copula ser and estar 'be', e.g., cpor que no esta xamuda^? 'why isn't it [the frog] cute?' (BS4d-4;l 1), where esta (from estar) is used in place of es (from ser), which denotes permanent attributes'-1;

//'/.

Errors in the use of the se particle, e.g., la ranita se_ solid de la caja 'the frog RFL-exited the box' (BS4c-4;9), where the particle se is inappropriately 15 used with the verb salir 'exit/go out', analogously to


427

the correct use of this particle with other verbs of motion such as bajar 'get down' or subir 'climb', in which case it indicates completion of the activity or episode boundary (Sebastian & Slobin 1994); iv.

Over-generalizations, such as adding the plural morpheme to the existential verb haber 'be', e.g., habian dos kocim16 '(there) were-PL two thistles' (BS4i-4), where the verb is marked with the plural morpheme -n to agree with the subject 'two thistles' while the normative form would be the singular habia for both singular and plural subjects; and

v.

Cases of verb regularizations, e.g., dijia for decia 'said-iPFv-3RD-siNG', from the form dijo 'said-PFV-3RD-siNG' (BS4c-4;9).

These errors reflect instances of "breaking points" in the grammar, that is, difficult, structurally opaque forms (as is the case for object clitics, whose morphology varies according to case and gender), as well as multi-functional grammatical devices (Karmiloff-Smith 1979; Slobin 1977). For example, the morpheme se is multi-functional, since it is used in different contexts with different functions: as a reflexive marker, e.g., sepeina el cabello 'for-himself/herself brushes the hair'; as a marker of reciprocality, e.g., los dos amigos se abrazaron 'the two friends eachother hugged'; to indicate inception in change-of-state verbs such as sentar-se 'sit down' or despertar-se 'wake up'; and as an optional marker of completion with verbs of motion, as in se bajo 'got down' and se subio 'climbed'17. Errors categorized as "developmental" occurred in the grammar of bilingual school-age children and adults, who, apparently, are not fully able to cope with these difficult grammatical paradigms even later on in their development. This is illustrated in [58] and [59], by a seven-year-old bilingual and a bilingual adult: [58] Y el nino la esta llamandoia. 'And the boy it-Acc-FEM is calling-it-ACC-FEM.' (BS7i-7;ll)

[59] Lo busco en la casa, adentro de los zapatos. '(He) it-Acc-MASC searched in the house, inside the shoes.' (BSadult-26)

Example [58] shows an error in a construction with an object clitic, in which the object clitic is inserted twice: preposed to a finite verb and postposed to an infinitive, which leads to a redundancy18. In Example [59], the object clitic lo (masculine) refers to a feminine referent (the frog), in which case the correct form would be la. These errors, which reveal juvenile strategies applied to handle difficult grammatical paradigms (Berman 1986), are considered as "transient" in normal

428

Judy Kupersmitt

language development. However, the results show that in this particular bilingual situation these same procedures become fossilized in the language of bilinguals. Transfer errors in bilinguals' texts were attributed to the influence of Hebrew grammar on the use of highly language-specific devices, such as prepositions, e.g., la llamaron de_ la ventana, corresponding to Hebrew kar 'u la me ha-xalon '(they) called it from [= through] the window' (BS7e-7;ll) for the correct la llamaron desde la ventana', or y lo agarro al nene en_ la cabeza, corresponding to the Hebrew tafas et ha-yeled ba-rosh 'and (he) caught the boy on the head" (BS11 f-12; 1) for the Spanish y lo agarro al nene de_ la cabeza. The likelihood that prepositions will be susceptible to influence from the dominant language derives from the fact that these are highly language-specific lexical elements which denote different perspectives in the expression of concepts such as movement and space (Bowerman 1985; Slobin 1996a). Other frequent transfer errors included omission of obligatory markers, as in [60], where the noun drbol 'tree' should be preceded (marked by "#") by the indefinite masculine article un ART:INDEF:MASC:SG. In Hebrew, indefinite nouns take zero marking, as in [61]: [60] El chico estaba buscando adentro de un tronco de # arbol. 'The boy was searching inside a branch of # tree.' (BS1 lc-12;l) [61] hu ra' a xor ba' adama 'he saw (a) hole in the ground' (H11 a-11 ;4)

Insertion of unrequired categories was less frequent. One example is a redundant use of the subordinator que corresponding to the subordinator she in Hebrew, e.g., en la noche, cuando que ya se fueron a dormir 'in the night, when that (they) already had gone to sleep' (BS1 le-12) (cf. cuando ya se fueron a dormir}, corresponding to Hebrew she kvar alxu lishon 'when (they) had already gone to sleep'. Other transfer errors were incorrect use of the infinitive after certain verbs of aspect, such as the verb seguir 'continue', e.g., y siguieron buscar a la rana 'and (they) continued to-search for the frog' (BSlla-11), instead of siguieron buscando 'continued searching' (see the discussion of this example in the section on non-finite forms). Another group of errors could be attributed to individual strategies used to cope with the exigencies of on-line text production, e.g., y el perro le esta siguiendo ladrando 'and the dog to-him is continuing barking.' (BS7i-7;ll). where two consecutive verbs - the aspectual verb and its complement - are in the progressive aspect, resulting in an ungrammatical construction. Besides errors in core grammar, lexical errors were found in almost all of the bilinguals' texts, indicating that subjects had poor access to specific lexical items. Pre-school and school-age subjects often replaced the unretrieved or


429

unknown word by a word or phrase that retained or referred to some of its semantic properties ([62-63]): [62] El nene busco en un pozo y el perro en una casa de abejas. The boy searched in a well and the dog in a house of bees [= beehive].' (BS7f-8;l)

[63] Entonces vieron miel. 'Then (they) saw honey [= beehive].' (BS7c-7;3)

They also replaced a word by a very general term (Example [64]), or by the Hebrew equivalent (marked HEB in Example [65]): [64] Y aca salio alguien. 'And here exited someone [= a squirrel].' (BS4d-4;l 1)

[65] Y el nene esta viendo el bor-HEB 'And the boy is looking at the hole.' (BS4a-4)

Older children and adults would sometimes use these same strategies, but they also opted to ask the interviewer about the exact lexical item. Olshtain (1989) reports similar strategies used by children and adults who speak English as a second language in a Hebrew-dominant context. Interestingly, most of the strategies used by bilinguals in this study were found also in the texts of young monolinguals, a fact that suggests another possible developmental lag in the language of bilinguals.

5. DISCUSSION 5.1 Bilinguals' versus monolinguals' form-function relations in narrative The results show that the differences in linguistic expression between monolinguals and bilinguals lie mainly in the extension and diversification of the uses of morpho-syntactic forms for discourse purposes. The development of form-function relations indicates two "breaking points" in the language of bilinguals as compared with monolinguals. One is detected in the route taken by forms in acquiring more mature and complex narrative functions. This appears to be "truncated" somewhere along the developmental path, so that forms which have already been acquired are generally used in their most basic function. An example is the use of the imperfective aspect. At a very early stage in development, monolingual children use the imperfective with inherently durative predicates, such as statives. Later, they are able to extend this semantic distinction to other

430

Judy Kupersmitt

types of predicates so as to denote the contrast between telicity and atelicity, e.g.. corrio (perfective) 'ran' versus corria (imperfective) 'ran', even with predicates that are inherently punctual such as salir 'exit'. However, bilinguals (in the context of this study) seem never to reach the later phases and so continue to use imperfective inflections mainly with states and activities. The second "breaking point" is reflected in the absence of later-acquired forms, analogously to the absence of later-acquired functions. One such form is the past perfect, used to refer to an event in the narrative that is anterior to the narrative time, which was found to be a late use in other Spanish monolingual samples (Sebastian & Slobin 1994). The development of form and function concerns also the embedding of forms in the on-going flow of discourse: with development, more forms are inserted into more complex syntactic and discursive contexts, to serve more globally oriented narrative purposes. The texts of bilinguals show little such mutual enrichment of form and context. There, forms appear in more isolated contexts and serve local rather than global functions, as in the cases of relative clauses, connectivity devices, and grammatical aspect. Clearly, in a languages-in-contact situation, this "under-development" is also influenced by the properties of Hebrew, the dominant language. The results show that this inter-language influence is not absolute in the sense that it eradicates nonequivalent constructions in the other language. Rather, it seems to "block" access to these forms for the speaker. The degree of accessibility of non-equivalent forms depends on several interacting factors. Non-equivalent grammatical categories that express different conceptual distinctions (e.g., temporal notions such as boundedness and progressivity) are accessible to the speaker with their core meaning - as in the case with imperfective aspect and gerunds - but they do not mature to serve as more widely used rhetorical devices. In contrast, categories that differ only at the structural level, such as null subjects, are less prone to influence by the dominant language, and they will generally be used in a nativelike way, as shown by the results in this study. Moreover, the degree of accessibility of certain form-function mappings is largely determined by the rhetorical characteristics of Hebrew as the dominant language. In this sense, Hebrew does not interfere with the grammar, but rather with the way a particular set of grammatical features in the language are used for rhetorical purposes. One example of this kind of influence is in the use of VS word order with different predicates for the expression of various functions in narrative. Again, bilinguals clearly know the grammar of VS constructions in Spanish. However, they prefer to adhere to the more conservative SV word order, in line with Hebrew, in which deviations from standard word order are mainly pragmatically determined and, overall, a less-favoured rhetorical option. The changes undergone by the Spanish of bilinguals. as reflected in their texts, can also be related to basic typological features of the (first) language. In


431

this analysis, it was assumed that "resilient" forms would be those that are: a) semantically transparent and easily processable; and/or b) pervasive in the sense that they are acquired early and are very central in the language, independently of their syntactic or semantic complexity. In contrast, complex or opaque linguistic structures would be candidates for being lost or replaced by simpler, more transparent forms. But these conditions would apply only to the acquisition of isolated grammatical forms. In the context of this study, in which forms are analysed in relation to the function they perform in narrative, this would yield erroneous interpretations. For this reason, the analysis must take into account the speaker's need to be expressive (Slobin 1977). This means that speakers will pay attention to typologically salient features in their language and develop a rhetorical set (Slobin 1996a) which includes more elaborate and complex grammatical structures. In view of this discursive perspective for language change, forms will be more "vulnerable" to change if they fulfil at least one of these conditions: a) they are multi-functional in the sense that they evolve to serve more narrative functions, and, as this happens, they become more detached of their core, unmarked meaning; and b) they appear in different semantic, syntactic, and/or lexical contexts. It is thus important to differentiate between forms that are grammatically resilient and forms that are rhetorically resilient. Null subjects, for example, were found to be very resistant to change, although they are differently constrained in the two languages. This can be explained in the following terms: They are acquired early and their use is pervasive (Austin 1997; Lopez-Ornat 1994). Their production can be considered a purely syntactic operation, since they do not represent any semantic content. Besides, they are easily processable since they are recoverable from the syntactic context in which they appear: adjacent to a verb marked for person and number. So, null subjects seem to fulfil all the conditions necessary to "survive" in the Spanish of bilinguals, both as a purely grammatical device and as a rhetorical option. Gerunds, in contrast, are maintained as grammatical but not as rhetorical devices. They are used in combination with an auxiliary to express the progressive aspect from an early age (Sebastian & Slobin 1994), and they are semantically transparent: the -ando/-endo suffix indicates progressivity and is added to the verb stem. But, in the course of development, this same form becomes structurally complex and semantically opaque, as it becomes detached from its core meaning and appears in different contexts: in combination with various verbs of aspect, without the auxiliary for the expression of manner and simultaneity, and as an adverbial participle. So, in the texts of bilinguals, this form appeared mainly with its core meaning, but was rarely used in varied syntactic environments, as a rhetorical option for the expression of simultaneity or manner, or for connectivity purposes.

432

Judy Kupersmitt

5.2 Bilinguals' versus monolinguals' language development The findings from the error analysis indicate two possible language-developmental profiles with respect to the acquisition and development of linguistic forms in bilinguals. One is characterized by a developmental deficit, in the sense that several linguistic forms become "fossilized" at a certain point in development - with the result that the grammar of bilinguals resembles the grammar of a monolingual in the early phases of acquisition. This means that, to a certain extent, the grammar of bilinguals is far from having reached anything like the grammar of fully proficient speakers (Berman 1986; Berman & Slobin 1994), who make skilful use of the full repertoire of linguistic forms in different discursive contexts and are able to apply grammatical rules constrained by the conventions and norms of adult usage. In this sense, I would suggest that while bilingual speakers in this study know the grammar of Spanish, they are far from being proficient speakers. The second profile indicates that bilinguals follow a qualitatively different path from monolinguals in the construction of grammar. In this sense, the construction of specific linguistic paradigms is affected by the influence of Hebrew grammatical features, leading to a kind of "interlanguage" (Andersen 1984; Olshtain 1989).

6. CONCLUSION This study aimed to examine to what extent bilingual speakers are sensitive to the availability of those language-particular devices at the levels of grammatical morphology, clause-internal structure, and inter-clausal connectivity in their nondominant language which are needed as rhetorical devices when constructing discourse. In other words, do bilingual speakers really develop a Spanish "typology of rhetoric" guided by the Spanish "typology of grammar" (Slobin 1996a)? In light of the results on the development of form-function relations and in light of the error analysis, I suggest the following: Bilingual speakers are guided by the salient typological features of the LI in the construction of grammar, as reflected in their knowledge of the "core grammar" of Spanish. They correctly use co-ordinate and subordinate constructions, almost the entire range of aspectual forms in the language, null subjects, word-order alternations, and relative clauses. However, most of these forms have a short-lasting developmental trajectory, in that they never evolve to become fully-fledged rhetorical means of expression. The findings of this study should clearly be situated in a multi-dimensional frame where age and linguistic background; psycholinguistic factors such as structural complexity and/or semantic transparency; and language typology are all v iewed as interacting with the forms and functions of the target language - the way these forms are selected as rhetorical options in the flow of discourse (Berman 1998b; Slobin 1997b). As a result, a theory on the acquisition and development of


433

language in bilinguals should take into consideration all these interactive levels of analysis.

NOTES 1.

This chapter is based on a previous work titled "Form-function relations in the narratives of Spanish-Hebrew bilinguals: What stays and what goes". I am deeply grateful to Prof. Dan Slobin for his thorough review of that work and for his insightful comments on the present paper. I also want to thank Prof. Ruth Berman for her invaluable comments.

2.

For example, Slobin has claimed that in bilingual-contact situations, Indo-European languages partially lose their inflectional systems over time; while in Turkish, the transparent analytical system of agglutinative inflectional morphology remains intact (1977:190-191). In the case of Spanish, subject-less constructions and verbal clitics seem to be less susceptible to change than other categories, as found by SilvaCorvalan in her (1994) study on language contact and change in a Spanish-English bilingual situation in Los Angeles.

3.

The elimination was done after carefully examining the questionnaire and listening to the narratives, especially in cases where the vocabulary test seemed not really to reflect the ability of the subject to handle the task. For example, there were a number of subjects who did well on the vocabulary test but whose questionnaire and/or text reflected a poor ability to perform the task. On the other hand, if subjects knew fewer words than expected, but succeeded in the task of narrating, I decided not to eliminate them from the study.

4.

For more comprehensive details of each sub-component, see Berman (1988a:476478).

5.

The examples - here and throughout the chapter - end with a tag identifying the subject/narrator. For example, in the code "MS4f-4;5", "MS4" reads "monolingual Spanish, belonging to the group of four-year-olds; "f' is a letter identifying the individual subject within that group; and "4;5" is the exact age of the subject: 4 years, 5 months. Similarly, "BS" reads "bilingual Spanish", "H" "Hebrew", etc.

6.

"Frogs does not" and similar ungrammaticalities in the translation are clumsy attempts to render in English the grammatical errors in the Spanish original.

7.

The figure does not include marking of the present tense, as this is not considered relevant for the present discussion of aspectual contrasts. For a fuller discussion of the choice of anchor tense in monolinguals' and bilinguals' narratives, see Kupersmitt (1999).

8.

The use of the imperfective was counted as a proportion of the total number of lexical predicates. This procedure could have biased the results, since, in texts anchored in the present tense, there is almost no chance for the imperfective to occur. In further research, it is theoretically relevant to examine the use of this aspect in relation to past-tense forms.

434

9.

Judy Kupersmitt

This example is taken from the Hebrew monolingual corpus in the Berman and Slobin (1994) cross-linguistic study.

10. VS cases involving the existential predicate haber 'be' and the copula ser 'be' used in this picture-book-based story for presentation (haber) and for presentation and/or identification (ser) were not counted, since these verbs always appear in a pre-nominal position in both the monolingual and the bilingual group, across all age groups. The verb estar 'be' usually appears in a pre-nominal position, but also allows a non-inverted SV word order, mainly with locatives (Kail & Sanchez y Lopez 1998). Thus it was of theoretical relevance to include this verb in the analysis, as it allows word-order variation. 11. According to the "unaccusative hypothesis", the subjects of this group of intransitive verbs have the properties of the objects of transitive verbs, and so they have no subject, only objects, in their deep-structure configuration (Borer & Grodzinsky 1986: Grimshaw 1987). 12. The unmarked relative pronoun que can sometimes be preceded by a preposition, as demanded by the subordinated predicate, e.g., el asunto del que hablamos parece interesante 'the issue about:DEF.ART:MASC which (we) talk seems interesting'. Also, the relative que can be replaced with more specific pronouns in other contexts, specifically in non-restrictive relative clauses, e.g., el presidente, quien pronuncio sit ultimo discurso, fue abucheado por el publico 'the president, who delivered his last speech, was booed by the audience'. 13. Note the use of a Hebrew word: xamuda 'cute'. 14. For a more comprehensive review on the use of ser and estar, see Sera (1992) 15. The use of the reflexive particle se in this particular example would be acceptable in Peninsular Spanish, but not in South American Spanish, which is the variety under consideration here. 16. The Hebrew word kocim 'thistles' is produced instead of the Spanish word cardos. 17. Note that Hebrew has no exact analogue to this marker. Rather, these functions are met in large part by morphological marking through the verb-pattern binyan conjugations which are typically intransitive and unaccusative, mainly the hitpa 'el and to a lesser extent the nif'al conjugation. 18. Clitics can be either preposed to a finite verb (la buscan a la rana 'it-Ace (they) look for the frog') or postposed to an infinitive (van a buscarla '(they) go to look for-itACC'), but they cannot appear in both positions in the same clause.

Bilingualism and Narrative Construction1

Ludo Verhoeven

1. INTRODUCTION Research on first-language (LI) acquisition has shown children to command many of the grammatical principles and rules governing their native language by the age of four years (e.g., Slobin 1985; Herman & Slobin 1994; Berman 1997; Radford 1995). However, language development both at the level of grammar and at that of preferences for language usage has also been shown to continue into the early school years (Hickmann 1995; Nippold 1998), and into the later school years (Berman & Slobin 1994; Berman & Verhoeven 2002a, 2002b). The narrative also constitutes a significant domain for later language acquisition. Around the age of five, for example, developmental shifts have been found to occur from the use of intra-sentential devices to the use of inter-sentential devices, basic structures to additional functions, and extra-linguistic abilities to intra-linguistic abilities (cf. Karmiloff-Smith 1992). In many parts of the world, children are part of an ethnic minority group whose language is different from the dominant or standard language of the society in which they live. The question, then, is how narratives develop for minority children living in a multilingual environment. In this chapter, an attempt will be made to shed light on this question with a special focus on minority populations in Western Europe. Most minority children learn two languages in a successive manner. They initially acquire the ethnic-group language (LI) within the home and immediate community, and the second language (L2) enters their lives via television, peer contacts, and occasionally day-care centers. When they enter school later on, the language input is almost always L2. While these children take part in all of the lessons in the L2, their L2 acquisition appears to occur very naturally and mainly via interaction with teachers and peers. In fact, previous studies of bilinguals' language development give us no reason to believe that the basic processes and

436

Ludo Verhoeven

features of bilinguals' and monolinguals' language development differ. The essential difference is that bilingual children are confronted with two sets of linguistic input and thus with the task of distinguishing between two different linguistic systems. Just how children do this and the sorts of operating principles (Slobin 1985) that they use to acquire two languages more or less simultaneously are not well understood, however. The extent to which language transfer occurs and the conditions under which such transfer occurs are also not well understood (cf. Verhoeven 1994). Furthermore, the extent to which children actually succeed in producing native-like narratives is completely open to question. As Berman (1998b, 200Ic) has made clear, children must go through several stages of language development in order to achieve native-like narrative production. To start with, the core grammatical principles needed to generate simple-clause structures must be mastered: the rules for word order, temporal inflection, verbal agreement, and case marking. At the same time, the process of lexical selection, involving the precise encoding of specific semantic categories and distinctions, must occur and requires a substantial stock of words. Thereafter, rhetorical expressiveness comes into play, which involves the selection of appropriate lexical and grammatical devices to establish connectivity, perspective via shifts in voice and valency, and foreground versus background information via tense and aspect switches. The final stage of narrative development involves so-called register appropriateness or the use of the cultural and genre distinctions related to the communicative setting. Studies of bilinguals 1 later language development are scarce. That is, just how minority children acquire the basic syntax, lexicon, rhetorical devices, and register appropriateness for their mother tongue and a second language is little understood. In only a very limited number of studies, moreover, has the topic of bilingual narrative development been investigated (see Verhoeven 1993; Extra & Verhoeven 1994, 1998; Verhoeven & Stromqvist 2001). In this chapter, the results of a number of empirical studies of the development of narrative construction within a multilingual context will therefore be reviewed, with special reference to the results of some recently completed studies of minorities living in the Netherlands. Throughout the chapter, the general issue of L1 and L2 acquisition will be explored using a range of data, including the frog story but also other types of narrative, conversation, etc. With respect to L1 development, the case of Turkish children living in the Netherlands will be highlighted, and attention will be paid to other regions both in Europe and elsewhere. With regard to L2 development, the focus will be on relevant Dutch-language skills of children from various ethnic groups. The role of interdependency in bilingual narrative development will also be explored. In addition, an attempt will be made to identify some socio-cultural predictors of bilingual narrative development. With respect to the course of bilingual narrative development, the following hypotheses are formulated. First of all. it is hypothesized that, owing to restricted

Bilingualism and Narrative Construction

437

channels of input, children's narrative development in both LI and L2 will be hampered. We expect children to experience difficulties with clause-packaging and clause-linking devices in narratives, given the complex hierarchical organization of events located in time and space with various event phases subordinated and inter-related. However, although a stagnation of development may occur, only a minor role for interference problems is predicted. In each of the two languages, we believe children's experiences to be filtered into verbalized events via the choice of a highly similar perspective taken, in addition to a set of structural options provided by each individual language. The latter may lead to some thinking-for-speaking effects (see Slobin 1991), but these are considered to be only marginal. Instead, it can be hypothesized that the acquisition of a form to express a particular function may provide an opening wedge for the acquisition of similar functions associated with that form in the other language. Thus, the acquisition of pragmatic devices in one language is expected to channel the attention of the learner towards particular functions that can trigger linguistic expressions in the other language. Interdependencies in bilingual narrative development can thus be the result. Finally, we assume factors related to input, socio-cultural orientation, and schooling to further predict children's narrative development in LI and L2.

2. LI DEVELOPMENT: TURKISH IN THE NETHERLANDS 2.1 Some trends and predictions In this section, the acquisition and use of the Turkish language by Turkish immigrants in the Netherlands will be discussed. It will be explored to what extent these children learn to master the basic core-grammar principles and the appropriate lexical and grammatical devices for rhetorical expressiveness, including clause linking, topic continuity, and temporality, necessary in order to achieve native-like narrative production in their native language. Coming from rural areas of Central Anatolia and the Black Sea region as a result of labour migration since the 1960s, the Turks now constitute the largest minority group in the Netherlands. As a result of net immigration and natural growth, their number has increased to about 210,000. The Turkish community shows an underrepresentation of females (45.6%) as well as older age groups; at this moment, more than 20% of all the Turks in the Netherlands are under the age often. Given the presence of a strongly dominant L2 (Dutch), a gradual shift in language-choice patterns can be expected to occur within the immigrant community. The first generation of immigrants is predominantly Turkishspeaking and has only limited Dutch competence. Owing to decreased use of the mother tongue or interference from the dominant contact language, some attrition

438

Ludo Verhoeven

of the mother tongue will take place. The reduced variant of the mother tongue thus evolving then serves as the Turkish model for the immigrant children, who participate in a very complex linguistic network. Their early language input is Turkish, but the Dutch language quickly enters their lives via playmates and day-care centers. By the age of six, these children are often bilinguals with their Turkish and Dutch language systems in a state of flux: at school, they are submerged in a Dutch curriculum; at home, they may still speak and hear Turkish for at least a few hours a week. The question, then, is: to what extent does attrition - and thereby reduced transmission of the mother tongue - continue to occur within the second and subsequent generations of Turkish immigrants? The results in various preliminary reports suggest that the development of the mother tongue among Turkish children in Western European countries is weakened only during the early stages. Oktem and Oktem (1986) for Germany and Boeschoten and Verhoeven (1986) for the Netherlands have concluded that the acquisition of Turkish by the children of Turkish migrants basically stagnates. Verhoeven and Boeschoten nevertheless found evidence of some structural as well as temporal differences when they compared the grammatical and pragmatic LI abilities of four- to eight-year-old Turkish children living in the Netherlands with those of five- and seven-year-old children living in Turkey. At the morphosyntactic level, strong differences with regard to the pattern of acquisition for non-finite verb forms were observed. The use of complex constructions requiring nominalized verb forms appeared to be extremely difficult for the children in the Netherlands. Factor analyses also showed the parameters underlying the children's language proficiency to differ for the two acquisition contexts. Whereas the mother-tongue proficiency of the children living in Turkey could be interpreted in terms of a single dimension or parameter, the mother-tongue proficiency of the children living in the Netherlands could be interpreted in terms of two underlying dimensions or parameters representing grammatical and pragmatic abilities. Apparently, the children in the Netherlands have not only a lower level of mother-tongue competence but also problems applying the knowledge that they do have. However, it was by no means clear that such deviations constituted a new language variety for the Turkish diaspora. The following linguistic domains will therefore be explored in greater detail: complex syntax, clause linking, topic continuity, and temporality. All of these linguistic domains are particularly relevant to later language development and part of the skills needed for narrative production (cf. Verhoeven 1993). 2.2 Acquisition of core grammar The acquisition of core grammar can be seen as an initial prerequisite for narrative construction. Children must master the core elements of simple-clause structure.


439

which include word order, case marking, verb agreement, and the marking of tense, mood, and aspect. In Verhoeven (199la), the typical grammatical errors made by Turkish children in the Netherlands were analysed. For four- and fiveyear-old children, the number of errors turned out to be relatively small and to clearly reflect the gradual expansion of their linguistic repertoire. For the sixto eight-year-old children, the number of grammatical errors was found to be quite large and quite variable. Problems were observed in the domain of lexical semantics and in the expression of grammatical relations within simple clauses. The sharp increase in the number of grammatical errors during a more advanced phase of development in a bilingual context corresponds to Ekmekfi's report (1986) on the development of Turkish in a monolingual context. It can thus be assumed that children progress from the success-oriented procedures during Phase 1 to largely organization-oriented procedures during Phase 2, which may also produce a temporary regression in their linguistic proficiency. Given that the increase in the number of errors was much larger for the children living in the Netherlands, it can be concluded that the temporary regression in LI linguistic proficiency is much greater in an L2 environment than in an LI environment. In none of the cases, moreover, could the occurrence of the errors be explained with reference to contact with the L2, which was Dutch in this case. 2.3 Acquisition of clause linking Clause linking is a crucial device for marking inter-clausal and inter-sentential relations within discourse. Dutch and Turkish differ greatly with regard to the devices used to link clauses. In Dutch, co-ranking structures consisting of several verbs with the same rank are formed to link clauses. In Turkish, clause-chaining devices typically end with a verb dominating over a fuller structure than any of the preceding verbs. With respect to the acquisition of complex sentences, Turkish children must learn to use a variety of complex constructions that generally require the insertion of non-finite verb forms. However, there are some conjoining devices in Turkish that leave the separate clauses intact and thereby virtually exclude nominalization of the verb. In a study by Verhoeven (1989, 199 la), the use of clause-linking devices by six- to eight-year-old children learning Turkish in the Netherlands was compared with that of their peers in Turkey, and clear evidence of stagnation on the part of the children learning Turkish in the Netherlands was found. At the age of seven, the children in Turkey were found to use four times as many non-finite verb forms in their spontaneous speech as the children in the Netherlands. The use of syntactic features to mark dependency and embeddedness appeared to pose major problems for the children in the Netherlands, moreover. A gradual progression in the linkage of independent clauses was found to occur for these children, and

440

Ludo Verhoeven

the use of such devices proved quite common at the age of eight. For dependent and embedded clauses, however, the situation proved very different. Among the six-year-olds, linkage of dependent or embedded clauses was very uncommon and no clear progress was observed as the children grew older. Less than half of the eight-year-old children were found to use any of the devices available to mark dependence. For embedded clauses, the situation proved even worse. There was a clear tendency for the children in the Netherlands to over-generalize analytic structures of clause combining. A clear stagnation in the use of clause-linking devices by the children in the Netherlands - as a consequence of the presumably restricted Turkish input that they hear - was the conclusion. Schaufeli (1991, 1993) compared the Turkish-language production of twelve-year-old Turkish children living in Turkey and the Netherlands, respectively, and found that the children living in the Netherlands used non-finite verb forms to link clauses to a much lesser extent than the children in Turkey. Once again, the differences were most striking for dependent and embedded clauses, with the children in the Netherlands having a higher frequency of analytic structures at the expense of nominalized verb forms. 2.4 Acquisition of topic continuity The representation of the major characters in a narrative largely determines the organization of the narrative. Protagonists can be described in various ways: using a full noun phrase, a reduced noun phrase, or zero marking. The marking of topic continuity involves the acquisition of an anaphoric strategy with nominal expressions used to reintroduce characters, and pronouns or zero marking used to maintain reference to characters. Dutch and Turkish differ with regard to how they conceptualize narrative entities and the linguistic devices they use to refer to these entities. Dutch makes use of a set of pronouns marking person (first/second/ third) and number (singular/plural). There is a gender distinction in the third person and a politeness form in the second person. In object position, most of the pronouns are inflected. Under specific conditions, demonstratives can be used to refer to entities. Pronominalization in discourse is required. Dutch has explicit markers for indefinite and definite reference. Turkish, in contrast, has a set of pronouns to mark person and number; depending on their grammatical role within the sentence, these pronouns are also inflected for case. There is no distinction for gender. Demonstratives can also be used to refer to an entity. Subject agreement in Turkish is marked on the verbal element by means of a person suffix. The use of pronouns is optional. When the subject serves an emphatic or contrastive function, the use of a pronominal form is obligatory. Given the more frequent use of overt pronouns in Dutch than in Turkish, it is very possible that Turkish children living in the Netherlands may over-generalize


441

the use of Turkish pronominals. In a number of studies, the use of devices for anchoring discourse cohesion has been investigated. In a small-scale study of how seven- and eight-year-old Turkish children living in the Netherlands indicate discourse cohesion, Verhoeven (1990a) found the referential devices used by these children to highly resemble the devices used by five-year-olds in Turkey. The initial introduction of characters was made using a full NP in less than two-thirds of the cases. In many cases, a deictic form was used to introduce the protagonist in a story. To reintroduce a protagonist, a full NP was used in only one-third of the cases. In most cases, a protagonist was reintroduced using a deictic pronoun or simply the agreement marking of the verb. In order to maintain reference to a character, verb agreement was used predominantly. However, the seven-year-olds living in the Netherlands still used pronouns with para-linguistic gestures to denote their deictic nature in many cases, in which they resembled the five-year-olds living in Turkey. Schaufeli (1991) compared the use of devices to mark topic continuity in the narratives of twelve-year-old Turkish children living in Turkey and the Netherlands, respectively, and found the devices used by the two groups of children to be highly comparable. The conclusion then drawn was that the Turkish children living in the Netherlands had developed a native-like level of narrative competence with no signs of interference from their L2, Dutch. Based on the results of a longitudinal study of frog stories in both Turkish and Dutch by four- to ten-year-old Turkish children living in the Netherlands, Aarssen (1996) also concludes that a high degree of similarity in the development of the processes of reference tracking appears to prevail between the two languages. New characters are introduced using the least presupposing device, which is an indefinite NP in both Turkish and Dutch. Subsequent references are made using less explicit devices, which are zero anaphors in Turkish and pronouns in Dutch. In addition, no evidence indicating over-representation of pronominalization in the Turkish data was found. 2.5 Acquisition of temporality To explore the acquisition of linguistic references to time, three fundamental categories of temporality can be considered: the marking of basic temporal relations, or tense; aspect; and internal temporal features. Tense is a deictic category and one function of the marking of temporal relations in narratives is to anchor events along the narrative time line. Aspect refers to the various perspectives that can be taken on an event (e.g., a perfective versus imperfective perspective). And internal temporal features are the quasi-objective temporal characteristics of an event, such as durativity or transformativity. Tense and aspect oppositions in narratives not only serve to locate events relative to the moment of speech; they also organize the structure of a narrative.

442

Ludo Verhoeven

The temporal systems of Dutch and Turkish are quite different. Dutch is a language with a restricted tense system and an even more restricted aspect system. Basically, there is a contrast between present, simple past, and perfect. Aspect plays a role in that progressive forms can be contrasted with imperfective forms through the use of a locative expression with durative aspect. Turkish grammaticizes many more temporal distinctions than Dutch. With respect to tense, there are different markers for present, past, and perfect. Within the past tense, a modal distinction is also made between the expression of direct and indirect evidence (cf. Slobin & Aksu 1982). Past events can be marked as imperfective or perfective. Imperfective reference to past events can be further specified with regard to factuality. When it comes to the acquisition of Turkish temporality within a Dutchlanguage environment, the question is to what extent children succeed in mastering the available tense and aspectual distinctions. From the narratives of six- to eight-year-old Turkish children living in the Netherlands. Boeschoten (1990) and Verhoeven (1990b) concluded that the acquisition of strategies for establishing discourse cohesion by the Turkish children living in the Netherlands had stagnated. Their findings provide evidence of a strong inclination towards unifunctionality: out of the many formal devices available in a language to express certain meanings, only a device that is relatively easy to process or one that has already been acquired to express a closely related meaning will be used. More importantly, none of the cases pointed to the transfer of Dutch structures. In his study of the development of temporal anchoring in the Turkish narrativ es of monolingual and bilingual Turkish children between the ages of four and ten years, Aarssen (1996, 2001) similarly concluded that no structural differences in the temporal distinctions marked in the older children's frog-story narratives could be detected. With respect to the expression of simultaneity, the children in both groups were found to acquire new functions for already-acquired linguistic devices over time. That is, the children in the different cohorts all started out with a limited set of forms to express a broad set of meanings. Over time, development towards a more restricted range of use for individual forms could be seen. Almost all of the forms used by the monolingual children were also used by the bilingual children. 2.6 Cross-linguistic trends The review of the relevant studies shows the development of the narrative skills of Turkish children living in a foreign-language environment, namely the Netherlands, to be at risk. Owing to restricted LI input, these children demonstrate a lower level of mother-tongue competence and also problems with the application of the knowledge that they do have. Despite this stagnation, there is ample evidence that the children in both the LI and L2 environments rely on


443

highly comparable intra-lingual strategies. That is, in the speech of the Turkish children living in an L2 environment, there is no clear evidence of inter-lingual patterns of language use. In several other studies, the LI development of Turkish children growing up in a foreign-language environment has been explored. Pfaff (1993, 1994, 1998, 2001) examined the acquisition of Turkish by Turkish children living in Berlin. With respect to core grammar, a relatively large number of errors was reported to occur among the six-year-old children. The inflectional morphology for case and number was found to be intact, but the relevant derivational morphology was weak. The acquisition of complex syntactic devices was quite late. Sentence modification was often done by paraphrasing. Pfaff's conclusion is that the children increasingly shift towards German and thus produce varieties of Turkish with clear evidence of extensive German contact. The acquisition aspect of L I and L2 frog-story narratives by Turkish children living in France was investigated by Akinci (1999) and by Akinci, Jisa, and Kern (2001). A delay was found in the number of macro-structures encoded in the Turkish narratives. Compared with their monolingual peers, moreover, the Turkish children living in France were found to produce a greater number of grammatical errors and to have a more restricted lexical repertoire. The above findings show the second generation of Turkish migrants living in the Netherlands not to attain native-like levels of proficiency, and a gradual language shift to be occurring among this diaspora. The relatively low level of loss observed for the Turkish linguistic features nevertheless shows the shift to be a rather slow one. The question, then, is what particular narrative structures and language varieties are produced by such restricted channels of input. Recent research on the code switching found to occur among Turkish teenagers living in the Netherlands, for example, provides ample evidence for the emergence of a mixed narrative code as part of the vernacular for this group (see Backus 1996). Inter-lingual interaction appears to play a major role here and to produce a high prevalence of grammatically reduced sentences. If such code switching comes to characterize the predominant variety of Turkish spoken in the Netherlands, and thus also the input model for a third generation of Turkish "immigrants", a substantial erosion of the grammatical and pragmatic sub-systems of the Turkish language can be expected to occur within the Netherlands in the near future, accompanied by a reduction in the level of Turkish narrative performance.

3. L2 DEVELOPMENT In several studies, the production of narratives in a second language has been examined and compared with the production of narratives by native speakers of the same language. The acquisition of Dutch as L2 has been examined in

444

Ludo Verhoeven

dissertations by Aarssen (1996) and Bos (1997), in connection with Turkish and Moroccan Arabic, respectively. For the theoretical background to these studies, see Verhoeven (1993). Frog-story narrative data were collected from groups of 50 Turkish children, 50 Moroccan children, and 50 Dutch children between the ages of four and eight years. Each group of informants was tested four times at oneyear intervals. All the children were of a poor socio-economic background. The analysis of the narrative data started with a quantitative examination of the distribution of various referential devices in the children's narratives. Of particular interest was how the bilingual children with lower L2 syntactic skills handled the introduction, reintroduction, and discussion of characters relative to their native-Dutch-speaking peers. Qualitative analyses of micro-level changes in the strategies used by the children for narrative cohesion were also conducted. For topic continuity, no differences in the patterns of development between the bilingual and the monolingual children were observed. The acquisition strategies of the children can thus be characterized as universal and not depending on their linguistic backgrounds or on their monolinguality or bilinguality. All of the children appeared to make the initial establishment of new characters using the least presupposing device, the indefinite NP. Subsequent references requiring maintenance of the character's identity in a chain of foregrounded activities were generally made using pronouns and independently of the specific device employed by the mother tongue (e.g., zero anaphors in Turkish). Reference shifts for when a character was temporarily out of focus were typically marked using nominal forms. The general conclusion is that children's narrative references are largely governed by pragmatic considerations and not by highly languagespecific syntactic rules. Analyses of temporality using the same data showed that both foreground and background elements were distinguished by the children. With regard to foreground elements, topic information was distinguished from focus information, and the strategies for time anchoring and foregrounding were found to be highly comparable for the monolingual and the bilingual children. The bilingual children were nevertheless slower than the monolingual children during the early stages of acquiring the relevant temporal devices, and the monolingual Dutch children appeared to use more forms than the bilingual children. The monolingual Dutch children also demonstrated better knowledge of how to map functions onto forms than the bilingual children. For instance, a more restricted register for the use of temporal adverbials was observed among the Arabic-Dutch children. The monolingual Dutch children showed a rather sophisticated use of adverbials, conjunctions, and particles, requiring detailed knowledge of the rules for subordination and inversion of basic word order. In conclusion, the data on the acquisition of Dutch as L2 shows ethnicminority children to have a strikingly lower level of narrative-production skills


445

than their native-Dutch-speaking peers. The differences in the children's mastery of the devices for narrative cohesion appear to decrease as the children grow older. From a structural point of view, the results also show the differences in narrative production to be largely explained by intra-lingual rather than interlingual developmental features on the part of the L2 learner. It was generally found that the structure of the target language, namely Dutch, and not so much the structure of the source language, can account for various narrative-production difficulties. However, in the narrative production of beginning L2 learners, in particular, some evidence of transfer from the LI was found. Research on L2 development in other linguistic environments provides additional evidence for the same conclusions. Lanza (2001) investigated temporality in the frog-story narratives of Norwegian- and English-speaking children living in Norway and found the bilingual children's narrative development to be similar to that of their monolingual peers. Certain differences that could be purely individual or indicative of transfer from the source language were also observed. Using a variety of picture stories, Nistov (2000, 2001) examined the development of topic continuity in the L2 narratives of Turkish adolescents living in Norway. The focus of the study was on the marking of reference continuation in short narratives written by the L2 learners, and the beginning L2 learners were found to use more zero markings than beginning Norwegian LI learners. Although the L2 learners were expected to use fewer pronouns for the shift function because this is not allowed by their LI, no empirical evidence for this was found. Considerable individual variation, with some evidence of LI influence only at the beginning of L2 narrative production, was observed. In a study with diverse elicitation procedures, Viberg (1998, 2001) explored a number of the agerelated features of narrative development in Swedish as L2 with a focus on the devices used to map forms onto functions in the domains of lexical processing, clause combining, and temporal relations. The overall conclusion was that agerelated differences exist at all levels of language structure, and particularly at that of narrative functions. In addition, Viberg claims that many L2-related features, such as neutralizations of specific functional or semantic contrasts in Swedish, also determine the course of narrative development. Indeed, a small number of the relevant features were clearly distinct from early LI, which leaves room for a conception of the L2 acquisition process as different from the LI acquisition process. Akinci, Jisa, and Kern (2001) recently examined the development of frogstory narrative production in school-aged Turkish children living in France. Comparison with monolingual French children allowed the authors to explore the influence of the bilingual children's LI Turkish on their L2 French narratives, and they did indeed find a delay in the acquisition of the L2 linguistic devices for encoding macro-structures in narratives. However, comparison of the LI

446

Ludo Verhoeven

and L2 texts produced by the Turkish children revealed no differences in the number of macro-structure components encoded. In addition, a disparity between the monolingual and bilingual groups with regard to social class suggests that the relative delay in the Turkish children's acquisition of certain L2 narrative devices may be due to differences in the amount of exposure to literacy-related activities. Kupersmitt (1999, this volume) asked bilingual (Spanish- and Hebrewspeaking) children from a middle-class background to tell the frog story in each of their two languages and examined the extent to which their knowledge of linguistic structure appears to develop in tandem with their ability to use the language for narrative construction within a discourse context. The use of particular linguistic categories within a narrative context was found to be determined by the availability of the linguistic forms to the speaker, whereas the global narrative structure in one language was found not to be markedly affected by the other language. Finally, Stavans (2001) explored the frog-story narratives of trilingual children (speaking Hebrew, English, and Spanish) in Israel. The development of the children's narrative construction was examined as a function of both language-specific issues and language-contact issues. The conclusion was that the children's manipulation of the form-function relationships in each language was similar to that evidenced by monolinguals. Differences in the form-function relationships could be attributed to the child's competence in the specific language and to socio-cultural contact between the various languages.

4. INTERDEPENDENCY IN BILINGUAL DEVELOPMENT An important question is to what extent learners succeed in acquiring L2 by generalization from LI (MacWhinney 1992). Besides some thinking-forspeaking effects that differentiate between LI and L2 narratives, the perspective for the construction of a narrative in the two languages can be considered highly similar. Therefore, one can hypothesize that the acquisition of a form to express a particular function in one language may facilitate the acquisition of a form to express that same function in another language. Such interdependency was recently sought in a number of empirical studies (cf. Verhoeven 1994). Verhoeven and Narain (1996) examined the patterns of LI and L2 development for 91 Turkish, 111 Moroccan, and 104 Antillean children living in the Netherlands. The latter group of children originate from the Netherlands Antilles, a former Dutch colony where Papiamentu is spoken as the native language. In a longitudinal design, oral-language data were collected at three moments in time: at the beginning of kindergarten, after one year of school, and


447

after two years of school. The Turkish language was found to be dominant for the Turkish children at all three measurement points. The same was found to be true for the Moroccan group, but to a lesser extent than for the Turkish group. The language-proficiency levels of the Antillean children tended to be more balanced. The differences in the L1 and L2 proficiency levels of the three groups could be explained in terms of the language input provided within the family and the wider community, on the one hand, and the cultural orientations of the children and their parents, on the other hand. Next, the specific interdependencies in the bilingual development of the Turkish, Moroccan, and Antillean children were explored in greater detail. For each minority group, a substantial amount of transfer from LI to L2 was found to occur at the beginning of kindergarten. For the Antillean children, the direction of transfer tended to reverse after two years of schooling. The children's language proficiency shifted from dominance by Papiamentu to dominance by Dutch - the latter part of this finding reflecting the more general assumption that transfer usually occurs from the more dominant language to the weaker language. Verhoeven (1994) also explored the possibility of transfer in the development of the LI and L2 pragmatic abilities of Turkish children between the ages of six and eight years living in the Netherlands. In order to measure the pragmatic abilities of the children in Turkish and Dutch, spontaneous speech was elicited using three different procedures: having the children provide a spatial description of what they saw in some pictures, having the children describe a series of events depicted in some pictures, and collecting spontaneous-speech samples. Two measures of pragmatic proficiency were then computed: the number of different content words (nominal and verbal types) occurring in 75 utterances of speech and the mean number of morphemes in the longest utterances (10% of each individual corpus). Multivariate analysis of variance on the number of different content words revealed a significant main effect of age and thus a general progression in the use of content words over time. The main effect of language did not prove significant, but the language-by-age interaction did. This shows that the older but not the younger children use more different content words in L2 than in L I . The multivariate analysis of variance on longest-utterance length revealed significant main effects of age and language, and also a significant age-by-language interaction. In other words, the increase in utterance length as the children grow older is greater for LI than for L2, with the longest-utterance length generally being greater in Turkish than in Dutch. In order to investigate the degree of interdependency between the pragmatic skills of the children in Turkish and Dutch, analyses of linear structural relations (LISREL, Joreskog & Sorbom 1996) were undertaken. The latent variables for pragmatic proficiency in Turkish and Dutch were defined in terms of number of different content words and longest-utterance length. The best-fitting structural model showed, in addition to

448

Ludo Verhoeven

the longitudinal effects of age, a strong effect of L1 pragmatic proficiency on L2 pragmatic proficiency. In other words, empirical evidence for the positive transfer of narrative skills from a dominant language to a second language was found. It is nevertheless possible that such underlying factors as the children's cognitive capacities may account for the relations between LI and L2. More research is clearly needed to confirm or reject the interdependency hypothesis.

