YEARBOOK OF MORPHOLOGY 1997 Yearbook of Morphology Editors: Geert Booij Jaap van Marie Consulting Editors: Stephen Anderson (Yale) Mark Aronoff (Stony Brook, N.Y.) Mark Baker (Montreal) Laurie Bauer (Wellington) Rudie Botha (Stellenbosch) Joan Bybee (Albuquerque, N.M.) Andrew Carstairs-McCarthy (Christchurch) Greville Corbett (Surrey) Wolfgang Dressler (Wien) Martin Haspelmath (Bamberg) Jack Hoeksema (Groningen) Rochelle Lieber (Durham, N.H.) Peter Matthews (Cambridge, U.K.) Franz Rainer (Wien) Sergio Scalise (Bologna) Henk Schultink (Utrecht) Andrew Spencer (Colchester) Editorial address: Editors, Yearbook of Morphology Vakgroep Taalkunde, Vrije Universiteit De Boelelaan 1105 1081 HV Amsterdam, The Netherlands e-mail: [email protected] YEARBOOK OF MORPHOLOGY 1997 Edited by GEERT BOOIJ General Linguistics, Free University, Amsterdam, The Netherlands and JAAP VAN MARLE Dutch Royal Academy of Sciences, Amsterdam, The Netherlands Springer-Science+Business Media, B.V. A CLP. Catalogue record for this book is available from the Library of Congress. ISBN 978-94-010-6098-1 ISBN 978-94-011-4998-3 (eBook) DOI 10.1007/978-94-011-4998-3 Printed on acid-free paper All Rights Reserved © 1998 Springer Science+Business Media Dordrecht Originally published by Kluwer Academic Publishers in 1998 No part of the material protected by this copyright notice may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording or by any information storage and retrieval system, without written permission from the copyright owner. Table of Contents ANDREW CARSTAIRS-MCCARTHY / How lexical semantics constrains inflectional allomorphy 1 EWA CZA YKOWSKA-HIGGINS / Verbalizing suffixes and the structure of the Polish verb 25 LISE M. DOBRIN / The morphosyntactic reality of phonological form 59 LAURA 1. DOWNING / Prosodic misalignment and reduplication 83 SHARON INKELAS / The theoretical status of morphologically conditioned phonology: a case study of dominance effects 121 LIZANNE KAISER / The interaction of noun incorporation and applicative formation in Ainu 157 CEMIL ORHAN ORGUN / Cyclic and noncyclic phonological effects in a declarative grammar 179 INGO PLAG / The'polysemy of -ize derivatives: On the role of semantics in word formation 219 ANGELA RALLI and MELITA STAVROU / Morphology-syntax interface: A-N compounds vs. A-N constructs in Modem Greek 243 BARBARA STIEBELS / Complex denominal verbs in German and the morphology~semantics interface 265 Book Notices GEERT BOOU I Grazia Crocco Galeas, Metafora Morfologica. Saggio di Morfologia Naturale 303 JAAP VAN MARLE I Georgette Dal, Grammaire du sufflXe -etC te) 303 JAAP VAN MARLE I Christiane Dalton-Puffer, The French Influence on Middle English Morphology. A Corpus-Based Study of Derivation 304 GEERT BOOU I Siebren Dijk, Noun incorporation in Frisian 304 GEERT BOOU / Wolfgang U. Dressler (ed.), Studies in Pre- and Protomorphology 305 GEERT BOOU / Alan Ford, Rajendra Singh, Gita Martohardjono, Pace Panini. Toward a Word-based Theory ofM orphology 305 GEERT BOOU / Sylvia C. Ltihken, Deutsche Wortprosodie. Abschwagungs- und Tilgungsvorgange 307 JAAP VAN MARLE / Jorma Luutonen, The Variation of Morpheme Order in Mari Declension 307 v GEERT BOOIJ I Tapani Salminen (ed.), Typological and historical studies in language by Mikko Korhonen. A memorial volume published on the 60th anniversary of his birth 308 GEERT BOOIJ I Rajendra Singh & Rama Kant, Hindi Morphology. A Word-based Description 309 VI How lexical semantics constrains inflectional allomorphy ANDREW CARSTAIRS-MCCARTHY 1. LEXICAL SEMANTIC CONSTRAINTS The view that inflectional affixes are lexical items or vocabulary items, with much the same morphological status as lexemes or stems belonging to lexical categories, has a respectable following. 1 Versions of it (let's call it the LEXICAL-AFFIX view) are expounded by Lieber (1981, 1992), Jensen and Stong-Jensen (1984) and Halle and Marantz (1993), and are also at the heart of approaches to syntax in terms of which inflectional affixes are seen as heads of phrases belonging to 'functional categories' . But this view has also been widely criticized by proponents of what one may call REALIZATIONAL views of inflection (e.g. Matthews 1972, Zwicky 1985, Anderson 1992, Aronoff 1994, Beard 1995, Stump 1991, 1997). The purpose of this article is not to rehash more or less familiar arguments on either side, but to address a so fa;" neglected question: what does the lexical-affix view imply about possible and im possible meanings, or morpho syntactic content, for inflectional affixes? I will pro pose a tentative answer, reinforcing the importance attributed to the paradigmati . dimension in inflectional morphology by Carstairs(-McCarthy) (1983, 1994). I de, not think that this answer resolves the old controversy, but it may refresh it by injecting a new ingredient. A conceivable answer might be that treating inflectional affixes as lexical items has no implications at all for what they might mean. That would be the only possible answer if lexical semantic structure were a purely language-particular affair, subject to no general restrictions whether from Universal