Saami Linguistics (Amsterdam Studies in the Theory and History of Linguistic Science Series IV: Current Issues in Linguistics Theory)

Saami Linguistics AMSTERDAM STUDIES IN THE THEORY AND HISTORY OF LINGUISTIC SCIENCE General Editor E.F.K. KOERNER (Ze...

Author: Ida Toivonen | Diane Nelson

438 downloads 1870 Views 2MB 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

Saami Linguistics

AMSTERDAM STUDIES IN THE THEORY AND HISTORY OF LINGUISTIC SCIENCE General Editor E.F.K. KOERNER (Zentrum für Allgemeine Sprachwissenschaft, Typologie und Universalienforschung, Berlin) Series IV – CURRENT ISSUES IN LINGUISTIC THEORY Advisory Editorial Board Lyle Campbell (Salt Lake City); Sheila Embleton (Toronto) Brian D. Joseph (Columbus, Ohio); John E. Joseph (Edinburgh) Manfred Krifka (Berlin); E. Wyn Roberts (Vancouver, B.C.) Joseph C. Salmons (Madison, Wis.); Hans-Jürgen Sasse (Köln)

Volume 288

Ida Toivonen and Diane Nelson (eds.) Saami Linguistics

Saami Linguistics

Edited by

Ida Toivonen Carleton University

Diane Nelson University of Leeds

JOHN BENJAMINS PUBLISHING COMPANY AMSTERDAM/PHILADELPHIA

4-

The paper used in this publication meets the minimum requirements of American National Standard for Information Sciences — Permanence of Paper for Printed Library Materials, ANSI Z39.48-1984.

Library of Congress Cataloging-in-Publication Data Saami linguistics / edited by Ida Toivonen, Diane Nelson. p. cm. -- (Amsterdam studies in the theory and history of linguistic science. Series IV, Current issues in linguistic theory, ISSN 0304-0763 ; v. 288) Includes bibliographical references and index. 1. Sami language. I. Toivonen, Ida. II. Nelson, Diane Carlita, 1968PH706.S23 2007 494'.55--dc22 2007028374 ISBN 978 90 272 4803 9 (Hb; alk. paper) © 2007– John Benjamins B.V. No part of this book may be reproduced in any form, by print, photoprint, microfilm, or any other means, without written permission from the publisher. John Benjamins Publishing Co. • P.O.Box 36224 • 1020 ME Amsterdam • The Netherlands John Benjamins North America • P.O.Box 27519 • Philadelphia PA 19118-0519 • USA

ACKNOWLEDGEMENTS

This collective volume would not have been possible without the support of our friends and colleagues. The editors would like to thank the contributors, Pekka Sammallahti (Oulu University), and the reviewers for their invaluable comments on previous versions of papers included in this volume. We also thank Myrdene Anderson (Purdue University) as well as two anonymous scholars for serving as external referees. We would like to thank the following individuals for their support in editing this volume: Cecile de Cat, Kester Clegg, Bill Palmer, Ruth Payne-Woolridge, Michael Rießler, Tamir Stulberg, Anita Szakay and Nina Widjaja. We are also grateful to Anke de Looper at John Benjamins. Special thanks go to Ash Asudeh for writing some of our crucial LATEX style files and for helping us solve many of our formatting problems. This research was supported by University of Canterbury Research Grant U6566, SSHRC Standard Research Grant 410-2006-1650, British Academy Grants SG31040 and LRG 31734, and the University of Leeds SMLC Research Strategy Committee. This financial support is gratefully acknowledged. Diane Nelson and Ida Toivonen June 2007

CONTENTS

List of Abbreviations Introduction Diane Nelson & Ida Toivonen Etymological Nativization of Loanwords: A case study of Saami and Finnish Ante Aikio

viii 1

17

Grade Alternation in Inari Saami and Abstract Declarative Phonology Patrik Bye

53

Productive Syncretism in Saami Inflectional Morphology Gunnar Ólafur Hansson

91

Roots and Verbs in North Saami Marit Julien The Instability of Systems with Ternary Length Distinctions: The Skolt Saami evidence Zita McRobbie-Utasi

137

167

Events and Case in Inari Saami Diane Nelson

207

Verbal Agreement in Inari Saami Ida Toivonen

227

Saami Linguistics Bibliography Compiled by Anita Szakay, Michael Rießler, Ida Toivonen, Diane Nelson, Zita McRobbie-Utasi & Nina Widjaja

259

Index

305

LIST OF ABBREVIATIONS

The following is a list of abbreviations that are used in examples. Other abbreviations are explained as they are presented.

abessive case ablative case accusative case adessive case accusative/genitive case category causative classifier comitative case dative case definite diminutive dual elative case essive case frequentative genitive case gender illative case imperative imperfective aspect indicative inessive case infinitive intransitive locative case masculine

neuter nominative nonfinite number oblique partitive case passive past tense participle person plural potential mood prepositional phrase preterite perfective aspect present tense prolative case pronoun question particle singular subjunctive subject tense transitive

INTRODUCTION

DIANE NELSON & IDA TOIVONEN University of Leeds — Carleton University

1. Introduction This volume is a collection of papers that present and investigate various linguistic aspects of the Saami languages.1 The purpose of the volume is twofold: the primary purpose is to create a venue for new research in Saami linguistics; in addition, the volume is intended to serve as an introduction to the Saami languages for linguists. Although there exists a rich literature on Saami (see the bibliography at the end of this volume), many of the publications are written in Nordic languages and so are not accessible to much of the international linguistics community. Also, a large part of the literature is quite old. This does not mean that this literature is not valuable, but the discussion is often cast against a background of outdated assumptions. Furthermore, even though the existing literature is rich and invaluable for anyone interested in these languages, many fascinating topics remain underexamined. This volume is not an introduction in the sense that it provides an overview of the structure of the languages (for an excellent overview, see Sammallahti 1998); instead, it includes a number of papers that each puts a single linguistic phenomenon under the microscope and illustrates that phenomenon with a large number of Saami examples. The papers in this volume draw attention to some of the main phenomena in the Saami languages which are of interest for current linguistic theory. The Saami languages divide into three groups: the northern group of Western Saami (Lule, Pite and North Saami); the southern group of Western Saami (South and Ume Saami); and Eastern Saami (Inari, Skolt, Akkala, Kildin and 1 In older literature, Saami is called Lapp or Lappish, but this is now dispreferred. Different authors adopt different spellings of the word Saami: Sámi and Sami are common variants.

2

DIANE NELSON & IDA TOIVONEN

Ter Saami). The Saami languages are shown with their regional distribution in Table 1, which also shows that North Saami has by far the most speakers. The number of speakers has not been indicated for languages where only a handful of speakers remain, if any.

Western, South South Saami Ume Saami Western, North Pite Saami Lule Saami North Saami Eastern Inari Saami Skolt Saami Akkala Saami Kildin Saami Ter Saami

400 Sweden, Norway Sweden, Norway Sweden, Norway 2,000 Sweden, Norway 17,000 Norway, Sweden, Finland 350 Finland 400 Finland, (Russia) Russia 700 Russia Russia

Table 1: The distribution of the Saami languages The geographical areas where the different languages are spoken are indicated on the map in Figure 1. Historically, speakers of the Saami languages occupied a much larger area, that extended into southern Finland and central Sweden until the Middle Ages. Most Saami speakers today are bi- or multilingual. The language contact situation is quite interesting, as the contact languages belong to different language families: Swedish and Norwegian are Germanic languages, Russian is Slavic, and Finnish is Finno-Ugric. Of the contact languages, only Finnish is related to Saami, as Saami is also Finno-Ugric, although only distantly related to Finnish. The Saami languages are thus of great interest for studies in language contact, and language variation and change. Ante Aikio’s paper ‘Etymological nativization of loanwords’ touches on these issues; however, most of the papers in this volume are limited to topics on linguistic structure. In the remainder of this introduction, we will briefly introduce some of the distinguishing characteristics of Saami. Specifically, the focus will be on three areas: consonant gradation, or qualitative and quantitative consonant alternations in certain morphological environments (section 2); the nature of morphological paradigms in Saami (section 3); and the marking of grammatical and semantic functions (section 4). Examples are drawn from several of the Saami languages.

Figure 1: The geographical distribution of the Saami languages

4


2. Consonant gradation The Saami languages display spectacular consonant gradation; that is, systematic alternations between two consonants within certain morphosyntactic environments. Consonant gradation is often associated with compensatory vowel lengthening and can be quantitative (1), qualitative (2), or both (3).2 (1)

(2)

(3)

a.

njun.e ‘nose. ’ – njune ‘nose.

’

b.

vavvoim ‘(we) suspected’ – vav.om ‘(I) suspect’

c.

kiellân ‘language.’ – kiel.â ‘language. ’ – kielâ ‘language. ’

d.

kumppi ‘wolf. ’ – kuumpi ‘wolf.

a.

vaho ‘s/he suspects’ – vav.om ‘(I) suspect’

b.

lahcâ ‘cream. ’ – laavcâ ‘cream.

’

c.

’ tupe ‘cabin. ’ – tuve ‘cabin. ’

a.

päikki ‘home. ’ – pääihi ‘home.

b.

saaˇccˇ âd¯ ‘push.’ – saaj.âm ‘(I) push’

’

Historically, gradation was phonologically predictable, but the system today is complex, and massive memorization of individual forms during acquisition may seem unavoidable. However, Patrik Bye (this volume) argues that the synchronic gradation patterns in Inari Saami are actually still phonologically (or ‘morphophonologically’) predictable. At least some of the Saami languages — for example, Inari Saami, Lule Saami, North Saami and Skolt Saami — have a three-way consonant length distinction, as discussed in Collinder (1952), Magga (1984), Sammallahti (1977), McRobbie-Utasi (1999), Engstrand (1987), Itkonen (1971), and elsewhere. Ternary length contrasts are very rare cross-linguistically and not easily accomodated in phonological theory; claims of three-way length distinctions have always been accompanied by controversy (as discussed by Zita McRobbie-Utasi, this volume). The data raise the question of whether the ternary length distinction is phonologically real, or whether the length distinctions are phonetic sideeffects of other distinctions. Bye’s paper treats the contrasts seen in the Inari Saami nominal paradigms entirely with reference to binary features. 2 The examples in (1–3) are Inari Saami. A letter with a dot represents a half-long sound; for example n. is a half-long [n]. The vowel â is a mid, central, unrounded vowel.

INTRODUCTION

5

We will here present the results of a recent phonetic study of Inari Saami which clearly reveals the phonetic reality of the ternary distinction. The results shown here are taken from Toivonen & Hay (2005) and Toivonen (2006). For this study, four Inari Saami speakers produced disyllabic words that were all of a CVCV(C) template. The words were recorded with a DAT-recorder. The words that were analyzed are of three main types: the first type has a short middle consonant, the second type has a half-long middle consonant and the third type has a long middle consonant. The speakers produced several words of each type, and all words were repeated at least four times in a carrier phrase. The length measurements were made using PRAAT acoustic analysis software. Figure 2 shows the average lengths of long, half-long and short consonants for all four speakers. The three-way distinction is highly significant in the speech of all four speakers (p < .001).3 Regardless of the phonological status of quantity in Inari Saami, our results show that the ternary length distinction is robust phonetically. Zita McRobbie-Utasi’s paper examines the ternary quantity distinction in Skolt Saami. McRobbie-Utasi focusses on the fact that quantity involves more than just duration. She points out the importance of other phonetic correlates alongside duration, and she also presents a cross-linguistic overview of ternary quantity contrasts. Quantity is of course only one interesting aspect of Saami consonant gradation. Though the details differ, each of the Saami languages displays an intricate system of gradation, which is interesting both from a synchronic and diachronic perspective. The consonant gradation is relevant for the phonetics, the phonology and the morphology of Saami, and is therefore of central importance.

3

For the statistical analysis, we used the R statistics software to fit a series of ordinary least square (OLS) regression models. The models were fit by hand, starting with a saturated model, and removing factors that did not reach significance. The words included all have phonologically short vowels. We controlled for the quality of the first and second vowels, as well as the quality of the consonant itself. We also controlled for whether the second syllable was open or closed.

lengthC

lengthC

300

250

200

150

100

50

0

200

150

100

long

−

●

−

long

−

●

−

Ctype

half-long

short

−

●

−

short

−

●

−

long

−

●

−

long

−

●

−

Figure 2: Consonant length in Inari Saami

Ctype

half-long

−

●

−

●

− −

lengthC lengthC

200 150 100 200 150 100 50

Ctype

half-long

−

●

−

Ctype

half-long

−

●

−

short

−

●

−

short

−

●

−

6 DIANE NELSON & IDA TOIVONEN

INTRODUCTION

7

guolli guolit guoli guliid guollái guliide guolis guliin guliin guliiguin ! guollin Table 2: The paradigm for guolli ‘fish’ (North Saami) 3. Morphological paradigms and processes The Saami languages have a rich morphology. The inflectional paradigms mark a large number of morphosyntactic distinctions, and they involve extremely complex morphophonology. Both nominal and verbal paradigms include large numbers of different stem classes. For example, Inari Saami has no fewer than 11 nominal and 15 verbal major stem classes according to Sammallahti & Morottaja (1993). The nominal paradigms display rich case marking and two declensions (absolute and possessive). An example absolute (that is, non-possessive) paradigm is given in Table 2 (North Saami, adapted from Sammallahti 1998:64). Pronouns inflect for three numbers (singular, dual and plural), but non-pronominal nouns only distinguish between singular and plural. In Saami, a possessor can be expressed as a prenominal genitive noun or pronoun (as in English). A pronominal possessor can also be expressed as a possessive suffix. For example, North Saami ‘my guest. ’ is guos"sán and ‘your. guest. ’ is guos"sát. A partial possessive paradigm from North Saami is given in Table 3 (adapted from Sammallahti 1998:65).4 The full possessive paradigm in North Saami marks the two possible numbers and six cases of the head noun, and the three numbers and three persons of the pronominal possessor. The possessor marking follows the case marking, except in comitative plural: ‘my guests. ’ is gussiidanguin, where –guin is the case ending. Nominative plural is uncommon in the possessive declension: it appears mainly in religious texts and sometimes as a vocative. Across the Saami languages the verbal inflectional paradigms contain a greater range of forms than the nominal ones. Verbs are conjugated for four tenses (past, present, perfect and pluperfect) but two of those tenses, the perfect and the 4

The dotted vowel e. marks final stress in the preceding diphthong.

8


1 guos"sán

2 3 1 2 3 1 2 3

guos"sát guos"sis guos"sáme guos"sáde guos"siska guos"sámet guos"sádet guos"set

guossistan guossistat

gussiidasan gussiidasas

guossisteame. guossisteatte.

gussiidasame gussiidasaska

guossisteame.t guossistette.t

guussiidasamet gussiidasaset

Table 3: A partial possessive paradigm for guos"si ‘guest’, North Saami pluperfect, are compound tenses. The verb is further conjugated for four moods (indicative, conditional, potential and imperative). The Lule Saami present and past tense indicative paradigms for the verb bårråt ‘to eat’ are given in Table 4, which is adapted from Spiik (1989:79). The table illustrates that the person endings are not identical in the different tenses.

1 båråv

bårriv 2 bårå bårri 3 bårrå båråj 1 bårrin båråjma 2 bårråbihtte båråjda 3 bårråba båråjga 1 bårråp båråjma 2 bårråbihtit båråjda 3 bårri bårrin

Table 4: The present and past tense paradigms for bårråt ‘to eat’, Lule Saami As is the case in other Finno-Ugric languages, Saami negation is realized syntactically as an auxiliary verb that hosts inflection. The languages fall into two groups on the basis of which verbal features are encoded by the negative auxiliary. Table 5 provides some examples. The western languages (except North Saami) mark the tense, mood and agreement on the negative auxiliary, while the main verb is uninflected. In the eastern languages and in North Saami, the nega-

9

INTRODUCTION

"

(Western) (manne) im daejrieh (datne) ih daejrieh (manne) idtjim daejrieh (datne) idtjih daejrieh

(Eastern) (mun) jie"m tie"de ¯ (tun) jie"h tie"de ¯ (mun) jie"m tiättám (tun) jie"h tiättám

‘(I) don’t know’ ‘(you.# don’t know’ ‘(I) didn’t know’ ‘(you.) didn’t know’

Table 5: Verbal negation, present and past tense tive auxiliary verb inflects for person and number agreement, and the tense and mood is shown on the main verb. The Saami languages also employ morphosyntactic processes that affect the expression of argument structure, including passive and causative. The productivity and syntax of these processes varies from language to language within the group. For example, in the Torne dialect of North Saami, causative morphology introduces a causer argument while the causee surfaces in accusative case (Svonni & Vinka 2003:343-344). This is illustrated in (4–5); Svonni & Vinka’s (1–2). (4)

cuvkii láse Máhtte Máhtte. break. $ window. “Máhtte broke the window.”

(5)

Máhte Mon cuvke-h-in I. break. % % & Máhtte. “I caused Máhtte to break the window.”

láse. window.

A causative verb can then form the input for passivization (Svonni & Vinka 2003:356); example (6) is Svonni & Vinka’s example (21b). (6)

cuvke-h-uvvu-i. Láse window. break. % % % “The window was caused to be broken.”

$

In (6), the underlying causee argument Máhtte is removed under passivization, and the underlying accusative object láse ‘window’ resurfaces in nominative case. Another set of verbal affixes affects a verb’s aspectual properties. The North Saami affix -stit, for example, when attached to a verb of accomplishment normally yields a reading equivalent to ‘a shorter time period than usual’ for the event (Hedlund 2006). This has the effect of restricting the range of temporal adverbials that can modify the predicate, as illustrated in (7).

10 (7)


a.

Máhtte logastii girjji guovtti diimmun. Máhtte read. $ book. two. hours. “Máhtte read the book in two hours.”

b.

*Máhtte logastii girjji guokte diimmu. Máhtte read. $ book. two. hours. “Máhtte read the book for two hours.”

In addition to the case paradigms discussed above, Saami has rich nominal derivational morphology, for example yielding diminutives (8) and augmentatives (9). Examples (8–9) are from Kildin Sami (Riessler 2006). (8)

(9)

a.

koabp" ‘ditch’

b.

koab"-a ‘small ditch’

c.

koab"-enˇc ‘dimple’

a.

perrt ‘house’ ˚˚ pert-pihk ‘large house’ ˚

b.

Some of the Saami languages allow reduplicated ‘double diminutives’. The examples in (10) are from North Saami (Nickel 1994:339). (10)

a.

nieidda-š girl- “little girl, little daughter”

b.

nieidda-ža-š girl- - “tiny little girl”

Other derivational affixes are associated with changes of major category. Julien’s paper in this volume offers an analysis of the formation of denominal and deadjectival verbs in North Saami. Saami verbs also host a wide range of non-finite forms and nominalizing affixes. Infinitive verb forms appear with the suffix /-t/ and are selected as complements by various classes of matrix verbs (11) (North Saami, Sammallahti 1998:85), or in copular predicates (12) (Kildin Saami, from Szabó 1985:57). (11)

m¯u boahtit. Dat doaivv¯ui S/he think.$ I. come. “S/he thought that I would come.”

INTRODUCTION

(12)

11

šieg˘a jiel"l"e Tann˘a lei then be.$ good live. “Then it was good to live.”

A range of non-finite verbs and gerunds are derived from verbs with participles and nominalizing morphology. Several of the participial affixes encode aspectual or temporal information relative to the main clause verb. Example (13) shows the Kildin Saami ‘action’ participle affix /-Em, -am/ (Szabó 1985:109). (13)

UjnsEn aijes" puEdd-em. see$ grandfather come- “They saw the grandfather coming.”

Some of these forms may host inflection for case and/or possessive agreement, as illustrated in the Inari Saami example in (14). (14)

Tun juurdáh cˇ uážž-um-âd lihâ-hánnáá. you. believe.' stand--2 move- “You believe you stood without moving.”

In the example above, the non-finite form cˇ uážžumâd includes a verb stem with a participial affix –um, plus the second person possessive affix âd. The form lihâhánnáá includes the affix -hánnáá, an abessive participle.5 The data presented in this section barely scratch the surface of Saami morphology. Its importance in Saami grammar is reflected by the fact that most of the papers of this volume are concerned with some aspect of morphology. The papers by Patrik Bye, Gunnar Hansson and Marit Julien make morphology their central focus: Bye’s paper addresses grade alternation in the nominal paradigms of Inari Saami; Hansson’s paper is concerned with syncretism in the North Saami case paradigms; and Julien’s paper looks at the syntax and semantics of verbal derivation in North Saami. 4. Grammatical and semantic functions Distinctions between separate grammatical functions among nominals are cross-linguistically marked with word order, agreement marking (head-marking) and/or case marking (dependent marking). All three types of grammatical marking are drawn upon in Saami, although case marking provides the most salient cue to the grammatical or semantic function of noun phrases. 5 Although paradigms exist for case and possessive agreement in non-finite verbs, these forms are not frequent in spoken Saami.

12


The basic word order for most Saami languages is SVO in finite clauses, but both SOV and SVO are common in non-finite clauses. In South Saami, SOV is the basic order in finite and non-finite clauses. However, word order is not rigid in Saami; instead, it tends to be conditioned by discourse function, and grammatical relations within the clause are signalled primarily by case and agreement marking. The Saami languages, with their rich verbal agreement, allow nonovert subjects (“pro-drop”). The different Saami languages have at most nine cases. The partitive and abessive have gone out of use in some of the languages; as a generalization, those two cases are still in regular use in the Eastern Saami languages, but not in the Western Saami. Across the Saami languages, the morphological marking for accusative and genitive has undergone syncretism, especially in the singular. For example, in Inari Saami, the distinction between accusative and genitive case can be seen in the plural only. The examples in (15–23) illustrate the nine cases in Inari Saami. (15)

vyej.i-min. Mun oinim suu I. saw him/her drive- “I saw him/her drive.”

(16)

Nieidâ ustev poodij. ¯ girl. friend. arrived “The girl’s friend arrived.”

(17)

Mun oinim kuumpijd. I. saw wolves. “I saw the wolves.”

(18)

Muoi adelijn hiäppušijd kandáid. we. gave horses. boys. “We gave horses to the boys.”

(19)

Kyesih cˇ okkájeh avdost. guests. sit car. “The guests are sitting in the car.”

(20)

Tuoi Ánnáin puátivettee meecist suuin. you. Anna. come forest. him/her. “You and Anna are coming from the forest with him.”

INTRODUCTION

(21)

poodijn Suoi ¯ maassâd nisonijttáá. women. they. came back “They returned without the women.”

(22)

Sij jurdii muorâ táállun. they. thought tree. house. “They thought the tree was a house.”

(23)

Muor.â lii táállud ucceeb. tree. is house. smaller “The tree is smaller than the house.”

13

Table 6 gives a rough outline of the main functions of the cases. The semantic and syntactic functions of case marking have received considerable attention in the Saamic literature (see Bartens 1972 and Sammallahti 2005, for example). The role of case marking is also discussed in several of the papers included in this volume, most notably in Diane Nelson’s paper, which concerns case marking on adverbials in Inari Saami. In addition, Ida Toivonen’s paper investigates the case marking of subjects and objects.

!

!(

)

mun ‘I’ nieidâ ‘girl’ kuumpijd ‘wolves’ kandáid ‘boys’ avdost ‘car’ Ánnáin ‘Anna’ nisonijttáá ‘women’ táállun ‘house’ táállud ‘house’

subject of finite verb prenominal possessor object to, into, onto on, at, in; from with without as X, becoming X after numbers 7+, comparatives

Table 6: The functions of the morphological cases, Inari Saami Toivonen’s paper also discusses the case marking of predicative possessors. Predicative possessors bear locative case, and the verb is a copula. This is shown in the Lule Saami example in (24), taken from Lindbach (2006). (24)

Máhtu-n li girje. Máhttu- be.$ book. “Máhttu has (got) books.”

14


In some of the Saami languages, predicative possession can also be signalled lexically with certain verbs. North, South, Pite, Ume and Lule Saami all use a form of the verb atnit ‘to use’ to signal possession (25), as illustrated by example (25) from Lule Saami (Lindbach 2006). (25)

adná bednagijt. Máhttu Máhttu. use.$ dog. “Máhttu has (got) dogs.”

Agreement marking has a more limited use than case marking in Saami, and the role and function of agreement marking has therefore received less attention in the literature. Finite verbs in Saami agree with the subject in three persons and three numbers; singular, dual and plural (see Table 4 above). Certain pronouns also inflect for three numbers, but non-pronominal nouns are marked for singular and plural only. Inanimate subjects cannot in general trigger dual agreement on the verb. This is discussed in Ida Toivonen’s paper, which also points to further restrictions on dual agreement marking. 5. Concluding remarks The papers in this volume discuss a number of Saami characteristics that are cross-linguistically unusual, for example partial agreement on verbs and a threeway quantity distinction in consonants. We hope that this volume may draw attention to this fascinating group of languages and inspire linguists to find out more, which is indeed possible: most of the languages have been described in grammars, textbooks and/or dictionaries. This volume includes a bibliography on much of the available literature on Saami linguistics.6 In addition to the growing literature, Saami language courses are offered throughout Scandinavia at universities and other institutions. There are also collections of texts and original literature, and some of the Saami languages are represented on radio and television. In sum, even though the languages are endangered, there is a growing body of Saami literature and other materials of interest to linguists. References Bartens, Raija. 1972. Inarinlapin, merilapin ja luulajanlapin kaasussyntaksi. (= Mémoires de la Société Finno-Ougrienne 148.) Helsinki: SuomalaisUgrilainen Seura. 6 The Saami bibliography was compiled by Anita Szakay, Michael Rießler, Ida Toivonen, Diane Nelson, Zita McRobbie-Utasi and Nina Widjaja.

INTRODUCTION

15

Collinder, Björn. 1952. “Three degrees of quantity”. Studia Linguistica 5.28–43. Engstrand, Olle. 1987. “Durational patterns in Lule Saami phonology”. Phonetica 44.117–128. Hedlund, Cecilia. 2006. “The North Sámi derivational suffix –stit”. Paper presented at The Saami Linguistics Symposium, University of Tromsø, November 2006. Itkonen, Erkki. 1971. “Ehdotus inarinlapin fonemaattiseksi transkriptioksi”. Lapin murteiden fonologiaa ed. by Erkki Itkonen, Terho Itkonen, Mikko Korhonen & Pekka Sammallahti. (= Castrenianumin toimitteita 1.), 43–67. Helsinki: Suomalais-Ugrilainen Seura. Lindbach, Kristin L. 2006. “The habitive construction”. Paper presented at The Saami Linguistics Symposium, University of Tromsø, November 2006. Magga, Tuomas. 1984. Duration in the Quantity of Bisyllabics in the Guovdageaidnu Dialect of North Lappish. Oulu: University of Oulu. McRobbie-Utasi, Zita. 1999. Quantity in the Skolt (Lappish) Saami Language: An acoustic analysis. Bloomington: Indiana University. Nickel, Klaus Peter. 1994. Samisk grammatikk. Karasjok, Norway: Davvi Girji. Riessler, Michael. 2006. “Russian Grammar in Kildin Saami”. Paper presented at The Saami Linguistics Symposium, University of Tromsø, November 2006. Sammallahti, Pekka. 1977. Norjansaamen Itä-Enontekiön murteen äänneoppi. (= Mémoires de la Société Finno-Ougrienne 160.) Helsinki: SuomalaisUgrilainen Seura. . 1998. The Saami Languages: An introduction. Karasjok, Norway: Davvi Girji. . 2005. Láidehus sámegiela cealkkaoahpa dutkamii. Karasjok, Norway: Davvi Girji. Sammallahti, Pekka & Matti Morottaja. 1993. Säämi-suomâ sänikirje. Utsjoki, Finland: Girjegiisá. Spiik, Nils Erik. 1989. Lulesamisk grammatik. London: George Allen and Unwin. Svonni, Mikael & Mikael Vinka. 2003. “Constraints on the morphological causatives in the Torne dialect of North Sámi”. Generative Approaches to Finnic and Saami Linguistics ed. by Diane Nelson & Satu Manninen, 343– 380. Stanford, Calif.: CSLI Publications. Szabó, László. 1985. “Infinitive and participial constructions in Kola Sami”. Nordlyd: Tromsø University Working Papers on Language and Linguistics 10.28–127.

16


Toivonen, Ida. 2006. “A phonetic study of consonant length in Inari Saami”. Paper presented at The Saami Linguistics Symposium, University of Tromsø, November 2006. Toivonen, Ida & Jen Hay. 2005. “The effect of vowel length on a ternary consonant length distinction: A phonetic study of Inari Sami”. Paper presented at The Annual Meeting of the Linguistic Society of America, Oakland, Calif., January 6–9 2005.

ETYMOLOGICAL NATIVIZATION OF LOANWORDS∗ A CASE STUDY OF SAAMI AND FINNISH

ANTE AIKIO University of Oulu Speakers bilingual in two genetically related languages may become aware of regular sound correspondences between the languages, and this gives an impetus to mimic the observed patterns by adapting loanwords to these correspondences. Due to this process of etymological nativization, even young loanwords can display correspondences that are not phonetically transparent. The study presented in this paper shows that loanwords between Saami and Finnish, two not very closely related branches of the Uralic family, show vowel substitution rules such as i > a and a > i, which lack a phonetic motivation but conform to correspondences in inherited vocabulary. Such phenomena require a revision of the criteria commonly applied in the dating and stratification of loanwords: in a case of intensive borrowing between two related languages, sound correspondences do not always allow one to consistently distinguish between older and younger loanwords or even between loanwords and cognate items.

1. Introduction It is well known that speakers acquainted with two dialects of their native language can acquire an intuitive understanding of the phonological correspondences between the varieties, with the result that they are able to regularly transform forms of one dialect into the other. In fact, some dialect jokes are based on this phenomenon. It has also been at least implicitly recognized in foreign language pedagogy that this phenomenon works not only between dialects but also between closely related languages. Estonian textbooks for native speakers ∗

This study is dedicated to the memory of the eminent comparative linguist Larry Trask. I remain obliged to him for advising me to write this paper and for pointing out valuable references. I also wish to thank Petri Kallio, Pekka Sammallahti, Jussi Ylikoski and the anonymous referees for valuable remarks on an earlier manuscript.

18

ANTE AIKIO

of Finnish explicitly instruct the student to form Estonian words on the basis of their Finnish cognates by applying processes such as apocopation, syncopation, shortening of unstressed long vowels, and abolition of vowel harmony, which directly correspond to sound changes that have taken place during the history of Estonian. For instance, Kasik’s Estonian textbook (1991) begins with a description of eight consonant and nine vowel rules by which Finnish word-forms can be regularly transformed into their Estonian cognates. After this there is an exercise where the student must convert entire Finnish sentences into Estonian by applying these rules, as in (1). (1)

Minä olen uusi eestin kielen opettaja. (Finnish) Mina olen uus eesti keele õpetaja. (Estonian) “I am the new Estonian teacher”

The rules applied in the (1) include abolition of vowel harmony (minä > mina), apocope after a long initial syllable (uusi > uus), deletion of the genitive singular ending -n (eestin kielen > eesti keele), monophthongization of falling diphthongs (kielen > keele), the vowel shift o > õ, and the shortening of geminates in suffixal syllables (opettaja > õpetaja). What is less well known is that similar recognition of regular sound correspondences is occasionally witnessed also among speakers bilingual in languages that are only remotely related, and that this tends to have an effect on how loanwords transferred between the two languages are phonologically nativized: borrowed items can be adopted in a form which partially or entirely conforms to the sound correspondences that occur in shared vocabulary. This phenomenon has been discussed in a couple of textbooks and case studies under terms such as ‘etymological nativization’ and ‘correspondence mimicry’, but the information is scattered and there exists no detailed account on the subject. This paper presents a case study on a historical language contact situation which reflects widespread recognition of etymological correspondences — the contact between Saami and Finnic, two rather remotely affiliated branches of the Uralic (Finno-Ugric) language family. The next section provides a summary of references about this phenomenon. Section 3 provides an analysis of the substitution patterns of vowels in loanwords between Saami and Finnish. The concluding section examines what implications the results of the present study have for the methodology of comparative linguistics.

ETYMOLOGICAL NATIVIZATION OF LOANWORDS

19

2. A summary of the literature The existence of etymological nativization has been noted in a couple of handbooks. The definition provided by Larry Trask in his Dictionary of Historical and Comparative Linguistics can be taken as a starting point: When there is widespread bilingualism between speakers of two closely related languages, speakers will often be keenly aware of the phonological and morphological correspondences holding between the two languages. In such circumstances, a loan word may be nativized replacing each of its segments with the regularly corresponding segment in the borrowing language [...] As a result, the borrowed items may be indistinguishable from native formations [...] (Trask 2000a s.v. loan nativization)

According to Trask, the names ‘loan nativization’, ‘loan adaptation’ and ‘correspondence mimicry’ have been applied to the phenomenon; the first two of these terms are ambiguous and therefore do not seem to be suitable. The same phenomenon has also been briefly discussed by H. H. Hock (1986:392-393), who calls it ‘etymological nativization’; the term coined by Hock seems particularly apt. Hock provides a couple of examples from Slavic and Celtic, including the substitution of Pre-Irish *q- for Pre-Welsh *p- in borrowed vocabulary, such as the personal name *Qatrikias < Pre-Welsh *Patrikios (< Latin Patricius). The model is provided by the correspondence q ∼ p in cognate items such as Pre-Irish *q¯e ‘who’ ∼ Pre-Welsh *p¯e id. Primary studies that treat etymological nativization appear to be scarce. Nash (1997) discusses lexical relations between various Australian aboriginal languages in the Northern Territory, providing also examples of etymological nativization — or ‘correspondence mimicry’ in Nash’s terminology. Loanwords between languages in the Pama-Nyungan family have become adjusted to consonant correspondences resulting from the so-called ‘initial dropping’ that has taken place in certain languages. The initial-dropping languages have lost all instances of original word-initial consonants. Speakers that are bilingual in an initial-dropping and a neighboring non-initial-dropping language tend to recognize the regular correspondence, and mimic it in loanwords so that word-initial consonants become dropped when borrowed into an initial-dropping language from a non-initial-dropping one. This example of etymological nativization in Australian languages is also mentioned by Koch (1997:35) in the same volume. A brief discussion on the phenomenon in certain Australian languages is also provided by Alpher & Nash (1999:14-15). Leer (1990:86-88) discusses the relation of Tlingit to its remote relatives, the Athabaskan languages and Eyak, and points out that certain words borrowed

20

ANTE AIKIO

from Athabaskan to Pre-Tlingit show puzzling consonant substitutions which have no phonetic or structural motivation. Referring to an unpublished paper by Michael Krauss, Leer suggests that etymological nativization (which he calls ‘loan adaptation’) may be responsible for the phenomenon. However, because Leer’s explanation involves the positing of unattested prehistoric Tlingit dialects, etymological nativization in Tlingit is presumably best regarded as a hypothesis rather than a solidly documented case. Grace (1996) provides an interesting description of how the application of the comparative method to certain apparently closely related Austronesian languages of New Caledonia has ‘failed’; the lexical material shared by these languages contains so many recurring sound correspondences that it is not possible to set up sound laws and reconstruct the proto-language. Grace considers this state of affairs the result of a Sprachbund with widespread bi- and multilingualism, where boundaries of linguistic communities have not coincided with language boundaries, and suggests that also etymological nativization (even though he does not use this term) has taken place during the course of extensive borrowing and interference: What I want to propose is that to the extent that there is diffusion of signantia from any particular phonological dialect to any other, there will be regular sound correspondences which provide the basis of conversion of such signantia from one to the other. And this will be true either if the two phonological dialects are virtually indistinguishable ones spoken, say, by next-door neighbors or if they are fundamentally different ones — for example, belonging to mutually unintelligible languages. (Grace 1996:175)

Interestingly, there is also evidence suggesting that etymological nativization can occur between unrelated languages. Comrie (1993) briefly mentions what he calls ‘traditions of sound correspondence’ in loanwords adopted to Persian from Arabic. As the languages in question are not genetically related, the patterns of etymological nativization can only be modeled after words reflecting an earlier period of borrowing succeeded by further sound changes in the source and target languages. Mistakenly, though, Comrie considers these kinds of patterns to be influenced by literary language; but the existence of similar processes in Australian and other languages demonstrates that etymological nativization is a result of phonological processing and not attributable to the influence of archaic spelling. Regrettably, Comrie provides no concrete examples or further references. Another case of etymological nativization between unrelated languages is reported by Trask (2000b:53-54). He analyses the Basque word zeru ‘sky’ as


21

a borrowing from an Early Romance development of Latin caelum id.; the borrowing is clearly post-Latin, as it shows a reflex of the spirantization of Latin /k/ before a front vowel. However, Latin u was lowered into o in Romance before the spirantization took place, and hence the final -u in zeru is unexpected, especially as there are loanwords older than zeru that display Basque -o in the place of Latin u (for example, Basque bago ∼ pago ‘beech’, cf. Spanish haya < Latin fagus). The substitution of Basque u for Romance o in zeru has no phonetic or structural explanation, as final -o is quite normal in Basque. Trask proposes that the substitution was modeled after the correspondences in the oldest layer of loanwords which were adopted before the shift u > o in Romance (for example, Basque liburu ‘book’, cf. Spanish libro). A further example provided by Trask is the substitution of Basque final -oi for Spanish -ón in recent loanwords such as Basque abioi ‘airplane’ < Spanish avión, Basque kamioi ‘lorry, truck’ < Spanish camión. This substitution is not phonetically or structurally motivated, as word-final -on is fully permitted in Basque. Instead, this pattern has also arisen due to the analogical influence of earlier borrowings. Spanish -ón reflects Early Romance *-one (< Latin -onem), which was borrowed into Pre-Basque as *-one and then underwent regular loss of intervocalic *n: Pre-Basque *-one > *-oe > Modern Basque -oi. The development is attested in such Romance loanwords as Basque arratoi ‘rat’ < *arratone < Romance *ratone (> Spanish ratón), Basque arrazoi ‘reason’ < *arrazone < Romance *ratsone (> Spanish razón). Finally, we can turn to the case of Saami. Erkki Itkonen provides a vivid description of how the regular correspondences between Finnish and Saami are recognized by bilingual persons: The relationship between Finnic and Lapp is special in that without particular practice the representatives of these language groups cannot understand each others’ speech at all, but the identity of hundreds of common Finnic-Lapp words and tens of common inflectional forms can immediately be recognized by one acquainted with the sound correspondences between the languages, as if deciphering some code sound by sound into another cipher system. To one’s astonishment, among Lapps one occasionally gets to notice how intelligent bilingual individuals intuitively master this system since childhood. In the year 1931 in Lemmenjoki, Inari, a 12-year-old boy remarked that Lapp gâpper [= gahpir] ‘cap, hat’ is the same word as Finnish kypärä [‘helmet’]. (Itkonen 1961:53; translated from Finnish)

Itkonen does not explicitly discuss how this intuitive understanding of the relationship between Finnish and Saami affects the phonological nativization of

22

ANTE AIKIO

loanwords. But as might be expected, the sound substitutions in Finnish loanwords in Saami are significantly influenced by the perceived correspondences in cognate vocabulary, as the data presented in Section 3 will reveal. Even though this phenomenon has hardly ever been explicitly discussed in the literature, it is clear that specialists in Saami linguistics have implicitly recognized it. For example, Korhonen (1981:105) notes that certain Finnish second-syllable vowels have regularly undergone etymological nativization in borrowings into Saami (see 3.6 for examples and discussion), and calls it “predictable” that the substitutions have been modeled after the correspondences in cognate vocabulary. To sum up, there is evidence of etymological nativization in a number of languages around the world — at least Saami, Slavic, Celtic, Pama-Nyungan, Basque and Persian, possibly also Austronesian and Tlingit. The fact that there is scattered information from diverse languages in many parts of the world suggests that the phenomenon may in fact be relatively common, even though its existence has not been widely recognized. On the basis of the examples discussed, one adjustment can be made to Trask’s definition of etymological nativization quoted above. It is not necessary for the contacting languages to be closely related — for example, Saami and Finnish cannot be considered closely related — or even genetically related at all, as demonstrated by the treatment of Spanish loanwords in Basque. All that is needed is a sufficient number of shared vocabulary displaying a given sound correspondence that provides a model for sound substitution. Whether this vocabulary reflects genetic relationship or merely an earlier period of borrowing is of no significance. It must be noted that while ‘etymological nativization’ as defined here is a phonological phenomenon, the other terms applied in research have not been used in exactly the same sense. Leer defines ‘loan adaptation’ so that it comprises also morpheme-to-morpheme substitutions and calques. Also ‘correspondence mimicry’ has been applied to morphological processes: Evans (1998) discusses the origin of initial consonant mutation in the Australian language Iwaidja, and proposes a development which involves an “extreme case of [...] correspondence mimicry” (p. 143). However, what he refers to is apparently an instance of crosslanguage morphosyntactic analogy: Evans concludes that the distribution of the so-called ‘miscellaneous prefix’ in Iwaidja has been remodeled according to the corresponding possessor-gender forms in the neighboring languages Ilgar and Maung.


23

Spanish: ratón razón avión Basque: arratoi arrazoi x (= abioi) Table 1: Spanish -ón > Basque -oi It is necessary, though, to make a distinction between two types of substitutions: those that mimic etymological correspondences between the contacting languages on the one hand, and those based on the semantic-functional correspondences of morphemes on the other. ‘Etymological nativization’ must be defined as the first type of case, and from this it follows that etymological nativization can only occur in the realm of phonology. Between morphemes there are always also semantic-functional correspondences due to which it is often impossible to decide whether a given morpheme substitution results from etymological or semantic-functional factors. A single example will illustrate this. Finnish kelvoton ‘worthless, useless’ (consisting of kelpo ‘good, fine, decent’ + the caritive suffix -ton) has been borrowed into North Saami as gealbboheapme, the Saami caritive suffix -heapme replacing the corresponding Finnish suffix -ton. Despite the phonological dissimilarity the suffixes are indeed cognate: both reflect the Proto-Uralic caritive ending *-ptama. Nevertheless, this is not an instance of etymological nativization because there is also a full semanticfunctional correspondence between the two suffixes. Hence, it is impossible to know whether their cognation has played any role in the adoption of this word. Now that etymological nativization has been defined, it can be placed in a general linguistic framework. The phenomenon can be described as a quite normal instance of analogy, as the Basque example discussed above reveals (Table 1). The concept of analogy is implicitly present in Hock’s (1986) treatment of etymological nativization. The only notable difference from typical examples of analogy is that etymological nativization involves forms belonging to two distinct languages instead of one language. 3. The nativization of vowels in the Finnic loanwords in Saami 3.1 The contact setting This section presents a study of the substitution patterns of vowels in Finnic / Finnish loanwords in Saami. First, it is useful to briefly summarize the main outline of the duration and extent of the language contact between Finnic and Saami. Finnic is a group of closely related languages spoken around the Gulf of

24

ANTE AIKIO

Finland. The total number of Finnic languages is between six and eight, depending on where one is inclined to draw the line between a dialect and a distinct language. For the purposes of this presentation it is sufficient to reckon with the phonologically most conservative Finnic language; that is, Finnish, and its immediate reconstructed predecessor, Proto-Finnic. Saami, likewise, is a group of languages that are approximately as closely related as the Finnic languages. It is customary to distinguish between ten Saami languages, which are spoken in a chain-like continuum along an area stretching from central Scandinavia in the southwest to the tip of the Kola Peninsula in the East. In addition, unknown varieties of Saami were widely spoken also further south in Finland and Karelia in the Middle and Early Modern Ages, but they became extinct under the pressure of the Finno-Karelian agricultural expansion. It is likely that many Finnic loanwords in Saami have been mediated to the surviving Saami languages by their extinct sister languages further south, just as more recent borrowings from Finnish and Karelian have been mediated from the central and eastern Saami area further to southwestern Saami on the Scandinavian peninsula. This presentation mainly concentrates on North Saami and its predecessor Proto-Saami, but examples from other Saami languages will be drawn where they provide relevant additional information. Saami is known to have been in intensive contact with Finnic for a long time. The number of Finnish and older Finnic loanwords in present-day North Saami is probably over a thousand. The majority of these borrowings are quite recent. Lehtiranta’s (1989) comparative vocabulary of the Saami languages lists 153 Finnic loanwords in Proto-Saami; that is, slightly over 10% of the total of 1479 reconstructed basic stems. The actual number of Finnic loans in Proto-Saami must have been higher, though, as only a part of the vocabulary can be reliably reconstructed; the 1479 stems can naturally not represent the full inventory of Proto-Saami lexical roots. Moreover, core vocabulary is overrepresented in a reconstructed corpus, whereas peripheral lexical items have probably included more borrowings. Etymological research has revealed that even the earliest Finnic loanwords in Proto-Saami can be divided into several consecutive strata. On phonological grounds some loanwords must have been adopted during the Pre-Finnic phase before such Proto-Finnic sound changes as *š > *h, for instance, had taken place — cf. North Saami vašši ‘hatred’ < Proto-Saami *vëš¯e Finnish viha ‘hatred’).1 Thus, Saami and Finnic must have been in contact for 1

The transcriptions in this paper follow standard orthographies of the Saami languages, except


25

a very long time. During known history this contact has been characterized by bilingualism among the Saami in the contact zone, and it is likely that conditions have been similar in the more remote past, too. The bilingual speakers in the contact zone have probably been the ones who have adopted the majority of Finnic loanwords and mediated them to the rest of the Saami area; thus, the sociolinguistic conditions must have favored the emergence of etymological patterns of nativization. On the other hand, Saami has also contributed loanwords to Finnish, albeit to a lesser extent. In the northernmost dialects of Finnish the number of Saami loanwords amounts to hundreds, but there are also dozens of borrowings adopted from extinct Saami languages in the Finnish dialects spoken in central and southern Finland. Because the loans in both directions have been adopted in the same contact situation, the vowel substitution patterns of the Saami loanwords in Finnish will also be briefly discussed below. 3.2 A summary of historical phonology As the main aim of this study is to examine etymological nativization, the substitution patterns of all Finnish vowels are not taken into consideration. With regard to stressed vowels, the ones occurring in the first syllable, only the substitution patterns of Finnic short vowels are discussed; the Finnic long vowels show a low frequency in inherited vocabulary, and they partially represent the result of secondary Proto-Finnic developments such as loss of intervocalic *w, *j, *x or *N. Also the stressed short vowel ö is left outside the treatment, as it is a Proto-Finnic innovation which appears mainly in sound-symbolic vocabulary and has no regular Proto-Uralic source. The treatment of unstressed vowels is restricted to stem-final position in the second syllable.2 First, it is necessary to summarize the main lines of Finnic and Saami vowel history. As regards short stressed vowels, the present day Finnish system reflects the common Proto-Uralic paradigm almost unchanged (see Table 2). The only exceptions are the addition of ö to the vowel paradigm and the merger of ProtoUralic *ï and *a into *a. The latter change is apparently very early, as it is for Kildin and Ter Saami, which are transcribed phonologically. 2 The treatment is based on generally accepted Uralic reconstructions and sound laws. Detailed accounts of historical phonology can be found in handbooks such as Korhonen (1981) and Sammallahti (1998) (on Saami) and Laanest (1982) (on Finnic). For more information on the phonological reconstruction of Proto-Uralic in general see Sammallahti (1988). Additional examples of each vowel correspondence can be found in Sammallahti (ibid.), Korhonen (1981), Itkonen & Kulonen (1992–2000), and Lehtiranta’s (1989) comparative Saami vocabulary.

26

ANTE AIKIO

Proto-Uralic i e ä

ü

ï u o a

i e ä

Proto-Finnic/ Finnish y u ö o a

Table 2: Proto-Uralic and Finnic vowels Proto-Uralic i ü ï u e o ä a

Proto-Saami ie i u ea ë o a¯

uo oa

ie ea

North Saami i u (e) o á a

uo oa

Table 3: Proto-Uralic and Saami shared with two other branches, Saami and Mordvin.3 Because neither Saami nor Finnish show any difference in the reflexes of Uralic *ï and *a, this distinction is of no concern to the present study. in the Finnish orthography stands for /ü/. In contrast to Finnic, Proto-Saami completely reorganized the Uralic vowel system via a complex series of developments involving shifts, mergers, and splits. Also the vowel harmony characteristic of Finnic and many other Uralic languages was lost, and many stressed vowels were lengthened and subsequently diphthongized. The result of this “great Saami vowel shift” was radically different from both the Uralic and the Finnish vowel paradigm (see Table 3). In North Saami the Proto-Saami system of stressed vowels was preserved as such, save for the unconditioned shift *ë > a (the symbol indicates a mid central unrounded vowel) and a few minor combinatory developments (see Table 3). The letter stands for long /¯a/. The correspondences between the Finnic and Saami stressed vowels, together with an example word of each correspondence, are listed in Table 4. As the table shows, all the differences in the reflexes of first syllable short vowels in the Finnish and Saami cognate pairs are due to sound changes that have taken place in Saami, whereas Finnic has preserved the original vowel unchanged.

3 According to Sammallahti (1988) *ï and *a have merged also in Mari. However, the evidence is ambiguous: Proto-Uralic *ï seems to have developed into Mari ö or ü under unclear conditions, whereas PU *a is always reflected either as a or o.


F ä käsi

SaaN ie giehta

PS *ie *kietë

PU *ä *käti ‘hand, arm’

conditions in PU *i-stems

1b)

ä äijä

á áddjá

*¯a *¯ajj¯a

*ä *äjjä ‘old man’

in PU *ä-stems

2a)

e a *ë mene- manna- *mënë-

*e *meni- ‘to go’

in PU *i-stems

2b)

e elä-

ea ealli-

*ea *eal¯e-

*e *elä- ‘to live’

in PU *ä-stems

3)

i nimi

a namma

*ë *nëmë

*i *nimi ‘name’

4)

y kynsi

a gazza

*ë *këncë

*ü *künˇci ‘nail’

5)

a kala

(v)uo guolli

*(v)uo *kuol¯e

*a (and *ï) *kala ‘fish’

6a)

o koski

(v)uo guoika

*(v)uo *o *kuoškë *ko´ski ‘rapids’

6b)

o olka

oa oalgi

*oa *oalk¯e

*o in PU *a-stems *wolka ‘shoulder’

7)

u muna

o monni

*o *mon¯e

*u *muna ‘egg’

1a)

27

vuoin initial position in PU *i-stems; vuoin initial position

Table 4: The correspondences between Finnish and Saami stressed vowels (F = Finnish; SaaN = North Saami; PS = Proto-Saami; PU = Proto-Uralic.)

28

ANTE AIKIO

3.3 The substitution of first syllable high vowels 3.3.1 The treatment of i and u. The Finnish stressed high vowels i and u show two patterns of nativization in loanwords. A number of cases display etymological substitution: the regular diachronic correspondents of F i and u — SaaN a and o, respectively — have been substituted for them. On the other hand, F i and u have frequently also been rendered with their nearest phonetic equivalents, SaaN i and u. The examples in (2) illustrate this dual patterning (a more extensive list of examples can be found in Appendix A).4 (2)

a.

Etymological substitution: F i > SaaN a F hinta ‘price’ > SaaN haddi id. F u > SaaN o F surma ‘bane, death’ > SaaN sorbmi id.

b.

Phonetic substitution: F i > SaaN i F hirsi ‘timber’ > SaaN hirsa id. F u > SaaN u F surkea ‘miserable’ > SaaN surgat id.

One might hypothesize that the dual correspondences result from two chronologically distinct periods of borrowing, in which case it would not be necessary to postulate two alternative strategies of phonological nativization. Indeed, such an explanation has been attempted. Korhonen (1981) has treated the phonological nativization of Finnic loanwords in Saami in his excellent handbook of Saami historical linguistics. He sees the two different substitutes for F i as due to distinct periods of borrowing: the loans showing SaaN a (< PS *ë) would reflect an older period, whereas the words that have retained i would be younger loans (ibid.:80-82). However, the stratification is actually circular because it is only based on the particular vowel correspondence itself, and a closer examination reveals that this correspondence is not a valid criterion for determining the age of borrowings. Let us first consider phonetic nativization. Unsurprisingly, there are plenty of young-looking loans showing the substitutions i > i and u > u (see Appendix A for examples). But there are also loans showing the same treatment which on account of their distribution must have been adopted considerably earlier; examples include SaaN ihtit ‘to come in sight’ (cf. F itää ‘to germinate, sprout’), girjjat ‘spotted, mottled, multi-colored’ (F kirjava id.), gurra ‘gorge; cleft’ (F kuru id.), 4 I apply the term ‘phonetic substitution’ to any sound substitution that has a straightforward phonetic motivation (for example, i > i), as opposed to phonetically unmotivated nativization strategies such as the etymological substitution i > a.


29

muiti- ‘to remember’ (F muista- id.), uksa ‘door’ (Estonian uks id.). The distribution of these words reaches the southwesternmost and easternmost Saami languages, and hence Lehtiranta (1989) reconstructs them into Proto-Saami. Thus, the correspondences F i ∼ SaaN i and F u ∼ SaaN u are not reliable indicators of late borrowing. Turning to etymological nativization, the substitutions F i > SaaN a and F u > SaaN o have occurred in many loanwords that are demonstrably younger than Proto-Saami. For instance, loans that retain initial h- must belong to this category, because Proto-Saami lacked the phoneme h and this foreign sound became established only later via loanwords (earlier loans show Ø- as the substitute for foreign h-). Examples of retention combined with etymological nativization include SaaN haddi ‘price’ (< F hinta id.), hapmu ‘craving (for a particular food)’ (< F himo ‘lust, desire, craving’), SaaN holbi ‘selvage’ (< F hulpio id.) and SaaN hohpi ‘scanty, short-lasting’ (< F hupa id.). In some cases the Finnish loan original itself can be shown to be a rather recent loanword, as is the case with SaaN barta ‘cabin’ (< F pirtti id.); the Finnish item derives from dialectal Russian; nepm ‘a peasant’s hut’ (< p˘ırt˘ı). On the other hand, there are items showing etymological nativization in much older layers of borrowings as well, such as SaaN vašši ‘hatred’ which must have been borrowed from Pre-Finnic *viša before the shift *š > Proto-Finnic *h (cf. F viha ‘hatred’). Thus, the two Saami reflexes of F i and u do not consistently fall into chronologically distinct layers of loans. Both correspondences are attested in younger and older borrowings alike, and hence there is no alternative to postulating two alternative strategies of sound substitution. This interpretation is further supported by occasional dialectal oscillation between SaaN i and a in Finnic loans: cf. SaaN hirbmat ∼ harbmat ‘horrible’ ( SaaN i and F u > SaaN u. On the other hand, the analogy of existing cognate items suggests that SaaN a and o ought to be substituted for F i and u, respectively. Needless to say, not only true cognates exert this analogical influence. The loanwords that are adapted to the regular sound correspondences start serving as new models, upholding and strengthening the pattern. Similar influence may be exerted even by quasi-cognates such as F kulta ‘gold’ ∼ SaaN golli


31

id. and F lukko ‘lock’ ∼ SaaN lohkka id., which have been separately borrowed into Finnic and Saami from Scandinavian. The dual treatment of Finnic stressed i and u in loanwords shows that in this case etymological nativization must be characterized as a tendency rather than a norm. Neither etymological nor phonetic substitutions clearly dominate in the material. Apparently, even the existence of a large number of ‘counterexamples’ to a given substitution model does not need to result in the disruption of patterns of etymological nativization; the speakers are able to detect correspondences such as F u ∼ SaaN o even despite the fact that there are numerous lexical pairs of the type F kuru ‘gorge’ ∼ SaaN gurra id. which contradict this correspondence. 3.3.2 The treatment of y. We can now take a look at the treatment of F y in loans, which is markedly different from that of i and u. Considering phonetic substitution, either the labiality or the palatality of the vowel has had to be compromised in borrowings because Saami has no phoneme y. The phonetically predictable substitutes are thus SaaN i and u, both of which occur. Saami i is much more common; examples include F kylä ‘village’ > SaaN gilli id., F myrkky ‘poison’ > SaaN mirku id., F pysy- ‘to stay’ > SaaN bissu- id., F (obsolete) yrkä ‘bridegroom’ > SaaN irgi id. The substitution F y > SaaN u is rarer, but there are a couple of examples: F tyhjä ‘empty; trifle’ (Pre-Finnic *tüšjä) > SaaN dušši ‘trifle, nothing’, F ystävä ‘friend’ > SaaN ustit id. In contrast with the treatment of F i and u, etymological nativization of F y has been rare. There appears to be only one indisputable loan item in North Saami where SaaN a (< Proto-Saami *ë) has been substituted for F y, namely SaaN šadda- ‘to be born; to grow; to become’ < F synty- ‘to be born’; even this seems to be a very old borrowing, to judge from its uniform distribution in Saami and its basic vocabulary status. Apparently, in all the other instances, quite regular phonetically motivated nativization has taken place. In the eastern Saami languages there are more cases showing a reflex of Proto-Saami *ë (> SaaK e / ¯ id.,˘ F a) in the place of Finnic y in loanwords: F kylvä- ‘to sow’ > SaaK ka´lve˘ ˘ ¯ ¯rt ‘sauna bath’, F kynnys (dialectal) kyly ‘bath’ (+ pirtti ‘cabin’) > SaaK kel-pe ˘ ˘ ‘threshold’ > SaaK kendas id., F lypsä- ‘to milk’ > SaaK la´pse- id., F pysy- ‘to ˘ ˘ ˘ stay’ > SaaK pešše- id. Note that some of these words have also been borrowed ˘ ˘ into North Saami, but show phonetic nativization (F y > SaaN i): cf. SaaN gilvi‘to sow’, bissu- ‘to stay’. There are also some similar examples in Inari Saami, cf. SaaI kopšâ- ‘to cook (intr.)’ (< *këpšë-) vs. SaaN giksa- id. (< *kipsë-) < F kypsy- id. However, phonetic nativization of F y has still been more common than etymological nativization also in the eastern Saami languages.

32

ANTE AIKIO

It is not clear why etymological nativization of F y has been so rare. However, one explanation can be tentatively suggested. In addition to etymological nativization, Hock (1986:393-394) discusses another type of sound substitution where the principle of phonetic nearness is broken, which he terms ‘systembased substitution’. This means that the phonetically closest phoneme in the target language is not used as the substitute for a given sound in the source language because it is already “reserved” as the substitute for another sound. For example, in the English loanwords of Hindi, plain stops have been substituted for English aspirated stops because the Hindi aspirated stops are already “reserved” as substitutes for English unvoiced fricatives. This results in nativization patterns such as English proof [ph r¯uf] > Hindi pr¯uph ‘proof’ (example taken from Hock). Thus, it can be suggested that North Saami a is perceived as “reserved” as the substitute for F i, and that this resulted in the tendency to avoid substituting the same sound for another foreign phoneme, F y. However, this explanation is not entirely satisfactory, because the question of why etymological nativization of F y has been much more common in eastern Saami than in North Saami remains unaccounted for. 3.4 The substitution of first syllable mid vowels In contrast to the treatment of Finnish high vowels, the mid vowels e and o show simple patterns of nativization. Leaving aside a couple of sporadic exceptions, Finnish e and o are rendered with the Saami diphthongs ea and oa in borrowings of all ages. A couple of random examples illustrate this: F pelto ‘field’ > SaaN bealdu id., F merkki ‘sign’ > SaaN mearka id., F merta ‘fish trap’ > SaaN meardi id., F leski ‘widow’ > SaaN leaska id.; F loppu ‘end’ > SaaN loahppa id., F oppi- ‘to learn’ > oahppa- id., F morsian ‘bride’ > SaaN moarsi id., F sota ‘war’ > SaaN soahti id. The only systematic exception to these substitutions is F o > SaaN uo after the initial glide v-: F voitta- ‘to win’ > SaaN vuoiti- id., F voima ‘power’ > SaaN vuoibmi id., F voihki- ‘to moan, groan’ > SaaN fuoiku- id. (< *vuoiku-). These cases receive a natural explanation from Saami phonotaxis: SaaN v- can never be followed by the diphthong oa, whereas the sequence vuo- is common. Thus, the substitution o > uo has been necessary in order to preserve the initial v- of the donor language. The substitution F e > SaaN ea can be easily explained as motivated by purely phonetic factors. Proto-Saami had no vowel e, and even today e occurs in North Saami only in conditioned environments as a morphophonological alternant of the diphthong ea (excluding only partially nativized internationalisms). Thus, ea is in fact the phonetically closest generally applicable substitute for F e;


33

it is the only structurally unrestricted vowel in the system that has a qualitatively identical initial component. It is in principle possible, though, that the substitution e > ea has also an etymological component to it. Proto-Uralic *e has metaphonically split in ProtoSaami according to the original height of the second syllable vowel, so that in high-vowel stems *e developed into Proto-Saami *ë but in low-vowel stems into Proto-Saami *ea (see 3.2). Thus, the cognate items of the latter type may also have contributed to the popularity of the substitution pattern F e > SaaN ea. It must be kept in mind, though, that the substitution e > ea is attested also in words of the type merkki ‘sign’ and leski ‘widow’ (> SaaN mearka, leaska) which show non-low vowels in the second syllable, and in such cases the substitution is not etymologically adequate. An argument against the relevance of the etymological correspondence Saami ea ∼ F e must be mentioned, though. The substitution F e > Proto-Saami *ë (> SaaN a) is, in contrast, almost never attested in borrowings even though it occurs in cognate items. The loanword SaaN darvi ‘tar’ (< PS *tërv¯e ? < F terva id.) may be the only exception to this rule, and even here the Saami word might perhaps instead be a separate borrowing from Indo-European (cf. English tar and its cognates). If other cases showing this substitution exist, they must be very rare indeed. It must be noted that even in the case of SaaN darvi the substitution e > *ë is not in fact etymologically adequate, because the correspondence F e ∼ PS *ë never occurs in cognate items that have a Uralic low vowel stem. The reason for the lack of the substitution F e > PS *ë could perhaps be that the metaphonic split of the Uralic vowel *e in Proto-Saami has obscured the patterns of sound correspondence and thus blocked the operation of etymological nativization. Sound correspondences that are environmentally conditioned are naturally more difficult for the speakers to notice than unconditioned correspondences of the type F i ∼ SaaN a. On the other hand, this explanation is not very satisfying, because it was already demonstrated that the occurrence of a relatively high number of apparent ‘counterexamples’ to a given sound correspondence does not need to prevent etymological nativization from taking place (see 3.3.1). The background of the substitution of F o > SaaN oa is evidently more complex than that of F e > SaaN ea. The first thing that must be noted is that it can be based on no phonetic or phonotactic motivation whatsoever. Stressed o is a completely unmarked vowel in Saami and there is thus no structural reason why it could not have been substituted for F o. On phonetic grounds one would expect

34

ANTE AIKIO

to find loanwords showing this substitution, but curiously, it is in fact never attested in borrowings save for a handful of recent adoptions such as SaaN horbmá ‘willow herb’ < F horsma ∼ (dialectal) horma id., SaaN bojá ‘boy (pejorative)’ < F poju id. and SaaN somá ‘fun, nice’ < F soma ‘pretty, neat’ (note that the last two words lack consonant gradation, which demonstrates that they are very young loans). Thus, there exists a case where the ‘principle of phonetic nearness’ is apparently almost completely overridden by other contradicting factors that influence sound substitution. Etymological nativization could in principle partially account for the phenomenon. Proto-Uralic *o has — quite like Proto-Uralic *e — undergone a metaphonic split in Proto-Saami: in stems with low vowels in second syllable it was diphthongized to Proto-Saami *oa, but before second syllable high vowels to *uo (see 3.2). But the substitution F o > SaaN oa is attested in all stem types regardless of their stem vowel, whereas the substitution F o > SaaN uo has very rarely taken place save for after initial v- where it has a special phonotactic explanation. There are a couple of sporadic exceptions, such as SaaN juovka ‘horse hair’ < F jouhi id. and SaaN ruoivvis ‘sheaf of hemp or flax’ < F roivas id., but in general the substitution o > uo has been so rare that it is almost unknown. The situation is thus rather similar to the treatment of F e. If one invoked etymological nativization, the rarity of the substitution F o > SaaN uo would remain a mystery — exactly as the lack of the substitution F e > PS *ë (> SaaN a). On the other hand, one cannot resort to phonetic nativatization, because the lack of the phonetically unmarked substitution F o > SaaN o violates the principle of phonetic nearness. Evidently, some other explanation is needed. It could be surmised that the substitution pattern o > oa has been analogically generalized to all stems from the low-vowel stems where it was originally regular. But this hardly amounts to a genuine explanation, because it leaves two questions unanswered. First, it was shown in section 3.3.1 that etymological nativization does not need to prevent phonetic nativization from taking place; instead, two rival patterns of sound substitution can coexist, and hence one is left with no explanation for the lack of the phonetically unmarked substitution F o > SaaN o. Second, it remains unaccounted for why the substitution o > uo could not have been generalized instead. To solve the problem, two additional factors must be taken into account. First, analogical influence from the substitution pattern F e > SaaN ea likely plays a role in the phenomenon. If F e is rendered in Saami as an opening diphthong with an identical first component, this provides a model for treating its


35

velar pair o the same way — hence, F o > SaaN oa. Second, keeping Hock’s concept of ‘system-based substitutions’ in mind, it can be suggested that the substitution o > oa is motivated by the fact that the phonetically unmarked substitute, SaaN o, is considered already “reserved” as the substitute for F u. This would also neatly explain why the phonetically least marked alternative F o > SaaN o is almost never attested in loans. However, this explanation suffers from the weakness that there is no absolute need for keeping the substitutes for two vowels separate; this is demonstrated by the fact that both F e and ä show identical substitutes in Saami, namely SaaN ea (see 3.5). It is not necessary to assume that any single one of the factors discussed above plays the decisive role in determining the treatment of F o in borrowings. If anything, the examples above demonstrate that sound substitution can be influenced by a multitude of factors, which may either produce contradicting patterns and rival tendencies (as in the case of F i and u) or coincide to strengthen an already existing tendency (as may be the case with the treatment of F o). 3.5 The substitution of first syllable low vowels The Finnish low vowels ä and a, too, show only a single substitute each. Save for a couple of isolated exceptions, SaaN ea (< PS *ea) has always been substituted for F ä, and SaaN á (< PS *¯a) for F a. A couple of random examples serve to illustrate this: F välttä- ‘to avoid’ > SaaN vealti- id., F väki ‘crowd, people’ > SaaN veahka id., F käy- (preterite stem käv-) ‘to go; to visit; to happen’ > SaaN geavva- ‘to happen’, F hätä ‘distress, danger, need’ > SaaN heahti id.; F alka- ‘to begin’ > SaaN álgi- id., F kaima ‘namesake’ > SaaN gáibmi id., F tajua- ‘to understand, realize’ > SaaN dádjad- ‘to find one’s way’, F vastaa- ‘to answer’ > SaaN vástid- id. Both of these patterns are phonologically motivated, or at least they have been so until very recently. In Proto-Saami *¯a was the only low vowel in the system and thus the only phonetically natural substitute for F a. Its later descendant, North Saami á, is still a natural substitute for foreign a; in most dialects is pronounced as long /aa/, but in certain dialect areas it has been umlauted to /ä(ä)/ in conditioned environments. North Saami short /a/ (dialectally weakly labialized /â(â)/), which developed through lowering from Proto-Saami *ë, would provide a phonetically sensible alternative, but nevertheless, the substitution F a > SaaN a (< *ë) is surprisingly rare even in late loanwords (but cf. F paha ‘bad; evil’ > SaaN bahá id. and F saha ‘saw’ > SaaN sahá id.). This might perhaps be due to system-based substitution once again: SaaN a may be perceived as “reserved” as the substitute for F i (cf. 3.3, 3.4).

36

ANTE AIKIO

However, the explanatory power of system-based substitution restrictions is weakened by the case of F ä. As noted in section 3.4, F e is always rendered as SaaN ea in loanwords, but despite this SaaN ea is at the same time used as the exclusive substitute for F ä as well. Thus, the phonological distinction between F ä and e is categorically lost in loanwords, and there seems to be no tendency, not even a weak one, which would aim at preserving this opposition. Another curious feature of the treatment of Finnish low vowels is that they never appear to become subject to etymological substitution. The development of Uralic *ä in Saami also involves a metaphonic split into Proto-Saami *ie in high vowel stems and Proto-Saami *¯a (> SaaN á) in low vowel stems (see 3.2). However, the substitution F ä > SaaN ie is not attested in loanwords and even examples of the substitution F ä > SaaN á are extremely rare (but cf. F häipy‘to disappear’ > SaaN áiba- ‘to be missing’, F (obsolete) änki ‘force, power’ > SaaN ágga ‘objection, excuse; hindrance’). One could hypothesize here, too, that the split of the vowel has caused the etymological sound correspondences to become more difficult to recognize, but this does not explain why both of the regular correspondences have been abandoned as models of nativization. It is true that there are also a couple of words where Proto-Uralic *ä irregularly developed into PS *ea (for example, PU *pälä ‘side; half’ > SaaN bealli id.), but a change which is irregular to begin with cannot offer a plausible model for etymological nativization. Moreover, Uralic *a has not undergone a metaphonical split, as it has regularly developed into the Saami diphthong uo in all contexts. This correspondence is attested in numerous basic vocabulary items which could be imagined becoming easily associated by bilingual speakers: cf. F pala- ‘to burn’ ∼ SaaN buolliid., F kala ‘fish’ ∼ SaaN guolli id., F jalka ‘foot, leg’ ∼ SaaN juolgi id., F kaksi ‘two’ ∼ SaaN guokte id., F sata ‘hundred’ ∼ SaaN cˇ uodi ¯ id., F kanta- ‘to carry’ ∼ SaaN guoddi- id. One would expect that the substitution of F a had been modeled after such examples in at least some loanwords, but, for unknown reasons, the relation F a ∼ SaaN uo is entirely unattested in loan vocabulary. 3.6 The substitution of unstressed stem vowels We can now turn to the treatment of unstressed second-syllable vowels in borrowings. First, the reflexes of the primary Proto-Uralic stem types will be taken into consideration. The bulk of the Proto-Uralic word roots were bisyllabic and had a stem ending either in the low vowel *a or *ä (frontness depending on vowel harmony) or in the high vowel *i (by some scholars reconstructed as *e instead). In Finnish a low stem vowel is preserved as such, whereas the reflexes


F 1) a muna 2) ä silmä 3) -e- ∼ -i käsi (käte-)

SaaN i monni i cˇ albmi a giehta

37

PS PU *¯e *a *mon¯e *muna ‘egg’ *¯e *ä *ˇcëlm¯e *´silmä ‘eye’ *ë *i *kietë *käti ‘hand, arm’

Table 5: The Finnish and Saami reflexes of Proto-Uralic unstressed stem vowels of the *i-stems now end in -e-, which alterates with -i in final position: cf. F käsi ‘hand, arm’ : GenSg käde-n < PU *käti : *käti-n. In contrast, Proto-Saami completely reorganized also the unstressed vowel system. Vowel harmony was abolished, and the *a- and *ä-stems coalesced into Proto-Saami *¯e-stems. The Uralic *i-stems became Proto-Saami *ë-stems. In North Saami PS *¯e became i, whereas PS *ë was lowered to a. The resulting correspondences can be seen in Table 5. These correspondences have given rise to a highly systematic pattern of etymological nativization. In borrowings of all ages the Finnic stem vowels a, ä and e have usually been converted into their regular correspondents in Saami, as illustrated by the following cases: F a > SaaN i F ä > SaaN i F -e- (∼ -i) > SaaN a

F hinta ‘price’ > SaaN haddi id. F silmä ‘eye’ > SaaN šalbmi ‘ax eye, needle eye’ F arki (arke-) ‘workday’ > SaaN árga id.

More examples of each substitution can be found in Appendix B (the list is far from exhaustive, though, as a complete corpus would amount to hundreds of words). In fact, the etymological nativization of the Finnish stem vowels a, ä and e has been so regular that it can be regarded as a rule. There is only one recurring exception, a group of borrowings exhibiting the substitution F a / ä > SaaN á: cf. F paha ‘bad; evil’ > SaaN bahá id., F palkka ‘salary’ > SaaN bálká id., F vielä ‘still, yet’ > SaaN velá id. However, phonological features such as intervocalic -h- and the stressed vowel -e- show that these loans are quite young, and in the older strata of borrowings exceptions are rare in the extreme. Moreover, even in young loanwords etymological nativization abounds. The regularity is especially striking when one considers the actual correspondences involved. In the nominative singular of nouns there is a crosswise pattern where F -i corresponds to SaaN -a, whereas F -a corresponds to SaaN -i. One

38

ANTE AIKIO

would expect the mimicry of this type of correspondence to be less common because it so directly contradicts the principle of phonetic nearness. But curiously, examples of phonetically unmarked substitutions of the type F i > SaaN i and F a > SaaN a are practically nonexistent, young borrowings included. Nevertheless, a case worth noting is SaaN govva ‘picture’ < F kuva id., where the exception regarding the stem vowel is especially puzzling because the word still shows the etymological substitution F u > SaaN o in the stressed syllable (see 3.3). Notably, even this word shows the expected *¯e-stem in eastern Saami, cf. SaaI kove ‘picture’, SaaT kov– v– e ‘figure, shape’ (< PS *kov¯e), and in Lule Saami there is oscillation between the two stem types: SaaL gåvvå (< *kovë) ∼ gåvve (< *kov¯e). Considering this, SaaN -a and SaaL -å might perhaps be due to some sort of irregular secondary development instead of sound substitution. Thus, there is only one somewhat dubious example of the substitution F -a > SaaN -a. The substitution F i > SaaN i remains entirely hypothetical; I have been able to find no examples, and if such exist they must be very rare indeed. (The retrogradical dictionary of North Saami (Sammallahti 2002) was employed in the search for examples of second-syllable vowel substitutions.) An interesting question emerges: why has the etymological nativization of Finnish stem vowels been so pervasive that it has managed to almost completely abolish all contradicting tendencies? The number of potential models seems to provide the answer. Because the frequency of the Finnish stem vowels a, ä and e is very high in inherited vocabulary, there are many more Finnish-Saami word pairs which serve as models of these correspondences than there are of any correspondence involving first syllable vowels. Finnish also has other stem types which have emerged through secondary developments. Of these, i-stems can be taken into consideration next. This group involves mainly loanwords and derivatives where -i- developed through a fusion of the Proto-Uralic stem vowel and a following glide *j, and hence, the Finnish istems do not have a single regular diachronic correspondent in Saami. Therefore, it is interesting to note that loanwords quite systematically display the substitution F -i > SaaN -a, as in SaaN báhppa ‘priest’ < F pappi id. (for more examples see Appendix B). This substitution pattern can have no straightforward phonetic or etymological motivation. Instead, it seems to have emerged on the basis of analogy. Finnish e-stem nouns show -i in final position, in the nominative singular. Thus, the distinction between e- and i-stem nouns is not visible in the nominative singular: cf. F arki ‘workday’ (GenSg arje-n) vs. pappi ‘priest’ (GenSg papi-n). This


39

provides a basis for analogy, and once the substitution pattern was established for nouns, it is easy to imagine that it was generalized to verbs as well where the stem vowel can never in fact occur in word-final position. Moreover, it is likely that also the first-syllable substitution pattern F i > SaaN a (see 3.3.1) has influenced the treatment of F unstressed i in loanwords. The development of the stem vowel substitution i > a through analogy illustrates how speakers can perceive the underlying sound correspondences as “wrong” and generalize them into positions where they are not actually etymologically valid. But even this finding does not seem to provide an explanation for all vowel substitutions in borrowings from Finnish into Saami. I shall discuss one more example which illustrates how the morphology of the recipient language can influence sound substitution. There are many loanwords where SaaN a (< PS *ë) has been substituted for F u or its harmonic pair y in stem-final position, e.g., S gáhta- ‘to regret’ < F katu- id., SaaN áiba- ‘to be missing’ < F häipy- ‘to disappear’. In fact, such cases are more common than ones showing the phonetically expected substitution F u / y > SaaN u (see the examples in Appendix C). Etymological nativization does not provide an explanation for this pattern, because no analogous regular correspondence exists in cognate vocabulary. It is true, there are a couple of words where a Finnish u- or y-stem word has a Saami cognate with an a-stem, such as F kysy- ‘to ask’ ∼ SaaN gahˇca- id., F suku ‘family, kin’ ∼ SaaN sohka id., F syksy ‘autumn’ ∼ SaaN cˇ akˇca id. But these cases are irregular and so rare that they do not provide plausible models for sound substitution. On the other hand, nothing suggests that analogy or ‘system-based substitutions’ would play a role here either. Instead, the phenomenon may be a morphological adjustment by origin. A clue is provided by the fact that most items showing this substitution are verbs. While verb stems ending in -u- are phonotactically entirely normal in Saami, it seems that the need to disassociate borrowed verbs from certain derivative types hinders the application of the phonetically predictable substitution F u / y > SaaN u. The majority of SaaN u-stem verbs contain either the frequentativecontinuative or the translative verb suffix -u-. The former suffix appears in such verbs as luodd-u- ‘to split (many objects)’ (cf. ludd-e- ‘to split (one object)’) and cˇ uorv-u- ‘to shout (continuously or many times)’ (cf. cˇ urv-e- ‘to shout (once)’). Examples of the latter include buoid-u- ‘to become fat’ (< buoidi ‘fat’) and stuorr-u- ‘to grow bigger’ (< stuoris ‘big’). Thus, frequentative or translative semantics would easily become associated with a new u-stem verb, and it is rational for the speakers to try to avoid this when the borrowed verb is not

40

ANTE AIKIO

inherently frequentative or translative. After the substitution pattern had been established for verbs, it could have been analogically applied to some nouns, too, where no comparable structural motivation exists. Admittedly, the explanation remains somewhat hypothetical because the tendency to avoid preserving the labial vowel in non-frequentative and non-translative verbs has not been exceptionless, as demonstrated by such loanwords as SaaN áigu- ‘to intend to’ < F aiko- id., SaaN bissu- ‘to stay’ < F pysy- id. (see Appendix C). But on the other hand, a quite different example of nativization involving stem vowels offers some support for this explanation. Finnish verbs with frequentative semantics have often taken the corresponding suffix -u- when borrowed into Saami. Examples of this include F huiski- ‘to whisk’ > SaaN hušku‘to beat’, F poimi- ‘to pick up’ > SaaN boaibmu- ‘to peck (of birds)’, F raapi‘to scratch, scrape’ > SaaN ráhpu- id., F tiuski- ‘to snap at’ > SaaN divsku- id., F voihki- ‘to moan, groan’ > SaaN fuoiku- id. In these cases the transformation of the stem vowel is not, strictly speaking, a sound substitution but rather a morpheme substitution: F -i- is a suffix which forms frequentative-continuative verbs, and it has been replaced with its functional equivalent, SaaN -u-. Even this pattern is far from regular, though, as there are also borrowed frequentative verbs showing the expected sound substitution F i > SaaN a: cf. F (dialectal) tärppi- ‘to bang, batter’ > SaaN dearpa- id. 3.7 Etymological nativization in the Saami loanwords of Finnish Lastly, a brief discussion of the nativization of borrowings adopted from Saami to Finnish is in order. There are numerous such loanwords in the northernmost dialects of Finnish, and as might be expected, the etymological substitutions that have been discussed above are mirrored in borrowings in the opposite direction. This can be illustrated with the treatment of the representatives of Proto-Saami stressed *ë and o. Loanwords from Saami have been adapted to the regular sound correpondences PS *ë ∼ F i and SaaN o ∼ F u. Compare the following cases (for more examples see Appendix D): PS *ë > F i PS *o > F u

PS *sëp¯ekk¯e ‘ski’ (SaaN sabet) > F sivakka id. PS *moršë ‘walrus’ (SaaN morša) > F mursu id.

In addition to such northerly dialect words there are also more ancient Saami loan items in the dialects of southern Finland. These must have been adopted in the Middle Ages or even in the prehistoric period when now-extinct Saami languages were still spoken in these areas. These loans also show etymological


41

substitutions identical to those attested in later borrowings in the north. As examples of such more southerly loans one can mention F vinka ∼ vinkka ‘hook (for hanging a cauldron over the fire)’ < PS *vëNk¯e (> SaaN faggi ‘hook’), F viti ‘freshly fallen snow’ < PS *vëcë (> SaaN vahca id.), and F ume ‘fog’ < PS *omV- (> SaaI omo id.). Due to the regular vowel correspondences, such southerly Saami loanwords have often mistakenly been considered cognate in etymological references. For instance, in Itkonen & Kulonen (1992–2000) the words vinka and ume have been analyzed as cognate with the corresponding Saami items, and also the cognation of F viti and SaaN vahca is considered possible. However, in each case there is proof that the Finnish items must have been borrowed. First, vinka ∼ vinkka ‘hook’ shows an extremely narrow dialect distribution in Finnish, in addition to which the oscillation between a single and geminate stop is irregular and typical of loanwords. Second, viti ‘freshly fallen snow’ has failed to participate in the Proto-Finnic sound change *ti > si, and hence it must be a loanword. Lastly, ume ‘fog’ can only derive from Saami because the Saami item is not inherited either, but a borrowing from Scandinavian (cf. Old Norse húm ‘dusk, half-dark’). The classification of these and other similar Finnish words as inherited has perhaps been influenced by the tendency of previous research to underestimate the number of Saami borrowings in the Finnish dialects. Nevertheless, Finnish dialectal vocabulary contains surprisingly many Saami loanwords, and the influence of Saami can also be clearly seen in the numerous Saami substrate toponyms in southern Finland. Many of these borrowed place names have also undergone etymological nativization; cf. Ilomantsi < PS *ël¯em¯an´ cˇ ë ‘uppermost’ (> SaaN alimuš) and Pisa < PS *pës¯e ‘sacred’ (> SaaN bassi) (for more examples see Aikio In press). 4. Discussion According to the traditional view, the phonological nativization of loanwords involves substituting the phonetically or perceptually nearest native equivalents for foreign sounds (Paul 1909:394-396; Bloomfield 1935:445-450). Modern textbooks describe the phenomenon in essentially the same terms, even though it is now more clearly recognized how structural differences between the donating and receiving languages on the phonotactic, morphological and morphosyntactic levels can restrict the possibilities of sound substitution. In addition to ‘pure’ sound substitution textbooks mention a variety of phenomena observed in the adaptation of loanwords, including phonological simplification, tone assignment, morphological and morphosyntactic adjustments, gender assignment,

42

ANTE AIKIO

and reassignment of part of speech (Anttila 1989:156-158; Bynon 1977:225-231; Campbell 1998:60-64). Still, it is maintained (at least implicitly) that nativization aims at preserving the phonetic shape of the loan item as close as possible to its model in the source language. Occasional deviations from this prediction have been explained on the basis of factors such as folk etymology (Bloomfield 1935:450). It has been pointed out, though, that this kind of ‘principle of phonetic nearness’ does not necessarily allow one to predict sound substitution because “neither the speaker himself nor the linguist who studies his behavior is always certain as to just what sound in his native tongue is most nearly related to the model” (Haugen 1950:215). Despite this, sound substitution often seems to be quite consistent within a linguistic community even in cases where multiple strategies are theoretically possible. To cite a well-known example, speakers of French substitute /s/ and /z/ for English /T/ and /D/, whereas speakers of German tend to apply /t/ and /d/, even though the phoneme inventories of both languages have /t d s z/ (McMahon 1994:206). To explain this kind of consistency it has to be assumed that sound substitution strategies can become fixed through analogy. Notably, Heath (1984) has argued that the adaptation of loanwords often takes place according to such established ‘routines’, even though he employs this term primarily in reference to morphosyntactic adjustments. The data treated in this paper demonstrates that also sound substitution rules which are neither phonetically motivated nor connected with the structure of the recipient langauge per se can become analogically fixed. The factors responsible for the emergence of such substitution models can be tentatively divided into three groups: (a) The perception of existing sound correspondences between the contacting languages can provide an impetus to mimic those correspondences in loanwords, resulting in etymological nativization. (b) The wish to maintain non-native phonological oppositions of the source language, i.e. a need to keep the substitutes for two phonemes separate, can produce system-based substitution patterns. (c) Certain sound substitutions appear to be based on systemic analogies; the possibility of the substitution o > oa in the Finnic loanwords of Saami being modeled after e > ea was discussed in section 3.4.


43

As we are primarily concerned with etymological nativization here, we can now examine what kind of methodological implications the existence of this phenomenon has in the field of comparative linguistics. The received methods of lexical stratification can be first taken into consideration. It is a standard procedure to rigorously divide loanwords into consecutive strata according to their phonological shape, as the following quotes from Theodora Bynon sum up: ...where loan-words are concerned phonological correspondences may be set up by comparing the segments of a word in the donor language with the corresponding segments in the recipient language and... for any specific point of time, these correspondences are quite as regular as those between cognate words in related languages. (Bynon 1977:221; emphasis in the original) We arrive at the form at the time of borrowing by reconstructing backwards, using the rules of the historical grammar of the recipient language, until we reach a form as close as possible to that of the source in the donor language. (ibid.: 224-225; emphasis added)

As the data discussed in this paper should demonstrate, this classical picture is somewhat idealized and does not hold in every instance. Loanwords can display contradicting patterns of nativization so that the substitutes for a given phoneme are not necessarily regular during any specific period of borrowing. Moreover, etymological nativization can deceptively cause individual loanwords to look older than they are. Reconstructing backwards until the ‘closest phonetic match’ may thus produce too early a date of borrowing, which could in turn lead to the accumulation of errors if the sequencing of lexical strata is used as key evidence in determining the relative order of sound changes in two languages, or in the dating of reconstructed proto-languages. Thus, where there is a long history of intensive contact between two languages, it is necessary to take the possibility of etymological nativization into consideration. For example, it was pointed out above that the chronology of Finnic borrowings in Saami has often been misinterpreted because of the failure to pay due attention to etymological sound substitutions (see 3.3). Likewise, some Saami loanwords in the Finnish dialects have been mistakenly analyzed as Finnish-Saami cognate items because they display regular vowel correspondence (see 3.7). Etymological nativization might also turn out to clarify certain kinds of data which pose particular problems for linguistic reconstruction. In the case of Finnic and Saami, one is in the fortunate position of being able to operate with precise sound laws and proto-language reconstructions when stratifying loan vocabulary, as the sound correspondences between the languages and the diachronic changes underlying them have been worked out in detail. However, the

44

ANTE AIKIO

situation would be quite different if one did not already know which correspondences are indicative of inheritance and which of borrowing, and if the contact setting were considerably more complex. It is easy to imagine a situation with more than two branches of the same family in intimate contact, each of which has borrowed heavily from the others. Clearly, in the absence of historical documents, such a situation would be very difficult to interpret in diachronic terms. This brings us to the type of case which is illustrated by the Austronesian languages discussed by Grace (1996). One can deduce that extensive borrowing combined with etymological nativization could indeed result in just the kind of multitude of correspondences Grace observes between the New Caledonian languages. The main reason for this is that etymological nativization does not necessarily affect every phonological segment in the word, and as a result, loan items may display contradicting phonological features diagnostic of both inheritance and borrowing. If the contact history is long and each language in the Sprachbund contains several layers of borrowings from more than one source, the data can indeed become so saturated with sound correspondences that the underlying linguistic history is very difficult to reconstruct. Whether such processes really can account for the linguistic development of the New Caledonian languages is, of course, a matter for Austronesian specialists to decide. Hopefully, though, a thorough analysis of easier cases such as the contact history of Saami and Finnic helps to reveal how complex sound correspondences can arise through a combination of inheritance and borrowing between related languages. Even in the case of Saami, painstaking comparative analysis was initially needed to distinguish between loanwords and inherited vocabulary, and the setting up of a detailed lexical stratification has required us to broaden our conception of the processes involved in loanword nativization. Hopefully, this knowledge will then allow us to sharpen the tools of comparative linguistics in order to tackle some of the more difficult cases.


