Meaning Through Language Contrast, Volume 1 (Pragmatics & Beyond New Series)

Meaning Through Language Contrast

Pragmatics & Beyond New Series Editor Andreas H. Jucker University of Zurich, English Department Plattenstrasse 47, CH-8032 Zurich, Switzerland e-mail: [email protected]

Associate Editors Jacob L. Mey University of Southern Denmark

Herman Parret Belgian National Science Foundation, Universities of Louvain and Antwerp

Jef Verschueren Belgian National Science Foundation, University of Antwerp

Editorial Board Shoshana Blum-Kulka

Catherine Kerbrat-Orecchioni

Hebrew University of Jerusalem

University of Lyon 2

Jean Caron

Claudia de Lemos

Université de Poitiers

University of Campinas, Brazil

Robyn Carston

Marina Sbisà

University College London

University of Trieste

Bruce Fraser

Emanuel Schegloff

Boston University

University of California at Los Angeles

Thorstein Fretheim

Deborah Schiffrin

University of Trondheim

Georgetown University

John Heritage

Paul O. Takahara

University of California at Los Angeles

Kansai Gaidai University

Susan Herring

Sandra Thompson

University of Texas at Arlington

University of California at Santa Barbara

Masako K. Hiraga

Teun A. Van Dijk

St.Paul’s (Rikkyo) University

Pompeu Fabra, Barcelona

David Holdcroft

Richard J. Watts

University of Leeds

University of Berne

Sachiko Ide Japan Women’s University

Volume 99 Meaning Through Language Contrast: Volume 1 Edited by K.M. Jaszczolt and Ken Turner

Meaning Through Language Contrast Volume 1 Edited by

K.M. Jaszczolt University of Cambridge

Ken Turner University of Brighton

John Benjamins Publishing Company Amsterdam
/
Philadelphia

8

TM

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 Meaning through language contrast / edited by K.M. Jaszczolt and Ken Turner. p. cm. (Pragmatics & Beyond, New Series, issn 0922-842X ; new ser. 99-100) Includes bibliographical references and indexes. 1. Contrastive linguistics. 2. Semantics. 3. Pragmatics. 4. Grammar, Comparative and general. I. Jaszczolt, Katarzyna. II. Turner, Ken, 1956- III. Series. P134 M35 2002 401’.43--dc21

2002074427

isbn 90 272 5119 3 (Eur.) / 1 58811 206 3 (US) (Hb; alk. paper, vol. 1. 99) isbn 90 272 5120 7 (Eur.) / 1 58811 207 1 (US) (Hb; alk. paper, vol. 2. 100) isbn 90 272 5349 8 (Eur.) / 1 58811 332 9 (US) (Hb; alk. paper, set volumes 1-2.) © 2003 – 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

Table of contents

Acknowledgements

ix

Editorial preface Katarzyna M. Jaszczolt and Ken Turner

xi

Negation Distributional restrictions on negative determiners L. M. Tovena

3

Towards a comprehensive view of Negative Concord João Peres

29

Temporality On temporal constructions involving counting from anchor points – Semantic and pragmatic issues Telmo Móia

45

On the semantics and pragmatics of situational anaphoric temporal locators in Portuguese and in English Ana Teresa Alves

61

Remarks on the semantics of eventualities with measure phrases in English and Romanian Ilinca Cr˘ainiceanu

75

The present perfect in English and in Catalan Hortènsia Curell A contrastive reading of temporal-aspectual morphemes in Swahili: The case of ‘-li’ and ‘-me’ Frederick Kang’ethe Iraki

101

117



Table of contents

Modality Semantic and pragmatic constraints on mood selection Rui Marques Dilemmas and excogitations: Further considerations on modality, clitics and discourse A. Capone

129

147

Evidentiality Inferred evidence: Language-specific properties and universal constraints Sergei Tatevosov

177

Extension of meaning: Verbs of perception in English and Lithuanian Aurelia Usonien˙e

193

Perspectives on eventualities Information structure, argument structure, and typological variation Márta Maleczki

223

The network of demotion: Towards a unified account of passive constructions Andrea Sansò

245

Valence change and the function of intransitive verbs in English and Japanese Mayumi Masuko

261

The transitive/intransitive construction of events in Japanese and English discourse Patricia Mayes

277

Topics in grammar and conceptualization Towards a universal DRT model for the interpretation of directional PPs within a reference frame Didier Maillat

295

Table of contents 

The interaction of syntax and pragmatics: The case of Japanese ‘gapless’ relatives Akiko Kurosawa

307

Constraint interaction at the semantics/pragmatics interface: The case of clitic doubling Javier Gutiérrez-Rexach

335

Cross-language commutation tests and their application to an error-prone contrastive problem – Ger. einige, Fr. quelques, Sp. algunos Eva Lavric

355

Language index

371

Name index

373

Subject index

375

Contents of Volume 2

381

TSL[v.20020404] Prn:23/08/2002; 12:08

F: PBNSACK.tex / p.1 (ix)

Acknowledgements

We are indebted to a number of people. Firstly, we would like to express our thanks to the Conference Secretary Marilyn Dowling for her invaluable help with organizing the conference at Newnham College, Cambridge of which this volume is an outcome. To Patrick Welche we are grateful for designing and maintaining the conference web site. Our thanks also go to Heather Wynn for her administrative assistance. To the Ph.D. students Luna Filipovic, Rachel Smith and Marina Terkourafi we are grateful for their help with running the event. The co-opted members of the Programme Committee, Robyn Carston and Victoria Escandell-Vidal contributed their excellent work selecting conference contributions: thank you! To Elsevier Science we are grateful for their sponsorship of a social event accompanying the conference. We would also like to thank Newnham College at Cambridge for hosting the conference and all the College staff for their contribution to the tremendous success of the event. Last but not least, our gratitude goes to Isja Conen of J. Benjamins for her encouragement and invaluable editorial assistance.

TSL[v.20020404] Prn:23/08/2002; 12:18

F: PBNSPR.tex / p.1 (xi)

Editorial preface Katarzyna M. Jaszczolt and Ken Turner

This collection brings together selected papers from the Second International Conference in Contrastive Semantics and Pragmatics that took place in Cambridge from 11 to 13 September 2000 and was organized by the editors. The five-year period that separated this event from the first conference on this theme (see Jaszczolt & Turner 1996) witnessed a significant turn of orientation. There is a conspicuous shift of emphasis from speech-act based studies predominant in the past to semantics-based approaches, of the truth-conditional or cognitive orientation. Research that adheres to speech act theory is in a significant minority and appears progressively ‘detheorized’ in its selective utilization of those aspects of the theory that facilitate discourse about language use. Furthermore, this collection testifies to the growing importance of diachronic analysis. There are numerous contributions in the category of historical semantics and pragmatics and for those, predictably, grammaticalization is the focal object of research. The contrastive, cross-linguistic perspective proves to be a successful way of researching historical processes in that it sheds more light on the universality and language-relativity of aspects of language change (see e.g., Nicolle, Visconti, this collection). The historical perspective is also documented in speech act theory research (see Jung & Schrott, this collection). Another significant shift of emphasis can be observed at the lexical/sentence semantics interface. The generally acknowledged return to studies of word meaning is corroborated in this collection in the papers that approach the lexicon from the cognitive perspective or the perspective of a generative lexicon, frequently through the use of electronic lexical databases (see e.g., Viberg). The role of information structure for truth conditions is the cutting edge of contemporary semantic research. This avenue is also represented in the collection (see e.g., Maleczki). Descending to more detailed shifts of emphasis, the noun phrase seems to be gradually giving way to the studies of the verbal phenomena such as modality, tense, aspect and voice. These topics are numerously and strongly represented (see e.g., Marques, Masuko, Sanso, Usoniene).

TSL[v.20020404] Prn:23/08/2002; 12:18

F: PBNSPR.tex / p.2 (xii)

 Katarzyna M. Jaszczolt and Ken Turner

The discussion of the boundary between semantics and pragmatics has also undergone various changes of emphasis and style. In the 1970s, sense-generality and pragmatic inference were brought to the fore (see e.g., Cole 1981; Atlas 1989; Turner 1999; Jaszczolt 1999). Almost two decades later, the dynamic perspective in semantics allowed for contextual information to be semanticized (see Kamp & Reyle 1993). Subsequent developments of the idea of underspecification (see e.g., van Deemter & Peters 1996) demonstrated that pragmatics, intentions and intentionality are frequently irreducible and do not yield to formalizations (see Blutner & van der Sandt 1998; van Deemter 1998; Dekker 2002; Jaszczolt & Turner 2002). The predominance of semantic analyses strongly suggests that (i) contrasting meaning in various natural languages requires firm foundations, strict modelling and some degree of formalization; (ii) both (a) cognitive semantics and (b) Tarskian, post-Montagovian semantics supplemented with post-Gricean pragmatics are more productive than the offshoots of the ordinary language philosophy. Finally, to address the empiricism-rationalism dilemma, it can be observed that inferring from quantitative analyses and supporting theories by unquantified data constitute equally successful directions in semantic and pragmatic research.

References Atlas, J. D. (1989). Philosophy Without Ambiguity: A Logico-Linguistic Essay. Oxford: Clarendon Press. Blutner, R., & R. van der Sandt (1998). Editorial Preface to Special Issue on Underspecification and Interpretation. Journal of Semantics, 15, 1–3. Cole, P. (Ed.). (1981). Radical Pragmatics. New York: Academic Press. Dekker, P. (2002). Pronouns in a pragmatic semantics [Special Issue on Conceptual Contours at the Semantics/Pragmatics Interface]. Journal of Pragmatics, 34, 815–827. Jaszczolt, K. M. (1999). Discourse, Beliefs, and Intentions: Semantic Defaults and Propositional Attitude Ascription. Oxford: Elsevier Science. Jaszczolt, K. M., & Turner, K. (Eds.). (1996). Contrastive Semantics and Pragmatics, 2 volumes. Oxford: Pergamon Press. Jaszczolt, K. M., & Turner, K. (2002). Editorial Introduction to Special Issue on Conceptual Contours at the Semantics/Pragmatics Interface. Journal of Pragmatics, 34, 811–813. Kamp, H., & Reyle, U. (1993). From Discourse to Logic: Introduction to Modeltheoretic Semantics of Natural Language, Formal Logic and Discourse Representation Theory. Dordrecht: Kluwer. Turner, K. (Ed.). (1999). The Semantics/Pragmatics Interface from Different Points of View. Oxford: Elsevier Science. van Deemter, K. (1998). Ambiguity and idiosyncratic interpretation. Journal of Semantics, 15, 5–36. van Deemter, K., & Peters, S. (Eds.). (1996). Semantic Ambiguity and Underspecification. Stanford: CSLI Publications.

TSL[v.20020404] Prn:23/08/2002; 12:04

P I

Negation

F: PBNSP1.tex / p.1 (1)

Distributional restrictions on negative determiners L. M. Tovena Université de Lille

This paper discusses restrictions in the distribution and interpretation of negative determiners in English, Italian and French. The typology of determiners defined by Chierchia (1998) is taken as a starting point, and it is tested against new data. Several shortcomings are highlighted. In particular, it is shown that some determiners are sensitive to differences among members of the class of uncountable nouns, and that the distribution of a number of singular determiners cannot be captured by restricting their domain of quantification to all and only singular countable nouns.

.

The issue

This paper discusses restrictions in the distribution and interpretation of negative determiners in English, Italian and French. We take the typology of determiners defined by Chierchia (1998) as our starting point, and we test it against new data. In Chierchia (1998) a unique lattice model is built for the denotation of countable and uncountable nouns, under the name of Inherent Plurality Hypothesis. The main thesis of this unified analysis of nouns is that mass nouns differ from count ones (only) insofar as they come out of the lexicon with plurality already built in. A singular countable noun is associated with a set of atoms, and the set-forming operator PL is used to enable us to talk about sets of them. The basic lexical entry of a mass noun does not single out the set of atoms, but a sublattice. In this case, the difference between singular and plural is neutralised, for the noun applies equally to atoms and sets thereof. The atomic texture is foregrounded in countable nouns, as by definition its extension singles out a set of atoms. On the contrary, in uncountable nouns this structure is present but only implicitly. This gives one a single structure for describing the denotation of countable and mass nouns. This unified model is then used by Chierchia to analyse the distribution of determiners. Attention is paid to the issue of how to compose information coming



L. M. Tovena

from the morphological number of the N, the count/noncount distinction and the determiner characteristics in order to obtain truth conditional satisfaction in the appropriate cases. The structure of the paper is as follows. We will first look at the case of the English unrestricted determiner no, then at the singular Italian determiner nessuno (no) together with French aucun (no). Then, determiner-like uses of Italian niente (no) will be discussed. Finally, this type of occurrence of niente will be contrasted with uses of nessuno-phrases in verbless clauses.

. Unrestricted determiners The standard definition of the English determiner no in Generalized Quantifier Theory is as in (1). (1) NO (X)(Y) = 1 iff X ∩ Y = Ø This determiner is called unrestricted (Chierchia 1998) because it applies to singular and plural count nouns, as well as to uncountable nouns. In this case, the variation in number and mass/countable nature is said not to affect the interpretation of the determiner, which is always expressed as a constraint of empty intersection of the first argument with the second. In contrast, in the case of the, another unrestricted determiner, the same variation has a broader impact. There is a uniqueness presupposition when this determiner combines with a singular noun, and a presupposition that there be more than one element when it combines with a plural noun. Chierchia ascribes to Schwarzschild (1991) the claim that a theory of plurality which uses set theory gives the wrong results in the case of (1). For a sentence such as no men lifted the piano, such an analysis would require that no men did, but leave open the possibility that single men did. Similar considerations apply for cases where no combines with a singular noun. In sum, if the head noun of the quantified NP is singular and count, as in (2), the proposition will be false when a student hit a lorry, but it should also turn out to be false when a group of students does. Similarly, (3) is expected to be false also when a single student hit the lorry. The two sentences are truth conditionally equivalent, and sentence (4), for instance, is contradictory. (2) No student hit a lorry (3) No students hit a lorry (4) *No students hit a lorry but Daniel did In order to treat the question of the unrestricted distribution of no and its seeming insensitivity to morphological information, that he believes to be the cause of the

Distributional restrictions on negative determiners

contradictoriness of (4), Chierchia (1998) puts forward the definition in (5), which is cast under his Inherent Plurality Hypothesis. (5) NO (X)(Y) = π(X) ∩ Y = Ø For any set u in U, the ideal π(u) generated by u is {x : x ≤ u}, where ≤ is the ‘component of ’ relation that orders the structure used to represent nouns.1 For any set X, π(X) is the set of all the elements which are components of the supremum of X. So, if X is the denotation of a mass noun, π(X) = X, because X already contains all the components of its supremum. On the contrary, if X is singular count, π(X) yields the closure of X under the operation of ‘sum’ defined in terms of ≤. Finally, if X is plural, π(X) will add the atoms to it. In any case, π(X) yields a complete atomic sublattice of the domain. In this way, (2) and (3) are accounted for as well as the fact that no works on uncountable, see (6) and (7). As a short aside, note that Chierchia touches only very briefly on the distinction between concrete and abstract mass nouns. For the latter, he proposes quanta identified with instantaneous states in the case of ‘honesty’ or with unspecified units in the case of ‘sense’, and says that abstract mass nouns denote sets of these quanta closed under sum. (6) No water flooded this area (7) There was no honesty in his words In other words, π(u) allows one to get all the components of u independently of the constraints that number and the count/uncountable distinction impose on the denotation of N. It is as if information provided by morphological number were no longer available at the level of the quantified NP. Actually, Delfitto (1998) goes as far as claiming that the difference in number in a no-phrase is devoid of any interpretive import. However, there is some evidence supporting the interpretive relevance of morphological number in unrestricted negative determiners such as no.

. The relevance of number In this section, we review some evidence in favour of the interpretive relevance of number information. First, it is true that no combines with singular and plural count nouns, but it is also true that the resulting quantified NPs are not always free alternates. (8) No man is immortal (9) No buses are running today Examples (8) and (9) are presented to highlight the fact that there is a preference for using no Nsingular when making a generic statement, whereas no Nplural is used in





L. M. Tovena

contingent statements. It is not obvious how this situation may be captured if one takes (5) as saying that the denotation of the first argument is always ‘normalised’ by π and number differences are hidden as a result. A second case where number differences cannot be claimed to be devoid of interpretive import has to do with the individuation of references and presupposition. Consider examples (10)–(11). (10) There will be no train tomorrow (11) There will be no trains tomorrow Sentence (10) contains a no-phrase with a singular countable noun and can be used felicitously only to deny tomorrow’s running of a single train. Furthermore, the existence of a train is presupposed and its ‘identity’ is discourse old and accessible. On the contrary, (11) cannot be used to deny the running of a single train, and although it may be taken to presuppose the existence of trains, as it is the case for the restrictor of a generalised quantifier, it does not impose requirements with respect to their identification. A third case is about coreference and discourse anaphora. Consider the sentences in (12). (12) a.

No students attended the meeting. *Each of them had something else to do. b. No student attended the meeting. Each of them had something else to do. c. No students attended the meeting. All of them had something else to do.

The anaphoric link established by the quantified NP in the second clause cannot be resolved if no coreference can be established with respect to the denotation of the N in the quantified NP in the first clause. In (12a) coreference does not seem to be possible, whereas it does in (12b, c). This variation in acceptability confirms the observation that number remains visible. Only no Nsingular gives access to the set of individual students, which is the restricted complement of the denotation of the quantified NP. Finally, no-phrases behave differently with respect to collective predicates, as shown in (13). As discussed in Löbner (2000), quantified NPs with plural noncollective nouns admit collective predicates in the second argument, as shown in (13b), whereas quantified NPs with singular nouns do not, as shown in (13a). In Löbner’s analysis this difference follows from the fact that only the former type of phrase yields a quantification for which the predicate in the second argument is defined.

Distributional restrictions on negative determiners

(13) a. *No student gathered in the courtyard. b. No students gathered in the courtyard. The visibility of number information from the head noun could be enhanced within Chierchia’s framework by assuming that the part of structure made visible by π remains backgrounded, and that the original denotation is the only foregrounded part. However, giving different statuses to the original denotation of the noun and to the part added by π goes against the original idea of ‘normalising’ all the cases, and nullifies the purpose of π.

. Singular determiners Not all negative determiners are unrestricted. Some of them exhibit restrictions related to the structure of the domain. The class of singular determiners gathers together determiners that require a domain composed only of atoms. Chierchia defines the function SG to check whether a predicate foregrounds a set of atoms or not. SG maps each possible noun denotation into its presupposed singularities. If A is already a set of singularities, then SG(A) = A. Otherwise, SG(A) is the set of atoms At that generates A via the set-forming operator PL, if such a set exists. Therefore, since SG(A) is undefined if A is the denotation of a mass noun, SG works as a tool for setting apart countable from uncountable nouns. The function SG acts as domain regulator for determiners according to the schema in (14a), abbreviated as (14b). (14) a. D(SG(X))(Y) b. DSG (X)(Y) Then, S is defined as a restriction of SG to atoms. It sets apart singular from plural countable nouns. For any subset X of the domain, S(X) = X if X ⊆ At, and otherwise S(X) is undefined. Using S, one can capture the distribution of determiners such as the English every, classified as an instance of singular determiner and defined as follows. (15) EVERYS (X)(Y) = S(X) ⊆ Y Chierchia assigns also the Italian negative determiner nessuno2 (no) to the group of singular determiners, together with positive qualche (some), but does not provide a definition for the former. The data in (16) show that nessuno combines with singular count nouns but not with plurals or uncountables and seem to offer support to his classification.





L. M. Tovena

(16) a.

nessuno studente no student b. *nessun libri no books c. *nessun vino no wine

Considering the data in (16) and the contradictory status of (17), and working from (5) and (15), one can get (18) as a potential definition for nessuno. (17) *Nessuno studente ha colpito il camion ma Mario e Daniele sì no student hit the lorry but Mario and Daniel did (18) NESSUNOS(X)(Y) = π(S(X)) ∩ Y = Ø

(putative)

However, as soon as one takes into account a little more Italian data, it appears not only that (18) is problematic, but also that a characterisation of nessuno as standard singular determiner is not as straightforward as it would appear from the discussion in Chierchia (1998).

. Nessuno with mass nouns First, contra Chierchia’s claim, nessuno applies to uncountable as well as count nouns, cf. (19). (19) Non ha mostrato nessun coraggio She didn’t show any courage at all (20) ?Nessun’acqua laverà il suo sangue dalle tue mani no water will wash her blood from your hands Sentence (20) is admittedly marginal, and has required a careful wording in order to reach so much as this level of acceptability. On the contrary, (19) is perfect in the reading ‘not the least bit of ’, i.e. a truly mass reading. In both cases we leave aside the taxonomic readings, which seem always to be possible, although rather awkward in the case of (19). Therefore, the restriction S does not seem to be appropriate for capturing the distribution of nessuno, because it rules out acceptable cases such as (19), where nessuno combines with an uncountable noun. The problem is that S cannot be defined in a different way, in the system as it is. This is not a minor point. Indeed, Chierchia claims that there are no functions restricted to singular count and mass noun denotations in natural language, and says that the existence of such a gap follows from his system because there is no natural domain regulator that would have the effect of restricting the left argument of a determiner in this way. Recall that nessuno does not combine with plurals,

Distributional restrictions on negative determiners

and that in his analysis plural and uncountable nouns come down to the same for the issue in hand. In the light of the evidence provided in (16) and (19), his classification appears to break down. Second, at first sight the distribution of nessuno does not seem to correspond to a consistent behaviour. This determiner sometimes combines with a mass noun, as in (19), and sometimes does not, as in (16c). This variation rules out the possibility of capturing the distribution of nessuno by referring to a constraint expressed purely in terms of morphological number. However, on a closer look, it can be seen that the line separating acceptable and unacceptable instances follows rather closely the abstract vs concrete divide. There is a preference for abstract mass nouns. As noted above, (20) is not perfect. In general nessuno does not combine with concrete mass nouns of the substance mass type, e.g. ‘milk’ and ‘water’, or of the collective mass type (Bosveld-de Smet 1998), i.e. ‘mail’ and ‘furniture’. (21) a. *Non ha messo nessuno zucchero nella spremuta she didn’t put any sugar in the juice b. *Non ha ricevuto nessuna posta She didn’t get any mail However, this distinction among types of mass nouns cannot find a place in Chierchia’s classification. Note that, with respect to uncountable nouns, the behaviour of nessuno clearly differs from qualche, the other Italian singular determiner mentioned in Chierchia (1998), as the latter cannot combine with them at all, as shown in (22). (22) a. *Questa impresa ha richiesto qualche coraggio this entreprise required some courage b. *Ha messo qualche sale nella minestra she put some salt in the soup It could be objected that these data do not really falsify Chierchia’s classification. It could be said that in (19) there is an implicit occurrence of a type-shifting device, that makes available a reading of the type ‘degree-of ’ or ‘act-of ’. However, this remark would not give any contribution towards clarifying the contrast between (19) and (22), since it would not explain why a similar device is not equally available for (22a). Third, note also that nessuno is not the only determiner to exhibit a non homogeneous behaviour with respect to uncountable nouns, to split this class into two groups, and to cluster abstract uncountable nouns together with singular countable nouns. French provides more examples, as shown in (23) for the negative determiner aucun.3





L. M. Tovena

(23) a.

Il n’a pris aucun livre He did not take any book b. Il n’a montré aucune pitié He didn’t show any mercy c. *Il n’a vu aucun étudiants He didn’t see any students d. *Il n’utilise aucun sable He uses no sand

Furthermore, this sensitivity to the typology of uncountable noun is not a peculiarity of negative determiners. As shown in (24), the negative polarity and free-choice French item le moindre4 (the least) behaves in the same way in this respect. (24) a. b. c. d. e.

Il n’a pas lu le moindre livre he did not read any books Il connaît le moindre recoin du village he knows any corner Il n’a pas montré le moindre courage He did not show the least bit of courage *Il n’a pas lu le moindre livres/les moindres livres He didn’t read any books *Il n’a pas bu la moindre eau He did not drink a drop of water

A first conclusion that can be drawn from the preceding discussion is that Chierchia’s domain restrictor S does not work for the cases in hand, while a restrictor predicted to be impossible seems to be needed. Then, this need for an impossible S could be taken to point out that the whole model does not work since it is unable to distinguish among different types of mass nouns. Otherwise, less drastically, it could be taken to suggest that the interaction with the mass/count distinction – i.e. Chierchia’s noun properties 4 to 7 – is not sufficient for a proper classification of determiners. The idea of modifying the basic ontology is too costly a solution, in our opinion, to be further explored as the first choice, and we would rather leave it as a last resort option. The definition of S and the interaction between determiners and the mass/count distinction can be considered as two sides of the same coin. This section closes with the question of how to characterise the subset of uncountable nouns that combines with nessuno, aucun and le moindre. The idea of abstract noun, that provides a first intuition for singling out this subset, cannot be captured simply by requiring that the members of this subset be devoid of the spatial dimension, as shown by the unacceptability of (25).

Distributional restrictions on negative determiners

(25) a. *Non ho nessun tempo I have no time b. *Je n’ai aucun temps Tovena (2001) puts forwards an account for the split within mass nouns and the association of part of it with count nouns which can be exploited for the data in (16)–(19) and would come at a minimal cost to Chierchia’s model. We discuss it in the next section.

. Abstract nouns The subset of mass nouns compatible with nessuno and aucun can be characterised via the notion of grandeur intensive (intensive quantity)5 (Van de Velde 1996), i.e. entities that manifest themselves in degrees. Van de Velde provides a set of contexts where this type of nouns behave differently from the rest of the class. For instance, they are compatible with un certain (a certain), like countable nouns but unlike the rest of uncountable nouns. They cross over the quantity/quality distinction, cf. (26), – which suggests that degrees do not qualify as traditional quanta and are not like measures that map parts into numbers. (26) a.

quanto coraggio! = che coraggio! what courage b. quanto burro! = che burro! what a lot of butter = what (good) butter

Furthermore, they are only indirectly quantized. For instance, courage is a property expressed via courageous acts, which, in turn, are quantized into eventualities instantiating them. Similarly, the quanta ‘instantaneous state’ proposed by Chierchia applies to ‘honest behaviour’, i.e. a manifestation of ‘honesty’, and not to ‘honesty’ itself. Tovena (2001) explains the double behaviour of intensive quantities by building on these observations. The classification of intensive quantities as mass nouns is derived from the fact that their lexical entries are not directly associated with the atoms in their denotations. Degrees of intensive quantities are quanta that do not discretise the domain into individuals, nor into measurable parts. As Kant says: Thus a certain expansion which fills a space, for instance, heat, and every other kind of phenomenal reality, may, without leaving the smallest part of space empty, diminish by degrees in infinitum, and nevertheless fill space with its smaller, quite as much as another phenomenon with greater degrees. (Kant 1881: 153)





L. M. Tovena

So, a lot of courage is not a bigger (extensive) quantity of courage, it may well be the same quantity at that, but it is a bigger intensive quantity, i.e. a higher degree of intensity of courage. At the same time Tovena also derives their distribution somehow close to countable nouns, as far as the combination with determiners is concerned, because manifestations, such as courageous acts, are sort of instances of the entity ‘courage’. A high degree or, for that, any given degree of courage are different types of courage. The domain is discretised by weakly discrete units interpreted as degrees, and this is the default way for intensive quantities. For countable nouns the default way is a discretisation by individuals, which are strongly discrete units. For the rest of uncountable nouns, the default strategy is by parts, defined with respect to units of measure, and by species only as a secondary option that applies to rescue a phrase when no measure is provided or recoverable from the context. For intensive quantities, no units of measures are available, because measures apply to extensions. It appears that the level of discreteness provided by weakly discrete units is enough to satisfy the requirement of a singular determiner such as nessuno and aucun. On the one hand, it provides a partitioning that can be exploited by the domain restrictor S. On the other hand, strongly discrete units are not needed because we are dealing here with negative quantifiers, whose witness set (Barwise & Cooper 1981; Szabolcsi 1997) is empty by definition. Hence, at no stage of the interpretation will a particular occurrence have to be singled out. Therefore, the acceptability of (19) and (23b) is the result of the interaction between a weakly discretised domain and a determiner whose denotation can be represented by a constant function. The situation changes if the determiner is not a constant function. In the case of the positive singular Italian determiner qualche, for instance, some occurrences would have to be singled out despite the fact that they are not lexically accessible, because the witness sets have cardinality greater than zero. As a result the quantified NP cannot be built, cf. (27). (27) *qualche coraggio some courage

. First partial conclusion In conclusion, by assuming that the mass/count distinction is based on strongly discrete units, and that quantifiers may be satisfied with weakly discrete units in certain cases, one can rescue Chierchia’s domain restrictor S. However, his description of the interaction of the determiner system with the mass/count distinction is confirmed to be insufficient for a proper classification of determiners.

Distributional restrictions on negative determiners

The line of reasoning adopted in the present paper seems promising also with respect to the problem raised by sentences such as (28), predicted to be impossible by Chierchia’s account. (28) She still had every confidence in him as a mathematician. In this case, the quantified NP has a unique witness set that includes the whole denotation of the noun. Such a set may be identified without making use of strongly discrete units as it doesn’t partition the denotation of the noun. With respect to le moindre, cf. the sentences in (24), although this determiner is not a constant function strictly speaking, it points to an end-of-scale position and can be equated with a universal quantification on the whole scale and, depending on the directions of the inferences on the scale, turns out to be a sort of universal or negative quantifier.

. Niente N phrases In the remainder of this paper we will discuss more data on negative determiners in Italian. We will start off from a remark on nessuno made by Chierchia while discussing the predictions of his classification of determiners. He observes that there is no principled way to predict if a language has an unrestricted function or a set of restricted functions to cover a given space. Furthermore, there may be gaps. He says that Italian has no mass or plural counterpart of nessuno. The ‘mass side’ of nessuno has been touched upon in Section 3, where it has been shown that the distribution of this determiner is broader than predicted by Chierchia. However, a second option is sometimes exploited in Italian. The expression niente (nothing), usually working as a quantifier, in certain cases seems to work also as a determiner. In these cases it is interpreted as no. Some examples are provided in (29) to (31). (29) a.

Niente età limite per le adozioni no maximal age threshold for adoptions b. Niente pensione a chi risiede all’estero no pension to the citizens living abroad

(30) a.

(25/7/1996IM)

Per questa torta ci vogliono tre etti di farina, un uovo, due mele e niente zucchero for this tart one needs 300gr of flour, one egg, two apples and no sugar b. Daniele mangia pesche, pere, niente mele e poca uva Daniel eats peaches, pears, no apples and a few grapes





L. M. Tovena

(31) a.

Non ho niente voglia I have no desire (i.e. I don’t feel like it at all) b. Non fa niente freddo it is not the least bit cold

In these cases, niente combines with bare nouns or N , as shown in (32) with an adjective in pre and post nominal position, and does not tolerate the presence of any other determiner, see the selection of determiners in (33). (32) a.

niente piccoli passi no small steps b. niente palloni rossi no red balls

(33) a. *niente l’acqua no the water b. *niente un libro no a book c. *niente questo libro no this book d. *niente tre libri no three books The examples in (29) to (31) might involve some more structure than a simple DP. The literature, to the best of our knowledge, has largely ignored these cases. In this paper, pending a syntactic analysis of these structures, we will adopt the minimal option consisting of treating them as quantified NPs, i.e. DPs. In this view, niente behaves as an unrestricted determiner, as it combines with singular and plural countable nouns as well as uncountable ones. In these uses, it may be taken to complete the coverage of negative quantification in Italian. Examples (29) to (31) instantiate three different cases. By far, the most common case is the one presented in (29). A first important question raised by this type of occurrences of niente is why should the distribution of this phrase be restricted almost exclusively to verbless sentences. Again, the literature does not offer an explanation for this behaviour. Examples such as (29a) and (29b) are very shortly, if ever, discussed. It has been proposed in Manzotti (1991) to treat them as cases of ellipsis of the verb. It is said that (29a) is equivalent to something close to (34), but the reason motivating the relation is not discussed. This analysis would require the additional step represented by establishing a relation of transformation or deletion that would hold between niente and the determiner nessuno or a surface null element, so that (29a) can be related to (34), or to (35) where there is no determiner. This is a crucial step which is left unexplained.

Distributional restrictions on negative determiners

(34) Non ci sarà nessuna età limite per le adozioni there will be no limit of age for adoptions (35) Non ci sarà età limite per le adozioni there will be no limit of age for adoptions It is also not explained why verb ellipsis is obligatory in (29). For these ‘reduced’ clauses no ‘full’ clauses are possible. Furthermore, an ellipsis implies that the thing being omitted has already been expressed, which does not seem to be the case for the verbs in these clauses. Third, niente and nessuno phrases in verbless clauses undergo a particular thematic restriction which so far seem to have gone unnoticed. Neither of these phrases can discharge what would be the agent role in the corresponding clause with no verb ellipsis. When needed, this role is discharged by a PP, as shown in (36) and (37), which makes us suppose that a passive verb form was ‘left out’. Such a restriction does not follow from ellipsis. (36) Dal Csm nessuna censura al procuratore Vigna no reproof [is made] to attorney Vigna by the CSM

(19/2/1998IM)

(37) Niente scherzi dalla Bundesbank no tricks [were played] by the Bundesbank

(22/3/1998IM)

The second case, exemplified in (30)–(31), involves niente-phrases inside a conjunction. Depending on the preferred analysis of coordination, they might also be taken as more instances of ellipsis, if one stress the following fact. On the one hand, in both (30a) and (30b) niente occurs within a sequence of NP conjuncts6 that are all acting as direct objects of the verb occuring in the same sentence. On the other hand, however, it is the whole conjunct structure that instantiates the argument position and not the single conjuncts, and there is no verb inside the sequence. Note that there is no constraint on the linear position of the conjunct containing niente in the sequence. Middle and final position are possible, as we have seen in (30b) and (30a) respectively, but the initial position is not excluded either, as shown in (38). (38) Se si passa poi dalla quantità dei lavoratori alla qualità della loro vita, ci si addentra in un roseto niente petali e tutto spine: è impossibile programmarsi serenamente la vita sui tempi lunghi, essendo il rinnovo dei loro contratti – in media ogni anno – legato necessariamente agli appalti che le imprese o cooperative da cui dipendono riescono a vincere. (13/12/2000IM) Then, if [the discussion] moves from the number of workers to the quality of their lifes, we enter a rose garden [where there are] no petals and all thorns: it is impossible for them to plan their lives calmly on a long time





L. M. Tovena

span, as the renewal of their contract – about once a year – is necessarily linked to the contracts by tender that the companies or cooperatives employing them manage to get. The sentences in (30) and (38) are special insofar as they do not satisfy the constraint of negative concord. Simplifying a little, negative concord in Italian requires negation to be expressed in a preverbal position for the sentence to be allowed to contain N-words in postverbal positions (Corblin & Tovena 2001). However, it appears that this constraint is waived when a conjunction in postverbal position contains positive and negative conjuncts, in which case the verb would be expected to be both negated and not negated depending on the polarity of the conjunct, as discussed in Tovena (1996). Example (38) brings in a novelty element in the discussion because it contains an instance of negative conjunct occurring as the first element of a sequence. A constraint that appears to be satisfied in all these cases of conjunction, and for which we cannot offer an explanation so far, applies to the type of noun that enters these niente-phrases. This type of context does not seem to be suitable for a niente Nsingular count phrase, as shown in (39), contrary to the general situation described in Section 6. (39) ??Daniele mangia pesche, pere, niente mela e poca uva Daniel eats peaches, pears, no apple and a few grapes If interpretable at all, (39) talks about the type of food ‘apple’, and not about the fruits, as it is done for pears and peaches. Finally, as for examples such as (31), they could be dismissed as cases where niente works as a degree adverb that strengthens the predication, in alternative to the analysis as determiner. In other words, (40a) would be considered as a paraphrase of (40b) – a hypothesis that would go against Molinelli’s (1988) claim that modern Italian does not have this type of adverbial use – rather than of (40c). So far, we do not see strong evidence in favour of either analyses. (40) a.