5. BILINGUAL DEVELOPMENT: SOCIO-CULTURAL FACTORS 5.1 Predicting language and literacy development Depending on the cultural values and language input within the home environment, considerable variation in the LI and L2 acquisition patterns observed for minority groups can be expected. In another study (Narain & Verhoeven 1993), we examined the patterns of LI and L2 development in 91 Turkish, 111 Moroccan, and 104 Antillean children living in the Netherlands. In a longitudinal design, oral-language data were collected at three moments in time: at the beginning of kindergarten, after one year of school, and after two years of school. The Turkish language was found to be dominant for the Turkish children at all three measurement points. The same was found to hold for the Moroccan group, but to a lesser extent. The language-proficiency levels of the Antillean children, however, tended to be more balanced. The differences in the L1 and L2 proficiency levels of the three groups could be explained in terms of the language input provided within the family and the wider community, on the one hand, and the cultural orientations of the children and their parents, on the other hand. Different socio-cultural factors can explain the patterns of development observed. In a study by Verhoeven (1991b), an attempt was made to identify those factors that predict the LI and L2 proficiency of Turkish children at the age of six. The sample consisted of 72 Turkish children, living in the Netherlands since infancy, who were interviewed prior to their entrance into the first grade. Information on each child, his or her family, and any institutional care received by the child was collected along with a number of direct and indirect measures of the child's LI and L2 proficiency. Three socio-cultural factors were found to be related to the children's pragmatic skills in Turkish: the children's orientation towards the Turkish way of life, the presence of Turkish-speaking peers in institutional care, and parental involvement in institutional care. Three sociocultural factors were also found to be clearly related to the grammatical skills of the children in their mother tongue: the children's orientation towards the Turkish way of life, the cultural behaviour of their parents, and the extent of caretaker interaction in Turkish. With respect to the children's acquisition of Dutch as L2, the following pattern emerged: Three socio-cultural factors were found to


449

relate to the pragmatic skills of the children in Dutch: the children's orientation towards the Dutch way of life, the extent of family interaction in L2, and the extent of caretaker interaction in LI (i.e., an indirect measure). No fewer than nine socio-cultural factors yielded a significant correlation with the level of Dutch grammatical skills for the children. For eight of these factors, the correlations were positive: the children's cognitive capacities, the parents' attitudes towards the Dutch culture, the extent of L2 literacy within the family, the extent of family interaction in L2, the parents' cultural behaviour, the extent of peer interaction in L2, the extent of caretaker interaction in LI, and the extent of special involvement of a parent in day care. A negative correlation was found for the factor "caretaker interaction in L2", which could be explained by the fact that - in cases of L2 learning problems for some of the groups - extra caretaker interaction was provided via day care. In summary, the results of the above-mentioned research show multiple factors to account for the individual differences observed in the bilingual proficiency of Turkish-minority children at the age of six. The prediction of children's LI and L2 proficiency is thus complicated. The present study clearly shows different child- and family-related and institutional factors to play a role in children's LI versus L2 development. The correlation of such socio-cultural factors with the children's LI proficiency turned out to be relatively low, which can be explained in part by the fact that first-generation immigrants constitute a relatively homogenous group with respect to their attitudes towards the home culture and their LI interaction patterns (cf. Tosi 1998). Leseman (1997) examined the effects of the socio-cultural context, and in particular the style of parent-child storybook reading, on home literacy in the Netherlands, and found the development of literacy to be determined by opportunities for literacy-related interactions, by the kinds of guidance and informal instruction provided, and by the affective experiences accompanying the literacy interactions. Examination of the effects of literacy practices on children's literacy development in a variety of communities also showed that the differences between families and their children's literacy development arise primarily from the socio-cultural context and, in particular, the everyday cultural and religious practices of the family and the type of parental employment. Furthermore, Leseman and De Jong (2001) investigated the impact of home literacy on school literacy development in a multicultural sample, comparing children of low socio-economic status and minority children with middle-class children in the Netherlands. They found small but significant effects of home literacy practices on vocabulary and word-decoding development. Moreover, they found home literacy and informal education at home to be rather strongly related to the family's socio-economic and cultural context. Boyd and Naucler (2001) investigated how Turkish and Swedish parents and Swedish pre-school teachers carry out the co-construction of stories for a number of monolingual and bilingual children. The groups under consideration

450

Ludo Verhoeven

were the children and adults in eight Turkish families living in Sweden and seven Swedish working-class families. The children were all five to six years old and attending Swedish pre-school. The authors found different patterns of interactive storytelling for the two groups. Whereas the Turkish children were given a limited number of turns in such activities, the Swedish children were given a much greater number of turns, which led to a real co-construction of narratives. The Turkish and Swedish mothers reported different orientations towards language and literacy learning within their families. The Swedish mothers attached greater importance to reading and other literacy activities for the language development of their children than the Turkish mothers. Conversely, the Turkish mothers stressed the importance of interaction in social activities as beneficial for the language development of their children. Interestingly, the teachers' interactions with children paralleled those of the mothers. When interacting with Swedish children, a co-operative strategy was adopted, with many turns on the part of the children. When interacting with the Turkish children, the teachers told the story with a minimum of turns on the part of the child. These differences can be explained in terms of the expectations that the teachers have with regard to the socio-cultural norms and linguistic proficiency of the two groups of children in their classrooms. A more in-depth analysis of the co-construction of narratives in the mother tongue by Turkish children living in Germany was conducted by Pfaff (2001). The parent and teacher responses to two children of four and seven years of age were examined, and the interlocutors were found to be concerned primarily with the content, not the form, of the children's narratives. In many cases, for example, mixed-language use or the use of non-standard forms by the child was simply accepted or even repeated as part of the process of narrative co-construction. And to the extent that the teachers or parents reacted to the more formal aspects of the children's language, use of vocabulary from the other language was the feature most likely to receive explicit comment from the interlocutor. Morphological or syntactic errors on the part of the children rarely received explicit comment. 5.2 Effects of schooling There is reason to believe that the acquisition of narrative competence depends on the development of critical literacy skills (cf. Verhoeven 1996a. 1997; Durgunoglu & Verhoeven 1998). Multicultural studies of early literacy show that, despite differences in cultural background and language diversity, children are able to learn the essentials of literacy at a very early age. Studies by Moore (1990), Galambos and Goldin-Meadow (1990), Bruck and Genesee (1995), Verhoeven (1996b), and Bialystok (1997) have all shown that bilingual children, in collaboration with parents or older peers, are capable of acquiring the meta-linguistic skills required for reading at a very young age. One tentative


451

explanation for this is that bilingual children analyse and compare the structural features of different languages in a more advanced manner than their monolingual peers. As a result, bilingual children may actually become more proficient at the meta-linguistic tasks than monolingual children. However, large differences in the relevant knowledge and desire for literacy among minority children entering school have been found to exist; as already mentioned, these differences depend on the amount and type of literacy support provided in the home environment. Scarborough and Dobrich (1994) have clearly shown early literacy acquisition to depend on both the values attached to literacy at home and the steps that parents take to explain the value of literacy to their children. Literacy is a very culture-specific activity. According to Serpell (2000), becoming literate means becoming a member of a community of practice and thereby mastering the cultural-meaning system that informs the literate activities of the community. Serpell claims that such routines as joint storybook reading are deeply rooted in particular sub-cultures and a highly efficient means of establishing literacy. In addition, the different communities within a society may have very different motives for using written language and therefore engage in very different literacy practices. Snow et al. (1991) found the language and literacy development of children from varying socio-economic backgrounds to be clearly related to such home and school factors as the literacy environment within the home, the mother's education, the mother's expectations for the child, and the contact between parents and teachers regarding academic matters. In multicultural societies, thus, the socio-linguistic position of various minority groups should be recognized along with their differing literacy practices (Durgunoglu & Verhoeven 1998). In addition, minority groups may use the written codes available to them for at least partially distinct functions. The written code with the highest status will be used primarily in connection with societal institutions while the minority written code will be used mostly for intra-group communication and to express one's ethnicity. According to Au (1998), many of the gaps in the literacy achievement of the different cultural groups within a society can be explained in terms of the relevant groups' having linguistic and cultural norms that are in conflict with those of mainstream education. Au has also argued that poverty and school failure are both manifestations of historical and systemic conditions with their roots in discrimination: the communities in which ethnic-minority pupils live and the material circumstances of their schools often produce educational inequalities.

6. CONCLUSIONS AND DISCUSSION From the research on narrative development in a bilingual context, several conclusions can be drawn. The overall conclusion is that the narrative development

452

Ludo Verhoeven

of bilingual children resembles that of monolingual children, with filtering and packaging serving as the guiding principles (cf. Berman & Slobin 1994). The filtering principle holds that experiences are filtered into verbalized events via the choice of perspective taken on the experiences and the set of linguistic operations provided by the language in question. The packaging principle holds that a narrative in LI or L2 does not involve a linear chain of successive events but a hierarchical set of events located in time and space with various event phases subordinated and inter-related. Research on the developmental relations between form and function within a multilingual context has shown that the L1/L2 grammatical forms available to learners channel their attention towards particular functions. That is, frequently used forms direct attention to the functions associated with those forms. In addition, the acquisition of a form to express a particular function has been shown to provide an opening wedge for the acquisition of the more advanced functions associated with the same form. In other words, universal strategies appear to underlie the development of narrative cohesion in LI and L2. In whatever language, children must learn to elaborate inter-sentential devices to connect independently represented entities into a whole (see Karmiloff-Smith 1992). And via the development of such a control process, bilingual children gradually learn to constrain the production of cohesive markers across related clauses in both their LI and L2 narratives (cf. Bialystok 1992). The research conducted so far clearly shows the development of the narrative skills of bilingual children growing up in an L2-submersion environment to lag behind that of their monolingual peers. In order to become competent storytellers, children must first acquire a sufficient level of lexical knowledge and knowledge of the obligatory linguistic rules. In order to overcome grammatical shortcomings, bilingual learners may try to maximize top-down comprehension processes and thereby acquire a minimal ability to construct simple clauses. However, it may well be the case that the number of errors that one can detect in one's own production is insufficient to arrive at a full reorganization of one's linguistic system and thereby attain the standard of a native speaker. Even if errors are detected by the learner, efforts to reorganize the system to decrease the number of such errors may simply lose out to the strength of already highly automated procedures (cf. MacWhinney 1992). A breakdown of the learning processes needed to acquire rhetorical expressiveness may then occur. In order to prevent or stop such a breakdown, the learner must again focus on the features of basic clause structure and clause combining rather than the macro-structure of the narrative. Given that the narratives of bilingual children are rooted in the socio-cultural context of their existence, the bilingual narrative data can be expected to mirror the socialization patterns of the narrator. With the onset of language attrition, a progressive decline in the native language and fragmentation of its linguistic


453

features and word-formation paradigms can be expected. As a consequence, the lexical and structural choices of the narrator will be increasingly constrained, with restricted narrative complexity and rhetorical style in the end. Language shift and even the loss of minority languages may occur across generations. Boyd (1985, 1993) has identified six major determinants of the rate of shift. First, the process of shift is influenced by such historical factors as the socio-linguistic background of the minority groups in the home country, the relations between the home country and the host country, and the social and linguistic situations of those minority groups having arrived earlier. Second, demographic factors, such as group size, birth rate, marriage patterns, and the time that has elapsed since the immigration to the new country turn out to be influential. Third, such demographic factors as the distance from the home country, the concentration of settlement, and the self-sufficiency of the group in question are clearly relevant. Fourth, social factors related to the socio-economic status of the group play a role. Fifth, institutional factors related to culture, religion, mass media, and so forth can play a role. Finally, numerous attitudinal factors can play a role. Given that these factors often overlap and are not mutually exclusive, their relative importance is difficult to determine. Whether bilingual children can keep up with monolingual children with regard to the construction of narratives in a second language is still unclear. Schools must help children acquire native-like levels of narrative competence in the L2. Teachers must recognize that the context of textbooks is not equally familiar to all children. The use of pre-reading activities - such as discussion of the content of the story, provision of background information, building of shared experiences, and explanation of difficult lexical items - can therefore be recommended. Special attention to the structure of the story can also facilitate narrative comprehension and production. That is, children can presumably produce coherent narratives when the structural cues necessary are already available to them. Well-organized story grammars can help children identify, summarize, and outline main ideas (Garcia 2000). Research has also shown explicit instruction with regard to the structure of a text to facilitate the (re)production of narratives (Dickson, Simmons & Kameenui 1995). Reading and writing instruction should therefore focus on the use of signal words, topic sentences, and organizational structures to identify and/or encode the main ideas and relations in a narrative. And in order to become competent readers and storytellers, LI and L2 learners should also be taught to apply in a flexible manner the meta-cognitive strategies they acquire (cf. Bialystok & Ryan 1985). Finally, ethnographic studies have shown that minority-group members often master two (or more) written codes for complementary sets of purposes. The primary function of the majority language may be for inter-group communication within the community as a whole, and the primary function of

454

Ludo Verhoeven

the minority language may be for intra-group communication and for expressing one's ethnicity (cf. Fishman 1980; Extra & Verhoeven 1993). With respect to the use of language and literacy in general, two worldwide trends can be noted. On the one hand, increased globalization and internationalization are occurring as a result of the mass media, trade, labour migration, and tourism. On the other hand, increased awareness and recognition of the cultural and linguistic diversity in the world are emerging. A basic policy question in the area of institutional support, then, is just how to reconcile these trends towards unification and diversification within the context of a multicultural and multilingual society.

NOTES 1.

The author is grateful to Ruth Berman and Dan Slobin for their invaluable comments on an earlier version of this chapter.

Epilogue


Variations on a Ranarian Theme

Herbert H. Clark

1. INTRODUCTION Telling stories is hard. First you need something to tell - a genuine story. You must be acquainted with a happening, or series of events, about which you can make a point. Then you need an occasion to tell the story. Someone has just brought up a topic of conversation, and you have a story that is relevant (Jefferson 1978). Or you are participating in a round of storytelling, and it is now your turn (Kirshenblatt-Gimblett 1974). Or someone has asked for a particular story. Even before you start, you need an overall plan - where to begin, what events and evaluations to include or emphasize, and where to end. In the actual telling, you need to formulate utterances one by one. For each utterance, you must select the right words and gestures from the repertoire available to you - your language and culture - and you must do so in a timely fashion. Your selections must be both true to the events in the story and understandable to your audience. It is no wonder that telling stories is such an art - how many good storytellers do you know? - and that it takes children so long to acquire. Many of these issues have been addressed in research on Mercer Mayer's (1969) story Frog, where are you? In these studies, people are asked to look through Mayer's wordless picture book (see Appendix I) and then, while paging through the book again, retell the story to someone else. The landmark collection of these studies is Relating events in narrative: A cross-linguistic developmental study, edited by Ruth Berman and Dan Slobin (1994). It set out the basic issues for the development of storytelling, and showed how children tell the frog story in many languages and cultures. The current collection is a follow-up with an even wider range of languages and cultures. Let me call the two volumes Frog I and //. My aim is to take up three variations on this ranarian theme: story versus telling, diegesis versus mimesis, and conception versus description. If you listen

458

Herbert H. Clark

closely, you will also hear a leitmotiv that recurs throughout the collection itself- Slobin's (1996b) notion of thinking-for-speaking.

2. VARIATION 1: STORY VERSUS NARRATIVE There is a traditional distinction in storytelling between the story someone tells and the telling of that story (see Berman & Slobin 1994). As Chatman (1978) put it: [E]ach narrative has two parts: a story (histoire), the content or chain of events (actions, happenings), plus what may be called the existents (characters, items of setting); and a discourse (discours), that is, the expression, the means by which the content is communicated. In simple terms, the story is the what in a narrative that is depicted, discourse the how.

With Chatman, I will distinguish between the same two parts, calling them the story and the narrative (or telling). Most work on the frog stories has focused on the narrative. And rightly so, for that is what changes most from language to language, and from childhood to adulthood. Always the contrarian, however. 1 want to take a closer look at the story. 2.1 Story types One of the progenitors of the modern work on narratives was William Labov's work on narratives of personal experience (Labov 1972; Labov & Waletzky 1967). In one study, Philadelphia pre-adolescents, teenagers, and adults were asked such questions as (Labov 1972:354): "Were you ever in a situation where you were in serious danger of being killed, where you said to yourself 'This is it'?" or "Were you ever in a fight with a guy bigger than you?" When the respondents said "yes," they were asked "What happened?" and out came a narrative. The stories behind these narratives were experiences the narrators had lived, and telling the stories required them to reconstruct those experiences from memory. Indeed, Labov spoke of the narrators becoming "deeply involved in rehearsing or even reliving events of [their] past," and the narratives he cites suggest that the narrators got into them with heart and soul. The frog stories are very different from the Philadelphia stories. Some of the differences make storytelling easier, but others make it harder. /.

Factual versus fictional stories. Labov's narrators were telling stories of fact, of events that actually happened. The frog narrators were telling stories that were pure fiction, events they knew hadn't actually happened. Now, fact and fiction place different requirements on storytellers. In


factual accounts you are constrained by what actually happened, and your story cannot include impossible events or non-existent people or settings. With fiction, you have license to make things up and to create people and settings in any way you like. A simple example is proper names. In the frog stories, the boy was given such invented names as Piita, Mikel, Adoni, Aitor, Torsten, Ali, Cookie, Pelle, Bert, Mans, Pepe, Pedro, Dani, and Tomasito, and the dog, Jensi, Zikin, Txuri, Zikina, Lasse, Bonkie, and Plutt. These names were created out of the language and culture of the narrator. The names used in the Philadelphia stories were presumably the characters' actual names, which the narrators tried to get right. First- versus third-person stories. The frog stories were told in the third person, where the narrator was distinct from the main protagonist. Here is an example from a nine-year-old narrating in English: "And he starts running. And he tips him off over a cliff into the water. And he lands."1 The Philadelphia stories were told, instead, in the first person, where the narrator was also the main protagonist. Example: "An' then, three weeks ago I had a fight with this other dude outside. He got mad 'cause I wouldn't give him a cigarette." First- and third-person perspectives require very different modes of thinking and linguistic resources (see, e.g., Segal et al. 1997). It may be easier to formulate certain utterances - say, about what protagonists are thinking, feeling, seeing, or trying to say - from the perspective of the protagonist than from the perspective of an observer, that is, in the first person rather than the third person. And different aspects of the story may get emphasized in the first person - excitement, fear, first impressions - than in the third person - the protagonist's appearance, location, and surroundings. Yet, other features of a story may be easier to formulate from the perspective of an observer, who can take in the broad sweep of a scene and not worry about the perspectives of different characters. Creating a story versus retelling a story. The Philadelphia narrators had to create both the stories and the narratives from scratch. They had to decide how to orient their audience to the who, what, where, why, and when of the story, which events to treat as the complicating action and which to treat as the resolution, what background commentary to include, and more. In Labov's analysis, the narratives they created emerged in six main sections: abstract, orientation, complicating action, evaluation, result or resolution, and coda.

459

460

Herbert H. Clark

The frog narrators, in contrast, had the stories handed to them in a book, and they had only the narratives to create. Their job was to retell the story as they understood it. Retelling a story has its own burdens. You are not free to make up this or that episode. Instead, you have to get the episodes of the story right, including the order of events and who is doing what. You must also be accurate on the point of the original narrator - why he or she was telling that story. This was a bit easier for the frog stories because the picture book gave the narrators leeway in the description of each event and the construal of Mayer's point. And the narrators were also reminded page by page, episode by episode, of how the story went. One result was that these narrators didn't create the sections that were characteristic of the Philadelphia stories. Creating sections was another burden lifted from their shoulders. Plainly, then, storytelling is not all alike, a point that has been made by many others (e.g., Berman, this volume; Hickmann, this volume). It just isn't the same to tell a factual versus a fictional story, to tell a story from a first- versus a thirdperson perspective, or to create versus retell a story. Surprisingly little is known about these differences. And yet, if we are interested in the skill of storytelling, or the age at which it is mastered, these are issues we must understand. 2.2 The same story? The research on the frog stories has relied on a tried-and-true experimental method: Have everyone retell the same story - as represented in Mayer's Frog, where are you? -, then attribute any differences in the narratives to the language, culture, age, or skill of the narrator. But is everyone really retelling the same story? The contributions to Frog II suggest that the answer is "no." First, the youngest children didn't seem to understand the story fully, or had alternative interpretations of the same pictures. And second, what the story was taken to be changed from culture to culture. Where does the frog story come from? The answer, of course, is Mayer's Frog, where are you? But does everyone understand the story as Mayer intended? Clearly not. Many of the youngest narrators described what was depicted in each picture, and didn't seem to connect successive pictures. They might have recognized the boy, dog, and frog as the same in successive pictures, but they often failed to see a causal or temporal relation between the pictures. This point was made in a number of the contributions to Frog I and II. If this is true, what are we to make of the problems that children have in narrating the frog story? Is it that the children are unable to interpret the picture-book as intended, or is it that they are immature in their ability to describe cause, temporal relations, motion events, and other features of the basic story? Indeed, Mayer's picture book isn't all that easy to interpret. In Picture 1. we meet a frog I will call Snag, and it leaves the house in Picture 2. In the penultimate


461

picture, the boy discovers a family of frogs - a father frog, a mother frog, and baby frogs —, and in the final picture, he leaves with one of the baby frogs, giving us the resolution of the story. But what has happened to Snag? I assumed that Snag was the father frog, and the boy was leaving with one of his babies. But as a friend pointed out, how could Snag have found a mate and sired a family overnight? Why didn't 1 assume that Snag was the mother who had gone back home just in time to deliver her babies? A number of the tellings suggest that I wasn't the only reader with problems. It is no surprise that children and adults read Frog, -where are you? against the common ground of their own culture. As an American, Mercer Mayer probably intended his depictions to be interpreted against American presuppositions about boys, dogs, and frogs, but narrators from other cultures clearly read it differently. This point is illustrated by how people identified the several features of the story • the gopher, the owl, the bees, the deer, and the body of water: •

the gopher was sometimes seen as a groundhog (Tzeltal), a mole, chipmunk, or squirrel (Japanese), or a squirrel or mouse (Spanish); the owl was sometimes a butterfly (West-Greenlandic) or bat (Japanese);

•

the bees were sometimes flies (Icelandic), mosquitoes, or wasps (Spanish);

•

the deer was sometimes a reindeer (West-Greenlandic) or elk (English), a gazelle or goat (Turkish), and sometimes even a horse (Spanish - do horses ever have antlers?);

•

the body of water was a puddle or lake (Tzeltal), a river, lake, or sea (West-Greenlandic), a sea (Icelandic), a river or pond (Basque), or a stream or little puddle (American Sign Language); the frog was considered a pet in most cultures, but prey in Arrernte;

•

the boy and dog were always a boy and a dog, as Mayer presumably intended.

These categories, surely, reflect cultural and not linguistic differences. The WestGreenlandic or Icelandic choices of "sea" for the body of water are dictated not by the syntax of Greenlandic or Icelandic, but by the narrators' familiarity with bodies of water. And the Turkish children may be more familiar with gazelles and goats than with deer - at least compared with American children. Ultimately,

462

Herbert H. Clark

familiarity is a cultural, not a linguistic issue. And with these simple categorical differences, we have barely scratched the surface of the cultural influence on the interpretation of Mayer's story. One more point. It seems only common sense that narrators start with a story and end with the narrative. After all, there can be no telling of a story without a story to tell. But this is much too simple a conception of storytelling, a point demonstrated again and again in Frog I and //. Narrators don't construct a narrative simply to fit a story, selecting their words, phrases, and rhetorical devices to express the elements of a pre-determined conceptualization of events. They also pick and choose among the potential elements of a story to find ones they are capable of expressing. That, of course, is one of the ideas behind thinking-forspeaking. You see a guy leaving a room in a hurry. In English, you might speak about him running out of a room (The man is running out of the room), whereas in Spanish, you would speak about him going out of a room (El hombre solid de la sala) - with no mention of speed (see, e.g., Slobin, this volume). Viewed this way. English and Spanish encourage you - perhaps even require you - to tell different stories of what you saw. Each of these choices may seem inconsequential, but in the aggregate they can lead to very different stories. In brief, if we are to understand the process of narrating, it is essential to understand the story being narrated - how it was created, where it came from, what its features are. It makes a difference whether the story is fact or fiction, third-person or first-person, created from scratch or retold. And it makes a difference what common ground - largely cultural - the narrator presupposes in narrating the story.

3. VARIATION 2: MIMESIS VERSUS DIEGESIS Another distinction in storytelling, dating back to Plato, is between diegesis and mimesis. With diegesis, narrators describe things, and with mimesis, they show things, as in direct quotation. Both devices are useful in narratives. As David Lodge (1990:144) put it, "Roughly speaking, mimesis gives us the sense of reality in fiction, the illusion of access to the reality of personal experience, and diegesis conveys the contextualising information and framework of values which provide thematic unity and coherence." And both devices are common in narrative. As Lodge went on, "[The] alternation of authorial description and characters' verbal interaction remains the woof and warp of literary narration to this day". The frog stories offer a striking example of Lodge's observations. Mercer's picture book has only one character who can talk, so the frog stories should contain no conversation at all. And yet they are filled with mimesis. But how and why?


463

3.1 What is mimesis? Showing, in its everyday sense, takes two main forms. The first is indicating. I can show you my car by pointing at it - by indicating it as my car. "There it is," I say, using the direction of my finger as an index to the car. The second form of showing is demonstrating. I can show you how my brother walks by doing a limp across the floor. "Here's how my brother walks with his sprained ankle," I say, while limping. By doing the limp, I demonstrate something for you, namely how my brother moves. What Plato and Lodge mean by mimesis, pretty clearly, is demonstrating and not indicating. Indicating and demonstrating deploy what the philosopher Charles Sanders Peirce called indexes and icons. Indicating is a method of signaling by which a person means something by creating an index for an addressee, and demonstrating is a method by which a person means something by creating an icon for an addressee (see Clark 1996, Chapter 6). For Peirce, indexes and icons contrast with symbols (e.g., words and emblematic gestures), which are used in pure descriptions. That is, indicating and demonstrating contrast with describing-as. As Peirce himself noted, most signals rely on a combination of these methods. It is rare to find a pure indication, demonstration, or description. How people demonstrate is often treated as self-evident, but it is hardly that. I will illustrate with the analysis of quotations as a species of demonstration (Clark & Gerrig 1990). Demonstrations, Gerrig and I argued, are selective depictions that are intended to "enable others to experience [in part] what it is like to perceive the things depicted" (Clark & Gerrig 1990:765). Take this excerpt from a frog story by a Turkish adult I will call Eda, as translated into English (Aksu-Koc & Tekdemir, this volume): AH started climbing on the rock to escape the owl. But what did he see? Actually he didn't see anything. He thought what he saw was a tree and holding onto a branch he started yelling, "Little frog! Little frog!"

Suppose that in producing "Little frog! Little frog!" Eda cupped her hands around her mouth and spoke slowly and loudly ("Li-i-tt-le fro-o-g! Li-i-tt-le fro-o-g!"). With this entire performance, she intended to create a selective depiction of what the boy, Ali, did. She tried to depict not only his words (little frog, little frog), but also the loudness and intonation of his voice, and the motion of his hands. And yet she was not trying to depict, for example, his gender or the direction in which he yelled. Demonstrations therefore partition into four separable aspects:

464

Herbert H. Clark

;'.

Depictive aspects. Only some aspects of a demonstration are intended to depict features of the referent. Eda's demonstration was intended to depict the words, intonation, loudness, and hand formation of Ali's original. These constitute the demonstration proper.

ii.

Supportive aspects. Other aspects of a demonstration are not themselves depictive, but are necessary as support in the performance of the depictive aspects. The gender and direction of Eda's voice were needed for her demonstration, but she didn't intend them to be depictive.

//'/'.

Annotative aspects. Still other aspects of a demonstration are added as commentary on what is being depicted. These might include concurrent laughter or sneers, or exaggerated intonation.

iv.

Incidental aspects. The remaining aspects of a demonstration are incidental, ones the speaker has no specific intentions about. These might include the way Eda sat, blinked her eyes, or pursed her lips.

Eda intends only part of her full demonstration to be depictive. Her addressees would have misunderstood her if they had construed her supportive or incidental aspects as depictive too. In general, people intend different aspects of their demonstrations to be depictive, supportive, and annotative. Gerrig and I called this the decoupling principle. The issue is how speaker and audience coordinate on this decoupling. Producing and understanding demonstrations relies on imagination and pretense. Eda must begin by imagining Ali calling the frog and ask. "How can I get my audience to imagine his actions?" One way is by describing (diegesis). "... he started yelling for his frog," but that leaves all the work of imagining to the audience. Another way is to create a selective depiction that enables the audience to experience, in part, what it is like to see, hear, and feel Ali calling the frog. That experience should make it easier to imagine the right scene (see Clark & van der Wege 2001). Another element needed for interpreting demonstrations is joint pretense. Eda's audience must join with her in the belief that, in producing "Li-i-tt-le fro-o-g! Li-i-tt-le fro-o-g!" with cupped hands, she herself isn't calling a frog. She is pretending to be the boy calling the frog, a type of play-acting, and she expects her audience to recognize the pretense. Like all fiction, quotations are to be understood against a joint pretense by speaker and addressee (Clark 1996, Chapter 12). Eda's thinking-for-speaking, then, includes imagining the right scene and determining which aspects to depict and how. It also includes establishing the right joint pretense.


465

3.2 Mimesis in the frog stories The frog narratives contain a remarkable variety of mimesis, or demonstrations. It is even more remarkable how often the different varieties were used in combination. Pictures Mayer's picture book is itself entirely mimesis. It is a series of drawn depictions, and it takes skill to read these as a story. Indeed, the very title of Mayer's book is a bit of mimesis, the quotation "Frog, where are you?" Understanding depictions isn't easy. To interpret Vincent van Gogh's painting The potato-eaters, according to Walton (1990), we must use the oil shapes on the canvas to help us imagine a fictional world in which there are two men and three women, poor Dutch farmers, sitting around a table eating potatoes and drinking tea. We must imagine ourselves in that world because we are viewing this tableau from within the room over the left shoulder of one of the women. We must also attribute the appropriate motivations to these farmers. We might assume, for example, that they are eating and drinking because they are hungry and thirsty. But are we to infer that they are starving, or that they have nothing to eat but potatoes? In interpreting this painting, then, a) we see the shapes on canvas as depictions of things; b) we imagine a coherent world in which those things reside; c) we understand that that world is fictional; and d) we infer states and attributes that are not directly depicted. Interpreting Frog, where are you? requires even more. When we view Picture 1, we imagine a fictional world with a boy, a dog, a frog in a jar, and a bed, all in a bedroom at night. But when we view Picture 2, we don't create a new fictional world: we simply add a next scene to the fictional world of Picture 1. We see the boy, dog, frog, jar, bed, and bedroom of Picture 2 as identical to those of Picture 1, but at a later time. For other pictures, we are supposed to add attributes and motivations - curiosity, fear, happiness - that are not directly depicted in the drawings. So in interpreting the picture book, e) we see the successive pictures as successive states of the same world; andy) we see them as representing a story - a complicating action followed by its resolution. Mayer's pictures enable us, in short, to "experience [in part] what it is like to perceive the things depicted" - the story of the boy and his frog. Narrators of the frog stories often used the pictures as props for their storytelling as well. As many authors in this volume noted, narrators often pointed at the pictures. They indicated things for their audience, showing them in the indicative sense of show. These acts piggy-back on the pictorial demonstrations. Suppose Eda points at the owl as she says, "Ali started climbing on the rock to escape the owl." Eda isn't pointing at the referent of the owl, which is fictional and nowhere to be found. She is pointing at a depiction of the owl, which is to be interpreted against the joint pretense that the thing depicted is part of an actual world.

466

Herbert H. Clark

Quotation Quotations appear in one frog narrative after another in the examples cited in this collection - in Tzeltal, Icelandic, Swedish, West-Greenlandic, Spanish, Turkish, Japanese, Basque, and American Sign Language. What is more, many of these quotations are particularly sophisticated, as these examples illustrate (in English translation): /.

Saying versus thinking. A Tzeltal-speaking adult (Brown) used quotations not only for speech, but for thoughts: "Maybe he says to himself ..." and "they are thinking ..." An Icelandic nine-year-old (Ragnarsdottir & Stromqvist) did the same: " 'Where has it gone?' she thinks."

//.

Manner of speaking. Swedish and English narrators (Stromqvist, Nordqvist & Wengelin) embedded direct quotations in a great variety of verbs, including ask, mutter, exclaim, whisper, and call. A WestGreenlandic narrator said: "Then he shouts 'Mummy.' " Swedish narrators also used a variety of adverbs with their verbs of speaking (e.g., "asked confusedly").

/'/;'.

Animal quotations. An American Sign Language narrator (Galvan & Taub) quoted the dog speaking (or rather signing): "The concerned dog barks up at the boy and says, 'Get down, get down.' " So did a Tzeltal narrator (Brown), " 'There's something here. There's something squatting here,' he says." A seven-year-old Spanish narrator, in contrast, quoted the dog barking ("The dog was au au au au au au!") as well as the bees buzzing ("and the bees were bbbbzzzzzzzz"). So the narrators sometimes depicted the dog's message and other times the dog's sounds. These reflect two distinct reasons for demonstrating.

iv.

Sound quotations. A West-Greenlandic narrator said "Shh maybe they/ some are there," quoting not only the boy's words but his saying of "shh."

v.

Prosody. Although one set of Swedish narrators were congenitally deaf (Stromqvist, Nordqvist & Wengelin), they used quotations of speech in their typewritten stories, complete with orthographic representations of prosodic effects. They typed "grodaaaaaa" to represent a prolonged "froooooog," and "H-A-L-L-A" and "H-J-A-L-P" to represent the shouting of hello and help.


467

All of these features are common for spoken and written quotations (Clark & Gerrig 1990). There are probably many more such features in a larger corpus of the languages represented here. Mimetics Most languages have linguistic techniques for representing the sounds and other features of animals, things, and events. In English, for example, James Joyce used pitapat in Ulysses, "And while she gazed her heart went pitapat," and a newspaper columnist used boom in a story, "So we have these Trident 2 missiles fall down and go boom." Words like these are variously called ideophones, onomatopoeia, mimetics, and sound symbolism, and they take different forms in different languages. Many of these words are conventional, such as pitapat, boom, cockadoodle doo, clunk, and ticktock in English. But narrators also make up their own words, as in the following English examples (Clark & Gerrig 1990): The pounding of the cylinders increased: ta-pocket-pocketa-pocketa-pocketapocketa. (James Thurber, The secret life of Walter Mitty) The room reeked of camphor. "Ugf, ahfg," choked Briggs ... (James Thurber, The night the bed fell] Everywhere you went you could hear the awful brrrpppbrrrpppbrrrppp saws. (Charles McGrath, Id)

of their

Among the languages represented in this volume, it is Japanese and Basque that make the most of these expressions. Mimetics are common in Japanese, as in these English translations "The bees flew buzzing" "Abat flew out, so [the boy] fell down thud" and "The frog sneaked out hop" (Kiintay & Nakamura). In the Basque stories, there were mimetics for quickly and suddenly in "And 'quickly' [they] jump onto the log," and "And 'suddenly' a mouse [comes] out" (IbarretxeAntunano). These mimetics work something like splat in "The tomato hit him splat in the face," in which splat describes the hit as done "with a smacking or splashing noise." Mimetics are also a type of demonstration (Clark & Gerrig 1990:788). When Joyce writes "her heart went pitapat," he doesn't mean that her heart produced literally these sounds. He intends only selected aspects of pitapat to be depictive - namely the rhythm. He could just as well have said dum-de-dum, ratatat, or even Peter Pan, and if he were speaking, he could have said it quickly or slowly to depict the heart's speed. The other features of pitapat are supportive aspects. In English, pitapat happens to be a conventional word for demonstrating a heartbeat, just as there are conventional words for demonstrating other sounds. So, like any demonstration, mimetics are selective depictions of their referents.

468

Herbert H. Clark

Iconic gestures Narrators often use iconic gestures, and although gestures weren't recorded in most languages of this volume, they were in Basque (Ibarretxe-Antunano) and American Sign Language (Galvan & Taub). Consider an example from a Basque adult: "And, 'suddenly,' a bird like this, em, an owl came [flew] out to him." Over the word flew, the narrator spread out his arms and flapped them like wings. We don't take the narrator to be trying to fly. We take him to be trying to depict an ow 1 flying, and to get his audience to use that selective depiction to imagine the owl flying. The Basque narrators used many such gestures. According to Galvan and Taub, signers use not only conventional signs, but iconic gestures or modifications that depict selective aspects of the referent. As they note, "Even lexical verbs will sometimes be spatially modulated to show specific paths and movements." And in a phenomenon called role shift (or referential shift} in sign language, "the signer's own face and body are taken to represent the face and body of some entity being described: the signer takes on that entity's 'persona' and demonstrates the entity's actions or experiences." These are demonstrations, and instances of joint pretense, for depicting entities. As Galvan and Taub noted, citing work by Engberg-Pedersen, signers have a narrative ideal: "In a signed story, wherever possible, narrators should 'show' the information using iconic forms, rather than 'describe' it using lexical forms." It is as if these narrators had read David Lodge's advice on how to write good fiction. Mimetics are often accompanied by iconic gestures - as if the two were of a piece. The Japanese speakers recorded by Kita (1997) all gestured whenever they used mimetics. Much the same is found in Basque (Ibarretxe-Antunano. this volume). When one Basque adult says, "And there goes the dog, very fast, running running running," he gestures a running motion in rhythm over the repeated word running. The repeated word is itself a type of sound demonstration (see Clark & Gerrig 1990:789), and the narrator fills it out with a second type of demonstration - an iconic gesture. Kita's and Ibarretxe-Antunano's evidence suggests that the type of thinking needed for speaking here is a perceptual, imaginal representation. Without that, it would be impossible to formulate these composite auditory and visual demonstrations. 3.3 Imagining-for-speaking Demonstrations add a notable wrinkle to Slobin's thinking-for-speaking. First, demonstrations are ubiquitous in narratives. They take the form of illustrations, quotations (both direct and free indirect quotations), mimetics of various types, iconic gestures, and forms I haven't even mentioned. To narrate, as Lodge noted, is both to tell and to show. Second, demonstrations are a central component of effective narratives. In an unpublished study of my own, I had 30 students listen


469

to a brief story and then retell it to a partner. Later, I asked other students to rate the quality of these 30 retellings. Rated quality rose with the amount of quotation used: the percentage of words that were in quotations accounted for about 25% of the variance in the judged quality of the narratives. And, third, child narrators use demonstrations - quotations, iconic gestures, mimetics - from the very start. (In a diary study by Eve Clark (unpublished), one child used his first quotation at age 2.) Oddly enough, the use of quotations and mimetics in narratives decreases over age in Japanese and Turkish (Kttntay & Nakamura). These cultures appear to place a value on diegesis as the proper form of narration, as do some Englishspeaking cultures. It is as if too much mimesis is bad. Still, children are fluent in its use. Demonstrations, or mimesis, require a radically different form of thinkingfor-speaking, or so I have argued (Clark 1996, Chapter 6). To formulate a gardenvariety quotation, you must imagine a person speaking and gesturing - the person's appearance, or what he or she looks and sounds like - and then select which aspects of that scene you are going to depict. You must select aspects not merely that you wish to depict, but that you are able to depict with devices you have at hand for demonstrating - tone of voice, intonation, concurrent gestures, the ability to lisp, slur words, or speak in dialect. Only then can you formulate the quotation itself. The selection process is no different, really, from the selection process behind describing motion events. There, too, you must select those features - path, manner, tense - not merely that you wish to express, but that your language - Tzeltal, Basque, Spanish - allows you to express.

4. VARIATION 3: CONCEPTION VERSUS DESCRIPTION Motion events can be described in many ways. In English, I can describe one and the same physical event in these two ways: [1]

Ken ran out of the room.

[2]

Ken left the room at a run.

Intuitively, these differ in how I conceptualize the event. In [1], I think about Ken trotting, sprinting, moving fast, and I think about this motion in relation to its direction - out of the room. In [2], I think about Ken in relation to the room - that he is no longer in it - and I think about that change of state in relation to the manner in which it happened - at a run. According to these intuitions, in formulating the descriptions in [1] and [2], I conceptualize the event in two distinct ways. What I have just described is, of course, an example of Slobin's thinking-for-speaking.

470

Herbert H. Clark

Here I take up four issues on this topic: a) no matter what the language, speakers must analyze scenes in order to describe them; b) formulating descriptions is a matter of distribution and detail from these analyses; c) analyzing scenes depends on community expertise; and d) analyzing scenes is incompatible with certain views of embodied cognition. 4.1 Scene analysis Every description requires an analysis of the scene being described. Suppose I am looking at five goats. In order to describe them in a single phrase, I must analyze the scene into certain features, dimensions, aspects, parts, relations. For English, I might identify three elements: a} the individual animals; b) the classification of each individual as a goat; and c) the number of those individuals. The result: five goats. For Mandarin, I would analyze the scene into four elements: a) goatiness; b) the individuals; c) the classification of each individual as an animal; and d) number. The result (in its English equivalent): five head of goats. If I know the goats by name, I might analyze the scene into the five individuals: Billy, Gruff, Beardo, Wilma, and Speedy. It would be strange, though possible, to analyze the scene as follows: a mammal, a thing, two goats, and Billy. People can conceptualize the same scene in perhaps an unlimited number of ways. What are the components, parts, dimensions, aspects, or relations of such an analysis - of such a process of abstraction? One of the only ways to answer the question is to look at what is required for descriptions. Here are a few examples for spatial descriptions (see, e.g., Talmy 2000a, 2000b): Dimensions We cannot describe a door as tall, wide, or thick, or a soup as thick or thin, or a town as to the north, without analyzing the door, soup, and local area into physical dimensions such as height, width, thickness, and north-southness (Bierwisch 1967). Directions Nor can we describe a person as tall versus short, or a lake as deep versus shallow, without knowing the positive and negative directions on these dimensions. Tall and deep are measures in the positive direction, and short and shallow in the negative direction, presumably for perceptual reasons (Clark 1973). Trajector and landmark To describe a cup as on a table, we must conceptualize the cup in relation to the table, and not vice versa. We must see the cup as trajector and the table as landmark - as thing related versus relatum (or as Talmy's Figure versus Ground).


471

Some relations seem easier to see as trajector and landmark than vice versa. These include: /. ii. Hi.

small objects in relation to large objects, as in the cup is on the table; moving objects in relation to static ones, as in the man ran by that chair, objects in the focus of attention in relation to objects not in the focus of attention.

All spatial prepositions presuppose trajector and landmark, and so do many spatial verbs, adjectives, and even nouns. Spatial relations In English, we cannot select among at or on or in without analyzing a scene into such spatial relations as abutment, support, and containment, notions that themselves may need to be analyzed further. With-respect-to relations The relation of trajector to landmark is merely the prototype of a more general relation I will call with-respect-to, or wrt. We ordinarily see: /. ii. Hi. iv. v.

parts with respect to wholes, as in the handle of the cup; dimensions with respect to wholes, as in width of the door, directions with respect to origins or goals, as in north of London, northerly; attributes with respect to what they are attributes of, as in the flavor of your coffee, or the color of that cloth; actions with respect to actors, as in the singing of the contralto.

Chains of relations. With-respect-to relations can be combined into chains. In describing a room, for example, I see a desk chair next to a desk, and the desk in the corner. I don't see the desk chair in the corner as such (see Shannon 1984). And I see the brown color of the chair as belonging to the chair, which is next to the desk in the corner of the room. The brown doesn't belong to the desk or to the corner of the room. All of this is common sense, but only because we are so used to such analyses. It is often assumed that features of the scene itself, or our perception of those features, dictate how we conceptualize it, but that isn't entirely so. We are readily able, for example, to see a table with respect to a cup - as supporting, or under, the cup - although that may take more time (Clark & Chase 1974). We can also see objects with respect to their parts, dimensions, directions, attributes, and actions,

472

Herbert H. Clark

even if that, too, may take more time. The with-respect-to relation appears to be universal - although this is always a dangerous claim. To describe the relation between two things, we are usually forced to conceptualize one of them with respect to the other. Symmetrical relations are rare. 4.2 Distribution and detail Speakers must analyze motion events in order to describe them. According to Talmy (2000a, 2000b), the main elements of such an analysis are Motion, Path, and Manner, and they get combined in characteristic ways. Verb-framed languages prefer expressions like Ken left the room running, in which Motion and Path are combined in the verb (left}, and Manner is elsewhere (running). Satellite-framed languages prefer expressions like Ken ran out of the room, in which Motion and Manner are combined in the verb (run) and Path is elsewhere (out of the room). Other languages keep all three components separate, as in Ken went running out of the room (see Slobin, this volume). These patterns may only be preferences. English, for example, allows all three major patterns. But motion descriptions often consist of much more than Motion, Path, and Manner, a point demonstrated in many of the contributions to Frog I and //. Motion descriptions are always to be understood against presuppositions about the cultural situation - against people's common ground. And although some elements of these events are expressed in the description, other pieces are to be inferred. That is, languages differ in their distribution of information among: a) basic elements such as verb, object, and satellites to express Motion, Path, and Manner; b) other content of these expressions; c) auxiliary expressions; and d) inferences from common ground. Let me illustrate these points with the expression on the sofa. If I were to ask a friend what it means to be "on the sofa," she might reply, "Is this a trick question?" The answer seems obvious. But consider: the old woman is on the sofa', the couple was making love on the sofa', I looked for the manufacturer s tag on the sofa; she put antimacassars on the sofa', there are oak legs on the sofa; the sun is on the sofa; we put new upholstery on the sofa (cf. Searle 1978). We cannot locate the woman, the couple, the tag, or the upholstery simply by understanding the "locative" on the sofa. Information about their location is distributed among several elements: a) the meanings of on; b) presuppositions we make in our culture about sofas; c) presuppositions we make in our culture about women, couples, tags, and upholstery (e.g., where manufacturer's tags are found and where antimacassars go); and d) the cw-relations afforded by sofas. And when we need to, when on the sofa wouldn't be enough, we add auxiliary information, as in there is a smudge on the left arm of the sofa. Talmy's schema, then, represents only part of the information we normally use in describing motion events. For example, individual languages have devices


473

for enriching these descriptions, as illustrated again and again in this collection. Here are three such devices. Tzeltal would be classified as a verb-framed language, and indeed, narrators use a lot of verbs such as enter, exit, and arrive. But that doesn't do justice to their descriptions of motion events. As Brown documents, Tzeltal has a large inventory of positional verb roots, such as lie-face-upward—spread-eagled, low-crouch, and insert-tightly, and narrators use them to imply manner. When a narrator says "He [dog] looks like he's low-crouching walking [= he's limping]," he describes the position of the dog in a low-crouch, which implies a manner of walking. Tzeltalspeakers exploit their large inventory of positional verbs for work that narrators of other verb-framed languages (such as Spanish) might neglect or that narrators of satellite-framed languages like English would do with motion verbs (such as limp). Japanese is also a verb-framed language, but that doesn't explain the common use of mimetics in describing motion events. Recall the four- and five-year-olds who used examples such as "The bees flew buzzing" and "The frog sneaked out with a hop" where the italicized words are mimetics analogous to English thud or splat. Mimetics were also common in Basque. In contrast, Turkish-speakers availed themselves of all of the same resources as Japanese-speakers - except for mimetics, which they almost never used (see Kiintay & Nakamura). American Sign Language, as I noted earlier, makes liberal use of another type of demonstration called role shift or referential shift. In this device, "the signer's own face and body are taken to represent the face and body of some entity being described; the signer takes on that entity's 'persona' and demonstrates the entity's actions or experiences." Signers used these shifts to enrich their motion event descriptions beyond the bare bones of Talmy's Motion-Path-Manner schema. 4.3 Community expertise One of the oldest, and best documented, observations about language is that vocabulary size reflects shared community expertise (see Clark 1998). Lawyers have a large vocabulary of legal terms, doctors a large vocabulary of medical terms, and car enthusiasts a large vocabulary of cars and car parts. If a culture deals in chili peppers or rice, it will develop specialized terms for types and subtypes of chili peppers and rice, where other cultures will not (Berlin 1972; Burling 1970). Note that legal, medical, and car terms aren't part of the vocabulary of the general English-speaker, even though they are English words. It is the shared expertise of the community that counts. It takes a special community expertise to learn and maintain some of the enrichment devices just mentioned. The large number of positionals in Tzeltal, for example, suggests that Tzeltal-speakers are experts on the shapes and orientations of natural objects and artifacts in daily life. Indeed, as Berlin (1968) documented,

474

Herbert H. Clark

Tzeltal also has over 500 nominal classifiers. These are words comparable to English sheet and head in r\\'o sheets of paper and five head of goats. They classify natural objects and artifacts mainly on the basis of animacy, shape, and orientation - much as the positional do. Other enrichment devices, in contrast, seem unrelated to community expertise. Japanese maintains a large number of mimetics, but are the Japanese in general any more expert on the referents of these mimetics than, say, the Turks? According to Slobin (this volume), English maintains a larger vocabulary of manner verbs (e.g., limp, amble] than Spanish, but is this a reflection of differences in cultural expertise? The obvious alternative is that the Japanese language maintains mimetics, and the English language manner verbs, precisely for enriching descriptions of motion events in Japanese and English. The conclusion might be this: In descriptions of motion events, narrators enrich the skeletal Motion-Path-Manner schema with devices that are unique to their language. Some of these reflect an expertise that is special to that community, but others reflect a communicative expertise that is maintained simply for its utility in communication. 4.4 Embodied cognition There is much talk nowadays of embodied cognition, the idea that people think about entities as physical objects - things that are seen, felt, and experienced holistically. Theories of embodied cognition come in two main varieties. One variety assumes that people think about dogs, owls, bees, boys, running, falling, and flying as wholes and do not analyze them into parts, features, components, aspects, or relations (e.g., Glenberg 1997). The second variety, which I myself have championed (see Clark 1973; Clark & Chase 1972), assumes that even if people can and do think about objects and events as wholes for some purposes, they must also analyze them into aspects, features, dimensions, parts, and relations for other purposes (see, e.g., Barsalou 2003). Let me call these two views the holistic and analytic versions of embodied cognition. If we have learned one thing from Frog I and //, it is that the holistic version of embodied cognition is incomplete. For people to describe a motion event, they must, at some point, analyze that event into parts, features, components, aspects, relations - at least the elements of Motion, Path, and Manner. They need other elements if they are ever to formulate descriptions that include dimensions (e.g., the large deer}, directions (off over the cliff), trajectors and landmarks (owl in the hole}, as well as other parts, features, and attributes of objects (startled dog). We may well be able to view a scene holistically. But in preparation for speaking, we cannot stop there. We must abstract out all those elements that are needed in formulating effective descriptions in our particular language.