Appendix A: Examples of both etymological and phonetic nativization of Finnish i and u in North Saami

i > a:

i>i

u>o

Finnish himo ‘craving, desire’

>

hinta ‘price’ hirsi(-puu) ‘gallows’ hirveä- ‘to dare to’ ilman ‘without’ linna ‘castle’ pirtti ‘cabin’ sileä ‘smooth’ silta ‘bridge’ viha ‘hatred’ (< *viša) himo ‘craving, desire’ hirsi ‘timber’ ilma ‘air; weather’ ilo ‘joy’ kilju- ‘to scream’ niska ‘back of the neck’ sinku- ‘to scold, brawl’ tila ‘space; condition’

> > > > > > > > > > > > > > > > >

tiuku ‘small bell’ > vika ‘fault, defect’ > hulpa ∼ hulpio ‘selvage’ > hupa ‘short-lasting’ > kuva ‘picture’ > murhe ‘sorrow’ (< *mureš) > muta ‘mud’ > rukka ‘poor creature’ > rukoile- ‘to pray’ suitsu ‘thick smoke’

> >

surma ‘bane, death’ tapaturma ‘accident’

> >

North Saami hapmu ‘craving (for a particular food)’ haddi id. harca(-muorra) id. arva- id. almmá id. ladni id. barta id. šallat id. šaldi id. vašši id. hipmu id. (cf. hapmu) hirsa id. (cf. harca-muorra) ilbmi ‘weather; world’ illu id. gillju- id. niski id. šiggu- id. dilli ‘time; opportunity; condition’ divga id. vihki id. holbi id. hohpi id. govva id. moraš id. mohti id. -rohkki ‘late, deceased’ (in compounds) rohkadalla- id. soica ‘thick driving snow; thick smoke’ sorbmi id. dáhpedorbmi id.

45

46

ANTE AIKIO

u > u huiski- ‘to whisk’ kulma ‘eyebrow; corner’ kuru ‘gorge, ravine’ luhta ‘sedge; flood meadow’ muista- ‘to remember’ mulko- ‘to glare’ suitta- ‘to afford’ surkea ‘miserable’ uksi ‘door’ (dialectal) urpu ‘catkin’

> hušku- ‘to beat’ > gulbmi ‘eyebrow’ > gurra id. > lukti ‘sedge’ > muiti- id. > mulgu- id. > suiti- id. > surgat id. > uksa id. > urbi id.


47

Appendix B: Examples of the nativization rules of Finnish stem vowels a, ä, e and i in North Saami

a > i:

ä > i:

Finnish arka ‘shy, timid’ hinta ‘price’ liika ‘surplus, too much’ luotta- ‘to trust in’ matka ‘trip, journey’ muista- ‘to remember’ niska ‘back of the neck’ paina- ‘to press’ rauha ‘peace’ velka ‘debt’ estä- ‘to prevent’ kylä ‘village’ kylvä- ‘to sow’ köyhä ‘poor’ metsä ‘forest’

> > > > > > > > > > > > > > >

pitä- ‘to hold’

>

päästä- ‘to release’ silmä ‘eye’

> >

tyhmä ‘stupid’ yrkä ‘bridegroom’ (obsolete) e (∼ -i) > a arki (arke-) ‘workday’ juoni (juone-) ‘intrigue, plot’ hake- ‘to fetch; to seek’ hoke- ‘to harp’ huoli (huole-) ‘worry; care’ kanki (kanke-) ‘handspike’ meri (mere-) ‘sea’ olki (olke-) ‘straw’ pääse- ‘to get somewhere’ Suomi (Suome-) ‘Finland’

> > > > > > > > > > > >

North Saami árgi id. haddi id. liigi id. luohtti- id. mátki id. muiti- id. niski id. báidni- ‘to dye’ ráfi id. vealgi id. easti- id. gilli id. gilvi- id. geaffi id. meahcci ‘wilderness, wilds’ bihti- ‘to have the strength to’ beasti- ‘to rescue’ šalbmi ‘ax eye, needle eye’ diihmi ∼ duihmi id. irgi id. árga id. juotna id. háhka- ‘to get, find, procure’ hoahka- id. fuolla id. gágga id. mearra id. oalga id. beassa- id. Suopma id.

48

ANTE AIKIO

i > a:

huoli- ‘to accept’ merkki ‘sign, mark’ pappi ‘priest’ pirtti ‘cabin’ sopi- ‘to fit; to agree upon’ soti- ‘to wage war’ tauti ‘illness’ toimi- ‘to function’ viitsi- ‘to bother to’

> > > > > > > > >

fuolla- id. mearka id. báhppa id. barta id. soahpa- id. soahta- id. dávda id. doaibma- id. višša- id.


49

Appendix C: Examples of the nativization of the Finnish stem vowels u and y in North Saami Finnish u / y > a: asu- ‘to dwell’ haju ‘smell’ häipy- ‘to disappear’ katu- ‘to regret’ kuru ‘gorge, ravine’ kypsy- ‘to cook (intr.)’ liikku- ‘to move (intr.)’ loppu ‘end’ luopu- ‘to give up’ paisu- ‘to swell’ riehu- ‘to rage’

> > > > > > > > > > >

suuttu- ‘to get angry’ synty- ‘to be born’

> >

tiuku ‘small bell’ tottu- ‘to get used to’ tyyty- ‘to be satisfied with’ vaipu- ‘to sink to the ground’ veny- ‘to stretch (intr.)’

> > > > >

viipy- ‘to be delayed’ viru- ‘to lie dead or ill’

> >

u / y > u: ampu- ‘to shoot’ hoppu ‘hurry’ häijy ‘wicked, mean’ niitty ‘meadow’ parku- ‘to cry, bawl’ pyssy ‘gun’ pysy- ‘to stay’ riepu ‘rag; poor thing’ rääky- ‘to shriek, squall’ säily- ‘to be kept, preserved’ sumu ‘mist, fog’

> > > > > > > > > > >

North Saami ássa- id. hádja id. áiba- ‘to be missing’ gáhta- id. gurra id. giksa- id. (-ks- < *-ps-) lihkka- ‘to wake up’ loahppa id. luohpa- id. báisa- id. riedja- ‘to brawl, make a racket’ suhtta- id. šadda- ‘to be born; to grow; to become’ divga id. doahtta- id. duhta- id. váiba- ‘to get tired’ vietna- ‘to get strained (of muscles)’ vihpa- id. virra- ‘to lie dead (of slaughter animals)’ ábbu- ‘to boil over’ hoahppu id. headju ‘weak, bad’ niitu id. bárgu- id. bissu id. bissu- id. riehpu ‘poor thing’ reahku- id. seailu- id. sopmu id.

50

ANTE AIKIO

Appendix D: Examples of etymological nativization in the Saami loanwords in the Far-Northern dialects of Finnish

a > i:

o > u:

North Saami caggi ‘prop’ cˇ ahki ‘lump of frozen snow’ cˇ avil ‘mountain ridge’ fasti ‘ugly, disgusting’ navvi- ‘to unhair (a skin)’ njalla ‘storehouse on a pillar’ njavvi ‘small rapids’ sabet ‘ski’ doggi ‘rennet bag’ jorbmi ‘whirlpool’ goldi- ‘to fish with a drift-net’ goksi- ‘to take s.o.’s share’ (-ks- < *-ps-) gorži ‘waterfall’ gorut ‘carcass; body’ morša ‘walrus’ njolgi ‘trot (of reindeer)’ noras ‘marrow bone’ rohtu ‘thicket’

> > > > > > > > > > > >

Finnish sinka id. kika id. kivalo id. visto id. nivo- id. nili id. niva id. sivakka id. tunka id. jurmu id. kulta- id. kupso- id.

> > > > > >

kurkkio id. kurento id. mursu id. nulkka id. nurus id. ruto id.

References Aikio, Ante. In press. “The Study of Saami Substrate Toponyms in Finland”. The Borrowing of Place Names in the Finno-Ugric Languages ed. by Ritva Liisa Pitkänen & Janne Saarikivi. (= Onomastica Uralica 4.), 21–75. Debrecsen & Helsinki: Debreceni Egyetem Magyar Nyelvtudományi Intézete. Alpher, Barry & David Nash. 1999. “Lexical replacement and cognate equilibrium in Australia”. Australian Journal of Linguistics 19:1.5–56. Anttila, Raimo. 1989. Historical and Comparative Linguistics (2nd ed.). Amsterdam and Philadelphia: John Benjamins. Bloomfield, Leonard. 1935. Language. London: George Allen & Unwin Ltd. Bynon, Theodora. 1977. Historical Linguistics. London, New York & Melbourne: Cambridge University Press. Campbell, Lyle. 1998. Historical Linguistics: An introduction. Cambridge, Mass.: MIT Press.


51

Comrie, Bernard. 1993. “A review of S. A. Starostin, Altaskaja problema i proiskhozhdenie japonskogo jazyka”. Language 69.828–832. Evans, Nicholas. 1998. “Iwaidja mutation and its origins”. Case, Typology and Grammar. In Honor of Barry J. Blake ed. by Anna Siewierska & Jae Jung Song, 115–149. Amsterdam and Philadelphia: John Benjamins. Grace, George. 1996. “Regularity of change in what?” Regularity and Irregularity in Language Change ed. by Mark Durie & Malcolm Ross, 157–179. Oxford: Oxford University Press. Haugen, Einar. 1950. “The analysis of linguistic borrowing”. Language 26.210– 231. Heath, Jeffrey G. 1984. “Language contact and language change”. Annual Review of Anthropology 13.367–384. Hock, Hans Heinrich. 1986. Principles of Historical Linguistics. Berlin: Mouton de Gruyter. Itkonen, Erkki. 1961. Suomalais-ugrilaisen kielen- ja historiantutkimuksen alalta. Helsinki: Suomalaisen Kirjallisuuden Seura. Itkonen, Erkki & Ulla-Maija Kulonen, eds. 1992–2000. Suomen sanojen alkuperä. Etymologinen sanakirja. Helsinki: Suomalaisen Kirjallisuuden Seura & Kotimaisten Kielten Tutkimuskeskus. Kasik, Reet. 1991. “Hakkame rääkima! Viron kielen peruskurssi”. Publications of the Department of Finnish and General Linguistics of the University of Turku 37. Turku. Koch, Harold. 1997. “Archaeology and linguistics: Aboriginal Australia in a global perspective”. In McConvell & Evans 1997, 27–43. Korhonen, Mikko. 1981. Johdatus lapin kielen historiaan. Helsinki: Suomalaisen Kirjallisuuden Seura. Laanest, Arvo. 1982. Einführung in die ostseefinnische Sprachen. Hamburg: Helmut Buske. Leer, Jeff. 1990. “Tlingit: A portmanteau language family?” Linguistic Change and Reconstruction Methodology ed. by Philip Baldi, 73–98. Berlin: Mouton de Gruyter. Lehtiranta, Juhani. 1989. Yhteisaamelainen sanasto. (= Mémoires de la Société Finno-Ougrienne 200.) Helsinki: Suomalais-Ugrilainen Seura. McConvell, Patrick & Nicholas Evans, eds. 1997. Comparative Linguistics and Australian Prehistory. Oxford: Oxford University Press. McMahon, April. 1994. Understanding Language Change. Cambridge: Cambridge University Press.

52

ANTE AIKIO

Nash, David. 1997. “Comparative flora terminology of the central Northern Territory area”. In McConvell & Evans 1997, 187–206. Paul, Hermann. 1909. Prinzipien der Sprachgeschichte. Halle an der Saale, Germany: Max Niemeyer. Sammallahti, Pekka. 1988. “Historical phonology of the Uralic languages”. The Uralic Languages ed. by Denis Sinor, 478–554. E.J. Brill: Leiden. . 1998. The Saami Languages: An introduction. Karasjok, Norway: Davvi Girji. . 2002. North Saami Resource Dictionary. Oulu: Publications of the Giellagas Institute. Trask, Larry. 2000a. The Dictionary of Historical and Comparative Linguistics. Edinburgh: Edinburgh University Press. . 2000b. “Some issues in relative chronology”. Time Depth in Historical Linguistics ed. by Colin Renfrew, April McMahon & Larry Trask. 1, 45–58. Cambridge: The McDonald Institute for Archaeological Research.

GRADE ALTERNATION IN INARI SAAMI AND ABSTRACT DECLARATIVE PHONOLOGY∗ PATRIK BYE University of Tromsø/CASTL This article provides a declarative phonological analysis of the intricate morphophonology of disyllabic nouns in Inari Saami. In traditional process morphophonemics, the view underlying the vast bulk of generative work, morphophonemes are phonetically coherent elements drawn from the same universal alphabet as discrete phonetic segments. However, morphophonemes are converted to discrete phonetic segments by destructive processes which cannot be accommodated within a monostratal approach. I propose to solve this problem for Declarative Phonology by representing morphophonemes as language-specific and phonetically non-coherent ‘hypercategories’ containing more than one ‘exemplar cloud’. Assuming that drawing up the articulatory plan for a segment depends on identifying a unique exemplar cloud from which to sample, a hypercategory is by itself phonetically uninterpretable. Purely information-adding operations in the phonological component serve to specify a particular exemplar cloud for the purposes of rendering the hypercategory phonetically interpretable.

1. Introduction Inari Saami evinces an extremely intricate and largely unnatural system of entirely regular morphophonemic alternations in the quantity and quality of vowels and consonants in the stem known as ‘gradation’ or ‘grade alternation’ (Itkonen 1939, 1946, 1971, 1973, 1986–1991, 1992; Sammallahti 1998b,a; Sammallahti & Morottaja 1993). These alternations attend morphosyntactic distinctions within both the nominal and verbal paradigms. The phenomenon may be exemplified by comparing the and accusative forms of ‘stack’: p´ınnoo ( ; ∗

Thanks to Sylvia Blaho, Paul de Lacy, Paul Kiparsky, Ove Lorentz, Tore Nesset, Richard Ogden, Curt Rice, Pekka Sammallahti, Peter Svenonius, and Ida Toivonen for feedback and discussion on various aspects of this work. Any remaining errors of fact and interpretation are my own responsibility.

54

PATRIK BYE

strong grade) ∼ p´ınoo ( ; weak grade). Here the strong grade form surfaces with a medial geminate, the weak grade with a medial singleton. In addition, grade alternation interacts with an alternation in the location in which secondary stress resumes following the main word stress. (The main stress itself is always word-initial.) This alternation is dubbed ‘perturbation’ here. In the non-perturbed pattern, secondary stress resumes on the postpeninitial syllable, whereas in the perturbed pattern it resumes on the peninitial syllable, immediately following the main stress of the word. This constrast is exemplified by the pair p´ınnoon [p´ın;o;n] (; non-perturbed) ∼ p´ınnòon [p´ınnoon] (; perturbed). The cross-classification of perturbation and grade gives a three-way phonetic distinction in consonant quantity: short [C] • half-long [C;] • long [CC]. The intervocalic geminate /nn/ is longer in p´ınnòon, where it straddles a foot boundary, than in p´ınnoon, where it is foot-medial. Foot-medial geminates are traditionally described as ‘half-long’. Some researchers such as Sammallahti & Morottaja (1993) take the phonetic three-way distinction as phonemic. Here it is assumed to result from the interaction of moraic and prosodic structure. Baroque phonological systems seem to be common in small close-knit societies, which seem to have the kind of social structure which disfavours the loss of marked phonological structures during transmission to the next generation. However, this is apparently no barrier to phonetic innovation, which with time increases the opacity of the phonology. The phonetic motivation of many of the regular morphophonemic patterns we see in Inari Saami is synchronically opaque. For reasons of space we’ll restrict our focus in this paper to the absolute (non-possessed) declension of vowel-final disyllabic stems. 1.1 Abstract Declarative Phonology This paper applies Abstract Declarative Phonology (ADP) to the analysis of the Inari Saami data. ADP is based on the Declarative Phonology (DP) framework (Scobbie 1991; Bird 1995; Scobbie et al. 1996; Coleman 1998; Ogden 1999). In contrast to classical DP, which adopts the representational assumptions of mainstream phonological theory, ADP posits morphophonemes as phonetically non-cohering. DP is constraint-based and arguably conceptually minimal, because generalizations are stated directly over a unique level of representation: there are no intermediate levels or strata. Monostratalism entails that operations must be at least structure-preserving (Ogden 1999:57) and (possibly vacuously) structurebuilding (Bird & Klein 1994).

GRADE ALTERNATION IN INARI SAAMI

55

A constraint in DP is a partial description D of a surface representation. As such it is non-defeasible, unlike constraints in Optimality Theory. Since a partial description may contribute either lexical or grammatical information to the surface representation, there is no distinction between ‘rules’ and representations. Constraints in DP may not conflict, but they may overlap. In fact, the issue of conflict is apparently more complex than this, since some writers, such as Scobbie (1993), allow for conflicts between constraints which may be resolved by appeal to intrinsic principles, in particular the Elsewhere Principle. This refers to cases of conflict in which a constraint C1 whose antecedent (structural description) SD1 stands in the relationship general to specific with the antecedent SD2 of some other constraint C2 . For reasons of expository convenience I will also avail myself of this principle here, although it must be stressed that the priority of application in these cases need not be taken literally. This is because the conflict may easily be eliminated by taking SD1 to include the complement of SD2 . Transformational approaches such as Optimality Theory (OT) and Derivational Rule-Based Theory (DT) are able to reconcile incompatible specifications by appeal to distinct input and output forms. In transformational phonology, the input is mapped onto an output by applying potentially destructive processes. In DT, these processes may be crucially ordered, while in Classical OT, they apply simultaneously to the candidate output. DP’s monostratalism, however, entails the impossibility of phonological neutralization (feature change; deletion). In classical DP the underlier of a neutralized segment is therefore standardly represented as a disjunction of segments or features. By way of illustration, consider the lexical partial description of an underlying ‘voiced’ stop in a language with Final Obstruent Devoicing such as German in (1). In Derivational Rule-Based Theory, this relationship is understood in terms of a feature-changing rule mapping the systematic phonemic (underlying) level specification [+voice] preceding a word boundary to its opposite [−voice] on the discrete phonetic. In DP, however, the mapping cannot be feature-changing. We have to reconstrue the distinction between underlyingly [−voice] and [+voice] as obtaining between [+voice] [−voice] and a disjunction . The truth of a disjunction p ∨ q is [−voice] compatible with the truth of either disjunct, p or q. It follows that it is possible to unify a specification of the form p ∨ q with either of its disjuncts. Thanks to these logical properties, the demands on the word-final distribution of voicing are rendered compatible with lexical requirements.

56

PATRIK BYE

(1)

Final Obstruent Devoicing in DT and DP a.

DT [+voice]# → [−voice]#

b.

DP [+voice] # → [−voice] [−voice]

However, the disjunction may be seen as an artefact of classical DP’s concrete morphophonemic approach, which entails that phonological units have coherent intrinsic content. ADP makes it possible to dispense with such disjunctions precisely because it does not require phonological categories to have coherent intrinsic phonetic content. In the absence of coherent content, the conflict doesn’t arise, and so logical consistency is achievable within the restrictions imposed by the theory. To handle the present case we set up two features: one corresponding to the discrete phonetic level, which we can call [voice], and one arbitrary voice]. For reaand purely systematic, which I will render in outline shape as [voice voice sons which will become clearer in the next section, I will refer to categories in outline shape as ‘hypercategories’. In addition to phonetically coherent features and segments, I will assume the existence of both arbitrary hyperfeatures and hypersegments.1 In the pair [ Ra:t] (Rat), ‘counsel’ • [ Ra:t] (Rad), ‘wheel’ there is a systematic distinction between [−voice] and [+voice voice voice] which is visible to the computational system but, since the phonetic content of [+voice voice voice] is not coherent, it cannot be interpreted by the articulatory and perceptual systems. This hyperspecification is ‘filled out’ with the phonetically interpretable feature [voice] in the relevant environments according to the constraints in 2. Note that (2a) has priority over (2b) by the Elsewhere Condition. (2)

1

Final Obstruent Devoicing in Abstract Declarative Phonology a.

voice [+voice voice]# →[−voice]#

b.

[+voice voice voice]→[+voice]

I use these terms here rather than ‘underlying’ and ‘surface’ in an effort to stave off any unnecessary confusion, since DP countenances only a unique surface level. This unique surface representation contains both unpredictable (lexical) and predictable information. In ADP, the terms ‘surface representation’ and ‘phonetic representation’ cannot be equated as they can in the Process or Concrete Morphophonemic view. Phonetically interpretable symbols are rather included within the surface representation.


57

Since only [+voice] has coherent phonetic content and is phonetically interpretable, [+voice voice voice] and [−voice] are perfectly compatible and so nothing prevents them from cooccurring within the same phonological representation. The distinction does not correspond to a level distinction, covert or otherwise. Let us now delve a little more deeply into the notion that phonological categories may lack phonetic coherence. From the point of view of the phonological computation, all phonological categories are on a par irrespective of whether they have coherent phonetic content or not. The difference between the categories we are writing in outline shape and those we are rendering in solid shape has to do with their phonetic interpretability. The content of phonological features is traditionally taken to relate fairly directly to articulatory and acoustic-auditory properties of the exponents of those categories. However, this is no longer the only way to construe the content of phonological categories after the advent of ‘exemplar theory’ (Johnson 1997; Smith 1998; Pierrehumbert 2001). According to exemplar theory, the content of a given phonological category may be equated with a set of phonetically detailed ‘memory traces’ deposited each time that category is perceived or pronounced.2 For example, the category /s/ in a given speaker labels all the memories corresponding to articulatory and perceptual events which cluster according to certain properties, such as a grooved lamino-alveolar constriction or having a spectral peak around 8 kHz. Exemplars of a given category will generally be phonetically similar to each other, and so they will cluster together in clouds in the phonetic parameter space. Given this memory deposit of phonetically detailed representations, the next question concerns production, or more specifically how the articulatory plan is drawn up. Exemplars may encode types of articulatory and auditory events which are more or less frequent as realizations of that category. Many variants of exemplar theory assume that exemplars differ individually in terms of their activation potential as a function of the probability of occurrence of the corresponding stimulus in the ambient speech environment. With this as her point of departure, Pierrehumbert (2001) proposes that the articulatory plan for a given category φ is drawn up by selecting a random point p in the corresponding exemplar cloud. A window of fixed size is computed with p at its centre. All exemplars inside this window contribute to the ultimate articulatory plan, which is drawn up by calculating the activation-weighted average of all the exemplars in the window. Notice 2

However, the degree of detail encoded is subject to perceptual limitations (just noticeable differences) and the number of exemplars stored at any one time is presumably subject to memory limitations. Each exemplar will thus represent an equivalence class of highly similar events, rather than individual events.

58

PATRIK BYE

that this view assumes that each category is associated with a unique exemplar cloud. However, this cannot be the case for certain phonological categories, given the existence of intrinsic (linguistically governed) allophony. Kingston & Diehl (1994) propose that distinctive features may have phonetically polymorphous or non-coherent content. (See also Walsh-Dickey 1997.) For example, it is apparent that ‘light’ and ‘dark’ allophones of /l/ in RP English will cluster in two clouds, albeit these clouds will be relatively close to one another in the parameter space. Indeed if we admit that categories may be polymorphous, then there is no formal principle which would prohibit the formation of such multinebular categories. Morphophonemic alternations, such as that between /k/ and /s/ in English (telic∼telicity) are I suggest no different in kind. This case involves an abstract k morphophoneme which is associated with two exemplar clouds located at some distance from each other in the parameter space. Given the phonetic interpretation algorithm adumbrated above, such multi-nebular categories are not directly phonetically interpretable, not because they have no phonetic content, but because they have too much — in the sense that the category does not uniquely define an exemplar cloud from which to draw up the articulatory plan. For a hypercategory C to be interpretable, the phonetic interpretation component has to know which exemplar cloud to sample. Therefore some additional phonology has to ‘fill out’ C by unifying it with some additional category C which does have a unique exemplar cloud. The phonetic interpretation then takes the intersection of C and C. The function of the phonological computation can thus be construed as rendering lexical phonological categories (that is, those which encode systemic phonematic contrasts) interpretable at the interface to the phonetics. The phonology does this by supplying additional information which guides phonetic interpretation to particular exemplar clouds. 2. Empirical background In the interests of accessibility, all the examples in this paper have been reinterpreted in IPA from the Finno-Ugric Transcription (FUT) used in the primary sources. Anyone who has had any experience dealing with FUT will realize that this is no trivial task. Since IPA and FUT slice up phonetic continua slightly differently, there is no guarantee that something of potential significance is not lost in the translation. The difficulties are clear when attempting to assign IPA symbols to vowels: IPA operates with four reference points in the vowel aperture/height dimension, but FUT operates with five. Nevertheless, I have been considerably aided in the task of translation thanks to the very useful synopses in Eriksson (1961) and Sammallahti (1998a). A couple of additional comments in


59

respect of the difficulties of translation are in order. Compared with the low level of phonetic detail phonologists generally work with, FUT is frequently overdetermined, distinguishing up to 5 degrees of phonetic length. From the vantage point of modern phonological and phonetic theory, FUT transcriptions equivocate in respect of the source of duration, and this compounds the problem of interpretation. Phonetic duration and its representation in FUT is determined by a number of linguistic and non-linguistic factors, including lexical quantity contrast, quantitative trade-offs within prosodic domains, such as the foot or prosodic word, and (less importantly) low-level segmental factors. Also, since Itkonen’s FUT transcriptions purport to be records of actual token utterances, they may also occasionally reflect errors in performance on the part of the informant. Nevertheless, despite these problems, I am confident that I have been able to arrive at a sound and coherent interpretation of Inari Saami phonological structure on the basis of the substantial amount of material collected and published by Itkonen. It is also worth stressing that the phonological analysis proposed here takes as its point of departure Itkonen’s findings, which were recorded well over fifty years ago. It does not set out to account for any restructuring in the phonological system that may have taken place over the last half century or so. Analysis of the results of more recent fieldwork by Pekka Sammallahti and others will therefore plausibly uncover significant differences in phonological structure. 2.1 Phonological inventory The Inari Saami sound inventory is given in Table 1. Inari Saami essentially has a phonetic six-vowel system, although in addition to the set of six ‘full’ vowels, Inari Saami also has what we can refer to as a ‘reduced’ back vowel, transcribed here as 2, and a set of diphthongs. Both full vowels and diphthongs may appear either long or short. Short diphthongs are transcribed with an acute accent over the second component. The reduced vowel occurs in a short and an ultrashort variant. A full vowel may be either long (bimoraic) or short (monomoraic). (This range is abbreviated below as μ21 .) The reduced vowel /2/, however, can never be realized as bimoraic, but only as short or ultrashort (amoraic) (μ10 ). When ultrashort, the vowel is transcribed with a brève accent: 2˘. At the discrete phonetic level, both full monophthongs and diphthongs evince a contrast between monomoraic and bimoraic. However, there is no corresponding systematic contrast. The stop system is characterized by a contrast between plain [−spread glottis] and (post)aspirated [+spr.gl.] stops. The [spr.gl.] contrast also extends to the

60

PATRIK BYE

p ph f v m

t th D n l r

ts tsh s z

tˇs tˇsh ˇs ˇz ñ

j

k kh h N

i e a

u o A

ii ee aa 2 2˘ ié ué uó ie iá uá ia eá oá ea

uu oo AA ue

uo ua oa

Table 1: Discrete phonetic segments of Inari Saami sibilants. The segments transcribed here as z and ˇz are described by Itkonen as only partially voiced, which is interpreted here as [−spr.gl.]. (The continuants /v/ and /D/ are fully voiced, but may be better treated as approximants.) The contrast between [+spr.gl.] and [−spr.gl.] obstruents is restricted to word-medial position. Word-initially, the contrast is suspended since only [−spr.gl.] plosives and [+spr.gl.] fricatives are found.3 All consonants may appear as phonetically short (amoraic) or geminate (moraic). At the hypersegmental level, things look somewhat different. Table 2 sets out the hypersegmental inventory. In the analysis, hypersegments play the same functional role as units on the systematic phonematic level in traditional generative phonology. The difference between systematic phonematic segments and hypersegments is that, while systematic phonemes are endowed with coherent phonetic content, this is not necessarily the case for hypersegments. For example, the hypersegment ph maps onto the set {ph pph v}, and so includes geminate and non-geminate as well as continuant and non-continuant exemplars. In the limiting case, a systematic unit may evince no alternation, in which case the corresponding hypersegment will be phonetically coherent. One example of such a case is f (= /f/), which does not participate in grade alternation. Hyperconsonants evince a contrast between hypergeminate and hypersingleton. Both hypergeminates and hypersingletons map onto both phonetic geminates and singletons, but the functions in each case are different. This is addressed in detail below. For the purposes of the analysis, Section 4, the most important hyperfeatures are [geminate geminate geminate] and [sonorant sonorant sonorant]. Also note that tš and k are grouped under the same hyperfeature [dorsal dorsal dorsal]. This stands in contrast to the division into natural 3 The [−spr.gl.] obstruents may be realized as voiced between an unstressed and a stressed vowel, but this appears to be subject to dialect variation which we abstract away from here.


pp ph pph f

tt th tth

D v vv dd n m mm nn l ll r rr

61

ttˇ tts tt ˇs kk h ts tsh kh h tts ttsh kkh s h ˇs ˇsˇs ss

ñ ññ

j jj Table 2: Inari Saami hyperconsonants classes of the corresponding discrete phonetic segments, tˇs ( ) and k∼h ( ∼ placeless). We also need to distinguish tš and k by means of a back back back [back back] hyperfeature such that tš is [−back back] and k [+back back]. For certain patterns the distinction between hyperocclusives (occlusive = plosive or affricate) and continuants is important. In certain conditions, hyperocclusives may be filled out as [+continuant] spirants or fricatives phonetically. The following hypersegment variables are used in this paper: C = any hyperconsonant; T = hyperocclusive (plosive/affricate); S = hypersibilant; B = hyperobstruent; N = hypersonorant; K = hyperdorsal. 2.2 The morphology of the disyllabic noun This section provides some basic descriptive terminology of the kind used in traditional accounts of Saami phonology (Bergsland 1948; Sammallahti 1998a). The noun in Inari Saami inflects for number (singular, plural) and case (nominative, genitive, accusative, illative (‘to’ or ‘into’), locative, comitative (‘with’), abessive (‘without’), essive (‘as’), and partitive), as illustrated in the nominal paradigm for p´ınnoo, ‘stack, pile’ in Table 3. The forms are rendered in a variety of scriptfaces. The reason for this will be explained in just a moment. A complete overview of the nominal morphology, let alone the morphology, of Inari Saami is

62

PATRIK BYE

(p´ınnoo) (p´ınoo-h) (p´ınoo) (p´ınoo-jj) (p´ınoo) (p´ıi)(nò-jj-t) (p´ınnoo-n) (p´ıi)(nòo-j-t) (p´ıi)(nòo-st) (p´ıi)(nòo-jj-n) (p´ıi)(nòo-jn) (p´ınoo-jj)-(kùjjm) h (p´ınoo-t)(t àa) (p´ınoo-jj-t)(thàa) (p´ın)(n` oo-n) (p´ın)(n` oo-t)

Table 3: Paradigm for pínnoo ‘stack, pile’ beyond the scope of the present paper. Readers who would like more details than I can provide here are referred to Sammallahti & Morottaja’s dictionary (1993) and Volume IV of Itkonen’s Inarilappisches Wörterbuch (1986-1991). Answers to other questions, including synopses of the historical development of the grade alternation and metrical systems and extended exemplification, may be found on my website at http://www.hum.uit.no/a/bye. Notice first of all that disyllabic nominal stems may be stressed according to one of two prosodies. In Table 3, the first two syllables may either be parsed separately, giving resumption of stress on the second syllable as in [(´ ν )(` ν . . .) . . .], or together as a syllabic trochee, giving [(´ ν ν) . . .]. Let us designate these prosodies by the terms ‘perturbed’ and ‘non-perturbed’ respectively. As shown in Table 3, perturbed and non-perturbed prosody alternates within the paradigm. Metrically perturbed items are rendered in slanted script, while non-perturbed items are rendered upright. Synchronically, the distribution of perturbed and non-perturbed prosody in the disyllabic nominal paradigm is morphologically conditioned: perturbed metrical structure is associated with the locative and comitative singular, the accusative, illative and locative plural, and the numerically uncharacterized essive and partitive forms. Non-perturbed metrical structure is found elsewhere in the disyllabic paradigm. In our description it will be crucial to distinguish between the first and second nucleus (ν) of the disyllabic stem. Following Sammallahti (1998a), I will refer to the first nucleus ν1 as the ‘vowel centre’ and to the second nucleus ν2 as the ‘latus’. The vowel centre ν1 always receives main word stress. The latus may either receive secondary stress (in the case of perturbed prosody) or be unstressed (in the case of non-perturbed prosody).


Hypersingleton Hypergeminate

Strong C¯ ¯CC

63

Weak ˘ C ˘ CC

Table 4: Hyper-geminate/hyper-singleton and grade The position between the vowel centre and the latus is occupied by a consonant or consonant cluster. This position is phonologically central in Saami, and is called the ‘consonant centre’ (Sammallahti 1998a). Variation in morphosyntactic environment is accompanied by alternation in the quantity and/or quality of one or more of the consonants within the consonant centre. These alternations I will refer to as ‘grade alternation’. Grade alternation centres on a distinction between the so-called ‘strong’ and ‘weak’ grades. In Table 3, items in the strong grade are rendered in bold, while items in the weak grade are rendered non-bold. Within the disyllabic nominal paradigm, the distributions of the strong and weak grades is morphologically, but not phonologically, predictable. The strong grade is associated with the nominative singular, illative singular, essive and partitive forms of the disyllabic noun. The weak grade occurs elsewhere. Grade is a property of hypersegments. I will follow the FUT convention of abbreviating the strong grade with a macron (¯) and the weak grade with a brève (˘) over the relevant symbol. The cross classification of the hypergeminate–hypersingleton dimension with grade then yields the four abstract terms in Table 4. The quantity and quality of the vowel centre and latus are also subject to alternation. These alternations may be related to grade, metrical structure, or other local segmental effects, but we leave fuller discussion of these phenomena to section 4.4. 3. Morphophonological Patterns This section lays out the data in the form of relatively abstract morphophonological patterns rather than concrete data. In terms of nuclear quantity, disyllables may be quantitatively ‘trochaic’ or ‘iambic’ as a function of vowel quality, fuller treatment of which awaits in section 4.4. In this paper, these terms refer exclusively to the relative quantitative prominence of two adjacent syllables, although to avoid confusion, I will adapt them as ‘q-trochees’ and ‘qiambs’. Cross-classifying grade (strong/weak) and perturbation (perturbed/nonperturbed) gives four environments for each hypersegment. Schemas are provided for both the q-trochaic and q-iambic environments.

64

PATRIK BYE

v

S W D S W m S W n S W ñ S W l S W r S W j S W

Q-trochaic −P +P ´ ´ ` VVvvV VVvv V ´ ´ V ` ∼VVv ´ VV ` com.sg VVvV VVv ´ ´ ` VVDDV VVDD V ´ ´ ` ´ VV ` com.sg VVDV VVDV∼VVD ´ ´ ` VVmmV VVmm V ´ ´ ` ´ ` com.sg VV VVmV VVmV∼VVm ´ ´ ` VVnnV VVnn V ´ ´ V ` ∼VVn ´ VV ` com.sg VVnV VVn ´ ´ ` VVññV VVññ V ´ ´ V ` ∼VVñ ´ VV ` com.sg VVñV VVñ ´ ´ ` VVllV VVllV ´ ´ V ` ∼VVl ´ VV ` com.sg VVlV VVl ´ ´ ` VVrrV VVrrV ´ ´ V ` ∼VVr ´ VV ` com.sg VVrV VVr ´ ´ ` VVjjV VVjjV ´ ´ V ` ∼VVj ´ VV ` com.sg VVjV VVj

Q-iambic −P +P ´ ´ VV ` VvvVV Vvv ´ ´ VV ` VvVV VVv ´ ´ VV ` VDDVV VDD ´ ´ ` VDVV VVDVV ´ ´ ` VmmVV Vmm VV ´ ´ ` VmVV VVmVV ´ ´ VV ` VnnVV Vnn ´ ´ ` VnVV VVnVV ´ ´ VV ` VññVV Vññ ´ ´ ` VñVV VVñVV ´ ´ VV ` VllVV Vll ´ ´ VV ` VlVV VVl ´ ´ ` VrrVV VrrVV ´ ´ VV ` VrVV VVr ´ ´ ` VjjVV VjjVV ´ ´ VV ` VjVV VVj

Table 5: Hypersingletons I: sonorants 3.1 Hypersingletons C) correspond to discrete phonetic singletons or geminates Hypersingletons (C N ), only grade is as a function of grade and perturbation. For the hypersonorants (N relevant to filling out predictable phonological features. N is a discrete phonetic geminate in the strong grade, but a singleton in the weak grade. This is shown in Table 5. (Key: S[trong], W[eak]; −P = non-perturbed, +P = perturbed.) B), on the other hand, is more The general picture for the hyperobstruents (B complex (Table 6). Not only do both grade and perturbation have an impact on fill-out, but other information is relevant as well. For one thing, it matters whether B is a hyperdorsal or not. We also have to know whether B is a hyperocclusive T or a hyperfricative F. When strong grade and non-perturbed, B¯ is filled out as a singleton. When strong and perturbed B¯ is filled out as a geminate. To exemplify the relevance of hyperdorsality, when weak and non-perturbed, hyperdorsals are geminates. Nonhyperdorsals in the weak grade and the non-perturbed prosody are singletons.


ph th tsh tšh kh s š

S W S W S W S W S W S W S W

−P ´ hV VVp ´ VVvV ´ hV VVt ´ VVDV h ´ VVts V ´ VVzV ´ sh V VVtˇ ´ VVjjV ´ VVhV ´ VVvvV ´ VVsV ´ VVzV ´ sV VVˇ ´ zV VVˇ

Trochaic +P h` ´ VVpp V ´ ` ´ VV ` com.sg VVvV∼VVv h ´ ` VVtt V ´ ` ´ VV ` com.sg VVDV∼VVD h ´ ` VVtts V ´ ` ´ VV ` com.sg VVzzV∼Vzz ´ sh V ` VVttˇ ´ V ` ∼Vjj ´ VV ` com.sg VVjj ´ ` VVhh V ´ ` ∼Vvv ´ VV ` com.sg VVvv V ´ ` VVssV ´ ` ∼Vzz ´ VV ` com.sg VVzz V ´ ` VVˇsˇsV ´ zˇzV ` ∼Vˇ´ zˇzVV ` com.sg VVˇ

65

Iambic −P +P h ´ h VV ´ ` Vp VV Vpp ´ ´ VV ` VvVV VVv h h ´ VV ´ VV Vtt ` Vt ´ ´ ` VDVV VVDVV h h ´ ´ VV Vtts ` Vts VV ´ ´ ` VzVV VVzzVV h h ´ s VV ´ s VV Vttˇ ` Vtˇ ´ ´ ` VjjVV VVjjVV ´ ´ VV ` VhVV Vhh ´ ´ ` VvvVV VVvvVV ´ ´ VV ` VsVV Vss ´ ´ ` VzVV VVzz VV ´ ´ ` VˇsVV VˇsˇsVV ´ zVV ´ zˇzVV ` Vˇ VVˇ

Table 6: Hypersingletons II: obstruents 3.2 Hypergeminates In the environment of the strong grade form of any hypergeminate CC CC, the vowel centre and the latus are short. There is one exception to this generalization in the illative singular, which, as we shall see, comes with special templatic requirements. Historically, the strong grade form of the geminate was overlong (bimoraic) as is still the case in North Saami (Sammallahti 1977, 1984; Magga 1984; Bye 1997a,b, 2001). Inari Saami subsequently lost the distinction between bimoraic and monomoraic geminates and clusters, but only after shortening of the vowel centre and latus had applied adjacent to a bimoraic consonant centre. Synchronically, the reflexes of vowels are short when the consonant is a hypergeminate in ¯ the strong grade (CC CC). Otherwise, there are similar differences in complexity between the hyperBB BB) of this class as we saw earlier for the NN) and hyperobstruents (BB sonorants (NN NN hypersingletons. For hypersonorants, only grade is relevant to fill-out. In the ¯ is filled out as a discrete phonetic geminate. This is strong and weak grades, NN shown in Table 7. ¯ is The hyperobstruents are again more complex. In the strong grade, BB filled out as a discrete phonetic geminate regardless. In the weak grade, however,

66

PATRIK BYE

vv

S W DD S W mm S W nn S W ññ S W ll S W rr S W S jj W

Trochaic −P +P ´ ´ V ` VvvV Vvv ´ ´ ` VVvvV VVvv V ´ ´ V ` VDDV VDD ´ ´ ` VVDDV VVDDV ´ ´ ` VmmV Vmm V ´ ´ ` VVmmV VVmmV ´ ´ V ` VnnV Vnn ´ ´ ` VVnnV VVnnV ´ ´ V ` VññV Vññ ´ ´ ` VVññV VVññV ´ ´ V ` VllV Vll ´ ´ V ` VVllV VVll ´ ´ ` VrrV VrrV ´ ´ ` VVrrV VVrr V ´ ´ ` VjjV VjjV ´ ´ V ` VVjjV VVjj

Iambic −P +P ´ ´ V ` VvvV Vvv ´ ´ ` VvvVV VVvv VV ´ ´ V ` VDDV VDD ´ ´ ` VDDVV VVDDVV ´ ´ ` VmmV Vmm V ´ ´ ` VmmVV VVmmVV ´ ´ V ` VnnV Vnn ´ ´ ` VnnVV VVnnVV ´ ´ V ` VññV Vññ ´ ´ ` VññVV VVññVV ´ ´ V ` VllV Vll ´ ´ VV ` VllVV VVll ´ ´ ` VrrV VrrV ´ ´ ` VrrVV VVrr VV ´ ´ ` VjjV VjjV ´ ´ VV ` VjjVV VVjj

Table 7: Hypergeminates I: sonorants the difference between perturbed and non-perturbed prosody becomes relevant. ¯ is filled out as a singleton, but as a geminate In a non-perturbed environment, BB when perturbed. This is summarized in Table 8. 3.3 Clusters Clusters behave like hypergeminates (Table 9). 4. Analysis The fill-out of the consonant centre, vowel centre and latus are determined by a combination of morphophonological and lexical factors. The lexical factors are the identity of the segments and hypersegments making up the lexical representation. This section will address the morphophonological factors in the following order. Section 4.1 introduces the constraints governing the distribution of grade and perturbation. Section 4.2 turns to the quantitative fill-out of the consonant centre and 4.3 their qualitative fill-out. Finally, section 4.4 addresses vowel quantity. In the interests of space, qualitative aspects of the vowel centre and latus will not be addressed in the analysis but are left to future research.


pph tth ttsh ttšh kkh ss šš

S W S W S W S W S W S W S W

67

Trochaic Iambic −P +P −P +P h h h h ´ ´ hV ´ ` ´ ` ´ Vpp V∼VVp Vill.sg Vpp V Vpp V Vpp h` h` ´ ´ ´ h VV VVpp ´ hV VVpp V Vp VV VVp h h h h ´ ´ V ´ ` ´ V ` ´ h Vill.sg VVtt V∼VVt Vtt VVtt V Vtt ´ hV ´ h VV ` ´ h VV VVtt ` ´ hV VVtt Vt VVt h h h h h ´ ´ ´ ` Vtts ´ ` ´ VVtts V∼VVts Vill.sg V Vtts VVtts V V h h` h ´ ´ ´ ` ´ VVtts V Vts VV VVttsh VV VVts V h h h h ´ s V∼VVtˇ ´ sV ´ sV ` Vttˇ ´ s V Vttˇ ` ´ sh Vill.sg VVttˇ VVttˇ h h` h ´ ´ ´ ` ´ VVttˇs V Vtˇs VV VVttˇsh VV VVtˇs V h h h h ´ ´ ´ ` Vkk ´ ` ´ h Vill.sg VVkk V∼VVk V Vkk VVkk V V ´ ´ ` ´ ´ ` VVhV VVhhV VhVV VVhhVV ´ ` ´ ´ V ` ´ ´ ∼VVsV VVss V VssV Vss VVssV ill.sg ´ ´ ` ´ ´ ` VVsV VVss V VsVV VVss VV ´ ` ´ ´ ` ´ ´ VVˇsˇsV VˇsˇsV VˇsˇsV VVˇsˇsV∼VVˇsVill.sg ´ sV ´ sˇsV ` ´ sVV ´ sˇsVV ` VVˇ VVˇ Vˇ VVˇ Table 8: Hypergeminates II: obstruents

Trochaic −P +P ´ ´ ` ´ LK S VVlkkV ∼VVlkV VVlkk V ill.sg ´ ´ ` ´ VV ` com.sg W VVlkV VVlkV∼Vlk

Iambic −P +P ´ ´ V ` VlkkV Vlkk ´ ´ ` VlkVV VVlkVV

Table 9: Clusters 4.1 Morphological conditioning Synchronically, the distribution of the strong and weak grades in the disyllabic nominal paradigm is morphologically governed. The strong grade is required in the nominative singular, the illative singular and the numerically noncharacterized essive and partitive forms. These morphosyntactic specifications do not form a natural class. Hence they are simply disjunctively listed in the antecedent of the constraint below. The rest of the singular paradigm and all of the plural take the weak grade. The distinction between strong and weak grade can be captured in terms of strong an abstract binary feature [strong strong]. The distribution of the values of this feature can be characterized in terms of the constraints in (3a) and (3b). Together, these constraints stand in the relation of specific to general. Where conflict arises,

68

PATRIK BYE

(3a) has priority over (3b) by the Elsewhere Condition. Recall, however, that the appearance of conflict is easily eliminated by taking, in this case, the antecedent of (3b) to be the complement of the antecedent of (3a). (3)

a.

Strong ⎡ ⎛Grade ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣

b.

⎜ ⎜ ⎜ ⎝

*#

⎞

⎤

⎟ ⎟ ⎟ ⎠

⎡ ⎥ ⎥ ⎥ ⎥−→ ⎢ ⎣ ⎥ ⎥ ⎦

ν C0 ν ]Stem

Weak Grade ⎡ ⎢ ν C0 ν ]Stem −→ ⎣

ν

ν

⎤ C0 ν ]Stem ⎥ | ⎦ [+strong strong strong]

⎤ C0 ν ]Stem ⎥ | ⎦ [−strong strong strong]

Perturbed metrical structure is also morphologically conditioned. The distribution of perturbed metrical structure within the paradigm is morphologically speaking fortuitous and is found in the locative and comitative singular, the accusative, illative and locative plural, and the numerically non-characterized essive and partitive. Again, there doesn’t appear to be any principled reason why precisely these morphosyntactic environments should pattern as a class. Elsewhere, non-perturbed metrical structure is the rule. The distribution of perturbed and non-perturbed in disyllabic stems is captured by the constraints in (4a) and (4b). (4a) requires the second syllable of the stem to be stressed, while (4b) requires the second syllable of the stem to be unstressed, by parsing it as the weak syllable of a foot whose left edge is aligned with the left edge of the word. (4a) and (4b) may be interpreted as requiring the building of the first layer of metrical grid structure. Another constraint, not given here, requires the first syllable of the word to carry the main stress.

69


(4)

a.

b.

Perturbed metrical structure ⎡ ⎛ ⎞ ⎤ * # ⎥ ⎢ ⎜ ⎢ ⎜ * # ⎟ ⎟ ⎥ ⎥ ⎢ ⎜ ⎢ ⎜ ⎟ ⎟ ⎥ ⎢ ⎜ ⎟ ⎥ ⎢ ⎜ ⎟ ⎥ ⎢ ⎜ ⎟ ⎥ −→ ⎥ ⎢ ⎜ ⎟ ⎥ ⎢ ⎜ ⎟ ⎥ ⎢ ⎝ ⎠ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ ⎦ ⎣ [Stem σσ ]Stem Non-perturbed metrical structure −→ [Stem σσ ]Stem

∗

[W d

σ(σ . . . ]W d

[W d

(σσ) . . . ]W d

∗

4.2 Consonant centre quantity CC CC) may be filled out as discrete 4.2.1 Hypergeminates. Hypergeminates (CC phonetic geminates (CC) or singletons (C) depending on their morphosyntactic and phonological environment. There is only one case where morphosyntactic environment has direct impact on fill-out, and this is the illative singular form, which is addressed first here. The illative singular form of disyllabic vowel-final nouns is formed by adding the suffix -n to the strong grade form stem. In addition, there is a templatic condition on the shape of the stem as a whole: the disyllabic stem must contain precisely three nuclear moras; that is, fill out as [σ νμμ ][σ νμ ] or [σ νμ ][σ νμμ ]. The quantitative profile is determined in part by the quality of the vowel centre and latus and whether the consonant centre is a hypergeminate or cluster. This will be addressed in detail in section 4.4. Where the consonant centre is a hypersingleton C¯ in the strong grade, the quantity profile of the disyllable may be either q-trochaic or q-iambic. However, when the consonant centre is a geminate ¯ or cluster C ¯ 1 C2 in the strong grade, the quantity profile is always q-trochaic, CC irrespective of vowel quality. The possible patterns are exemplified in Table 10. ¯ is spelt out as a In the illative, a strong grade obstruent hypergeminate BB singleton by (5a). (5b) furnishes additional examples of the pattern. The requirements of the illative singular force the vowel centre to be bimoraic in this environment. Historically, and dialectally too, we also find forms like vúakkun with the phonetic geminate that by default fills out the strong grade form of a hypergeminate. Following a bimoraic nucleus, geminate obstruents were degeminated giving the synchronic pattern represented here.

70

PATRIK BYE

a.

b.

Trochaic illative singular C¯ ‘money *#’ rúuth 2n kÁAsin ‘lamp oil *#’ ‘hand *#’ k´ıeth 2n Iambic illative singular C¯

lÁsAAn ‘door-post *#’ mÁnnAAn ‘egg *#’ vÁˇsoon ‘strong wind *#’ c. Trochaic illative singular C¯ 1 C2 A´ Aston ‘leisure,*#’ A´ AlkAn ‘son *#’ ¯ d. Trochaic illative singular CC m´ıillon ‘mill *#’ lÁAnnAn ‘prison *#’ vúakun ‘hole *#’

( = rútth 2˘ ) ( = kÁAsi) ( = k´ıeth 2˘ ) ( = lÁsAA) ( = mÁnnee) ( = vÁˇsoo) ( = A´stto) ( = A´lkke) ( = m´ıllo) ( = lÁnne) ( = vuákku)

Table 10: Templatic effects in the illative singular (5)

a.

⎡ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣

b.

¯ BB→B in illative singular BB

μμ \/ ν´

C0 | ⎡ ⎤ +geminate geminate ⎣ +strong ⎦ strong −sonorant sonorant

skÁAph in kúatan súukh an

μ | ν ]Stem

⎤ ⎥ ⎥ ⎥ ⎥ ⎥ ⎡ ⎥ ⎥ ⎢ ⎥ ⎥−→ ⎣ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦

‘cupboard *#’ ‘pillow *#’ ‘stocking *#’

⎤ μμ ¬μ μ ⎥ \/ | | ⎦ ν´ C0 ν ]Stem

( = skÁpph i) ( = kuátta) ( = súkkh a)

Outside of the illative singular environment, any strong grade hypergeminate ¯ is filled out as a phonetic geminate Cμ as enforced by (6a). As (6b) shows, it CC ¯ is sonorant or obstruent. is of no importance whether CC

71


(6)

a. ⎡ ⎢ ⎢ ⎢ ⎣

b.

¯ → Cμ CC ν´

C0 ν ]Stem | +geminate geminate +strong strong

¯ NN jóll˘2 lÁnnèn p´ırràn ¯ ii. BB skÁpph i m´ıtth o`n vuákku

i.

⎤

⎡ ⎥ ⎥ ⎢ ⎥−→ ⎣ ⎦

‘crazy * #’ ‘prison *#’ ‘spinning top *#’

⎤ μ ⎥ | ⎦ ν´ C0 ν ]Stem

( = lÁnne) ( = p´ırra)

‘cupboard * #’ ‘measure *#’ ( = m´ıtth o) ‘hole * #’

Any hypergeminate sonorant NN is filled out as a phonetic geminate by (7a). ¯ or weak NN ˘ as shown in (7b). (6a) and (7a) This is true whether it is strong NN NN, thus partially overlap, since both require that a strong grade sonorant hypergeminate surface as a phonetic geminate. (7) ⎡ ⎢ ⎢ ⎢ ⎣

a. ν´

b.

NN → Nμ C0 ν ]Stem | +geminate geminate +sonorant sonorant i.

⎤ ⎥ ⎥ ⎥−→ ⎦

⎡ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣

μ | ν C0 ν ]Stem | [+son]

⎤ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦

In strong grade jóll˘2 m´ıllo kÁAnnun p´ırràn

‘crazy * #’ ‘mill * #’ ‘jug *#’ ‘spinning top *#’

( = kÁnnu)

72

PATRIK BYE

ii.

In weak grade jóoll2 m´ılloo kÁAnnùst p´ırràajn

‘crazy * #’ ‘mill * #’ ‘jug * #’ ‘spinning top * #’

When we turn to obstruent hypergeminate BB BB, an additional variable is relevant to fill-out. The realization depends on whether the obstruent in question is in a position of juncture between two feet or in foot-medial position. This effect is completely absent in hypergeminate sonorants NN NN. In (8a), I give the fill-out constraint for foot-medial position only and provide some relevant examples of the pattern in (8b). We’ll return to the details of fill-out in foot-juncture position in section 4.2.3. (8)

⎡ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣

a.

˘ →B BB

(F t ν´

b.

C0 ν )F t ]Stem | ⎡ ⎤ +geminate geminate ⎣ −sonorant sonorant ⎦ −strong strong

túuph 2 súhaajj-tthàa kúata kúesi

⎤

⎡ ⎥ ⎥ ⎥ ⎢ ⎥−→⎣ ⎥ ⎦

‘sheath * #’ ‘stocking ( )’ ‘pillow * #’ ‘guest * #’

⎤ ¬μ ⎥ | ⎦ ν C0 ν ]Stem

( = túpph 2) ( = súkkh a) ( = kuáttha) ( = kuéssi)

4.2.2 Hypersingletons. This section addresses the fill-out of morphophonemic singletons C. As in the case of the hypergeminates CC CC, it is the sonorants in this class that behave most straightforwardly. In the strong grade, these are filled out as phonetic geminates according to the constraint (9a), as exemplified in (9b).


(9)

a. ⎡ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣ b.

¯ → Nμ N ν´

C0 ν ]Stem | ⎡ ⎤ −geminate geminate ⎣ +sonorant sonorant ⎦ +strong strong

⎤

⎡ ⎥ ⎥ ⎢ ⎥ ⎥−→ ⎣ ⎥ ⎦

73

⎤ μ ⎥ | ⎦ ν´ C0 ν ]Stem

‘stack, pile * #’ ‘fungus growth * #’ ‘half * #’

p´ınnoo tˇsaánna péelli

˘ fills out as phonetic singletons (10a–b). In the weak grade, N (10)

a.

⎡

˘ →N N ν´

⎢ ⎢ ⎢ ⎢ ⎢ ⎣

b.

C0 ν ]Stem | ⎡ ⎤ −geminate geminate ⎣ +sonorant sonorant ⎦ −strong strong

‘stack, pile *

p´ınoo

⎤

⎡ ⎥ ⎥ ⎢ ⎥ ⎥ −→ ⎣ ⎥ ⎦

⎤ ¬μ ⎥ | ⎦ ν´ C0 ν ]Stem

#’

In the context of the reduced vowel 2, hypersingletons are associated with particular phonological effects. In this environment, a hypersingleton is always filled out as geminate, irrespective of obstruent or sonorant status, as described in (11a) ¯ B, and exemplified in (11b). In the case of the hyperobstruents in the strong grade B this is the only context where they surface as phonetic geminates foot-medially at all. (11)

a. ⎡ ⎢ ⎣ b.

¯ → Cμ / V C V

2˘

⎤ C0 2˘ ]Stem ⎥ | ⎦ [+strong strong strong]

lúttˇsh 2˘ sóll˘2 kúss˘2

⎡ ⎢ −→ ⎣

V

⎤ μ ⎥ | ⎦ C0 2˘ ]Stem

‘diarrhoea * #’ ‘fathom * #’ ‘cow * #’

Under the assumption that the reduced vowel in the nominative singular is indeed amoraic, this state of affairs presumably reflects a bimoraic minimum on word

74

PATRIK BYE

size. A non-occurring form such as *lútˇsh 2˘ would on these assumptions fall foul of this requirement, since it contains a sole mora. A separate constraint ensures that the reduced latus vowel surfaces as amoraic domain-finally in the nominative singular. For this constraint, see below, (32) and (33). Preceding a moraic vowel, a hyperobstruent in the strong grade B¯ is filled out as a phonetic singleton B. The relevant constraint is given in (12a) along with examples in (12b). (12)

a.

⎡ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣

b.

¯ →B B ν´

C0 ν ]Stem | | ⎡ ⎤ −geminate geminate ⎣ −sonorant sonorant ⎦ Vμ21 +strong strong

r´ıth aa páath i t´ıath u

⎤

⎡ ⎥ ⎥ ⎢ ⎥ ⎥ −→ ⎣ ⎥ ⎦

⎤ ¬μ ⎥ | ⎦ ν C0 ν ]Stem

‘trap * #’ ‘fathom * #’ ‘knowledge * #’

The reflexes of the weak grade hyperdorsals k˘ and ˘tˇs require some diachronic comment. At some point in their historical development, the weak grade realizations of k˘ and ˘tˇs were, respectively, phonetic singletons *G and *J. Subsequently, however, these both underwent gemination to *GG and *JJ. The reason for this is not clear, although the [+distributed] nature of the articulation may have been a factor in motivating the change. Following gemination, *GG underwent labiodentalization to vv. In this way, the reflex of k˘ is distinct from that of p˘ , which is realized as singleton v. Because k˘ and ˘tˇs pattern together, they are grouped ˘ The relevant constraint is given in (13a) together to form a hyperdorsal class K K. along with examples in (13b). (13)

⎡ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣

a.

˘ → Cμ K ν´ ⎡ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣

C0 | −geminate geminate −strong strong −sonorant sonorant −continuant continuant +dorsal dorsal

ν ]Stem ⎤ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦

⎤ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦

⎡ ⎢ −→ ⎣

⎤ μ ⎥ | ⎦ ν C0 ν ]Stem


b.

k´ıevv2 tÁvvòost mávvòojn éejji skójjèest

‘cuckoo * #’ ‘deed, act * #’ ‘taste * #’ ‘father * #’ ‘useless knife * #’

75

( = k´ıeh2) ( = tÁhoo) ( = máahu) ( = éetˇsh i) ( = skótˇsh ee)

˘ is drawn from the set {p˘ , ˘t , ˘tšš}, it is filled ˘ is not hyperdorsal K ˘ i.e. B Where B K, out in the weak grade as a singleton. The relevant constraint is given in (14a) along with examples in (14b). (14)

a. ⎡ ⎢ ⎢ ⎢ ⎣

b.

˘ →C C ν

C0 ν ]Stem | −geminate geminate −strong strong

luévee kóovèest ´ıDoo pÁAD`ıijn

⎤

⎡ ⎥ ⎥ ⎢ ⎥ −→ ⎣ ⎦

‘shoulder * #’ ‘valley * #’ ‘sprout * #’ ‘pot * #’

⎤ ¬μ ⎥ | ⎦ ν C0 ν ]Stem ( = lúeph i) ( = kóph ee) ( = ´ıth oo) ( = páath i)

4.2.3 Foot-junctural effects. As pointed out earlier, foot position is relevant to the fill-out of morphophonemic obstruents B. Historically, the lengthening of consonant obstruents at the juncture between feet (Foot-junctural Obstruent Gemination, FOG) applied in two separate periods. First, FOG only applied before a bimoraic nucleus. Synchronically, before a bimoraic nucleus, fill-out as geminate is obligatory under foot-juncture. Examples are given in (15b). (15)

a.

Foot-junctural Obstruent Gemination (before bimoraic nucleus) ⎤ ⎡ ⎤ ⎡ C0 νμμ ) F t (F t μ ⎥ ⎢ ⎥ ⎢ | | ⎦ ⎦−→ ⎣ ⎣ C0 −sonorant sonorant

b.

Obligatory foot-junctural gemination of obstruent preceding bimoraic nucleus kuózzàajn túrkk`ıijn skÁpph`ıijn

( = kúos˘2) ‘spruce * #’ ‘messy pig * #’ ( = túrkki) ‘cupboard * #’ ( = skApphi)

76

PATRIK BYE

Later and, crucially, after the application of a process which lengthened stressed short vowels before a stressed syllable (see section 4.4.3 below), FOG was optionally applied preceding a monomoraic nucleus. Synchronically, before a monomoraic nucleus, fill-out as geminate applies optionally, as in (16b). (16)

a.

Foot-initial Obstruent Gemination (before monomoraic nucleus — optional) ⎤ ⎡ ⎡ ⎤ μ C0 νμ )F t (F t ⎥ ⎢ ⎢ ⎥ | | ⎦ ⎦ −→ ⎣ ⎣ C0 −sonorant sonorant

b.

Optional foot-junctural gemination of obstruent preceding monomoraic nucleus kúo(z)z`2st túur(k)k`ıst skAA(p)ph`ıst

‘spruce * #’ ‘messy pig * #’ ‘cupboard * #’

4.3 Consonant quality Originally grade alternation involved nothing more than allophonic variation in duration. However, this allochrony set the stage for a number of adjustments in consonant quality. The most wide-reaching change of this type was the historical spirantization of singleton occlusives (plosives and affricates) in the weak grade ˘ In the weak grade, obstruent hypersingletons generally (both occlusives T˘ T. T ˘ also underwent deaspiration and/or voicing. The synchronic S) and the sibilants S result is that the weak grade forms of obstruent hypersingletons are generally unaspirated continuants. ˘ historically simply k, underwent spiAs already mentioned, the reflex of kk, rantization and voicing to begin with to *G. The dorsal spirants then underwent secondary gemination and the output of this process, *GG, was labiodentalized to ¯ vv. Once k˘ had been spirantized, the strong grade form of the hypersingleton, kk, ˘ were both debuccalized as well as the weak grade form of the hypergeminate, kk kk, to h. ˘ are filled out as 4.3.1 Hypersingletons. In the weak grade, hyperobstruents B unaspirated continuants. (17a) gives the constraint with examples illustrating the generalization in (17b).

77


(17)

⎡ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣

a. ν

b.

˘ → unaspirated continuant B C0 ν ]Stem | ⎡ ⎤ −geminate geminate ⎣ −strong ⎦ strong −sonorant sonorant

luévee kóovèest ´ıDoo pÁAD`ıijn tˇsh éezi éejji A´zee

⎤

⎡

⎥ ⎥ ⎥ ⎥−→ ⎥ ⎦

⎢ ⎢ ⎢ ⎢ ⎢ ⎣

ν

C0 ν ]Stem | ⎡ ⎤ −son ⎣ +cont ⎦ −spr.gl.

⎤ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦

‘shoulder * #’ ( = lúeph i) ( = kóph ee) ‘valley * #’ ( = ´ıth oo) ‘sprout * #’ ( = páath i) ‘pot * #’ ‘paternal uncle * #’ ( = tˇseétsh i) ( = éetˇsh i) ‘father * #’ ‘skinside (of hide) * #’( = A´see)

In the strong grade, information on place becomes relevant, since the reflex of k˘ is not actually filled out as a plosive but as a voiceless laryngeal resonant h as shown in (18a) and (18b). (18)

a. ⎡ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣

b.

k˘ → h ν ⎡ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣

C0 | −geminate geminate +strong strong −sonorant sonorant −continuant continuant +dorsal dorsal

k´ıeh˘2 A´ hee máahu

ν ]Stem ⎤ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦

⎤ ⎥ ⎥ ⎡ ⎥ ⎥ ⎥ ⎢ ⎥ −→ ⎣ ⎥ ⎥ ⎥ ⎥ ⎦

⎤ ν C0 ν ]Stem ⎥ | ⎦ h

‘cuckoo * #’ ‘age * #’ ‘taste * #’

¯ in general are filled out as [+spr.gl.] In the strong grade, hyperobstruents (B B) obstruents as required by (19a). Representative examples are given in (19b).

78

PATRIK BYE

(19)

⎡

¯ → Bh B

a. ν

⎢ ⎢ ⎢ ⎢ ⎢ ⎣

b.

C0 ν ]Stem | ⎡ ⎤ −geminate geminate ⎣ +strong ⎦ strong −sonorant sonorant

lúeph i ´ıth oo tˇsh éetsh i éetˇsh i A´see

⎤

⎡

⎥ ⎥ ⎥ ⎥−→ ⎥ ⎦

⎢ ⎢ ⎢ ⎢ ⎢ ⎣

ν

C0 ν ]Stem | ⎡ ⎤ −son ⎣ −cont ⎦ +spr.gl.

⎤ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦

‘shoulder * #’ ‘sprout * #’ ‘paternal uncle * #’ ‘father * #’ ‘skinside (of hide) * #’

The hypersegment k is filled out as labiodental vv in the weak grade for the historical reasons already explored above. For the gemination seen here, see constraint (13a) above. (20)

a. ⎡ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣

b.

k → vv ν ⎡ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣

C0 | −geminate geminate −strong strong −sonorant sonorant −continuant continuant +dorsal dorsal back +back

k´ıevv2 A´ vvee máavvu

ν ]Stem ⎤ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦

⎤ ⎥ ⎥ ⎥ ⎡ ⎥ ⎥ ⎥ ⎢ ⎥ −→ ⎣ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦

‘cuckoo * #’ ‘age * #’ ‘taste * #’

⎤ ν C0 ν ]Stem ⎥ | ⎦ v

( = k´ıeh˘2) ( = A´ hee) ( = máahu)

4.3.2 Hypergeminates. The hypergeminates CC harbour far fewer surprises ˘ is filled out as h as in (21a) and (21b). than the hypersingletons. kk

79


(21)

˘ →h kk

a.

⎡

ν

⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣

b.

⎡ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣

C0 | +geminate geminate −strong strong −sonorant sonorant +dorsal dorsal back +back

káaha tséehi

ν ]Stem ⎤ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦

⎤ ⎥ ⎥ ⎡ ⎥ ⎥ ⎥ ⎥−→ ⎢ ⎣ ⎥ ⎥ ⎥ ⎥ ⎦

⎤ ν C0 ν ]Stem ⎥ | ⎦ h

‘throttle * #’ ( = kákkh a) ‘notch, dent * #’( = tsékkh i)

More generally, however, obstruent hypergeminates BB are filled out as [+spr.gl.] obstruents in both the strong and weak grades. (22)

⎡

BB → (B)Bh

a. ν

⎢ ⎢ ⎢ ⎣

b.

C0 ν ]Stem | +geminate geminate −sonorant sonorant

i.

ii.

skÁpph i m´ıtth o tsékkh i kuéssi skÁAph i m´ıth oo kúesi

⎤

⎡

⎥ ⎥ ⎥ −→ ⎦

⎢ ⎢ ⎢ ⎣

ν

C0 ν ]Stem | −son +spr.gl.

⎤ ⎥ ⎥ ⎥ ⎦

‘cupboard * #’ ‘measure * #’ ‘notch, dent * #’ ‘guest * #’ ‘cupboard * #’ ‘measure * #’ ‘guest * #’

4.4 Vowel quantity The fill-out of vowel quantity is partly a function of the grade and/or quantity of the consonant centre, partly a function of the qualities of the vowel centre and latus. 4.4.1 Interactions between nuclei and consonant centre. When adjacent to a hypergeminate or cluster in the strong grade, the vowel centre is filled out as short or long, and the latus as short or ultrashort. Examples are given in (23). In the strong grade, bimoraic vowel centres are found exclusively in the illative singular, where the consonant centre is a strong grade geminate or cluster.

80 (23)

PATRIK BYE

a.

kÁnnu túpph 2˘ tsékkh i

‘jug * #’ ‘sheath * #’ ‘notch, dent * #’

b.

ts´ıjkkh o lÁvNNe tˇsu´lkk˘2

‘bitch * #’ ‘turf * #’ ‘spittle * #’

c.

tsúumman ts´ıakh an ts´ıijkh on

‘kiss *#’ ‘notch *#’ ‘bitch *#’

( = tsúmme)

The illative singular takes the strong grade. As we have already seen, the illative singular has a particular templatic requirement of three moras. When the consonant centre is a hypersingleton, the quantity profile of the disyllable stem varies between trochaic and iambic as a function of vowel quality. However, when the consonant centre of the illative singular form is a hypergeminate, we only find a trochaic quantity profile (24). (24)

¯ Fill-out of q-trochee in illative singular with CC ⎡ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣

ν

C0 ν ]Stem | +geminate geminate +strong strong

⎤

⎡ ⎥ ⎥ ⎥ ⎥ −→ ⎢ ⎣ ⎥ ⎥ ⎦

⎤ μμ μ ⎥ \/ | ⎦ ν C0 ν ]Stem

Where the consonant centre is a strong grade hypergeminate, but the word is not in the illative singular, neither nucleus has more than one mora (25). (25)

Fill-out of vowel centre and latus adjacent to CC ⎤ ⎡ ⎡ ⎤ ν C0 ν ]Stem μ μ10 ⎥ ⎢ | ⎥ ⎢ ⎢ ⎥ | | ⎥ −→ ⎣ ⎢ ⎦ +geminate geminate ⎦ ⎣ ν C0 ν ]Stem +strong strong

Where C0 is neither strong grade nor hypergeminate, the vowel centre may be short or long and the latus short, long or ultrashort.


(26)

81

Vowel quantity when consonant centre not strong hypergeminate ⎡ ⎢ ⎢ ⎢ ⎣

ν

C0 ν ]Stem | +geminate geminate ¬ +strong strong

⎤

⎡ ⎥ ⎥ −→ ⎢ ⎥ ⎣ ⎦

⎤ μ21 μ20 ⎥ | | ⎦ ν C0 ν ]Stem

In the last part of this section, we turn to the reduced vowel, which presents some interesting additional complications. According to the Itkonen’s data, the reduced vowel transcribed here as 2 surfaces as ultrashort (interpreted here as amoraic) in a latus following a consonant in the strong grade.4 This unusual phonology characterizes the nominative singular form of disyllabic vowel stems ending in -2, as shown in (27). (27)

a. ⎡ ⎢ ⎣

b.

Fill-out of /2/ as amoraic following strong grade consonant centre ⎤ ⎡ ⎤ ν ]Stem ν´ C0 ¬μ ⎥ ⎢ ⎥ | | | ⎦ −→ ⎣ ⎦ strong 2 ν C0 ν ]Stem +strong i.

sóll˘2 kúss˘2 lúttˇsh 2˘ rútth 2˘

ii.

k´ıell˘2 k´ıeh˘2 rúotˇsh 2˘

‘fathom * #’ ‘cow * #’ ‘thin diarrhoea * #’ ‘money * #’ ‘language * #’ ‘cuckoo * #’ ‘crash, clatter * #’

Disyllables with a reduced vowel latus are also associated with special quantity profiles, an issue which is addressed in the next section. Pekka Sammallahti (p.c.) informs me that the distinction between 2 and 2˘ is no longer attested, at least he has been unable to find it in his own informants. For this reason he suspects the distinction may be nothing more than a theoretical artefact from Itkonen’s work. My own feeling is that, at least at the time Itkonen carried out his fieldwork in the first half of the twentieth century, this distinction was real. My basis for saying this is the fact that hyperobstruents must always be geminate preceding 2˘ , which I take to reflect a bimoraic minimum on lexical stem size. The assumption that the reduced vowel was indeed amoraic in the nominative singular meshes well with this interpretation. Indeed, it is difficult to see what alternative explanation there might be for this pattern. The loss of the distinction observed by Sammallahti, if such it is, brings still further opacity to the phonology of Inari Saami. 4

82

PATRIK BYE

4.4.2 Vowel quantity as a function of vowel quality. The second major determiner of phonetic vowel quantity is the quality of the vowel centre and latus. As discussed in the last section, both vowel centre and latus surface with a moraicity less than or equal to 1 in the context of a consonant centre containing a cluster or hypergeminate in the strong grade. However, in the complement of this environ +geminate geminate , the disyllable may surface with one of two quantitative ment, ¬ +strong strong (q-)profiles, which we may dub q-trochaic and q-iambic. A quantitative trochee (q-trochee) is a disyllabic unit in which the nucleus of the first syllable is bimoraic and the nucleus of the second syllable is monomoraic: [σ νμμ ][σ νμ ]. A quantitative iamb is a disyllabic unit in which these relations are reversed. In a q-iamb the nucleus of the first syllable is monomoraic, while that of the second is bimoraic: [σ νμ ][σ νμμ ]. In either case, the moraicity of the consonant centre is irrelevant to the fill-out of vowel quantity. Simplifying for the moment, a qtrochee generally results if the latus is a high or reduced vowel (28a), while, if the latus is non-high, the disyllable is filled out as a q-iamb (29a). Note that the q-iamb is not found with the reduced latus vowel 2, even though it is non-high. Instead, the presence of a reduced vowel in the latus results in a q-trochee. There are a number of departures from these basic patterns, which we shall look at in more detail below. (28)

a. ⎡ ⎢ ⎢ ⎢ ⎣

b.

Trochaic quantitative profile ν

i.

C0 ν ]Stem | +geminate geminate ¬ +strong strong

VVC0 i káarri páath i tˇséetsh i éetˇsh i péelli tˇsuáhi lúeph i

⎤

⎡ ⎥ ⎥ ⎢ ⎥ −→ ⎣ ⎦

⎤ μμ μ ⎥ \/ | ⎦ ν C0 ν ]Stem

‘wooden trough * #’ ‘pot * #’ ‘paternal uncle * #’ ‘father * #’ ‘half * #’ ‘low tide * #’ ‘shoulder * #’


ii.

VVC0 u p´ıath u k´ıatˇsh u máaˇsu máahu

iii. VVC0 2 kúuz2 jóoll2 rúojj2 (29)

⎡ ⎢ ⎢ ⎢ ⎣

a. ν

b.

83

‘wild beast * #’ ‘shepherding * #’ ‘composure * #’ ‘taste * #’ ‘cow * #’ ‘crazy * #’ ‘clatter * #’

( = kúss˘2) ( = jóll˘2) ( = rúotˇsh 2˘ )

Iambic q-profile with non-high latus C0 ν ]Stem | | +geminate geminate ¬ [−high] +strong strong

⎤

⎡ ⎥ ⎥ ⎢ ⎥ −→ ⎣ ⎦

i.

AC0 AA lÁsAA spÁhAA

‘door-post * #’ ‘saddlepack * #’

ii.

iC0 aa st´ıhaa r´ıth aa

‘boot * #’ ‘trap * #’

iii. VC0 ee tˇsu´DDee mÁnnee kóph ee ´ıˇsee A´see

‘foe * #’ ‘egg * #’ ‘valley * #’ ‘help * #’ ‘skinside * #’

iv. VC0 oo p´ınnoo tÁhoo vÁˇsoo

‘pile * #’ ‘deed, act * #’ ‘strong wind * #’

v.

VC0 aa tˇsuájaa tˇsánaa tˇs´ıavvaa

⎤ μ μμ ⎥ | \/ ⎦ ν C0 ν ]Stem

( = tˇsu´ ajja) ‘bay * #’ ‘birch fungus * #’ ( = tˇsáanna) ( = tˇs´ıaha) ‘cache * #’

84

PATRIK BYE

There are two interesting reversals of these expectations. One of these involves glide-initial clusters. Rimal glides are well-known to be ambiguous with regard to their role as codas or as part of the nucleus. In this case, the glide seems to lend extra prominence to the first nucleus. At the same time, the glide apparently doesn’t contribute a mora to ν1 , but conditions a bimoraic vowel centre, which results in q-trochees instead of the expected q-iambs. This is shown in (30). (30)

Trochaic q-profile with non-high latus and glide-initial cluster in C centre a.

⎡ ⎢ ⎣

b.

ν

⎤ ⎡ GC ν ]Stem ⎥ ⎢ | | ⎦ −→ ⎣ [−high] −strong strong

k´ıivse lÁAvNe ts´ıijho

⎤ μμ μ ⎥ \/ | ⎦ ν GC ν ]Stem

‘annoyance * #’ ‘turf * #’ ‘bitch * #’

The second reversal is far more unexpected and arises where both the vowel centre and the latus contain a low vowel (either as a monophthong or as the second component in a rising diphthong), and the latus is in addition [−back] a. Where the consonant centre is a strong grade cluster or geminate, the usual constraints fill out the vowel centre and latus as short. However, across a hypersingleton in the strong grade, this combination of vowel centre and latus qualities yields a qtrochee rather than the expected q-iamb. Thus tˇsáanna (‘birch fungus’ #, not *tˇsańnaa. However, this reversal does not apply in the weak grade, e.g., tˇsańaa * #, which gives the expected q-iamb.5 This is shown in (31a–b). Trochaic profile with low vowel centre and low latus across C¯

(31) ⎡ ⎢ ⎢ ⎢ ⎣

5

a. ν |

C0 ν ]Stem | | −geminate geminate [+low] a +strong strong

⎤

⎡ ⎥ ⎥ ⎢ ⎥ −→ ⎣ ⎦

⎤ μμ μ ⎥ \/ | ⎦ ν C0 ν ]Stem

Here I am taking the accusative-genitive singular form as reverting to the iambic pattern. However, the situation described by Itkonen (1946) is more complex than this. For some speakers, the weak grade form is filled out as a q-spondee: tˇsáanaa. This q-profile is unique to words containing low vowels.


b.

tˇsuájja tˇsaánna tˇs´ıaha

85

‘bay ( #’ ‘fungus growth on base of birch ( #’ ‘hiding-place, cache ( #’

The reduced vowel is also associated with special q-profiles, although these simply involve variation on the trochaic theme. In general, the trochaic q-profile is [σ νμμ ][σ νμ ]. Where the latus is ultrashort 2˘, however, we find [σ νμμ ][σ ν˘ ] and [σ νμ ][σ ν˘ ]. If the vowel centre is a non-low monophthong, it is filled out as monomoraic (32). Otherwise, if the vowel centre is a diphthong or a low monopthong, it is filled out as bimoraic (33). (32)

Non-low nucleus preceding ultrashort vowel a.

⎡ ⎢ ⎣

(33)

⎤ ⎡ 2˘ ]Stem ν C0 ⎥ ⎢ | | ⎦ −→ ⎣ −low +strong strong

‘diarrhoea ( #’ ‘fathom ( #’ ‘cow ( #’

b.

lúttˇsh 2˘ sóll˘2 kúss˘2

a.

Nucleus preceding ultrashort latus ⎡ ⎢ ⎣

b.

ν

k´ıell˘2 ˇsAÁll˘2 k´ıeth 2˘

⎤ μ ⎥ | ⎦ ν C0 2˘ ]Stem

⎤ ⎡ 2˘ ]Stem C0 ⎥ ⎢ | ⎦ −→ ⎣ +strong strong

⎤ μμ ⎥ \/ ⎦ ν C0 2˘ ]Stem

‘language ( #’ ‘pity ( #’ ‘hand ( #’

4.4.3 Vowel quantity as a function of foot-structure. So far we have looked at the impact of vowel quality on vowel length within the foot. The vowel centre may also be subject to lengthening where the syllable which dominates it is followed immediately by a syllable carrying secondary stress. Historically, this configuration triggered lengthening of a short vowel centre resulting in palliation ´˘ σ ´¯ σ of the stress clash: cross-linguistically σ ` is preferred over σ ` . This process

86

PATRIK BYE

was crucially ordered following the application of FOG preceding a stressed bimoraic nucleus but before FOG preceding a stressed monomoraic nucleus. FOG before a stressed bimoraic nucleus destroyed the environment for Vowel Lengthening under Foot Minimality, but Vowel Lengthening or Foot Minimality did not bleed the later application of FOG preceding a stressed monomoraic nucleus. Synchronically, vowel length is blocked preceding a moraic consonant centre followed by a stressed bimoraic nucleus. (34)

Vowel length blocked ⎤ ⎡ ⎡ μ21 μ μμ ⎥ ⎢ ⎢ | | \/ ⎦−→ ⎣ ⎣ ν´ C0 ν` ]Stem

(35)

kúzzàajn ruójjàajn skújj`ıijn mávvòojn lÁzzÀAst st´ıvvàast

⎤ μ μ μμ ⎥ | | \/ ⎦ ν C0 ν ]Stem

‘cow * #’ ‘crash, clatter * #’ ‘useless knife * #’ ‘taste * #’ ‘door-post * #’ ‘boot * #’

If one or the other, or both, of these two conditions is lacking, the vowel centre fills out as long preceding a stressed latus. (36)

Vowel length

(37)

ν´ C0 ν` ]Stem

⎡ ⎢ −→ ⎣

⎤ μμ ⎥ \/ ⎦ ν´ C0 ν` ]Stem

a.

sóolàajn péel`ıijn tˇsuúD`ıijn r´ıiDàast tˇsuúDèest móon`ıijn p´ıaDòojn k´ıelàajn

‘fathom * #’ ‘half * #’ ‘foe * #’ ‘trap * #’ ‘foe * #’ ‘egg * #’ ‘wild beast * #’ ‘language * #’

b.

kúuzz`2jjn lúujj`2st

‘cow * #’ ‘diarrhoea, mud * #’


87

From a rule-ordering perspective, such lengthening under foot-minimality may be transparent or opaque (37). In (37b), there is no surface-apparent reason for the vowel centre to be long, since the consonant centre has been geminated. Historically, lengthening under foot-minimality must have applied before the secondary gemination of j and z to jj and zz. 5. Conclusions The analysis accounts for a significant proportion of the morphophonemics of Inari Saami using inviolable constraints and shows, hopefully, that Abstract Declarative Phonology (ADP) is equal to the task of describing both the kind of abstract systematic relationships and multiple opacity phenomena which escape a more concrete declarative approach. The present contribution makes no pretence at being a complete analysis of the morphophonemics of Inari Saami. The briefest perusal of some of the sources cited at the beginning of this article will confirm how far it falls short of that. In its classic formulation, Declarative Phonology builds on a concrete conception of morphophonemics. This concreteness, combined with DP’s monostratal architecture brings with it some challenging descriptive problems. I have proposed a way of reconstruing ‘abstract morphophonemes’ not as empty diacritics, but as categories whose content is detailed phonetic memories, or exemplars. If the content of phonological categories in general may be described in terms of exemplars, there is no reason why phonological categories should have coherent phonetic content. Phonetic coherence of a category is identified here with the property of having a single well-defined exemplar cloud. Coherence in this sense is by hypothesis necessary in order to draw up an articulatory plan, i.e. for the phonetic interpretability of the category. However, interpretability is a requirement of the interface to phonetic interpretation — it is not a requirement that holds of category formation in memory, and hence the systemic categories of the lexicon may be phonetically non-coherent; that is, hypercategories. Inari Saami provides an example of a language whose systemic categories evince considerable phonetic non-coherence. The function of the phonology may be understood as ensuring that phonological representations meet the requirements of phonetic interpretation. References Bergsland, Knut. 1948. Røros-lappisk grammatikk. Oslo: Instituttet for Sammenlignende Kulturforskning.

88

PATRIK BYE

Bird, Steven. 1995. Computational Phonology: A Constraint-based approach. Cambridge: Cambridge University Press. Bird, Steven & Ewan Klein. 1994. “Phonological analysis in typed feature systems”. Computational Linguistics 20.455–491. Bye, Patrik. 1997a. “A generative perspective on ‘overlength’ in Estonian and Saami”. Estonian Prosody: Papers from a symposium ed. by Ilse Lehiste & Jaan Ross. 36–70. Tallinn: Eesti Keele Instituut. . 1997b. “Representing overlength”. Phonology in Progress – Progress in Phonology ed. by Geert Booij & Jeroen van de Weijer. (= HIL Phonology Papers 3.). 61–101. The Hague: Holland Academic Graphics. . 2001. Virtual Phonology. Rule Sandwiching and Multiple Opacity in North Saami. Ph.D. thesis, University of Tromsø. Coleman, John. 1998. Phonological Representations. Cambridge: Cambridge University Press. Eriksson, Manne. 1961. “Svensk Ljudskrift 1878–1960. En översikt över det svenska landsmålsalfabetets utveckling och användning huvudsakligen i tidskriften Svenska Landsmål”. Svenska Landsmål och Svenskt Folkliv B62. Stockholm. Itkonen, Erkki. 1939. Der ostlappische Vokalismus vom qualitativen Standpunkt aus mit besonderer Berücksichtigung des Inari- und Skoltlappischen. (= Mémoires de la Société Finno-Ougrienne 79.) Helsinki: Suomalais-Ugrilainen Seura. . 1946. Struktur und Entwicklung der ostlappischen Quantitätssysteme. (= Mémoires de la Société Finno-Ougrienne 88.) Suomalais-Ugrilainen Seura. . 1971. “Ehdotus inarinlapin fonemaattiseksi transkriptioksi”. Lapin murteiden fonologiaa ed. by Erkki Itkonen, Terho Itkonen, Mikko Korhonen & Pekka Sammallahti. (= Castrenianumin toimitteita 1.), 43–67. Helsinki: Suomalais-Ugrilainen Seura. . 1973. “Phonetische und phonologische Betrachtung der ostlappischen Dialekte”. Journal de la Société Finno-Ougrienne 72.129–149. . 1986–1991. Inarilappisches Wörterbuch. (= Lexica Societatis FennoUgricae 20.) Helsinki: Suomalais-Ugrilainen Seura. 4 volumes. . 1992. Inarinsaamelaisia kielennäytteitä — Aanaarkiela cˇ a˙ jttuzeh. (= Mémoires de la Société Finno-Ougrienne 213.) Suomalais-Ugrilainen Seura. Edited by Lea Laitinen.


