Prepositions in their Syntactic, Semantic and Pragmatic Context
Typological Studies in Language (TSL) A companion s...
424 downloads
2018 Views
3MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
AUTHOR ""
TITLE "Prepositions in their Syntactic, Semantic and Pragmatic Context"
SUBJECT "Typological Studies in Language, Volume 50"
KEYWORDS ""
SIZE HEIGHT "220"
WIDTH "150"
VOFFSET "4">
Prepositions in their Syntactic, Semantic and Pragmatic Context
Typological Studies in Language (TSL) A companion series to the journal Studies in Language
General Editor
Michael Noonan
Assistant Editors
Spike Gildea, Suzanne Kemmer
Editorial Board Wallace Chafe (Santa Barbara) Bernard Comrie (Leipzig) R. M. W. Dixon (Melbourne) Matthew Dryer (Buffalo) John Haiman (St Paul) Bernd Heine (Köln) Paul Hopper (Pittsburgh) Andrej Kibrik (Moscow) Ronald Langacker (San Diego)
Charles Li (Santa Barbara) Edith A. Moravcsik (Milwaukee) Andrew Pawley (Canberra) Doris Payne (Eugene, OR) Frans Plank (Konstanz) Jerrold Sadock (Chicago) Dan Slobin (Berkeley) Sandra Thompson (Santa Barbara)
Volumes in this series will be functionally and typologically oriented, covering specific topics in language by collecting together data from a wide variety of languages and language typologies. The orientation of the volumes will be substantive rather than formal, with the aim of investigating universals of human language via as broadly defined a data base as possible, leaning toward cross-linguistic, diachronic, developmental and live-discourse data.
Volume 50 Prepositions in their Syntactic, Semantic and Pragmatic Context Edited by Susanne Feigenbaum and Dennis Kurzon
Prepositions in their Syntactic, Semantic and Pragmatic Context Edited by
Susanne Feigenbaum Dennis Kurzon University of Haifa
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 Prepositions in their Syntactic, Semantic and Pragmatic Context / edited by Susanne Feigenbaum and Dennis Kurzon. p. cm. (Typological Studies in Language, issn 0167–7373 ; v. 50) Includes bibliographical references and indexes. 1. Grammar, Comparative and general--Prepositions. 2. Grammar, Comparative and general--Syntax. 3. Semantics. 4. Pragmatics. I. Feigenbaum, Susanne. II. Kurzon, Dennis. III. Series. P285 .P74 2002 415-dc21 isbn 90 272 2956 2 (Eur.) / 1 58811 172 5 (US) (Hb; alk. paper)
2002018541
© 2002 – 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
Preface Susanne Feigenbaum and Dennis Kurzon Instability and the theory of semantic forms: Starting from the case of prepositions Yves-Marie Visetti and Pierre Cadiot Schematics and motifs in the semantics of prepositions Pierre Cadiot The theoretical status of prepositions: The case of the “prospective use” of in Franck Lebas
Temporal semantics of prepositions in context David S. Brée and Ian E. Pratt-Hartmann
Prepositions and context Ian Pratt-Hartmann and Nissim Francez
Prepositional phrases as noun modifiers in contemporary Hebrew: Grammatical, semantic and pragmatic motivations Esther Borochovsky-Bar Aba and Hava Reppen The Hebrew prepositions mi-/min ‘from, of ’: Same or different? Yishai Tobin
A contrastive analysis of French and Hebrew prepositions: The case of sans, bli-belo and lelo Susanne Feigenbaum
A language in change: Declined prepositions in spoken Modern Hebrew as a case study Inbar Kimchi-Angert
The French preposition in contact with Hebrew Miriam Ben-Rafael
Table of contents
‘Preposition’ as functor: The case of long in Bislama Dennis Kurzon
Prepositions in modern Judeo-Greek (JG) Biblical translations Julia G. Krivoruchko
Quddiem and some remarks on grammatical aspects of Maltese prepositions Rami Saari
Locative prepositions in language acquisition and aphasia Mark Leikin
Index
Preface Susanne Feigenbaum and Dennis Kurzon1
.