Non ho niente fame I am not a bit hungry b. Non ho affatto fame I am not hungry at all c. Non ho neanche un po’ di fame I am not a bit hungry

We leave structural restrictions on the distribution of the determiner-like niente for further research. In the remainder of this paper we concentrate on verbless clauses, because they are the most frequent cases. They are often used in headings and titles, but may occur in the body of an article too, as shown in (41).

Distributional restrictions on negative determiners

(41) Il testo più bello comunque è quello che elogia il cattivo carattere sul luogo di lavoro. Niente sorrisi e cortesie, ma solo espressioni di malcelato fastidio per un ambiente dove le buone maniere sono solo la manifestazione di un’adesione alla gerarchia e l’accettazione di una espropriazione del proprio tempo da parte del capitale. (12/4/1998IM) Anyway, the best text is the one that celebrates the fact of being an illnatured person on the working place. No smiles or courtesies, but only expressions of half hidden annoyance for an environment where good manners are only a way to show that one endorses the hierarchy and accepts to lose control over one’s own time in favour of the capital. Furthermore, verbless clauses are the type of construction where niente can occur in partial alternation with nessuno. In the examples seen so far, both forms have been glossed as ‘no’, and are interpreted as imposing a constraint of empty intersection between the first and second argument. For want of a full analysis of niente-phrases that would expose its peculiarities, the study of its alternance with nessuno-phrases may provide some help to understand its functioning.

. Negative phrases in verbless clauses As it will soon be apparent, there are partial overlaps in the distribution and use of nessuno N and niente N phrases. One of the aims of this section is to discuss the cases for which often native speakers point out that occurrences of nessuno can be replaced with niente and vice-versa, with no apparent change in meaning, despite possible number modifications, and cases where the two types of phrase are not alternate. What makes this relatively free alternation of particular interest for us is the fact that it does not show up where one might expect on the basis of the equivalence between uncountable and countable plural nouns expressed in Chierchia’s system. In the following, first we will review the cases where there are overlaps, and then we will discuss the differences that are perceived in the use of the remainder of the forms.

. Overlaps in the distribution We saw that nessuno combines only with singular nouns. Niente is unrestricted. Tests on native speakers have shown a moderate possibility of replacing these Nwords with one another in their uses as determiner in verbless clauses. The examples are taken from the national press. Let us start from occurrences of nessuno that can be substituted by niente. In (42) the substitution is possible provided the





L. M. Tovena

countable noun is changed into plural form. In case there is an uncountable noun, such as in (43), the replacement does not force a change in morphological number, but there is a further restriction discussed below with respect to examples (65) and (66). (42) Nessuna reazione invece dalla presidenza della Repubblica(19/4/1998IM) but no reactions from the president of the republic (43) Nessuna meraviglia quindi se il discorso ha mandato su tutte le furie i repubblicani e si è guadagnato le lodi di non pochi esperti che hanno parlato di capolavoro politico concepito per la conquista del centro. (29/01/1997L’U) Then no surprise that the speech infuriated the republicans and succeeded in being praised by not a small number of experts who said it was a political masterpiece conceived for conquering the centre. The picture that emerges from the answers provided by native speakers can be summarised via the following schema. Nessuno-phrases might be changed into – niente+plural N – and into niente+singular N if the head noun is abstract (some type) Next, let us consider substitution in the opposite direction. Native speakers do not accept substituting nessuno for any naturally occurring instances of niente combined with singular nouns. The substitution in (44) would cause a shift from a sentence that conveys a command forbidding the envoy of the army, to a sentence that would record a situation of absence of any of the armies, with a partitive interpretation not available in the original form. In the case of (45), substituting the forms is not possible simply because nessuno does not combine with concrete uncountable nouns. (44) ANDREATTA Niente esercito in Puglia, altri blitz se sarà necessario (7/3/1997IM) Andreatta [said]: no army in Puglie, more blitz if needed (45) Ieri è stata rispettata la tregua, niente letame sulle autostrade o trattori sui binari (29/11/1997IM) [...] no manure on the motorways nor tractors on the rails In the case of niente combined with a countable plural noun, the substitution is possible. Example (46) would be equally possible with nessuno. (46) ONU Francia, Russia e Cina: niente mozioni contro l’Iraq (7/9/1996IM) Onu France, Russia and China: no motions against Iraq

Distributional restrictions on negative determiners

The picture that emerges from the answers of native speakers with respect to this second type of replacement can be summarised via the following schema. – –

Niente+singular N cannot be replaced by nessuno-phrases, except for abstract Ns (some type) Niente+plural N can be changed into nessuno-phrases

The first striking observation that can be made on the basis of these schemata is that meaning overlaps do not concern niente-phrases containing singular countable nouns. This is all the more surprising if one recalls that singular countable nouns make up the largest area of overlap in the distribution of the two items from the morphosyntactic point of view. A second observation, connected with the previous one, is that, when applied to a domain of atoms, these two negative quantifiers contribute different facets of meaning. In the following subsection, we give a closer look at the differences in the distribution to see if they provide hints for the characterisation of such facets. But, before that, let us add just one more observation on coordination. It was noted above, with respect to (30), that niente-phrases can occur as conjuncts. In the case where the coordination is inside the phrase, that is to say that the coordinated elements are the nouns niente applies to, as in (41), we observe a shift from conjunction to disjunction whenever niente is replaced with nessuno, as shown in (47). (41) Niente sorrisi e cortesie, ma solo espressioni di malcelato fastidio [. . . ] (12/4/1998IM) No smiles or courtesies, but only expressions of half hidden annoyance (47) Nessun sorriso o cortesia . . . At the present moment, it is not clear whether this shift is due to the constraint that nessuno, as a singular determiner, does not combine with a plurality – a hypothesis that requires the supplementary assumption that conjunctions denote only pluralities. It could also be ascribed to structural constraints, given the unacceptability of (48), but note that (49) is a little marginal but still possible, and has only the interpretation whereby there isn’t an entity that shares both properties of being white and being black, with e expressing set intersection as expected. On the contrary, no noticeable differences are reported between the cases of conjunction and disjunction with niente, cf. (41) and (50), as if intersection and union made no difference. Thus, nessuno appears to sharpen the logical interpretation of the connectives. (48) *nessun sorriso e cortesia (49) ??nessun bianco e nero no white and black (e.g. chequered or striped combination) (50) niente sorrisi o cortesie





L. M. Tovena

. Differences This section looks at the differences in the distribution of nessuno and niente in verbless clauses. We begin by comparing phrases with countable nouns, then we look at cases with countable nouns with both countable and uncountable readings, and we conclude with mass nouns.

.. Countable nouns When discussing sentence (44), it has been pointed out that different presuppositions arise from the use of nessuno and niente. The examples in (51) and (52) replicate the same opposition. The difference is not simply between presence versus absence of a presupposition of existence. Rather, in the case of presence, the presupposition of existence seem to be more precisely characterised as a presupposition of plurality, in the sense that the cardinality of the domain of quantification is expected to be > 1. In (44), as well as in (51), there is a presupposition of unicity with respect to the denotation of the noun. No such presupposition arises in the case of (52). On the contrary, here nessuno is used to exclude any interpretation of the resolution that could be taken as providing a mandate for military intervention in (52a) and (52b). (51) Bossi: “Niente Lega alle elezioni padane” Bossi said: No Lega at the elections in ‘Padania’

(20/7/1997IM)

(52) a.

KOFI ANNAN: Nessun mandato per colpire l’Iraq. Il segretario generale dell’Onu ha gettato acqua sugli ardori bellici degli Stati Uniti sostenendo che in caso di una infrazione irachena dell’accordo del 22 febbraio scorso sulle ispezioni ai siti presidenziali, la risoluzione approvata dal Consiglio di sicurezza non concede affatto agli Usa il cosiddetto “grilletto automatico”. (10/3/1998IM) Kofi Annan says that there is no mandate for hitting Iraq. The UN secretary-general threw some water on the United States’ burning war fervour, claiming that were Iraq to infringe the agreement on the inspections to the presidential sites dated 22 February, the resolution approved by the Security Council does not give the USA any so-called ‘automatic trigger’ b. Russia, Francia e Cina: “Nessun automatismo” (4/3/1998IM) Russia, France and China: No automatic right [to intervene]

Second, nessuno is marginal with nouns identifying unique entities, what we could call Russellian iotas. Sentence (53a) is the standard way of talking about a given type of weather. In our galaxy there is only one sun, and niente can be used to deny the presence of this entity in a particular context. On the contrary, (53b)

Distributional restrictions on negative determiners

brings in the presupposition that there is more than one sun, which, in order to be accommodated, forces a reading whereby the sentence conveys a claim about an astronomical situation that exceeds our galaxy. (53) a.

Niente sole quest’estate no sunshine this summer b. ??Nessun sole quest’estate none of the suns this summer

There is a similar contrast in French. (54) a.

??Aucun soleil cet été

none of the suns this summer b. Pas de soleil cet été no sunshine this summer It could be objected that (53b) and (54b) are not real occurrences of a countable noun, since their meaning is close to ‘sunshine’ or ‘sunny weather’. But even if one assumes that they contain instances of an uncountable noun, this does not explain why the interpretation as uncountable is the only one available. Furthermore, it would leave unexplained the fact that the shift to this type of reading cannot be escaped. The acceptability of (55) suggests that the presupposition of non unicity need not be strictly enforced considering the spatial and temporal dimensions. In fact, (55), like (53a), introduces a unique entity in the domain of discourse, but, unlike (53a), it is possible to exploit a variation along the temporal dimension. Since the pope is an entity who has been ‘instantiated’ by more than one individual across the centuries, there is no strict uniqueness and nessuno becomes acceptable. (55) Nessun papa ha difeso questa causa no pontiff defended this cause If the temporal dimension is also closed, as in (56a) where the progressive verbal form blocks the generic reading, then we record a difference in the function of nessuno. In (56a), it adds emphatic force, and the sentence is used felicitously as a reply. In (56b), which is a verbless clause, it is interpreted as a quantifier with its presupposition of non-uniqueness, and it is awkward. (56) a.

Nessun papa sta difendendo questa causa no pontiff is defending this cause b.??*Nessun papa per questa causa no pontiff for this cause





L. M. Tovena

The felicity of (56a) suggests that the presupposition of non-uniqueness conveyed by nessuno, what may also be called its partitive reading, is a pragmatic constraint, possibly triggered by the verbless construction. It is interesting to note that (19), the other instance of emphatic interpretation of nessuno we have met in this paper, also occurs in a tensed clause. Finally, niente is selected for non referential uses and event names. For instance, the issue in (57) is whether the person who is asked has or hasn’t an instance of authorisation, knowing that any instances would do. The question in the negative form is asked if no authorisation is produced (about presuppositions in negative questions see Ladd 1981; Gutiérrez-Rexach 1996), but only niente can be used felicitously in this context. (57) A policeman asking people at the entrance of a building under surveillance: a. ??Nessun permesso? none of the authorisations? b. Niente permesso? don’t you have the authorisation? In (58) and (59) too the effect of the use of niente is to block any specific reading of the noun. (58) a.

Decisione d’un parroco. Gay in chiesa: niente Messa (13/01/1997ICDS) the decision of a priest: if there are gay people in the church, there won’t be Mass

(59) a.

Il Papa è malato, niente udienza the Pope is ill, no audience

(06/02/1997LS)

Niente-phrases can be perceived as emphatic in the sense that the denial of existence that they can express may be used to convey an interdiction. Nessuno-phrases are preferred for expressing contingent information. This difference is highlighted by the minimal pair presented in (60). Sentence (60a) reports Cernomyrdin’s successful opposition to Lebed’s getting an office of vice-president in Russia. Recall that the office of vice-president does not exist in Russia. On the contrary, (60b) just says that the (existing) presidency of the committee has not (yet) been assigned. (60) a.

Cernomyrdin a Lebed: niente vicepresidenza (05/07/1996LU) Cernomyrdin [said] to Lebed: no vice-presidency b. Commissione di Vigilanza: nessuna presidenza (06/08/1997RR1) [as for the] Commissione di Vigilanza: no presidency [has been assigned]

Distributional restrictions on negative determiners

.. Double readings As we have seen, the substitution of nessuno for niente, and vice versa, may produce a switch between two possible interpretations, namely an interpretation of the noun as countable or uncountable. A regular feature in this alternation is the constant pairing of niente with the uncountable reading, as shown in (61a), and of nessuno with the countable reading, as shown in (61b). (61) a.

Niente formaggio alla fine del pasto no cheese at the end of the meal b. Nessun formaggio tra i cibi vietati no type of cheese among the food that has been banned

The difference in the reading associated with the noun has consequences on the interpretation of the whole clause. If we replace the determiners in (61) as done in (62), this causes a change in the interpretation that is not always easily rendered by the glosses. The English gloss of (62a) is slightly imprecise, because the example is not really about types of cheese, rather it is about instances of cheese, which possibly but not necessarily correspond to different types. While (61a) is used to exclude the presence of cheese, (62a) is used to report on its absence. Conversely, while (61b) reports the fact that the interdiction of eating certain instances/types of food does not concern cheese, (62b) forbids the interdiction of eating cheese. (62) a.

Nessun formaggio alla fine del pasto no type of cheese at the end of the meal b. Niente formaggio tra i cibi vietati no type of cheese among the food that has been banned

The restrictions coming from the determiners are so strong that, when they are not met, they may cause the discourse to become incoherent. So, only nessuno, that reports the absence of any degree/type of fear conceived as possible in a given situation, is compatible with the continuation proposed in (63). On the contrary, niente, that seems to negate the type itself, is unfelicitous in a contingent statement. (63) a.

nessuna paura tra gli ostaggi, l’operazione si è svolta con ordine no fear among the hostages, the operation progressed in an orderly way b. niente paura tra gli ostaggi, *l’operazione si è svolta con ordine

The interaction between the lack of existential import proper to niente-phrases and the properties of uncountable abstract nouns such as names of feelings gives rise to another phenomenon, discussed in the following subsection.





L. M. Tovena

.. Uncountable In Sections 3 and 4 the restricted possibilities for nessuno to combine with uncountable nouns were discussed. It has appeared that only intensive quantities are suitable. On the contrary, niente can combine with all sorts of uncountable nouns. When discussing (63), it was noted that names of feelings are a subset of abstract uncountable nouns with which both items can combine. It was also noted that the interpretation of the combination differs, in particular the different presuppositions associated with the items are satisfied in different contexts. Nientephrases have an intensional flavour that tends to make them infelicitous in contingent or episodic statements, where the entities in the restriction of the generalized quantifier are presupposed to exist or a specific spatio-temporal location is identified. In other words, the requirements coming from the item must not mismatch with the constraints coming from the interpretation of the predicate or VP. A better continuation for niente paura than the one in (63) is provided in (64), where the event described by the second clause is distinct. Actually, in (64) the first clause does not describe an eventuality, rather it expresses an exhortation, as captured by the imperative negative in the gloss. This wish-form applies to the situation holding at utterance time. (64) niente paura, sta arrivando un mediatore della nostra ambasciata do not be afraid, a mediator from our embassy is on her way Examples of this exhortative function of niente-phrases are frequent, cf. (65). (65) Niente paura Lazzeretti, la missiva craxiana ha raggiunto tutti i 108 delegati al consiglio generale dell’Internazionale. (23/1/1997IM) Do not be afraid Lazzaretti, the message from Craxi reached all the 108 delegates in the general council of the International Going back to example (43), the possibility of replacing niente with nessuno in this particular case, with no relevant shift in meaning, can be ascribed to the fact that the nessuno-phrase describes a state that also holds at utterance time. The difference between the two items is mainly a matter of whether the hearer/addressee is included in the set of people being surprised, as in the case with niente, or the inclusion is left unspecified, as in the case with nessuno. (43) Nessuna meraviglia quindi se il discorso ha mandato su tutte le furie i repubblicani [. . . ] Then no surprise that the speech infuriated the republicans Accordingly, (66a) is interpreted as providing an account of facts and (66b) describes the absence of feelings of sorrow by the terrorist.

Distributional restrictions on negative determiners

(66) a.

Un microscopico mondo popolato da feroci clan: nessuna pietà per i nemici, ma dedizione totale alla propria colonia. (31/7/1997IM) a microscopic world inhabited by fierce clans: no mercy towards the enemies, but total dedication to their own colony b. Nessun pentimento per Noam Friedman, l’attentatore israeliano. (3/7/1997IM) no regret by Noam Friedman, the Israeli terrorist

Finally, as a last difference between the two items with respect to the spectrum of uncountable nouns, let us recall that niente can combine with concrete mass nouns, cf. (67), while nessuno cannot, cf. (16c). (67) a.

niente latte no milk b. niente posta no mail

. Second partial conclusions If we consider the distribution of nessuno and niente from the point of view of the constraints imposed by the domain of quantification, we can recall the following points. –

–

–

When the structure of the domain does not contain atoms, or they are not clearly individuated, the use of niente allows the possibility of expressing a negation on the whole structure via a ‘negator of essence’ (Horn 1989). The expression can be interpreted as future oriented and having nomic character. It can have a modal flavour and it offers the possibility of eliminating the presupposition of existence that usually characterises the first argument of a generalised quantifier. Actually, niente allows the possibility of linking the ‘existence’ of entities in the restrictor with the will of the speaker. With respect to the same domain, the use of nessuno forces individuation, or at least suggests the possibility of a spatio-temporal localisation of discrete entities. Otherwise it prompts a taxonomic reading, which also implies reference to a set of discrete entities, as with concrete uncountable nouns. The ‘degree/quantity’ reading available with intensive quantities might fall in the same group, in the sense that degrees are closer to species than to extensive quantities. When the structure of the domain contains atoms, the use of nessuno allows the possibility of individuating them. This may motivate the contingent character of this type of quantification, and its use in many cases describing eventualities





L. M. Tovena

located in the past. In many cases, it gives the clause an orientation towards the past even in the absence of overt information. If we consider the constraints that seem to come from the determiners, we can recall the following points. –

–

nessuno seems to suggest a cardinality > 1 of the domain, i.e. in verbless clauses it has a partitive reading. From there it may follow the presupposition of existence that is particularly evident in verbless clauses, and the sensitivity to the presence of weakly or strongly discrete entities in the quantificational domain. niente does not seem to impose particular constraints. In return, it allows a greater freedom with respect to the presuppositional status of the entities in the quantificational domain.

In the light of this short recapitulation, we can look at the mixed sequence contained in example (68). (68) Berlusconi invece [. . . ] ha annunciato “vado ai Caraibi”. Ma niente bermuda alle Bermuda, niente Galliani, Confalonieri e compagnia in calzoncini corti, come in passato. Niente jogging con i collaboratori più stretti. Nessun rischio di fotografi nelle vicinanze. (28/12/1996LS) ‘Instead, Berlusconi announced “I go to the Caraibes”. But no bermuda shorts at the Bermuda, no Galliani, Confalonieri and the company with short trousers, as in the past. No jogging with the closest members of his team. He won’t run the risk of finding photographers nearby’ Niente-phrases can have a volitional flavour. The items in (68), be they things or people, are ruled out by Berlusconi himself. Their presence is intentionally barred as a consequence of the will of the speaker, all of them except for the last one which is beyond his direct control. In this particular case, the entity is introduced by nessuno.

. Summing up In this paper, we have discussed restrictions in the distribution and interpretation of negative determiners, in particular English no, Italian nessuno and French aucun, and determiner-like uses of Italian niente. The typology of determiners defined by Chierchia (1998) has been shown to be too coarse-grained for an accurate classification of negative determiners. In the case of no, evidence has been provided in support of the relevance of morphological number in its interpretation and use, while Chierchia acknowledges this relevance just for the.

Distributional restrictions on negative determiners

Then, it has been shown that some determiners are sensitive to differences among members of the class of uncountable nouns, and that the distribution of a number of singular determiners cannot be captured by restricting their domain of quantification to all and only singular countable nouns. The notion of intensive quantity (Van de Velde 1996) is used in Tovena (2001) to characterise the subset of uncountable nouns with which these determiners can combine. Finally, the behaviour of nessuno in verbless clauses has been compared with determiner-like occurrences of niente. Although both items could be defined as generalised quantifiers imposing a constraint of empty intersection between their first and second arguments, this characterisation is not sufficient. There are other features of their interpretation that must be captured if one wants to account for the differences in their use. In particular, they seem to carry different presuppositions on their quantificational domain.

Notes . In short, this relation covers the two cases of ‘greater or equal’ and ‘part of ’. It is used to define the operation of ‘sum’ (union or join). . The same form can be used also as a quantifier, cf. (i). (i)

Nessuno ha parlato. nobody spoke

. This form can be used also as a quantifier, in which case it is usually followed by a partitive, cf. (i). (i)

Aucun de nous n’a parlé. none of us spoke

It is analysed as anaphoric, a point made to explain the unacceptability of (ii) when uttered out of the blue. (ii) #Aucun n’a parlé. nobody spoke . For a discussion of the polarity sensitivity and free-choiceness of this item, see Tovena and Jayez (1999); Jayez and Tovena (2000). . Van de Velde borrows this term and the notion it names from Kant’s Critique of pure reason. The English rendering given in the text comes from F. Max Müller’s translation dated 1881. . We can call them DPs, the difference is not relevant for the issue under discussion.





L. M. Tovena

References Barwise, J., & Cooper, R. (1981). Generalized quantifiers and natural language. Linguistics and Philosophy, 4, 159–219. Bosveld-de Smet, L. (1998). On Mass and Plural quantification. Ph.D. thesis, University of Groningen. Chierchia, G. (1998). Plurality of mass nouns and the notion of ‘semantic parameter’. In S. Rothstein (Ed.), Events and Grammar (pp. 53–103). Dordrecht: Kluwer. Corblin, F., & Tovena, L. (2001). On the multiple expression of negation in Romance. In Y. D’Hulst, J. Rooryck, & J. Schroten (Eds.), Romance Languages and Linguistic Theory 1999 (pp. 87–115). Amsterdam: John Benjamins. Delfitto, D. (1998). When plural does not mean plural: Some remarks on ‘notional’ and ‘grammatical’ plurality. In Recherches de linguistique française et romane d’ Utrecht XVII (pp. 26–39). OTS. Gutiérrez-Rexach, J. (1996). Negative polarity licensing and the rhetorical interpretation of questions. In Proceedings of the Western States Conference on Linguistics. Horn, L. R. (1989). A natural history of negation. Chicago: Chicago University Press. Jayez, J., & Tovena, L. M. (2000). Strategies of Free–Choiceness. Ms., EHESS and Université de Lille. Kant, I. (1881). Critique of Pure Reason (translated by F. Max Müller). London: MacMillan. Ladd, R. D. (1981). A first look at the semantics and pragmatics of negative questions and tag questions. In Seventeenth Regional Meeting of the Chicago Linguistics Society. Löbner, S. (2000). Polarity in natural language: Predication, quantification and negation in particular and characterizing sentences. Linguistics and Philosophy, 23, 213–308. Manzotti, E. (1991). La negazione. In L. Renzi & G. Salvi (Eds.), Grande grammatica italiana di consultazione, Vol. 2, Ch. 5 (pp. 245–318). Bologna: il Mulino. Molinelli, P. (1988). Fenomeni della negazione dal latino all’italiano. Firenze: La nuova Italia. Schwarzschild, R. S. (1991). On the meaning of definite plural phrases. Ph.D. thesis, University of Massachusetts. Szabolcsi, A. (1997). Ways of scope taking. Dordrecht: Kluwer. Tovena, L. M. (1996). Studies on polarity sensitivity. Ph.D. thesis, University of Edinburgh, published by Garland 1998. Tovena, L. M. (2001). Between mass and count. Proceedings of WCCFL 20 (pp. 565–578). Cascadilla Press. Tovena, L. M., & Jayez, J. (1999). Any: From scalarity to arbitrariness. In F. Corblin, J.-M. Marandin, & C. D.-Sorin (Eds.), Empirical issues in formal syntax and semantics II (pp. 39–57). The Hague: Holland Academic Graphic. Van de Velde, D. (1996). Le spectre nominal. Paris: Peeters.

TSL[v.20020404] Prn:29/08/2002; 15:24

F: PBNS02.tex / p.1 (29)

Towards a comprehensive view of Negative Concord João Peres Faculdade de Letras da Universidade de Lisboa

.

Refreshing the notion of Negative Concord

The nuclear fact that characterises the notion of negative concord (henceforth, NC) is illustrated in the sentences in (1): (1) a. Pietro *(non) ha visto nessuno. b. O Pedro *(não) viu ninguém. c. Pedro *(no) vio a nadie. ‘Peter not has seen/saw (at) nobody’ Peter saw nobody/didn’t see anyone.

[italian] [portuguese] [spanish]

The salient phenomenon in these Romance data is the co-occurrence of a standard (arguably sentential) negation operator and a distinguished expression (nessuno, ninguém, nadie – English nobody), commonly named “n-word” (after Laka 1990). The well-known peculiarity of such kind of expressions resides in that, in contexts like (2) (2) a. Nessuno (*non) ha visto Pietro. b. Ninguém (*não) viu o Pedro. c. Nadie (*no) vió a Pedro. ‘Nobody (not) has seen/saw Peter’ Nobody saw Peter.

[italian] [portuguese] [spanish]

they can be proven to express by themselves, without any overt negation operator, some sort of negative value, while in contexts like (1), if, following the standard analyses of such sentences, the sentential negation operator is to be assigned its full semantic capacity, then the negative value of the n-word is, so to speak, inert.

TSL[v.20020404] Prn:29/08/2002; 15:24



F: PBNS02.tex / p.2 (30)

João Peres

This is the core notion of what might be called strong negative concord: an expression that is negative in some sense – in the cases under scrutiny, in the sense of involving contradiction and being antiadditive1 – is deprived of its negative value, while keeping its form, even if it involves a (more or less recognisable by present day speakers) negative morpheme. As is well known, this is not the exact pattern found in French, where the weak negative particle (ne) is always required, and, more especially, in Slavic languages or Hungarian, where a strong negation particle always accompanies n-words. Of course, the significance of a notion of NC where the concordant element does not autonomously possess some degree of negative capacity deserves serious reflection. I will skip this issue respecting the variety of patterns covered by the label NC in the literature. At this initial stage, it should also be stressed that more relaxed patterns of NC can be found in several languages (including those being considered here), namely cases where the negative value that justifies concordant elements like n-words is not an overt negative operator or even a negative operator as was the case in the data in (1). Witness examples like the following, which will not be scrutinised in this paper: (3) Credi che si presenterà nessuno? IT ‘believe that himself show-up nobody?’ Do you believe that anyone will show up?

[Rizzi 1982: 126]

(4) Dubito che venga nessuno. IT ‘doubt that comes nobody’ I doubt that nobody comes. (5) A Ana sabe o que fazer melhor do que ninguém. PT ‘the Ana knows the what do better than what nobody’ Ana knows what to do better than anyone. (6) Es la ultima vez que te digo nada.2 SP ‘is the last time that you say1st sg say nothing’ This is the last time I tell you anything. (7) En lugar de habérselo dicho a nadie, te lo deberías haber callado. SP ‘in place of have-it said to nobody, you it should have concealed’ Instead of having said it to someone, you should have kept it secret.

TSL[v.20020404] Prn:29/08/2002; 15:24

F: PBNS02.tex / p.3 (31)

Towards a comprehensive view of Negative Concord

. The extended domain of Negative Concord . The licensees of Strong Negative Concord At least in the restricted group of languages where the patterns in (1) and (2) can be found – which includes Italian, Portuguese and Spanish – NC is a much more widespread phenomenon than what is generally acknowledged in the literature. In fact, different subtypes of this construction can be found. The subsequent data are from Portuguese, but most of the constructions are available in the three mentioned Romance languages, with minor variation.

A. Non-coordinate (implicative) operators – even type A.1. Implicative NP’s/PP’s/AdvP’s/. . . ([not even XP] subtype). This subtype comprises at least different definite or generic NP’s, argument PP’s and different sorts of adverbials. The particular negative construction conveys an implicature that is abundantly described in the literature, namely as regards its positive counterpart (with even). A.1.1. Implicative definite NP’s (8) a.

Nem o Pedro respondeu à pergunta. ‘not-even the Pedro answered to-the question’ b. Não respondeu à pergunta nem o Pedro. ‘not answered to-the question not-even the Pedro’ Not even Pedro answered the question. Este Verão nem em Londres choveu. ‘this summer not-even in London rained’ b. Este Verão não choveu nem em Londres. ‘this summer not rained not-even in London’ This summer it didn’t rain even in London.

[NC version]

(9) a.

[NC version]

A.1.2. Implicative generic NP’s (10) a.

Nem um elefante derrubava este muro. ‘not-even an elephant would-knock-down this wall’ b. Não derrubava este muro nem um elefante. [NC version] ‘not would-knock-down this wall not-even an elephant’ Not even an elephant would knock this wall down.

A.1.3. Implicative NP’s with numerals (11) a.

Nem vinte estudantes foram à festa. ‘not-even twenty students went to-the party’



TSL[v.20020404] Prn:29/08/2002; 15:24



F: PBNS02.tex / p.4 (32)

João Peres

b. Não foram à festa (nem) vinte estudantes. [NC version; ambiguous without nem] ‘not went to-the party not-even twenty students’ Not even twenty students attended the party. (12) a.

Nem vinte estudantes couberam na sala. [simple past; episodic reading 3 ] ‘not-even twenty students fit in-the room’ b. Não couberam na sala (nem) vinte estudantes. [NC version; ambiguous without nem] ‘not fit in-the room not-even twenty students’ Not even twenty students fit in the room.

(13) a.

Nem vinte estudantes cabem na sala. [present tense; generic reading; implicatum: 20 students would fit; inference: possibly, < 20 will fit] ‘not-even twenty students fit in-the room’ b. Não cabem na sala (nem) vinte estudantes. [NC version; ambiguous without nem] ‘not fit in-the room not-even twenty students’ Not even twenty students fit in the room.

(14) a.

Nem vinte homens levantam esta pedra. [present tense; generic reading] ‘not-even twenty men can-lift this stone’ b. Não levantam esta pedra (nem) vinte homens. [NC version; ambiguous without nem; implicatum: 20 men would lift; inference: possibly, > 20 can lift] ‘not can-lift this stone not-even twenty men’ Not even twenty men can lift this stone.

A.1.4. Implicative NP’s with a universal quantifier plus group and generic reading (15) a.

Nem todos os membros desta equipa levantariam esta pedra. [ambiguity DR/GR] ‘not-even all the members of-this team would-lift this stone’ b. Não levantariam esta pedra (nem) todos os membros desta equipa. [GR; preferred NC; implicatum: all the members of the team together would lift] b . Não levantariam esta pedra (*nem) todos os membros desta equipa. [DR; *NC; implicatum: all the members of the team would each lift] ‘not would-lift this stone not-even all the members of-this team’ Not even all the members of this team would lift this stone.

TSL[v.20020404] Prn:29/08/2002; 15:24

F: PBNS02.tex / p.5 (33)

Towards a comprehensive view of Negative Concord

A.2. Implicative sentences (a subgroup of concessive conditionals – even if type) (16) a.

Nem que me matem eu conto tudo o que sei. ‘not-even that kill3rd pl me I tell all the what [I] know’ b. Eu não conto tudo o que sei, nem que me matem. [NC version] ‘I not tell all the what [I] know not-even that kill3rd pl me’ I won’t tell everything I know, even if they kill me.

B. Disjunctive coordinate structures (arguably NP, PP, AdvP, . . . coordinations) (17) a.

Nem o Pedro nem a Ana foram à festa. ‘neither the Pedro nor the Ana went to-the party’ b. Não foram à festa (nem) o Pedro nem a Ana. [NC version] ‘not went to-the party (neither) the Pedro nor the Ana’ Neither Pedro nor Ana went to the party.

. Identifying the class of expressions licensed under Strong Negative Concord Table 1 groups all the concordant NP types presented in the previous data, with a cross-reference to the sentences. Since I assume as an empirical evidence that this Table exhausts the set of subclasses of NP’s that enter in strong negative concord,4 the obvious question to be raised at this point concerns the characterisation of the group. From direct observation, the first striking fact is the absence of a number of NP types containing a negative element, namely, among others, those formed with the equivalents of not many, not few, not less than n, not more than n and not all (in the latter case, in the distributive reading usage) – which are frequently called Table 1. NP’s involved as concordant elements in NC in Portuguese5

a b c d e f g h i

Portuguese structures

English equivalents

nenhum nem [o]DEF N nem [o/um/os N]generic [nem n N]dr [nem n N]gr/episodic (1st arg. of fit V’s) [nem n N]gr/generic (1st arg. of fit V’s) [nem n N]gr/generic (1st arg. of lift V’s) [nem todos os N]gr (1st arg. of lift V’s) nem o X nem o Y

no not even [the] + N/Proper Noun not even [[the/a] + N]/bare plural not even n not even n not even n not even n not even all neither X nor Y

cf. (20) cf. (8)–(9) cf. (10) cf. (11) cf. (12) cf. (13) cf. (14) cf. (15) cf. (17)



TSL[v.20020404] Prn:29/08/2002; 15:24



F: PBNS02.tex / p.6 (34)

João Peres

“negated phrases”. In Horn’s terms applied to not many, these are forms of “special negation, despite their association with the semantics of contradictory opposition” (1989: 507). In the face of this partition, three questions, which I cannot address here directly, deserve a systematic scrutiny, since, to my knowledge, they have not yet been considered together in an integrated manner: a.