475

One argument often offered for the holistic version is that iconic gestures are based on holistic representations of objects and actions. But imagine June moving her right hand in small circles along a downward trajectory as a way of describing a ball rolling down a hill. Are June's gestures based on a holistic representation of the rolling ball? Not at all. Her gestures are a selective depiction - a depiction of only selected elements of the rolling-ball scene - and it takes an analysis of the scene to select those elements. Iconic gestures require just as much abstraction of elements as descriptions do. Recall the Basque narrator who demonstrated the flying owl. He depicted only selected aspects of the owl - that it was an animated being with projections at its side that went up and down. The Basque audience couldn't have understood his gesture without decoupling those aspects from the rest of his performance and using them to help imagine the scene the narrator was trying to depict. The same goes for quotations, mimetics, and drawings. Mimesis, like diegesis, is based on elements abstracted from the scene being described.

5. SUMMING UP At a party, 1 told a friend that I was reviewing two volumes of research on kids telling stories about frogs. Two volumes? About frogs? And this is research? I realized too late how absurd it all must have sounded. But, in fact, the project has been a remarkable success. It is a model of how to carry out comparative research on language use. The recipe is simple. Find a rich, but circumscribed, event that people can describe in any language. Record people describing that event in many languages and at many ages. Then compare. Some years ago, Wallace Chafe followed this recipe with a dialogue-free film about pear pickers, and the results have been influential (see Chafe 1980). Mayer's picture book is not only more portable than Chafe's film, but also more accessible to children. Already, it has elicited a remarkable line of research, and its future looks even rosier. I have considered three variations on our ranarian theme. What is the story in contrast to the telling of the story? What is mimesis, or demonstration, in contrast to diegesis, or description, and how is this distinction played out in the development of storytelling? And how do we conceptualize a motion event on our way to describing it? These questions have brought home - for me at least - an important point: Thinking-for-speaking is more than preparing for syntax, morphology, and lexicon. It is preparing for a factual versus fictional story, for a first- versus third-person story, and for a story told from scratch versus a retelling. It is also preparing to demonstrate versus describe - and I haven't even mentioned how we might prepare to indicate things. As it happens, demonstrating, describing, and indicating are normally found in combination - in composite signals - so preparing to describe, demonstrate, and indicate will not be simple.

476

Herbert H. Clark

For any piece of a story, narrators must select which of the three methods to use and how to combine them. This research reflects an odd paradox about speaking and thinking. We seem to know much more about speaking than we do about thinking. The problem is that much of what we do know about thinking has come, explicitly or implicitly, from what we have discovered about thinking-for-speaking. How will we ever break this cycle?

NOTES 1.

Narrators could have told the frog story in the first person - say, from the boy's point of view - but as far as I can tell, none did.

Appendices


Appendix I: Frog, where are you?

Pictures reproduced from Mayer (1969), with permission of the author/artist and publisher. Original format: 25 x 14.5 cm, sepia-tone, one single panel or one-half double panel per page, no text.

Appendix I: Mercer Mayer. The frog stoty

480

-4-

-5-

Appendix I: Mercer Mayer, The frog story

-6-

481

-7-

-8-

482

Appendix I: Mercer Mayer, The frog ston-

-9-

-10-

-11-


483

-12-

-13-

-14-


484

-15-

-16-

-17-


485

-18-

-19-

-20-

-21-

Appendix I: Mercer Mayer. The frog ston-

486

-22-

Appendix IIa: Frog-Story Research as per 1994 Research using Mercer Mayer's Frog, where are you? as per 1994. This is a reformatted and slightly updated reprint of Appendix III of the 1994 volume please note that the contact information may no longer be valid. Published papers listed below have been included in the list of References of the 2004 volume.

FIRST LANGUAGE The Berkeley Project: English, German, Hebrew, Spanish, Turkish Language(s) dealt with in each paper are indicated by the letters E, G, H, S, T. Aksu-Ko? (1988a, 1992, manuscript: T) Bamberg(1985, 1987, 1990, 1992:G; 1991:E; 1994a: E, G); Bamberg, Budwig & Kaplan (1991: G); Bamberg & Damrad-Frye (1991: E); Bamberg & Marchman (1990, 1991: E, G) Berman (1988, 1990, 1993c, 1996: H); Berman & Slobin (1987: E, G, H, S, T) Marchman (1989: E) Renner(1988: E) Sebastian (1989: S) Slobin (1987, 1991, 1996b: E, G, H, S); Slobin (1988, 1993a: T); Slobin, (1989, 1990: E, G, H, S, T); Slobin (1993b: E, G, S, T); Slobin (1996a: E, S); Slobin & Bocaz (1988: E, S) American Sign Language Bellugi, Ursula, Laboratory for Language and Cognitive Studies, The Salk Institute for Biological Studies, P.O. Box 85800, San Diego, CA 92186-5800 ([email protected]) (with Edward S. Klima, Judy Snitzer Reilly) Gal van, Dennis, Dept. of Psychology, Gallaudet University, 800 Florida Ave., N.E., Washington, DC 20002 ([email protected]) [Age range: 3, 5, 7, 9 (native ASL), 5, 9, adult (late signers)] (Ref.: Galvan, 1988, 1989) Arabic (Moroccan) Kail, Michele, Laboratoire de Psychologie Experimental, CNRS, 28 rue Serpente, 75006 Paris, France ([email protected]) (with Mohammed Bamhamed, Abdelkebir Drissi) [Age range: 3, 4, 5] (Ref: Kail & Drissi, 1990)

488

Appendix Ha

Arandic Languages (Australia) Wilkins, David P., Dept. of Linguistics, SUNY at Buffalo, Buffalo, NY 14260 (lindav [email protected]) Languages: Mparntwe Arrernte, Western Arrernte. Alyawarra [Age range: adult]

Bulgarian Olszewski, Krzysztof, Dept. of General Linguistics, Jagiellonian University, Mickiewicza 9, Krakow, Poland (with Magdalena Smoczyriska, [email protected]) [Age range: child, adult]

Dutch Dijkhuis, Karin, Marja Roelofs ([email protected]), Carla Zijlenmaker: Dept of Linguistics, University of Amsterdam, Spuistraat 210, 1012 VT, Amsterdam. Netherlands (with Anne Mills, [email protected]. uva.nl) [Age range: 4-9. adult] (Ref: Dijkhuis, manuscript.; Zijlenmaker, 1992, 1993)

English (American) Cowan, Carolyn, Philip Cowan, Institute of Human Development, University of California. Berkeley, CA 94720 ([email protected]) [Age range: child] Donahue, Mavis, College of Education, m/c 147, Box 4348, University of Illinois. Chicago, IL 60680 ([email protected]) [Age range: adult (mother-child)] Gemsbacher, Ann, Dept. of Psychology, University of Wisconsin. WI 53706 ([email protected]) [Age range: adult 3. (comprehension test)] (Ref: Gemsbacher, 1985) Harkins, Debra A., Dept. of Psychology, Suffolk University, Beacon Hill, 41 Temple St.. Boston, MA 02114-4280 [Age range: adults (mothers to 4-5-year-olds)] (Ref: Harkins, 1993) Hemphill, Lowry, Projects in Language Development, 306 Larsen Hall, Harvard Graduate School of Education, Cambridge, MA 02138 ([email protected]) [Age range: 5-8] (Ref.: Hemphill, Wolf, Camp & Griffin, 1993; Miranda. Camp. Hemphill & Wolf, 1992) Barzilay, Margaret, Elite Olshtain, School of Education, Tel Aviv University. Ramat Aviv. Israel 66978 [Age range: adult (special condition: native Americans compared with American immigrants in Israel; study of language attrition)] (Ref.: Olshtain & Barzilay, 1991)

Frog story research as for 1994

489

Reilly, Judy Snitzer, Department of Psychology, San Diego State University, 6363 Alvarado Court #221, San Diego, CA 92120-4913. ([email protected]) [Age range: 3-4, 7-8, 10-11] (Ref.: Reilly, 1992) Trabasso, Tom, Department of Psychology, University of Chicago, 5848 S. University Ave., Chicago, IL 60637 ([email protected]) (with Camille Baughn, Margret Park Munger, Margaret Nickels, Philip C. Rodkin, Nancy Stein) [Age range: 3, 4, 5, 9] (Ref.: Trabasso & Nickels, 1992; Trabasso, Stein, Rodkin, Munger & Baughn, 1992) Manhardt, Joan, Lowry Hemphill, Helen Tager-Flusberg (see below under Language/ Developmental Impairment)

English (Australian) Wigglesworth, Gillian, Department of Applied Linguistics, University of Melbourne, Parkville, Victoria 3052, Australia ([email protected]. au) [Age range: 4, 6, 8, 10, adult] (Ref.: Wigglesworth, 1992) Hebert, Julie (see below under Second Language/Bilingual)

English (British) Lieven, Elena, Dept. of Psychology, University of Manchester, Manchester Ml3 9PL, U.K. ([email protected]) [Age range: 3-5] Martin, J.R., Dept. of Linguistics, University of Sidney, Sydney, Australia [Age range: 6-7, 8-9, 10-11] (Ref.: Martin, 1977, 1983)

Finnish Dasinger, Lisa, Dept. of Psychology, University of California, Berkeley, CA 94720 (dasin [email protected]) [Age range: 3-9, adult] Lauren, Christer, Karita Mard, Swedish Immersion Research Project, Dept. of Scandinavian Languages, University of Vaasa, P.O. Box 297, SF-65101 ([email protected] wasa.fi) [Age range: 6-8] Toivainen, Jorma, Dept. of Finnish and General Linguistics, University of Turku, 20500 Turku, Finland ([email protected]) (with Riitta Jokinen, Kati Juhola, Elisa Koskinen, Annikka Kotiranta, Susanna Leveelahti, Sirpa Niiniviita, Liisa Ollila, Risto Palttala, Marja Raukola, Henna Seppa, Kirsti Toivainen) [Age range: 3, 5, 7, 9 1 0 , adult] (Ref.: Toivainen, 1989, 1992)

490

Appendix I la

Finno-Permic Languages Toivainen, Jorma, Dept. of Finnish and General Linguistics, University of Turku, 20500 Turku, Finland ([email protected]). Estonian - Paul Alvre, Ago Kunnap. Tiina Mihailov (Dorpat) Finnish (Sweden) - Raija Kangassalo (Umea), Komi (or Zyrian) - Jevgeni Igushev (Syktyvkar), Mari (or Cheremis) - Valentin Vasiljev (Joshkar-Ola), Mordvinian - Mihail Mosin (Saransk), Udmurt (or Votyak) Alexandr Sutov (Izhevsk), Veps - Nina Zaiceva (Petrozavodsk) [Age range: 3, 5, 7, 9]

French (France) Hickmann, Maya, Michele Kail, Laboratoire de Psychologic Experimental, CNRS, 28 rue Serpente, 75006 Paris, France ([email protected]) (with Francoise Roland) [Age range: 6, 9, 11] (Ref.: Hickmann, Kail & Roland, 1989; Hickmann. Kail & Roland, 1993; Kail & Hickmann, 1992) Jisa, Harriet, Centre de Recherches Linguistiques et Semiologiques, Universite Lumiere - Lyon 2, 69500 BRON, France ([email protected]) (with Sophie Kem) [Age range: 3-12, adult] (Ref.: Jisa & Kern, manucript)

Galician Garcia Soto, Jose Ramon, Dept. of Developmental and Educational Psychology, University of Santiago de Compostela, Campus Universitario, 15707 Santiago, Spain (with Miguel Perez-Pereira, Urzaiz 77, 2° B, 36201 Vigo, Spain; [email protected]) [Age range.: 4-2, adult] (Ref.: Garcia Soto, 1993) Rodriguez-Trelles, Ana, Patio de Madres 16, Santiago de Compostela (La Coruna). Spain (with Miguel Perez-Pereira, Urzaiz 77, 2° B, 36201 Vigo, Spain; [email protected]) [Age range.: 4, 5, 7, 9, 11, adult] (Ref.: Rodriguez-Trelles, 1991)

Guugu Yumithirr (Pama-Nyungan: Australia) Levinson, Stephen, Cognitive Anthropology Research Group, Max-Planck-Institute for Psycholinguistics, PB 310, NL-6500 AH Nijmegen, Netheralands ([email protected]) [Age range: adult]

Icelandic Ragnarsdottir, Hrafnhildur, University College of Education, v/Stakkahlid, 105 Reykjavik. Iceland ([email protected]) [Age range: 3^4, 5, 7, 9, adult] (Ref.: Ragnarsdottir. 1987, 1988, 1991, 1992)


Italian Orsolini, Margherita, Istituto di Pedagogia e Psicologia, Via Madonna degli Angeli 30, 66100 Chieti, Italy; Clotilde Pontecorvo and Franca Rossi, Dip. Psicologia, Via de Marsi 78, 00185 Rome, Italy [Age range: 4, 5 6, 8 10, adult] (Ref.: Orsolini, Rossi & Pontecorvo, manuscript; Rossi, manuscript) Bottari, Piero, Anna Maria Chilosi, Paola Cipriani, Elisabetta Lanzetta, Lucia Pfanner: Laboratory for the Study of Normal and Pathological Child - Language Acquisition, Stella Maris Foundation and Institute of Child Neuropsychiatry of the University of Pisa, Viale del Tirreno 331, 56018 Calambrone - Pisa, Italy ([email protected] e.cnr.it / [email protected]) [Age range: 4-11]

Japanese Fujii, Seiko Yamaguchii, Dept. of East Asian Languages and Cultures, and Dept. of Linguistics, 608 S. Matthews Ave., University of Illinois, Champaign-Urbana, IL 61801 ([email protected]) [Age range: adult] (Ref.: Fujii, 1992, 1993) Nakamura, Keiko, Dept. of Psychology, University of California, Berkeley, CA 94720 ([email protected]) [Age range: 3, 4, 5, 7, 9, adult] (Ref: Kiintay & Nakamura, 1993; Nakamura 1993, manuscript) Takahashi, Kazumi, 550 N. Harvard Ave., Claremont, CA 91711 ([email protected] aremont.edu) [Age range: adult] (Ref.: Takahashi, manuscript) Uchida, Nobuko, Bunkyoiku Gakbaku, Ochanomizu Women's University, 2-1-1 Otsuka, Bunkyo-yo, Tokyo 112, Japan [Age range: 3-5] (Ref.: Uchida, 1982, 1983)

Kgalagadi (Bantu: Botswana) Neumann, Sabine, Cognitive Anthropology Research Group, Max-Planck-Institute for Psycholinguistics, PB 310, N-6500 AH, Nijmegen, Netherlands ([email protected]) [Age range: adult]

Kickapoo Gomez de Garcia, Jule, Campus Box 295, Department of Linguistics, University of Colorado, Boulder, CO 80309 ([email protected]) [Age range: child, adult]

Kilivila (Austronesian: Trobriand Islands, Papua New Guinea) Senft, Gunter, Cognitive Anthropology Research Group, Max-Planck-Institute for Psycholinguistics, PB 310, NL-6500 AH Nijmegen, Netherlands ([email protected]) [Age range: 12 - adult]

492

Appendix I la

Lakhota van Valin, Robert D. Jr., Dept. of Linguistics, SUNY at Buffalo, Buffalo. NY 14260 ([email protected]) [Age range: adult]

Longgu (Austronesian: Solomon Islands) Hill, Deborah, Cognitive Anthropology Research Group, Max-Planck-Institute for Psycholinguistics, PB 310, NL-6500 AH Nijmegen, Netherlands (dhillfampi.nl) [Age range: 12, 16, adult]

Malay Peng, Yap Siew, 64 Pinngir Zaaba, Taman Tun Drive, Ismail 6000, Kuala Lumpur. Malaysia [Age range: 5, 7, adult] (Ref.: Peng. 1992)

Mandarin (Peoples Republic of China) Guo, Jiansheng, Dept. of Psychology, University of California, Berkeley. CA94720 ([email protected]) [Age range: 3, 4, 5. 9, adult]

Mopan (Mayan: Belize) Danziger, Eve, Cognitive Anthropology Research Group, Max-Planck-Institute for Psycholinguistics, PB 310, NL-6500 AN Nijmegen, Netherlands (danzigerf2mpi.nl) [Age range: child, adult]

Myene (Bantu: Gabon) Ogouamba, Pierrette Andree, Centre de Recherches Linguistiques et Semiologiques. Universite Lumiere - Lyon 2, 69500 BRON. France [Age range: 6.8,9. 10. 11, 12]

Nilotic Languages Bavin, Edith. Dept. of Linguistics, La Trobe University, Bundoora. Victoria forado.edu) 3083, Australia ([email protected]) Western Nilotic Languages: Acholi. Alur, Dhopodhola, Lango [Age range: adult]

Norwegian Lanza, Elizabeth, Dept. of Linguistics, University of Oslo, Post Box 1102. Blindern. 0317 Oslo, Norway ([email protected]) [Age range: 4-8]


493

Polish Smoczynska, Magdalena, Dluga 50-5, 31-146 Krakow, Poland ([email protected]) [Age range: child, adult]

Portuguese Hub Faria, Isabel, Departamento de Linguistica, Faculdade de Letras, Universidade de Lisboa, 1699 Lisboa codex, Portugal [Age range: child, adult]

Russian Durova, N. V., Institute of the National Problems of Education, Moscow, Russia; N. M. Yurieva, Institute of Linguistics, Russian Academy of Sciences, ul. Semashko, dom 1/12, 103009 Moscow K-9, Russia ([email protected]) [Age range: 3-6] (Ref.: Durova & Yurieva, manucript) Slobin, Dan I., Dept. of Psychology, University of California, Berkeley, CA 94720 (slobin @cogsci.berkeley.edu) (with Yana Mirsky) [Age range: 4-10, adult] Ushakova, Tatiana, Institute of Psychology, Russian Academy of Sciences, Moscow (with Michele Kail: see above under French) [Age range: 5, 9, 11, adult]

Sign Language of the Netherlands Hoiting, Nini, Royal Institute for the Deaf "H. D. Guyot", Rijksstraatweg 63, 9752 AC Haren, Netherlands (with Dan I. Slobin) [Age range: child, adult] (Ref.: Hoiting & Slobin, 1993)

Spanish (Argentina, Chile) Bocaz, Aura, Universidad de Chile, Versailles 2977, Depto 94, Las Condes, Santiago, Chile (with Dan I. Slobin) [Age range: 3, 4, 5, 7, 9, 11, adult] (Ref.: Bocaz, 1989a, 1989b, 1989c, 1991a, 1991b, 1991c, 1991d, 1992a, 1992b; Slobin & Bocaz, 1988)

Spanish (Mexico) Montes, Rosa Graciela, Institute de Ciencias Sociales y Humanidades, Universidad Autonoma de Puebla, Apdo. Postal 1356, 72001 Puebla, Pue, Mexico (montes@cca .pue.udlap.mx) [Age range: child, adult]

494

Appendix Ha

Spanish (Spain) Lopez-Ornat, Susana, Dpto. Psicologia Cognitiva, Campus de Somosaguas, Universidad Complutense de Madrid 28223, Spain ([email protected]) (with Javier del Castillo; Michele Kail, Maya Hickmann: see above under French) [Age range: 3, 4, 5,6,7,8,9, 10, 11, adult] Sera, Maria D., Institute of Child Development, University of Minnesota, 51 East River Rd., Minneapolis, MN 55455 [Age range: 3, 4, 5, 9, adult (data of Sebastian)] (Ref.: Sera, 1992)

Swedish Lauren, Christer, Karita Mard (see above, under Finnish) [Age range: 11-12] Stromqvist, Sven, Dept. of Linguistics, University of Goteborg, S-412 98 Goteborg, Sweden ([email protected]) [Age range: 14—15 (videotaped 94720 (oral stories and computer-logged written stories)]

Tamil (Dravidian: India) Pederson, Eric, Cognitive Anthropology Research Group, Max-Planck-Institute for Psycholinguistics, PB 310, NL-6500 AH Nijmegen, Netherlands ([email protected]) [Age range: 10, adult]

Totonac (Totonacan: Mexico) Levy, Paulette, UNAM, Seminario de Lenguas Indigenas, Instituto de Investigaciones Filologicas, Circuito Mario de la Cueva, 04510 Mexico, D.F., Mexico [Age range: adult]

Turkish Kiintay, Aylin, Dept. of Psychology, University of California, Berkeley, CA 94720 (kuntay @cogsci.berkeley.edu) [Age range: 3, 4, 5, 7, 9, adult] (Ref.: Kiintay, 1992; Kiintay &Nakamura, 1993) 1988)

Tzeltal (Mayan: Mexico) Brown, Penelope, Cognitive Anthropology Research Group, Max-Planck-Institute for Psycholinguistics, PB 310, NL-6500 AH Nijmegen, Netherlands ([email protected]) [Age range: 5, 6, 7, 12, adult]


495

Tzotzil (Mayan: Mexico) Gutierrez, Esteban, Linguistics Dept., Reed College, Portland, OR 97202 osaguas, ([email protected]) (with: John Haviland, Lourdes de Leon) [Age range: child, adult] de Leon, Lourdes, Linguistics Dept., Reed College, Portland, OR 97202 7, 8, 9, 10, ([email protected]) [Age range: 3, 5, 8, 10] (Ref.: de Leon, manuscript)

Warlpiri (Australia) Bavin, Edith, Dept. of Linguistics, La Trobe University, Bundoora, Victoria 3083, Australia ([email protected]) [Age range: 3—4, 6-7, 11-13, adult]

Yiddish (USA) Slobin, Dan I., Dept. of Psychology, University of California, Berkeley, CA 94720 (slobin @cogsci.berkeley.edu); Judith L. Slobin, Norval L. Slobin, 22801 Essex Way, Apt. 511, Southfield, MI 48034 [Age range: child, adult]

Yucatec Mayan (Mexico) Lucy, John, Suzanne Gaskins, Dept. of Anthropology, University of Pennsylvania, Philadelphia, PA ([email protected])

Yupno (Papua New Guinea) Wassmann, Jurg, Cognitive Anthropology Research Group, Max-Planck-Institute for Psycholinguistics, PB 310, NL-6500 AH Nijmegen, Netherlands ([email protected]) [Age range: adult] 4-8]

496

Appendix I la

SECOND LANGUAGE/BILINGUAL Arabic (Moroccan)-Dutch Verhoeven, Ludo, Linguistics Department, Universiteit Brabant, PB 90153. NL-5000 LE Tilburg, Netherlands ([email protected]) (with Petra Bos) [Age range: 4-12] [Language of narration: Arabic, Dutch] (Ref.: Verhoeven, 1993)

Finnish-Swedish Lauren, Christer, Karita Mard (see above, under Finnish) [Age range: 11-12 (Finnish LI)]

German-Russian Dittmar, Jeannette, FB-16, Germanistik, Freie Universitat Berlin, Germany [Age range: 5;1 - 6;5 (one child, longitudinal)] [Language of narration: German, Russian]

Greek-English Haritos, Calliope, Katherine Nelson: Developmental Psychology, City University of New York Graduate Center, 33 W. 42nd St., New York, NY 10036 [Age range: adult]

Guugu Yumithirr-Australian Aboriginal English de Leon, Lourdes, Dept. of Linguistics, Reed College, Portland, OR 97202 ([email protected]) [Age range: 6-14, adult] [Language of narration: Guugu Yimithirr, Australian Aboriginal English] Norwegian—English Lanza, Elizabeth, Dept. of Linguistics, University of Oslo, Post Box 1102, Blindern. 0317 Oslo, Norway ([email protected]) [Age range: 4—8] [Language of narration: Norwegian, English]

Spanish-Basque (Spain) Urrutia Cardenas, Hernan, Fac. de Ciencias de la Communicacion, Universidad del Pais Vasco, Bilbao, Spain; Carmen Silva-Corvalan, Dept. of Spanish and Portuguese. University of Southern California, Los Angeles, CA 90089 ([email protected]) [Age range: 5-6] [Language of narration: Spanish, Basque] (Ref: Urrutia Cardenas. 1992)


497

Spanish-English (U.S.A.) Silva-Corvalan, Carmen, Dept. of Spanish and Portuguese, University of Southern California, Los Angeles, CA 90089 ([email protected]) [Age range: 5-7] [Language of narration: Spanish, English] (Ref: Silva Corvalan, 1992)

Turkish-English (Australian) Hebert, Julie, 18 Evans St., Moonee Ponds 3039, Australia [Age range: 6-8] [Language of narration: English (English LI, Spanish LI, Turkish LI)] (Ref.: Hebert, 1990) Howell, Christine, Adult Migrant Education Service, Melbourne, Victoria, Australia [Age range: adult] [Language of narration: English (English LI, Spanish LI, Turkish LI)] (Ref: Howell, 1993)

Turkish-Dutch Verhoeven, Ludo, Linguistics Dept., Universiteit Brabant, PB 90135, NL-5000 LE Tilburg, Netherlands ([email protected]) (with Jeroen Aarssen) [Age range: 4-12] [Language of narration: Turkish, Dutch] (Ref: Verhoeven, 1993)

Turkish-French (France) narration Jisa, Harriet, Centre de Recherches Linguistiques et Semioiogiques, Universite Lumiere - Lyon 2, 69500 BRON, France ([email protected]) (with Sophie Kern) [Age range: 6-11, adult] [Language of narration: French (Turkish LI)]

498

Appendix Ha

LANGUAGE/ DEVELOPMENTAL IMPAIRMENT English (American) Bates, Elizabeth, Center for Research in Language, C-008, University of California at San Diego, La Jolla, CA 92093 ([email protected]) (with Virginia Marchman, Judy Snitzer Reilly, Joan Stiles, Beverly Wulfeck) [Type of impairment: specific language impairment, early focal brain injury] (Ref.: Reilly, Marchman & Bates, in press) Bellugi, Ursula, Laboratory for Language and Cognitive Studies, The Salk Institute for Biological Studies, P.O. Box 85800, San Diego, CA 92186-5800 (llcs(Ssalksci.sdsc.edu) (with Edward S. Klima, Judy Snitzer Reilly) [Type of impairment: Down Syndrome, Williams Syndrome] [Age range: 10-18] (Ref: Reilly, Klima & Bellugi, in press) Chapman, Robin, Department of Communicative Disorders, University of Wisconsin, 1975 Willow Dr., Madison, WI 53706 ([email protected]) [Type of impairment: Down Syndrome] [Age range: child, adolescent] Feldman, Heidi, Child Development Unit, Children's Hospital of Pittsburgh, Pittsburgh, PA 15213 (hfOa+ @andrew.cmu.edu) [Type of impairment: perinatal brain injury] Fine,

Jonathan, Dept. of English, Bar-Han University, Ramat-Gan, Israel ([email protected]) [Type of impairment: reading disability] [Age range: 10-12] (Ref: Fine, 1985, in press)

Hemphill, Lowry, Graduate School of Education, Harvard University, Roy E. Larsen Hall. Appian Way, Cambridge, MA 02138 (with Linda Camp and Dennis Palmer Wolf) [Type of impairment: mildly mentally retarded (and nonretarded)] [Age range: 11 (retarded), 7 (nonretarded)] (Ref.: Wolf, Hemphill & Camp, 1991 ) Manhardt, Joan, Dept. of Human Development, Bryn Mawr College, Bryn Mawr, PA 19010 ([email protected]) (with Leslie A. Rescorla) [Type of impairment: specific expressive language delay] [Age range: 8 (compared with 8 normally developing children)] (Ref: Rescorla & Manhardt, manuscript) Oram, Janis, School of Communication Sciences and Disorders, McGill University. Montreal, Quebec, Canada H3G 1AB ([email protected]) (with Martha B. Crago, Myrna Gopnik) [Type of impairment: familial specific language impairment] [Age range: Tager-Flusberg, Helen, Dept. of Psychology, University of Massachusetts, Boston. MA 02125-3393 (htagerf%umbsky.dnet@ns. umb.edu) (with Lowry Hemphill, Nancy Picardi) [Type of impairment: mildly mentally retarded (and nonretarded), autistic] [Age range: 10;11 (mean age, retarded), 7;8 (mean age, nonretarded)] (Ref: Hemphill, Picardi & Tager-Flusberg, 1991; Tager-Flusberg, 1993)


499

van Valin, Robert D. Jr., Dept. of Linguistics, SUNY at Buffalo, Buffalo, NY 14260 ([email protected]); David P. Wilkins, Dept. of Linguistics, SUNY at Buffalo, Buffalo, NY 14260 ([email protected]); Nina Dronkers, University of California at Davis and VA Medical Center, Martinez, CA ([email protected] eley.edu); Jeri Jaeger, Dept. of Psychology, SUNY at Buffalo, Buffalo, NY 14260 ([email protected]) [Type of impairment: aphasia] [Age range: adult]

Dutch Blankenstijn, Claudia, Annette Scheper: Department of General Linguistics, University of Amsterdam, Spuistraat 210, 1012 VT Amsterdam, Netherlands (with Anne Mills, [email protected]) [Age range: 4-9] [Type of impairment: psychiatric disturbance]

French Kail, Michele (see above under French) [Type of impairment: mentally retarded (and nonretarded)] [Age range: 17 (retarded), 9 (nonretarded)]

Hebrew Schiff, Rachel, Dept of Education, Bar-Han University, Ramat Gan, Israel. Bryn [Type of impairment: reading disability (proficient and non-proficient readers)] [Age range: 10-13 (oral and written stories)] (Ref: Schiff, 1991)

Italian Bottari, Piero, Anna Maria Chilosi, Paola Cipriani, Elisabetta Lanzetta, Lucia Pfanner (see above, under Italian) [Type of impairment: Down Syndrome, Williams Syndrome, SLI, dysphasic] [Age range: 4-16]

Turkish Slobin, Dan I. (with Dr. Oget Oktem-Tanor, Dept. of Neurology, Capa Medical School, Istanbul University, Istanbul, Turkey) [Type of impairment: aphasia] [Age range: adult]


Appendix lib: Frog-Story Research after 1994 Research using Mercer Mayer's Frog, where are you? after 1994. The supplementary information presented in this appendix is derived from an webbased questionnaire administered to the research community in 2002. It presents for each project: •

Project title;

•

Language(s) under investigation;

•

Acquisition type (First Language Acquisition [FLA], Second Language Acquisition/Bilingualism [SLA/BL], Language/ Developmental Impairment [LDI]);

•

Age of subjects (in years, or years and months);

•

Registration technique (usually audio, video, or both);

•

Name of principal investigator(s) [PI] and co-researcher(s) [CR];

•

Contact information;

•

Project-related publications; and

•

Accessibility of data.

The list is roughly sorted according to acquisition type and language under investigation; note, however, that some projects cover more than one language or acquisition type. Papers that has been published or accepted for publication, or otherwise is cited in the chapters of this volume, is included in the References section. In the list below, such material is referred to with author(s) and year only. Manuscripts in preparation or under review are detailed with titles as reported by the investigators.

Appendix lib

502

Languages)

Project title

Acq. type

Other/ Remark

Age range

Registr. technique

Principal investigator Co-researchers

Cantonese

Development of temporality in Cantonese children's stories

FLA

3- 11, adult

Audio

PI: S. Leung

English

A new perspective on the relation between preschoolers' narrative ability and later academic achievement

FLA

3.4

Video

PI: D. K. O'Neill CR: M. J. Pearce

English

Young preschoolers' ability to reference story characters: the contribution of gesture and voice

FLA

3.4

Video

PI: D. K. O'Neill C R : A . C . Holmes

Finnish. ErzyaMordvinian

Acquisition of Finnic languages

FLA

3-9. adult

Audio

PI: J. Toivainen CR: S. VuonokariKamarainen: CR: N. Mosina (Erzya)

German

Development of narrative competence in German

FLA

5-8. adult

Audio

PI: U. Stephany CR: A. Weiss

German, French, Italian, Romansh

Spatial expressions in satellite-framed and verbframed Languages

FLA

Adult

Audio

PI: R. Berthele

Greek

Development of narrative competence in Greek

FLA

3- 11. adult

Audio

PI: U. Stephany

Japanese

First language acquisition of Japanese

FLA

4. 9, adult

Audio

PI: Ch. Sakurai

Japanese

Narrative development in Japanese

FLA

3-9, adult

Audio

PI: K. Nakamura

Korean

Frog stories in Korean

FLA

5. 7. adult

Audio

P I : K . Oh

Mandarin

Becoming a better narrator FLA in Mandarin Chinese

3.4.5.9. adult

Audio

PI: J. Chen-Wilson

Spanish

Peruvian mothers' elicitation and storytelling styles

3.5

Audio

P I : G . Melzi

FLA

Mother-child conversations

503

Frog-Story Research after 1994

Contact person

Project related publications

Accessibility of data

Sam Leung Division of Speech & Hearing Sciences Faculty of Education University of Hong Kong [email protected] Daniela O'Neill Dept. of Psychology University of Waterloo, Canada [email protected]

Manuscript in preparation

Transcripts on request, CHILDES archive in future

Daniela O'Neill Dept. of Psychology University of Waterloo, Canada [email protected]

O'Neill & Holmes (2002)

Transcripts on request, CHILDES archive in future

Jorma Toivainen Dept. of Finnish and Gen. Ling. University of Turku, Finland [email protected] Ursula Stephany Inst. of Linguistics University of Cologne, Germany [email protected] Raphael Berthele Departement fur Germanistik Universitat Freiburg/Fribourg Switzerland [email protected] Ursula Stephany Inst. of Linguistics University of Cologne, Germany Stephany @uni-koeln.de Chikako Sakurai [email protected] Kei Nakamura Inst. of Cognitive & Linguistic Studies Keio University, Tokyo, Japan [email protected]

Toivainen (1992); Mosina (2002)

Finnish transcripts on request.

Stephany (1997a, 1998); Weiss (2002)

Transcripts on request

Kyung-ju Oh Psychology Dept. University of California, USA [email protected] Josephine Chen- Wilson Newman College of Higher Education Birmingham [email protected] Gigliana Melzi Dept. of Applied Psychology New York University, US [email protected]

On request

Stephany (1997a, 1997b, 1998)


Transcripts on request Nakamura (1993) (Ed.): Japanese/Korean Linguistics Volume 3 (pp. 84-99). Stanford, CA: CSLI. Transcripts on request

Part of the study presented as a poster at the Child Language Seminar at the University of Sheffield, September 1998


504

Language(s)

Appendix lib

Project title

Acq. type

Other/ Remark Information structure. word order. prosody

Age range

Registr. technique


Adults. elder children

Audio. video

PI: A. Lindvall PI: M. Tronnier

FLA

3;0. 3:6. 4:0. 4:6

Audio

PI: M. Eriksson CR: E. Berglund CR: A. Nordquist

First language acquisition of Thai

FLA

4, 6. 9. 1 1 . adults

Audio

PI: J. Zlatev CR: P. Yangklang

Tzeltal (Mayan)

First language acquisition ofTzeltal

FLA

3 to 16, adults

Basque

Motion events in Basque and Spanish

FLA SLA/ BL

Adults

PI: P. Broun Audio, some video, computer transcripts PI: I. IbarretxeAudio. video Antunano

British, American, and Nicaraguan Sign Language Dutch

Cross linguistic sign language development

FLA SLA/ BL

Samoan, Fijian

Information structure in a typological perspective. with special reference to Samoan

FLA

Swedish

Validation of SECDI

Thai

Dutch, English Dutch, English, Spanish Dutch, English, Spanish

Language deprivation (homesign)

3 - adult

Video

PI: G. Morgan CR: J. K e g l e t a l .

Children with specific language impairment

5,7.9

Audiotape

PI: J. Steenge CR: L. Verhoeven CR: Hans van Balkom

SLA/ BL SLA/ BL

Adult (12+) Adult

Audio + video Audio video

PI: E. Kellerman CR: R. V'ermeulen PI: E. Kellerman CR: A-M. van Hoof

SLA/ BL

22-46

Video

PI:A-M. van Hoof CR: E. Kellerman (supervisor)

FLA SLA/ BL

12. 15. 18

Audio

PI: A. Housen PI: M. Pierrard CR: L. Mettewie CR: S. Janssens CR: N. Costers

Language impairment in FLA Turkish, Moroccan and SLA/ Surinamese children in the BL Netherlands

Gesture placement in motion events: A crosslinguistic study of LI and L2 'thinking-forspeaking' patterns Dutch & The development of Dutch French (L2); and French language French, Dutch proficiency in Dutch (LI) language schools in Brussels


Contact person

505


Accessibility of data Transcripts on request

Ann Lindvall/MechtildTronnier Dept. of Linguistics and Phonetics Lund University, Sweden [email protected] [email protected] Marten Eriksson Dept of Education and Psychology University of Gavle, Sweden

Eriksson (2001); Berglund (2000); Nordqvist(2001)


Jordan Zlatev Dept. of Linguistics and Phonetics Lund University, Sweden [email protected]

Zlatev & Yangklang (in press); Zlatev & Yangklang (in preparation). The Thai frog story corpus: Transcription and computerized analysis of 50 Thai narratives from five age groups. Brown (2000)

CHILDES

Iraide Ibarretxe-Antunano Faculty of Philosophy and Letters Deusto University Bilbao, Basque Country [email protected]

Ibarretxe-Antunano (under review-a); Ibarretxe-Antunano (under review-b). Language typologies in our language use: The case of Basque motion events in oral narratives; Ibarretxe-Antunano (under review-c). Basque: A verb-framed or a satellite-framed language?; barretxe-Antunano (in preparation) Linguistic typology in motion events: Path and manner


Gary Morgan LCS City University, London, UK

Morgan (1996a, 1996b, 1999a, 1999b, 2002)

Videoarchive restricted; some transcripts available

J.Steenge Dept. of Educational Science University of Nijmegen Nijmegen, The Netherlands J. [email protected]

Steenge, Verhoeven & van Balkom (2002)

CHILDES after end of project 2004

Eric Kellerman [email protected]

Vermeulen & Kellerman (1999)

On request

Eric Kellerman [email protected]

Kellerman & van Hoof (2002)

On request

Anne-Marie van Hoof Max Planck Inst. for Psycholinguistics Nijmegen, The Netherlands [email protected]

van Hoof (2000, 2001)

Not accessible

Penelope Brown Max Planck Inst. for Psycholinguistics Nijmegen, The Netherlands [email protected]

Alex Housen Dept. of Germanic Languages Vrije Universiteit Brussel, Belgium A lex. Housen@ vub.ac.be

Transcripts (in Tzeltal) on request

Recordings and transcripts on request

Appendix lib

506

Languages)

Project title

Acq. type

Dutch (LI), Spanish (LI), Papiamentu (LI andL2) English, Arabic

Part of undergraduate final thesis work on gesture in motion events

FLA SLA/ BL

Other/ Remark

Age range

Registr. technique

Principal investigator Co-researchers PI: E. Kellerman CR:A. Sisilia

FLA Sociolinguistic SLA/ competence and the bilingual, s adoption of BL phonetic variants: auditory and instrumental data from English-Arabic bilingual children.

5,7, 10, adult

Audio

PI: G. KJiattab

PI: J. Cenoz CR: L. Garcia Lecumberri CR: M. Pilar Garcia Mavo CR: Y. Ruiz CR: A. Doiz CR: D. Lasagabaster CR: R. Jimenez CR: P. Gallardo CR: J. Gutierrez CR: E. Gomez PI: A. Housen PI: M. Pierrard CR: S. Janssens

The influence of age on third language acquisition

SLA/ BL

8-18

Audio, video

French and English as foreign languages in secondary schools in Flanders Dutch, French The development of Dutch (L2); French, and French language proficiency in Dutch Dutch ( L I ) language schools in Brussels Second language English acquisition of English in (L2); Italian, French. Dutch educational contexts (LI) Bilingual narrative English and development in English Japanese and Japanese

SLA/ BL

18

Audio

FLA SLA/ BL

12, 15, 18

Audio

SLA/ BL

9. 11, 13, 15, 17

Audio

SLA/ BL

6-12

Audio

PI: M. Minami CR: E. Fujiyama CR: S. Fukuda CR: H. Mogi

7-8; 10-11

Audio

PI: B. Z. Pearson CR: O.K. Oiler. CR: A. B. Cobo-Lewis CR: V. Umbel CR: M. Fernandez CR: R.E. Filers CR: V.C. Muller Gathercole

English, Basque, Spanish

English & French (L2); Dutch ( L I )

English, Spanish

Language and literacy in bilingual children; narrative probe study

FLA SLA/ BL

Different stages of second language acquisition, in a factorial design

PI: A. Housen PI: M. Pierrard CR: L. Mettewie CR: S. Janssens CR: N. Costers PI: A. Housen CR: G. Pallotti


507



Eric Kellerman e.kellerm an (2ilet.kun.nl

Khattab (2002) Ghada Khattab Dept. of Speech, School of Education Commnication and Language Sciences University of Newcastle-upon-Tyne, UK ghadakhattab@hotmail .com


Jasone Cenoz [email protected]

Cenoz (2001, 2002a, 2002b, in press); Garcia Mayo (in press); Garcia Mayo & Garcia Lecumberri (in press); Garcia Lecumberri & Gallardo (in press); Lasagabaster & Doiz (in press)

In preparation for CHILDES

Alex Housen Dept. of Germanic Languages Vrije Universiteit Brussel, Belgium [email protected] Alex Housen Dept. of Germanic Languages Vrije Universiteit Brussel, Belgium [email protected]

Housen, Janssens & Pierrard (2001)


Alex Housen Dept. of Germanic Languages Vrije Universiteit Brussel, Belgium [email protected]

Housen (1998, 2000, in press-a, in press-b, in press-c, in press-d)


Masahiko Minami Dept. of Foreign Languages and Literatures San Francisco State University, US [email protected] Barbara Zurer Pearson Dept. of Communication Disorders University of Massachusetts, US [email protected]

Minami (2002); Minami (in preparation): Holding on to a native tongue: Retaining bilingualism for school-age children of Japanese heritage

CHILDES

Pearson (200 1,2002)

CHILDES; Miami corpus


Appendix lib

508

Language(s)

Project title

Acq. type

Ewe

Ewe grammar and discourse

Finnish, Swedish

Sprakrum

SLA/ BL

Hebrew

Hebrew language development and attrition among Israeli immigrant children in the US

FLA SLA/ BL

Hebrew

Emergent literacy in Hebrew as a foreign language among native speakers of English in the US. The frog story narratives of Irish-English bilinguals

SLA/ BL

Irish, English

Other/ Remark

SLA/' BL

Japanese

Oral and written narratives FLA SLA/ in Japanese BL

Japanese

Thinking-for-speaking in Japanese

Spanish, English

Identification of language impairment in SpanishSpeaking children

FLA SLA/ BL FLA SLA/ BL LDI


Audio

PI: F. K. Ameka CR: J. Essegbey

6-8

Audio

PI: G. Oker-Blom CR: V. Franzen CR: A. Palojam

First language development and attrition

6-13

Audio (oral) and writing

PI: D. Kaufman

Development of Hebrew as a foreign language

6-12

Writing

PI: D. Kaufman

Adults

Audiotaped and transcribed

PI: T. Bennett-Kastor

1;7^;0

Audio

PI: M. Shibata

Adults

Handwritten narratives. audio Audio, video

PI: K. Nakamura

SLA/ BL

Comparing lexical aspect and narrative discourse in second language learners' tense-aspect morphology

Registr. technique

Adults (12+)

Spatial language research; possible for codeswitching/ bilingualism

Japanese

Age range

L I is English

4, 6, adults

4-8

Audio

PI: K. Satoh

PI: V. F. GutierrezClellen CR: M. Adelaida Restrepo


Contact person



Felix Ameka Max Planck Inst. for Psycholinguistics Nijmegen, The Netherlands [email protected] [email protected]

Ameka & Essegbey (in press, to appear)


Vivan Franzen vi van. franzen@ling. lu.se

In preparation

In preparation

Dorit Kaufman Dept. of Linguistics Stony Brook University State University of New York, US [email protected] Dorit Kaufman Dept. of Linguistics Stony Brook University State University of New York, US [email protected] Tina Bennett-Kastor Dept. of English Wichita State University, US [email protected] Miki Shibata Okinawa University, Japan shibata@)okinawa-u. ac.jp

Kaufman (1995, 1998, 2000, 200 la, 200 Ib)

On request

In preparation

On request

Bennett-Kastor (2002).


Kei Nakamura Inst. of Cognitive & Linguistic Studies Keio University, Tokyo, Japan [email protected] Kyoko Satoh Yokohama City University, Japan [email protected] V. F. Gutierrez-Clellen Dept. of Communicative Disorders San Diego State University, US [email protected]

Transcripts on Shibata (manuscript-a). Linguistic marking of request perspective taking in L2 narrative discourse (1998); Shibata (manuscript-b). The use of Japanese tense-aspect morphology in L2 discourse narratives (1998); Shibata (manuscript-c). The use of the Japanese non-finite form -te in second language narrative discourse (2000); Shibata (manuscript-d). Comparing lexical aspect and narrative discourse in second language learners' tense-aspect morphology: a cross sectional study of Japanese as a second language (Ph.D. dissertation) (2000)

Satoh, K (200 la, 200 Ib, in press)

Gutierrez-Clellen (2001); Gutierrez-Clellen & DeCurtis (2001)


510

Languages)

Appendix Mb

Project title

Acq. type

Other/ Remark LI andL2 acquisition in a bilingual school, in a Spanish monolingual environment

Age range

Registr. technique


15

Audio

PI: C.L.Ordonez

Spanish, English

Oral bilingual proficiency of Colombian adolescents

FLA (Sp) SLA/ BL (En)

Spanish, Moroccan Arabic

The development of narrative abilities in Moroccan Arabic-Spanish bilingual children.

SLA/ BL

5-10

Audio

PI: N. Nouaouri

Turkish, Swedish

Language socalization in immigrant families and its relation to language learning in the Swedish preschool

5-6

Audio

PI: S. Boyd CR: K. Naucler

Dutch

The development of morphosyntactic and semantic-pragmatic abilities in Dutch-speaking children with psychiatric disorder

FLA Comparison (Sw) of adult child SLA;' interaction in different BL (Tu& activities, frog Sw) story, meal and play Psychiatric FLA LDI disordered

4-9

Audio and video

PI: C. Blankenstijn PI: A. Scheper

Dutch (as 1 st and 2nd language)

Language impairment in Turkish, Moroccan and Surinamese children in the Netherlands A cognitive neuropsychological approach to specific language impairment in children. Phase 1 and 2. Linguistic abilities in children with Williams syndrome and specific language impairment (PhD project) Specific language impairment: A syntactic or a processing deficit? Narrative skills in children w ith pragmatic language impairment, compared to autism and typical SLI Narrative capability in children with a history of autism

5.7.9

Audiotape

PI: J. Steenge CR: L. Verhoeven CR: H. van Balkom

6;4-9;8 (normal) 10;2-13;11 (SLI)

Audio (DAT) Transcriptions

PI: H. 1C. J. vd Lely CR: L. Stollwerck

7;6-12;0

Audio and video

PI: V. Stojanovik CR: M. Perkins CR: S. Howard (PhD supervisors)

13. 14. 15

Audio

PI: D. Wetherell CR: G. Conti-Ramsden

7-12

Audio and video

PI: D. Bishop CR: C. Norbury

5-9

Video

PI: E. Kelley CR: L. Naigles CR: D. Fein

English

English

English

English

English

FLA SLA,' BL

Children with specific language impairment

FLA LDI

Williams syndrome and specific language impairment Specific language impairment FLA LD!