89

Johnson, Keith. 1997. “Speech perception without speaker normalization”. Talker Variability in Speech Processing ed. by Keith Johnson & John W. Mullennix, 145–166. San Diego: Academic Press. Kingston, John & Randi L. Diehl. 1994. “Phonetic knowledge”. Language 70.419–454. Magga, Tuomas. 1984. Duration in the Quantity of Bisyllabics in the Guovdageaidnu Dialect of North Lappish. (= Acta Universitatis Ouluensis. Series B Humaniora 11.) Oulu: University of Oulu. Ogden, Richard. 1999. “A declarative account of strong and weak auxiliaries in English”. Phonology 16.55–92. Pierrehumbert, Janet. 2001. “Exemplar dynamics: Word frequency, lenition and contrast”. Frequency Effects and Emergent Grammar ed. by Joan Bybee & Paul Hopper, 137–157. Amsterdam and Philadelphia: John Benjamins. Sammallahti, Pekka. 1977. Norjansaamen itä-enontekiön murteen äänneoppi. (= Mémoires de la Société Finno-Ougrienne 160.) Helsinki: SuomalaisUgrilainen Seura. . 1984. “The phonology of the Guovdageaidnu dialect of North Saami: Some notes”. Riepmoˇcála. Essays in Honor of Knut Bergsland ed. by Bernt Brendemoen, Even Hovdhaugen & Ole Henrik Magga, 136–152. Oslo: Novus. . 1998a. The Saami Languages. An Introduction. Karasjok, Norway: Davvi Girji. . 1998b. “Saamic”. The Uralic Languages ed. by Daniel Abondolo, 43–95. London: Routledge. Sammallahti, Pekka & Matti Morottaja. 1993. Säämi-suomâ sänikirje. Inarinsaamelais-suomalainen sanakirja. Utsjoki, Finland: Girjegiisá. Scobbie, James M. 1991. Attribute Value Phonology. Ph.D. thesis, University of Edinburgh. . 1993. “Constraint violation and conflict from the perspective of Declarative Phonology”. Canadian Journal of Linguistics 38.155–168. Scobbie, James M., John S. Coleman & Steven Bird. 1996. “Key aspects of declarative phonology”. Current Trends in Phonology: Models and methods ed. by Jacques Durand & Bernard Laks, 685–710. Salford: ESRI. Smith, Eliot R. 1998. “Mental representation and memory”. The Handbook of Social Psychology ed. by Daniel T. Gilbert, Susan T. Fiske & Gardner Lindzey, 391–445. Boston: McGraw-Hill.

90

PATRIK BYE

Walsh-Dickey, Laura. 1997. The Phonology of Liquids. Ph.D. thesis, University of Massachusetts, Amherst.

PRODUCTIVE SYNCRETISM IN SAAMI INFLECTIONAL MORPHOLOGY∗ GUNNAR ÓLAFUR HANSSON University of British Columbia Syncretism, whereby a single form represents several ‘paradigm cells’, is a topic of controversy in current morphological theory. A key question is whether such form-to-form identity relations constitute grammatical facts in their own right (represented by ‘rules of referral’ or analogous devices). I argue for the importance of diachronic evidence in this context, bringing data from Saami to bear on the debate. Two developments are examined: (a) the conflation of the previously separate inessive and ablative into a unified ‘locative’ case; (b) the identity between comitative singular and locative plural forms. I demonstrate how the = syncretism was extended from polysyllabic to monosyllabic stems in eastern Finnmark dialects of North Saami, indicating that the identity pattern constituted a productive aspect of the inflectional system. The non-contiguity of the paradigm cells involved argues in favour of ‘referrals’; the Saami facts are simply incompatible with more restrictive theories.

1. Introduction Syncretism is the cross-linguistically widespread phenomenon whereby a single inflected form fits more than one morphosyntactic description (Spencer 1991:45). In Latin, for example, where dative and ablative are typically distinct, and are identical in each and every inflectional class (serv-¯ıs of serv-us “slave”, mar-ibus, of mar-e “sea”). As a general phenomenon, syncretism is of great importance to morphological theory. For example, syncretism is allotted chapter-length treatment in the studies of inflectional morphology by Carstairs (1987) and Stump (2001), and is ∗

I would like to thank Matthew Baerman, Andrew Garrett, Sharon Inkelas, the editors, and five anonymous reviewers for discussion, comments and suggestions which greatly helped to improve this work at various stages of its development. Needless to say, any remaining errors are my own.

92

GUNNAR ÓLAFUR HANSSON

the topic of several contributions in Plank (1991a). Syncretism has been the central focus of much recent work by Greville Corbett, Matthew Baerman and other associates in the Surrey Morphology Group, culminating in the first booklength study entirely devoted to the topic (Baerman et al. 2005; see also Baerman et al. 2001; Corbett et al. 2001; Evans et al. 2001; Baerman 2004). The group’s website (http://www.surrey.ac.uk/LIS/SMG) contains several useful resources on syncretism, including an annotated bibliography and two web-searchable typological databases.1 Despite its importance, syncretism remains rather poorly understood, and works on this topic have espoused a wide range of different, often conflicting, views regarding its nature and formal analysis. A common view is that syncretism arises through the neutralization of morphosyntactic oppositions (Jakobson 1936; Williams 1981; Neidle 1988; Blevins 1995; Johnston 1997; Noyer 1997, 1998; Kiparsky 2001). A competing view allows for the alternative possibility that syncretism be taken at face value, as a purely morphological fact, whereby the realization of one inflectional form is systematically ‘referred to’ that of another form within the same paradigm (Zwicky 1985; Corbett & Fraser 1993; Stump 1993, 2001; Evans et al. 2001). This paper contributes to the ongoing debate about syncretism as a morphological phenomenon, its typological properties cross-linguistically, and its formal analysis. As such, the paper has a twofold goal. The first is to emphasize the relevance of diachronic evidence, especially as regards the difficult task of distinguishing between systematic inflectional homonymy patterns (genuine syncretism) and ones that can legitimately be called accidental (Zwicky 1991). Secondly, various diachronic changes affecting the inflectional morphology of the Saami languages, in particular North Saami and its dialects, are examined from the perspective of syncretism as a (potentially productive) synchronic morphological phenomenon. The main focus is on inflectional marking involving the locative (inessive/elative) case, including a pervasive identity pattern, observed in North Saami and elsewhere, whereby comitative singular = locative plural (thus biillain means either ‘by car, with the car’ or else ‘in (the) cars, from (the) cars’).2 Dialectal 1

Baerman et al. (2005) appeared shortly before the present volume went to press; as that monograph was unavailable to me, none of its content is reflected here. 2 All forms cited are in the standard orthography for North Saami (and for the other Saami languages, where relevant); phonemic transcription is provided where necessary. Note that postvocalically, denotes a glide /j/ (biillain = /pijl:ajn/). Note also that stands for /a:/ (dialectally /æ:/), and , , represent voiceless unaspirated stops /p t k/.

PRODUCTIVE SYNCRETISM IN SAAMI INFLECTIONAL MORPHOLOGY

93

differences within North Saami with respect to the scope of this identity pattern are argued to be the result of the analogical extension of the syncretism from polysyllabic to monosyllabic stems in certain dialects. The syncretism is thus shown to be a productive aspect of North Saami inflectional morphology, which in turn has serious implications for our understanding of syncretism as such. The Saami facts are powerful evidence in favour of frameworks which allow for syncretisms to be stipulated by means of ‘rules of referral’ or analogous devices. Furthermore, the Saami evidence directly contradicts proposals which limit syncretism to geometrically contiguous regions within a paradigm (McCreight & Chvany 1991; Johnston 1997). The paper is organized as follows: some of the main alternatives that have been pursued in the formal analysis of syncretism are introduced in section 2, with particular attention to inferential-realizational theories of morphology (to use the terminology of Stump 2001); the problem of distinguishing between systematic and accidental inflectional homonymy is also addressed. The notion of productive syncretism in diachronic change — the analogical extension of an identity pattern from one paradigm to another — is introduced in section 3, and illustrated with detailed discussion of the historical development of locative case marking in the Saami languages. Proposals regarding potential constraints on syncretism patterns are considered in section 4, in particular constraints that would prohibit syncretisms relating disjoint (non-contiguous) paradigm cells. In section 5, the = syncretism found in many Saami languages is examined in detail, focusing on its development in the Finnmark dialects of North Saami, and the theoretical implications of these facts are considered. The general conclusions are summarized in section 6. 2. Syncretism and its analysis 2.1 Syncretism as inflectional homonymy In a broad sense, the term syncretism simply refers to an identity relation, that might be called inflectional homonymy. From this perspective, any instance of phonologically identical forms occurring within the same inflectional paradigm would count as an instance of syncretism. Consider the ambiguity of the English form called: it can be either the finite past tense form of the verb or the past participle of that same lexeme. Under this simplistic view, the ambiguity of a form like called is essentially no different from that of, say, the form match (as in tennis match versus box of matches). That is, syncretism is merely a special case of a many-to-one relationship between meaning and form, one where such a relationship happens to hold within the inflectional paradigm of individ-

94


ual words. In its most general form, this view is of little theoretical interest, as it ignores the systematic character of syncretism patterns, treating them as being on a par with accidentally homophonous word pairs. Consequently, no predictions are made about what types of syncretism patterns are possible in human language, whether there are any constraints on their realizations, and so forth. 2.2 Syncretism and feature neutralization One commonly held position on the nature of syncretism is that it arises from the neutralization or underspecification of morphosyntactic features (Jakobson 1936; Bierwisch 1967; Williams 1981; Neidle 1988; Blevins 1995; Johnston 1997; Noyer 1997, 1998; Kiparsky 2001). The underlying assumption is that the syncretized forms share an array of morphosyntactic and/or semantic feature specifications separating them from other cells in the paradigm. The syncretism itself is a matter of neutralizing, or leaving unspecified, the feature (or features) distinguishing the two cells from one another. For example, Bierwisch (1967) uses underspecification to capture the pervasive / syncretism found throughout much of German nominal inflection: / diese ‘this’, / dieses, (all genders) diese, and so on. Bierwisch interprets the German four-way case contrast in terms of two binary features, essentially [±oblique] ( versus ) and [±object] ( versus ). Syncretism arises because among the transformational rules responsible for suffixation there happen not to be any that distinguish between [-obl, -obj] (= nominative) and [-obl, +obj] (= accusative). For example, the rules adding -es in neuter singulars (dieses) or -e in feminine singulars (dies-e) refer only to [-obl], without mentioning the feature [±obj]. An alternative perspective, which is identical for all intents and purposes, would hold that a suffixal morpheme like -es is, as a lexical entry, specified as [+neuter], [+singular], [-oblique], and so on, but carries no value at all for [±object]. An actual syncretic form like dieses is, on this view, neither nominative nor accusative, strictly speaking, but simply ‘direct’ (as opposed to oblique). This is essentially what Blevins (1995) refers to as ‘artifactual’ syncretism: rather than a homophony relation between two inflected forms, we are in fact dealing with a single inflected form whose content is to some extent featurally impoverished or underspecified. A related approach, which takes the notion of impoverishment a step further, is advocated by Noyer (1997, 1998). He suggests that when two paradigm cells are linked through syncretism, the (marked) feature value distinguishing one cell from the other actually undergoes outright deletion by a so-called im-


95

poverishment rule. As a result, the set of morphosyntactic properties for that cell is defective, and is filled in with default (unmarked) values supplied by redundancy rules. The German example above would then, given the featural system assumed by Bierwisch (1967), be the result of an impoverishment rule along the lines of (1a), combined with a redundancy rule like (1b). (1)

a.

[+object] →

Ø/ [+N, -oblique, . . . ]

b.

[] →

[-object]

On this interpretation, we might again say that a syncretic form like dieses is actually a single form that does double duty, rather than two separate but homophonous forms. It should be noted that although underspecification plays a central role in Noyer’s approach, underspecified representations only occur as an intermediate stage: after the application of a neutralizing impoverishment rule but before that of the relevant redundancy rule.3 All accounts of syncretism relying on underspecification, neutralization or related notions share a fundamental assumption: that the contrast being neutralized reflects underlying (covert, not necessarily apparent) affinities at the morphosyntactic and/or semantic level. Gvozdanović (1991:153) articulates this assumption by stating that “syncretism as an identity in form presupposes an identity at some level in meaning” (though she hastens to add, as an exception clause, “unless this formal identity is due to phonological developments or distributional restrictions”), and a section heading is particularly telling of this perspective: “Patterning of grammatical meaning as revealed by syncretism”. On this view, the fact that two paradigm cells display syncretism is evidence that they form a neutralizable opposition (specified with ‘+’ versus ‘−’ for some binary feature, for example) at that hidden level of grammatical structure. The assumption that syncretism patterns reflect underlying patterns of grammatical organization formed the basis of much theoretical work in morphology in the structuralist tradition (Hjelmslev 1935; Jakobson 1936; see Plank 1991b on the ancient provenance of this view). 3

As pointed out by one reviewer (see also Noyer 1998), impoverishment rules become very similar to rules of referral (see section 2.3) as soon as the reference to markedness is removed. While this is true, complications arise when the two paradigm cells are disjoint, as in some of the crucial examples discussed below. For example, in order to syncretize with in North Saami (see section 5), one needs two separate but crucially interdependent impoverishment rules: one eliminating the feature for comitative case, but only in the singular, the other rule eliminating the singular feature, but only in the comitative.

96


2.3 Syncretism and rules of referral The neutralization approach derives syncretism from the organization of paradigms at the level of morphosyntactic or semantic oppositions. A competing view allows for syncretism to be taken at face value as a purely morphological fact, essentially as an outright stipulation. The realization of a particular inflectional form is systematically ‘referred to’ that of another form within the same paradigm (Zwicky 1985; Stump 1993; note also the notion of ‘take-over’ in Carstairs 1984, 1987). This alternative to the neutralization approach is compatible with most theories of inflectional morphology in Hockett’s (1954) ‘word-and-paradigm’ category, such as those of Matthews (1972); Zwicky (1985); Anderson (1992); Corbett & Fraser (1993); and Stump (2001). These are all, in Stump’s (2001) classification, inferential-realizational theories of inflectional morphology. They are inferential, rather than lexical, in that they see the relation between a lexeme’s root and its fully inflected word forms as being expressed by rules or formulas, not by the concatenation of lexically listed objects each carrying their own morphosyntactic properties. In other words, they assume that affixes do not exist as lexical entries (morphemes) in their own right. Moreover, such theories are realizational, rather than incremental, in that an inflected word does not acquire its morphosyntactic properties in a piecemeal fashion by accumulating the elements expressing those properties (for example, affixes). Instead, the presence of those elements is simply licensed by the presence of the properties in question. As it is the most formally elaborated of all the frameworks incorporating rules of referral, the Paradigm Function Morphology of Stump (2001) will serve as an illustrative example. In this framework, each cell in the inflectional paradigm of some lexeme L is defined as a form/property-set pairing , where Y is the inflected form as such (its phonological representation) and σ is its complete set of morphosyntactic properties. The inflectional morphology of a given language is simply a function in the set of such pairings, a ‘paradigm function’ (PF), which serves as a well-formedness condition. Given a lexeme L, whose root is X, a pairing is licensed as a cell in L’s inflectional paradigm if and only if applying the paradigm function to the root pairing <X, σ> yields as its result, that is, if PF(<X, σ>) = . For example, if σ stands for the morphosyntactic property set { :comitative,

:pl}, then (2) is true given the PF that constitutes North Saami inflectional morphology (mánná “child”, mánáiguin “with children”). (2)

PF(<mánná, σ>) =<mánáiguin, σ>


97

The paradigm function is in turn defined in terms of realization rules (RR): the individual morphological rules of the language. Typically these take the form of rules of exponence, formalized as in (3), where τ stands for the set of morphosyntactic properties the rule is expressing and C the class of lexemes encompassed by the rule. (3)

RRτ,C (<X, σ>) =def

Here and throughout this chapter, the statement of realization rules has been greatly simplified (relative to Stump 2001), for example by ignoring the notion of rule blocks. Each rule is here assumed to apply directly to the root, and an inflected form is defined by a single rule, not a nested series of rules affiliated with separate rule blocks (such as RR ,φ,C (RR,υ,C (RR,τ,C <X, σ>)) for rule blocks through ). If we ignore the vagaries of Saami consonant gradation and other morphophonological alternations, and also assume for the sake of the argument that the plural marker -i- (/-j-/) and the following case ending are introduced jointly by a single rule (rather than two rules residing in different blocks), then (4) is not implausible as one of the rules of exponence in the North Saami inflectional system. (4)

RR{ :com, :pl},N (<X, σ>) =def <Xiguin, σ>

Inflected forms whose shapes are affected by syncretism derive from realization rules of a second type: rules of referral, a term originally introduced by Zwicky (1985). Such rules refer the realization of a given cell in a lexeme’s paradigm to the realization of another cell in that same paradigm. In Zwicky’s original formulation, these are essentially statements like the one in (5), describing the German / syncretism discussed earlier. Zwicky’s ‘Nounal’ covers nouns, adjectives and determiners; note that (5) expresses a default, which is overridden by a more specific rule in the masculine singular (where we find -en versus -er). (5)

In the context of [ [ :nom].

:Nounal], [ :acc] has the same realization as

The format for rules of referral in the Paradigm Function Morphology of Stump (1993, 2001) is shown, somewhat simplified, in (6).4 4

As before, the concept of rule blocks has been left out, and with it the ability to distinguish

98 (6)


RRτ,C (<X, σ>) =def , where PF(<X, σ/ρ>) =

The expression σ/ρ stands for the set of morphosyntactic properties associated with the other paradigm cell in the referral relation: the one to which the realization of <X, σ> is being ‘referred’. In other words, σ/ρ is identical to σ except for whatever properties are mentioned explicitly in ρ. For example, if the realization of some case form is being referred to that of the nominative form, then σ/ρ should read ‘σ/{ :nom}’. Translated into this formalism, the statement in (5) reads approximately as (7). (7)

RR{ acc},N ounal (<X, σ>) =def , where PF(<X, σ/{ :nom}>) =

The second line of (7) simply states that the phonological realization of a nominative form of a word with root X is independently defined as Y by the inflectional system (PF) of the language in question. In the case of a German adjective like jung “young”, X is jung- and the PF of German independently defines Y as junges when σ contains the properties {+neut,

:sg, :strong}. To put it differently, a separate rule of exponence exists in German which suffixes -es in the strong forms of adjectives. The rule in (7) simply states that the corresponding accusative form — that is, the paradigm cell which contains all the same feature values except for having accusative instead of nominative case — uses the same phonological form Y (here, junges) as its realization.5 between whole-word syncretisms and so-called block syncretisms (Stump 2001:217-218). Also omitted is the notion of rule competition by P¯anini’s principle; in Stump’s formulation, the righthand clause in (6) does not refer to PF(<X, σ/ρ>) directly, but to Narn (<X, σ/ρ>), where ‘Narn ’ stands for the ‘narrowest applicable rule’ in rule block n given the property set σ/ρ. 5 Syncretism and referrals are approached in a similar manner in Network Morphology (Corbett & Fraser 1993; Evans et al. 2001). Here lexical information is organized as a network in the form of an inheritance hierarchy (usually formalized in the knowledge-representation language DATR; Evans & Gazdar 1996), where all facts — including relations between form and morphosyntactic properties (morphological ‘rules’ in the usual sense) — are distributed across the individual nodes in the hierarchy. In a Network Morphology implementation, Zwicky’s rule in (5) would likely be captured in the form of the following declaration, situated at the ‘Nounal’ node in the hierarchy and thus inherited by default by all lower nodes, including all lexemes belonging to the Nounal category (the attribute label ‘mor’ simply stands for morphology): (i) <mor acc> == “<mor nom>” Here the value assigned to the attribute-path <mor acc>, and any extension of that path such as <mor acc sg neut>, is declared to be identical to the value assigned to the corresponding path


99

The very notion of referral entails a specific directionality (the realization of one form is referred to that of the other form, not vice versa), yet nothing has yet been said about directionality effects or how the directionality implicit in a syncretism can be discovered. As an example, consider the = syncretism in animate nouns in Russian, which is illustrated in Table 1 in the following section; for example, -ov marks both forms in many masculine nouns. We are able to determine that this suffix (and all other = suffix variants) in fact originates in the by examining the paradigms of inanimate nouns, where the is instead syncretic with , and the suffix in question is thus exclusively a marker. Hence it must be the whose realization is being referred to the , not the other way around. In many cases, evidence of this kind is hard to come by, and the choice of directionality appears arbitrary. In yet other cases, the identity relation in fact seems to be bidirectional. See Baerman (2004) for extensive discussion of directionality and related issues. The question of directionality will arise again in section 5.3 in connection with the North Saami = syncretism. It is important to note that in approaches which make use of rules of referral, these are not the only possible source of syncretism. Syncretism may also arise in situations where the relevant rules of exponence are general enough to cover both of the syncretized paradigm cells. Consider a language whose verbal morphology includes separate rules of exponence applying, respectively, in the contexts {:1,

:sg}, {

:sg} and {

:pl}, but no rules explicitly mentioning combinations like {:2,

:pl} or {:3,

:sg}. The result is the complete syncretism of 1st/2nd/3rd person in the plural (all being defined by the single rule referencing {

:pl}), and the complete syncretism of ' and $ forms (defined by the single rule for {

:sg}; this is overridden in the 1st person by the specific rule for {:1,

:sg}). Here syncretism derives from the relative lack of specificity of some of the rules of exponence, not from any referral. Stump (2001) refers to this phenomenon as unstipulated syncretism, to be contrasted with the stipulated syncretisms produced by rules of referral. Unstipulated syncretisms are a close parallel to the idea that syncretism arises through neutralization or underspecification (in the above example, one might speak of a single plural form, not specified for person). What distinguishes inferentialcontaining <nom> instead of . In our German example, the value of the full path <mor nom sg neut> at the lexeme node is junges, and the value of the path <mor acc sg neut> at that same node is, then, likewise junges by the above declaration.

100


realizational approaches like those of Zwicky (1985); Corbett & Fraser (1993) or Stump (2001) is their insistence that a considerable number of attested syncretisms are not of this type, but must instead be seen as stipulated outright; consequently, morphological theory must have the formal resources to encode such stipulations.6 2.4 Systematic versus accidental homonymy Any theory that incorporates a principled treatment of syncretism must have some means of distinguishing between inflectional homonymies that are genuinely systematic and ones that are better seen as purely accidental. Only the former would need to be accounted for in morphological terms; the latter would presumably either be treated as lexical exceptions or as genuine homophony (in the traditional sense) at the level of affixes or the rules which introduce them. For example, consider the fact that in (North) Saami, demonstratives like da-t “that, it” have identical and forms, both happening to end in -t. The suffix -t, unique to these demonstratives and a complete anomaly in the overall inflectional system, happens to be identical to the regular suffix -t. Should this identity relation be taken seriously as a (systematic) syncretism within North Saami inflectional morphology? Or should it be treated as a pure accident that the affix introduced by the special rule determining the of demonstratives is phonologically identical to that introduced by the regular rule? The question may be turned around, and stated more generally: when can an inflectional homonymy be plausibly interpreted as genuinely systematic rather than merely accidental? The clearest cases are where the identity pattern repeats itself across many distinct paradigms, and where the actual affixal material in6

In a recent paper, Arnold Zwicky (who coined the very term ‘rule of referral’; Zwicky 1985) advocates an alternative to referrals which is nonetheless stipulative in the same sense (Zwicky 2000; see Baerman 2004 for discussion). He suggests that the two paradigm cells are defined as a class (by a declaration like { ∪ } = X), and that this class is then referenced by an ordinary rule of exponence (such as X = stem + -ne). Formally speaking, this is close in spirit to approaches using neutralization or underspecification, though it is far less restrictive; any two feature combinations can in principle be conflated into a single class through this kind of set union. However, it is not quite equivalent to referrals in terms of formal power: Baerman (2004) demonstrates that Zwicky’s solution is incapable of capturing so-called bidirectional syncretisms (where the directionality of referral is different for different inflectional classes). With respect to the North Saami = case examined in section 5 below, the fact that the ending -in /-jn/ is arguably bimorphemic (plural /-j/ + locative /-n/) is also a problem for Zwicky’s proposal.


byk (m) “bull” byki bykov bykov

uˇcitel’ (m) “teacher” uˇcitelja uˇcitelej uˇcitelej

ženšˇcin-a (f) “woman” ženšˇciny ženšˇcin ženšˇcin

101

cˇ udovišˇc-e (n) “monster” cˇ udovišˇca cˇ udovišˇc cˇ udovišˇc

Table 1: Accusative/genitive syncretism in Russian animate nouns volved is different for each paradigm. Consider the inflection of animate nouns in Russian, whose is always identical to their as illustrated in Table 1. Here the = pattern holds across the various inflectional classes, even though the actual suffixes involved are different in each case (-ov, -ej, -Ø). This makes it highly implausible that the identity relation is accidental within each class. However, evidence of this sort is circumstantial, as it appeals primarily to the common sense of the morphologist: the facts simply suggest that the state of affairs is probably more than a mere accident. The criteria are not clear-cut; one might ask just how many distinct paradigm classes are required, what exactly constitutes different affixal material (for example, does zero marking really count?), and so forth. It should be mentioned at this point that syntactic evidence can sometimes be brought to bear on the issue of determining whether an inflectional homonymy pattern is systematic or merely accidental. In the phenomenon known as syncretic resolution, a form which is ambiguous between conflicting morphosyntactic properties is able to represent both of these at once (Zwicky & Sadock 1975; Taraldsen 1981; Dyła 1984; Zaenen & Karttunen 1984; Pullum & Zwicky 1986; McCreight & Chvany 1991; Zwicky 1991). Consider the German sentences in (8), where the pronominal subject of the headless relative clause is simultaneously required to carry nominative case (as subject of ärgern “annoy”) and accusative case (the relative clause being the object of zerstören “destroy”); the examples are from Taraldsen (1981). (8)

a.

mich ärgert]. Ich zerstöre [was I destroy what-{ , } me annoys “I destroy what annoys me.”

b.

Ich zerstöre [*wer/*wen mich ärgert]. I destroy who- me annoys “?I destroy who(m) annoys me.”

Because of the systematic

= syncretism in German neuters, the

102


form was “what” is used in accusative as well as nominative syntactic contexts. This is apparently enough to resolve the case-marking conflict inherent in the construction in (8), making (8a) acceptable. In (8b), on the other hand, where a syncretic form is not available, the sentence is entirely unacceptable with either accusative wen or nominative wer. It is generally assumed that the ability of a form to appear in syncretic resolution contexts is a sign that the homonymy in question is systematic rather than accidental. This would suggest that resolution can be used as a diagnostic to determine whether an identity pattern is a genuine syncretism (and thus part of the grammar) or not. Things are unfortunately not so simple. As noted by Zwicky (1991; see also Pullum & Zwicky 1986), the availability of syncretic resolution crucially depends on whether the conflicting feature values are ‘free’ (for example, tense in a main-clause verb) or ‘imposed’ by the syntactic context (for example by subcategorization). In the latter case, resolution appears to be possible even when the identity pattern is quite clearly an idiosyncrasy of individual lexical items: a truly accidental homonymy with no systemic status in the inflectional morphology. The complications noted by Zwicky (1991) make syncretic resolution far less feasible than one might think as a diagnostic for genuine syncretism qua systematic inflectional homonymy. It will frequently be extremely difficult, or even impossible, to devise syntactic contexts where ‘free’ values of the feature(s) in question are in conflict. Alternative sources of evidence for the systematicity of inflectional homonymies would thus be desirable. One such alternative, drawing on historical changes in inflectional systems, is introduced in the following section. 3. Diachronic change and productive syncretism 3.1 Productive syncretism as analogical extension A great many syncretism patterns appear to have their origin outside of morphology, in sound changes that are ‘destructive’ in the sense that they happen to collapse inflectional forms that were previously distinct. For example, modern Finnish uses the same ending, -n, to mark the accusative and genitive singular of most nominals, even though these reflect nasals with distinct place of articulation at earlier stages of the language (Proto-Uralic *-m versus *-n). Here the syncretism has been caused by a phonological change whereby non-coronal consonants became eliminated from word-final position. Although the ultimate origin of syncretism is often non-morphological in this sense, it may subsequently take on a life of its own. Frequently an identity


103

pattern of this type is extended, by analogy, to environments where it was never motivated — for example, to words belonging to a different inflectional class. In order for such extension to be possible, speakers-learners must first have noticed the identity pattern and assigned it some kind of systematic status within their grammar. This opens up the possibility that syncretisms (as genuinely systematic inflectional homonymies) could be identified post hoc, relying on comparativehistorical evidence. I will henceforth refer to this kind of analogical extension as productive syncretism: a particular identity pattern acting as a productive model which is extended to new parts of the grammar and/or lexicon. Consider the historical development of the accusative and genitive in Saami. Most Saami languages share with Finnish the = syncretism mentioned above, due to a sound change by which word-final nasals were lost. This collapse of and is shared by North Saami and all the Eastern Saami languages (Inari, Skolt, Akkala, Kildin and Ter Saami): North Saami guoli “fish ( )” < Proto-Saami (PS) *kuol¯e-m versus *kuol¯e-n; cf. Finnish kala-n.7 By contrast, in the southwestern part of the Saami language area (South, Ume, Pite and Lule Saami) the versus distinction is upheld: South Saami gueliem, guelien. In the plural, North Saami also conflates accusative and genitive (guliid “fish. /”), but this is a later development. The two were originally distinct, going back to Proto-Fenno-Saami (PFS) *-j, the bare plural marker (originally a ‘connective’, reanalyzed to form the basis of the entire oblique plural paradigm), and going back to PFS *-j-ta/tä (plural marker + partitive ending). Where modern North Saami has / guliid, PS thus had, on the one hand, *kuolij e¯ (< PFS *kala-j-ta) and, on the other hand, *kuolij (< PFS *kala-j). Evidently the form which was originally confined to the (now guliid in North Saami), has ousted the historical form. However, this appears to be a relatively recent innovation. In the orthography devised by Konrad Nielsen (Nielsen 1979[1932–1962]), the two were still kept distinct as g¯uliid versus g¯ulii. It does not appear likely that the collapse of accusative and genitive in the plural paradigm is the consequence of any sound change, as its counterpart in the singular clearly was. Instead, as pointed out by Sammallahti (1998:67), the development in the plural is, at least in part, based on analogy with the syncretism pattern found in the singular. 7

Here and throughout this article, the transcription used for rendering Proto-Saami and ProtoFenno-Saami reconstructed forms is that of Sammallahti (1998). When labelling such forms, Proto-Saami will henceforth be abbreviated PS, and Proto-Fenno-Saami (also known as ‘Early Pre-Finnic’, a direct translation of Finnish varhaiskantasuomi) will be abbreviated PFS.

104


PS

*-(s˘a)n *-sn¯e *-st¯e

< <
*-sn > *-sn > *-st (Wiklund 1902:57; Ravila 1960:32; Korhonen ˚ 1981:224; Sammallahti 1998:66). lowing the loss of the intervocalic *s. Thus guollá- < late PS *kuol;¯an < early PS *kuol;˘a-s˘an < PFS *kala-sin (cf. guolli < PS *kuol;¯e < PFS *kala). An alternative reconstruction (see Korhonen 1981:217-18) posits two distinct illative endings for Proto-Saami, *-s˘an in nonpersonal pronouns (North Saami ma-sa “what ”, cf. Finnish mi-hin < PFS *mi-sin) but *-j˘an otherwise. If so, then the contraction would be due to loss of an intervocalic j rather than s; but even on this account, the /-j/ ending of modern North Saami needs to be interpreted as a secondary phenomenon, not as a direct descendant of the *j of the *-j˘an ending itself. 9 In North Saami, the ending was rendered as -st in the Nielsen orthography (Nielsen 1979[1932–1962]), but the t is typically not pronounced; in the modern orthography the ending is spelled -s, reflecting current pronunciation. However, the /t/ reappears whenever a form is followed by a possessive suffix (goadi-st-an “in/from my hut”), and it may therefore be ¯ reasonable to assume that the ending is underlyingly /-st/.

106


Lule Saami Torne North Finnmark North Saami Saami goade-n ( ) goadi-n ¯ ¯ ( ) goadi-s(t) goade-s

Table 3: Locative singular in Torne (North) Saami dialects A likelier hypothesis is that *sn > st occurred as a more general sound change, not confined to word-final position. This is suggested by North Saami sisti “tanned reindeer hide” < PS *sis;n¯e, a borrowing from (early) Proto-Fennic *šišna (cf. Finnish hihna “strap, [leather] band”). The Proto-Fennic word in question goes back to PFS *šišna, ultimately itself a loanword from (Proto-) Baltic. Cognates of this word in South and Lule Saami appear to be direct reflexes of that original word, rather than mediated by Proto-Fennic through borrowing: PFS *šišna > PS *s˘as;n¯e > Lule Saami sasne “thin hide, shammy”, South Saami sesnie “hide, from which the hair has been removed by soaking” (Korhonen 1981:179). Another cognate pair supporting a general *sn > st sound change in North Saami is South Saami gesnedh versus North Saami gastit “sneeze”.10 The correspondence between North Saami VstV and South Saami VsnV provides a striking parallel to the one holding between North Saami locative -st and South Saami inessive -snV. In any case, it seems clear that the syncretism in the inessive/elative singular was ultimately caused by purely phonological factors rather than morphological, syntactic or semantic ones. The transitional dialects located on the border between North Saami and Lule Saami present an interesting compromise between the two systems. In Torne Saami, the southernmost dialect group within North Saami, the locative singular ending is -n rather than -s(t), and the same is also true of the northernmost dialects of Lule Saami (Korhonen 1981:224; Sammallahti 1998:47). The correspondences are shown in Table 3. The -n used to mark the syncretic locative in the Torne dialects cannot be the result of any phonological merger of earlier -st(V) and -sn(V), but must be a reflex of inessive *-sn¯e alone, corresponding to Lule Saami -n. It appears that 10

I am grateful to one of the anonymous reviewers for pointing out this pair. The same reviewer also notes that there is reason to believe that the same denasalization sound change affected all /sN/ clusters in North Saami, not just *sn: cf. South Saami gasngese (with /sN/) versus North Saami gaskas “juniper” and Inari Saami lusme versus North Saami luspi “outlet of lake”, though the *sm > sp change appears somewhat erratic (witness South Saami luspi “id.”, as well as the North Saami back-formation bisma alongside bispa “bishop”).


107

the syncretism pattern as such — the identity relation between forms expressing “in X” and “from X” — has been extended from the Finnmark Saami dialects (where it was the fortuitous result of sound change) to neighbouring dialects which at the time had an inessive -n versus elative -s opposition. As a result, Torne Saami and Finnmark Saami share the same syncretism pattern, and hence neither dialect maintains a case distinction which is absent in the other dialect. The syncretic locative case is simply signalled by different segmental material in the two dialects. At the level of morphological structure, the local case system is thus identical across North Saami as a whole. Returning to Finnmark North Saami and the Eastern languages, the conditions that resulted in the phonological merger of inessive and elative were only relevant in the singular paradigm. However, the syncretism has also been extended by analogy to other parts of the grammar and lexicon, not just the plural paradigm but also adverbs and postpositions with spatial semantics. (Sammallahti 1998:66 includes the possessive declension in this list, cf. goadi-st-an ¯ “in/from my hut”, but if *sn > st was a regular phonological development, as suggested above, the inessive-elative syncretism need not be attributed to analogical extension in such forms.) This extension is a clear indication that the inessive=elative identity relation had attained systemic status within the grammar. In other words, the development constitutes a case of productive syncretism. The situation in the plural paradigm is complicated somewhat by the fact that Proto-Fenno-Saami did not have any specific local cases (illative, inessive or elative) in the plural paradigm, only in the singular; the plural counterparts seem to have developed sometime during the Proto-Saami period (Sammallahti 1998:67-69). What Proto-Saami did inherit from Proto-Fenno-Saami was a general partitive/ablative plural, PFS *-j-ta/tä > PS *-j e¯ (surviving as the in most modern Saami languages), and a general essive/locative plural, PFS *-jna/nä > PS *-j-n¯e. The latter became the plural counterpart of inessive singular *-sn¯e. For the elative plural, a new ending *-j-st¯e was created by combining the plural marker and the ending. In South, Ume, Pite and Lule Saami, where the inessive versus elative distinction is upheld in the singular, the plural paradigm reflects this (late) Proto-Saami situation quite closely, as the partial paradigm of South Saami gåetie “hut” in Table 4 illustrates. In North Saami and the whole Eastern branch — that is, in precisely those languages where inessive and elative were merged through sound change in the singular — the inessive versus elative distinction is absent from the plural as well. Here it is clear that the inessive plural form has taken over the function of

108


singular

plural

gåetesne gåetine (< *-j-n¯e) gåeteste gåetijste (< *- j-st¯e) Table 4: Inessive and elative marking in South Saami the elative plural: North Saami godiin ¯ “in/from the huts” (< PS *koa ijn¯e “in the huts”), gielain “in the snares/out of the snares”. The syncretism found in the singular has thus been extended to the plural paradigm. Interestingly, the transitional dialects on the border between Lule Saami and North Saami again present a slightly different picture.11 In the southernmost varieties of the Torne dialect of North Saami (as spoken in the Kaitum and Jukkasjärvi areas), the ending is -is and not -in (Sammallahti 1998:47; according to Korhonen 1981:223-24, the same is also true of the northernmost dialects of Lule Saami); see Table 5. Lule Saami

goadin goadis

Torne North Saami North Saami (Kaitum/Jukkasjärvi) (other dialects) godiis ¯

godiin ¯

Table 5: Locative plural in transitional Torne (north) Saami dialects Here it is the that has ousted the , rather than vice versa. What makes this even more curious is the fact that in precisely these dialects (as well as other varieties of Torne North Saami) the marker is -n, a reflex specifically of the inessive, as was shown in Table 3 above. Again, it appears that the syncretism pattern as such has been modelled after the neighbouring North Saami dialects, but that the segmental material used to mark the syncretic locative form is different from that used in those dialects. Once the syncretism pattern has been extended from singular to plural in this manner, there are no longer any grounds for positing inessive and elative as separate morphosyntactic categories within the inflectional system. Instead of the 11

It should be noted that, despite the fact that the transitional dialects diverge in various ways, all varieties nevertheless appear to show inessive/elative syncretism either in singular and plural alike or not at all. In other words, there are to the best of my knowledge no dialects with syncretism only in the singular but not in the plural (or vice versa, for that matter). The dialectal variations mentioned here and in Table 3 concern the source of segmental material (the suffix) used for realizing the syncretic inessive/elative form.


109

a. Spatial postpositions: siste alde bealde luhtte duohkin

“in(side) X; out of X” “on (top of) X; down from X” “at X”s side; from X”s side” “at X, with X; from X” “behind X; from behind X”

sisa ala beallái lusa duohkái

“into X” “onto X” “to X”s side” “to X” “(to) behind X”

olggos bajás nuorttas guhkás

“out” “up(wards)” “to the east” “far (away)”

b. Spatial adverbs: olgun badjin nuortan guhkkin

“outside; from outside” “above; from above” “in the east; from the east” “in the distance; from afar”

Table 6: Spatial adverbs and postpositions in North Saami original system of three local cases, all varieties of North Saami and the Eastern Saami languages now have a two-way system opposing ‘illative’ (really a general lative, indicating orientation towards something or someone) with ‘locative’ (indicating location in/at X, or motion out of/away from X). This two-way system contrasting a lative and a (semantically dual-purpose) ‘locative’ has also been carried over into the domain of spatial adverbs and postpositions. As a result, a single form with both Location and Source functions now stands in opposition to another signifying Direction; examples from North Saami are shown in Table 6. The dual-purpose ‘locative’ forms typically end either in -de/-te or in -n. Most spatial adverbs and postpositions developed historically out of what were originally case forms of locational nouns. Those words that end in -n generally go back to forms with the PFS essive (general locative) suffix *-na/nä, and may safely be assumed to have had a purely locative (inessive/adessive) function. The ones ending in -de or -te go back to forms containing PFS partitive/ablative *-ta/tä; these presumably had a purely elative/ablative function, with no locative meaning (in the narrower sense). For example, North Saami olggun “outside; from outside” can be compared with the Lule Saami pair ålggon “outside” versus ålggot “from outside” (Sammallahti 1998:66). The short list in Table 6 reveals that in North Saami, the surviving form expressing the syncretic ‘locative’ function is sometimes the original ablative form and sometimes the original locative form; there seems to be very little regularity, except perhaps that adverbs appear to favour -n. Note that the suffix-like

110


elements found in spatial adverbs and postpositions, though they are reflexes of Proto-Fenno-Saami case endings, were not directly identifiable with the inessive or elative case endings involved in the inflectional syncretism patterns described earlier. The basis for extending the syncretism to these lexical items appears to be the semantic-functional relationship between spatial adverbs and postpositions on the one hand and local case forms of nouns on the other. 4. Constraints on syncretism patterns 4.1 Is any syncretism pattern possible? In theoretical frameworks that use referrals to capture syncretism (see section 2.3), there are few, if any, inherent limitations on which two paradigm slots can be linked. The formalism in which rules of referral (or their equivalent) are expressed is sufficiently powerful that it can easily connect the realization of one set of morphosyntactic features to that of any other set. To pick a random example, a syncretism in verbs between the first person singular present subjunctive form and the third person plural past indicative form could, in Paradigm Function Morphology, be captured by a rule like (11). (11)

RR{ :prs, :sub, :1, :sg},V (<X, σ>)=def , where PF(<X, σ/{ :past, :ind, :3,

:pl}>) =

The ability of the formalism to express syncretism between radically different pairs of paradigm cells, which share virtually no morphosyntactic properties at all, might easily be seen as a weakness of the theories in question. Since they appear to place no limits on what constitutes a possible syncretism pattern, are they not excessively powerful?12 12

One reviewer notes that it is still possible in principle, and probably necessary in practice, to formulate restrictions on which paradigm cells can be linked through referral. The reviewer points out that in (11), the two forms involve “syncretized combinations of values which belong to the same features”, in that each is defined in terms of values for , , and

, as opposed to a hypothetical (and probably impossible) syncretism linking accusative singular and 1st person subjunctive. The general point is certainly valid: referrals are a very powerful construct, and nothing rules out the existence of (as yet undiscovered or unstated) substantive universal restrictions on their form or extent in natural languages. On the other hand, the specific point about values of the same versus distinct features is, I believe, somewhat misguided. The reason why all four features are mentioned for each of the two forms in (11) is simply that each constitutes an individual (unique) paradigm cell, the definition of which thus requires a full set of feature specifications. It would have been equally possible to syncretize, say, the set of all present subjunctive singular forms with the set of all past indicative 3rd person


111

At the opposite end on the scale of empirical restrictiveness are theories that account for syncretism by way of neutralization or underspecification (see section 2.2). Consider again the hypothetical & =$ example. Let us assume, for the sake of the argument, that the first versus third person opposition is to be represented featurally as [+speaker, -hearer] versus [-speaker, -hearer], and that the other oppositions are a matter of the binary features [±subjunctive], [±plural] and [±past]. The two inflectional cells would then, given full specification, be represented as in (12). (12)

a.

[-past, +sub, -pl, +speaker, -hearer] (= & )

b.

[+past, -sub, +pl, -speaker, -hearer] (= $

)

An underspecified representation covering both would contain no information beyond [-hearer], or ‘non-2nd person’. Consequently, the neutralization of the oppositions distinguishing (12a) from (12b) would encompass all other first and third person forms as well, throughout the entire paradigm, be they singular or plural, past or present, indicative or subjunctive, and so forth. In sum, ‘disjoint’ pairings of the sort in (12) cannot be related through neutralization, and a strong prediction is made: syncretism patterns relating such pairings cannot exist in any inflectional system.13 Ultimately the choice between alternative theories must be based not so much on their inherent restrictiveness as on available empirical evidence. Greater restrictiveness is preferable only insofar as predictions are upheld by the known facts. If syncretisms like the hypothetical & = $ do in fact occur, theories unable to express them must be discarded on the grounds that forms. That referral would mention tense, mood and number on one side but tense, mood and person on the other. Secondly, syncretism is by definition a relation between forms within the same inflectional paradigm. To the extent that no single lexeme has a paradigm containing two inflected forms describable as ‘accusative singular’ and ‘1st person subjunctive’, a syncretism linking such a pair of forms will of course be inconceivable. One could however imagine a syncretism between a non-finite verb form (with no values for tense, person or number) and a finite form like (with specific values for those features), and this would be a close analogue of the = scenario. Though I am unaware of any cases of this type, they might well be possible. 13 There is in fact a solution available within the underspecification/neutralization approach, albeit an unsavoury one. The affix which marks the paradigm cells in (12a-b), but no others, could be treated as a generic default, specified only as non-2nd person ([-hearer]), or even left entirely unspecified for all features. The failure of that affix to appear in 1st and 3rd person forms other than those in (12) would then be ascribed to override (blocking) by other affixes realizing more specific morphosyntactic representations.

112


they are too restrictive. In practice, resolving this issue by empirical observation is not straightforward, as one must take care to demonstrate that the crucial data constitute a genuine syncretism rather than an accidental homonymy (see section 2.4). That difficult empirical issue is the focus of the remainder of this paper. It is ultimately argued (in section 5) that North Saami provides unequivocal evidence of precisely the crucial kind of ‘disjoint’ syncretism. However, we must first consider a third alternative, less restrictive than theories based on neutralization or underspecification but considerably less powerful than rule-of-referral theories, as the Saami evidence bears directly on the empirical predictions of this theory as well. 4.2 Geometric aspects of syncretism McCreight & Chvany (1991) propose a theory of syncretism which formalizes the traditional notion that an inflectional paradigm is a geometrically arranged group of related forms. The features distinguishing individual forms constitute the dimensions of the geometry, and combinations like or & thus exist as locations in the space defined by these dimensions. A paradigm is then simply “the partitioning of grammatical space by a set of related forms” (McCreight & Chvany 1991:94). In this model, which continues a tradition initiated by Roman Jakobson in his famous paper on Russian case inflection (Jakobson 1936; cf. Neidle 1988), syncretism is seen as the conflation of two or more paradigm cells which are contiguous: adjacent along some dimension (see also Plank 1991b; Johnston 1997). This has typological implications: two forms which by themselves constitute a geometrically non-contiguous pair can be syncretized, but crucially only by way of a chain of adjacency relations. For example, ' and & may be syncretized as long as & (or ') is included as well (it is adjacent to one form in the person dimension and the other in the number dimension). It is helpful to consider the various logical possibilities of identity patterns in a simple 2 × 2 paradigm, modelled in (13) with two binary features, [±F] and [±G]. Neutralization approaches predict that (13a-c) will be possible syncretism patterns, but not (13d-e). The former three can be characterized as [+F, 0G], [0F, +G] and [0F, 0G], respectively, while the last two would require devices such as disjunction or set union.


(13)

+F −F

a.

b.

+G −G

−F

+F

−F

+G −G

+F −F

c.

+F

113

+G −G

d. +G −G

+F −F

e. +G −G

The geometric conflation model advocated by McCreight & Chvany (1991), on the other hand, allows syncretism to affect an irregularly shaped (non-‘rectangular’) area of the paradigm, as long as it is contiguous as in (13d). The authors clearly hold non-contiguous syncretisms like (13e) to be impossible in principle, separating “conflatable syncretisms” from “identical endings that are not contiguous” (citing in this context the distinction between systematic and accidental homonymy as drawn by Zwicky 1991). McCreight & Chvany (1991) adduce evidence from syncretic resolution (see section 2.4 above) in support of their proposal, arguing that geometric contiguity is a necessary condition (though not a sufficient one; see their footnote 7) for syncretic resolution. However, most of their evidence is highly problematic. For example, in their interpretation of Finnish examples cited from Zaenen & Karttunen (1984), McCreight & Chvany fail to recognize the existence of a separate accusative case in Finnish (thus sinu-t contrasts with sinu-n for the ' pronoun sinä “you”). A possessed form like kirjansa “his/her book” thus covers not only , and , but also crucially and , and this completely undermines their analysis of the syntactic distribution of such forms. Secondly, they do not control for the distinction between ‘free’ and syntactically ‘imposed’ feature values (see section 2.4). In fact, none of the negative evidence cited by McCreight & Chvany (that is, ungrammaticality of sentences with allegedly impossible syncretic resolution patterns) is expected on Zwicky’s (1991) account. If nothing else, all of this highlights the difficulties inherent in using syncretic resolution as a diagnostic for separating genuine syncretism from accidental homonymy.

114


Problems aside, one of the virtues of the geometric conflation model is that it makes an important empirical prediction about the typology of syncretism patterns. Non-contiguous identity patterns within a paradigm must in every instance be a matter of accidental rather than systematic inflectional homonymy, and hence do not count as syncretisms in any meaningful sense. Unfortunately, there is reason to believe that this prediction is too strong. Some well-known counterexamples are considered in the next section, but are shown to be inconclusive. In section 5 evidence from North Saami dialects is shown to provide a far stronger case against the geometric conflation approach, as well as against the even more restrictive neutralization/underspecification approach. 4.3 Apparent challenges to the geometric model On a purely observational level, there is no shortage of apparent counterexamples to the geometrical contiguity claim in the world’s languages. There is a fair number of cases where it seems, at least superficially, that an identity relation links two inflectional forms that do not form a contiguous region under any conceivable geometric construal of the paradigm in question. These, then, would seem to be instances of syncretism patterns with the structure in (13e) above, precisely the type that is predicted not to be possible in the model advocated by McCreight & Chvany (1991; cf. Johnston 1997). A striking case is agreement marking in Hua, a Papuan language of New Guinea (Haiman 1980, 1998; Stump 2001). Here the ' and & of verbs carry the same suffix, while all other person-number combinations have a different suffix. The distribution is illustrated in Table 7; the dual marker -P- (preceding the suffix) is omitted for simplicity.14

1 2 3

-e -ne -e

-e -e -e

-ne -e -e

Table 7: Agreement markers (indicative mood) in Hua 14

Hua is different from other cases considered in this paper, in that syncretism does not result in complete identity of word-forms, due to alternations within the stem. For example, the stem of a (monosyllabic) verb typically contains a rounded vowel (u or o) in 1. person forms, but an unrounded vowel (i or a) in 2. person forms. The and forms are thus not homonymous as such: they merely contain identical endings. This has implications for the formal expression of rules of referral (see the discussion of ‘whole-word syncretism’ versus ‘block syncretism’ in Stump 2001:217ff), but these complications need not concern us here.


Mood Indicative Interrogative Relative Purposive Concessive Inconsequential Medial (a) Medial (b) Exclamatory (a) Exclamatory (b) Counterfactual protasis Counterfactual apodosis

115

'/& default -ne -pe -paP -piP -pa -pana -na -pa -pane -pae -sipana -sine

-e -ve -maP -miP -va -mana -ga -ma -mane -mae -hipana -hine

Table 8: Agreement markers across mood paradigms in Hua Each mood uses different agreement markers, but always with the same general pattern, juxtaposing a '/& variant with a default one. The pairs are summarized in Table 8. In every mood paradigm, Hua affixal morphology thus enforces an identity linking two blatantly non-contiguous paradigm cells, ' and &. There would seem to be no way of relating these geometrically without incorrectly implicating & or ' as well (leaving aside the dual sub-paradigm). Another intriguing case, also involving agreement morphology, is the NorthEast Caucasian language Khinalug. In this language, which has four distinct genders (labelled I-IV), the affixes used to mark gender-number agreement fall in three separate sets (Corbett 1991:119-123); these are shown schematically in Table 9. The members of the three sets occur in different slots in a verb’s morphological structure — for example, a single verb form will often simultaneously include Set 1 and Set 2 markers of a given gender-number combination. The Set 3 markers are restricted to a small set of irregular verbs. Each of the three paradigms (the ‘sets’) has only three distinct exponents of the eight possible gender-number combinations. One of these is non-syncretic, uniquely marking II., whereas the other two are syncretic, marking I./II. /III. on the one hand and I./III./IV./IV. on the other. As in Hua, the same distribution of identity relations (here considerably more intricate) repeats itself over three separate paradigms. One of the two syncretisms in Table 9 is problematic for the geometric conflation model: the I.=II.=III.

116


Set 1

I II III IV

Ø- -b-z- -b-b- -Ø-Ø- -Ø-

Set 2

Set 3

-j- -v-z- -v-v- -j-j- -j-

-h- -f-s- -f-f- -h-h- -h-

Table 9: Gender-number agreement marking in Khinalug identity. It is impossible to rearrange the paradigms geometrically (even in a 3dimensional diagram) such that the III. cell can be linked contiguously with either I. or II. without erroneously implicating some additional paradigm cell. By contrast, the other syncretism pattern (I.=III.=IV.=IV.) can straightforwardly be construed as contiguous; this is easily visualized by shifting the row for gender I to the bottom of Table 9 instead of the top. Unfortunately, such apparent cases of geometrically non-contiguous syncretism are not always sufficiently convincing. One fundamental problem concerns the exponents (affixes, in these cases) which mark the syncretic forms. The strength of the Hua and Khinalug cases derives primarily from the fact that the identity pattern repeats itself across several distinct paradigms (see also the Russian example in Table 1 above). The syncretic 2./1. in Hua is marked with -ne, -pe, -pana, and so forth, depending on the mood of the verb, and the syncretic I./II./III. in Khinalug is marked with -b-, -v- or -f- depending on the identity of the verb and its internal structure. But how distinct are these paradigms, really? In Khinalug, the -b-, -v- and -f- which mark I./II./III. are conspicously all labial obstruents, just as the -z- and -s- marking II. are alveolar sibilants; the series -Ø-, -j-, -h- which marks I./III./IV./IV. might also be connected phonologically without any great stretch of the imagination. This raises the legitimate question whether the three ‘sets’ might not be reduced to a single paradigm, given a sufficiently elaborate analysis of Khinalug (morpho)phonology. (As noted by Corbett 1991:120, the picture in Table 9 already masks a great range of predictable variation; for example, the set 1 marker shown here as -b- is realized as -b-, -ph -, -bi- or even -Ø-, depending on context.) Such a reduction would, in turn, make it harder to argue that the disjoint I./II. = III. identity is anything more than an accidental homonymy. Similar reservations could be voiced about Hua. Within each mood paradigm, the syncretic 2./1. suffix and the default suffix are always eerily similar to each other, and consistently differ only in their initial segment. A p-initial


117

2./1. suffix always corresponds to an m- or v-initial default suffix; all three are labials. Another regular correspondence is s versus h: needless to say, the latter is cross-linguistically often a lenited counterpart of the former. A third one is n versus either g (before a back vowel) or Ø (before a front vowel). It is not at all difficult to envisage an analysis which does not posit separate affix shapes for the syncretic 2./1. category, but instead posits, for each of the twelve moods, a single underlying affix for all person-number combinations. On such an analysis, what the 2. and 1. have in common — and what sets them apart from other cells — is then some relatively abstract phonological property (for example a floating feature, or perhaps a prespecified prosodic-constituent edge), which triggers a slightly different surface realization of the mood affix. This makes an accidental homonymy analysis of the 2.=1. pattern far less awkward than it would appear at first glance. A second major problem with such cases as these is that they involve static distributional patterns. It is always possible in principle — especially if one is willing to play devil’s advocate — to write such cases off as accidental homonymy rather than genuine syncretism. (Note that ‘accidental’ does not imply that the pattern came about by pure chance, but merely that it has no status as an element of organization in the synchronic grammar internalized by speakers of the language.) This is a persistent problem, as syncretism is by definition manifested in distributional terms. By contrast, the diachronic processes discussed in section 3 under the rubric of ‘productive syncretism’ are, as evidence, not fraught with the same problems. These are dynamic processes (in the real-time dimension of diachronic change), and the syncretism patterns involved therefore cannot be explained away as mere observational curiosities. In the next section I argue that a case of precisely this kind is found in certain dialects of North Saami, where a geometrically non-contiguous syncretism pattern has been extended by analogy from one inflectional class to another. 5. Non-geometric productive syncretism in North Saami 5.1 Syncretism of comitative singular and locative plural In the inflectional paradigms of all North Saami polysyllabic nominals (that is, nouns, non-attributive adjectives and participles, and certain pronouns), the forms encoding comitative singular and locative plural are always identical, both ending in -in (/jn/).15 This is illustrated by the noun paradigms in (10), with 15 The same identity pattern is observed in most other Saami languages as well; the following discussion is concerned exclusively with North Saami and its dialects.

118


cˇ albmi cˇ almmi-t cˇ almmi cˇ almmi-i-d cˇ albmá-i cˇ almmi-i-de cˇ almmi-s cˇ almmi-i-n cˇ almmi-in cˇ almmi-i-guin cˇ albmi-n

viessu viesu vissu-i viesu-s viesu-in

nisu nissona nissoni-i nissoni-s nissoni-in

muitalus muitalusa-t muitalusa muitalusa-i-d muitalussi-i muitalusa-i-de muitalusa-s muitalusa-i-n muitalusa-in muitalusa-i-guin muitalus-in

nissona-t nissoni-i-d nissoni-i-dda nissoni-i-n nissoni-i-guin nisu-n

viesu-t viesu-i-d viesu-i-de viesu-i-n viesu-i-guin viessu-n

Table 10: Polysyllabic nominals in North Saami the relevant forms indicated in boldface (ˇcalbmi “eye”, viessu “house”, nisu “woman”, muitalus “story, tale”). In the forms, the string /jn/ is best interpreted synchronically as a sequence of two affixes, the (oblique) plural marker /-j-/ and a locative case ending /-n/ (this is certainly true historically, see discussion below and in section 3.2 above). The morpheme boundaries in Table 10 reflect this assumption. However, regardless of whether the ending is analyzed as internally complex or not, the ending can only be interpreted synchronically as a unitary suffix, /-jn/. The question which immediately arises is whether or not the = identity observed in North Saami nominals is a fact to be captured in the synchronic grammar of the language. In all relevant words, the same string is involved, /-jn/, and this makes it perfectly plausible to treat the homonymy as accidental rather than systematic. The fact that both and happen to be expressed by affixing the string /jn/ is then about as interesting as the fact that in English, plural in nouns and $ in verbs are both expressed by affixing /-z/. If this is a valid interpretation of the North Saami facts, they can have no implications whatsoever for the formal analysis of syncretism. On the face of it, there does not seem to be anything in the synchronic grammar of North Saami that could be used as a counterargument against this view. That is, there is little in the way of hard evidence that the =


da-t da-t da-n da-i-d da-sa da-i-dda da-s da-i-n da-inna da-i-guin da-nin

119

gii gea-t gea-n gea-i-d gea-sa gea-i-dda gea-s gea-i-n gea-inna gea-i-guin gea-nin

Table 11: Monosyllabic pronouns in (western) North Saami identity relation is in fact part of the North Saami morphological system. For one thing, there are no distinct inflectional classes, with separate sets of suffixes, across which the same abstract identity pattern repeats itself. The main class distinction that exists in Saami is that between imparisyllabic stems (ones with an odd number of syllables), such as nisu “woman”, and parisyllabic stems (with an even number of syllables), such as the other nouns in Table 10. But even there the sole divergence in terms of affixal material is found in the forms (-idda versus -ide). To make matters even worse, when we look at the paradigms of monosyllabic pronominal stems, the = identity pattern appears to break down. This is illustrated in Table 11 with the paradigms of dat “that, it” and gii “who”. Here the phonological material used to mark (-inna = /-jn:a/) is distinct from that used to mark locative plural (-in = /-jn/ just as in polysyllabic stems), even though the two are admittedly highly similar. The /-jn:a/ version of the ending is also found in the singular paradigm of personal pronouns (mu-inna “with me”).16 Here, however, the locative utilizes singular morphology (-s) throughout all three numbers ( mu-s, munnos, mi-s), and this precludes anything corresponding to the = syncretism of non-pronominal paradigms. In the dual paradigm, munnu-in shows the same /-jn/ allomorph as do (parisyllabic) nominals like viessu “house”, viesu-in. In the plural, the form is minguin, formed along the same lines as the of nouns and non-personal pronouns; the origin of that structure is a (genitive) postpositional phrase, with -guin related to the noun guoibmi “companion”. Phonologically, minguin is /mi:nkujn/ with a long root vowel; as suggested by one reviewer, this is likely a prosodic word minimality effect (cf. min /mi:n/, mis /mi:s/) and as such attests to the clitic status of /-kujn/. One might speculate that the original form was perhaps */mi-jn/ (cf. munnu-in /mun:o-jn/), and that the eventual homophony of this form with the form further facilitated the introduction of the + /-kujn/ construction by analogy with nouns and non-personal pronouns. 16

120


da-t da-t da-n da-i-d da-sa da-i-dda da-s da-i-nna da-inna da-i-guin da-nin

gii gea-t gea-n gea-i-d gea-sa gea-i-dda gea-s gea-i-nna gea-inna gea-i-guin gea-nin

Table 12: Monosyllabic pronouns in eastern Finnmark dialects The mismatch between the and forms of pronouns in Table 11 above would seem to be sufficient evidence against treating the identity pattern observed in polysyllabic nominals as a genuine syncretism rather than an accidental homonymy. However, this is not the whole story. The fact is that the paradigm in Table 11, though it is the one typically presented in reference works on North Saami (Bartens 1989; Nickel 1990; Sammallahti 1998), does not hold true throughout the entire North Saami dialect area. It is based specifically on the dialects and western Finnmark dialects, that is, those varieties of North Saami spoken in Alta, Kautokeino, Enontekiö, Sodankylä and further to the south and west of these areas.17 The eastern Finnmark dialects on the other hand — as spoken, for example, in Karasjok and Utsjoki — show a slightly different pattern, which is shown in Table 12. Note that works published in Finland by Finnish scholars do frequently show these eastern Finnmark forms, sometimes to the exclusion of the western ones (Itkonen 1960). Here we find the same inflectional homonymy pattern as in polysyllabic stems; the -inna ending, which in western dialects is confined to the , is here used in the as well. In eastern Finnmark North Saami, we thus find that two separate (albeit similar) affixal strings are each used to encode two apparently unrelated cells within the paradigm. The string /-jn/ marks = on polysyllabic stems, whereas /-jn:a/ does the same for monosylI have been unable to confirm whether the same pattern ( dainna, geainna versus dain, geain) also extends to the Sea North Saami dialects. As these are known to have 17

been conservative in many respects (for example the treatment of Proto-Saami geminate nasals and *št/šk clusters), such comparative evidence might shed light on the discussion below. As for Torne dialects, note that the -is found in their southernmost varieties (see Table 5) extends to monosyllabic pronouns as well; thus “in these mountains” is dá-is vári-is as opposed to dá-in vári-in in the other Torne dialects (Sammallahti 1998:47).


121

labic stems. In the western dialects, by contrast, only /-jn/ has this dual function (in polysyllabic stems, specifically) whereas the /-jn:a/ of monosyllabic stems is strictly a marker. In what follows, I argue in favour of the hypothesis (articulated originally by Korhonen 1981:222) that the eastern Finnmark pattern represents a secondary innovation. Specifically, the claim is that the eastern pattern in Table 12 developed out of the system in Table 11 through analogical extension of the = syncretism pattern found in polysyllabic stems. (It should be kept in mind that the polysyllabic class represents the vast majority of nominals in the language, whereas there are extremely few monosyllabic stems.) The very fact that this historical development occurred entails, in turn, that the identity relation between the two categories must have had systemic status within the grammar of North Saami at the time when the change occurred, and presumably still does. The following section examines in detail the comparative-historical background of the comitative singular and locative plural morphology in Saami, in support of the claim that the eastern Finnmark system is indeed an innovation. We then turn to the broader implications of the North Saami facts for the analysis of syncretism as such, bearing in particular on the theoretical alternatives discussed in section 2 and section 4. 5.2 The historical evidence The historical origin of the locative plural morphology is quite clear and unequivocal. As already explained in section 3.2 above, it is to be identified with the inessive plural in those languages which still retain an inessive versus elative case distinction: South, Ume, Pite and Lule Saami. North Saami -in /-jn/ is thus a direct reflex of PS *-j-n¯e, which in turn goes back to PFS *-j-na/nä, a combination of the (oblique) plural marker *-j and the essive/locative case suffix *-na/nä. Unlike the elative plural, which is demonstrably an innovation that developed after Proto-Saami split off from Proto-Fennic (see Sammallahti 1998:68-69), the inessive plural in PS *-j-n¯e goes back to a form existing already in Proto-Fenno-Saami (though its semantics may have been somewhat different). In terms of its phonotactic makeup and subsequent development through Proto-Saami and into the individual daughter languages, the inessive/locative plural closely parallels another form which also goes back to the Proto-FennoSaami period. This is the plural partitive/ablative in PFS *-j-ta/tä, which survives as the accusative-genitive plural in the modern Saami languages. The parallel is shown in Table 13 with the relevant forms of the noun “hut” (North Saami goahti, cf. Finnish kota).

122


PFS

PS > *koa -i-j e¯ > > *koa i-j-n¯e >

/ *kota-j-ta / *kota-j-na Table 13: Origins of

and endings in North Saami

expected dat(?) dan dasa **dastti

**dainni

North Saami godiid ¯ ( /) godiin ¯ ( )

(or dainna) **dani

actual dat dan dasa das

expected dat **daiddi daidda(?) **dainni

dainna

daiguin

actual dat daid daidda dain ∼ dainna daiguin

danin

Table 14: ‘Expected’ vs. observed forms for pronouns in North Saami Turning now to the monosyllabic pronouns, such as dat “it, that”, whose stem appears to go back to PFS *ti- (cf. Finnish se- ∼ si-), it is not quite clear exactly what the expected phonological development ought to be. If a morphologically complex form like / PFS *ti-j-nä (the ancestor of North Saami dain) were to be treated exactly like a morphologically simplex word, the regular outcome ought to have been PS *t˘aj;n¯e > North Saami **daidni, or perhaps **dainni with weak grade generalized due to relative lack of stress, the pronoun being a function word. (Note that from this point on, a double asterisk will be used to indicate incorrect/unattested forms, while a single asterisk marks reconstructed forms at earlier diachronic stages.) The form PFS *ti-j-tä (the ancestor of North Saami daid) should by the same token have come out as **daidi or **daiddi. For expository purposes, and as a tentative thought experiment, Table 14 presents a full paradigm where such ‘expected’ forms are juxtaposed with the ones actually observed (in western Finnmark dialects). The additional complication of grade alternations has been factored out by assuming a generalized weak grade throughout the paradigm. Those paradigm slots where the attested form is (or may be) different from what is expected are indicated in boldface. The essive has simply been augmented by addition of -(i)n by analogy with polysyllabic nouns (cf. guolli “fish”, ess. guolli-n; eallin “life”, ess. eallim-in).


123

The expected outcome in the is somewhat unclear (either unattested **dainni or attested dainna) because the phonological shape of the suffix in Proto-Fenno-Saami is not quite certain; this issue will be taken up again below. This leaves three paradigm slots where the attested form deviates from the one that the regular sound changes in the history of North Saami ‘ought to’ have given rise to: locative singular, accusative/genitive plural, and locative plural. In each of these three categories, we find a monosyllabic form (das, daid, dain) instead of a disyllabic one (**dastti, **daiddi, **dainni), where a wordfinal i (< PS *¯e < PFS *a/ä) has been apocopated. This may have happened either as a phonological process (perhaps again due to lack of stress, although this seems to be contradicted by disyllabic forms such as and ) or through analogy with polysyllabic stems, where the corresponding endings are vowelless ( -s, -id, -in). There are three things in particular that are worth noting in this context. Firstly, in all those inflectional forms that are securely reconstructed as having contained word-final PFS *a/ä (> PS *¯e), this word-final vowel is absent from the modern North Saami reflexes.18 Secondly, whatever the regular outcome of the (in)essive/locative plural PFS *ti-j-nä > PS *t˘aj;n¯e may be expected to be, it is certainly not dainna as attested in the eastern Finnmark dialects; were a disyllabic form to be expected, the second-syllable vowel should certainly be i (= /i:/) and not a. Thirdly, even if dainna were somehow the result of regular changes, this would leave the monosyllabicity of das and / daid unaccounted for, since all three are reconstructed with roughly the same phonological structure in PFS (*CiCCä). The hypothesis that dialectal dainna is a secondary development — replacing earlier dain by analogical extension of the = syncretism — circumvents this problem, since all three forms (das, daid, dain) then reflect the exact same historical development. Let us turn now to the comitative singular, expressed by -in on polysyllabic stems and -inna on monosyllabic ones throughout all the dialects. Here the comparative evidence from the individual Saami and Fennic languages is somewhat contradictory. There are two alternative reconstructions for this suffix in PFS, either *-jna/jnä or *-jni (Korhonen 1981; Häkkinen 1985; Sammallahti 1998). The former of these reconstructions is typically interpreted as the combination of a 18 The essive (going back ultimately to PFS *ti-nä) is not necessarily an exception to this generalization; the modern form danin may either go back to an earlier *dani (the expected form), augmented with -n, or else to monosyllabic *dan (cf. das), augmented with -in.

124


derivational suffix *-j (forming possessive adjectives) and the essive case marker *-na/nä (Korhonen 1981:225) The comitative singular would then have been completely identical in its (morpho)phonological structure to the essive/locative plural, both being marked with bimorphemic PFS *-j-na/nä. The second alternative reconstruction, PFS *-jni, does not have any adequate interpretation, except perhaps as a secondary (and irregular) development from earlier *-j-na/nä; that is, with *-ni as a variant of essive *-na/nä, which existed already in Proto-FennoSaami or perhaps even earlier (Sammallahti 1998:66). Some languages clearly point towards PFS *-jna/jnä (> PS *-jn¯e). For example, the second-syllable vocalism in North Saami parisyllabic i-stem and u-stem nouns (guolli “fish”, buod¯du ¯ “dam”) indicates that the third syllable contained PS *¯e (< PFS *a/ä) rather than PS *˘a (< PFS *i). What we find is forms like guliin (< PS *kuol;i-jn¯e < PFS *kala-jna) and buoduin (< ¯ Had the PFS forms instead been *kala-jni PS *puo o¯ -jn¯e < PFS *pado-jna). ¯ and *pado-jni, they would have come out as PS *kuol;¯a-jn˘a and *puo u-jn˘a, ¯ eventually yielding incorrect **guoláin and **buduin in North Saami. Com¯ pare guolli-n “as a fish” (< PS *kuol;¯e-n¯e < PFS *kala-na) with 1. possessed guollá-n “my fish” (< PS *kuol;¯a-m˘a < PFS *kala-mi). Even more direct evidence for *-jna/jnä is found in South Saami, where the vowel of the ending has not been apocopated and clearly betrays its origin in PS *¯e, such as in gieriehtsinie “sled” (Korhonen 1981:225). Finally, there is a limited amount of evidence from Fennic languages in support of *-jna/jnä, specifically in Votic and certain Estonian dialects, where forms with -ina(a), -inä(ä) and the like are attested. However, the body of evidence pointing towards *-jni is equally robust. In Inari Saami, the qualitative and quantitative makeup of the stem-final vowel preceding the ending unequivocally indicates that the apocopated suffix vowel was PS short *˘a (< *i) rather than long *¯e < *a/ä (Korhonen 1981:225, 293; cf. Itkonen 1939). According to Sammallahti (1998:66), Lule Saami likewise presupposes *-jn˘a < *-jni, and the same is also true of Fennic in general (outside of the varieties mentioned above), for example Finnish -ine. Finally, Mordva has a comitative suffix -nj ek which, if cognate with the *nV component of the Fenno-Saami comitative, also speaks in favour of PFS *-jni over *-jna/jnä. There seems to be no other option available than to reconstruct both variants of the comitative suffix, *-jna/jnä and *-jni, as having coexisted in PFS, and to assume that PS retained the reflexes of both variants, *-jn˘a and *-jn¯e. The disagreement among the individual daughter languages may then reflect dialectal


125

variation existing in the Proto-Saami period. Korhonen (1981:222) assumes that North Saami dainna, mainna, geainna and so forth are regular forms: the expected reflexes of regular phonological changes. This is supported by the fact that -inna is found across all dialects of North Saami, and that it is independently attested in the (rather less regular) paradigm of personal pronouns (muinna, duinna, suinna). However, it is not clear how the North Saami outcome /CVjn:a/ fits with the general diachronic correspondences otherwise posited by Korhonen (1981) and other sources on Saami historical phonology, unless the ending is assumed to have been PS *-jn˘a < PFS *-jni and not PS *-jn¯e < PFS *-jna/jnä. But as we saw above, parisyllabic i- and u- stems in North Saami unequivocally point towards PS *-jn¯e < PFS *-jna/jnä. When suffixed to a monosyllabic stem such as PFS *ti- > PS *t˘a- > North Saami da-, the latter would presumably have yielded **dainni, or perhaps apocopated **dain as in the locative plural, but most certainly not the attested form dainna. Since it seems clear that the two versions of the suffix (PS *-jn˘a and *-jn¯e) must have coexisted in Proto-Saami, it is possible that modern North Saami reflects both variants, each in a different word class. In monosyllabic pronouns the then goes back to forms in PS *-jn˘a whereas in polysyllabic nominals it reflects PS *-jn¯e. Given the evidence available at present, this must be considered the most plausible hypothesis. Just as -inna cannot reflect PS *-jn¯e < PFS *-jna/jnä, given what is known about Saami historical phonology, the homophonous ending found in the of the eastern Finnmark dialects (dainna, geainna) cannot be a regular reflex of what we know for certain was the ending for that form in Proto-Saami: *-j-n¯e < PFS *-j-na/nä. It is true that the development of the western variants dain, geain, is not completely clear either. Nevertheless, the other two inflectional forms where the ending also had the structure *-C-C¯e in Proto-Saami, inessive/elative singular (PS *-s-n¯e/*-s-t¯e) and accusative plural (PS *-j- e¯ ), show the exact same monosyllabism across western and eastern dialects alike: das(t), daid. The western ending -in is therefore not in itself problematic in the same way as its eastern counterpart -inna. The conclusion to draw from the comparative-historical evidence is, then, that the eastern Finnmark dialects, with dainna, geainna, matching dainna, geainna, represent a secondary state of affairs. The paradigm must originally have been as it is today in the remaining dialects, with monosyllabic forms (dain, geain) distinct from the disyllabic forms. Since they are not the result of regular phonological changes, the eastern disyl-

126


a. Model: polysyllabic stems (viessu “house”, beana “dog”)

vissui viesus viesuin

viesuide viesuin viesuiguin

beatnagii beatnagis beatnagiin

beatnagiidda beatnagiin beatnagiiguin

b. Target: monosyllabic pronouns (dat “that, it”)

dasa das dainna

daidda dain daiguin

dasa ⇒ das dainna

daidda dainna daiguin

Table 15: Extension of = syncretism in eastern Finnmark dialects labic forms must have come about by some analogical process. The most plausible explanation — indeed the only feasible one — is that suggested by Korhonen (1981:222). The monosyllabic pronouns simply copied the = syncretism pattern which was ubiquitous throughout the nominal inflection. 5.3 Theoretical implications of the North Saami evidence The diachronic development of the of monosyllabic pronouns in the eastern Finnmark dialects of North Saami, schematized in Table 15, constitutes evidence that the inflectional homonymy linking and in polysyllabic stems — and, in eastern Finnmark dialects, of monosyllabic stems as well — is in fact a genuine syncretism. A prerequisite for the analogical extension of this homonymy pattern from one word class to another is that the syncretism was encoded in the grammar. In other words, the = relation must be a systematic element of North Saami morphology and not a mere accidental homonymy (along the lines of, say, 1. -n in verbs and essive -n in nouns). It should be clear that this syncretism pattern is precisely of the sort that ought to be ruled out given the underspecification/neutralization approach. To begin with, it would be hard to argue that the categories Locative(/Ablative) and Comitative(/Instrumental) somehow form a minimal opposition, distinguished by some feature [+F] versus [-F], while at the same time forming a ‘natural class’ within the case system based on their specifications for all other attributes ([±G], [±H], and so on). Aside from the difficulty in identifying any such commonali-


127

ties, the Locative(/Ablative) category obviously enters into a much more closely knit two-way subsystem with the Illative, in which the Comitative(/Instrumental) takes no part. But even if such an analysis in terms of abstract semantic features were feasible, a thornier problem remains: the two forms are also opposed to each other as [+plural] versus [-plural], and neutralizing this opposition will inevitably implicate and as well. A similar problem arises for the geometric conflation approach of McCreight & Chvany (1991): ‘diagonal’ pairs of disjoint cells like and can only be syncretized through the mediation of either or (or both), as this is required in order to yield a geometrically contiguous partitioning of the case/number space. A further problem, shared by both approaches, involves the fact that the ending is arguably internally complex (plural -i- + locative -n), while the ending is not. As long as we want to analyze the as containing the very same number-marking suffix -i- as other oblique plural case forms do, then there would appear to be no viable way of expressing an identity relation between the form and any singular case form ending in -in — even if this happened to be true of the geometrically contiguous rather than the non-contiguous Approaches that express syncretism through referral, by contrast, encounter no problems with the North Saami pattern. Recall that in inferential-realizational theories such as Paradigm Function Morphology (Stump 2001) or Network Morphology (Corbett & Fraser 1993), inflectional morphemes do not exist as such; rather, the segmental string which makes up an inflectional affix is introduced by a rule making reference to the relevant morphosyntactic properties (those marked by that affix). In North Saami, the plural marker -i- is accounted for by a rule of exponence like (14), and the rule responsible for locative -n (in the plural) is something along the lines of (15). (14)

RR{:pl},N (<X, σ>) =def <Xi, σ>

(15)

RR{ :loc, :pl},N (<X, σ>) =def <Xn, σ>

Let us assume that the rules assigning case suffixes, such as (15), apply to the output of rules assigning number suffixes, like (14). The realization of a word W in the —in other words the realization of <W, ρ>, where ρ is a complete extension of the set { :loc,

:pl}—will be determined as in (16). The result is <Win, ρ>, which is thus also the result of applying the

128


paradigm function as a whole (North Saami inflectional morphology as such) to the pairing <W, ρ>. In other words, PF(<W, ρ>) = <Win, ρ>. (16)

RR{ :loc, :pl},N (RR{:pl},N (<W, σ>)) =<Win, σ>

We can now make use of this fact in the account of the = syncretism pattern. If we stipulate that it is which is ‘copying’ the realization of , rather than the other way around, the syncretism can be captured by the rule of referral in (17) or something like it (the issue of directionality will be addressed below). (17)

RR{ :com, :sg},N (<X, σ>) =def , where PF(<X, σ/{ :loc,

:pl}>) =

The rule of referral in (17) will determine the content of the paradigm cell (of the same word-root W), that is, the pairing <W, τ> where τ is a complete extension of the set { :com,

:sg}. The rule does so by referring it to the realization of the cell in the same paradigm (that is, of the pairing <W, ρ> dealt with above). Since PF(<W, ρ>) = <Win, ρ> by (16), the effect of rule (17) is that PF(<W, τ>) = <Win, τ>. We have arrived at the desired result: (a) that and are both marked by the suffixation of -in; (b) that this commonality is not accidental but explicitly encoded in the morphological system; and (c) that the -i- of the is the same as that found in other oblique plural forms (in that a single rule determines the occurrence of -ithroughout the plural). Something like (14–17), then, constitutes part of the inflectional morphology of North Saami, and provides the basis for the analogical extension that occurred in the eastern Finnmark dialects. In the western dialects, the rule of referral is restricted to the inflectional class of polysyllabic nominals: nouns, adjectives, participles, and those indefinite pronouns that are not monosyllabic (and which occur in the plural, such as soames ∼ soamis “someone”). There are two ways of incorporating this limitation into the above analysis. One is to explicitly restrict the class of words to which the rule of referral in (17) applies (indicated here by the subscript ‘N’, intended to stand for all nominals). Another, more attractive, solution is to maintain the generality of (17) as encompassing all nominal categories, but to have it be overridden by a more specific rule, applying to the of (monosyllabic) pronouns alone as shown in (18). The need for specific rules for this class of lexemes is independently motivated in some of the


129

other case forms, for example being marked with -sa (instead of the general -i) as captured by (19). (18)

RR{ :+, :com, :sg},N (<X, σ>) =def <Xinna, σ>

(19)

RR{ :+, :ill, :sg},N (<X, σ>) =def <Xsa, σ>

Being more narrowly defined, the rule of exponence in (18) overrides the rule of referral in (17). Consequently, the of a pronoun like dat (with the root da-) is dainna by (18), instead of being defined by (17) as having an identical realization to the form dain (whose realization is in turn defined by the joint action of the rules of exponence in (14) and (15) above). This is precisely what we find in the western dialects. At this point the reader may be puzzled by an apparent contradiction. The rule of referral in (17) states that, synchronically, the ‘copies’ the realization of the , whereas what happened in the eastern Finnmark dialects was that, diachronically, the instead copied the form. However, a crucial distinction must be upheld between a synchronic ‘process’ (essentially an implicational relationship defined by a synchronic grammar G) and a diachronic process (the relationship between a synchronic grammar G at time t and another, slightly different synchronic grammar G at time t+1). Assuming that analogical change typically occurs through imperfect learning, that is, in the process of language transmission across generations (Kiparsky 1992), grammar G results from a learner’s attempt to ‘reconstruct’ grammar G (imperfectly, hence restructuring it in the process) using the output of G as his/her primary input data on which to base that (re)construction. Given the existence of rules (14), (15), (17) and (18) in grammar G (essentially, the grammar currently found in western Finnmark dialects), the eastern Finnmark system with dainna instead of dain (and no other overt differences) could have arisen as grammar G by the following conjectural scenario. Imagine a speaker-learner S who has, on the basis of observed data, correctly posited the general rules for plural and locative (plural) marking in (14–15). Let us assume that S has also correctly posited the general rule of referral in (17), deriving the shape of the from that of the corresponding form. S would have done so on the basis of (a) the persistently recurring = matches in the input data, and (b) the insight that the -in must be the morphologically complex -i-n, based on comparison with the other oblique plural forms (-id, -ide/-idda, -iguin). Finally, S is in the process of establishing special rules — overriding these more general ones — which s/he

130


has discovered are necessary to account for certain inflectional forms of monosyllabic pronouns, including -sa (for non-personal pronouns), -t (for demonstratives), -nnje/-tnje (for personal pronouns), and so forth. Upon encountering dainna as the form of dat “that, it” (and similarly mainna from mii “what”, geainna from gii “who”), and noticing that this contains an ending different from the usual one (that is, -inna rather than -in), S is faced with a choice between two alternatives. One is the ‘correct’ one, namely to assume that -inna should be accounted for by a special rule of exponence for the form of pronouns, analogous to the special rule required, for example, for -sa. This would result in G = G, the successful ‘replication’ by S of the previous generation’s grammar (at least with respect to this part of the inflectional morphology). The other alternative interpretation of dainna which S must consider is an ‘incorrect’ one, namely that this form, just like all other forms in the language, is an exact replica of the corresponding form. Consequently, S will infer that, given dainna, it must be the case that the of dat is also dainna. This too is a slightly unexpected form, since the general rule in (15) would predict dain (da-i-n). S must therefore assume that there is a special rule of exponence accounting for the occurrence of -nna (instead of -n) as the locative marker in the plural paradigm of pronouns — just like there is such a special rule for -sa (instead of -i) as the illative marker in the singular paradigm of these same words, and so forth. The end point of this alternative learning path is a system where the of a pronoun like dat is da-i-nna, by the general rule suffixing plural -i and a special pronoun rule suffixing locative (plural) -nna, and where the of dat is exactly identical to this form, due to the general rule of referral in (17). In other words, we now have G’=G, where the grammar of S at time t+1 is subtly different from the grammar of older speakers at time t. A diachronic change has occurred, both at the covert level of the grammar itself (the realization rules and their definitions) and at the overt level of language output (the observable inflectional forms). At the level of the grammar, there is now a special pronoun rule for locative case marking in the plural (-nna rather than -n), instead of there being a special pronoun rule for comitative case marking in the singular (-inna rather than -in). At the level of observable language output, the form of monosyllabic pronouns has the shape CVinna instead of CVin. As long as our language-learner S is the least bit uncertain of what exactly the correct form of monosyllabic pronouns like dat is supposed to be,


131

there is a distinct possibility that s/he will choose the wrong one of these two learning paths. This will ultimately result in the transformation of the pronominal paradigm (relative to that of the generation preceding S) exactly as shown in Table 15b above. As a diachronic event, the change dain ⇒ dainna will thus have occurred by means of productive syncretism: an analogical restructuring based on the existence of the rule of referral (17) in the grammar. This is precisely what took place in the diachronic development of the eastern Finnmark dialects of North Saami. The apparent contradiction in the directionality of the referral relation ( → versus ← ) is a mere illusion, which evaporates as soon as one is careful to keep the dimensions of synchrony and diachrony separate. The directionality inherent in rule (17), whereby the realization of is dependent on that of but not vice versa, is thus fully compatible with the historical change that took place in the North Saami dialects in question. 6. Summary The proper analysis of syncretism in inflectional morphology is a controversial topic, and one that bears on fundamental differences between competing formal-theoretical frameworks. In this paper I have attempted to elucidate what the main issues are in this context, and what kinds of evidence can be adduced to determine whether a given inflectional homonymy pattern does in fact count as a syncretism (and can thus be used as empirical evidence for or against a given theory). A central argument was that diachronic changes of a particular kind — so-called productive syncretisms — provide a particularly valuable type of evidence in this regard. Three broadly defined approaches to syncretism were contrasted with each other with respect to their predictions regarding possible versus impossible syncretism patterns. These were referred to as the feature neutralization (or underspecification) approach, the closely related geometric conflation approach, and the rule of referral approach. A central claim was that certain subtle differences among modern North Saami dialects, when taken together with comparativehistorical evidence and current knowledge about Saami historical phonology and morphology, constitute unequivocal evidence in favour of the rule of referral approach over other alternatives. In the eastern Finnmark dialects, the pervasive identity pattern = has been extended from the inflectional paradigm of polysyllabic nominals to that of monosyllabic ones. As the two paradigm cells are disjoint, the rule of referral approach is the only alternative

132


which is capable of expressing the = relation. A tentative analysis was developed within one such model, the Paradigm Function Morphology of Stump (1993, 2001). Finally, a diachronic scenario was outlined, invoking analogical restructuring through imperfect learning, to account for the change distinguishing the eastern Finnmark dialects. Baerman et al. (2001) propose a typology of case syncretism patterns, in which ‘Type 5’ is the use of a single form for one case in one number and a different case in a different number. Citing an earlier version of the present work, they refer to North Saami as “the most convincing example” of such syncretism existing as “a stable morphological relationship” between the two paradigm cells in question (the North Saami facts are again cited in Baerman 2004 in a similar context). Other cases of disjoint syncretism patterns are most certainly attested (cf. section 4.3; see also Stump 2001: ch. 7). However, what sets the North Saami case apart as a particularly persuasive one is the fact that the syncretism is revealed through an active (diachronic) process, supported by clear comparativehistorical evidence, involving the analogical extension of a productive identity pattern. In conclusion, the rather innocuous morphological reanalysis that happened at some point in the recent history of the eastern Finnmark dialects of North Saami has profound implications for morphological theory, and constitutes strong evidence in favour of rules of referral, or their equivalent, as a component of inflectional systems in the world’s languages. References Anderson, Stephen R. 1992. A-morphous Morphology. Cambridge: Cambridge University Press. Andronis, Mary, Christopher Ball, Heidi Elston & Sylvain Neuvel, eds. 2001. CLS 37: The Main Session. Papers from the 37th meeting of the Chicago Linguistic Society, vol. 1. Chicago: Chicago Linguistic Society. Baerman, Matthew. 2004. “Directionality and (un)natural classes in syncretism”. Language 80.807–827. Baerman, Matthew, Dunstan Brown & Greville Corbett. 2001. “Case syncretism in and out of Indo-European”. In Andronis et al. 2001, 15–28. . 2005. The Syntax-Morphology Interface: A study of syncretism. Cambridge: Cambridge University Press. Bartens, Hans-Hermann. 1989. Lehrbuch der saamischen (lappischen) Sprache. Hamburg: Helmut Buske.