Prepositions and linguistic theory
Growing interest among linguists in the word-class of prepositions (PP) has been seen in a number of fields. First, researchers have tended to develop their insight into this category by centering on syntax and case, on semantics or on the interaction between the modules, implicitly following the distinction between colorless case PPs, colorful semantic PPs and mixed PPs, along the lines of Spang-Hanssen (1963). The root of this approach derives from structuralism and the search for a distinctive meaning of the preposition, with many of the researchers referring to the semantic treatment by Brøndal (1950) or to the psycho-mechanic theory of Guillaume, developed in Pottier (1962, 1997). In these works, one finds the hypothesis that the semantics of PPs is above all relational and that the basic meaning of PPs lies in the fact that they orient an object in space, be it local or temporal. For syntax and semantics, Jackendorf (1973, 1977) marks a new area by claiming that prepositions form phrases with the same structure – in terms of X-bar syntax – as the other major phrase classes (VPs, NPs and APs); they, therefore, are no longer a small class of functors like conjunctions or case markers (see Fillmore 1968), but fully fledged contentives. This hypothesis has been developed by Emonds (1985), who distinguishes three categories on the basis of the obligatory or optional occurrence of NPs as complement, and by Rauh (1993), who establishes new syntactical and semantic criteria for the division between lexical and non-lexical PPs, the latter either occurring in phrases or being assigned by another word-class. Recently, the study of prepositions has emerged in the context of morpho-syntax. As an example, Englebert (1993), Kupferman (1996) and Lagae (1998) deal with French de and en respectively, verifying their qualification as PPs or quantifiers. Melis (1992) investigates a/pour in dative function and Bat-Zeev Shyldkrot (1995) studies grammati-
Preface
calization. Furthermore, syntactic valency has proven to be a productive field of interest, as shown in the analysis of simple and compound PPs in various languages, an undertaking in which Gaatone (1976), Wunderlich (1984), Anscombre (1993), Gross (1995, 1996), Zifonun (1996, 1997), Borillo (1997), and many others have participated. Secondly, cognitive theories have approached PPs in the context of lexical semantics. On the one hand, the hypothesis initiated in the eighties, maintaining that the basic spatial function of locative prepositions allows for a separation of the semantic and the conceptual meaning, is still accepted by many scholars today (cf. Taylor 2000). On the other hand, the pragmatic view that the various uses of the preposition are first and above all captured by world knowledge, using cognitive processes like the metaphorical domains advocated by Lakoff and Johnson (1980), or the schematic network by Langacker (1987), is still defended in present research. Furthermore, cognitive models close to discourse semantics, with the focus on deixis or prototypes, have taken up the study of PPs, describing the structure of local and temporal prepositions in several languages, as one may find in Schwarze (1981), Jackendoff (1983, 1987), and Vandeloise (1986, 1996). An additional purpose to the study of prepositions in the cognitive domain is to contribute to the theory of generative semantics and to artificial intelligence, as Herskovits (1986) and ZelinskyWibbelt (1993) show. Also, the universality of linguistic theories on prepositions have been tested in the field of language acquisition since the seventies (Bowerman 1994) up till now (Barbe and Tourrette 1995). PPs have thus become a coherent semantic set of linguistic units, interacting with the lexical context of the sentence. Third, a number of publications have appeared during the last decade, in which PPs have become the showcase for the pragmatic vision of language. Among these, the monographs by Cervoni (1991), by Cadiot (1991, 1997) and by Berthonneau & Cadiot (1992) are of primary interest. The main point made by pragmaticists is that the structural frame for the study of PPs should not be the grammatical unit of the sentence but the discursive unit of statements. Naturally, a pragmatic study of PPs is inherently related to encyclopedic knowledge. Some scholars blend the encyclopedic with the lexical context. For example, Cervoni combines Pottier’s relational hypothesis with that of speech acts. He maintains that the insight into PPs comes, on the one hand, from pragmatic competence and, on the other hand, from the basic logic and the intuitional mechanics of the mental operation. Another example of a discourse-oriented study is the semiotic approach of Tobin (1994) within textual linguistics, where
Preface
the reference of the prepositions is determined according to continuous or discontinuous space. A recent development in pragmatic research on PPs comes from scholars whose scientific background is influenced by the theory of dynamics (for a specific reading of this theory and its application to linguistics one may refer to Fauconnier 1984, Nicolis 1991, or Visetti and Cadiot, this volume). A linguistic analysis relying on dynamics is inductive rather than deductive, allowing for the fact that the outcome may be multiple. Thus, linguists working within this framework display scepticism towards the cognitivist approach, that local or temporal space is a semantic primitive on which any study of PPs can be based. On the contrary, their common claim is that prepositions have to be accessed by their function and that the concept of space has to be redefined. Based on the assumption that memory is dynamic and that patterns are produced from basic algorithms and external stimuli, the extralinguistic context is of extreme importance. For example, Cadiot’s stand is that prepositions are gestalts and that the study of PPs should cross all disciplinary boundaries in order to capture the tension between the well-established signification of the word and the effects of its meanings. The idea that conceptual space is just one of the components of mental space is well supported by the visible polysemy of prepositions and by the constant difficulty to verify antonymic relationships inherent to conceptual space. While non-dynamic research considers words as representations of extralinguistic objects of reference, which exists independently of our experience, mental space consists of verbal scenes leading to schematic forms that allow us to apprehend and organize the world (cf. Victorri 1999). Thus, this constant development in linguistic theory allows us to tackle new problems and to raise new disputes. In a conference held at the University of Haifa, in early June 2000, a group of scholars with various scientific orientations, from both Israel and Europe, came together and presented papers on aspects of prepositions which are at the center of any serious discussion of the topic, as suggested above.2 For this volume, most of the participants have further elaborated their studies. Although the contextual element was not given as a prerequisite for the conference, it is explicitly or implicitly included in the articles included here.