FLA

Highfunctioning autism

511


Contact person


Claudia Lucia Ordonez Universidad de los Andes Bogota, Colombia [email protected]. co

Transcripts on Shiroetal. (2001); request Ordonez et al. (2001); Ordonez (to appear) Oral Bilingual Proficiency of Colombian Adolescents, Dissertation Abstracts International

Nadi Nouaouri Facultad de Filosofia y Letras Universidad de Cadiz, Spain [email protected] [email protected] Kerstin Naucler [email protected]

Not yet

In preparation for CH1LDES

Naucler &Boyd( 1996, 1997, 1999); Boyd& Naucler (2001); Naucler (2002)


Annette Scheper [email protected]

Blankenstijn & Scheper (2003)

No

J.Steenge Dept.of Educational Science Univ. of Nijmegen, The Netherlands J. [email protected] Heather K. J. van der Lely Dept. of Human Communication Science University College London, UK [email protected] Vesna Stojanovik Dept. of Human Communication Sciences University of Sheffield, UK. [email protected]

Steenge, Verhoeven & van Balkom (2002)

CHILDES archive after finishing the project (end of 2004)

van der Lely (1997)

Not publically available

None as yet


None as yet, ongoing PhD.

CHILDES

None as yet


Presentation at International Study for Research in Child language-Madison, WN, July 2002

Transcripts on request, not stored in CHILDES format but can be converted

Danielle Wetherell University of Manchester, England [email protected] Dorothy Bishop Dept. of Experimental Psychology University of Oxford, UK [email protected] Elizabeth Kelley Psychology Dept. University of Connecticut, US [email protected]


Appendix lib

512

Language(s)

Project title

Acq. type

Other/ Remark

Age range

Registr. technique


English

Narrative as a clinical tool

FLA

Pragmatic difficulties, SLI

8

Audio

PI: N. Boning

Typology, particularly rhetorical

Adult

Audio

PI: M. Noonan CR:H. Smith CR:K. Grunow-Harsta CR:M. Brehm

Adult

Audio

PI: E. Schultze-Berndt

LDI

Chantyal, Nar-Phu, Ghale, Magar, Baragaunle Jaminjung

Linguistic descriptions of Tibeto-Burman languages of Nepal

Jaminjung grammar and lexicography (ongoing project)

English

NICHD study of early child care and youth development

English

Narrative development in aggressive boys

Kilivila

The frog story in Kilivila

Swedish

PI: S. Friedman PI: K.. Hirsh-Pasek

Audio tape

PI: D. E. Cole

Audio and video

PI: G. Senft

Spoken versus written narrations

Adult

Audio. computerlogged writing

P I : R . Alves

Saliba, an Oceanic lanuage of Papua New Guinea

Stimulus for elicitation of narratives

10-45

Audio/ video

PI: A. Margerts

Speaker age in acoustics, perception, and synthesis

Phonetic voice-related correlates of age Controls in four age groups, congenitally deaf, dyslexic, aphasic subjects

6-91

Audio

PI: S. Schotz

9, 12, 15. adults

Audio. video. computerlogged writing

PI: S. Stromqvist CR:E. Ahlsen. CR:A. N'ordqvist. CR:A. Wengelin

Portuguese

Saliba (Oceanic, Papua New Guinea) Swedish

FLA SLA/ BL

9-11 Part of longitudinal large scale study (1300 children) Males at risk 8-12 for aggressive behavior Native speaker 12-55 language production

Reading and writing of functionally disabled groups

513


Contact person



Nicola Dotting School of Education University of Manchester, UK [email protected]

Botting (2002)


Michael Noonan Dept. of English University of Wisconsin Milwaukee, US [email protected] Eva Schultze-Berndt Max Planck Inst. for Evolutionary Anthropology Leipzig, Germany Sarah Friedman [email protected] Kathy Hirsh-Pasek [email protected]

Transcriptions of two Chantyal frog stories have been published in Noonan (1999)

Transcriptions on request, including morpheme glosses and translations

Schultze-Berndt (2000)


See website for list (nichd.nih.gov/crmc/secc)

Forthcoming public data set

Doris Cole Valparaiso University [email protected]

Cole (2001)


Gunter Senft Max Planck Inst. for Psycholinguistics Nijmegen, The Netherlands [email protected]

Senft (to appear).


Rui Alves Faculdade de Psicologia e de Ciencias da Educacao Universidade do Porto, Portugal [email protected] see also www.ScriptLog.net Anna Margetts Monash University, Australia [email protected]

Alves (2002); Alves et al. (in press)

On request

Margetts (1999a, 1999b, to appear)


Susanne Schotz [email protected]

Coming PhD diss.

On request

Sven Stromqvist Dept. of linguistics Lund University, Sweden [email protected] see also www.ScriptLog.net

Ahlsen(1999); Behrns (2002); Holmqvist, Johansson, Stromqvist & Wengelin (2002); Nordqvist(2001); Stromqvist, Hellstrand & Nordqvist (2000); Wengelin (2002); Wengelin & Stromqvist (2000)

Selected data available on CHILDES; release of more data in preparation


Appendix He: Typological Overview of Frog-Story Research The classification is made according to Ruhlen's A Guide to the World's Languages (1987, Vol. I: Classification. London: Edward Arnold) and with the help of information available at www.ethnologue.com and the ASIATIS-ASEDA Catalogue (coombs.anu.edu.au/SpecialProj/ASEDA). Altogether, there are 72 different languages and 13 phyla (language families) covered by research using Frog, where are you? (Phyla not represented in the research data are Khoisan, Caucasian, Chukchi-Kamchatkan, and Na-Dene.) In addition to the typological phyla, the three categories "Language isolates", "Pidgins and Creoles", and "Other" (sign languages are categorized here, for the time being) are all represented in the data. Languages to which papers in the present volume are dedicated are indicated with an asterisk.

Phylum

Group

Language

Niger-Congo

Ewe Kgalagadi Myene

Nilotic

Acholi Alur Dhopodhola Lango

Semitic

Arabic Hebrew*

Greek Italic

Greek French Galician Italian Portuguese Romansh Spanish* Irish

Niger-Kordofanian

Nilo-Saharan

Afro-Asiatic

Indo-Hittite

Celtic

516

Phylum

Appendix He

Group

Language

Germanic

Dutch English* German Icelandic* Norwegian Swedish* Yiddish Bulgarian Polish Russian

Indo-Hittite (cont.)

Balto-Slavic

Uralic-Yukaghir Finno-Ugric

Estonian Finnish Komi Mari Mordvinian Udmurt Veps

Turkic Korean Japanese-Ryukyuan

Turkish* Korean Japanese*

Inuit

Greenlandic"

Dravidian

Tamil

Sinitic

Cantonese Mandarin Baragaunle Chantyal Ghale Magar

Altaic

Eskimo-Aleut

Elamo-Dravidian

Sino-Tibetan

Tibeto-Burman

517

Typological Overview of Frog-Story Research

Phylum

Group

Language

Austronesian

Daic

Fijian Kilivila Longgu Malay Samoan Thai*

Trans-New Guinea

Yupno

JaminJungan Pama-Nyungan

Jaminjung Alyawarra Arrernte* Guugu Yumithirr Warlpiri*

Almosan-Keresiouan

Kickapoo Lakhota Saliba Mopan Totonac Tzeltal* Tzotzil Yucatec

Austric

Indo-Pacific

Australian

Amerind

Andean Penutian

Language isolates Basque* Pidgins and Creoles Portuguese-based

Papiamentu

Other American Sign Language* British Sign language Nicaraguan Sign Language Sign Language of the Netherlands Swedish Sign Language


Appendix III: Format and Abbreviations for Glosses* In running text, all foreign-language examples are given in italics, with translation equivalents in single quotes: kanunju 'down'; kanpora 'outside: ALL'. If necessary, a free translation may be added (within parentheses, preceded by an equal sign): gel-me-di-n 'come-NEO-PAST-2sc' (= 'you didn't come'). In interlinear format, foreign-language examples are numbered and given in ordinary font. A free translation in single quotes always follows: [1]

II traverse la riviere en nageant. 'He crosses the river swimming.'

When translation equivalents are given for each foreign element, they appear directly below it: [2]

Wati man

ka IPFV

nyina-mi sit-NONPAST

ngurra-ngka. camp-toe

'The man is in his camp.' Very long examples may be arranged column-wise instead, for readability. Both in running text and in interlinear format, grammatical codes are given in small caps (see list below). If particular words or phrases are of current interest to the exposition, they are highlighted by underlining, if nothing else is stated, in original and translation (when not aligned): y el nene siguio buscando a la rana 'and the boy continued searching for the frog'. The degree of precision of segmentation and glossing of an example depends, of course, on the role it plays in the exposition, and may occasionally vary within an example. [3]

og svo datt hundurinn og strakurinn 'and then fell the dog and the boy

ofan-i above-into

sjo sea:Acc'

Hyphens in a gloss always correspond to hyphens in the foreign example. If one foreign element corresponds to more than one English element and/or grammatical code, the collection of meaning equivalents is joined by colons; e.g.,

* The format used here is adapted from that presented in Slobin (1985-97); see also Comrie(1981).

Appendix III

520

gel-medin 'come-NEG:pAST:2so', or gelmedin 'come:NEG:pAST:2sc'. The use of the colon does not exclude the possibility of segmentation. If a grammatical code consists of two words or abbreviations, they are joined by a period; e.g., DEF.ART means "definite article". Combining the principles for use of colons and periods, consider the gloss for the German definite article in its masculine singular accusative form: den 'DEF.ART:MASC:SG:ACC'.

1 2 3 ABESS ABL ABS ACC ACT ADESS ADJ ADMON ADV AFFIRM AGR AGENT ALLAT AN ANTI AORIST APL ART ASP AUG AUX BEN CAUS CL CLASS CMPLR CNTR COMIT COMM COMPAR COMPL CONC COND

First Person Second Person Third Person Abessive ('without X') Ablative ( 'from X') Absolutive Accusative Active Adessive ('towards X') Adjective, Adjectival Admonitive Adverb(ial) Affirmative Agreement Agent Allative ('to( wards) X') Animate Antipassive Aorist Applicative Article Aspect Augmentative Auxiliary 1 Benefactive Causative Clitic Classifier Complementizer Contrastive Comitative ('(together) with X') Common Comparative Completive Concessive Conditional

CONJ CONN CONS EC CONT CONTEMP COP DAT DECL DEF DEICT DEM DER DESID DIM DIREC DO DU DYN FLAT EMPH EQU ERG ESS EVID EXCL EXIS EXP EXT FACT FEM FIN FOC FUT GEN GER HAB

Conjunction Connective Consecutive Continuous, Continuative Contemporative Copula Dative Declarative Definite Deictic Demonstrative Derived, Derivation Desiderative Diminutive Directional Direct Object Dual Dynamic (Nonstative) Elative ('out of X') Emphatic Equative Ergative Essive ('as X') Evidential Exclusive Existential Experiential Extension Factive Feminine Finite Focus Future Genitive Gerund Habitual

521

Format and Abbreviations for Glosses

HON HUM ILL IMP INAN INCH INCL INCOMPL INDEF INDIC INESS INF INFER INSTR INT INTENT INTERJ INTRANS 10 1PFV 1RR ITER LOG MASC MKR MOD N NEC NEUT NEUTRAL NOM NOME NONPAST NONVIR NUM OBJ OBL OBLIG OPT PART PARTIT PASS PAST PAT

Honorific Human Illative ('into X') Imperative Inanimate Inchoative Inclusive Incompletive Indefinite Indicative Inessive ('in X') Infinitive Inferential Instrumental Interrogative Intentive Interjection Intransitive Indirect Object Imperfective Irrealis Iterative Locative Masculine Marker Modal Noun Negative Neuter Neutral Nominative Nominal Non-past Non-virile Numeral, Numeric Object Oblique Obligatory Optative Participle Partitive Passive Past Patient

PERF PFV PL POL POSS POST POT PP PRE PREP PRES PRESUM PRET PRO PROG PROL PROLOG PTL PURP PVB Q QUANT QUOT RECENT RECIP REFL REL REM REPET REPORT RES SG SIMUL STAT SUBJ SUBJV SUBL SUFF SUPER SUPERL TAGQ TEMP TNS TOP

Perfect Perfective Plural Polite Possessive Postposition Potential Past Participle Prefix Preposition Present Presumptive Preterite Pronoun Progressive Prolative ('along X') Prolocative Particle Purposive Preverb Question Quantifier Quotative Recent Reciprocal Reflexive Relative Remote Repetition Reportative Resultative Singular Simultaneous Stative Subject Subjunctive Sublative ('onto X') Suffix Superessive ('on X') Superlative Tag Question Temporal Tense Topic

522

TRANS TRANSL V VIR VN VOC

Appendix III

Transitive Translative ('becoming X') Verb Virile Verbal Noun Vocative

References

Aarssen, J. (1996). Relating events in two languages: Acquisition of cohesive devices by Turkish-Dutch bilingual children at school age. Tilburg studies in multilingualism 2. Tilburg University Press. Aarssen, J. (2001). Development of temporal relations in narratives by Turkish-Dutch bilingual children. In L. Verhoeven & S. Stromqvist (Eds.), Narrative development in a multilingual context (pp. 209-232). Amsterdam/Philadelphia: John Benjamins. Aarssen, J. & Bos, P. (1999). Cohesive devices in bilingual development. In G. Extra & L. Verhoeven (Eds.), Bilingualism and migration (pp. 143-164). Berlin-New York: Mouton de Gruyter. Abelin, A. (1999). Studies in sound symbolism. Gothenburg, Sweden: Gothenburg Monographs in Linguistics 17. Ahlsen, E. (1999). Aphasia and text writing - an overview based on 9 cases. In S. Stromqvist & E. Ahlsen (Eds.), The process of writing- a progress report (pp. 121-129). Gothenburg papers in theoretical linguistics, Goteborg University, Department of Linguistics. [Also in Proceedings of the ISAAC 1998 Biennial Conference, University College, Dublin, 1998, pp. 4-5.] Aisenman, R. (Ed.) (1999a). Developing literacy across genres, modalities, and languages, vol. 1. Working papers, International Literacy Project, Tel Aviv University. Aisenman, R. (Ed.) (1999b). Developing literacy across genres, modalities, and languages, vol. 2: Relevant features of spoken and written language. Working papers, International Literacy Project, Tel Aviv University. Aisenman, R. & Assayag, N. (1999). Fact and fiction in evaluating narrative evaluation. In R. Aisenman (Ed.) Developing literacy across genres, modalities, and languages, vol. I (pp. 38-57). Working papers, International Literacy Project, Tel Aviv University. Akinci, M. A. (1999). Developpement des competences narratives des enfants bilingues Turc-Francais en France ages de 5 a 10 ans [The development of narrative competence in Turkish-French bilingual children in France aged 5 to 10 years]. (PhD. thesis). Universite Lumiere, Lyon 2. Akinci, M. A., Jisa H. & Kern, S. (1996). Developing overallplotline. Paper presented at the Workshop on Narrative Development in a Bilingual Context. Tilburg, February 1996. Akinci, M. A., Jisa, H. & Kern, S. (2001). Influence of LI Turkish on L2 French narratives. In L. Verhoeven & S. Stromqvist (Eds.), Narrative development in a multilingual context (pp. 189-208). Amsterdam/Philadelphia: John Benjamins.

524

References

Aksu-Ko9, A. (1988a). Simultaneity in children's narratives: The development of cohesion in discourse. In S. Ko9, (Ed.) Studies on Turkish linguistics (pp. 55-78). Ankara: Middle East Technical University Press. Aksu-Koc, A. (1988b). The acquisition of aspect and modality: The case of past reference in Turkish. New York: Cambridge University Press. Aksu-Ko9, A. (1992). The development of complement constructions. Paper presented at the 6th International Conference on Turkish Linguistics, Anadolu University, Eskiehir, Turkey. Aksu-Ko9, A. (1994). Development of linguistic forms: Turkish. In R. A. Berman & D. I. Slobin (Eds.), Relating events in narrative: A crosslinguistic developmental study (pp. 329-392). Hillsdale, NJ: Lawrence Erlbaum Associates. Aksu-Ko9, A. (1996). Frames of mind through narrative discourse. In D. I. Slobin, J. Gerhardt, A. Kyratzis & J. Guo (Eds.), Social interaction, social context, and language: Essays in honor of Susan Ervin-Tripp (pp. 309-328). Mahwah, New Jersey: Lawrence Erlbaum Associates, Inc. Aksu-Koc, A. (manuscript). A developmental analysis of temporality in narratives. Unpublished manuscript (1991), University of Istanbul. Aksu-Koc, A. & Kiintay, A. C. (2001). Reformulating causal relations while retelling narratives: Evidence from Turkish. Paper presented at the International Symposium on the Acquisition and Construction of Meaning in Crosslinguistic Perspective, Paris, France. Aksu-Ko9, A. & Tekdemir, G. (this volume). Interplay between narrativity and mindreading. A comparison between Turkish and English. In S. Stromqvist & L. Verhoeven (Eds.) (2004), Relating events in narrative: Typological and contextual perspectives. Mahwah, NJ: Lawrence Erlbaum Associates. Aksu-Ko9, A. & von Stutterheim, C. (1994). Temporal relations in narrative. In R. A. Berman & D. I. Slobin (Eds.), Relating events in narrative: A crosslinguistic developmental study (pp. 393—456). Hillsdale, NJ: Lawrence Erlbaum Associates. Allen, M. S., Kentoy, M. K., Sherblum, J. C. & Petit, J. I. (1994). Children's narrative productions: A comparison of personal events and fictional stories. Applied Psycholinguistics 15, 149-176. Allwood, J. (1986). Some perspectives on understanding in spoken interaction. In M. Furberg, T. Wetterstrom, & C. A. Berg (Eds.), Logic and abstraction (pp. 13-59). Acta Philosophica Gothoburgensia, vol. 1. University of Gothenburg: Department of Philosophy. Allwood, J. (Ed.) (1996), Talspraksfrekvenser [Frequency dictionary of spoken language]. Gothenburg papers in theoretical linguistics S 20. University of Goteborg: Department of Linguistics. Allwood, J., Nivre, J. & Ahlsen, E. (1992). On the semantics and pragmatics of linguistic feedback. Journal of Semantics 9, 1-26. Alves, R. A. (2002). Limites e processes cognitivos na composicao escrita [Cognitive constraints and processes in written compositions]. Master's thesis. Universidade do Porto: Faculdade de Psicologia e de Ciencias da Educacao.

References

525

Alves, R. A. Castro, S. L., de Sousa, L. & Stromqvist, S. (in press). The cognitive cost of execution in written composition. In D. Galbraith, L. van Waes & M. Torrance (Eds.), Recent developments in writing process research, vol. I: Basic processes and word-level effects. Dordrecht: Kluwer. Ameka, F. K. (2001). Ideophones and the nature of the adjective word class in Ewe. In F. K. E. Voeltz & C. Kilian-Hatz (Eds.), Ideophones (pp. 25^48). Amsterdam/ Philadelphia: John Benjamins. Ameka, F. K. & Essegbey, J. (in press). Serialising languages: Satellite-framed, verbframed or neither. In L. Hyman & I. Maddieson (Eds.), African comparative and historical linguistics: Proceedings of the 32nd Annual Conference on African Linguistics. Lawrenceville, NJ: Africa World Press. Ameka, F. K. & Essegbey, J. (to appear). Elements of Ewe grammar of space. In S. C. Levinson & D. P. Wilkins (Eds.), Grammars of space. Cambridge University Press. Andersen, R. W. (1984). The one to one principle of interlanguage construction. Language Learning 34, 4, 77-95. Andersen, R. W. (1989). The 'up' and 'down' staircase in second language development. In N. Dorian (Ed.), Investigating obsolescence. Studies in language contraction and death (pp. 385-394). Cambridge: Cambridge University Press. Andersson, L. (1992). Visuella korrelat till svenska vokalers rundning [Visual correlates of rounding in Swedish vowels]. In D. Huber (Ed.), Papers from the 6th Swedish Phonetics Conference (pp. 103-106). Chalmers University of Technology: Department of Information Theory. Aoki, H. (1970). Nez Perce grammar. University of California Publications in Linguistics, no. 62. Berkeley: University of California Press. Applebee, A. N. (1978). The child's concept of a story. Chicago: Chicago University Press. Argyle, M. (1967). The psychology of interpersonal behaviour. Penguin. Argyle, M. (1975). Bodily communication. London: Methuen. Armon-Lotem, S. (1996). The minimalist child: Parameters and functional heads in the acquisition of Hebrew. Doctoral dissertation, Tel Aviv University. Aske, J. (1989). Path predicates in English and Spanish: A closer look. Proceedings of the 15"' Annual Meeting of the Berkeley Linguistics Society (pp. 1-14). Berkeley, CA: Berkeley Linguistics Society. Astington, J. W. & Gopnik, A. (1991). Theoretical explanations of children's understanding of the mind. In G. E. Butterworth, P. L. Harris, A. M. Leslie & H. M. Wellman (Eds.), Perspectives on the child's theory of mind (pp. 7-32). Oxford: Oxford University Press, BPS Books. Astington, J. W. & Jenkins, J. M. (1999). A longitudinal study of the relation between language and theory-of-mind development. Developmental Psychology 35(5), 1311-1320. Au, K. H. (1998). Social constructivism and the school literacy learning of students of diverse backgrounds. Journal of Literacy Research 30, 297-319.

526

References

Au, T. (1993). Literacy instruction in multicultural settings. Fort Worth, TX: Harcourt Brace Jovanovich. Austin, J. (1997). The status of pro-drop in the initial state: Results from new analyses of Spanish. In A. T. Perez-Leroux & W. R. Glass (Eds.), Contemporary perspectives on the acquisition of Spanish, vol. 1: Developing grammars. Somerville: Cascadilla Press. Backus, A. (1996). Two in one: Bilingual speech among Turkish immigrants in the Netherlands. Tilburg: Tilburg University Press. Bamberg, M. (1985). Form and function in the construction of narratives: Developmental perspectives. Unpublished doctoral dissertation, University of California, Berkeley. Bamberg, M. (1987). The acquisition of narratives: Learning to use language. Berlin: Mouton de Gruyter. Bamberg, M. (1990). The German perfekt: Form and function of tense alternations. Studies in Language 14, 253-290. Bamberg, M. (1991). Narrative activity as perspective taking: The role of emotionals, negations, and voice in the construction of the story realm. Journal of Cognitive Psychotherapy 5, 275-290. Bamberg, M. (1992). Binding and unfolding: Establishing viewpoint in oral and written discourse. In M. Kohrt & A. Wrobel (Eds.), Schreibprozesse - Schreibprodukte. Festschrift fur Gisbert Keseling (pp. 1-24). Hildesheim: Georg Olms Verlag. Bamberg, M. (1994a). Actions, events, scenes, plots and the drama: Language and the constitution of part-whole relationships. Language Sciences 15, 6, 39-79. Bamberg, M. (1994b). [Review]. Why narrativizing a wordless picture book is not artificial. A response to Jim Gee's review of "Relating events in narrative ", by one of the collaborating authors. TESL-EJ 1(2), Review 1-2. Bamberg, M. (Ed.) (1997a). Narrative development: Six approaches. Mahwah, New Jersey: Lawrence Erlbaum Associates. Bamberg, M. (1997b). Positioning between structure and performance. Journal of Narrative and Life History 7, 335—342. Bamberg, M., Budwig, N. & Kaplan, B. (1991). A developmental approach to language acquisition. First Language 11, 121-141. Bamberg, M. & Damrad-Frye, R. (1991). On the ability to provide evaluative comments: Further explorations of children's narrative competencies. Journal of Child Language 18(3), 689-709. Bamberg, M. & Marchman, V. (1990). What holds a narrative together? The linguistic encoding of episode boundaries. Papers in Pragmatics 4, 58-121. Bamberg, M. & Marchman, V. (1991). Binding and unfolding: Towards the linguistic construction of narrative discourse. Discourse Processes 14, 277-305.

References

527

Bamberg, M. & Marchman, V. (1994). Foreshadowing and wrapping up in narrative. In R. A. Berman & D. I. Slobin (Eds.), Different says of relating events in narrative: A crosslinguistic developmental study (pp. 555-590). Hillsdale, NJ: Lawrence Erlbaum Associates. Bamberg, M. & Reilly, J. (1996). Emotion, narrative, and affect: How children discover the meaning of what to say and how to say it. In D. I. Slobin, J. Gerhardt, A. Kyratzis & J. Guo (Eds.), Social interaction, social context, and language: Essays in honor of Susan Ervin-Tripp (pp. 329-341). Mahwah, New Jersey: Lawrence Erlbaum Associates, Inc. Barbieri, M. S. (1995). [Review of R. A. Berman & D. I. Slobin (Eds.) (1994), "Relating events in narrative: A crosslinguistic developmental study", Hillsdale, NJ: Erlbaum.] First Language 15, 357-361. Barsalou, L. W. (2003). Abstraction as dynamic interpretation in perceptual symbol systems. In L. Gershkoff-Stowe & E. Rakison (Eds.), Building object categories. Mahwah, NJ: Erlbaum. Bauer, P. J. & Mandler, J. M. (1989). One thing following another: Effects of temporal structure on one- and two-year-olds' recall of events. Developmental Psychology 25,197-206. Bauman, R. (1977). Verbal art as performance. Newbury House. Bavin, E. (1990). Locative terms and Warlpiri acquisition. Journal of Child Language 17, 43-66. Bavin, E. (1991). The acquisition of Warlpiri kin terms. Pragmatics 1, 319-344. Bavin, E. (1992). The acquisition of Warlpiri. In D. I. Slobin (Ed.), The cross-linguistic study ofkanguage acquisition, vol. 3 (pp. 309-372). Hillsdale, NJ: Lawrence Erlbaum Associates. Bavin, E. (1993). Language and culture: Socialisation in a Warlpiri Community. In M. Walsh & C. Yallop (Eds.), Language and culture in Aboriginal Australia (pp. 85-96). Canberra: Aboriginal Studies Press. Bavin, E. (1995). Inflections and lexical organisation: some evidence from Warlpiri. In H. Pishwa & K. Marold (Eds.), The development of morphological systematicity (pp. 39-53). Tubingen: Gunter Narr Verlag. Bavin, E. (1998). Factors of typology in language acquisition: Some examples from Warlpiri. In A. Siewierska & J. Song (Eds.), Case, typology and grammar (Typological studies in language) (pp. 37-55). Amsterdam/ Philadelphia: John Benjamins. Bavin, E. (2000). Ellipsis in Warlpiri children's narratives: An analysis of Warlpiri frog stories. Linguistics 38, 569-589. Bavin, E. (this volume). Focusing on "where": An analysis of Warlpiri frog stories. In S. Stromqvist & L. Verhoeven (Eds.) (2004), Relating events in narrative: Typological and contextual perspectives. Mahwah, NJ: Lawrence Erlbaum Associates.

528

References

Bazzanella, C. & Calleri, D. (1991). Tense coherence and grounding in children's narratives. Text 11, 175-187. Beattie, G. (1980). Encoding units in spontaneous speech. In H. W. Dechert & M. Raupach (Eds.), Temporal variables in speech (pp. 131-143). The Hague: Mouton. Behrens, H. (1997). How difficult are complex verbs? Linguistics 36, 679-712. Behrns, I. (2003). The process of writing in persons with aphasia. Licentiate thesis. Goteborg University: Dept of Linguistics. Bennett-Kastor, T. (2002). The frog story narratives of Irish-English bilinguals. Bilingualism: Language and Cognition 5, 2, 131-146. Berglund, E. (2000). Language assessment of Swedish 3-year-olds using a protocol of natural speech: PLUS-3 productive language use at 3. Log Phon Vocol 25, 51-62. Berlin, B. (1968). Tzeltal numeral classifiers. A study in ethnographic semantics. The Hague, Paris: Mouton. Berlin, B. (1972). Speculations on the growth of ethnobotanical nomenclature. Language in Society; 1,51-86. Berman, R. A. (1978). Modern Hebrew structure. Tel Aviv: Universities Publishing. Berman, R. A. (1980). The case of an (S)VO language: Subjectless constructions in Modern Hebrew. Language 56, 759-776. Berman, R. A. (1984). From nonanalysis to productivity: Interim schemata in child language. Working paper no. 31. Tel Aviv University, Unit for Human Development and Education. Berman, R. A. (1985). Interference in FLT. In T. Husen & T. N. Hustlethwaite (Eds.), International encyclopaedia of education (pp. 1943-1945). Oxford: Pergamon Press. Berman, R. A. (1986). A step-by-step model of language learning. In I. Levin (Ed.), Stage and structure: Reopening the debate (pp. 191-219). Norwood, NJ: Ablex. Berman, R. A. (1987). A developmental route: Learning about the form and use of complex nominals. Linguistics 27, 1057-1085. Berman, R. A. (1988a). On the ability to relate events in narrative. Discourse Processes 11,469-497. Berman, R. A. (1988b). Word-class distinctions in developing grammars. In Y. Levy, I. M. Schlesinger & M. D. S. Braine (Eds.), Categories and processes in language acquisition (pp. 45-72). Hillsdale, NJ: Lawrence Erlbaum Associates. Berman, R. A. (1990). Acquiring an (S)VO language: Subjectless sentences in children's Hebrew. Linguistics 28, 1135-1166. Berman, R. A. (1993a). Crosslinguistic perspectives on native language acquisition. In K. Hyltenstam & A. Viberg (Eds.), Progression and regression in language (pp. 245-266). Cambridge University Press. Berman, R. A. (1993b). Developmental perspectives on transitivity: A confluence of cues. In Y. Levy (Ed.), Other children, other languages: Issues in the theory of language acquisition (pp. 189-241). Hillsdale, NJ: Lawrence Erlbaum Associates.

References

529

Berman, R. A. (1993c). The development of language use: Expressing perspectives on a scene. In E. Dromi (Ed.), Language and cognition: A developmental perspective (pp. 172-201). Norwood NJ: Ablex. Berman, R. A. (1994). The developmental paradox: How much and how little five-year olds know about language structure and language use. In L. Verhoeven & A. Teberosky (Eds.), Proceedings of the Workshop on Understanding Early Literacy in a Developmental and Cross-linguistic Approach (pp. 17—40). Wassenaar, The Netherlands. Berman, R. A. (1995). Narrative competence and storytelling performance: How children tell stories in different contexts. Journal of Narrative and Life History 5, 285-313. Berman, R. A. (1996). Form and function in developing narrative skills. In D. I. Slobin, J. Gerhardt, A. Kyratzis & J. Guo (Eds.), Social interaction, social context, and language (pp. 343-368). Hillsdale, NJ: Lawrence Erlbaum Associates. Berman, R. A. (1997a). Preschool knowledge of language: What five-year-olds know about language structure and language use. In C. Pontecorvo (Ed.), Writing development (pp. 61-76). Amsterdam: John Benjamins. Berman, R. A. (1997b). Narrative theory and narrative development: The Labovian impact. Journal of Narrative and Life History 7, 235-244. Berman, R. A. (1998a). Typological perspectives on connectivity. In N. Dittmar & Z. Penner (Eds.), Issues in the theory of language acquisition (pp. 203-224). Bern: Lang. Berman, R. A. (1998b). Bilingual proficiency/proficient bilingualism: Insights from narrative texts. In G. Extra & L. Verhoeven (Eds.), Bilingualism and migration (pp. 187-208). Berlin: Mouton de Gruyter. Berman, R. A. (1999). Genre and modality distinctions in developing discourse abilities. In R. Aisenman (Ed.) Developing literacy across genres, modalities, and languages, vol. 1 (pp. 93-113). Working papers, International Literacy Project, Tel Aviv University. Berman, R. A. (2000). Thematic perspectives on how children talk about interpersonal conflict. In M. Aparici, N. Argerich, E. Rosado & L. Tolchinsky (Eds.), Developing literacy in different contexts and different languages (pp. 65-76). Working papers, International Literacy Project, vol. 3. Barcelona: University of Barcelona. Berman, R. A. (200la). Setting the narrative scene: How children begin to tell a story. In A. Aksu-Ko?, C. Johnson & K. Nelson (Eds.), Children s language, vol. 10 (pp. 1-31). Mahwah, NJ: Erlbaum. Berman, R. A. (200 Ib). New perspectives in lexical categorization. Paper presented at a conference on Lexical and Morphological Processing in Spoken and Written Language. Bar-Han University, Ramat Gan. Berman, R. A. (200Ic). Narrative development in multilingual contexts: a cross-linguistic perspective. In L. Verhoeven & S. Stromqvist (Eds.), Narrative development in a multilingual context (pp. 419^28). Amsterdam/Philadelphia: John Benjamins.

530

References

Berman, R. A. (in press). Development of discourse stance in different languages and communicative contexts. Introduction to Journal of Pragmatics (special issue, edited by R. A. Berman). Berman, R. A. (this volume). The role of context in developing narrative abilities. In S. Stromqvist & L. Verhoeven (Eds.) (2004), Relating events in narrative: Typological and contextual perspectives. Mahwah, NJ: Lawrence Erlbaum Associates. Berman, R. A. & Katzenberger, I. (1998). Cognitive and linguistic factors in development of picture-series narration. In A. G. Ramat & M. Chini (Eds.), Organization of learners 'texts. Special issue ofStudia italiani i linguistica teorica e applicato 27, 21-47. Berman, R. A. & Katzenberger, I. (in press). Form and function in introducing narrative and expository texts: A developmental perspective. Discourse Processes. Berman, R. A. & Neeman, Y. (1994). Development of linguistic forms: Hebrew. In R. A. Berman & D. I. Slobin (Eds.), Relating events in narrative: A cross-linguistic developmental study (pp. 285-328). Hillsdale, NJ: Lawrence Erlbaum Associates. Berman, R. A. & Nir, B. (in press). Linguistic indicators of inter-genre differentiation in later language development. Journal of Child Language. Berman, R. A., Ragnarsdottir, H. & Stromqvist, S. (2002). Discourse stance. Written Language and Literacy 5, 255-290. Berman, R. A. & Slobin, D. I. (1987). Five ways of learning how to talk about events: A crosslinguistic study of children's narratives. Cognitive science report no. 46, Institute of Cognitive Studies, University of California, Berkeley. Berman, R. A. & Slobin, D. I. (1994). Narrative structure. In R. A. Berman & D. I. Slobin (Eds.), Relating events in narrative: A crosslinguistic developmental study (pp. 39-84). Hillsdale, NJ: Lawrence Erlbaum Associates. Berman, R. A. & Slobin, D. I. (Eds.) (1994). Relating events in narrative: A crosslinguistic developmental study. Hillsdale, NJ: Lawrence Erlbaum Associates. Berman, R. A. & Verhoeven, L. (2002a). Cross-linguistic perspectives on the development of text-production abilities in speech and writing. Written Language and Literacy 5 (1/2), 1—44 (special issue on developing text production abilities in speech and writing). Berman, R. A. & Verhoeven, L. (Eds.) (2002b). Written Language and Literacy 5, parts 1-2 (special issue on developing text production abilities in speech and writing). Bialystok, E. (1992). Language processing in bilingual children. Cambridge, NY: Cambridge University Press. Bialystok, E. (1997). Effects of bilingualism and biliteracy on children's emerging concepts of print. Developmental Psychology 33, 429^440. Bialystok, E. & Ryan, E. B. (1985). Metacognitive framework for the development of first and second language skills. In D. L. Forrest-Pressley, G. E. MacKinnon & T. G. Waller (Eds.), Metacognition, cognition and human performance. New York: Academic Press.

References

531

Bierwisch, M. (1967). Some semantic universals of German adjectivals. Foundations of Language, 3, 1-36. Billman, D., Swilley, A. & Krych, M. (2000). Path and manner priming: Verb production and event recognition. Proceedings of the 22nd Annual Meeting of the Cognitive Science Society (pp. 615-620). Philadelphia, Pennsylvania. Bisang, W. (1995). Verb serialization and coverbs - difference and similarities. In M. Haspelmath & E. Konig (Eds.), Adverbial participle, gerunds. Berlin: Mouton de Gruyter. Blankenstijn, C. J. K. & Scheper, A. R. (2003). The development of morphosyntactic and semantic-pragmatic abilities in Dutch-speaking children with psychiatric disorder. Dissertation, University of Amsterdam, Amsterdam: University Press. Bocaz, A. (1989a). Desarrollo de la referencia temporal adverbial [Development of adverbial temporal reference]. Lenguas modernas 16, 23-39. Santiago, Chile. Bocaz, A. (1989b). Estudio evolutive de la marcacion aspectual de los sucesos en la produccion de discurso narrative [Developmental study of aspectual marking of events in the production of narrative discourse]. In Actas del III Congreso Internacional sobre el Espanol de America (pp. 1311-1320). Salamanca, Spain: Graficas Varona. Bocaz, A. (1989c). Los marcadores de expresion de la simultaneidad en el desarrollo de estructuras sintacticas y textuales complejas [Markers of the expression of simultaneity in the development of syntactic structures semantics and textual complexes]. Revista de linguistica tedricay aplicada 27, 5-22. Concepcion, Chile. Bocaz, A. (199la). Desarrollo de construcciones sintacticas temporales y sus funciones en el discurso narrative [Development of temporal syntactic structures and their functions in narrative discourse]. In Actas del simposio sobre cognicion, lenguaje y cultura: didlogo transdisciplinario en ciencia cognitiva (pp. 169-176). Santiago, Chile: Editorial Universitaria. Bocaz, A. (1991 b). El caso de 'estar subido': Una antesala en el desarrollo deperfecto? [The case of estar subido ('to be ascended'): A language waiting room in the development of the perfect?]. Paper presented at the Seminario Nacional de Investigacion y Ensenanza de la Linguistica, Universidad de Concepcion, Chile. Bocaz, A. (1991 c). Esquematizacion espacial y temporal de escenas narrativas y su proyeccion en el espanol [Spatial and temporal schematization of narrative scenes and their linguistic projection in Spanish]. Lenguas modernas 18, 47-62. Santiago, Chile. Bocaz, A. (199Id). Marcacion de la organization secuencial de los sucesos en la produccion de discurso narrative [Marking of sequential organization of events in the production of narrative discourse]. In Actas del VIII Seminario Nacional de Investigacion y Ensenanza de la Linguistica (pp. 249-255). Santiago, Chile: Universidad de Santiago de Chile.

532

References

Bocaz, A. (1992a). Funciones de las clausulas relativas en la production de relates en dos variedades del espanol de America [Functions of relative clauses in the production of stories in two varieties of present American Spanish]. Paper presented at the conference "Cuarto congreso internacional sobre el espanol en America", Pontificia Universidad Catolica de Chile. Bocaz, A. (1992b). Procesos inferenciales abductivos en la interpretation de escenas narrativas complejas [Abductive inferential processes in the interpretation of complex narrative scenes]. Lenguas modernas 19, 99-106. Santiago, Chile. Boeschoten, H. E. (1990). Acquisition of Turkish by children in the Netherlands. Wiesbaden: Harrassowitz. Boeschoten, H. E. & Verhoeven, L. (1986). Turkish language development of Turkish children in the Netherlands. In A. Aksu & E. Erguvanh (Eds.), Proceedings of the 2"J Conference on Turkish linguistics (pp. 269-280). Istanbul: Bogazici Press. Bohnemeyer, J. (in press-a). The unique vector constraint: the impact of direction changes on the linguistic segmentation of motion events. In E. van der Zee & J. Slack (Eds.), Representing direction in language and space. Oxford: Oxford University Press. Bohnemeyer, J. (in press-b). The pitfalls of getting from here to there: Bootstrapping the syntax and semantics of motion event coding inYukatek Mayan. In M. Bowerman & P. Brown (Eds.), Crosslinguisticperspectives on argument structure: Implications for learnability. Mahwah, NJ: Erlbaum. Bohnemeyer, J. (manuscript). Positions on positional. Language and Cognition Group, Max Planck Institute for Psycholinguistics. Nijmegen. Bohnemeyer, J. & Brown, P. (to appear) Standing divided: Dispositional verbs and locative predications in two Mayan languages. In F. Ameka & S. Levinson (Eds.), Linguistics (special issue). Borer, H. (1984). Parametric syntax: Case studies in Semitic and Romance languages. Dordrecht: Foris. Borer, H. & Grodzinsky, Y. (1986). Syntactic cliticization and lexical cliticization: The case of Hebrew dative clitics. In H. Borer (Ed.), Syntax and semantics, vol. 19: The syntax of pronominal clitics. New York: Academic Press. Bos, P. (1997). Development ofbilingualism: A study of Moroccan children in the Netherlands. Doctoral dissertation, Tilburg University. Sotting, N. (2002). Narrative style across children with different communication impairments. Child Language Teaching and Therapy 18(1), 1-22. Bowerman, M. (1985). What shapes children's grammar? In D. I. Slobin (Ed.), The crosslinguistic study of language acquisition, vol. 2: The data (pp. 1257-1320). Hillsdale, NJ: Lawrence Erlbaum Associates. Bowerman, M., de Leon, L. & Choi, S. (1995). Verbs, particles, and spatial semantics: Learning to talk about spatial actions in typologically different languages. In E. V. Clark (Ed.), Proceedings of the 27th Annual Child Language Research Forum (pp. 101-110). Stanford, CA: CSLI.

References

533

Boyd, S. (1985). Language survival. A study of language contact, language shift and language choice in Sweden. Doctoral dissertation, University of Gothenburg. Boyd, S. (1993). Immigrant minority languages and education in Sweden. In G. Extra & L. Verhoeven (Eds), Immigrant languages in Europe (pp. 273-284). Clevedon: Multilingual Matters. Boyd, S. & Naucler, K. (2001). Sociocultural aspects of bilingual narrative development. In L. Verhoeven & S. Stromqvist (Eds.), Narrative development in a multilingual context (pp. 129-152). Amsterdam/Philadelphia: John Benjamins. Boyes Braem, P. (1981). Features of the handshape in American Sign Language. Unpublished doctoral dissertation, University of California, Berkeley. Brennan, M. (1990). Word formation in BSL. Stockholm: University of Stockholm Press. Brewer, W. F. (1985). The story schema: universal and culture-specific properties. In D. R. Olson, N. Torrance, A. Hildyard & R. Scollon (Eds.), Literacy, language, and learning: The nature and consequences of reading and writing (pp. 167-194). Cambridge: Cambridge University Press. Bronckart, J.-P. & Schneuwly, B. (1991). Children's production of textual organizers. In G. Pieraut-Le Bonniec & M. Dolitsky (Eds.), Language bases ... discourse bases (pp. 143-156). Amsterdam: Benjamins. Brown, P. (1994). The INs and ONs of Tzeltal locative expressions: The semantics of static descriptions of location. Linguistics 32 (4/5), 743-790. Brown, P. (2000). 'He descended legs-upwards': Position and motion in Tzeltal frog stories. In E. V. Clark (Ed.), Proceedings of the 30th Annual Child Language Research Forum (pp. 67-75). Stanford: CSLI. Brown, P. (in press). The grammar of Tzeltal space. In S. C. Levinson & D. P. Wilkins (Eds.), The grammar of space. Cambridge University Press. Brown, P. (this volume). Position and motion in Tzeltal frog stories: The acquisition of narrative style. In S. Stromqvist & L. Verhoeven (Eds.) (2004), Relating events in narrative: Typological and contextual perspectives. Mahwah, NJ: Lawrence Erlbaum Associates. Bruck, M. & Genesee, F. (1995). Phonological awareness in young second language learners. Child Language 22, 307-324. Bruner, J. (1986). Two modes of thought. Actual minds possible worlds (pp. 11-43). Cambridge, Mass: Harvard University Press. Burling, R. (1970). Man s many voices: Language in its cultural context. New York: Holt, Rinehart and Winston. Butterworth, B. & Beattie, G. (1978). Gesture and silence as indicators of planning in speech. In R. N. Campbell. & P. T. Smith. (Eds.), Recent advances in the psychology of language: Formal and experimental approaches (pp. 347-360). New York: Plenum Press.

534

References

Cenoz, J. (2001). The effect of linguistic distance, L2 status and age on crosslinguistic influence in the third language. In J. Cenoz; B. Hufeisen & U. Jessner (Eds.), Cross-linguistic influence in third language acquisition: Psycholinguistic perspectives. Clevedon: Multilingual Matters. Cenoz, J. (2002a). Age differences in foreign language learning. ITL Review of Applied Linguistics, 135/36, 125-142. Cenoz, J. (2002b). Three languages in contact: Language attitudes in the Basque country. In D. Lasagabaster & J. Sierra (Eds.), Language awareness in the foreign language classroom. Bilbao: University of the Basque Country. Cenoz, J. (in press). The influence of age on the acquisition of English: General proficiency, attitudes and code mixing. In M. P. Garcia Mayo & M. L. Garcia Lecumberri (Eds.), Age and the acquisition of English as a foreign language: Theoretical issues and field work. Clevedon: Multilingual Matters. Chafe, W. (1976). Givenness, contrastiveness, definiteness, subjects, topics and point of view. In C. N. Li (Ed.), Subject and topic (pp. 25-55). NY: Academic Press. Chafe, W. (1977). The recall and verbalization of past experience. In R. W. Cole (Ed.), Current issues in linguistic theory* (pp. 215-246). Bloomington: Indiana University Press. Chafe, W. (1979). The flow of thought and the flow of language. In T. Givon (Ed.), Syntax and semantics, vol. 12: Discourse and syntax (pp. 159-182). New York: Academic Press. Chafe, W. (1980). The pear stories: Cognitive, cultural, and linguistic aspects of narrative production. Norwood, NJ: Ablex. Chafe, W. (1982). Integration and involvement in speaking, writing and oral literature. In D. Tannen (Ed.), Coherence in spoken and written discourse. Norwood, NJ: Ablex. Chafe, W. (1998). [Review of R. A. Berman & D. I. Slobin (Eds.) (1994). "Relating events in narrative: A crosslinguistic developmental study". Hillsdale, NJ: Erlbaum.] Anthropological Linguistics 40, 155-158. Chatman, S. (1978). Story and discourse: Narrative structure in fiction and film. Ithaca NY: Cornell University Press. Choi, S. & Bowerman, M. (1991). Learning to express motion events in English and Korean: The influence of language-specific lexicalization patterns. Cognition 41, 83-121. Chuwicha, Y. (1993). Clausehood in serial verb construction in Thai. Unpublished PhD. dissertation, Chulalongkorn University. Clancy, P. M. (1986). The acquisition of communicative style in Japanese. In B. B. Schieffelin & E. Ochs (Eds.), Language socialization across cultures (pp. 213250). Cambridge: Cambridge University Press. Clark, H. H. (1973). Space, time, semantics, and the child. In T. Moore (Ed.), Cognitive development and the acquisition of language (pp. 27-63). New York: Academic Press. Clark, H. H. (1996). Using language. New York: Cambridge University Press.