133

Bierwisch, Manfred. 1967. “Syntactic features in morphology: General problems of so-called pronominal inflection in German”. To Honor Roman Jakobson: Essays on the occasion of his seventieth birthday, 239–270. The Hague: Mouton. Blevins, James P. 1995. “Syncretism and paradigmatic opposition”. Linguistics and Philosophy 18.113–152. Carstairs, Andrew. 1984. “Outlines of a constraint on syncretism”. Folia Linguistica 18.73–85. . 1987. Allomorphy in Inflexion. London: Croom Helm. Corbett, Greville, Matthew Baerman & Dunstan Brown. 2001. “Domains of syncretism: A demonstration of the autonomy of morphology”. In Andronis et al. 2001, 385–398. Corbett, Greville G. 1991. Gender. Cambridge: Cambridge University Press. Corbett, Greville G. & Norman M. Fraser. 1993. “Network Morphology: A DATR account of Russian nominal inflection”. Journal of Linguistics 29.113– 142. Dyła, Stefan. 1984. “Across-the-board dependencies and case in Polish”. Linguistic Inquiry 15.701–705. Evans, Nicholas, Dunstan Brown & Greville Corbett. 2001. “Dalabon pronominal prefixes and the typology of syncretism: A Network Morphology analysis”. Yearbook of Morphology 2000 ed. by Geert Booij & Jaap van Marle, 187–231. Dordrecht: Kluwer. Evans, Roger & Gerald Gazdar. 1996. “DATR: A language for lexical knowledge representation”. Computational Linguistics 22.167–216. Gvozdanović, Jadranka. 1991. “Syncretism and the paradigmatic patterning of grammatical meaning”. In Plank 1991a, 133–160. Haiman, John. 1980. Hua: A Papuan language of the Eastern Highlands of New Guinea. Amsterdam and Philadelphia: John Benjamins. . 1998. “Hua (Papuan)”. Handbook of Morphological Theory ed. by Andrew Spencer & Arnold M. Zwicky, 539–562. Oxford: Blackwell. Häkkinen, Kaisa. 1985. Suomen kielen äänne- ja muotorakenteen historiallista taustaa. Turku: Åbo Akademi. Hjelmslev, Louis. 1935. La catégorie des cas. Aarhus: Universitetsforlaget. Hockett, Charles F. 1954. “Two models of grammatical description”. Word 10.210–231.

134


Itkonen, Erkki. 1939. Der ostlappische Vokalismus vom qualitativen Standpunkt aus mit besonderer Berücksichtigung des Inari- und Skoltlappischen. (= Mémoires de la Société Finno-Ougrienne 79.) Helsinki: Suomalais-Ugrilainen Seura. . 1960. Lappische Chrestomathie mit grammatikalischem Abriss und Wörterverzeichnis. (= Suomalais-Ugrilaisen Seuran Apuneuvoja 7.) Helsinki: Suomalais Kirjallisuuden Seura. Jakobson, Roman. 1936. “Beitrag zur allgemeinen Kasuslehre: Gesamtbedeutungen der russischen Kasus”. Travaux du Cercle Linguistique de Prague 6.240–288. Johnston, Jason. 1997. Systematic Homonymy and the Structure of Morphological Categories: Some lessons from paradigm geometry. Ph.D. thesis, University of Sydney. Kiparsky, Paul. 1992. “Analogy”. International Encyclopedia of Linguistics ed. by William Bright, 56–61. Oxford: Oxford University Press. . 2001. “Structural case in Finnish”. Lingua 111.315–376. Korhonen, Mikko. 1981. Johdatus lapin kielen historiaan. Helsinki: Suomalaisen Kirjallisuuden Seura. Matthews, Peter H. 1972. Inflectional Morphology: A theoretical study based on aspects of Latin verb conjugation. Cambridge: Cambridge University Press. McCreight, Katherine & Catherine Chvany. 1991. “Geometric representation of paradigms in a modular theory of grammar”. In Plank 1991a, 91–111. Neidle, Carol. 1988. The Role of Case in Russian Syntax. Dordrecht: Kluwer. Nickel, Klaus Peter. 1990. Samisk grammatikk. Oslo: Universitetsforlaget. Nielsen, Konrad. 1979[1932–1962]. Lappisk ordbok: grunnet på dialektene i Polmak, Karasjok og Kautokeino [ = Lapp Dictionary: Based on the dialects of Polmak, Karasjok and Kautokeino]. (= Instituttet for sammenlignende kulturforskning, serie B 17.) Oslo: Aschehoug. 5 volumes. Co-author for vols. 4-5: Asbjørn Nesheim. Noyer, Rolf. 1997. Features, Positions and Affixes in Autonomous Morphological Structure. New York: Garland. . 1998. “Impoverishment theory and morphological markedness”. Morphology and its Relation to Phonology and Syntax ed. by Steven Lapointe, Diane Brentari & Patrick Farrell, 264–285. Stanford, Calif.: CSLI. Plank, Frans, ed. 1991a. Paradigms: The economy of inflection. Berlin: Mouton de Gruyter. Plank, Frans. 1991b. “Rasmus Rask’s dilemma”. In Plank 1991a, 264–285.


135

Pullum, Geoffrey K. & Arnold M. Zwicky. 1986. “Phonological resolution of syntactic feature conflict”. Language 62.751–773. Ravila, Paavo. 1960. “Adjektiiviattribuutin kongruenssin synty suomen kielessä”. Sananjalka 2.28–35. Sammallahti, Pekka. 1998. “Saamic”. The Uralic Languages ed. by Daniel Abondolo, 43–95. London: Routledge. Spencer, Andrew. 1991. Morphological Theory: An introduction to word structure in generative grammar. Oxford: Blackwell. Stump, Gregory T. 1993. “Rules of referral”. Language 69.449–479. . 2001. Inflectional Morphology. Cambridge: Cambridge University Press. Taraldsen, Knut Tarald. 1981. “Case-conflict in Norwegian topicalization”. NELS 11: Proceedings of the 11th Annual Meeting of the North Eastern Linguistic Society ed. by Victoria Burke & James Pustejovsky, 377–398. Amherst, Mass.: University of Massachusetts Department of Linguistics. Wiklund, K. B. 1902. “Zur Geschichte des urlappischen a¯ und u¯ in unbetonter Silbe”. Finnisch-Ugrische Forschungen 2.41–71. Williams, Edwin. 1981. “On the notions ‘lexically related’ and ‘head of a word”’. Linguistic Inquiry 12.245–274. Zaenen, Annie & Lauri Karttunen. 1984. “Morphological non-distinctiveness and coordination”. Proceedings of the 1st Eastern States Conference on Linguistics ed. by Gloria Alvarez, Belinda Brodie & Terry McCoy, 309–320. Columbus: Ohio State University. Zwicky, Arnold M. 1985. “How to describe inflection”. Proceedings of the 11th Annual Meeting of the Berkeley Linguistics Society ed. by Mary Niepokuj, Mary Van Clay, Vassiliki Nikiforidou & Deborah Feder, 372–386. Berkeley: Berkeley Linguistics Society. . 1991. “Systematic versus accidental phonological identity”. In Plank 1991a, 113–131. . 2000. “Describing syncretism: Rules of referral after fifteen years”. 26th Annual Meeting of the Berkeley Linguistics Society, February 18-21, 2000, University of California, Berkeley. Zwicky, Arnold M. & Jerrold M. Sadock. 1975. “Ambiguity tests and how to fail them”. Syntax and Semantics, vol. 4 ed. by John P. Kimball, 31–36. New York: Academic Press.

ROOTS AND VERBS IN NORTH SAAMI∗

MARIT JULIEN Lund University Although it has been argued lately that roots have no lexical category, a close look at deadjectival and denominal verbs in North Saami reveals that roots nevertheless differ with respect to their semantic type, and that this semantic contrast between roots leads to systematic syntactic and semantic differences between derived verbs. More specifically, state-denoting (‘adjectival’) roots can combine directly with a verbalizer, yielding verbs that mean ‘be Root’, ‘become Root’ or ‘cause to be Root’. Entity-denoting roots, on the other hand, must combine with a (possibly abstract) preposition before the verbalizer is merged, and because of the obligatory presence of the preposition, the result is a verb that means ‘have Root’, ‘get Root’ or ‘cause to have Root’. Hence, it is not the case that any root can appear in just any syntactic environment.

1. Introduction Over the last decade many researchers have entertained the idea that the lexical-conceptual structure of a word is encoded in the syntax (see, for example, Hale & Keyser 1993, 1998, 2002; Arad 1998; Travis 1992, 2000; Pylkkänen 2002, and Ramchand 2002a). Applied to verbs, this idea implies that if a verb denotes a semantically complex event, each part of the event is represented in the syntax. Alternations with respect to event structure and argument structure should then be tightly connected to syntactic alternations. In this paper, I use a framework of the type just mentioned to investigate North Saami deadjectival and denominal stative, inchoative and causative verbs. ∗

The research for this paper was financially supported by the Norwegian Research Council, grant no. 141687/540. In developing the ideas presented here, I have received helpful input from Berit Anne Bals, Harald Gaski, and Peter Svenonius. Ollu giitu! I also thank the audiences at Grammar in Focus 2003, Lund, BLS 29, Berkeley, the Thursday Night Linguistics Seminar, University of Tromsø, and three anonymous reviewers for their feedback.

138

MARIT JULIEN

In addition to suggesting how these verbs should be decomposed syntactically, I address the proposal put forth in Marantz (1997) and Borer (2000) that roots have no category, and that the lexical categories noun, adjective, and verb result from combining a category-less root with functional projections of a nominal, adjectival and verbal nature, respectively. I argue that roots are nevertheless differentiated semantically. As Marantz (2001) also observes, some roots denote entities while others denote states. In my view, this distinction has syntactic consequences: an entity-denoting root, for example, only needs to be combined with nominal functional elements to become a noun in the syntactic sense, but in order to make an adjective out of it, a category-changing element is also necessary. Hence, it is not the case that any root can become a verb, a noun, or an adjective at the same cost. Instead, each semantic class corresponds most closely to one particular lexical category. The semantic differentiation of roots is also seen in derived words, where the roots appear without the functional apparatus that otherwise give them a syntactic category (the same point is made in Baker 2002, 2003). In North Saami, the meaning of stative, inchoative and causative verbs built on adjectival roots contrasts systematically with the meaning of verbs built on nominal roots. Deadjectival verbs mean ‘be Root’, ‘become Root’, or ‘cause to be/become Root’; in other words, the meaning of the root is ultimately predicated of the lower argument. This is not true of denominal verbs, which (with some exceptions) mean ‘have Root’, ‘get Root’, or ‘cause to have Root’. Thus, the semantic class membership of the root affects the computation of the meaning of the derived verb as a whole. The paper is organized as follows. In section 2, I present the basic facts about deadjectival and denominal stative verbs in North Saami, and I put forth a proposal concerning the syntactic representation of these verbs. In section 3, I do the same for deadjectival and denominal inchoative verbs. In section 4, I address the relation between roots and syntactic categories. In section 5, I discuss the analysis of derived verbs proposed by Baker (2002, 2003), and in section 6, I present my own analysis of deadjectival and denominal stative and inchoative verbs in North Saami. Some proposals concerning causative verbs follow in section 7. In section 8, I add some more comments on Marantz (1997), and my conclusions are summed up in section 9. 2. Stative verbs Let us start with the semantically simplest type of derived verbs, stative verbs. These verbs denote eventualities that are devoid of dynamics altogether.

139

ROOTS AND VERBS

a. b. c. d. e.

ruoksat bahˇca bahˇca liekkas liekkas

‘red’ ‘bitter’ ‘bitter’ ‘warm’ ‘warm’

> > > > >

ruoks-á-t bahˇc-ist-it bahˇc-iid-it liekk-ist-it likk-iid-it

‘appear red’ ‘smell or taste bad’ ‘smell or taste bad’ ‘feel warm’ ‘feel warm (by direct contact)’

Table 1: Stative verbs derived from adjectives In the terms of Parsons (1990), they simply hold or do not hold. As illustrated in Table 1, stative verbs can be built from adjectival roots in North Saami.1 Note that the verbs are given in their infinitival form, which includes the infinitival marker -t (after vowels) or -it (after consonants). The actual derivational marker is the boldfaced element between the root and the infinitival marker.2 The citation forms of the adjectives also include more than just the root. In ruoksat ‘red’, for example, the final -at can be analyzed as a marker of nominative singular. Since adjectives in North Saami agree with their subjects when they are used predicatively, the nominative singular form of the adjective shows up in constructions like (1a). When adjectives are used attributively, on the other hand, they do not agree with the noun. Instead, they show up in an invariant attributive form, which often involves a suffix that is not seen elsewhere in the adjectival paradigm. An example is the -es suffix in rukses ‘red’ in (1b).3

1

These verbs are called essive in Nielsen (1979[1926]) and Nickel (1990). My data are partly drawn from Nielsen (1979[1926]), Nielsen (1979[1932–1962]), and Nickel (1990), and partly elicited from native speakers. The reader should, however, note that not all the constructions discussed in this paper would be used by every present-day speaker of North Saami. It appears that under the influence of the much more analytic Scandinavian languages, several of the derivational patterns seen in the language of older generations are now about to disappear in the North Saami of younger speakers (see Sara 2002). 3 The North Saami examples in this paper are written in the orthography that was agreed upon by the joint Saami conference in 1978 and later adopted as the official orthography for North Saami in Norway, Sweden and Finland (the three countries where this variety is spoken). Note the following characters: represents [a], represents [tS], represents [S], represents [ts], and represents [dz]. 2

140 (1)

MARIT JULIEN

a.

Bireh-a biila lea ruoks-at. Biret- car. is red- “Biret’s car is red.”

b.

Bireh-is lea ruks-es biila. Biret- is red- car. “Biret has a red car.”

We can conclude that the root meaning ‘red’ is ruoks-, and that only the root of the adjective, not the suffix, takes part in verb formation. (The monophthongization /uo/>/u/, triggered by the unstressed /e/ of the suffix, is one instance of a more general phonological phenomenon in the language.) As for the verbalizing suffixes that we see in Table 1, their meanings can be characterized as follows. The suffix -á- indicates a state or appearance. Arguably, the same element is found in some non-derived verbs describing postures, such as ceaggát ‘stand upright’, cˇ ohkkát ‘sit’ and veallát ‘lie’ (see Nielsen 1979[1926]:234).4 The suffixes -ist- and -iid- both mean ‘appear to the senses’, apparently with the minor difference that -iid- but not -ist- indicates direct contact; compare (d) and (e) of Table 1. Thus, the meanings of these suffixes are more specific than the meaning of the copula leat ‘be’, which can also be used to make verbal predicates out of adjectives, as in leat ruoksat ‘be red’, leat liekkas ‘be warm’ (see (1a)).5 In North Saami, stative verbs can also be formed from nominal roots. Some examples are given in Table 2.6 (In Table 2a-b we see another pervasive feature of North Saami morphophonology, namely, stem consonant gradation, as exemplified by the alternation tn ∼ dn.) Note, firstly, that the choice of verbalizing suffix is idiosyncratic for each root. Traditionally, such idiosyncratic combinations are seen as being typical of ‘lexical’ derivation. However, Marantz (1997) argues that the properties associated with ‘lexical’ word formation are restricted to a specific syntactic domain — namely, the domain below the syntactic head that introduces the agent. None of the verbs we have seen so far involves an agent. Hence, on Marantz’s approach, the observed idiosyncrasies are not unexpected. 4

In addition to the stative -á- North Saami has an -á- that marks inceptive aspect, as illustrated by the verb pair doarrut ‘fight’ vs. doarrát ‘begin to fight’. A few verbs in -á- are neither stative nor inceptive, examples are buiját ‘cheat, swindle’ and cummát ‘kiss’ (cf. Sammallahti 2002). 5 The copula is a relatively recent innovation in Saami (see Korhonen 1981, for example). 6 As an anonymous reviewer points out, bártidit in Table 2d can also be a causative verb. However, here I am dealing with the stative verb bártidit.

ROOTS AND VERBS

a. b. c. d. e.

botnji ritni ballu bárti illu

‘twist’ ‘rime’ ‘fear’ ‘trouble’ ‘joy’

> > > > >

141

bodnj-á-t ridn-á-t ball-a-t bárti-d-it illu-d-it

‘be twisted, bent’ ‘be covered with rime’ ‘fear, be afraid’ ‘be in trouble’ ‘be happy’

Table 2: Stative verbs derived from nouns Secondly, there is a systematic meaning difference between the verbs in Table 1 and those in Table 2. While the verbs in Table 1 essentially mean ‘be Adjective’, the verbs in Table 2 do not mean ‘be Noun’, they mean roughly ‘have Noun’. But there is no obvious reason why the derived verbs in Table 2 should not mean ‘be Noun’. From a pragmatic point of view, we would expect a verb meaning ‘be joy’, for example, to be perfectly acceptable. Moreover, the nominals in Table 2 may well combine with the copula leat ‘be’ to give expressions like leat ritni ‘be rime’, leat illu ‘be joy’. Hence, the source of the meanings that we see in Table 2 must be sought in the grammar, or more precisely, in the syntactic and semantic relations holding between the elements that the complex verbs are made up of. On the assumption that morphologically and/or semantically complex predicates are built in the syntax, all the verbs shown in Table 1 and Table 2 must minimally have the syntactic representation sketched in (2). Here the root, which is not verbal, has combined with a verbal element, designated as VS(T AT IV E) , which projects a VS P and thereby accommodates in its Spec the subject of the state, that is, the argument that the state holds of. The verbalizing suffixes in Table 1 and Table 2 are then all realizations of this VS element. On top of VS P there will be inflectional elements, but I will ignore this here and in the following, since the focus is on the formation of the verb itself. VS P

(2)

VS P

DP VS

Root

Postulating the structure in (2) is however only the first step towards providing an explanation of the meaning contrast between deadjectival stative verbs on the one hand and denominal stative verbs on the other. But since other classes

142

MARIT JULIEN

a. b. c. d. e. f. g.

dimis goikkis obba guhkki lossat stuoris bahˇca

‘soft’ ‘dry’ ‘cloudy’ ‘long’ ‘heavy’ ‘big’ ‘bitter’

h. ruoksat ‘red’ i. seavdnjat ‘dark’

> > > > > > >

" dipm-a-t goik-a-t obb-e-t guhkk-u-t loss-u-t stuorr-u-t bahˇca-g-it

‘become soft’ ‘become dry’ ‘become cloudy’ ‘become long(er)’ ‘become heavy/heavier’ ‘become big(ger)’ ‘become bitter, get a bad taste’ > ruvss-od-it ‘turn red, blush’ > sevnnj-od-it ‘become dark’

Table 3: Inchoative verbs derived from adjectives of derived verbs in North Saami show parallel meaning differences, I will postpone further discussion of this contrast until some more derived verbs have been presented. 3. Inchoative verbs In North Saami we also find derived inchoative verbs.7 Following Parsons (1990), I define inchoative verbs as intransitive verbs that denote an event where an entity (the subject) undergoes a change of state. In Table 3, I show some North Saami inchoative verbs formed from adjectival roots; and in Table 4, I show some that are formed from nominal roots. It appears that these verbs are telic, in the sense that they are interpreted as describing bounded events. Notably, they are all much better with in x time adverbials than with for x time adverbials. I illustrate this with the two examples in (3). (3)

a.

goikka-i {* golbma beaivvi / Biergu three. day. meat. dry- $ beaivvi-s}. day- “The meat dried *for three days/OK in three days.”

OK

golmma three.

7 These verbs are called translative by Nielsen (1979[1926]) and Nickel (1990), who reserve the term inchoative for verbs denoting the initial point of an action.

143

ROOTS AND VERBS

a. b. c. d. e. f. g. h. i. j.

cˇ áhci ruobbi bárti vašši dállu idja vealgi bihkka dálvi sálti

‘water’ ‘rash’ ‘trouble’ ‘hatred’ ‘farm’ ‘night’ ‘debt’ ‘tar’ ‘winter’ ‘salt’

> > > > > > > > > >

" cˇ áhc-u-t ruobb-u-t bárttá-sk-it vašá-sk-it dálui-duvva-t idja-duvva-t vealggái-duvva-t bihka-huvva-t dálvvá-huvva-t sálta-šuvva-t

‘become wet’ ‘get a rash’ ‘get into trouble’ ‘get furiously angry’ ‘get a farm’ ‘be overtaken by night’ ‘get (deeply) into debt’ ‘get smeared with tar’ ‘be overtaken by winter’ ‘absorb salt; become salt enough’

Table 4: Inchoative verbs derived from nouns b.

{* guokte sekundda / Elle vašáski-i two. second. E. get.angry- $ sekundda-s}. second- “Elle got angry *for two seconds/OK in two seconds.”

OK

guovtte two.

Hence, the semantics of derived inchoative verbs in North Saami involves not only the process that their subject undergoes, but also the resulting state. If verbal meaning is built compositionally in the syntax, as I am assuming here, it follows that the verbs presently under discussion must involve a verbal head that encodes the process as well as a syntactic correlate of the result state. Following Ramchand (2002a,b) I take the result state to be represented by a projection whose head — the ‘R head’ in Ramchand’s terminology — is verbal in nature. It can have a verbal realization; for example, Ramchand (2002b) argues that R is spelled out as a verb in resultative light verb constructions, exemplified by Bengali in (4). (4)

kh eye ph ellam. Ami am-Ta I. mango- eaten. throw. & “I ate up the mango.”

On Ramchand’s analysis, kh eye ‘eaten’ in (4) is the spellout of R, while ph ellam ‘threw’ is the spellout of a higher verbal head with processual content. Thus, the verbal nature of R seems clear, and I think it is justified to replace the designation

144

MARIT JULIEN

R with VS . The syntax of derived inchoative verbs in North Saami is then as shown in (5).8 (5)

VD P VD P

DP

VS P

VD DP

VS P VS

Root

Here a processual verbal head VD (D for ‘dynamic’) is merged over the projection of the stative VS . The argument DP that starts out in Spec-VS P, where it is interpreted as the subject of the state, raises to Spec-VD P and gets the additional interpretation that it is also the undergoer of the process (see Ramchand 2002a). The fact that stuorrut in Table 3f, for example, can mean both ‘get big’ and ‘get bigger’ does not necessarily contradict my analysis. As Parsons (1990) points out, when an inchoative verb is formed from a gradable adjectival root, such as ‘big’, the resulting state need not involve a predefined or absolute degree to hold true of its subject; it suffices that in the resulting state, the subject is bigger than it was before the change of state took place (see Parsons 1990). Hence, either meaning of stuorrut should be compatible with the structure in (5), and thereby also with a telic reading.9 Interestingly, the suffixes -(i)duvva-, -huvva-, and -šuvva-, seen in Table 4e– j, are arguably morphologically complex. The element -d-, for example, appears on its own as a verbalizer in Table 2d–e, where I analyzed it as a realization of VS . I therefore take -d- to represent VS also in Table 4e–g, and -h- and -š- in Table 4h–j to be like -d-. The -uvva- part of the suffix is then a realization of the processual head VD in these examples. As for the non-complex inchoative verbalizers in Table 3 and Table 4, the -g-, -od-, and -sk- verbalizers only derive inchoative verbs. These elements might 8

In this paper I use strikethrough notation to indicate elements that have moved. Another relevant proposal is found in Hay et al. (1999), where it is argued that deadjectival verbs will be telic if the degree to which the affected argument changes with respect to the property associated with the adjectival root can be interpreted as bounded. In the case under discussion, the upper bound can in principle be any state where the undergoer is bigger than it was before. 9

ROOTS AND VERBS

145

therefore be seen as realizations of VD . The elements -a-, -e-, and -u- have other uses as well. For example, we have already seen in Table 2c that -a- can form stative verbs. Another example of a stative verb with -a- is astat ‘have time to’. Stative verbs with -e- also exist, such as heivet ‘be suitable’. Hence, it is not clear exactly which head these elements spell out in derived inchoative verbs, and I will not try to give a more specific suggestion here. Another point worth noting about the verbs shown in Table 3 and Table 4 is that we find a meaning contrast between Table 3 on the one hand and Table 4 on the other that is parallel to the contrast between Table 1 and Table 2. While the verbs in Table 3 all mean ‘become Adjective’, the verbs in Table 4 do not mean ‘become Noun’, they mean ‘get Noun’ or ‘be affected by Noun’ (see Nickel 1990:281). The observed parallelism calls for a unified explanation of the semantic differences between verbs built on adjectival roots and verbs built on nominal roots. In the following sections, I will try to develop such an explanation. 4. Roots and categories As we have just seen, intransitive verbalizations in North Saami fall into two classes. In the first class, exemplified in Table 1 and Table 3, the root of the verb is ultimately predicated over the subject. In the second class, exemplified in Table 2 and Table 4, there is no such predication. One might want to argue that the verbalizing suffixes are the source of the meaning difference between Table 1 and Table 2, and between Table 3 and Table 4. However, the fact that there is overlap between the set of verbalizers employed in Table 1 and the set of verbalizers employed in Table 2 and between the set of verbalizers employed in Table 3 and the set of verbalizers employed in Table 4, poses a problem for such an approach. Moreover, a similar meaning contrast is seen in English zero causative derivations, as (6) and (7) illustrate. (6)

a. b

clear (a screen) ‘make clear’ dry (some tomatoes) ‘make dry’

(7)

a. saddle (a horse) ‘attach a saddle to’ b. man (the torpedoes) ‘put men to’

Instead of postulating two different phonologically empty causativizers in English, and various pairs of homophonous verbalizers in North Saami, I will argue that roots fall into different classes, and that the class membership of the root matters for the result of the verbalization. This claim is seemingly in conflict

146

MARIT JULIEN

with the proposal put forth by Marantz (1997) that roots have no category and that nouns, verbs, and adjectives are the result of combining a root with functional elements of the nominal, verbal, and adjectival type, respectively (see also Borer 2000). If this holds without qualification, we would expect any root to in principle be able to surface as a noun, an adjective, or a verb, without any systematic difference in meaning. It should then be possible to derive verbs meaning ‘be Root’ from the roots in Table 2 as well from the roots in Table 1, for example. The fact that this is not what we get suggests that roots are differentiated even before they are inserted into syntactic structures. In fact, Marantz (1997) acknowledges this when he points out that in nominalizations, roots like destroy behave differently from roots like grow, in that roots of the former class may include an agent in their nominalizations, while this is impossible for roots of the latter class. Marantz attributes the difference to the semantics of the roots: destroy denotes an externally caused change of state, while grow denotes an internally caused change of state. This difference has consequences for the interpretation of a possessor occurring with the nominalization: only if the root belongs to the destroy class can the possessor be interpreted as a causer (compare John’s destruction of the tomatoes to *John’s growth of tomatoes). Similarly, the basis of the contrast between the two classes of verbalizations can be analyzed as a semantic distinction between roots of different types. But note that saying that roots are of different semantic types does not amount to claiming that the syntactic categories as such are given in the lexicon. Rather, a state-denoting root becomes an adjective in the syntactic sense only when appropriate functional categories are added, and an entity-denoting root becomes a nominal in the syntactic sense only after appropriate functional categories are added. For North Saami, this is transparently true. Consider, for example, the constructions with adjectives given in (1), where we see that some marker must be added to the adjectival root whenever it is to appear in a syntactic structure. Likewise, a verbal root has to combine with verbal functional elements to become a verb in the syntactic sense. In North Saami, these elements are arguably visible. All verbs in this language, with the exception of leat ‘be’, have at least two syllables. In (8) I give some non-derived disyllabic verbs as an illustration.

ROOTS AND VERBS

(8)

147

% a. barg-a-t ‘work’ b. skuhrr-a-t ‘snore’ c. bahkk-e-t ‘push oneself forward’

d. e. f.

fill-e-t ‘deceive’ átn-u-t ‘ask, beg’ skierbm-u-t ‘limp’

For all these verbs, it can be argued that the vowel in the second syllable, the ‘theme’ vowel, is not part of the root nor of the infinitival marker. That it is not a part of the infinitival marker is seen from the fact that it is retained in the so-called ‘connegative’ form of the verb, that is, the form that appears with the negation. Modulo possible gradation effects, the connegative is identical to the verb stem. Now consider (9), where I show the infinitive and the connegative of bargat ‘work’ and muitalit ‘tell’. As we see, compared to the infinitive bargat drops only the final -t in the connegative, while muitalit drops -it. Hence, these elements must be the infinitival markers. (9)

a. bargat ‘work’ b. muitalit ‘tell’

bargga muital

As for the claim that the theme vowels in (8) do not belong to the root either, one can for example compare the verb bargat ‘work’ to the noun bargu ‘work’. As we see, the second a in bargat contrasts with an u in bargu. Thus, one could take these vowels to represent a verbal and a nominal head, respectively, so that the root is just barg–. More generally, the theme vowels could be seen as the phonological realizations of verbalizing and nominalizing heads. Nevertheless, while relatively little syntactic machinery is required when a root is used in what we might call the prototypical way (see Josefsson 1998) — for example, when a state-denoting root functions as an adjective — it takes more for a state-denoting root to function as a noun, or for an entity-denoting root to function as an adjective. In North Saami, a visible derivational element must be added in either case. This element gives a change of semantic type, even if it does not alter the core meaning of the root. Some examples are shown in Table 5 and Table 6. I take this to mean that each of the semantic types that roots fall in is closely related to one particular syntactic category, so that a state-denoting root, for example, can be inserted directly into an adjectival syntactic environment, but it must be manipulated in certain ways before it can be inserted into a nominal or verbal syntactic frame.

148

MARIT JULIEN

a. b. c. d. e.

buorre guhkki jállu okto nuorra

‘good’ ‘long’ ‘bold’ ‘alone’ ‘young’

> > > > >

burro-dat guhkko-dat jállo-dat okto-vuohta nuorra-vuohta

‘goodness’ ‘length’ ‘boldness’ ‘loneliness’ ‘youth’

Table 5: Nouns derived from adjectives

a. b. c. d. e. f.

boallu geaągi vuodja fápmu mánná dárbbaš

‘button’ ‘stone’ ‘fat’ ‘power’ ‘child’ ‘need’

> > > > > >

boallu-i geaągá-i vuodjá-i fámo-laš máná-laš dárbbaš-laš

‘having many buttons’ ‘stony’ ‘fatty’ ‘powerful’ ‘childish’ ‘necessary’

Table 6: Adjectives derived from nouns Note that even in English, it is not the case that any root can freely appear as a noun, a verb or an adjective. Instead, some roots, like saddle, appear either as nouns or verbs, while others, like dry, appear as adjectives or verbs (the noun dry is very specialized in meaning and clearly derived). This suggests that adjectival and nominal roots can sometimes be verbalized by means of a phonologically empty verbalizer (in other cases the verbalizer is visible, as in brighten), but beyond phonology, roots are not more flexible in English than in North Saami. 5. The proposal in Baker (2002, 2003) The observation that deadjectival verbs allow the root to be predicated of the subject while denominal verbs do not is discussed in Baker (2002, 2003). Baker relates the contrast to the proposal that nouns are characterized by having a referential index, even when they appear as predicates or form the base of a derived verb. As evidence for his claim that nouns always have a referential index, Baker points to the English constructions in (10), which he takes to show that even as a predicate, a noun can establish a discourse referent, while a predicate adjective cannot (examples from Baker 2002:30):10 10 But as Peter Svenonius points out to me, the following is acceptable: ‘We are industrious. It keeps us warm.’

ROOTS AND VERBS

(10)

149

a.

In the winter, Merlin is a wolf. It (the wolf) has a brown coat and sharp teeeth.

b.

We are industrious. #It (being industrious) helps John, too.

The characteristic property of a verb, according to Baker, is that it assigns a theta role to its specifier. An adjective, by contrast, has no referential index and assigns no theta role. He then explains the contrast between deadjectival and denominal verbs as follows. When an adjective is incorporated into a verb meaning ‘be’ or ‘become’, the adjective and the verb are fused into one node, because of the poor lexical content of the verb. Due to the verbal component, the resulting node assigns a theta role to a DP in its Spec, which yields a construction where the adjectival meaning is predicated of that DP. This is illustrated in (11a). (11)

a.

VP VP

DPi V A+be b.

*

AP A

VP VP

DPj V Ni +be c.

NP Ni

VP VP

DPi V Nj

NP V

Nj

150

MARIT JULIEN

In the (hypothetical) case of a denominal verb, shown in (11b), the node consisting of the noun and the verbal element will have a referential index, which comes from the noun, and it will also have a theta role to assign, because of its verbal part. But this is impossible, according to Baker, who postulates that no syntactic node can bear a referential index and at the same time have a theta role to assign. A noun can incorporate into a verb only if the two remain separate nodes, as in (11c). But what we get then is an instance of incorporation of a nominal object into a full lexical verb. Baker further argues that verbalizing a predicative noun requires that its referential index is suppressed, so that we have a situation like (11a) rather than the one in (11b). That is, the noun will get an adjective-like reading. He goes on to suggest that in English, certain verbs in -ize are formed in this way, such as crystallize. As support for the claim that crystallize involves suppression of the referential index of the nominal base, Baker points to the following contrast in English. While the nominal predicate in (12a) establishes a discourse referent, the verbalized noun in (12b) does not (Baker 2002:30). (12)

a.

The solution became a crystal. It was two inches long...

b.

The solution crystallized. # It (the crystal) was two inches long...

However, the contrast between (12a) and (12b) is not necessarily due to the presence or absence of verbalization. Consider the Norwegian examples in (13). We see here that singular nominals used as predicates in Norwegian can have a determiner, as in (13a), or appear without one, as in (13b). Crucially, only when there is a determiner does the nominal have an individual reference that can be picked up by a pronoun. (13)

a.

Den har brun pels Om vinteren er Merlin ein ulv. during winter is M. a wolf. it. has brown fur og skarpe tenner. and sharp teeth “In the winter, Merlin is a wolf. It has brown fur and sharp teeth.”

b.

Om vinteren er Merlin ulv. # Den har brun pels og during winter is M. wolf. it. has brown fur and skarpe tenner. sharp teeth “In the winter, Merlin is wolf. It has brown fur and sharp teeth.”

ROOTS AND VERBS

151

Moreover, the same contrast that we see in (13) is also found with argument nominals in Norwegian, as (14) demonstrates: (14)

a.

Den hadde brun pels og skarpe tenner. Vi såg ein ulv. we saw a wolf. it. had brown fur and sharp teeth “We saw a wolf. It had brown fur and sharp teeth.”

b.

Vi såg ulv. hadde brun pels og skarpe # Den/dei we saw wolf. it. /they had brown fur and sharp tenner. teeth. “We saw a wolf/wolves yesterday. # It. /they had brown fur and sharp teeth”

The problem with the second clause in (14b) is that the nominal ulv ‘wolf’ in the preceding clause does not have a quantified reference. What the first clause means is that we saw at least one individual of the species, but possibly several. That is, the determinerless singular nominal focuses the type (see Borthen 2003 for a detailed discussion of determinerless singular nominals in Norwegian). This is why it does not introduce an individuated discourse referent.11 But then it could also be the absence of a determiner, or of nominal functional elements altogether, that makes crystal in crystallize a bad antecedent for it in (12b). It does not follow that the verbalization in itself precludes the use of the pronoun. Thus, this part of Baker’s argumentation is not entirely convincing. Another problematic aspect of Baker’s analysis is the idea that an adjective or noun that combines with a verbalizing element fuses with that element so that a single syntactic node results. Baker makes use of this idea to explain why adjectives yield an acceptable result but nouns do not. However, we have seen that in North Saami, nominal roots sometimes combine with the same verbalizers that adjectival roots combine with, and the result is clearly not as in (11b), since we get a grammatical verb. Moreover, denominal verbs in North Saami are not like (11a) either, since the nominal roots do not get a predicative reading. Thus, although Baker mentions the existence of denominal verbs of the types shown in Tables 2 and 4, it is not clear how they are derived in his system. Further, if the fusion that Baker proposes did occur, we might expect the resulting node to be spelled out as a unit, with the nominal or adjectival root 11

(i)

Bare nominals are much better as antecedents of type anaphors (see Borthen 2003): Vi såg ulv. Det såg dei andre òg. we saw wolf. it. saw others too “We saw one or more wolves. That the others saw too.”

152

MARIT JULIEN

morphologically inseparable from the verbalizing element. But in Baker’s own examples, the verbalizer is morphologically separate from the root. And as we have seen, in North Saami deadjectival and denominal verbs the verbalizers can still be identified. Finally, Baker suggests that the phonological weight of the verbalizer might play a role, such that a phonologically heavy verbalizer is less likely to fuse syntactically with the base than a phonologically light verbalizer. Now the North Saami verbalizers -duvva-, -huvva-, and -šuvva-, shown in Table 4, are of wordsize phonologically. If Baker’s suggestion is correct, these verbalizers ought to retain their independence vis-à-vis the root, and to give verbs meaning ‘become Root’. But as we have seen, they do not. In the next section, I will present a slightly different analysis of deadjectival and denominal verbalizations, which makes room for the Saami denominal verbs that I have presented above. My analysis retains from Baker’s analysis the idea that the semantics of the derivational base matters in word formation, but it deviates from it in certain other respects. 6. Denominal and deadjectival stative and inchoative verbs As I have already argued, roots with substantial content can be of different semantic types. I suggested in section 4 that nominal roots denote entities, while adjectival roots denote states. This needs some refinement, since the examples in (13b) and (14b) demonstrated that a nominal root in itself does not denote an entity, at least not in Norwegian. Still, it can be argued that nominal roots are of the semantic type <e>, as Chierchia (1998) claims to be true for some languages, and Baker (2003) takes to hold for all languages. More precisely, nominal roots denote kinds, that is, classes of entities. When a nominal root appears in a nominal phrase, its reference will be quantified and possibly specified by functional elements such as determiners and quantifiers (cf. Zamparelli 2000; Longobardi 2001). Only then can the nominal be used to refer to an individual. For simplicity, I will however continue to refer to nominal roots as entity-denoting. Now let us consider the derivation of stative verbs, starting with deadjectival stative verbs, for which I have already proposed that they have the syntactic structure shown in (15a). I assume that VS requires a state-denoting element as its complement, and that VS together with its complement also denotes a state. Thus, when the complement of VS is a state-denoting (i.e. adjectival) root, as in (15a), no problems arise. However, if VS has an entity-denoting (i.e. nominal) root as its complement, the result is ungrammatical (15b). Since the root is of the wrong semantic type, the combination of VS with that root cannot be interpreted.

ROOTS AND VERBS

(15)

VS P

a.

VS P

DP

RootA

VS b.

*

VS P VS P

DP

RootN

VS c.

153

*

VS P VS P

DP

PP

VS P

RootN

One way to save the derivation is to merge the entity-denoting root with a P element, as in (15c).12 Now VS has a complement of the right type, and a state reading results.13 But because of the presence of the P, the root is not directly predicated of the subject of the state. That is, instead of the meaning ‘be Root’, we get ‘be P Root’, where ‘be’ plus P may come out as ‘have’ (Kayne 1993), or else the exact relation that the P represents will be determined in a pragmatic fashion (see, for example, the verbs in Table 4). Exactly the same reasoning applies to denominal inchoative verbs: there is no problem with forming a complex predicate from a state-denoting root and a 12 The representation of the PP in (15c) is somewhat simplified. The actual PP structure is probably more complex, and it is also possible that the DP subject originates inside the PP (see Svenonius 2003). 13 A similar effect is seen with resultatives in English. An adjective can be used as a resultative predicate directly, as in (ia), but a noun cannot, not even in cases where the intended meaning should be clear enough, as in (ib). As we see, a noun requires a preposition with it in order to be a resultative predicate. (i) a. I hammered the metal flat. b. I hammered the metal *(into) a saucer.

154

MARIT JULIEN

verbalizer, so that the combined meaning of these elements comes out as ‘become Adjective’. But if the root is entity-denoting, the resulting structure cannot mean ‘become Noun’. Instead we get a reading that can be rendered as ‘become P Noun’, as we have seen. Hence, I take the inchoative verbalizers to be like stative verbalizers in that they require a complement of state-denoting type, and I further conclude that a nominal root must normally combine with a P before it can be verbalized. That denominal verbs involve an abstract P was originally proposed by Hale & Keyser (1993), mainly on the basis of English verbs like saddle, whose meaning appears to involve a prepositional element (cf. (7a)). Now while the P element is invisible in English, it is sometimes detectable in North Saami. Consider the denominal verbs dáluiduvvat (Table 4e) and vealggáiduvvat (Table 4g), formed from the same roots as dállu ‘farm’ and vealgi ‘debt’. The -i- that follows the root in the denominal verbs is not a part of the root, and arguably not a part of the verbalizer either. In fact, it is identical to the marker of illative case, which can be analyzed as a reflex of a P. The regular illative singular forms of dállu and vealgi are dállui and vealgái, and we see that the only difference between the forms used in verbalization and the regular illative singular forms is that the stem consonants appear in the weak grade in the former but in the strong grade in the latter. Historically, consonant gradation in North Saami was determined by the phonology, such that the weak grade would appear before a closed syllable and the strong grade would appear before an open syllable (see, e.g., Sammallahti 1998). The loss of many syllable-final consonants has later complicated the system considerably, to the extent that the consonant gradation now seems to be conditioned more by the morphology than by the phonology. I am therefore not sure what to make of the observed grade contrast, but I nevertheless believe that associating the -i- with the P element is not entirely without justification, especially since the -i- in question is only found with nominal roots (see Nielsen 1979[1926]:274-275).14 But note that, since the verbalized nominal roots do not project nominal phrases, they do not establish discourse referents. For example, the verb dáluiduvvat in Table 4e does not involve reference to an individual farm. Conse14

In the framework of Hale & Keyser (2002), verbs like saddle involve a DP generated in the Spec of the preposition. This DP, which appears because the preposition requires it, surfaces as the object of the derived verb, while the surface subject is truly external and supplied by a higher verbal projection. But Hale & Keyser also mention cases of ‘delayed gratification’, as they call it, of the specifier requirement of the preposition. This is exactly what we see in (15c).

ROOTS AND VERBS

"

‘winter’ ‘summer’ ‘fine weather’

> dálv-a-t > geass-u-t > firt-e-t

d. gállu

‘cold weather’

>

e.

mosku

‘fog in the dark’ >

f.

duottar

‘bare mountain’

a. b. c.

dálvi geassi fiertu

g. šalka

155

‘firm track in ’ snow’ h. cˇ ahki ‘hard lump of snow’ i. geaągi ‘stone’ j. hilla ‘ember’ k. olmmái ‘man’

> > > > > >

‘turn winter’ ‘turn summer’ ‘become fine weather’ gálu-d-it ‘become cold weather’ moskku-d-it ‘become foggy in the dark’ duottar-duvva-t ‘become bare mountain’ šalka-luvva-t ‘become firm track in snow’ cˇ ahki-luvva-t ‘form into lumps (of snow)’ geaąg-u-t ‘turn into stone’ hilla-luvva-t ‘turn into embers’ olmmái-duvva-t ‘become a man’

Table 7: Inchoative verbs derived from nouns quently, following the verb dáluiduvvat, the farm can only be referred to by means of a full non-pronominal DP, as shown in (16). (16)

{#Dat/ OK dat dállu} lei oalle stuoris. Dalle son dálu-i-duva-i. that farm was quite big then he/she farm-i-verb- it “Then he/she ‘farm-got’. *It/OK that farm was quite big.”

It is now time to reveal that North Saami does in fact have a few denominal inchoative verbs that mean ‘become Noun’. Some examples are given in Table 7. If my account of the verbs meaning ‘become P Noun’ is correct, the existence of the verbs in Table 7 is surprising. It is striking, though, that the majority of verbs of this type are based on roots that refer to meteorological phenomena, as in the examples Table 7a–e, or to the terrain, as in 7f, or to the conditions for travelling, as in 7g and 7h. As Josefsson (1998) points out, some roots denote concepts that do not clearly belong to one or the other ontological category — that is, their semantic

156

MARIT JULIEN

class membership is somewhat vague — and consequently, the grammatical behaviour of these roots is compatible with more than one semantic type. For the roots that we see in Table 7a-h, it does not seem far-fetched to claim that they can be conceptualized as entities, which gives rise to the nouns seen in the left-hand column, or as states, which allows them to combine directly with an inchoative verbalizer and yield a verb which will be used to predicate the state of a subject (which in the case of a meteorological verb will be phonologically null.) For example, Nielsen (1979[1932–1962]:584) says that duottarduvvat (in Table 7f) means ‘assume the character of bare mountain’. I take this to mean that in this verb, the root duottar is seen from a state perspective rather than from an entity perspective.15 For the verb in Table 7k, on the other hand, the analysis I have just sketched may seem less plausible. However, an informant tells me that the precise meaning of olmmáiduvvat is ‘come to be like a man’. Hence, it is arguably the case even here that the root is conceptualized as state-denoting, although it normally would be entity-denoting. It should be noted that very few inchoative denominal verbs in North Saami do not describe meteorological phenomena or other aspects of the natural environment. Thus, there are not many verbs like olmmáiduvvat in the language. They cannot be freely formed, in contrast to the denominal verbs involving a P.16 7. Causative verbs We will now move on to look at North Saami derived verbs of an even more complex type, namely, causative verbs, of which there are two main types. The ‘syntactic’ causative, which has been analyzed quite extensively by Julien (1996) and Vinka (2002), will be left out of the discussion here. I will focus instead on the so-called ‘lexical’ causatives, that is, causatives that do not embed agents or other external arguments of the base verb. What these verbs represent semantically is a change of state on the part of an argument and a causing of the change of state (see Parsons 1990; Pylkkänen 2002).17 15

Note that my analysis of these cases capitalizes on the fuzziness that Baker (2003) tries to get rid of. 16 The following authentic (American) English example, on the other hand, appears to be a productively formed exception to the generalization that a denominal verb cannot mean ‘be/become Noun’: (i) He has not grandfathered me yet. (said by a man of his son, meaning ‘He has not made me a grandfather yet.’) 17 Pylkkänen (2000) shows that in addition to non-stative causatives, Finnish has stative causative verbs where the Cause element is interpreted as a perception eventuality, as in the

157

ROOTS AND VERBS

a. b. c. d. e. f.

buorre cˇ uorbi guhkki hálbi liekkas stuoris

‘good’ ‘clumsy’ ‘long’ ‘cheap’ ‘warm’ ‘big’

> > > > > >

buori-d-it cˇ uorbbi-d-it guhki-d-it hálbbi-d-it liekka-d-it stuori-d-it

‘improveT RAN S ’ ‘make clumsy’ ‘lengthen’ ‘make cheap(er)’ ‘warm up, heat’ ‘enlarge’

Table 8: Causative verbs derived from adjectives In North Saami, many ‘lexical’ causative verbs involve a suffixed -d-, as illustrated in Table 8. According to Vinka (2002), these and other ‘lexical’ causatives in North Saami are formed by replacing the inchoative verbalizer with a causative verbalizer, called Cause. In other words, he takes them to belong to the class that Pylkkänen (2002) refers to as root-selecting causatives. Vinka further argues, like Pylkkänen, that Cause is separate from Voice, the element that introduces the external argument. Thus, Vinka proposes that the causatives in Table 8 have the syntactic structure shown in (17), where the higher DP is the causer and the lower DP is the undergoer. (17)

VoiceP DP

VoiceP Voice

CauseP DP

CauseP Cause

Root

However, the claim that causatives can be formed directly from (adjectival) roots is problematic from a semantic point of view. Causatives built from adjectival roots do not mean ‘cause Adjective’; they mean ‘cause to be/ become Adjective’. Hence, it seems that the Cause element must have an eventive complement, so following: (i) Hyttyset inho-tta-vat Mikko-a. mosquitos. find.disgusting- M.- ‘Mosquitoes disgust Mikko.’ It is not clear whether Saami has verbs of this type, and in any case, I am leaving them out of the discussion.

158

MARIT JULIEN

a. bihkka b. ruhta c. vearju

‘tar’ > ‘money’ > ‘weapon’ >

bihkka-d-it ‘cover with tar’ ruhta-d-it ‘finance’ vearju-d-it ‘arm’

Table 9: Causative verbs derived from nouns

a. b. c.

buori-d-it ‘improveT RAN S ’ (Table 8a) liekka-d-it ‘warm up’ (Table 8e) ruhta-d-it ‘finance’ (Table 9b)

>

buori-d-uvvo-t

>

liekka-d-uvvo-t ‘be warmed up’

>

ruhta-d-uvvo-t

‘be improved’

‘be financed’

Table 10: Passives derived from causative verbs that there is a caused event and a causing event (see Parsons 1990; Pylkkänen 2002). Since a bare adjectival root does not represent an event, it is surprising that a causative should be able to take a bare adjectival root as its complement. To get the whole picture we need to consider the fact that causative verbs can also be formed from nominal roots. I show some examples of this in Table 9. As for the causative element itself, Vinka (2002) is clearly right in saying that it is not identical to the Voice head. If a causative verb is passivized, the causative marker -d- is retained and appears inside the passive marker: If the North Saami passive marker is the realization of a [−active] Voice head (Julien 1996; Vinka 2002), it follows that the causativizer is separate from and below the Voice head. Another interesting observation is that neither the verbs in Table 8 nor the verbs in Table 9 appear to encode a resulting state. I show this below with liekkadit ‘warm up, heat’, which was given in Table 8e, and ruhtadit ‘finance’, which was given in Table 9b. (18)

cˇ ázi { njeallje minuhta Liekka-d-in warm- - .& water. four. minute. minuhta-s}. minute- “I heated the water for four minutes/ *in four minutes.”

/* njealje four.

159

ROOTS AND VERBS

(19)

ruhta-d-ii prošeavtta guokte Gielda local.council. money- - .3 project. two. jagi. year. “The local council financed the project for two years.”

Given that inchoative verbs can encode a resulting state, the judgements in (18) and (19) are surprising. In (18), it ought to be possible to infer an upper bound to the heating up of the water: either the boiling point or some other pre-determined temperature. In (19), we might expect it to be possible for the PP to be interpreted as representing the result state. My proposal is that in ‘lexical’ causative verbs in North Saami, the causative element has a non-stative verbal complement, and because of this, adverbials that combine with these verbs do not have access to any resulting state. More specifically, I take the verbs in Table 8 to have the syntactic structure shown in (20a), while the verbs in Table 9 have the syntactic structure shown in (20b). In either case, VD stands for a non-stative verbal head. In (20a) this head is interpreted as ‘become’, and the whole verb means ‘cause to become Adjective’, while in (20b), VD and P together come out as ‘get’ (‘get’ being the non-stative counterpart of ‘have’), and the meaning of the whole verb is ‘cause to get Noun’. (20)

a.

VoiceP DP

VoiceP Voice

CauseP VD P

Cause

VD P

DP VD

Root

160

MARIT JULIEN

b.

VoiceP DP

VoiceP Voice

CauseP VD P

Cause

VD P

DP

PP

VD P

RootN

That the Cause head can have a non-stative verbal complement in North Saami is also seen from verbs like those in Table 11, which are causativizations of non-stative non-derived unaccusative verbs.

- a. dušša-t ‘perish’ b. hávka-t ‘drownIN T R’ c. šadda-t ‘growIN T R ’

> > >

dušša-d-it ‘annihilate’ hávka-d-it ‘drownT RAN S ’ šadda-d-it ‘growT RAN S ’

Table 11: Causative verbs derived from unaccusative verbs I take these verbs to have a syntax like (20a), only with a verbal root at the bottom instead of an adjectival root. In North Saami, it is however also possible, from adjectival roots, to derive causative verbs where the -d- causativizer appears outside an inchoative verbalizer. I show some examples of this in Table 12. Note that some roots appear both in Table 8 and in Table 12. For example, we have guhkidit in Table 8c and guhkkudit in Table 12c, stuoridit in Table 8f and stuorrudit in Table 12e. For each pair, there is a subtle meaning difference between the two verbs. The meaning of guhkidit is given as ‘lengthen’ (Nielsen 1979[1932–1962]:Vol. 2, 198), while the meaning of guhkkudit is ‘cause to get longer’ (Nielsen 1979[1932–1962]:Vol. 2, 200). Similarly, stuoridit means ‘enlarge’ (Nielsen 1979[1932–1962]:Vol. 3, 571), while stuorrudit means ‘cause to get bigger’ (Nielsen 1979[1932–1962]:Vol. 3, 572). Thus, both the morphology and the semantics suggest that that the verbs in Table 12 are formed by adding a

161

ROOTS AND VERBS

" a. b. c. d. e.

dipm-a-t (Table 3a) goik-a-t (Table 3b) guhkk-u-t (Table 3d) loss-u-t (Table 3e) stuorr-u-t (Table 3f)

‘softenIN T R ’

> dipm-a-d-it

‘dryIN T R ’

> goik-a-d-it

‘get long(er)

> guhkk-u-d-it

‘get heavy’

> loss-u-d-it

‘get big(ger)’

‘softenT RAN S , make softer, soak to soften’ ‘dryT RAN S , cause to dry’ ‘cause to get longer’

‘cause to be heavy/heavier’ > stuorr-u-d-it ‘cause to get bigger, cause to grow’

Table 12: Causative verbs derived from inchoative verbs Cause and a Voice head on top of the inchoative structure shown in (5). That is, their syntactic structure is as sketched in (21). (21)

VoiceP DP

VoiceP Voice

CauseP VD P

Cause

VD P

DP

VS P

VD DP

VS P VS

RootA

In other words, it seems that ‘lexical’ causative verbs in North Saami can have a considerably more complex syntax than what Vinka (2002) claims. 8. Marantz (1997) again If we now return to Marantz (1997), it seems to me that my claim that roots are of different semantic types is not necessarily in conflict with Marantz’s argu-

162

MARIT JULIEN

mentation. Marantz’s core observation is that external arguments are not syntactically included in nominalizations, hence the ungrammaticality of (22). (22)

* John’s growth of tomatoes.

The explanation that Marantz offers is that the root has no category, and that when it appears as a transitive verb, it is combined with a v head that verbalizes the root and also introduces the external argument in its Spec, as sketched in (23a). An unaccusative verb would involve a v head that does not allow an external argument to appear in its Spec. But when the root is inserted in a nominal environment, as in (23b), there is no v and consequently no position where an external agent could appear unless the existence of that agent is implied by the root itself.18 Since grow implies internal causation, growth does not allow an external agent to be realized. (23)

a.

v Ext.arg.

v v

Root Root

b.

Int.arg.

n n

Root Root

Int.arg.

If a root of the destroy type is nominalized, on the other hand, the possessor can be interpreted as the agent, since the semantics of the root itself implies the presence of an agent. Thus, the analysis in Marantz (1997) gives the desired result concerning the distribution of external agents in nominalizations. But as far as I can see, so could an analysis based on the view that roots are specified for category, or at least that the semantic type of the root matters for word formation. After Rivero (1990), Travis (1992), and Kratzer (1996), the idea that the external argument is introduced by a syntactic head distinct from V has become a fairly standard assumption. It follows from this assumption 18 In (23b) I have replaced Marantz’s D with n since I believe that all nominal phrases contain several functional projections below D (see Julien 2005).

163

ROOTS AND VERBS

that the agent-introducing head, that is Voice, must be above VD in non-stative transitive verbs and above VS in stative transitive verbs. That is, verbs with external arguments involve at least the structure shown in (24a), where V stands for VS or VD . (24)

a.

VoiceP Ext.arg.

VoiceP Voice

VP (Int.arg.)

VP V

b.

NP N

VP Int.arg.

VP V

To keep the external argument out of nominalized verbs, all we need to assume in addition is that the nominalizer combines with a constituent that is smaller than VoiceP. In the spirit of Alexiadou (2001) we could say that eventive nominalizations like destruction and growth contain a VD projection, and that the nominalizer combines with VD P. Result state nominals, such as break, involve a VS head. In short, nominalizations minimally have the structure in (24b).19 In this structure, V can stand for a verbal root, or it could stand for a verbalizer that has a root of some other class as its complement. 9. Conclusions Firstly, the preceding discussion shows that derived stative, inchoative, and causative verbs in North Saami can successfully be analyzed within a framework based on syntactic decomposition of complex events. Secondly, the properties of these verbs indicate that roots belong to different semantic classes, corresponding 19 In -ing nominalizations, which allow much more verbally related material inside them, including agents, the nominalizer combines at least with a VoiceP (see van Hout & Roeper 1998).

164

MARIT JULIEN

to syntactic categories, and that the semantic class membership of the root can have syntactic and semantic consequences in verbalization. References Alexiadou, Artemis. 2001. Functional Structure in Nominals. Amsterdam and Philadelphia: John Benjamins. Arad, Maya. 1998. VP-structure and the Syntax-Lexicon Interface. Ph.D. thesis, University College London. Baker, Mark C. 2002. “On category asymmetries in derivational morphology”. Morphology 2000 ed. by Sabrina Bendjaballah, Wolfgang U. Dressler, Oskar E. Pfeiffer & Maria D. Voeikova, 17–36. Amsterdam and Philadelphia: John Benjamins. . 2003. Lexical Categories. Cambridge: Cambridge University Press. Borer, Hagit. 2000. “Exo-skeletal vs. endo-skeletal explanations: Syntactic projections and the lexicon”. Unpublished ms., University of Southern California. Borthen, Kaia. 2003. Norwegian Bare Singulars. Ph.D. thesis, NTNU, Trondheim. Chierchia, Gennaro. 1998. “Reference to kinds across languages”. Natural Language Semantics 6.339–405. Hale, Kenneth & Samuel Jay Keyser. 1993. “On argument structure and the lexical expression of syntactic relations”. The View from Building 20 ed. by Kenneth Hale & Samuel Jay Keyser, 53–109. Cambridge, Mass.: MIT Press. . 1998. “The basic elements of argument structure”. MIT Working Papers in Linguistics 32.73–118. . 2002. Prolegomena to a Theory of Argument Structure. Cambridge, Mass.: MIT Press. Hay, Jennifer, Christopher Kennedy & Beth Levin. 1999. “Scalar structure underlies telicity in ‘Degree Achievements”’. Proceedings of SALT IX ed. by Tanya Matthews & Devon Strolovitch. 127–144. Ithaca, New York: CLC Publications. Josefsson, Gunlög. 1998. Minimal Words in a Minimal Syntax: Word formation in Swedish. Amsterdam and Philadelphia: John Benjamins. Julien, Marit. 1996. Syntactic Word Formation in Northern Sámi. Oslo: Novus. . 2005. Nominal Phrases from a Scandinavian Perspective. Amsterdam and Philadelphia: John Benjamins. Kayne, Richard. 1993. “Toward a modular theory of auxiliary selection”. Studia Linguistica 47.3–31.

ROOTS AND VERBS

165

Korhonen, Mikko. 1981. Johdatus lapin kielen historiaan [ = Introduction to the History of the Saami Language]. Helsinki: Suomalaisen Kirjallisuuden Seura. Kratzer, Angelika. 1996. “Severing the external argument from its verb”. Phrase Structure and the Lexicon ed. by Johan Rooryck & Laurie Zaring, 109–137. Dordrecht: Kluwer. Longobardi, Giuseppe. 2001. “How comparative is semantics? A unified parametric theory of bare nouns and proper names”. Natural Language Semantics 9.335–369. Marantz, Alec. 1997. “No escape from syntax”. University of Pennsylvania Working Papers in Linguistics 4:2.201–225. . 2001. “Words”. Paper presented at the West Coast Conference of Formal Linguistics 20, University of Southern California, Los Angeles, February 2001. Nickel, Klaus Peter. 1990. Samisk grammatikk [=Saami grammar]. Oslo: Universitetsforlaget. Nielsen, Konrad. 1979[1926]. Lærebok i lappisk (samisk), vol. 1, Grammatikk [Lapp (Saami) Primer, vol. 1, Grammar]. Oslo: Universitetsforlaget. . 1979[1932–1962]. Lappisk ordbok: grunnet på dialektene i Polmak, Karasjok og Kautokeino [ = Lapp Dictionary: Based on the dialects of Polmak, Karasjok and Kautokeino]. (= Instituttet for sammenlignende kulturforskning, serie B 17.) Oslo: Aschehoug. 5 volumes. Co-author for vols. 4-5: Asbjørn Nesheim. Parsons, Terence. 1990. Events in the Semantics of English. Cambridge, Mass.: MIT Press. Pylkkänen, Liina. 2000. “On stativity and causation”. In Tenny & Pustejovsky 2000, 417–444. . 2002. Introducing Arguments. Ph.D. thesis, MIT. Ramchand, Gillian. 2002a. “Aktionsart, L-syntax and selection”. Proceedings of the Perspectives on Aspect Conference ed. by Henk J. Verkuyl, 1–15. Utrecht: OTS. . 2002b. “A particle theory of light verbs”. Paper presented at the Workshop of Complex Predicates, Particles, and Subevents, University of Konstanz, October 2002. Rivero, María-Luisa. 1990. “The location of nonactive voice in Albanian and Modern Greek”. Linguistic Inquiry 21.135–146. Sammallahti, Pekka. 1998. The Saami Languages: An introduction. Karasjok, Norway: Dávvi Girji.

166

MARIT JULIEN

. 2002. North Saami Resource Dictionary. Oulu: Giellagas Institute. Sara, Laila Susanne. 2002. Suorggádusaid geavaheapmi: Guovdageainnu nuoraid giellageavaheapmi [=The use of derivations: The language use of young people in Guovdageaidnu]. Master’s thesis, University of Tromsø. Svenonius, Peter. 2003. “Limits on P: Falling in holes vs. filling in holes”. Nordlyd 31:2.431–445. Tenny, Carol & James Pustejovsky, eds. 2000. Events as grammatical objects. Stanford, Calif.: CSLI Publications. Travis, Lisa. 1992. “Inner Aspect and the structure of VP”. Cahiers de Linguistique de l’UQAM 1.130–146. . 2000. “Event structure in syntax”. In Tenny & Pustejovsky 2000, 145–185. van Hout, Angeliek & Thomas Roeper. 1998. “Events and aspectual structure in derivational morphology”. MIT Working Papers in Linguistics 32.175–200. Vinka, Mikael. 2002. Causativization in North Sámi. Ph.D. thesis, McGill University. Zamparelli, Roberto. 2000. Layers in the Determiner Phrase. New York: Garland.

THE INSTABILITY OF SYSTEMS WITH TERNARY LENGTH DISTINCTIONS∗ THE SKOLT SAAMI EVIDENCE

ZITA MCROBBIE-UTASI Simon Fraser University Research on the issue of phonological systems with ternary length distinction has not yet provided answers to two related questions of theoretical importance. First, although there are languages with three degrees of length, why are they so few in number? Second, why do three-way distinctions appear to be transitory? This study aims at examining (i) the instability of ternary length distinctions, (ii) the development of systems with three-way length contrast into systems with a more complex ternary contrast, and (iii) the tendency to restore binary contrasts via prosodic changes. In addition to a case study of the Skolt Saami quantity system, a brief cross-linguistic survey on languages with ternary length distinction will be presented.

1. Introduction Research on the intriguing issue of phonological systems with ternary length distinction has not yet provided answers to two related questions of considerable theoretical importance: Why are languages with three degrees of length so few in number? This is an issue that is largely ignored in linguistic theories based on recognizing binary distinctions. Also, why do three-way distinctions appear to be transitory? Indeed, it has been observed that languages with this type of contrastive system (all known such systems having resulted from strikingly similar historical developments, discussed below) undergo significant changes in their prosody in the direction of re-establishing binary distinctions. This study aims at ∗

I gratefully acknowledge Ilse Lehiste’s encouragement over the course of many years, and thank her for the comments on an earlier draft of this paper. I also wish to acknowledge the many suggestions given by the anonymous reviewers. Naturally, responsibility for all claims and such errors as may exist rests with me.

168

ZITA MCROBBIE-UTASI

examining (i) the instability of ternary length distinctions, (ii) the development of systems with three-way length contrast into systems with a more complex ternary contrast (that is, in addition to duration, at least one more prosodic factor needs to be considered), and (iii) the tendency to restore binary contrasts via prosodic changes. Discussion of these issues is based on a survey and evaluation of research on languages with alleged or existing ternary length distinction, and an acoustic analysis of the Skolt Saami quantity system. It will be shown that in connection with most languages claimed to have a ternary length distinction, research has failed to provide sufficient evidence of the existence of such a system. These languages include German, four Amerindian languages, and perhaps Scottish Gaelic (see sections 3.1–3.3), as being the most frequently cited languages with regard to phonological systems with a possible ternary length distinction. It will further be discussed that those languages in which a ternary system undoubtedly does exist — Estonian and Skolt Saami will be examined here — have undergone significant changes in their prosody. It will be argued that these prosodic changes have caused the three-way length distinction to develop into a more complex ternary quantity contrast, thus providing evidence for the instability of systems with ternary length opposition. The theoretical approach of laboratory phonology employed here is similar to that utilized in the research on ternary contrasts by Lehiste (1960, 1965, 1966, 1970–1975, 1989). This research is summarized in Lehiste (1997), where she argues for the advantages of aiming at least a degree of objectivity by way of emphasizing the importance of analyzing the acoustic signal followed by subjecting the results to systematic listening tests (Lehiste 1997:12–13). In generative phonology, it has been an accepted practice to distinguish between length as a durational property of segments and quantity as a durational property of the syllable (Lass 1984). In Laver (1994:436) and Kenstowicz (1994: 44) the term quantity refers to contrastive duration, that is, when duration functions as an independent variable in the phonological system (Lehiste 1970:42). In a more recent article this approach was modified in light of extensive research over the last three decades on identifying the independent variable relevant to phonemic contrasts in Estonian (Lehiste 1997). It must be acknowledged that the relevance of prosodics other than duration has already been recognized long before the advancement in experimental phonetic techniques during the past three decades (Trubetzkoy 1936, 1969[1939]; Põldre 1937; Ravila 1962; among others). Although there is no agreement as yet concerning the phonetic realization

THE INSTABILITY OF TERNARY LENGTH DISTINCTIONS

169

of the independent variable responsible for the signalling of contrasts in Estonian (see section 3.4), recognition of the need for changing the former claim with regard to the definition of quantity is clearly evident in Lehiste’s study (1997:12). In that paper she states that identifying only one single feature is not sufficient for establishing a quantity degree. All three prosodic dimensions — duration, fundamental frequency and intensity — as well as the position of syllables within the word have to be taken into consideration. This statement is important in two respects. First, it separates instances where duration alone is responsible for the realization of phonological contrasts, from those where other prosodic factors have also to be taken into consideration to account for degree alternations relevant to the phonology of the language. Estonian has provided evidence for the recognition of this latter point, and, as will be shown later in this study, Skolt Saami also confirms the validity of viewing ternary contrasts from this more complex perspective, involving all three prosodic factors. Second, the domain of phonologically relevant degree alternations must also be considered important, for there are languages where the domain of contrastive prosodics may extend beyond the segment to the syllable or the foot (as in Estonian and Skolt Saami, where the relevant phonological units are disyllabic feet). As Lehiste has shown in several studies on Estonian quantity (for example, Lehiste 1970–1975, 1978a, 1983, 1989, 1997), pitch movement, in addition to durational differences, may play a role in perceiving different phonological quantities (see section 3.4). Recent research on Estonian quantity has also taken into account the role of accent in the realization of different quantity degrees (Eek 1983, 1987a,b,c, 1994; Eek & Help 1986, 1987; Eek & Meister 1997). Thus, determining what constitutes the independent variable in the occurrence of different quantity degrees represents somewhat of a challenge. In other words, analyzing the role of duration in quantity contrast may prove to be but the first step towards establishing the complex configuration of prosodic features relevant to distinguishing between different quantity degrees. The position adopted in the present study is that the terms length and overlength will be employed only in cases where the independent variable in the realization of prosodic opposition is duration. Thus, the term quantity will be used when the relevance of other prosodic factors (in addition to duration) need to be considered in the manifestation of contrast. Further, because the realization of prosodic oppositions cannot be understood without reference to their domain, discussion of alternating quantity contrasts (in Estonian and Skolt Saami) will proceed within the context of disyllabic metrical feet.

170

ZITA MCROBBIE-UTASI

As mentioned above, there are few known languages with three degrees of quantity distinction in their phonological systems. Furthermore, very little information exists on the phonetic correlates of the presumed three-way opposition. Lehiste (1970) summarizes the results of previous findings claiming to identify three-way contrasts in languages such as Hopi (a Uto-Aztecan language), Mixe, (a Zoquean language) and German. After discussing the Skolt Saami case, a brief survey of more recent research in connection with those languages suspected of having three degrees of length will be presented. Three additional languages will be touched upon, which are assumed to have three contrastive lengths, Wichita (a Caddoan language), Rood (1975, 1976, 2001); Sarcee (an Athapaskan language), Cook (1971, 1984), Cook (p.c.); and Scottish Gaelic, Ternes (1973). Languages with surface ternary length (such as the St. Lawrence Island dialect of Central Siberian Yupik, Hayes 1995; Krauss 1985; Jacobson 1983) will not be considered in this study of contrastive quantity. 2. The Skolt Saami case Skolt Saami is spoken by about 300–500 speakers (Sammallahti 1998). The two dialects from which the material for the present study was collected are the Suonikylä and the Paatsjoki-Petsamo dialects. The Skolt Saami people live around the Sevettijärvi community in North-East Finland close to the Norwegian border, and between the Ivalo and Nellimö areas close to the Russian border. In my research on Skolt Saami quantity, I utilized the laboratory phonology approach and techniques employed also in researching Estonian quantity (see section 3.4). The reason for this is that Skolt Saami — like Estonian — is a language in which the realization of quantity degrees requires consideration of other prosodic factors in addition to duration (Korhonen 1971, 1975). Further, it is a language in which the domain of quantity is larger than the segment (Itkonen 1946), one in which there is three-way quantity opposition (Itkonen 1946; Korhonen 1971, 1975). However, despite the apparent similarities of the quantity systems in Estonian and Skolt Saami, research on the latter has so far produced results significantly different from those on Estonian quantity with regard to the configuration of the prosodic factors manifested in the realization of quantity alternations (Korhonen 1971, 1975; McRobbie-Utasi 1999).


171

2.1 Methods and materials Two native speakers of Skolt Saami1 participated in the experiment that resulted in an acoustic analysis of quantity on the basis of speech production. The material analyzed consisted of recordings of over 3000 disyllabics from 550 test words of a quality appropriate for acoustic phonetic measurement. In selecting the test material the objective was a balanced representation of each of the structural types and sub-types. In employing the term ‘structural types’, I follow the tradition of classifying Skolt Saami disyllabics on the basis of their prosodic patterns (Itkonen 1946). The two subjects were asked to place each word in the sentence frame. They were given a group of 100 randomized test words at a time, and were requested to maintain a steady speech tempo. Each test word was recorded three times (in different recording sessions) by both speakers. The recording was made with a Scully Full-Track machine tape-recorder and re-analyzed for the purpose of this paper with the Praat program, Version 4.0.23 (Boersma & Weeninck 2003). Measurements of duration, fundamental frequency and intensity were carried out in accordance with standard procedures (McRobbie-Utasi 1999:33–35, 39–47). 2.2 Analysis of ternary quantity Research on Saami quantity, similar to that on Estonian, has to take into consideration (i) the domain, and (ii) the phonetic components of quantity. 2.2.1 The domain of quantity. Previous studies, in particular Itkonen’s dissertation on East Lappish quantity (Itkonen 1946) and Bergsland’s grammar of Røros Lappish (Bergsland 1948), have established that, in analyzing quantity in Saami, one must regard disyllabic units — that is, stress-groups — as the domain of quantity. There appear to be two arguments supporting the existence of disyllabic stress-groups in Saami. First, identical consonant segments behave differently and may or may not undergo gradation, depending on their position in the word (Korhonen 1981:149–153). This may be seen, for example, in connection with the so-called suffixal gradation. Here the occurrence of the strong versus weak grade depends on whether the suffix follows an odd- or even-numbered syllable (for example, læˇZˇZap ‘be. &’ versus mânâˇZæp ‘go. &’).2 Second, as has been recognized from as early as the beginning of the last century, 1

The two (female) subjects are speakers of the Suonikylä and dialect the Paatsjoki-Petsamo dialect, respectively. Regarding the acoustic correlates of quantity, the differences found in the material analyzed were not statistically significant (McRobbie-Utasi 1999). 2 In North Saami orthography â = IPA [O].

172

ZITA MCROBBIE-UTASI

there are interdependencies between segments of the first and second syllables (Äimä 1918). Experimental phonetic analyses were carried out by Äimä on the Inari Saami language, and his measurements were subsequently cited and served as arguments for establishing the stress-group concept, most notably by Itkonen (1939, 1946). Results of modern acoustic analyses have confirmed the findings of earlier research concerning the existence of the presumed interdependencies between segments of the first and second syllables in North Saami (Magga 1984) and in Skolt Saami (McRobbie-Utasi 1999); thus, the stress-group concept has proved to be an accepted point of departure for analyzing quantity in Saami. An acoustic analysis of duration for a Saami dialect (the Guovdageaidnu dialect of North Saami) was pertained to confirm experimentally, by employing a segmental approach, the relevance of the disyllabic stress-group through demonstrating the existence of durational interdependencies within this rhythmic unit (Magga 1984). Magga assumes the validity of the analysis in Sammallahti (1977) according to which vowels in the first syllable may occur in three quantities, intervocalic consonants may also have three quantities, and second syllabic vowels may occur in two quantities. As will be pointed out below in section 3.4, in the realization of Estonian quantity the hierarchy of segment – syllable – disyllabic foot needs to be recognized. With regard to Saami, there has been a reluctance to consider the role of the syllable in the manifestation of quantity. In fact, the syllable has not been regarded as a quantity unit at all within the disyllabic stress-group. In his influential work, Itkonen (1946:14), clearly stated that the syllable has no significance within the stress-group (Sprechtakt). Sammallahti (1977:109) states that because all structural restrictions seem to centre in syllable boundaries, syllables therefore play no role in the stress-group. While this approach did reveal durational interdependencies within the stress-group, it did not result in recognizing the possibility of a prosodic hierarchy involving the syllable as a quantity unit in Saami. As will be shown below, it is precisely these durational interdependencies that imply recognition of the syllable as a significant unit — at least the relevance of a prosodic hierarchy involving the syllable cannot be ignored in Skolt Saami. In concluding this section, it can be stated that the disyllabic stress-group concept has to be granted. Acceptance of this concept may be justified on the basis of both suffixal gradation (see above) and durational interdependencies between the relevant segments within this disyllabic unit. The latter has been supported by both earlier and more recent research (Itkonen 1946; Sammallahti 1977; Magga 1984). It appears, though, that the role of the syllable in the hi-


173

erarchy of prosodic levels remains uncertain.3 In the next section, the issue of hierarchy of prosodic units within the disyllabic stress-group (= foot) in Skolt Saami will be discussed, in the context of the acoustic analysis supporting the acceptance of a prosodic hierarchy where the syllable appears to play a significant role. 2.2.2 Hierarchy of prosodic units. Acoustic analysis of the prosodic correlates of quantity (McRobbie-Utasi 1999) confirmed the relevance of disyllabic rhythmic units in the realization of quantity alternation in Skolt Saami. Furthermore, it proposed the recognition of a hierarchy of segment, syllable and disyllabic stress-group. In particular, the relevance of the syllable in this hierarchy was argued for on the basis of analyzing the durational interdependencies apparent between the main stress-group locations (Sammallahti 1977): vowel centre (first syllabic vowel), consonant centre (intervocalic consonant or consonants) and the latus (second syllabic vowel). 2.2.2.1 Segmental quantity. In reviewing segmental quantity in Skolt Saami, I shall summarize Korhonen’s analysis (Korhonen 1971, 1975). Length is marked in accordance with the IPA convention, thus [;] and [:] indicate half-long and long segments respectively. Raised capital letters designate reduced vowels. Examples are cited in Skolt Saami orthography (Korhonen et al. 1973). The [´] indicates a more forward articulation when it follows a back vowel; when it follows a front vowel, the symbol indicates a less forward articulation of that vowel. Vowels in the vowel centre (both monophthongs and diphthongs) occur in two distinctive quantities, short and long. A short vowel is always phonetically realized as short, for example, jokk ‘river. ’. A long vowel may be realized phonetically either as long (in monosyllabic and disyllabic words always, and in polysyllabic words where the syllable following has a vowel bearing secondary stress), for example, mÂAˆńn [â:] ‘egg. ’, ää´võõllad [ä:] ‘open.’; or half-long (in words with three or more syllables where the second syllable contains a reduced vowel; that is, when the second syllable is stressless), for example mainstam [ma;inA stam] ‘tell.&’. Vowels in the latus (only monophthongs are encountered here) may occur 3

In his phonological analysis employing a metrical approach to North Saami quantity, Bye (1997:52–53) argues for Q3 as an extrasyllabic mora, differing from the Estonian in terms of the parsing status of the third quantity degree, the centre of Q3 in Estonian functioning as a syllable. Bye’s analysis therefore suggests a prosodic hierarchy different from the one presumed for Estonian (Lehiste 1965, 1997).

174

ZITA MCROBBIE-UTASI

in two quantities, short and long. The phonetic realization of short vowels in this position depends on (i) the presence of secondary stress in the syllable in question, and (ii) whether the vowel is followed by a consonant. Thus, when the syllable is stressed, the short vowel is realized phonetically as short; when the syllable is unstressed and the vowel is not followed by a consonant, the short vowel is realized phonetically as being reduced, for example, kuätta ‘Lappish tent.’, and vuâsk [vuâskA ] ‘catfish’. Vowels in unstressed syllables, when followed by a consonant, are phonetically realized as short, for example, mõõnam ‘go.&’. The long vowel is realized phonetically either as long in open syllables, for example, ää´võõlam ‘open.&’, or as half-long in open syllables when the word consists of at least two more syllables after the syllable in question, for example, ää´võõlažem ‘open. &’. Single consonants may be realized as half-long or short. They are half-long after short stressed vowels, for example, ää´ve´ked [k;] ‘open. '’; they are short after a long vowel, for example, mõõnam ‘go.&’, or after a reduced vowel between the second and third syllables, for example, tuoddri [tuoddA ri] ‘hill.’. Geminates may be realized phonetically as follows: long (after short vowels in the first syllable and before reduced vowels in the second syllable), for example, eˇccˇ a ‘father.’, kue´ss [kue´ssE ] ‘guest. ’; half-long (after short vowels in all other instances), for example, vä´ZZal [Z;] ‘walk.$’; and short (after long vowels), for example, viiv [vi:vA ] ‘son-in-law. , ’. Consonant clusters may be realized phonetically as follows: after short vowels the first member of the cluster is long and the second member half-long, for example, kuä´stte [s:t;] ‘appear.$’; after a long vowel in the first syllable or after a reduced vowel between the second and third syllables the consonant cluster is short, for example, joordak ‘think.'’, maainsted [ma;inA sted] ‘tell.’. 2.2.2.2 Duration within the stress-group. In my analysis, I distinguish between six main structural types representing different distributions of duration. In Type 1, the consonants occurring are long geminates (for example, cˇ uoppâd ‘cut.’). In Type 2, there is a long consonant cluster (for example, ju´rdded ‘think.’). Type 3 has a short consonant following the first syllabic vowel (for example, mõõni ‘go. $’). Type 4 contains a short geminate consonant. In Type 4a, the relevant consonants are liquids, nasals or voiced fricatives (for example, kue´ll ‘fish. ’); in Type 4b, the consonant is a plosive, affricate or voice-


175

Figure 1: Mean durations (¯x) of the first syllabic vowel and the consonant(s) following, illustrating a compensatory relationship between the vowel centre and the consonant centre less fricative for example, kue´ss ‘guest. , ’). In Type 5, there is a short consonant cluster (for example, vue´sti ‘buy. $’). Type 6 has a historically contracted vowel in the second syllable; this type is divided into two sub-groups: Type 6a contains geminates intervocalically and Type 6b contains consonant clusters intervocalically (for example, vä´ldde ‘take.$’). Type 6a is further subdivided into two groups: Type 6a1 has liquids, nasals or nonsibilant fricatives intervocalically (for example, jäurra ‘lake.’); Type 6a2 has plosives, affricates or sibilant fricatives in the same position (for example, i´tte ‘appear. $). For further details and justification of dividing Skolt Saami disyllabics into six structural types, see McRobbie-Utasi (1999:35–38). The distribution of duration within the disyllabic stress-group forms a complex pattern. Measurement of the main stress-group locations reveals that (i) there exists a compensatory relationship between the vowel centre and the consonant centre, and (ii) the duration of the latus is short in connection with all structural types (and sub-types). First syllabic vowels may occur in two different durations, half-long and long; the consonant(s) following the first syllabic vowel occur in three different durations: short, half-long and long. The compensatory relationship between the two main stress-group locations is illustrated in Figure 1. As can be observed there, longer durations associated with consonant(s) occur together with shorter durations associated with the first syllabic

176

ZITA MCROBBIE-UTASI

Type 1 2 3 4a 4b 5

Sums of duration 73 (V = 34; C = 39) 64 (V = 29; C = 35) 71 (V = 66; C = 5) 65 (V = 29; C = 36) 68 (V = 31; C = 37) 74 (V = 33; C = 41)

Table 1: Sums of durational increase in V1 and consonant(s) vowel in structural types 1, 2 and 6; long first syllabic vowel durations co-occur with short durations associated with the consonant following (structural type 3); again, structural types 4 and 5 have a distinct compensatory pattern between the first syllabic vowel and the consonants following. Second syllabic vowels are short in all structural types (¯x=87 msec); thus, it is the first syllabic vowel and the consonant(s) following that play a determining role in the distribution of durational patterns which distinguish between the structural types. In Skolt Saami, the word-final vowel in the second syllable may be dropped or become reduced.4 In this context only the duration of vowel reduction is relevant, though only historically non-contracted vowels undergo this optional phonological process (a fact of importance that will be made explicit below). In analyzing the consequences of the optional second syllabic vowel drop (or vowel reduction), it was observed that the sum of durational increases in the first syllabic vowel and the consonant(s) following — as a result of compensatory lengthening — is approximately that of the duration of the second syllabic vowel that is missing or became reduced in duration. The measurement values presented in Table 1 show that these durational increases are distributed nearly evenly in the first syllabic vowel and the consonant(s) following. Figure 2 summarizes the mean V/C ratios in disyllabics with a full vowel in the second syllable, and with no vowel or a vowel with reduced duration in the second syllable. The manifestations of compensatory lengthening shown in the figure reveal another fact that has significant implications for the domain of quantity in Skolt Saami. Although it has been pointed out that durational increase is realized in both the first syllabic vowel and the consonant(s) following, there is one structural type in which this pattern does not apply. If we consider the results of the durational measurements in Type 3, we observe that reduced duration of 4

The characteristics of latic vowel reduction are discussed in McRobbie-Utasi (2000, 2001).