Grammar or elsewhere. So is there any evidence for such restrictions? One might expect that locating a received opin ion on this question, at least in outline, would be a simple matter of looking up a handbook or textbook. Semantics beats all other linguistic titles in terms of the numbers of books bearing it (Pullum 1988), and the lexicon-syntax interface and issues of conceptual structure have received considerable attention from linguists and philosophers in recent years (e.g. the work cited by Jackendoff and by Levin and Rappaport Hovav in Lappin 1996). Yet, as Fass (1993) notes, there is surprisingly little in this literature which bears on the issue of possible and impossible meanings at a level of resolution appropriate for discussing individual lexical items and at the same time not sidetracked towards the interface with syntax through argument struc ture and thematic relations. The closest approaches to what we are looking for in the linguistic literature are in the pioneering work of Lehrer (1974), in Cruse's elaborate taxonomy of lexical-semantic relationships (1986), and in Talmy's analysis of dif ferent languages' lexicalization habits in regard to semantic features such as PATH Geen Booij and Jaap van Marie (eds.), Yearbook of Morphology 1997, 1-24. © 1998 Kluwer Academic Publishers. 2 Andrew Carstairs-McCarthy and MANNER (1985). But all of these are in the nature of extended case studies rather than expositions of theoretical frameworks intended to make predictions about what can and cannot happen in lexical semantics. Psychological and psycholinguistic research fills the gap to some extent. There is a considerable literature on the psychology of concept formation (e.g. Bruner et al. 1956; Keil1979, 1989) as well as on children's acquisition of concepts at the level of lexical meanings (e.g. Markman 1989; Clark 1993). The Prague School notion of markedness, although its compass extends well outside lexical semantics, is also relevant here (for references see Vachek 1966, Battistella 1990). It is on these that I will mainly draw for examples of parallels between lexical semantics and the morphosyntax of inflectional affixes. The lack of a generally agreed and integrated theory of possible lexical meanings admittedly entails a risk, namely that I will tend to magnify the importance of those lexical-semantic phenomena which seem to have parallels in affixal meaning and ignore ones which do not. But it is better to run this risk than to not to embark on the inquiry at all. If there is clearcut prima facie evidence for generalizations which span both lexical and inflectional meaning, then a direction is indicated for further re search. Even if this research shows that that the parallels are not exact, solid discov eries about both lexical semantics and inflectional morphology may yet result. 1.1. The Principle of Contrast Young children learning their first language are guided by several pragmatic princi ples which narrow the range of possible meanings for any new lexical item and so help them pick up new vocabulary accurately and quickly. One is the Mutual Exclu sivity Principle, on the basis of which a child may (for example) go through a stage of being reluctant to accept that the words tree and sycamore can both apply to the same plant (Merriman & Bowman 1989; Markman 1989). Children learn in due course that the Mutual Exclusivity Principle can be overridden, to the extent of allowing two terms to be related as superordinate and hyponym; but we preserve into adulthood an expectation that two terms applying to objects at the same level of generality (two species of tree, for example) will not overlap in meaning. Apparent counterexamples to this, such as the application of both plane and sycamore to the species acer pseudoplatanus, generally tum out to be indeed only apparent: thus, the name plane is used in Scotland whereas sycamore is used in England (where plane means rather platanus orientalis or platanus x hispanica) (Mitchell 1974). Somewhat less restrictive than the Mutual Exclusivity Principle is the Principle ef Contrast (Clark 1993): 'Every two forms contrast in meaning.' This Principle is not violated by tree and sycamore, since the wider extension of tree provides the requisite contrast. But pairs of exact synonyms (e.g. aubergine and eggplant, or nearly and almost) do violate it. The existence of such pairs shows that the Principle cannot be inviolable. But evidence amply confirms two expectations which the Principle generates: that there should be very few exact synonyms among the vo- How lexical semantics constrains inflectional allomorphy 3 cabulary learned in infancy when vocabulary acquisition is most rapid, and that when children under experimental conditions are faced with ascribing a suitable meaning to an invented word they will not interpret it as synonymous with any word they already know. 1.2. Dislike of disjunctive meanings English vocabulary provides a generic term fruit and species-level terms such as apple, pear, orange, plum and banana. It also provides terms intermediate between the generic and species levels, such as