References

535

Clark, H. H. (1998). Communal lexicons. In K. Malmkjaer & J. Williams (Eds.), Context in language learning and language understanding (pp. 63-87). Cambridge: Cambridge University Press. Clark, H. H. & Chase, W. G. (1972). On the process of comparing sentences against pictures. Cognitive Psychology 3, 472-517. Clark, H. H. & Chase, W. G. (1974). Perceptual coding strategies in the formation and verification of descriptions. Memory and Cognition 2, 101-111. Clark, H. H. & Clark, E. V. (1977). Psychology and language. New York: Harcourt Brace Jovanovich. Clark, H. H. & Gerrig, R. (1990). Quotations as demonstrations. Language 66, 764-805. Clark, H. H. & Haviland, S. E. (1977). Comprehension and the given-new contrast. In R. O. Freedle (Ed.), Discourse processes: Advances in research and theory, vol. I: Discourse production and comprehension (pp. 1-40). Hillsdale, NJ: Lawrence Erlbaum Associates. Clark, H. H. & van der Wege, M. (2001). Imagination in discourse. In D. Schiffrin & D. Tannen (Eds.), Handbook of discourse. Oxford: Basil Blackwell. Clark, H. H. & Wasow, T. (1998). Repeating words in spontaneous speech. Cognitive Psychology 37, 201-242. Clark, M. (1992). Serialization in Mainland Southeast Asia. Pan-Asiatic linguistics. Proceedings of the 3rd International Symposium on Language and Linguistics, vol. I (pp. 145-160). Chulalongkorn University, Bangkok, January 8-10, 1992. Cole, D. E. (2001). Narrative development in aggressive boys. Behavioral Disorders 26(4). Cook, H. M. (1997). The role of Japanese masu form in caregiver-child conversation. Journal of Pragmatics 28, 695-718. Croft, W., Barddal, J., Hollman, W., Sotirova, V. & Taoka, C. (2002). Revising Talmy's typological classification of complex events. Conference of the Linguistic Society of America, San Francisco. Cummins, J. (1991). Conversational and academic language proficiency in bilingual contexts. AILA Review 8, 75-89. Dahl, O. (1985). Tense and aspect systems. Oxford: Basil Blackwell. Dasinger, L & Toupin, C. (1994). The development of relative clause functions in narrative. In R. A. Berman & D. I. Slobin (Eds.), Relating events in narrative: A crosslinguistic developmental study (pp. 457-514). Hillsdale, NJ: Lawrence Erlbaum Associates. de Houwer, A. (1995). Bilingual language acquisition. In P. Fletcher & B. MacWhinney (Eds.), Handbook of child language (pp. 219-250). Cambridge, MA: Basil Blackwell. de Leon, L. (manuscript). Observations on the use of directionals by Tzotzil children in the frog stories. Unpublished draft manuscript (1992). Cognitive Anthropology Research Group, Max-Planck-Institute for Psycholinguistics, Nijmegen.

536

References

de Week, G. (1991). La cohesion dans les narrations d'enfants [Cohesion in children's narrations]. Neuchatel: Delachaux andNiestle. de Week, G. & Schneuwly, B. (1994). Anaphoric procedures in four types of texts written by children. Discourse Processes 17, 465—477. DeLancey, S. (1989). Klamath stem structure in genetic and areal perspective. Papers from the 1988 Hokan-Penutian languages workshop (pp. 31-39). Eugene, OR: University of Oregon. DeLancey, S. (1996). Argument structure of Klamath bipartite stems. SSILA Conference, San Diego, CA. Denny, J. P. (1982). Semantics of the Inuktitut (Eskimo) spatial deictics. International Journal of American Linguistics 48, 4, 359-384. Dickson, S. V., Simmons, D. C. & Kameenui, E. J. (1995). Text organization: Curricular and instructional implications for diverse learners. Technical report, no. 18. National Center to Improve the Tools of Educators (NCITE). Dijkhuis, K. (manuscript). Coherentie in vertellingen en taalvaardigheid: Een onderzoek metpsychisch gestoorde kinderen [Coherence in narratives and linguistic capacity: An investigation of mentally disturbed children]. Unpublished master's thesis, University of Amsterdam: Department of General Linguistics. Doi, T. (1971). The anatomy of dependence. New York: Harper & Row. Donaldson, M. (1978). Children's minds. New York: W. W. Norton. Dromi, E. & Berman, R. A. (1986). Language particular and language general in developing syntax. Journal of Child Language 14, 371-387. Durgunoglu, A. & Verhoeven, L. (1998). Literacy development in a multilingual context. Mahwah, NJ: Lawrence Erlbaum. Durie, M. (1997). Grammatical structures in verb serialization. In A. Alsina, J. Bresnan & P. Sells (Eds.), Complex predicates, Stanford: Stanford University Press. Durova, N. V. & Yurieva, N. M. (manuscript). Sostavlenie rasskaza po serii illjustracij det'mi doskol'nikami [Production of a story according to a series of illustrations by preschoolers]. Unpublished paper (1993). Institute of Linguistics, Russian Academy of Sciences, Moscow. Eaton, J. H., Collis, G. M. & Lewis, V. A. (1999). Evaluative explanations in children's narratives of a video sequence without dialogue. Journal of Child Language 26, 699-720. Ehrich, V. (1982). Discourse organization and sentence form in child language: How children describe rooms. Papers and Reports on Child Language Development 21, 55-62. Ehrlich, M.-F., Remond, M. & Tardieu, H. (1999). Processing of anaphoric devices in young skilled and less skilled comprehenders: Differences in metacognitive monitoring. Reading and writing: An interdisciplinary journal 11, 29-63. Eisenberg, A. R. (1985). Learning to describe past-experiences in conversation. Discourse Processes 8, 177-204.

References

537

Eklundh, K. S. & Kollberg, P. (1995). Translating key strokes records into a general notation for the writing process. In G. Rijlaarsdam, H. van der Bergh & M. Couzijn (Eds.), Theories, models and methodology: Current trends in writing on writing. Amsterdam: Amsterdam University Press. Ekmek9i, O. (1986). Creativity in the language acquisition process. In H. E. Boeschoten & L. Verhoeven (Eds.), Studies on modern Turkish (pp. 177-188). Tilburg: University Press. Emmorey, K. & Reilly J. S. (1995). Theoretical issues relating language, gesture and space: An overview. In K. Emmorey & J. Reilly (Eds.), Language, gesture, and space (pp. 1-16). Hillsdale, NJ: Lawrence Erlbaum Associates. Emslie, H. C. & Stevenson, R. J. (1981). Pre-school children's use of the articles in definite and indefinite referring expressions. Journal of Child Language 8, 313328. Engberg-Pedersen, E. (1993). Space in Danish Sign Language: The semantic and morphosyntax of the use of space in a visual language. Hamburg: SIGNUM-Press. Engberg-Pedersen, E. (1999) Path- and ground-denoting expressions in descriptions of motion events in Danish Sign Language: A typological perspective. Paper presented at the 6th International Cognitive Linguistics Conference, Stockholm, Sweden. Engberg-Pedersen, E. (2000). The interrelation of grammar, text type, and age: Expression of time relations and the pragmatic status of event participants in Danish narratives. Acta Linguistica Hafniensia 32, 121-141. Engberg-Pedersen, E. & Trondhjem, F. B. (this volume). Focus on action in motion descriptions: The case of West-Greenlandic. In S. Stromqvist & L. Verhoeven (Eds.) (2004), Relating events in narrative: Typological and contextual perspectives. Mahwah, NJ: Lawrence Erlbaum Associates. Enkvist, N., Spencer, J., & Gregory, M. (1964). Linguistics and style. London. Eriksson, M. (2001). Narratives validate communicate development inventories, Applied Psycholinguistics 22, 45-60. Ervin-Tripp, S. & Kuntay, A. (1996). The occasioning and structure of conversational stories. In T. Givon (Ed.), Conversation: Cognitive, communicative and social perspectives (pp. 133-166). Amsterdam: John Benjamins. Esperet, E. (1991). The development and role of narrative schema storytelling. In G. Pieraut-Le Bonniec & M. Dolitsky (Eds.), Language bases ... discourse bases (pp. 209-227). Amsterdam: Benjamins. Evans, G. (1982). The varieties of reference. Oxford: Oxford University Press. Extra, G. & Verhoeven, L. (1993). Immigrant languages in Europe. Clevedon, UK: Multilingual Matters. Extra, G. & Verhoeven, L. (1994). The cross-linguistic study of bilingual development. Amsterdam: North Holland. Extra, G. & Verhoeven, L. (1998). Bilingualism and migration. Berlin: Mouton de Gruyter.

538

References

Farrar, M, Friend, M. & Forbes, J. (1993). Event knowledge and early language acquisition. Journal of Child Language 20, 591-606. Fayol, M. (1985). Le recit et sa construction: Une approche depsvchologie cognitive [The story and its construction: A psychological approach]. Neuchatel: Delachaux and Niestle. Fayol, M. (1991). Stories: A psycholinguistic and ontogenetic approach to the acquisition of narrative abilities. In G. Pieraut-Le Bonniec & M. Dolitsky (Eds.), Language bases ... discourse bases (pp. 229-243). Amsterdam: Benjamins. Fernandez Lagunilla, M. & Anula Rebollo, A. (1995). Sintaxis y cognicion. Introduccion al conocimiento, elprocesamiento y los deficits sintdcticos [Syntax and cognition. Introduction to syntactic consciousness, processing, and deficits]. Madrid: Editorial Sintesis. Fernandez-Ramirez, S. (1986). Gramatica espanola [Spanish grammar], vol. 4. Madrid: Arco/Libro, S. A. Filipovic, L. (1999). Language-specific expression of motion and its use in narrative texts. Unpublished Mphil. dissertation, University of Cambridge, UK. Filipovic, L. (in preparation). Verbs in motion expressions: Structural perspectives. Doctoral dissertation, University of Cambridge, UK. Fillmore, C. (1966). Deictic categories in the semantics of "come". Foundations of Language 2,219-227. Fine, J. (1985). Cohesion as an index of social cognitive factors: Oral language of the reading disabled. Discourse Processes 8, 91-112. Fine, J. (1994). How language works: Cohesion in normal and nonstandard communication. Norwood, NJ: Ablex. Fischer, K. (1980). A theory of cognitive development: The control and construction of hierarchies of skills. Psychological Review 87, 477-531. Fishman, J. (1980). Ethnocultural dimensions in the acquisition and retention of biliteracy. Journal of Basic Writing 3, 48-61. Fivush, R., Gray, J. T. & Fromhoff, F. A. (1987). Two-year-olds' talk about the past. Cognitive Development 2, 393-^09. Flavell, J. H. (1999). Cognitive development: Children's knowledge about the mind. Annual Review of Psychology 50, 21^5. Flavell, J. H., Flavell, E. R. & Green, F. L. (1987). Young children's knowledge about the apparent-real and pretend-real distinctions. Developmental Psychology- 23, 816-822. Fong, V. & Poulin, C. (1998). Locating linguistic variation in semantic templates. In J.-P. Koenig (Ed.), Discourse and Cognition. Stanford, CA: CSLI. Fortescue, M. (1980). Affix ordering in West-Greenlandic. International Journal of American Linguistics 46, 4, 259-278. Fortescue, M. (1984a). West-Greenlandic. Croom Helm Descriptive Grammars. London: Croom Helm.

References

539

Fortescue, M. (1984b). Learning to speak Greenlandic: A case study of a two-year-old's morphology in a polysynthetic language. First Language 5, 101-114. Fortescue, M. (1988). Eskimo orientation systems. Meddelelser om Gr0nland [Communications about Greenland], Man & Society 11, 3-30. Fortescue, M. & Lennert Olsen, L. (1992). The acquisition of West-Greenlandic. In D. I. Slobin (Ed.), The crosslinguistic study of language acquisition, vol. 3 (pp. 111219). Hillsdale, NJ: Lawrence Erlbaum Associates. Fox Tree, J. E. & Clark, H. (1999). Pronouncing "the" as "thee" to signal problems in speaking. Cognition 62, 151-167. French, L. A. & Nelson, K. (1982). Taking away supportive context: Preschoolers' talk about the then-and-there. Quarterly Newsletter of Comparative Human Cognition 4, 1-6. French, L. A. & Nelson, K. (1985). Children's acquisition of relational terms: Some ifs, ors, and buts. NY: Springer Verlag. Fridjonsson, J. (1989). Samsettar myndir sagna [The compositional structure of verbs]. Reykjavik: Malvisindastofnun Haskola Islands. Fujii, S. Y. (1992). The use and learning of temporal clause-linkage in Japanese and English. In B. Kettemann & W. Wieden (Eds.), L2 acquisition research in Europe: Selected papers from the 1991 European Second Language Association Conference. Tubingen: Giinter Narr. Fujii, S. Y. (1993). The use and learning of clause-linkage: A case study of conditionals in Japanese and English. Unpublished doctoral dissertation, University of California, Berkeley. Galambos, S. J. & Goldin-Meadow, S. (1990), The effects of learning two languages on levels of metalinguistic awareness. Cognition 34, 1-56. Galvan, D. (1988). The acquisition of three morphological subsystems in American Sign Language by deaf children with deaf or hearing parents. Unpublished doctoral dissertation, University of California, Berkeley. Galvan, D. (1989). A sensitive period for the acquisition of complex morphology: Evidence from American Sign Language. Papers and Reports on Child Language Development 28, 107-114. Galvan, D. & Taub, S. (this volume). The encoding of motion information in American Sign Language. In S. Stromqvist & L. Verhoeven (Eds.) (2004), Relating events in narrative: Typological and contextual perspectives. Mahwah, NJ: Lawrence Erlbaum Associates. Gao, H. (2001). The physical foundation of the patterning of physical action verbs: A study of Chinese verbs. Lund, Sweden: Lund University Press. Garcia, G. E. (2000). Bilingual children's reading. In M. L. Kamil, P. B. Mosenthal, P. D. Pearson & R. Barr (Eds.), Handbook of reading research, vol. 3 (pp. 813-834). Mahwah, NJ: Lawrence Erlbaum.

540

References

Garcia Lecumberri, M. L. & Gallardo, F. (in press). English FL pronunciation in school students of different ages. In M. P. Garcia Mayo & M. L. Garcia Lecumberri (Eds.), Age and the acquisition of English as a foreign language: Theoretical issues and field work. Clevedon: Multilingual Matters. Garcia Mayo, M. P. (in press). Age, length of exposure and grammaticality judgments in the acquisition of English as a foreign language. In M. P. Garcia Mayo & M. L. Garcia Lecumberri (Eds.), Age and the acquisition of English as a foreign language: Theoretical issues and field work. Clevedon: Multilingual Matters. Garcia Mayo, M. P. & Garcia Lecumberri, M. L. (Eds.) (2003). Age and the acquisition of English as a foreign language: Theoretical issues and field work. Clevedon: Multilingual Matters. Garcia Soto, J. R. (1993). Evolucion da cohesion nos relatos infantis [The development of cohesion in children's stories]. Unpublished doctoral dissertation. University of Santiago, Spain. Garnham, A., Oakhill, J. & Johnson-Laird, P. (1982). Referential continuity and the coherence of discourse. Cognition 11, 29-46. Gayraud, F. (2000). Developpement de la differentiation oral/ecrit vu a trovers le lexique [The developmental differentiation between speech and writing from a lexical point of view]. PhD. dissertation, Universite Lyons 2, Maison Rhone-Alpes de Sciences de 1'Homme: Dynamique du Langage. Gee, J. P. (1991). Memory and myth: A perspective on narrative. In A. McCabe & C. Peterson (Eds.), Developing narrative structure (pp. 1-25). Hillsdale, NJ: Lawrence Erlbaum Associates. Gee, J. P. (1994). [Review of R. A. Berman & D. I. Slobin (Eds.) (1994), "Relating events in narrative: A crosslinguistic developmental study". Hillsdale. NJ: Erlbaum.] TESL-EJ 1(2), Review 1-3. Geeraerts, D. (1993). Vagueness's puzzles, polysemy's vagaries. Cognitive Linguistics 4. 223-272. Gelman, R. (1978). Cognitive development. Annual Review of Psychology- 29, 297-332. Gennari, S. P., Sloman, S. A., Malt, B. C. & Tecumseh Fitch, W. (2002). Motion events in language and cognition. Cognition 83, 49-79. Gernsbacher, M. A. (1985). Surface information loss in comprehension. Cognitive Psychology- 17,324-363. Gerrig, R. J. (1995). [Review of R. A. Berman & D. I. Slobin (Eds.) (1994), "Relating events in narrative: A crosslinguistic developmental study". Hillsdale, NJ: Erlbaum.] Language 71, 806-808. Glass, A. D. (1980). Cohesion in Ngaanyatjarra discourse. Master's thesis, Australian National University. Glenberg, A. M. (1997). What memory is for. Behavioral & Brain Sciences 20, 1-55. Goldberg, A. E. (1995). Constructions. A construction grammar approach to argument structure. University of Chicago Press.

References

541

Goldman-Eisler, F. (1968). Psycholinguistics: Experiments in spontaneous speech. London: Academic Press. Gombrich, E. (1960). Art and illusion. London: Phaidon Press. Gopnik, M. (1989). The development of text competence. In M. E. Conte, J. S. Petofi & E. Sozer (Eds.), Text and discourse connectedness. Studies in language companion series XXIV (pp. 225-244). Amsterdam: Benjamins. Grice, H. (1975). Logic in conversation. In P. Cole & J. Morgan (Eds.), Speech acts. Syntax and semantics, vol. 3 (pp. 41-58). New York: Academic Press. Grimshaw, J. (1987). Unaccusatives - an overview. North East Linguistic Society 17, 244-259. Gumperz, J. & Levinson, S. (Eds.) (1996). Rethinking linguistic relativity. Studies in the Social and Cultural Foundations of Language, vol. 17. Cambridge: Cambridge University Press. Gutierrez-Clellen, V. F. (2001). Narratives in two languages: Assessing performance of bilingual children. Linguistics and Education 13, 175-197. Gutierrez-Clellen, V. F. & DeCurtis, L. (2001). Narrative assessment: A qualitative approach. Seminars in Speech and Language 22, 79-89. Haberland, H. (1986). Reported speech in Danish. In F. Coulmas, (Ed.), Direct and indirect speech. Berlin: Mouton de Gruyter. Hale, K. (1982). Some essential features of Warlpiri verbal clauses. In S. Swartz (Ed.), Papers in Warlpiri grammar: In memory of Lothar Jagst (pp. 217-315). Work Papers of SIL-AAB:Series A, 6, Darwin: SIL. Halliday, M. A. K. & Hasan, R. (1976). Cohesion in English. London: Longman. Hamano, S. (1998). The sound-symbolic system of Japanese. Stanford, CA: CSLI Publications. Harkins, D. A. (1993). Parental goals and styles of storytelling. In J. Demick, K. Bursik & R. Dibiase (Eds.), Parental development (pp. 61-74). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc. Harris, P. (1989). Young children's theory of mind and emotion. Cognition and Emotion 3,379-400. Haugen, E. (1987). Danish, Norwegian and Swedish. In B. Comrie (Ed.), The world's major languages, (pp. 157-179). London: Croom Helm. Hauser, M. D. (1996). The evolution of communication. Cambridge, Massachusetts: MIT Press. Haviland, J. (1993). Anchoring, iconicity, and orientation in Guugu Yimidhirr pointing gestures. Journal of Linguistic Anthropology 3, 3-45. Haviland, J. (1994). Te xa setel xulem [The buzzards were circling]: Categories of verb roots in Zinacantec Tzotzil. Linguistics 32 (4/5), 691-741. Heath, S. B. (1983). Ways with words: Language, life, and work in communities and classrooms. Cambridge: Cambridge University Press.

542

References

Hebert, J. (1990). Language-specific differences in the encoding of events in tincti narrative discourse. Unpublished B. A. honours thesis, Department of Linguistics. La Trobe University, Melbourne, Victoria, Australia. Hecht, B. F. (1985). Situations and language: Children s use of plural allomorphs in familiar and unfamiliar settings. Unpublished doctoral dissertation, Stanford University, Stanford, CA. Hemphill, L., Picardi, N. & Tager-Flusberg, H. (1991). Narrative as an index of communicative competence in mildly mentally retarded children. Applied Psycholinguistics 12, 263-279. Hemphill, L., Wolf, D., Camp, L. & Griffin, T. (1993). Narrative competence: Local patterns of usage. Paper presented at the Biennial Meeting of the Society for Research in Child Development, New Orleans, LA. Hendriks, H. (1993). Motion and location in narrative discourse: A crosslinguistic and developmental perspective. Doctoral dissertation, Rijksuniversiteit Leiden. Hickmann, M. (1980). Creating referents in discourse: A developmental analysis of linguistic cohesion. In J. Kreiman & E. Ojeda (Eds.), Papers from the parasession on pronouns and anaphora (pp. 192-203). Chicago Linguistic Society. Hickmann, M. (1991). The development of discourse cohesion: Some functional and crosslinguistic issues. In G. Pieraut-Le Bonniec & M. Dolitsky (Eds.), Language bases ... discourse bases (pp. 157-185). Amsterdam: Benjamins. Hickmann, M. (1995). Discourse organization and the development of reference to person, space, and time. In P. Fletcher & B. MacWhinney (Eds.), The handbook of child language (pp. 194-218). Oxford: Blackwell. Hickmann, M. (1996). Information status and grounding in children's discourse: A crosslinguistic perspective. In J. Costermans & M. Fayol (Eds.), Processing interclausal relationships in the production and comprehension of text (pp. 221243). Mahwah, NJ: Lawrence Erlbaum Associates. Hickmann, M. (1998). Form, function, and context in narrative development [Review of R. A. Berman & D. I. Slobin (Eds.) (1994), "Relating events in narrative: A crosslinguistic developmental study", Hillsdale, NJ: Erlbaum.] Journal of Pragmatics 29, 33-56. Hickmann, M. (2003a). Children's discourse: Person, time, and space across languages. Cambridge: Cambridge University Press. Hickmann, M. (2003b). The relativity of motion in first language acquisition. Paper presented at the conference Space in Languages: Linguistic Systems and Cognitive Categories. Paris, February 2003. Hickmann, M. & Hendriks, H. (1999). Cohesion and anaphora in children's narratives: A comparison of English, French, German, and Chinese. Journal of Child Language 26,419^52. Hickmann, M., Hendriks, H. & Roland, F. (1998). Reference spatiale dans les recits d'enfants francais [Reference to space in stories by French children: An interlanguage perspective]. Languefrancaise 118, 104-123.

References

543

Hickmann, M., Hendriks, H., Roland, F. & Liang, J. (1996). The marking of new information in children's narratives: A comparison of English, French, German, and Mandarin Chinese. Journal of Child Language 23, 591-619. Hickmann, M., Kail, M. & Roland, F. (1989). The referential organization of children s narrative discourse as a function of mutual knowledge. Paper presented at the 10th Biennial Meeting, International Society for the Study of Behavioural Development, Jyvaskyla, Finland. Hickmann, M., Kail, M. & Roland, F. (1993). Cohesive anaphoric relations in French children s narratives as a function of mutual knowledge. Paper presented at the 12th Biennial Meeting, International Society for the Study of Behavioural Development, Recife. Hickmann, M., Kail, M. & Roland, F. (1995). Cohesive anaphoric relations in French children's narratives as a function of mutual knowledge. First Language 15, 277-300. Hickmann, M. & Liang, J. (1990). Clause-structure variation in Chinese narrative discourse: A developmental analysis. Linguistics 28, 1167-1200. Hickmann, M. & Schneider, P. (1993). Children's ability to restore the referential cohesion of stories. First Language 13, 169-202. Hickmann, M. & Schneider, P. (2000). Cohesion and coherence anomalies and their effects on children's referent introduction in narrative retell. In M. Perkins & S. Howard (Eds.), New directions in language development and disorders (pp. 251260). New York, NY: Kluwer Academic/ Plenum Publishers. Hildyard, A. & Hidi, S. (1985). Oral-written differences in the production and recall of narratives. In D. R. Olson, N. Torrance & A. Hildyard (Eds.), Literacy, language, and learning: The nature and consequences of reading and writing. Cambridge. Cambridge University Press. Hinton, L., Nichols, J. & Ohala, J. (1994). Sound symbolism. Cambridge: Cambridge University Press. Hohenstein, J. M. (2001). Motion event similarities in English- and Spanish-speaking children. Unpublished doctoral dissertation, Yale University. Hoiting, N. & Slobin, D. I. (1993). Reference to movement in speech and sign. Paper presented at the 6th International Congress for the Study of Child Language, Trieste, Italy. Holmqvist, K., Johansson, V., Stromqvist, S. & Wengelin, A. (2002). Studying reading and writing on-line. In Stromqvist, S. (Ed.), The diversity of languages and language learning. Lund Lectures in Languages and Literature (pp. 103-123). Lund University: Centre for Languages and Literature. Hopper, P. J. (1979a). Aspect and foregrounding in discourse. In T. Givon (Ed.), Syntax and semantics, vol. 12: Discourse and syntax (pp. 213-241). New York, NY: Academic Press. Hopper, P. J. (1979b). Some observations on the typology of focus and aspect in narrative language. Studies in Language 3, 37-64.

544

References

Hopper, P. J. (Ed.) (1982). Tense and aspect: Behveen syntax and semantics. Amsterdam: Benjamins. Home, M., Hansson, P., Bruce, G., Frid, J. & Filipsson, M. (1999). Discourse markers and the segmentation of spontaneous speech: The case of Swedish men 'but/so*. In Working papers 47 (pp. 123-129). Lund University: Department of Linguistics. Hottenroth, P.-M. (1985). Die italienischen Ortsadverbien [The Italian place adverbs]. In C. Schwarze (Ed.), Bausteine fur eine ilalienische Grammatik. vol. 2 (pp. 385462). Tubingen: Gunter Nam Housen, A. (1998). Le processus de grammaticalisation dans le domaine de la temporalite et le developpement de 1'interlangue [The process of grammatical ization in the domain of temporality and the development of an interlanguage]. Travaux de Linguistique 36, 209-222. Housen, A. (2000). Verb semantics and the acquisition of tense-aspect morphology in English. Studio Linguistica 54, 2, 249-259. Housen, A. (in press-a). Facteurs contextuels en apprentissage guide d'une langue seconde [Contextual factors in supervised learning of a second language]. Acquisition et interaction en langue etrangere (AILE). Berlin: Mouton de Gruyter. Housen, A. (in press-b). Second language achievement in the European school system of multilingual education. In D. So & G. Jones (Eds.), Education and society in plurilingual contexts. Longman. Housen, A. (in press-c). A corpus-based study of the L2-acquisition of the English verb system. In S. Granger, J. Hung & S. Petch-Tyson (Eds.), Computer learner corpora, second language acquisition and foreign language learning. London: John Benjamins. Housen, A. (in press-d). The development of tense-aspect in English as a second language and the variable influence of inherent aspect. In Y. Shirai & R. Salaberry (Eds.), Description and explanation in the acquisition offense-aspect. John Benjamins. Housen, A., Janssens, S. & M. Pierrard. (2001). Frans en Engels als vreemde talen in Vlaamse scholen [French and English as foreign languages in Flemish schools]. Brussel: VUB Press. Howell, C. (1993). Aspects of second language narratives. Unpublished master's thesis. Department of Linguistics, La Trobe University, Melbourne, Victoria, Australia. Hualde, J. I. (1991). Basque phonology. London and New York: Routledge. Hualde, J. I., Lakarra, J. A. & Trask, R. L. (Eds.). (1995). Towards a history of the Basque language. Amsterdam and Philadelphia: John Benjamins. Hudson, J. A. (1986). Memories are made of this: General event knowledge and development of autobiographic memory. In K. Nelson (Ed.), Event knowledge: Structure and function in development (pp. 97-118). Hillsdale, NJ: Lawrence Erlbaum Associates. Hudson, J. A. (1993). Reminiscing with mothers and others: Autobiographical memory in young two-year-olds. Journal of Narrative and Life History 3, 1-32.

References

545

Hudson, J. A. & Nelson, K. (1983). Effects of script structure on children's story recall. Developmental Psychology 19, 625—635. Hudson, J. A. & Shapiro, L. R. (1991). From knowing to telling: Children's scripts, stories, and personal narratives. In A. McCabe & C. Peterson (Eds.), Developing narrative structure (pp. 89-136). Hillsdale, NJ: Lawrence Erlbaum Associates. Hyams, N. (Ed.) (1986). Language acquisition and the theory of parameters. Dordrecht: Reidel. Hyams, N. (1989). The null subject parameter in language acquisition. In O. Jaeggli & K. J. Safir (Eds.), The null subject parameter (pp. 215-238). Dordrecht: Kluwer. Hymes, D. (1975). Breaking into performance. In D. Ben-Amos & K. Goldstein (Eds.). Folklore: Performance and communication (pp. 11-74). The Hague: Mouton. Hyona, J. & Olson, R. K. (1994). Dyslexic and normal readers' eye movement patterns in reading, visual search, and tracking. In J. Ygge & G. Lennerstrand (Eds), Eye movements in reading (pp. 233-244). Oxford: Pergamon. Ibarretxe-Antunano, I. (2001). Thinking for speaking in Basque. Evidence from oral and written narratives. Paper presented at the Eleventh Annual Meeting of the Society for Text and Discourse. Santa Barbara, July 12-14, 2001. Ibarretxe-Antunano, I. (in preparation). Tipi-tapa, tipi-tapa korrika!!! Motion and sound symbolism in Basque. Ibarretxe-Antunano, 1. (manuscript). Path and manner in motion events: A comparative study. Unpublished manuscript (2001). University of California at Berkeley. Ibarretxe-Antunano, I. (this volume). Motion events in Basque narratives. In S. Stromqvist & L. Verhoeven (Eds.) (2004), Relating events in narrative: Typological and contextual perspectives. Mahwah, NJ: Lawrence Erlbaum Associates. Ibarretxe-Antunano, I. (under review-a). Some notes on the semantic properties of "entrar en/a" 'enter in/to'in Spanish. Submitted (2002). Iverson, G. R. & Norpoth, H. (1987). Analysis of variance. Quantative applications in the social sciences, series no. I . Thousand Oaks, CA: Sage Publications. Jakobson, R. (1941). Kindersprache, Aphasie und allgemeine Lautgesetze. Uppsala: Almqvist och Wiksell. [English translation (1968): Child language, aphasia and phonological universals. The Hague: Mouton.] Jakobson, R., Fant, C. G. M. & Halle, M. (1952). Preliminaries to speech analysis. Cambridge, Massachusetts: MIT Press. Jefferson, G. (1978). Sequential aspects of storytelling in conversation. In J. Schenkein (Ed.), Studies in the organization of conversational interaction (pp. 219—248). New York: Academic Press. Jendraschek, G. (2002). Semantische Eigenschaften von Ideophonen im Turkischen [Semantic properties of ideophones in Turkish]. Munich: Lincom Europa.

546

References

Jisa, H. & Kern, S. (manuscript). La cognition, la langue elle aime ca: Developpement cognitifet langagier chez I 'enfant [Cognition, language she loves it: Cognitive and language development in the child]. Unpublished paper (1993). Universite Lumiere, Lyon. Jisa, H., Reilly, J., Verhoeven, L., Baruch, E. & Rosado, E. (2002). Passive voice constructions in written texts: Crosslinguistic and developmental perspectives. Written Language and Literacy 5, 163-182. Johansson, V. (1999). In Aisenman, R. (Ed.), Working papers in developing literacy across genres, modalities and languages, vol. 1 (pp. 182-198). Tel AvivUniversity: Department of Linguistics. Jonsson, J. G. (1992). The two perfects of Icelandic. Islenskt mat og almenn malfrcedi \ 4. 129-146. Joreskog, K. G. & Sorbom, D. (1996). LISREL-8 User's reference guide. Chicago: Scientific Software International. Kail, M. & Drissi, A. (1990). Acquisition des indices linguistiques et temporels du traitement: Recherches interlangues [Acquisition of linguistic and temporal indices of processing: Crosslinguistic investigation]. In Z. Zavialoff (Ed.), La lecture: De la neurobiologie a la pedagogic, vol. 1 (pp. 348-352). Paris: Editions L'Harmattan. Kail, M. & Hickmann, M. (1992). French children's ability to introduce referents in narratives as a function of mutual knowledge. First Language 12, 73-94. Kail, M. & Sanchez y Lopez, I. (1998). Referent introductions in Spanish children's narratives as a function of contextual constraints: A Crosslinguistic perspective. First Language 17, 103-130. Kamio, A. (1979). On the notion of speaker's territory of information: A functional analysis of certain sentence-final forms in Japanese. In G. Bedell, E. Kobayashi & M. Muraki (Eds.), Explorations in linguistics: Papers in honor ofKazuko Inoue (pp. 213-231). Tokyo: Kenkyusha. Karmiloff-Smith, A. (1979). A functional approach to child language: A study of determiners and reference. Cambridge: Cambridge University Press. Karmiloff-Smith, A. (1981). The grammatical marking of thematic structure in the development of language production. In W. Deutsch (Ed.), The child's construction of language (pp. 121-147). NY: Academic Press. Karmiloff-Smith, A. (1985). Language and cognitive processes from a developmental perspective. Language and Cognitive Processes 1, 61-83. Karmiloff-Smith, A. (1986a). From meta-processes to conscious access: Evidence from children's metalinguistic and repair data. Cognition 23, 95-147. Karmiloff-Smith, A. (1986b). Stage/structure versus phase/process in modelling linguistic and cognitive development. In I. Levin (Ed.), Stage and structure: Reopening the debate (pp. 164-190). Norwood, NJ: Ablex. Karmiloff-Smith, A. (1987). Function and process in comparing language and cognition. In M. Hickmann (Ed.), Social and functional approaches to language and thought (pp. 185-202]. Orlando: FL, Academic Press.

References

547

Kamiiloff-Smith, A. (1992). Beyond modularity: A developmental perspective on cognitive science. Cambridge, Ma: Bradford. Katzenberger, I. (1994). Cognitive, linguistic, and developmental factors in the narration of picture series. Unpublished doctoral dissertation, Tel Aviv University [in Hebrew]. Katzenberger, 1. (in press). Expository text organization. Discourse studies. Katzenberger, 1. & Cahana-Amitay, D. (in press). Segmentation in text production: A developmental perspective. Linguistics. Kaufman, D. (1995). Where have all the verbs gone: Autonomy and interaction in attrition. Southwest Journal of Linguistics 14 (1-2), 43-66 (special issue on language loss and public policy). Kaufman, D. (1998). Children's assimilatory patterns and LI attrition. Proceedings of the Annual Boston University Conference on Language Development (pp. 409-420). Sommerville, MA: Cascadilla Press. Kaufman, D. (2000). Attrition of Hebrew in the United States: Sociolinguistic perspectives. In Olshtain, E. (Ed.), Immigration, identity and language (pp. 173196). Magnes Publ., Hebrew University. Kaufman, D. (200la). Narrative development in Hebrew and English. In Verhoeven, L. & Stromqvist, S. (Eds.), Narrative development in a bilingual context (pp. 319-340). John Benjamins. Kaufman, D. (200 Ib). Tales of LI attrition: Evidence from pre-puberty children. In T. Ammerlaan, M. Hulsen, H. Strating & K. Yagmur (Eds.), Sociolinguistic and psvcholinguistic perspectives on maintenance and loss of minority languages (pp. 185-202). Munster, Germany: Waxmann. Kearins, J. M. (1981). Visual spatial memory in Australian Aboriginal children of desert regions. Cognitive Psychology 13, 434-460. Kellerman, E. & van Hoof, A.-M. (2002). Manual accents. In H. Jacobs & L. Wetzels (Eds.), Liber Amicorum BernardBichakjian. Maastricht: Shaker Press. Kendon, A. (1982). The study of gesture: Some observations on its history. Recherches semiotiques/Semiotic Inquiry 2.1, 45—62. Kern, S. (1997). Comment les enfants jonglent avec les contreaintes communicationnenelles, discursives et linquistiques dans la production d'une narration [How children juggle with communicative, discursive and linguistic constraints in the production of a narrative]. Doctoral dissertation, Universite Lumiere, Lyon 2. Kernan, K. T. (1977). Semantic and expressive elaboration in children's narratives. In S. Ervin-Tripp & C. Mitchell Kernan (Eds.), Child discourse. New York: Academic Press. Khattab, G. (2002). /!/ production in English-Arabic bilingual speakers. International Journal ofBilingualism 6.

548

References

Kirshenblatt-Gimblett, B. (1974). The concept and varieties of narrative performance in east European Jewish culture. In R. Bauman & J. Sherzer (Eds.), Explorations in the ethnography of speaking (pp. 283-308). Cambridge: Cambridge University Press. Kita, S. (1997). Two-dimensional semantic analysis of Japanese mimetics. Linguistics 35, 379^15. Kita, S. (1999). Japanese enter/exit verbs without motion semantics. Studies in Language 23,317-340. Klich, L. Z. (1988). Aboriginal cognition and psychological nescience. In S. H. Irvine & J. W. Berry (Eds.), Human abilities in cultural context (pp. 421-452). Cambridge: Cambridge University Press. Klima, E. & Bellugi, U. (1979). The signs of language. Cambridge, Mass.: Harvard University Press. Koch, H. (1984). The category of'associated motion' in Kaytej. Language in Central Australia 1, 23-34. Kollberg, P. (1998). S notation - a computer based method for studying and representing text composition. Royal Institute of Technology: Department of Numerical Analysis and Computing Science, Stockholm. Koopmans, M. (1995). [Review of R. A. Berman & D. I. Slobin (Eds.) (1994), "Relating events in narrative: A crosslinguistic developmental study". Hillsdale, NJ: Erlbaum.] Studies in Second Language Acquisition 17, 529-530. Kramer, J. (1981). Die Ubernahme der deutschen und der niederlandischen Konstruktion Verb + Verbzusatz durch die Nachbarsprachen [The adoption of the German and Dutch construction verb + verb particle through the neighbouring languages]. In W. Meid & K. Heller (Eds.), Sprachkontakt als Ursache von Verdnderungen der Sprach- und Bewusstseinsstruktur: Eine Sammlung von Studien zur sprachlichen Interferenz (pp. 129-140). Innsbruck, Austria: Institut fur Sprach wissenschaft der Universitat Innsbruck. Kress, B. (1982). Islandische Grammatik [Icelandic grammar]. Miinchen: Cax Hueber Verlag. Kuno, S. (1987). Functional syntax: Anaphora, discourse and empathy. Chicago: University of Chicago Press. Kiintay, A. (1997). Extended discourse skills of Turkish preschool children across shifting contexts. Unpublished doctoral dissertation, University of California, Berkeley. Kiintay, A. & Ervin-Tripp, S. (1997). Narrative structure and conversational circumstances. Journal of Narrative and Life His ton- 7, 113-120. Kiintay, A. & Ervin-Tripp, S. (1998). Conversational narratives of Turkish children: Occasions and structures. In A. Aksu-Koc, E. Erguvanh-Taylan, A. S. Ozsoy & A. Kiintay (Eds.), Perspectives in language acquisition: Selected papers from the International Congress for the Study of Child Language (pp. 237-250). Bogazici University Press: Istanbul.

References

549

Kuntay, A. & Nakamura, K. (1993). Evaluative strategies in elicited Turkish and Japanese narratives. Paper presented at the 6th International Congress for the Study of Child Language, Trieste, Italy. Kuntay, A. & Nakamura, K. (this volume). Linguistic strategies serving evaluative functions in Japanese and Turkish picture book narratives. In S. Stromqvist & L. Verhoeven (Eds.) (2004), Relating events in narrative: Typological and contextual perspectives. Mahwah, NJ: Lawrence Erlbaum Associates. Kupersmitt, J. (1999). Development of form/function relations in the narratives of Spanish-Hebrew bilinguals: What stays and what goes? M. A. thesis, Tel Aviv University. Kupersmitt, J. & Berman, R. A. (2001). Linguistic features of Spanish-Hebrew children's narratives. In L. Verhoeven & S. Stromqvist (Eds.), Narrative development in a multilingual context (pp. 277-318). Amsterdam/Philadelphia: John Benjamins. Kyratzis, A. (1993). Pragmatic and discursive influences on the acquisition of subordination-coordination. In E. V. Clark (Ed.), Proceedings of the 25th Annual Child Language Research Forum (pp. 324-332). Stanford University: Center for the Study of Language and Information. Labov, W. (1972). Language in the inner city. Philadelphia, PA: University of Pennsylvania Press. Labov, W. (1997). Some further steps in narrative analysis. Journal of Narrative and Life History 7, 395-415 (special issue, edited by M. Bamberg). Labov, W. & Waletzky, J. (1967). Narrative analysis: oral versions of personal experience. In J. Helm (Ed.), Essays on the verbal and the visual arts (pp. 12-44). Seattle, London: U. Washington Press. Lambrecht, K. (1981). Topic, antitopic, and verb agreement in non-standard French. Series Pragmatics and Beyond. Amsterdam: Benjamins. Langacker, R. (1987). Foundations of cognitive grammar I: Theoretical prerequisites. Stanford: Stanford University Press. Lanza, E. (2001). Temporality and language contact in narratives by children bilingual in Norwegian and English. In: L. Verhoeven & S. Stromqvist (Eds.), Narrative development in a multilingual context (pp. 15-50). Amsterdam/Philadelphia: John Benjamins. Lasagabaster, D. & Doiz, A. (in press). Maturational constraints on foreign language written production. In M. P. Garcia Mayo & M. L. Garcia Lecumberri (Eds.), Age and the acquisition of English as a foreign language: Theoretical issues and fieldwork. Clevedon: Multilingual Matters. Laughren, M. (1978). Directional terminology in Warlpiri. Working papers in language and linguistics 8 (pp. 1-16). Tasmanian College of Advanced Education, Launceston. Lee, P. (Ed.) (1996). The Whorf theory complex. A critical reconstruction. Amsterdam Studies in the Theory and History of Linguistic Science. Amsterdam: John Benjamins.

550

References

Lennert Olsen, L. & Hertling, B. (1988). Gronlandsk tilhcengsliste [A list of Greenlandic affixes]. Pilersuiffik. Leseman, P. (1997). Cultural dimensions of storybook reading. Paper presented at the 2nd International Conference on Functional Literacy, Noordwijkerhout, June 5-6, 1997. Leseman, P. & de Jong, P. (2001). How important is home literacy for acquiring literacy in school? In L. Verhoeven & C. E. Snow (Eds.), Literacy and motivation (pp. 71-94). Mahwah, NJ: Lawrence Erlbaum. Levelt, W. (1983). Monitoring and self repair in speech. Cognition 14, 41-104. Levelt, W. (1989). Speaking. Cambridge, Mass.: The MIT Press. Levinson, S. (2000). Presumptive meanings. The theory of generalized conversational implicature. Cambridge, Mass.: MIT Press. Levy, Y. (1987). The wug technique revisited. Cognitive Development 2, 71-87. Lewis, D. (1976a). Observations on route-finding and spatial orientation among the Aboriginal peoples of the Western Desert region of central Australia. Oceania 46. 249-282. Lewis, D. (1976b). Route finding by desert aborigines in Australia. The Journal of Navigation 29, 21-38. Lewis, G. L. (1986). Turkish grammar. Oxford: Clerendon Press. Liddell, S. (1995). Real, surrogate, and token space: Grammatical consequences in ASL. In K. Emmorey & J. Reilly (Eds.), Language, gesture, and space (pp. 19^2). Hillsdale, NJ: Lawrence Erlbaum Associates. Lieven, E. (1996). [Review of R. A. Berman & D. I. Slobin (Eds.) (1994), "Relating events in narrative: A crosslinguistic developmental study", Hillsdale, NJ: Erlbaum.] Functions of Language 3, 127-143. Linell, P. (1982). The written language bias in linguistics. University of Linkoping: Department of Communication Studies. Lodge, D. (1990). Narration with words. In H. Barlow, C. Blakemore & M. WestonSmith (Eds.), Images and understanding (pp. 141-153). Cambridge: Cambridge University Press. Lopez-Ornat, S. (1994). La adquisicion de la lengua espanola. [The acquisition of the Spanish language]. Siglo Veintiuno Editores, Espafia. Lotan, O. (2003). The effect of elicitation procedure on children's narrative production. Master's thesis, Tel Aviv University [in Hebrew]. MacWhinney, B. (1992). Transfer and competition in second language learning. In R. J. Harris (Ed.), Cognitive processes in bilinguals (pp. 371-390). Amsterdam: Elsevier. MacWhinney, B. (2000). The CHILDES project: Tools for analyzing talk. Hillsdale, NJ: Lawrence Erlbaum. Mandler, J. M. (1978). A code in the node: The use of a story schema in retrieval. Discourse Processes 1, 14—35.

References

551

Mandler, J. M. & Deforest, M. (1979). Is there more than one way to recall a story? Child Development 50, 886-889. Mandler, J. M. & Johnson, N. S. (1977). Remembrance of things passed: story structure and recall. Cognitive Psychology 9, 111-151. Maratsos, M. P. (1976). The use of definite and indefinite reference in young children. Cambridge: Cambridge U. Press. Marchman, V. (1989). Episodic structure and the linguistic encoding of events in narrative: A study of language acquisition and performance. Unpublished doctoral dissertation, University of California, Berkeley. Marcus, H. & Kitayama, S. (1991). Culture and the self: Implications for cognition, emotion, and motivation. Psychological Review 98, 224-253. Margetts, A. (1999a). Valence and transitivity in Saliba, an Oceanic language of Papua New Guinea. MPI Series in Psycholinguistics, MPI Nijmegen. Margetts, A. (1999b). Negation in Saliba. In U. Mosel & E. Hovdhaugen (Eds.), Negation in Oceanic languages: Typological studies. Studies in Austronesian Linguistics, 2. Mtinchen and Newcastle, LINCOM Europa. Margetts, A. (to appear). The linguistic encoding of three-participant events in Saliba. Studies in Language 26:3, 613—636. Martin, S. (1975). A reference grammar of Japanese. Tokyo: Tuttle. Mayer, M. (1969). Frog, where are you? New York: Dial Press. Maynard, S. K. (1990). An introduction to Japanese grammar and communication strategies. Tokyo: The Japan Times. Maynard, S. K. (1993). Discourse modality: Subjectivity, emotion and voice in the Japanese language. Amsterdam, Netherlands: John Benjamins. McCabe, A. & Peterson, C. (1988). A comparison of adults' versus children's spontaneous use of 'because 'and 'so'. Journal of Genetic Psychology 149, 257268. McCabe, A. & Peterson, C. (Eds.) (1991). Developing narrative structure. Hillsdale, NJ: Lawrence Erlbaum Associates. McCartney, K. A. & Nelson, K. (1981). Children's use of scripts in story recall. Discourse Processes 4, 59-70. McDonald, B. (1982). Aspects of the American Sign Language predicate system. Unpublished doctoral dissertation, State University of New York, Buffalo, Buffalo, NY. McGurk, H. & MacDonald, J. (1976). Hearing lips and seeing voices. Nature 264, 746-748. McLaughlin, B. (1985). Second language acquisition in childhood. Hillsdale, NJ: Lawrence Erlbaum Associates. McNeill, D. (1992). Hand and mind: What gestures reveal about thought. Chicago: University of Chicago Press.

552

References

McNeill, D. (1999). An ontogenetic universal and several cross-linguistic differences in thinking for speaking. Paper presented at the 6th International Cognitive Linguistics Conference, Stockholm, Sweden. McNeill, D. (2000). Analogic/analytic representations and cross-linguistic differences in thinking for speaking. Cognitive Linguistics 11. 1/2: 43-60. McNeill, D. & Duncan, S. D. (2000). Growth points in thinking-for-speaking. In D. McNeill (Ed.), Language and gesture: Window into thought and action (pp. 141161). Cambridge: Cambridge University Press. Menig-Peterson, C. L. (1975). The modification of communicative behavior in preschoolaged children as a function of the listener's perspective. Child Development 46, 1015-1018. Menig-Peterson, C. L. & McCabe, A. (1978). Children's orientation of a listener to the context of their narratives. Developmental Psychology 14, 582-592. Michaels, S. (1981). 'Sharing time': Children's narratives styles and differential access to literacy. Language in Society 10, 423^442. Michaels, S. & Collins, J. (1982). Oral discourse style: Classroom interaction and the acquisition of literacy. In D. Tannen (Ed.) 1982, Coherence in spoken and written discourse. Norwood, NJ: Ablex. Michelena, L. (1990) [1961]. Fonetica Historica Vasca [Basque Historical Phonetics]. Donostia: Gipuzkoako Foru Aldundia. Miller, P. J. & Sperry, L. L. (1988). Early talk about the past: The origins of conversational stories of personal experience. Journal of Child Language 15, 293-315. Minami, M. (1995). [Review of R. A. Berman & D. I. Slobin (Eds.) (1994), "Relating events in narrative: A crosslinguistic developmental study". Hillsdale, NJ: Erlbaum.] Journal of Narrative and Life History 5, 191-202. Minami, M. (1996). Japanese children's personal narratives. First Language 16, 339363. Minami, M. (1998). Politeness markers and psychological complements: Wrapping-up devices in Japanese oral personal narratives. Narrative Inquiry 8(2), 1-20. Minami, M. (2002). Vocabulary development in English-Japanese bilingual children: Correspondence of achievement levels in first and second languages. Studies in Language Sciences 2, 261-278. Minami, M. & McCabe, A. (1995). Rice balls and bear hunts: Japanese and North American family narrative patterns. Journal of Child Language 22, 423-445. Moore, A. (1990). A whole language approach to the teaching of bilingual learners. Illinois: Center for the Study of Reading. Mora Gutierrez, J. P. (1998). Directed motion in English and Spanish. Unpublished doctoral dissertation, University of Seville, Spain. [Available as Estudios de linguistica espanola, 2001, vol. 11 - http://elies.rediris.es/eliesl I/] Morgan, G. (1996a). Discourse organisation in a British Sign Language narrative. In E. Pedro (Ed.), Discourse analysis (pp. 235-251). Lisbon: Colibri.