. The contributions First, the mental schemes of PPs are presented by a group of scholars from France, who have integrated the theory of dynamics into their research. These
Preface
papers analyze hypotheses with various semantic conjectures, as the distinction between content and structure, or the axiom of the prototypical locative and temporal content of prepositions. Yves-Marie Visetti and Pierre Cadiot have joined their efforts in order to discuss some aspects of the theory of catastrophe. They analyze the principles according to which it is possible to build an analogy, or even continuity, between language and perception. Several misleading options are identified, arising from (i) erroneous models of perception, (ii) a disregard for polysemy as a fundamental property of language, and (iii) the inability to allow the necessary interactions between discourse and the most interior level of ‘linguistic schemes’. Starting from the example of prepositions, they challenge all these difficulties, in order to put forward general semantic principles applicable to all categories of words and constructions. On this basis, the construction of semantic forms can be distributed between three layers of meaning, or ‘stabilization and development phases’, termed motifs, profiles, and themes. In an additional article, focusing on the French language, Pierre Cadiot argues that PPs are understood on the basis of experimental types, demonstrating the implications of his epistemological views with the help of the colorful and colorless use of en (in), par (through), sur (on), sous (under) and contre (against). The spatial language is re-examined and new, thematized categories are proposed. Finally, Franck Lebas, too, questions the validity of the spatial theory. He bases his arguments on Guillaume’s psychomechanics and on the ideas outlined by Cadiot, insisting that the semantic material of prepositions is provided by their use. His discussion centers thus on the extrinsic properties of dans (in), concentrating on its temporal prospective reading. Next, a group of scholars from England and from Israel take up the discussion of cognition and natural language processing in English. They verify to what extent adverbial modifiers add to the meaning of unmodified sentences in the context of verbal aspect. Where temporal PPs are concerned, David Brée and Ian Pratt-Hartmann conclude that they are additive in meaning, independently of the verbal aspect in the core sentence, thus providing cognitive simplicity. With regard to locative PPs, Pratt-Hartmann and, this time, Nissim Francez discover a more complex relationship. Locative PPs behave differently with events and states, having an additive effect with the former and a restrictive one with the latter. Third, several contributions deal with the preposition in Hebrew. With regard to morpho-syntax, Esther Borochowski and Hava Reppen present a detailed study of why certain modifiers of nouns admit one preposition and not another. It is true that this problem has been studied in other languages as well,
Preface
for instance in French, where the noun modifier may be introduced by de, à or by en (see Bartning 1993 or Cadiot 1993). Yet, these prepositions are normally called colorless, whereas Hebrew uses morphemes that would normally be considered lexical, such as mi (from), be (in) or locative le (to), and never a genitive morpheme like shel (of). The authors consider the prepositions in question to be grammatical and not lexical, but they show that the head of the PP furnishes the semantic motif for the use of either of them. The next two items also take into account the prefixed prepositions be, le and mi, although from a sign oriented perspective. To begin with, mi has a free form min. Yishai Tobin shows that the pair mi/min is far from being a free variation, as some believe. According to him, the prefixed morpheme mi is marked for Semantic Integrality in continuous space; it reveals more complex morphological characteristics than the unmarked form min, and it will replace the latter once particular themes are at work within the structure of a text. The second item concerns the interaction of be and le with the negative morpheme lo, crucial to the formation of the Hebrew equivalent of “without”. Susanne Feigenbaum discusses the concurrent forms bli/lelo/blo according to the criterion of semiotic valency defined by Peirce. Based on data gathered from a large corpus, she shows that three semiotic categories are responsible for its syntactical and lexical use and that the relatively new lelo is most productive, whereas the classical forms have gained expressive force. The two subsequent articles are in search for normativity concerning the PPs in spoken or written Modern Hebrew, exemplifying thus the dynamics in this language. For one, morphological aspects like affixation and inflection have an impact on the development of prepositions. Inflectional forms have changed in the course of time, with some of the biblical forms being maintained, whereas others have become archaic. In the sociolinguistic study by Inbar Kimchi-Angert, it is shown that the non-normative use does not correlate with a lack of knowledge or education and that archaic forms tend to be preferred by educated people. Miriam Ben-Rafael reports on the code-switching used by French immigrants to Israel involving French and Hebrew prepositions and compares it to the errors made by L2 learners of French. She suggests that the colorless prepositions à or en belong to the grammatical part of the lexicon for L1 speakers and that they are, therefore, error resistant. At the same time, L2 speakers, who lack the linguistic knowledge concerning these prepositions, miscalculate their use on the basis of conceptual knowledge. The final group includes treatments on Bislama, Russian, Judeo-Greek and Maltese, taking into account the parameters these speech communities have accepted for the semantic space of a preposition. A pertinent example that space