Given some language, which quantifying expressions can become negative (that is, acquire the status of n-words) and/or negated? b. What is the status, in terms of degree of negativity, of the Italian neanche, Portuguese nem, and Spanish ni (English not even) subclass and the Italian non, Portuguese não, and Spanish no (English not) type of constructions? c. Which are the relevant syntactic differences between the two groups (part of the answer presumably being that the not type always corresponds to constituent negation)?

Engaging in a more modest endeavour than trying to answer these questions, I will just attempt to draw a semantic dividing line between the expressions that enter NC chains as licensing elements6 and those that don’t, in the languages under analysis. Before that, let me point out two interesting syntactic facts. The first is that the two classes are in complementary distribution with respect to their occurrence as verb complements in the presence or the absence of standard negation: (18) a.

O Pedro (*não) escreveu não mais de cinco/não muitos artigos. [non-licensee] ‘the Pedro wrote not more of five/not many papers’ Pedro wrote not more than five/not many papers. b. O Pedro *(não) escreveu nem cinco artigos. [licensee] ‘the Pedro not wrote not-even five papers’ Pedro wrote not even five papers.

In contrast to this absolute syntactic discrepancy, the second interesting fact is that some of the NC licensees are totally incompatible with a computable negation operator to their right – cf. (19a–b) – while others share this possibility with non-licensees – cf. (19c–f): (19) a. *Nenhum estudante NÃO foi à festa. ‘no student not went to-the party’ b. *Nem o Pedro nem a Ana NÃO foram à festa. ‘neither the Pedro nor the Ana not went to-the party’ c. Nesta festa, nem o Pedro NÃO dançou. ‘in-this party, not-even the Pedro not danced’ At this party, not even Pedro didn’t dance.

[licensee] [licensee] [licensee]

TSL[v.20020404] Prn:29/08/2002; 15:24

F: PBNS02.tex / p.7 (35)

Towards a comprehensive view of Negative Concord

d. Nem cinco estudantes NÃO foram à festa. ‘not-even five students not went to-the party’ Not even five students didn’t attend he party. e. Nem todos os estudantes NÃO foram à festa. ‘not all the students not went to-the party’ Not all the students didn’t attend the party. f. Não poucos estudantes NÃO foram à festa. ‘not few students not went to-the party’ Not few students didn’t attend the party.

[licensee]

[non-licensee]

[non-licensee]

It is quite obvious that at least the syntactic property just considered doesn’t constitute solid grounds for defining the two classes of operators we have been observing. I will not tackle these issues here, concentrating instead on some semantic properties that present themselves as candidates for characterising the class of NC licensees. The first semantic property that comes to mind is decreasing monotonicity, a property that became of paramount importance in the treatment of negation related issues in particular after the work of Ladusaw (1979). In fact, the NP’s that are licensed in NC involve quantifying operations that result in left downward entailment (that is, DE with respect to the first argument of the quantifier), except for the special case of definite descriptions, which can only trivially be so considered. In order to facilitate the computation of the monotonicity of the classical n-words, a sentence like (20) below, where a determiner n-word (nenhum/nessun/ningún – English no) is used instead of a full NP (ninguém/nessuno/nadie – English nobody) should be observed instead of (2) above: (20) O Pedro não leu nenhum livro. ‘the Pedro not wrote no book’ Pedro didn’t write any book. The uniformity of left monotonicity could give support to Dowty’s (1994) idea that NC acts as a special marker for DE positions, to which I adhered in Peres (1997). However, it can and has been objected that several DE positions, namely those that are not induced by a standard negation operator, are not assigned any special marking. Accordingly, it seems more reasonable to think of left DE as a necessary but not a sufficient condition for NC licensing. The second noticeable semantic fact concerns the logical properties that have been associated with the different manifestations of negation and the degrees of negativity. In this respect, the first important verification is that both the expressions in Table 1 and those that reject NC – namely, expressions like not many, not few, not less than n, not more than n and not all (in DR) – are associated with con-



TSL[v.20020404] Prn:29/08/2002; 15:24



F: PBNS02.tex / p.8 (36)

João Peres

tradiction, inasmuch as all of them play a crucial role in the generation of the contradictory of another proposition. Accordingly, this feature, exactly like decreasing monotonicity, is not responsible for concordant elements in NC, other ingredients of negation having to be scrutinised. The most obvious candidates are the properties that De Morgan’s laws reveal about negation. A standard negative operator of the kind of not and its equivalents is, as is well-known from the literature, antimorphic, which means that it constitutes a function f obeying the two De Morgan laws generalised from set theory and the propositional calculus. This is their set-theoretical formulation: (21) a. f (X ∪ Y) ↔ f (X) ∩ f (Y) b. f (X ∩ Y) ↔ f (X) ∪ f (Y)

(anti-additivity) (anti-multiplicativity)

In Table 2, the two groups of expressions under scrutiny are classified, in columns 2–4, with respect to monotonicity and the two above properties, regarding both arguments of the quantifying expression. The group of expressions in lines 8–12 including only non-licensees of NC, once again it becomes indisputable that in neither argument decreasing monotonicity is a distinctive factor in the matter. As for anti-additivity and anti-multiplicativity, it is also observable that the first argument position is totally unrevealing. However, the same cannot be said about anti-additivity in the second argument: except for the vexing case of nem n N (not even n N) in distributive reading, all the licensed n-phrases are anti-additive in the second argument, which means that, besides “their association with the semantics of contradictory opposition” (Horn 1989: 507), they hold a considerable share of the logical properties of what Zwarts (1986) termed classical negation (epitomised by the not and without sort of operators). To this group of weaker (than classical negation) negative operators, the same author assigned the label regular negation. Table 2. Negation related properties of (Portuguese) negative and negated n-phrases

nenhum N nem o X nem o Y nem [o]DEF N nem [o/um/os N]GENERIC [nem n N]DR [nem n N]GR [nem todos os N]GR [nem todos os N]DR não muitos (not many) N não poucos (not few) N não mais de (not more than) n N não menos de (not less than) n N

Monot.

Anti-addit.

Anti-mult.

Anti-verid.

↓ MON ↓ n.a. MON ↓ Ø MON ↓ ↓ MON ↓ ↓ MON ↓ ↓ MON ↓ ↓ MON ↓ ↑ MON ↓ Ø MON ↓ Ø MON ↑ ↓ MON ↓ ↑ MON ↑

+ A-ADD + n.a. A-ADD + – A-ADD + + A-ADD + – A-ADD – – A-ADD + – A-ADD + – A-ADD – – A-ADD – – A-ADD – – A-ADD – – A-ADD –

– A-MULT – n.a. A-MULT – – A-MULT + – A-MULT – – A-MULT – – A-MULT – – A-MULT + – A-MULT + – A-MULT – – A-MULT – – A-MULT – – A-MULT –

+ A-VERID + A-VERID + A-VERID + A-VERID + A-VERID + A-VERID + A-VERID – A-VERID – A-VERID – A-VERID – A-VERID – A-VERID

TSL[v.20020404] Prn:29/08/2002; 15:24

F: PBNS02.tex / p.9 (37)

Towards a comprehensive view of Negative Concord

We can therefore state that, except for the already mentioned exception, regular negation operators are licensees in NC. Giannakidou 1998 has brought considerable attention to the properties of veridicality, non-veridicality and anti-veridicality in relation with the licensing of the expressions that have for long been called “negative polarity items”, showing that this kind of expressions can be sensitive to that variety of values. The value that matters here is anti-veridicality, which grosso modo is the property of a context where some proposition is assumed to be false by some entity relevant for the discourse, either directly, by means of an overt antiveridical operator (like classical negation) or indirectly, by implicature. It is the latter case that may shed some light on the picture given in Table 2. In fact, all the operators in lines 3–7 of the Table are implicative in nature, each conveying the denial of an implicit proposition (for instance, not even twenty people attended the meeting, carries the implicatum that at least twenty people would attend the meeting, which is being denied, the same going for the equivalents in the NC languages being considered). Given the account of the relevant facts just sketched, it now becomes quite clear why nem todos os N (English not all N, in distributive reading), contrary, for instance, to nem n N (English not even n N), is not licensed in a NC chain: (22) a.

Nem todos os estudantes foram à festa. ‘not all the students went to-the party’ b. Não foram à festa (*nem) todos os estudantes. [*NC version; ambiguous without nem] ‘not went to-the party not twenty students’ Not all the students attended the party.

In fact, this operator not only lacks the apparently necessary condition of decreasing monotonicity in its first argument – which precludes the emergence of a stronger property of negation like anti-additivity – but also lacks the either direct or implicative anti-veridical effect that other operators exhibit. The claims that were just made should not be interpreted as sustaining the idea that only under anti-veridicality – verified in the contexts analysed here – can n-phrases occur in the languages under consideration. This is not the case, as I have tried to prove in my 1998 Salford paper and in Peres (2000), where I paid close attention to the cross-sentential licensing of n-words, trying to prove that Portuguese n-words are endowed with a multifactorial sensitivity, an idea that, at least in my reading of her work, plays a central role in the view of the field that was offered in Tovena (1998).



TSL[v.20020404] Prn:29/08/2002; 15:24



F: PBNS02.tex / p.10 (38)

João Peres

. Mutual licensing of n-phrases It is a well-known fact that classical n-words (of the Italian nessun, Portuguese nenhum, and Spanish ningún – English no – type) and n-phrases (of the Italian nessuno, Portuguese ninguém, and Spanish nadie – English nobody – or Italian niente, Portuguese and Spanish nada – English nothing – type, etc.) are mutual licensers. Having now contemplated a wider domain of n-phrases, the question is in order whether or not all the operators that engage in a NC chain headed by anti-morphic negation can license each other. The observation of the data leads to the following description: A. In general, classical n-phrases can act as a licensers of one another, as is abundantly documented in the literature: (23) a. a . b. b . c. c .

Nessuno ha fatto niente. IT Nessun studente conosce né Parigi né Londra. Ninguém fez nada. PT Nenhum estudante connhece (nem) Paris nem Londres. Nadie hizo nada. SP Ningún estudiante conoce (ni) Paris ni Londres. Nobody did anything. No student knows Paris or London.

B. Disjunctive n-phrases, exemplified above in (17), operate as NC licensers just like classical n-phrases: (24) a. a . b. b . c. c .

Né Pietro né Maria hanno letto nessun libro. IT Né Pietro né Maria sono stati né a Parigi né a Londra. Nem o Pedro nem a Maria leram nenhum livro. PT Nem o Pedro nem a Maria estiveram (nem) em Paris nem em Londres. Ni Pedro ni Maria han leído ningún libro. SP Ni Pedro ni Maria estuvieron (ni) en Paris ni en Londres. Neither Peter nor Mary read any book. Neither Peter nor Mary were in Paris or London.

C. Definite n-phrases are in general accepted as licensers, apparently with some variation: (25) a.

Neanche Pietro há fatto niente per aiutarmi/avveva letto *nessun libro. IT ‘not-even Pietro has done nothing for help-me/had read no book’

TSL[v.20020404] Prn:29/08/2002; 15:24

F: PBNS02.tex / p.11 (39)

Towards a comprehensive view of Negative Concord

b. Nem (sequer) o Pedro fez nada para me ajudar/tinha lido nenhum livro. PT ‘not-even the Pedro did nothing for me help/had read no book’ c. Ni (siquiera) Pedro hizo nada para ayudarme/leyó ningún libro. SP ‘not-even Pedro did nothing for help-me/read no book’ Not even Peter did anything to help me/(had) read any book. D. Also group reading implicative operators are in general accepted as licensers, again apparently with some variation. (26) a.

Nemmeno cinque uomini potrèbbero sollevare *nessuna/neanche una di queste pietre. IT ‘not-even five men could lift none/not-even one of these stones’ b. Nem cinco homens levantariam nenhuma destas pedras. PT ‘not-even five men would-lift none of-these stones’ c. Ni (siquiera) cinco hombres levantarían ninguna de estas piedras. ES ‘not-even five men would-lift none of these stones’ Not even five men could lift any of these stones.

(27) a.

Nemmeno tutti i lavoratori potrèbbero sollevare *nessuna di queste pietre. IT ‘not-even all the workers could lift none of these stones’ b. Nem todos os operários levantariam nenhuma destas pedras. PT ‘not-even all the workers would-lift none of-these stones’ c. Ni (siquiera) todos mis obreros (juntos) serian capaces de levantar ninguna de estas piedras. ES ‘not-even all my workers (together) would-be able of lift none of these stones’ Not even all my workers together would be able to lift any of these stones.

E. Numeral distributive n-phrases don’t play the licenser role: (28) a. *Neanche due studenti hanno fatto niente per aiutarmi/hanno letto nessun libro. IT ‘not-even two students have done nothing for help-me/have read no book’ b. *Nem dois estudantes fizeram nada para me ajudar/leram nenhum livro. PT ‘not-even two students did nothing for me help/read no book’



TSL[v.20020404] Prn:29/08/2002; 15:24



F: PBNS02.tex / p.12 (40)

João Peres

c. *Ni diez estudiantes hicieron nada para ayudarme/han leído ningún libro. SP ‘not-even two students did nothing for help-me/have read no book’ Not even two students did anything to help me/have read any book. F. Finally, any combination of implicative operators is rejected, that is, nothing of the following format is accepted in any of the languages under scrutiny: (29) not-even X – V – not-even Y (30) *Nem vinte estudantes fizeram nem dois trabalhos. ‘not-even twenty students made not-even two papers’ *Not even twenty students wrote not even two papers. I assume this fact corresponds to a very general – possibly, universal – constraint on the implicative even-constructions at stake. Apparently, and in accordance with all the facts that were discussed before, the only possible explanation for the distribution shown by the data in A–E above is that, in order to be able to license other n-phrases, these expressions must have the capacity to generate a context that shares with classical negation not only a contradictory semantics, but also other logical properties. This is precisely the case with the operators in A–D, which are all anti-additive in their second argument, inside which the licensee is supposed to occur. Consistently, this is not the case with the negated numerals in E, which only carry a weaker negative value, namely, decreasing monotonicity. This is shown in Table 3, where values are marked for n-phrases that act as licensers in Portuguese (and, for that matter, for their equivalents in other languages), regarding both arguments and with respect to monotonicity, anti-multiplicativity and anti-additivity.

Table 3. Negation related properties of (Portuguese) licenser n–phrases

nenhum N nem o X nem o Y nem [o]DEF N nem [o/um/os N]GENERIC [nem n N]GR [nem todos os N]GR [nem n N]DR

Monot.

Anti-addit.

Anti-mult.

↓ MON ↓ n.a. MON ↓ Ø MON ↓ ↓ MON ↓ ↓ MON ↓ ↓ MON ↓ ↓ MON ↓

+ A-ADD + n.a. A-ADD + – A-ADD + + A-ADD + – A-ADD + – A-ADD + – A-ADD –

– A-MULT – n.a. A-MULT – – A-MULT + – A-MULT – – A-MULT – – A-MULT + – A-MULT –

TSL[v.20020404] Prn:29/08/2002; 15:24

F: PBNS02.tex / p.13 (41)

Towards a comprehensive view of Negative Concord

. General conclusion The domain of analysis of the present work is defined by the structures that were classified in Part 1 as instances of “strong negative concord”, that is, those structures where an expression that can per se convey negative values occurs under the licensing effect of another negative expression, whose negative value is the only one to be computed. Two issues were addressed: firstly, which phrases are licensed in a Negative Concord chain (the analysis having been restricted to noun phrases), and, secondly, which noun phrases that are licensed in negative concord chains can assume the role of licensers in the same kind of configuration. The conclusions that were reached can be summarised as follows: (i) as should be expected, all licensees and licensers in NC chains involve a semantics of contradiction; (ii) a strong logical property of classical negation (again, resorting to Zwarts’ 1986 terminology), namely anti-additivity, is a sufficient condition for the fulfilment of both roles, and a necessary condition only for the licenser role; (iii) the licensee role is less constrained than that of licenser, inasmuch as the conveyance of anti-veridicality, either directly or by implicature, acts as a sufficient condition. Crucially, while the licensed elements can have the required negative values realised indirectly, via implicature, licensers must directly communicate strong negative values.

Notes . For an n-word like English no, asserting that this operator is anti-additive is tantamount to stating that the following equivalence (where N1 and N2 stand for any nouns and X stands for any predicative expression) holds (see Zwarts 1995 for deeper analysis): (ii) [no (N1 or N2 ) X] if and only if [no (N1 ) and no (N2 ) X] . To my knowledge, this kind of structure as well as the subsequent one was not yet acknowledged in the literature. I am indebted to León Acosta for having called my attention to them. . If subject to appropriate lexical selection or given an adequate context, all the sentences in (12)–(14) are ambiguous between a distributive (DR) and a group reading (GR). Witness the following clearly ambiguous sentence (given a situation where someone is testing the capacity of a briefcase for containing different folders separately or as a group): (i)

Nem cinco cadernos couberam na minha pasta. ‘not-even five folders fit in-the my briefcase’ Not even five folders fit in my briefcase.

However, since this variation does not impinge on the licensing of NC, it is being skipped here. Quite differently, it is relevant with the universal quantifier – cf (15). . Several authors – among them van der Wouden (1994) – take phrases like almost nobody to be quantifiers. This is a debatable view, to which I prefer considering almost as an operator



TSL[v.20020404] Prn:29/08/2002; 15:24



F: PBNS02.tex / p.14 (42)

João Peres

on propositions. If this trend is adopted, then the concordant element is still just nobody and related expressions. . In lines e–h in Table 1, the reference to specific verbs implies the acknowledgement that a semantic difference of the kind established between upward and downward scalar predicates in Beck and Rullmann (1996) may play an important role in the semantic characterisation of nem (not even) phrases. . I won’t discuss here the status of licensers and licensees, since I assume that in sentences like (1) above the n-word is the licensed element. Accordingly, I reject the idea – advocated in some literature – that the standard negative operator can be a doubled element triggered by the n-word. As I tried to prove elsewhere, such an idea is not compatible with NC chains where a standard negative element occurs, as a licenser, in a sentential domain higher than that of the licensed n-word and cannot be dispensed with in the meaning computation.

References Beck, S., & Rullmann, H. (1996). Degree Questions, Maximal Informativeness, and Exhaustivity. In P. Dekker & M. Stokhof (Eds.), Proceedings of the Tenth Amsterdam Colloquium, December 18–21, 1995 (pp. 73–92). Amsterdam: ILLC. Dowty, D. (1994). The Role of Negative Polarity and Concord Marking in Natural Language Reasoning. In M. Harvey & L. Santelmann (Eds.), Proceedings from Semantics and Linguistic Theory IV (pp. 114–144). Ithaca, NY: DMLL Publications. Giannakidou, A. (1998). Polarity Sensitivity as (Non)Veridical Dependency. Amsterdam: John Benjamins. Horn, L. (1989). The Natural History of Negation. Chicago: University of Chicago Press. Ladusaw, W. (1979). Polarity Sensitivity as Inherent Scope Relations. Ph.D. dissertation, University of Texas at Austin. Laka, I. (1990). Negation in syntax: on the nature of functional categories and projections. Ph.D. dissertation, MIT, Cambridge. Peres, J. A. (1997). Extending the Notion of Negative Concord. In D. Forget, P. Hirschbühler, F. Martineau, & M.-L. Rivero (Eds.), Negation and Polarity, Syntax and Semantics, Selected Papers from the Colloquium Negation: Syntax and Semantics, Ottawa, 11–13 May, 1995 (pp. 289–310). Amsterdam: John Benjamins. Peres, J. A. (1998). On Romance Sensitivity to Non-veridicality. Ms., paper presented to the Negation: Syntax, Semantics and Pragmatics Conference, University of Salford, Manchester, October 30–November 1, 1998. Peres, J. A. (2000). On the Nature and Licensing Conditions of N-phrases in Portuguese. DELTA (J. of the Brazilian Association of Linguistics), 16 (special issue), 165–199. Rizzi, L. (1982). Issues in Italian Syntax. Dordrecht: Foris. Tovena, M. L. (1998). The Fine Structure of Negative Polarity. New York: Garland Publishers. Wouden, T. van der (1994). Negative Contexts. Ph.D. dissertation, Rijksuniversiteit Groningen. Zwarts, F. (1986). Categoriale Grammatica en algebrasche semantiek. Eeen studie naar negatie en polariteit in het Nederlands. Ph.D. dissertation, Rijksuniversiteit Groningen. Zwarts, F. (1995). Nonveridical Contexts. Linguistic Analysis, 25, 286–312.

TSL[v.20020404] Prn:23/08/2002; 12:19

P II

Temporality

F: PBNSP2.tex / p.1 (43)

TSL[v.20020404] Prn:29/08/2002; 15:26

F: PBNS03.tex / p.1 (45)

On temporal constructions involving counting from anchor points Semantic and pragmatic issues Telmo Móia Universidade de Lisboa, Portugal

Introduction This paper focuses on two closely related subtypes of temporal expressions, illustrated in (1) and (2) below. The property that brings them together is the fact that they both involve counting of temporally ordered entities (such as weekdays – Sundays, in the examples below – calendar years or sport seasons, for instance). I will concentrate on Portuguese and English, although parallel expressions in other languages, such as French or Spanish, seem to behave similarly with respect to the issues at stake. (1) a.

O Paulo não vai à igreja há três domingos. “the Paulo not goes to-the church there-is three Sundays” b. Paulo hasn’t been to church for three Sundays (now).

(2) a.

O Paulo não vai à igreja desde há três domingos (atrás). “the Paulo not goes to-the church since there-is three Sundays (behind)” b. Paulo hasn’t been to church since three Sundays ago.

Notice that the two Portuguese sentences above – sentences a – contain the expression há (an inflected present tense verb form, similar to French il y a), differing only with respect to the temporal preposition desde (the counterpart of English since).1 However, their approximate English counterparts – sentences b – symptomatically involve different temporal particles: for in the first case, and a combination of since and ago in the second case.

TSL[v.20020404] Prn:29/08/2002; 15:26



F: PBNS03.tex / p.2 (46)

Telmo Móia

Now, the sentences in (1) and (2) – which describe a negative habitual state – are not equivalent. In a scenario where Paulo went to church on a Wednesday the week before the utterance time, for instance, the statements in (1) would not necessarily be false, whereas those in (2) would. The reason is that the assertions in (1) merely involve the three Sundays that precede the utterance time, while those in (2) involve the whole interval between the utterance time and the third Sunday in its past, including all weekdays (cf. Section 3 for elaboration on this issue). In schema (3), the relevant intervals for the location of the described state, which are discontinuous for (1), are represented by the thicker lines below the time axis: (3)

há três domingos / three Sundays ago, in (2) Sunday3

Sunday2

utterance time Sunday1 há três domingos / for three Sundays, in (1)2 desde há três domingos / since three Sundays ago, in (2)

Given the asymmetry portrayed in (3), I will assume that the relevant temporal expressions in (1) and (2) belong in different semantic categories, Portuguese háexpressions being genuinely ambiguous,3 approximately in the following terms: i.

In sentences like (1a), Portuguese há-expressions are associated with the sum of n periods of the stated type (in this case, three Sundays) in the past of the utterance time, that is, they involve a set of location times, rather than a single location time. In this respect, they are akin to the so-called ‘adverbials of temporal quantification’ like todos os domingos (every Sunday), em três domingos (on three Sundays) and nos últimos três domingos (on the last three Sundays), and directly express what could perhaps be termed location relative to a set of intervals (cf. Móia 2000). In the use at stake, Portuguese há-expressions behave as the counterpart of English for-expressions (like those in (1b)). ii. In sentences like (2a) – or (4a) below – Portuguese há-expressions represent the nth period of the stated type (in this case, the third Sunday) in the past of the utterance time;4 in fact, they are time-denoting expressions, which can be combined – though need not be, as illustrated in (4a) – with a locating preposition like desde to form a (strict) temporal locating adverbial, i.e. a phrase defining a single location time (cf. end of Section 1 for a development on this issue). In the use at stake, Portuguese há-expressions behave as the counterpart

TSL[v.20020404] Prn:29/08/2002; 15:26

F: PBNS03.tex / p.3 (47)

Temporal counting from anchor points

of English ago-expressions (like those in (2b) or (4b)), or before-expressions, in other contexts (such as Paulo had been in church three Sundays before (that)). (4) a.

O Paulo esteve na igreja há três domingos (atrás). “the Paulo was in-the church there-is three Sundays (behind)” b. Paulo was in church three Sundays ago.

As said above, the property shared by the temporal expressions in (1) and (2) is the involvement of an operation of counting temporally ordered entities from an anchor point. Before proceeding to the semantic analysis of these expressions, which is the focus of this paper, it is worth noticing that the same temporal particles – há in Portuguese, for or ago in English – may be used with predicates of amounts of time (e.g. three hours, three months) as their complements, in which case an operation of time measurement (rather than counting) is involved. Witness the following examples: (5) a.

O Paulo não vai à igreja há três meses. “the Paulo not goes to-the church there-is three months” b. Paulo hasn’t been to church for three months (now).

(6) a.

O Paulo não vai à igreja desde há três meses (atrás). “the Paulo not goes to-the church since there-is three months (behind)” b. Paulo hasn’t been to church since three months ago. [odd, or not used]

Sentences (5a) and (6a) arguably contain two different há-homonyms (parallel to those in (1a) and (2a), respectively). However, these two Portuguese sentences – contrary to what is the case with (1a) and (2a) – are equivalent. In Section 2, these facts will be considered, together with the possible oddity of English sentences like (6b).

.

Time-denoting expressions involving counting from anchor points

Time-denoting phrases which define intervals by measurement from an anchor point, like those in (6) (há três meses/three months ago), have often been mentioned in the literature. For instance, Bras (1990) considers a class of “adverbiaux qui désignent la zone temporelle en operánt un report de mesure” (p. 199, my emphasis), and Asher et al. (1995) – referring back to Molinès (1989) and Bras and Molinès (1993) – state that (. . . ) [locating time adverbials] fall into (. . . ) classes, depending on (. . . ) whether or not the identification of the referent depends on the projection of



TSL[v.20020404] Prn:29/08/2002; 15:26



F: PBNS03.tex / p.4 (48)

Telmo Móia

a length of time on the temporal axis (from some given point). For example, for the adverbials il y a 3 jours, 3 jours avant Noël, depuis 3 jours (3 days ago, 3 days after Christmas, for the last 3 days) the measure “3 days” is projected on the temporal axis. (p. 109, my emphasis)

To my knowledge, however, the possibility illustrated in (2), where the same temporal particles occur in constructions involving counting rather than measuring (há três domingos/three Sundays ago), has not been discussed in the literature (in connection with particles of the type at stake). From a descriptive point of view, it is worth noticing that at least three different sorts of entities – to wit: intervals, eventualities and “ordinary” individuals/objects – may be involved in the counting operation associated with these time-denoting expressions (see Table 1). In connection, three different types of expressions may occur, in the relevant position, in combination with the temporal particles under analysis: (common) predicates of times, situational predicates, and object/individual-denoting predicates. In general, the use of this relatively complex type of temporal expressions appears to be particularly constrained, their counterparts with predicates of amounts of time being possibly more common. One fact that seems to favour the use of these expressions (though it is not crucial) is the relevance of the mentioned type of interval in connection with the type of described eventuality (world knowledge obviously interfering). So, for instance, sentences (7) and (8) are possibly less likely to appear with Tuesdays than with Sundays or weekends, in (7) and (8) respectively (or, for that matter, with measure nouns like weeks).

Table 1. Time-denoting expressions involving counting from anchor points Type of counted entity A intervals

B eventualities

C ordinary individuals/objects

Type of relevant com- Illustrative sentences in Portuguese and plement English predicates of times (three weekends/nights/ Sundays/summers. . . ) situational predicates (three classes/elections/ meals/Olympic Games. . . ) individual/objectdenoting predicates (three cigarettes/ papers. . . )

O Paulo esteve nos Alpes há três fins-desemana (atrás). Paulo was in the Alps three weekends ago. O professor começou a analisar a obra de Goethe há três aulas (atrás). The teacher started to analyse Goethe’s work three classes ago. O Paulo está a fumar cigarros uns atrás dos outros. Há três cigarros (atrás) começou a tossir. Paulo is smoking cigarettes one after the other. Three cigarettes ago he started to cough.

TSL[v.20020404] Prn:29/08/2002; 15:26

F: PBNS03.tex / p.5 (49)

Temporal counting from anchor points

(7) a.

O Paulo não esteve na igreja há três {domingos/terças-feiras} (atrás). “the Paulo not was in-the church there-is three {Sundays/Tuesdays} (behind)” b. Paulo was not in church three {Sundays/Tuesdays} ago.

(8) a.

O Paulo teve de trabalhar há três {fins-de-semana/terças-feiras} (atrás). “the Paulo had to work there-is three {Sundays/Tuesdays} (behind)” b. Paulo had to work three {weekends/Tuesdays} ago.

The first group of expressions in Table 1 – that is, those involving (common) predicates of times – essentially involve what some authors term “calendar nouns”, i.e. expressions which represent recursive intervals of the time axis, like parts of the day (afternoons, evenings, nights), days of the week, weekends, months or seasons of the year, for instance. Searching the British National Corpus (henceforth, BNC), numerous instances of these expressions were found – e.g. nights ago, summers ago or Saturdays ago (39, 10, and 5 instances, respectively). For obvious pragmatic reasons, the commonest forms are those with low cardinals (two, three, or four) and vague quantifiers (like a couple of or a few). There are, however, mainly in sport reports, a few cases with higher exact numerals, like the following: (9) a.

“Forty three summers ago, England suffered their greatest humiliation when the likes of Tom Finney, Billy Wright and Stan Mortensen lost 1-0 to the USA in the 1950 World Cup.” (K3A 1256) b. “If those boys in Karachi had found an 8 to hook on the scoreboard those 33 long winters ago, instead of rummaging fruitlessly about on all fours (. . . ).” (CU1 534)

It must be noted that common nouns like semana/week, mês/month or ano/year, when used as calendar terms (and not as measure nouns), form expressions of this type as well. The difference is that – contrary to what happens with nouns like domingo/Sunday or fin-de-semana/weekend – the relevant counted intervals are adjacent. Thus, time-denoting expressions like há três anos/three years ago are ambiguous: they may involve time measurement – just like in (6) – or counting (of calendar years) – just like in (2) – the latter case being illustrated in (11) below: (10) há três anos/three years ago – time measurement – “a moment surrounding the 1095th day (approximately) in the past of the utterance time”; – counting – “the third calendar year in the past of the utterance time.”



TSL[v.20020404] Prn:29/08/2002; 15:26



F: PBNS03.tex / p.6 (50)

Telmo Móia

(11) a.

No ano passado, foram construídos 250 km de novas estradas, em contraste com 110 km há dois anos e 95 km há três anos. “in-the year past, were built 250 km of new roads, in contrast with 110 km there-is two years and 95 km there-is three years” b. Last year, 250 km of new roads were built, in contrast with 110 km two years ago, and 95 km three years ago.

Schematically, the difference can be depicted as follows: (12)

há três anos / three years ago

COUNTING

[ano / year as CALENDAR NOUN]

utterance time

| (calendar) year3 || (calendar) year2 || (calendar) year1 |

|

1095 days há três anos / three years ago

MEASUREMENT

[ano / year as MEASURE NOUN]

As for the cases in line B of Table 1, the relevant situational predicates obviously refer to recurrent eventualities. These eventualities may occur with a (more or less fixed) regular cycle or not – compare, for instance, the differences between Jogos Olímpicos/Olympic Games, refeições/meals, eleições/elections, aulas/classes or viagens/trips, all of which may occur in the relevant environment. The use of this type of complements in the time-denoting expressions under discussion seems to be strongly restricted, although it may take place if an adequate context is supplied, as in the examples given in Table 1, or in the following ones taken from the BNC (where 85 and 6 instances of seasons ago and games ago, respectively, are registered): (13) a.

“Allison, who took over Rovers four games ago after the sacking of Dennis Rofe, has already hauled them off the bottom of the table (. . . ).” (CBG 8775) b. “It was Robson’s return four games ago that helped spark a maximum haul of 12 points.” (CBG 10414)

As for the (apparently, not very frequent) expressions in line C of Table 1, the designated objects are also associated with given intervals of the time axis (in the examples given, cigarettes with the moment they are smoked) and, by way of this association, ordered in time. Obviously, a special context, where the counted objects are involved in some form of temporal ordering, is required, as in the following example from the BNC:

TSL[v.20020404] Prn:29/08/2002; 15:26

F: PBNS03.tex / p.7 (51)

Temporal counting from anchor points

(14) “There had been no boys left to help Jimmy because since five minutes and three hundred pounds ago, three of them had started working for us and a fourth one was lying in the toilet presently not working for anybody.” (H80 1899) Before proceeding to a brief discussion of some restrictions affecting the use of these expressions, two side-notes are in order at this point. The first is that, although I am concentrating here on expressions with há in Portuguese and ago in English, many other phrases (with different particles) belong in this group of timedenoting expressions as well – see (15)–(16) below. In fact, as can be easily seen, the class of expressions under analysis may be subdivided according to different parameters such as (i) the direction of the operation – backwards or forwards in time – and (ii) the type of anchor point – a deictic or anaphoric temporal perspective point, or a time set by an explicit (referentially independent) complement (see Móia 2000, Chapter 7, for a more thorough analysis): –

forward counting or measurement from a deictic temporal perspective point: (15) a.

daqui a três {fins-de-semana/horas} “from-here to three {weekends/hours}” b. three {weekends/hours} from now

–

backward or forward counting or measurement from a time set by an explicit (referentially independent) complement: (16) a.

três {fins-de-semana/meses} {antes/depois} das eleições “three {weekends/months} {before/after} of-the elections” b. three {weekends/months} {before/after} the elections

The second note is to briefly justify why I am classifying all the relevant Portuguese há-expressions and the English ago-expressions as time-denoting phrases, and not as locating adverbials, contrary to the usual categorisation in the literature. As is known, a clear dividing line between time-denoting expressions and temporal locating adverbials is not easy to draw, given that some expressions – like yesterday or last week, for instance – may occur with the same superficial form in the typical contexts of both categories (cf. e.g. yesterday was a nice day vs. John left yesterday). The relevant há- and ago-expressions belong in this group of ambivalent phrases, as illustrated in (17)–(19). The first two of these examples contain occurrences of such phrases in the typical contexts of time-denoting expressions – verb complement position, in (17), and complement of an explicit temporal preposition (até/until), in (18). Conversely, in (19), they surface as ‘full’ adverbials.



TSL[v.20020404] Prn:29/08/2002; 15:26



F: PBNS03.tex / p.8 (52)

Telmo Móia

(17) a.

A reunião estava marcada para há duas horas (atrás). “the meeting was scheduled for there-is two hours (behind)” b. The meeting was scheduled for two hours ago.

(18) a.

O Paulo esteve na sala até há três horas (atrás). (cf. (2a)) “the Paulo was in-the room until there-is three hours (behind)” b. Paulo was in the room until three hours ago. (cf. (2b))

(19) a.