References

553

Morgan, G. (1996b). Learning to use referential space in sign language discourse. In K. Bernhardt (Ed.), Papers in linguistics from the University of Manchester, vol. 1. (pp. 145-156). Morgan, G. (1999a). Event packaging in British Sign Language discourse. In E. Winston (Ed.), Story telling & conversation: Discourse in deaf communities, (pp. 27-58). Washington D. C: Galluadet University press. Morgan, G. (1999b). Narrative development in British Sign Language. In C. Schelletter, C. Letts & M. Garman (Eds.), Issues in child language: From phonology to narrative. Reading: FECS press. Morgan, G. (2002). The development of classifier constructions in British Sign Language. In K. Emmorey (Ed.), Classifiers in signed and spoken language. MIT press. Mosina, N. M. (2002). Erzyanskaya retsh detei ot 3 do 7 let [Erzyan children's speech from 3 to 7 years]. Morphological and syntactic analysis. Finno-Ugrian transcripts including glosses and translations in English. Saransk. Msimang, C. T. & Poulos, G. (2001). The ideophone in Zulu. In F. K. E. Voeltz & C. Kilian-Hatz (Eds.), Ideophones (pp. 235-249). Amsterdam/Philadelphia: John Benjamins. Muansuwan, N. (2000). Directional serial verb construction in Thai. In D. Flickinger & A. Kathol (Eds.), Proceedings of the 7th International HPSG Conference. UC Berkeley July 22-23, 2000. Munn, N. (1973). Warlpiri iconography. Cornell: Cornell University Press. Munn, N. D. (1986). Walbiri iconography: Graphic representation and cultural symbolism in a Central Australian society. Chicago: University of Chicago Press. Myers, F. (1986). Pintupi country, Pintupi self. Canberra: AIAS. Naigles, L. R., Eisenberg, A. R., Kako, E. T., Highter, M. & McGraw, N. (1998). Speaking of motion: Verb use in English and Spanish. Language and Cognitive Processes 13,521-549. Naigles, L. R. & Terrazas, P. (1998). Motion verb generalizations in English and Spanish: Influences of language and syntax. Psychological Science 9, 363-369. Nakamura, K. (1993). Referential structure in Japanese children's narratives: The acquisition of wa and ga. In S. Choi (Ed.), Japanese/Korean linguistics, vol.3 (pp. 84-99). Stanford, CA: CSLI. Nakamura, K. (1999). The acquisition of formal and informal language by Japanese preschool children. In M. Bernstein (Ed.), Proceedings of the 23rd Annual Boston University Conference on Language Development (pp. 507-518), Boston: Cascadilla Press. Nakamura, K. (manuscript). The acquisition of referential structure: The development of the usage ofwa and ga in Japanese oral narratives. Unpublished manuscript (1992), Department of Psychology, University of California, Berkeley. Narain, G. & Verhoeven, L. (1993). Ontwikkeling van vroege tweetaligheid [Development of early bilingualism]. Tilburg: University Press.

554

References

Narasimhan, B. (1998). The encoding of complex events in Hindi and English. Unpublished doctoral dissertation, Boston University. Nash, D. (1982). Warlpiri preverbs and verb roots. In S. Swartz (Ed.), Papers in Warlpiri grammar: In memory of Lothar Jagst (pp. 165-216). Work Papers of SIL-AAB: Series A, 6, Darwin: SIL. Nash, D. (1986). Topics in Warlpiri grammar. New York: Garland Publishing Inc. Nash, D. (manuscript). A traditional land claim by the Warlmanpa, Warlpiri. Mudburu and Waramungu traditional owners. Manuscript (1980). Naucler, K. (2002). Tell me why! Explanatory talk with Turkish and Swedish children at home and in the Swedish pre-school. In Researching early childhood, vol. 4: Social competence and communication. Early Childhood Research and Development Centre, University of Gothenburg. Naucler, K. & Boyd, S. (1996). Turkish and Swedish pre-school children's language socialization. Implications for literacy development. In L. Huss (Ed.), Manga vagar till tvasprakighet [Many roads to bilingualism]. Uppsala: Centre for Multiethnic Research, Uppsala University. Naucler, K. & Boyd, S. (1997). Telling a story. Norms for interaction with Turkish and Swedish children at home and in pre-school. In A. Sjogren (Ed.), Language and environment: A cultural approach to education for minority and migrant students. Botkyrka. Multicultural Centre. Naucler, K. & Boyd, S. (1999). Crosscultural perspectives on interaction with minority and majority children at home and in pre-school. In TASG News (Newsletter of the Turkish Area Study Group), 49/1999. Nelson, K. (1996). The emergence of the projective mind. Language in cognitive development: The emergence of the mediated mind (pp. 292-322). Cambridge: Cambridge University Press. Nelson, K. (Ed.) (1986). Event knowledge: Structure and function in development. Hillsdale, NJ: Lawrence Erlbaum Associates. Nelson, K. & Gruendel, J. M. (1981). Generalized event representations: Basic building blocks of cognitive development. In M. E. Lamb & A. L. Brown (Eds.), Advances in developmental psychology, vol. I (pp. 131-158). Hillsdale, NJ: Lawrence Erlbaum Associates. Nicolopolou, A. (1996). Narrative development in social context. In D. I. Slobin, J. Gerhardt, A. Kyratzis & J. Guo (Eds.), Social interaction, social context, and language. Hillsdale, NJ: Lawrence Erlbaum Associates. Nippold, M. A. (1998). Later language development. Austin, Texas: Pro-ed. Nistov, I. (2000). Referential choice in L2 narratives. A study of Turkish adolescent immigrants learning Norwegian. Doctoral dissertation, University of Oslo. Nistov, I. (2001). Reference continuation in L2 narratives. In L. Verhoeven & S. Stromqvist (Eds.), Narrative development in a multilingual context (pp. 51-86). Amsterdam/Philadelphia: John Benjamins. Noonan, M. (1999). Chantyal dictionary and texts. Berlin: Mouton de Gruyter.

References

555

Nordqvist, A. (1998). Projecting speech to protagonists in oral and written narratives: A developmental study. Psychology of Language and Communication 2, 37-46. Nordqvist, A. (1999). The development of (in)direct speech in spoken and written narratives. In S. StrSmqvist & E. Ahlsen (Eds.), The process of writing - a progress report (pp. 141-153). Gothenburg papers in theoretical linguistics 82. University of Goteborg: Department of Linguistics. Nordqvist, A. (2001). Speech about speech — A developmental study on form and function of direct and indirect speech. PhD. dissertation. Goteborg University, Sweden. O'Connell, B. & Gerard, A. (1985). Scripts and scraps: The development of sequential understanding; Child Development 56, 671-681. O'Neill, D. K. & Holmes, A. C. (2002). Young preschoolers' ability to reference story characters: The contribution of gesture and character speech. First Language 22, 73-103. Ochs, E. & Schieffelin, B. (1984). Language acquisition and socialization: Three developmental stories. In R. Shweder & R. LeVine (Eds.), Culture theory: Essays on mind, self, and emotion (pp. 276-320). Cambridge: Cambridge University Press. Oh, K. (in preparation). Language, cognition, and development: Motion events in English and Korean, Doctoral dissertation, University of California, Berkeley. Ohala, J. (1984). An ethological perspective on common cross-language utilization of F of voice. Phonetica4\: 1-16. Ohara, K. H. (2000). Cognitive and structural constraints on motion descriptions: Observations from Japanese and English. Proceedings of the 2nd International Conference on Cognitive Science and the 16lh Annual Meeting of the Japanese Cognitive Science Society (pp. 994-997). Ohman, S. (1979). Introduction. In Lindblom, B. & Ohman, S. (Eds.), Frontiers of speech communication research (pp. xv-xxv). London etc: Academic Press. Oktem, A. & Oktem, O. (1986). Kulturelle Identitat, Sozialisation und Sprache bei turkischen Arbeiterkindern in der Bundesrepublik Deutschland [Cultural identity, socialization, and language in Turkish working-class children in Western Germany]. In J. Rehbein (Ed.), Interkulturelle Kommunikation (pp. 70-102). Tubingen: Narr. Olsen, D. R. (in press). What writing is. Pragmatics and cognition. Olshtain, E. (1989). Is second language attrition the reversal of second language acquisition? Studies in Second Language Acquisition 11, 151-165. Olshtain, E. & Barzilay, M. (1991). Lexical retrieval difficulties in adult language attrition. In H. W. Seliger & R. M. Vago (Eds.), First language attrition (pp. 139150). Cambridge: Cambridge University Press. Ong, W. J. (1982). Orality and literacy: The technologizing of the word. London: Methuen.

556

References

Ordonez, C. L. (to appear). Oral bilingual proficiency of Colombian adolescents. Special issue of the International Journal of Bilingual Education and Bilingualism: Bilingual Education in South America. Ordonez, C. L., Barriga Villanueva, R., Snow, C., Uccelli, P., Shiro, M. & Schnell, B. (2001). Sintaxis y discurso: dos areas de investigacion en la adquisicion del espanol oral [Syntax and discourse: Two research areas in the acquisition of spoken Spanish]. Revista latino depensamiento y lenguaje 9:2, 131-163. Orsolini, M. (1990). Episodic structure in children's fantasy narratives: "Breakthrough" to decontextualised discourse. Language and Cognitive Processes 5, 53-79. Orsolini, M., Rossi, F. & Pontecorvo, C. (manuscript). Re-introduction of reference in Italian children's narratives. Manuscript (1993). Ozyiirek, A. & Oz9ahskan, S. (1998). How do children learn to conflate manner and path in their speech and gestures? Differences in English and Turkish. In E. V. Clark (Ed.), Proceedings of the 30th Stanford Child Language Forum. Stanford: CSLI. Ozfahskan, S. (2002). Metaphors we move bv: A crosslinguistic-developmental analysis of metaphorical motion events in English and Turkish. Unpublished doctoral dissertation, University of California, Berkeley. Ozcahskan, S. & Slobin, D. I. (1999). Learning how to search for the frog: Expression of manner of motion in English, Spanish, and Turkish. In A. Greenhill, H. Littlefield & C. Tano (Eds.), Proceedings of the 23rd Annual Boston University Conference on Language Development, vol. 2 (pp. 541-552). Somerville, MA: Cascadilla Press. Oz9ahskan, S. & Slobin, D. I. (2000a). Climb up vs. ascend climbing: Lexicalization choices in expressing motion events with manner and path components. In S. Catherine-Howell, S. A. Fish & T. K. Lucas (Eds.), Proceedings of the 2 4'h Annual Boston University Conference on Language Development, vol. 2 (pp. 558-570). Somerville, MA: Cascadilla Press. Oz9ahskan, S. & Slobin, D. I. (2000b). Expression of manner of movement in monolingual and bilingual adult narratives: Turkish vs. English. In A. Goksel & C. Kerslake (Eds.), Studies on Turkish and Turkic languages. Wiesbaden: Harrasowitz. Ozcahskan, S. & Slobin, D. I. (in press-a). Manner of movement in monolingual and bilingual adult narratives: Turkish vs. English. Proceedings of the 14lh International Conference on Turkish Linguistics. London: Holland Academic Graphics. Oz9ahskan, S. & Slobin, D. I. (in press-b). Codability effects on the expression of manner of motion in Turkish and English. In E. Erguvanh-Taylan & A. S. Ozsoy (Eds.), Proceedings of the 10th International Conference on Turkish Linguistics. Wiesbaden: Harrasowitz. Ozyiirek, A. & Kita, S. (1999). Expressing manner and path in English and Turkish: Differences in speech, gesture, and conceptualization. In M. Hahn & S. C. Stoness (Eds.), Proceedings of the 27" Annual Conference of the Cognitive Science Society (pp. 507-512). Mahwah, NJ: Lawrence Erlbaum Associates.

References

557

Papafragou, A., Massey, C. & Gleitman, L. (2001). Motion events in language and cognition. In A. H.-J. Do, L. Dominguez & A. Johansen (Eds.), Proceedings of the 25th Annual Boston University Conference on Language Development (pp. 566-574). Sommerville, MA: Cascadilla Press. Pearson, B. Z. (2001). Language and mind in the stories of bilingual children. In L. Verhoeven & S. Stromqvist (Eds.), Narrative development in a multilingual context. Amsterdam, NL: John Benjamins. Pearson, B. Z. (2002). Narrative competence among monolingual and bilingual children in Miami. In D. K. Oiler & R. E. Eilers (Eds.), Language and literacy in bilingual children. Clevedon, UK: Multilingual Matters. Pearson, B. Z. & Umbel, V. M. (1996). The relationship between narrative and linguistic skills in English and Spanish among Hispanic schoolchildren in Miami. Paper presented at Vllth International Congress for the Study of Child Language, Istanbul. Pederson, E., Danziger, E., Wilkins, D., Levinson, S., Kita, S. & Senft, G. (1998). Semantic typology and spatial conceptualization. Language 74(3), 557-589. Peled, N. & Blum-Kulka, S. (1992). Between speaking and writing: Narratives of 1st- to 3rd-grade children. Maagaley Kria 21, 111-127 [in Hebrew]. Pellegrini, A. D. (1984). The effect of dramatic play on children's generation of cohesive text. Discourse Processes 7, 57-67'. Peng, Y. S. (1992). Developmental study of aspects of Malay narratives. Unpublished B. A. honours thesis. Department of Linguistics, La Trobe University, Melbourne, Victoria, Australia. Perner, J. (1993). Understanding thinking and belief. Understanding the representational mind (pp. 171-204) Cambridge, MA: The MIT Press. A Bradford Book. Peterson, C. (1990). The who, when and where of early narratives. Journal of Child Language 17, 433^55. Peterson, C. & Dodsworth, P. (1991). A longitudinal analysis of young children's cohesion and noun specification in narratives. Journal of Child Language 18, 397-415. Peterson, C. & McCabe, A. (1983). Developmentalpsycholinguistics: Three ways of looking at a child's narrative. New York: Plenum. Peterson, C. & McCabe, A. (1997). Extending Labov and Waletzky. Journal of Narrative and Life History 7, 251-258. Pfaff, C. (1993). Turkish language development in Germany. In G. Extra & L. Verhoeven (Eds.), Immigrant languages in Europe (pp. 119-146). Clevedon: Multilingual Matters. Pfaff, C. (1994). Early bilingual development of Turkish children in Berlin. In G. Extra & L. Verhoeven (Eds.), The cross-linguistic study of bilingual development (pp. 97-122). Amsterdam: Royal Academy of Sciences.

558

References

Pfaff, C. (1998). Changing patterns of language mixing in a bilingual child. In G. Extra & L. Verhoeven, Bilingualism and migration (pp. 97-122). Berlin: Mouton de Gruyter. Pfaff, C. (2001). The development of co-constructed narratives. In L. Verhoeven & S. Stromqvist (Eds.), Narrative development in a multilingual context (pp. 153-188). Amsterdam/Philadelphia: John Benjamins. Pierrard, M., Janssens, S. & Housen, A. (in press). Le francais face a /'anglais dans les ecoles secondaires en Flandre [French versus English in secondary schools in Flanders]. Serie Francais et Societe, Ministere de la Communaute Francaise de Belgique. Brussels, Duculot. Power, R. J. D. & Dal Martello, M. F. (1986). The use of the definite and indefinite articles by Italian preschool children. Journal of Child Language 13, 145-154. Preece, A. (1987). The range of narrative forms conversationally produced by young children. Journal of Child Language 14, 353-373. Rabinowitch, S. (1985). Proficiency in Hebrew of English-Hebrew bilingual 3- to 5year-olds compared with their monolingual Hebrew-speaking peers. Doctoral dissertation (in Hebrew), Tel Aviv University. Radford, A. (1995). Phrase structure and functional categories. In P. Fletcher & B. MacWhinney (Eds.), Handbook of child language (pp. 483-507). Cambridge, MA: Basil Blackwell. Ragnarsdottir, H. (1987). The development of narrative structure and its relation to the use offenses and aspects in Icelandic children s stories. Paper presented at the conference of the Icelandic Linguistic Society in honor of the 200th anniversary of Christian Rask, Reykjavik. Ragnarsdottir, H. (1988). Soguprddur og setningagerdir i sogum barna ogfullordinna [Narrative structure and sentence types in children's and adults' stories]. Paper presented at the conference on Icelandic in the primary school, organized by the Icelandic Institute of Educational Research, Reykjavik. Ragnarsdottir, H. (1991). What does it take to tell a story? Paper presented at the Scandinavian Congress of Logopedics and Phoniatrics, Reykjavik. Ragnarsdottir, H. (1992). Episodic structure and interclausal connectives in Icelandic children's narratives. In R. Soderbergh (Ed.), Proceedings of Colloquium Paedolinguisticum Lundensis (pp. 33^45). Lund University, Sweden: Dept. of Linguistics. Ragnarsdottir, H. & Stromqvist, S. (1997), The linguistic encoding of spatial relations in Scandinavian child language development. In E. V. Clark (Ed.), Proceedings of the 28th Annual Child Language Research Forum (pp. 271-282). Stanford: CSLI. Ragnarsdottir, H., Stromqvist, S. & Nordqvist, A. (2000). Completed and progressive action in Swedish and Icelandic child language. In E. V. Clark (Ed.), Proceedings of the 301'' Annual Child Language Research Forum (pp. 95-104). Stanford: CSLI. Ravid, D. (1988). Transient and fossilized phenomena in inflectional morphology-: varieties of spoken Hebrew. Doctoral dissertation, Tel Aviv University.

References

559

Ravid, D. (1995). Language change in child and adult Hebrew: A psycholinguistic perspective. Oxford University Press. Ravid, D. & Avidor, A. (1998). Acquisition of derived nominals in Hebrew: Developmental and linguistic principles. Journal of Child Language 25, 229-266. Ravid, D. & Cahana-Amitay, D. (in press). Verbal and nominal expression in narrating conflict situations. Journal of Pragmatics (special issue on discourse stance). Ravid, D. & Tolchinsky, L. (2002). Developing linguistic literacy: A comprehensive model. Journal of Child Language 29, 419^48. Ravid, D., van Hell, J., Rosado, E. & Zamora, A. (2002). Subject NP patterning in the development of text production. Written Language and Literacy 5, 69-93. Reilly, J. (1992). How to tell a good story: The intersection of language and affect in children's narratives. Journal of Narrative and Life History 2, 355-377. Reilly, J., Bates, E. & Marchman, V. (1998). Narrative discourse in children with early focal brain injury. In M. Dennis (Ed.), Brain and Language 61(3), 335-375 (special issue on discourse in children with anomalous brain development or acquired brain injury). Reilly, J., Jisa, H., Baruch, E. & Berman, R. A. (2002). Prepositional attitudes: Development of modal expression in three languages at different ages. Written Language and Literacy 5, 183-218. Reilly, J., Klima, E. S., Bellugi, U. (1990). Once more with feeling: Affect and language in atypical populations. Development and Psychopathology 2, 367-391. Reilly, J., Zamora, A. & McGivern, R. (in press). Acquiring perspective in English: The development of stance. Journal of Pragmatics (special issue on discourse stance). Reinhart, T. (1984). Principles of gestalt perception in the temporal organization of narrative texts. Linguistics 22, 779-809. Renner, T. (1988). Development of temporality in children s narratives. Unpublished doctoral dissertation, University of California, Berkeley. Rescorla, L. & Manhardt, J. (manuscript). Narrative competencies of children with a history of specific expressive language delay. Unpublished manuscript (1993), Department of Human Development, Bryn Mawr College. Rizzi, L. (1986). Null subjects and the theory of "pro". Linguistic Inquiry 17, 501-557. Rodriguez-Trelles Astruga, A. (1991). Expresion de las relaciones temporales y aspectuales en los relatos: Una perspectiva evolutiva [Expression of temporal and aspectual relations in stories: A developmental perspective]. Unpublished doctoral dissertation, University of Santiago de Compostela, Galicia, Spain. Rose, S. A. & Blank, M. (1974). The potency of context in children's cognition: An illustration through conservation. Child Development 45, 499-502. Rossi, F. (manuscript). Forms and functions of reference in spoken and written narratives. Unpublished manuscript (1993), Department of Developmental and Social Psychology, University of Rome, Italy.

560

References

Rubino, C. (2001). Iconic morphology and word formation in Ilocano. In F. K. E. Voeltz & C. Kilian-Hatz (Eds.), Ideophones (pp. 304-320). Amsterdam/Philadelphia: John Benjamins. Sachs, J. (1983). Talking about the there and then: The emergence of displaced reference in parent-child discourse. In K. Nelson (Ed.), Children's language, vol. 4. Hillsdale, NJ: Lawrence Erbaum Associates. Sacks, H., Schegloff, E. & Jefferson, G. (1974). A simplest systematics for the organization of turntaking in conversation. Language 50, 696-735. Sak-Humphry, C., Indambraya, K. & Starosta, S. (1997). Flying 'in' and 'out' in Khmer and Thai. In A. S. Abramson (Ed.), South Asian linguistics studies in honour of Vichin Panupong. Bangkok: Chulalongkom University Press. Sapir, E. (1958) [1924]. Selected writings of Edward Sapir in language, culture, and personality. Ed. by D. G. Mandelbaum. Berkeley, Los Angeles: University of California Press. Satoh, K. (200 la). Thinking for speaking in Japanese narrative. In Proceedings of the 1" Seoul International Conference on Discourse and Cognitive Linguistics. Satoh, K. (200Ib). Katari-ni-okeru tense-no koutai to aspect [Tense shift and aspectuality in narrative]. Bulletin of Yokohama City University. Satoh, K. (in press). Katari-no kouseiyouso-toshite no aspect [Aspect of narrative structure]. Saussure, F. de. (1916). Cours de linguistique generate. Ed. by C. Bally, A. Sechehaye & A. Riedlinger. Lausanne: Payot. [English translation (1966): Course in general linguistics. New York: McGraw-Hill.] Sauvaire, V. & Vion, M. (1989). Expression of the given/new contract in referential communication: A study of seven- and nine-year-old children. European Bulletin of Cognitive Psychology 9, 431-449. Scarborough, H. S. & Dobrich, W. (1994). On the efficacy of reading to preschoolers. Developmental Review 14, 245-302. Schaefer, R. P. (2001). Ideophonic adverbs and manner gaps in Emai. In F. K. E. Voeltz & C. Kilian-Hatz (Eds.), Ideophones (pp. 339-354). Amsterdam/Philadelphia: John Benjamins. Schank, R. & Abelson, R. (1977). Scripts, plans, goals, and understanding: An inquiry into human knowledge structures. Hillsdale, NJ: Lawrence Erlbaum Associates. Schaufeli, A. (1991). Turkish in an immigrant setting. Amsterdam: University Press. Schaufeli, A. (1993). Turkish language development in the Netherlands. In G. Extra & L. Verhoeven (Eds.), Immigrant languages in Europe (pp. 147-158). Clevedon: Multilingual Matters. Schick B. (1987). The acquisition of classifier predicates in American Sign Language. Unpublished doctoral dissertation, Purdue University.

References

561

Schiff, R. (1991). Hashva 'a beyn ha-yexolot ha-morfologit, ha-taxbirit ve haSilva narativit ecel kor 'im tovim le 'umat mitkashim [A comparison of the morphological, syntactic, and narrative ability of proficient and poor SI readers]. Unpublished master's thesis, Tel Aviv University. Schneider, P., Williams, B. & Hickmann, M. (1997). Awareness of referential cohesion in stories: A comparison of children with and without language/learning disabilities. Journal of Speech-language Pathology andAudiology 21 (1), 8-16. Schultze-Berndt, E. (2000). Simple and complex verbs in Jaminjung: A study of event categorisation in an Australian language. MPI Series in Psycholinguistics, no. 14. Wageningen, Netherlands: Ponsen & Looijen. Schwarze, C. (1985). "Uscire" e "andare fuori": struttura sintattica e semantica lessicale ["Exit" and "go out": Syntactic structure and lexical semantics]. Societa di Linguistica Italiana 24, 355-371. Searle, J. R. (1978). Literal meaning. Erkenntnis 13, 207-224. Sebastian, E. (1989). Tiempoy aspecto verbal en el lenguage infantil [Verbal tense and aspect in child language]. Unpublished doctoral dissertation, Universidad Autonoma de Madrid. Sebastian, E. & Slobin, D. I. (1994). Development of linguistic forms: Spanish. In R. A. Berman & D. I. Slobin (Eds.), Relating events in narrative: A cross-linguistic developmental study (pp. 239-284). Hillsdale, NJ: Lawrence Erlbaum Associates. Segal, E. M., Miller, G., Hosenfeld, C., Mendelsohn, A., Russell, W., Julian, J., Greene, A. & Delphonese, J. (1997). Person and tense in narrative interpretation. Discourse Processes 24(2-3), 271-307. Seidman, S., Nelson, K. & Gruendel, J. (1986). Make-believe scripts: the transformation of ERS in fantasy. In K. Nelson (Ed.), Event knowledge: Structure and function in development (pp. 161-188). Hillsdale, NJ: Lawrence Erlbaum Associates. Selinker, L. (1974). Interlanguage. In J. Richards (Ed.), Error analysis: Perspectives on second language acquisition (pp. 31-54). London: Longman. Senft, G. (2001). What do we really know about serial verbs constructions in Austronesian and Papuan languages? Paper presented at the 3rd European Meeting on Oceanic linguistics, Paris, March 9-10, 2001. Senft, G. (to appear). Prolegomena to a Kilivila grammar of space. In S. C. Levinson & D. Wilkins (Eds.), Grammars of space. Cambridge: Cambridge University Press. Sera, M. D. (1992). To be or to be: Use and acquisition of the Spanish copulas. Journal of Memory and Language 31, 408-427. Serpell, R. (2000). Cultural dimensions of literacy promotion and schooling. In L. Verhoeven & C. E. Snow (Eds.), Literacy and motivation (pp. 243-274). Mahwah, NJ: Lawrence Erlbaum. Shank, R. C. & Abelson, R. P. (1977). Scripts, plans, goals, and understanding. Hillsdale, NJ: Lawrence Erlbaum Associates. Shannon, B. (1984). Room descriptions. Discourse Processes 7, 225-255.

562

References

Shapiro, L. R. & Fernald, A. (1998). Connectedness and agency in mother-child play: A Japan-U. S. comparison. Poster presented at the Biennial Meeting of the Society for Research in Child Development, Albuquerque, New Mexico. Shapiro, L. R. & Hudson, J. A. (1991). Tell me a make-believe story: Coherence and cohesion in young children's picture-elicited narratives. Developmental Psychology 27, 960-974. Shatz, M. (1985). A song without music and other stories: How cognitive process constraints influence children's oral and written narratives. In D. Schiffrin (Ed.), Meaning, form, and use in context: Linguistic applications (pp. 313-324). 35th Annual Georgetown University Round Table on Languages and Linguistics. Washington, DC: Georgetown University Press. Shen, Y. (1988). Schema theory and the processing of narrative texts: The X-bar story grammar and the notion of discourse topic. Journal of Pragmatics 12, 639-676. Shen, Y. (1990). Centrality and causal relations in narrative texts. Journal of Literary Semantics 19, 1-29. Shen, Y. & Berman, R. A. (1997). From isolated event to action structure: Stages in narrative development. In Y. Shimron (Ed.), Mexkarim ba-psixologiya shel halashon be-Yisrael: Rexishat lashon, kri'a u-xtiva [Psycholinguistic studies in Israel: Language acquisition, reading, and writing] (pp. 119-145). Jerusalem: Magnes Press. Shibatani, M. (1990). The languages of Japan. New York: Cambridge University Press. Shiro, M., Uccelli, P., Ordonez, C, Barriga Villanueva, R., Snow, C. & Schnell, B. (2001). Learning to talk: A partial review of research on Spanish language development. In M. Almgren, A. Barrena, M.-J. Ezeizabarrena, I. Idiazabal & B. MacWhinney (Eds.), Research on Child Language Acquisition: Proceedings of the 8th Conference of the International Association for the Study of Child Language (pp. 582-604). Somerville, MA: Cascadilla Press. Sigurdsson, H. A. (1989). Verbal syntax and case in Icelandic. University of Lund: Department of Nordic languages. Silva-Corvalan, C. (1983). Tense and aspect in oral Spanish narrative: Context and meaning. Language 59, 60-80. Silva-Corvalan, C. (1992). Algunos aspectos de la gramatica de los ninos bilingiies de Los Angeles [Some aspects of the grammar of bilingual children of Los Angeles]. In H. Urrutia Cardenas & C. Silva-Corvalan (Eds.), Bilinguismoy adcquisicion del espanol (pp. 227-239). Bilbao, Spain: Institute Horizonte. Silva-Corvalan, C. (1994). Language contact and change: Spanish in Los Angeles. Oxford: Clarendon Press. Sinha, C. & Kuteva, T. (1995). Distributed spatial semantics. Nordic Journal of Linguistics 18(2), 167-199. Slobin, D. I. (1977). Language change in childhood and in history. In Macnamara, J. (Ed.), Language learning and thought: Perspectives in neurolinguistics and psycholinguistics (pp. 185-214). New York: Academic Press.

References

563

Slobin, D. I. (1985). Crosslinguistic evidence of the language-making capacity. In D. I. Slobin (Ed.), The crosslinguistic study of language acquisition, vol. 2 (pp. 11571256). New Jersey: Lawrence Erlbaum. Slobin, D. I. (Ed.) (1985-1997). The crosslinguistic study of language acquisition, vol. 1-5. Hillsdale, NJ: Lawrence Erlbaum Associates. Slobin, D. I. (1987). Thinking for speaking. Proceedings of the 13th Annual Meeting of the Berkeley Linguistics Society (pp. 435^144). Berkeley, CA: Berkeley Linguistics Society. Slobin, D. I. (1988). The development of clause chaining in Turkish child language. In S. Ko? (Ed.), Studies on Turkish linguistics (pp. 27-54). Ankara: Middle East Technical University. [Circulated as Berkeley cognitive science report no. 55, Institute of Cognitive Studies, University of California, Berkeley, 1989.] Slobin, D. I. (1989). Factors of language typology in the crosslinguistic study of acquisition. Paper presented at the Tenth Biennial Meeting of the International Society for the Study of Behavioural Development, Finland. [Circulated as Berkeley cognitive science report no. 66, Institute of Cognitive Studies, University of California, Berkeley, August 1990.] Slobin, D. I. (1990). The development from child speaker to native speaker. In J. W. Stigler, G. Heerdt & R. A. Shweder (Eds.), Cultural psychology: Essays on comparative human development (pp. 235-256). Cambridge: Cambridge University Press. Slobin, D. I. (1991). Learning to think for speaking: Native language, cognition, and rhetorical style. Pragmatics 1, 7-26. Slobin, D. I. (1993a). Converbs in Turkish child language: The grammaticalization of event coherence. In M. Haspelmath & E. Konig (Eds.), Converbs (adverbial participles, gerunds) in crosslinguistic perspective. Berlin: Mouton de Gruyter. Slobin, D. I. (1993b). Passives and alternatives in children's narratives in English, Spanish, German, and Turkish. In B. Fox & P. J. Hopper (Eds.), Voice: Form and function. Amsterdam: John Benjamins. Slobin, D. I. (1996a). Two ways to travel: Verbs of motion in English and Spanish. In M. Shibatani & S. A. Thompson (Eds.), Grammatical constructions: Their form and meaning (pp. 195-220). Oxford: Clarendon Press. Slobin, D. I. (1996b). From "thought and language" to "thinking for speaking". In J. Gumperz & S. Levinson (Eds.), Rethinking linguistic relativity. Studies in the social and cultural foundations of language, vol. 17 (pp. 70-96). Cambridge: Cambridge University Press. Slobin, D. I. (1997a). Mind, code, and text. In J. Bybee, J. Haiman & S. A. Thompson (Eds.), Essays on language function and language type: Dedicated to T. Givon (pp. 437-467). Amsterdam/Philadelphia: John Benjamins. Slobin, D. I. (1997b). The universal, the typological and particular in acquisition. In D. I. Slobin (Ed.), The crosslinguistic study of language acquisition, vol. 5: Expanding the contexts. Mahwath, NJ, Lawrence Erlbaum Associates.

564

References

Slobin, D. I. (1998). A typological perspective on learning to talk about space. In H. Ragnarsdottir & S. Stromqvist (Eds.), Learning to talk about time and space. Proceedings of the 3rd NELAS Conference. Gothenburg papers in theoretical linguistics 80 (pp. 1-30). Reykjavik & Goteborg: University College of Education and Department of Linguistics, University of Goteborg. Slobin, D. I. (2000). Verbalized events: A dynamic approach to linguistic relativity and determinism. In S. Niemeier & R. Dirven (Eds.), Evidence for linguistic relativity (pp. 107-138). Amsterdam/Philadelphia: John Benjamins. Slobin, D. I. (2001). Form-function relations: how do children find out what they are? In M. Bowerman & S. Levinson (Eds.), Language acquisition and conceptual development. Language, culture & cognition, vol. 3. Cambridge: Cambridge University Press. Slobin, D. I. (2002a). Cognitive and communicative consequences of linguistic diversity. In S. Stromqvist (Ed.) The diversity of languages and language learning (pp. 1-14). Lund: Centre for Languages and Literature. Slobin, D. I. (2002b). Language and thought online: Cognitive consequences of linguistic relativity. In D. Centner & S. Goldin-Meadow (Eds.), Advances in the investigation of language and thought. Cambridge, MA: MIT Press. Slobin, D. I. & Aksu, A. (1982). Tense, aspect and modality in the use of the Turkish evidential. In P. J. Hopper (Ed.), Tense-aspect: Between semantics and pragmatics. Typological studies in language, vol. 1 (pp. 185-200). Amsterdam: Benjamins. Slobin, D. I. & Bocaz, A. (1988). Learning to talk about movement through time and space: The development of narrative abilities in Spanish and English. Lenguas modernas 15, 5-24. Santiago, Chile. [Circulated as Berkeley cognitive science report no. 55, Institute of Cognitive Studies, University of California, Berkeley, January 1989.] Slobin, D. I. & Hoiting, N. (1994). Reference to movement in spoken and signed languages: Typological consideration. Proceedings of the 201'' Annual Meeting of the Berkeley Linguistics Society (pp. 487-503). Berkeley, CA: Berkeley Linguistics Society. Snow, C., Barnes, W., Chandler, J., Goodman, I & Hemphill, L. (1991). Unfulfilled expectations: Home and school influences on literacy. Boston: Harvard University Press. Stahlke, H. (1970). Serial verb. Studies in African Linguistics 1, 60-99. Stavans, A. (2001). Trilingual children narrating in Hebrew, English and Spanish. In L. Verhoeven & S. Stromqvist (Eds.), Narrative development in a multilingual context (pp. 341-372). Amsterdam/Philadelphia: John Benjamins. Steenge, J., Verhoeven, L. & van Balkom, H. (2002). Language impairment in bilingual children: Narrative data from bilingual Turkish and Moroccan children with language impairment in the Netherlands. Paper presented at the 23rd Annual Symposium on Research in Child Language Disorders and the 9Ih Congress of the International Association for the Study of Child Language (SRCLD-IASCL 2002). Madison, WI.

References

565

Stein, N. L. (1982). The definition of a story. Journal of Pragmatics 6, 487-507. Stein, N. L. & Albro, E. R. (1997). Building complexity and coherence: Children's use of goal-structured knowledge in telling stories. In M. Bamberg (Ed.), Narrative development: Six approaches (pp. 5—44). Mahwah, New Jersey: Lawrence Erlbaum Associates. Stein, N. L. & Glenn, C. G. (1979). An analysis of story comprehension in elementary school children. In R. Freedle (Ed.), Advances in discourse processes, vol. 2: New directions in discourse processing (pp. 53-120). Norwood, NJ: Ablex. Stein, N. L. & Glenn, C. G. (1982). Children's concepts of time: The development of a story schema. In W. Friedman (Ed.), The developmental psychology of time (pp. 255-281). NY: Academic Press. Stein, N. L. & Nezworski, M. T. (1978). The effects of organization and instructional set of story memory. Discourse Processes 1, 177-193. Stein, N. L. & Trabasso, T. (1981). What's in a story: Critical issues in story comprehension. In R. Glaser (Ed.), Advances in the psychology of instruction (pp. 213-267). Hillsdale, NJ: Lawrence Erlbaum Associates. Stenning, K. & Mitchell, L. (1985). Learning how to tell a good story: The development of content and language in children's telling of one tale. Discourse Processes 8, 261-279. Stephany, U. (1997a). Endophorische Anapher und Textkoharenz in deutschen und griechischen Kindergeschichten [Endophoric anaphora and textual coherence in German and Greek children's narratives]. In C. Diirscheid, K. H. Ramers & M. Schwarz (Eds.), Sprache im Fokus. Festschrift fitr Heinz Vater zum 65. Geburtstag (pp. 473-494). Tubingen: Niemeyer. Stephany, U. (1997b). The development of textual coherence in Greek children s narratives. Paper presented at the Vlllth European Conference on Developmental Psychology, Rennes, France, Sept. 3-7, 1997. Stephany, U. (1998). Referent tracking in Greek and German children's narratives. In A. Aksu-Koc, E. Erguvanli-Taylan, A. S. Ozsoy & A Kuntay (Eds.), Perspectives on language acquisition: Selected papers from the Vllth International Congress for the Study of Child Language (pp. 277-291). Istanbul: Bogazici University Press. Strauss, S. (Ed.) (1982). U-shaped behavioral growth. New York: Academic Press. Stromqvist, S. (1994). Language acquisition, processing and change - perspectives on speech, sign, and writing. In Ahlgren, I. & Hyltenstam, K. (Eds.), Bilingualism in deaf education (pp. 219-238). Hamburg: Signum. Stromqvist, S. (1996). Discourse flow and linguistic information structuring: Explorations in speech and writing. Gothenburg papers in theoretical linguistics 78. Stromqvist, S. (1997). Understanding descriptions of motion — a crosslinguistic perspective on narrative discourse. In G. Hakansson & U. Nettelbladt (Eds.), Sprakforstaelse [Language understanding]. Proceedings of the 1995 ASLA Conference (pp. 207-222). Lund: University of Lund.

566

References

Stromqvist, S. (1998). Tal, skrift och tanke i sprakinlarningsprocessen [Speech, writing, and thought in the language acquisition process]. In K. Hainenen & M. Lehtihalmes (Eds.), Proceedings of the 7th Nordic Child Language Symposium (pp. 12-22). Oulon Yliopisto: Publications of the Department of Finnish, Saami and Logopedics. Stromqvist, S. & Ahlsen, E. (Eds.) (1999). The process of writing - a progress report. Gothenburg papers in theoretical linguistics 82. University of Goteborg: Department of Linguistics. Stromqvist, S., Ahlsen, E. & Wengelin, A. (1999). The production process in speech and writing. In S. Stromqvist & E. Ahlsen (Eds.), The process of writing - a progress report. Gothenburg papers in theoretical linguistics 82 (pp. 9-24). University of Goteborg: Department of Linguistics. Stromqvist, S. & Day, D. (1993). On the development of narrative structure in child LI and adult L2 acquisition. Journal of Applied Psycholinguistics 14, 135-158. Stromqvist, S. & Hellstrand, A. (1994). Tala och skriva i lingvistiskt och didaktiskt perspektiv: en projektbeskrivning [Speech and writing from linguistic and didactic perspectives: A project description]. Didaktisk tidskrift 1-2. Didaktiska sallskapet, Goteborg. Stromqvist, S., Hellstrand, A. & Nordqvist, A. (2000). From child speech to literary art. In P. Breeder & J. Murre (Eds.), Language and cognition (pp. 5-28). Tubingen: Gunter Narr. Stromqvist, S., Johansson, V., Kriz, S., Ragnarsdottir, H. & Ravid, D. (2002). Towards a crosslinguistic comparison of lexical quanta in speech and writing. Written Language and Literacy 5, 45-68. Stromqvist, S. & Karlsson, H. (2002). ScriptLogfor Windows - user's manual. Technical report. University of Lund: Department of Linguistics/University College of Stavanger: Centre for Reading Research. Stromqvist, S. & Malmsten, L. (1998). ScriptLogfor Macintosh. Technical report. University of Goteborg: Department of Linguistics. Stromqvist, S., Nordqvist, A. & Wengelin, A. (this volume). Writing the frog story: Developmental and cross-modal perspectives. In S. Stromqvist & L. Verhoeven (Eds.) (2004), Relating events in narrative: Typological and contextual perspectives. Mahwah, NJ: Lawrence Erlbaum Associates. Stromqvist, S., Ragnarsdottir, H., Engstrand, O., Jonsdottir, H., Lanza, E., Leiwo, M., Nordqvist, A., Peters, A., Plunkett, K., Richthoff, U., Simonsen, H. G., Toivainen, J. & Toivainen, K. (1995). The inter-Nordic study of language acquisition. Nordic Journal of Linguistics 18, 3-29. Supalla, T. (1982). Structure and acquisition of verbs of motion and location in American Sign Language. Unpublished doctoral dissertation, University of California, San Diego. Supalla, T. (1990). Serial verbs of motion in ASL. In S. D. Fischer & P. Siple (Eds.), Theoretical issues in sign language research, vol. 1: Linguistics (pp. 127-152). Chicago: University of Chicago Press.

References

567

Sutton, P., Anderson, C., Jones, P., Dussart, F. & Hemming, S. (1988). Dreamings. The art of Aboriginal Australia. NY: The Asia Society Galleries in association with George Braziller Publishers. Svavarsdottir, A. & Jonsdottir, M. (1998) [1988]. Islenskafyrir utlendinga: kennslubok i mdlfrcedi. [Icelandic for foreigners: Manual in Icelandic grammar]. 2nd edition. Reykjavik: Malvisindastofnun Haskola islands. Tager-Flusberg, H. (1993). Once upon a ribbit: Stories narrated by autistic children. Paper presented at the Annual Meeting of the Developmental Section of the British Psychological Society, Cambridge. Takahashi, K. (1997). Verbs of global locomotary body motion in Thai. In B. Caron (Ed.), Proceedings of the 16th International Congress of Linguists, Paris, July 20-25, 1997. Takahashi, K. (manuscript). Tense variation in Japanese narrative. Unpublished manuscript (1993), University of California at Los Angeles. Talmy, L. (1972). Semantic structures in English andAtsugewi. Unpublished PhD. dissertation. University of California, Berkeley. Talmy, L. (1983). How languages structure space. In H. Pick & L. Acredolo (Eds.), Spatial orientation: Theory, research and application (pp. 225-282). New York: Plenum. Talmy, L. (1985). Lexicalization patterns: Semantic stucture in lexical forms. In T. Shopen (Ed.), Language typology and syntactic description, vol. 3: Grammatical categories and the lexicon (pp. 36-149). Cambridge: Cambridge University Press. Talmy, L. (1991). Path to realization: A typology of event conflation. Proceedings of the 17"' Annual Meeting of the Berkeley Linguistics Society (pp. 480-519). Berkeley, CA: Berkeley Linguistics Society. Talmy, L. (2000a). Toward a cognitive semantics, vol. 1. Cambridge, Mass.: MIT Press. Talmy, L. (2000b). Toward a cognitive semantics, vol. 2. Cambridge, Mass.: MIT Press. Taub, S. (2001). Language from the body: Iconicity and conceptual metaphor in American Sign Language. Cambridge: Cambridge University Press. Taub, S. & Galvan, D. (2000). Patterns of conceptual encoding in ASL motion descriptions. Sign Language Studies 1, 175-200. Taub, S., Pinar, P. & Galvan, D. (2002). Comparing spatial information in speech/gesture and sign language. Paper presented at the 1st Congress of the International Society for Gesture Studies, Austin, TX. Tesniere, L. (1959). Elements de syntaxe structural [Elements of structural syntax]. Paris: Klincksieck. Thepkanjana, K. (1986). Serial verb construction in Thai. Unpublished PhD. dissertation, University of Michigan. Thorell, O. (1973). Svenskgrammatik [Swedish grammar]. Stockholm: Esselte Studium. t>rainsson, H. (1990). Setningafrcedi [Syntax]. Reykjavik: Malvisindastofnun Haskola Islands.

568

References

Toivainen, J. (1989). Situationally determined linguistic differences between narratives produced by a five-year-old Finnish child: Demonstratives and tenses in face-toface and telephone narratives. Paper presented at the 10th Biennial Meeting of the International Society for the Study of Behavioural Development, Jyvaskyla, Finland. Toivainen, J. (1992). Vieri- ja puhelinkerronnan erot erddn 5-vuotiaan kuvakirjakertomuksessa: demonstratiivit ja tempukset. [Face-to-face vs. telephone narration in a 5-years-old's speech: Finnish demonstratives and tenses]. Virittajd 96,60-71. Tolchinsky, L., Johansson, V. & Zamora, A. (2002). Text openings and closings in writing and speech: Autonomy and differentiation. Written Language and Literacy 5, 219-254. Tolchinsky, L. & Rosado, E. (in press). Effect of age, modality, and text type on use of middle, passive voice and impersonal constructions for agency alternation in Spanish. Journal of Pragmatics (special issue). Tomlin (Ed.) (1987). Coherence and grounding in discourse. Amsterdam: Benjamins. Tosi, A. (1998). The notion of community in language maintenance. In G. Extra & L. Verhoeven (Eds.), Bilingualism and migration (pp. 325-344). Berlin: Mouton de Gruyter. Trabasso, T. & Nickels, M. (1992). The development of goal plans of action in the narration of a picture story. Discourse Processes 15, 249-611. Trabasso, T. & Rodkin, P. C. (1994). Knowledge of goals/plans: A conceptual basis for narrating Frog, where are you? In R. A. Berman & D. I. Slobin (Eds.), Relating events in narrative: A crosslinguistic developmental study (pp. 85-106). Hillsdale. NJ: Lawrence Erlbaum Associates. Trabasso, T., Stein, N. L., Rodkin, P. C., Munger, M. P., & Baughn, C. (1992). Knowledge of goals and plans in the on-line narration of events. Cognitive Development 1', 133-170. Trask, R. L. (1997). The history of Basque. London and New York: Routledge. Trondhjem, F. Blytmann. (2000). Gronlandske mundtlige og skriftligefrohistorier: En sammenligning afligheder ogforskelle udfra 3 grupper mundtlige og 3 grupper skriftlige fortcellinger [Greenlandic oral and written frog stories: A comparison of similarities and differences based on 3 groups of oral and 3 groups of written stories]. Master's thesis, Department of Greenlandic Language and Culture, Ilisimatusarfik, Nuuk. Tsitsipis, L. D. (1983). Narrative performance in a dying language: Evidence from Albanian in Greece. Word'34, 25-36. Tyler, L. (1984). Integration of information during language comprehension: A developmental study. Papers and Reports on Child Language Development 23, 125-133. Uchida, N. (1982). Yooji wa ika ni monogatari o tsukuru ka? [How do children make narratives?]. Kyooiku shinrigaku kenkyuu 30(3), 212-222.