stonefruit, pipfruit and citrus, and terms for varieties below the species level, such as Cox's, braeburn and red delicious for apples. What English does not provide is a word such as blum with a disjunctive meaning 'plum banana' , applicable to plums and bananas and nothing else. And OR this fact is hardly isolated. Disjunctive denotations of this kind seem equally hard to find in other areas of vocabulary and in other languages.2 At first sight, there is nothing surprising in this. After all, if fruit-eating English speakers are familiar with both plums and bananas and have no difficulty in distin guishing between them, it is only natural that they should have separate words for them, and it is hard to see what purpose would be served by a word designating both of them indiscriminately. In the context of our present discussion, however, the oddness of meanings such as 'plum banana' is worth pointing out because OR disjunctions do crop up in the meanings ascribed to some inflectional affixes by morphological theorists of various persuasions:3 (1) Georgian person-number suffixes (Jensel'1 & Stong-Jensen 1984:492): a. -t1 [ {+I, +II} +pl [ ]] b. -0 [ {+I, +II}[ ]] (2) Vogul person-number suffix (Lieber 1992:105): -uw { [+1 +PI-Du], [+11 -PI]! [+11 +Du] } (3) Potawatomi person-number suffixes: a. (Anderson 1992: 179): -n [ -me -you {+PI, +Obv} ] b. (Halle & Marantz 1993:151): -nan [+1 +PI {ACC, GEN} ] Does this mean, then, that the lexical-affix position is weakened by a divergence between the kinds of meaning which affixes and nonaffixallexical items can have? Before we say yes, we must acknowledge that disjunctive meanings do indeed exist, albeit ones which differ in crucial respects from that of the hypothetical blum. I am not referring to terms such as citrus or stonefruit, whose meanings are not disjunctive in the relevant sense. That is, citrus does not mean simply 'orange OR lemon OR grapefruit OR ... ', and stonefruit does not mean simply 'plum OR apricot OR 4 Andrew Carstairs-McCarthy peach OR .. .'. What makes a fruit a citrus fruit is not simply membership in a disjunc tive list but close genetic relationship and shared characteristics of rind, pith and acidity; and similar considerations apply to stonefruit. On terms whose meanings really are disjunctive in the relevant sense, some light is shed by the study of concept formation by Bruner et al. (1956). Given a stimulus array of rectangles and triangles of which some are black and some yellow, some large and some small, people find it relatively easy to learn conjunctively defined concepts (e.g. 'small triangle AND large rectangle') but relatively hard to learn disjunctively defined ones (e.g. 'small trian gle OR large rectangle OR both'). An example of a real-life term which is defined disjunctively is strike in baseball: 'a pitch that is across the plate and between the batter's knees and shoulders or ... any pitch at which the batter strikes but fails to send the ball into the field' (Bruner et al. 1956:43). Strike in this sense differs from most words in a crucial respect: it is a technical term defined explicitly in the framework of a codified set of rules for a game, and so is arguably not part of the 'natural' vocabulary of English. Yet, even if such technical terms are excluded from consideration, there are other examples which are harder to dismiss. lackendoff (1985) points out that the meaning of climb contains PREFERENCE FEATURES 'upward' and 'clambering', but that these are disjunct, so as to yield the following pattern of acceptability judgments: (4) a. The boy climbed up the tree. b. The boy climbed down the tree. c. The snake climbed up the tree. d. *The snake climbed down the tree. Only (4d), where both 'upward' and 'clambering' are excluded, is illformed. A cluster of disjunct semantic features seems also to underlie the meaning of game, famously discussed by Wittgenstein (1953:31-5). And the same applies to any word whose meaning has undergone extension by metonymy. For example, crown can mean 'headgear with points around the rim' (as in He put on the paper crown which fell out of his Christmas cracker) OR '(associated with the) state or government' (as in crown land, crown prerogative). What differentiates terms with acceptably disjunctive meanings, like climb and crown, from ones whose disjunctive meanings are either unacceptable or narrowly technical, such as blum and strike, is the compatibility of the disjuncts. The disjunc tion 'plum OR banana' is incompatible (or exclusive) because a fruit cannot be both a plum and a banana. On the other hand, 'clambering OR upward' is compatible (or inclusive) because it is possible to clamber upwards - indeed, one may well regard clambering upward as prototypical climbing. Likewise, an article of headgear with points around the rim can well be associated with a state function, as in the corona tion of a monarch - and again, it is that sort of crown which is prototypical. In fact, it appears that all or most of the disjunctions which do occur in the meanings of lexical items are of the inclusive kind, better represented by AND/OR than by OR. What makes