Preface
is perpetually negotiated comes from Dennis Kurzon’s inquiry into the use of the preposition long in Bislama, one of the English-lexified Melanesian Pidgins. The meaning of this preposition has expanded from locative and temporal usages (e.g. “on”, “at”) to extensive polysemy (e.g. “about”), and finally to zero meaning in its function as a complement marker. A further point to be examined is language change, with its implicit link to the theory of dynamics (cf. Cooper 1999). Contributions within this area come from two languages: Julia G. Krivoruchko deals with translations from the Bible into modern JudeoGreek. She describes the impact the Hebrew language has had on the development of the prepositional system in the target language. By challenging the norm of the latter, prepositions become signals of systematic and pragmatic change. Rami Saari analyzes in detail the local PP quddiem ‘in front of ’ in Maltese. It is exemplary for the dynamic changes going on in this language, which has been affected both by the etymology of Classical Arabic and by the system of European languages. The last contribution, by Mark Leikin, comes from the field of language acquisition. The difficulties observed among children acquiring locative prepositions in their first language (in this case Russian) are paralleled by similar problems among Russian-speaking aphasics in handling such prepositions. The findings add a chronological dimension to the grammar – context discussion: conceptual knowledge precedes grammatical knowledge, in so far as children manipulate locative PPs on the basis of their spatial cognition, while adults also rely on language experience.
Notes . The editors acknowledge with thanks the contributions of the French Institute – Embassy of France, Tel Aviv; of the Research Authority and of the Center of French Civilization at the University of Haifa. . The year 2000 became definitely the year of the preposition. The University of Tel Aviv organized its own international conference Prep An 2000: La préposition française dans tous ses états on 3–9th September, 2000. We would like to extend our professional and personal esteem to its organizer, Lucien Kupferman, for considering the two events complementary, as we did.
Preface
References Anscombre, J. C. (1993). Sur/sous: From spatial to temporal location. Lexique, 11, 111–145. Barbe, V., & Tourrette, C. (1995). The acquisition of spatial-position markers and field dependence/independence in young children. Année psychologique, 95, 425–440. Bartning, I. (1993). The preposition de and the possible interpretation of complex binominal phrases. Lexique, 11, 163–191. Bat-Zeev Shyldkrot, H. (1995). La grammaticalisation des prépositions. Revue Romane, 30 (2) 205–225. Berthonneau, A. M., & Cadiot, P. (Eds.). (1992). Les prépositions: méthodes d’analyse. Lille: Presses universitaires de Lille. Borillo, A. (1997). An aid in the identification of compound prepositions of time and place. Faits de Langues, 9, 175–184. Bowerman, M. (1994). Learning a semantic system: What role do cognitive predispositions play? In P. Bloom (Ed.), Language Acquisition (pp. 329–363). Cambridge, MA: MIT Press. Brøndal, V. (1950). Théorie des prépositions: introduction à une sémantique rationnelle. Copenhague: E. Munksgaard. Cadiot, P. (1991). De la grammaire à la cognition: la préposition POUR. Paris: CNRS. Cadiot, P. (1993). The preposition à between two nouns: Toward nominal compounding. Lexique, 11, 193–240. Cadiot, P. (1997). Les prépositions abstraites en français. Paris: Colin. Cervoni, J. (1991). La préposition. Louvain-la-Neuve: Duculot. Cooper, D. L. (1999). Linguistic Attractors: The Cognitive Dynamics of Language Acquisition and Change. Amsterdam: John Benjamins. Emonds, J. E. (1985). A Unified Theory of Syntactic Categories. Dordrecht: Foris. Englebert, A. (1993). The grammatical status of de. Journal of French Language Studies, 3 (2), 127–144. Fauconnier, G. (1984). Espaces mentaux. Paris: Minuit. Fillmore, C. (1968). The Case for Case. In E. Bach & R. T. Harms (Eds.), Universals in Linguistic Theory (pp. 1–91). New York: Holt, Rinehart and Winston. Gaatone, D. (1976). Locutions prépositives et groupes prépositionnels. Linguistics, 167, 15–33. Gross, M. (1995). The notion of place argument of the verb. In H. Bat-Zeev Shyldkrot & L. Kupferman (Eds.), Tendances récentes en linguistique française et générale: volume dédié à David Gaatone (pp. 173–200). Amsterdam: John Benjamins. Gross, M. (1996). The French forms être Prep X. Lingvisticae Investigationes, 20, 217–270. Herskovits, A. (1986). Language and Spatial Cognition: An Interdisciplinary Study of the Prepositions in English. Cambridge: Cambridge University Press. Jackendoff, R. S. (1973). The base rules of prepositional phrases. In S. Anderson & P. Kiparsky (Eds.), A Festschrift for Morris Halle (pp. 345–356). New York: Holt, Rinehart and Winston. Jackendoff, R. S. (1977). X-bar Syntax: A Study of Phrase Structure. Cambridge, MA: MIT Press. Jackendoff, R. S. (1983). Semantics and Cognition. Cambridge, MA: MIT Press.