O Paulo saiu da sala há três horas (atrás). “the Paulo left of-the room there-is three hours (behind)” b. Paulo left the room three hours ago.

Now, by postulating a null locating preposition with a value close to that of in, on or at, in these latter contexts – as I advocate in Móia (2000) and assume here – a uniform analysis is achieved, in which the relevant há- and ago-expressions are always taken as time-denoting expressions. Accordingly, (19) will be analysed as follows: (19 ) a. O Paulo saiu da sala Øem há três horas (atrás). b. Paulo left the room Øat three hours ago.

. Some particular restrictions on the combination desde há in Portuguese and since . . . ago in English In this section, I will concentrate on a particular instance of the time-denoting expressions described in the previous section, namely that with the combination desde há in Portuguese, and since. . . ago in English, which seems to be subject to specific restrictions. As was said before, concerning sentences (1) and (2), the difference between há (n periods) and desde há (n periods) in Portuguese, as well as that between for (n periods) and since (n periods) ago in English (where “n periods” stands for a predicate of the groups described in Table 1 above), lies essentially in the relevance of the intervals between the counted periods. Now, this difference vanishes – as was briefly mentioned, in connection with sentences (5) and (6) – in structures where predicates of amounts of time (x-time, henceforth) occur in the same context as “n periods”. This seems due to the fact that continuous intervals are associated with predicates of amounts of time in these contexts. Observe the following sentences (which are of the same type as (5) and (6)): (20) a.

O Paulo está no hospital há dois meses. “the Paulo is in-the hospital there-is two months” b. O Paulo está no hospital desde há dois meses. “the Paulo is in-the hospital since there-is two months”

⇔

TSL[v.20020404] Prn:29/08/2002; 15:26

F: PBNS03.tex / p.9 (53)

Temporal counting from anchor points

(21) a. Paulo has been in the hospital for two months (now). ⇔ b. Paulo has been in the hospital since two months ago.[odd, or not used] The two Portuguese sentences in (20) – the second containing counterparts of since and ago – are equivalent and equally grammatical. The sentences in (21) illustrate a curious fact about English, to wit: that the logically admissible combination of since and an ago-expression with predicates of amounts of time is normally not used and/or accepted. In fact, sentence (21b) is considered odd by most speakers. In order to express the temporal relations at stake, English normally resorts to a forconstruction, as in (21a), which, though formally involving a temporal expression of a different category, conveys the same temporal information. Now, the oddity of the combination since. . . ago in (21b) is somehow unexpected, given that, once we assume (as I do) a time-denoting analysis for ago-phrases, they are expected to freely combine with temporal locating prepositions, such as since or until. As for the combination until. . . ago, it seems unproblematic in English, like in Portuguese for the counterparts of these particles: (22) a.

O Paulo esteve no hospital até há dois meses. “the Paulo was in-the hospital until there-is two months” b. Paulo was in the hospital until two months ago.

As for the combination since x-time ago, its oddity seems, at a closer look, more a matter of use than of strict grammaticality. In fact, the BNC contains several records of this combination: (23) a.

“With exasperation, he said, ‘Since when? Since Christmas?’ Doone said stolidly, ‘Since ten days ago.’ ” (ADY 679) b. “Nothing in the metal’s fundamentals has changed since a month ago, when the price languished at a seven-year low of $126 an ounce.” (CR7 2794) c. “I haven’t played with anyone for like since months ago!” (KE1 805, KE1 1418) d. “Every room here has been booked since a year ago, and I was dearly hoping Donna would screw up the nerve to send her packing (. . . )” (JY6 2805) e. “(. . . ) and they’ve probably been isolated since fourteen thousand years ago, with that every population has gone slightly different to, to the next one.” (F8H 74) f. “(. . . ) I well recall the Second Reading debate of the Shops Bill 1986 which followed upon the discussions and deliberations of the Auld committee which, since as long ago as 1985, had been engaged in a series of debates about Sunday trading.” (HHX 11597)



TSL[v.20020404] Prn:29/08/2002; 15:26



F: PBNS03.tex / p.10 (54)

Telmo Móia

g. “Animals have been anthropomorphised since as long ago as Egyptian times.” (BMD 337) In contrast with these cases, others exist where (according to my informants) the combination of since and ago appears to be generally accepted. This is the case of (2b) above, for instance. As far as I can judge from the information I gathered, the non-problematic cases in English seem to require the concurrence of three facts:5 i.

involvement of discontinuous intervals (e.g. Sundays, weekends, nights, seasons), corresponding to temporal predicates like those in Table 1 above (given that the combination with predicates of amounts of times, which are associated with continuous intervals, is seldom used) – cf. the contrast between since three Sundays ago, in (24a) and since three weeks ago, in (24d) below; ii. relevance of the periods between those intervals (otherwise, a for-construction is preferred) – cf. since three Sundays ago, in (24a) vs. for three Sundays, in (24b); iii. (contextual) salience of the period marking the starting point of the location time, which justifies the choice of the relatively complex expression “n periods ago” (otherwise, a construction with “for x-time” is preferred) – cf. since three Sundays ago, in (24a) vs. for three weeks, in (24c). (24) a.

Paulo hasn’t been to church since three Sundays ago. (This was the last time the priest saw him.) b. Paulo hasn’t been to church for three Sundays (now).[different truthb . Paulo hasn’t been to church for the last three Sundays. conditions] c. Paulo hasn’t been to church for three weeks (now). [relevant Sunday c . Paulo hasn’t been to church for the last three weeks. is not salient] d. (?)Paulo hasn’t been to church since three weeks ago.

When these facts concur, there seems to be no alternative construction to “since n periods ago” that is formed by a single temporal preposition and a time-denoting complement, of the type “prep x-time” or “prep the last x-time”. The nonexistence of such an alternative apparently closely correlates with the acceptance of the combination since. . . ago. The examples given above contain atelic descriptions. In combination with telic descriptions, constructions with since. . . ago behave similarly, the main difference being that the temporal preposition now occurring in the approximately equivalent constructions with predicates of amounts of time is not for, but rather within, in/on or during (together with the last). Compare, for instance, (25) below, which represents a sum of telic events, with (26), which basically differs in that it does not make the mentioned third Sunday salient:

TSL[v.20020404] Prn:29/08/2002; 15:26

F: PBNS03.tex / p.11 (55)

Temporal counting from anchor points

(25) Thirty weddings have been celebrated in this church since three Sundays ago. (This was the day weddings restarted to be celebrated here, after a break of nearly three years.) (26) Thirty weddings were celebrated in this church {within/in/during} the last three weeks. (27) Thirty weddings were celebrated in this church since three weeks ago. [odd, or not used] Likewise, we observe that constructions with “since n periods ago” – as (25) – and those with e.g. “within the last n periods” – as (28) below – are not equivalent, differing – like (2) differs from (1) – in that the latter only involves the mentioned three Sundays and not the intervening periods:6 (28) Thirty weddings have been celebrated in this church {within/on/during} the last three Sundays.

. Semantic-pragmatic differences between constructions involving counting of ordered entities The constructions involving counting of ordered entities of the types illustrated in (1) and (2) exhibit several semantic-pragmatic differences. In this section, I will elaborate a bit more on two of them. One difference, which has already been mentioned several times, concerns the relevance of periods intervening between the counted intervals. The combination with ‘adverbs of temporal quantification’ – illustrated in (29) and (30) – shows that different intervals are involved, in accordance with schema (3): (29) a.

Um avião etíope aterra em Lisboa todos os dias desde há três domingos (atrás). “an aeroplane Ethiopian lands in Lisbon all the days since there-is three Sundays (behind)” b. An Ethiopian aeroplane has been landing in Lisbon everyday since three Sundays ago.

(30) a. *Um avião etíope aterra em Lisboa todos os dias há três domingos. “an aeroplane Ethiopian lands in Lisbon all the days there-is three Sundays” b. *An Ethiopian aeroplane has been landing in Lisbon everyday for (the last) three Sundays.



TSL[v.20020404] Prn:29/08/2002; 15:26



F: PBNS03.tex / p.12 (56)

Telmo Móia

It was also briefly mentioned that if the periods between the discontinuous mentioned intervals are immaterial (world knowledge, for instance, interfering), then constructions with for (in English) and há (in Portuguese) are preferred, this preference being possibly stronger in English than in Portuguese.7 Therefore, assuming – in accordance with current knowledge – that soup eating only takes place during meals, while wine drinking may occur both during and between meals, the contrasts in (31) and (32) are predicted. Sentences in (31) are slightly odd with desde há and since. . . ago, while those in (32) are not. This seems due to the fact that sentences in (32) with desde há and since. . . ago are not equivalent to those with há or for, respectively, whereas in (31) this equivalence exists (given the irrelevance of the periods between meals). (31) a.

O Paulo não come sopa {?desde há duas refeições (atrás)/há duas refeições}. “the Paulo not eats soup {since there-is two meals (behind)/there-is two meals” b. Paulo hasn’t eaten soup {?since two meals ago/for two meals}.

(32) a.

O Paulo não bebe vinho {desde há duas refeições (atrás)/há duas refeições}. “the Paulo not drinks wine {since there-is two meals (behind)/there-is two meals” b. Paulo hasn’t drunk wine {since two meals ago/for two meals}.

A second difference between the constructions like those in (1) and (2) has to do with the pertinence – for counting purposes – of the period containing the utterance time. Take for example the Portuguese expressions desde há três domingos/há três domingos, or their English counterparts since three Sundays ago/for three Sundays, and consider a scenario where the utterance takes place on a Sunday. In the constructions with desde há and since. . . ago, the general consensus among speakers seems to be (despite some hesitation) that the “Sunday of utterance” is never one of the three that are being counted. Thus, sentences like (33) below always count back three entire Sundays (irrespective of whether any landing occurred on the Sunday of utterance or not): (33) a.

Um avião etíope aterra neste aeroporto desde há três domingos. “an aeroplane Ethiopian lands at-this airport since there-is three Sundays” b. An Ethiopian aeroplane has been landing at this airport since three Sundays ago.

Constructions with há and for are different, one of two situations possibly holding: (i) if the “Sunday of utterance” contains (prior to the utterance) an instance of the

TSL[v.20020404] Prn:29/08/2002; 15:26

F: PBNS03.tex / p.13 (57)

Temporal counting from anchor points

events of the type described in the main clause (landings of Ethiopian aeroplanes, here), then it is taken as one of the three that are being counted – see (34), and its possible continuation in a and b . Otherwise, that is, if no landing has yet occurred on that Sunday, by the utterance time, then it is irrelevant for counting purposes – see (34) again, and its continuation in a and b . (34) a.

Um avião etíope aterra neste aeroporto há três domingos. “an aeroplane Ethiopian lands in-this airport there-is three Sundays” b. An Ethiopian aeroplane has been landing at this airport for three Sundays (now).

a . . . . Hoje chegou mais cedo que das outras duas vezes. “today [it] arrived more soon than of-the other two times” b . . . . Today it arrived sooner than the other two times. a . . . . Vamos ver se hoje também aterra e no mesmo sítio das outras três vezes. “let-us see if today [it] also lands and at-the same place of-the other three times” b . . . . Let us see if it also lands today, and at the same place as the other three times.

. Conclusion In this paper, two different temporal constructions involving counting from anchor points – which, to my knowledge, have not been analysed in the literature – were discussed, their semantic and pragmatic specificity being partially explored. Concurrently, we argued for the need to distinguish between two categories of temporal phrases occurring in these constructions (which, interestingly, have homonymous elements in Portuguese): (i) one whose members count the number of contiguous intervals of a given type containing instances of given events, illustrated in (1); (ii) another one whose members are merely time-denoting phrases, illustrated in (2). The semantic and pragmatic differences that were pointed out here should be taken as a contribution to a more thorough study – pending further research – of the vast subclass of anchor-dependent temporal expressions in natural languages, of which they are distinguished instances.



TSL[v.20020404] Prn:29/08/2002; 15:26



F: PBNS03.tex / p.14 (58)

Telmo Móia

Notes . In European Portuguese, the adverbial form atrás (literally ‘behind’) optionally co-occurs with há – without any semantic variance – in sentences of type (2a). In Brazilian Portuguese, there seems to be a growing trend to avoid the verbal form (há) and only use the adverbial atrás. This is more conspicuous in sentences like (2a), where the temporal expression is dependent on an explicit preposition (e.g. desde [since], até [until] or de [from/of ]), than in sentences like (4a) – cf. Móia and Alves (2000). In this paper, only European Portuguese is taken into account. . As will be mentioned in Section 3, the utterance time may overlap “Sunday1 ” in structures like (1), but not in those like (2). . Portuguese time-denoting há-expressions – as those in (2a) and (4a) – contrast with the type of há-expressions in (1a) in at least the following three properties: (i) they optionally co-occur with the adverbial form atrás, (ii) they may be preceded by explicit temporal prepositions, like the counterparts of since and until, and (iii) they combine both with atelic and telic descriptions (whereas those in (1a) only combine with atelics). . I take these expressions to denote, as a whole, the entire nth instance of the relevant period, counting backwards from the anchor point, given the interpretation of sentences like the following: (i)

O número de bilhetes (para este museu) vendidos aos fins-de-semana é o seguinte: 1200 no fim-de-semana passado, 1100 há dois fins-de-semana, 800 há três fins-desemana. “the number of tickets (for this museum) sold on weekends is the following: 1200 in-the weekend past, 1100 there-is two weekends, 800 there-is three weekends”

(i ) The number of tickets (for this museum) sold on weekends is as follows: 1200 last weekend, 1100 two weekends ago, 800 three weekends ago. (ii) 115 aviões aterraram neste aeroporto há três fins-de-semana, o que constituiu um novo recorde. “115 planes landed in-this airport there-is three weekends, the what constitutes a new record” (ii ) 115 planes landed at this airport three weekends ago, which sets a new record. The processing of these sentences requires consideration of all the relevant events (ticketsales or landings) that occurred within a given stretch of time. In these cases, this stretch is obviously a whole weekend (no part of it being irrelevant). . The requirements (ii) and (iii) apply, with the relevant adaptations, to Portuguese constructions with “desde há n periods” as well (although the second one – discussed below, in Section 3 – appears to be less compulsory in Portuguese). . Facts in Portuguese are similar for desde há constructions. See the examples below (where se is a clitic pronoun, marking a passive construction, which is not translated in the glosses):

TSL[v.20020404] Prn:29/08/2002; 15:26

F: PBNS03.tex / p.15 (59)

Temporal counting from anchor points

(i)

Celebraram-se trinta casamentos nesta igreja desde há três domingos (atrás). [like (25)] “Celebrated se thirty weddings in-this church since there-is three Sundays (behind)”

(ii) Celebraram-se trinta casamentos nesta igreja nas últimas três semanas. “Celebrated se thirty weddings in-this church in-the last three weeks”

[like (26)]

(iii) Celebraram-se trinta casamentos nesta igreja desde há três semanas (atrás). [not odd, unlike (27)] “Celebrated se thirty weddings in-this church since there-is three weeks (behind)” (iv) Celebraram-se trinta casamentos nesta igreja nos últimos três domingos. [like (28)] “Celebrated se thirty weddings in-this church in-the last three Sundays” . See, however, the following example from the BNC: “I’ve been coming to games at the Manor since I was 10 (. . . ) since three seasons ago I haven’t missed a match – I’ve been to them all” (ECN 1624).

References Asher, N. et al. (1995). Spatial, Temporal and Spatio-Temporal Locating Adverbials in Discourse. In P. Amsili, M. Borillo, & L. Vieu (Eds.), Workshop Notes of the 5th International Workshop on Time, Space and Movement TSM’95 (pp. 101–119). Bras, M. (1990). Calcul des Structures Temporelles du Discours. Ph.D. thesis, Université Paul Sabatier de Toulouse. Bras, M., & Molinès, F. (1993). Adverbials of Temporal Location: Linguistic Description and Automatic Processing. In Proceedings of the 26th Colloquium of Linguistics. Niemeyer, Tübingen: Linguistiche Arbeiten. Kamp, H., & Reyle, U. (1993). From Discourse to Logic. Introduction to Modeltheoretic Semantics of Natural Language, Formal Logic and Discourse Representation Theory. Kluwer: Dordrecht. Móia, T. (2000). Identifying and Computing Temporal Locating Adverbials with a Particular Focus on Portuguese and English. Ph.D. thesis, Universidade de Lisboa. Móia, T., & Alves, A. (2000). Sobre a Expressão de Distâncias Temporais no Português Europeu e no Português Brasileiro. Actas do XVI Encontro Nacional da Associação Portuguesa de Linguística (Colóquio Português Europeu e Português Brasileiro Unidade e Diversidade na Passagem do Milénio – PEPB-2000) (pp. 699–713). Lisboa: APL. Molinès, F. (1989). Acceptabilité et Acceptation des Adverbiaux de Localisation Temporelle: Grammaire ou Distionnaire. Mémoire de DEA en Sciences du Langage, Université de Toulouse Le Mirail.



TSL[v.20020404] Prn:17/10/2002; 10:36

F: PBNS04.tex / p.1 (61)

On the semantics and pragmatics of situational anaphoric temporal locators in Portuguese and in English Ana Teresa Alves Universidade dos Açores, Portugal

Introduction Several authors, among whom Kamp and Reyle (1993) and Móia (2000), distinguish between two types of temporal expressions: those that directly represent intervals of time, e.g., 2001, next week, last month, today, and those that represent intervals of time indirectly, such as eventuality descriptions like John visited Paris, Mary was hospitalised, or the school will hire a new teacher. In the Kamp and Reyle (1993) framework, the former directly introduce discourse referents for temporal expressions, i.e., discourse referents of type t, while the latter introduce discourse referents for eventualities, that is, discourse referents of type e(vent) or s(tate). To account for the fact that eventuality descriptions also represent time, Kamp and Reyle introduce a function loc, which associates each eventuality with the smallest time interval it occupies. The distinction between these two types of temporal expressions – I’ll dub the former as direct and the latter as indirect – proved to be important in the study of temporal anaphors, inasmuch as some of these constrain their antecedents to be either a direct or indirect temporal expression. In this paper, I’ll focus on anaphoric temporal expressions that refer back to time intervals that are defined by eventuality descriptions, and try to describe their semantic and pragmatic licensing conditions.

TSL[v.20020404] Prn:17/10/2002; 10:36



F: PBNS04.tex / p.2 (62)

Ana Teresa Alves

.

A tripartite classification of anaphoric temporal locators: Direct, indirect, and ambivalent

Regarding the type of temporal expression by means of which anaphoric temporal locators define a location interval, these locators might be sub-classified as direct, indirect and ambivalent, as illustrated in the Table 1. Table 1. Subcategorisation of anaphoric locators wrt. the type of possible antecedent Direct nesse mês (in) that month no mesmo ano/ (in) the same year ...

{depois/antes} disso {after/before} that nessa altura/ at that time até {lá, então}/ until then ...

Indirect durante esse tempo/ during that time entretanto/ (in the) meantime enquanto isso/ (in the) meanwhile ...

examples to be considered in this paper

Direct temporal locators define a location interval by means of direct timedenoting expressions only. On the other hand, indirect locators define location intervals by means of eventuality descriptions only. Ambivalent locators are those that might define location intervals by means of expressions of both kinds. Direct and indirect anaphoric locators are illustrated in appropriate context (1) and (2), respectively. (1) O Paulo chegou a Paris em [Maio de 1999]i . A Maria também chegou a Paris n[esse mês]i . Paul arrived in Paris in [May, 1999]i. Mary also arrived in Paris [that month]i . (2) a.

[O Paulo visitou Paris]i em [1980]*i. Durante [esse tempo]i , a Maria tomou conta do bebé. [Paul visited Paris]i in [1980]*i . During [that time]i Mary looked after the baby. b. [A noite passada]*i o [Paulo corrigiu dez testes]i . Enquanto [isso]i a Maria escreveu duas cartas. [Last night]*i [Paul corrected ten tests]i . [Meanwhile]i Mary wrote two letters.

In (1), the interval associated with esse mês/that month coincides with the time interval associated with the direct temporal expression Maio de 1999/May, 1999. In (2a) and (2b), the location interval associated with durante esse tempo/during that

TSL[v.20020404] Prn:17/10/2002; 10:36

F: PBNS04.tex / p.3 (63)

On the semantics and pragmatics of situational anaphoric temporal locators

time and enquanto isso/meanwhile corresponds, respectively, to the running time of the eventualities ‘Paul visit Paris’ and ‘Paul correct ten tests’; Clearly, the expressions 1980 and a noite passada/last night do not provide appropriate antecedents for the anaphoric expressions. See, for the sake of illustration, the simplified DRSs for sentences (1) – where, besides other simplifications, the contribution of also is ignored – and (2a), respectively. x y t n e e1 t 1 Paul (x)

x y t n e s t1

Mary (y)

Paul (x)

May, 1999 (t)

Mary (y)

e Ty(t), }

(b) Scanning ringo-o ‘the apple-ACC’ The case, -o, indicates it must have a node of Ty(e→t) (a predicate node) as its mother in the final tree. Again the node itself is introduced as unfixed in relation to the root. {? Ty(t)}

{Fo(Ringo), Ty(e), ? <  > Ty(e  t), }

(c) Scanning tabeta ‘ate’ Verbs are a driving force in articulating a tree structure. A two-place predicate such as tabe ‘eat’ creates two argument-nodes and projects a requirement for a VP node of Ty(e→t). A sequence of actions is defined by the verb, with the requirement of Ty(t) at the root node as their trigger: {? Ty(t)} {? Ty(t)} {Ty(e), Fo(Upro)}

{? Ty(e  t)}

{Ty(e), Fo(Vpro)}

{Ty(e (e  t)), Fo(Tabe), }

(d) Merge If nothing follows, then two place-holders, Upro on the subject node and Vpro on the object node in (c), are merged with John-ga and ringo-o respectively. This is the same process as in the regular Left-Dislocation Structure (see (2)).

Japanese ‘gapless’ relatives {? Ty(t)}

John

{? Ty(t)}

Ringo

{? Ty(e  t)}

{Ty(e), Fo(Upro)} MERGE MERGE

{Ty(e), Fo(Vpro)}

{Ty(e (e  t)), Fo(Tabe), }

(e) Compilation The string is compiled at the root node as a proposition:1 {Ty(t), Fo (Tabeta (Ringo) (John)), ◊} ate apple John

. Relatives in English In DS, relatives are analysed as a pair of linked structures: a tree is defined as LINKed to another if it has a copy of an element in the main tree. In English, a relative pronoun introduces the copy in one such LINKed tree. (7) Johni , whoi I like ei , . . .

(I = the speaker = Bill)

(a) a LINK transition – result of processing who {? Ty(t)}

the head

{Ty(e), Fo(John)} {? Ty(e  t)} LINK

the new tree

{? Ty(t)} {Ty(e), Fo(John), }

Linked node Unfixed node

The relative pronoun projects an annotation. The new tree has a requirement ? Ty(t) and a copy of Fo(John) at node n somewhere in the new tree. As the node location for Fo(John) is not known yet, it is represented as an unfixed node as in (a).



 Akiko Kurosawa

(b) Scanning like

(c) Result of processing the relative {? Ty(t)} {? Ty(t)}

{Ty(e), Fo(John)} {? Ty(e  t)} {Ty(e), Fo(John)} {? Ty(e  t)}

LINK

LINK

{? Ty(t)}

Like(John)(Bill),  {? Ty(e  t)}

Bill/Upro

Bill Like {Ty(e), Fo(John)}

Like(John)

Vpro MERGE

Like

John

The rest of the updating process till the unfixed node is identified as the object node of the LINKed tree in (b) is exactly the same as that of the left dislocation case as in 2.3. Then the proposition of relative clause is compiled as (Like(John)(Bill)) as in (c).

. Relative clauses (with a gap) in Japanese As Japanese is a head-final language, the head nominal is preceded by the relative clause. An initially identical string may or may not involve an embedded clause as illustrated in examples (4)–(6). Modern Japanese lacks a relative marker which ensures the LINK transition as in languages such as English (e.g. who) and Arabic (e.g. illi), so each predicate merely shows that there is a disjunction: this is the end of a sentence OR the string is followed by something.2 Then how do we interpret (8) below as containing a LINKed structure? Are all the constituents initially introduced as unfixed nodes, with verbs constructing tree structures, just like a simple sentence in (4)? At which point is the LINK transition made? Consider the following example: (8) John-wa [ei okane-o tukanda] dorobooi -o tukamaeta. John-top money-acc grabbed thief -acc caught ‘John caught the thief who grabbed the money’

Japanese ‘gapless’ relatives

(a) John-wa okane-o tukanda | doroboo-o . . . {? Ty(t)} LINK {Ty(e), Fo(John)}

{Ty(e), Fo(Doroboo) ?<  > Ty(e  t), } ‘thief ’

{Ty(e), Fo(Okane) {? Ty(t)} ?<  > Ty(e  t)} ‘money’ {Ty(e), {? Ty(e  t)} Fo(Upro)}

MERGE

{Ty(e), Fo(Vpro)}

{Ty(e (e  t)), Fo(Tukanda)} ‘grabbed’ MERGE

The above figure illustrates the processing of doroboo-o. The first predicate, tukanda, does not work as a relativiser, as it no longer has a distinction as the adnominal form. The string John-wa okane-o tukanda might be interpreted as a simple sentence, ‘John grabbed the money’. The next lexical input, doroboo, is a good candidate for the head on the basis of our general knowledge in which a thief is associated with an action of grabbing money. However, it is not guaranteed structurally, as doroboo can belong to another embedded clause.3 What indicates that doroboo is interpreted as the head is phonological information. Okane-o tukanda | doroboo has a downstep after the verb, showing this string constitutes a phrase, ‘the thief who grabbed the money’.4 Using this information, the string is interpreted on-line. Phonological information interacts with lexical information in structure building and plays an important role when it reflects complex syntactic structure and has a distinctive function in case the structure is not yet fully specified. The downstep expresses a great deal how two adjacent constituents are related to each other; whether they form one domain or not. In the above case, if doroboo is not the head of a relative clause but a nominal that belongs to another embedded clause, then the downstep is blocked and doroboo is realised as high pitched as tukanda, indicating that this is the beginning of a new domain.5,6 Thus the downstep serves as a disambiguation device.7 So Doroboo is introduced as the head of the relative clause. Okane-o is merged with Vpro on the object node instructed by its case information, and the head, doroboo, provides a value to Upro decorating the subject node, hence providing the shared element in the LINKed tree. In (b) below, after the last constituent, tukamaeta ‘caught’, is scanned, dorobooo is slotted into the sister node projected by the matrix verb. Then the first constituent, John, is finally incorporated into the structure, merging with a placeholder on the matrix subject node.



 Akiko Kurosawa

(b) Scanning tukamaeta {? Ty(t)} {Ty(e), Fo(John)} {? Ty(e  t)}

{Ty(e), Fo(Doroboo) ‘thief ’ LINK

{Ty(e (e  t)), Fo(Tukamaeta), } ‘caught’

{Ty(t), Fo(Tukanda(Okane)(Doroboo/Upro))}

{Ty(e), Fo(Doroboo/Upro)}

{Ty(e  t), Fo((Tukanda)(Okane))} {Ty(e), Fo(Okane)} ‘money’

{Ty(e (e  t)), Fo(Tukanda)} ‘grabbed’

. Gapless relatives – problems for the LINK analysis? In defining LINK structure as having a common element in the structure built from the two clauses, the DS analysis reflects the construal of regular relatives, in which some item needed to complete the interpretation of the relative is taken to be provided by a copy of the head nominal. However there are so-called gapless relatives which do not seem to have an empty node to be fulfilled, and this seems to pose a problem. Gapless relative clauses in Japanese are divided into three categories: a. ‘Pseudo relatives’ b. Appositional Gapless Relatives (‘Content’ clauses) c. Gapless relatives whose head leads an adverbial clause8 This paper focuses on the first one, ‘pseudo relatives’, followed by a brief look at the appositional gapless relatives in 6.6, which head-initial language such as English also have.

. ‘Pseudo relatives’ as gapless relatives ‘Pseudo-relatives’,9 are found in languages like Japanese:10

Japanese ‘gapless’ relatives

(17) [Keeki-ga yakeru] nioi-wa subarasii. cake-nom bake smell-top is-lovely (lit.) ‘The smell that the cake is baking is lovely’ ‘The smell of the cake baking is lovely’ (18) [Seetaa -ni wain-o kobosita] simi-ga oti -nai. Jumper-loc wine-acc spilled stain-nom come_off-neg ‘the stain that (I made by) spilling wine on the jumper doesn’t come off ’ In these examples, there is no ‘gap’ in the relative clause that is co-indexed with the italicised head noun. Contrast (18) and (19): (19) [ ei Seetaa -ni wain-o kobosita] hitoi jumper-loc wine-acc spilled person ‘the person who spilled wine on the jumper’ Teramura (1975–1978/1992) turns his attention to the range of nouns that are typically used as a head of this type of relative clause: ‘perception’ nouns such as nioi, ‘smell’, oto, ‘sound’, kookei, ‘sight’, koe, ‘voice’ and so-called relational nouns such as usiro ‘back’, yokuzitu ‘the next day’, and gen’in ‘cause’. Examples of relational nouns are given below: (20) [Humiko-ga suwatta ] usiro-no mado -ni -wa momizi-ga Humiko-nom sat back-gen window-loc-top maple-nom aokatta. green_was ‘Behind where Humiko was sitting, there were green maple leaves outside the window’ (Yasunari Kawabata Senba-zuru ‘The Thousand Cranes’) (21) [kodomo-ga umareta ] yokuzitu -wa Kurisumasu-datta. Child -nom was_born next_day -top Christmas-was (lit.) ‘The next day that the child was born was Christmas’ (22) [Kazi-ga hirogatta] gen’in-wa hakkiri_si-nai. fire-nom spread cause- top is_clear -neg (lit.) ‘the cause that the fire spread is unclear’ Relative clauses in (20)–(22) express what is the converse of the concept of the head; in (20), while the head noun, usiro, means back, the position where Humiko sat is in the front of the window. In (21), the meaning of the head, yokuzitu, is the next day, and the time when the child was born is the day before. In (22), whereas the head, gen’in, means cause, the fact that the fire spread is the consequence. This relation expressed with gapless relatives is often found in locational and temporal expressions, including nouns such as mae, ‘front’, usiro, ‘back’, tonari,



 Akiko Kurosawa

‘next (position)’, hidari, ‘left’, migi, ‘right’, ue, ‘upper position’, sita, ‘lower position’, mae, ‘earlier time’, ato, ‘later time’, and so on.11 As all these relatives have no gap, a question arises: how do we interpret the relation between the head nominal and the relative clause? In the following subsections from 6.2 to 6.5, I will examine several possible answers to this question.

. The two-place relation analysis The first possible analysis for pseudo relatives is the two-place relation analysis. In examples from (23) to (25), the head noun seems to be in the two-place relation, ‘Y is the . . . of X’, with the NP in the relative clause: X-gen Y (23) keeki-no nioi

Y of X ‘the smell of the cake’

(24) wain-no simi

‘the stain of the wine’

(25) tabako-no kemuri

‘the smoke of the cigarette’

This interpretation involves construing the head nominal as a relation between two arguments. This adds an argument to the representation of the head nominal (e.g. ‘the cake’ is added to ‘the smell’ to show the meaning relation of the relative and the head). Although this analysis applies to a number of cases, it is inadequate to explain some others such as in (26): (26) [Dareka-ga otiba -o taku ] kemuri-ga someone-nom fallen_leaves-acc burn smoke-nom tati_nobotte iru. rising -is ‘The smoke of someone burning fallen leaves is rising’ (27) *otiba-no kemuri ‘the smoke of fallen leaves’ ‘The smoke’ and ‘the fallen leaves’ do not seem to have the relation that holds between ‘the smoke’ and ‘the cigarette’. Among Japanese linguistic literature for machine translation, Narita (1994) proposes the following treatment for pseudo relatives: (28) Change the structure into the regular relative clause which contains the gap, and then translate it into English. (Narita 1994: 86) For example, (29) [gasu-ga morete-iru ] nioi gas-nom leaking-is smell (lit.) ‘the smell that the gas is leaking’

Japanese ‘gapless’ relatives 

is transformed to (30) whose structure allows a more natural translation: (30) [ei morete-iru] gasui -no nioi leaking-is gas-gen smell ‘the smell of the gas that is leaking’ With this treatment, (26), repeated here as (31), is converted to (32): (31) [Dareka -ga otiba -o taku ] kemuri-ga tati_nobotte iru. someone-nom fallen_leaves-acc burn smoke-nom rising -is (lit.) ‘The smoke that someone is burning fallen leaves is rising’ (32) [Dareka-ga ei taku] otibaI -no kemuri-ga someone-nom burn fallen_leaves-gen smoke-nom tati_nobotte iru. rising -is ‘The smoke of the fallen leaves that someone is burning is rising’ This manipulation of transforming the structure, motivated by technical needs for machine translation, resolves certain problematic cases like (26)/(31). The main factor that makes the interpretation acceptable is the verb in the relative clause that supplies the missing link in the meaning relation between ‘the smoke’ and ‘the fallen leaves’. However Narita’s treatment fails to give the right interpretation for such examples as in (33) and (38): (33) [Kuruma-ga torakku-to butukatta] oto -wa sugokatta. car -nom lorry -with crashed sound-top was-dreadful ‘The sound of the car crashing with the lorry was dreadful’ (34) ?[ei torakku-to butukatta] kurumai -no oto -wa sugokatta. lorry -with crashed car -gen sound-top dreadful ‘The sound of the car that crashed with the lorry was dreadful’ (35) ?[Kuruma-ga ei butukatta] torakkui -no oto -wa sugokatta. car -nom crashed lorry -gen sound-top dreadful ‘The sound of the lorry that the car crashed with was dreadful’ In this case, supplying the (regular) relative clause is not sufficient, as changing the structure into the regular relative does not change the respect that the relation is set between ‘the sound’ and an individual picked on by the NP (either ‘the car’ or ‘the lorry’). (36)

?kuruma-no oto

‘the sound of the car’

(37)

?torakku-no oto

‘the sound of the lorry’

 Akiko Kurosawa

The argument to be added to the head nominal cannot be clearly specified in this case. Intuitively we know it is wrong to attribute the source of the sound just to one of them. Another counter example is: (38) [Dareka -ga to -o tataku ] oto-ga sita. someone-nom door-acc knock sound-nom did (lit.) ‘I heard the sound that someone was knocking on the door’ ‘I heard the sound of someone knocking on the door’ Changing the structure into a regular relative gives: (39) *[ ei to-o tataku ] dareka i -no oto door-acc knock someone-gen sound ‘the sound of someone who knocks on the door’ (40) ?[ dareka -ga ei tataku ] to i -no oto someone-nom knock door-gen sound ‘the sound of the door that someone knocks on’ From these observations, it is now clear that there is not a single NP (even with a regular relative clause) that has a meaning relation with the head. It is the whole gapless relative clause rather than an NP that projects the implied relation and that has to project the required argument. (41) [Someone is burning the fallen leaves] The smoke . . . (42) [The car crashed with the lorry] The sound . . . (43) [Someone knocked on the door] The sound . . . This view also accommodates examples (17) and (18) that seem to fit the two-place relation analysis that we have looked at above: (44) [The cake is baking] The smell . . . (45) [I spilled wine on the jumper] The stain . . . The two-place relations, ‘X-no Y’ or ‘Y of X’, are in fact semantically indeterminate in nature like the possessive construction,12 whose meaning relation is interpreted contextually.13 In (17), ‘the smell of the cake’ does not necessarily mean ‘the smell of the cake baking’; it can be used to describe ‘the smell of the cake’ which is baked this morning and now on the plate. Unlike the two-place relations, gapless relatives in (17) and (18) describe how the smell and the stain are made. What is not linguistically encoded is how these relatives are related to the head as relatives. In the next sub-section, I will take up a parallelism found between gapless relatives and bridging reference.