177

Figure 2: Mean V/C ratios of disyllabics the second syllabic vowel results in compensatory lengthening in the first syllabic vowel only. There is practically no durational increase in the consonant following this vowel (see Table 1, where the increase averages only 5 msec). Consequently, even though the V/C ratios are almost identical in the other four structural types — regardless of the durational status of the second syllabic vowel — they are different in Type 3. Type 3 has a short consonant intervocalically, with an average duration of 92 msec. That there is no durational increase apparent in the onset of the second syllable implies that compensatory lengthening is manifested in the four other structural types in the first syllabic vowel, as well as in that part of the consonant that is to the left of the syllable boundary (that is, in the first syllable of the disyllabic stress-group). The implications of the above findings are two-fold: (i) the distribution of durational increases in the relevant stress-group locations indicates that, although absolute durations have changed, the ratios remain the same; (ii) the fact that durational increases are realized in the first syllable implies the relevance of the syllable as being a prosodic level within the disyllabic stress-group, thus a hierarchy of segment, syllable and disyllabic stress-group (=foot) needs to be recognized in connection with the domain of quantity in Skolt Saami. These implications concerning the consequences of the compensatory lengthening phenomenon may

178

ZITA MCROBBIE-UTASI

be interpreted as showing that the presence versus absence (or reduced status) of the second syllabic vowel in the VC(C)V sequence determines the duration of the preceding segments, providing evidence for a cohesive relationship between the first and second syllable. What remains constant is the V/C ratio which preserves the identity of the structural type. In summary, the acoustic analysis of duration, in addition to confirming the significance of the disyllabic stress-group as the basic domain of quantity, has resulted in two important findings. First, on the basis of the manifestation of duration in connection with compensatory lengthening, the analysis provides evidence for recognizing the role of durational ratios between two main stress-group locations: between the first syllabic vowel and the consonant(s) following. Second, the syllable appears to be a distinctly discernible quantity unit within the disyllabic stress-group. 2.3 The phonetic correlates of the ternary quantity distinction When describing the characteristics of the ternary quantity distinction, reference has to be made first to consonant gradation in relation to the development of the overlong grade. Consonants participating in grade alternation have traditionally been grouped into three series: x-series, xx-series and xy-series. Members of the first series contain a single consonant in the weak grade, and these alternate with short geminates in the strong grade; those of the xx-series contain short geminates in the weak grade, and alternate with long geminates in the strong grade; members of the xy-series have short consonant clusters in the weak grade that alternate with long consonant clusters in the strong grade. In Skolt Saami, consonant gradation has become entirely morphologized. The original phonetic motivation (that is, the closed versus open status of the syllable) has disappeared. Historical processes, the loss of many word-final consonants being the most important, have rendered gradation rules (a lenition process) opaque. The overlong grade in Skolt Saami has developed as a result of historical contraction (Korhonen 1981:153–157), whereby original trisyllabic words became disyllabics and compensatory lengthening caused the original strong grade of consonants of the x-series to become similar to the strong grade of consonants in the xx-series. Earlier studies on quantity have described the complex length alternations by relating them to consonant gradation in many Saami dialects (Nielsen 1902; Collinder 1929; Wiklund 1914, 1915, 1919, 1921; Itkonen 1916, 1946, among others). More recent works (Magga 1984; Engstrand 1987) have concentrated on the role of duration in the quantity of the Saami languages examined. Thus far, very few studies have taken other prosodic features into consideration (the most


179

important being Ravila 1962; Korhonen 1971, 1975; and Sammallahti 1977, 1982). Ravila recognized that durational differences alone between consonants may not reflect what are perceived as differences in quantity by speakers of Saami. For example, in the North Saami minimal triplet bore ‘edge. ’ /Q1/, borre ‘edge. ’ /Q2/, and bor"re ‘eat.’ /Q3/,5 the difference in duration of the consonants is only partially responsible for the difference between Q2 and Q3. There exist differences also in the pitch movement of the first syllabic vowel as well as in the stress on the second syllable. While in Q2 the vowel has a rising pitch movement, that of Q3 is characterized by a falling pitch movement (Ravila 1962:492). Further, he argues that the three-way quantity opposition occurs only in connection with the syllable carrying the main stress. The following examples illustrate this: goaro ‘sew. '’ /Q1/, goarro.$ /Q2/, and goar"ro. [r:r] /Q3/. According to Ravila (1962:491), quantity in the above triplet has to be associated with a distinction between vowel-peaked (Q1), balanced (Q2), and consonant-peaked (Q3) structures. Ravila claims that, while the sum of syllable quantities is the same, the distributions of quantity and articulatory intensity are different. I am not aware of any empirical evidence that Ravila may have used in order to establish this theory on the roles of pitch movement and stress in the realization of quantity degree differences. In Korhonen (1971, 1975), second syllabic stress is claimed to be significant in distinguishing between Q2 and Q3 in Skolt Saami (see discussion in section 2.3.1). In order to determine the quantity degrees of the key stress-group locations (vowel centre, consonant centre and the latus), according to Sammallahti (1977) it is necessary to assume the following: (i) Consonants in the consonant centre may glottal arrangements: (a) a single sub-glottal stretch or pulse (b) two sub-glottal stretches (c) three sub-glottal stretches

represent three different sub(C) (C;C) (C:C)

In justifying the dividing of consonants into the three groups indicated above, Sammallahti (1982) refers to Harms’s spectrographic analysis of data (Harms 1975) from Utsjoki (another North Saami dialect) where the existence of the extra syllabic pulse in C:C was shown. As far as the acoustic correlates of this extra pulse are concerned, Harms observes: “In Utsjoki this extra pulse is clearly discernible on spectrograms as an additional increase in amplitude” [my emphasis] (1975:436). 5

In earlier North Saami orthography ["] indicates the strong grade of consonants.

180

ZITA MCROBBIE-UTASI

(ii) Vowels in the vowel centre may be distinguished by the peak of intensity. Only the vowel /a/ in the vowel centre may have contrastive quantities before three-pulsed consonants; all the other vowels, monophthongs and also diphthongs, are short before three-pulsed consonants (Magga 1984:16). Long vowels or diphthongs may be of three types: (a) initially stressed long vowels (V) (b) evenly stressed long vowels (V"V) (c) finally stressed long vowels (VV") These assumptions, as to the presence of another prosodic factor being significant in the surfacing of the third quantity degree, have so far remained at the level of hypothesis; no experiments designed to test the linguistic significance of the presumed acoustic differences between Q2 and Q3 have as yet been undertaken. 2.3.1 Configuration of prosodic components: An acoustic template of quantity. Korhonen (1971, 1975) has proposed the following theory for understanding the configuration of the prosodic components in Skolt Saami quantity (Suonikylä dialect). He argues that there are only two different quantities in connection with the first syllable. However, he asserts that stress plays an important role together with duration in the realization of the quantity degrees in the second syllable. In Skolt Saami, primary stress is always in the first syllable of the word, with secondary stress falling on the third syllable. When the second syllable is in Q3, it is also associated with secondary stress (see discussion below). Korhonen claims that, in the triplet (a) sä´lbb [sælb:] or [sælb:ˇe] ‘bolt. ’ (the last ˚ ˚ vowel may be deleted or become reduced, see the discussion in section 2.2.2); (b) sä´lbbe [sælb:e] ‘put.$ up [the framework of] a log house’ and (c) sä´lbbe ˚ [sælb:e:] ‘bolt.$’, stress is what distinguishes (a) from (b), length is what ˚ distinguishes (b) from (c), and both stress and length have to be considered in the contrast between (a) and (c). My acoustic analysis of all three prosodic parameters was aimed at testing Korhonen’s hypothesis experimentally. What follows is a discussion of the relevance of duration, fundamental frequency, and intensity realizations in the production of the three alternating quantities related to Korhonen’s hypothesis. Concerning the durational patterns associated with disyllabics such as those in the above triplet, the following facts have been confirmed by acoustic analysis: (i) the second syllabic vowel in disyllabics such as (a) in the triplet may become deleted or, as the phonetic transcription shows, reduced (see discussion in section


181

2.2.2); (ii) there is a short vowel in the second syllable of disyllabics such as (b) in the triplet; (iii) the first syllabic vowel in both (a) and (b) is half-long, while the consonant is long; (iv) disyllabics such as (c) in the triplet are contracted stems, representing only a small group of non-productive words (originally trisyllabic stems, developing into disyllabics via historic contraction). The duration of the second syllabic vowel of these disyllabics averages 240 msec — considerably longer than the majority of disyllabics in Skolt Saami (McRobbie-Utasi 1999:133–134). The most important aspect of Korhonen’s hypothesis is doubtless its recognition of the configuration of more than one prosodic factor in the ternary quantity contrast in Skolt Saami. As indicated above, he claims that it is stress that distinguishes (a) from (b). This claim implies the presence of a secondary stress immediately following the first syllable having a primary stress. He further argues that all observations confirm the stressed status of the second syllabic vowel in disyllabics of the [sælb:e] ‘put.$ up [the framework of] a log house’ ˚ type (Korhonen 1975:32). Results from the acoustic analysis of intensity associated with the second syllabic nucleus of disyllabics indicate that Korhonen’s hypothesis with regard to the existence of stress on the second syllable of disyllabics such as (b) can be supported by empirical evidence. When examining intensity in connection with Korhonen’s hypothesis, three types of measurements of the data were carried out by means of the RMS amplitude envelope (the window was set for 40 msec and for dB differences from 1, with F0 range between 90 and 365 Hz). First, the amplitude peak was measured (the maximum amplitude value within the syllable nucleus). Second, the “total amplitude” value — an approximation of the intensity integral — was calculated. Finally, values for the “average amplitude” were obtained; this measurement is independent of duration, and was derived by dividing the total amplitude by the duration and then multiplying by ten (a method similar to that employed by Beckman 1986). Further, in the present analysis, when measuring peak amplitude values, the point in duration when this value is reached was also computed. In summing up the findings of the acoustic analysis relevant to Korhonen’s hypothesis, the following tendencies have to be recognized: (i) the realization of peak amplitude values — in terms of both amplitude rise time and the point of duration when the peak values start to decrease — clearly distinguishes between disyllabics belonging to structural types associated with (a) and (b) in the minimal triplet; (ii) the difference is even more apparent in open second syllables; (iii) although peak amplitude values are the same for the two types of disyllab-

182

ZITA MCROBBIE-UTASI

Closed syllable Amplitude rise time : Amplitude increase: End of peak: Amplitude decrease:

Disyllabics of type (a) Disyllabics of type (b) x¯ x¯ 33 msec 39 msec 4 dB 3 dB 70 msec 80 msec 6 dB 4 dB

Open syllable Amplitude rise time: Amplitude increase: End of peak: Amplitude decrease:

x¯ 36 msec 4 dB 70 msec 5 dB

Peak amplitude value difference between open/closed syllables: ≤ 1dB Average amplitude value difference between open/closed syllables: ≤ 1dB

x¯ 23 msec 2 dB 80 msec 2 dB

≤ 1dB ≤ 2dB

Table 2: Summary of intensity realizations in the second syllable of disyllabics ics, average amplitudes are different (and this difference is more evident in open syllables). These tendencies are summarized in Table 2. As far as fundamental frequency patterns are concerned, there has been no reference in the literature to whether they may play a role in the realization of quantity alternation in Skolt Saami. According to the only available hypothesis (Korhonen 1971, 1975) concerning the configuration of prosodic factors in Skolt Saami quantity, it is intensity together with durational difference that are considered relevant to the three-way quantity contrast. Because no acoustic analysis of the prosodics of Skolt Saami quantity has been undertaken as yet, the assumption with regard to the configuration of the latter two prosodic components of quantity is no more than a hypothesis. Therefore it was justified to undertake a detailed analysis of fundamental frequency patterns in order to establish whether the main structural types are associated with different pitch contour characteristics (that is, whether or not fundamental frequency patterns play a role in the realization of quantity alternation). As will become apparent below, fundamental frequency patterns do play a role in distinguishing between members of such triplets as in the example presented, inasmuch as F0 appears to be one of the phonetic components of stress present in disyllabics such as those in (b). It is important to recognize this fact when considering the realization of intensity distribution in the second syllabic vowel, for it signifies that in Skolt Saami the


183

configuration of all three prosodic factors need to be taken into account when describing three-way quantity contrast. The analysis of fundamental frequency movement, with regard to patterns relevant to disyllabics such as (a) and (b) in the triplet, points to differences that may prove to be relevant to the contrast. There is a clearly discernible step-down pattern of fundamental frequency movement associated with Skolt Saami disyllabics in all structural types. This step-down pattern is realized in the frequency difference of 89 Hz (¯x). However, the realization of this pattern divides disyllabics such as the ones in (b) from the rest — that is, from the ones such as in (a) that are relevant to the discussion here. First syllabic F0 movements show a level pattern in disyllabics of both types; this level pattern is established at around 60 msec (¯x) into the vowel articulation; the peak value is reached after a rise of about 50 Hz (¯x) and the fall at the end of the nucleus averages 24 Hz (¯x). In the second syllable, a falling F0 pattern is attested in disyllabics such as (a) — with 30 Hz (¯x) F0 difference between the peak and the end of the nucleus — and a level F0 pattern for disyllabics such as (b). The implications of these differences are important, for they point to the presence of an F0 component in the manifestation of secondary stress in the second syllable of disyllabics such as in (b). In other words, the phonetic realization of phonemic stress in disyllabics of the (b) type includes, in addition to intensity differences, a fundamental frequency component as well. The coinciding differences with regard to fundamental frequency and intensity patterns associated with disyllabics such as in (b) in comparison with those in (a) strongly suggests that a configuration of these two prosodic elements — intensity and fundamental frequency realizations — has to be taken into account when describing the phonetic manifestation of the contrast evident in the [sælb:] or [sælb:ˇe] ‘bolt. ’ and [sælb:e] ‘put.$ up [the framework ˚ ˚ ˚ of] a log house’ pair. The fundamental frequency patterns associated with the quantity contrast analyzed here are summarized in Tables 3 and 4. The analysis of the acoustic correlates of linguistically significant quantity has resulted in the following findings: (i) the role of durational ratios between two main stress-group locations — between the first syllabic vowel and the consonant(s) following — has to be recognized on the basis of the manifestation of duration in connection with compensatory lengthening; (ii) differences in the realizations of the general step-down pattern appear to divide all disyllabics into two clearly recognizable groups (see above), thus providing a strong argument for taking the F0 component into consideration in the realization of stress in Skolt Saami; (iii) disyllabics form two clearly distinguishable groups on the basis of

184

ZITA MCROBBIE-UTASI

Word:

Step-down patterns (∼ 91 Hz F0 difference)

First syllable: Peak is reached:

Level pattern F0 Within the first 60-65 msec x¯ F0 rise (from the fifth glottal cycle to peak): 58 Hz Duration of maintaining peak: 81 msec F0 fall: 25 Hz Second syllable: Peak is reached: Duration of maintaining peak: F0 fall

Level pattern x¯ 22 msec 57 msec 30 Hz

Table 3: Fundamental frequency patterns for disyllabics of type (a)

Word:

Step-down patterns (∼ 88 Hz difference)

First syllable: Peak is reached:

Level pattern F0 Within the first 60-65 msec x¯ F0 rise (from the fifth glottal cycle to peak): 60 Hz Duration of maintaining peak: 53 msec F0 fall: 26 Hz Second syllable: Peak is reached: Duration of maintaining peak: F0 fall

Level pattern x¯ 23 msec 62 msec 2 Hz

Table 4: Fundamental frequency patterns for disyllabics of type (b)


185

the different realizations of intensity. Such results provide empirical support for the relevance of stress assumption. The above summary of results for the acoustic analysis of duration, intensity, and fundamental frequency patterns, confirms Korhonen’s hypothesis with regard to taking into account the configuration of duration and stress in the realization of phonologically relevant quantity alternations in Skolt Saami. Although he did not consider the role of pitch movement in his hypothesis, the tendencies evident in the realization of F0 patterns also appear to play a role in distinguishing between disyllabics of types (a) and (b) in the minimal triplet (as shown in the tables above). In connection with disyllabics of type (c) the following points can be made: (i) Korhonen’s assumption (Korhonen 1971, 1975) with regard to the difference between disyllabics such as (b) and (c) — [sælb:e] ‘put.$ ˚ up [the framework of] a log house’ versus [sælb:e:] ‘bolt.$’ — is upheld ˚ in view of length being the obvious distinguishing factor (¯x = 87 msec versus 240 msec); (ii) his statement concerning the relevance of both length and stress in the contrast between disyllabics of types (a) and (c) is substantiated by the acoustic analysis of the configuration of prosodic components in the realization of quantity (McRobbie-Utasi 1999). It has to be pointed out, though, that because of the low number of type (c) disyllabics, the ternary quantity contrast as stated in Korhonen’s hypohesis has a restricted function in Skolt Saami phonology. This will be further discussed in section 4. 3. Research on ternary quantity in other languages: A brief overview This section presents a brief overview of the most frequently referred to languages assumed to have ternary length distinction. The cross-linguistic survey provides additional support for considering such systems uncommon and transitory. 3.1 German German is one of the few languages which has received attention regarding the existence of three degrees of length. Research on German vowels has a long history during which examination of the presumed three-way contrast proceeded from both phonetic and phonological perspectives. Influential studies, such as von Essen (1957, 1966); Martens & Martens (1961); Hildebrandt (1963); Hanhard et al. (1965); Weiss (1976); and Wodarz (1979) represent a phonetic approach to the ternary length phenomenon; phonological analyses of the three-way contrast include studies by Mueller (1958); Bluhme (1970); Dinnsen & Garcia-Zamor (1971), among others. The large volume of earlier studies de-

186

ZITA MCROBBIE-UTASI

voted to the issue of Überlänge [overlength] has failed to provide convincing arguments for the claim that a third degree of length is relevant in the phonology of Standard German. In Wodarz (1979:31-37, 39-40) there is a critical account of the most influential earlier studies, including Martens & Martens (1961); Mueller (1956, 1958); von Essen (1957, 1966); and Pilch (1966) which advocate the legitimacy of distinguishing between long and overlong vowels. Among the most frequently cited opponents of the three-way quantity are Mangold (1974); Hanhard et al. (1965); and Bluhme (1970). There is a brief account of their views in Wodarz (1979:37-39, 40-41). On the basis of his acoustic analysis, Wodarz concludes that the so-called Überlänge does not play a role in Standard German phonology (1979:285). In a recent study of German phonology (Wiese 1996), the three-way quantity contrast is mentioned only in passing with reference to the study presented in Kohler (1977) accepting the latter’s conclusion, that there exists no empirical evidence justifying the claim of a three-way distinction. It has long been acknowledged that German dialects display patterns different from those of Standard German with regard to the number of phonologically relevant durations. In his study, Weiss (1976) refers to Niederdeutsch where the three contrastive degrees are associated with four phonetic realizations of duration. Further, in Ternes (1973:107) and Kohler (1977:125-126) reference is made to the so-called Rheinischen Schärfung, a phenomenon characteristic of some dialects of Rhineland, where the distinction between short, half-long and long vowels is achieved by (i) durational differences, and (ii) a falling F0 pattern associated with the half-long vowel. The presence of Überlänge in the North Saxon dialects has been examined by studies such as Bremer (1928); Keller (1961) and Ruscher (1983). In a survey of German dialects (Russ 1989) the Überlänge phenomenon is no longer considered relevant in the phonological system of North Saxon. Russ refers to Kohler’s study, in which several arguments are presented for the lengthening effect of the lenis consonant, a process thought to result in Überlänge of the preceding vowel (Kohler 1986). Kohler concludes that the three-way contrast is so sporadic that it has no role in the language. 3.1.1 Schleifton. It has long been acknowledged that the phonetic realization of Überlänge (vowels in the third degree) may include a characteristic pitch movement in addition to an increase in duration. Thus the term Schleifton has been employed in several studies to describe these vowels (Bremer 1928; Schmitt 1950; Wängler 1960; von Essen 1958; Martens & Martens 1961; Ruscher 1983, among others).


187

The Schleifton phenomenon has previously been analyzed mainly from a historical perspective. Bremer (1928) attributes the development of Schleifton to the deletion of word-final unstressed vowels. The consequence of apocope on the vowel in the preceding syllable was two-fold: (i) increase in duration, and (ii) a falling pitch movement. According to Bremer, both the durational increase and the falling pitch apparent in the resulting vowel conform to the duration and pitch pattern of the original disyllabic form. In other words, he considers the development of Schleifton as a compensatory process as far as duration is concerned, and, with regard to pitch movement, he emphasizes the preserving of the characteristic pattern associated with the relevant disyllabics. Bremer’s explanation of the Schleifton phenomenon is to be interpreted as being characteristic of German in general rather than as a special feature of certain northern dialects. Keller (1961) restricts the relevance of Schleifton to the North Saxon dialects. In summing up the research on the three-way quantity contrast, it may be concluded that (i) a presumed ternary length distinction in Standard German is no longer held to be a valid assumption, (ii) the claim that Überlänge is a dialectal feature relevant in the phonological system has similarly been refuted, and (iii) the existence of Schleifton in some German dialects — most notably the North Saxon dialects — requires further study in order to establish whether the occurrence of Schleifton does indeed result in a ternary quantity distinction in the phonological system. 3.2 Amerindian languages: Mixe, Hopi, Wichita, and Sarcee 3.2.1 Mixe (Zoquean). At present, the status of the three-way vowel contrast in Mixe remains unresolved. In view of the differing approaches — Hoogshagen (1959); Hoogshagen & Hoogshagen (1993) versus Wichmann (1995) — it is unclear whether the domain of ternary length alternation is the syllable nucleus alone or the syllable nucleus together with coda. The other question that awaits clarification concerns the phonetic correlates of the three degrees under investigation. Namely, is the phonetic manifestation of length only duration, or is there another prosodic factor relevant to the perceiving of contrast? To my knowledge no acoustic analysis has addressed this question. 3.2.2 Hopi (Uto-Aztecan). The assumption that Hopi is a language with three contrastive lengths has never been validated. Whorf (1946) suggested in his grammar that Hopi has a ternary vowel length distinction, though very little is known of the specifics of this claim. Citing Whorf, Trubetzkoy (1969[1939]) emphasizes that the ternary distinction does not occur in all positions: “three

188

ZITA MCROBBIE-UTASI

lengths do not occur in a word-final vowel ... If such a vowel is accentuated, its length is medium, with very few cases of long.” [emphasis in the original] (Trubetzkoy 1969[1939]:224). In a recent dictionary containing an outline of Hopi grammar, the vowel inventory of Hopi is given as having a binary vowel contrast (Hill et al. 1998:863); the description of vowel length alternation also refers to short and long vowels. No reference is made to a possible ternary length contrast. 3.2.3 Wichita (Caddoan). Rood (1975) claimed that Wichita has three degrees of vowel length. In describing the phonology of Wichita, he points out that of the three vowel phonemes in Wichita, /i/ and /a/ may occur short or long; overlong /i/ is rare, overlong /a/ is even more rare. Short /e/ is rare, while overlong /e/ is common (Rood 1976:230). According to a more recent report (Rood 2001), the phonological inventory of Wichita contains three vowels with two length contrasts and one pitch contrast. 3.2.4 Sarcee (Athapaskan). Cook (1971) cites Sapir (1922, 1925) and Li (1931), who proposed a three-way vowel length distinction for Sarcee. In their analysis, Hoijer & Joël (1963) upheld the view concerning the relevance of short, long and overlong distinction for vowels by postulating two “phonemes of duration” in order to adhere to the bi-uniqueness principle (referred to in Cook 1971). The validity of the ternary length distinction in Sarcee was disputed and rejected by Cook (1971, 1984), and Cook (2004, p.c.) who claims that tone should be considered the relevant feature in the apparent vowel contrast. Regarding the above survey on the presumed three degrees of length in the four Amerindian languages, the existing research provide no more than inconclusive evidence. With few acoustic analyses available to identify the role of duration in the three-way distinction, the history and present of the ternary length contrast remains uncertain. 3.3 Scottish Gaelic In Ternes (1973), the surface three-way vowel length distinction in Scottish Gaelic is interpreted by using the mora as a phonemic unit. Minimal triplets, such as /S`ın´/ ‘we’ (stressed pronoun), /S`ı;ń/ ‘venison’, /S`ı:ń/ ‘to sing’ indicate /V/ versus /V;/ versus /V:/ contrasts; this ternary distinction is relevant to monosyllabics only in the dialect analyzed (Applecross, Ross-shire dialect of Scottish Gaelic). The author’s moraic phonological approach does not make explicit


Q1 sada Q2 saada! Q3 saada [a:a]

189

‘hundred. ’ taba ‘lock. ’ tapa! ‘kill. '’ ‘send. '’ tappa [p:p] ‘kill.’ ‘get.’

Table 5: Ternary quantity contrast in Estonian whether the surface contrast represents durational differences only in the realization of three morae. 3.4 Estonian Estonian is the language that has been most thoroughly researched with regard to a ternary quantity contrast (see examples in Table 5).6 Originally, it contained two contrastive lengths in the phonological system; later, three degrees of length came into existence due to a compensatory lengthening process (Lehiste 1978a). In describing Estonian quantity, reference has to be made to (i) the domain of quantity, and (ii) the configuration of prosodic factors relevant in the realization of quantity alternation in the disyllabic metric unit. 3.4.1 The domain of quantity. Lehiste has argued in a number of publications (Lehiste 1960, 1965, 1966, 1970, 1989, 1997), that quantity in Estonian cannot be adequately described without reference to disyllabic units (that is, metric feet) within the phonological word. The three degrees are contrastive only within these units; thus the domain of quantity in Estonian is the disyllabic metric foot (Lehiste 1997). In addition to recognizing the role of the disyllabic foot, it is essential to establish a phonological hierarchy of the segment, syllable, metric foot and word in order to understand quantity patterns in Estonian. It is quantity that relates these units to one another within this hierarchy (Lehiste 1997:16). The relationship between segmental and syllabic quantity is quite complex. The vowel in open Q1 syllables is always short, but when segmental quantity is examined in relation to Q2 and Q3 quantities this relationship is not so straightforward, for syllable quantity is not the sum of segmental quantities present in the syllable. For example, even though the sum of segmental quantities may be the same, as in saate ‘get.'’ and saade ‘hayrick. ’ (2 + 2 in saate and 3 + 1 in saade), syllabic quantities are nevertheless different (Q2 and Q3 respectively).7 6

The abbreviations Q1, Q2, and Q3 for the three degrees of quantity follow established practice (see, for example, the relevant papers in Lehiste & Ross 1997). 7 Surveys of segmental quantity in relation to syllabic quantity may be found in Posti (1950);

190

ZITA MCROBBIE-UTASI

In connection with syllabic quantity, the following facts have to be taken into consideration: (i) three contrastive quantities may occur in the first syllable of a polysyllabic word; (ii) the number of possible contrasts in subsequent syllables is determined by the quantity of the odd-numbered syllable (when the odd-numbered syllable is in quantity three, the syllable following may occur in quantity three as well; after a Q1 or Q2 syllable, short or non-contrastive long syllables may occur). The prosodic levels in the proposed phonological hierarchy of segment, syllable, disyllabic foot, and word are summarized in the model developed in Lehiste (1965) and reiterated in Lehiste (1997). According to this model, segmental quantity cannot be described without reference to the syllable and its position within the word. Quantity alternation can thus only be described by referring to units larger than the segment (such as syllables which are constituents of disyllabic units). This is not to say, though, that segmental quantity is not relevant: the same quantity degree of the syllable can be associated with different segmental quantities that are contrastive. However, segmental quantity is not predictable on the basis of syllabic quantity (Lehiste 1997:16). More recent research on Estonian quantity appears to give further weight to the role of stress (=accent) in quantity. Approaching the issue of the domain of quantity from the perspective of the “accent theory”, Eek & Meister (1997) also recognize the relevance of disyllabic units in Estonian. In their study, a model of prosodic hierarchy is proposed in which disyllabic accent feet are designated as minimal prosodic units in Estonian: All three accents (traditionally called syllabic quantity degrees) can be identified only when information about the second syllable vowel has been delivered to listeners. The smallest prosodic unit in Estonian does not coincide with the syllable. Disyllabic accent feet are minimal prosodic units in Estonian. (Eek & Meister 1997:71)

In an experimental phonetic study (Lehiste 1970–1975), the validity of the disyllabic foot as being the domain of quantity was confirmed. The results of the experiment clearly indicate that with triplets such as taba ‘lock. ’ /Q1/, tapa! ‘kill. '’ /Q2/, tappa ‘kill.’ /Q3/, the duration of the second syllable is relevant in identifying the three quantities. Shorter duration in the second syllable (90 msec) favours assigning it to quantity three, while longer duration (180 msec) favours assigning it to quantity one and quantity two. The second syllabic duration is determined by the ratio between the durations of the Lehiste (1960, 1965, 1978b, 1997); and Krull (1992).


191

first and second syllables, therefore it is not independently contrastive. Lehiste’s experiment thus reveals the manifestation of a three-way contrast at the level of the disyllabic sequence. In summing up the issue of the prosodic hierarchy in relation to the domain of quantity, the three-way contrast manifests only at the prosodic level of the disyllabic foot. 3.4.2 The phonetic correlates of the ternary quantity distinction. Most early studies describing quantity alternation attribute segmental durational differences to the manifestation of different quantity degrees (for example, Ariste 1933, 1939, 1953; Sõster 1938). There were, however, some early attempts to relate factors other than duration to the manifestation of quantity degrees. For example, Põldre (1937) examines pitch and intensity in relation to quantity, and suggests the relevance of this relationship in the realization of contrastive quantities in Estonian. In another early study, Durand (1939) argues that the distinction between long and overlong grades should be analyzed as a tonal difference. All these early studies adopt a segmental approach. Posti (1950) proposed that differences between the second and third quantities are determined by syllabic rather than segmental properties. This approach was also implicit in a much earlier study (Ariste 1933), in which it was noted that, while segmental consonant quantities display the same duration, the respective vowel quantities are different — thus suggesting that quantity should be analyzed as a syllabic phenomenon (Harms 1962:148-149). The concept of syllabic quantity was later adopted and developed in greater detail by Tauli (1973). More recent research directed towards identifying the phonetic correlates of Q1, Q2 and Q3 has confirmed the relevance of another prosodic factor, in addition to duration, in the alternating quantity degrees. With regard to duration, it is the role of durational ratios between the first and second syllables in the metric foot that has been seen as relevant to the realization of the three quantity degrees (implying that the duration of the first syllable is not sufficient for signaling the degree of quantity, Eek 1980a,b; Fox & Lehiste 1987, 1989; Lehiste 1989, 1997). In Hint (1998) a different view is expressed, according to which the first syllable alone determines the surfacing quantity degree. Further, Krull & Traunmüller (2000, 2002) consider the duration of the segment preceding and following the segment carrying a quantity distinction important in the perceiving of different quantity degrees. This latter approach is not confirmed by perception experiments, which support the perception of quantity degrees as a hierarchical process (Eek & Meister 1997).

192

ZITA MCROBBIE-UTASI

In addition, the relevance of another prosodic factor — either F0 or accent — has been acknowledged. In her experimental phonetic analysis referred to above Lehiste (1970–1975) examined the acoustic components of the sada ‘hundred. ’ /Q1/, saada! ‘send. '’ /Q2/, saada ‘get.’ /Q3/ triplet. The results of the experiment (consisting of synthetic stimuli administered for evaluation to a large group of listeners) indicate that, although the distinction between Q1 and Q2 quantities is made on the basis of durational differences alone, the distinction between Q2 and Q3 is made primarily on the basis of pitch contour, and, to a far lesser degree duration in the second syllable: These results leave no doubt of the perceptual significance of F0. [...] It is important to recall here that there were some durations which could be changed from Q2 and Q3 by the pitch contour alone. In these intermediate durations, then, it was the pitch contour that the listeners used for making the distinction between the words. But pitch is not the only distinction; overlong Q3 is normally longer than long Q2. (Lehiste 1983:18)

Research on Estonian quantity undertaken from a different perspective appears to give more weight to the role of stress in quantity (Eek & Help 1987). The so-called “accent theory” has been developed further in a recent paper (Eek & Meister 1997), and the hypothesis concerning the possible realization of articulatory energy was experimentally tested. In this paper, the term accent (A1, A2 and A3) is employed in place of the traditional Q1, Q2 and Q3 symbols. The authors distinguish between A2 and A3 by associating the perception of A3 with some kind of peakedness that is lacking in A2. They acknowledge the existence of several acoustic features in the realization of balanced versus unbalanced articulatory energies. An earlier paper by Eek (Eek 1986, cited in Eek & Meister 1997) hypothesized that the articulatory energy associated with A2 is balanced between the two syllables, whereas in an A3 foot the articulatory energy is unbalanced. Further, in addition to the validity of the peakedness concept, they emphasize the relevance of both temporal and accentual cues (F0 and intensity) in the realization of these prosodic units. Tests conducted in two experimental studies (Fox & Lehiste 1987, 1989) confirm the results of earlier analyses with regard to the role of fundamental frequency in the perceiving of the third degree of quantity. However, they also raise certain doubts: [...] [the] intersyllabic ratio is sufficient for discrimination between Q3 and Q1. In case of identical overall durations [...] the ‘same’ judgements of ratios 2:1 and 2:1 were 96.9 [%]; in the case of different overall durations, [...] the recognition


193

score dropped to 82.2%, which is still relatively high. In the identification of meaningful words, however, correct identification scores never rose above 51%. [...] Thus the addition of the falling contour to the first syllable of an actual word failed to provide the necessary cue for identification at a level better than chance. (Lehiste 1989:85)

That the addition of F0 did not result in higher scores in the judging of Q3 may suggest that another cue is needed. Lehiste does not exclude the possibility that such an additional cue may prove to be one of the features proposed by Eek & Help (1987) in their theory outlined above. The possibility of employing another cue has received further substantiation by the research of Krull (1993a,b). While acknowledging the results of research on the influence of intonation on pitch contour (Krull 1997, 1998; Asu & Nolan 1999), Lehiste argues that pitch remains “quite stable if words are compared when they occur in sentence-medial position in identical context” (Ross & Lehiste 2001:49). Recent research on Estonian quantity acknowledges the existence of two types of distinctions. First, the distinction between Q1 and Q2 is affirmed as being durational (as has been argued most notably by Lehiste 1989, 1997; Fox & Lehiste 1987, 1989; Eek 1980a,b; Krull 1992). As far as the second distinction is concerned, there appears to be no agreement as yet on what constitutes the additional cue which, in addition to duration, signals the distinction between Q2 and Q3. 4. Theoretical implications At the outset of this study, two questions of considerable theoretical significance were posited. First, why are there very few languages with ternary length contrast? Second, why are these systems with three-way length distinction unstable and transitory? The survey of research in the preceding sections highlights some of the main tendencies — which appear to be strikingly similar in all the languages analyzed — with regard to these two questions. The discussion that follows attempts to provide insight into these issues, with the objective of initiating a scholarly dialogue aimed at resolving the problems associated with the ternary length (or quantity) contrast encountered in natural languages. In Lehiste (1970) only three languages (Hopi, Mixe, and German) are mentioned — in addition to the two quantity languages, Estonian and Saami — as perhaps having a three-way contrast. As described in sections 3.1–3.3, the relevance of the ternary length contrast in the phonological system of the six languages presented (German, Mixe, Hopi, Wichita, Sarcee and Scottish Gaelic) has proved to be questionable. There exists no decisive evidence to substantiate the

194

ZITA MCROBBIE-UTASI

existence in Mixe or Hopi of three distinctive lengths. The presumed existence of the ternary length distinction in Mixe is badly in need of substantiating by acoustic analysis. As it now stands, the issue of ternary contrast in that language is still unresolved. Similarly, while Whorf (1946) may have been correct in describing Hopi as a language with three degrees of length, subsequent research has yet to confirm the validity of his assumption. In Standard German the existence of Überlänge has been refuted (Kohler 1977; Wiese 1996). Research has tended to invalidate the assumption of three degrees of length in German dialects as well (Russ 1989). Earlier claims asserting the relevance of ternary length distinction in Wichita and Sarcee are no longer accepted (Rood 2001; Cook 1984, Cook 2004 p.c.). Estonian and Saami are the two languages where ternary contrast is incontrovertibly relevant. However, as discussed in sections 2 and 3.4, it is not legitimate to consider these two languages as having a ternary length contrast in their phonological system. In the realization of three-way distinction, the relevance of other prosodic factors have to be taken into account, in addition to duration. The three-way quantity contrast attributed to Scottish Gaelic (Ternes 1973) makes no definite claims with regard to the prosodic correlates of the phonemic length distinction. The reference to an earlier study of another Scottish Gaelics dialect (Borgstrøm 1941:104) suggest that Ternes was considering pitch movement as well in the realization of half-long and long vowels. That languages with ternary length distinction are rare, may explain why research has adduced no substantive evidence supporting the existence of phonological systems containing three-way length contrast. However, it has been convincingly argued (see Korhonen 1981; Lehiste 1978a, 1997) that the original binary length contrast has developed into ternary contrast on the basis of historical processes bringing about compensatory lengthening either from loss of a segment (as was the case in Estonian) or by contraction (as in Saami). Compensatory lengthening resulted in overlength in both languages. The loss of a syllable, however, was not only compensated for in terms of the additional duration in the preceding syllable. The changes that have developed in the prosody in the overlong syllable are clear in the manifesting of Estonian quantity (whether we accept pitch or stress as being the independent variable, or consider both as relevant to the resulting ternary contrast). Lehiste has convincingly argued that in Estonian a prosodic change is in progress (Lehiste 1983, 1997), whereby the change is from a “primarily quantitative language to a basically accentual” one (Lehiste 1997:26). The direction of development in Estonian is thus as follows:


195

binary length contrast → ternary quantity contrast (that is, duration plus pitch and/or accent) → and, possibly, a phonological system in which stress, quantity, and tone all play a part. Development from a language originally with binary length contrast to a ternary quantity language appears to be also the case with North Saami. When binary length contrast via compensatory lengthening gave rise to the overlong degree, the subsequent stage was associated with prosodic change: a configuration of prosodic components had to be taken into account (Ravila 1962; Harms 1975; Sammallahti 1982). Similarly, the analysis of Skolt Saami shows that a combination of prosodic components is relevant when accounting for the three-way quantity contrast in that language. It is significant that the ternary distinction in Skolt Saami came into existence as a result of the historical compensatory process that, with the development of overlength, changed the original binary length distinction into a ternary one. However, as in Estonian, the compensatory process involved changes in the prosody of the disyllabic stress-group, causing stress and pitch to play a role in the contrast. As mentioned in section 3.1.1, in some German dialects — notably North Saxon — the so-called Schleifton phenomenon has to be taken into account. Although it is uncertain whether Schleifton plays a role in the realization of contrast at present, one aspect of the Schleifton phenomenon needs to be acknowledged here in connection with the three-way quantity contrast: namely, that its development was tied to a historical compensatory lengthening process (von Essen 1957, 1958, 1966). This process resulted in the evolving of Überlänge, and additional prosodic factors were associated subsequently with the alternating grades. This development is strikingly similar to that described above in connection with Estonian and Saami quantity, In Hoogshagen (1959), it was indicated that the development of the overlong segment in Mixe had been brought about by a compensatory lengthening process resulting from the loss of a segment. Ternes (1973:104) states that the trimoraic vowel has developed from disyllabics sequences of two short vowels. No information on the acoustic properties relevant to the presumed three-way length alternation in Mixe is available to date. To summarize, the following tendencies were observed with regard to the acknowledged cases of ternary length distinction: (i) the evolving of overlength that brought about ternary length contrast can be traced back to historical changes that resulted in compensatory lengthening; (ii) in languages in which the ternary system has become stabilized, the three-way length contrast has given way to a more complex three-way quantity contrast (in addition to duration, other proso-

196

ZITA MCROBBIE-UTASI

dic factors need to be taken into account); (iii) in languages with ternary quantity distinctions, it is durational ratios, not absolute durations, that are relevant in the contrast, together with at least one more prosodic feature; (iv) the domain of quantity is larger than the segment: in fact, the only languages known for having a complex quantity system with ternary distinction (Estonian and Saami) have the disyllabic foot (or stress-group) as the domain of quantity; (v) the existence of a prosodic hierarchy needs to be recognized (as shown by both Estonian and Skolt Saami); and (vi) there are indications that primarily quantity languages may become accentual languages. The implications of these tendencies clearly point to instability as being characteristic of a system with three-way length distinction. Further, the last point suggests that three-way quantity systems may also prove to be not stable either — an issue that needs to be tested in subsequent research. In order to achieve an understanding of this latter point, first the differences between the Estonian and Skolt Saami systems have to be taken into account. In Estonian, the prosodic change in progress appears to proceed from a quantity language to an accentual language, but in Skolt Saami the situation is different. The group of disyllabics of the (c) type [sælb:e:] ‘bolt.$’ is small, and comprises disyllabics that ˚ are not productive in the language, rendering the occurrence of minimal triplets a restricted one. It is doubtful that disyllabics of this type will ever become productive again, due to the severely endangered status of Skolt Saami. It seems more likely that instead of developing into an accentual language (like Estonian), Skolt Saami may revert to a binary system by virtue of the restrictive status of disyllabics such as those in the (c) type. Because the language is severely endangered, it may not be possible to draw additional general conclusions regarding the direction of change indicated. 5. Conclusions This study provides a survey of the languages known to have a three-way quantity contrast in their phonological systems. While research is not conclusive as to the specific issues, the general trends associated with ternary quantity distinctions appear to be strikingly similar. The acoustic analysis of Skolt Saami described here provides additional evidence substantiating the temporary state of the three-way length distinction. The results emerging from my analysis of Skolt Saami data provide support for the six tendencies listed. In Skolt Saami, the third degree of length came into existence through historical changes that resulted in compensatory lengthening (section 2.3). The initial three-way length contrast gave way to three-way quantity con-


197

trast where both stress and fundamental frequency patterns appear to be relevant to the ternary contrast. The domain of quantity is the disyllabic stress-group, and durational ratios rather than absolute durational values are significant in the phonological system of the language. It is also clear that the three-way length distinction constituted no more than a transitory stage in the phonology of Skolt Saami. The results of the present study may contribute to our knowledge of human languages through showing that although phonological systems can function perfectly well with ternary distinction, there appears to be a counter-tendency, as soon as the system becomes ternary, to re-establish the binary system or to readjust the system to render another prosodic feature distinctive (such as tone in Estonian, Lehiste 1997). As few languages are known to have ternary length or quantity distinction, it is difficult to come to a definitive conclusion regarding the issue of phonological systems containing this type of contrast. Ternary length contrasts came into existence through historical processes; systems in which overlength has developed have simply “got out of balance”, as Lehiste (1997) puts it with regard to Estonian. It seems clear that such systems strive to readjust themselves along the lines described in section 4. As presented in this study, Skolt Saami represents a case in point in two respects: (i) the development of its complex quantity system corresponds in many ways to that attested in other languages, most notably Estonian, and (ii) the differences point to a hitherto unexplored instance of reestablishing the binary distinction: that is, the status of the so called supistumavartalot [contracted stems] renders the ternary contrast restrictive in the language, as analyzed in Korhonen (1971, 1975) and re-analyzed in McRobbie-Utasi (1999). References Äimä, Frans. 1918. Phonetik und Lautlehre des Inarilappischen. (= Mémoires de la Société Finno-Ougrienne 42.) Helsinki: Suomalais-Ugrilainen Seura. Ariste, Paul. 1933. “Eestil sulghäälikud k p t ja b d g”. Eesti Keel 12.73–82; 180–81. . 1939. “A quantitative language”. Proceedings of the Third International Congress of Phonetic Sciences. 276–280. Ghent. . 1953. “Foneem eesti keeles”. Eesti NSV Teaduste Akadeemia Toimetised, Ühiskonnateaduste 3.357–367.

198

ZITA MCROBBIE-UTASI

Asu, Eva Liina & Francis Nolan. 1999. “The effect of intonation on pitch cues to the Estonian quantity contrast”. Proceedings of the XIVth International Congress of Phonetic Sciences, San Francisco ed. by John J. Ohala, Yoko Hasegawa, Manjari Ohala, Daniel Granville & Ashlee C. Bailey. 3. 1873– 1876. University of California, Berkeley. Beckman, Mary E. 1986. Stress and Non-Stress Accent. Dordrecht, Netherlands and Riverton, NJ: Foris. Bergsland, Knut. 1948. Røros-lappisk grammatikk. Oslo: Instituttet for Sammenlignende Kulturforskning. Bluhme, Hermann. 1970. “Das phonologische System des Deutschen”. Lingua 25.358–380. Boersma, Paul & David Weeninck. 2003. “Praat (Version 4.0.23)”. University of Amsterdam: Institute of Phonetic Sciences. Borgstrøm, Carl Hj. 1941. The Dialects of Skye and Ross-shire. Oslo: Norwegian Universities Press. Bremer, Otto. 1928. “Der Schleifton im Nordniedersächsischen”. Niederdeutsches Jahrbuch 1:3.1–32. Bye, Patrik. 1997. “A generative perspective on ‘overlength’ in Estonian and Saami”. Estonian prosody: Papers from a Symposium ed. by Ilse Lehiste & Jaan Ross. 36–70. Tallinn: Institute of Estonian Language. Collinder, Björn. 1929. Über den finnisch-lappischen Quantitätswechsel: Ein Beitrag zur finnisch-ugrischen Stufenwechsellehre. Uppsala: Uppsala Universitets Årsskrift. Cook, Eung-Do. 1971. “Vowels and tones in Sarcee”. Language 47.164–179. . 1984. A Sarcee Grammar. Vancouver, B.C.: University of British Columbia Press. Dinnsen, Daniel & Maria Garcia-Zamor. 1971. “The three degrees of vowel length in German”. Journal of Linguistics 4:1.111–126. Durand, Marguerite. 1939. “Durée phonéetique et durée phonologique”. Proceedings of the Third International Congress of Phonetic Sciences. 80–100. Ghent, Belgium. Eek, Arvo. 1980a. “Estonian quantity: Notes on the perception of duration”. Estonian Papers in Phonetics, 5–29. Academy of Sciences of the Estonian S.S.R. Tartu. . 1980b. “Further information on the perception of Estonian quantity”. Estonian Papers in Phonetics, 31–56. Tartu: Academy of Sciences of the Estonian S.S.R.


199

. 1983. “Kvantiteet ja rõhk eesti keeles”. Keel ja Kirjandus 26:9.481–559. . 1987a. “Kvantiteet ja rõhk eesti keeles”. Keel ja Kirjandus 30:3.153–213. . 1987b. “Kvantiteet ja rõhk eesti keeles”. Keel ja Kirjandus 30:5.277–284. . 1987c. “Kvantiteet ja rõhk eesti keeles”. Keel ja Kirjandus 30:6.344–352. . 1994. Studies on Quantity and Stress in Estonian. Ph.D. thesis, Tartu University, Tartu. Eek, Arvo & Toomas Help. 1986. “Rütminihked eesti keele kujunemisloos”. Academy of Sciences of the Estonian SSR, Section of Social Sciences 1–39. Tartu. Preprint KKI-45. . 1987. “The interrelationship between phonological and phonetic sound changes: A great rhythm shift of Old Estonian”. Proceedings of the XIth International Congress of Phonetic Sciences ed. by Arvo Eek & Antti Iivonen. 6. 218–233. Tallinn. Eek, Arvo & Einar Meister. 1997. “Simple perception experiments on Estonian word prosody: Foot structure vs. segmental quantity”. Estonian Prosody: Papers from a Symposium ed. by Ilse Lehiste & Jaan Ross, 71–99. Tallinn: Institute of Estonian Language. Engstrand, Olle. 1987. “Durational patterns of Lule Sami phonology”. Phonetica 44.117–128. Fox, Robert A. & Ilse Lehiste. 1987. “Discrimination of duration ratios by native English and Estonian listeners”. Journal of Phonetics 17.349–363. . 1989. “Discrimination of duration ratios in bisyllabic tokens by native English and Estonian listeners”. Journal of Phonetics 17.349–363. Hanhard, Arthur M., Dean H. Obrecht, William R. Babcock & John B. Delack. 1965. “A spectrographic investigation of the structural status of Überlänge in German vowels”. Language and Speech 8.214–218. Harms, Robert. 1962. Estonian Grammar. (= Uralic and Altaic Series 12.) Bloomington: Indiana University Publications. . 1975. “Problems in Lapp phonology”. Congressus Tertius Internationalis Fenno-Ugristarum, pars I 433–440. Hayes, Bruce. 1995. Metrical Stress Theory: Principles and case studies. Chicago: Chicago University Press. Hildebrandt, Bruno. 1963. Experimentalphonetische Untersuchungen zur Nordniedersächsischen. Hamburg: Universität Hamburg.

200

ZITA MCROBBIE-UTASI

Hill, Kenneth C., Emory Sekaquaptewa, Mary E. Black & Ekkehart Malotki, eds. 1998. Hopi Dictionary [= Opìikwa lavàytutuveni]: A Hopi-English dictionary of the Third Mesa dialect, with an English-Hopi finder list and a sketch of Hopi grammar. Tucson: University of Arizona Press. Hint, Mati. 1998. “Why syllabic quantity? Why not the foot?” Linguistica Uralica 34.172–177. Hoijer, Harry & Janet Joël. 1963. “Sarcee nouns”. Studies in the Athapaskan Languages ed. by Harry Hoijer, 62–75. Berkeley: University of California Press. Hoogshagen, Searle. 1959. “Three contrastive vowel lengths in Mixe”. Zeitschrift für Phonetik und allgemeine Sprachwissenschaft 12.111–115. Hoogshagen, Searle & Hilda Hoogshagen. 1993. Diccionario mixe de Coatlán Oaxaca. Mexico City: Instituto Lingüístique de Verano. Itkonen, Erkki. 1939. Der ostlappische Vokalismus vom qualitativen Standpunkt aus mit besonderer Berücksichtigung des Inari- und Skoltlappischen. (= Mémoires de la Société Finno-Ougrienne 79.) Helsinki: Suomalais-Ugrilainen Seura. . 1946. Struktur und Entwicklung der ostlappischen Quantitätssysteme. (= Mémoires de la Société Finno-Ougrienne 88.) Helsinki: Suomalais-Ugrilainen Seura. Itkonen, Toivo Immanuel. 1916. Venäjänlapin konsonanttien astevaihtelu koltan, kildinin ja turjan murteiden mukaan. (= Mémoires de la Société FinnoOugrienne 39.) Helsinki: Suomalais-Ugrilainen Seura. Jacobson, Steven. 1983. “Siberian Yupik and Central Yupik prosody”. Yupik Eskimo Prosodic Systems: Descriptive and comparative studies ed. by Michael Krauss, 25–46. Fairbanks: Alaska Native Languages Center, University of Alaska. Keller, Rudolf E. 1961. German Dialects: Phonology and morphology with selected texts. Manchester: Manchester University Press. Kenstowicz, Michael. 1994. Phonology in Generative Grammar. Oxford: Blackwell. Kohler, Klaus J. 1977. Einführung in die Phonetik des Deutschen. Berlin: Erich Schmidt. . 1986. “Überlänge und Schleifton im Niederdeutschen”. Arbeitsberichte 23:5.5–17.


201

Korhonen, Mikko. 1971. “Ehdotus koltanlapin Suonikylän (nyk. Sevettijärven) murteen fonemaattiseksi transkriptioksi”. Lapin murteiden fonologiaa ed. by Erkki Itkonen, Terho Itkonen, Mikko Korhonen & Pekka Sammallahti, 69–86. Helsinki: Suomalais-Ugrilainen Seura. . 1975. “Zur Phonologie des Skoltlappischen”. Phonologische Analyse der uralischen Sprachen, 11–34. Berlin: Akademie der Wissenschaften der DDR. . 1981. Johdatus lapin kielen historiaan. Helsinki: Suomalaisen Kirjallisuuden Seura. Korhonen, Mikko, Jouni Mosnikoff & Pekka Sammallahti. 1973. Koltansaamen opas. (= Castrenianumin toimitteita 4.) Helsinki: Suomalais-Ugrilainen Seura. Krauss, Michael, ed. 1985. Yupik Eskimo Prosodic Systems: Descriptive and comparative studies. Fairbanks: Alaska Native Languages Center, University of Alaska. Krull, Diana. 1992. Temporal and Tonal Correlates to Quantity in Estonian. Perilus XV, Stockholm: Institute of Linguistics, Stockholm University. . 1993a. “Temporal and tonal correlates to quantity in Estonian spontaneous speech”. Papers from the seventh Swedish Phonetics Conference, Uppsala, May 12–14. Reports from Uppsala University Linguistics 23. Uppsala: Uppsala University. . 1993b. “Word-prosodic features in Estonian conversational speech”. Proceedings of an ASCA Workshop on prosody, September 27–29, Lund. Working Papers 41. Lund: Department of Linguistics, Lund University. . 1997. “Prepausal lengthening in Estonian: Evidence from conversational speech”. Estonian Prosody: Papers from a Symposium ed. by Ilse Lehiste & Jaan Ross, 136–148. Tallinn: Institute of Estonian Language. . 1998. “Perception of Estonian word prosody: A study of words extracted from conversational speech”. Linguistica Uralica 34.167–171. Krull, Diana & Hartmut Traunmüller. 2000. “Perception of quantity in Estonian”. Proceedings of Fonetik 2000 ed. by Antonis Botinis. 85–88. Skövde, Sweden: University of Skövde. . 2002. “Perception of quantity in Estonian (Part II)”. Fonetik 2000, the XVth Swedish Phonetics Conference, Quarterly Progress and Status Report 44. 57– 60. Stockholm: Department of Speech, Music and Hearing and the Centre for Speech Technology, KTH. Lass, Roger. 1984. Phonology: An introduction to basic concepts. Cambridge: Cambridge University Press.

202

ZITA MCROBBIE-UTASI

Laver, John. 1994. Principles of Phonetics. Cambridge: Cambridge University Press. Lehiste, Ilse. 1960. “Segmental and syllabic quantity in Estonian”. American Studies in Uralic Linguistics ed. by Thomas A. Sebeok. (= Uralic and Altaic Series 1.), 21–82. Bloomington: Indiana University Publications. . 1965. “The function of quantity in Finnish and Estonian”. Language 41.447–456. . 1966. Consonant Quantity and Phonological Units in Estonian. (= Uralic and Altaic Series 65.) Bloomington: Indiana University Publications. . 1970. Suprasegmentals. Cambridge, Mass.: MIT Press. . 1970–1975. “Experiments with synthetic speech concerning quantity in Estonian”. Congressus Tertius Internationalis Fenno-Ugristarum Tallinæ Habitus, 17–23 August 1970, Volume I ed. by Valmen Hallap, 254–269. Tallinn: Valgus. . 1978a. “Polytonicity in the area surrounding the Baltic Sea”. Nordic Prosody: Papers from a Symposium ed. by Eva Gårding, Gösta Bruce & Robert Bannert, Travaux de l’institut de linguistique de Lund, 237–247. Lund: Department of Linguistics, Lund University. . 1978b. “The syllable as a structural unit in Estonian”. Syllables and Segments ed. by Alan Bell & Joan B. Hooper, 75–83. Amsterdam: North Holland Publishing Company. . 1983. “Prosodic change in progress: Evidence from Estonian”. Language Change ed. by Irmengard Rauch & Gerald F. Carr, 10–27. Bloomington: Indiana University Press. . 1989. “Current debates concerning Estonian quantity”. FUSAC ’88: Proceedings of the Sixth Annual Conference of Finno-Ugric Studies Association of Canada ed. by Joel Nevis. 77–86. Lanham, Md.: University Press of America Inc. . 1997. “Search for phonetic correlates in Estonian prosody”. Estonian Prosody: Papers from a Symposium ed. by Ilse Lehiste & Jaan Ross, 11–35. Tallinn: Institute of Estonian Language. Lehiste, Ilse & Jaan Ross. 1997. Estonian Prosody: Papers from a Symposium. Tallinn: Institute of Estonian Language. Li, Fang-Kuei. 1931. “A study of Sarcee verb-stems”. International Journal of American Linguistics 6.3–27.


203

Magga, Tuomas. 1984. Duration in the Quantity of Bisyllabics in the Guovdageaidnu Dialect of North Lappish. (= Acta Universitatis Ouluensis. Series B Humaniora 11.) Oulu: University of Oulu. Mangold, Max. 1974. Duden Aussprachewörterbuch: Wörterbuch der deutschen Standardaussprache. Mannheim, Germany: Bibliographisches Institut. Martens, Peter & Carl Martens. 1961. Phonetik der deutschen Sprache. Münich: Max Hueber. McRobbie-Utasi, Zita. 1999. Quantity in the Skolt (Lappish) Saami Language: An Acoustic Analysis. (= Uralic and Altaic Series 165.) Bloomington: Indiana University. . 2000. “Acoustic characteristics of word-final vowel reduction in Skolt Saami”. Proceedings of the Ninth International Finno-Ugric Congress. 366– 371. Tartu. . 2001. “Vowel reduction in Skolt Saami in connected speech”. Folia Uralica Debreceniensia 8.397–403. Mueller, Hugo J. 1956. “Hinweis für einen deutschen Phonologiekurs”. Deutschunterricht für Ausländer 6:6.166–170. . 1958. “Length as a phoneme in the German vowel system”. Canadian Journal of Linguistics 4.35–37. Nielsen, Konrad. 1902. Die Quantitätsverhältnisse im Polmaklappischen. (= Mémoires de la Société Finno-Ougrienne 20.) Helsinki: Suomalais-Ugrilainen Seura. Põldre, Elin. 1937. “Intonatsiooni, kvantiteedi ja dünaamilise rõhu suhteist eesti keeles”. Eesti keel 16.164–181. Pilch, Herbert. 1966. “Das lautsystem der hochdeutschen Ungangsprache”. Zeitschrift für Mundartforschung 33.247–266. Posti, Lauri. 1950. “On quantity in Estonian”. Journal de la Société FinnoOugrienne 54.1–14. Ravila, Paavo. 1962. “Quantity and phoneme analysis”. Proceedings of the IVth International Congress of the Phonetic Sciences ed. by Antti Sovijärvi & Pentti Aalto. 490–493. The Hague: Mouton. Rood, David S. 1975. “The implications of Wichita phonology”. Language 51.315–337. . 1976. Wichita Grammar. New York: Garland. . 2001. “Wichita”. http://www.mpi.nl/DOBES/projects/wichita. Ross, Jaan & Ilse Lehiste. 2001. The Temporal Structure of Estonian Runic Songs. Berlin: Mouton de Gruyter.

204

ZITA MCROBBIE-UTASI

Ruscher, Maria D. H. 1983. On the Phenomenon of Schleifton in the Dialect of Heikendorf. Master’s thesis, Simon Fraser University. Russ, Charles V. J., ed. 1989. The Dialects of Modern German: A linguistic survey. London: Routledge. Sammallahti, Pekka. 1977. Norjansaamen Itä-Enontekiön murteen äänneoppi. (= Mémoires de la Société Finno-Ougrienne 160.) Helsinki: SuomalaisUgrilainen Seura. . 1982. “On grade alternation and the illative plural in Lappish”. FinnischUgrische Forschungen 44.110–113. . 1998. The Saami Languages: An introduction. Karasjok, Norway: Davvi Girji. Sapir, Edward. 1922. “Sarcee grammar files and stem files”. Unpublished data. University of Calgary Library Archives. Originally from Harry Hoijer’s data collection. . 1925. “Pitch accent in Sarcee, an Athabaskan language”. Journal de la Sociéteé des Américanistes de Paris 17.185–205. Schmitt, Alfred. 1950. “Schleifton und Stosston”. Zeitschrift für Phonetik 4.91– 95. Sõster, Õie. 1938. “Teise silbi poolpikk vokaal eesti ühiskeeles”. Eesti keel 17.213–223. Tauli, Valter. 1973. “Quantity and accent in Estonian”. Mémoires de la Société Finno-Ougrienne 150.390–403. Ternes, Elmar. 1973. The Phonemic Analysis of Scottish Gaelic: Based on the dialect of Applecross Ross-shire. Hamburg: Helmut Buske. Trubetzkoy, Nikolai S. 1936. “Die phonologischen Grundlagen der sogenannten ‘Quantität’ in verschiedenen Sprachen”. Scritti in onore di Alfredo Trombetti, 155–176. Milan: Ulrico Houpli. . 1969[1939]. Principles of Phonology [Grundzüge der Phonologie 1939]. Travaux du cercle linguistique de Prague VII, Berkeley: University of California Press. Translated from the German by Christine Baltaxe. von Essen, Otto. 1957. “Überlange Vokale und gedehnte Konsonanten des Hochdeutschen”. Zeitschrift für Phonetik und allgemeine Sprachwissenschaft 10.239–244. . 1958. “Die Vokale der niederdeutschen Mundart von Kirchwerder”. Zeitschrift für Phonetik und allgemeine Sprachwissenschaft 11:2/3.105–118. . 1966. Allgemeine und angewandte Phonetik. Berlin: Akademie-Verlag.


205

Wängler, Hans-Heinrich. 1960. Grundriss einer Phonetik des Deutschen. Marburg: N. G. Elwert. Weiss, Rudolf. 1976. The Perception of Vowel Length and Quality in German: An experimental-phonetic investigation. Hamburg: Helmut Buske. Whorf, Benjamin. 1946. “The Hopi Language”. Linguistic Structures of Native America ed. by Cornelius Osgood, Viking Fund Publications in Anthropology, 159–183. New York: Viking Fund. Wichmann, Søren. 1995. The Relationship among the Mixe-Zoquean Languages of Mexico. Salt Lake City: University of Utah Press. Wiese, Richard. 1996. The Phonology of German. Oxford: Clarendon Press. Wiklund, Karl Bernhard. 1914. “Stufenwechselstudien”. Le Monde oriental 7.197–239. . 1915. “Stufenwechselstudien”. Le Monde oriental 9.171–239. . 1919. “Stufenwechselstudien”. Le Monde oriental 13.55–121. . 1921. “Kolmiasteinen astevaihtelu”. Virittäjä 19.7–14. Wodarz, Hans Walter. 1979. Phonetische Untersuchungen zum Problem der Überlänge. Hamburg: Helmut Buske.

EVENTS AND CASE IN INARI SAAMI∗

DIANE NELSON University of Leeds This paper presents new data from Inari Saami in order to throw light on the relationship between case and event structure in a language with case-marked temporal adverbials. First, a link is shown between the type of case that appears (oblique vs accusative) and whether or not the adverbial delimits the event by signalling an aspectual bound. Next, Inari Saami data is contrasted with similar examples from Finnish and Russian to test several predictions from the literature about the relationship between accusative case and event delimitation. The data show that although Inari Saami allows mulitple accusative expressions for certain classes of predicate, the adverbial and not the direct object delimits the event in such cases, as predicted.

1. Introduction The aim of this paper is to explore the relationship between case-marked adverbials and event structure by looking at new data from Inari Saami, a FinnoUgric language spoken in northern Finland. In particular, this paper will focus on two relations on the interface between morphosyntax and event semantics: (a) the type of case (grammatical versus oblique) associated with particular types of temporal adverbials (event delimiting versus non-delimiting); and (b) the relationship between accusative adverbials and aspect. Section 1 will provide a general overview of the case system of Inari Saami, including a brief outline of the morphology and phonotactics. Section 2 will introduce Inari Saami case-marked ∗

This paper is part of an ongoing joint project with Ida Toivonen. Thanks are due to our Inari Saami informants Anna Kuuva, Sammeli Kuuva, Kaarina Mattus and Iisakki Mattus; and thanks also to Pekka Sammallahti, Satu Manninen, Asya Pereltsvaig, Myrdene Anderson and three anonymous referees for valuable comments. All errors are definitely the responsibility of the author. This work was supported by British Academy research grants SG31040 and LRG31734.

208

DIANE NELSON

nominative genitive accusative illative locative comitative abessive essive partitive

litto lito lito liiton littoost littooin litottáá litton littod

litoh litoi littoid littooid littoin litoiguim litoittáá

Table 1: Case paradigm for litto ‘alliance’ (Sammallahti & Morottaja 1993:125) adverbials in a cross-linguistic context. Section 3 focuses on non-delimiting adverbials in Inari Saami in locative, prolative and genitive case. Section 4 presents the Inari Saami data on accusative adverbials and gives an overview of the relationship between accusative case and delimiting expressions in theories of event structure, supported by data from Finnish and Russian. Section 5 places the Inari Saami accusative adverbial data within a theoretical context. Two main conclusions are drawn from the data. Firstly, for Inari Saami it is shown that oblique case-marked adverbials do not delimit events. Secondly, the data suggest that event measurers, both direct objects and delimiting adverbials, do occur in accusative case as predicted by Arad (1998). However, it is shown that accusative case and event delimitation do not always have a one-toone relationship at the morphosyntactic level. Delimiters in the three languages examined are invariably accusative, but not all accusative expressions are delimiters. Furthermore, unlike in Finnish and Russian, accusative direct objects in Inari Saami are not linked with the aspectual feature [+ ]. 1.1 Case in Inari Saami Inari Saami (henceforth IS) is currently spoken by approximately 250-400 people around the shores of Lake Inari in Finnish Lapland.1 Word order in IS is relatively free. Like other members of the Finno-Ugric family, IS has a rich system of inflection, including nine productive morphological cases. Examples are given in Table 1. A tenth case, the (non-productive) prolative, will also be discussed in Section 3.2. 1 Inari Saami is listed as Seriously Endangered in the Unesco Red Book of Endangered Languages. No descriptive grammar has yet been published for this language.

EVENTS AND CASE IN INARI SAAMI

a.

b.

c.

d.

209

n > n with geminate in strong stem: ˙ njune nose. ˙ njune nose. njuuneest nose. njunneen nose. kk > h with stem variants: okko week. oho week. oohoid week. ˙ ohhoost week. okkod week. p > v with geminate in strong stem: tupe cottage. tuve cottage. tuveh cottage. tuuveest cottage. tuppeed cottage. kyes si guest. kyesi guest. kyessist guest.

Table 2: IS stem phonotactics for case 1.2 IS stem phonotactics for case Before the syntactic properties of the case system of IS are discussed, a brief overview of the stem phonotactics for case in IS will be presented to aid the reader in interpreting the data. IS has a three-way system of consonant quantity, with contrastive short, half-long and long consonants. Half-long consonants are marked with a subscript dot (C) in the orthography. Examples are presented in ˙ Table 2. Depending on the noun class and the case, stems may vary with respect to consonant quantity (a-c), quality (b, c), or related syllabic properties (d). IS morphophonology for case is therefore more complex than the system of Finnish, which despite having a greater number of morphological cases has only two consonant quantities and fewer stem classes for nouns (the examples in Table 2 illustrate a subset of these for IS). 2. Case-marked adverbials in IS: An overview Along with Korean, Russian, Chinese, Arabic, Warumungu, Finnish and many other languages, IS has case-marked temporal adverbials. Both oblique

210

DIANE NELSON

and structural cases may appear with temporal expressions, but accusative case specifically is associated cross-linguistically with duration, frequency, and measure phrases while oblique cases are associated with other adverbial functions (Maling 1993). Following Borer (1994); Tenny (1994); Wechsler & Lee (1996); Arad (1998); and Pereltsvaig (2000), certain accusative adverbials2 will here be assumed to bear structural case and to function as event delimiters, which signal a temporal endpoint to an otherwise atelic event. Arad’s aspectual terminology will be adopted here: telicity is taken to be a property of events at VP-level, while boundedness is taken to be a property of sentences, including aspectual morphology and adverbials (Arad 1998:279-81). (1)

lavluh uppâ peivi. Tun you. sing. ' whole day. “You sang the whole day.”

In (1), the core predicate lavluh ‘(you) sang’ is atelic (lacks an endpoint); the addition of the accusative expression uppâ peivi ‘the whole day’ to the VP makes the predicate bounded and provides an endpoint to it, because the end of the day also marks the end of the singing event. The syntactic and semantic properties of accusative adverbials like uppâ peivi are discussed further in section 4.1. In contrast, IS time adverbials marked with other cases, especially the locative case, are non-delimiting in the sense that they locate a situation in time, but do not specify an endpoint to the event (2). (2)

livvuadeh táán ohhoost. Poccuh ¯ reindeer. lie.down.$ this week. “The reindeer (will) lie down this week.”

As in example (1), the IS verb livvuadeh ¯ ‘(they) lie down’ in (2) is interpreted as atelic. However, the locative case-marked adverbial ohhoost ‘this week’ does not delimit the event the same way as the accusative uppâ peivi ‘the whole day’ in (1): the end of the week does not pick out the same point in time as the end of the lying down. All that is entailed is that some lying down will occur at some point during the week. Based on these intuitions, temporal adverbials will be classified in the following sections into two types, non-delimiting and delimiting. It will be argued that IS oblique case-marked adverbials are non-delimiting, while accusative adverbials invariably delimit events where they occur. 2 This paper will focus on temporal adverbials only and ignore other accusative-marked delimiters, for instance distance and location adverbials. More research is needed to look at the case marking and interpretation of other classes of adverbials in IS.


211

3.

Non-delimiting expressions in IS IS temporal adverbials in a variety of cases locate an event in time, but do not delimit the event as accusative adverbials do. Non-delimiting adverbials in IS appear in locative, genitive and prolative cases. (For comparison, Finnish employs a different group of cases in these environments, including adessive, inessive and essive forms. Essive adverbials do occur in IS, but so rarely that no systematic description of their distribution can be provided here). 3.1 Locative case Locative case occurs on NPs like okko ‘week’ and iideed ‘morning’, and ¯ signals location in time. The modifiers pyätte ‘next’ and moonnâm ‘last’ do not inflect for case (3) and (4). (3)

(4)

a.

lávluh pyätte ohhoost. Tun you. sing.' next week. “You will sing next week.”

b.

Muoi kuáláástáin moonnâm ohhoost. we. fish. & last week. “We (two) fished last week.”

Muoi puurrain kuolijd onne ide ¯ dist. ¯ we. eat. & fish. today morning. “We two ate fish this morning.”

In the examples above, the locative expressions are non-delimiting because they do not signal the endpoint of the atelic event. The addition of these adverbials to an atelic predicate yields an unbounded atelic predicate; the aspectual properties of the predicate are therefore unaffected by the temporal expression. Locative expressions in IS can also occur with telic events. A standard test for delimitedness is the appropriateness of in/for expressions (Vendler 1967): delimited events are usually felicitous with in expressions, while non-delimited events are usually felicitous with for expressions (Tenny 1994:13-4). (5)

a.

Susan shook the tree (*in an hour/for an hour).

b.

Mary built a house (in a day/*for a day).

In the IS examples below, the delimited events co-occur with in expressions in locative case; the direct object, as the delimiting NP, bears accusative case.

212 (6)

DIANE NELSON

a.

Almai vuájuttij kárbá tiijmest. man sink. $ boat. hour. “The man sank the boat in an hour.”

b.

Sun raahtij táálu kuulmâ ivveest. s/he build. $ house. three year. “S/he built the house in three years.”

In (6) the expressions tiijmest ‘in an hour’ and kulmâ ivveest ‘in three years’ specify the duration of the event, but the direct object itself is the delimiter. Not surprisingly, IS has no restriction on multiple locative case-marked expressions within the same sentence; this is as expected, since as non-delimiters they do not affect the aspectual status of the predicate as in (7). (7)

meecist moonnâm ohhoost. Kuobžâh lijjii week. bear. be. $ forest. last “Bears were in the forest last week.”

In this section data has been presented to show that locative temporal expressions in IS do not affect the aspectual status of the predicate, in other words do not serve as delimiters. In the next section, other types of non-delimiting temporal expressions are examined. 3.2 Genitive and prolative case One interesting feature of the IS case system is that some non-delimiting time adverbials appear marked with genitive case (8)–(10). (8)

keccin. Muoi tanssajen oho we. dance.& week. following “We will dance a week from now.”

(9)

Tái ivii mun rahtin ˙ this. year. I build. “During these years I built houses.”

(10)

ohoi lii arvam ennuu. Tái this. week. is.$ rain. much “These weeks it has rained a lot.”

tááluid.

& house.

Interestingly, there is some speaker variation with respect to the genitive in time expressions in IS; genitive sometimes alternates freely with locative (11).


213

keesi tälvi cˇ ohˇcâ kid¯dâ ¯

keessiv tälviv cˇ ohˇcuv kid¯duv ¯

‘summer’ ‘winter’ ‘autumn’ ‘spring’

Table 3: Prolative case in a paradigm for seasons (11)

Táán oho /ohhoost luámáneh láttih. this week. /week. cloudberry. ripen.$ “During this week the cloudberries ripen.”

These data also suggest that genitive in these environments is not necessarily structurally assigned, since it occurs in a range of syntactic environments (both with demonstrative pronouns (9-11), where it alternates with locative case, and with postpositions (keccin in 8)) with no apparent effect on its status as an adverbial.3 Although more work is needed to establish the syntactic properties of the genitive in IS temporal expressions, including word order effects if any, it will be assumed that genitive here has the status of an inherent rather than a structural case. Another unusual pattern to emerge from the IS data involves an archaic prolative case. The prolative case is no longer fully productive in any of the Saami languages (see Table 3). For this reason it may be more appropriate to analyse these ‘prolative’ forms as lexicalised adverbials. One argument against doing so, however, is that this form is distributed across whole paradigms of days and seasons, not just isolated lexical items. The prolative has been retained in IS associated with a few non-delimiting time expressions, for example seasons (Pekka Sammallahti pc)(12)–(13). (12)

muu enni áásá meecist. Keesiv summer. my mother live.$ forest. “In the summer my mother lives in the forest.”

(13)

Kid¯duv párnáh siârâdeh ¯ ¯ olgon. spring. child. play.$ outside “In the spring the children play outside.”

The IS prolative case also occurs with days of the week (14). 3 An anonymous referee points out that the genitive case in (10) may indeed be structurally assigned, by the postposition keccin.

214

DIANE NELSON

(14)

Purruij-uv âânnâp lávurtuv ko moonnâm turastuv? eat-

more Saturday. than last Thursday. “Did one eat more on Saturday than last Thursday?”

Genitive and prolative adverbials, then, fall into the same class of non-delimiting expressions as locative adverbials. In the next section, the distribution of accusative adverbials in IS and other languages will be examined. 4.

Case and delimited events The data presented above show that when temporal adverbials are marked with oblique cases (locative, genitive and prolative), the adverbial does not delimit the event. In other words, an atelic event with an oblique case-marked adverbial is still unbounded and non-delimited. As mentioned in the previous section, IS also marks duration and measure adverbials with accusative case; these expressions are potential event delimiters. In this section, an overview of the syntactic and semantic properties of IS accusative adverbials is presented. Then data from Finnish and Russian are presented in the context of several hypotheses which have been made about the distribution of accusative adverbials. We will then examine the relationship between accusative case and events in IS in light of these generalisations. 4.1 Accusative adverbials as delimiting expressions Example (15), given previously as (1) in section 2 above, shows an accusative temporal adverbial of duration in IS. (15)

Tun lavluh uppâ peivi. you. sing. ' whole day. “You sang the whole day.”

Here the duration adverbial uppâ peivi ‘the whole day’ signals an endpoint to an atelic event: the end of the day also marks the end of the singing event. The accusative expression in (15) is therefore an example of an event delimiter in the sense of Tenny (1994). Even though they are not arguments of the verb, accusative adverbials like uppâ peivi share certain syntactic properties with direct objects cross-linguistically. In Finnish (Hakulinen & Karlsson 1979:216) they occur in partitive under negation (16 a,b) and passivize (16c), and in Russian (Pereltsvaig 2000:16-17) they occur in genitive under negation (17).


(16)

(17)

a.

matkalla viikon. Viivyin stayed.& trip. week. “I stayed on the trip for a week.”

b.

En viipynyt matkalla viikoa. neg.& stay. trip. week. “I didn’t stay on the trip for a week.”

c.

Siellä viivyttiin kokonainen viikko. there stayed. whole. week. “There they stayed for the whole week.”

a.

Ty rabotaeš u nas uzhe mesjac. you work for us already month. “You have already been working for us for a month.”

b.

Ty u nas esche i mesjaca ne rabotaeš. you for us yet even month. not work “You haven’t been working for us for even a month.”

215

Accusative adverbials also display aspectual properties in common with internal arguments of certain classes of verbs because they both delimit events (Vendler 1967; Dowty 1979; Tenny 1994; Wechsler & Lee 1996; Arad 1998; Pereltsvaig 2000). For example, with verbs of creation and consumption such as raahti ‘(she) built’, the building event has an endpoint in time that is directly linked to the direct (internal) object: once the house is built, the event is finished (18).4 (18)

Sun raahti táálu. s/he build. $ house. “S/he built a/the house.”

Accusative adverbials such as uppâ peivi ‘the whole day’ thus delimit situations in a related fashion, and, significantly, they also bear accusative case morphology. However, there is a difference in the aspectual status of the predicates containing táálu ‘house’ in (18) and uppâ peivi ‘the whole day’ in (15). The direct object táálu ‘measures out’ the event: as the house is built part by part, the event nears completion part by part, that is, it is an incremental theme in the sense of Dowty (1991). But according to Tenny (1994:10-18), uppâ peivi ‘the whole day’ is not an event measure, because the predicate lavluh ‘(you) sang’ does not entail 4 This sentence has another interpretation, ‘S/he was building a/the house’, in which the direct object does not delimit the event. These two readings will be discussed in section 5.

216

DIANE NELSON

motion along a path, incremental themehood or change of state; the day itself does not undergo some transformation which brings the singing event to a close. In Tenny’s taxonomy of aspectual roles, expressions that measure out are in fact a subset of those that delimit. A different position is adopted by Pereltsvaig (2000), following Wechsler & Lee (1996), who argues that accusative adverbials should be analysed as extensive measure functions which measure out an event. These two approaches to accusative adverbials are related to proposed constraints on which type of element can measure out an event: Tenny (1994) and Arad (1998) argue that only internal arguments/direct objects can measure out events, while Pereltsvaig (2000) proposes no such constraint. Given the number of syntactic and aspectual properties shared by accusative adverbials and direct objects, several researchers have attempted to capture the facts by positing a formal link between event measures or delimiters and accusative case. These include Arad’s (1998) generalisation in (19). (19)

All measurers are (universally) marked with accusative case. (Arad 1998:73)

Both Arad and Pereltsvaig (2000) point out, however, that this is not a symmetric condition, in other words, not all accusative expressions measure out events. Another set of predictions has been made which relate to the distribution of delimiting expressions, for example Tenny’s Single Delimiting Constraint in (20). (20)

The event described by a verb may have only one measuring-out and be delimited only once. (Tenny 1994:79)

If event delimitation is associated with accusative case, then this could potentially be seen as an aspectual constraint which has the syntactic effect of limiting the number of accusative expressions in a VP to one. This may in turn be related to syntactic-aspectual constraints on the co-occurrence of resultatives and path phrases (such as the Unique Path Constraint, Goldberg 1991; see also Fong 2003 for an analysis of resultatives, depictives and case in Finnish). Evidence against this hypothesis is presented in a later section. Other hypothetical restrictions on multiple accusative expressions are purely morphosyntactic; the Case-OCP (Mohanan 1994) prevents adjacent elements from being assigned the same case. The next section will examine data from Finnish and Russian within the context of these proposed constraints.


217

4.2 Accusative adverbials in Finnish The distribution of objective case in Finnish as it relates to event structure has attracted a fair amount of attention in the literature (Heinämäki 1984; de Hoop 1992; Kiparsky 1998). Generally speaking, accusative case on the direct object signals that the event is . at VP-level (21a), while partitive case signals that it is % (21b).5 (21)

a.

Aili rakensi veneen. Aili built.$ boat. “Aili built a/the boat.”

b.

Aili rakensi venettä. Aili built.$ boat. “Aili was building a/the boat.”

The addition of an accusative adverbial delimits an otherwise unbounded event (22a, b). However, Finnish disallows accusative case on both the direct object and the adverbial (22c). (22)

a.

He lauloivat laulut. they sang.$ song. “They sang (the) songs.”

b.

He lauloivat lauluja koko yön. they sang.$ song. whole night. “They sang songs the whole night.”

c.

*He lauloivat laulut they sang.$ song.

koko yön. whole night.

These facts show that for Finnish, as is expected cross-linguistically, only one delimiting expression is available per VP (Arad 1998; Tenny 1994). In syntactic terms, this looks like a restriction on more than one accusative-marked NP occurring in a given predicate. If the adverbial bears accusative case, the direct object must appear in partitive case (22b, 23a). Conversely, if the direct object is accusative, case-marked temporal adverbials must bear oblique (inessive) case, not accusative case (23b). 5

This is admittedly an oversimplification of the system in Finnish, since the partitiveaccusative case alternation is also associated with NP semantics (mass/count, definiteness, specificity). Objects appear in the accusative only if both the NP and the VP are (Kiparsky 1998). The notions of VP-boundedness and delimitedness are here taken to be very closely related, if not equivalent.

218 (23)

DIANE NELSON

a.

Matti osti maitoa tunnin. Matti bought.$ milk. hour. “Matti bought milk for an hour.”

b.

Matti osti maidon tunnissa / *tunnin. Matti bought.$ milk. hour. / hour. “Matti bought the milk (with)in an hour.”

This shows that inessive time expressions such as Finnish tunnissa ‘in an hour’ in (23b) do not function as event delimiters, but accusative objects do. These examples support the proposed constraints mentioned above: only one accusative delimiting expression is allowed. However, one class of stative predicates (what Kiparsky 1998 terms ‘bounded states’; see Itkonen 1976) unexpectedly allow both an accusative direct object and an accusative adverbial (24)– (25). (24)

naisen koko tuon tunnin. Näin saw.& woman. whole that. hour. “I saw the woman for that whole hour.”

(25)

vastauksen koko viikon. Tiesin knew.& answer. whole week. “I knew the answer all week.”

Other stative verbs in Finnish (for example psych predicates like rakastaa ‘to love’ and pelottaa ‘to fear’) select partitive objects, not accusative. Kiparsky (1998:284) explains the contrasting aspectual and case properties of the two classes of stative verbs as a function of gradability; ‘bounded’ states like fear and know do not allow modifiers like very much in either language (26). (26)

a.

I fear dogs very much.

b.

*I know dogs very much.

In his model, this property of certain states contributes the feature [+ ] to the event structure of the predicate, which in turn yields accusative internal arguments. Being stative, however, they are still atelic predicates and allow accusative adverbials as delimiters. The fact that certain classes of predicate license multiple accusative expressions means that purely morphosyntactic constraints (such as that suggested by Mohanan 1994) cannot adequately capture the crosslinguistic data. Instead, the acceptability of (26) points to an explanation related to the interaction between case and event structure. Unlike other event types,

219


states are not measured out by their direct objects (Tenny 1987, 1994). A duration adverbial, however, can delimit a stage-level state because it signals that the state has an endpoint. As expected, the adverbial bears accusative case, but because the internal argument itself is not a delimiter, the aspectual constraint on multiple delimiting accusatives is not violated. Given a sentence with two accusative expressions in Finnish (25), only the adverbial will delimit the event. Importantly, stative predicates such as (24-25) show that a symmetric relation between accusative case and event delimitation is untenable; although it may be true that event delimiters or measurers are always accusative (Arad 1998), accusatives are clearly not always event delimiters. The question remains as to whether other languages with accusative casemarked adverbials show the same constraints on accusative expressions as Finnish. Russian also has accusative adverbials. In Russian, it turns out that the distribution of these adverbials is tied to verbal aspect; an accusative adverbial may co-occur with an accusative internal argument only when the verb is imperfective, or where an accusative direct object is non-delimiting (27) (Pereltsvaig 2000:11-12). (27)

a.

knigu ves’ vecher. Maria taskala Mary carried. book. all. evening. “Mary carried the book for the whole evening.”

b.

*Maria pritaschila knigu ves’ vecher. Mary carried. book. all. evening. “Mary brought the book for the whole evening.”

This pattern falls out of the morphology of Russian, where aspectual boundedness is signalled via V not NP (Kiparsky 1998). Pereltsvaig (2000) argues that although Russian appears to violate the constraint on multiple accusatives seen in Finnish, Finnish (23a) and Russian (27a) are actually structurally identical. In other words, not all accusative internal objects are also event delimiters, and the [% ] feature of (23a) is encoded in the verbal morphology of Russian in (27a), not through case morphology. Again, this suggests that Arad’s (1998) generalisation is not symmetric. One fact that emerges from the data so far is that accusative adverbials always delimit the event, and when they do, no other internal arguments may do so. In the next section, the distribution of accusative adverbials in IS will be explored in light of these generalisations.

220

DIANE NELSON

5. Accusative adverbials in Inari Saami Two patterns have emerged from the cross-linguistic data so far. In Finnish, there is a strong correspondence between accusative case and delimited events. Internal arguments are accusative only when both the event and the NP are . (Kiparsky 1998), and multiple accusative expressions are disallowed except with certain stative predicates. In Russian, on the other hand, an accusative internal argument may co-occur with an accusative adverbial in non-stative predicates, but only when the verb is imperfective. However, neither language violates Tenny’s constraint on delimiting expressions (20) because accusative case on an object does not entail event delimitation. Instead, the apparent contrast in the two languages is a by-product of the way aspectual properties are signalled morphologically in different languages. The question remains as how case interacts with event structure in IS. As in Finnish and Russian, IS duration adverbials bear accusative case when they delimit events. Certain verbs select accusative duration adverbials as arguments (28b and 29b), or non-argument accusative adverbials may delimit an otherwise unbounded event (30, 31). (28a) and (29a) show nominative forms for contrast. (28)

(29)

a.

lâi kukke. Tiime hour. was.$ long “The hour was long.”

b.

Lavlum piištij tiijme. singing lasted.$ hour. “The singing lasted an hour.”

a.

lâi sâvnâd. Ijjâ night. was.$ dark “The night was dark.”

b.

Juhle piištij uppâ iijâ. party last. $ whole night. “The party lasted all night.”

(30)

vaccih tiijme. Tun you. walked hour. “You walked for an hour.”

(31)

illâvaje uppâ cˇ oovˇca. Sun lâi whole autumn. s/he was.$ ill “S/he was ill all autumn.”

221


Like Finnish, IS allows more than one accusative expression in ‘bounded states’ (stage-level verbs of perception, mental states, and related classes of predicate), as shown in (32) and (33). (32)

tuu ubbâ ton Mun oinim I see. & you. whole that. “I saw you that whole hour.”

tiijme. hour.

(33)

mainâs tiijme. Mun muštim I remember. & tale. hour. “I remembered the tale for an hour.”

This shows that IS has no morphosyntactic restriction on multiple accusative expressions. However, IS does not show the same restriction as Finnish when it comes to non-stative predicates. Like Russian, IS allows an accusative direct object to co-occur with an accusative adverbial in non-stative predicates (34)– (36). (34)

(35)

(36)

a.

vuojih tuâid avdoid uppâ táid iivijd. Tun you. drove.' those. cars. whole these years. “You drove those cars all these years.”

b.

*Sinä ajoit nuo autot koko nämä vuodet. (F) you drove.2 those car. whole these year.

a.

Sij lavluu lavluid uppâ iijâ. they sang.3 song. whole night. “They sang (the) songs all night.”

b.