Preface
Jackendoff, R. S. (1987). The status of thematic relations in linguistic theory. Linguistic Inquiry, 18, 369–411. Kupferman, L. (1996). De: de la préposition au mode de quantification. Langue Française, 109, 104–125. Lagae, V. (1998). Quantitative en: Pronoun linked to object function or position? Travaux de Linguistique, 35, 103–114. Lakoff, G., & Johnson, M. (1980). Metaphors We Live By. Chicago: University of Chicago Press. Langacker, R. W. (1987). Foundations of Cognitive Grammar, Vol 1. Stanford: Stanford University Press. Melis, L. (1992). Prepositions in alternation or fortuitous encounter: The case of à and pour in competition with lui as expressions of the ‘dative’ in French. Leuvense Bijdragen, 81, 311–326. Nicolis, J. S. (1991). Chaos and Information Processing. Singapore: World Scientific. Pottier, B. (1962). Systématique des éléments de relation. Etude de morphosyntaxe structurale romane. Paris: Klincksieck. Pottier, B. (1997). Cognitive and linguistic aspects of the expression of relations. Faits de Langues, 9, 29–38. Rauh, G. (1993). On the grammar of lexical and non-lexical prepositions in English. In Zelinsky-Wibbelt (Ed.) (pp. 99–150). Schwarze, Ch. (Ed.). (1981). Analyse des prépositions. III. Colloque Franco-Allemand de Linguistique Théorique à Constance. Tübingen: Niemeyer. Spang-Hanssen, E. (1963). Les Prépositions incolores en français moderne. Copenhagen: Gads Forlag. Taylor, J. R. (2000). The network model and the two-level model in comparison. In B. Peeters (Ed.), The Lexicon-Encyclopedia Interface (pp. 115–141). Oxford: Elsevier Science. Tobin, Y. (1994). Invariance, Markedness and Distinctive Feature Analysis. Amsterdam: John Benjamins. Vandeloise, C. (1986). L’espace en français. Paris: Le Seuil. Vandeloise, C. (1996). Meronomy, topological inclusion and the preposition dans. Faits de Langues, 7, 81–90. Victorri, B. (1999). Grammatical meaning. Langages, 34 (136), 85–105. Wunderlich, D. (1984). Zur Syntax der Präpositionalphrase im Deutschen. Zeitschrift für Sprachwissenschaft, 3, 65–99. Zelinsky-Wibbelt, C. (Ed.). (1993). The Semantics of Prepositions. Berlin/New York: Mouton de Gruyter. Zifonun, G. (1996). Ungewöhnliche Verwendungen von mit (1). Deutsch als Fremdsprache, 4 (33), 218–222. Zifonun, G. (1997). Ungewöhnliche Verwendungen von mit (2). Deutsch als Fremdsprache, 1 (34), 20–25.
Instability and the theory of semantic forms Starting from the case of prepositions Yves-Marie Visetti and Pierre Cadiot CNRS-Lattice, and Université Paris 8, CNRS-Lattice
.