Japanese ‘gapless’ relatives

. Bridging reference14 and gapless relatives Interestingly Japanese gapless relatives look like bridging reference in English. (46) I walked into the room. The chandeliers sparkled brightly. (47) Peter went to a Japanese restaurant. The waitress was from Osaka. In interpreting The chandeliers and The waitress whose referents are not explicitly mentioned, the hearers add the (bridging) assumptions in (48) and (49) below to the context: Bridging assumptions: (48) The room had chandeliers. (49) The waitress was working at the Japanese restaurant. Now if a Japanese gapless relative clause and the matrix clause are divided into two simple sentences, it yields a bridging reference construction as follows (shown in English): (50) The cake is baking. The smell is lovely. (51) I spilled wine on the jumper. The stain doesn’t come off. (52) Someone is burning the fallen leaves. The smoke is rising. (53) The car crashed with the lorry. The sound was dreadful. (54) Someone knocked on the door. I heard the sound. (55) The fire spread. The cause is unclear. The italicised NPs which are the head in the gapless relative construction are bridging NPs in bridging reference.15 The referent of the bridging NP is pragmatically inferred through a bridging assumption. Likewise the relation between the head nominal and the pseudo relative clause is retrieved by pragmatic inference via a contextual assumption which is based on general knowledge schemas. Contextual assumptions: (56) If the cake is baking, there is a smell of it. (57) If one spills wine, it makes a stain. (58) If someone burns fallen leaves, it makes smoke. (59) If the car crashes with the lorry, it causes a (dreadful) sound. (60) If someone knocks on the door, there is a sound of it. (61) If the fire spread, there must be a cause.



 Akiko Kurosawa

Contextual assumptions, being combined with the proposition of the relative clause as another premise, yield contextual implications in the following manner: (62) If the cake is baking, there is P→Q contextual assumption a smell of it. The cake is baking. P proposition in the relative clause There is a smell of it.

Q

contextual implication (head)

(63) If the car crashes with the lorry, it causes a (dreadful) sound. The car crashes with the lorry. It causes a (dreadful) sound. In (62) and (63), the conclusion part includes the head nominal, ‘smell’ and ‘sound’ respectively, showing how the relation which is not encoded is provided through inference. The assumption that allows bridging reference and pseudo gapless relatives is generally the one that is readily accessible in our general knowledge schemas. Accessibility of assumption affects the acceptability of bridging and gapless relatives. It is argued in Matsui (1995, 1998) that the accessibility factors involve a pragmatic criterion of consistency with the principle of relevance proposed by Sperber and Wilson (1986/1995) which hearers can evaluate the resulting interpretation with. Criterion of consistency with the principle of relevance An utterance, on a given interpretation, is consistent with the principle of relevance if and only if the speaker might rationally have expected it to be optimally relevant to the hearer on that interpretation. Communicative principle of relevance Every act of ostensive communication communicates a presumption of optimal relevance. Sperber and Wilson’s claim is that an utterance is optimally relevant if and only if it has enough cognitive effects with no unjustifiable effort for the hearer to obtain them. Conversely less effort ensures higher acceptability.16 To demonstrate this point, compare (64) with (65) and (66): (64) a.

[Dareka -ga otiba -o taku ] kemuri someone-nom fallen_leaves-acc burn smoke ‘The smoke of someone burning fallen leaves’

Japanese ‘gapless’ relatives 

b. [Dareka -ga otiba -o taku (toki deru) ] someone-nom fallen_leaves-acc burn (when is-produced) kemurii smoke ‘The smoke (that is produced when) someone burns fallen leaves’ (65) a. ?*[Hito -ga ie -o tateta ] gomi people-nom house-acc have_built rubbish ‘the rubbish of people having built a house’ b. [Hito -ga ie -o tateta (toki deru) ] gomi people-nom house-acc have_build (when is-produced) rubbish ‘the rubbish (that is produced when) people have built a house’ (66) a. ?*[Hito -ga sekitan-o hotta] gasu people-nom coal -acc have_dug gas ‘the gas of people having dug coal’ b. [Hito -ga sekitan-o hotta (toki deru) ] gasu people-nom coal -acc have_dug (when is-produced) gas ‘the gas (that is produced when) people have dug coal’ (examples slightly altered from those in Abe 1994) In (64), the interpretation which is the explicature of (64a) is given in (64b). However, in (65a) and (66a), the gapless relatives are NOT acceptable, while (65b) and (66b) show the same kind of relation. Building a house may produce rubbish, and digging coal may produce gas, but the association between them is not as strong as the one between burning the fallen leaves and the smoke, or the cake baking and the smell, and not accessible enough to allow them taking the form of a gapless relative clause and its head. It is presumably due to the low frequency of use (of those combinations) and complexity of the assumption. The contextual assumption needed for (65a) has two steps; when a house is built, an old building on the site has to be demolished, and if an old building is demolished, it produces rubbish. If the step is reduced to one, the acceptability increases a great deal as in (67) below:17 (67) [Hito -ga ie -o kowasita] gomi people-nom house-acc demolished rubbish ‘the rubbish of people having demolished a house’ (65a) and (66a) are unacceptable as the contextual assumption is considered not accessible enough given normal expectation of effect.

 Akiko Kurosawa

. Saturation of hidden indexicals and ‘free’ enrichment processes We assume that there is a pragmatic instruction interacting with computational rules and lexical actions in interpreting utterances, and Dynamic Syntax assumes Relevance Theory developed by Sperber and Wilson (1986/1995) as its background. In Relevance Theory, examples of ‘and’ conjunction such as (69) and (70) are argued as a case of pragmatic enrichment (Carston 1998, 2000): (68) It’s summer in England and it’s winter in New Zealand. (P and Q = Q and P) (69) Sue got a PhD and (then) became a lecturer.

(P and Q = Q and P)

(70) She shot him in the head and (as a consequence) he died instantly. (P and Q = Q and P) In (68), changing the order of two conjuncts/propositions does not affect the truthconditional meaning, whereas in (69) and (70) it does. A temporal relation in (69) and a cause-consequence relation in (70) are NOT linguistically encoded and determined but pragmatically inferred. The pragmatically inferred relation enriches the logical form which then develops into an explicature, i.e. a communicated assumption. While the enrichment is the process of recovering unarticulated constituents, there is a process of supplying contextual values to indexicals which is termed ‘saturation’ (Recanati 1993: 233–268). A saturation process does not only supplies values to overtly marked indexicals such as pronouns and demonstratives but also supplies values contextually to ‘hidden indexicals’, which are invisible but linguistically present constituents. Examples include certain ‘relational’ terms and genitive constructions (Carston 2000): (71) The winners each get £1,000.

[winners of what?]

(72) I like Sally’s shoes.

[shoes in what relation to Sally?]

There is a covert slot to be filled, and these are semantically incomplete until the constituent is contextually supplied in answering the bracketed question, and there is a lexical item which requires completion (e.g. winner, genitive marker). The question is whether the pseudo gapless relative is a case of enrichment or saturation. In case of pseudo gapless relatives whose head is a relational noun such as usiro ‘back’, yokuzitu ‘next day’ and gen’in ‘cause’ as in (20)–(22), they have a covert slot to be filled in like winner in (71), which calls for saturation (e.g. the back of what?). Then the proposition of the relative clause is there to fill the slot.18

Japanese ‘gapless’ relatives 

However, no lexical item dictates ‘completion’ in, for example, (17), repeated here as (73), whose explicature is shown in (74): (73) [Keeki-ga yakeru] nioi -wa subarasii. Cake -nom bake smell-top is_lovely (lit.) ‘The smell that the cake is baking is lovely’ (74) [Keeki-ga yakeru (-toki suru) ] nioi-wa subarasii. cake-nom bakes (when it_smells) smell-top is-lovely ‘The smell (that is made when) the cake is baking is lovely’ None of the lexical items in Keeki-ga yakeru ‘the cake bakes’, and nioi ‘the smell’ has such a requirement. The grammatical structure, the relative clause in this case, constrains the interpretation, forcing the hearer to look for a certain meaning relation between the relative clause and the head, as between them there must be some relation as long as they are adjacent in the form of the relative clause and the head. In this sense, a hidden slot is present though in a very abstract way. One might say that nioi ‘the smell’ evokes a question, ‘the smell of what?’, and that there is a covert slot as well. But if so, it seems that more and more nonrelational expressions are subject to having this type of slot: The sound of what and caused by what, the rubbish of what, and the gas of what, and so on. Even for the nouns such as book and door, one can ask for slots: the book of whom, on what, the door of which room, etc. Then we are not sure how many ‘slots’ we will have. In this sense, words are ‘relational’ to some extent, and it is undetermined how many slots we want.19 However, more importantly, the pragmatic inference that supplies the value for the meaning relation to the pseudo gapless relatives is based on the contextual assumption accessible from encyclopaedic knowledge schemas, which is the mechanism of interpretation also for enrichment. As Carston (1998) puts it: . . . these relations being supplied perhaps by highly accessible general knowledge schemas concerning relevant ways in which events connect up. (Carston 1998: 225)

It is presumably general knowledge schemas on the basis of which the interpretation of (17)/(73) is determined, and this non-linguistic aspect of interpretation processes is what is shared by ‘and’ conjunction, bridging reference, and pseudo relatives. In the next subsection, I turn to the Dynamic Syntax proposal in which I shall argue that an event variable is employed in modeling the gapless relatives. The event variable stands for the proposition in the relative clause. However, unlike a bound variable in quantification, it is not sufficient to fully determine the interpretation itself; its role is to express that the whole relative clause has a meaning relation with the head. How it is related is only pragmatically inferred.

 Akiko Kurosawa

. The event variable analysis As we have seen, the relation between the head and the gapless relative clause may vary. My proposal is to represent it employing an event variable so that these various and complex relations can be expressed in a unified way. The notion of event variable is not new. The idea that, the referential relation between a proposition and the event it denotes is one of reference, was introduced by Davidson (1967). Davidson directed his attention to the logical (entailment) relation between (a) and the others in (75) below: (75) a. b. c. d. e.

Jones buttered the toast in the bathroom with a knife at midnight. Jones buttered the toast in the bathroom. Jones buttered the toast with a knife. Jones buttered the toast at midnight. Jones buttered the toast.

Davidson proposes that (75a) has a logical form as in (76): (76) ∃x (butter (x, Jones, toast) & in (x, bathroom) & with (x, knife) & at (x, midnight)) The verb butter stands for an event, and the event variable, x, is bound by an existential quantifier, asserting the existence of the event. (76) is paraphrased as an event of Jones’ buttering the toast took place, its place was the bathroom, the instrument used for it was a knife, and the time of it was at midnight. Prepositions are treated as a predicate, with the event variable and a nominal as their arguments. Separating each prepositional phrase from the verb as a conjunct using the event variable, now it is possible to show the entailment relation between (a) and the others in (75), as a conjunction entails each of its conjuncts: (P & Q) → P. Davidson’s proposal is extended by Parsons (1990) to account for various phenomena such as the relation between ‘Agatha saw Brutus stab Caesar’ and ‘Agatha saw the stabbing of Caesar by Brutus’. Anaphoric properties of tense are expressed through the use of event variables by Kamp and Reyle (1993: 483–690), and Steedman (1997). In Dynamic Syntax, a metavariable is used as a label si for a propositional formula P as shown below: (77) si : P This variable can be used as an argument for any relation/predicate of appropriate type. Perrett (2000) gives an account of temporal and causal relations in Hadiyya, which is a verb-final language of Cushitic group of the Afro-Asiatic family:

Japanese ‘gapless’ relatives 

(78) S1 precedes the Utterance time

Precede (SUTT ) (S1 )

(79) a. S1 precedes S2 b. S1 causes S2

Precede (S2 ) (S1 ) Cause (S2 ) (S1 )

Past tense

Tense has anaphoric properties in that it expresses a temporal relation between the event variable associated with the current sentence and some other event time, and, which event time that is, is determined in context. This anaphoric relation of tense is inferred like the case of pronouns, whose interpretation is fixed by other representations in context, e.g. from expressions in the preceding clauses and sentences. This observation underlies the DS representation of tense. The temporal metavariable in (78), like the Formula metavariables, [Fo(U)], has no denotational content and a value will have to be supplied through the interpretation process. Si is an argument of a tense relation indicated by the verbal morphology in (78) above, e.g. the suffix -ed in English, and -u (or -o), a simple perfect morpheme in Hadiyya. The time value for S1 , which interval at some time in the past of the utterance time, is contextually inferred by non-linguistic knowledge. In (79), the particular interpretation of the temporal relation in (a) and the causal relation in (b) between conjoined clauses may be pragmatically achieved. Similarly, the event variable that stands for the proposition is used to represent the whole event in the gapless relative clause.20 The anaphoric relation (i.e. there is a common element) between two clauses is expressed by using a metavariable. The variable provides the argument for the relation between the gapless relative clause and the head. The structure-building process of (17)/(73) is demonstrated below: (80) [Keeki-ga yakeru] nioi-wa subarasii. cake-nom bakes smell-top is-lovely ‘The smell of the cake baking is lovely’ As seen in Section 5, phonological information plays a role in determining the structure on-line. A downstep between yakeru and nioi provides information that they belong to the same phonological domain reflecting the structure, which indicates that nioi is the head of a LINKed structure constructed from the immediately previous clause rather than a constituent of some separate structure. At this juncture, keeki-ga yakeru nioi can be unambiguously identified as a relative clause and its head. However, unlike the regular relatives, the tree constructed from keeki-ga yakeru has no outstanding requirement and no node for nioi to merge.

 Akiko Kurosawa { Ty(t)}

{Fo(S1, Nioi (S1)), {Fo(Subarasii) Ty(e)} ‘smell’ Ty(e  t)} ‘lovely’ LINK {Ty(t), Fo(S1: (Yakeru(Keeki))}

{Fo(Keeki), ‘cake’ Ty(e) }

{Fo(Yakeru), ‘bake’ Ty(e  t)}

S1 = bake (cake)

The event of the relative clause is expressed with S1 , which is the label for the proposition yakeru (keeki), ‘bake (cake)’, and it appears as an event variable on the node for the head, nioi, ‘smell’, showing that this is the common element between the two clauses. S1 is used as an argument of the head, nioi, expressing that the event of cake baking, S1 , has a property of smell. In this way the relation which is pragmatically inferred is represented. The event variable does not supply the interpretation of the relation. It is from a contextual assumption, ‘if the cake bakes, there is a smell of it’ and the relative clause, ‘the cake bakes’, as another premise, that the conclusion, ‘there is a smell of it’, is drawn. This inference provides the basis for understanding the relation between the proposition of the relative clause and the individual denoted by the head.

. Appositional gapless relatives (‘content’ clauses) Another form of gapless relative clause, the appositional relative is known as a ‘content clause’ in the Japanese linguistic literature (Masuoka 1994): (81) [ John-ga Mary-o aisiteiru ] (to iu ) zizitu John-nom Mary-acc love comp fact ‘the fact that John loves Mary . . . ’ All the argument nodes in the relative clause being filled with formulae, the head, zizitu ‘fact’, cannot have an argument-predicate relation with the verb, aisiteiru ‘love’, hence no ‘gap’. The head, as given by zizitu ‘the fact’, is identified with the structure projected by the whole relative clause; i.e. the relative clause gives the ‘content’ of the ‘fact’, and is in apposition to the head in this sense. The DS analysis is that the shared element is the label s1 for the formula at the root node projected from the relative clause, in the same manner as for pseudo relatives.

Japanese ‘gapless’ relatives  ? Ty(t)

Ty(e), Fo(S1, Zizitu(S1)) ‘fact’ LINK S1 : Aisiteiru (Mary) (John)

John

Aisiteiru (Mary)

Mary zizitu (S1) ‘fact’

Aisiteiru ‘love’ S1 is the fact

The difference between the pseudo relatives and the appositional relatives is that pragmatic inference processes which are requisite for the former are not necessary for the latter. Despite differences between the two structures, we can nevertheless maintain the definition of LINK for both pseudo and appositional ‘gapless’ relative clauses. There is a common element present in both LINKed and main trees – the event variable. That the whole event expressed by the relative clause has a meaning relation with the individual denoted by the head is shown through it. Looking back and comparing head-initial (e.g. English) and head-final (e.g. Japanese) relative clauses, we see that there is asymmetry between them. In headinitial relative structures, only appositional (i.e. fact-that) constructions provide an event variable as the common element. In head-final relative constructions, the head dictates that an event-variable interpretation is available in both types of gapless relatives.

. Conclusion In this paper, Japanese relatives are analysed in the Dynamic Syntax perspective as linked structures sharing a variable. The process of projection is the reverse of English, an argument variable in the LINKed structure copied over as the head. On the assumption that the structure projected is semantically transparent, event variables labelling the propositional formula are used as a common element. This manipulation of event variables allows a natural characterisation of the structure required for relative clause construal while respecting the pragmatic nature of the inference process. The projection of structure for natural languages needs to allow integration between structural and pragmatic factors.

 Akiko Kurosawa

Notes * This paper is an extension of a talk given jointly by myself and Ruth Kempson at the Second International Conference on Contrastive Semantics and Pragmatics in Cambridge. I am indebted to Ruth Kempson for insightful comments and constructive suggestions on earlier drafts of the paper; all remaining errors are mine. I have also benefited greatly from discussions with Tomoko Matsui. . In Japanese, nominative and accusative case markers are optionally omitted (e.g. John ringo tabeta). Each argument node is always introduced as initially ‘unfixed’ in the structure (see 2.3). Furthermore, arguments themselves need not be explicitly expressed. (e.g. e e tabeta). Even in such a case, the tree structure is articulated by the verb as in (c) in 3.2. Then metavariables on two argument nodes are substituted with John and ringo. This anaphora resolution may involve a pragmatic process. . In Classical Japanese, the adnominal form (rentai-kei, e.g. aru ‘exist’) of a predicate which is distinctive from the conclusive form (shushi-kei, e.g. ari ‘exist’) in most cases played a role of the relativiser, indicating that some kind of nominal should follow. However, in Modern Japanese, the distinction is lost from the majority of predicates, and its role is much weaker. Note that Modern Korean has the adnominal form which is distinctive from the conclusive form and works as a relativiser. . Strictly it is not the noun which is the head but the variable X which is an argument of the predicate projected by the noun, Doroboo, (X, Doroboo(X)). I sidestep this distinction in what follows, referring loosely to the noun as providing the head (see Kempson et al. 2001: 35, 115, 137 for the internal structure of the noun). . See Kubozono (1987: Chap. 5; 1995: 44, 98–116) for the detailed discussion on phonological phrasing reflecting structure. . For instance, (ii) below shows how the downstep is blocked after aisita while it occurs in (i): (i)

Watasi-ga aisita | supai I -nom loved spy ‘the spy I loved’

(ii) Watasi-ga aisita supai-ga katuyaku-suru monogatari I -nom loved spy -nom play_an_active_role story ‘the story I loved, in which a spy plays an active role’ In (ii), supai ‘spy’ is a constituent of the second relative clause, ‘(in which) a spy plays an active role’. . In Japanese, a downstep occurs before the head nominal if at least one of the two constituents of the relative clause is lexically accented. In Okane-o tukanda | doroboo, tukanda has a lexical accent on ka. If both constituents in the relative clause are accentless as in (i) below, the three constituents merge, forming a single intonational phrase. The second and the third constituents lose their original low pitch on the first nuclei in (ii), leaving the domain-initial nucleus, o, as low pitched as in (iii):

Japanese ‘gapless’ relatives

(i)

Okane-o suteta doroboo money-acc threw_away thief ‘the thief who threw away the money’

(ii) [o ka ne o] [su te ta] [do ro bo o] (iii) [o ka ne o su te ta do ro bo o] This intonational phrasing is blocked if doroboo is not the head nominal but a constituent of another syntactic domain (e.g. another relative clause), indicating that there is a border between okane-o suteta and doroboo. . This paper discusses phonological information confining it to the one which is crucial in structure building of LINK on given examples. . There are a considerable number of nouns which are the head of a relative clause and at the same time function as a (part of) subordinator leading an adverbial clause in Japanese. Among them are: toki ‘time/when’; mae ‘front/before (temporal adverbial), front/in front of (locational adverbial)’; ato ‘back/after’ (temporal); usiro ‘back/behind’ (locational); tokoro ‘place, point, scene/as’; tame ‘reason/because (causal), purpose/in order to (purpose)’; baai ‘case/if ’; ue ‘top/in addition to’. Each word has a different degree of grammaticalisation, but all belong to both lexical and functional categories to some extent. . Teramura (1975–1978/1992) used the term Soto-no Kankei ‘the outer relation’ for the relations between the head noun and the modifying clause for both pseudo and appositional (content) gapless relatives while that of the regular relatives is Uchi-no Kankei ‘the inner relation’. Two sub-types with nouns of perception and relation as the head are termed Pseudo relatives in Abe (1994), and the other three sub-types are called Content clauses in Masuoka (1994). . Korean also has this type of gapless relatives. . The same noun may be used in regular relatives and in appositional gapless relatives as well as in pseudo relatives as in the examples below. So it is NOT the noun itself that determines the type of relatives: (i)

[ei koohyoo sareta] kekkai was_published result ‘the result i that ei was published’

(regular relative clause)

(ii) Syusyoo -ga hatugen_si-te, [sizi -ritu -ga sagaru] (to iu) prime minister-nom speak -and approval rate-nom drop (comp) kekka-to natta. result-ptl became (appositional gapless relative clause) ‘The prime minister made a remark and (it) brought about the result that the approval rate (for the government) dropped’



 Akiko Kurosawa

(iii) [Syusyoo -ga hatugen_sita] kekka, sizi -ritu -ga sagatta. prime minister-nom spoke result approval rate-nom dropped (pseudo relative clause) ‘As a result of the prime minister’s having made a remark, the approval rate (for the government) dropped’ . See Kempson (1977: 123–138) for a discussion. . See Blakemore (1992: 83) for examples in a context. . Cf. Matsui 1992, 1995, 1998, 2000; Wilson & Matsui 1998. . The pseudo relatives can invariably be paraphrased as bridging reference sentences. However the reverse is not always true; e.g. (46) and (47) are not acceptable as gapless relatives in Japanese. A general association between two NPs such as a room and chandeliers, or a restaurant and a waitress may allow a bridging reference interpretation as well as the relation between an event in the first clause and an NP in the second clause as in (50)–(55). However, in pseudo relatives, the relation is restricted to the one between the event described in the relative clause and the head NP. . This paper does not discuss an interesting case of gapless relatives seen in Japanese haiku poetry, each consisting only of 17 syllables: The contextual assumption which the reader is required to make in order to connect the head and the relative clause is sometimes far less accessible than in the examples we have seen, but nevertheless it is acceptable (examples below). Relevance Theory explains this in terms of the balance of the processing effort and the cognitive effect; extra processing effort demanded should be offset by extra effect, like any element of indirectness seen in the case of metaphor, repetition, etc. puts the hearer to extra effort, and promises extra effects. In case of haiku poetry, unusual, less frequently used general knowledge schemas (hence low accessibility) encourages a number of weak implicatures as well, (readers are responsible for their recovery), hence poetic effects. (i)

[Choo -no sita zenmai -ni niru ] atusa-kana butterfly-gen tongue flowering_fern-to resemble heat -ptl (lit.) ‘the heat that the butterfly’s tongue resembles a flowering fern’ (Ryunosuke Akutagawa)

(ii) [Uki tosi -no ima aratamaru] kiteki -kana sorrowful year-gen/nom now is-renewed steam_whistle -ptl The sorrowful year is now being renewed, which (is known by the sound of) the steamwhistle (that all the ships in port blow at midnight) (Tae Koizumi) The expected cognitive effects are more specific in these haiku examples, and the interpretation process is effect-driven, while the interpretation process with more accessible contextual assumptions in examples like (17) and (18) is effort-driven. Approaches such as cognitive grammar (Yamanashi 1995) and Frame Semantics (Matsumoto 1997) might not be able to provide very adequate accounts in this respect, as the head noun of the pseudo relatives is explained either as ‘a default value drawn from the event schema’ (Yamanashi 1995: 180–182) or a member which bears a role in a frame, where

Japanese ‘gapless’ relatives 

the frame is associated with prototypical scenes (Matsumoto 1997). These descriptions do not seem to accommodate unconventional cases of gapless relatives. . This example is due to Jun’ichi and Chihiro Tsuchiya (personal communication). . To be exact, there needs to be the bracketed constituent with which the proposition in the relative clause fills the slot: (i)

[ Humiko-ga suwatta ] (basho -no) usiro Humiko-nom sat place-gen back ‘Behind (the place) Humiko was sitting’

(ii) [Kodomo-ga umareta ] (hi -no ) yokuzitu child -nom was_born day -gen next_day ‘the next day (of the day ) the child was born’ . It is of some interest that Marten (1999, 2002) advocates that ALL verbs underspecify their type. The specific choice of type is selected according to the number of optional arguments. This argument is well suited to explain this phenomenon. Carston (2000) argues that not all the ‘hidden indexicals’ receive a contextual value on a particular occasion of use, and only those constituents which are relevant are recovered (e.g. I’ve eaten might contain four hidden indexicals as in I’ve eaten [x] [in manner y] [at location l] [within time span t], but in I must wash my hands: I’ve eaten, only the manner constituent, using my hands, is necessary to be recovered and the others are redundant). . The motivation for the use of the event variable in DS is different from that of Davidson’s and Parsons’, and there is no need to separate prepositional phrases as conjuncts.

References Abe, Y. (1994). Rentai Shushoku no Sho-Mondai. [Issues on Noun Modification.] In Y. Takubo (Ed., 1994) (pp. 153–171). Blakemore, D. (1992). Understanding Utterances. Blackwell. Carston, R. (1998). Pragmatics and the Explicit – Implicit distinction. Ph.D. thesis, University College London. Carston, R. (2000). Explicature and Semantics. UCL Working Papers in Linguistics, 12, 1– 44. To appear in S. Davis & B. Gillon (Eds.), Semantics: A Reader. Oxford: Oxford University Press. Davidson, D. (1967). The logical form of action sentences. In N. Rescher (Ed.), The Logic of Decision and Action (pp. 81–95). Pittsburgh: University of Pittsburgh Press. Kamp, H., & Reyle, U. (1993). From discourse to Logic. Dordrecht: Kluwer Academic Publications. Kempson, R. M. (1977). Semantic Theory. Cambridge University Press. Kempson, R. M., Meyer-Viol, W., & Gabbay, D. (2001). Dynamic Syntax: The Flow of Natural Language Understanding. Oxford: Blackwell. Kubozono, H. (1987). The Organization of Japanese Prosody. Ph.D. thesis, University of Edinburgh.

 Akiko Kurosawa

Kubozono, H. (1995). Go-keisei to On’in Kozo. [Word Formation and Phonological Structure]. Tokyo: Kuroshio Shuppan. Kurosawa, A. (2000). ‘Soto-no kankei’-no rentai shushoku ni tsuite. [On Gapless Relatives.] Paper presented at the 13th conference of Japanese Language Teaching, Vienna. Kurosawa, A. (in prep.). Japanese Relative Clauses from a Dynamic Syntax Perspective. Ph.D. thesis, King’s College London. Marten, L. (1999). Syntactic and Semantic Underspecification in the Verb Phrase. Ph.D. thesis, SOAS, University of London. Marten, L. (2002). At the Syntax-Pragmatics Interface: Verbal Underspecification and Concept Formation in Dynamic Syntax. Oxford: Oxford University Press. Masuoka, T. (1994). Meishi Shushoku Hyogen no Setsuzoku Keishiki -Naiyo-setsu o Chushi ni. [Optional Complementisers for Content Clauses in Noun Modification.] In Y. Takubo (Ed., 1994) (pp. 5–27). Matsui, T. (1992). Bridging Reference and the Notions of ‘Topic’ and ‘Focus’. In UCL Working Papers in Linguistics, 4 (pp. 239–258). University College London. Matsui, T. (1995). Bridging and Relevance. Ph.D. thesis, University College London. Matsui, T. (1998). Pragmatic criteria for reference assignment: A relevance-theoretic account of the acceptability of bridging. Pragmatics & Cognition, 6, 47–98. Matsui, T. (2000). Bridging and Relevance. Amsterdam: John Benjamins. Matsumoto, Y. (1997). Noun-modifying Constructions in Japanese – A Flame-semantic approach. Amsterdam: John Benjamins. Narita, H. (1994). Rentai Shushoku-setsu no Kozo Tokusei to Gengo Shori – Nihongo rashii Hyogen no Kikai Hon’yaku to Oyo Gijutsu. [Syntactic Idiosyncracy and Processing of Noun Modification – The Machine Translation of Idiosyncratic Japanese Expressions and Applied Technology.] In Y. Takubo (Ed., 1994) (pp. 67–126). Parsons, T. (1990). Events in the Semantics of English – A Study in Subatomic Semantics. Cambridge, MA: The MIT Press. Perrett, D. L. (2000). The Dynamics of Tense Construal in Hadiyya. Ph.D. thesis, SOAS, University of London. Recanati, F. (1993). Direct Reference – from Language to Thought. Oxford: Blackwell. Sperber, D., & Wilson, D. (1986/1995). Relevance [Second edition]. Oxford: Blackwell. Steedman, M. (1997). Temporality. In J. van Benthem & A. ter Meulen (Eds.), Handbook of Logic and Language, Chap. 16 (pp. 895–938). Amsterdam: Elsevier Science B. V. Takubo, Y. (Ed.). (1994). Nihongo no Meishi Shushoku Hyogen. [Noun Modification in Japanese.] Tokyo: Kuroshio Shuppan. Teramura, H. (1975–1978/1992). Rentai Shushoku no Shintakusu to Imi. [Syntax and Semantics of Noun Modification in Japanese.] In H. Teramura (Ed., 1992), Teramura Hideo Ronbun-shu I (pp. 157–336). Tokyo: Kuroshio Shuppan. Wilson, D., & Matsui, T. (1998). Recent approaches to bridging: Truth, coherence, relevance. In UCL Working Papers in Linguistics, 10, 173–200. University College London. Yamanashi, M. (1995). Ninchi Bunpo-ron [Cognitive Grammer]. Tokyo: Hitsuji Shobo.

TSL[v.20020404] Prn:29/08/2002; 16:38

F: PBNS18.tex / p.1 (335)

Constraint interaction at the semantics/pragmatics interface The case of clitic doubling Javier Gutiérrez-Rexach The Ohio State University, USA

Introduction Clitic pronouns are prosodically weak proforms which are morpho-syntactically dependent and normally attach to a verbal host. The study of these proforms faces two main challenges. First, it is necessary to provide a formal semantics of clitics and clitic doubling constructions in a manner that goes beyond the impressionistic characterizations which are often found in the literature. Second, cross-linguistic variation has to be addressed in a way in which semantic and pragmatic factors are factored in as the main triggers of variation. In Gutiérrez-Rexach (1999a, 2001a), a strong correspondence hypothesis at the syntactic and semantic levels is proposed. In the syntax, clitics are specified as definite determiners, as proposed by Torrego (1995) and Uriagereka (1995). This assimilation of clitics to determiner heads is in the spirit of Postal’s (1966: 203) proposal: “My basic claim is that the so-called pronouns I, our, they, etc. are really articles, in fact types of definite articles.” As a matter of fact, English pronouns exhibit overt determiner-like behavior in some cases: (1) a. Did you see us guys? b. Who insulted you men? The phenomenon of clitic doubling, also called pronominal reduplication, can be seen as a clear instantiation of Postal’s idea. The difference between Spanish-type languages, which allow clitic doubling, and French-type languages, which do not, can be cast in the following terms: Spanish-type clitics are specified as selecting for

TSL[v.20020404] Prn:29/08/2002; 16:38

F: PBNS18.tex / p.2 (336)

 Javier Gutiérrez-Rexach

an argumental DP (the doubled element), whereas French or Italian clitics lack this specification. (2) a. Juan le ha dado libros a Isabel b. *Jean lui a donné des livres a Isabelle ‘Jean has given books to Isabelle’ In Gutiérrez-Rexach (1999a, 2001a), in accordance with the proposed strong correspondence hypothesis, it is also proposed that clitic pronouns denote determiner functions inherently restricted to context sets. Westerståhl (1985) proposed that pronouns should be treated as functions from context sets to generalized quantifiers, where a generalized quantifier is a function from sets to truth values. For example, the denotation of the pronoun they is, for a universe of individuals E and all sets A, C ⊆ E, the function THEY ∈ [P(E) → [P(E) → 2]] such that THEY(C)(A) = 1 iff C ⊆ A & |C| ≥ 2. The set C is the context set of the pronominal function. In (3), C is the set of two old men who are walking in the park. (3) Two old men are walking in the park. They are bald. Clitics are, thus, pronominal determiners. As is well known, clitic pronouns are normally morphologically specified as accusative or dative forms. Following Keenan’s (1989) semantic case theory, we can take the case feature to indicate the type of the function with which clitics combine. Let α be a clitic expression, E a universe, and [A → B] the set of functions from A to B. Then, (i) if α is [+nom.], [[α]] is a function F ∈ [P(E) → [P(E) → 2]]; (ii) if α is [+acc.], [[α]] is a function F ∈ [P(E) → [P(E2 ) → P(E)]]; and finally (iii) if α is [+dat.], [[α]] is a function F ∈ [P(E) → [P(E3 ) → P(E2 )]]. The derivation of the truth conditions of (4) is as in (5), where R ⊆ E2 , C ⊆ E, a, b ∈ E, Ip is the individual generated by Pedro (Keenan 1996) and LO(C)(R) = {a|C ⊆ {b| < a, b >∈ R} & |C| = 1}. (4) Pedro lo compró Pedro it bought ‘Pedro bought it’ (5) Ip (IT)(C)(BOUGHT) = 1 iff Ip ( {a|C ⊆ {b| < a, b >∈ BOUGHT} & |C| = 1}) = 1 iff p ∈ {a|C ⊆ {b| < a, b >∈ BOUGHT} & |C| = 1} Similarly, the derivation of the truth conditions of (6), where a dative pronominal clitic occurs, is as in (7), where R ⊆ E3 , C ⊆ E, a, b, c ∈ E, and LES(C)(R) = {< a, b > | C ⊆ {c| < a, b, c >∈ R} & |C| ≥ 2}.