*He lauloivat laulut they sang.3 song.

a.

uppâ iijâ. Sun raahtij kárbá s/he built.3 boat. whole night. “They built the boat all night.”

b.

*Hän rakensi veneen koko yön. s/he built.3 boat. whole night.

koko yön. whole night.

(IS)

(IS)

(Finnish)

(IS)

(Finnish)

The explanation for the acceptability of the IS examples above lies in the relationship between morphological case and event structure. In contrast to Finnish,

222

DIANE NELSON

objective case in IS does not encode the feature .% , and there is no equivalent of the partitive-accusative alternation reflected in the case morphology. Although IS has a partitive form (as illustrated in Table 1), its distribution is extremely limited, mainly appearing in noun phrases with numerals 7 and higher (Nelson & Toivonen 2003). Instead, accusative is normally the default structural case for direct objects. (37)

kirje. Mun luuhim I read. & book. “I read the book / I was reading the book.”

IS informants provide (37) as a translation from Finnish with both accusative and partitive objects, and confirm that an accusative object in IS is actually ambiguous between a . and a % interpretation. Unlike Russian, the IS verbal paradigm does not signal im/perfectivity morphologically either, as (37) shows. The aspectual feature . , then, is not morphologically formalised in relation to accusative expressions in IS. Since neither the objective case morphology nor the verbal paradigm of IS signal aspectual contrasts, the question remains as to which accusative expression (or both) in a potentially bounded predicate like (36a) actually delimits the event. If both NPs are interpreted as event delimiters, then the IS data pose a potential challenge to the generalisations mentioned above by Arad (1998) and Tenny (1987) that each predicate may only contain one delimiter. In fact, informant intuitions support their proposals. In sentences where an accusative direct object occurs with a locative adverbial, the object is interpreted as the event delimiter (38a, 39a). In sentences with two accusative expressions, the event is interpreted as delimited by the temporal adverbial but not by the direct object (38b, 39b). (38)

a.

tiijmest. Sun raahtij kárbá s/he built.3 boat. hour. “They built the (whole) boat in an hour.”

b.

Sun raahtij kárbá uppâ iijâ. s/he built.3 boat. whole night. “They built/were building the boat all night.”


(39)

a.

uádimviste tiijmest. Eeˇci cˇ urkij ¯ father clean. $ bedroom. hour. “Father cleaned (the whole) bedroom in an hour.”

b.

Eeˇci cˇ urkij uádimviste tuon iideet. ¯ ¯ father clean. $ bedroom. that. morning. “Father cleaned/was cleaning the bedroom that morning.”

223

(38a) and (39a) entail that the entire boat was built and the entire room was cleaned, whereas (38b) and (39b) do not entail such a result, only that the whole span of time (morning or night) has elapsed. In this respect, IS patterns in the same way as Finnish and Russian: given a predicate with two accusative expressions, it is the adverbial rather than the direct object that delimits the event. The data here provide further evidence against a symmetric mapping between accusative case and event delimitation: all event delimiters are accusative, but not all accusatives are event delimiters. 6. Conclusion Inari Saami is a language with rich case morphology which employs a variety of grammatical and oblique cases with temporal adverbials. The data presented here first support a broad taxonomy of case with respect to event delimitation. Oblique case-marked adverbials (locative, prolative and genitive) are non-delimiting; they may locate the event in time, but they do not signal an aspectual bound. Conversely, objects or adverbials that do delimit events are marked for accusative case. This is consistent with Arad’s proposals about the link between accusative case and event structure cross-linguistically. However, accusative case in Inari Saami turns out to have a less constrained distribution than either Finnish or Russian in . events. While Finnish disallows an accusative direct object with an accusative adverbial, and Russian disallows the same case pattern with perfective verbs, IS shows no such restriction. This is shown to be the result of different relationships between case and event structure in the three languages. While accusative case in Finnish is closely associated with event delimitation and the aspectual feature . , no isomorphic association holds in IS or Russian: accusative case for direct objects does not entail event delimitation. References Arad, Maya. 1998. VP-structure and the Syntax-lexicon Interface. Ph.D. thesis, University College London.

224

DIANE NELSON

Borer, Hagit. 1994. “The projection of arguments”. Functional Projections ed. by E. Benedicto & Jeff Runner. (= UMass Occasional Papers 17.), 19–48. Amherst: University of Massachusetts. de Hoop, Helen. 1992. Case Configuration and Noun Phrase Interpretation. Ph.D. thesis, University of Groningen. Dowty, David. 1979. Word Meaning and Montague Grammar. Dordrecht: Reidel. . 1991. “Thematic roles and argument selection”. Language 67.547–619. Fong, Vivenne. 2003. “Resultatives and depictives in Finnish”. In Nelson & Manninen 2003, 201–233. Goldberg, Adele. 1991. “It can’t go down the chimney up: Paths and the English resultative”. Berkeley Linguistics Society 17.368–378. Hakulinen, Auli & Fred Karlsson. 1979. Nykysuomen lauseoppia [=Modern Finnish Syntax]. Helsinki: Suomalaisen Kirjallisuuden Seura. Heinämäki, Orvokki. 1984. “Aspect in Finnish”. Aspect Bound ed. by Casper de Groot & Hannu Tommola, 153–177. Dordrecht: Foris. Itkonen, Terho. 1976. Erään sijamuodon ongelmia. (= Opusculae Instituti Lingua Fennicae Universitas Helsingiensis 53.) Helsinki: University of Helsinki. Kiparsky, Paul. 1998. “Partitive case and aspect”. The Projection of Arguments: Lexical and compositional factors ed. by Miriam Butt & William Geuder, 265–308. Stanford, Calif.: CSLI Publications. Maling, Joan. 1993. “Of nominative and accusative: the hierarchical assignment of grammatical case in Finnish”. Case and Other Functional Categories in Finnish Syntax ed. by Anders Holmberg & Urpo Nikanne, 49–74. Berlin: Mouton de Gruyter. Mohanan, Tara. 1994. “Case OCP: A constraint on word order in Hindi”. Theoretical Perspectives on Word Order in South Asian Languages ed. by Miriam Butt, Tracy Holloway King & Gillian Ramchand, 185–216. Stanford, Calif.: CSLI Publications. Nelson, Diane & Satu Manninen, eds. 2003. Generative Approaches to Finnic and Saami Linguistics. Stanford, Calif.: CSLI Publications. Nelson, Diane & Ida Toivonen. 2003. “Counting and the grammar: Case and Numerals in Inari Sami”. In Nelson & Manninen 2003, 321–341. Pereltsvaig, Asya. 2000. “On accusative adverbials in Russian and Finnish”. Adverbs and Adjunction ed. by Alexandra Alexiadou & Peter Svenonius. (= Linguistics in Potsdam 6.), 155–176. Potsdam University.


225

Sammallahti, Pekka & Matti Morottaja. 1993. Säämi-suomâ sänikirje = Inarinsaamelais-suomalainen sanakirja. Utsjoki, Finland: Girjegiisá. Tenny, Carol. 1987. Grammaticalizing Aspect and Affectedness. Ph.D. thesis, MIT. . 1994. Aspectual Roles and the Syntax-Semantics Interface. Dordrecht: Kluwer. Vendler, Zeno. 1967. Linguistics and Philosophy. Ithaca: Cornell University Press. Wechsler, Stephen & Yae-Sheik Lee. 1996. “The domain of direct case assignment”. Natural Language and Linguistic Theory 14.629–664.

VERBAL AGREEMENT IN INARI SAAMI∗

IDA TOIVONEN Carleton University This paper concerns agreement marking in Inari Saami. Inari Saami verbs agree in three persons and three numbers. In addition to the full agreement paradigm, there is a partial agreement paradigm. The partial agreement paradigm contains two forms: one form agrees with singular nouns, and the other form agrees with dual and plural nouns. It is proposed here that nouns marked as human and specific trigger full agreement, whereas other nouns trigger partial agreement. It is further argued that the agreement trigger is necessarily the subject. Finally, this paper compares the Inari Saami agreement system with verbal agreement in Finnish.

1. Introduction Inari Saami has a split agreement system: a finite verb is conjugated for full or partial agreement. The choice of paradigm is determined mainly by animacy: the verb fully agrees with nouns that denote animates and partially with nouns that denote inanimates. This paper describes the Inari Saami agreement system and provides a lexicalist analysis of the partial and the full agreement. Another goal of this paper is to determine the exact nature of the agreement trigger. Consider the Inari Saami sentence in (1). ∗

The following people have provided helpful comments on earlier drafts of this paper: Ash Asudeh, Myrdene Anderson, Stephen Anderson, Andrew Carstairs-McCarthy, Greville Corbett, Mary Dalrymple, Tibor Laczkó, Diane Nelson, Pekka Sammallahti, Anita Szakay, and three anonymous reviewers. I also received useful feedback from audiences at Rochester University, University of Canterbury, and the 19th Scandinavian Conference in Linguistics in Tromsø. I am grateful to everybody mentioned above, and I am also indebted to Anna Kuuva, Sammeli Kuuva, Iisakki Mattus, Kaarina Mattus and Pekka Sammallahti for helping me with the Inari Saami data. The research presented in this paper was made possible by the British Academy research grants SG-31040 and LRG-317345, University of Canterbury internal research grant U6566, and SSHRC Standard Research Grant 410-2006-1650.

228

IDA TOIVONEN

1 2 3

kuá"lástam kuá"lástah kuáláást

kuá"lásteen kuá"lástep kuá"lástvettee kuá"lástvetted¯ kuá"lástava kuá"lásteh

Table 1: The verb ‘to fish’, present indicative (1)

sierâdeh Párnááh ¯ meecist. children. play. forest. “The children are playing in the forest.”

In (1), the noun párnááh triggers agreement. This is thought of as subject agreement: párnááh is clearly the subject since it agrees with the verb and bears nominative case. It is also in the typical subject position: Inari Saami has relatively free word order, but is canonically SVO. For word order, agreement and case marking as standard subjecthood diagnostics, see Keenan (1975) and others. In section 3 below, I introduce examples where it is not so clear which element is the subject. I argue that the NP which triggers agreement is always the subject, regardless of case marking and linear order. The data examined are relevant for the question of whether grammatical functions have an independent status in the grammar. The Inari Saami agreement facts indicate that it is indeed necessary to posit entities like subjects in our grammatical models. Finally, this paper briefly compares the agreement marking systems in Inari Saami and the related language Finnish. In particular, the possessive constructions will be compared. I conclude that although the possessive constructions of the two languages are superficially very similar, they are in fact fundamentally different, as distinct NPs are grammaticalized as the subject. 2. Verbal agreement in Inari Saami 2.1 Full and partial agreement Inari Saami finite verbs agree in three persons and three numbers, as is shown in the present tense paradigms for kuá"lástâd¯ ‘to fish’ in Table 1. Personal pronouns also appear in three persons and three numbers. Table 2 gives the nominative case forms of the personal pronouns. Non-pronominal nouns only inflect for singular and plural; they do not have a dual form. Consider the noun almai ‘man’ in the examples in (2).

VERBAL AGREEMENT IN INARI SAAMI

229

1 mun 2 tun 3 sun

muoi tuoi suoi

mij tij sij

Table 2: Personal pronouns, nominative case (2)

a.

Almai kuáláást onne. man. fish.3 today “The man is fishing today.”

b.

Almah kuá"lástava onne. man. fish.3 today “The two men are fishing today.”

c.

Almah kuá"lásteh onne. man. fish.3 today “The men are fishing today.”

The examples in (2) shows that non-pronominal nouns do not have a distinct dual form. However, the morphologically plural subject is interpreted as dual when the verb is marked for dual agreement (2b). In addition to the full agreement paradigm exemplified in Table 1, Inari Saami also has a partial agreement paradigm. The full and the partial agreement paradigms for the verb lede ¯ ‘to be’ in present indicative are given in Table 3. The third person singular verb can occur with any singular subject, and the third person plural verb form can occur with any dual or plural subject. However, unlike many other languages, Inari Saami has no default form. In other words, there is no single form (such as 3, for example) which can be used with all persons and numbers. The distribution of the partial agreement forms is restricted by animacy. Consider examples (3–4). (3)

Animate subject: a.

já viljá. Meecist lava uábbi forest. are. sister. and brother. “In the forest are my sister and brother.”

b.

Kyehti almaa láin meecist. two man were.3 forest. “Two men were in the forest.”

230

IDA TOIVONEN

1 lam 2 3 1 2 3 1 2 3

lii lah lii lii lii láán láá leppee láá lava láá lep láá lepped¯ láá láá láá

Table 3: Agreement paradigms for ‘to be’ (4)

Inanimate subject: a.

kyehti keedgi. Riddoost láá ¯ beach. are.3 two rock “On the beach are two rocks.”

b.

Táálust láá kyehti vi’ste. house. are.3 two room “There are two rooms in the house.”

c.

Kyehti stuorra keedgi meecist. ¯ láá two big rock are.3 forest. “Two big rocks are in the forest.”

The subject of each example in (3–4) refers to two people or things; yet only the verbs in (3), display dual agreement. The crucial difference between the two sets of examples is that the subjects in (3) are animate, whereas the subjects in (4) are inanimate. These examples illustrate a general pattern: animates trigger full agreement and inanimates trigger partial agreement in Inari Saami. Example (5) further illustrates the difference in agreement as it is triggered by animacy (example (5) is taken from Mattus’s 1987 translation into Inari Saami of Johan Jernsletten and Sverre Hatle’s introductory North Saami reading book). (5)

Ohtii láin kyehti viiljaš. Sunnuu noomah lijjii Juvvá já once were.3 two brother their.3 name. were.3 J. and Päˇccˇ in. P. “Once upon a time, there were two brothers. Their names were Juvvá and Päˇccˇ in.”


231

The animate NPs in (3) and (5) are all human. NPs that denote animals are compatible with both full and partial agreement. This is shown in (6). (6)

a.

Kyehti poccuu ruáttáin/ryettih meecist. two reindeer ran.3 /ran.3 forest. “Two reindeer ran in the forest.”

b.

Puásui já peenuv lava/láá meecist. reindeer and dog are.3 /are.3 forest. “The reindeer and the dog are in the forest.”

The subjects in (6) are reindeer and dogs, and both full and partial agreement are possible. This is typical for animacy splits: animals often have the option of patterning with humans, especially if they are considered personified. Crosslinguistically, ‘higher-level’ animals (such as pets and livestock) tend to pattern with humans, whereas ‘lower-level’ animals (such as mice) pattern with inanimates. This does not appear to be the case in Inari Saami, where all animals have the option of patterning with humans. Sometimes partial agreement is used even when the subject is human, as illustrated in (7). (7)

a.

kyehti viiljâ. Muste lava/láá I. are.3 /are.3 two brother “I have two brothers.”

b.

Táálust lava/láá kyehti ulmuu. house. are.3 /3 two person “There are two people in the house.”

c.

Kyehti almaa poodijn/pottii. ¯ two man came.3 /3 “Two men arrived.”

Most examples of partial agreement with humans in our data collection occur in existential constructions and possessive constructions. This indicates that there is a correlation between agreement marking and the specificity of the subject.1 Existential and possessive constructions provide frames where non-specific readings are likely. It is notoriously difficult to elicit judgments on specificity from native speakers, but some additional evidence comes from examples such as (8). 1 Thanks to Pekka Sammallahti for helping me investigate the relevance of specificity. For specificity and verbal agreement in North Saami, see Sammallahti In preparation.

232 (8)

IDA TOIVONEN

Must lava kyehti cˇ eesi kiäh lava pargoost Helsigâst. I. are.3 two uncle who. are. work. H. “I have two uncles who work in Helsinki.”

A specific reading tends to be favored when the noun in question is modified by a relative clause. The fact that speakers strongly prefer the dual form (lava) in (8) thus supports the hypothesis that specificity is relevant for agreement. Also relevant is the fact that personal pronouns, which are inherently specific, almost always appear with full agreement, as in (9). (9)

a.

lah meecist. Tun you. are.2 forest. “You are in the forest.”

b.

Mun tuubdâm maηgâ ulmuu. I know many person “I. know.& many people.”

The fact that personal pronouns normally trigger full agreement adds further evidence to the relevance of specificity. On rare occasions, personal pronouns do, however, occur with partial agreement: (10)

a.

jie"h lah ohtuu, tust lam/lii Tun you. not.2 be. alone you. am.1/is.3 mun. I. “You are not alone, you’ve got me.”

b.

Ovdil Piäkkâ lâi ohtuu, mutâ tääl sust lepped/láá ¯ before P. was.3 alone, but now he. are.3 /are.3 tij. you. “Pekka was alone before, but now he’s got you.”

Partial agreement with pronouns only occurs in expressions where the NP triggering agreement follows the verb. NPs often receive non-specific readings in such contexts, but only specific readings should be available for personal pronouns. It may be that the possibility of partial agreement arises with the lack of topicality associated with the post-verbal NP. This claim gains support from the difference in grammaticality judgments in (11).


full

default on

olet on olemme olette ovat

1 olen 2 3 1 2 3

233

Table 4: Present tense paradigm for Finnish olla ‘to be’ (11)

a.

tun. ?Meecist lâi forest. was.3 you. “In the forest was you.”

b.

*Tun lâi meecist. you. was.3 forest.

Topicality has previously been argued to be directly connected to specific readings (Portner 2002). Another hypothesis for why partial agreement is sometimes possible with personal pronouns is that the speakers are influenced by their knowledge of Finnish. Finnish does not have full agreement in the possessive and existential constructions. All speakers of Inari Saami are also fully competent speakers of Finnish. However, note that Finnish has default agreement, with one default form, unlike the Inari Saami partial agreement paradigm. The Finnish system is illustrated with the present tense indicative paradigm for the verb olla ‘to be’ in Table 4. Inari Saami speakers never use the third person singular verb form for dual or plural subjects. In other words, Inari Saami always has partial agreement as opposed to default agreement, which would perhaps be expected if the variation with pronominals (or human subjects in general) were to be explained as a language contact phenomenon.2 2

A third possible hypothesis is that definiteness is relevant. Inari Saami does not mark definiteness on its NPs, but the idea would be that human nominals with a definite interpretation trigger full agreement, and human nominals with an indefinite interpretation do not. This hypothesis is supported by the fact that animate and definite nominals often pattern together crosslinguistically (see e.g. Comrie 1981: Chapter 6,). However, the fact that personal pronouns may trigger partial agreement (see (10)) shows that this hypothesis is incorrect, as personal pronouns have a definite interpretation.

234

IDA TOIVONEN

In sum, Inari Saami finite verbs agree fully or partially with the subject nominal. In the partial agreement paradigm, the third person singular form covers all singular subjects, and the third person plural form occurs with all dual and plural subjects. Whether the full or partial agreement paradigm is employed is determined first and foremost by animacy: inanimate subjects trigger partial agreement. This generalization is not sufficient to explain all the data, as animate subjects occasionally trigger partial agreement: even subjects as high up on the animacy hierarchy as personal pronouns can take partial agreement. The available data indicate that specificity explains the possibility of partial agreement with animates: only NPs with specific readings trigger full agreement. The fact that personal pronouns can under particular circumstances trigger partial agreement (see 10) indicates that topicality may also be a relevant factor. This topic is not pursued further here, but is left more as a suggestion than a firm claim; more research is necessary in order to properly understand the data in (10). However, it is important to note that Inari Saami verbs always display some agreement; the language does not have default agreement. 2.2 A lexical analysis This section provides a lexicalist analysis of the generalizations presented in the previous section. For the analysis, I employ the framework of LexicalFunctional Grammar (LFG; Kaplan & Bresnan 1982; Dalrymple et al. 1995; Bresnan 2001; Dalrymple 2001), since LFG provides an explicit and formally well-understood theory of the lexicon. The discussion is kept quite general and makes minimal use of LFG-specific formal devices (for more sophisticated treatments of agreement within LFG, see Dalrymple & Kaplan 2000 and others). Before I turn to the Inari Saami analysis, I briefly go over some notational and theoretical points that are adopted in LFG. The main focus here is on the lexical entries. Each entry includes a specification of the lexical form (for example, cats), the syntactic category (for example, N), and the necessary feature specifications (for example,

). The lexical entries will be presented as in (12). (12)

cats: N

*↑ # = ‘cat’ *↑

# =

The arrow (‘↑’) refers to the functional structure (to be discussed directly below) of the mother node in the tree and will not be relevant to any of the points made below. The feature is an identifier and a pointer to the semantics of a given


235

word. The value of the feature also contains the subcategorization frame of the predicate. Words are syntactically combined in c(onstituent)-structure, which is modelled with phrase structure trees. The lexical information of the words in the tree is then mapped onto another level of syntactic information: f(unctional)structure. F-structure is formally modelled with feature structures. The f-structure representation for the sentence in (13) is given in (14). (13) (14)

Cats purr. ⎡

⎤

‘purr * # ’ ⎥ ⎢ ⎥ ⎢ ⎢ ⎥ ⎥ ⎢ ⎥ ⎢ ‘cat’ ⎦ ⎣

As illustrated in (14), an f-structure can contain other f-structures, namely the feature structures containing information about syntactic functions such as * #, * #, * # and * #. This is the level of grammar where it is possible to make direct reference to syntactic functions for relations such as binding, control, and — crucially for this paper — agreement. The features of lexical entries can merge or unify in a single functional structure. For example, a verb may contribute information about its subject, which maps onto the f-structure. It is then crucial that the information the verb provides is compatible with the information contributed by the subject itself. Conflicting information leads to an ill-formed functional structure by the principle of Uniqueness: (15)

Uniqueness Every attribute has a unique value.

All identical features can unify. Every feature has a unique value, and so features can never unify. Let us now turn to the Inari Saami data. Consider first the lexical entry for láán in (16). (16)

láán: V

*↑ # = ‘be’ *↑ # = *↑ # = *↑

# =

*↑ # = 1 *↑ "

# = +

236

IDA TOIVONEN

The lexical entry for láán is only compatible with first person dual subjects that are human. Note that it can be argued that the human specification is superfluous for first and second person subjects. This feature can be removed from the first and second person lexical entries given here without further consequences for the analysis. The lexical entry for the first person dual pronoun muoi is given in (17). (17)

muoi: D

*↑ # = ‘pro’ *↑

# =

*↑ # = & *↑ # = *↑ "

# = +

The featural representation of a sentence like muoi láán meecist “we are in the forest” is given in (18). ⎡ ⎤ (18) ‘be’ ⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ ⎡ ⎤ ⎥ ⎢ ‘pro’ ⎥ ⎢ ⎥ ⎢ ⎥ ⎥ ⎢ ⎢

⎥ ⎥ ⎢ ⎢ ⎥ ⎥ ⎢ ⎢ ⎥ ⎥ ⎢ & ⎢ ⎥ ⎥ ⎢ ⎢ ⎥ ⎥ ⎢ ⎢"

+ ⎥ ⎥ ⎢ ⎢ ⎣ ⎦ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ ‘forest’ ⎥ ⎣ ⎦

As we see, the feature information in (16) does not conflict with the information in (17), so the feature structure in (18) is well-formed. The lexical entry for the third person dual form lava is given in (19). (19)

lava: V

*↑ # = ‘be’ *↑ # = *↑ # = *↑

# =

*↑ # = 3 *↑ "

# = +


237

The lexical entry for lava is only compatible with third person dual subjects that are positively specified for the feature "

* #. Let us turn for a moment to the choice of the attribute label "

for the relevant animacy feature. The label "

seems a natural choice for personal pronouns and first and second person verbs (as in (17–18) above), since they will always refer to humans (setting aside the possibility of personification). However, recall from the previous section that lava can be used with animals (although animals may also trigger partial agreement in which case the verb would be láá). If we maintain the feature "

, lexical entries for animals will have to be specified with a disjunctive feature value for "

: (↑ "

) = + ∨ −. Another option is an analysis where animals lack specification for the feature "

. As will be discussed below, this is not a viable option, since the present analysis appeals to morphological blocking. The feature "

(as other features) should be interpreted as a formal feature. The specific labels that features hold are intended to reflect the relevant linguistic distinctions. This practice is upheld for the purpose of getting the linguistic intuitions across and also for the sake of readability. Another possibility would be to observe linguistic distinctions and label them according to the order in which we discover those distinctions. We could then label the attributes with numbers and the values with alphabetic characters or positive/negative specifications: (↑ 1) = ; (↑ 2) = ; (↑ 3) = +/−; etc. In a sense, then, it does not matter if we choose the label "

, or perhaps /: the point is that there is a distinction here, and we use features as formal tools to capture that distinction. In actual practice, linguists of course make use of features that are as close as possible to those used by traditional grammarians. The lexical entry for láá is given in (20). (20)

láá: V

*↑ # = ‘be’ *↑ # = *↑ # =

As we have seen, the subjects which can cooccur with láá in the partial paradigm are not restricted to a single person or number value. The lexical entry for láá in (20) does not contribute any information which can conflict with any of its possible subjects, as it lacks subject specifications completely. The lexical entries given above cover the following two generalizations: 1) only human subjects fully specified for and

can take verbs from the full agreement paradigm; and 2) all dual and plural subjects can cooccur with

238

IDA TOIVONEN

láá. However, the discussion so far leaves two obvious questions unanswered: First, why don’t human subjects freely cooccur with láá? Second, what prevents singular subjects from occurring with láá? In order to answer these questions, we need to appeal to the notion of morphological blocking, which states that a more specific lexical entry is always chosen over a less specific one, all other things being equal. For a thorough discussion of morphological blocking as it is employed in LFG, see Andrews (1990). All other things being equal, morphological blocking picks out lava over láá. However, conflicting feature information is of course ruled out by Uniqueness: lava cannot take a first person subject, for example. It should now be clear why animals cannot be unmarked for the feature "

: if they were, the dual animal subjects would always have full agreement, as lava is more specific than láá. Now consider two further lexical entries; the entry for lam in (21) and the entry for lii in (22). (21)

lam: V

(22)

lii: V

*↑ # = ‘be’ *↑ # = *↑ # = *↑

# = *↑ # = 1 *↑ "

# = + *↑ # = ‘be’ *↑ # = *↑ # = *↑

# =

The lexical entry for lam in (21) specifies that its subject is first person singular. The entry for lii is specified for a singular subject, but there is no person specification. The form lii is therefore compatible with all singular subjects, but not with dual and plural subjects. Since lii is more specific than láá (it has a number specification), it will be chosen over láá whenever the subject is singular. The entry for lii is not specified for , as it can co-occur with all three persons in the partial paradigm. The sample lexical entries above are all forms of the verb lede ¯ ‘to be’. It should be understood that the forms of other verbs have lexical entries parallel to those seen above. For illustration, the second person dual entry for the verb moonnâd¯ ‘to go’ is given in its past tense form in (23).


(23)

moonaid: V

239

*↑ # = ‘go’ *↑ # = *↑ # = *↑

# =

*↑ # = 2 *↑ "

# = +

By Uniqueness, only second person dual subjects can cooccur with the verb form in (23). One piece of the puzzle is still unaccounted for. Recall from the previous section that human subjects can trigger partial agreement. This is rare, but possible, and so must be taken into account. The entries for lii and láá given above are left sufficiently underspecified to allow first and second person singular pronouns to occur with lii and all dual and plural subjects to occur with láá. It was argued above that the full agreement paradigm is only compatible with NPs with specific readings. If this is the case, it can easily be captured in our current analysis with an additional lexical specification on the full agreement forms. This is illustrated with the lexical entry for cˇ okkáám in (24). (24)

cˇ okkáám: V

*↑ # = ‘sit’ *↑ # = *↑ # = *↑

# = *↑ # = 1 *↑ "

# = + *↑ # = +

The feature *↑ # = + ensures that the subject has a specific reading. If the subject is not specific, this lexical entry cannot be used. The specificity information should perhaps be included in the semantic representation only, rather than in the f-structural information. However, it is included in the f-structure here for the sake of simplicity. In sum, section 2.1 showed that Inari Saami verbs always have overt agreement morphology, and the verb may agree fully or partially. It was further shown that animacy crucially determines which agreement paradigm is employed: inanimate subjects never trigger full agreement. This section has shown how the data of the previous section can be captured by a lexical feature analysis and the adoption of Uniqueness and morphological blocking. Note that these aspects of the analysis are widely employed outside the grammatical framework of LFG: Most

240

IDA TOIVONEN

current theories put heavy emphasis on the lexicon, morphological blocking has been referred to in various syntactic theories, and Uniqueness is a general property of feature logics. 3. Subject agreement In the previous sections, it was assumed without discussion that the verb agrees (fully or partially) with the subject. This claim deserves further attention, as there are some possible objections. One objection is theory-dependent, as some theories do not adopt grammatical functions (for example, subjects and object) as part of their theoretical apparatus. For example, in transformational grammar (for example, Standard Theory or Principles and Parameters), grammatical functions are simply assumed to be convenient labels for elements that are actually formally defined by their phrase structural position (but see the interesting discussions in Davies & Dubinsky 2001). Other theories allow direct reference to grammatical functions; either as primitives as in Relational Grammar, or as derived entities, as in LFG and Head-driven Phrase Structure Grammar (HPSG).3 Here, I will maintain the assumption that reference to subjecthood is legitimate. Readers more comfortable with transformational assumptions can think of ‘subject’ as a given phrase structural position; be it [Spec, IP], [Spec, AgrSP], or some other phrasal position. There is a further objection to the claim that the agreement trigger is the subject, and this objection is theory-independent: even if we allow reference to subjecthood, it is not necessarily the case that agreement is dependent on grammatical function at all. Consider the examples in (25-26).4 (25)

3

Suoi lava tyebbin. they. are.2 there “They (the two of them) are over there.”

In LFG’s Lexical-Mapping Theory (LMT; Bresnan & Kanerva 1989; Bresnan & Moshi 1990; Bresnan & Zaenen 1990), GFs are composed of more primitive features, and so GFs are not themselves primitives of the theory. With LMT, GFs can be derived from argument structure features. However, GFs are still ‘visible’ at a separate syntactic level (f-structure), where they are governed by constraints and principles distinct from those that govern other levels of information (see the appendix). Similarly, the list values of HPSG’s features are dependent on the ordering of items on the list, which in turn is determined by the obliqueness hierarchy. The obliqueness hierarchy makes reference to thematic roles (see Pollard & Sag 1994: Chapter 9, and Manning & Sag 1999). 4 In this section, it does not matter whether the agreement is full or partial. The main point is that the verb always agrees and we are trying to determine what triggers the agreement.


(26)

241

tun. Muste lah I. are.2 you. “I’ve got you.”

In example (25), it is natural to think of the agreement trigger suoi as the subject: it is in the canonical subject position (it appears before the verb), it is the topic, it has the highest thematic role, and it bears nominative case. However, the subject status of tun in (26) is not as clear. The pronoun tun bears nominative case, but it is not in the canonical subject position, it is not the topic and it does not have the highest thematic role: the possessed entity is lower on the thematic hierarchy than the possessor. The postverbal NP in the possessive construction thus displays only two of the prototypical subjecthood characteristics: it agrees with the verb and it bears nominative case. An alternative take on the possessive construction would be to analyze the NP bearing the highest thematic role (muste in (26)) as the subject. A third possibility is, of course, to reject the notion of subjecthood completely. Examples such as (26) thus call into question the claim that the NP which triggers agreement is the subject. In other words, we need to establish exactly what it is that triggers agreement in Inari Saami. It is clear from (26) that agreement is not determined by the highest thematic role. Nor is it determined by the word order: in (25) the verb agrees with the nominal on its left, and in (26) the verb agrees with the nominal on its right. The unmarked word order in Inari Saami is SVO, but the word order is generally quite free. Even so, it is not possible to switch the order of the NPs in possessive constructions such as (26) without significantly changing the meaning. The remaining possibilities are that the verb agrees with the subject, or else that the verb agrees with the NP which bears nominative case. The nominative case hypothesis turns out to be problematic, due to the effect numerals have on case marking. When a noun is modified by a number, it is always in singular and it bears either genitive or partitive case, as illustrated in Table 5. Table 5 shows that the numbers 2–6 select for an accusative/genitive noun and numbers 7 and higher select for a partitive noun. For further discussion of the numerals and case marking in Inari Saami, see Nelson & Toivonen (2003). The verb can agree with a nominal which is assigned a case other than nominative by a numeral. This is shown in (27–28).5

5

The verb ruáttáin is used for animals and kaˇccˇ áin is used for people.

242

IDA TOIVONEN

one reindeer two reindeer three reindeer four reindeer five reindeer six reindeer seven reindeer eight reindeer

ohtâ kyehti kulmâ nelji vittâ kuttâ cˇ iˇccˇ âm käävci

puásui poccuu ” ” ” ” poccud ”

lii lava/láá láá ” ” ” ” ”

⇓ twenty-one reindeer ⇓

kyehtlovohtâ ”

”

... Table 5: Noun modified by numerals (27)

ruáttáin meecist. Kyehti poccuu two reindeer. ran.3 forest. “Two reindeer ran in the forest.”

(28)

Meecist kaˇccˇ áin kyehti almaa. forest. ran.3 two man. “Two men ran in the forest.”

In order to rescue our hypothesis that the verb always agrees with the nominative element, we could try to argue that the verb actually agrees with the number itself, which is indeed in nominative case in examples (27-28). There are four problems with this proposal. First, we saw in the previous section that agreement is influenced by animacy: an inanimate noun triggers partial agreement and an animate noun triggers full agreement. This effect remains when the nominal is modified by a number, as becomes clear when we compare the examples in (27– 28) to the examples in (4) above. In (27–28), poccuu and almaa are animate and the verb is dual. The noun keedgi ¯ in (4), however, is inanimate and the verb is plural. As animacy is a feature of nouns, not numerals, it seems that the nouns and not the numbers agree with the verb.


243

Second, the nominal and not the number is the semantic head, and there is often a one-to-one correspondence between syntactic and semantic heads. A third piece of evidence that the noun, not the number, displays the relevant case marking comes from examples with lexical and semantic case: (29)

lii ennuu purrâmuš. Kuulmâ poccust three. reindeer. is.3 much food. “The three reindeer have a lot of food.”

In (29), the NP following the numeral bears locative case. Importantly, the noun alone displays this case marking and the numeral does not. There is some variation in the case marking on the number across the nominal paradigm. For example, the number in (29) is not in nominative case, unlike the numbers we have seen above. However, the numeral does not carry the crucial locative case, which is what marks the function of the NP in (29). As this is typical for NPs with lexical and semantic case marking, there is no reason to assume that the number is crucial for displaying the case of the full NP. A fourth reason to reject the proposal that the verb actually agrees with the numerals and not the nouns is that the postverbal numerals in (30–31) bear nominative case, and yet they do not trigger agreement: (30)

puurráin vittâ porkanâ. Kyehti poccuu two. reindeer. ate.2 five. carrot. “Two reindeer ate five carrots.”

(31)

Mun oinim kyehti stuorrâ poccuu. I. saw.1 two. big reindeer. “I saw two big reindeer.”

If nominative case is the single factor that is relevant for agreement, we would expect it to be possible for the verb to agree with vittâ porkanâ in (30) and kyehti stuorrâ poccuu in (31). In other words, there should be some optionality in agreement in these examples; the verbs should be able to agree with either NP. However, there is no such optionality; only kyehti poccuu and mun can trigger agreement in (30–31). In an attempt to rescue the hypothesis that verbs always agree with an element bearing nominative case, one might propose that nominative and accusative numbers are homophonous. The idea would then be that vittâ and kyehti in (30– 31) are in accusative case, which happens to be homophonous with nominative

244

IDA TOIVONEN

vittâ vittâ (= ) viidâ ¯

kedgi keedgi ¯ (= ) keedgi ¯

mun muu (= ) muu

Table 6: Numerals and case case for numerals. This is unlikely, as genitive numbers are not homophonous with nominatives, and accusative and genitive normally pattern together morphologically. The proposal that nominative and accusative numerals are homophonous leads to the morphological pattern in Table 6. Table 6 shows that the hypothesis that nominative and accusative numbers are homophonous results in one pattern for nouns and pronouns, and another for numerals. We are forced to resort to this hypothesis if we wish to maintain that nominative numerals trigger agreement, but there is no independent evidence that numerals pattern differently from nominals in the manner illustrated in Table 6. Furthermore, the hypothesis of homophony of accusative and nominative cases in numerals alone is rendered quite unlikely if we also consider the case marking of the words maηgâ ‘many’, kalle ‘how many’, and muád"di ‘a couple, several’. Like the numbers 2–6, the words maηgâ, kalle and muád"di demand that the noun they modify be genitive singular, regardless of whether the NP is a subject or an object. Following our discussion for the numerals, we could posit that these three words are actually heads and they thus show the crucial nominative/accusative case marking. However, like numerals, these words stand in the same form regardless of whether they are part of the subject or the object: (32)

cˇ okkájii tobbeen. Maηgâ almaa many man. sat.3 there “Many men were sitting over there.”

(33)

Mun uáinám maηgâ almaa. I. see.1 many man. “I see many men.”

It is clear that both numbers and the triplet maηgâ, kalle and muád"di demand that the noun they modify bear a specific case (genitive). In order to rescue our hypothesis that the case marking of the entire NP shows up on the modifier, not the noun, we must in addition assume that nominative and accusative case just happen to be homophonous in precisely these groups of words. Recall that


245

conflating nominative/accusative case is not the norm in Inari Saami (see Table 6). The case marking that a word assigns and the case marking that a word bears should be completely independent factors, and it seems a highly unlikely coincidence that numerals and the three words mentioned here would just happen to pattern the same in both regards. I want to conclude this brief discussion of numerals and case marking with a few final comments. I have argued here that NPs with numbers provide an argument against the claim that agreement is directly predictable from nominative case marking. A noun that would normally have nominative case bears another case when modified by a numeral. I have further argued that it is not the case that the numeral then bears the relevant case marking, which can then trigger agreement. Although I have argued against these possibilities, I do not mean to say that numerals are completely irrelevant for agreement. They are in fact crucial in that they provide information about number: all numerals other than the number one ensure that the NP has the number specification (for the number two) and (for all other numbers).6 In sum, there is no one-to-one correspondence between nominative case and agreement marking in Inari Saami. It is also not the case that agreement is determined by word order or semantic role. As neither of these factors can be singled out as the relevant linguistic trigger for agreement, I conclude that direct reference to grammatical functions, or more specifically, to subjecthood, is needed in order to capture the correct generalization which governs Inari Saami agreement. In other words, Inari Saami verbs agree with their grammatical subjects.7 Before turning to the next section, I want to briefly address the issue of subjecthood tests. Although such tests work very well for some languages (see the discussion of Icelandic in Zaenen et al. 1985, for example), these tests are notoriously difficult to apply to Finno-Ugric languages in general (Sands & Campbell 2001). I have not been able to successfully apply any of the traditional tests in Inari Saami – the results do not clearly identify one phrase as the subject. For example, the possessive suffix diagnostic that Magga (1978) applies in North 6

This can be captured formally in the following way: only singular nouns which have nominative case are specified for . All other singular nouns are unspecified for number. The numerals are inserted under a co-head (possibly Det) which is annotated (↑ = ↓), and lexically specified for the feature ↑ = or ↑ = . The other information is specified under an (or ) function:↑ = ‘two’; ↑ = ; etc. 7 For present purposes, we can assume that verbs identify the appropriate syntactic function for each participant. The interested reader is referred to the Appendix for a discussion of how the semantic arguments of a verb are mapped onto syntactic functions in Lexical-Mapping Theory. The appendix specifically refers to some of the Inari Saami data discussed above.

246

IDA TOIVONEN

Saami does not work for Inari Saami, where possessive suffixes are falling out of use completely. Where possessive suffixes are still used in Inari Saami, they do not necessarily corefer with the subject. In fact, they can refer to someone not mentioned at all in the sentence. In order to illustrate the difficulties with syntactic subjecthood tests in Inari Saami, let us consider verbs that correspond to English raising verbs. The verb nevttid¯ ‘look like; seem like’ will be used here. When examining a verb like nevttid, ¯ it may at first appear as if raising shows that the possessor in the tricky possessive construction is the subject (which would go against what is assumed in this paper): (34)

nevtih lemin ud¯dâ Uábistâm ¯ kammuuh. sister.1% seem.3 be. new shoes. “My sister seems to have new shoes.”

In (34), uábistâm, not kammuuh, precedes the supposed raising verb nevtih, but note that nevtih agrees with kammuuh. Recall also that word order is relatively free in Inari Saami and so cannot be taken as a reliable diagnostic of grammatical function. However, if we assume that agreement is a reliable subjecthood test, then (34) in fact indicates that kammuuh, the possessed NP, is the subject. Consider further (35). (35)

nevtih lemin ennuu poccuh. Meecist forest. seem.3 be. much reindeer. “There seem to be a lot of reindeer in the forest.”

The word order in (35) would indicate that meecist is the subject in (35), but we know that word order is not a reliable diagnostic, and there is no other reason to assume that the place expression meecist in (35) is a subject. (The place expression meecist does not bear nominative case, it does not hold the highest thematic role, and it does not agree with the verb.) In fact, if topicalized, any phrase can precede the verb in Inari Saami: this is true in raising expressions as well as other sentences. Again, nevtih agrees with ennuu poccuuh, which is the logical subject, and, I claim, also the grammatical subject. Raising expressions cannot be said to show anything based on their word order. If we instead rely on verbal agreement, these expressions offer support for the analysis adopted in this paper, where subjecthood and agreement go hand in hand: the raising verbs agree with the same NP that the main verb would agree with in a corresponding non-raising sentence. In other words, if anything, Inari


247

1 minä menen

2 3 1 2 3

sinä hän me te he

menet menee menemme menette menevät

Table 7: Nominative pronouns and present paradigm for mennä ‘to go’ Saami raising shows that the current hypothesis is correct in assuming that the possessed phrase is the subject of the possessive construction. There are various potential problems with using data such as (34–35) as subjecthood tests in Inari Saami, so I will not dwell on raising here. I will simply conclude by admitting that I have not found any syntactic diagnostics that work well for determining grammatical functions in Inari Saami. 4. A comparison with Finnish It is instructive to compare the Inari Saami data with data from Finnish. In the previous section, I concluded that verbal agreement is always triggered by grammatical subjects in Inari Saami. This section shows that this is not the case in Finnish, where verbs agree only if several independent requirements are met at the same time. Finnish grammatical functions, subjects especially, have been discussed widely in the literature (see, for example, Hakanen 1972; Hakulinen & Karlsson 1979; Vilkuna 1989; Nelson 1998; Kiparsky 2001; Helasvuo 2001). The purpose of this section is to bring out the main characteristics of agreement in Finnish, and not to provide a full discussion of case and grammatical function. For more detail, examples, and important insights, I refer the interested reader to the works listed above. Finnish verbs agree in three persons and two numbers, as illustrated in Table 7. Three differences between Finnish and Inari Saami are especially relevant here. First, Finnish does not have a grammatical dual. Second, Finnish does not have partial agreement. Third, animacy is not grammmaticalized in Finnish (but see Laitinen & Vilkuna 1993 and references cited there for dialects of Finnish where animacy does seem to play a role). Finnish verbs normally agree with nominative NPs only. Consider examples (36-37), which are taken from Stenberg (1971).

248

IDA TOIVONEN

(36)

Autot ajavat yleensä kovaa moottoriteillä. cars. drive.3 generally hard motorways. “Cars generally drive fast on the motorways.”

(37)

Linja-autoja kulkee nykyisin joka sunnuntai. buses. run.3 nowadays every Sunday “Nowadays, buses run every Sunday.”

In example (36), the subject carries nominative case and the verb agrees with it. In (37), the subject is in partitive case, and the verb has third person singular default agreement. The verb also displays default agreement in existential and possessive constructions, as shown in (38–39), which are taken from Nelson (1998: 55). (38)

Perheeseen syntyi kauniit tytöt. family. born. .3 beautiful. girls. “To the family were born beautiful daughters.”

(39)

Koulussa on uudet opettajat. school. is.3 new. teachers. “The school has new teachers.”

In both (38) and (39), the (post-verbal) subject is nominative, but the verb does not agree. Instead, it displays default third person singular agreement. A further point worth noting is that possessed pronouns in the possessive construction are in accusative case in Finnish: (40)

on meidät. Pekalla Pekka. is.3 us. “Pekka has us.”

In (40), the possessed entity meidät has accusative case and the possessor has adessive case. As exemplified in (26) above, pronominals in the corresponding Inari Saami possessive construction have nominative case. Table 8 summarizes the characteristics of Inari Saami and Finnish agreement. One of the main differences between the two languages follows from the fact that Inari Saami has partial agreement: Inari Saami always displays agreement, but Finnish does not. The correlation between case, agreement marking and grammatical function is less straightforward in Finnish than in Inari Saami. In Inari Saami, the verb


Partial agreement Default agreement Animacy effects Agreement in possessive construction Agreement in existential construction Possessed nouns in nominative case Possessed pronouns in nominative case

249

Inari Saami Finnish √ √ √ √ √ √ √ √

Table 8: Agreement in Inari Saami and Finnish always agrees with the subject, which bears nominative case unless the regular subject case marking is overridden by a case assigned by a numeral. In Finnish, however, the situation is more complicated. In Finnish as well as in Inari Saami the verb never agrees with a non-subject. However, in Finnish the agreement trigger is also required to be in the prototypical subject position and to bear nominative case in order to trigger agreement.8 In other words, several requirements need to be met in order for Finnish verbs to display agreement. The difference between Inari Saami and Finnish is highlighted by a comparison of the possessive construction in the two languages. Compare the Inari Saami sentences in (41) to the Finnish examples in (42). (41)

Inari Saami: a.

puásui já peenuv. Muste lava I. are.3 reindeer. and dog. “I have a reindeer and a dog.”

b.

Muste lah tun. I. are.2 you. “I have you.”

8 There are also some cases where the verb can agree with a partitive subject, if the subject is interpreted as definite.

250 (42)

IDA TOIVONEN

Finnish: a.

Minulla on poro ja koira. I. is.3 reindeer. and dog. “I have a reindeer and a dog.”

b.

Minulla on sinut. I. is.3 you. “I have you.”

Examples (41–42) illustrate two previously mentioned differences between the possessive constructions in the two languages. First, the verb in Inari Saami always agrees with the possessed item, whereas in Finnish it does not. Second, in Inari Saami, the possessed item is always nominative, but in Finnish, it is accusative if it is a personal pronoun and possessed nouns are often partitive, especially if plural. These data are consistent with the conclusion that the verb always agrees with the subject in Inari Saami, and the subject normally has nominative case. The possessed item thus seems to be the subject in Inari Saami. Which argument is the subject in the possessive construction in Finnish? The possessed entity does not seem to be a good candidate for a subject in Finnish: it does not trigger agreement, it does not appear in the prototypical subject position, it is not necessarily in nominative case, and it is not the highest role on the thematic hierarchy (see Nikanne 1990 for the relevance of the thematic hierarchy for subjecthood in Finnish). Moreover, there is evidence which suggests that the possessor is the subject, since the possessor shares the normal discourse properties of subjects (Helasvuo 2001), and it also shares some syntactic properties with other subjects; for example, it can bind anaphoric possessive suffixes (Hakanen 1972; Kiparsky 2001). There are thus good reasons to accept the possessor as the subject in the possessive construction. However, the verb does not agree with the possessor in adessive case; the examples in (43) are ungrammatical. (43)

a.

ja koira. *Minulla olen poro I. is.1 reindeer. and dog.

b.

*Minulla olen sinut. I. is.1 you.

In sum, the verb in the possessive construction does not agree with either argument in Finnish. The verb cannot agree with the possessed NP because that


251

NP is not the subject, and it cannot agree with the possessor because the possessor does not carry nominative case. This again illustrates that several conditions must hold in order for the verb to agree with the subject in Finnish, and if those conditions do not hold, the verb displays default agreement. Although the possessive construction in Inari Saami is superficially very similar to the same construction in Finnish, the two are actually quite different, since they differ in case marking, agreement marking, and grammatical function. It is commonly assumed that possessive constructions that are similar to locative constructions derive historically from a type of locative construction. An expression meaning something like ‘on/with me is a book’ can be reanalyzed to come to mean ‘I have a book’. The possessed item was originally the subject, but it has come to be analyzed as the object, with the possessor taking over the subject role (Comrie 1981). Since this is a common historical development, it is not surprising to find two related languages where one has a grammaticalized possessor subject and the other still treats the possessed item as the subject.9 5. Conclusion This paper has examined verbal agreement marking in Inari Saami. Inari Saami has both full and partial agreement marking, and animacy is an important factor in determining which paradigm is employed. The data are captured in this paper with a lexical analysis cast in LFG. The analysis makes crucial use of lexical blocking. An important question posed in this paper concerns the agreement trigger. It was argued in Section 3 that verbal agreement in Inari Saami is not dependent on word order, thematic roles or case marking. I concluded that reference to grammatical functions, specifically subjects, is necessary in order to understand the verbal agreement system in Inari Saami. The Inari Saami agreement system was then briefly compared to agreement marking in Finnish. The Inari Saami system at first seems more intriguing and complex than the Finnish one, as Inari Saami displays the cross-linguistically unusual characteristic of partial agreement. Finnish instead has full and default agreement, which is not rare cross-linguistically. However, a closer examination reveals that Finnish agreement is in a sense more complex than Inari Saami agreement. Inari Saami verbs always agree, and they always agree with the subject. In Finnish on the other hand, several conditions must be met in order for a 9 Comrie discusses Maltese and Classical Arabic (1981, Section 10.4), and the parallels to Finnish and Inari Saami are striking. See also Freeze (1992) for an interesting discussion of the parallels between locative, possessive and existential constructions.

252

IDA TOIVONEN

−r

−o +o

+r

θ θ

Table 9: LMT argument-function mapping verb to inflect for agreement. The generalization for Finnish is, roughly, that the verb only agrees with the grammatical subject when the subject is in its prototypical phrase-structural position and when it bears nominative case. In a comparison of Finnish and Inari Saami, the possessive construction is especially noteworthy: in Inari Saami, the verb agrees with the nominative possessed NP, whereas the verb displays default agreement in Finnish. I suggested that this is because the possessor has been grammaticalized as the subject in Finnish, but not in Inari Saami. As the Finno-Ugric languages tend to have rich morphological case marking, it is common for scholars interested in grammatical relations to focus on case. However, in this paper I hope to have shown that examining agreement marking may also prove illuminating. Appendix: Lexical Mapping Theory We saw in section 2.2 that LFG adopts a level of syntactic structure which makes direct reference to syntactic functions, such as and . The generalization that governs Inari Saami agreement is therefore straightforwardly accounted for within this framework. In early LFG, syntactic functions were assumed to be primitive syntactic building blocks in the grammar (see Bresnan 1982, and others). Modern LFG adopts Lexical Mapping Theory (LMT; Bresnan & Kanerva 1989; Bresnan & Moshi 1990; Bresnan & Zaenen 1990; Alsina 1996; Bresnan 2001, and others), where the basic argument functions are defined in terms of the features [±r] and [±o]. The feature [+r] singles out the grammatical functions that are semantically restricted. The syntactic functions that are marked [−r] are not semantically restricted; in fact, they have the option of being associated with no semantic role at all. The feature [+o] refers to the objective functions, i.e., functions that complement transitive V or P. The basic argument functions are thus grouped as shown in Table 9. The patientlike roles are classified as [−r], secondary patientlike roles are classified as [+o] and other semantic roles are [−o]. This classification can be overridden by lexical specifications. LMT provides a theory for the mapping between a(rgument)-structure and

253


f-structure. A-structure is a syntactic level of representation which links the lexical semantics of predicates and their arguments to f-structure. The arguments of a given predicate are ordered for their relevant prominence according to the thematic hierarchy: (44)

Thematic Hierarchy: agent beneficiary experiencer/goal instr patient/theme loc

The mapping between a-structure and f-structure is governed by the principles in (45), adapted from Bresnan (2001: 311). The logical subject is defined as the most prominent semantic role of a predicator. (45)

Mapping Principles: a.

Subject roles: i.

ii. b.

The logical subject marked [−o] is mapped onto initial in the a-structure; otherwise:

when

The semantic role marked with [−r] is mapped onto .

Other roles are mapped onto the lowest compatible function (according to the following partial ordering: , θ θ ).

One of the results of LMT is that the argument which is the highest on the thematic hierarchy will be the , unless some lexical specification interferes. For a simple example, let us consider the Inari Saami verb puurrâd¯ ‘to eat’. (46) puurrâd¯

The first argument of puurrâd¯ is an agent and thus not a patientlike role. It is specified [−o] and gets mapped onto by mapping principle (45ai). The second argument is a patientlike role. It is classified as [−r] and gets mapped onto the object function.

254

IDA TOIVONEN

For the sake of concreteness, I assume the following straightforward nominative-accusative default case marking principles for Inari Saami:10 (47)

Default case marking: a. b.

is assigned case. is assigned case.

The principles in (47) can be overridden by lexical specifications and case assigned by numerals. In example (46), the will bear nominative case, unless there is a numeral within the NP. I set aside here the interesting issues that are raised by semantic case. It should now be clear how LMT together with the principles in (47) account for simple examples involving verbs such as puurrâd. ¯ We will now consider a more intricate case, namely the possessive construction, exemplified by (26) above, repeated here as (48). (48)

tun. Muste lah I. are.2 you.

“I’ve got you.”

This is an interesting construction, as the subject is the possessed item, which is less prominent thematically than the possessor (see Mohanan 1994:177-182, for example). Any analysis of the Inari Saami possessive construction must account for the generalizations in (49). (49)

• The possessed item is the subject. • The possessor must precede the verb. • The possessed item bears nominative case. • The possessor bears locative case.

Recall that the possessive construction parallels the existential construction in Inari Saami: the possessor mirrors the properties of the location and the possessed item mirrors the theme. I will therefore adopt an analysis where the possessive construction is an instance of the existential construction, which is in 10 This is by no means the only possible way to handle morphological case marking in LFG. In fact, there is a rich literature on case in this framework, see Zaenen et al. (1985); Mohanan (1994); and Nordlinger (1998), for example.


255

turn a kind of locative construction (see Freeze 1992 for a discussion of crosslinguistic parallels between existentials, locatives and possessives).11 The argument structure for lede ¯ ‘to be’ will be mapped out as in (50). (50) lede ¯

theme location < x y > [−r] [−o]

The theme is a patientlike role and is specified [−r]. The location is not a patientlike role and so receives the specification [−o]. As [−o] is not initial in the a-structure, [−r] is mapped onto the subject role. The lowest compatible function for [−o] is oblique, and so the theme/possessed will be mapped onto subject and the location/possessor will be mapped onto an oblique role. The mapping in (50) follows from the independently motivated principles of , and it accounts for most of the characteristics of the possessive/existential construction. However, we still must account for the trait that distinguishes possessives/existentials from other locatives: the oblique NP precedes the verb. This can be done lexically. Consider the lexical entry for existential/possessive lah in (51). (51)

lah: V

*↑ # = ‘be * , #’ *↑ # = *↑ # = *↑

# = *↑ # = 2 *↑ "

# = + *↑ # =c *↑ # *↑ # =

In sum, this analysis of the possessive/existential construction assumes the oblique must be a topic and thus in topic position, which is at the beginning of the clause (for concreteness, we can assume that this position is [Spec, CP]). Finally, possessive/existential lede ¯ specifies that its bears locative case.

11 Thanks to an anonymous reviewer for helpful comments on the struction.

account of this con-

256

IDA TOIVONEN

References Alsina, Alex. 1996. The Role of Argument Structure in Grammar. Stanford, Calif.: CSLI Publications. Andrews, Avery. 1990. “Unification and morphological blocking”. Natural Language and Linguistic Theory 8:4.507–558. Bresnan, Joan. 1982. “Control and complementation”. Linguistic Inquiry 13.343–434. . 2001. Lexical-Functional Syntax. Oxford: Blackwell. ˆ A case Bresnan, Joan & Jonni Kanerva. 1989. “Locative inversion in Chichewa: study of factorization in grammar”. Linguistic Inquiry 20:1.1–50. Bresnan, Joan & Lioba Moshi. 1990. “Object asymmetries in comparative Bantu syntax”. Linguistic Inquiry 21.147–85. Bresnan, Joan & Annie Zaenen. 1990. “Deep unaccusativity in LFG”. Grammatical Relations: A cross-theoretical perspective ed. by Katarzyna Dziwirek, Patrick Farrell & Errapel Mejias-Bikandi, 45–57. Stanford, Calif.: CSLI Publications. Comrie, Bernard. 1981. Language Universals and Linguistic Typology. Chicago: University of Chicago Press. Dalrymple, Mary, ed. 2001. Lexical Functional Grammar. (= Syntax and Semantics 34.) New York: Academic Press. Dalrymple, Mary & Ronald M. Kaplan. 2000. “Feature indeterminacy and feature resolution”. Language 76.759–798. Dalrymple, Mary, Ronald M. Kaplan, John T. Maxwell, III & Annie Zaenen, eds. 1995. Formal Issues in Lexical-Functional Grammar. Stanford, Calif.: CSLI Publications. Davies, William D. & Stanley Dubinsky, eds. 2001. Objects and Other Subjects: Grammatical functions, functional categories and configurationality. Dordrecht: Kluwer. Freeze, Ray. 1992. “Existentials and other locatives”. Language 68:3.553–595. Hakanen, Aimo. 1972. “Normaalilause ja eksistentiaalilause”. Sananjalka 14.36–76. Hakulinen, Auli & Fred Karlsson. 1979. Nykysuomen lauseoppia. Helsinki: Suomalaisen Kirjallisuuden Seura. Helasvuo, Marja-Liisa. 2001. Syntax in the Making: The emergence of syntactic units in Finnish conversation. Amsterdam and Philadelphia: John Benjamins.


257

Kaplan, Ronald M. & Joan Bresnan. 1982. “Lexical-Functional Grammar: A formal system for grammatical representation”. The Mental Representation of Grammatical Relations ed. by Joan Bresnan, 173–281. Cambridge, Mass.: MIT Press. Keenan, Edward. 1975. “Some universals of passive in Relational Grammar”. Papers from the Eleventh Annual Meeting of the Chicago Linguistic Society 11. 340–352. Chicago: Chicago Linguistic Society. Kiparsky, Paul. 2001. “Structural case in Finnish”. Lingua 111.315–376. Laitinen, Lea & Maria Vilkuna. 1993. “Case-marking in necessive constructions and split intransitivity”. Case and Other Functional Categories in Finnish Syntax ed. by Anders Holmberg & Urpo Nikanne, 23–48. Berlin: Mouton de Gruyter. Magga, Ole Henrik. 1978. Lokative Læt-Sætninger i Samisk. Kautakeino, Norway: Sámi Institut’ta. Manning, Christopher & Ivan Sag. 1999. “Dissociations between argument structure and grammatical relations”. Lexical and Constructional Aspects of Linguistic Explanation ed. by Gert Webelhuth, Jean-Pierre Koenig & Andreas Kathol, 63–78. Stanford, Calif.: CSLI Publications. Mohanan, Tara. 1994. Argument Structure in Hindi. Stanford, Calif.: CSLI Publications. Nelson, Diane. 1998. Grammatical Case Assignment in Finnish. New York: Garland. Nelson, Diane & Ida Toivonen. 2003. “Counting and the grammar: Numerals in Inari Sami”. Generative Approaches to Finnic and Saami Linguistics ed. by Diane Nelson & Satu Manninen, 321–341. Stanford, Calif.: CSLI Publications. Nikanne, Urpo. 1990. Zones and Tiers: A study of thematic structure. Helsinki: Suomalaisen Kirjallisuuden Seura. Nordlinger, Rachel. 1998. Constructive Case. Evidence from Australian Languages. Stanford, Calif.: CSLI Publications. Pollard, Carl & Ivan A. Sag. 1994. Head-Driven Phrase Structure Grammar. Stanford, Calif.: CSLI Publications. Portner, Paul. 2002. “Topicality and (non-)specificity in Mandarin”. Journal of Semantics 19.275–287. Sammallahti, Pekka. In preparation. “Course reader”. Oulu University.

258

IDA TOIVONEN

Sands, Kristina & Lyle Campbell. 2001. “Non-canonical subjects and objects in Finnish”. Non-Canonical Marking of Subjects and Objects ed. by Alexandra Y. Aikhenvald, R.M.W. Dixon & Masayuki Onishi. (= Typological Studies in Language 46.), 251–305. Amsterdam and Philadelphia: John Benjamins. Stenberg, Anne-Marie. 1971. Finsk satslära. Helsinki: Söderström & Co. Vilkuna, Maria. 1989. Free Word Order in Finnish. Helsinki: Suomalaisen Kirjallisuuden Seura. Zaenen, Annie, Joan Maling & Höskuldur Thráinsson. 1985. “Case and grammatical functions: The Icelandic passive”. Natural Language and Linguistic Theory 3.441–483.

SAAMI LINGUISTICS BIBLIOGRAPHY COMPILED BY ANITA SZAKAY, MICHAEL RIESSLER, IDA TOIVONEN, DIANE NELSON, ZITA MCROBBIE-UTASI & NINA WIDJAJA

Abercromby, John. 1895. “The earliest list of Russian Lapp words”. Journal de la Société Finno-Ougrienne 13:2.1–8. Aikio, Ante. 2004. “An essay on substrate studies and the origin of Saami”. Etymologie, Entlehnungen und Entwicklungen: Festschrift für Jorma Koivulehto zum 70. Geburtstag ed. by Irma Hyvärinen, Petri Kallio & Jarmo Korhonen. (= Mémoires de la Société Néophilologique de Helsinki 63.), 5–34. Helsinki: Société Néophilologique. . 2006. “On Germanic-Saami contacts and Saami prehistory”. Journal de la Société Finno-Ougrienne 91.9–55. Aikio, Marjut. 1986. “Some issues in the study of language shift in the Northern Calotte”. Journal of Multilingual and Multicultural Development 7:5.361– 378. . 1988. Saamelaiset kielenvaihdon kierteessä: Kielisosiologinen tutkimus viiden saamelaiskylän kielenvaihdosta 1910-1980 [The Cycle of Language Shift among the Sami: A sociolinguistic study of language shift in five Sami villages, 1910-1980]. (= Suomalaisen Kirjallisuuden Seuran Toimituksia 479.) Oulu: Suomalaisen Kirjallisuuden Seura. . 1991. “The Sami language: Pressure of change and reification”. Journal of Multilingual and Multicultural Development 12:1/2.93–103. . 1992. “Are women innovators in the shift to a second language? A case study of Reindeer Sami women and men”. International Journal of the Sociology of Language 94.43–61. Aikio, Samuli. 1971. “Mitä nimi maksaa? Nuorgamin porvarin ‘kauppakirjan’ nimistöä”. Kalevalaseuran Vuosikirja 51.28–35. . 1987. “Yksi mieli, yksi kieli: Pohjoissaamen kirjakielen yhtenäistyminen”. Virittäjä 91:4.457–490. . 1993. Sámegiel loddenamahusat. Helsinki: Ruovttueatnan gielaid dutkanguovddáš.

260

SAAMI LINGUISTICS BIBLIOGRAPHY

Äimä, Frans. 1902. “Tietoja Inarin kalastajalappalaisten rakennuksia”. Virittäjä 6.33–38; 61–64. . 1906a. “Die Hypothese von einem postkonsonantischem Wechsel k ∼G, t ∼D, p ∼B im Urlappischen”. Finnisch-Ugrische Forschungen 6.181–211. . 1906b. “Eräs lappalainen astevaihtelutapaus”. Journal de la Société FinnoOugrienne 23.1–23. . 1908. “Lappalaisia lainasanoja suomen murteissa”. Journal de la Société Finno-Ougrienne 25.1–64. . 1918a. Phonetik und Lautlehre des Inarilappischen. I. Beobachtungsphonetik und deskriptive Lautlehre. (= Mémoires de la Société FinnoOugrienne 42.) Helsinki: Suomalais-Ugrilainen Seura. . 1918b. Phonetik und Lautlehre des Inarilappischen. II. Instrumentale Versuche und Messungen. (= Mémoires de la Société Finno-Ougrienne 43.) Helsinki: Suomalais-Ugrilainen Seura. . 1919a. Astevaihtelututkielmia. 1. osa: Lapin murteiden välisiä konsonanttivaihteluita. (= Mémoires de la Société Finno-Ougrienne 45.) Helsinki: Suomalais-Ugrilainen Seura. . 1919b. “Eine Gruppe von vokalwechselfällen im Inarilappischen”. Finnisch-Ugrische Forschungen 14.1–11. . 1922. “Prof. Wiklundin viimeisten astevaihtelututkimusten johdosta”. Virittäjä 26.1–29. Äimä, Frans & Björn Collinder. 1928. “Zur Lautkombinationslehre des Lappischen”. Finnisch-Ugrische Forschungen 19.36–81. Äimä, Frans & Toivo Immanuel Itkonen. 1918. “Jacob Fellmanin muistiinpanot Sompion ja Kuolajärven lapin murteista”. Journal de la Société FinnoOugrienne 30.1–91. Alvre, Paul. 1991. “Anmerkungen zum qualitativen Stufenwechsel mit besonderer Berücksichtigung der alten estnischen Schriftsprache”. Linguistica Uralica 27:4.241–250. Anderson, Hans & Olavi Korhonen. 1994. Vad betyder namnen? Luleå, Jokkmokk, Sarek: en liten ordlista. 2. uppl. (2nd rev. ed.). Jokkmokk, Sweden: H. Anderson. Anderson, Myrdene. 1984. “Proper names, naming, and labeling in Saami”. Anthropological Linguistics 26:2.186–201. Atányi, István. 1942. “A lapp melléknevek attributív-predikatív alakpárjainak kérdéséhez”. Finnugor értekezések 6.


261

Bakro-Nagy, Marianne. 1996. “Über die lappisch-samojedischen Sprachkontakte: Auf Grund des Uralischen Etymologischen Wörterbuchs”. In Larsson 1996. 11–25. Bals, Berit Anne. 2004. “The acquisition of grade alternation in North Saami”. Nordlyd: Tromsø University Working Papers on Language and Linguistics 32:1.1–27. Bańczerowski, Jerzy. 1967. Alternacja konsonantyczna we wschodniolaponskim. Ph.D. thesis, Adam Mickiewicz University, Poznań. . 1969. Konsonantenalternation im Ostlappischen unter dem Aspekt der Verstärkung-Lenierung. Versuch einer strukturell-phonetischen Analyse. (= Prace wydział u filogicznego. Seria filologia ugrofińska 1.) Poznań: Adam Mickiewicz University. Bartens, Hans-Hermann. 1979. “Zur Verwendung der Modalverben im Lappischen”. Festschrift für Wolfgang Schlachter zum 70. Geburtstag ed. by Christoph Glaser & Janos Pusztay, 25–48. Wiesbaden: Harrassowitz. . 1980. Die Verwendung von Potential und Konditional im Lappischen. (= Mémoires de la Société Finno-Ougrienne 177.) Helsinki: SuomalaisUgrilainen Seura. . 1981. “Der Finalsatz im Lappischen”. Congressus Quartus Internationalis Fenno-Ugristarum: Budapestini Habitus 9.-15. Septembris 1975 ed. by Gyula Ortutay, Gábor Bereczki & János Gulya. 3. 325–330. Akadémiai Kiadó. . 1986. “Die Stellung der Supinumformen im lappischen Sprachsystem”. Festschrift für Istvan Futaky Zum 60. Geburtstag am 12. Mai 1986 ed. by Wolfgang Veenker, 17–40. Hamburg: Helmut Buske. . 1989. Lehrbuch der saamischen (lappischen) Sprache. Hamburg: Helmut Buske. . 1996. “Zur Verwendung des Imperativs im Saamischen”. In Larsson 1996. 27–41. Bartens, Hans-Hermann & Karl-Heinz Best. 1997. “Word-length distribution in Sami texts”. Journal of Quantitative Linguistics 4:1–3.45–52. Bartens, Raija. 1962. “Über die Illative tasä, tasi im Lulelappischen”. Commentationes Fenno-Ugricae in honorem Paavo Ravila. (= Mémoires de la Société Finno-Ougrienne 125.), 171–175. Helsinki: Suomalais-Ugrilainen Seura. . 1971. “Zur Kongruenz des lappischen Adjektivattributs”. FinnischUgrische Forschungen 39.31–40. . 1972. Inarinlapin, merilapin ja luulajanlapin kaasussyntaksi. (= Mémoires de la Société Finno-Ougrienne 148.) Helsinki: Suomalais-Ugrilainen Seura.

262


. 1974a. “Inarilapin, merilapin ja luulajanlapin post-ja prepositiorakenteiden käytöstä”. Journal de la Société Finno-Ougrienne 73.156–167. . 1974b. “Lapin ablatiivista ja sen taustasta”. Journal de la Société FinnoOugrienne 73.138–155. . 1978. Synteettiset ja analyyttiset rakenteet lapin paikanilmauksissa. (= Mémoires de la Société Finno-Ougrienne 166.) Helsinki: SuomalaisUgrilainen Seura. Bátori, István. 1976. “Miért nincs a lapp nyelvben magánhangzó-harmónia?” Nyelvtudományi Közlemények 78.273–283. Bergsland, Knut. 1942. “Det samiske slektskaps- og svogerskapsordsystem”. Norsk tidsskrift for sprogvidenskap 13.148–198. . 1943. Røros-samiske tekster. Oslo: A. W. Brøggers Boktrykkeri. . 1945. “L’alternance consonantique date-t-elle du lapon commun?” Festskrift til Konrad Nielsen på 70-årsdagen. (= Studia Septentrionalia 2.), 1–53. Oslo: Aschehoug. . 1946. Røros-lappisk grammatikk. Oslo: Instituttet for Sammenlignende Kulturforskning. . 1948. “Om språket i den svensk-samiske ABC fra 1726”. Lapponica. (= Studia Septentrionalia 3.), 40–44. Oslo: Aschehoug. . 1950a. “Norwegian research on the language and folklore of the Lapps; Part 1: Language”. Journal of the Royal Anthropological Institute of Great Britain and Ireland 80:1/2.79–88. . 1950b. “Remarques sur les pronoms démonstratifs lapons”. Commentationes Fenno-Ugricae in honorem Y. H. Toivonen. (= Mémoires de la Société Finno-Ougrienne 98.), 27–35. Helsinki: Suomalais-Ugrilainen Seura. . 1952. “Hvordan den nye samiske rettskrivningen ble til”. Sameliv 1.28–52. . 1953a. Grammatisk innföring i det fellessamiske skriftspråk. Oslo: Kirkeog undervisningsdepartement. Bergsland, Knut, ed. 1953b. Lapponica. (= Studia Septentrionalia 3.) Oslo: Aschehoug. . 1953c. Numeral constructions in Lapp. (= Studia Septentrionalia 5.) Oslo: Aschehoug. . 1958a. Mui’talusat, cáliiga Lars Haetta ja Anders Baer. Oslo: Universitetsforlaget.


263

Bergsland, Knut. 1958b. “Mui’talusat Guov’dagaeno dilálasvuodaid birra åvdal d.j. 1850. Beretninger om forholdene i Kautokeino før år 1850”. Mui’talusat, cáliiga Lars Haetta ja Anders Baer ed. by Knut Bergsland, 7–8. Oslo: Universitetsforlaget. . 1960. “Lapin kieltoverbin taivutuksesta”. Virittäjä 64.327–330. . 1961. Samisk grammatikk med øvelsesstykker. Oslo: Kirke- og undervisningsdepartement. . 1962. “The Lapp dialects south of Lappland”. Commentationes FennoUgricae in honorem Paavo Ravila. (= Mémoires de la Société FinnoOugrienne 125.), 27–39. Helsinki: Suomalais-Ugrilainen Seura. . 1963. “Suomalais-skandinaavinen *rebäs kettu”. Virittäjä 2.149–154. . 1964. “Suomen soihtu lapin valossa”. Virittäjä 2.144–147. . 1965. “Finno-Skandinavian rebas fox”. Norsk tidsskrift for sprogvidenskap 20.243–248. . 1967. “Lapp dialectal groups and problems of history”. Lapps and Norsemen in Olden Times ed. by Ingeborg Hoff et al., 32–53. Oslo: Instituttet for sammenlignende kulturforskning. . 1968. “The grouping of Lapp dialects as a problem of historical linguistics”. Congressus secondus internationalis Fenno-Ugristarum Helsingiae habitus ed. by Paavo Ravila et al. 1, 77–85. Helsinki. . 1973. “Simplification of the Finno-Ugric transcription: Lapp”. FUtranskription yksinkertaistaminen. Helsinki: Suomalais-Ugrilainen Seura. . 1976. Lappische Grammatik mit Lesestücken. (= Veröffentlichungen der Societas Uralo-Altaica 11.) Wiesbaden: Otto Harrassowitz. . 1977. “Saamen kieli ja naapurikielet”. Virittäjä 81.1–11. . 1983. “Southern Lapp and Scandinavian quantity patterns”. Symposium saeculare Societatis Fenno-Ugricae. (= Mémoires de la Société FinnoOugrienne 185.), 73–87. Helsinki: Suomalais-Ugrilainen Seura. . 1994. Sydsamisk grammatikk. Karasjok, Norway: Davvi Girji. . 1995. “Bidrag til sydsamenes historie”. Senter for samiske studier, University of Tromsø. . 1998. “Two cases of language endangerment: Aleut and Sámi”. Papers From the International Symposium on Endangered Languages, Tokyo, November 18-20 ed. by Kazuto Matsumura, ICHEL Linguistic Studies Vol. 1. 33–48. Tokyo: Hituzi Shobo. Bergsland, Knut & Lajla Mattsson Magga. 1993. Åarjelsaemien-daaroen baakoegærja / Sydsamisk-norsk ordbok. Indre Billefjord, Norway: Idut. ¯

264


Bergsland, Knut & Hans Vogt. 1962. “On the validity of glottochronology”. Current Anthropology 3:2.115–154. Beronka, Johan. 1922. Syntaktiske iagttagelser fra de finske dialekter i Vadsø og Porsanger. (= Videnskapsselskapets skrifter. II, Hist.-filos. klasse 1921 10.) Kristiania [Oslo]: Jacob Dybwad. . 1925. Iagttagelser fra orddannelses- og formlæren i de finske dialekter i Vadsø og Porsanger. (= Skrifter / Det norske videnskaps-akademi i Oslo. II, Hist.-filos. klass; 1925 1.) Oslo: Jacob Dybwad. . 1937a. “Lappische Kasusstudien: zur Geschichte des KomitativInstruktivs und des Genitivs im Lappischen”. Oslo Etnografiske Museums skrifter 2:2.56–128. . 1937b. “Lappische Kasusstudien: zur Geschichte des KomitativInstruktivs und des Genitivs im Lappischen 2”. Oslo Etnografiske Museums skrifter 2:3.132–225. . 1937–1940. Lappische Kasusstudien. Oslo: A. W. Brøggers Boktrykkeri. Blomstedt, Oskar Aukusti Frithiof. 1869. Halotti beszéd ynnä sen johdosta wertailevia tutkimuksia Unkarin, Suomen ja Lapin kielissä: yli-opistollinen väitöskirja. Ph.D. thesis, Helsinki University. Boon, Inez. 1968. “Bilingualism and education in Lappish speaking districts”. Inter-Nord 10.285–292. Börretzen, J. 1966. Liten samisk grammatik. Trondheim: Norges samemisjons forlag. Bradean-Ebinger, Nelu & Hans-Jürgen Hube. 1991. Sprachkontakte und Zweisprachigkeit in Fennoskandinavien: Soziolinguistische Aspekte der Zweisprachigkeit im nordlichen Areal. Budapest: Akadémiai Kiadó. Bull, Tove. 1990. “The influence of multilingualism on a Northern Norwegian Dialect”. Learning, Keeping, and Using Language: Selected papers from the 8th World Congress of Applied Linguistics, Sydney, 16-21 August 1987 ed. by M. A. K. Halliday, John Gibbons & Howard Nicholas. 51–61. Amsterdam and Philadelphia: John Benjamins. . 1991. “Språklig identitet i ei nordnorsk fjordsamebygd”. Danske Folkemål 33.23–35. . 1992. “A contact feature in the phonology of a northern Norwegian dialect”. Language Contact: Theoretical and empirical studies ed. by Ernst Håkon Jahr, 17–36. Berlin: Mouton de Gruyter. . 1993. “Forholdet mellom norsk, finsk og samisk i skolen”. Mål og makt 22:3.3–9.


265

. 1995. “Language maintenance and loss in an originally trilingual area in North Norway”. International Journal of the Sociology of Language 115.125– 134. . 2004. “Samisk-nordisk språkkontaktforskning - eit statusoversyn”. In Kusmenko 2004b, 183–198. Bye, Patrik. 1992. “A phonological perspective on gradation”. BA thesis, University of Reading. . 1997a. “A generative perspective on ‘overlength’ in Estonian and Saami”. Estonian Prosody: Papers from a Symposium ed. by Ilse Lehiste & Jaan Ross. 36–70. Tallinn: Eesti Keele Instituut. . 1997b. “Representing overlength”. Phonology in Progress – Progress in Phonology ed. by Geert Booij & Jeroen van de Weijer. (= HIL Phonology Papers 3.), 61–101. The Hague: Holland Academic Graphics. . 2001. Virtual Phonology. Rule Sandwiching and Multiple Opacity in North Saami. Ph.D. thesis, University of Tromsø. Castrén, Matthias Alexander. 1839. De affinitate declinationum in lingua Fennica, Esthonica et Lapponica. Helsingfors: Frenckell. . 1845. Vom Einfluss des Accents in der lappländischen Sprache. St. Petersburg: Kaiserliche Academie der Wissenschaften. . 1858. “Om accentens inflytande i Lappska språket”. Nordiska Resor och Forskningar V: Smärra Afhandlingar 62–125. Cheboksarov, N. N. 1962[1952]. “Questions concerning the origins of the FinnoUgrian language group”. Studies in Siberian Ethnogenesis. Anthropology of the North: Translations from Russian Sources 2 ed. by Henry N. Michael, 207–219. University of Toronto Press for The Arctic Institute of North America. [Translated from Sovetskaya etnografiya 1: 36-50]. Christiansen, Reidar. 1950. “Norwegian research on the language and folklore of the Lapps; part II: Mythology and folklore”. Journal of the Royal Anthropological Institute of Great Britain and Ireland 80:1.89–95. Collinder, Björn. 1928a. “Die Wörter für fünf, sechs und sieben im Lappischen”. Festskrift til Rektor J. Qvigstad; 1853–4 April 1928. (= Tromsø Museums Skrifter 2.), 356–374. Oslo: A.W. Brøggers Boktrykkeri. . 1928b. “Kantasuomen s ∼ h: kantalapin s ∼ š toisen tavun vokaalin jäljessä”. Virittäjä 32.237–251. . 1928c. “Über den Begriff der Quantität mit besonderer Rücksicht auf die lappische Sprache”. Journal de la Société Finno-Ougrienne 42.1–24.

266


. 1929. Über den finnisch-lappischen Quantitätswechsel. I. Uppsala: Almqvist & Wiksell. . 1932. Lapparna, deras kultur och arbetsliv. (= Studentföreningen Verdandis Småskrifter 352.) Stockholm: Albert Bonniers. . 1934. Indo-uralisches Sprachgut: die Urverwandtschaft zwischen der indoeuropäischen und der uralischen (finnischugrisch-samojedischen) Sprachfamilie. (= Uppsala universitets årsskrift 1.) Uppsala: Uppsala universitet. . 1938. Lautlehre des Waldlappischen Dialektes von Gällivare. (= Mémoires de la Société Finno-Ougrienne 74.) Helsinki: Suomalais-Ugrilainen Seura. . 1940. Jukagirisch und Uralisch. (= Uppsala universitets årsskrift 8.) Uppsala: Uppsala Universitet. . 1942. Lappische Sprachproben aus Härjedalen: nach Nils Axman, Torkel Larsson, Lars Renfelt, Elisabeth Rensberg und Nils Rutfjäll: mit einem Abriss der Nominal- und Verbalbeugung. Uppsala: Almqvist & Wiksell. . 1943. Lappisches Wörterverzeichnis aus Härjedalen. Uppsala universitets årsskrift, Uppsala: Uppsala universitet. . 1943–1945. The Affinities of Lapp. Uppsala: Språkvetenskapliga sällskapet i Uppsala. . 1949a. The Lappish Dialect of Jukkasjärvi. A Morphological Survey. (= Skrifter utgivna av K. Humanistiska Vetenskaps-Samfundet i Uppsala 37.) Uppsala: Almqvist & Wiksell. . 1949b. The Lapps. New York: Princeton University Press. . 1952. “Three degrees of quantity”. Studia Linguistica 5.28–43. . 1954. Proto-Lappish and Samoyed. Uppsala: Språkvetenskapliga sällskapet i Uppsala. . 1955. Fenno-Ugric vocabulary: An etymological dictionary of the Uralic languages. (= A Handbook of the Uralic languages 1.) Stockholm: Almqvist & Wiksell. . 1957. Survey of the Uralic languages. Stockholm: Almqvist & Wiksell. . 1960. Comparative Grammar of the Uralic languages. Stockholm: Almqvist & Wiksell. . 1962. Introduktion till de uraliska språken: Finskan och dess frändespråk. Stockholm: Natur och kultur. . 1964. Ordbok till Sveriges lapska ortnamn. Uppsala: Kungl. ortnamnskommissionen. . 1965a. “Hat das Uralische Verwandte? Eine sprachvergleichende Untersuchung”. Acta Societatis linguisticae Upsaliensis. Nova series. 1:4.110–180.


267

. 1965b. An Introduction to the Uralic Languages. Berkeley: University of California Press. . 1967. “Om nordiska topografiska termer i lapskan”. Namn och Bygd 55.155–168. Corson, David. 1995. “Norway’s ‘Sami Language Act’: Emancipatory implications for the world’s aboriginal peoples”. Language in Society 24:4.493–514. Dahlstedt, Karl-Hampus, Åke Hansson & Siiri Sahlman-Karlsson. 1982. Språkhistoria och språkkontakt i Finland och Nord-Skandinavien: Studier tillägnade Tryggve Sköld den 2 november 1982 [= Language History and Language Contact in Finland and Northern Scandinavia: Studies dedicated to Tryggve Sköld 2. November 1982]. Umeå: Skytteanska samfundet, Länsmuseet. Décsy, Gyula. 1987. “Lappisch eadne und das Suomi-Problem: Hajdú Péternek december 27-ére”. Ural-Altaic Yearbook 59.45–48. Delaporte, Yves & Michele Roue. 1974. “Le Système nominal des Lapons Kautokeino”. Anthropos: International Review of Ethnology and Linguistics 69.843–857. Demant-Hatt, Emilie, ed. 1918–1919. Lappish Texts, by Johan Turi and Per Turi. (= Det Kongelige Danske Videnskabernes Selskabs Skrifter 7.) Copenhagen: A.F. Høst och Søn, Kungliga Hof-boghandel. Donner, Otto. 1876. Lieder der Lappen. Helsingfors: Buchdruckerei der Finnischen Litteratur-Gesellschaft. Eira, John Henrik. 1984. “Notes on the sound system of the dialect of Skatnie”. Riepmoˇcála. Essays in Honour of Knut Bergsland ed. by Bernt Brendemoen, Even Hovdhaugen & Ole Henrik Magga. Oslo: Novus. Eira, Nils Isak. 1984. Boazobargi giella. Kautokeino, Norway: Sámi instituhtta. Ejerhed, Eva. 1978. “Strukturella aspekter på sydsamisk tvåspråkighet”. Tvåspråkighet. Elsvatn, Leif. 1989. Samiske stedsnavn i Hattfjelldal. Mosjøen, Norway: Sijti Jarnge / Joh. Petersen bokhandel. Engstrand, Olle. 1987a. “Durational patterns of Lule Sami phonology”. Phonetica 44.117–128. . 1987b. “Preaspiration and the voicing contrast in Lule Sami”. Phonetica 44.103–116. . 1987c. “Salient features of Lule Sami pronunciation”. Department of Phonetics, Umeå University.

268


Ernic, Enn. 1977. “Predvaritel’nye dannye o sv’az’ax meždu saamskim i tunguzo-mandžurskimi jazykami”. Soviet Finno-Ugric Studies 13.20–24. Fernandez, M. M. Jocelyne. 1983. “L’unification de la langue lapone (‘same’): Le crépuscule des lapologues, ou une aube nouvelle pour la lapophonie?” Language Reform: History and future ed. by Istvan Fodor & Claude Hagege, 1– 25. Hamburg: Helmut Buske. . 1986. “Word order and enunciative particles in the Sami (‘Lappish’) dialogue: A contrastive and panchronic approach”. Festschrift für Istvan Futaky Zum 60. Geburtstag am 12. Mai 1986 ed. by Wolfgang Veenker, 113–125. Hamburg: Helmut Buske. . 1987. Le Discours des Sames: Oralité contrastes énonciation. Paris: Didier Érudition. Fernandez-Vest, M. M. Jocelyne. 1992. “Particles as fundaments of discourse structuring”. Nordic Research on Text and Discourse ed. by Ann Charlotte Lindeberg, Nils Erik Enkvist & Kay Wikberg, 45–59. Turku: Åbo Akademi. . 1996. “Du Mediatif finno-ougrien: Mode oblique en estonien, particules en finnois et en same”. L’Énonciation Médiatisée ed. by Zlatka Guentcheva, 169–182. Louvain: Peeters. . 1997. Parlons Lapon. Les Sames, langue et culture. Paris: L’Harmattan. . 2002. “Translating orality: The case of Northerrn Sami, from oral tradition to written and literary language”. Proceedings of the XIIIth Conference of the Finno-Ugric Studies Association of Canada ed. by Zita McRobbie-Utasi & Connie K. So, 31–41. Vancouver: Simon Fraser University Publications. Frette, Thor. 1975a. Norsk-samisk ordbok / Dárugie-sámigiel sádnigir’ji. Oslo: Universitetsforlaget. . 1975b. “Samiske stedsnavn”. Norske stedsnavn/stadnamn ed. by Botolv Helleland, 105–118. Oslo: Gröndahl. Friis, Jens Andreas. 1856a. Lappisk grammatik: udarbeidet efter den finmarkiske hoveddialekt eller sproget, saaledes som det almindeligst tales i norsk Finmarken. Christiania [Oslo]: Cappelen. . 1856b. Lappiske sprogprøver: en samling af lappiske eventyr, ordsprog og gaader: med ordbog. Christiania [Oslo]: Cappelen. . 1887. Ordbog over det lappiske sprog: med latinsk og norsk forklaring: samt en oversigt over sprogets grammatik. Christiania [Oslo]: Jacob Dybwad. Fromm, Hans. 1971. “Zur Klassenbildung der frühen nordischen Lehnwörter im Lappischen”. Zeitschrift für Deutsches Altertum und Deutsche Literatur 100.243–49.