Introduction
Phenomenological and Gestalt perspectives have become increasingly important in linguistics, which should lead to an improved interrelationship with semiotics and cognitive sciences. Cognitive linguistics, and to a certain extent what is known as linguistique de l’énonciation, have led the way.1 They have each in their own way established something of a Kantian schematism at the center of their theoretical perspective, developing on this basis what we might call a theory of semantic forms. They have introduced genuine semantic topological spaces, and attempted to describe the dynamics of the instantiation and transformation of the linguistic schemes they postulate. As a result, the idea of grammar itself has been modified, and at the same time we have obtained a better understanding of the phenomenon of polysemy, at least as far as this grammatical level is concerned. However, a closer analysis reveals a number of difficulties, which call for a better understanding of what a genuine phenomenological and gestalt framework should be in semantics. First, if we agree with the fact that there is a privileged relation, or some kind of similar organization, between language and perception, we should make more precise the general theory of perception (and jointly of action!) which we take as a reference. Secondly, if we also agree with the idea of a specifically linguistic schematism, analogous to, but different from, what is needed for ‘external’ perception-and-action, its realm of dimensions should be determined: but we note here that there is a real, important disagreement between the authors. Thirdly, if we view language activity as a construction of genuine, ‘internal’ semantic forms based on linguistic schemes,
Yves-Marie Visetti and Pierre Cadiot
it is obvious that polysemic words should correspond to deformable schemes: but the works we have just referred to remain very vague on this point, and only propose lists of cases rather than genuine transformation processes. As a matter of fact, very few authors consider polysemy as a fundamental property of language which should be taken into account by linguistics from the very beginning. Starting from the example of prepositions, we shall try to challenge all these difficulties, in order to put forth general semantic principles, applicable to all categories of words and constructions. To this extent, a key question will be the relation between spatial and less or non-spatial uses of words. After that (Section 2), we shall come back very briefly to gestalt and phenomenological theories of perception, stressing the fact that they are semiotic theories, and not only morphological or ‘configurational’ theories of perception. As an immediate application to semantics, we will show the interest of this kind of approach to clarify the meaning of other categories of polysemic words (e.g. nouns). We shall then propose (Section 3) – but in a very sketchy way – some general postulates for a theory of the stabilization of semantic forms, based upon the mathematical notion of instability. The theory postulates three layers of meaning (or ‘phases’ of stabilization), called motifs, profiles, and themes. Taken together, they shape linguistic structure and semantic activity. They apply in exactly the same way in lexical as well as in grammatical semantics. And actually, they are conceived in the perspective of being integrated more tightly into a global ‘textual’ semantics, very similar to the one developed by Rastier (1987, 1989, 1994, 2000). This paper sketches a theory of semantic forms, which is a joint work arising from our common interest for semantics, gestalt theory, phenomenology, and complex dynamical models (e.g. Visetti 1994, 2001; see also Rosenthal and Visetti 1999). Examples have been taken – sometimes slightly reformulated – from P. Cadiot’s previous works. The semantics of prepositions, and more generally grammatical semantics, should be considered here as a very important starting point, and a first application of our theory. However our real purpose is much more global, and goes beyond that: we try to put from the very beginning – at least at a theoretical level – the whole semantics under the pressure of a fully dynamical, discursive, and diachronic perspective. The interested reader will find a much more detailed presentation in our recent book.2
Instability and the theory of semantic forms
. Linguistic schemes, and their ‘perceptive model’ With the semantics of prepositions, we find in a particularly striking form the problem of the relation to space and to the physical world. Most often there is a trend towards relying on a very general psychological prototype, according to which language, at its most fundamental level, encodes tangible and/or physical structures. Therefore, in order to describe prepositions, as well as other categories of words, linguistics should favor spatial and/or concrete uses, and even take them as a primary basis for all the other ones. This idea leads in cognitive semantics, and also in theories of grammaticalization, to a hierarchy of meanings, which starts from spatial or physical values, taken as literal meanings, up to temporal or abstract meanings, which are supposed to be derived from the previous ones by some kind of metaphorical transfer process. However, authors like Lakoff (1987), Langacker (1999), Talmy (2000) or Vandeloise (1999) underline that these primary values proceed from specifically linguistic schemes, which should not be confused with perceptive ‘external’ structures: indeed they are far more schematic, and at the same time genuinely linguistic, since for example they shape space by introducing ‘fictive’ contours or ‘fictive’ motions (Talmy 2000). But in spite of these very important addings, the primacy (and/or the prototypical status) of a certain kind of spatial and physical meanings is not really questioned. Furthermore, schematical relations between language and perception often rely on a very peculiar conception of the spatial and physical experience, which fails to appreciate the true nature of what the phenomenological tradition names the ‘immediate experience’ of subjects. It amounts to a reduction of this ‘immediate experience’ to a purely external space, and to a purely externalized physics of ‘forces’, both separated from their motor, intentional and intersubjective (even maybe social and cultural) sources. In this external space, language would identify relations between ‘trajectors’ and ‘landmarks’, conceived as independent, separate, individuals or places, entirely pre-existing to the relations they enter in. We think that this type of analysis extends to semantics a very questionable conception of perception, which stems from ontological prejudices, and not from rigorous descriptions. As a consequence of this wrong starting point, some works in the field of grammar retain only a very poor and abstract schematism; while others, or even sometimes the same works, address only the spatial or physical uses, hoping that the gap thus created between these uses and all the others will be filled by an appeal to the magical notion of metaphor. The approach we advocate is considerably different.3 It aims at going beyond this kind of schematism, while keeping some of its ‘good’ properties. The