TSL[v.20020404] Prn:29/08/2002; 16:38

F: PBNS18.tex / p.3 (337)

Constraint interaction at the semantics/pragmatics interface 

(6) Juan les dio algunos libros Juan them gave some-pl books ‘Juan gave them some books’ (7) JUAN ((LES(C)(GAVE))(SOMEpl BOOK)acc) = 1 iff Ij ( {a | | BOOK ∩ {b | C ⊆ {c | < a, b, c >∈ GAVE} & |C| ≥ 2}} | ≥ 2})= 1 iff j ∈ {a | | BOOK ∩ {b | C ⊆ {c | < a, b, c >∈ GAVE} & |C| ≥ 2}} | ≥ 2}. The determination of the relevant context set in cases such as the ones illustrated above is constrained by standard restrictions on pronominal anaphora resolution, such as the following ones: salience, parallelism, linear precedence in a text, focal stress assignment, and common sense reasoning. For instance, the discourse in (8) makes the woman who was walking salient and this singleton set is the context set for the interpretation of the pronoun. On the other hand, in (9a) there are two possible “antecedents” which are equally salient. A common sense reasoning constraint is the one that determines the context set of la in (9a). In (9b), where there would be two context sets available, the dominant constraint to determine the content of the context set is focal stress assignment. (8) Una mujer estaba andando. Pedro la vio a woman was walking. Pedro her saw ‘A woman was walking. Pedro saw her.’ (9) a.

María dijo que Luisa estaba enferma, pero Pedro la vio esta mañana ‘María said that Luisa was sick, but Pedro saw her this morning’ b. Pedro dijo que María y LUISA estaban enfermas, pero Pedro la vio esta mañana ‘Pedro said that María and LUISA were sick, but Pedro saw her(Luisa) this morning’

Finally, a parallelism constraint on ellipsis resolution is responsible for the determination of the context set in (10), where the singleton set whose only element is Juan cannot be the context set of the accusative pronoun. Here I am neutral about the concrete mechanism that derives this constraint, but see Fox (1995), Dalrymple, Shieber and Pereira (1991), and Gutiérrez-Rexach (1999b) for a thorough discussion of the issues involved. (10) Juan odia a Pedro y Luis lo odia también ‘Juan hates Pedro and Luis hates him too’ The main claim of this paper is that pronominal doubling is associated with a series of constraints that operate at the semantics/pragmatics interface and that are different from the ones described above. What this means is that when a pronominal

TSL[v.20020404] Prn:29/08/2002; 16:38

F: PBNS18.tex / p.4 (338)

 Javier Gutiérrez-Rexach

clitic is doubled, it has to satisfy stricter semantic/pragmatic requirements than its non-doubled counterpart. In recent years, Optimality Theory (Prince & Smolensky 1993, 1997) has developed as a formal theory of constraint interaction and satisfaction in several domains of linguistic inquiry. Applications to syntax and, particularly to phonology and morphology, are well-known (see Kager 1999; Archangeli & Langendoen 1997; etc. for introductory overviews). Recently, an Optimality Theoretic approach to semantic/pragmatic constraints has been defended in work by Hendricks and de Hoop (2001), Zeevat (1999), Blütner (1999), etc. What I will argue is that an optimality-theoretic approach to constraint interaction in the semantics/pragmatics interface can successfully account for dialectal and cross-linguistic variation in clitic-doubling constructions.

Semantic and pragmatic constraints on clitic doubling Clitic doubling constructions differ from regular pronominal constructions in the way in which the content of the context set is determined. Without exception, in clitic doubling structures the context set is retrieved from the denotation of the doubled expression. Consider the correlate of (9b) in (11), where the only difference is that in the latter the pronoun doubles the accusative argument. No matter what the focus structure of (11) is, i.e. no matter whether María or Luisa receive focal stress, the context set for la has to be the singleton set with the individual denoted by María. (11) Pedro dijo que María y Luisa estaban enfermas, pero Pedro Pedro said that María and Luisa were sick, but Pedro la vio a María esta mañana her saw A María this morning ‘Pedro said that María and Luisa were sick, but Pedro saw her(María) this morning’ The relevant conclusion is that there has to be “semantic matching” between the context set required by the pronoun and one of the elements of the associated generalized quantifier. Formally: (12) Let αd/acc be an accusative doubling clitic expression and αacc its nondoubling counterpart. Then, for all C ⊆ E, R ⊆ E2 , and quantifiers Q (with accusative extension Qacc ): [[αd/acc ]](C)(R)(Qacc) = [[αacc ]](C)(R) & C ∈ Q

TSL[v.20020404] Prn:29/08/2002; 16:38

F: PBNS18.tex / p.5 (339)

Constraint interaction at the semantics/pragmatics interface 

This characterization does not impose any restriction on the nature of the associated quantifier, so in principle it predicts that an accusative clitic may double any expression with the matching case feature. Nevertheless, this is not correct. I claim that accusative clitic doubling in Spanish obeys four semantic/pragmatic constraints. The first one is the Principal Filter/Referentiality Constraint (REF). In the terms of generalized quantifier theory, we can say that the generalized quantifier associated with an accusative clitic has to be a principal filter. A generalized quantifier Q over E is a principal filter iff there is a non-empty set A ⊆ E, such that for all B ⊆ E, Q(B) = 1 iff A ⊆ B. The set A is called the generator of Q. When a DP denotes a principal filter it is referential. It refers to a particular group. The context set of a clitic pronoun is the group referred to by the associated DP. The contrast between the sentences (13) and (14) follows from the REF constraint. (13) Los engañaron a los/esos/mis/todos los estudiantes them fooled A the/those/my/all the students ‘They fooled those/my/all the students’ (14) *Los engañaron a una mayoría de them fooled A a majority of los/varios/pocos/aproximadamente tres estudiantes the/several/few/approximately three students ‘They fooled most/several/few/approxiamtely three students’ The REF constraint applies not only when the complement of a verb is a DP but also applies to propositional attitude verbs. In this case, the relevant criterion is the distinction between those verbs that select propositions as their complements, such as know, and those that select sets of propositions, such as question embedding verbs of the wonder type (Karttunen 1977). Typically, the complements of the verbs in the first class satisfy REF vacuously and can be accusatively doubled (15a), whereas those in the second class do not satisfy REF and are not doubled (15b). (15) a.

Lo sé que te sientes mal it know-I that you feel bad ‘I know that you feel bad’ b. *Se lo preguntó si te sentías mal refl it wonder if you felt bad ‘He wondered whether you felt bad’

The second constraint that has to be satisfied in Spanish accusative clitic doubling constructions is the Presuppositionality Constraint (PRES). The generator of the generalized quantifier associated with an accusative clitic is a presupposed set (See Capone 2000 for an analysis of presuppositionality and non-doubling clitics). The relevant presupposition is an existence presuposition, so the existence of the indi-

TSL[v.20020404] Prn:29/08/2002; 16:38

F: PBNS18.tex / p.6 (340)

 Javier Gutiérrez-Rexach

viduals in the set is presupposed. The constraints REF and PRES are not in conflict. Actually, as formulated, PRES represents a specialization of REF, since PRES requires that the group referred to by the associated DP has to be presupposed. The effects of this constraint are clearly shown in the following examples. Typically, focused elements cannot be clitic doubled. All of the “clitic . . . DP” associations in (16) would satisfy REF because Juan denotes a principal filter. Critically, none of them satisfies PRES because Juan is focused and does not belong to the presuppositional content of the sentence. Thus, prosodically focused doubled constituents (16a), or doubled constituents occurring as answers to a question (16b), in cleft and pseudocleft constructions (16c, d), or in the scope of the focus operator sólo ‘only’ (16e) violate PRES. (16) a. *Yo lo ví a [Foc Juan] I him saw A [Foc Juan] ‘I saw [Foc Juan]’ b. ¿A quién viste? (*Lo) vi a Juan A who saw-you him saw A Juan c. A quien (*lo) vi es a Juan A who him saw-I is A Juan ‘Who(ever) I saw is Juan’ d. Es Juan a quien (*lo) vi is Juan A who him saw-I ‘It is Juan who I saw’ e. (*Lo) vi sólo a Juan him saw-I only A Juan ‘I saw only Juan’ Other potential semantic/pragmatic constraints are first, a Salience/Emphasis Constraint (EMPH). The context set of the clitic (the group referred to by the doubled quantifier expression) is salient in a scale ordering available context sets. Again, this constraint does not conflict with the REF constraint. On the other hand, a conflict may arise with PRES, when the context sets available are focused and the associated order reflects the salience order among novel discourse elements. In this case, EMPH may override PRES, and the above sentences are perceived as felicitous. They are also usually associated with a specific intonational contour (“emphatic” intonation). Nevertheless, the default scenario is that PRES and EMPH do not conflict and the element emphasized by accusative clitic doubling is a discourse topic. In many Spanish dialects there is an additional constraint that we call the Animacy Constraint (ANIM). ANIM requires that the doubled quantifier live on the set ANIMATE/HUMAN. The “live on” property is understood in the terms of Bar-

TSL[v.20020404] Prn:29/08/2002; 16:38

F: PBNS18.tex / p.7 (341)

Constraint interaction at the semantics/pragmatics interface 

wise and Cooper (1981), namely the intersection of the denotation of the head noun with the set of animate/human entities is non-empty. What this amounts to is that only expressions that denote animate/human entities may be doubled. In most Spanish dialects, ANIM conflicts with REF. Even if an expression denotes a principal filter, it cannot be doubled by an accusative clitic if the generator of that principal filter is not a set of animate individuals. The ungrammaticality of (17) shows, then, that ANIM is ranked higher than REF (ANIM  REF), and that a violation of ANIM becomes fatal. (17) *Las limpié las mesas them cleaned-I the tables ‘I cleaned the tables’

Existential quantifiers and interactions with modality As has been observed in the literature (Suñer 1988, 1991; Sportiche 1996; Uriagereka 1995; etc.), specific objects tend to favor accusative doubling whereas non-specific ones block it. Thus, the examples in (18) are only grammatical/felicitous if the object is understood as “specific.” (18) a.

Las he visto a tres monjas/tres de las monjas them-fem have-I seen A three nuns/three of the nuns ‘I have seen three (specific) nuns’ b. Los enojó a algunos profesores them-pl annoyed A some-pl teachers ‘He annoyed some teachers’

What I want to defend here is that the “specificity effect” on doubled objects reduces to the combination of the REF and PRES constraints. REF requires that a doubled object denote a group and PRES requires that this group be presupposed. When applied to existential quantifiers, the combined satisfaction of both constraints is only compatible with “specific” readings. Furthermore, the specificity effect is in reality a referentiality effect. Something stronger than mere specificity, which equals presuppositionality in Diesing’s (1992) terms, is required. Scope marking is a property which is sensitive to the combination of the strong specificity attributes of a DP. In Spanish, there are two scope marking possibilities. In the neutral default case, word order determines the scopal order of constituents. Thus, (19a) is only compatible with the SUBJECT > OBJECT scopal order. Inverse scope readings, in which the indefinite object scopes over the subject, require either overt focusing of the object, as in (19b), or clitic doubling, as in (19c).

TSL[v.20020404] Prn:29/08/2002; 16:38

F: PBNS18.tex / p.8 (342)

 Javier Gutiérrez-Rexach

(19) a.

Tres three b. Tres three c. Tres three

estudiantes students estudiantes students estudiantes students

leyeron dos libros read two books leyeron [Foc dos libros] read two books los leyeron dos libros them read two books

The object wide scope reading of (19c) is a byproduct of the satisfaction of REF and PRES. A weaker notion of specificity based only on the satisfaction of PRES would not suffice. For example, it has been claimed that overt partitives are specific (De Hoop 1992; Enç 1991). The object DP in sentence (20a) is weakly specific. The speaker is asking the addressee to bring a group of several books from a presupposed set. In this respect, more than one group could satisfy the speaker’s request, as long as they belong to the presupposed set of books. Sentence (20b) requires a stronger notion of specificity, namely the combination of referentiality and presuppositionality and, as a consequence, the speaker is requesting a unique group of books. (20) a.

Quiero que me traigas varios de los libros want-I that me bring-you several of the books ‘I want you to bring me several of the books’ b. Quiero que me los traigas varios de los libros want-I that me them bring-you several of the books ‘I want you to bring me (a group of) several of the books’

Accusative doubling has also a “scope freezing” effect, so sentence (21) can only be construed in the object wide scope interpretation. (21) Tres profesores los engañaron a muchos de los estudiantes three professors them fooled A many of the students ‘Three professors fooled a group consisting of many of the students’ As observed by Montague (1969), the objects of intensional or opaque verbs are ambiguous. They can have a non-transparent or a transparent reading, the latter arising when the object scopes over the verb. (22) a. I am looking for a unicorn b. I want two secretaries In Spanish, the transparent reading requires the presence of the accusative doubling clitic. For example, (23) lacks the opaque reading. Similarly accusative doubling of a DP modified by a subjunctive relative clause is not possible. This follows from the fact that this type of subjunctive modification is intrinsically non-transparent (Rivero 1977).

TSL[v.20020404] Prn:29/08/2002; 16:38

F: PBNS18.tex / p.9 (343)

Constraint interaction at the semantics/pragmatics interface 

(23) Lo busco a un unicornio him look-for-I A a unicorn ‘I am looking for a unicorn’ (24) a.

Las busco a dos secretarias que saben them-fem look-for-I A two secretaries that know-ind-they Francés French ‘I am looking for two secretaries who speak (indicative) French’ b. *Las busco a dos secretarias que sepan Francés them look-for-I A two secretaries that know-subj-they French ‘I am looking for two secretaries who speak (subjunctive) French’

Variation and constraint satisfaction Dative doubling In the previous sections we have studied several constraints on Spanish accusative doubling. Dative clitic doubling is not sensitive to the constraints REF, and PRES, so there is not a contrast between the examples (25a, b). None of them violate the relevant semantic constraints. (25) a.

Les dieron un libro a los/esos/mis/todos los to-them gave a book to the/those/my/all the ‘They gave a book to the/those/my/all the students’ b. Les dieron regalos a muchos/varios/pocos to-them gave-they presents to many/several/few ‘They gave presents to many/several/few students’

estudiantes students estudiantes students

The animacy and the emphasis constraint do not have to be satisfied by dative doubling constructions either. Thus, dative clitic doubling can be considered a mark of syntactic agreement (case matching), wheras accusative clitic doubling is a mark of true semantic agreement. This property is related to two facts. First, since accusative clitics semantically select certain features, they are a marked and more stringent choice. Dative clitic doubling has become obligatory as the unmarked option in many dialects: sentences in which dative indirect objects are clitic doubled can be considered as the neutral variants. The only exception to this generalization is the use of le for certain types of semantic marking, such as “intensive” le in Mexican Spanish (Torres 2000), illustrated in (26). In this case, dative doubling satisfies REF, PRES, EMPH.

TSL[v.20020404] Prn:29/08/2002; 16:38

F: PBNS18.tex / p.10 (344)

 Javier Gutiérrez-Rexach

(26) Ándele, pues walk-cl-dat thus

Accusative doubling in Romanian and Greek The constraints that have been shown to be active in Spanish also have to be considered in the study of clitic doubling in Romanian and Greek. In Romanian, accusative doubling is obligatory when the object is preceded by the preposition pe. It always triggers also semantic agreement, in other words, it is not semantically inert. The constraints REF and PRES are satisfied. (27) L/I-am v˘azut pe Popescu/to¸ti/to¸ti copii him/them-have(I) seen PE Popescu/all/all children-the ‘I have seen Popescu/all/all the children’ Generalized existential quantifiers satisfying the intersectivity condition (Keenan 1996) may be doubled only when they are referential, that is to say, when they denote principal filters and satisfy REF. In the following examples from DobrovieSorin (1994), the objects are understood as referential. (28) a.

I-am dus pe dou˘a fran¸tuozaice la gar˘a them-have-I taken PE two Frenchwomen to-the station ‘I have taken two Frenchwomen to the station’ b. I-am p˘ac˘alit pe mul¸ti copii them-have-I fooled PE many children ‘I have fooled many children’

A final parallelism with the Spanish case is the doubling of indefinites in their strongly specific or principal filter denoting reading, as in (29), and their incompatibility with the subjunctive mood in a modifying relative clause (30). (29) O caut pe fat˘a de la noi din sat her am-I looking-for a girl from our village ‘I am looking for a girl from our village’ (30) a.

Caut un elev care s˘a-¸stie engleze¸ste look-for-I a student which speak-subj English ‘I am looking for a student who would be able to speak English’ b. *Il caut pe un elev care s˘a-¸stie engleze¸ste him look-for-I PE a student which speak-subj English

An important contrast between Spanish and Romanian surfaces in dative clitic doubling. Recall that we have shown above that Spanish dative clitic doubling dif-

TSL[v.20020404] Prn:29/08/2002; 16:38

F: PBNS18.tex / p.11 (345)

Constraint interaction at the semantics/pragmatics interface 

fers from accusative doubling in not being sensitive to any concrete semantic constraint. In Romanian, dative doubling is obligatory if the associated expression is referential and presuppositional. Thus, it also has to satisfy REF and PRES. The dative object in (31) lacks the weakly specific (partitive) and non-specific readings. (31) le caut˘a leac unor bloi incurabile cl-dat.pl search-he remedy some illness incurable ‘He searches remedies for incurable diseases’ In Greek, only expressions that are always referential may be doubled by an accusative clitic. Consequently, a contrast arises between Greek and Spanish existential quantifiers. In Spanish, these expressions may be doubled but in Greek they may not. According to Gutiérrez-Rexach (1999a), this shows that the referentiality restriction on accusative doubling is actually stronger in Greek and amounts to a co-intersectivity condition. Only co-intersective generalized quantifiers (universals, definites, proper names, etc. See Keenan 1996) are doubled in Greek. Thus, we may split the referentiality/principal filter constraint in two subconstraints: REFCOINT and REFINT/EXIST . In Greek, *REFEXIST is operative, and this explains the ungrammaticality of the doubled constructions in (32). (32) a. *Tin psaxno mia/kapja grammatea her look-for-I one/a secretary ‘I look for one/a secretary’ b. *Tous eksetase merikous apo tous asthenis them examined-he several of the patients ‘He examined several of the patients’ Anagnostopoulou and Giannakidou (1995) propose that clitic doubled objects have to be prominent, in Heim’s (1982) sense. In (33a), the non-doubled ton sigrafea is ambiguous. It may either refer to Arthur Miller (k = i) or to the (accommodated) author of the book that John read (k = j). The second option is not available under clitic doubling. Ton sigrafea in (33b) can only be understood as referring to Arthur Miller (k = i). (33) a.

O Jannis diavase [j ena vivlio jia ton [i Arthur Miller]], enthusiastike, ke thelise na gnorisi [k ton sigrafea] apo konta John read [j a book about [i Arthur Miller]], he got very enthusiastic, and he wanted to get to know [k the author] b. O Jannis diavase [j ena vivlio jia ton [i Arthur Miller]], enthusiastike, ke thelise na ton gnorisi [i ton sigrafea] apo konta John read [j a book about [i Arthur Miller]], he got very enthusiastic, and he wanted to get to cl know [k the author]

TSL[v.20020404] Prn:29/08/2002; 16:38

F: PBNS18.tex / p.12 (346)

 Javier Gutiérrez-Rexach

In sum, clitic doubling in Greek requires that the entity referred to by the noun phrase have been explicitly introduced in the previous discourse. We can call this constraint PROM. Since what is asserted cannot be presupposed, this condition is in conflict with the Presuppositionality condition that we formulated before. In Greek, PROM is ranked higher than PRES (PROM  PRES) but in Spanish, PRES is ranked higher (PRES  PROM), as was discussed above. Thus, in (34) the anaphoric linking in which j = k is the preferred one. The author that Juan wants to get to know is the author of the book, not Arthur Miller. (34) Juan leyó [j un libro sobre [i Arthur Miller]], se entusiasmó, y quiso conocerlo [k al autor] John read [j a book about [i Arthur Miller]], he got very enthusiastic, and he wanted to get to cl know [k the author] One consequence of the above contrast is that in Greek doubling of accommodative definites (35a), weak definites (Poesio 1994) (35b), novel proper names/definites (those mentioned for the first time) (35c) is not possible. All these expressions relate to presupposed elements which are not prominent. efaga tis sokolates (35) a. *I Maria mou estile glika; tis the Mary to-me sent sweets; them ate-I the chocolates amesos immediately ‘Mary sent me sweets; I immediately ate the chocolates’ b. *Xtes ton gnorisa ton fititi enos diasimou glossologou yesterday him met-I the student of-a famous linguist ‘ I met the student of a famous linguist yesterday’ c. *Xtes onirevtika oti ton gnorisa ton Clinton s’ena party yesterday dreamt-I that him met the Clinton in-a party ‘Last night I dreamt that I met Clinton at a party’ In Spanish, where (PRES  PROM), doubling of all these types of definites is possible. They all require that the doubled element refer to a presupposed entity. (36) a.

María me envió dulces; me los comí los chocolates Mary to-me sent sweets; cl-dat them ate-I the chocolates inmediatamente immediately ‘Mary sent me sweets; I immediately ate the chocolates’ b. Ayer lo conocí al estudiante de un famoso linguista yesterday him met-I A-the student of a famous linguist ‘I met the student of a famous linguist yesterday’

TSL[v.20020404] Prn:29/08/2002; 16:38

F: PBNS18.tex / p.13 (347)

Constraint interaction at the semantics/pragmatics interface 

c.

Fíjate, ayer soñé que lo conocí a Clinton en look, yesterday dreamt-I that him met A Clinton in una fiesta a party ‘Look, yesterday I dreamt that I met Clinton at a party’

Economy of context There is an additional constraint on dative doubling which accounts for some puzzling patterns in Spanish doubling. We can formulate this Economy of Context (ECON) constraint as follows: Let Qi , Qj be two quantifiers satisfying REF such that they would yield an identical interpretation in a clitic doubling construction. Thus, Qi , Qj have to be principal filters and denote the same function. Then, if one is an inherently contextually restricted function and the other is not, only the contextually restricted quantifier can be doubled. A contextually restricted function is one that is specified as being restricted to a context set. This property is normally highlighted by morphological marking. In Spanish, the inherently restricted determiners are typically complex determiners formed by a basic determiner function and the determiner the. The existence of the constraint ECON explains why several Spanish quantifiers, apparently denoting the same function, nevertheless exhibit differential patterns with respect to doubling. The contrasting complex and simple determiners are todos los ‘all the’ and todo ‘every’; cada uno de los ‘each one of the’ vs. cada ‘each’; and algunos ‘some pl’ vs. unos ‘a-pl.’ Let us consider for example the universal determiners todos los ‘all the’ and todo ‘every’. We have claimed that the first one is inherently contextually restricted whereas the second one is not. What this means is that no matter what the nature of the surrounding discourse is, todos los is understood to be a contextually or discourse linked expression whereas todo can only be contextually dependent when it is coerced (Pustejovsky 1995) under very strict conditions. For instance, both sentences in (37) express a universal statement about men. Sentence (37a) can also be used when the speaker is referring to a particular group of men whereas this is not possible in the case of (37b). The latter sentence is understood as a generic statement about males. (37) a.

Todos los hombres aman a una mujer all-masc.pl the men love A a woman ‘All the men love a woman’ b. Todo hombre ama a una mujer every man loves A a woman ‘Every man loves a woman’

TSL[v.20020404] Prn:29/08/2002; 16:38

F: PBNS18.tex / p.14 (348)

 Javier Gutiérrez-Rexach

The discourses in (38) show the contrast between these determiners even more clearly. Recall that todos los denotes a function inherently restricted to a context set. In (38a), the set of twelve students who came becomes the context set of the determiner and the discourse is felicitous. On the other hand, todo in (38b) lacks a context set argument. The discourse becomes infelicitous because the context provided by the first sentence is not compatible with a universal/generic statement about Europeans. (38) a.

Vinieron doce estudiantes. Todos los europeos came-they twelve students. All the Europeans sonreían were-smiling ‘Twelve students came in. All the European ones were smiling’ b. Vinieron doce muchachos. *Todo europeo sonreía came-they twelve students. Every European were-smiling ‘Twelve students came in. Every European was smiling’

The other contrasting couples of simple and complex determiners considered above exhibit the same behavior with respect to these tests. According to the constraint ECON, only the quantifiers headed by the complex determiners may be doubled. This prediction is borne out, as the following constrasts show: (39) a.

Los vi a todos los hombres them saw-I A all-masc.pl the men ‘I saw all the men’ b. *Lo vi a todo hombre him saw-I A every man

(40) a.

Lo vi a cada uno de los estudiantes him saw-I A each one of the students ‘I saw each one of the students’ b. *Lo vi a cada estudiante him saw-I A each student

Similarly, algunos ‘some pl’ is a contextually restricted indefinite determiner, but unos ‘a-pl’ is not (Gutiérrez-Rexach 2001b). As expected, only DPs headed by algunos can undergo accusative doubling: (41) a.

Los vi a algunos hombres them saw-I A some-masc.pl men ‘I saw some men’ b. *Los vi a unos hombres them saw-I A a-pl men

TSL[v.20020404] Prn:29/08/2002; 16:38

F: PBNS18.tex / p.15 (349)

Constraint interaction at the semantics/pragmatics interface 

The constraint ECON is ranked lower than REF. In all the above examples, the doubled quantifiers are principal filters. If ECON were ranked higher than REF, it would be possible to double contextually restricted expressions even if they do not denote principal filters. In Porteño Spanish (Suñer 1988), *REFEXIST is active and doubling of quantifiers headed by intersective or generalized existential determiners is not possible (42a). Nevertheless, ECON is ranked higher than *REFEXIST (ECON  *REFEXIST ), so a violation of *REFEXIST is possible in (42b) in order to prevent a violation of the higher ranked ECON. (42) a. *Los vi a tres them saw-I A three ‘I saw three men’ b. Los vi a tres them saw-I A three

hombres men de los hombres of the men

The activity of the ECON constraint can also be detected in the domain of negative quantifiers. The determiners ninguno de los ‘none of the’ and ningún ‘no’ and the quantifiers ninguno ‘nobody’ vs. nadie ‘nobody’ contrast again in that the former denotes a contextually restricted function. By applying the same reasoning as before, we explain the following contrasts as the result of the satisfaction or violation of ECON. (43) *No lo he visto a ningún político not him have-I seen A no politician ‘I have not seen any politician’ (44) No lo he visto a ninguno de los políticos not him have-I seen A none of the politician ‘I have not seen any of the politicians’ (45) a.

No lo he visto a not him have-I seen A ‘I have not seen anybody’ b. ??No lo he visto a not him have-I seen A ‘I have not seen anybody’

ninguno nobody nadie noone

Clitic doubling of interrogative quantifiers In this section, I will consider accusative doubling in Spanish and Romanian questions. I will treat wh-words in general as interrogative generalized quantifiers or determiners (Gutiérrez-Rexach 1997). Consequently, clitic doubling in questions

TSL[v.20020404] Prn:29/08/2002; 16:38

F: PBNS18.tex / p.16 (350)

 Javier Gutiérrez-Rexach

has to be treated as a case of pronoun-quantifier interaction in parallel with doubling in declarative sentences. Evidence for this parallelism comes from the fact that Spanish dative clitic doubling does not impose any semantic restriction on the associated wh-word: (46) a.

¿A quién le diste un libro? to who to-him gave-you a book ‘To whom did you give a book?’ b. ¿A cuál le molestan estos ruidos? A whicho to-him bothers these noises ‘Which one is bothered by these noises?’

Accusative doubling shows sensitivity to the nature of the associated interrogative quantifier. Spanish wh-words quién ‘who’ and qué ‘what’ contrast with cuál ‘which’ in that the latter denotes a contextually restricted determiner or generalized quantifier function (Gutiérrez-Rexach 1997). In (47), the variant with cuál expresses a query about a presupposed set of individuals (the context set); the variant with quién is not subject to this restriction. (47) ¿Quién/Cuál vino a la fiesta? who/which-one came to the party? Only the interrogative expression inherently restricted to a context set can be doubled by an accusative clitic, as (48) shows: (48) a. *¿A quién lo viste? A who him saw-you ‘Who did you see?’ b. *¿Qué lo compraste? what it bought-you ‘What did you buy?’ c. ¿A cuál de ellos lo viste? A which of them him saw-you ‘Which one of them did you see?’ d. ¿A cuál lo han matado? A which him have-they killed ‘Which one did they kill?’ This contrast constitutes strong evidence of the activity of ECON in the interrogative domain. The functions denoted by qué and quién only differ from cuál in that they are not restricted to a context set. Thus, if they are associated with an accusative clitic, a violation of ECON arises. The constraint PRES is also satisfied vacuously when ECON is satisfied because the context set represents pre-

TSL[v.20020404] Prn:29/08/2002; 16:38

F: PBNS18.tex / p.17 (351)

Constraint interaction at the semantics/pragmatics interface

supposed information. The nature of the information exchange that is typical in question-answer dialogues renders the constraint REF inactive. Interrogative generalized quantifiers cannot be principal filters because in information questions the speaker is not referring to a particular individual or group of individuals. Rather, he is asking to identify one. There is dialectal variation with respect to how strong ECON is. For instance, in Porteño Spanish, ECON requires overt partitivity in all doubled quantifiers: (49) a.

¿A cuál de ellos viste? A which of them saw-you ‘Which one of them did you see?’ b. ??¿A cuál han matado? A which have-they killed ‘Which one did they kill?’

Free choice wh-elements also satisfy ECON. The free choice item cualquiera ‘whichever’ may be doubled by an accusative clitic whereas quienquiera ‘whoever’ may not. (50) a.

En este departamento, lo admiten a in this department, him admit A ‘In this department, they admit anyone’ b. *En este departamento, lo admiten a in this department, him admit A ‘In this department, they admit anyone’

cualquiera whichever quienquiera who-pl-ever

Accusative doubling in Romanian interrogatives also satisfies ECON. The interrogative quantifier expressions cine ‘who’ and ce ‘what’ do not denote contextually restricted functions, whereas care ‘which’ does. As was the case in Spanish, only the latter quantifier may be doubled, as shown in (52). (51) a. *Pe cine l-ai v˘azut? PE who him-have(you) seen ‘Who did you see?’ b. *Ce (roman) l-ai citit? what (novel) it-have(you) read ‘What (novel) did you read?’ (52) a.

Pe care l-ai v˘azut? PE which him-have(you) seen ‘Which did you see?’ b. Pe care b˘aiat l-ai v˘azut? PE which boy him-have(you) seen ‘Which boy did you see?’



TSL[v.20020404] Prn:29/08/2002; 16:38

F: PBNS18.tex / p.18 (352)

 Javier Gutiérrez-Rexach

Conclusions In this paper, I have argued that the problem of the association between a clitic and a generalized quantifier can be succesfully accounted for in an Optimality Theoretic framework. The essence of my proposal is that the nature of the constraints at play is semantic, and that the predicted restrictions take place in the semantics/pragmatics interface. Taking the arguments and data considered in GutiérrezRexach (1999a) as a departure point, I have formulated five main constraints on accusative doubling (REF, PRES, EMPH, ANIM, and ECON) and have shown that this type of marking can be considered truly semantic. On the other hand, dative clitic doubling is not sensitive to these constraints and cannot be considered a proper instance of semantic marking. The account extends successfully to clitic doubling in interrogative environments, where ECON is claimed to play a more significant role. Finally, I have discussed clitic doubling in Greek and Romanian and argued that cross-linguistic variation follows from different patterns of activity in the set of constraints discussed or by differences in constraint ordering. This confirms some of the main tenets of Optimality Theory and shows that the explanation of linguistic variation based on constraint interaction and satisfaction postulated by this approach is relevant for the analysis of semantic/pragmatic phenomena.

References Anagnostopoulou, E., & Giannakidou, A. (1995). Clitics and prominence, or why specificity is not enough. Chicago Linguistic Society 31: Parasession on Clitics. Archangeli, D., & Langendoen, T. (1997). Optimality Theory. An Overview. Oxford: Basil Blackwell. Barwise, J., & Cooper, R. (1981). Generalized quantifiers and natural language. Linguistics and Philosophy, 4, 159–219. Blütner, R. (1999). Some aspects of optimality in natural language interpretation. Ms. Berlin University. Capone, A. (2000). Dilemmas and Excogitations. An essay on modality, clitics and discourse. Messina: Armando Siciliano. Dalrymple, M., Shieber, S., & Pereira, F. (1991). Ellipsis and higher-order unification. Linguistics and Philosophy, 14, 399–452. De Hoop, H. (1992). Case Configuration and Noun Interpretation. Ph.D. dissertation, University of Groningen. Diesing, M. (1992). Indefinites. Cambridge, MA: MIT Press. DobrovieSorin, C. (1994). The Syntax of Romanian. Berlin/New York: Walter De Gruyter. Enç, M. (1991). The semantics of specificity. Linguistic Inquiry, 22, 1–26. Fox, D. (1995). Economy and scope. Natural Language Semantics, 3, 283–341.