References

569

Uchida, N. (1990). Kodomo no bunshoo: kaku koto, kangaeru koto [Children's narratives: writing and thinking]. Tokyo: Tokyo University Press. Umiker-Sebeok, D. J. (1979). Preschool children's interconversational narratives. Journal of Child Language 6, 91-109. Ure, J. N. (1971). Lexical density and register variation. In G. E. Perren & J. L. M. Trim (Eds.), Applications of linguistics: Selected papers of the 2nd International Congress of Applied Linguistics (pp. 443-452). Cambridge: Cambridge University Press. Urrutia Cardenas, H. (1992). Competencia lexica y sintactica en ninos monolingiies y bilingiies en el Pais Vasco [Lexical and syntactic competence in monolingual and bilingual children in the Basque Country]. In H. Urrutia Cardenas & C. SilvaCorvalan (Eds.), Bilinguismo y adquisicion del espanol (pp. 17-26). Bilbao, Spain: Institute Horizonte. van der Lely, H. K. J., (1997). Narrative discourse in grammatical specific language impaired children: A modular language deficit? Journal of Child Language 24, 221-256. van Dijk, R. T. (1980). Semantic macrostructures. New York, NY: Lawrence Erlbaum Associates. van Hoof, A. M. T. G. (2000). Gesture placement in motion events: A crosslinguistic study of LI and L2 'thinking for speaking'patterns. M. A. thesis, Katholieke Universitet, Nijmegen, The Netherlands. van Hoof, A. M. T. G. (2001). Als gebaren spreken [Speaking gestures]. Toegepaste Taalwetenschap in Artikelen 67(1). Verhoeven, L. (1989). Acquisition of clause linking in Turkish. In H. Bennis & A. van Kemenade (Eds.), Linguistics in the Netherlands 1989 (pp. 153-162). Dordrecht: Foris. Verhoeven, L. (1990a). Acquisition of discourse cohesion in Turkish. In S. Koc (Ed.), Proceedings of the 4th Conference on Turkish Linguistics. Ankara: Orta dogu. Verhoeven, L. (1990b). Temporality in Turkish narratives: A developmental perspective. Paper presented at the 5th Conference on Turkish Linguistics, London, August 1990. Verhoeven, L. (199 la). Acquisition of Turkish in a mono- and bilingual context. In H. Boeschoten & L. Verhoeven (Eds.), Turkish linguistics today (pp. 113-149). Leiden: Brill. Verhoeven, L. (1991b). Predicting minority children's bilingual proficiency: Child, family and institutional factors. Language Learning 41, 205-233. Verhoeven, L. (1993). Acquisition of narrative skills in a bilingual context. In B. Kettemann & W. Wieden (Eds.), Current issues in European second language research (pp. 307-323). Tubingen: G. Narr. Verhoeven, L. (1994). Transfer in bilingual development: The linguistic interdependency hypothesis revisited. Language Learning 44, 381-415.

570

References

Verhoeven, L. (1996a). Literacy outside the mainstream. InternationalJournal of the Sociology of Language 119, 1-12. Verhoeven, L. (1996b). Early bilingualism, cognition, and assessment. In M. Milanovic & N. Saville (Eds.), Performance testing, cognition and assessment (pp. 276-291). Cambridge: Cambridge University Press. Verhoeven, L. (1997). Acquisition of literacy by immigrant children. In C. Pontecorvo (Ed.), Writing development (pp. 219-240). Amsterdam: John Benjamins. Verhoeven, L. & Boeschoten, H. (1986). First language acquisition in a second language environment. Journal of Applied Psycholinguistics 7, 241-256. Verhoeven, L. & Narain, G. (1996). Ontwikkeling van tweetaligheid bij allochtone kleuters [Bilingual development among minority children in kindergarten]. Tijdschrift voor onderwijsresearch 21, 54—80. Verhoeven, L. & Strb'mqvist, S. (2001). Development of narrative production in a multilingual context. In L. Verhoeven & S. Stromqvist (Eds.), Narrative development in a multilingual context (pp. 1-13). Amsterdam/Philadelphia: John Benjamins. Vermeulen, R. & Kellerman, E. (1999). Causation in narrative: The role of language background and proficiency in two episodes of the 'frog story'. In Albrechtsen, D., Henriksen, B., Mees, I. & Poulsen, E. (Eds.), Perspectives on foreign and second language pedagogy (pp. 161-176). Odense, DK: Odense University Press. Viberg, A. (1993). Crosslinguistic perspectives on lexical organization and lexical progression. In K. Hyltenstam & A. Viberg (Eds.), Progression and regression in language (pp. 340-385). Cambridge: Cambridge University Press. Viberg, A. (1998). Lexical development in a second language. In G. Extra & L. Verhoeven, Bilingualism and migration (pp. 165-186). Berlin: Mouton de Gruyter. Viberg, A. (2001). Age-related and L2-related features in bilingual narrative development. In L. Verhoeven & S. Stromqvist (Eds.), Narrative development in a multilingual context (pp. 233-254). Amsterdam/Philadelphia: John Benjamins. Voeltz, F. K. E. & Kilian-Hatz, C. (Eds.) (2001). Ideophones. Amsterdam/Philadelphia: John Benjamins. Walchli, B. (2001). A typology of displacement. Sprachtypol. Univ. Forsch. (STUF) 54(3), 298-323.' Walton, K. L. (1990). Mimesis as make-believe: On the foundations of the representational arts. Cambridge, MA: Harvard University Press. Warden, D. (1976). The influence of context on children's uses of identifying expressions and reference. British Journal of Psychology 67, 101-112. Warden, D. (1981). Learning to identify referents. British Journal of Psychology 72, 93-99.

References

571

Weiss, A. (2002). Zur Versprachlichung des Raums in Bildergeschichten deutschsprachiger Vor- und Grundschulkinder [On the verbalization of space in German-speaking pre-school and school-age children's picture stories]. Arbeitspapier Nr. 43 (Neue Folge), Institut fur Sprachwissenschaft, Universitat zu Koln. Weissenborn, J. (1986). Learning how to become an interlocutor: The verbal negotiation of common frames of reference and actions in dyads of 7-14 year-old children. In J. Cook-Gumperz, W. Corsaro & J. Streek (Eds.), Children's world and children's language (pp. 377-404). Berlin: Mouton de Gruyter. Weissenborn, J. (1992). Null subjects in early grammars: Implications for parameter setting. In J. Weissenborn, H. Goodluck & T. Roeper (Eds.), Theoretical issues in language acquisition: Continuity and change in development (pp. 269-299). Hillsdale, NJ: Lawrence Erlbaum Associates. Weist, R. M. (1986). Tense and aspect. In P. Fletcher & M. Garman (Eds.), Language acquisition: Studies in first language development (2 nd ed., pp. 356—374). Cambridge: Cambridge University Press. Wellhousen, K. (1993). Eliciting and examining young children's storytelling. Journal of Research in Childhood Education 7, 62-66. Wellman, H, M., Phillips, A. T. & Rodriguez, T. (2000). Young children's understanding of perception, desire and emotion. Child Development 71 (4), 895-912. Wellman, H. M. & Sommerville, S. C. (1980). Quasi-naturalistic tasks in the study of cognition: The memory-related skills of toddlers. In M. Perlmutter (Ed.), Children's memory (pp. 33-48). San Francisco: Jossey-Bass. Wengelin, A. (2002). Textwriting in adults with reading and writing difficulties. PhD. dissertation. Goteborg University: Department of Linguistics. Wengelin, A. & Stromqvist, S. (2000). Discourse level writing in dyslexics - methods, results, and implications for diagnosis. Log Phon Vocol 25, 22-28. Westermann, D. (1930). A study of the Ewe language. London: Oxford University Press. Whorf, B. L. (1956) [1940]. Language, thought, and reality: Selected writings of Benjamin Lee Whorf. Ed. by J. B. Carroll. New York: Technology Press of MIT and John Wiley & Sons. Wide, C. (manuscript). Vera buinn ad + inf. in dialogue: Form and functional potential of a grammatical construction in contemporary Icelandic. Manuscript (2002). University of Helsinki. Wienold, G. (1995). Lexical and conceptual structures in expression for movement and space. In U. Egli, P. Pause, C. Schwarze, A. Stechow & G. Wienold (Eds.), Lexical knowledge in the organization of language. Amsterdam: Benjamins Press. Wienold, G., Dehnhardt, A., Kim, C-D. & Yoshida, M. (1991). Lexikalische und syntaktische Strukturenjapanischer und koreanischer Bewegungsverben [Lexical and syntactic structures of Japanese and Korean motion verbs]. Working paper no. 29, Fachgruppe Sprachwissenschaft der Universitat Konstanz, Germany.

572

References

Wienold, G. & Schwarze, C. (1989). Lexical structure and the description of motion events in Japanese, Korean, Italian and French. Working paper no. 5, Fachgruppe Sprachwissenschaft der Universitat Konstanz, Germany. Wigglesworth, G. (1990). Children's narrative acquisition: A study of some aspects of reference and anaphora. First Language 10, 105-125. Wigglesworth, G. (1992). Investigating cognitive and linguistic development through narrative. Doctoral disseration, La Trobe University. Wigglesworth, G. (1997). Children's individual approaches to the organization of narrative. Journal of Child Language 24, 279-310. Wilkins, D. P. (1989). Mparntwe Arrernte (Aranda): Studies in the structure and semantics of grammar. Doctoral dissertation, Australian National University. Wilkins, D. P. (1991). The semantics, pragmatics and diachronic development of 'associated motion' in Mpamtwe Arrernte. Buffalo papers in linguistics, no. 1 (pp. 207-257). Wilkins, D. P. (1997). The verbalization of motion events in Arrernte (Central Australia). In E. V. Clark (Ed.), Proceedings of the 28th Annual Child Language Research Forum (pp. 295-308). Stanford: The Center for the Study of Language and Information. Wilkins, D. P. (this volume). The verbalization of motion events in Arrernte. In S. Stromqvist & L. Verhoeven (Eds.) (2004), Relating events in narrative: Typological and contextual perspectives. Mahwah, NJ: Lawrence Erlbaum Associates. Wolf, D. P., Hemphill, L. & Camp, L. (1991). A contrastive analysis of narrative performance in a child with mild mental retardation and a nonretarded child. Paper presented at the Gatlinburg Conference on Language Disorders, Key Biscayne, FL. Wolf, D. P., Moreton, J. & Camp, L. (1994). Children's acquisition of different kinds of narrative discourse: Genres and lines of talk. In J. Sokolov & C. E. Snow (Eds.), Handbook of research in language development using CHILDES (pp. 286-323). Hillsdale, NJ: Lawrence Erlbaum Associates. Ying, C. (1988). Deutsche und chinesische Bewegungsverben: Bin sprachdidaktischer Vergleich ihrer Semantik und Valenz [German and Chinese motion verbs: A comparison of semantics and valence]. Berlin/New York: de Gruyter. Zijlenmaker, C. (1992). Het gebruik van referentie in het vertelten van een verhaal "The frog story " bij kinderen en volwassenen [The use of reference in telling a story "The frog story" by children and adults]. Unpublished master's thesis, General Linguistics, University of Amsterdam. Zijlenmaker, C. (1993). Anafoorgebruik van kinderen in een narratieve setting [Use of anaphora by children in a narrative setting]. Unpublished doctoral manuscript, Institute of General Linguistics, University of Amsterdam. Ziv, Y. (1988). Word order in children's literature: FSP and markedness. In Y. Tobin (Ed.), The Prague School and its legacy (pp. 123-144). Amsterdam: John Benjamins.

References

573

Zlatev, J. (1997). Situated embodiment. Studies in the emergence of spatial meaning. PhD. thesis, Department of Linguistics, Stockholm University. Stockholm: Gotab Press. Zlatev, J. (in press-a) Spatial prepositions: polysemous or general? In H. Cuyckens & D. Sandra (Eds.), Cognitive linguistic approaches to lexical semantics. Berlin: Mouton de Gruyter. Zlatev, J. (in press-b) Holistic spatial semantics of Thai. In E. Casad & G. Palmer (Eds.), Cognitive linguistics and non-Indo-European Languages. Berlin: Mouton de Gruyter. Zlatev, J. & Yangklang, P. (this volume). A third way to travel: The place of Thai in motion event typology. In S. Stromqvist & L. Verhoeven (Eds.) (2004), Relating events in narrative: Typological and contextual perspectives. Mahwah, NJ: Lawrence Erlbaum Associates.


Subject Index

The subject index reflects the broad range of disciplines and expertise brought together in this volume: cognitive linguistics, psycholinguistics, developmental psychology, language typology, and narratology. A guiding principle has been to detail keywords and terms referring to functional and semantic distinctions to a greater degree than terms referring to syntactic and morphological distinctions. Detailed descriptions of morphological categories and syntactic forms used for the linguistic retelling of the frog story in different languages can be found in the individual chapters. See also Appendix 3.

Aboriginal mythology 18 absolutely-oriented gestures. See gesture(s) abstract [story grammars]. See story grammars access, accessed, accessing 11, 19, 138, 197,238,252,384,428,430 accessible, accessibility 7, 9, 38, 56,138, 139, 223, 227, 237, 238, 245, 250, 251, 252, 253, 284, 364, 368, 406, 430,475 Acholi 492, 515 adverb(s) 49, 84, 93, 102, 103. See also misrepresentations adverbial(s) 40, 71, 72, 73, 74, 76, 80, 82, 93,94, 104, 107, 108, 109 affect. See manner affective (experiences). See manner affective (expressions). See manner affective (states). See manner affect verb affixes. See verb Afro-Asiatic 515 age groups 3 yrs 307-328 3-5 yrs 37-58 4 yrs 159-190, 307-328, 329-358

4-5 yrs 17-36,395^34 5 yrs 59-88, 113-142, 307-328, 329-358 6 yrs 159-190 6-7 yrs 17-36,37-58 7 yrs 307-328,329-358 7-8 yrs 395^34 7-10 yrs 143-158 8 yrs 17-36 8-13 yrs 37-58 9yrs 113-142,159-190,307-328, 329-358, 359-394 9-10 yrs 17-36,261-280 10 yrs 59-88 11 yrs 159-190 11-12 yrs 395-434 12 yrs 359-394 12-13 yrs 261-280 15 yrs 113-142,359-394 adults 17-36,37-58,59-88,89-112, 113-142, 143-158, 159-190, 191218, 261-280, 307-328, 329-358, 359-394, 395-434 agglutinative. See typology aggressive boys 513 Akan 160, 168, 188 Algonquian 247, 249 Almosan-Keresiouan 517

Subject Index

576

Altaic 516 Alur 492,515 Alyawarra 488, 517 American English. See English American Sign Language [ASL] 12, 797277, 236, 243, 245, 246, 254, 365, 461, 466, 468, 473, 487, 504, 517 Amerind 517 analytic(al), analytically 308, 333, 433, 440,474 anaphor(s), anaphora, anaphoric 205, 291, 301,440 anaphoric demonstratives 69, 71, 72, 76 anchor tense 120, 121. See also tense Andean 517 animal quotations. See quotations annotative aspect. See demonstration(s) aphasia, aphasic. See impairment Arabic 398, 444, 487, 496, 507, 508, 515 Aranda. See Arrernte Arandic 488 areal cultural factors 156, 243 Arrernte 12, 57, 85, 88, 143-157, 242, 243, 254, 256, 355, 357, 461, 488, 517 ASL. See American Sign Language [ASL] aspectual 113, 116, 247, 284, 286, 287, 291, 298, 299, 331, 349, 396, 407, 408,409,411,412,413,414,428, 432, 433, 442 aspect [grammatical] 8, 10, 11, 103, 113, 120, 121, 122, 124, 137,209, 257, 276, 277, 290, 327, 329, 406, 407-412, 413, 414, 415, 425, 428, 429, 430, 431, 436, 439, 441, 442 associated motion 57, 85, 147, 150, 153, 156,242,243 atelic, atelicity 410,430 Athabaskan 247,249 Atsugewi 190, 192

attempt [story grammars]. See story grammars attention 7, 11, 17,49, 55, 72, 103, 144, 155, 157,211,213,279-255,285, 301, 304, 338, 339, 362, 410, 431, 453 channeling of attention 9,37,38,41, 149, 156,437,452 focus of attention 471 narrative attention 48, 92, 98, 100, 109, 145, 152 visual attention 390 attitudinal stance. See stance attrition 437, 438, 452, 488, 508 Australian (Aboriginal) languages, 143, 146, 156,243,247,336,517 Australian Aboriginal English. See English Australian desert languages. See Australian (Aboriginal) languages Australian English. See English Austric 516 Austronesian 5, 160, 249, 491, 492, 516 autistic, autism. See impairment

B back-channelling 356 background 6, 10, 38, 41, 54, 122, 196, 199, 242, 266, 268, 269, 270, 271, 284, 285, 286, 302, 306, 333, 356, 404, 407, 408, 411, 436, 444, 453, 459 backgrounding 5, 9, 125, 139, 257, 364, 410,411 backtracking 123, 124, 125. See also retrospective function Balto-Slavic 515 Bantu 159 Baragaunle 510, 516 bare motion verbs. See verb Basque 12, 59-777, 249, 254, 466, 469, 496,504,507,517

Subject Index

beginning [story grammars]. See story grammars beliefs 307,353,355 bilingual, bilingualism. See second 1 anguage/b i 1 i ngual i sm bilingual society. See second language/ bilingualism binary typology. See typology bipartite typology. See typology bipartite verbs. See verb Black English. See English both hands. See hand(s) bottom-up 7. See also top-down boundary-crossing(s) 61, 70, 71, 168, 182, 206, 207, 208, 209, 210, 211, 212, 225, 226, 228, 229, 239, 251, 257, 298 boundary-crossing constraint 61, 168, 170, 183, 186, 187,225-226,236, 243, 253 boundary-crossing events 186 boundary focus 61 brain injury. See impairment British English. See English British Sign Language 504, 517 Bulgarian 488, 515

Caddo 159 Cantonese 502, 516 case marker(s) 21,27,28,241,256 case marking 111,118,119,147,245, 250, 426, 436, 439 causal connectors. See evaluative devices cause-of-motion verb(s). See verb Celtic 515 chaining structures 10 chains of relations. See relations change of direction(al). See direction(s) change of location. See location(s)

577

change of state 226,421 change-of-state verbs. See verb channeling of attention. See attention Chantyal 510,516 characteristic expression of motion. See motion character speech. See evaluative devices Cheremis 490 Chinese 160, 176, 180, 192, 255, 293, 294, 295, 297, 299, 303, 306. See also Mandarin [Chinese] classifier(s) 193, 194, 196, 199, 201, 202, 203,204,205,210,217,474 clause-linking 10,279,437,439 clause package, clause packaging 277, 437. See also packaging cline of grammaticalization 113, 180 cline of manner salience. See manner co-event 90, 162, 163, 187 co-ranking structure(s) 10,439 co-reference. See reference codability 237. See also readily encodable coda [story grammars]. See story grammars coherence 9, 50, 54, 214, 281-306, 311, 317,332,462 cohesion 9, 83, 84, 281-306, 388, 407, 441,442,444,445,452 cohesive(ly) 123,312,395,396,418 collocations 377, 379, 388 common ground 461,462,472 communicative conditions 13, 359, 368 community expertise 473 comparators [internal evaluations]. See internal evaluations complete-path hypothesis 89, 97, 241 completed action 113-116 complex paths 51, 61, 76, 192, 208, 211, 212,240 complex verb-framed (languages). See typology

578

complicating action(s) [story grammars]. See story grammars computer-logged narratives. See narrative conception versus description 469—475 conflate, conflation(s), conflation pattem(s) 91, 93, ] 76-180, 190, 205, 230-231, 235,250 conflict(s) 271,274,275,278,279 confluence of cues 278 conformation. See motion congenitally deaf 359, 369, 386-390, 392, 466,513 conjoined clauses 149 connectivity 9, 10, 270, 379, 393, 402^06, 412, 413, 415, 416, 418, 430,431,432,436 constellations of (spatial) morphemes 134, 135, 136,240 constraints 363 contact situations 251,433 context 261-280,281-306 contextual variation(s), contextual variability 11,273,281,293,299, 302 conventional sound symbolism. See sound symbolism converbs. See verb core-grammar errors. See errors core schema. See motion corporeal sound symbolism. See sound symbolism correlatives [internal evaluations]. See internal evaluations creating a story versus retelling a story. See narrative Creoles 517 cultural practices 219,220,223,326 cultural styles 334

D Daic 516

Subject Index

Danish 62 decoupling principle 464 defective-tense hypothesis 286, 299 deictic 71, 72, 75, 85, 160, 163, 165, 168, 227, 229, 247, 248, 250, 253, 255, 257, 269, 285, 386, 403, 408, 416, 418, 441. See also motion Deictic Center 165 deictic path verb(s). See verb deixis 177, 179, 189, 228, 229, 269, 272, 273 demonstration(s) 463-469, 473, 475 annotative aspects 464 depictive aspects 464, 467 incidental aspects 464 supportive aspects 464, 467 depictive aspect. See demonstration(s) deplacement 189 development 3-14,261-280,281-306 developmental hypothesis 145,156 later development 275-277, 393 U-shape(d) 48, 49, 50, 273 developmental errors. See errors developmental hypothesis. See development developmental impairment. See impairment Dhopodhola 492,515 diachronic consequences 251 diegesis 462—469 direction(s) 10, 17, 19, 21, 22, 27, 30, 34, 35, 40, 45, 46, 47, 48, 56, 70, 71, 74, 85,97, 113, 117, 119, 138, 144, 164, 167, 198, 245, 246, 256, 297, 463, 464,469,470,471,474 change of direction(al) 182,253 directional(s) 27, 28, 29, 31, 39, 42, 45, 47, 48, 49, 52, 53, 54, 55, 56, 57, 160, 242, 247 directional adverbs 242,248,251 directional affix(es) 21,27,28,29 directional morphemes 233, 242 directional prepositions 245

Subject Index

directional satellites 95 directional terminology 144 directional(s). See direction(s) directional adverbs. See direction(s) directional affixes. See direction(s) directional morphemes. See direction(s) directional prepositions. See direction(s) directional satellites. See direction(s) direct speech. See projected speech discourse anchoring 284 discourse frame 91, 109 discourse-organizing processes 288 discourse stance. See stance discourse topic(ality) 288, 294, 311, 321 displace reference. See reference distributed spatial semantics 188, 248, 250 Djamindjung. See Jaminjung Djamindjungan. See Jaminjungan Down syndrome. See impairment Dravidian 494, 516 duration 80,88,360,361,390,391 Dutch 182, 223, 224, 225, 227, 229, 240, 244, 248, 251, 254, 256, 264, 274, 277, 295, 303, 306, 436, 437, 438, 439, 440, 441, 442, 443, 444, 445, 446, 447, 448, 449, 465, 488, 496, 497,499,504,507,510,515 dyslexia, dyslexic. See impairment dysphasia, dysphasic, dysphatic. See impairment

E ease of processing 223, 229, 230, 231, 234,237,241,245 editing [in writing]. See writing effort(s) 189, 198, 212, 332, 359, 379, 394,452 Elamo-Dravidian 516

579

elicitation(s) 11, 13, 56, 92, 261, 262-274, 271, 278, 279, 331, 390, 398, 445 elicitation procedures 274 elicitation settings 262 Emai 235,255 emblematic gestures. See gesture(s) embodied cognition 470, 474-475 emotion(s) 204, 307, 333, 336, 357 emotional stance. See stance emotive stance. See stance ending [story grammars]. See story grammars endophora 4 English vii, 4, 10, 12, 23, 34, 37, 38, 51, 57, 60, 61, 62, 70-80, 85, 90, 91, 95, 98,99, 100, 104, 105, 106, 114, 115, 116, 117,118, 119, 132,133, 138, 143, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 161, 167, 168, 177, 182, 183, 184, 187, 189, 190, 191, 192, 194, 195, 197, 198, 199, 200, 201, 202, 203, 204, 206, 212, 213, 220, 221, 222, 223, 225, 227, 229, 230, 231, 232, 233, 235, 236, 237, 239, 240, 241, 243, 244, 245, 246, 249, 254, 255, 256, 257, 264, 270, 271, 274, 276, 277, 279, 284, 292, 293, 294, 295, 296, 297, 298, 299, 303, 306, 307-327, 330,333,336,350,351,353,354, 355, 363, 364, 365, 398, 429, 433, 445, 446, 459, 461, 462, 466, 467, 469, 470, 471, 472, 473, 474, 487, 496, 497, 502, 504, 507, 508, 510, 513,515 American English 111, 146, 221, 222, 332, 365, 488-500, 497, 498 Australian Aboriginal English 496 Australian English 232, 489, 497 Black English 262 British English 489 enrichment expressions 338, 341, 343, 344, 347, 349, 350 episodic boundary 300

580

epistemic modality 80 epistemic stance. See stance equipollently-framed (languages). See typology ergative 32, 86 errors core-grammar errors 426 developmental errors 397, 426 error analysis 426-429 fossilized errors 397 grammatical errors 433, 439, 443 lexical errors 428 spelling errors 372, 374. See also writing: spelling transfer errors 397, 428 error analysis. See errors Erzya-Mordvinian 502 Eskimo-Aleut 516 Estonian 490,516 ethnographic 143, 145, 155 ethnolinguistic 143, 145 evaluation(s) [story grammars]. See story grammars evaluative [functions] 10,307-327, 329-358 evaluative devices 330-358 causal connectors 326, 332, 338, 341, 342,343,347,348,351,353,355, 394 character speech 332, 338, 339, 341, 342, 345, 346, 347, 348, 350, 354, 393 frames of mind 326,332,333,337, 340, 341, 345, 346, 347, 350, 352, 353,354 hedges 327, 332, 337, 340, 341, 346, 347, 350 negative qualifiers 332, 338, 341, 345, 346, 347, 348 evaluative factor. See manner event component(s) 8, 37, 88, 755, 245. See also path: path components

Subject Index

event conflation 9. See also conflate, conflation(s), conflation pattern(s) event granularity 187, 239 event time. See time evidential(s), evidentiality 115, 308, 310, 326, 327, 350 Ewe 160, 188, 235, 251, 255, 507, 515 existential-relative structures 301 exophora 4 expletives [internal evaluations]. See internal evaluations expository 274-279,300 expressive features (distribution of) 361, 363, 384, 385, 393 external evaluations 331. See also story grammars: evaluation(s) eye-tracking 390, 394

face-to-face interaction 360, 361, 363, 392 facial expression(s) 193, 195, 196, 325, 333, 366, 382 factual, factuality 333, 335, 442, 458-459, 460, 475 factual versus fictional 458 fairytales 300 false beliefs. See misrepresentations familial SLI. See impairment familial Specific Language Impairment. See impairment familiar event sequences 267, 287 feedback 356, 360, 361 fiction, fictional 5, 14, 229, 271, 458-459, 460, 462, 464, 465, 468, 475 figure. See motion figure-type. See typology Fijian 504,516 film(s) 262, 266, 267, 268, 272, 475 film retellings 335 filtering 9, 11, 12, 126,365,452

581

Subject Index

finalistic model of communication 362 final edited version. See writing fingerspelling 204, 205 Finnish 489, 490, 496, 502, 516 Finno-Permic 490 Finno-Ugric 160,249,516 first language [acquisition] 487^95, 502-513 first-person narratives. See narrative first- versus third-person stories 459 fixation(s) 390,391,392 fluent 365 focus of attention. See attention force dynamics 226, 232, 236 foreground 6, 10, 38, 284, 285, 286, 289, 295, 302, 306, 404, 407, 408, 411, 436, 444 foregrounding 5, 9, 125, 257, 364, 444 formal [genre/register/style] 115, 125, 129, 137, 300, 301, 326, 339, 345, 350,406,413 fossilized errors. See errors frames of mind. See evaluative devices free direct speech. See projected speech French 94, 163, 177, 182, 189, 207, 223, 224, 225, 240, 244, 251, 252, 254, 264, 270, 274, 277, 294, 295, 297, 298, 299, 300, 301, 302, 303, 304, 306, 356, 398, 445, 490, 497, 499, 502,504,507,515 frequency [semantic category] 80 frequency [usage] 27, 34, 125, 127, 128, 130, 132, 137, 138, 179,204,237, 252,311,313,332,345,347,351, 377, 440 frog story 4-8, 479-486. See also narrative functionally disabled groups 513

gaze 193, 195,392 generic characters 357 genre(s) 4, 11, 13,41, 115, 137, 179,254, 262-280, 299, 299-302, 327, 350, 351,358,436 German vii, 4, 10, 12,34, 118, 119, 182, 189, 220, 223, 224, 225, 227, 229, 239, 240, 244, 246, 251, 254, 294, 295, 296, 297, 299, 303, 306, 325, 326, 398, 443, 487, 496, 502, 515 Germanic 20,21,34, 117, 163, 182, 185, 225, 227, 229, 231, 240, 249, 255, 515 gerunds 253,406,412,413,414,415, 430,431 gesture(s) 13, 97, 104, 106, 107, 108, 111, 193, 197,212,214,220,232, 235-236, 237, 255, 333, 342, 362, 363,366,384,441,457,469 absolutely-oriented gestures 157 emblematic gestures 463 iconic gestures 13, 89, 468, 469, 475 manner fogs 105 manner gesture(s) 105, 213, 235, 255 path fogs 108 path gesture(s) 213 Ghale 510,516 global cohesion 123 goal. See motion grammar typology. See typology grammatical errors. See errors Greek 496,502,515 Greenlandic 516. See also WestGreenlandic ground(s). See motion grounding of) information. See information grounding Guugu Yimidhirr. See Guugu Yumithirr Guugu Yumithirr 490, 496, 517

H Galician 490,515

habitual expression(s) 237, 249, 250

582

Subject Index

hand(s) [as articulators] 65, 106, 107, 108, 109, 193, 195, 196, 205, 217, 235, 236, 237, 243, 463, 464 both hands 105, 106, 108, 111, 195, 196,208,210,215,216 handshape 196,210,217,236 left hand 108,111,195,196,215,216, 217

right hand 106, 108, 109, 111, 195, 196, 215,216,217 handshape. See hand(s) handwriting. See writing Hebrew vii, 4, 10, 12, 34, 181, 182, 220, 223, 224, 225, 229, 232, 240, 246, 251, 262, 264, 266, 267, 268, 271, 274, 277, 279, 280, 296, 301, 303, 332, 395, 396, 397, 398, 402, 403, 405, 406, 408, 411, 412, 413, 414, 415, 417, 419, 422, 423, 424, 426, 428, 429, 430, 432, 433, 434, 446, 487,499,508,515 hedges. See evaluative devices hesitant phase 365 hesitation markers 362 hierarchical organization 288, 437 Hmong-Mien 160,249 Hokan 192,247,249 holistic, holistically 474, 475 holistic spatial semantics 188 hunter-gatherers, hunting-gathering society 19,85,246 I

Icelandic 12, 113-141, 184, 240, 245, 254, 274, 278, 461, 466, 490, 515 icon(s), iconic, iconicity. See mimesis iconic gestures. See gesture(s) ideophone(s) 232, 233-235, 237, 250, 251,467 Ilocano 235, 255 imagining-for-speaking. See thinking-forspeaking

imitative sound symbolism. See sound symbolism impairment 498-499,508-513 aphasia, aphasic 499, 513 autistic, autism 498,510 brain injury 498 developmental impairment 489 Down syndrome 498, 499 dyslexic 513 dysphasic 499 familial Specific Language Impairment [SLI] 498 language delay 498 mentally retarded 498, 499 pragmatic language impairment 510 psychiatric disturbance 499 reading disability 498, 499 Specific Language Impairment [SLI] 498,499,510 Williams syndrome 498, 499, 510 impersonal constructions 276. 277, 278 inaccessible 402 incidental aspect. See demonstration(s) independent clauses 415,417,439 index [mimesis]. See mimesis indicating [mimesis]. See mimesis indirect speech. See projected speech Indo-European 433 Indo-Hittite 515 Indo-Pacific 516 informal 115,345 information grounding, grounding of) information 284, 285, 288, 289 inner state 94, 109,232 input 118, 266, 273, 435, 436, 437, 438, 440, 442, 443, 447. 448 intensifies [internal evaluations]. See internal evaluations intentions 6, 137, 329, 351, 464 interactive storytelling. See narrative interactivity 361

Subject Index

internal evaluations 331 comparators 331,332 correlatives 331 expletives 331 intensifies 331,332,338,349 intonation, intonational 13, 101, 174, 201, 338, 363, 393, 463, 464, 469 Inuit 516 Irish 508,515 Italian 223, 224, 225, 226, 248, 251, 254, 491,499,502,507,515 Italic 515

Jaminjung 247,510,517 Jaminjungan 249 Japanese 12,111,159, 207, 234, 235, 244, 249, 255, 325, 326, 329-358, 461, 466, 467, 468, 469, 473, 474, 491, 502,507,508,516 Japanese-Ryukyuan 516 joint construction. See narrative joint pretense 464 journey(s) 5, 20, 21, 26, 30, 34, 37, 38, 39, 40, 41, 44, 53, 54, 55, 76, 77, 82, 84,88,91, 101, 111, 144, 146, 147, 149, 150, 151, 157 journey from the jar 147-150 journey off the cliff 150-155 journey segments 152,154

K Kgalagadi 491,515 Kickapoo 491,517 Kilivila 491,513,516 Klamath-Takelman 247, 249 Komi 490,516 Korean 207, 249, 250, 253, 502, 516

583

Lahu 160, 180 Lakhota 492,517 landmark(s). See motion landscape of action 307, 315, 316, 324, 326, 330 landscape of consciousness 307, 308, 326, 330 Lango 492,515 language-choice 437 language delay. See impairment language deprivation 504 language group(s). See typology language isolates 517 language transfer 436 later development. See development Latin 189 left hand. See hand(s) lexical density 275, 276, 277, 393 lexical diversity 252, 275, 277 lexical errors. See errors lexicography 510 linear chain(ing) [of events] 9, 264, 266, 406,452 linear distribution [of information] 288, 361,363,384,392,393 linear structural relations [LISREL] 447 linear text [on-line] 370, 371, 373, 387 linguistic relativity 8, 12, 59, 89, 91, 159 linguistic typology. See typology LISREL. See linear structural relations literacy 19, 56, 276, 397, 446, 448-451, 454,508 literary [style] 177, 420, 423 literary narratives 271 literary traditions 244 location(s) 17, 18, 19, 20-23, 26, 28, 29, 34, 38, 85, 97, 98, 191, 193, 196, 198, 216, 221, 222, 234, 241, 243, 244,

584

Subject Index

246, 255, 289, 294, 295, 296, 297, 302,304,321,419,459,472 change of location 30, 34, 89, 185, 220 static location 254 stationary location 89, 90, 162 locational case(s) 95, 100, 241 locative 77-35, 37, 38, 40, 54, 60, 71, 72, 73, 74, 76, 80, 82, 85, 89, 95, 96, 162, 167, 175, 222, 240, 242, 244, 245, 256, 296, 303, 442, 472 locative case(s) 21, 27, 28, 29, 30, 70, 86, 88, 106, 184, 188,242 Longgu 492,516 loudness 463, 464

M Magar 510,516 maintain(ing) reference, reference maintenance. See reference make-believe 5, 7, 265, 272 Malay 492,516 Mandarin 212, 223, 224, 225, 227, 228, 235, 251, 254, 255, 306, 336, 470, 492,502,516 manner affect 110, 195, 197-204, 217, 236, 254,329-358 affective experiences 449 affective expressions 340, 346 affective states 199,200,332,337 cline of manner salience 220, 228, 250 evaluative factor 254 motor pattern(s) 93, 109, 110, 111,226, 254 posture(s) 21, 125, 140, 179, 195,232, 233, 236, 254 rate 94, 107, 109, 110, 111,232,236, 254

rhythm 13,61,254,467,468 manner fogs. See gesture(s) manner gesture(s). See gestures manner of speaking. See quotations manner + path verb(s). See verb

manner preverb 250 manner-type. See typology manner verb(s), manner-of-motion verb(s). See verb manner verbs versus path verbs 181 Mari 490,516 Mayan 38, 56, 159, 232, 494, 495, 504 Medieval ages 113 medium 11, 274, 278, 306, 325. See also modality; speech; writing; sign language; gestures; mimetics; pictures medium(s) [air, water] 103, 167, 244 mentally retarded. See impairment mental state(s) 13, 307, 308, 313,315, 318, 324, 325, 330, 332, 333, 336, 337,340,352,355,356,357 mental verb(s). See verb Middle Ages 386 mimesis 339,341,344,347,349,351, 462-469. See also demonstration(s); iconic gestures icon(s), iconic, iconicity 194, 463 index,indexes 463 indicating 463 mimetics 13,234,467,468 selective depictions 463, 467 mimetics. See mimesis minimally different languages 119 minus-ground clause(s) 74, 95 misrepresentation 308, 309-327, 314 adverb(s) 310 false beliefs 308 modality, modalities 3,137,220, 236, 237, 243, 275, 278, 299. See also epistemic modality; medium Mon-Khmer 160,249 monolingual(s) 395, 396, 398, 436, 446 monolingual peer(s)/child(ren) 442, 443, 444, 445, 449, 451, 452, 453 monolingual speaker(s)/subject(s) 4, 264, 336, 365, 367, 369, 395-434 monolingual community 38

585

Subject Index

monolingual condition/context 439 monolingual narrative(s) 13 monolingual peer(s)/child(ren) See monolingual(s) monolingual speaker(s)/subject(s) See monolingual(s) monologic narrative. See narrative Mopan 492,517 Mordvinian 490, 516 mother-tongue competence 438, 442 motion characteristic expression of motion 93, 109, 177 conformation 255 core schema 89,90, 159, 161, 163, 176, 180, 186, 192,206 deictic 227 figure(s) 39, 40, 41, 47, 54, 60, 61, 90, 91,93, 117, 165, 176, 190, 191, 192, 195, 197, 198, 199, 200, 201, 202, 204, 205, 207, 208, 209, 210, 211, 212, 213, 223, 236, 470. See also motion: trajector(s) goal 17,21,27,29,30,34,47,51,71, 73, 74, 75, 95, 96, 97, 100, 103, 107, 110, 111, 117, 126, 147,237,240, 241, 244, 245, 246, 268, 296, 303, 364,365,471 ground(s) 54, 61, 70, 71, 72, 73, 75, 76, 77, 80, 84, 85, 86, 88, 90, 92, 95, 96, 97,98,99, 100, 103,109,117,118, 119, 126, 139, 159, 161,168,170, 182, 183, 184, 185, 186,187, 189, 190, 191, 192, 195, 197, 198, 199, 200, 201, 202, 204, 208, 209, 210, 211,213,220,221,222,223,236, 244-247, 296, 470. See also motion: landmark(s) landmark(s) 18, 117, 165, 167, 168, 177, 189, 190, 198,208,209,210, 211, 220, 223, 244, 246, 470, 471, 474. See also motion: ground self-contained motion 162, 254 source(s) 17, 22, 27, 28, 29, 30, 31, 34, 47, 51, 71, 73, 74, 75, 95, 96, 97, 100,

107, 108, 110, 111, 117,240,241, 244, 246, 296, 303 trajector(s) 116, 117, 165, 167, 168, 169, 179, 470, 471, 474. See also motion: figure(s) translational motion 97, 107, 110, 762, 241 vector 253, 255 motion-event typology. See typology motion verbs. See verb motor pattern(s). See manner mouth adverbs 204 mouvement 189 movement imitatives 102, 109, 111, 233 Myene 492,515 N

Nar-Phu 510 narrative computer-logged narratives 367 creating a story versus retelling a story 459 first-person narratives 290, 304, 331. See also personal experience(s) interactive storytelling 450 joint construction 353 monologic narratives 356 narrative competence 9, 11, 272, 273, 274, 302, 305, 308, 311, 312, 322, 450 narrative construction 4, 11, 113, 119, 219,334,353,438,446 narrative development 9, 10, 11, 19,38, 56, 113,263,264,265,281,287,293, 299, 336, 436, 451. See also development narrative discourse 11, 263, 264, 275, 304, 306, 365 narrative ideal 468 narrative organization 38,219,263, 264,281,287,305,309,396 narrative style 9, 20, 37, 38, 41, 42, 50, 55, 56, 59, 61, 83, 85, 229, 230, 231,

586

239, 240, 243, 245, 246, 247, 254, 322, 326 personal experience(s) 262, 263, 272, 273, 274, 280, 300, 304, 330, 331, 332, 458, 462. See also first-person narratives point(s) 6,331 retelling(s) 5, 42, 60, 87, 191, 353, 469 storytelling performance 272 story components 120, 121,351. See also plot structure story versus narrative 458^462 third-person narratives 290, 300, 304, 332 narrative attention. See attention narrative competence. See narrative narrative construction. See narrative narrative development. See narrative narrative discourse. See narrative narrative ideal. See narrative narrative organization. See narrative narrative style. See narrative narrative time. See time Navajo 190 negative qualifiers. See evaluative devices NezPerce 159,247 ngurra 17, 18, 21 Nicaraguan Sign Language 504,517 Niger-Congo 160,249,515 Niger-Kordofanian 515 Nilo-Saharan 515 Nilotic 35,492,515 nomadic people 144 non-boundary-crossing. See boundary crossing non-deictic. See deictic non-intentional 310-321 non-literate 38, 41 Norwegian 445, 492, 496, 515 nuclear verbs. See verb null subjects 406, 415-419, 430, 431, 432

Subject Index

o occasion 457 Ojibwa 160 omoiyari 352 on-going action 113,125 on-line 89,91,252,268,272,361 on-line communication 392 on-line constraints 360, 379, 388, 392, 393 on-line factors 243 on-line feedback 360, 361 on-line interaction 367 on-line packaging 234 on-line playback 370 on-line process 369 on-line processing 219,253,291 on-line production [and transfer] 415 on-line recording 369 on-line text production 428 on-line use 89,91,92,272 on-line process of writing. See writing onomatepoeia, onomatopoeic 89, 101, 111,339,341,344,347,349,351, 402,467 onset [plot structure]. See plot structure operating principles 392 organization 288 orientation(s) [spatial] 19, 21, 22, 29, 34, 39, 40, 47, 55, 85, 144, 145, 147, 236. 473, 474 orientation [story grammars]. See story grammars outcome [story grammars]. See story grammars

packaging 9, 11, 38, 238, 242, 244, 276, 277, 298, 312, 363, 402, 403, 404, 452. See also clause package, clause packaging Pama-Nyungan 490, 517 paper and pencil 377, 383

Subject Index

Papiamentu 447,504,517 passive(s) 38, 115, 275, 276, 277, 311, 314,316,322,323,324,325,421 path 17,34 path components 29, 238-239, 249 path focus, path-focused 61,207 path segment(s) 61, 88, 148, 149, 151, 152, 153, 156, 186, 187, 198, 239-244, 256. See also journey segments path components. See path path fogs. See gestures path gesture(s). See gestures path segments. See path path-type. See typology path verbs. See verb pause(s) 173, 174, 175, 190, 214. See also writing pause criterion. See writing pausing (rate). See writing Penutian 247, 517 persona(s) 193, 195, 196, 217, 468, 473 personal experience(s) [in narrative]. See narrative pet ownership 355 phrasal stress 116, 118, 141 phylum. See typology picture(s) 3, 197, 325, 465. See also mimesis pidgins 517 pitch 236 plain verbs. See verb plot 63,312 plotline 60, 84, 275, 284, 329, 333, 335, 337, 396, 400, 408 plot structure onset 63,284,309,311,396 resolution 284,309,311 unfolding 63,284,309,311 plus-ground clause(s) 74, 75, 95 point(s) [narrative]. See narrative

587

Polish 240,254,493,515 polymorphemic verbs. See verb Polynesian 159, 192 polysemous, polysemy 132, 162, 170, 175 polysynthetic. See typology Portuguese 240,254,493,513,515 Portuguese-based 517 position 38 positionals 474 postpositions 60, 95, 241 postural bonus effect 140 posture(s). See manner posture verbs, verbs of posture. See verb pragmatic language impairments. See impairment pre-digested 243 preposition(s) 51, 60, 118, 175, 177, 244, 250, 424, 426, 434 presuppositions 295,461,472 pretending 464 pretense 327 preverbs. See verb principle of chronological order 289, 291 produce 268 production rate 377-377, 379, 386, 386-388, 388 progressive [aspect] 10, 114, 116, 124, 137, 139,407,428,431,442 progressivity 116, 125, 138,430,431 projected speech 382-390 direct speech 382, 383, 384, 423 free direct speech 382, 383, 384 indirect speech 382, 383, 384 proper names 6, 357, 459 propositional attitudes 275, 277, 279 proposition rate 366, 367, 394 prosodic, prosodically 38,57, 116, 117, 140, 174, 236, 331, 333, 357, 407, 466 prosody 236, 357

588

prosody [in quotations]. See quotations proverbs 35 psychiatric disturbance. See impairment pure manner verbs. See verb pure path verbs. See verb purpose 7, 39

Q quotation(s) 13, 463, 466-467. See also mimesis animal quotations 466 manner of speaking 466 prosody 466 saying versus thinking 466 sound quotations 466

R randomized 292 rate. See manner rational stance. See stance reaction [story grammars]. See story grammars readily encodable 59, 60, 86, 246 reading disability. See impairment recall [of story] 261, 283, 286, 287, 291, 292, 305, 306 recall [of topography or scenes] 18, 267, 268 redundancy 205,427 reduplication(s) viii, 40, 104, 173,390 reference 10, 123, 126, 139, 187, 222, 223, 245, 262, 263, 268, 284, 444 co-reference, co-referential 70, 87, 294, 300,301 displace reference 4 maintain(ing) reference, reference maintenance 5,269,284,288,291,294, 300,301,417,440,441 referential chain 292 referential function(s) 307, 330, 333 referential implicitness 184, 186

Subject Index

referential shift(s) 193, 202, 203, 204, 468, 473. See also role shift(s) referential specificity 85, 246 referential strategy(ies), reference strategy(ies) 269, 272, 273 switch reference 86, 157, 263 reference strategy, reference strategies. See reference reference time. See time referential. See reference referential chain. See reference referential function(s). See reference referential implicitness. See reference referential shift(s). See reference referential specificity. See reference referential strategy, referential strategies. See reference register(s) 172, 273, 275, 278, 299, 300, 301,302,406,413,444 register appropriateness 436 rehearsing 458 related versus relatum. See relations relations chains of relations 471 related versus relatum 470 spatial relations 471 with-respect-to relations 471 relative clause(s) 10, 38, 174, 329, 402, 403, 404, 405, 406, 421. 424-426, 430, 432, 434 reliving 458 repetition(s) 20, 30, 83, 84, 88, 173, 175, 200, 201, 205, 207, 208, 211, 213. 245,249,291,334,338 resolution [plot structure]. See plot structure resolution [story grammars]. See story grammars restored randomized 292 result 459 retelling(s). See narrative

Subject Index

retrospective function 425. See also backtracking revisions [in writing]. See writing rhetorical style(s) 18, 35, 85, 86, 92, 100, 109, 143, 144, 145, 156, 161, 193, 220, 223, 225, 227, 243, 246, 248, 254, 257, 270, 334, 336, 417, 453 rhetorical typology, typology of rhetoric. See typology rhythm. See manner right hand. See hand(s) role(s) [character] 201,270 roleshift(s) 193-217, 468, 473. See also referential shift(s) role switches 382 role types 357 role(s) [semantic]. See semantic role role shift(s). See role(s) [character] role switches. See role(s) [character] role types. See role(s) [character] Romance 21,60, 117, 159, 160, 163, 176, 177, 181, 185, 186, 192,240,248, 249,251,253 Romansh 502,515 room descriptions 295, 303 route-direction (task) 295, 303, 387 Russian 181, 182, 189, 223, 224, 225, 227, 229, 231, 239, 240, 244, 254, 336,363,364,493,496,515

Saliba 513,517 Samoan 504, 516 sand drawing(s) 19,157 satellite-framed (languages). See typology saying versus thinking. See quotations scaffolding 268 scene analysis 470-^172 scene setting(s) 41, 42, 45, 54, 92, 98, 100, 161, 185, 187, 244, 246, 269, 270,271,401