269

. 1972. “Zur Klassenbildung der frühen nordischen Lehnwörter im Lappischen”. Mélanges offerts à Aurelien Sauvageot pour son soixante-quinzième anniversaire ed. by Jean Gergely, L. Moreau, Jean Perrot & Jozsef Erdodi, 81–87. Budapest: Akadémiai Kiadó. Galkin, I. S. 1987. “Neskol’ko slov o mariisko-saamskix leksicheskix parallel’ax”. Soviet Finno-Ugric Studies 23:4.252–255. Ganander, Henry. 1743. Grammatica lapponica. Stockholm: Holmiae. Genetz, Arvid. 1891. Wörterbuch der Kolalappischen Dialekte nebst Sprachproben. (= Bidrag till kännedomen af Finlands natur och folk 50.) Helsinki. . 1899. “Vieläkin lapin sabme sanan mahdollista suomalaisista vastineista”. Virittäjä 3.107–108. Gordon, Matthew. 1992. “Consonant gradation in Lappish: A study in syllable theory”. Honours Thesis, Harvard University. . 1995. The mechanism of consonant gradation in Saami, Finnish and Estonian. Master’s thesis, UCLA. . 1998. “A fortition-based approach to Balto-Fennic-Sámi consonant gradation”. Folia Linguistica Historica 18:1/2.49–79. Greller, Wolfgang. 1996. Provision and Regulation of the Sámi Languages. Aberystwyth: Centre for Educational Studies. Grundström, Harald. 1928. “Ett lapskt ordstav om pesten”. Festskrift til Rektor J. Qvigstad. (= Tromsø Museums Skrifter 2.), 59–62. Oslo: A.W. Brøggers Boktrykkeri. . 1950. “Lapskt namnskick”. Arv; Tidskrift För Nordisk Folkminnesforskning 6.60–61. Oslo. . 1952–1954. Lulelapsk ordbok. Uppsala: Lundequistiska Bokhandeln. Guttorm, Seija. 1987. Dutkanraporta sámemánáid guovttegielalasvuoda ¯ birra ohcejoga vuolledásis (1-6 L K). Kautokeino, Norway: Sámi instituhtta. Guttorm, Vesa. 1993. Gii, gea- ja mii, maa-relatiivamáddaga válljasupmi “olmmos” korreláhtaid oktavuodas. ¯ Kautokeino, Norway: Sámi instituhtta. . 1998. Giellaˇcállosat 3. Kautokeino, Norway: Sámi instituhtta. Guttorm, Vesa, Nils Helander & Kjell Kemi. 1992a. Giellaˇcállosat. Kautokeino, Norway: Sámi instituhtta. . 1992b. Muhtun giellagažaldagat. Kautokeino, Norway: Sámi oahpahusrád¯di. ¯ . 1994. Giellaˇcállosat 2. Kautokeino, Norway: Sámi instituhtta.

270


Gyarmathi, Sámuel. 1983[1799]. Grammatical Proof of the Affinity of the Hungarian Language with Languages of Fennic Origin. Amsterdam and Philadelphia: John Benjamins. Hætta, Inga Laila & Inga Hætta Skarvik. 1996. Sánit min beaivválas eallimis ja ruovttudoalus: dáru-sámi sátnegirjjáš / Sámi-dáru sátnegirjjáš [= Norsksamisk ordliste / Samisk-norsk ordliste]. Kautakeino, Norway: Sámi oahpahusrád¯di ¯ / Sámi oahpponeavvoguovddáš. Haetta, Lars Jakobsen. 1958. “Mui’talusat”. Mui’talusat, cáliiga Lars Haetta ja Anders Baer ed. by Knut Bergsland, 9–39. Oslo: Universitetsforlaget. Hajdú, Péter. 1975. Finno-Ugrian Languages and Peoples. London: Andre Deutsch Limited. Translated and adapted by G. F. Cushing. Halász, Ignácz. 1885. “Lapponica. 1, Sprachtexte im schwedisch-lappischen Dialecte der Lule- und Pite-lappmark: gesammelt und zum Teil mit ungarischer Übersetzung, sowie einem lappisch-ungarisch und deutschen Glossar”. Nyelvtudományi Közlemények 19. . 1886. Svéd-lapp nyelv 2, Jemtlandi lapp nyelvmutatványok. (= Ugor füzetek 8.) Budapest: Magyar Tudományos Akadémia. . 1887. Svéd-lapp nyelv 3, Ume- és tornio-lappmarki nyelvmutatványok. (= Ugor füzetek 9.) Budapest: Magyar Tudományos Akadémia. . 1890. “A svédországi lapp nyelvjárások”. Nyelvtudományi közlemények 22.161–259. . 1891. Svéd-lapp nyelv 4, Déli-lapp szótár. (= Ugor füzetek 10.) Budapest: Magyar Tudományos Akadémia. . 1893. “A déli lapp nyelv hangjairól”. Nyelvtudományi közlemények 23.206–215. . 1896. Svéd-lapp nyelv 6, Pite-lappmarki szótár és nyelvtan: rövid karesuandói lapp szójegyzékkel. Budapest: Magyar Tudományos Akadémia. Hallan, Nils. 1976a. “Annermannen i norsk och sörsamisk”. Arkiv för Nordisk Filologi 91.192–201. . 1976b. “Stadnamn i grenseland”. Maal og Minne 1/1976.30–45. Hämäläinen, Antti. 1938. Koltta-lappia sanoin ja kuvin. Helsinki: Werner Söderström Oy. Hansegård, Nils Erik. 1963. Sámi-giella, gål’li-giella; lapsk kursbok. Stockholm: Sveriges Radios Förlag. . 1965. “Sea Lappish and Mountain Lappish”. Journal de la Société FinnoOugrienne 66:6.1–91.


271

. 1967. Recent Finnish Loanwords in Jukkasjärvi Lappish. (= Studia Uralica et Altaica Upsaliensia 3.) Uppsala: Uppsala Universitet. . 1968. Giellamet; kursbok i lapska skriftspråket för Nordens samer. Stockholm: Sveriges Radios Förlag. . 1971. Finnish Loanwords in Jukkasjärvi Lappish Classified According to Their Meaning. Supplement to recent Finnish loanwords in Jukkasjärvi Lappish. (= Studia Uralica et Altaica Upsaliensia 6.) Stockholm: Uppsala Universitet and Almqvist & Wiksell. . 1978. “Om samiskans ordskatt”. Kultur på karrig jord. Festskrift til Asbjørn Nesheim, 215–224. Oslo: Norsk folkemuseum. . 1979. “Om samiskans ordskatt”. By og bygd 26.215–223. . 1982. “Some figures concerning the lexicon of Northern Lappish”. FennoUgrica Suecana 5.93–110. Harms, Robert T. 1975. “Problems in Lapp phonology”. Congressus Tertius Internationalis Fenno-Ugristarum pars I 433–440. Hasselbrink, Gustav. 1944. Vilhelmina-Lapskans Ljudlära. Uppsala: Almqvist & Wiksells. . 1965a. Alternative Analyses of the Phonemic System in Central South Lappish. (= Indiana University Uralic and Altaic Series 49.) Bloomington: Indiana University. . 1965b. “Die konsonanten im zentralen Südlappischen unter dem Einfluss von Akzent und Silbenstellung”. Ural-Altaische Jahrbücher 36.72–103. . 1967. “Homonyme Wortformen und Minimalpaare im Südlappischen”. Zeitschrift für Mundartforschung 3.319–331. . 1969–1970. “Südlappische Verbalnomina”. Actes du Xe Congreès International des Linguistes ed. by Alexandru Graur, 583–587. Bucarest: Acad. R.S.R. . 1981. Oårj’elsaamien baaguog’ärjaa – Südlappisches Wörterbuch. Uppsala: AB Lundequistska Bokhandeln. . 1981–1985. Südlappisches Wörterbuch. (= Schriften des Instituts für Dialektforschung und Volkskunde in Uppsala. Lapskt språk och lapsk kultur c.4.) Uppsala: Lundequistiska Bokhandeln. 3 volumes. Haugen, Einar. 1979. “An ecological model for bilingualism”. Notes on Linguistics 12.14–20. Helander, Elina. 1984. Om trespråkighet: en undersökning av språkvalet hos samerna i Övre Soppero. (= Umeå Studies in the Humanities 67.) Stockholm: Universitetet i Umeå and Almqvist & Wiksell.

272


Helander, Nils Øyvind. 1988. “Samisk språk i dag – hvordan er situasjonen?” Norsklæraren 12:4.36–38. Helimski, E. 1996. “Protolappisch und Samojedisch: die vorgeschlagenen Wortgleichungen im Lichte der heutigen etymologischer Forschung - Lapponica et Uralica”. Studia Uralica Upsaliensia 26.51–69. Hertzen, Erik von. 1970. “Ikivanha vesistonnimi. (Ein uralter Gewassername)”. Virittäjä 74.39–44. Hjorth, Poul Lindegård. 1958. Nordsamisk bøjningslære. (= UniversitetsJubilæets danske Samfunds Skriftserie 387.) Copenhagen: Schultz. Holm, Gösta. 1980. “Nordiskt och lapskt i Lapplands ortnamn”. Ortnamn och språkkontakt: Handlingar från NORNA: Sjätte Symposium i Uppsala 5-7 Maj 1978 ed. by Thorsten Andersson, Eva Brylla & Allan Rostvik, 28–62. Uppsala: Nordiska samarbetskommittén för namnforskning. Holmberg-Harva, Uno. 1912. “Lappalaisten saivo”. Virittäjä 16.86–88. . 1928. “Petsamonmaan kolttain pyhät paikat”. Juhlakirja Yrjö Wichmann kuusikymmenvuotispäväksi. (= Mémoires de la Société Finno-Ougrienne 58.), 12–25. Helsinki: Suomalais-Ugrilainen Seura. Houlihan, Kathleen & Charlotte Webb. 1972. “Syllable shaping in Lapp”. From Soundstream to Discourse: Papers from the 1971 Mid-America Linguistics Conference ed. by Daniel G. Hays & Donald M. Lance. 55–61. Columbia: University of Missouri. Hunfalvi, Pál. 1866. “A lapp nyelv”. Nyelvtudományi közlemények 5.286–320. Huss, Leena. 1996a. “Att rädda ett hotat språk”. Invandrare & Minoriteter 5– 6.30–31. . 1996b. “Erste Hilfe für eine bedrohte Sprache: Wiederbelebungsmassnahmen bei den norwegischen Lulesamen”. In Larsson 1996, 71–78. Hyltenstam, Kenneth. 1990. Språkbyte och språkbevarande i ett internationellt perspektiv med särskilt beaktande av situationen för samiskan i Sverige: Underlagsrapport. Stockholm: Allmänna förlaget. Hyltenstam, Kenneth & Christopher Stroud. 1991. Språkbyte och språkbevarande: om samiskan och andra minoritetsspråk. Lund: Studentlitteratur. Hyltenstam, Kenneth, Christopher Stroud & Mikael Svonni. 1999. “Språkbyte, språkbevarande, revitalisering: samiskans ställning i svenska Sapmi”. Sveriges sju inhemska språk ed. by Kenneth Hyltenstam, 41–97. Lund: Studentlitteratur.


273

Ihre, Johannis, Eric Lindahl & Johanne Öhrling. 1780. Lexicon lapponicum: cum interpretatione vocabulorum sveco-latina et indice svecano lapponico. Holmiæ [Stockholm]: Johan Georg Lange. Ijäs, Johanna. 1995. Sirpmá luossagilvvuin goargnun- ja suhkangilvvuide Buolbmágis: goallossániid ja sátnelihtuid partiála ja temporála gaskavuodat. ¯ Master’s thesis, University of Tromsø. Ingold, Tim. 1976. The Skolt Lapps Today. Cambridge: Cambridge University Press. Itkonen, Erkki. 1934a. “Lp. aibmo”. Virittäjä 38.147–151. . 1937. “Pieniä lisiä suomalais-lappalaiseen yhteissanastoon”. Virittäjä 41.299–311. . 1939a. Der ostlappische Vokalismus vom qualitativen Standpunkt aus mit besonderer Berücksichtigung des Inari- und Skoltlappischen. (= Mémoires de la Société Finno-Ougrienne 79.) Helsinki: Suomalais-Ugrilainen Seura. . 1939b. “Eräs lappalainen kuukaudennimi”. Virittäjä 43.52–256. . 1939–1940. “Über den Ursprung einiger lappischen Wörter”. FinnischUgrische Forschungen 26.192–197. . 1941a. “Lappalaisista merkitsevistä nimityksistä ja niiden historiallisesta taustasta”. Virittäjä 45.76–93. . 1941b. “Über den Charakter des ostlappischen Stufenwechselsystems”. Finnisch-Ugrische Forschungen 27.137–167. . 1946a. “Lappalaisten esihistoriaa valaisevia sanoja”. Virittäjä 50.401–421. . 1946b. Struktur und Entwicklung der ostlappischen Quantitätssysteme. (= Mémoires de la Société Finno-Ougrienne 88.) Helsinki: Suomalais-Ugrilainen Seura. . 1947. “Kuolanlapin vokatiivimuodoista”. Virittäjä 51.192–197. . 1950a. “Das Perfekt des Partizips im Lappischen”. Commentationes Fenno-Ugricae in honorem Y. H. Toivonen. (= Mémoires de la Société FinnoOugrienne 98.), 271–304. Helsinki: Suomalais-Ugrilainen Seura. . 1950b. “Varhaiskantasuomen a:n ja ä-n kehitys lapissa”. Virittäjä 54.69– 74. . 1951. “Suomen tunturilapin kirjakielen kehitysvaiheet”. Virittäjä 55.169– 183. . 1955a. “Die Herkunft und Vorgeschichte der Lappen im Lichte der Sprachwissenschaft”. Ural-Altaische Jahrbücher 27.32–44. . 1955b. “Petsamon Suonikylän nimi ja siihen liittyvä tarinas”. Kalevalaseuran vuosikirja 35.166–174.

274


. 1955c. “Über die Betonungsverhältnisse in den finnisch-ugrischen Sprachen”. Acta Linguisticae Academiae Scientiarum Hungaricae 5.21–34. . 1959a. “Kolttien satu revontulien synnystä”. Virittäjä 59.96–109. . 1959b. “Suomen höyhen sanasta ja sen lappalaisesta vastineesta”. Virittäjä 59.306–312. . 1960. Lappische Chrestomathie mit grammatikalischem Abriss und Wörterverzeichnis. (= Suomalais-Ugrilaisen Seuran Apuneuvoja 7.) Helsinki: Suomalaisen Kirjallisuuden Seura. . 1962a. “Eräistä Knud Leemin lapin sanakirjoissa mainitusita sanoista”. Virittäjä 66.261–267. . 1966a. “Lappalaisten esihistoria kielitieteen valossa”. SuomalaisUgrilaisen kielen- ja historiantutkimuksen alalta. (= Tietolipas 20.), 85–131. Helsinki: Suomalaisen Kirjallisuuden Seura. . 1967. “Eräiden latiivin ja prolatiivin päätteiden tarkastelua”. Virittäjä 71.249–255. . 1968. “Zwei Andenken an die Zeit von Karelien und Haalogaland im Lappischen”. Fenno-Ugrica. Juhlakirja Lauri Postin kuusikymmenvuotispäiväksi 17.3.1968. (= Mémoires de la Société Finno-Ougrienne 145.), 42– 64. Helsinki: Suomalais-Ugrilainen Seura. . 1969a. “Über einige Formen der dritten Person in der lappischen Konjugation”. Finnisch-Ugrische Forschungen 37.98–117. . 1969b. “Zur Wertung der finnisch-ugrischen Lautforschung”. UralAltaische Jahrbücher 41.76–111. . 1971a. “Ehdotus inarinlapin fonemaattiseksi transkriptioksi”. In Itkonen et al. 1971, 43–67. . 1971b. “Ehdotus kildininlapin Šongujn murteen fonemaattiseksi transkriptioksi”. In Itkonen et al. 1971, 87–110. . 1973a. “Phonetische und phonologische Betrachtung der ostlappischen Dialekte”. Journal de la Société Finno-Ougrienne 72.129–149. . 1973b. “Zur Geschichte des Partitivs”. Finnisch-Ugrische Forschungen 40.278–339. . 1975. “Etymologisia lisiä”. Virittäjä 79.168–177. . 1977a. “Betrachtungen zum lappischen Stufenwechsel”. Journal de la Société Finno-Ougrienne 75.13–30. . 1977b. “Lapin kieli suomen kielen historian valaisijana”. Academia Scientiarum Fennica Yearbook 1/1977.121–128.


275

. 1980a. “Einige Gesichtspunkte zur Frühgeschichte der Lappen und des lappischen”. Journal de la Société Finno-Ougrienne 76.5–21. . 1980b. “Über einige lappische Verbalsuffixe”. Journal de la Société FinnoOugrienne 76.23–41. . 1982. “Die naso-oralen Vokale im Inarilappischen”. Voces amicormum Sovijärvi. In honorem Antti Sovijärvi septuagesimum annum agentis die XXII mensis aprilis anno MCMLXXXII. (= Mémoires de la Société FinnoOugrienne 181.), 143–150. Helsinki: Suomalais-Ugrilainen Seura. . 1986–1991. Inarilappisches Wörterbuch. (= Lexica Societatis FennoUgricae 20.) Helsinki: Suomalais-Ugrilainen Seura. 4 volumes. . 1992. Inarinsaamelaisia kielennäytteitä — Aanaarkiela cˇ a˙ jttuzeh. (= Mémoires de la Société Finno-Ougrienne 213.) Helsinki: Suomalais-Ugrilainen Seura. Edited by Lea Laitinen. . 1993a. “Allegromuodoista Pohjois-Inarin tunturilapin murteessa”. Festschrift für Raija Bartens zum 25.10.93. (= Mémoires de la Société FinnoOugrienne 215.). Helsinki: Suomalais-Ugrilainen Seura. Itkonen, Erkki, Terho Itkonen, Mikko Korhonen & Pekka Sammallahti, eds. 1971. Lapin murteiden fonologiaa. (= Castrenianumin toimitteita 1.) Helsinki: Suomalais-Ugrilainen Seura. Itkonen, Erkki & Juhani Lehtiranta. 1985. Kildinlappische Sprachproben. Gesammelt von Erkki Itkonen; Herausgegeben von Juhani Lehtiranta. (= Mémoires de la Société Finno-Ougrienne 191.) Helsinki: Suomalais-Ugrilainen Seura. Itkonen, Terho. 1962b. “Lp. biNâl”. Commentationes Fenno-Ugricae in honorem Paavo Ravila. (= Mémoires de la Société Finno-Ougrienne 125.), 139–146. Helsinki: Suomalais-Ugrilainen Seura. . 1965. “Über einige Sandhi-Erscheinungen im Finnmarklappischen, Finnischen und Finlandschwedischen”. Finnisch-Ugrische Forschungen 35.218–263. . 1993b. “Lapin perua Sisä-Suomen sanastossa ja paikannimissä”. Festschrift für Raija Bartens zum 25.10.1993. (= Mémoires de la Société Finno-Ougrienne 215.), 159–168. Helsinki: Suomalais-Ugrilainen Seura. Itkonen, Toivo Immanuel. 1920. “Lappalaisperäisiä paikannimiä suomenkielen aluella”. Virittäjä 124.1–11. . 1928. “Lappalais-suomalaisia sanavertailuja”. Juhlakirja Yrjö Wichmann kuusikymmenvuotispäväksi. (= Mémoires de la Société Finno-Ougrienne 58.), 26–56. Helsinki: Suomalais-Ugrilainen Seura.

276


. 1932. “Lappische Lehnwörter im Russischen”. Annales Academiae Scientiarum Fennicae B 27.47–65. . 1933. “Petsamon kartan nimistö”. Virittäjä 37.487–488. . 1934b. “Uuttu”. Virittäjä 38.216–221. . 1936. “Muutamia suomalais-lappalaisia kalastustermejä”. Virittäjä 40.290–295. . 1939–40. “Über die wurfschlinge der Lappen”. Finnisch-Ugrische Forschungen 26.217–247. . 1942. “Suomen-, kuolan- ja ruijanlappalaisten vanhat henkilönnimet”. Virittäjä 46.60–89. . 1950c. “Ostjakisch-lappische Wortvergleichungen”. Commentationes Fenno-Ugricae in honorem Y. H. Toivonen. (= Mémoires de la Société FinnoOugrienne 98.), 133–136. Helsinki: Suomalais-Ugrilainen Seura. . 1962c. “Kuivi ein heiliger Ort der Lappen”. Commentationes FennoUgricae in honorem Paavo Ravila. (= Mémoires de la Société FinnoOugrienne 125.), 133–136. Helsinki: Suomalais-Ugrilainen Seura. . 1962d. “Lappalaisia henkilönnimiä 1500-ja 1600-luvulta”. Virittäjä 66.368–381. . 1966b. “Lappalaisia paikannimiä 1500-luvulta”. Journal de la Société Finno-Ougrienne 67.3–32. Itkonen, Terho. 1955–1956. “Outakosken lapinmurteen vokaalisto”. Journal de la Société Finno-Ougrienne 58. . 1971. “Itäruijanlapin foneemianalyysin pulmia Outakosken murteen valossa”. In Itkonen et al. 1971, 23–41. . 1973a. “Lisiä erääseen lapin vokaaliston ongelmaan”. Commentationes Fenno-Ugricae in Honorem Erkki Itkonen. (= Mémoires de la Société FinnoOugrienne 150.), 75–124. Helsinki: Suomalaisen Kirjallisuuden Kirjapaino. . 1973b. “Über eine germanische Lehnwortsippe im Finnischen und Lappischen”. Finnisch-Ugrische Forschungen 40.42–59. . 1977. “Zur Vereinfachung der FU-Transkription. Ergebnisse der Tätigkeit des internationalen Transkriptionskomitees”. Journal de la Société FinnoOugrienne 75. . 1983. “Välikatsaus suomen kielen juuriin”. Virittäjä 87.190–229 & 349– 386. Itkonen, Toivo Immanuel. 1916. Venäjänlapin konsonanttien astevaihtelu Koltan, Kildinin ja Turjan murteiden mukaan. (= Mémoires de la Société Finno-Ougrienne 39.) Helsinki: Suomalais-Ugrilainen Seura.


277

. 1918. “Lappalais-suomalaisia sanavertailuja”. Journal de la Société FinnoOugrienne 32:3. . 1921. Lappalaisten Ruokatalous. (= Mémoires de la Société FinnoOugrienne 51.) Helsinki: Suomalais-Ugrilainen Seura. . 1942. Lappalaisten vanhat henkilönnimet. (= Lapin sivistysseuran julkaisuja 10.) Helsinki: Suomalaisen Kirjallisuuden Seura. . 1946. “Über die lappischen Masseinheiten”. Finnisch-Ugrische Forschungen 29.52–68. . 1958a. Koltan- ja kuolanlapin sanakirja [Wörterbuch des Kolta- und Kolalappischen]. (= Lexica Societatis Fenno-Ugricae 15.) Helsinki: SuomalaisUgrilainen Seura. . 1958b. “Über die skandinavischen Lehnwörter im Kolta- und Kolalappischen”. Journal de la Société Finno-Ougrienne 60.3–21. Jahr, Ernst Håkon. 1982. “Language contact in northern Norway: Adstratum and substratum in the Norwegian, Lappish and Finnish of northern Norway”. Die Leistung der Strataforschung und der Kreolistik: Typologische Aspekte der Sprachkontakte ed. by P. Sture Ureland, 307–320. Tübingen: Niemeyer. . 1993. “The Sami language in northern Scandinavia - A language maintenance perspective”. Language: A doorway between human cultures ed. by Øyvind Dahl, 238–254. Oslo: Novus. . 1996. “A note on language preservation – With special reference to Sami in northern Scandinavia”. Suvremena Lingvistika 22:1/2 (41–42).255–264. Jefremoff, Irja. 2001. Vaaristâllâm. Inarisaamelaisten vuotiismuutto. (= Inarin saamelaismuseon julkaisuja 4.) Vammala, Finland: Vammalan Kirjapaino Oy. . 2005. Kolttasaamelaiset nyt. Tutkimus kotoutetunkansan elämäntilanteesta uuden vuosituhannen alussa. Jyväskykä, Finland: Gummerus Kirjapaino Oy. Jernsletten, Nils, ed. 1973. Sámegiela skuv’lasádnelis’to: mas sánit laet jår’galuvvun dárugillii. Samisk-norsk ordliste. Oslo: Grunnskolerådet, Kirkeog Undervisningsdepartementet, Universitetsforlaget. Jernsletten, Nils. 1982. “Språket i samiske samfunn”. Nordnorsk ed. by T. Bull & K. Jetne, 101–117. Oslo: Det Norske Samlaget. . 1997. “Sami traditional terminology: Professional terms concerning salmon, reindeer and snow”. Sami Culture in a New Era: The Norwegian Sami experience ed. by Harald Gaski, 86–108. Karasjok, Norway: Davvi Girji.

278


Jernsletten, Per John. 1988. Merrosat nominála gihpuin: iskkadeapmi makkár merrosat substantiivála gihpuin leat buolbmát sámegielas. Master’s thesis, University of Tromsø. Johnsen, Rita M. 1997. Fornorskingspolitikken av samene sett i et språkteoretisk lys. Master’s thesis, University of Oslo. Julien, Marit. 1996. Syntactic Word Formation in Northern Sámi. Oslo: Novus. . 2003. “On the negated past in Finnic and Saami”. In Nelson & Manninen 2003, 419–446. Kecskeméti, István & Gábor Zaicz. 1977a. “Norvég lapp a tergo toldaléktár”. Nyelvtudományi Közlemények 79.79–103. . 1977b. “Repertoire de terminaisons a tergo du lapon norvégien”. Études finno-ougriennes 14.87–113. Kert, Georgij Martynoviˇc. 1958. “Znaˇcenie saamskogo jazyka dl’a finnougorskogo jazykoznanija”. Pribaltijsko-finskoe jazykoznanie ed. by G. N. Makarov, 104–117. Petrozavodsk: Gosudarstvennoe izdatel’stvo Karel’skoj ASSR. . 1960. “Nekotorye saamskie toponimiˇceskie nazvanija na territorii Karel’skoj ASSR”. Voprosy jazykoznanija 2.86–92. . 1961. Obrazcy saamskoj reˇci: materialy po jazyku i fol’kloru saamov Kol’skogo poluostrova (kil’dinskij i iokan’gskij dialekty). Moscow: Nauka. . 1962. “Imennaja i glagol’naja osnovy v kil’dinskom dialekte saamskogo jazyka”. Voprocy finno-ugorskogo jazykoznanija, 143–153. Moscow and Leningrad. . 1964. “Dolgota soglasnyx i glasnyx zvukov v kil’dinskom dialekte saamskogo jazyka”. Voprosy finno-ugorskogo jazykoznanija. Grammatika i leksikologia, 44–50. Moscow and Leningrad. . 1966. “Saamskij jazyk”. Finno-Ugorskie i samodijskie jazyki ed. by V. I. Lytkin, 155–171. Moscow: Nauka. . 1967a. “Fonetiˇceskie izmenenija i fonologiˇceskie cˇ eredovanija (na materiale kil’dinskogo dialekta saamskogo jazyka)”. In Mullonen & Ollykajnen 1967, 20–26. . 1967b. “Saamskaja pis’mennost”’. In Mullonen & Ollykajnen 1967, 110– 115. . 1971. Saamskij jazyk (kil’dinskij dialekt): fonetika, morfologija, sintaksis. Leningrad: Nauka. . 1975a. “Nekotorye osobennosti leksiki saamskix dialektov Kol’skogo poluostrova”. Fenno-Ugristica 1.159–166.


279

. 1975b. “Saamskij jazyk”. Osnovy finno-ugorskogo jazykoznanija: pribaltijsko-finskie, saamskie i mordovskie jazyki ed. by V. I. Lytkin, K. E. Majtinskaja & K. Rédei, 203–247. Moscow: Nauka. . 1978–1979. “Sravnitel’noe issledovanie leksiki karel’skogo, vepskogo i saamskogo jazykov s pomošˇc’u perfokart”. Études finno-ougriennes 15.211– 215. . 1986. Slovar’ saamsko-russkij i russko-saamskij. Leningrad: Prosvešˇcenie. . 1987. “O glagol’nyx slovoobrazovatel’nyx suffiksax v kil’dinskom dialekte saamskogo jazyka”. Soviet Finno-Ugric Studies 23:2.92–100. . 1994. “Saamsko-russkie jazykovye kontakty”. Pribaltijsko-finskoe jazykoznanie ed. by Petr Mefodieviˇc Zajkov, 99–116. Petrozavodsk: Karel’skij nauˇcnyj centr RAN. . 2002a. Primenenie kompjuternyx texnologij v issledonvanii toponimii (pribaltijsko-finskaja, russkaja). Petrozavodsk: Karel’skij nauˇcnyj centr RAN. . 2002b. “Saami toponymy of the Kola Peninsula as a research object”. Onomastica Uralica 2.121–132. . 2005. “Saami village names in the Kola Peninsula”. Onomastica Uralica 3.35–44. Keskitalo, Alf Isak. 1978. “Kritiske faktorer i samisk språkutvikling”. Språkvård 4.15–20. . 1981. “The status of the Sami Language”. Minority Languages Today ed. by Einar Haugen, J. Derrick McClure & Derick Thomson, 152–162. Edinburgh: Edinburgh University Press. Kintel, Anders. 1991. Syntaks og ordavledninger i lulesamisk. Kautokeino, Norway: Samisk utdanningsråd. Koivulehto, Jorma. 1988. “Lapin ja itämerensuomen suhteesta: Ieur. -tr- yhtyman korvautuminen lainoissa”. Virittäjä 92:1.26–51. . 1992. “Germanisch-lappische Lehnbeziehungen”. Finnisch-Ugrische Sprachen zwischen dem germanischen und dem slavischen Sprachraum ed. by László Honti, Sirkka-Liisa Hahmo, Tette Hofstra, Jolanta Jastrzebska & Osmo Nikkilä, 85–95. Amsterdam: Rodopi. Koizumi, Tamotsu. 1993. Rappugo nyumon [= A Lappish (Sami) grammar]. Tokyo: Daigaku Shorin. Koponen, Eino. 1988. “Ovatko suomen longota ja lonsota lappalaisia lainoja?” Virittäjä 92:1.92–99.

280


. 1996. “Lappische Lehnwörter im Finnischen und Karelischen”. In Larsson 1996. 83–98. Korhonen, Mikko. 1962. “Zum syntaktischen Gebrauch des lappischen Verbs gâl’gât "sollen, müssen"”. Commentationes Fenno-Ugricae in honorem Paavo Ravila. (= Mémoires de la Société Finno-Ougrienne 125.), 233–251. Helsinki: Suomalais-Ugrilainen Seura. . 1964. “Lapin murteiden keskinäisistä suhteista”. Lapin tutkimusseura vuosikirja 5.49–64. . 1967a. Die Konjugation im Lappischen. Morphologisch-historische Untersuchung. I. Die finiten Formkategorien. (= Mémoires de la Société FinnoOugrienne 143.) Helsinki: Suomalais-Ugrilainen Seura. . 1967b. “Lapin kielen vartalonsisäistä äännevaihtelujen morfologisesta funktiosta”. Virittäjä 71.353–358. . 1969a. “Die Entwicklung der morphologischen Methode im Lappischen”. Finnisch-Ugrische Forschungen 37.203–362. . 1969b. “Lapin metafoniasta”. Juhlakirja Paavo Siron täyttäessä 60 vuotta 2.8. 1969. (= Acta Universitatis Tamperensis A 26.), 79–89. Tampere: University of Tampere. . 1970. “Der Ausdruck des indefiniten Subjekts im aktivischen Satz des Lappischen”. Symposion über Syntax der uralischen Sprachen, 15.-18. Juli 1969 in Reinhausen bei Göttingen, 142–153. Göttingen: Vandenhoeck & Ruprecht. . 1971. “Ehdotus koltanlapin Suonikylän (nyk. Sevettijärven) murteen fonemaattiseksi transkriptioksi”. In Itkonen et al. 1971, 69–86. . 1973a. “Ikivanhan opposition heijastusta vai myöhaista palautusta?” Journal de la Société Finno-Ougrienne 72.175–183. . 1973b. “Zur Geschichte des negativen Präteritums und Perfekts im Ostseefinnischen und Lappischen”. Commentationes Fenno-Ugricae in Honorem Erkki Itkonen. (= Mémoires de la Société Finno-Ougrienne 150.), 174–184. Helsinki: Suomalaisen Kirjallisuuden Seura. . 1974. Die Konjugation im Lappischen. Morphologisch-historische Untersuchung. II. Die nominalen Formkategorien. (= Mémoires de la Société FinnoOugrienne 155.) Helsinki: Suomalais-Ugrilainen Seura. . 1975. “Zur Phonologie des Skoltlappischen”. Symposion Phonologische Analyse der uralischen Sprachen, Berlin, 1974. (= Linguistische Studien. Reihe A. Arbeitsberichte 22.), 11–34. Berlin: Akademie der Wissenschaften der DDR.


281

. 1977. “Lapin kielen varhaisvaiheista”. Academia Scientiarum Fennica Yearbook 129–141. . 1979a. “Neues zur Phonologie des Lappischen”. Finnisch-Ugrische Forschungen 43.236–247. . 1979b. “Über die vorgeschichtlichen Beziehungen der Lappen und Ostseefinnen”. Fenno-Ugrica Suecana 2.59–73. . 1981. Johdatus lapin kielen historiaan. (= Suomalaisen Kirjallisuuden Seuran Toimituksia 370.) Helsinki: Suomalaisen Kirjallisuuden Seura. . 1983. “Kolttalaulujen fonologiaa”. Journal de la Société Finno-Ougrienne 78. . 1988b. “The history of the Lapp language”. In Sinor 1988, 264–287. Korhonen, Mikko, Jouni Mosnikoff & Pekka Sammallahti. 1973. Koltansaamen opas. (= Castrenianumin toimitteita 4.) Helsinki: Suomalais-Ugrilainen Seura. Korhonen, Olavi. 1973. Sádnigir’ji: sámigielas ruotagilli, ruotagielas sámigilli [= samisk-svensk, svensk-samisk ordbok]. Jokkmokk, Sweden: Sámiid Ál’bmutallaskuv’la. . 1974. “Den samiska språkvårdens problem”. Språkvård 2.11–18. . 1976. “Linguistic and cultural diversity among the Saamis and the development of Standard Saamish”. Linguistics 183.51–66. . 1979a. Bákkogir’je: julevusámes dárrui, dáros julevusámái [= Lulesamisk svensk, svensk lulesamisk ordbok]. Uppsala / Länsskolnämnden and Luleå: Almqvist & Wiksell. . 1979b. “Lappische Lehnwörter im ältesten Einödgebiet Finnlands”. Finnisch-Ugrische Forschungen 43.175–207. . 1984. “Language contact as reflected in the migration of the Finns, the Saamis and the Vikings”. Scandinavian Language Contacts ed. by P. Sture Ureland & Iain Clarkson, 67–96. Cambridge: Cambridge University Press. . 1994. “Några termer i den samiska renmjölkningen”. Festskrift til Ørnulv Vorren. (= Tromsø Museums Skrifter 15.), 254–268. Tromsø: Tromsø museum. . 1997. “Hur samiskan blev samiska”. Mer än ett språk ed. by Eva Westergren & Hans Åhl, 79–115. Stockholm: Norstedt. Koski, Mauno, ed. 1992. Kontrastiivista kielentutkimusta 2. (= Fennistica 10.) Turku: Åbo Akademi, Finska institutionen. Koskimies, A.V. & Toivo Immanuel Itkonen. 1917. Inarinlappalaisista kansantietoutta. (= Mémoires de la Société Finno-Ougrienne 40.) Helsinki: SuomalaisUgrilainen Seura.

282


Koskinen, Arja. 1994. “Deonttinen modaalisuus saamessa ja suomessa: teoriakehys ja alustavia tuloksia”. Nordlyd: Tromsø University Working Papers on Language and Linguistics 20.182–192. . 1998. Toiminnan välttämättömyys ja mahdollisuus. Pohjoissaamen modaalisten ilmausten semantiikkaa ja syntaksia. (= Mémoires de la Société Finno-Ougrienne 231.) Helsinki: Suomalais-Ugrilainen Seura. Krohn, Kaarle. 1906. “Lappische beiträge zur germanischen mythologie”. Finnisch-Ugrische Forschungen 6.155–180. Kulonen, Juha Pentikäinen, Ulla-Maija & Irja Seurujärvi-Kari, eds. 1994. Johdatus saamentutkimukseen. (= Tietolipas 131.) Helsinki: Suomalaisen Kirjallisuuden Seura. Kuoljok-Angéus, Susanna. 1997. Nominalavledningar på ahka i lulesamiskan [Nominal Derivatives in ahka in Lule Sami]. Ph.D. thesis, Uppsala University. Kuruˇc, Rimma. 1981. “O kategorii cˇ isla sušˇcestvitel’nyx v kil’dinskom dialekte saamskogo jazyka”. Soviet Finno-Ugric Studies 17.266–272. Kuruˇc, Rimma, Nina Afanas’jeva & Jekatarina Meˇckina. 1985. Saamsko-russkij slovar’ [= S¯am’-r¯ušš soagknehk’]. Moscow: Russkij jazyk. Kuruˇc, Rimma, Nina Afanas’jeva & Iraida Vinogradova. 1995. Pravila orfografii i punktuacii saamskogo jazyka. Murmansk and Moscow. Kusmenko, Jurij. 2004a. “Skandinavisch-samischer Sprachkontakt. Gegenseitiger Einfluss”. In Kusmenko 2004b, 215–235. Kusmenko, Jurij, ed. 2004b. The Sami and the Scandinavians. Aspects of 2000 years of contact. (= Schriften zur Kulturwissenschaft 55.) Hamburg: Kovaˇc. Kusmenko, Jurij. 2005. “Scandinavian umlaut, Sami metaphony and SwedishNorwegian levelling (jämning): A floating Scandinavian-Sami isogloss”. Mot rikare mål å trå. Festskrift til Tove Bull ed. by G. Alhaug, E. Mørck & A.-K. Peders, 118–126. Oslo: Novus. Kylstra, Andries D. 1972. “Die Präaspiration im Westskandinavischen und im Lappischen”. Orbis 21.367–382. . 1983. “Skandinavisch-lappische Paralellen”. Symposium saeculare Societatis Fenno-Ugricae. (= Mémoires de la Société Finno-Ougrienne 185.), 159– 177. Helsinki: Suomalais-Ugrilainen Seura. Lagercrantz, Eliel. 1920. “Kalastajalappalaisten elämästä”. Virittäjä 24.119– 121. . 1923. Sprachlehre des Südlappischen nach der Mundart von Wefsen. (= Bulletin 1.) Kristiania [Oslo]: Kristiania Etnografiske Museum.


283

. 1926a. Sprachlehre des Westlappischen nach der Mundart von Arjeplog. (= Mémoires de la Société Finno-Ougrienne 55.) Helsinki: SuomalaisUgrilainen Seura. . 1926b. Wörterbuch des Südlappischen nach der Mundart von Wefsen. (= Serie B 4.) Oslo: Instituttet for Sammenlignende Kulturforskning. . 1927. Strukturtypen und Gestaltwechsel im Lappischen. (= Mémoires de la Société Finno-Ougrienne 57.) Helsinki: Suomalais-Ugrilainen Seura. . 1928. “Die Geheimsprachen der Lappen”. Journal de la Société FinnoOugrienne 42.1–13. . 1929. Sprachlehre des Nordlappischen nach den seelappischen Mundarten. (= Bulletin 3.) Oslo: Oslo Etnografiske Museum. . 1939. Lappischer Wortschatz I – II. (= Lexica Societatis Fenno-Ugricae 6.) Helsinki: Suomalais-Ugrilainen Seura. . 1941. “Synopsis des lappischen”. Oslo Etnografiske Museums Skrifter 2:4.227–350. . 1947. Lapin kielen opas. Helsinki: Suomalaisen Kirjallisuuden Seura. . 1957. Lappische Volksdichtung. (= Mémoires de la Société FinnoOugrienne 112.) Helsinki: Suomalais-Ugrilainen Seura. . 1958. Lappische Volksdichtung. (= Mémoires de la Société FinnoOugrienne 115.) Helsinki: Suomalais-Ugrilainen Seura. . 1959. Lappische Volksdichtung. (= Mémoires de la Société FinnoOugrienne 117.) Helsinki: Suomalais-Ugrilainen Seura. . 1960. Lappische Volksdichtung. (= Mémoires de la Société FinnoOugrienne 120.) Helsinki: Suomalais-Ugrilainen Seura. . 1961. Lappische Volksdichtung V. See- und Skolte-lappische Texte des südlichen Varangergebiets. (= Mémoires de la Société Finno-Ougrienne 124.) Helsinki: Suomalais-Ugrilainen Seura. Lakó, György. 1978. “Zu den Beziehungen der Lappen mit den Samojeden”. Ural-Altaische Jahrbücher 50.91–97. . 1986. Chrestomathia Lapponica. Budapest: Tankönyvkiadó. Larsson, Lars-Gunnar. 1985. “Kriterien zur Klassifizierung der lappischen Dialekte in Schweden”. Dialectologia Uralica ed. by Wolfgang Veenker, 159– 171. Wiesbaden: Harrassowitz. Larsson, Lars-Gunnar, ed. 1996. Lapponica et Uralica: 100 Jahre finnischugrischer Unterricht an der Universität Uppsala: Vorträge am Jubiläumssymposium 20.-23. April 1994. (= Studia Uralica Upsaliensia 26.) Uppsala: Acta Universitatis Upsaliensia and Almqvist & Wiksell.

284


Larsson, Lars-Gunnar. 1998. “Suomen -hän ja saamen -son I: Oekeeta asijoo”. Commentationes Fenno-Ugricae in honorem Seppo Suhonen sexagenarii 16. V. 1998. (= Mémoires de la Société Finno-Ougrienne 228.), 276–279. Helsinki: Suomalais-Ugrilainen Seura. . 2001. “Lappische Sprachforschung an der Schwelle zum 21. Jahrhundert”. Finnisch-Ugrische Forschungen 56.79–94. Leem, Knud. 1748. En lappisk Grammatica efter den Dialect, som bruges af Field-Lapperne udi Porsanger-Fiorden: samt et Register over de udi samme Grammatica anførte Observationers Indhold. Copenhagen: Gottman Friderich Kisel. . 1756. En lappesk nomenclator efter den dialect, som bruges af fjeldlapperne i Porsanger-fjorden. Trondheim: Jens Christensen Winding. Lehtiranta, Juhani. 1982. “Eine Beobachtung über die Gründe der raschen Veränderung des Grundwortschatzes im Lappischen”. Finnisch-Ugrische Forschungen 44:1–3.114–118. . 1986a. “K voprosu o sozdanii pis’mennosti dl’a kol’skix saamov”. Soviet Finno-Ugric Studies 22:1.25–29. . 1986b. “Zur Verwertung ‘protolappischer’ Elemente”. Soviet Finno-Ugric Studies 22:4.249–252. . 1989. Yhteissaamelainen sanasto. (= Mémoires de la Société FinnoOugrienne 200.) Helsinki: Suomalais-Ugrilainen Seura. . 1992. Arjeploginsaamen äänne- ja taivutusopin pääpiirteet. (= Mémoires de la Société Finno-Ougrienne 212.) Helsinki: Suomalais-Ugrilainen Seura. . 1993. The Fundamentals of Arjeplog Saami (Lappish) Phonology and Inflection. Helsinki: Teknillinen Korkeakoulu. Leino, Pentti. 1971. “Norjanlappalaisen sananparsiston alkusointuisuus”. Kalevalaseuran Vuosikirja 51.171–188. Leskinen, Heikki. 1966. “Die Konstruktionstypus anna∼laske + finite Verbform in den osfi. Sprachen und im Lappischen”. Journal de la Société FinnoOugrienne 67.1–160. Leskinen, Juha. 1991. “Historiallisen lauseopin ongelmia instruktiiviaineiston valossa”. Virittäjä 95:1.25–32. Leskinen, V. 1967. “O nekotoryx saamskix gidronimax Karelii”. In Mullonen & Ollykajnen 1967, 79–88. Liberman, A.S. 1971. “The voiceless vowel in Lapp again”. Soviet Finno-Ugric Studies 7.269–271.


285

Lindahl, Erik & Johan Öhrling. 1780. Lexicon lapponicum, cum interpretatione vocabulorum Sveco-Latina, et indice Svecano Lapponico. Holmiæ [Stockholm]: Johan Georg Lange. Lindberg, Carl. 1941. Terrängordet köl (käl, kielas): en betydelsehistorisk och -geografisk undersökning. Uppsala: Almqvist & Wiksell. Lindgren, Anna-Riitta. 1994. Vähemmistökieli muuttuvassa maailmassa. Pohjois-Suomen tutkimuslaitos, Oulu: University of Oulu. Linkola, Anni. 1996. Koltansaamen nykytilanne vähemmistökielenä Suomessa. Master’s thesis, Helsinki University. Linkola, Anni & Martti Linkola. 2000. Kolttasaamelaiset [The Skolt Sámi]. (= Inarin saamelaismuseon julkaisuja 2.) Inari, Finland: Siida. Lukkari, Pekka. 1974. Sami-suoma satnegirji. (= Lapin Sivistysseura 34.) Helsinki: Sami Cuvgetusseärvvi. . 1977. Suoma-sami sadnekirji. Suomalais-saamelainen sanakirja. (= Lapin Sivistysseura 38.) Helsinki: Sami Cuvgetusseärvi. . 1990. Eallid ja šattuid namahusaid sámi-suoma sátnegirji. Suomi-saame sanakirja eläinten ja kasvien nimistä. Inari, Finland: Sami Girji. Määttä, Taisto. 1981. Topics of Interest for Experimental Phonetic Studies in Lappish. (= The Sounds of Lappish 1.) Umeå: Department of Phonetics, Umeå universitet. Magga, Ole Henrik. 1974. “ABC 1619 - dem første bok på samisk”. Ottar 81.32– 35. Oslo: Universitetsforlaget for Tromsø Museum. . 1976. Samisk fonetikk: dialektlære. Alta, Norway: Lærerutdanningen. . 1978. Lokative læt-sætninger i samisk. Kautakeino, Norway: Sámi Instituhtta. . 1980. Giellaoahppa: jietna-, hápme- ja cealkkaoahppa oanehˇccˇ at cˇ ilgejuvvon. Kautakeino, Norway: Sámi instituhtta. . 1982. Modalverb og infinitiv innen verbalet: prosjektrapport. Kautakeino, Norway: Sámi instituhtta. . 1985. “Språkforskning og språkplanlegging i det samiske samfunnet (Linguistics and language planning in the Sami community)”. Acta Borealia 2.71– 79. . 1986. Studier i samisk infinitivsyntaks. Kautakeino, Norway: Sámi instituhtta. Magga, Tuomas. 1984. Duration in the Quantity of Bisyllabics in the Guovdageaidnu Dialect of North Lappish. (= Acta Universitatis Ouluensis. Series B Humaniora 11.) Oulu: University of Oulu.

286


. 1999. “Orthographic overloading and morphological cues in Saami”. Diedut ¯ 3/99.13–25. Magiste, Julius. 1975. “Zu einigen estnisch-livisch-lappischen Etymologien”. Ural-Altaische Jahrbücher 47.124–127. Martinussen, Bente. 1989. “Språkskiftet på 1800-tallet”. Nordlyd: Tromsø University Working Papers on Language and Linguistics 15.123–131. . 1992. “Forholdet mellom samisk og norsk”. The Nordic Languages and Modern Linguistics ed. by J. Louis-Jensen & J.H. Poulsen. 7/2, 401–407. Tórshavn, Faroe Islands: Føroya Fródskaparfelag. ¯ . 1997. “Norske tallord i samisk”. Nordiska dialektstudier ed. by Maj Reinhammar, 195–201. Stockholm: Språk- och Folkminnesinstitutet. Mattisson, Ann-Christin. 1996. “K. B. Wiklund als Ortsnamenforscher”. In Larsson 1996, 367–372. Matveev, A. K. 1974. “Toponimiˇceskie etimologii, VI: Dve saamskie osnovy v substratnoj toponimike russkogo Severa”. Soviet Finno-Ugric Studies 10.239– 240. . 1976. “Toponimiceskie etimologii, VIII”. Soviet Finno-Ugric Studies 12.182–186. . 1977. “Ob odnoi fonetiˇceskoi osobennosti drevnix saamskix dialektov”. Études finno-ougriennes 14.41–43. McRobbie, Zita. 1987. “The Foot-Formation Rule in relation to stress placement in Lappish”. Studien zur Phonologie und Morphonologie der uralischen Sprachen ed. by Károly Rédei, 320–329. Vienna: Verband der Wiss. Gesell. Österreichs. . 1989. “Durational measurements of disyllabics in the Suonikylä dialect of Skolt Sami; With special session on folk music of Finno-Ugric Peoples”. FUSAC ’88 ACEFO: Proceedings of the Sixth Annual Meeting of the FinnoUgric Studies Association of Canada / Actes du Sixième Colloque Annuel de l’Association Canadienne des Études Finno-Ougriennes ed. by Joel Ashmore Nevis. 87–94. Lanham, Maryland: UP of America. McRobbie-Utasi, Zita. 1990. An Acoustic Analysis of Duration in Skolt Sami Disyllabics. Ph.D. thesis, University of Manitoba. . 1991. “Preaspiration in Skolt Sámi”. Simon Fraser University Working Papers in Linguistics 1.77–87. . 1995. “Language planning, literacy and cultural identity: The Skolt Sámi case”. Zur Frage der uralischen Schriftsprachen, 31–41. Budapest: A Magyar Tudományos Akadémia Nyelvtudományi Intézete.


287

. 1999. Quantity in the Skolt (Lappish) Saami Language: An acoustic analysis. (= Uralic and Altaic Series 165.) Bloomington: Indiana University. . 2000. “Acoustic characteristics of word-final vowel reduction in Skolt Saami”. Proceedings of the Ninth International Finno-Ugric Congress. 366– 371. Tartu. . 2001. “Vowel reduction in Skolt Saami in connected speech”. Folia Uralica Debreceniensia 8.397–403. . 2003. “Normative preaspiration in Skolt Sami in relation to the distribution of duration in the disyllabic stress-group”. Ünnepi kötet Honti László tiszteletére, 291–300. Budapest: A Magyar Tudományos Akadémia Nyelvtudományi Intézete. Mikalsen, Anne Dagmar Biti. 1993. Reksuvdnavearbbat. Master’s thesis, University of Tromsø. Mikkola, J. J. 1941. Kolttakylän arkisto. (= Lapin Sivistysseuran julkaisuja 8.) Porvoo: Werner Söderstrom Oy. Mikola, Tibor. 1996. “Zur Geschichte der auslautenden Tenues im Lappischen”. In Larsson 1996, 99–102. Moreau, Jean-Luc. 1977. “Réflexions sur quelques désinences casuelles du finnois et du lapon”. Études finno-ougriennes 14.19–32. Mosnikoff, Jouni & Pekka Sammallahti. 1988. U’cc sääm-lää’dd sää’nnkeârjaz [= Pieni koltansaame-suomi sanakirja]. Utsjoki, Finland: Jorgaleaddji. Mullonen, M. & V. Ollykajnen, eds. 1967. Pribaltijsko-finskoe jazykoznanie: voprosy fonetiki, grammatiki, i leksikologii. Leningrad: Nauka. Närhi, E. M. & R. Korhonen. 1992-1993. “Onomastic Bibliography 1990-1991. Lappish”. Onoma 31.332. Nelson, Diane & Satu Manninen, eds. 2003. Generative Approaches to Finnic and Saami Linguistics. Stanford, Calif.: CSLI Publications. Nelson, Diane & Ida Toivonen. 2003. “Counting and the grammar: Numerals in Inari Saami”. In Nelson & Manninen 2003, 321–341. Nesheim, Asbjørn. 1942. Der lappische Dualism it Berücksichtigung finnischugrischer und indo-europäischer Verhältnisse. (= Hist.-Filos. Klasse 5.) Oslo: Norske Videnskaps Akademi II. . 1947a. “Lappisk fiske og fisketerminologi”. Lapponia. (= Studia Septentrionalia 3.), 69–210. Oslo: Aschehoug. Summary in English in Nordnorske tidsskrift for Sprogvidenskap 14: 352. Oslo.

288


. 1947b. “Russiske og russisk-finsk lånord i Finnmark-lappisk”. Festskrift för Professor Olaf Broch ed. by C. Strang, E. Krag & A. Gallis, 155–167. Oslo: Jacob Dybwad. . 1949. “Finnish hiisi and Lappish sii’dâ”. Finnisch-Ugrische Forschungen 30.292–302. . 1962. “The Lappish dialect of Ullsfjord and its relations to other Lappish dialects”. Commentationes Fenno-Ugricae in honorem Paavo Ravila. (= Mémoires de la Société Finno-Ougrienne 125.), 333–360. Helsinki: SuomalaisUgrilainen Seura. . 1964. “The Lapp fur and skin terminology and its historical background”. Lapponica; Ethnographica Upsaliensia 21.199–218. Uppsala. . 1967. “Duol’lje - coar’ve k- muorra: samisk sløydterminologi og dens opprinnelse”. Norrbottensläns Hembygdsförenings Årsbok, 23–30. Luleå: Norrbottensläns Hembygdsförening. . 1973. “Bynames and names of domestic animals in Ullsfjord, NorthTroms”. Commentationes Fenno-Ugricae in Honorem Erkki Itkonen. (= Mémoires de la Société Finno-Ougrienne 150.), 222–237. Helsinki: SuomalaisUgrilainen Seura. Nevis, Joel Ashmore. 1986. “Decliticization and deaffixation in Saame: Abessive taga”. Ohio State University Working Papers in Linguistics 34.1–9. . 1988. “A morphotactic paradox in Northern Saame: Comitative –guim”. Ural-Altaische Jahrbücher, Neue Folge 8.38–50. Nickel, Klaus Peter. 1990. Samisk grammatikk. Karasjok, Norway: Davvi Girji. Nickul, Karl. 1970. Saamelaiset kansana ja kansalaisina. (= Suomalaisen Kirjallisuuden Seuran Toimituksia 297.) Helsinki: Suomalaisen Kirjallisuuden Seura. . 1973. “Suenjelin kolttayhdyskunnan elämästä”. Commentationes FennoUgricae in honorem Erkki Itkonen. (= Mémoires de la Société FinnoOugrienne 150.), 238–244. Helsinki: Suomalais-Ugrilainen Seura. Nielsen, Konrad. 1902a. Die Quantitätsverhältnisse im Polmaklappischen. (= Mémoires de la Société Finno-Ougrienne 20.) Helsinki: Suomalais-Ugrilainen Seura. . 1902b. “Zur Aussprache des Norwegisch-Lappischen”. Journal de la Société Finno-Ougrienne 20.1–72. . 1903. “Lappalaisia murteita tutkimassa: matkakertomus Norjan ja Suomen Lapista”. Journal de la Société Finno-Ougrienne 22:3.


289

. 1905. Die Quantitätsverhältnisse im Polmaklappischen. II. Nachtrag und Register. (= Mémoires de la Société Finno-Ougrienne 24.) Helsinki: Suomalais-Ugrilainen Seura. . 1912. Lappisk som gjenstand for videnskabelig forskning: tiltrædelsesforelæsning holdt ved Kristiania universitet 25. mars 1912. Kristiania [Oslo]: Aschehoug. . 1913a. “Die 3.p.sg. praes.–form im Lappischen”. Journal de la Société Finno-Ougrienne 30–28. . 1913b. “Die wissenschaftliche bedeutung des lappischen”. FinnischUgrische Forschungen 13.186–206. . 1913c. Lappisches Wörterbuch nach den Dialekten von Polmak, Karasjok und Kautokeino. (= Lexica Societatis Fenno-Ugricae 1.) Helsinki: SuomalaisUgrilainen Seura. . 1920. Til undersøkelsen av lappiske stedsnavn. Videnskapsselskapets skrifter, 2. Hist.-filos. klasse, Kristiania [Oslo]: Jacob Dybwad. . 1922. Palatogrammes de deux patois lapons. (= Oslo Etnografiske Museums Skrifter 2/1.) Oslo: A. W. Brøggers Boktrykkeri. . 1926. Lærebok i Lappisk. Oslo: A. W. Brøggers Boktrykkeri. . 1928. “Et tredelt bidrag: 1. “Fjelldialekten” i Varanger: 2. Partitiv i finnmark-lappisk: 3. Litt om nordiske lånord i den lappiske ren- og rendriftsterminologi”. Festskrift til rektor J. Qvigstad. (= Tromsø Museums skrifter 2.), 154–183. Tromsø: Tromsø Museum. . 1931. “Lappische lexicographie”. Suomalais-Ugrilaisen Seuran Aikakauskirja 46.39–51. . 1932–1962[1979]. Lappisk ordbok: Grunnet på dialektene i Polmak, Karasjok og Kautokeino. (= Skrifter / Instituttet for sammenlignende kulturforskning, serie B 17.) Oslo: Aschehoug. . 1933. “A note on the origin of attributive forms in Lapp”. Liber semisaecularis Societatis Fenno-Ugricae. (= Mémoires de la Société FinnoOugrienne 67.), 296–307. Helsinki: Suomalais-Ugrilainen Seura. . 1945. Lappiske studier. (= Studia Septentrionalia 1.) Oslo: A. W. Brøggers Boktrykkeri. . 1946. “The use of pronominal adverbs of place with the termination –ggo in Finnmark Lapp”. Finnisch-Ugrische Forschungen 29.32–36. . 1950. “A curious word in earlier Lapp dictionaries”. Commentationes Fenno-Ugricae in honorem Y. H. Toivonen. (= Mémoires de la Société FinnoOugrienne 98.), 245–254. Helsinki: Suomalais-Ugrilainen Seura.

290


Nielsen, Konrad & Asbjørn Nesheim, eds. 1962. Lappisk ordbok, grunnet på dialektene i Polmak, Karasjok og Kautokeino (Lapp dictionary, based on the dialects of Polmak, Karasjok and Kautokeino). (= Insituttet for Sammenlignende Kulturforskning, Skrifter B. 17.) Oslo: Universitetsforlaget. 5 volumes. Nilsen, Steinar. 1998. Porsánggu suopmana guolástanterminologiija: semánttalas, fonologalas, morfologalas ja historjjálas suokkardallan. Master’s thesis, University of Tromsø. Nilsson-Mankok, Erik. 1970. “Hur sydsamiskan försvinner”. Umeå. . 1975. Systematik i sydsamiskan Vilhelmina-Vefsen Vol. Umeå: Centraltryckeriet. . 1981. Samiska ortsnamn och topografiska ord inom Vilhelmina och Dorotea. Umeå: Centraltryckeriet. Nissen, Kristian. 1921. “Lapponian lichen names”. Studies on the Lichen Flora of Norway ed. by Bernt Lynge. (= Videnskapsselskapets Skrifter 1, M.-N. Klasse 7.), 238–247. Oslo: Jacob Dybwad. Nissilä, V. 1967. “Die lappischen Ortsnamen Schwedens”. Ural-Altaische Jahrbücher 39.118–119. Nystad, Pia Savonius. 1994. “Lapin läänin etunimistä”. Nordlyd: Tromsø University Working Papers on Language and Linguistics 20.174–181. Öhrling, Johan. 1742. “Exercitium academicum nonnullas in orthographiam Lapponicam observationes sistens”. Upsaliae impressum. Olsen, Alf J. 1984. “Samisk, finsk og norsk språk i Lyngen rundt arhundreskiftet”. Heimen 21:3.171–179. Olsen, Magnus & Knut Bergsland. 1943. Lappisk i en islandsk runeinnskrift. Avhandlinger / Det Norske Videnskap-akademi i Oslo. II. Hist.-filos. klasse, Oslo: Jacob Dybwad. Olthuis, Marja-Liisa. 2000. Kielâoppâ [Inarinsaamen kielen kielioppi]. Inari, Finland: Sämitigge. . 2001. “Kyevtstaavvâl pajekielâ -mi- já anarâškielâ -mi- ∼ -me-nominih já toi álgupuáttim”. Lisensiaatti thesis, Oulu University. . 2003. “Sanaston aktiivisen kartuttamisen metodiikkaa: Mallina inarinsaameen luodut biologian uudissanat”. Virittäjä 4.529–544. Outakoski, Hanna. 2003. “On reflexive binding in North Sami”. Nordlyd: Tromsø University Working Papers on Language and Linguistics 31:4.723– 739. Panfilov, Vladimir Zinoveviˇc, Georgij Martynoviˇc Kert & Petr Mefodieviˇc Zajkov. 1988. Obrazcy saamskoj reˇci. Petrozavodsk: Karelskij filial AN SSSR.


291

Pedersen, Christer. 1997. “De snakker samisk i alle barnehager”. Norsk Førskolelærerblad 5.12–15. Pellijeff, Gunnar. 1980. “Ortnamn i språkkontakt”. Ortnamn och språkkontakt: Handlingar från NORNA: Sjätte Symposium i Uppsala 5-7 maj 1978 ed. by Thorsten Andersson, Eva Brylla & Allan Rostvik. 13–27. Uppsala: Nordiska samarbetskommitten för namnforskning. Pineda, David. 2004. “Nekotorye nabljudenija nad zaimstvovanijami v russkom govore Kol’skogo poluostrova”. Poljarnyj vestnik 7.25–40. Pitkänen, Ritva Liisa, Jorma Rekunen & Jyrki Kalliokoski. 1988. Lapinjoen varrelta. (= Castrenianumin toimitteita 30.) Helsinki: Suomalais-Ugrilainen Seura. Posti, Lauri. 1954. “On the origin of the voiceless vowel in Lapp”. Scandinavica et Fenno-Ugrica ed. by Dag Strömbeck, 199–209. Stockholm: Almqvist & Wiksell. Posti, Lauri & Terho Itkonen. 1973. FU-transkription yksinkertaistaminen. (= Castrenianumin toimitteita 7.) Helsinki: Suomalais-Ugrilainen Seura. Qvigstad, Just Knud. 1876. “Beiträge zur Vergleichung des verwandten Wortvorrathes der lappischen und der finnischen Sprache”. Acta Societatis Scientarium Fennicae 7.113–240. . 1893. Nordische Lehnwörter im Lappischen. Christiania [Oslo]: Christiania Videnskabs–Selskabs forhandlinger. . 1899. “Übersicht der Geschichte der lappischen Sprachforschung”. Journal de la Société Finno-Ougrienne 16:3. . 1902. “Lappiske fuglenavne”. Nyt magazin for naturvidenskab 40:3.268– 291. . 1920–1934. Zur Sprach- und Volkskunde der Norwegischen Lappen. (= Oslo Etnografiske Museums Skrifter 1.) Oslo: Oslo Etnografiske Museum. . 1922. Lappische Sprichwörter und Rätsel. Bind I. Hefte 3 Kristiania [Oslo]: Etnografiske Museums skrifter. . 1925. Die lappischen Dialekte in Norwegen. (= Oslo Etnografiske Museums Skrifter 1–1.) Oslo: Oslo Etnografiske Museum. . 1932. Lappische Heilkunde mit Beiträgen von K.B. Wiklund. (= Serie B XX.) Oslo: Instituttet for sammanlignende Kulturforksning. . 1935. De lappiske stedsnavn i Troms fylke. (= Instituttet for sammenlignende kulturforskning. Serie B 28.) Oslo: Aschehoug. . 1938. De lappiske stedsnavn i Finnmark og Nordland fylker. (= Instituttet for sammenlignende kulturforskning. Serie B 33.) Oslo: Aschehoug.

292


. 1944. De lappiske appellative stedsnavn. (= Instituttet for sammenlignende kulturforskning. Serie B 42.) Oslo: Aschehoug. Qvigstad, Just Knud & G. Sandberg. 1888. “Lappische Sprachproben. Einige Bemerkungen von der Zaubertrommel der Lappen”. Journal de la Société FinnoOugrienne 3.1–104. Räisänen, Alpo. 1995. “Kainuun saamelaisperäisiä paikannimiä”. Virittäjä 99:4.532–544. Rankama, Tuija. 1993. “Managing the landscape: A study of Saami place-names in Utsjoki, Finnish Lapland”. Inuit Studies 17.47–69. Rantala, Leif. 1975. De samiska ortnamnen i Polmak: pro gradu uppsats i finskugrisk språkforskning. Master’s thesis, University of Helsinki. Rask, Rasmus. 1832. Ræsonneret lappisk sproglære efter den sprogart, som bruges af fjældlapperne i Porsangerfjorden i Finmarken. Copenhagen: Schubothes Boghandling. Ravila, Paavo. 1931. Ruijanlappalaisia kielennäytteitä Petsamosta ja eteläVarangista. (= Mémoires de la Société Finno-Ougrienne 61.) Helsinki: Suomalais-Ugrilainen Seura. . 1932. Das Quantitätssystem des seelappischen Dialektes von Maattivuono. (= Mémoires de la Société Finno-Ougrienne 62.) Helsinki: Suomalais-Ugrilainen Seura. . 1934. Reste lappischen Volksglaubens. (= Mémoires de la Société FinnoOugrienne 68.) Helsinki: Suomalais-Ugrilainen Seura. . 1935. “Die Stellung des lappischen Innerhalb der finnische-ugrischen Sprachfamilie”. Finnisch-Ugrische Forschungen 23.20–65. . 1956. “Der sog. Stimmlose Vokal im Lappischen”. Ural-Altaische Jahrbücher 28.184–185. . 1960. “Probleme des Stufenwechsels im Lappischen”. Finnisch-Ugrische Forschungen 33.285–325. . 1961. “Kvantiteetti distinktiivisenä tekijänä”. Virittäjä 65.345–350. Rédei, Károly. 1988–1989. “On the Fennic-Lapp Geminate Affricates”. Finnisch-Ugrische Mitteilungen 12/13.63–67. Rießler, Michael. 2001a. “North Scandinavian and Sámi: Two morphosyntactic parallels”. Proceedings of the 18th Scandinavian Conference in Linguistics ed. by Arthur Holmer, Jan-Olof Svantesson & Åke Viberg. 2. 168–179. Lund: Lund University.


293

. 2001b. “Vårvinter – kevättalvi – gid¯dadálvi: om en inlånad årstid i svenska ¯ och norska”. Nordiska språk: Insikter och utsikter. Föredrag från det 6:e studentkollokviet i nordisk språkvetenskap ed. by Jurij Kusmenko & Sven Lange, 54–68. Berlin: Nordeuropa-Institut. . 2002. “Samiskt och finskt substrat i nordskandinaviska dialekter”. Strövtåg i nordisk språkvetenskap. Föredrag från det 5:e studentkollokviet i nordisk språkvetenskap ed. by Jurij Kusmenko & Sven Lange, 78–90. Berlin: Nordeuropa-Institut. Ruong, Israel. 1943. Lappische Verbalableitung dargestellt auf Grundlage des Pitelappischen. (= Uppsala universitets årsskrift 10.) Uppsala and Leipzig: AB Lundequistska Bokhandeln and Otto Harrassowitz. . 1962. “Der Name Arjeplog”. Commentationes Fenno-Ugricae in honorem Paavo Ravila. (= Mémoires de la Société Finno-Ougrienne 125.), 463–467. Helsinki: Suomalais-Ugrilainen Seura. . 1970. Min Sámigiella. Lärobok i samiska. Uppsala: Almqvist and Wiksells Boktryckeri A.B., for Svenska Utbildningsförlaget Liber A.B. Ruoppila, Veikko. 1961. “Lappalainen vai saamelainen?” Virittäjä 65.163–164. Rydving, Håkan. 1986. “Neskol’ko zamechanii o rezul’tatakh issledovaniia saamskoi dialektologii v predelakh LAE”. Soviet Finno-Ugric Studies 22:3.198–202. . 1995a. “Namnbyten i samisk tradition”. Studia Anthroponymica Scandinavica 13.89–99. . 1995b. “Samiska ortnamn: en litteraturöversikt och bibliografi”. Studia Anthroponymica Scandinavica 13.37–62. Saarikivi, Janne. 2004. “Über das saamische Substratnamengut in Nordrussland und Finnland”. Finnisch-Ugrische Forschungen 58.162–234. Saba, Isak. 1920. Om sammensatte lappiske stedsnavn. Tromsø: Tromsø Museum. Sadovszky, Otto L. 1968. “The cardinal-point terms in the Finno-Ugric languages”. Ural-Altaische Jahrbücher 40.229–234. Saitton Burman, Ellen. 1999. Grammatik. Samisk - del 1, svensk - del 2. Jokkmokk, Sweden: Ellens förlag. Sajnovics, János. 1770. Joannis Sajnovics Demonstratio idioma Ungarorum et Lapponum idem esse. Tyrnaviæ [Trnava, Slovakia]: Collegii Academici Societatis Jesu. . 1972. Beweis, daß die Sprache der Ungarn und Lappen dieselbe ist. Wiesbaden: Otto Harrassowitz.

294


Sammallahti, Pekka. 1971. “Ehdotus norjanlapin länsimurteiden fonemaattiseksi transkriptioksi”. In Itkonen et al. 1971, 7–22. . 1975. “Grade alternation in Lappish”. Congressus Tertius Internationalis Fenno-Ugristarum pars I. 449–452. Tallin. . 1977a. “Bridging the gap between two inchoative suffixes: -skande- and Lapp -skoatte-”. Finnisch-Ugrische Forschungen 42.192–194. . 1977b. Norjansaamen itä-enontekiön murteen äänneoppi. (= Mémoires de la Société Finno-Ougrienne 160.) Helsinki: Suomalais-Ugrilainen Seura. . 1980. “Jietnadatoahppa: Fonetihka vuod¯doáššit I”. Diedut ¯ 2. ¯ . 1982. “On grade alternation and the illative plural in Lappish”. FinnischUgrische Forschungen 44.110–113. . 1984a. “The phonology of the Guovdageaidnu dialect of North Saami: Some notes”. Riepmoˇcála. Essays in Honor of Knut Bergsland ed. by Bernt Brendemoen, Even Hovdhaugen & Ole Henrik Magga, 136–152. Oslo: Novus. . 1984b. “Saamelaisten esihistoriallinen tausta kielitieteen valossa”. Tvärminnen symposiumi 17-19.1.1980. 137–155. Helsinki: Societas Scientiarum Fennicae. . 1985. “Die Definition von Sprachgrenzen in einem Kontinuum von Dialekten: Die lappischen Sprachen und einige Grundfragen der Dialektologie”. Dialectologia Uralica ed. by Wolfgang Veenker, 149–158. Wiesbaden: Harrassowitz. . 1989a. “A linguist looks at Saami prehistory”. Acta Borealia 6.3–11. . 1989b. Sámi-suoma sátnegirji = Saamelais-suomalainen sanakirja. Utsjoki, Finland: Jorgaleaddji. . 1998a. The Saami Languages: An introduction. Karasjok, Norway: Davvi Girji. . 1998b. “Saamic”. The Uralic Languages ed. by Daniel Abondolo, 43–95. London: Routledge. . 2002. North Saami Resource Dictionary. (= Publications of the Giellagas Institute 1.) Oulu: University of Oulu. . 2005. Láidehus sámigiela cealkkaoahpa dutkamii. Karasjok, Norway: Davvi Girji. Sammallahti, Pekka & Anastasija Chvorostuchina. 1991. Unna sámi-sam’ sátnegirjjás. Udc’ sám’-sámi soagknegka. Utsjoki, Finland: Girjegiisá.


295

Sammallahti, Pekka & Matti Morottaja. 1983. Säämi-suoma-säämi skovlasänikirje. Inarinsaame-suomi-inarinsaame koulusanakirja. Helsinki: Ruovttueatnan gielaid dutkanguovddas. . 1993. Säämi-suomâ sänikirje. Inarinsaamelais-suomalainen sanakirja. Utsjoki, Finland: Girjegiisá. Sammallahti, Pekka & Jouni Mosnikoff. 1991. Suomi-koltansaame sanakirja = Lää’dd-sää’m sää’nnke’rjj. Utsjoki, Finland: Girjegiisá. Sammallahti, Pekka & Helena Valkeapää. 1984. 1400 sáni. Giellaoahppi sátnelistu: Sámegiela 1400 dábáleamos sáni. Helsinki: Kouluhallitus, Valtion painatuskeskus. Sara, Laila Susanne. 2002. Suorggádusaid geavaheapmi: Guovdageainnu nuoraid giellageavaheapmi [=The use of derivations: The language use of young people in Guovdageaidnu]. Master’s thesis, University of Tromsø. Sauvageot, Aurelien. 1977. “Où situer le lapon?” Études finno-ougriennes 14.7– 18. Scheller, Elisabeth. 2004a. “Paren’ – ett samiskt lånord i ryskan?” In Kusmenko 2004b, 173–182. . 2004b. Samernas språksituation i Sovjetunionen och i dagens Ryssland. Master’s thesis, Umeå universitet. Schiefer, Erhard. 1977. “Zur semantischen Differentia Specifica zwischen Lapp. vuole almai ‘Bier-Mann’ und Deutsch ‘Biermann”’. Études finno-ougriennes 14.169. Schiemenz, Günter Paulus. 1959. “Die Behandlung des Stufenwechsels in den älteren ostseefinnischen und lappischen Grammatiken bis zum Anfang des 19. Jahrhunderts”. Ural-Altaische Jahrbücher 31.362–374. Schlachter, Wolfgang. 1941. “Ein lappisches Kulturdokument”. Ural-Altaische Jahrbücher 21.251–258. . 1942. “Wirkungen von Sprachrhythmus und Satzmelodie im Lappischen”. Ungarische Jahrbücher 22.212–256. . 1953. “Lappische Passivsyntax I”. Ural-Altaische Jahrbücher 25.179– 208. . 1954. “Lappische Passivsyntax II”. Ural-Altaische Jahrbücher 26.149– 171. . 1954–55. “Lappisches im Lappischen Stufenwechsel I-II”. Münchener Studien zur Sprachwissenschaft 4.2–72. . 1955. “Lappisches im Lappischen Stufenwechsel III-IV”. Münchener Studien zur Sprachwissenschaft 7.5–43.

296


. 1957. “Lappische Passivsyntax III”. Ural-Altaische Jahrbücher 29.2–32. . 1958a. Wörterbuch des Waldlappendialektes von Malå und Texte zur Ethnographie. (= Lexica Societatis Fenno-Ugricae 19.) Helsinki: SuomalaisUgrilainen Seura. . 1958b. Wörterbuch des Waldlappendialekts von Malå und Texte zur Ethnographie. (= Lexica Societatis Fenno-Ugricae 14.) Helsinki: SuomalaisUgrilainen Seura. . 1960. “Lappische Passivsyntax IV”. Ural-Altaische Jahrbücher 32.25–64. . 1970. “Das lappische Tempussystem”. Symposion über Syntax der uralischen Sprachen, 15.-18. Juli 1969 in Reinhausen bei Göttingen ed. by Wolfgang Schlachter, 178–195. Göttingen: Vandenhoeck & Ruprecht. . 1973a. “Das erzählende Perfekt im Lappischen”. Journal de la Société Finno-Ougrienne 72.360–369. . 1973b. “Eine Mischkategorie im lappischen Tempussystem”. Commentationes Fenno-Ugricae in Honorem Erkki Itkonen, Mémoires de la Société Finno-Ougrienne, 358–368. Helsinki: Suomalais-Ugrilainen Seura. . 1976. “Anaptyxe im Malålappischen”. Nyelvtudományi Közlemények 78.425–434. . 1983. “Eine grammatische Kategorie im Kontext”. Nyelvtudományi Közlemények 85:2.421–433. . 1991a. Geschichte und Sprachgeschichte betrachtet am Beispiel des Lappischen. Göttingen: Vandenhoeck & Ruprecht. . 1991b. Stufenwechselstörungen im Malålappischen - Aufbau oder Abbau eines Systems? Wiesbaden: Otto Harrassowitz. Sebestyén, I. 1953. “Beiträge zum Problem der protolappischen Sprache”. Acta Linguistica Hungarica 3.291–322. Senkeviˇc-Gudkova, Viktorija V. 1959. “K voprosy o fonetiˇceskix i nekotoryx morfologiceskix obšˇcnostjax kol’sko-saamskix dialektov o samodijskimi jazykami”. Sovešˇcanie po voprosam istoriˇceskoj grammatiki i istoriˇceskoj dialektologii finno-ugorskix jazykov. Moscow: Akademija nauk SSSR. . 1960. “Ustinja Pavlovna Tarunova lapi muinasjutte”. Emakeele seltsi aastaraamat 6.210–219. . 1963. “Morfologiˇceskaja struktura saamskogo jazyka”. Morfologiˇceskaja struktura slova v jazykax razliˇcnyx tipov ed. by V. M. Žirmunskij, 244–251. Moscow: Akademija nauk SSSR.


297

. 1965a. “O nekotoryx elementax analitizma v saamskom jazyke (na materiale notozerskogo dialekta)”. Analitiˇceskie konstrukcii v jazykax razliˇcnyx tipov ed. by V. M. Žirmunskij & O. P. Sunik, 333–340. Moscow: Nauka. . 1965b. “Osobennosti aglutinacii v saamskom jazyke (na materiale sklonenija sušˇcestvitel’nyx notozerskogo dialekta)”. Morfologiˇceskaja tipologija i problema klassifikacii jazykov ed. by B. A. Serebrennikov. Moscow: Nauka. . 1966. “Zakon kontrastnoj korrel’acii koliˇcestva i ego vlijanie na systemu fonem notozerskogo dialekta saamskogo jazyka”. Issledovanija po fonologii, 368–375. Moscow: Nauka. . 1968. “O differenc’alnyx morfologiˇceskix priznakax imeni prilagatelynogo v saamskom jazyke (ne materiale notozerskogo dialekta)”. Voprocy teorii cˇ astej reˇcj na materiale jazykov pazliˇcnyx tipov ed. by B.M. Žimunskij & O.P. Sunik. Leningrad: Nauka. . 1971. “Fonologiˇceskaja struktura russkix saimstvovanij v notozerskom dialekte saamskogo jazyka”. Voprosy vzaimodeijstvija Pribaltiijsko-Finskix jazykov s inostemnymi jazykami. K 80-letiju so dn’a roždenija D. V. Bubrixa ed. by G. M. Kert & M. N. Mullonen. (= Pribaltijsko-finskoe jazykoznanie 5.), 46–51. Leningrad: Akademia nauk SSR Karel’skij filial. . 1978. “K voprosy o passivnyx konstrukcijax saamskogo jazyka (na materiale saamskix dialektov kol’skogo poluostrova)”. Problemy teorii grammatiˇceskogo zaloga ed. by B.S. Xrakovskij, 274–277. Leningrad: Nauka. Setälä, Emil Nestor. 1890. “Lappische lieder aus dem XVII:ten Jahrhundert”. Journal de la Société Finno-Ougrienne 8.105–123. . 1899. “Sabme–Häme”. Virittäjä 3.109–110. Sinor, Denis, ed. 1988. The Uralic Languages: Description, history, and foreign influences. Leiden: E. J. Brill. Sköld, Tryggve. 1961. Die Kriterien der urnordischen Lehnwörter im Lappischen. 1, Einleitende Kapitel anlautender Konsonantismus, Vokalismus der ersten Silbe. (= Skrifter utgivna av Institutionen för nordiska språk vid Uppsala universitet 8.) Uppsala: Uppsala universitet. . 1964. “The Scandinavian noður and the Lappish system of orientation”. Lapponica: Ethnographica Upsaliensia 21.267–283. Uppsala. . 2004. “Ömsesidig kolartro”. In Kusmenko 2004b, 35–39. Söderholm, Eira. 1994. “Topografiset appellatiivit kielikontaktissa - esimerkkinä suomalais-lappalainen sanapari korva-goarvi”. Nordlyd: Tromsø University Working Papers on Language and Linguistics 20.98–109.

298


Söderström, Sven. 1982. “Från en sydsamisk ordbok”. Fenno-Ugrica Suecana 5.305–323. Solbakk, Aage. 1997. Sámi cˇ állingiela historjá. (= Gielas gillii mielas millii 1.) Karasjok, Norway: Davvi Girji. Spiik, Nils Eric. 1989. Lulesamisk Grammatik. Jokkmokk, Sweden: Sameskolstyrelsen. Steckzén, Birger. 1969. “Studier i lapska ortnamn”. Norrbottenfångstmän, fiskare, jägare, Norrbottens Museum, årsbok 1969, 13–48. Århus: Klicheer Buchtrups Cliche och Offsetanstalt. Steen, Adolf. 1963. “Et samisk navnesystem”. Sameliv 5.108–117. Oslo. Stipa, Günter J. 1983. “Die Stellung des Lappischen in Skandinavien in sprachkundlicher Sicht”. Sprache und Volk im 18. Jahrhundert ed. by HansHermann Bartens, 139–152. Frankfurt: Peter Lang. Stockfleth, Nils Vibe. 1840. Grammatik i det lappiske Sprog, saaledes som det tales i norsk-Finmarken. 1, Bogstav- og Formlæren. Christiania [Oslo]: Grøndahl. . 1852. Norsk-lappisk ordbog. Christiania [Oslo]: Cappelen. Storjord, Thorbjørn. 1998. Tysfjord: Divtasvuodna. Samiske og norske stadnamn. Drag, Norway: Árran. Sverloff, Matti. 2003. Suenjelin saamelaisten perintö. Vammala, Finland: Vammalan Kirjapaino Oy. Svonni, Mikael. 1990. Sámi-ruota, ruota-sámi sátnegirji = Samisk-svensk, svensk-samisk ordbok. Jokkmokk, Sweden: Sámi Girjjit. . 1993. Samiska skolbarns samiska: en undersökning av minoritetsspråksbehärskning i en språkbyteskontext. (= Umeå Studies in the Humanities 113.) Umeå: Umeå universitet. . 1996. “The future of Saami. Minority language survival in circumpolar Scandinavia”. Northern People, Southern States: Maintaining ethnicities in the circumpolar world ed. by R.P. Wheelersburg, 13–42. Umeå: SERUM. . 1997. “Att lära sig samiska - ett minoritetsspråk”. Mer än ett språk: en antologi om två- och trespråkigheten i norra Sverige ed. by Eva Westergren och Hans Åhl, 130–152. Stockholm: Norstedt. . 2006. “Umesamiskan - Det gåtfulla språket”. Sápmi Y1K - Livet i samernas bosättningsområde för ett tusen år sedan ed. by Andrea Amft & Mikael Svonni. (= Sami Studies 3.), 151–170. Umeå: Umeå universitet. Svonni, Mikael & Mikael Vinka. 2002. Nordsamisk grammatik. Del 1. Fonologi och morfologi. Preliminär version 2002. Umeå: Umeå universitet.


299

. 2003. “Constraints on the Morphological Causatives in the Torne Dialect of North Sámi”. In Nelson & Manninen 2003, 343–380. Swedell, Ulla. 1992. “Om lånenamn i en Gällivareby”. Ortnamnssällskapets i Uppsala Årsskrift 36–42. Szabó, László. 1968a. “Faktitive Verben im Kildinschen Dialekt des Kolalappischen”. Ural-Altaische Jahrbücher 40.18–21. . 1968b. Kolalappische Volksdichtung. Teil I: Texte aus den Dialekten in Kildin und Ter. Teil II: Grammatische Aufzeichnungen. Göttingen: Akademie der Wissenschaften in Göttingen. . 1969a. “Das Referat im Kolalappischen”. Glossa 3.165–189. . 1969b. “Referaatti kuolanlapissa. Das Referat im Kolalappischen”. Virittäjä 73.164–178. . 1983. “Depalatalization or palatalization in Russian loanwords in KolaLappish”. Papers from the Annual Meetings of the Atlantic Provinces Linguistic Association (PAMAPLA) 7.77–83. . 1984a. “The function of the essive, comitative and abessive in Kola Lappisch”. Nordlyd: Tromsø University Working Papers on Language and Linguistics 9.156–171. . 1984b. “The function of the inessive-elative and the dative-illative in Kola Lappisch”. Nordlyd: Tromsø University Working Papers on Language and Linguistics 8.4–52. . 1984c. “The Function of the Nominative, Accusative, Genitive and Partitive in Kola Lappisch”. Nordlyd: Tromsø University Working Papers on Language and Linguistics 9.71–155. . 1985a. “The inchoative in Kildin-Sami sentences contrasted with their Russian translations”. Papers from the Annual Meetings of the Atlantic Provinces Linguistic Association (PAMAPLA) ed. by Lilian Falk, K. Flikeid & M. Harry. 9. 137–144. Canada: Atlantic Provinces Linguistic Association. . 1985b. “Infinitive and participial constructions in Kola Sami”. Nordlyd: Tromsø University Working Papers on Language and Linguistics 10.28–127. . 1987a. “Peculiarities of negation in language: The example of Kola-Lapp”. Papers from the Annual Meetings of the Atlantic Provinces Linguistic Association (PAMAPLA) 11.130–135. . 1987b. “The use of the inchoative in Kola-Sami sentences”. Nordlyd: Tromsø University Working Papers on Language and Linguistics 13.70–103. . 1988. “Possession in Ter Sami”. Papers from the Annual Meetings of the Atlantic Provinces Linguistic Association (PAMAPLA) 12.91–97.