TSL[v.20020404] Prn:29/08/2002; 16:38

F: PBNS18.tex / p.19 (353)

Constraint interaction at the semantics/pragmatics interface 

Gutiérrez-Rexach, J. (1997). Questions and generalized quantifiers. In A. Szabolcsi (Ed.), Ways of Scope Taking (pp. 409–452). Dordrecht: Kluwer. Gutiérrez-Rexach, J. (1999a). The formal semantics of clitic doubling. Journal of Semantics, 16, 315–380. Gutiérrez-Rexach, J. (1999b). Scope parallelism and the interpretation of ellipsis at the syntax/semantics interface. Syntaxis, 2, 51–79. Gutiérrez-Rexach, J. (2001a). Interface conditions and the semantics of argument clitics. In J. Gutiérrez-Rexach & L. Silva-Villar (Eds.), Current Issues in Spanish Syntax and Semantics (pp. 107–142). Berlin/New York: Mouton-De Gruyter. Gutiérrez-Rexach, J. (2001b). The semantics of Spanish plural existential determiners and the dynamics of judgment types. Probus, 13, 113–154. Heim, I. (1982). The Semantics of Definite and Indefinite Noun Phrases. Ph.D. dissertation, University of Massachusetts, Amherst. Hendricks, P., & de Hoop, H. (2001). Optimality theoretic semantics. Linguistics and Philosophy, 24, 1–32. Kager, R. (1999). Optimality Theory. Cambridge: Cambridge University Press. Karttunen, L. (1977). The syntax and semantics of questions. Linguistics and Philosophy, 1, 3–44. Keenan, E. (1989). Semantic case theory. In R. Bartch et al. (Eds.), Semantics and Contextual Expression (pp. 33–57). Dordrecht: Foris. Keenan, E. (1996). The semantics of determiners. In S. Lappin (Ed.), Handbook of contemporary semantic theory (pp. 41–64). Oxford: Basil Blackwell. Montague, R. (1969). On the nature of certain philosophical entities. The Monist, 53, 159–194. Poesio, M. (1994). Weak definites. In Proceedings of Semantics and Linguistics Theory, IV (pp. 282–299). Ithaca, NY: CLC Publications. Postal, P. (1966). On so-called “pronouns” in English. In D. Reibel & S. Schane (Eds.), Modern studies in English (pp. 201–224). Englewood Cliffs, NJ: Prentice Hall. Prince, A., & Smolensky, P. (1993). Optimality Theory. (MIT Press, forthcoming). Prince, A., & Smolensky, P. (1997). Optimality: From neural networks to universal grammar. Science, 275, 1604–1610. Pustejovsky, J. (1995). The Generative Lexicon. Cambridge, MA: MIT Press. Rivero, M. (1977). Estudios de Gramática Generativa del Español. Madrid: Cátedra. Sportiche, D. (1996). Clitic constructions. In J. Rooryck & L. Zaring (Eds.), Phrase Structure and the Lexicon (pp. 213–276). Dordrecht: Kluwer. Suñer, M. (1988). The role of agreement in clitic-doubled constructions. Natural Language and Linguistic Theory, 6, 391–434. Suñer, M. (1991). Two properties of clitics in clitic-doubled constructions. In J. Huang & R. May (Eds.), Logical Structure and Linguistic Structure: Crosslinguistic Perspectives (pp. 233–251). Dordrecht: Kluwer. Torrego, E. (1995). On the nature of clitic doubling. In H. Campos & P. Kempchimsky (Eds.), Evolution and Revolution in Linguistic Theory. Washington, DC: Georgetown University Press. Torres, R. (2000). Intensive “le” usage in Mexican Spanish. Talk at LSRL 30.

TSL[v.20020404] Prn:29/08/2002; 16:38

F: PBNS18.tex / p.20 (354)

 Javier Gutiérrez-Rexach

Uriagereka, J. (1995). Aspects of the syntax of clitic placement in Western Romance. Linguistic Inquiry, 26, 79–124. Westerståhl, D. (1985). Determiners and context sets. In J. van Benthem & A. ter Meulen (Eds.), Generalized Quantifiers in Natural Language (pp. 45–71). Dordrecht: Foris. Zeevat, H. (1999). Explaining presupposition triggers. Proceedings of the Twelfth Amsterdam Colloquium, 19–24.

TSL[v.20020404] Prn:4/09/2002; 13:56

F: PBNS19.tex / p.1 (355)

Cross-language commutation tests and their application to an error-prone contrastive problem Ger. einige, Fr. quelques, Sp. algunos Eva Lavric Institut für Romanische Sprachen, Wirtschaftsuniversität Wien

As a linguist, and particularly as a contrastive linguist, I would like to start this contribution with three cheers to our students: Thank God we have students; thank God the students ask questions and thank God the students make errors! Because the errors and the questions about what would have been right and why, not only provide the contrastive linguistics branch of research with a first-class right to exist but also with a continuous stream of ideas of what to examine more minutely from what point of view. The investigation at hand thus starts with errors, errors of German-speaking students producing texts in business French, errors in particular areas of nominal determination and especially in the field of indefinite plural determiners. Those errors serve as the starting point of an investigation into the three languages German, French and Spanish, with a short look at the English language.1 The question how to carry out such a contrastive semantic investigation, i.e. what methods to apply to accomplish a satisfactory description of a particular area allowing for the deduction of rules suitable for teaching – that is a question which exceeds mere error analysis and which calls for the whole range of linguisticsemantic methodology. Personally, I see myself as part of a structuralist tradition as e.g. represented by Coseriu,2 Vater,3 Schifko.4 In addition to standard linguistic methods, there are of course special contrastive methods, in particular the translation comparison, which yields valuable results but which, however, did not seem sufficient for my question concerning determiner semantics. I would like to exemplify why this is the case. And my first example just happens to be an error example, and a three-fold one at that:5

TSL[v.20020404] Prn:4/09/2002; 13:56

F: PBNS19.tex / p.2 (356)

 Eva Lavric

(1) Le premier impact [d’un accident] est celui du point de vue de *ces institutions qui doivent financer *ces frais qui sont directement liés aux accidents. *Ces frais qui se trouvent à gauche sont les frais directement liés à l’accident This is of course about determiners in noun phrases with restrictive relative clauses. Both the explanation and the respective rule are simple; what I would like to point out in this example (which is located halfway between syntax and semantics) is the way how to get there. This way is via another example, and it is that example by means of which I want to exemplify my semantic method: (2) J’étais entré dans le bonheur, qui est de faire le métier qu’on aime Ich genoß nun das Glück, das darin besteht, daß man jene Arbeit macht, die man gerne tut Comencé a disfrutar de la felicidad que supone dedicarse a la profesión que a uno le gusta (HPP 180–183) The example is taken from the corpus of my postdoctoral thesis,6 and one can see that this corpus is trilingual: It consists of texts with translations (half of them literary and the other half everyday texts). In this case, the original text is in French – and therefore comes first – the translations are into German and Spanish. (The other two directions of translation are of course also represented in the corpus.) The applied method is first and foremost a translation comparison.7 Example (2) is relevant for our first error example in so far as it includes a noun phrase with a restrictive relative clause and in its present form also provides the explanation of the error: In German, a restrictive relative clause can be signaled by a demonstrative determiner, namely by jener; the student has by analogy tried to do the same in French. One can already provide the learner with a rule of thumb here: In French, restrictive relative clauses cannot be signaled by a demonstrative determiner, one has to use the definite article. Linguistically, however, this rule is still unsatisfactory because it does not completely and explicitly represent the situation in both languages; moreover, Spanish is still missing to render the comparison a triple one. How is a restrictive relative clause signaled in Spanish? Example (2) shows: by the definite article, as in French. There is a parallel between the two Romance languages, which are in opposition to German, at least in example (2). Unfortunately however, this result is totally incorrect: mere translation comparison has led us astray in this case. With a mere translation comparison, one is far too much subject to the coincidences of translation, and that is why I system-

TSL[v.20020404] Prn:4/09/2002; 13:56

F: PBNS19.tex / p.3 (357)

Cross-language commutation tests 

atically supplement the translation comparisons by commutation tests in all three languages. This means that in the same slot, I test all other determiners of the relevant language and note those which lead to the same meaning as the original text. I adapted this method from Vater (1963/1979), adding the contrastive dimension myself.8 The commutation tests restore the whole range of paradigmatic possibilities in all languages involved; under this method, one not only compares what is there in the text, but also what could have been there in this slot of the text while maintaining its meaning. The result for our example (2): (2 ) J’étais entré dans le bonheur, qui est de faire le métier qu’on aime Ich genoß nun das Glück, das darin besteht, daß man jene Arbeit macht, die man gerne tut daß man diejenige Arbeit macht, die man gerne tut daß man die Arbeit macht, die man gerne tut Comencé a disfrutar de la felicidad que supone dedicarse a la profesión que a uno le gusta a aquella profesión que a uno le gusta There are three forms yielding the same meaning in the same slot in German, two in Spanish and only one in French. And thus we can with the help of this characteristic example lay down all possibilities to signal a restrictive relative clause in the systems of the three languages: 1. In all three languages, there is the option of the definite article. 2. In addition, there are possibilities with demonstrative determiners in German and Spanish: Sp. aquel, Ger. jener and derjenige.9 3. In French, there is no possibility to signal a restrictive relative clause by a demonstrative determiner.10 This constraint is probably related to the fact that French does not really have a distal demonstrative,11 while both Ger. jener and Sp. aquel are distal demonstratives.12 So this would be the precise result of an empirically-based interlingual systems comparison, which the combination of translation comparison and commutation tests allows for. Although the rule of thumb for the student remains the same, we can now predict that e.g. German-speaking learners of Spanish will not make any interference errors with such examples, but that their transfer attempts will be successful.13 This brings us to the main part of this paper, the plural indefinites used for small numbers, i.e. Fr. quelques, Sp. algunos and Ger. einige. Again, we start with a few typical error examples:

TSL[v.20020404] Prn:4/09/2002; 13:56

F: PBNS19.tex / p.4 (358)

 Eva Lavric

(3) Dans *quelques classe(s) d’école 70% des élèves ont besoin de leçons particulières (4) Son électorat est très hétérogène, mais *quelques groupes sont particulièrement attirés (5) *Quelques villes ont un certain pouvoir d’attraction, c’est la raison pour laquelle beaucoup d’entreprises s’y installent volontairement What these examples have in common is that the learners use the Fr. quelques in contexts in which it is not ‘appropriate’. By way of correcting, one could insert Fr. certains in all the examples; however, what the students actually ‘meant’ was not its German equivalent manche, but the German form einige, which was rashly equated with quelques. It is furthermore noteworthy that similar errors do not occur with Sp. algunos. We will therefore have to look at the French and Spanish equivalents of Ger. einige. Our ‘pre-investigation’ by means of error analysis suggests that Sp. algunos and Ger. einige could be synonyms, whereas Fr. quelques shows a narrower range of possible applications. The relation to certains furthermore suggests that the opposition [+/–specificity] plays a certain role regarding these differences. (What I mean by [+/–specificity] is the opposition between e.g. Engl. certain and any, i.e. the opposition constituted by random selection versus a selection of the referent which is governed by (knowledge of) its identity).14 In the linguistic literature, Fr. quelques is often compared with plusieurs (e.g. Gondret 1976; K˛esik 1978; Gaatone 1991). This, however, is of no help with our error examples, since it is not the interlingual problem concerning German. Here, errors occur because of the learners’ hypothesis that Fr. quelques is the equivalent of Ger. einige. This hypothesis is not totally unfounded, however. It is definitely valid in a number of cases (and the Spanish versions show that algunos is indeed the Spanish equivalent): (6) Nach hinten wird das Gelände von einigen Eichen begrenzt Vers le fond, quelques chênes bornent l’espace visible Hacia detrás, el terreno está delimitado por algunos robles (Wal 22/25/20) (7) Quelques femmes célèbres sont censées symboliser au Parlement et au gouvernement l’égalité des droits entre les sexes Einige berühmte Frauen gelten in Parlament und Regierung als symbolhafte Verkörperung der Gleichberechtigung von Mann und Frau Algunas mujeres célebres simbolizan en el Parlamento y en el gobierno la igualdad de derechos entre los sexos (Hen 9–16)

TSL[v.20020404] Prn:4/09/2002; 13:56

F: PBNS19.tex / p.5 (359)

Cross-language commutation tests 

The forms quelques, algunos and einige have in common the semantic feature [small number],15 which means that the number of referents is lower than expected, lower than an objective or at least intersubjective standard. Since we are moreover interested in the aspect of [+/–specificity], I have chosen examples 6 and 7 to illustrate both poles of the opposition: (6) is non-specific and (7) is specific. Surprisingly, however, Fr. quelques can be found in both examples, which means it is not marked for the opposition [+/–specific]. The semantic restrictions of Fr. quelques compared with Ger. einige can thus not be located in this dimension. Therefore, we still have to look for authentic examples in which Ger. einige can occur, while Fr. quelques cannot. In my corpus, few but very characteristic examples can be found, the most meaningful of which is the following: (8) Ce truculent article ravira bien des présidents d’organisations agricoles! *quelques présidents Dieser gesalzene Artikel wird einige Präsidenten landwirtschaftlicher Organisationen erfreuen . . . ! ¡Este artículo encantará a algunos presidentes de organizaciones agrícolas . . . ! (Pre) A particular semantic feature can be tested against this example, namely the feature [considerable number]. While the opposition [small number] versus [large number] represents the objective aspect of the ‘number’ dimension, the opposition [considerable number] versus [inconsiderable number] denotes a subjective evaluation of the number on part of the speaker.16 The two aspects can be well differentiated in the semantics of the forms Fr. plus d’un / Sp. más de un, since both aspects are in a state of conflict here: Plus d’un /más de un denote an objectively small but subjectively considerable number. Fr. bien des signifies a large and at the same time considerable number.17 Example (8) shows that Fr. quelques cannot be used to refer to a considerable number and will therefore be substituted with bien des; quelques thus shows the feature [inconsiderable number], and that is exactly the difference to Ger. einige and Sp. algunos. Because einige and algunos are neutral concerning the opposition [+/–considerable], they can, as in the majority of examples (6)–(7), denote an inconsiderable number, but are also used to refer to a considerable number, as in example (8).

TSL[v.20020404] Prn:4/09/2002; 13:56

F: PBNS19.tex / p.6 (360)

 Eva Lavric

[SMALL NUMBER] [CONSIDERABLE NUMBER]

einige algunos

[INCONSIDERABLE NUMBER]

quelques

From a didactic viewpoint, it may make sense to ask whether there is a determiner in German which, just like Fr. quelques, represents a combination of the features [small number] and [inconsiderable number]. There is indeed such a determiner: It is the form ein paar. This enables us to include Spanish again, because Sp. un par de is a synonym of Ger. ein paar, i.e. they can be applied in exactly the same contexts; e.g. in the following, which also confirms the equivalence with Fr. quelques: (9) Combien de prétendus savants sont aujourd’hui capables de faire une règle de trois ? Quelques mathématiciens ! Wie viele, die sich Wissenschaftler nennen, sind heutzutage noch fähig, einen Schluß zu rechnen? Ein paar Mathematiker! ¿Cuántos pseudo-científicos son capaces hoy en día de hacer una regla de tres? ¡Un par de matemáticos! (Wil 92) One can see that ein paar / un par de are German and Spanish determiners which represent the same features as quelques. Is it therefore possible to tell the learners to use Fr. quelques only when they could also use Ger. ein paar? It is possible, albeit not one hundred percent true. Because there is a certain group of examples by which quelques exceeds ein paar / un par de. The commutations for examples (6) and (7) are particularly meaningful here: (6 ) Nach hinten wird das Gelände von einigen Eichen begrenzt ein paar Eichen Vers le fond, quelques chênes bornent l’espace visible Hacia detrás, el terreno está delimitado por algunos robles un par de robles (7 ) Quelques femmes célèbres sont censées symboliser au Parlement et au gouvernement l’égalité des droits entre les sexes

TSL[v.20020404] Prn:4/09/2002; 13:56

F: PBNS19.tex / p.7 (361)

Cross-language commutation tests 

Einige berühmte Frauen gelten in Parlament und Regierung als symbolhafte Verkörperung der Gleichberechtigung von Mann und Frau ??Ein paar berühmte Frauen Algunas mujeres célebres simbolizan en el Parlamento y en el gobierno la igualdad de derechos entre los sexos ??Un par de mujeres célebres

Example (6) is [–specific], and ein paar / un par de fit well, example (7) is [+specific], and it excludes ein paar / un par de. Ein paar / un par de are thus marked as [–specific]. We therefore have to enlarge our table by the ‘specificity’ dimension: [SMALL NUMBER]

‘number’ dim.



‘specificity’ dim.  [+ SPEC.] [– SPEC.]

[CONSIDERABLE NUMBER]

einige algunos

[INCONSIDERABLE NUMBER]

quelques ein paar un par de

If we now want to briefly include the English language, we can try to translate our example sentences into this language and test certain commutations. The obvious form for a comparison with einige/quelques/algunos and ein paar / un par de resp., is of course Engl. a few, which signifies [small number] just like all the other examined forms. But in which variant? Is a few perhaps an equivalent of ein paar / un par de? Example (9) at least would suggest just that: (9 ) How many of those who call themselves scientists are nowadays able to do a rule-of-three? A few mathematicians! We see that a few can occur in [–considerable] examples. To be equivalent to ein paar / un par de, it would also have to signify [–specific]. So let’s check examples (6) and (7): (6 ) The back of the area is bordered by a few oaks (7 ) A few famous women in Parliament and in Government are deemed as token embodiements of equal rights of men and women Engl. a few is possible in both examples and is thus not marked for the opposition [+/–specific]. It could therefore be an equivalent of Fr. quelques or Ger. einige.

TSL[v.20020404] Prn:4/09/2002; 13:56

F: PBNS19.tex / p.8 (362)

 Eva Lavric

Let’s check its behaviour with regard to example (8), which tests compatibility with [+considerable]: (8 ) This malicious article will be to the liking of *a few presidents of agricultural organisations A few completely changes the meaning and can thus not be inserted here. From this it can be deduced that it behaves more or less like Fr. quelques, which represents the features [small number] and [–considerable]. However, a few can also express [+considerable] if a relevant modifier is added:18 (8 ) This malicious article will be to the liking of quite a few presidents of agricultural organisations Views may differ as to how to interpret this possibility: A few can either be regarded as [–considerable] and quite a few as [+considerable], so that the two terms are in privative opposition; or a few is (like Ger. einige and Sp. algunos) not marked for the opposition [+/–considerable], so that quite a few only topicalises a possible facet of meaning – which would mean that this is a case of an inclusive opposition. In this case, however, one would have to explain why a few alone ‘fits’ well in example (9) and not at all in example (8), i.e. why without a modifier it tends towards [–considerable], whereas Ger. einige and Sp. algunos actualize [+considerable] in the respective examples. It can be seen that even with such seemingly simple (pseudo-)equivalences each language triggers completely new questions. After this short excursus into English we now return to German, French and Spanish and will, in the last part of this paper, enlarge our semantic table by several additional categories, which enable us to include further forms of determiners: If we e.g. want to locate the specific determiners Fr. certains / Sp. ciertos / Ger. manche we have to exceed the realms of [small number], because certains/ ciertos/manche are not specified according to number.

‘number’ dim. [LARGE NUMBER]

[SMALL NUMBER]



‘specificity’ dim.  [+ SPEC.] certains / ciertos manche [– SPEC.]

[CONSIDERABLE NUMBER]

einige algunos

[INCONSIDERABLE NUMBER]

quelques ein paar un par de

TSL[v.20020404] Prn:4/09/2002; 13:56

F: PBNS19.tex / p.9 (363)

Cross-language commutation tests 

In this context, I provide my students with the following rule of thumb: Einige in the sense of manche = certains; einige in the sense of ein paar = quelques. This rule of thumb is valid for certain areas and would e.g. have been sufficient to avoid the errors in examples (3) to (5). To be precise, however, one would have to say that Fr. quelques as opposed to Ger. einige conveys, as an additional feature, a subjectively negative evaluation of the number, i.e. the feature [inconsiderable number]. For the sake of completeness, we should now closer examine Sp. algunos. It showed solidarity with Ger. einige in all previous examples and could thus be considered a synonym of einige. This is not exactly true, however, because the range of algunos is actually broader than that of einige. Algunos can also occur in examples in which the referents exist merely hypothetically, i.e. in a possible world, whereas einige is restricted to the existence of referents in the real world, i.e. to actual existence:19 (10) Bringt er irgendwelche Ideen vor, ist man automatisch dagegen *einige Ideen S’il présente des idées, on est automatiquement contre *quelques idées Si expresa algunas ideas, se le lleva automáticamente la contraria (Pre) One can see that in the relevant examples, the German equivalent of algunos is the form irgendwelche, which refers to [–specific] referents, no matter whether they exist in possible worlds or simply in the real world.20 For a final table, the ‘specificity’ dimension therefore has to be broadened by a worlds aspect ([RW] = existence in the real world; [PW] = existence only in a possible world):

‘number’ dim. [LARGE NUMBER]

[SMALL NUMBER]



‘specificity’/ ‘worlds’ dim.  [+ SPEC.] certains / ciertos manche

[CONSIDERABLE NUMBER]

einige

[RW]

quelques ein paar un par de

[– SPEC.] irgendwelche [PW]

[INCONSIDERABLE NUMBER]

algunos

TSL[v.20020404] Prn:4/09/2002; 13:56

F: PBNS19.tex / p.10 (364)

 Eva Lavric

In all its complexity, this table shows how real equivalences – e.g. certains/ciertos/ manche or ein paar / un par de – are actually an interlingual exception. The plural indefinites of the three examined languages overlap in specific functional areas, but their borders are divergent in the different languages. In particular, there is actually a relation of inclusion between the pseudo-equivalents algunos, einige and quelques: quelques ⊂ einige ⊂ algunos. The investigation clearly shows how the three languages structure reality in a comparable, but not in the same way. The table developed here shows the limits of certain contrastive rules of thumb, because in the subtlety of analysis, it far exceeds the necessities of error explanation. It shows the problematic forms to be parts of a complex system, parts which can be compared in minute detail by means of a table divided according to semantic features and oppositions. By applying the combined contrastive method, it is possible to generate a comprehensive map of determiner meanings in three different languages. French ‘number’ dim. [LARGE NUMBER]

[SMALL NUMBER]



‘specificity’/ ‘worlds’ dim. 

[CONSIDERABLE NUMBER]

[+ SPEC.] certains

[INCONSIDERABLE NUMBER]

quelques

[RW] [– SPEC.] [PW]

Spanish ‘number’ dim. [LARGE NUMBER]

[SMALL NUMBER]



‘specificity’/ ‘worlds’ dim. 

[CONSIDERABLE NUMBER]

[INCONSIDERABLE NUMBER]

[+ SPEC.] ciertos [RW]

un par de

[– SPEC.] [PW]

algunos

TSL[v.20020404] Prn:4/09/2002; 13:56

F: PBNS19.tex / p.11 (365)

Cross-language commutation tests 

German ‘number’ dim. [LARGE NUMBER]

[SMALL NUMBER]



‘specificity’/ ‘worlds’ dim.  [+ SPEC.] manche

[CONSIDERABLE NUMBER]

[INCONSIDERABLE NUMBER]

einige

[RW]

ein paar

[– SPEC.] irgendwelche [PW]

French, Spanish and German ‘number’ dim. [LARGE NUMBER]

[SMALL NUMBER]



‘specificity’/ ‘worlds’ dim.  [+ SPEC.] certains / ciertos manche

[CONSIDERABLE NUMBER]

einige

[RW]

quelques ein paar un par de

[– SPEC.] irgendwelche [PW]

[INCONSIDERABLE NUMBER]

algunos

Sources of the examples:21 Hen

Käthe Henschelmann: Technik des Übersetzens Französisch-Deutsch, Heidelberg: Quelle & Meyer 1980 (+ Spanish translation by Ángel Borda)

HPP Herbert Huber, Henri Perrin, Alain Pacthod: Deutsch-französische Übersetzungen mit Stilübungen. Thèmes et versions, Vienna: Österreichischer Bundesverlag 1973 (+ Spanish translation by Ángel Borda) Pre

“Pas facile d’être président” / “Von der Kunst, Präsident zu sein” (translated by Susanne Auer) / [“No es fácil ser presidente”], journal article of unknown origin, which has been taken from the internal bulletin of the SFG, “Lausanne Bourgeoise”, November 1983 (+ Spanish translation by Ángel Borda)

TSL[v.20020404] Prn:4/09/2002; 13:56

F: PBNS19.tex / p.12 (366)

 Eva Lavric

Wal

Martin Walser: Eiche und Angora. Eine deutsche Chronik, Frankfurt a.M.: Suhrkamp 1963 Fr.: Chêne et lapins angora. Chronique allemande (traduit de l’allemand par Gilbert Badia), Paris: Gallimard 1968 Sp.: Roble y conejos de Angora (traducción Heleno Saña Alcón), Madrid: Edicusa 1970 (Cuadernos para el diálogo. Libros de teatro 19)

Wil

Marc Wilmet: La détermination nominale. Quantification et caractérisation (Linguistique nouvelle), Paris: Presses universitaires de France 1986 (German and Spanish translation by me, E. L.)

Notes . For error analysis and the notion of interference, see e.g. Rattunde 1977; FlamentBoistrancourt 1985; James 1990; Vogel 1990; Henrici & Zöfgen 1993. . See e.g. Coseriu 1973. . See the famous dissertation, Vater 1963/1979. . See e.g. Schifko 1975 and 1992. . The fact that the same mistake occurs three times in this student’s text shows that it is a competence mistake, i.e. that there is a wrong rule stored in the interlanguage. . Lavric 2001a. . The founder of this method is of course Mario Wandruszka (1969 and 1971). See also the recent symposium held in the course of the Romanistentag Osnabrück 1999 (Albrecht & Gauger 2001). . For a justification and explanation of the combined method (translation comparison + commutation tests), see among others Lavric 2001b. . Derjenige is actually confined to this function and cannot be used otherwise. (Although sometimes it also marks restrictive attributes in a form other than that of the relative clause.) . There are exceptions to this rule which correlate with clearly describable stylistic effects; see Lavric 1996, 1997 and 1998. . Ce. . . -là can only be used with explicit or implicit contrast. . Strictly speaking, the proximate demonstratives Ger. dieser and Sp. este are marked for non-restrictivity (= appositionality) of the following relative clause. The distal demonstratives are actually unmarked and thus theoretically comprise both possibilities (restrictivity or appositionality); in fact, however, they are used to mark restrictivity in the overwhelming majority of cases. (Strictly speaking, the definite article also embodies both possibilities). See Lavric (2001a: 900–914). . In addition, one could predict that Spanish-speaking learners of French will face problems similar to those of German-speaking ones when confronted with this matter, while native speakers of French learning German or Spanish will probably tend to exclusively

TSL[v.20020404] Prn:4/09/2002; 13:56

F: PBNS19.tex / p.13 (367)

Cross-language commutation tests 

use the definite article with restrictive relative clauses and thus convey the so-called “overindulgence” phenomenon (Levenston 1972) – out of several varieties the one similar to the mother tongue is preferred – which, however, is obviously not too troublesome. . See e.g. Vater 1963/1979; Sroka 1983; Zhou 1985. This oposition should not be confused with the one between actually versus hypothetically existing referents, which is often referred to by the same term. For a more extensive discussion of the ambiguities of indefinite reference see Galmiche 1983; Martin 1983; and Lavric 1990. . Moreover, they of course show the features [indefinite], [more than one referent] and [–exact number], meaning that they are plural indefinites. (For the semantics of plural indefinites in French and Spanish, see Lavric 2000 and forthcoming). . A number of authors (e.g. Oomen 1977; Flückiger-Studer 1983; Zhou 1985) mix the two oppositions, which in fact can be distinguished in case of a few determiners only. Moxey & Sanford (1993), however, do distinguish between the two approaches to quantity, the objective and the subjective one. . Ger. so mancher and Fr. maint(s) refer to a number as being considerable without specifying whether it is large or small. . Of particular interest is the existence of a singular form many a (many a president . . . ), which is the exact equivalent of Fr. maint(s) and Ger. so manche(r). All three forms signify [+considerable] without specifying a large or small number. They are furthermore stylistically marked as slightly pompous. . For the concept of possible worlds see e.g. Martin 1983 and 1987, and subsequently Lavric 1990 and 1995. . Engl. a few is here subject to the same restrictions as Ger. einige; it has to be substituted with any (here equivalent of Ger. irgendwelche): (10 ) When he proposes any new idea(s), people are automatically against them *a few new ideas . All emphases in the examples are mine, E. L.

References Albrecht, J., & Gauger, H.-M. (Eds.). (forthcoming). Sprachvergleich und Übersetzungsvergleich. Leistung und Grenzen, Unterschiede und Gemeinsamkeiten. Frankfurt a.M. c.a.: Peter Lang. Coseriu, E. (1973). Probleme der strukturellen Semantik. Lecture given in the winter semester 1965/1966 at the University of Tübingen. Authorised and edited notes by Dieter Kastovsky (Tübinger Beiträge zur Linguistik 40). Tübingen: Narr. Flament-Boistrancourt, D. (1985). L’interférence : Un masque, une ambiguïté ? Acta universitatis wratislaviensis, 818, Romanica wratislaviensia, 23 (= Le français langue étrangère. Actes du colloque franco-polonais, Lille, avril 1983), 21–35.

TSL[v.20020404] Prn:4/09/2002; 13:56

F: PBNS19.tex / p.14 (368)

 Eva Lavric

Flückiger-Studer, T. (1983). Quantifikation in natürlichen Sprachen. Zur Semantik und Syntax französischer und deutscher Beschreibungen (Linguistische Arbeiten 132). Tübingen: Niemeyer. Gaatone, D. (1991). Les déterminants de la quantité peu élevée en français. Remarques sur les emplois de quelques et plusieurs. Revue romane, 26 (1), 3–13. Galmiche, M. (1983). Les ambiguïtés référentielles ou les pièges de la référence. In G. Kleiber & M. Riegel (Eds.), Grammaire et référence (= Langue française, 57) (pp. 60–86). Gondret, P. (1976). ‘Quelques’, ‘plusieurs’, ‘certains’, ‘divers’: Étude sémantique. Le français moderne, 44, 143–152. Henrici, G., & Zöfgen, E. (Eds.). (1993). Fehleranalyse und Fehlerkorrektur (= Fremdsprachen Lehren und Lernen, 22). James, C. (1990). Learner language. Language teaching, 23 (4), 205–213. K˛esik, M. (1978). Quelques vs. plusieurs. In Actes de la IIe Conférence des linguistes romanisants polonais (pp. 59–64). Warsaw: Editions de l’Université de Varsovie. Lavric, E. (1990). Mißverstehen verstehen: Opake Kontexte und Ambiguitäten bei indefiniten und definiten Nominalphrasen (Grazer Linguistische Monographien 7). Graz: Institut für Sprachwissenschaft der Universität Graz. Lavric, E. (1995). Referential ambiguities, possible worlds, and universes – The problem of substitution in contexts of propositional attitude. In R. A. Geiger (Ed.), Reference in multidisciplinary perspective. Philosophical object, cognitive subject, intersubjective process (pp. 45–62). Hildesheim/Zurich/New York: Georg Olms. Lavric, E. (1996). ‘Cet air doublement immobile qu’ont les choses mobiles quand elles ne bougent pas...’: Haloeffekt und Relativsatz-Ankündigung. Moderne Sprachen, 40 (2), 157–170. Lavric, E. (1997). ‘Ese reino movible’ – Spanische, französische und deutsche Demonstrativa. In G. Wotjak (Ed.), Studien zum romanisch-deutschen und innerromanischen Sprachvergleich. Akten der III. Internationalen Arbeitstagung zum romanischdeutschen Sprachvergleich (Leipzig, 9.10.–11.10.1995) (pp. 515–543). Frankfurt a.M./Berlin/Bern/New York: Peter Lang. Lavric, E. (1998). Este, ese y aquel en función determinativa. In G. Ruffino (Ed.), Atti del XXI Congresso Internazionale di Linguistica e Filologia Romanza (Palermo, 18–24 settembre 1995), Vol. III: Lessicologia e semantica delle lingue romanze (pp. 405–418). Tübingen: Niemeyer. Lavric, E. (2000). Indéfinis pluriels français et espagnols. In A. Englebert, M. Pierrard, L. Rosier, & D. Van Raemdonck (Eds.), Actes du XXIIe Congrès International de Linguistique et Philologie Romanes, Bruxelles, 23–29 juillet 1998, Vol. VII: Sens et fonction (pp. 377–386). Tübingen: Niemeyer. Lavric, E. (2001a). Fülle und Klarheit. Eine Determinantensemantik Deutsch–Französisch– Spanisch (Stauffenburg Linguistik 9), Vol. I: Referenzmodell; Vol. II: Kontrastivsemantische Analysen. Tübingen: Stauffenburg. Lavric, E. (2001b). Übersetzungsvergleich mit Kommutationsprobe – eine paradigmatische Tiefendimension. In J. Albrecht & H.-M. Gauger (Eds.), Sprachvergleich und Übersetzungsvergleich. Leistung und Grenzen, Unterschiede und Gemeinsamkeiten (pp. 97– 129). Frankfurt a.M. c.a.: Peter Lang.

TSL[v.20020404] Prn:4/09/2002; 13:56

F: PBNS19.tex / p.15 (369)

Cross-language commutation tests 

Lavric, E. (forthcoming). Indefinidos plurales españoles. In F. Sevilla & C. Alvar (Eds.), Actas del XIII Congreso de la Asociación Internacional de Hispanistas, Madrid 1998, Vol. III: Hispanoamericana. Lingüística. Teoría literaria. Madrid: Castalia. Levenston, E. A. (1972). Über- und Unterrepräsentation – Aspekte der muttersprachlichen Interferenz. In G. Nickel (Ed.), Reader zur kontrastiven Linguistik (pp. 167–174). Frankfurt a.M.: Fischer. Martin, R. (1983). Pour une logique du sens (Linguistique nouvelle). Paris: Presses universitaires de France. Martin, R. (1987). Langage et croyance. Les ‘univers de croyance’ dans la théorie sémantique (Philosophie et langage). Brussels: Pierre Mardaga. Moxey, L. M., & Sanford, A. J. (1993). Prior expectation and the interpretation of natural language quantifiers. European journal of cognitive psychology, 5 (1), 73–91. Oomen, I. (1977). Determination bei generischen, definiten und indefiniten Beschreibungen im Deutschen (Linguistische Arbeiten 53). Tübingen: Niemeyer. Rattunde, E. (1977). Transfer – Interferenz? Probleme der Begriffsdefinition bei der Fehleranalyse. In E. Rattunde (Ed.), Fehleranalyse/Fehlerbewertung (= Die Neueren Sprachen, 7 (1)) (pp. 4–14). Schifko, P. (1975). Bedeutungstheorie. Eine Einführung in die linguistische Semantik (Problemata 45). Stuttgart-Bad Cannstatt: Fromman-Holzbog. Schifko, P. (1992). Spanisch: Lexikologie und Semantik. In G. Holtus, M. Metzeltin, & C. Schmitt (Eds.), Lexikon der Romanistischen Linguistik (LRL) Vol. VI, 1: Aragonesisch/Navarresisch, Spanisch, Asturianisch/Leonesisch (pp. 132–148). Tübingen: Niemeyer. Sroka, K. A. (1983). Definiteness and truth relation. In R. Jongen, S. De Knop, P. H. Nelde, & M.-P. Quix (Eds.), Sprache, Diskurs und Text. Akten des 17. Linguistischen Kolloquiums, Brüssel 1982, Vol. I (Linguistische Arbeiten 133) (pp. 110–120). Tübingen: Niemeyer. Vater, H. (1963/1979). Das System der Artikelformen im gegenwärtigen Deutsch. Tübingen: Niemeyer, 1963. 2nd, revised edition (Linguistische Arbeiten 78). Tübingen: Niemeyer, 1979. Vogel, K. (1990). Lernersprache. Linguistische und psycholinguistische Grundfragen ihrer Erforschung (Tübinger Beiträge zur Linguistik 341). Tübingen: Narr. Wandruszka, M. (1969). Sprachen, vergleichbar und unvergleichlich. Munich: Piper. Wandruszka, M. (1971). Interlinguistik. Umrisse einer neuen Sprachwissenschaft. Munich: Piper. Zhou, H. (1985). Determination und Determinantien. Eine Untersuchung am Beispiel neuhochdeutscher Nominalsyntagmen (Bochumer Beiträge zur Semiotik 2). Bochum: Brockmeyer.