Yves-Marie Visetti and Pierre Cadiot
exact abstraction level as well as the interior diversity of each scheme are a first key matter. On the one hand, abstract topological and/or cinematic characterizations (let us call them ‘configurational’) are too poor. On the other hand, schemes weighted from the beginning by spatial or physical values are too specific, and furthermore rely on a very peculiar conception of spatial and physical experience. Actually, more ‘intentional’ or ‘praxeologic’ dimensions, intuitively related to ‘interiority’, ‘animacy’, ‘expressiveness’, ‘appropriation’, ‘control’, ‘dependence’, ‘anticipation’ etc. are needed. By entering in the process of discourse, all these dimensions – configurational or not – can be neatly put forward by speech, or alternately kept inside the dynamics of the construction of meaning as a more or less virtual aspect of what is thematized. In particular, configurational or morphological values are not a systematic basis: they may be pushed in the background, or even disappear, superseded by others, which are quite equally fundamental and grammatical. More generally, these motifs, as we shall call them as from now, to distinguish them definitely from the problematics we criticize, appear deformed, reshaped, in various profiles, abstract as well as concrete. A motif is a unifying principle for this diversity of uses, which can only be understood if one takes into account from the very beginning dimensions of meaning which cannot be integrated into the narrow frame of a schematism – at least if by a ‘schematism’ we mean something (still predominant in cognitive linguistics) which can be traced to kantian philosophy (Kant [1781–1787]; for a discussion on this point, cf. Salanskis 1994). Of course we have to consider all these fundamental dimensions at a very generic level, so as to assume that they are systematically put into play, and worked out by each use. But generic as they may be, our thesis is that these dimensions can be traced back to the immediate experience of perception, action and expression, if they are conveniently described in their social and cultural setting. This is why we decided to drop the designation of scheme, and to adopt the word motif to express the kind of ‘seed of meaning’ we wish to attribute to many linguistic units. Indeed, the word ‘scheme’ evokes a certain immanence, a restricted repertoire of categories not constituted by culture and social practices, and a privilege granted to a certain biased representation of the physical world. It is therefore a term not suitable for indicating an historical, cultural, ‘transactional’ unifying linguistic principle, whose function is to motivate the variety of uses of a grammatical or a lexical unit. What is then our own perspective? In summary, we advocate:
Instability and the theory of semantic forms
–
–
–
–
–
No privilege for spatial or physical usage of words (as conceived by current trends in cognitive linguistics), and consequently no doctrine of metaphorical transfer of meaning. Search for grammatical motifs, which are ways of giving/apprehending/ displaying, immediately available in all semantic domains, without any analogical or metaphorical transfer stemming from more specific values, allegedly conceived as the primitive ones. Rejection (most of the time) of purely configurational versions of those motifs: on the contrary, a motif, especially a grammatical one, is an unstable, and at the same time a strongly unified, means of building and accessing ‘semantic forms’; it ties together, and defines a kind of transaction between many dimensions which cannot be dissociated at its level, but at the level of profiling inside specific semantic domains. Rejection of an ‘immanent’ explanation of the variety of uses, based upon an identification of the motif with some kind of ‘autonomous’ potential; indeed, depending on the specific use, some dimensions of the motif can be further specified, enriched with other dimensions, or on the contrary virtualized, even completely neutralized. The parameters controlling the profiling dynamics are not an internal property of the motif : the relation between the motif and a particular profile has to be considered as a linguistic motivation, because profiling a motif consists of recovering it within other dynamics, brought about by the co-text and the context. A conception of the grammatical motifs (e.g. a motif of a preposition) as highly unstable ‘forms’ (or seeds of forms) which can be stabilized only by interaction with the other constituents of surrounding phrases, or even by more distant elements of the co-text: as we have said, this stabilization is not a ‘simple’ instantiation of the motif, but a recapture by other non immanent dynamics giving rise to the variety of its profiles.
Actually, this approach is very general, and applies both to grammar and to lexicon. It is strongly different from other approaches currently worked out by cognitive linguistics. We have already underlined some differences in the analysis of the grammatical expression of space, and in the assessment of its status relatively to the global functioning of the relevant units (cf. Cadiot, this volume, for the case of prepositions). But the situation is the same for grammar as a whole, and in particular regarding its difference with the lexical aspects of meaning. As a matter of fact, all the different trends in cognitive linguistics have severely criticized the autonomy of syntax postulated by generative linguistics.
Yves-Marie Visetti and Pierre Cadiot
But they have maintained a clear cut separation between structure and content: ‘structure’ refers to a central and universal schematic level of meaning, called grammatical, and ‘content’ refers to all the remaining dimensions (concepts, notions, domains, etc.) specifically brought by the lexicon. Grammar is therefore a kind of imagery, a way of structuring, of giving ‘configurations’ to all semantic domains, and also to the ‘scenes’ evoked by speech. Imagery includes: – – – –
structural organization of ‘scenes’ (space, time, movement, figure/ground or target/landmark structure, separation between entities and processes) perspective (point of view, ways of going over the scene) distribution of attention (focusing, stressing) and, for Talmy, Vandeloise, etc. (but not for Langacker) some less configurational dimensions, like the system of forces, or dimensions like control, or access.