589

Scheffe tests 24,25 schematization of experience 91, 110 schooling 42, 50, 56, 274, 398, 437, 447, 450-451 script(s) 20, 26, 34, 266, 267, 271, 272, 273, 282 second language/bilingual(ism) 395^34, 496^197,504-513 bilingual, bilingualism 11, 13, 146, 264, 365,366,435-454 bilingual society 62 second language acquisition 11 second language acquisition. See second language/bilingualism selective depictions [mimesis]. See mimesis self-contained motion. See motion self-repairs 362, 365 semantic role 310-327 semi-repetition(s) 82, 83, 84, 88 Semitic 159, 192, 240, 249, 253, 515 separation 205, 207, 208, 210, 211, 212, 213,214,245 Serbo-Croatian 240,254 serial-verb (languages). See typology serial-verb construction(s). See verb setting(s) [story grammars]. See story grammars setting felicitation] 11,13, 266, 272, 278, 279 setting [physical; communicative] 3, 4, 23, 263, 265, 274, 304, 332, 355, 358, 360,361,436 shift(s). See referential shift(s); role shift(s) sign language(s) 220, 246, 359, 360, 363, 366, 468 Sign Language of the Netherlands [SLN] 206,207,493,517 silent reading 386 simultaneity 10, 80, 236, 395, 406, 407, 412,414,431,442

590

Sinitic 516 Sino-Tibetan 160,249,516 Slavic 185, 227, 229, 231, 240, 249, 251, 255. See also Balto-Slavic SLI. See impairment: Specific Language Impairment SLN. See Sign Language of the Netherlands socio-economic status 335, 449, 453 sound quotations. See quotations sound symbolism 101-104, 233, 234, 235, 339, 349, 467 conventional sound symbolism 102 corporeal sound symbolism 101 imitative sound symbolism 101 synesthetic sound symbolism 101 source. See motion source location 250 Spanish vii, 4, 10, 12, 34, 37, 38, 49, 50, 56, 57, 60, 61, 62, 70-80, 84, 90, 91, 94,95,98, 101, 104, 105, 106, 111, 117, 119, 132, 133, 145, 147, 150, 161,163, 170, 182, 183, 184, 192, 193, 207, 212, 220, 222, 223, 224, 225, 229, 230, 231, 232, 233, 235, 240, 244, 246, 250, 251, 252, 254, 255, 256, 257, 270, 274, 277, 278, 279, 294, 295, 296, 299, 300, 302, 303, 304, 336, 395-434, 446, 461, 462, 466, 469, 473, 474, 487, 493, 494, 496, 497, 502, 504, 507, 508, 515 spatial anchoring 294, 295, 303 spatial morpheme(s) 117, 133, 135, 137, 138 spatial relations. See relations spatial verb(s). See verb Specific Language Impairment [SLI]. See impairment speech act verbs. See verb speech time. See time speech versus writing. See writing

Subject Index

speed 94, 163, 167, 359, 360, 361, 363, 462 spelling. See spelling errors; fingerspelling; writing spelling errors. See errors spontaneously produced 299 stance 351 attitudinal stance 363, 384 discourse stance 275 emotional stance 393. See also emotive stance emotive stance 6, 13. See also emotional stance epistemic stance 326, 327 rational stance 6 stasis 38, 39, 40, 45, 55, 57, 232 static location. See location(s) stationary location. See location(s) status of information 284 storytelling performance. See narrative story components. See narrative story grammars 282, 283, 288, 289, 453 abstract 459 attempt 283 beginning 283 coda 459 complicating action(s), complication 283,331,459,465 ending 283 evaluation(s) 283, 330, 331, 333, 334, 350, 459 orientation 288,459 outcome 283 reaction 283 resolution 63, 267, 268, 275, 283, 331, 396,401,459,461,465 setting(s) 283,288,289 story versus narrative. See narrative stylistic feature 85 subject-orientation 295 supportive aspect. See demonstration(s) Swedish 12, 113-141, 184, 240. 254. 277. 466, 494, 496. 504, 510, 513. 515

Subject Index

Swedish Sign Language 369, 386, 388, 517 switch reference. See reference synesthetic sound symbolism. See sound symbolism

Tai-Kadai 160,249 Tamil 159,245,494,516 telic, telicity 60, 70, 71, 225, 412, 430 temporality vii, 9, 17, 246, 270, 330, 437, 438,441,442,444,445 temporal anchoring 288, 299, 442 temporal contour(s) 302, 396 temporal profile 6, 8, 138 tense(s) 8, 10, 11,80, 103, 113, 114, 120, 121, 122, 124, 137,276,277,327, 329, 407, 408, 409, 415, 433, 436, 439, 441, 442, 469. See also anchor tense text telephone 360, 361, 388, 392 Thai 12, 759-790, 225, 227, 228, 254, 255,504,516 thematic subject 273, 286, 288, 300 thematic subject strategy 269 theory of mind 8, 11, 13,307 thinking-for-narrating. See thinking-forspeaking thinking-for-signing. See thinking-forspeaking thinking-for-speaking 7, 8, 11, 13, 37, 55, 59, 60, 86, 89, 90, 104, 109, 110, 156, 238, 305, 359, 367, 368, 392, 462 imagining-for-speaking 468 thinking-for-narrating 157 thinking-for-signing 386 thinking-for-writing 13, 359, 367, 368, 392 thinking-for-writing. See thinking-forspeaking third-person narratives. See narrative thoughts 6, 7, 11, 60, 204, 307, 466

591

Tibeto-Burman 516 time 4,9, 19,41,42,83, 104, 113, 114, 120-125, 137, 226, 360, 361, 408, 437, 452 event time 410 narrative time 430, 441 reference time 410, 425, 426 speech time 410 tone of voice 13,342,469 top-down 7, 286, 301,452. See also bottom-up topic maintenance 322, 323, 325, 326, 417,418,420,421 topic-orientation 295 topic shifting 322, 325, 326 Totonac 494,517 trajector. See motion trajectory, trajectories 37, 38, 47, 51, 52, 60, 90, 92, 98, 151, 185, 186, 206, 244, 295, 302, 303, 304, 475. See also trajector Trans-New Guinea 516 transfer 442, 445, 447, 448 transfer errors. See errors translational motion. See motion transliteration system 171,172 tripartite typology. See typology true 457 Turkic 249,253,516 Turkish vii, 4, 10, 12,34,96, 100, 111, 181, 182, 206, 207, 220, 223, 224, 225, 229, 230, 231, 232, 233, 234, 235, 242, 244, 246, 251, 254, 255, 264, 296, 303, 307-327, 329-355, 398, 433, 436, 437-443, 461, 463, 466, 469, 473, 487, 494, 497, 499, 510,516 turn 356, 362, 365, 388 typological frame. See typology typologies of language use. See typology typology 219-257,515-518 agglutinative 100,433

592

binary typology 184,233,237,247, 257. See also bipartite typology bipartite typology 250. See also binary typology complex verb-framed (languages) 160, 180,206,228,236 equipollently-framed (languages) 228, 251 figure-type 159, 192 grammar typology, typology of grammar 432 language group 515-517 language typology 114, 239, 395, 432 linguistic typology 11, 145, 156, 161, 253,515-518 manner-type 159, 165, 176, 192 motion-event typology 159, 160, 161, 162, 168, 176, 180, 186 path-type 159, 165, 176, 192, 206 phylum 515-517 polysynthetic 60, 80, 85, 246 rhetorical typology, typology of rhetoric 432 satellite-framed (languages) 10, 37, 147, 156, 182, 192,295 serial-verb (languages) 160, 161, 168, 181, 182, 188,227,228,229,234, 237, 247, 250, 255 tripartite typology 251 typological frame 91, 109, 110 typologies of language use 253 verb-framed (languages) 10, 37, 61, 74, 92, 101, 105, 109, 110, 117, 119, 145, 147, 156, 180, 181, 182, 192,206, 257, 295, 296, 304 Tzeltal 12, 37-57, 160, 232, 233, 236, 240, 242, 254, 461, 466, 469, 473, 474,494,504,517 Tzotzil 57,242,495,517

u U-shape(d). See development Udmurt 490,516 understandable 457

Subject Index

unfolding [plot structure]. See plot structure Unique Vector Constraint 256 Uralic-Yukaghir 516

V&V-construction(s). See verb vector. See motion; See also Unique Vector Constraint Veps 490,516 verb-framed (languages). See typology verb-style shifts 339, 344. See also verb verb affect verb affixes 40 bare motion verbs 61,74 bipartite verbs 247, 250 cause-of-motion verb(s) 168-170, 177, 178, 179, 180, 190. See also verb change-of-state verbs 422, 427 converbs 10,253 deictic-path verb(s) 164, 165, 167, 169, 177, 178, 179, 180, 182, 186 manner(-of-motion) verb(s) 50, 54, 55, 109, 110, 160, 161, 162, 163, 165, 166, 167, 168, 169, 170, 178, 179, 180, 181, 182, 186, 187,206,207, 224, 225, 227, 228, 229, 230, 231, 232, 233, 234, 237, 239, 242, 245, 248,251,257 manner + path verbs 167 mental verbs 308, 310. See also misrepresentations motion verbs 23, 32, 39, 40, 45, 48, 80. 119, 127-134, 159, 160. 168, 178, 182, 186,229,238,239,244 nuclear verbs 132, 137 path verb(s) 160, 163, 165, 167, 168, 178, 181, 182,229,248 plain verbs 193 polymorphemic verb 193 posture verbs, verbs of posture 34, 127. 179 preverbs 21,23,247,256 pure manner verbs 230

Subject Index

pure path verbs 230 serial-verb construction(s) 160, 163, 168, 180, 186,228,231,255 spatial verb(s) 126, 193 speech-act verbs 384,385 V&V-construction(s) 116,125,139, 140 verb gapping 402 verb morphology 21,84,86 verb prefixes 227, 229 verb gapping. See verb verb morphology. See verb verb prefixes. See verb visual attention. See attention vocabulary size 473 voice quality 220, 363, 382, 384, 386, 393 volume 236 Votyak 490

w Warlpiri 12, 17-35, 143, 144, 160, 242, 254,256,495,517 well-formed stories 282, 283, 292 West-Greenlandic 12, 59-88, 246, 254, 461,466 Western Nilotic languages. See Nilotic languages Williams syndrome. See impairment with-respect-to relations. See relations word frequencies 377 word order 10, 20, 67, 234, 293, 294, 406, 407, 419-423, 420, 430, 432, 434, 436, 439, 444 writing 220, 359-394. See also medium editing (operations) 361, 370, 372-374, 374, 382, 390, 392 final edited version 360 handwriting 367 on-line process of writing 359,370 pause(s) 362,363,365,370,371, 372-382, 386-387, 390, 392 pause criterion 372

pausing (rate) 366, 386 revisions 362, 391 speech versus writing 278, 359-369 spelling 367, 372, 379, 388, 394 writing on a computer 363, 367, 369, 372 writing with paper and pencil 369 writing on a computer. See writing writing with paper and pencil. See writing

Yiddish 495,515 Yucatec. See Yukatek Yukatek, Yucatec 57, 240, 495, 517 Yupno 495,516

zero anaphora, zero anaphors 184, 441. 444. See also anaphors, anaphora Zulu 235, 255


Person Index

A Aarssen 254, 256, 264, 395, 397, 398, 441,442,444,497,523 Abelin 101,523 Abelson 20,282,560,561 Abramson 560 Acredolo 567 AWgren565

Ahlsen 361,366, 367, 513, 523, 524, 555, ,,, j6o .. ,_„ - _ _ _ _ . .,.- » ft ,„, Aisenman 277, 333, 334, 523, c529, 546 Akinci 264, 395, 398, 443, 445, 523

Aparici 529 Applebee 264, 282, 284, 525 Areerich 529 g ^_ Argyle 363, 392, 525 ,, A T r Mr Armon-Lotem 415,525 _ Aske 60, 70, 71, 73, 225, 525 Assayag 333, 334, 523 . „_ _ _ _ A Astington 308, 525 A u , K . H . 451,525 ^ ^ ^ A Au, T. 334, 525 Austin 431,526,554,567 Avidor 413, 559

Aksu 308, 442, 532 Aksu-Koc 96, 242, 254, 256, 303, 307, 308,309,312,315,317,318,335, 353, 396, 407, 487, 523, 524, 529, 548,564,565

Backus 443,526 Balkom 504, 510 Bally 560

Albrechtsen 570 Albro 282, 283, 287, 565 Allen 157,265,524 Allwood 360, 364, 368, 524

Bamberg 8,254,256,286,290,299,300, 301, 304, 308, 324, 325, 326, 330, 332,333,336,337,350,351,352, 353, 354, 355, 398, 487, 526, 527,

Almgren562 Alsina 536 Alves 513,524 Alvre 490 Ameka 160, 161, 168, 186, 187, 188, 190, 228, 254, 255, 507, 525, 532 Ammerlaan 547 Andersen 396, 432, 525 Anderson 567 Andersson 363, 525 Anilovich 254 Anula Rebollo 424, 538 Aoki 247, 525

g

549 565

' Bamhamed 487 Barbieri 8 527

' °

Barlow 55

Barnes 564

Barr 539 Barrena 562 Barriga Villanueva 556, 562 Barsalou 474, 527 Baruch 546, 559 Barzilay 488, 555 Barddal 535 Bates 498, 559

596

Bauer 290, 527 Baughn 489, 568 Bauman 17,527,547 Bavin 17, 18, 28, 143, 159, 242, 245, 254, 256, 492, 495, 527 Bazzanella 286, 527 Beattie 365,527,533 Bedell 546 Behrens 118,527 Behrns 513,528 Bellugi 357, 365, 366, 367, 487, 498, 548, 559 Ben-Amos 545 Bennett-Kastor 508, 528 Bennis 569 Berg 524 Berglund 504, 528 Berlin 473, 528 Berman vii, viii, 3, 4, 8, 9, 11, 12, 13, 17, 20, 34, 37, 38, 41, 42, 49, 55, 56, 57, 59, 60, 62, 63, 64, 65, 66, 69, 84, 89, 92, 110, 120, 126, 144, 145, 146, 148, 149, 156, 161, 171, 173,220,223, 238, 246, 249, 251, 254, 256, 261, 262, 263, 264, 265, 266, 269, 270, 271, 272, 273, 274, 275, 276, 277, 278, 279, 281, 282, 284, 286, 287, 288, 290, 294, 295, 296, 297, 298, 299, 301, 303, 309, 327, 329, 332, 333, 334, 336, 364, 395, 396, 397, 398, 399, 401, 402, 403, 405, 407, 408, 411,413, 415, 418, 419, 424, 427, 432, 433, 434, 435, 436, 452, 454, 457, 458, 460, 487, 524, 526, 527, 528, 529, 530, 534, 535, 536, 540, 542, 548, 549, 550, 552, 559, 561,562,568 Bemhardt 553 Bernstein 553 Berry 548 Berthele 502 Bialystok 450, 452, 453, 530

Person Index

Bierwisch 470, 530 Billman 254,531 Bisang 160,531 Bishop 510 Blakemore 550 Blank 265, 559 Blankenstijn 499,510,531 Blum-Kulka 265, 557 Bocaz 144, 254, 397, 487, 493, 531, 532, 564 Boeschoten 438, 442, 532, 537, 569, 570 Bohnemeyer 56, 57, 240, 256, 532 Bolado 110 Borer 415,434,532 Bos 254, 256, 395, 397, 398, 444, 496, 523,532 Bottari 491,499 Botting 510,532 Bowerman 57, 157, 189, 242, 248, 254, 295, 364, 428, 532, 534, 564 Boyd 449, 453, 510, 532, 533, 554 BoyesBraem 194, 533 Braine 528 Brandt 87 Brehm 510 Brennan 193,533 Bresnan 536 Brewer 18,20,533 Broeder 566 Bronckart 284, 290, 299, 533 Brown, A. L. 554 Brown, P. 37, 38, 55, 56, 57, 159, 160, 232, 233, 240, 242, 254, 466, 473, 494, 504, 532, 533 Bruce 544 Bruck 450, 533 Bruner 307, 325, 330, 533 Budwig 256, 487, 526 Burling 473, 533 Bursik 541

Person Index

Butterworth, B. 365, 533 Butterworth, G. E. 525 Bybee 563

Cahana-Amitay 278, 279, 547, 559 Called 286, 527 Camp 265, 272, 488, 498, 542, 572 Campbell 533 Caron 567 Carroll 254, 571 Casad 573 Castro 524 Catherine-Howell 556 Cenoz 507, 533, 534 Chafe 8, 262, 268, 284, 365, 368, 475, 534 Chandler 564 Chapman 498 Chase 471,474,534,535 Chatman 458, 534 Chen-Wilson 502 Chilosi 491,499 Chini 530 Choi 248, 295, 364, 532, 534, 553 Chuwicha 160,534 Cipriani 254,491,499 Clancy 352, 353, 534 Clark, E. V. 140, 157, 365, 469, 532, 533, 535, 549, 556, 558, 572 Clark, H. 12, 13, 284, 360, 362, 363, 365, 457, 463, 464, 467, 468, 469, 470, 471,473,474,534,535,539 Clark, M. 184,535 Cobo-Lewis 507 Cole, D. E. 513,535 Cole, P. 541 Cole, R. W. 534 Collins 384, 552

597

Collis 355, 536 Comrie 519,541 Conte 541 Conti-Ramsden 510 Cook, H. M. 344,535 Cook-Gumperz, J. 571 Corsaro 571 Costermans 542 Costers 504, 507 Coulmas 541 Couzijn 536 Covey 254 Cowan, C. 488 Cowan, P. 488 Crago 498 Croft 254, 535 Cummins 535 Cuyckens 573

D Dahl 116,125,535 Dal Martello 285, 558 Damrad-Frye 308, 324, 325, 326, 330, 332,333,336,337,350,351,352, 353, 355, 487, 526 Danziger 157,492,557 Dasinger 424, 426, 489, 535 Day 377, 566 Dechert 527 DeCurtis 508,541 Deforest 282, 550 Dehnhardt 571 DeLancey 247, 536 Delphonese 561 del Castillo 494 Demick 541 Dennis 559 Denny 85,536 Deutsch 546

598

de Houwer 535 de Jong 449, 550 de Leon 495, 496, 532, 535 de Sousa 524 de Week 290, 299, 535 Dibiase 541 Dickson 453, 536 Dijkhuis 488, 536 Dirven 564 Dittmar, J. 496 Dittmar, N. 529 Do 557 Dobrich 451,560 Dodsworth 285, 557 Doi 352,536 Doiz 507,549 Dolitsky 533, 537, 538, 542 Dominguez 557 Donahue 488 Donaldson 265, 536 Dorian 525 Drissi 487, 546 Dromi 419,528,536 Dronkers 499 Drozd 157 Duncan 105,235,254,552 Durgunoglu 450, 451, 536 Durie 160,536 Durova 254, 493, 536 Diirscheid 565 Dussart 567

E Eaton 355, 536 Egli 571 Ehrich 295,536 Ehrlich 291,536 Eilers 507, 557 Eisenberg 284, 536, 553

Person Index

Eklundh 369,536 Ekmekci 439, 537 Emmorey 193,537,550,553 Emslie 285, 537 Engberg-Pedersen 59, 63, 87, 193, 205. 246, 254, 468, 537 Engstrand 566 Enkvist 368, 537 Erguvanh 532 Erguvanh-Taylan 548, 556, 565 Eriksson 504, 537 Ervin-Tripp 331, 332, 335, 524, 527, 537, 547, 548 Esperet 283, 284, 537 Essegbey 160, 161, 168, 186, 188, 190, 228, 254, 507, 525 Evans, G. 7,537 Evans, P. 157 Extra 436, 454, 523, 529, 532, 537, 557, 558, 560, 568, 570 Ezeizabarrena 562

Fant 545 Farrar 265, 537 Fayol 282, 287, 538, 542 Fein 510 Feldman 498 Fernald 352, 562 Fernandez, M. 507 Fernandez-Ramirez, S. 538 Fernandez Lagunilla, M. 424, 538 Filipovic 252, 254, 538 Filipsson 544 Fillmore 189,538 Fine 498, 538 Fischer, K. 265,538 Fischer, S. D. 566 Fish 556

Person Index

Fishman 454, 538 Fivush 284,538 Flavell, E. R. 538 Flavell, J. H. 308,538 Fletcher 535,542,558,571 Flickinger 553 Fong 254,538 Forbes 265, 537 Forrest-Pressley 530 Fortescue 67, 71, 72, 80, 84, 85, 86, 87, 246, 538, 539 Fox, B. 563 Fox Tree, J. E. 362, 539 Franzen 508 Freedle 535, 565 French 265, 282, 287, 299, 539 Frid 544 Friedman, S. 513 Friedman, W. 565 Friend 265, 537 Fridjonsson 113, 115,539 Fromhoff 538 Fujii 491,539 Fujiyama 507 Fukuda 507 Furberg 524

Galambos 450, 539 Galbraith 524 Gallardo 507, 539 Galvan 159, 191, 236, 237, 243, 245, 246, 254, 256, 466, 468, 487, 539, 567 Gao 255, 539 Garcia, G. E. 453, 539 Garcia Lecumberri, M. L. 507, 534, 539, 540, 549 Garcia Mayo, M. P. 507, 534, 539, 540, 549

599

Garcia Soto, J. R. 490, 540 Garman 553,571 Garnham 291,540 Gaskins 495 Gayraud 360,540 Gee 8,334,540 Geeraerts 162, 540 Gelman 265, 540 Genesee 450, 533 Gennari 254, 540 Gentner 564 Gerard 282, 555 Gerhardt 524, 527, 529, 554 Gernsbacher 488, 540 Gerrig 8, 363, 463, 464, 467, 468, 535, 540 Gershkoff-Stowe 527 Givon 534, 537, 543, 563 Glass, A. D. 144,540 Glass, W. R. 526 Gleitman 254, 557 Glenberg 474, 540 Glenn 282 Goksel 556 Goldberg 162,540 Goldin-Meadow 450, 539, 564 Goldman-Eisler 365, 540 Goldstein 545 Gombrich 7, 540 Gomez, E. 507 Gomez de Garcia, J. 491 Goodluck 571 Goodman 564 Gopnik,A. 308,498,525 Gopnik, M. 284, 541 Granger 544 Gray 284, 538 Green 538 Greene 561

600

Greenhill 556 Gregory 537 Grice 364, 541 Griffin 488, 542 Grimshaw 434, 541 Grodzinsky 434, 532 Gruendel 265,282,554,561 Grunow-Harsta 510 Gumperz 8,368,541,563 Guo 254, 492, 524, 527, 529, 554 Gutierrez, E. 495 Gutierrez, J. 507 Gutierrez-Clellen, V. F. 508,541

H Haberland 363, 541 Hahn 556 Haiman 563 Hainenen 566 Hakansson 565 Hale 21,541 Halle 545 Halliday 284,541 Hamano 101,102,111,234,255,541 Hansson 544 Haritos 496 Harkins 488, 541 Harris 355,525,541,550 Hasan 284, 541 Haspelmath 531,563 Haugen 113,541 Hauser 360, 541 Haviland, J. 19,57,495,541 Haviland, S. E. 284, 535 Hawk 206 Heath 334,541 Hebert 489,497,541 Hecht 265, 542 Heerdt 563

Person Index

Heffeman 156 Heller 548 Hellstrand 566 Helm 538,541,549 Hemming 567 Hemphill 488, 489, 498, 542, 564, 572 Hendriks 288, 294, 295, 297, 303, 306, 542 Henningsson 189 Henriksen 570 Hertling 86, 549 Hickmann 4, 8, 13, 261, 262, 263, 269, 281, 282, 283, 285, 286, 288, 290. 291, 292, 294, 295, 297, 298, 299, 300, 303, 306, 398, 435, 460, 490, 494, 542, 543, 546, 561 Hidi 384,543 Highter 553 Hildyard 384, 533, 543 Hill 488,492 Hinton 101, 102, 103, 104,543 Hohenstein 251,254,543 Hoiting 60, 70, 71, 159, 160, 161. 168, 170, 176, 177, 181, 186, 188, 192, 206,207,211,212,213,225,228, 236, 493, 543, 564 Hollman 535 Holmes 502,555 Holmqvist 394,513,543 Hoogenraad 156 Hopper 285, 408, 543, 563, 564 Home 363, 544 Hosenfeld 561 Hottenroth 251, 544 Housen 504,507,544,558 Hovdhaugen 551 Howard 543 Howell 497, 544 Hualde 103, 544 Huber 525

Person Index

Hub Faria 254, 493 Hudson 262, 263, 265, 272, 282, 287, 544, 545, 562 Hufeisen 533 Hulsen 547 Hung 544 Husen 528 Huss 554 Hustlethwaite 528 Hyams 140,415,545 Hyltenstam 528,565,570 Hyman 525 Hymes 17,545 Hyona 391,545 I

Ibarretxe-Antunano 89, 103, 111, 118, 159, 160, 183, 184, 187,233,235, 241, 250, 254, 467, 468, 504, 545 Idiazabal 562 Igushev 490 Indambraya 166, 560 Irvine 548 Iverson 24,410,545

Jacobs 547 Jacobsen 87 Jaeger 499 Jaeggli 545 Jagst 541,554 Jakobson 140,362,545 Janssens 504, 507, 544, 558 Jefferson 457, 545, 560 Jendraschek 255, 545 Jenkins 308, 525 Jessner 533 Jimenez 507

601

Jisa 254, 256, 275, 276, 277, 395, 398, 443, 445, 490, 497, 523, 545, 546, 559 Johansen 557 Johansson 377, 513, 543, 546, 566, 568 Johnson, C. 529 Johnson, N. S. 282,551 Johnson-Laird, P. 291,540 Johnson Herrera, C. 256 Jokinen 489 Jones, G. 544 Jones, P. 567 Jonsdottir, H. 566 Jonsdottir, M. 115, 567 Jonsson 546 Joreskog 447, 546 Jovanovic 252 Juhola 489 Julian 561

K Kail 269, 294, 300, 398, 434, 487, 490, 493, 494, 499, 543, 546 Kako 553 Kameenui 453, 536 Kamil 539 Kamio 352, 546 Kangassalo 490 Kaplan 487, 526 Karlsson 370, 390, 566 Karmiloff-Smith 3, 8, 120, 262, 264, 265, 269, 274, 279, 285, 288, 291, 300, 301,427,435,452,546 Kathol 553 Katzenberger 262, 263, 265, 266, 268, 269, 271, 273, 276, 278, 279, 401, 530, 547 Kaufman 508,547 Kearins 19,547 Kegl 504

602

Kellerman 504,547,570 Kelley 510 Kendon 104,547 Kentoy 524 Kern 254, 256, 264, 270, 356, 395, 398, 401, 443, 445, 490, 497, 523, 545, 547 Kernan 332, 547 Kerslake 556 Keseling 526 Kettemann 539, 569 Khattab 507, 547 Kilian-Hatz 234, 251, 525, 553, 560, 570 Kim 571 Kirshenblatt-Gimblett 457, 547 Kita 226, 234, 235, 468, 548, 556, 557 Kitayama 352,551 Klich 18,548 Klima 357, 365, 366, 367, 394, 487, 498, 548, 559 Kobayashi 546 Ko? 523,563,569 Koch 144,548 Koenig 538 Kohrt 526 Kollberg 369, 536, 548 Konig 531,563 Koopmans 8, 548 Koskinen 489 Kotiranta 489 Kramer 251,548 Kreiman 542 Kress 116,548 Kriz 566 Krych 254,531 Kunnap 490 Kuno 352, 548 Kuntay 254, 269, 308, 309, 316, 317, 319, 320, 324, 326, 329, 331, 332, 335,

Person Index

353, 358, 467, 469, 473, 491, 494, 524, 537, 548, 549, 565 Kupersmitt 395, 401, 402, 408, 415, 433, 446, 549 Kuteva 188,248,562 Kyratzis 351, 524, 527, 529, 549, 554

Labov 18, 262, 282, 288, 307, 325, 330, 331, 332, 333, 411, 458, 459, 549, 557 Lakarra 103,544 Lamb 554 Lambrecht 306, 549 Langacker 117,165,189,212,549 Lanza 445, 492, 496, 549, 566 Lanzetta 491,499 Lasagabaster 507, 534, 549 Latona viii Laughren 21,22,35, 144,549 Lauren 489,494,496 Lee 8,368,549,571 Lehtihalmes 566 Leiwo 566 Lennerstrand 545 Lennert Olsen 84, 85, 86, 246, 539, 549 Leseman 449, 550 Leslie 498, 525 Letts 553 Leung 502 Leveelahti 489 Levelt 362, 365, 550 Levin 528,546 LeVine 555 Levinson 8, 170, 368, 490, 525, 532, 533, 541,550,557,561,563,564 Levy, P. 494 Levy, Y. 265,528,550 Lewis, D. 19, 144,550 Lewis, G. L. 550

Person Index

Lewis, V. A. 355,536 Li 463,464,534 Liang 262, 542, 543 Liddell 193,194,213,550 Lieven 8, 489, 550 Lindblom 555 Lindvall 189,504 Linell 368, 550 Littlefield 556 Lizza viii Lodge 462, 463, 468, 550 Longacre 10 Lopez-Ornat 431,494,550 Lotan 262, 280, 550 Lucas 556 Lucy 495 Lynge 87

M MacDonald 363,551 MacKinnon 530 Macnamara 562 MacWhinney 171, 173, 446, 452, 535, 542, 550, 558, 562 Maddieson 525 Malmkjaer 534 Malmsten 370, 566 Malt 540 Mandelbaum 560 Mandler 282, 283, 288, 290, 527, 550, 551 Manhardt 489, 498, 559 Maratsos 285, 551 Marchman 214, 254, 286, 290, 332, 487, 498,526,551,559 Marcus 352,551 Mard 489, 494, 496 Margetts 513,551 Marold 527

603

Martin, J. R. 489 Martin, S. 339,551 Massey 254, 557 Mayer vii, 3, 4, 5, 17, 18, 37, 59, 60, 63, 89, 92, 119, 145, 157, 194, 281, 308, 309, 329, 330, 336, 370, 398, 460, 461,462,465,479,487,551 Maynard 335, 339, 344, 350, 551 McCabe 262, 284, 334, 335, 352, 353, 355, 356, 540, 545, 551, 552, 557 McCartney 282,551 McDonald 193,551 McGivern 559 McGrath 467 McGraw 553, 560 McGurk 363,551 McLaughlin 551 McNeill 104, 105, 106, 107, 111, 193, 197,212,213,235,254,551,552 Mees 570 Meid 548 Melzi 502 Mendelsohn 561 Menig-Peterson 284, 552 Mettewie 504, 507 Michaels 334, 384, 552 Michelena 103,552 Mihailov 490 Milanovic 570 Miller, F. J. viii Miller, G. 561 Miller, P. J. 284, 332, 552 Mills 488,499 Minami 8, 334, 335, 344, 350, 353, 356, 507,552 Mirsky 493 Mitchell, L. 287,565 Mitchell Kernan, C. 547 Mogi 507 Montes 493

Person Index

604

Moore, A. 450,552 Moore, T. 534 Mora Gutierrez 254, 552 Moreton 265, 272, 572 Morgan 504,541,552,553 Mosel 551 Mosenthal 539 Mosin 490 Mosina 502, 553 Motzfeldt 87 Msimang 255, 553 Muansuwan 164,165,553 Muller Gathercole 507 Munger 489, 568 Munn 19, 144,553 Muraki 546 Murre 566 Myers 144,553

N Naigles 254,510,543,553 Nakamura 308, 325, 326, 329, 335, 339, 344, 345, 357, 358, 467, 469, 473, 491, 494, 502, 508, 548, 549, 553 Napaljarri 35 Napanangka, C. 35 Napanangka, N. 35 Narain 446, 448, 553, 570 Narasimhan 254, 554 Nash 18,23,554 Naucler 449,510,533,554 Neeman 254,398,403,405,411,413, 424,530 Nelson 20, 262, 265, 266, 282, 287, 299, 308, 496, 529, 539, 544, 551, 554, 560, 561 Nettelbladt 565 Neumann 491 Nezworski 282 Nichols 101, 102, 103, 104,543

Nickels 264, 489, 568 Nicolopolou 263, 554 Niemeier 564 Niiniviita 489 Nippold 3, 435, 554 Nir 276, 277, 530 Nistov 445, 554 Nivre 524 Noonan 510, 554 Norbury 510 Nordqvist 115, 120, 140, 254, 256, 276, 359, 382, 386, 388, 466, 504, 513, 555, 558, 566 Norpoth 24,410,545 Nouaouri 508

o O'Connell 282,555 O'Neill 502,555 Oakhill 291,540 Ochs 353, 534, 555 Ogouamba 492 Oh 44, 250, 254, 257, 502, 555 Ohala 101, 102, 103, 104, 543, 555 Ohara 254, 555 Ohman 362,363,555 Ojeda 542 Oker-Blom 508 6ktem,A. 438,555 Oktem, 6. 438,499,555 Oiler 507, 557 Ollila 489 Olsen 386, 555 Olshtain 429, 432, 488, 547, 555 Olson 391,533,543,545 Olszewski 488 Ong 367, 382, 555 Oram 498 Ordonez 508, 556, 562

605

Person Index

Orsolini 254,287,491,556 Ovid viii Ozcahskan 37, 54, 56, 229, 230, 232, 254, 556 Ozsoy 548,556,565 Ozyurek 37, 556

Pallotti 507 Palmer 573 Palojarvi 508 Palttala 489 Papafragou 254, 557 Pause 571 Peacock vii Pearce 502 Pearson, B. Z. 256, 395, 398, 507, 557 Pearson, P. D. 539 Pederson 161,245,494,557 Pedro 552 Peirce 463 Peled 265, 557 Pellegrini 299, 557 Peng 492,557 Penner 529 Perez-Leroux 526 Perez-Pereira 490 Perkins 510,543 Perlmutter 571 Perner 308,557 Perren 569 Petch-Tyson 544 Peters 566 Peterson 262, 284, 285, 334, 352, 355, 540,545,551,557 Petit 524 Petofi 541 Petree viii Pfaff 443, 450, 557, 558

Pfanner 491,499 Phillips 307,571 Picardi 498, 542 Pick 567 Pieraut-Le Bonniec 533, 537, 538, 542 Pierrard 544, 558 Pinar 567 Pishwa 527 Plato 462, 463 Plunkett 566 Polinsky 397 Pontecorvo 491, 529, 556, 570 Poulin 254, 538 Poulos 255, 553 Poulsen 570 Power 285, 558 Preece 299, 558

R Rabinowitch 402, 558 Radford 3, 435, 558 Ragnarsdottir 113, 115, 118, 120, 125, 140, 159, 184, 240, 245, 254, 256, 277, 278, 466, 490, 530, 558, 564, 566 Rakison 527 Ramat 488, 498, 499, 529, 530 Ramers 565 Raukola 489 Raupach 527 Ravid 276, 277, 278, 279, 397, 413, 419, 558, 559, 566 Rehbein 555 Reilly 193, 256, 264, 275, 277, 278, 279, 332, 333, 357, 358, 487, 489, 498, 527, 537, 546, 550, 559 Reinhart 408, 559 Remond 291, 536 Renner 214, 254, 309, 315, 316, 318, 319, 320, 487, 559

606

Person Index

Rescorla 498, 559 Restrepo 508 Rhodes 247 Richards 561 Richthoff 566 Riedlinger 560 Rijlaarsdam 536 Rizzi 419,559 Rodkin 283, 284, 354, 355, 489, 568 Rodriguez,!. 307,571 Rodriguez-Trelles Astruga, A. 490, 559 Roelofs 488 Roeper 571 Roland 288, 294, 295, 297, 300, 303, 490, 542, 543

Rosado 275, 278, 279, 529, 546, 559, 568 Rose 265, 559 Rossi 491,556,559 Rubino 255, 560 Ruiz 507 Russell 561 Ryan, C. 146, 156 Ryan, E. B. 453, 530

Sachs 285, 560 Sacks 362, 560 Safir 545 Sak-Humphry 166,560 Sakurai 502 Salaberry 544 Sanchez y Lopez 269, 294, 300, 434, 546 Sandra 573 Sapir 11,91,560 Satoh 508,560 Saussure 101, 159,560 Sauvaire 285, 560 Savic 254,256 Saville 570

Scarborough 451,560 Schaefer 255, 560 Schank 20, 560 Schaufeli 440,441,560 Schegloff 560 Schelletter 553 Schenkein 545 Scheper 499,510,531 Schick 193,560 SchierTelin 353, 534, 555 Schiff 499,561 Schiffrin 535,562 Schlesinger 528 Schneider 291,292,543,561 Schnell 556,562 Schneuwly 284, 290, 299, 533, 535 Schotz 513 Schultze-Berndt 247,510,561 Schwarz 565 Schwarze 250, 251, 544, 561, 571, 572 Scollon 533

Searle 472, 561 Sebastian 49, 254, 296, 397, 408, 427, 430,431,487,494,561 Sechehaye 560 Segal 459, 561 Segerback viii Seidman 265, 561 Selinker 397, 561 Sells 536 Senft 160,491,513,557,561 Seppa 489 Sera 434,494,561 Serpell 451,561 Shank 282, 561 Shannon 471, 561 Shapiro 262, 265, 272, 287, 352, 545, 562 Shatz 265, 562 Shen 264, 562

Person Index

Sherblum 524 Sherzer 547 Shibata 508 Shibatani 344,562,563 Shimron 562 Shirai 544 Shiro 508,556,562 Shopen 567 Shweder 555, 563 Sierra 534 Siewierska 527 Sigurdsson 113,562 Silva-Corvalan 411, 433, 496, 497, 562, 569 Simmons 453, 536 Simonsen 566 Sinha 188,248,562 Siple 566 Sisilia 504 Sjogren 554 Slack 532 Slobin, D. I. vii, viii, 3, 4, 7, 8, 9, 11, 12, 13, 17,34,37,38,41,42,49,54,55, 56, 57, 59, 60, 61, 62, 63, 64, 65, 66, 69, 70, 71, 73, 74, 75, 76, 77, 78, 79, 80, 84, 86, 88, 89, 91, 92, 94, 95, 98, 100, 101, 103, 104, 109, 110, 111, 120, 126, 132, 138, 140, 144,145, 146, 148, 149, 150,151,156, 157, 159, 160, 161, 167, 168, 170, 171, 173, 176, 177, 181, 182, 183, 184, 185, 186, 187, 188, 189, 192, 193, 206,207,211,212,213,214,219, 220, 223, 225, 228, 229, 230, 232, 235, 236, 237, 238, 239, 244, 246, 249, 251, 254, 257, 261, 263, 264, 269, 270, 274, 275, 279, 281, 282, 284, 286, 287, 288, 290, 294, 295, 296, 297, 298, 303, 305, 308, 309, 327, 329, 332, 334, 336, 358, 359, 364, 368, 392, 395, 397, 399, 402, 408, 413, 427, 428, 430, 431, 432,

607

433, 434, 435, 436, 437, 442, 452, 454, 457, 458, 462, 472, 474, 487, 493, 495, 499, 519, 524, 526, 527, 529, 530, 532, 534, 535, 539, 540, 542, 543, 548, 550, 552, 554, 556, 561,562,563,564,568 Slobin, J. L. 495 Slobin, N. L. 495 Sloman 540 Smith, H. 510 Smith, P. T. 533 Smoczyriska 254, 256, 488, 493 Snow 451, 550, 556, 561, 562, 564, 572 So 544 Soderbergh 558 Sokolov 572 Sommerville 571 Song 527 Sorbom 447, 546 Sotirova 535 Sozer 541 Spencer 537 Sperry 284, 332, 552 Stahlke 160,564 Starosta 166, 560 Stavans 256, 446, 564 Stechow 571 Steenge 504,510,564 Stein 282, 283, 287, 288, 289, 489, 564, 565, 568 Stenning 287, 565 Stephany 502, 565 Stevenson 285, 537 Stigler 563 Stiles 498 Stojanovik 510 Stoll 256 Stollwerck 510 Stoness 556 Strating 547

Person Index

608

Strauss 273, 565 Streek 571 Stromqvist vii, viii, 3, 13, 113, 115, 118, 125,140, 159, 184, 189,240,245, 254, 256, 275, 276, 277, 278, 359, 360, 361, 363, 366, 367, 368, 369, 370, 371, 377, 386, 387, 390, 394, 436, 466, 494, 513, 523, 524, 527, 529, 530, 533, 537, 539, 543, 545, 547, 549, 554, 555, 557, 558, 564, 565,566,570,571,572,573 Supalla 193, 205, 206, 207, 211, 212, 213, 566 Sutov 490 Sutton 18, 567 Svavarsdottir 115, 567 Swartz 541,554 Swilley 254,531

Tager-Flusberg 489, 498, 542, 567 Takahashi 189,357,491,567 Talmy 8, 12, 13, 20, 37, 39, 60, 70, 89, 90,91, 103, 109, 110, 117, 138, 144, 147, 159, 160, 161, 162, 165, 176, 177, 180, 189, 191, 192,205,206, 212, 213, 219, 220, 223, 225, 227, 228, 230, 236, 237, 238, 247, 253, 255, 256, 257, 295, 413, 470, 472, 473, 567 Tannen 534,535,552 Tano 556 Taoka 535 Tardieu 291,536 Taub 159, 191, 194, 236, 237, 243, 245, 246, 254, 256, 466, 468, 539, 567 Teberosky 529 Tecumseh Fitch 540 Tekdemir 307, 463, 524 Terrazas 254, 553 Tesniere 176, 189,567

Thepkanjana 160, 164, 165,567 Thompson 563 Thorell 113,567 f>rainsson 113, 141, 567 Thurber 467 Tobin 572 Toivainen, J. 489, 490, 502, 566, 568 Toivainen, K. 489, 566 Tolchinsky 275, 276, 278, 279, 529, 559. 568 Tomlin 285, 568 Torrance, M. 524 Torrance,N. 533,543 Tosi 449, 568 Toupin 424, 426, 535 Trabasso 264, 282, 283, 284, 354, 355, 489, 565, 568 Trask 101, 103,544,568 Trim 569 Trondhjem 59, 82, 84, 87, 62, 246, 254, 537, 568 Tronnier 504 Tsitsipis 18, 568 Tyler 291,568

u Uccelli 556,562 Uchida 335,491,568,569 Umbel 395,398,507,557 Umiker-Sebeok 332, 569 Uneson viii Ure 368, 569 Urrutia Cardenas 496, 562, 569 Ushakova 493

van Balkom 564 van der Bergh 536 vanderLely 510,569

Person Index

van der Wege 464, 535 van der Zee 532 van Dijk 264, 282, 569 van Gogh 465 van Hell 559 van Hoof 256, 504, 547, 569 van Kemenade 569 van Valin 492, 499 van Wae 524 Vasiljev 490 Verhoeven vii, viii, 3, 13, 254, 256, 262, 273, 274, 275, 277, 278, 435, 436, 438, 439, 441, 442, 444, 446, 447, 448, 450, 451, 454, 496, 497, 504, 510, 523, 524, 527, 529, 530, 532, 533, 536, 537, 539, 545, 546, 547, 549, 550, 553, 554, 557, 558, 560, 561, 564, 566, 568, 569, 570, 572, 573 Vermeulen 504, 570 Viberg 132,445,528,570 Villanova 157 Vion 285, 560 Voeltz 234, 251, 525, 553, 560, 570 von Stutterheim 254, 396, 407, 524 Vuonokari-Kamarainen 502

609

Weist 410, 571 Wellhousen 265, 571 Wellman 265,307,525,571 Wengelin 256, 276, 359, 361, 386, 466, 513,543,566,571 Westermann 251,571 Weston-Smith 550 Wetherell 510 Wetterstrom 524 Wetzels 547 Whorf 11,91,364,549,571 Wide 115,571 Wieden 539, 569 Wienold 176,250,571,572 Wigglesworth 254, 270, 285, 398, 489, 572

Wilkins 37, 38, 57, 62, 77, 85, 88, 143, 144, 147, 159, 242, 243, 254, 256, 355, 357, 488, 499, 525, 533, 557, 561,572 Williams, B. 291, 561 Williams, J. 534 Winston 553 Wolf 265, 272, 488, 498, 542, 572 Wrobel 526 Wulfeck 498

w Walchli 189,570 Waletzky 18,282,283,288,307,325, 330,331,332,333,458,549,557 Waller 530 Walsh 527 Walton 465, 570 Warden 262, 285, 570 Wasow 362 Wassmann 495 Weinold 165 Weiss 502,571 Weissenborn 295,415,571

Yagmur 547 Yallop 527 Yangklang 159, 225, 228, 254, 504, 573 Ygge 545 Ying 255, 572 Yoshida 571 Yurieva 254, 493, 536

Zaiceva 490 Zamora 559, 568

610

Zavialoff 546 Zijlenmaker 488, 572 Ziv 423, 572 Zlatev 159, 162, 165, 170, 175, 177, 188, 189,225,228,254,504,573

Person Index

Relating Events in Narrative, Volume 2: Typological and Contextual Perspectives

Time and Narrative Volume 2(Time & Narrative)

Information Structure: Theoretical, Typological, and Experimental Perspectives

Benefactives and Malefactives: Typological Perspectives and Case Studies (Typological Studies in Language)

Nominalization in Asian Languages: Diachronic and typological perspectives

Adpositions: Pragmatic, Semantic and Syntactic Perspectives (Typological Studies in Language)

Civil War, A Narrative Volume 2

Perspectives On Learning Disabilities: Biological, Cognitive, Contextual

Ancient Narrative 2 (2002)

Modality in typological perspective

New Perspectives on Narrative and Multimodality (Routledge Studies in Multimodality)

Rethinking Grammaticalization: New perspectives (Typological Studies in Language)

Time and Narrative, Volume 3 (Time & Narrative)

ADVANCES IN CATALYSIS VOLUME 2, Volume 2

Plant Developmental Biology - Biotechnological Perspectives: Volume 2

Selves and Identities in Narrative and Discourse (Studies in Narrative)

Perspectives of Nonlinear Dynamics (Volume 2)

Time and Narrative, Volume 3 (Time & Narrative)

Time and Narrative Volume 1(Time & Narrative)

Contextual analysis

Contextual Computing: Models and Applications

Relating Software Requirements and Architectures

Syntax: A Functional-typological Introduction: Volume II

Contextual Computing: Models and Applications

ADVANCES IN GEOPHYSICS VOLUME 2, Volume 2 (v. 2)

Events Design and Experience (Events Management)

Poetry and Narrative in Performance

International Perspectives of Festivals and Events: Paradigms of Analysis

Events in computation

Beyond Narrative Coherence (Studies in Narrative)

Distributist Perspectives: Volume I

Relating Events in Narrative, Volume 2: Typological and Contextual Perspectives

Time and Narrative Volume 2(Time & Narrative)

Information Structure: Theoretical, Typological, and Experimental Perspectives

Benefactives and Malefactives: Typological Perspectives and Case Studies (Typological Studies in Language)

Nominalization in Asian Languages: Diachronic and typological perspectives

Adpositions: Pragmatic, Semantic and Syntactic Perspectives (Typological Studies in Language)

Civil War, A Narrative Volume 2

Perspectives On Learning Disabilities: Biological, Cognitive, Contextual

Ancient Narrative 2 (2002)

Modality in typological perspective

New Perspectives on Narrative and Multimodality (Routledge Studies in Multimodality)

Rethinking Grammaticalization: New perspectives (Typological Studies in Language)

Time and Narrative, Volume 3 (Time & Narrative)

ADVANCES IN CATALYSIS VOLUME 2, Volume 2

Plant Developmental Biology - Biotechnological Perspectives: Volume 2

Selves and Identities in Narrative and Discourse (Studies in Narrative)

Perspectives of Nonlinear Dynamics (Volume 2)

Time and Narrative, Volume 3 (Time & Narrative)

Time and Narrative Volume 1(Time & Narrative)

Contextual analysis

Contextual Computing: Models and Applications

Relating Software Requirements and Architectures

Syntax: A Functional-typological Introduction: Volume II

Contextual Computing: Models and Applications

ADVANCES IN GEOPHYSICS VOLUME 2, Volume 2 (v. 2)

Events Design and Experience (Events Management)

Poetry and Narrative in Performance

International Perspectives of Festivals and Events: Paradigms of Analysis

Events in computation

Beyond Narrative Coherence (Studies in Narrative)

Distributist Perspectives: Volume I

Recommend Documents