Language index

A Arabic 296, 314 B Bagwalal 178–185, 187, 188, 190, 191 C Catalan 101–113, 129, 130, 133, 140, 142–145 E English 3, 4, 7, 23, 26, 27, 29, 33, 35, 38, 41, 45–48, 51–54, 56, 61, 71, 75–78, 81–90, 93–95, 98, 99, 101–106, 108–114, 119, 120, 122, 124, 157, 158, 162–167, 169, 193–201, 203–205, 209–211, 213, 214, 216, 223–225, 228–231, 233–238, 241–243, 249, 250, 252, 261–265, 268–270, 272, 273, 277, 278, 280–288, 307, 308, 311, 313, 314, 316, 318, 321, 327, 329, 335, 344, 355, 361, 362 F Finnish 194 French 3, 4, 9, 10, 21, 26, 30, 45, 76, 84, 102, 104, 117–120, 122, 129–133, 140, 142–145, 194, 195, 262, 282, 336, 343, 355–358, 362, 364–367 G German 102, 104, 194, 355–358, 360, 362, 363, 365, 366

Greek 129, 130, 132, 133, 139, 140, 142–144, 344–346, 352 Guugu Yimithirr 297, 303 H Hadiyya 326, 327 Hausa 297, 303 Hungarian 30, 129, 130, 132, 133, 139, 140, 142–144, 194, 223–226, 228–231, 233–243 I Icelandic 195 Italian 3, 4, 7–9, 12–14, 16, 26, 31, 34, 38, 102, 104, 129, 130, 133, 140, 142–145, 147, 157, 160, 164–168, 171, 172, 194, 248–250, 252, 253, 255, 257, 258, 336 J Japanese 67, 261, 264, 265, 268, 269, 272, 273, 277, 278, 281–283, 285, 286, 288, 296, 307, 311, 314, 316, 318, 321, 328–332 K Khanty 183 L Lega 20, 190 Lithuanian 193–201, 208–217 M Mari 178–185, 187, 190, 191

 Language index

P Polish 194, 201 Portuguese 31, 33, 34, 36–38, 40, 45–48, 51–53, 56, 58, 61, 67, 72, 129, 130, 132–134, 140–145 Brazilian Portuguese 58 R Romanian 75–81, 83, 84, 86–93, 95, 96, 98–100, 344, 345, 349, 351, 352 Russian 194, 195, 262

S Sesotho 296 Spanish 31, 34, 38, 45, 102, 129, 130, 133, 140, 142–145, 194, 248, 249, 252, 255–258, 339–347, 349–351, 355–358, 360, 362, 364–367 Swahili 117–120, 122–124 Swedish 194, 214 T Tatar 178–185, 187, 190, 191 Turkish 177–179 Tzeltal 297, 303

Name index

A Abe, Y.

323, 331

B Blakemore, D.

332

C Carston, R. 9, 172, 325, 333 Chierchia, G. 3–5, 7–13, 17, 26, 148, 151, 154, 162 Cinque, G. 156, 158, 252 Cornelis, L. 249–252 Croft, W. 200, 250 D Dahl, Ö 102, 179 Davidson, D. 326, 333 de Swart, H. 75–77, 81–83 F Fellbaum, C. 270 G Gabbay, D. 307 Gazdar, G. 152–154 Givon, T. 212 Grice, P. 134, 162, 163

K Kamp, H. 61, 68, 71, 73, 75, 77–79, 81, 82, 84, 85, 88, 89, 93, 295, 300–302, 304, 326 Kempson, R. 330, 332 Kubozono, H. 330 L Langacker, R. 258, 277, 280 Lepschy, G. 252 Levin, B. 200, 263, 270 Löbner, S. 6 M Mann, W. 358, 361 Manzini, R. 252 Marten, L. 333 Masuoka, T. 331 Matsui, T. 322, 330, 332 Matsumoto, Y. 307, 332, 333 Meyer-Viol, W. 307 Michaelis, L. 101, 104, 106, 107, 109–111, 113 N Narita, H.

307, 318, 319

H Haspelmath, M. 248 Hidalgo, R. 256

P Parsons, T. 76, 326, 333 Perrett, D. 326 Prince, E. 247, 259

J Jackendoff, R.

R Recanati, F.

300

324

 Name index

Reyle, U. 61, 68, 71, 73, 75, 77–79, 81, 82, 84, 85, 88, 89, 93, 295, 300–302, 304, 326 Ritter, E. 224, 225, 230, 231, 233, 235, 236, 239, 242 Rosen, S. 224, 225, 230–236, 239, 242 Rosta, A. 262–264, 272 S Sansò, A. 245, 252, 258 Schwenter, S. 101, 109 Shibatani, M. 247, 248 Siewierska, A. 246 Soames, S. 148, 149, 151, 153, 154, 165 Sperber, D. 118, 322, 324

Stalnaker, R. 147, 149–151, 162, 167, 168 Steedman, M. 67, 326 Suga, K. 265, 266 T Teramura, H. 277, 307, 317, 331 Thompson, S. 248, 277, 280, 284, 285, 288, 289 W Wilson, D. 117, 118, 122, 124, 322, 324, 332 Y Yamanashi, M. 307, 332

Subject index

A accessibility 227, 246, 322 actions 102, 107, 267, 281–283, 286, 288, 308, 312, 324 activity verbs 232, 241 adverbial clause 331 ‘affectedness’ 273 agent 15, 185–188, 232, 246–252, 257, 258, 262–267, 269, 277, 280, 287 defocusing 248, 250 ago 45, 47–56, 58, 59, 90–93, 99, 107, 112, 118, 254, 267 aktionsart 65–68, 73, 77, 81, 88, 101, 113, 258 algunos 337, 341, 347, 348, 355, 357–360, 362–364 anaphora 6, 64–67, 70, 151, 307, 330, 337 see discourse anaphora anaphoric temporal locators 61, 62 anchor points 45, 47, 48, 57 animacy 246, 287, 340, 343 argument structure 223, 269 aspectual modifier 75, 76, 82 até 52, 53, 58, 62, 63, 141 atelic descriptions 54 aucune 9 B bare happening 251–253, 255, 257, 258 bare object complements 205, 217 bien des 359 bound duration adverbs 84

bridging reference 320–322, 325, 332 British National Corpus (BNC) 214 C certains 358, 362, 363 ciertos 362 clitic pronouns 335, 336 clitic doubling 335, 338–341, 343–347, 349, 350, 352 coercion 75, 76, 79, 80, 84, 91 cognitive grammar 258, 307, 332 collective predicates 6 combined contrastive method 364, 366 commutation tests 355, 357 complement clause 130, 132–136, 138–140, 143, 204, 210, 289, 311 see main clause complementation 193, 195–203, 205, 209, 210, 213, 214, 216 complex clauses 279, 289 compound verbs 283 computational action 308 conceptual information 118 see procedural information configurationality 223, 225, 226 constraint interaction 335, 338, 352 context 12, 16, 20, 22, 37, 40, 41, 50, 52, 62, 104, 105, 107, 122, 143, 144, 148–156, 160, 162–171, 179, 181, 184, 186–189, 216, 230, 241, 264, 268, 271, 281, 286, 288, 295, 296, 307, 321, 327, 332, 336–340, 347, 348, 350, 363

 Subject index

context sets 336, 337, 340 contextual information 118, 122 contingent statement 23 continuative perfect 104, 112 coordination 15, 19 counting of temporally ordered entities 45 D demonstrative determiner 356, 357 dependency 278, 307, 309 determiner semantics 355 see map of determiner meanings universal determiners 347 unrestricted determiners 4 detransitivization 247 direct impression 209 direct mental perception 204, 206 direct vs indirect perception 193, 195, 200, 204, 210 discourse anaphora 6, 307 discourse configurationality 225, 226 discourse representation structure (DRS) 302 Discourse Representation Theory (DRT) 75 discourse topic 229, 242, 246, 253, 340 durante esse tempo 64, 68, 70, 71 during that time 62, 64, 66–68, 70, 71 Dynamic Syntax 307, 308, 324–326, 329 E ein paar 360, 361, 363, 364 einige 355, 357–359, 361–364, 367 entretanto 64, 65, 68, 70, 71 enquanto isso 62–67, 70–72 epistemic attitude 133, 135, 138–140 possibility/uncertainty 202 error analysis 355, 358, 366 evaluation 132, 133, 188, 189, 207–209, 359, 363

events 54, 57, 58, 75, 77, 78, 80, 83, 88–91, 94, 99, 104, 105, 107–112, 117, 118, 122–124, 185, 189, 202, 224, 231, 234, 248, 262, 269, 272, 277, 284, 286–288, 295, 325 see intervals causation 277 role 232, 234, 238 scale 280, 281, 285 structure 72, 73, 89, 224, 250, 258 variable 325–329, 333 eventuality 24, 48, 61–67, 71–73, 75–77, 79–91, 93, 98, 99, 117, 118, 120–124, 300 descriptions 61, 62, 65, 66, 73, 75, 76, 79–81 every 7, 11, 13, 53, 71, 72, 179, 188, 217, 237, 258, 264, 303, 322, 347, 348 evidentiality 177, 178, 189, 190, 201, 210, 212, 213, 216 existential perfect 103, 104 experiential inferentials 178, 185, 190 see non-experiential inferentials exp-verbs 200, 201, 203, 204, 213 extension of meaning 193, 195, 200, 201, 204, 213 F a few 357, 361, 362, 367 factive verb 130, 162, 163, 169 finite complement clause 197 for 45–54, 56, 57 frame adverbs 84 frame anaphoric adverbs 84 Frame Semantics 307, 332 free choice 351 G gapless relatives 316, 317, 320–325, 328, 329, 331–333

Subject index 

appositional gapless relatives 316, 328, 331 generalized quantifier 4, 24, 336, 338, 339, 350, 352 Generalized Quantifier Theory 4, 339 generic statement 5, 347 genre 287 given-new 225–227, 229 goal-driven process 311

H há 38, 45, 47–53, 55–59 ‘hodiernal past’ 102 ‘hot news perfect’ 105, 109

I impersonalization 247 indefinite plural determiners 355 independent clause 278, 279 indicative 129–145, 343 see subjunctive inferred evidence 177, 189 infinitival complement 279 Infinitival Perception Verb Complements (IPVCs) 202 information structure 223–226, 229, 230, 233, 238, 241 intensive quantity 11, 27 interrogative quantifier 350, 351 intervals 46–50, 52, 54–57, 61, 62, 64, 73 see events intransitive clause 248, 277, 281, 285, 288 irgendwelche 363

J judgement types 224, 228, 241 categorical 224, 228–230, 238–240, 242 thetic 224, 228–230, 235, 238–240, 242

L le moindre 10, 13 left downward entailment 35 lexical instruction 308 LINK 313, 314, 316, 319, 329, 331 Lithuanian Language Corpus 214 long-distance dependency 307, 309 Longman Dictionary of Contemporary English 262 M main clause 57, 130, 132, 136, 143, 153, 155, 215, 279, 311 see complement clause maint(s) 367 más de un 359 manche 358, 362–364 (in the) meantime 62, 65, 70 map of determiner meanings 364 measure phrase 87, 88, 92, 93, 97–99 mental perception 202, 204, 206, 211, 212 merge 312, 327, 330 metalinguistic negation 137, 138 metavariable 308, 326, 327 modal particles 279 modal qualification 200, 201, 203, 211–213, 216 mood selection 129, 130, 132–134, 136–141, 144 N nadie 29, 30, 38, 349 natural discourse 277, 288 negative concord (NC) 16, 29–31, 33, 41 negative determiners 3, 5, 7, 10, 13, 26 negative polarity 10 negative quantifiers 12, 19, 349 N-words 16, 30, 35, 37, 38 nessuno 4, 7–15, 17–27, 29, 30, 38 network of demotion 245, 251, 252 newsworthiness 264, 270, 272

 Subject index

“newsworthy” 264, 269, 272 niente 4, 13–27, 38, 39 ninguém 29, 30, 35, 38 ningún 38–40, 349 no 4–6, 13, 26, 33–35, 38, 41 nobody 27, 29, 30, 38, 41, 42, 141, 181, 209, 349 nomic character 25 nominal determination 355 non-experiential inferentials 190, 191 see experiential inferentials noun phrase 271, 346, 356 abstract mass nouns 5, 9 concrete mass nouns 9, 25 countable nouns 3, 7, 9, 11, 12, 14, 19, 20, 27 definites 345, 346 indefinites 344, 357, 364, 367 plural indefinites 357, 364, 367 relational noun 324 uncountable nouns 3, 4, 7–12, 18, 24, 25, 27 O object manipulation 277 non-object manipulating 281–283, 286, 288 object-manipulating actions 281, 283, 286, 288 Optimality Theory (OT) 338, 339, 352 P parallelism 71, 124, 320, 337, 344, 350 parametrical variation 223, 243 parenthetical use 197–199, 213, 216 partitive interpretation 18, 240 passive constructions 245, 248, 249, 251, 252, 254, 257, 258 mediopassives 263 patient 245, 246, 249–254, 257, 262, 266, 277, 280, 287

-oriented process 251, 255, 257, 258 topicalization 245 pentru 88, 95, 96, 98, 99 PERF 75–79, 81–83, 85–89, 93–95, 99 perspective 51, 78, 88, 122, 227, 242, 249–252, 255, 258, 301, 329 plus d’un 359 pragmaticalization 198, 199, 213 present perfect (PrP) 101 presupposition 4, 6, 20–22, 25, 26, 106, 111, 147–155, 157, 159–172, 242, 339 presuppositionality 339, 341, 342, 346 PREZ 75–81, 84–88, 91, 92 principal filter 339–341, 344 procedural information 118 see conceptual information propositional content 211–213 prototype 280 pseudo relative clause 321, 332 pseudo relatives 316, 318, 325, 328, 329, 331, 332 punctual adverbs 84

Q qualche 7, 9, 12, 246 quantized eventualities 76 quelques 355, 357–364

R referentiality 251, 341, 342, 345 relative clause 279, 307, 311, 314, 315, 317–333, 342, 344, 356, 357, 366 restrictive relative clause 356, 357 Relevance, Principle of 322 Relevance Theory (RT) 117, 118, 121, 324, 332 resultative 103, 105, 106, 108–112

Subject index 

Rhetorical Structure Theory (RST) 248 S saliency 267, 273 saturation 324 Segmented Discourse Representation Theory (SDRT) 123 semantic case theory 336 semantic features and oppositions 364 semantic property 35, 243, 278 sentential negation 133, 136–138 simple clause 279, since 45, 47, 51–56, 58, 59, 91, 99 singular determiners 3, 7, 27 so mancher 367 spatial adverbs 296 spatial prepositional expressions 296 speaker’s assessment 212, 213 specific determiner 235 [+/–SPECIFICITY] 358, 359 specificity effect 341 speech time (ST) 76, 77 st-verbs 200, 201, 210, 211, 213 states 5, 20, 66, 75–77, 80, 82, 86, 88, 90, 94, 95, 99, 109, 113, 117, 118, 120, 122, 124, 262, 267, 269, 272, 284–286, 288 see events strongly discrete units 12, 13 structural underspecification 308, 309 subjunctive 129–145, 342–344 see indicative sufficient informativeness 272 T taxonomic readings 8 telic events 54, 90 tense operator 76, 77, 79, 80, 83, 84, 87, 88, 92 that-clause 156, 160, 199 theme-rheme 225–227

time measurement 47, 49 topic-comment 225–227 topic identification 247 see discourse topic transitive clauses 277, 278 transitivity 277–281, 284–288 translation comparison 355–357 type semantic 308, 309, 317, 325 types of entities 202 U un certain 11 un par de 360, 361, 364 unfixed node 309, 310, 313, 314 until 51, 53, 58, 62, 141, 287 V valence 247, 261 verb ellipsis 15 verb phrase bare infinitive 202 finite 196–199, 203, 205, 210, 212–214, 278, 310 full infinitive 209 non-finite 180, 198, 203, 204, 278 verbal modifiers 233, 235, 236, 242 verbless sentences/clauses 4, 14–17, 20, 26, 27 verbs of perception 193, 194, 200, 210 veridicality 37, 132, 139, 143 visual perception 193, 210, 217 volitionality 284 W weakly discrete units 12 wh-words 196, 349, 350 Z zero complementation 199, 216

Contents of Volume 2

Grammaticalization Distal aspects in Bantu languages Steve Nicolle From temporal to conditional: Italian qualora vs English whenever Jacqueline Visconti

3 23

Then – adverbial pro-form or inference particle? A comparative study of English, Ewe, Hungarian, and Norwegian Thorstein Fretheim, Stella Boateng, and Ildikó Vaskó

51

The polysemy of the Swedish verb komma ‘come’: A view from translation corpora Åke Viberg

75

Metaphor in contrast Studying metaphors using a multilingual corpus Kay Wikberg

109

Cross-language metaphors: Conceptual or pragmatic variation? Andreas Musolff

125

A contrastive cognitive perspective on Malay and English figurative language Jonathan Charteris-Black

141

Metaphorical expressions in English and Spanish stock market journalistic texts Anna Espunya and Patrick Zabalbeascoa

159

 Contents of Volume 2

Cross-cultural pragmatics and speech acts Directions of regulation in speech act theory Susumu Kubo

183

On Japanese ne and Chinese ba Mutsuko Endo Hudson and Wen-ying Lu

197

‘I am asking for a pen’: Framing of requests in black South African English Luanga A. Kasanga

213

Cultural scripts for French and Romanian thanking behaviour Tine Van Hecke

237

Sociocultural variation in native and interlanguage complaints Ronald Geluykens and Bettina Kraft

251

A cross-cultural study of requests: The case of British and Japanese undergraduates Saeko Fukushima

263

Questions as indirect requests in Russian and Czech Michael Betsch

277

The language of love in Melanesia: A study of positive emotions Les Bruce

291

Everyday rituals in Polish and English Ewa Jakubowska

331

A question of time? Question types and speech act shifts from a historical-contrastive perspective: Some examples from Old Spanish and Middle English Verena Jung and Angela Schrott The contrasts between contrasters: What discussion groups can tell us about discourse pragmatics Piibi-Kai Kivik and Krista Vogelberg

345

373

The semantics/pragmatics boundary: Theory and applications Cross-linguistic implementations of specificity Klaus von Heusinger

405

The semantics– pragmatics interface: The case of grounding Esam N. Khalil

423

Contents of Volume 2 

On translating ‘what is said’: Tertium comparationis in contrastive semantics and pragmatics K. M. Jaszczolt

441

Translation equivalents as empirical data for semantic/pragmatic theory Bergljot Behrens and Cathrine Fabricius-Hansen

463

Language index

477

Name index

481

Subject index

483

Contents of Volume 1

489

In the PRAGMATICS AND BEYOND NEW SERIES the following titles have been published thus far or are scheduled for publication: 1. WALTER, Bettyruth: The Jury Summation as Speech Genre: An Ethnographic Study of What it Means to Those who Use it. Amsterdam/Philadelphia, 1988. 2. BARTON, Ellen: Nonsentential Constituents: A Theory of Grammatical Structure and Pragmatic Interpretation. Amsterdam/Philadelphia, 1990. 3. OLEKSY, Wieslaw (ed.): Contrastive Pragmatics. Amsterdam/Philadelphia, 1989. 4. RAFFLER-ENGEL, Walburga von (ed.): Doctor-Patient Interaction. Amsterdam/Philadelphia, 1989. 5. THELIN, Nils B. (ed.): Verbal Aspect in Discourse. Amsterdam/Philadelphia, 1990. 6. VERSCHUEREN, Jef (ed.): Selected Papers from the 1987 International Pragmatics Conference. Vol. I: Pragmatics at Issue. Vol. II: Levels of Linguistic Adaptation. Vol. III: The Pragmatics of Intercultural and International Communication (ed. with Jan Blommaert). Amsterdam/Philadelphia, 1991. 7. LINDENFELD, Jacqueline: Speech and Sociability at French Urban Market Places. Amsterdam/Philadelphia, 1990. 8. YOUNG, Lynne: Language as Behaviour, Language as Code: A Study of Academic English. Amsterdam/Philadelphia, 1990. 9. LUKE, Kang-Kwong: Utterance Particles in Cantonese Conversation. Amsterdam/Philadelphia, 1990. 10. MURRAY, Denise E.: Conversation for Action. The computer terminal as medium of communication. Amsterdam/Philadelphia, 1991. 11. LUONG, Hy V.: Discursive Practices and Linguistic Meanings. The Vietnamese system of person reference. Amsterdam/Philadelphia, 1990. 12. ABRAHAM, Werner (ed.): Discourse Particles. Descriptive and theoretical investigations on the logical, syntactic and pragmatic properties of discourse particles in German. Amsterdam/Philadelphia, 1991. 13. NUYTS, Jan, A. Machtelt BOLKESTEIN and Co VET (eds): Layers and Levels of Representation in Language Theory: A functional view. Amsterdam/Philadelphia, 1990. 14. SCHWARTZ, Ursula: Young Children’s Dyadic Pretend Play. Amsterdam/Philadelphia, 1991. 15. KOMTER, Martha: Conflict and Cooperation in Job Interviews. Amsterdam/Philadelphia, 1991. 16. MANN, William C. and Sandra A. THOMPSON (eds): Discourse Description: Diverse Linguistic Analyses of a Fund-Raising Text. Amsterdam/Philadelphia, 1992. 17. PIÉRAUT-LE BONNIEC, Gilberte and Marlene DOLITSKY (eds): Language Bases ... Discourse Bases. Amsterdam/Philadelphia, 1991. 18. JOHNSTONE, Barbara: Repetition in Arabic Discourse. Paradigms, syntagms and the ecology of language. Amsterdam/Philadelphia, 1991. 19. BAKER, Carolyn D. and Allan LUKE (eds): Towards a Critical Sociology of Reading Pedagogy. Papers of the XII World Congress on Reading. Amsterdam/Philadelphia, 1991. 20. NUYTS, Jan: Aspects of a Cognitive-Pragmatic Theory of Language. On cognition, functionalism, and grammar. Amsterdam/Philadelphia, 1992. 21. SEARLE, John R. et al.: (On) Searle on Conversation. Compiled and introduced by Herman Parret and Jef Verschueren. Amsterdam/Philadelphia, 1992.

22. AUER, Peter and Aldo Di LUZIO (eds): The Contextualization of Language. Amsterdam/Philadelphia, 1992. 23. FORTESCUE, Michael, Peter HARDER and Lars KRISTOFFERSEN (eds): Layered Structure and Reference in a Functional Perspective. Papers from the Functional Grammar Conference, Copenhagen, 1990. Amsterdam/Philadelphia, 1992. 24. MAYNARD, Senko K.: Discourse Modality: Subjectivity, Emotion and Voice in the Japanese Language. Amsterdam/Philadelphia, 1993. 25. COUPER-KUHLEN, Elizabeth: English Speech Rhythm. Form and function in everyday verbal interaction. Amsterdam/Philadelphia, 1993. 26. STYGALL, Gail: Trial Language. A study in differential discourse processing. Amsterdam/ Philadelphia, 1994. 27. SUTER, Hans Jürg: The Wedding Report: A Prototypical Approach to the Study of Traditional Text Types. Amsterdam/Philadelphia, 1993. 28. VAN DE WALLE, Lieve: Pragmatics and Classical Sanskrit. Amsterdam/Philadelphia, 1993. 29. BARSKY, Robert F.: Constructing a Productive Other: Discourse theory and the convention refugee hearing. Amsterdam/Philadelphia, 1994. 30. WORTHAM, Stanton E.F.: Acting Out Participant Examples in the Classroom. Amsterdam/Philadelphia, 1994. 31. WILDGEN, Wolfgang: Process, Image and Meaning. A realistic model of the meanings of sentences and narrative texts. Amsterdam/Philadelphia, 1994. 32. SHIBATANI, Masayoshi and Sandra A. THOMPSON (eds): Essays in Semantics and Pragmatics. Amsterdam/Philadelphia, 1995. 33. GOOSSENS, Louis, Paul PAUWELS, Brygida RUDZKA-OSTYN, Anne-Marie SIMONVANDENBERGEN and Johan VANPARYS: By Word of Mouth. Metaphor, metonymy and linguistic action in a cognitive perspective. Amsterdam/Philadelphia, 1995. 34. BARBE, Katharina: Irony in Context. Amsterdam/Philadelphia, 1995. 35. JUCKER, Andreas H. (ed.): Historical Pragmatics. Pragmatic developments in the history of English. Amsterdam/Philadelphia, 1995. 36. CHILTON, Paul, Mikhail V. ILYIN and Jacob MEY: Political Discourse in Transition in Eastern and Western Europe (1989-1991). Amsterdam/Philadelphia, 1998. 37. CARSTON, Robyn and Seiji UCHIDA (eds): Relevance Theory. Applications and implications. Amsterdam/Philadelphia, 1998. 38. FRETHEIM, Thorstein and Jeanette K. GUNDEL (eds): Reference and Referent Accessibility. Amsterdam/Philadelphia, 1996. 39. HERRING, Susan (ed.): Computer-Mediated Communication. Linguistic, social, and cross-cultural perspectives. Amsterdam/Philadelphia, 1996. 40. DIAMOND, Julie: Status and Power in Verbal Interaction. A study of discourse in a closeknit social network. Amsterdam/Philadelphia, 1996. 41. VENTOLA, Eija and Anna MAURANEN, (eds): Academic Writing. Intercultural and textual issues. Amsterdam/Philadelphia, 1996. 42. WODAK, Ruth and Helga KOTTHOFF (eds): Communicating Gender in Context. Amsterdam/Philadelphia, 1997. 43. JANSSEN, Theo A.J.M. and Wim van der WURFF (eds): Reported Speech. Forms and functions of the verb. Amsterdam/Philadelphia, 1996.

44. BARGIELA-CHIAPPINI, Francesca and Sandra J. HARRIS: Managing Language. The discourse of corporate meetings. Amsterdam/Philadelphia, 1997. 45. PALTRIDGE, Brian: Genre, Frames and Writing in Research Settings. Amsterdam/Philadelphia, 1997. 46. GEORGAKOPOULOU, Alexandra: Narrative Performances. A study of Modern Greek storytelling. Amsterdam/Philadelphia, 1997. 47. CHESTERMAN, Andrew: Contrastive Functional Analysis. Amsterdam/Philadelphia, 1998. 48. KAMIO, Akio: Territory of Information. Amsterdam/Philadelphia, 1997. 49. KURZON, Dennis: Discourse of Silence. Amsterdam/Philadelphia, 1998. 50. GRENOBLE, Lenore: Deixis and Information Packaging in Russian Discourse. Amsterdam/Philadelphia, 1998. 51. BOULIMA, Jamila: Negotiated Interaction in Target Language Classroom Discourse. Amsterdam/Philadelphia, 1999. 52. GILLIS, Steven and Annick DE HOUWER (eds): The Acquisition of Dutch. Amsterdam/ Philadelphia, 1998. 53. MOSEGAARD HANSEN, Maj-Britt: The Function of Discourse Particles. A study with special reference to spoken standard French. Amsterdam/Philadelphia, 1998. 54. HYLAND, Ken: Hedging in Scientific Research Articles. Amsterdam/Philadelphia, 1998. 55. ALLWOOD, Jens and Peter Gärdenfors (eds): Cognitive Semantics. Meaning and cognition. Amsterdam/Philadelphia, 1999. 56. TANAKA, Hiroko: Language, Culture and Social Interaction. Turn-taking in Japanese and Anglo-American English. Amsterdam/Philadelphia, 1999. 57 JUCKER, Andreas H. and Yael ZIV (eds): Discourse Markers. Descriptions and theory. Amsterdam/Philadelphia, 1998. 58. ROUCHOTA, Villy and Andreas H. JUCKER (eds): Current Issues in Relevance Theory. Amsterdam/Philadelphia, 1998. 59. KAMIO, Akio and Ken-ichi TAKAMI (eds): Function and Structure. In honor of Susumu Kuno. 1999. 60. JACOBS, Geert: Preformulating the News. An analysis of the metapragmatics of press releases. 1999. 61. MILLS, Margaret H. (ed.): Slavic Gender Linguistics. 1999. 62. TZANNE, Angeliki: Talking at Cross-Purposes. The dynamics of miscommunication. 2000. 63. BUBLITZ, Wolfram, Uta LENK and Eija VENTOLA (eds.): Coherence in Spoken and Written Discourse. How to create it and how to describe it.Selected papers from the International Workshop on Coherence, Augsburg, 24-27 April 1997. 1999. 64. SVENNEVIG, Jan: Getting Acquainted in Conversation. A study of initial interactions. 1999. 65. COOREN, François: The Organizing Dimension of Communication. 2000. 66. JUCKER, Andreas H., Gerd FRITZ and Franz LEBSANFT (eds.): Historical Dialogue Analysis. 1999. 67. TAAVITSAINEN, Irma, Gunnel MELCHERS and Päivi PAHTA (eds.): Dimensions of Writing in Nonstandard English. 1999. 68. ARNOVICK, Leslie: Diachronic Pragmatics. Seven case studies in English illocutionary development. 1999.

69. NOH, Eun-Ju: The Semantics and Pragmatics of Metarepresentation in English. A relevance-theoretic account. 2000. 70. SORJONEN, Marja-Leena: Responding in Conversation. A study of response particles in Finnish. 2001. 71. GÓMEZ-GONZÁLEZ, María Ángeles: The Theme-Topic Interface. Evidence from English. 2001. 72. MARMARIDOU, Sophia S.A.: Pragmatic Meaning and Cognition. 2000. 73. HESTER, Stephen and David FRANCIS (eds.): Local Educational Order. Ethnomethodological studies of knowledge in action. 2000. 74. TROSBORG, Anna (ed.): Analysing Professional Genres. 2000. 75. PILKINGTON, Adrian: Poetic Effects. A relevance theory perspective. 2000. 76. MATSUI, Tomoko: Bridging and Relevance. 2000. 77. VANDERVEKEN, Daniel and Susumu KUBO (eds.): Essays in Speech Act Theory. 2002. 78. SELL, Roger D. : Literature as Communication. The foundations of mediating criticism. 2000. 79. ANDERSEN, Gisle and Thorstein FRETHEIM (eds.): Pragmatic Markers and Propositional Attitude. 2000. 80. UNGERER, Friedrich (ed.): English Media Texts – Past and Present. Language and textual structure. 2000. 81. DI LUZIO, Aldo, Susanne GÜNTHNER and Franca ORLETTI (eds.): Culture in Communication. Analyses of intercultural situations. 2001. 82. KHALIL, Esam N.: Grounding in English and Arabic News Discourse. 2000. 83. MÁRQUEZ REITER, Rosina: Linguistic Politeness in Britain and Uruguay. A contrastive study of requests and apologies. 2000. 84. ANDERSEN, Gisle: Pragmatic Markers and Sociolinguistic Variation. A relevance-theoretic approach to the language of adolescents. 2001. 85. COLLINS, Daniel E.: Reanimated Voices. Speech reporting in a historical-pragmatic perspective. 2001. 86. IFANTIDOU, Elly: Evidentials and Relevance. 2001. 87. MUSHIN, Ilana: Evidentiality and Epistemological Stance. Narrative retelling. 2001. 88. BAYRAKTAROG
LU, ArFn and Maria SIFIANOU (eds.): Linguistic Politeness Across Boundaries. The case of Greek and Turkish. 2001. 89. ITAKURA, Hiroko: Conversational Dominance and Gender. A study of Japanese speakers in first and second language contexts. 2001. 90. KENESEI, István and Robert M. HARNISH (eds.): Perspectives on Semantics, Pragmatics, and Discourse. A Festschrift for Ferenc Kiefer. 2001. 91. GROSS, Joan: Speaking in Other Voices. An ethnography of Walloon puppet theaters. 2001. 92. GARDNER, Rod: When Listeners Talk. Response tokens and listener stance. 2001. 93. BARON, Bettina and Helga KOTTHOFF (eds.): Gender in Interaction. Perspectives on femininity and masculinity in ethnography and discourse. 2002 94. McILVENNY, Paul (ed.): Talking Gender and Sexuality. 2002. 95. FITZMAURICE, Susan M.: The Familiar Letter in Early Modern English. A pragmatic approach. 2002. 96. HAVERKATE, Henk: The Syntax, Semantics and Pragmatics of Spanish Mood. 2002.

97. MAYNARD, Senko K.: Linguistic Emotivity. Centrality of place, the topic-comment dynamic, and an ideology of pathos in Japanese discourse. 2002. 98. DUSZAK, Anna (ed.): Us and Others. Social identities across languages, discourses and cultures. 2002. 99. JASZCZOLT, K.M. and Ken TURNER (eds.): Meaning Through Language Contrast. Volume 1. 2003. 100. JASZCZOLT, K.M. and Ken TURNER (eds.): Meaning Through Language Contrast. Volume 2. 2003. 101. LUKE, Kang Kwong and Theodossia-Soula PAVLIDOU (eds.): Telephone Calls. Unity and diversity in conversational structure across languages and cultures. 2002. 102. LEAFGREN, John: Degrees of Explicitness. Information structure and the packaging of Bulgarian subjects and objects. 2002. 103. FETZER, Anita and Christiane MEIERKORD (eds.): Rethinking Sequentiality. Linguistics meets conversational interaction. 2002. 104. BEECHING, Kate: Gender, Politeness and Pragmatic Particles in French. 2002. 105. BLACKWELL, Sarah E.: Implicatures in Discourse. The case of Spanish NP anaphora. 2003. 106. BUSSE, Ulrich: Linguistic Variation in the Shakespeare Corpus. Morpho-syntactic variability of second person pronouns. 2002. 107. TAAVITSAINEN, Irma and Andreas H. JUCKER (eds.): Diachronic Perspectives on Address Term Systems. n.y.p. 108. BARRON, Anne: Acquisition in Interlanguage Pragmatics. How to do things with words in a study abroad context. n.y.p. 109. MAYES, Patricia: Language, Social Structure, and Culture. A genre analysis of cooking classes in Japan and America. n.y.p. 110. ANDROUTSOPOULOS, Jannis K. and Alexandra GEORGAKOPOULOU (eds.): Discourse Constructions of Youth Identities. n.y.p. 111. ENSINK, Titus and Christoph SAUER (eds.): Framing and Perspectivising in Discourse. n.y.p. 112. LENZ, Friedrich (ed.): Deictic Conceptualisation of Space, Time and Person. n.y.p. 113. PANTHER, Klaus-Uwe and Linda L. THORNBURG (eds.): Metonymy and Pragmatic Inferencing. n.y.p.