300


Thomsen, Vilhelm. 1870. Über den Einfluss der germanischen Sprachen auf die finnisch-lappischen: eine sprachgeschichtliche Untersuchung. Halle: Verlag der Buchhandlung des Waisenhauses. . 1997[1869]. On The Influence of Germanic Language on Finnic and Lapp. (= Uralic & Altaic Series 87.) London: Routledge and Curzon. Thuen, Trond. 1985. “Samisk og norsk i en nordnorsk kystbygd”. Forskningsnytt 30:6/7.9–13. Tiesmaa, Erkki. 1937. “Lappalaisia lainoja Kittilän porosanastossa”. Virittäjä 41.168–171. Tjihkkom, Kåre. 1993. Rättelseblad för Nils Eric Spiiks Lulesamisk grammatik. Kristiansund: Th. Blaasværs forlag. Toivonen, Yrjö Henrik. 1949. “Protolapin ongelmasta”. Suomen Tiedeakatemian Esitelmät ja Pöytäkirjat 159–183. Tonstad, Per Lars. 1987. “Samisk i hjemmet, norsk i samfunnet”. Skoleforum 86:22.14–15. Torstensson, Eivor Nylund. 1982. “Lapin sananalkuisen cˇ –äänteen vastinesita suomalaisissa lainasanoissa ja paikannimissä”. Fenno-Ugrica Suecana 5.213– 223. Trosterud, Trond. 1991a. “Binding relations in two Finnmark Finnish dialects: A comparative syntactic study”. University of Trondheim Working Papers in Linguistics 12. . 1991b. “Lokalkasus og preposisjonar i finsk, kvensk, samisk og norsk”. Norsk Lingvistisk Tidsskrift 9:1.50–78. . 1994. “Auxiliaries, negative verbs and word order in the Sami and Finnic languages”. Minor Uralic Languages: Structure and development ed. by Ago Künnap, 173–181. Groningen: University of Groningen. . 1996. “Die südsamische Wortfolge als eine Kombination der deutschen und der marischen Wortfolge analysiert”. In Larsson 1996, 103–111. Tunkelo, E.A. 1902. “Elias Lönnrot Suomen lappalaisten äidinkielen puolesta”. Virittäjä 6.49–61. Turi, Else R. 1996. Kárásjoga mánáid vearbahámiid morfologalas variašuvnna guorahallan. Master’s thesis, University of Tromsø. . 1997. “Davvisámegiela morfologiija: kompendia”. UNIKOM, University of Tromsø. Turkin, Adolf. 1994. “O vzaimootnošenijax komi i saamskogo jazykov”. Linguistica Uralica 30:1.19–27.


301

Ulseth, Brit. 1981a. Duration Differences as a Cue for Consonant Gradation in Lappish. (= The Sounds of Lappish 1.) Umeå: Department of Phonetics, Umeå universitet. . 1981b. Varighetsforskjeller som grunnlag for stadieveksling i en samisk dialekt. (= The Sounds of Lappish 3.) Umeå: Department of Phonetics, Umeå universitet. Ureland, P. Sture. 1995. “Kontaktzonen zwischen Nordgermanisch und FinnoUgrisch”. Tijdschrift voor Skandinavistiek 16:2.45–83. Vest, Jovnna-Ánde. 1995. Synonymasátnegirji. Inari, Finland: Lappi leanaráehus. Vest, Jovnna-Ánde & Marsa Luukkonen. 1992. Matkalle mukaan: suomi-saamesuomi sanakirja. Jyväskylä, Finland: Gummerus. Viitso, Tiit Rein. 1996. “Ostseefinnisch und Lappisch”. In Larsson 1996, 113– 121. Vikør, Lars S. 2001. The Nordic Languages: Their status and interrelation (3rd ed). Oslo: Novus. Vinka, Mikael. 1999. “Causatives in Northern Sámi and the role of dative case”. Proceedings of WECOL 1998 ed. by E. van Gelderen & V. Samiian. Fresno: California State University, Fresno. . 2003. Causativization in North Sami. Ph.D. thesis, McGill University, Montreal. Vuolab, Kaarina. 1996. Man vai maid? Davvisámegiela relatiiva - ja interrogatiivapronomena mii ovttaidlogu akkusatiivahámid válljašupmi objeavtta funkšuvnnas. Master’s thesis, University of Oulu. Whitaker, Ian. 1953. “The Lapps and their names”. Man 53.164. London. . 1979. “The kinship and affinal terminology of the Karesuando Lapps (Sami)”. Finnisch-Ugrische Forschungen: Zeitschrift für Finnisch-Ugrische Sprach- und Volkskunde 43.80–110. Wickman, Bo. 1957. “Über den Ursprung des lappischen a”. Uppsala universitets årsskrift 12.43–53. . 1959. “Swedish contributions to Lapp linguistics”. Journal of the Royal Anthropological Institute of Great Britain and Ireland 89:2.149–153. . 1965. “The origins of the Lappish hunting and fishing terminology”. Hunting and Fishing: Nordic symposium on life in a traditional hunting and fishing milieu in prehistoric times and up to the present day ed. by Harald Hvarfner, 501–512. Luleå: Norrbottens Museum. Translated from Swedish by Donald Burton.

302


. 1975. “Phonologische Analyse des Lulelappischen”. Symposion: Phonologische Analyse der uralischen Sprachen. Linguistische Studien Reihe A Arbeitsberichte 22.1–5. . 1978. “Grammatical homonymy in Lappish”. Fenno-Ugrica Suecana 1.91–94. . 1980. “Some case forms in Lappish”. Lingua Posnaniensis 23.205–207. . 1982. “Nordisk påverkan på samisk syntax”. Språkhistoria och språkkontakt i Finland och Nord-Skandinavien. Studier tillägnade Tryggve Sköld ed. by K.-H. Dahlestedt, 279–285. . 1996. “Über den Ursprung des lappischen s”. In Larsson 1996, 123–126. Wiklund, Karl Bernhard. 1890. Lule-lappisches Wörterbuch. (= Mémoires de la Société Finno-Ougrienne 1.) Helsinki: Suomalais-Ugrilainen Seura. . 1891. “Laut- und Formenlehre der Lule-Lappischen Dialekte”. Göteborgs Kongl. Vetenskaps och Vitterhets Samhälles Handlingar. Ny Tidsföljd 25. . 1892a. “Die nordischen lehnwörter in den russisch-lappischen dialekten”. Journal de la Société Finno-Ougrienne 10–12.146–206. . 1892b. “Ein Beispiel des Lativs im Lappischen”. Suomalais-Ugrilainen Seura 10.207–209. . 1892c. “Nomen-verba im Lappischen”. Suomalais-Ugrilainen Seura 10.210–216. . 1894. Kleine lappische Chrestomathie mit Glossar. (= Apuneuvoja suomalais-ugrilaisten kielten opintoja varten 1.) Helsinki: Suomalaisen Kirjallisuuden Seura. . 1896. Entwurf einer urlappischen Lautlehre. I. Einleitung, Quantitätsgesetze, Accent, Geschichte der hauptbetonten Vokale. (= Mémoires de la Société Finno-Ougrienne 10.) Helsinki: Suomalais-Ugrilainen Seura. . 1901. Lärobok i lapska språket. Uppsala: Lundequistska bokhandeln. . 1902. “Zur Geschichte des urlappischen a¯ and u¯ in unbetoner Silbe”. Finnisch-Ugrische Forschungen 2.41–71. . 1905–1906. “Zur Geschichte der lappischen Affrikaten”. Journal de la Société Finno-Ougrienne 23. . 1906. “Zur Lehre vom Stufenwechsel im Lappischen”. Finnisch-Ugrische Forschungen 6.5–16. . 1911. “Lapparnas forna utbredning i Finland och Ryssland belyst af ortnamen”. Le Monde orientale 7.101–136, 175–196. . 1912. “Einige urnordische lehnwörter im Lappischen”. Finnisch-Ugrische Forschungen 12.30–37.


303

. 1913. “Stufenwechselstudien. I”. Le Monde orientale 7.197–239. . 1914. “Urnordiska ortnamn i de södra lappmarkerna”. Namn och bygd 2:105–120. . 1915a. Lärobok i lapska språket 2. revid. uppl. Uppsala: A.-B. Akademiska bokhandeln. . 1915b. “Stufenwechselstudien. II”. Le Monde orientale 9.171–239. . 1916. “Lapska namn på ren-oestriderna och deras larver”. Le Monde orientale 10:3.183–191. . 1917. “Die ältesten germanischen Lehnwörter im Finnischen”. Indogermanische Forschungen 38:1/2.48–115. . 1919. “Stufenwechselstudien. III”. Le Monde orientale 13.55–121. . 1921. “Kolmiasteiden astevaihtelu”. Virittäjä 25.7–14. . 1927. Lappische Studien. 1-3. Uppsala: Almquist & Wiksell. . 1928. “Das lappische Verbaladverbium und einige andere Kasus des Verbalstammes”. Festskrift til rektor J. Qvigstad. (= Tromsø Museums skrifter 2.), 316–353. Tromsø: Tromsø Museum. Work, Lisa An. 1993. Epenthesis in Saami. Master’s thesis, University of Trondheim. Ylikoski, Jussi. 2002. “Zu den adverbialen Nominalkonstruktionen im Nordlappischen”. Finnisch-Ugrische Forschungen 57.68–166. . 2004. “Zu den adverbialen Nominalkonstruktionen im Nordlappischen II. Finale Konstruktionen”. Finnisch-Ugrische Forschungen 58.57–161. . 2006a. “Davvisámegiela -nláhkai ∼ -nládje suffiksála ii-finihtta vearbaráhkadusat”. Sámi diedalaš áigeˇcála 1/2006.18–38. ¯ . 2006b. “Fuomášumit sámegiela adposišuvnnaid funkšuvnnain: ovdamearkan alde- ja ala-postposišuvnnaid ii-lokála geavaheapmi”. Sámi diedalaš ¯ áigeˇcála 1/2006.39–61. Ylikoski, Jussi & Ante Aikio, eds. 2007. Sámit, sánit, sátnehámit. Riepmoˇcála Pekka Sammallahtii miessemánu 21. beaivve 2007. [A Festschrift for Pekka Sammallahti on the Occasion of his 60th Birthday on May 21, 2007.]. (= Mémoires de la Société Finno-Ougrienne 253.) Helsinki: Suomalais-Ugrilainen Seura. Zajkov, Pekka (=Petr Mefodieviˇc). 1996. “Sprachliche Besonderheiten des Akkalalappischen”. In Larsson 1996, 137–142. Zajkov, Petr Mefodieviˇc. 1987. Babinskij dialekt saamskogo jazyka: fonologomorfologiˇceskoe issledovanie. Petrozavodsk: Karelija.

INDEX A abessive, 11–13, 61 ablative, 91, 104, 107, 109, 121, 126, 127 absolute declension, 7, 54 Abstract Declarative Phonology, 54–58, 87 accent, 169, 190, 192, 195 accent theory, 190, 192 accentual, 192, 194, 196 accomplishment verb of accomplishment, 9 accusative, 7, 9, 12, 13, 53, 61, 62, 68, 84, 94, 98, 101–103, 110, 111, 113, 121, 123, 125, 207, 208, 210, 211, 214–223, 241, 243–245, 248, 250, 254 acoustic, 57, 168, 171–173, 178–188, 192, 194–196 acquisition, see language learning activation, 57 active voice, 158 adessive, 109, 211, 248, 250 adjective, 97, 98, 128, 138–142, 145–149, 151, 153, 154, 156, 157, 159 adjectival, 138, 139, 142, 144–149, 151, 152, 157, 158, 160 adjective-like, 150 non-attributive, 117 possessive, 124 adjunct, 235, 245 ADP, see Abstract Declarative Phonology

adverb, 107, 109 spatial adverb, 109, 110 adverbial, 13, 142, 159, 207–223 duration adverbial, see duration temporal adverbial, 9, 207, 209–212, 214, 217, 222, 223 Äimä, Frans, 172 affricate, 61, 76 agent, 140, 146, 156, 162, 163, 253 agreement adjective agreement, 139 full agreement, 227–255 on negative auxiliary, 8, 9 partial agreement, 14, 227–255 possessive agreement, 11 split agreement, 227–255 verbal agreement, 8, 9, 11, 12, 14, 114–116, 227–255 Akkala Saami, 1–3, 103, 105 allophone, 58 allophonic, 76 allophony, 58 alveolar, 116 amplitude, 179, 181, 182 analogy, 22, 23, 30, 39, 42, 103, 107, 117, 119, 122, 123 analogical, 93, 102–104, 107, 121, 123, 126, 128, 129, 131, 132 analogical reanalysis, see reanalysis systemic analogy, 42 anaphor, 151 anaphoric, 250

306 animacy, 227, 229–231, 234, 237, 239, 242, 247, 249, 251 animacy split, 231 animate, 99, 101, 227, 230, 231, 233, 234, 237, 242 hierarchy, see hierarchy human (feature), 231, 233, 236–239 inanimate, 14, 99, 227, 230, 231, 234, 239, 242 aperture, 58 apocope, 18, 105, 123–125, 187 apocopation, 18 approximant, 60 Arabic, 20, 209 Classical Arabic, 251 archaic, 213 argument, 138, 141, 144, 151, 156, 214, 220 argument structure, 9, 137 external argument, 154, 156, 157, 162, 163 internal argument, 215, 216, 218–220 a(rgument)-structure, 252, 253, 255 articulation, 74 articulatory, 56, 57, 179, 192 articulatory energy, 192 articulatory plan, 53, 57, 58, 87 aspect, 207, 219 aspectual, 9, 11, 207, 208, 210–212, 215, 216, 218–220, 222, 223 inceptive aspect, 140 aspirated, 59 unaspirated, 76, 77, 92 Athapaskan, 170, 188

INDEX

attribute, see feature attributive, 139 auditory properties, 57 augmentative, 10 Austronesian, 20, 22, 44 B back-formation, 106 Baltic Proto-Baltic, 106 Basque, 20–23 Pre-Basque, 21 beneficiary, 253 Bergsland, Knut, 61, 171 bi-uniqueness, 188 bidirectional, see directionality bilingual, 18, 19, 21, 25, 36 bilingualism, 19, 25 binary, 94, 95, 111, 112, 167, 168, 188, 194–197 binding, 235 bisyllabic, see disyllabic blocking, 111, 237–240, 251 bound aspectual bound, 207, 223 bounded, 142, 144, 208, 210, 217, 218, 220–223 boundedness, 210, 217, 219 unbounded, 211, 214, 217, 219, 220, 222 C Caddoan, 170, 188 calque, 22 caritive, 23 case grammatical case, 207, 223 lexical case, 243

INDEX

oblique case, 94, 95, 103, 118, 121, 127–129, 207–210, 214, 217, 223 semantic case, 243, 254 structural case, 210, 213, 222 Caucasian, 115 causative, 9, 137, 138, 140, 145, 156–161, 163 causative verb, see verb causativizer, 145, 158, 160 Celtic, 19, 22 Central Siberian Yupik, 170 change of state, see state Chinese, 209 chronology, 43 co-head, 245 coda, 84, 187 cognate, 17–19, 22, 23, 26, 29, 30, 33, 39, 41, 43, 106, 124 quasi-cognates, 30 Collinder, Björn, 4, 178 comitative, 7, 13, 61, 62, 68, 92, 95, 96, 117–121, 123, 124, 126, 127, 130 comparative linguistics, 18, 19, 43, 44 comparative method, 20 compensatory, 4, 175–178, 183, 187, 189, 194–196 complex predicate, 141, 153 computation, 57, 58 computational system, 56 conflation, 91, 105, 112–115, 127, 131 connegative, 147 consonant centre, 63, 65, 66, 69–76, 79–82, 84, 86, 87

307 consonant cluster, 63, 65, 66, 69, 79, 82, 84, 105, 106, 120 consonant gradation, see grade alternation consonant peak, see peak consonant quantity, see quantity c(onstituent)-structure, 235 constraint-based, 54 continuant, 60, 61, 74, 76–78 non-continuant, 60 continuative, 39, 40 contracted, 175, 181, 197 non-contracted, 176 contrastive, 167–170, 180, 186, 187, 189–191 non-contrastive, 190 control, 235 coronal, 61 non-coronal, 102 correspondence mimicry, 18, 19, 22 covert affinities, 95 covert distinction, 57 covert level of grammar, 130 covert reanalysis, see reanalysis D dative, 91 DATR, 98 deadjectival, 10, 137, 138, 141, 144, 148, 152–156 debuccalization, 76 Declarative Phonology, 54, 55, 87 default, 95, 97, 98, 111, 115–117, 222, 229, 233, 234, 248, 251, 252, 254 definiteness, 217, 233 definite, 233, 249 indefinite, 233

308 degree alternation, 169 delimiter, see event delimiter, see event delimiter demonstrative, 100, 130, 213 denasalization, see nasal denominal, 10, 138, 141, 148–156 dependent marking, 11 depictive, 216 derivational affix, 10, 124 derivational base, 152 derivational element, 147 derivational marker, 139 derivational morphology, 10 derivational pattern, 139 Derivational Rule-Based Theory (DT), 55 desonorization, see sonorant determiner, 97, 150–152, 245 devoicing, see voicing diachronic, 5, 74, 92, 93, 102–110, 117, 122, 125, 126, 129–132, see also historical diachrony, 131 dialect, 17, 20, 24, 25, 35, 40, 41, 43, 47, 50, 92, 93, 106–108, 114, 117, 120, 121, 123–126, 128, 129, 131, 132 dialect joke, 17 dialectal, 29, 31, 34, 35, 40, 41, 92, 108, 123, 124 phonological dialect, 20 diminutive, 10 reduplicated, 10 diphthong, 7, 18, 32, 34, 36, 59, 84, 85, 173, 180 diphthongize, 26, 34

INDEX

direct object, see object directionality, 99, 100, 128, 131 bidirectional, 99, 100 discourse, 12, 250 discourse referent, 148, 151, 154 disjunction, 112 disyllable, 63, 69, 80, 82 disyllabic, 5, 54, 61–63, 67–69, 81, 82, 123, 125, 126, 169, 171–173, 175–178, 180–185, 187, 189–191, 195–197 dorsal, 60, 61, 74, 76, 77, 79 hyperdorsal, 61, 64, 74, 75 hyperdorsality, 64 DP, see Declarative Phonology DT, see Derivational Rule-Based Theory dual, 114, 115, 119, 121, 228–230, 232–234, 236–239, 242, 245, 247 duration, 5, 59, 76, 167–197, 210, 212, 214, 219, 220 ternary duration distinction, see ternary dynamic, 144 E East Lappish, 171 elative, 104–110, 121, 125 Elsewhere Condition, 56, 68 Elsewhere Principle, 55 endangered, 196 endpoint, 210, 211, 214, 215, 219 English, 32, 33, 42, 145, 148, 150, 153, 154, 156, 246 Engstrand, Olle, 4, 178

INDEX

entity-denoting, 138, 146, 147, 152–154, 156 essive, 7, 13, 54, 61–63, 67, 68, 109, 121–124, 126, 139, 209, 211 Estonian, 17, 18, 29, 124, 168–173, 189–197 etymological nativization, 17–50 event, 9, 137, 138, 142, 156, 158, 206–223 complex event, 137, 163 event delimiter, 206–223 event measure, 208, 210, 215, 216, 219 event structure, 137, 207, 208, 218, 220, 223 eventive, 157, 163 Single Delimiting Constraint, see Single Delimiting Constraint eventualities, see event exemplar, 57, 60, 87 exemplar cloud, 57, 58, 87 exemplar theory, 57 existential, 231, 233, 248, 254, 255 experiencer, 253 experimental, 168, 172, 180, 190, 192 exponence, see rule of exponence exponent, 115, 116 extinct, 24, 25, 40 Eyak, 19 F F0, 181–186, 192, 193 feature, 55–58, 64, 67, 94–95, 100, 110–112, 126, 127, 234–240, 245, 252

309

feature, 234–236 animate, see animacy attribute, 98, 126, 235, 237 binary feature, 4, 67, 94, 95, 111, 112 feature change, 55 feature logic, 240 feature structure, 235, 236 feature-changing rule, 55 floating feature, 117 human, see animacy hyperfeature, 56, 60, 61 neutralization, see neutralization primitive feature, 240 valence, 240 value, 94, 95, 98, 99, 102, 110, 111, 113, 235, 237, 240 feature neutralization, 94–95 Fennic, see Finnic Fenno-Saami Proto-Fenno-Saami, 103–107, 109, 110, 121–125 fill-out, 64–66, 69, 72, 75, 76, 79–82 final obstruent devoicing, see voicing Finnic, 18, 21, 23–44, 123, 124 Early Pre-Finnic, see Proto-Fenno-Saami Pre-Finnic, 24, 25, 29, 31 Proto-Fennic, see Proto-Finnic Proto-Finnic, 24–26, 29, 30, 41, 106, 121 Finnish, 2, 17–50, 102, 103, 105, 106, 113, 121, 122, 124, 156, 208, 209, 211, 214,

310 216–223, 228, 233, 247–252 Finnmark Saami, see North Saami Finno-Karelian, 24 Finno-Ugric, 2, 8, 18, 207, 208, 245, 252 Finno-Ugric Transcription, 58, 59, 63 FOG, see Foot-junctural Obstruent Gemination folk etymology, 42 foot, 59, 68, 75, 169, 172, 173, 177, 189–192, 196 foot boundary, 54 Foot Minimality, 86 foot-juncture, 72, 75–76 foot-medial position, 54, 72, 73 Foot-junctural Obstruent Gemination (FOG), 75, 76, 86 French, 42 frequency, 210 fundamental frequency, see fundamental frequency frequentative, 39, 40 frequentative-continuative, 39, 40 non-frequentative, 40 fricative, 60, 61 hyperfricative, 64 function word, 122 f(unctional)-structure, 235, 239, 240, 253 fundamental frequency, 169, 171, 180, 182–185, 192, 197 FUT, see Finno-Ugric Transcription

INDEX

G Gaelic, see Scottish Gaelic geminate, 18, 30, 41, 54, 60, 63–66, 69–76, 78–79, 81–84, 87, 120 gender, 22, 41, 94, 115, 116 genitive, 7, 12, 13, 61, 84, 102, 103, 119, 121, 123, 208, 211–214, 223, 241, 244 German, 42, 94, 95, 97–99, 101, 168, 170, 185–187, 193–195 North Saxon, 186, 187, 195 Germanic, 2 gerund, 11 goal, 253 gradation, see grade alternation grade alternation, 2, 4–5, 11, 53–87, 97, 122, 140, 147, 154 strong grade, 54, 63–65, 67, 69–74, 77, 79–82, 84, 154 weak grade, 54, 63–65, 67, 71, 73–76, 78, 79, 84, 122, 154 grammatical case, see case grammatical function, 2, 11–14, 228, 240, 246, 247, 251, 252 grammaticalized, 228, 251, 252 grammtical relation, see grammatical function H Head-driven Phrase Structure Grammar (HPSG), 240 head-marking, see agreement height, 58 hierarchy animacy hierarchy, 234

INDEX

inheritance hierarchy, 98 obliqueness hierarchy, 240 thematic hierarchy, 241, 250, 253 Hindi, 32 historical, 2, 4, 17–50, 93, 102–104, 109, 118, 121–126, 131, 132, 167, 175, 176, 178, 187, 194–197, 251, see also diachronic history, 123, 132 homonymy, 92–94, 100–103, 112–114, 116–118, 120, 126, 131 homonymous, 114 homophony, 94, 100, 119, 244 homophonous, 94, 95, 125, 243, 244 Hopi, 170, 187, 188, 193, 194 Hua, 114–116 hypercategory, 56, 58, 87 hyperdorsal, see dorsal hyperfeature, see feature hyperfricative, see fricative hyperobstruent, see obstruent hyperocclusive, see occlusive hypersegment, 56, 60, 61, 63, 66, 78 hypersegmental, 60 hypersingleton, see singleton hypersonorant, see sonorant I iamb, 82 q(uantitative)-iamb, 82 iambic, 63, 65–67, 70, 80, 84 q(uantitative)-iambic, 63, 64, 69, 82, 83 Icelandic, 245

311 Ilgar, 22 illative, 7, 13, 61–63, 65, 67–70, 79, 80, 104, 105, 107, 109, 119, 127, 130, 154 imparisyllabic, 119 imperfectivity, 222 imperfective, 219, 220 Inari Saami, 1–5, 7, 9, 11–13, 31, 38, 41, 53–87, 103, 105, 106, 124, 172, 206–223, 226–255 inchoative, 137, 138, 142–144, 152–157, 159–161, 163 incremental theme, see theme indefinite, see definiteness Indo-European, 33 inessive, 92, 104–110, 121, 125, 211, 217, 218 inferential, 96 inferential-realizational, 93, 96, 100, 127 infinitive, 10, 147 infinitival, 139 infinitival marker, 139, 147 inflectional, 7, 21, 91–132, 141 inheritance, 30, 44 inheritance hierarchy, 98 initial-dropping, 19 input, 55 instability, 167–197 instrument, 253 instrumental, 126, 127 intensity, 169, 171, 179–183, 185, 191, 192 International Phonetic Alphabet, see IPA

312 intervocalic, 21, 25, 37, 54, 105, 172, 173, 175, 177 intonation, 193 inventory phonological inventory, 188 vowel inventory, 188 IPA, 58 Irish Pre-Irish, 19 Itkonen, Erkki, 4, 21, 25, 41, 53, 59, 62, 81, 84, 120, 124, 170–172, 178 Iwaidja, 22 K Karelian, 24 Khinalug, 115, 116 Kildin Saami, 1–3, 10, 11, 25, 31, 103, 105 Korean, 209 Korhonen, Mikko, 22, 25, 28, 96, 104–106, 108, 121, 123–126, 140, 170, 171, 173, 178–182, 185, 194, 197 L labial, 116, 117 labiodental, 78 labiodentalization, 74 labiodentalized, 76 laboratory phonology, 168, 170 lamino-alveolar, 57 language learning, 4, 129–131 imperfect learning, 129, 132 language transmission, 129 laryngeal, 77 Latin, 19, 91

INDEX

lative, 104, 109 latus, 62, 63, 65, 66, 69, 74, 79–86 learning, see language learning length, 4–6, 85, 167–197 lengthening, 4, 75, 76, 85–87, 176–178, 183, 186, 189, 194–196 ternary length distinction, see ternary lenis, 186 lenited, 117 lenition, 178 lexeme, 93, 96–99, 111, 128 lexical, 94, 96, 98, 100, 102, 110, 234–240, 251–255 lexical-conceptual structure, 137 Lexical-Functional Grammar (LFG), 234, 238–240, 251, 252, 254 Lexical-Mapping Theory (LMT), 240, 245, 252–255 lexicalised, 213 lexicalist, 227, 234 light verb, 143 list value, 240 loan adaptation, 19, 20, 22 loan nativization, 19 loanword, 17–50 location, 104, 109, 112 locational, 109 locative, 7, 13, 61, 62, 68, 92, 93, 100, 104–110, 117–121, 123–127, 129, 130, 208–214, 222, 223, 243, 253–255 locative construction, 251, 255 logical subject, see subject

INDEX

Lule Saami, 1–4, 8, 14, 38, 103, 105–109, 121, 124 M Magga, Tuomas, 4, 65, 172, 178, 180 Maltese, 251 Mari, 26, 104 markedness, 95 mass/count, 217 Maung, 22 measure, see event measure medial, 54 foot-medial, 54, 72, 73 word-medial, 60 merge, 137, 144, 153, 235 merger, 105–107 metrical, 62, 63, 68, 69, 169, 173 minimal triplet, 179, 181, 185, 188, 196 Mixe, 170, 187, 193–195 monophthong, 59, 84, 85, 173, 180 monophthongization, 18 monostratal, 87 monostratalism, 54, 55 monosyllabic, 93, 114, 119–123, 125, 126, 128, 130, 131, 173, 188 monosyllabicity, 123 monosyllabism, 125 mora, 80, 84, 173, 188, 189 amoraic, 59, 60, 73, 74, 81 bimoraic, 59, 65, 69, 73, 75, 76, 79, 81, 82, 84–86 monomoraic, 59, 65, 76, 82, 85, 86 moraic, 54, 60, 74, 86, 188 moraicity, 82

313 trimoraic, 195 Mordva, see Mordvin Mordvin, 26, 104, 124 morphologized, 178 morphophonemic, 53, 54, 56, 58, 72, 75 morphophonemics, 87 motion, 109, 216 N nasal, 102, 103, 120 denasalization, 106 natural class, 61, 67, 126 negation, 8, 9, 147, 214 negative auxiliary, 8, 9 Network Morphology, 98, 127 neutralization, 55, 92, 94–96, 99, 100, 111, 112, 114, 126, 127, 131 New Caledonian, 20, 44 Nickel, Klaus Peter, 10, 120, 139, 142, 145 Niederdeutsch, 186 Nielsen, Konrad, 103, 105, 139, 140, 142, 154, 156, 160, 178 nominative, 7, 9, 13, 37, 38, 53, 61, 63, 67, 73, 74, 81, 94, 98, 101, 102, 228, 229, 241–252, 254 non-finite, 10–12 non-initial-dropping, 19 non-stative, see stative North Saami, 1–4, 7–11, 14, 23, 24, 26–41, 45, 46, 48–50, 65, 91–132, 137–164, 171–173, 179, 195, 230, 231, 246

314 Finnmark Saami, 93, 106, 107, 120–123, 125, 126, 128, 129, 131, 132 Torne Saami, 106–108, 120 North Saxon, see German Norwegian, 2, 150–152 noun, 7, 11, 14, 37–40, 61, 63, 69, 97, 99, 101, 104, 105, 109, 110, 117–119, 121, 122, 124, 128, 138–142, 145–148, 150, 151, 153–156, 158, 159, 227–229, 232, 241–245, 249 nominal, 4, 7, 10, 11, 53, 61–63, 67, 94, 102, 117–121, 125, 126, 128, 131, 138, 140–142, 145–148, 150–152, 154, 158, 162, 233, 234, 241–244 nominalization, 146, 162, 163 nominalize, 10, 11, 147, 162, 163 nounal, see nounal prenominal, 7, 13 nounal, 97, 98 nucleus, 62, 69, 75, 76, 79–82, 84–86, 181, 183, 187 numeral, 222, 241–245, 249, 254 O object, 235, 240, 244, 251–254 direct object, 208, 212, 214–219, 221–223 objectθ , 252, 253 objective function, 252 oblique (function), 235, 236, 255

INDEX

obliqueθ , 252, 255 oblique case, see case obliqueness hierarchy, see hierarchy obstruent, 55, 57, 60, 65, 67, 69, 70, 72, 73, 75–77, 79, 116 Foot-junctural Obstruent Gemination, see Foot-junctural Obstruent Gemination (FOG) hyperobstruent, 61, 64, 65, 73, 74, 76, 77, 81 occlusive, 61, 76 hyperocclusive, 61, 64 Old Norse, 41 onset, 177 ontological category, 155 opacity, 54, 81, 87, 178 Optimality Theory (OT), 55 optionality, 243 optional, 76, 176 orthography, 92, 103, 105, 209 OT, see Optimality Theory output, 55, 76 overlength, 169, 186, 194, 195, 197 Überlänge, 186, 187, 194, 195 overlong, 65, 178, 186, 188, 191, 192, 194, 195 P P¯anini’s principle, 98 Pama-Nyungan, 19, 22 Papuan, 114 paradigm, 2, 4, 7–11, 53, 61–63, 67, 68, 92–101, 103, 107, 108, 110–117, 119, 120, 122, 123, 125, 128, 130–132

INDEX

paradigm cell, 93–95, 98–100, 110–112, 115, 116, 120, 128, 131, 132 paradigm function, 96–98, 128 Paradigm Function Morphology, see Paradigm Function Morphology word-and-paradigm, 96 Paradigm Function Morphology (PFM), 96, 97, 110, 127, 132 parameter, 57, 58 parisyllabic, 119, 124, 125 partial agreement, see agreement partial ordering, 253 partitive, 61–63, 67, 68, 103, 104, 107, 109, 121, 208, 209, 214, 217, 218, 222, 241, 248–250 passive, 9, 158 passivization, 9 passivize, 214 path, 216 patient, 253 patientlike, 252, 253, 255 peak consonant-peaked, 179 peakedness, 192 vowel-peaked, 179, 192 perception, 191, 192 perceptual, 56, 57, 192 perfectivity, 222 imperfective, see imperfectivity perfective, 223 Persian, 20, 22 personification, 231, 237 perturbation, 54, 63, 64, 66

315 non-perturbed, 54, 62–64, 66, 68, 69 perturbed, 54, 62, 64, 66, 68, 69 phonemic, 54, 55, 168, 183, 188, 194 phonetic nearness, 30, 32, 34, 38, 42 phonological hierarchy, see prosody phonological simplification, 41 phonological weight, 152 phonotactics, 207, 209 pitch, 187, 188, 191–195 falling pitch, 187 pitch contour, 182, 192, 193 pitch movement, 169, 179, 185–187, 194 Pite Saami, 1–3, 14, 103, 105, 107, 121 place name, see toponym placeless, 61 plosive, 60, 61, 76, 77 polymorphous, 58 polysyllabic, 91, 93, 117, 119–123, 125, 126, 128, 131, 173, 190 possession, 14 predicative possessor, 13 non-possessed declension, 54 possessed form, 113, 124 possessive adjective, 124 possessive agreement, see agreement possessive construction, 228, 231, 233, 241, 246–252, 254, 255 possessive declension, 7, 107 possessive paradigm, 7

316 possessive suffix, 7, 11, 105, 245, 246, 250 possessor, 7, 13, 22, 146, 162, 241, 246, 250–252, 254, 255 postposition, 107, 109, 110, 119, 213 PRAAT, 5, 171 pragmatics, 141, 153 Pre-Basque, see Basque Pre-Finnic, see Finnic Pre-Irish, see Irish Pre-Tlingit, see Tlingit Pre-Welsh, see Welsh predicate copular predicate, 10 stage-level, see stage-level preposition, 137, 153, 154 prepositional, 154 primary stress, see stress Principles and Parameters Theory, 240 pro-drop, 12 prolative, 208, 211–214, 223 pronoun, 7, 14, 113, 117, 120, 122, 125, 126, 128–130, 213, 232, 236, 239, 241, 244, 247–249 indefinite, 128 non-personal, 105, 119, 130 non-pronominal, 7, 14, 119, 228, 229 personal pronoun, 119, 125, 130, 228, 229, 232–234, 237, 250 pronominal, 7, 101, 119, 131, 233, 248

INDEX

special pronoun rule, 130 prosody, 62, 64, 66, 167, 168, 194, 195 prosodic, 54, 59, 168–171, 173–178, 180–185, 187, 189–192, 194–197 prosodic change, 168, 194, 196 prosodic constituent edge, 117 prosodic hierarchy, 172–178, 189–191, 196 prosodic word, 119 prosodics, 168, 169 Proto-Baltic, see Baltic Proto-Fennic, see Finnic Proto-Fenno-Saami, see Fenno-Saami Proto-Finnic, see Finnic Proto-Saami, 24, 26–29, 31–41, 103–108, 120–125 Proto-Uralic, see Uralic pulse, 179 subglottal pulse, 179 syllabic pulse, 179 three-pulsed, 180 Q q(uantitative)-iamb, see iamb q(uantitative)-iambic, see iambic q(uantitative)-trochaic, see trochaic q(uantitative)-trochee, see trochee quantifier, 152 quantity, 59, 63, 167–197 consonant quantity, 2, 4, 5, 14, 53, 54, 63, 69–76, 167–197, 209 ternary quantity distinction, see ternary

INDEX

vowel quantity, 53, 63, 66, 79–87, 172–176, 178–180, 183, 185–191, 194, 195 R raising, 246, 247 ratio, 176–178, 183, 190–193, 196, 197 realizational, 96 inferential-realizational, 93, 96, 100, 127 reanalysis, 103, 132, 251 reduction, 59, 73, 74, 81–82, 85 redundancy rule, see rule referential index, 148–150 referral, see rule of referral Relational Grammar, 240 relative clause, 232 resonant, 77 restructuring, 59 result, 138, 143, 144, 153, 158, 159, 163, 216 resultative, see result Rheinischen Schärfung, 186 Romance, 21 root, 96–98, 119, 128, 129, 137–164 root-selecting, 157 rounded, 114 unrounded, 114 rule impoverishment rule, 95 redundancy rule, 95 rule block, 97, 98 rule of exponence, 97–100, 127, 129, 130 rule of referral, 93, 95–100, 110, 112, 114, 127–132

317 Russian, 2, 29, 30, 99, 101, 112, 116, 208, 209, 214, 216, 219–223 Røros Lappish, 171 S Sammallahti, Pekka, 1, 4, 7, 10, 13, 25, 26, 38, 53, 58, 59, 61–63, 65, 81, 103–109, 120, 121, 123, 124, 140, 154, 170, 172, 173, 179, 195, 208, 209, 231 Sarcee, 170, 187, 188, 193, 194 Scandinavian, 24, 31, 41 Schleifton, 186–187, 195 Scottish Gaelic, 168, 170, 188, 189, 193, 194 secondary stress, see stress segment, 116, 168–173, 178, 189–191, 194–196 segmental, 107, 108, 127, 172–174, 189–191 semantic function, 2, 11–14 semantic role, see thematic role set union, 112 shortening, 18 sibilant, 60, 76, 116 hypersibilant, 61 Single Delimiting Constraint, 216 singleton, 54, 60, 64, 66, 69, 72–76 hypersingleton, 60, 63–66, 69, 72–78, 80, 84 Skolt Saami, 1–5, 103, 105, 167–197 Slavic, 2, 19, 22 sonorant, 64, 66, 70–79 desonorization, 105 hypersonorant, 60, 61, 64, 65

318 sound change, 18, 20, 24, 26, 41, 43, 102, 103, 105–107, 123 source, 104, 109 South Saami, 1–3, 9, 12, 14, 103, 105–108, 121, 124 Spanish, 21–23 specificity, 217, 231, 232, 234, 239 non-specific, 232 specific, 232–234, 239 spectrogram, 179 spectrographic, 179 speech tempo, 171 spirant, 61, 76 spirantization, 21, 76 spirantized, 76 split agreement, see agreement spread glottis, 59, 60 stage-level, 219, 221 Standard Theory, 240 state, 138, 140–144, 146, 147, 152, 153, 156, 158, 159, 163, 218, 219, 221 change of state, 142, 144, 146, 156, 216 mental state, 221 state-denoting, 137, 138, 146, 147, 152–154, 156 stative, see stative statistically, 171 stative, 137–141, 144, 145, 152, 154, 156, 163, 218–220 non-stative, 156, 159, 160, 163, 220, 221 stem-final, 104, 124 step-down, 183, 184 stop, 41 aspirated stop, 32

INDEX

plain stop, 32 unaspirated stop, 92 stress, 54, 60, 62, 68, 76, 85, 86, 122, 123, 174, 179–183, 185, 188, 190, 192, 194, 195, 197 evenly stressed, 180 finally stressed, 180 initially stressed, 180 main stress, 179 primary stress, 180, 181 secondary stress, 54, 62, 85, 173, 174, 180, 181, 183 stress-group, 171–178, 183, 195–197 stressless, 173 unstressed, 60, 62, 68, 174, 187 stress-group, see stress structural case, see case structural description, 55 structural type, 171, 174–178, 181–183 structuralist, 95 structure-building, 54 structure-preserving, 54 subcategorization, 102, 235 subglottal, 179 subject, 139, 141–145, 148, 153, 154, 156, 228–231, 233–255 logical subject, 246, 253 non-subject, 249 subjecthood, 228, 240, 241, 245–247, 250 subjunctive, 110, 111 substitution, 19, 21–23, 40, 41 consonant substitution, 20

INDEX

morpheme substitution, 22, 23, 40 phonetic substitution, 28 sound substitution, 22, 28–30, 32, 34, 35, 38–43 system-based substitution, 32, 35, 36, 39, 42 vowel substitution, 17, 18, 21, 23, 25, 28–40 substrate, 41 suffix, 7, 11, 23, 39, 40, 99, 100, 104, 105, 108, 109, 114, 116, 117, 123–125, 127 possessive suffix, see possession suffix-like, 109 suffixal, 18, 94 suffixation, 94, 128 suffixing, 130 supistumavartalot, 197 Surrey Morphology Group, 92 SVO, see word order Swedish, 2 syllable, 5, 18, 22, 25, 26, 28, 32–36, 38, 39, 54, 62, 63, 68, 76, 82, 85, 123, 124, 146, 147, 154, 168, 169, 171–184, 187, 189–194 bisyllabic, see disyllable disyllabic, see disyllable disyllable, see disyllable extrasyllabic, 173 intersyllabic, 192 monosyllabic, see monosyllabic polysyllabic, see polysyllabic syllabic, 172–183, 189–191, 209

319 syllable boundary, 177 trisyllabic, see trisyllabic syncopation, 18 syncretism, 11, 12, 91–132 bidirectional, 100 geometric model, 112–117 productive , 91–132 stipulated, 93, 99, 100 syncretic resolution, 101, 102, 113 unstipulated, 99 synthetic, 192 systematic listening tests, 168 T taxonomy, 216, 223 telicity, 210 atelic, 210, 211, 214, 218 telic, 142, 144, 211 templatic conditions, 65, 69, 70, 80 temporal, 9, 11, 207, 209–214, 217, 222, 223 temporal adverbial, see adverbial Ter Saami, 2, 3, 25, 103, 105 ternary, 4, 5, 14, 54, 104, 167–197, 209 thematic hierarchy, see hierarchy thematic role, 149, 150, 241, 245, 246, 251–253 theme incremental theme, 215, 216 theta role see thematic role, 149 three-way, see ternary time, 210–213, 215, 218, 223 time adverbial, see adverbial Tlingit, 19, 20, 22

320 Pre-Tlingit, 20 tonal, 191 tone, 41, 188, 195, 197 topicality, 232–234 topic, 241, 255 topic position, 255 topicalized, 246 toponym, 41 Torne Saami, see North Saami transformational, 55, 94 transitory, 167, 185, 193, 197 translative, 39, 40, 142 non-translative, 40 triplet, see minimal triplet trisyllabic, 178, 181 trochaic, 63, 65–67, 70, 80, 85 q(uantitative)-trochaic, 63, 64, 69, 80, 82, 84, 85 trochee, 62 q(uantitative)-trochee, 63, 80, 82, 84 U Ume Saami, 1–3, 14, 103, 105, 107, 121 unaccusative, 160, 162 unbounded, see bounded underlying, 55, 56, 94, 95, 105, 117 underspecification, 94, 95, 99, 100, 111, 112, 114, 126, 131 underspecified, 94, 95, 111 unification unify, 55, 58, 235 Unique Path Constraint, 216 Uniqueness principle, 235, 238–240 universal, 110 unspecified, 94, 111 unstressed, see stress

INDEX

Uralic, 17, 18, 25, 26, 30, 33, 36, 37 Proto-Uralic, 23, 25–27, 30, 33, 34, 36–40, 102 Uto-Aztecan, 170, 187, 188 V value, see feature verb, 7–11, 13, 14, 39, 40, 137–164, 210, 214, 216, 218–221, 223, 227–255 auxiliary verb, 8, 9 causative, 137, 138, 140, 145, 156–161, 163 causative verb, 9 deadjectival verb, see deadjectival denominal verb, see denominal inchoative verb, see inchoative light verb, 143 matrix verb, 10 non-finite verb, see non-finite non-stative, 156, 159, 160, 163 perception verb, 221 root, 138–142, 144–148, 151–158, 160–163 stative verb, see stative verb classes, 215 verbal, 7–9, 11, 12, 14, 53, 138, 141, 143, 144, 146, 147, 150, 154, 159, 160, 163, 219, 222 verbal agreement, see agreement, 227–255 verbalization, 145, 146, 150–152, 164 verbalize, 154, 162 verbalizer, 144, 145, 148, 151, 152, 154, 156, 157, 163

321

INDEX

verbalizing, 140, 147, 150, 151 vocalism, 124 vocative, 7 voice, 157, 158, 161, 163 active, see active passive, see passive voiceless, see voicing voicing, 55, 76 devoicing, 105 final obstruent devoicing, 55, 57 voiced, 55, 60 voiceless, 77, 92 Volgaic, 104 Votic, 124 vowel centre, 62, 63, 65, 66, 69, 79, 80, 82, 84–87 vowel harmony, 18, 26, 36, 37 harmonic, 39 vowel peak, see peak vowel quantity, see quantity

W Warumungu, 209 well-formedness condition, 96 Welsh Proto-Welsh, 19 Wichita, 170, 187, 188, 193, 194 Wiklund, Karl Bernhard, 105, 178 word order free word order, 228, 241, 246 SVO, 228, 241 word-and-paradigm, 96 word-final, 102, 103, 105, 106, 123, 176, 178, 187, 188 Y Yupik, see Central Siberian Yupik Z zero marking, 101 Zoquean, 170, 187

CURRENT ISSUES IN LINGUISTIC THEORY

E. F. K. Koerner, Editor

Zentrum für Allgemeine Sprachwissenschaft, Typologie und Universalienforschung, Berlin [email protected] Current Issues in Linguistic Theory (CILT) is a theory-oriented series which welcomes contributions from scholars who have significant proposals to make towards the advancement of our understanding of language, its structure, functioning and development. CILT has been established in order to provide a forum for the presentation and discussion of linguistic opinions of scholars who do not necessarily accept the prevailing mode of thought in linguistic science. It offers an outlet for meaningful contributions to the current linguistic debate, and furnishes the diversity of opinion which a healthy discipline must have. A complete list of titles in this series can be found on the publishers’ website, www.benjamins.com 291 Baauw, Sergio, Frank Drijkoningen and Manuela Pinto (eds.): Romance Languages and Linguistic Theory 2005. Selected papers from ‘Going Romance’, Utrecht, 8-10 December 2005. viii, 332 pp. + index. Expected November 2007 290 Mughazy, Mustafa A. (ed.): Perspectives on arabic Linguistics XX. Papers from the twentieth annual symposium on Arabic linguistics, Kalamazoo, March 2006. xii, 248 pp. Expected December 2007 289 Benmamoun, Elabbas (ed.): Perspectives on Arabic Linguistics XIX. Papers from the nineteenth annual symposium on Arabic Linguistics, Urbana, April 2005. ca. 250 pp. Expected December 2007 288 Toivonen, Ida and Diane Nelson (eds.): Saami Linguistics. 2007. viii, 321 pp. 287 Camacho, José, Nydia Flores-Ferrán, Liliana Sánchez, Viviane Déprez and María José Cabrera (eds.): Romance Linguistics 2006. Selected papers from the 36th Linguistic Symposium on Romance Languages (LSRL), New Brunswick, March-April 2006. viii, 332 pp. Expected November 2007 286 Weijer, Jeroen van de and Erik Jan van der Torre (eds.): Voicing in Dutch. (De)voicing – phonology, phonetics, and psycholinguistics. viii, 186 pp. Expected October 2007 285 Sackmann, Robin (ed.): Explorations in Integrational Linguistics. Four essays on German, French, and Guaraní. ix, 217 pp. Expected November 2007 284 Salmons, Joseph C. and Shannon Dubenion-Smith (eds.): Historical Linguistics 2005. Selected papers from the 17th International Conference on Historical Linguistics, Madison, Wisconsin, 31 July - 5 August 2005. 2007. viii, 413 pp. 283 Lenker, Ursula and Anneli Meurman-Solin (eds.): Connectives in the History of English. 2007. viii, 318 pp. 282 Prieto, Pilar, Joan Mascaró and Maria-Josep Solé (eds.): Segmental and prosodic issues in Romance phonology. 2007. xvi, 262 pp. 281 Vermeerbergen, Myriam, Lorraine Leeson and Onno Crasborn (eds.): Simultaneity in Signed Languages. Form and function. 2007. viii, 360 pp. (incl. CD-Rom). 280 Hewson, John and Vit Bubenik: From Case to Adposition. The development of configurational syntax in Indo-European languages. 2006. xxx, 420 pp. 279 Nedergaard Thomsen, Ole (ed.): Competing Models of Linguistic Change. Evolution and beyond. 2006. vi, 344 pp. 278 Doetjes, Jenny and Paz González (eds.): Romance Languages and Linguistic Theory 2004. Selected papers from ‘Going Romance’, Leiden, 9–11 December 2004. 2006. viii, 320 pp. 277 Helasvuo, Marja-Liisa and Lyle Campbell (eds.): Grammar from the Human Perspective. Case, space and person in Finnish. 2006. x, 280 pp. 276 Montreuil, Jean-Pierre Y. (ed.): New Perspectives on Romance Linguistics. Vol. II: Phonetics, Phonology and Dialectology. Selected papers from the 35th Linguistic Symposium on Romance Languages (LSRL), Austin, Texas, February 2005. 2006. x, 213 pp. 275 Nishida, Chiyo and Jean-Pierre Y. Montreuil (eds.): New Perspectives on Romance Linguistics. Vol. I: Morphology, Syntax, Semantics, and Pragmatics. Selected papers from the 35th Linguistic Symposium on Romance Languages (LSRL), Austin, Texas, February 2005. 2006. xiv, 288 pp. 274 Gess, Randall S. and Deborah Arteaga (eds.): Historical Romance Linguistics. Retrospective and perspectives. 2006. viii, 393 pp. 273 Filppula, Markku, Juhani Klemola, Marjatta Palander and Esa Penttilä (eds.): Dialects Across Borders. Selected papers from the 11th International Conference on Methods in Dialectology (Methods XI), Joensuu, August 2002. 2005. xii, 291 pp.

272 Gess, Randall S. and Edward J. Rubin (eds.): Theoretical and Experimental Approaches to Romance Linguistics. Selected papers from the 34th Linguistic Symposium on Romance Languages (LSRL), Salt Lake City, March 2004. 2005. viii, 367 pp. 271 Branner, David Prager (ed.): The Chinese Rime Tables. Linguistic philosophy and historicalcomparative phonology. 2006. viii, 358 pp. 270 Geerts, Twan, Ivo van Ginneken and Haike Jacobs (eds.): Romance Languages and Linguistic Theory 2003. Selected papers from ‘Going Romance’ 2003, Nijmegen, 20–22 November. 2005. viii, 369 pp. 269 Hargus, Sharon and Keren Rice (eds.): Athabaskan Prosody. 2005. xii, 432 pp. 268 Cravens, Thomas D. (ed.): Variation and Reconstruction. 2006. viii, 223 pp. 267 Alhawary, Mohammad T. and Elabbas Benmamoun (eds.): Perspectives on Arabic Linguistics XVII–XVIII. Papers from the seventeenth and eighteenth annual symposia on Arabic linguistics. Volume XVII–XVIII: Alexandria, 2003 and Norman, Oklahoma 2004. 2005. xvi, 315 pp. 266 Boudelaa, Sami (ed.): Perspectives on Arabic Linguistics XVI. Papers from the sixteenth annual symposium on Arabic linguistics, Cambridge, March 2002. 2006. xii, 181 pp. 265 Cornips, Leonie and Karen P. Corrigan (eds.): Syntax and Variation. Reconciling the Biological and the Social. 2005. vi, 312 pp. 264 Dressler, Wolfgang U., Dieter Kastovsky, Oskar E. Pfeiffer and Franz Rainer (eds.): Morphology and its demarcations. Selected papers from the 11th Morphology meeting, Vienna, February 2004. With the assistance of Francesco Gardani and Markus A. Pöchtrager. 2005. xiv, 320 pp. 263 Branco, António, Tony McEnery and Ruslan Mitkov (eds.): Anaphora Processing. Linguistic, cognitive and computational modelling. 2005. x, 449 pp. 262 Vajda, Edward J. (ed.): Languages and Prehistory of Central Siberia. 2004. x, 275 pp. 261 Kay, Christian J. and Jeremy J. Smith (eds.): Categorization in the History of English. 2004. viii, 268 pp. 260 Nicolov, Nicolas, Kalina Bontcheva, Galia Angelova and Ruslan Mitkov (eds.): Recent Advances in Natural Language Processing III. Selected papers from RANLP 2003. 2004. xii, 402 pp. 259 Carr, Philip, Jacques Durand and Colin J. Ewen (eds.): Headhood, Elements, Specification and Contrastivity. Phonological papers in honour of John Anderson. 2005. xxviii, 405 pp. 258 Auger, Julie, J. Clancy Clements and Barbara Vance (eds.): Contemporary Approaches to Romance Linguistics. Selected Papers from the 33rd Linguistic Symposium on Romance Languages (LSRL), Bloomington, Indiana, April 2003. With the assistance of Rachel T. Anderson. 2004. viii, 404 pp. 257 Fortescue, Michael, Eva Skafte Jensen, Jens Erik Mogensen and Lene Schøsler (eds.): Historical Linguistics 2003. Selected papers from the 16th International Conference on Historical Linguistics, Copenhagen, 11–15 August 2003. 2005. x, 312 pp. 256 Bok-Bennema, Reineke, Bart Hollebrandse, Brigitte Kampers-Manhe and Petra Sleeman (eds.): Romance Languages and Linguistic Theory 2002. Selected papers from ‘Going Romance’, Groningen, 28–30 November 2002. 2004. viii, 273 pp. 255 Meulen, Alice ter and Werner Abraham (eds.): The Composition of Meaning. From lexeme to discourse. 2004. vi, 232 pp. 254 Baldi, Philip and Pietro U. Dini (eds.): Studies in Baltic and Indo-European Linguistics. In honor of William R. Schmalstieg. 2004. xlvi, 302 pp. 253 Caffarel, Alice, J.R. Martin and Christian M.I.M. Matthiessen (eds.): Language Typology. A functional perspective. 2004. xiv, 702 pp. 252 Kay, Christian J., Carole Hough and Irené Wotherspoon (eds.): New Perspectives on English Historical Linguistics. Selected papers from 12 ICEHL, Glasgow, 21–26 August 2002. Volume II: Lexis and Transmission. 2004. xii, 273 pp. 251 Kay, Christian J., Simon Horobin and Jeremy J. Smith (eds.): New Perspectives on English Historical Linguistics. Selected papers from 12 ICEHL, Glasgow, 21–26 August 2002. Volume I: Syntax and Morphology. 2004. x, 264 pp. 250 Jensen, John T.: Principles of Generative Phonology. An introduction. 2004. xii, 324 pp. 249 Bowern, Claire and Harold Koch (eds.): Australian Languages. Classification and the comparative method. 2004. xii, 377 pp. (incl. CD-Rom). 248 Weigand, Edda (ed.): Emotion in Dialogic Interaction. Advances in the complex. 2004. xii, 284 pp. 247 Parkinson, Dilworth B. and Samira Farwaneh (eds.): Perspectives on Arabic Linguistics XV. Papers from the Fifteenth Annual Symposium on Arabic Linguistics, Salt Lake City 2001. 2003. x, 214 pp. 246 Holisky, Dee Ann and Kevin Tuite (eds.): Current Trends in Caucasian, East European and Inner Asian Linguistics. Papers in honor of Howard I. Aronson. 2003. xxviii, 426 pp. 245 Quer, Josep, Jan Schroten, Mauro Scorretti, Petra Sleeman and Els Verheugd (eds.): Romance Languages and Linguistic Theory 2001. Selected papers from 'Going Romance', Amsterdam, 6–8 December 2001. 2003. viii, 355 pp.

244 Pérez-Leroux, Ana Teresa and Yves Roberge (eds.): Romance Linguistics. Theory and Acquisition. Selected papers from the 32nd Linguistic Symposium on Romance Languages (LSRL), Toronto, April 2002. 2003. viii, 388 pp. 243 Cuyckens, Hubert, Thomas Berg, René Dirven and Klaus-Uwe Panther (eds.): Motivation in Language. Studies in honor of Günter Radden. 2003. xxvi, 403 pp. 242 Seuren, Pieter A.M. and Gerard Kempen (eds.): Verb Constructions in German and Dutch. 2003. vi, 316 pp. 241 Lecarme, Jacqueline (ed.): Research in Afroasiatic Grammar II. Selected papers from the Fifth Conference on Afroasiatic Languages, Paris, 2000. 2003. viii, 550 pp. 240 Janse, Mark and Sijmen Tol (eds.): Language Death and Language Maintenance. Theoretical, practical and descriptive approaches. With the assistance of Vincent Hendriks. 2003. xviii, 244 pp. 239 Andersen, Henning (ed.): Language Contacts in Prehistory. Studies in Stratigraphy. Papers from the Workshop on Linguistic Stratigraphy and Prehistory at the Fifteenth International Conference on Historical Linguistics, Melbourne, 17 August 2001. 2003. viii, 292 pp. 238 Núñez-Cedeño, Rafael, Luis López and Richard Cameron (eds.): A Romance Perspective on Language Knowledge and Use. Selected papers from the 31st Linguistic Symposium on Romance Languages (LSRL), Chicago, 19–22 April 2001. 2003. xvi, 386 pp. 237 Blake, Barry J. and Kate Burridge (eds.): Historical Linguistics 2001. Selected papers from the 15th International Conference on Historical Linguistics, Melbourne, 13–17 August 2001. Editorial Assistant: Jo Taylor. 2003. x, 444 pp. 236 Simon-Vandenbergen, Anne-Marie, Miriam Taverniers and Louise J. Ravelli (eds.): Grammatical Metaphor. Views from systemic functional linguistics. 2003. vi, 453 pp. 235 Linn, Andrew R. and Nicola McLelland (eds.): Standardization. Studies from the Germanic languages. 2002. xii, 258 pp. 234 Weijer, Jeroen van de, Vincent J. van Heuven and Harry van der Hulst (eds.): The Phonological Spectrum. Volume II: Suprasegmental structure. 2003. x, 264 pp. 233 Weijer, Jeroen van de, Vincent J. van Heuven and Harry van der Hulst (eds.): The Phonological Spectrum. Volume I: Segmental structure. 2003. x, 308 pp. 232 Beyssade, Claire, Reineke Bok-Bennema, Frank Drijkoningen and Paola Monachesi (eds.): Romance Languages and Linguistic Theory 2000. Selected papers from ‘Going Romance’ 2000, Utrecht, 30 November–2 December. 2002. viii, 354 pp. 231 Cravens, Thomas D.: Comparative Historical Dialectology. Italo-Romance clues to Ibero-Romance sound change. 2002. xii, 163 pp. 230 Parkinson, Dilworth B. and Elabbas Benmamoun (eds.): Perspectives on Arabic Linguistics. Papers from the Annual Symposium on Arabic Linguistics. Volume XIII-XIV: Stanford, 1999 and Berkeley, California 2000. 2002. xiv, 250 pp. 229 Nevin, Bruce E. and Stephen B. Johnson (eds.): The Legacy of Zellig Harris. Language and information into the 21st century. Volume 2: Mathematics and computability of language. 2002. xx, 312 pp. 228 Nevin, Bruce E. (ed.): The Legacy of Zellig Harris. Language and information into the 21st century. Volume 1: Philosophy of science, syntax and semantics. 2002. xxxvi, 323 pp. 227 Fava, Elisabetta (ed.): Clinical Linguistics. Theory and applications in speech pathology and therapy. 2002. xxiv, 353 pp. 226 Levin, Saul: Semitic and Indo-European. Volume II: Comparative morphology, syntax and phonetics. 2002. xviii, 592 pp. 225 Shahin, Kimary N.: Postvelar Harmony. 2003. viii, 344 pp. 224 Fanego, Teresa, Belén Méndez-Naya and Elena Seoane (eds.): Sounds, Words, Texts and Change. Selected papers from 11 ICEHL, Santiago de Compostela, 7–11 September 2000. Volume 2. 2002. x, 310 pp. 223 Fanego, Teresa, Javier Pérez-Guerra and María José López-Couso (eds.): English Historical Syntax and Morphology. Selected papers from 11 ICEHL, Santiago de Compostela, 7–11 September 2000. Volume 1. 2002. x, 306 pp. 222 Herschensohn, Julia, Enrique Mallén and Karen Zagona (eds.): Features and Interfaces in Romance. Essays in honor of Heles Contreras. 2001. xiv, 302 pp. 221 D’hulst, Yves, Johan Rooryck and Jan Schroten (eds.): Romance Languages and Linguistic Theory 1999. Selected papers from ‘Going Romance’ 1999, Leiden, 9–11 December 1999. 2001. viii, 406 pp. 220 Satterfield, Teresa, Christina M. Tortora and Diana Cresti (eds.): Current Issues in Romance Languages. Selected papers from the 29th Linguistic Symposium on Romance Languages (LSRL), Ann Arbor, 8–11 April 1999. 2002. viii, 412 pp.

219 Andersen, Henning (ed.): Actualization. Linguistic Change in Progress. Papers from a workshop held at the 14th International Conference on Historical Linguistics, Vancouver, B.C., 14 August 1999. 2001. vii, 250 pp. 218 Bendjaballah, Sabrina, Wolfgang U. Dressler, Oskar E. Pfeiffer and Maria D. Voeikova (eds.): Morphology 2000. Selected papers from the 9th Morphology Meeting, Vienna, 24–28 February 2000. 2002. viii, 317 pp. 217 Wiltshire, Caroline R. and Joaquim Camps (eds.): Romance Phonology and Variation. Selected papers from the 30th Linguistic Symposium on Romance Languages, Gainesville, Florida, February 2000. 2002. xii, 238 pp. 216 Camps, Joaquim and Caroline R. Wiltshire (eds.): Romance Syntax, Semantics and L2 Acquisition. Selected papers from the 30th Linguistic Symposium on Romance Languages, Gainesville, Florida, February 2000. 2001. xii, 246 pp. 215 Brinton, Laurel J. (ed.): Historical Linguistics 1999. Selected papers from the 14th International Conference on Historical Linguistics, Vancouver, 9–13 August 1999. 2001. xii, 398 pp. 214 Weigand, Edda and Marcelo Dascal (eds.): Negotiation and Power in Dialogic Interaction. 2001. viii, 303 pp. 213 Sornicola, Rosanna, Erich Poppe and Ariel Shisha-Halevy (eds.): Stability, Variation and Change of Word-Order Patterns over Time. With the assistance of Paola Como. 2000. xxxii, 323 pp. 212 Repetti, Lori (ed.): Phonological Theory and the Dialects of Italy. 2000. x, 301 pp. 211 Elšík, Viktor and Yaron Matras (eds.): Grammatical Relations in Romani. The Noun Phrase. with a Foreword by Frans Plank (Universität Konstanz). 2000. x, 244 pp. 210 Dworkin, Steven N. and Dieter Wanner (eds.): New Approaches to Old Problems. Issues in Romance historical linguistics. 2000. xiv, 235 pp. 209 King, Ruth: The Lexical Basis of Grammatical Borrowing. A Prince Edward Island French case study. 2000. xvi, 241 pp. 208 Robinson, Orrin W.: Whose German? The ach/ich alternation and related phenomena in ‘standard’ and ‘colloquial’. 2001. xii, 178 pp. 207 Sanz, Montserrat: Events and Predication. A new approach to syntactic processing in English and Spanish. 2000. xiv, 219 pp. 206 Fawcett, Robin P.: A Theory of Syntax for Systemic Functional Linguistics. 2000. xxiv, 360 pp. 205 Dirven, René, Roslyn M. Frank and Cornelia Ilie (eds.): Language and Ideology. Volume 2: descriptive cognitive approaches. 2001. vi, 264 pp. 204 Dirven, René, Bruce Hawkins and Esra Sandikcioglu (eds.): Language and Ideology. Volume 1: theoretical cognitive approaches. 2001. vi, 301 pp. 203 Norrick, Neal R.: Conversational Narrative. Storytelling in everyday talk. 2000. xiv, 233 pp. 202 Lecarme, Jacqueline, Jean Lowenstamm and Ur Shlonsky (eds.): Research in Afroasiatic Grammar. Papers from the Third conference on Afroasiatic Languages, Sophia Antipolis, 1996. 2000. vi, 386 pp. 201 Dressler, Wolfgang U., Oskar E. Pfeiffer, Markus A. Pöchtrager and John R. Rennison (eds.): Morphological Analysis in Comparison. 2000. x, 261 pp. 200 Anttila, Raimo: Greek and Indo-European Etymology in Action. Proto-Indo-European *aǵ-. 2000. xii, 314 pp. 199 Pütz, Martin and Marjolijn H. Verspoor (eds.): Explorations in Linguistic Relativity. 2000. xvi, 369 pp. 198 Niemeier, Susanne and René Dirven (eds.): Evidence for Linguistic Relativity. 2000. xxii, 240 pp. 197 Coopmans, Peter, Martin Everaert and Jane Grimshaw (eds.): Lexical Specification and Insertion. 2000. xviii, 476 pp. 196 Hannahs, S.J. and Mike Davenport (eds.): Issues in Phonological Structure. Papers from an International Workshop. 1999. xii, 268 pp. 195 Herring, Susan C., Pieter van Reenen and Lene Schøsler (eds.): Textual Parameters in Older Languages. 2001. x, 448 pp. 194 Coleman, Julie and Christian J. Kay (eds.): Lexicology, Semantics and Lexicography. Selected papers from the Fourth G. L. Brook Symposium, Manchester, August 1998. 2000. xiv, 257 pp. 193 Klausenburger, Jurgen: Grammaticalization. Studies in Latin and Romance morphosyntax. 2000. xiv, 184 pp. 192 Alexandrova, Galina M. and Olga Arnaudova (eds.): The Minimalist Parameter. Selected papers from the Open Linguistics Forum, Ottawa, 21–23 March 1997. 2001. x, 360 pp. 191 Sihler, Andrew L.: Language History. An introduction. 2000. xvi, 298 pp.

190 Benmamoun, Elabbas (ed.): Perspectives on Arabic Linguistics. Papers from the Annual Symposium on Arabic Linguistics. Volume XII: Urbana-Champaign, Illinois, 1998. 1999. viii, 204 pp. 189 Nicolov, Nicolas and Ruslan Mitkov (eds.): Recent Advances in Natural Language Processing II. Selected papers from RANLP ’97. 2000. xi, 422 pp. 188 Simmons, Richard VanNess: Chinese Dialect Classification. A comparative approach to Harngjou, Old Jintarn, and Common Northern Wu. 1999. xviii, 317 pp. 187 Franco, Jon A., Alazne Landa and Juan Martín (eds.): Grammatical Analyses in Basque and Romance Linguistics. Papers in honor of Mario Saltarelli. 1999. viii, 306 pp. 186 Mišeska Tomić, Olga and Milorad Radovanović (eds.): History and Perspectives of Language Study. Papers in honor of Ranko Bugarski. 2000. xxii, 314 pp. 185 Authier, Jean-Marc, Barbara E. Bullock and Lisa A. Reed (eds.): Formal Perspectives on Romance Linguistics. Selected papers from the 28th Linguistic Symposium on Romance Languages (LSRL XXVIII), University Park, 16–19 April 1998. 1999. xii, 334 pp. 184 Sagart, Laurent: The Roots of Old Chinese. 1999. xii, 272 pp. 183 Contini-Morava, Ellen and Yishai Tobin (eds.): Between Grammar and Lexicon. 2000. xxxii, 365 pp. 182 Kenesei, István (ed.): Crossing Boundaries. Advances in the theory of Central and Eastern European languages. 1999. viii, 302 pp. 181 Mohammad, Mohammad A.: Word Order, Agreement and Pronominalization in Standard and Palestinian Arabic. 2000. xvi, 197 pp. 180 Mereu, Lunella (ed.): Boundaries of Morphology and Syntax. 1999. viii, 314 pp. 179 Rini, Joel: Exploring the Role of Morphology in the Evolution of Spanish. 1999. xvi, 187 pp. 178 Foolen, Ad and Frederike van der Leek (eds.): Constructions in Cognitive Linguistics. Selected papers from the Fifth International Cognitive Linguistics Conference, Amsterdam, 1997. 2000. xvi, 338 pp. 177 Cuyckens, Hubert and Britta E. Zawada (eds.): Polysemy in Cognitive Linguistics. Selected papers from the International Cognitive Linguistics Conference, Amsterdam, 1997. 2001. xxviii, 296 pp. 176 Van Hoek, Karen, Andrej A. Kibrik and Leo Noordman (eds.): Discourse Studies in Cognitive Linguistics. Selected papers from the 5th International Cognitive Linguistics Conference, Amsterdam, July 1997. 1999. vi, 187 pp. 175 Gibbs, Jr., Raymond W. and Gerard J. Steen (eds.): Metaphor in Cognitive Linguistics. Selected papers from the 5th International Cognitive Linguistics Conference, Amsterdam, 1997. 1999. viii, 226 pp. 174 Hall, T. Alan and Ursula Kleinhenz (eds.): Studies on the Phonological Word. 1999. viii, 298 pp. 173 Treviño, Esthela and José Lema (eds.): Semantic Issues in Romance Syntax. 1999. viii, 309 pp. 172 Dimitrova-Vulchanova, Mila and Lars Hellan (eds.): Topics in South Slavic Syntax and Semantics. 1999. xxviii, 263 pp. 171 Weigand, Edda (ed.): Contrastive Lexical Semantics. 1998. x, 270 pp. 170 Lamb, Sydney M.: Pathways of the Brain. The neurocognitive basis of language. 1999. xii, 418 pp. 169 Ghadessy, Mohsen (ed.): Text and Context in Functional Linguistics. 1999. xviii, 340 pp. 168 Ratcliffe, Robert R.: The “Broken” Plural Problem in Arabic and Comparative Semitic. Allomorphy and analogy in non-concatenative morphology. 1998. xii, 261 pp. 167 Benmamoun, Elabbas, Mushira Eid and Niloofar Haeri (eds.): Perspectives on Arabic Linguistics. Papers from the Annual Symposium on Arabic Linguistics. Volume XI: Atlanta, Georgia, 1997. 1998. viii, 231 pp. 166 Lemmens, Maarten: Lexical Perspectives on Transitivity and Ergativity. Causative constructions in English. 1998. xii, 268 pp. 165 Bubenik, Vit: A Historical Syntax of Late Middle Indo-Aryan (Apabhraṃśa). 1998. xxiv, 265 pp. 164 Schmid, Monika S., Jennifer R. Austin and Dieter Stein (eds.): Historical Linguistics 1997. Selected papers from the 13th International Conference on Historical Linguistics, Düsseldorf, 10–17 August 1997. 1998. x, 409 pp. 163 Lockwood, David G., Peter H. Fries and James E. Copeland (eds.): Functional Approaches to Language, Culture and Cognition. Papers in honor of Sydney M. Lamb. 2000. xxxiv, 656 pp. 162 Hogg, Richard M. and Linda van Bergen (eds.): Historical Linguistics 1995. Volume 2: Germanic linguistics.. Selected papers from the 12th International Conference on Historical Linguistics, Manchester, August 1995. 1998. x, 365 pp. 161 Smith, John Charles and Delia Bentley (eds.): Historical Linguistics 1995. Volume 1: General issues and non-Germanic Languages.. Selected papers from the 12th International Conference on Historical Linguistics, Manchester, August 1995. 2000. xii, 438 pp.

160 Schwegler, Armin, Bernard Tranel and Myriam Uribe-Etxebarria (eds.): Romance Linguistics: Theoretical Perspectives. Selected papers from the 27th Linguistic Symposium on Romance Languages (LSRL XXVII), Irvine, 20–22 February, 1997. 1998. vi, 349 pp. + index. 159 Joseph, Brian D., Geoffrey C. Horrocks and Irene Philippaki-Warburton (eds.): Themes in Greek Linguistics II. 1998. x, 335 pp. 158 Sánchez-Macarro, Antonia and Ronald Carter (eds.): Linguistic Choice across Genres. Variation in spoken and written English. 1998. viii, 338 pp. 157 Lema, José and Esthela Treviño (eds.): Theoretical Analyses on Romance Languages. Selected papers from the 26th Linguistic Symposium on Romance Languages (LSRL XXVI), Mexico City, 28–30 March, 1996. 1998. viii, 380 pp. 156 Matras, Yaron, Peter Bakker and Hristo Kyuchukov (eds.): The Typology and Dialectology of Romani. 1997. xxxii, 223 pp. 155 Forget, Danielle, Paul Hirschbühler, France Martineau and María Luisa Rivero (eds.): Negation and Polarity. Syntax and semantics. Selected papers from the colloquium Negation: Syntax and Semantics. Ottawa, 11–13 May 1995. 1997. viii, 367 pp. 154 Simon-Vandenbergen, Anne-Marie, Kristin Davidse and Dirk Noël (eds.): Reconnecting Language. Morphology and Syntax in Functional Perspectives. 1997. xiii, 339 pp. 153 Eid, Mushira and Robert R. Ratcliffe (eds.): Perspectives on Arabic Linguistics. Papers from the Annual Symposium on Arabic Linguistics. Volume X: Salt Lake City, 1996. 1997. vii, 296 pp. 152 Hiraga, Masako K., Christopher Sinha and Sherman Wilcox (eds.): Cultural, Psychological and Typological Issues in Cognitive Linguistics. Selected papers of the bi-annual ICLA meeting in Albuquerque, July 1995. 1999. viii, 338 pp. 151 Liebert, Wolf-Andreas, Gisela Redeker and Linda R. Waugh (eds.): Discourse and Perspective in Cognitive Linguistics. 1997. xiv, 270 pp. 150 Verspoor, Marjolijn H., Kee Dong Lee and Eve Sweetser (eds.): Lexical and Syntactical Constructions and the Construction of Meaning. Proceedings of the Bi-annual ICLA meeting in Albuquerque, July 1995. 1997. xii, 454 pp. 149 Hall, T. Alan: The Phonology of Coronals. 1997. x, 176 pp. 148 Wolf, George and Nigel Love (eds.): Linguistics Inside Out. Roy Harris and his critics. 1997. xxviii, 344 pp. 147 Hewson, John: The Cognitive System of the French Verb. 1997. xii, 187 pp. 146 Hinskens, Frans, Roeland van Hout and W. Leo Wetzels (eds.): Variation, Change, and Phonological Theory. 1997. x, 314 pp. 145 Hewson, John and Vit Bubenik: Tense and Aspect in Indo-European Languages. Theory, typology, diachrony. 1997. xii, 403 pp. 144 Singh, Rajendra (ed.): Trubetzkoy's Orphan. Proceedings of the Montréal Roundtable on “Morphonology: contemporary responses” (Montréal, October 1994). In collaboration with Richard Desrochers. 1996. xiv, 363 pp. 143 Athanasiadou, Angeliki and René Dirven (eds.): On Conditionals Again. 1997. viii, 418 pp. 142 Salmons, Joseph C. and Brian D. Joseph (eds.): Nostratic. Sifting the Evidence. 1998. vi, 293 pp. 141 Eid, Mushira and Dilworth B. Parkinson (eds.): Perspectives on Arabic Linguistics. Papers from the Annual Symposium on Arabic Linguistics. Volume IX: Washington D.C., 1995. 1996. xiii, 249 pp. 140 Black, James R. and Virginia Motapanyane (eds.): Clitics, Pronouns and Movement. 1997. 375 pp. 139 Black, James R. and Virginia Motapanyane (eds.): Microparametric Syntax and Dialect Variation. 1996. xviii, 269 pp. 138 Sackmann, Robin and Monika Budde (eds.): Theoretical Linguistics and Grammatical Description. Papers in honour of Hans-Heinrich Lieb. 1996. x, 375 pp. 137 Lippi-Green, Rosina L. and Joseph C. Salmons (eds.): Germanic Linguistics. Syntactic and diachronic. 1996. viii, 192 pp. 136 Mitkov, Ruslan and Nicolas Nicolov (eds.): Recent Advances in Natural Language Processing. Selected Papers from RANLP ’95. 1997. xii, 472 pp. 135 Britton, Derek (ed.): English Historical Linguistics 1994. Papers from the 8th International Conference on English Historical Linguistics (8 ICEHL, Edinburgh, 19–23 September 1994). 1996. viii, 403 pp. 134 Eid, Mushira (ed.): Perspectives on Arabic Linguistics. Papers from the Annual Symposium on Arabic Linguistics. Volume VIII: Amherst, Massachusetts 1994. 1996. vii, 261 pp. 133 Zagona, Karen (ed.): Grammatical Theory and Romance Languages. Selected papers from the 25th Linguistic Symposium on Romance Languages (LSRL XXV) Seattle, 2–4 March 1995. 1996. vi, 330 pp. 132 Herschensohn, Julia: Case Suspension and Binary Complement Structure in French. 1996. xi, 200 pp.

131 Hualde, José Ignacio, Joseba A. Lakarra and R.L. Trask (eds.): Towards a History of the Basque Language. 1996. 365 pp. 130 Eid, Mushira (ed.): Perspectives on Arabic Linguistics. Papers from the Annual Symposium on Arabic Linguistics. Volume VII: Austin, Texas 1993. 1995. vii, 192 pp. 129 Levin, Saul: Semitic and Indo-European. Volume I: The Principal Etymologies. With observations on Afro-Asiatic. 1995. xxii, 514 pp. 128 Guy, Gregory R., Crawford Feagin, Deborah Schiffrin and John Baugh (eds.): Towards a Social Science of Language. Papers in honor of William Labov. Volume 2: Social interaction and discourse structures. 1997. xviii, 358 pp. 127 Guy, Gregory R., Crawford Feagin, Deborah Schiffrin and John Baugh (eds.): Towards a Social Science of Language. Papers in honor of William Labov. Volume 1: Variation and change in language and society. 1996. xviii, 436 pp. 126 Matras, Yaron (ed.): Romani in Contact. The history, structure and sociology of a language. 1995. xvii, 208 pp. 125 Singh, Rajendra (ed.): Towards a Critical Sociolinguistics. 1996. xiii, 342 pp. 124 Andersen, Henning (ed.): Historical Linguistics 1993. Selected papers from the 11th International Conference on Historical Linguistics, Los Angeles, 16–20 August 1993. 1995. x, 460 pp. 123 Amastae, Jon, Grant Goodall, M. Montalbetti and M. Phinney (eds.): Contemporary Research in Romance Linguistics. Papers from the XXII Linguistic Symposium on Romance Languages, El Paso/Juárez, February 22–24, 1992. 1995. viii, 381 pp. 122 Smith, John Charles and Martin Maiden (eds.): Linguistic Theory and the Romance Languages. 1995. xiii, 240 pp. 121 Hasan, Ruqaiya, Carmel Cloran and David G. Butt (eds.): Functional Descriptions. Theory in practice. 1996. xxxvi, 381 pp. 120 Stonham, John T.: Combinatorial Morphology. 1994. xii, 206 pp. 119 Lippi-Green, Rosina L.: Language Ideology and Language Change in Early Modern German. A sociolinguistic study of the consonantal system of Nuremberg. 1994. xiv, 150 pp. 118 Hasan, Ruqaiya and Peter H. Fries (eds.): On Subject and Theme. A discourse functional perspective. 1995. xii, 414 pp. 117 Philippaki-Warburton, Irene, Katerina Nicolaidis and Maria Sifianou (eds.): Themes in Greek Linguistics. Papers from the First International Conference on Greek Linguistics, Reading, September 1993. 1994. xviii, 534 pp. 116 Miller, D. Gary: Ancient Scripts and Phonological Knowledge. 1994. xvi, 139 pp. 115 Eid, Mushira, Vicente Cantarino and Keith Walters (eds.): Perspectives on Arabic Linguistics. Papers from the Annual Symposium on Arabic Linguistics. VolumeVI: Columbus, Ohio 1992. 1994. viii, 238 pp. 114 Egli, Urs, Peter E. Pause, Christoph Schwarze, Arnim von Stechow and Götz Wienold (eds.): Lexical Knowledge in the Organization of Language. 1995. xiv, 367 pp. 113 Moreno Fernández, Francisco, Miguel Fuster and Juan Jose Calvo (eds.): English Historical Linguistics 1992. Papers from the 7th International Conference on English Historical Linguistics, Valencia, 22–26 September 1992. 1994. viii, 388 pp. 112 Culioli, Antoine: Cognition and Representation in Linguistic Theory. Texts selected, edited and introduced by Michel Liddle. Translated with the assistance of John T. Stonham. 1995. x, 161 pp. 111 Tobin, Yishai: Invariance, Markedness and Distinctive Feature Analysis. A contrastive study of sign systems in English and Hebrew. 1994. xxii, 406 pp. 110 Simone, Raffaele (ed.): Iconicity in Language. 1995. xii, 315 pp. 109 Pagliuca, William (ed.): Perspectives on Grammaticalization. 1994. xx, 306 pp. 108 Lieb, Hans-Heinrich: Linguistic Variables. Towards a unified theory of linguistic variation. 1993. xiv, 261 pp. 107 Marle, Jaap van (ed.): Historical Linguistics 1991. Papers from the 10th International Conference on Historical Linguistics, Amsterdam, August 12–16, 1991. 1993. xviii, 395 pp. 106 Aertsen, Henk and Robert J. Jeffers (eds.): Historical Linguistics 1989. Papers from the 9th International Conference on Historical Linguistics, New Brunswick, 14–18 August 1989. 1993. xviii, 538 pp. 105 Hualde, José Ignacio and Jon Ortiz de Urbina (eds.): Generative Studies in Basque Linguistics. 1993. vi, 334 pp. 104 Kurzová, Helena: From Indo-European to Latin. The evolution of a morphosyntactic type. 1993. xiv, 259 pp.

Saami Linguistics (Amsterdam Studies in the Theory and History of Linguistic Science Series IV: Current Issues in Linguistics Theory)

Linguistics and Philosophy: Essays in Honor of Rulon S. Wells (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Papers in theoretical linguistics (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Germanic linguistics: syntactic and diachronic (Current Issues in Linguistic Theory)

Historical Linguistics 2003: International Conference On Historical Linguistics, Copenhagen, 11-15 August 2003 (Amsterdam Studies in the Theory and History ... IV: Current Issues in Linguistic Theory)

Historical Linguistics 2003: International Conference On Historical Linguistics, Copenhagen, 11-15 August 2003 (Amsterdam Studies in the Theory and History ... IV: Current Issues in Linguistic Theory)

Emotion in Dialogic Interaction: Advances in the Complex (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Proto-Japanese; Issues and Prospects (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Grammar from the Human Perspective: Case, Space and Person in Finnish (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Verb Constructions in German and Dutch (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Subordination and Coordination Strategies in North Asian Languages (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Simultaneity in Signed Languages: Form and Function (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Representation Theory (Current Studies in Linguistics)

Morphological Analysis in Comparison (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

The Laws of Indo-European (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

The Paradox of Grammatical Change: Perspectives from Romance (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory, CILT 293)

Phonetics and Phonology: Interactions and interrelations (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Foundational Issues in Linguistic Theory: Essays in Honor of Jean-Roger Vergnaud (Current Studies in Linguistics)

Foundational Issues in Linguistic Theory: Essays in Honor of Jean-Roger Vergnaud (Current Studies in Linguistics)

Typological Models in Reconstruction (Amsterdam Studies in the Theory & History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Language Typology: A Functional Perspective (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Language Typology: A Functional Perspective (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Clinical Linguistics: Theory and Applications in Speech Pathology and Therapy (Current Issues in Linguistic Theory)

English Adjective Comparison: A Historical Perspective (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory, Volume 299)

English Adjective Comparison: A Historical Perspective (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory, Volume 299)

Constructions in Cognitive Linguistics: Selected Papers from the International Cognitive Linguistics Conference, Amsterdam, 1997 (Current Issues in Linguistic Theory)

Perspectives on Arabic Linguistics XVII-XVIII: Papers from the Seventeenth and Eighteenth Annual Symposia On Arabic Linguistics (Amsterdam Studies in the ... IV: Current Issues in Linguistic Theory)

Connectives in the History of English (Amsterdam Studies in the Theory and History of Linguistic Science)

Perspectives on Arabic Linguistics XIX: Papers from the nineteenth annual symposium on Arabic Linguistics, Urbana, April 2005 (Amsterdam Studies in the ... IV: Current Issues in Linguistic Theory)

Romance Languages and Linguistic Theory 2004: Selected Papers from 'Going Romance', Leiden, 911 December 2004 (Amsterdam Studies in the Theory and History ... IV: Current Issues in Linguistic Theory)

Loan Phonology (Current Issues in Linguistic Theory)

Saami Linguistics (Amsterdam Studies in the Theory and History of Linguistic Science Series IV: Current Issues in Linguistics Theory)

Linguistics and Philosophy: Essays in Honor of Rulon S. Wells (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Papers in theoretical linguistics (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Germanic linguistics: syntactic and diachronic (Current Issues in Linguistic Theory)

Historical Linguistics 2003: International Conference On Historical Linguistics, Copenhagen, 11-15 August 2003 (Amsterdam Studies in the Theory and History ... IV: Current Issues in Linguistic Theory)

Historical Linguistics 2003: International Conference On Historical Linguistics, Copenhagen, 11-15 August 2003 (Amsterdam Studies in the Theory and History ... IV: Current Issues in Linguistic Theory)

Emotion in Dialogic Interaction: Advances in the Complex (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Proto-Japanese; Issues and Prospects (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Grammar from the Human Perspective: Case, Space and Person in Finnish (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Verb Constructions in German and Dutch (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Subordination and Coordination Strategies in North Asian Languages (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Simultaneity in Signed Languages: Form and Function (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Representation Theory (Current Studies in Linguistics)

Morphological Analysis in Comparison (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

The Laws of Indo-European (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

The Paradox of Grammatical Change: Perspectives from Romance (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory, CILT 293)

Phonetics and Phonology: Interactions and interrelations (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Foundational Issues in Linguistic Theory: Essays in Honor of Jean-Roger Vergnaud (Current Studies in Linguistics)

Foundational Issues in Linguistic Theory: Essays in Honor of Jean-Roger Vergnaud (Current Studies in Linguistics)

Typological Models in Reconstruction (Amsterdam Studies in the Theory & History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Language Typology: A Functional Perspective (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Language Typology: A Functional Perspective (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory)

Clinical Linguistics: Theory and Applications in Speech Pathology and Therapy (Current Issues in Linguistic Theory)

English Adjective Comparison: A Historical Perspective (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory, Volume 299)

English Adjective Comparison: A Historical Perspective (Amsterdam Studies in the Theory and History of Linguistic Science, Series IV: Current Issues in Linguistic Theory, Volume 299)

Constructions in Cognitive Linguistics: Selected Papers from the International Cognitive Linguistics Conference, Amsterdam, 1997 (Current Issues in Linguistic Theory)

Perspectives on Arabic Linguistics XVII-XVIII: Papers from the Seventeenth and Eighteenth Annual Symposia On Arabic Linguistics (Amsterdam Studies in the ... IV: Current Issues in Linguistic Theory)

Connectives in the History of English (Amsterdam Studies in the Theory and History of Linguistic Science)

Perspectives on Arabic Linguistics XIX: Papers from the nineteenth annual symposium on Arabic Linguistics, Urbana, April 2005 (Amsterdam Studies in the ... IV: Current Issues in Linguistic Theory)

Romance Languages and Linguistic Theory 2004: Selected Papers from 'Going Romance', Leiden, 911 December 2004 (Amsterdam Studies in the Theory and History ... IV: Current Issues in Linguistic Theory)

Loan Phonology (Current Issues in Linguistic Theory)

Recommend Documents