For all these authors, this kind of schematism is specific to language (e.g. topological, not metric), but has many common properties with perception of external space. We have just criticized this schematism as well as the conception of perception to which it is correlated. Indeed, concerning the type of the grammatical schemes, and their relation to our external, everyday perception, we have seen that two main attitudes can be distinguished: sometimes, the schemes are from the very beginning merged with a certain conception of the physical world (Talmy, or Vandeloise 1991); and sometimes they are abstract, and purely configurational (Langacker). The reason for this false alternative is simple: there is no generic diagrammatic representation of action, animacy, interiority, expressivity, intentionality and anticipation, as they are constituted by their cognitive, social, cultural and linguistic modalities. So, whenever one tries to take some of these dimensions into account, the only way to recover some expressions of them is to resort to the physical experience – which is at the same time wrongly apprehended. Once again, such a conception of our ‘immediate experience’ not only provokes an impoverishment of the theory of grammar, it also introduces a gap between grammar and lexicon, as well as between the so-called literal meaning and the figurative ones. Finally, so to speak, the only relation between grammar and lexicon is grammar! And the only relation between the registered basic lexicon and the variety of uses is a metaphoric relation to space! In short, we think that cognitive linguistics has up to now too strongly dissociated ‘structure’ (identified to the schematical dimensions of meaning) from ‘content’. Therefore the very foundation of semantics is still grammar, understood as a fairly autonomous device, in spite of whatever these authors may say about the continuity between grammar and lexicon. In the
Instability and the theory of semantic forms
same way, there is a tendency to see grammaticalization as a pure bleaching process, which only retains values pertaining to a universal repertoire set once and for all. We think, and actually numerous linguistic analyses show, that we need a richer theoretical apparatus, inspired by an integrated theory of perception, action and expression, really susceptible to be transposed into grammatical and lexical studies, which would then become more tightly unified if we view them in this perspective. We look therefore towards a fully intentional theory of perception, a semiotic and ‘transactional’ theory of immediate experience, constituted by the simultaneous grasp of practical (praxis), axiological (ethics and esthetics), and subjective values. In order to recover such a theory, we would have to read carefully the gestaltist writings, notably those of the Berlin School (Wertheimer 1959; Koffka 1935; Köhler 1920), the message of which has been weakened by cognitive linguistics. Beyond that, we would have to come back to the phenomenological tradition (Husserl 1982, 1989, 1973; Gurwitsch 1964; Merleau-Ponty 1962), to Cassirer’s philosophy of symbolic forms (1953), and also to Vygotsky (1978)’s developmental psychology, which gives to social practices a constitutive role.4 Once a much more relevant model of perception-and-action is recovered in this way, we shall be in a position to transpose it into semantics, in order to provide for a more complex interplay between the dynamics of constitution and the constituted meanings than anticipated by current schematisms. Language activity will be described as a process analogous to what is called a complex system in other disciplinary areas. Notably, the construction of ‘semantic forms’ will appear as a kind of microgenetic developmental process, with concurrent unstable and stabilization ‘phases’. The description of the linguistic motifs as unstable seeds of forms (in a gestalt sense of the word ‘form’, transposed to semantics) is thus fundamental in our perspective. This will result in three semantics ‘modes’ or ‘phases’ in the dynamics of the construction of meaning, which we have called motifs, profiles, and themes.
. Gestalt and phenomenological theories of perception: Towards a theory of semantic forms . Gestalt, phenomenology, and language activity Among the several fundamental references quoted at the end of the preceding section, we shall limit ourselves, and even then in a sketchy manner, to
Yves-Marie Visetti and Pierre Cadiot
those that are from a Gestalt perspective.5 Gestalt psychology has often been reduced to its morphological and morphodynamical aspects (especially with the famous slogan ‘the whole is more than the sum of its parts’). Actually, it describes a much richer and deeper unity between perception, action and expression. It is precisely this kind of unity that we want to put at the core of the construction of meaning, seen as a construction of ‘semantic forms’. Under the expression ‘semantic forms’, we do not refer to a sensation conceived in isolation (even if the theme of the discourse resorts to our concrete, practical world), but to semiotic and multimodal ‘forms’ unfolding through language activity as units in all domains of thought and experience. We do not either take ordinary perception as a foundation for linguistics, but rather take it, when described according to the phenomenological style, as an essential correlate, and a particular illustration of the construction of meaning. Once again, the choice of a theoretical perspective on the perceptual experience is decisive for any linguistics which claims to find here a model, and perhaps an origin. For example, turning back to our fundamental relationship with space, we find currently in linguistics three main conceptions of this reference space: a. physical, objective space, with a universal geometry, and objective, universal categories of ‘objects’; b. perceived, psychological space (still independent of culture and language diversity as a general framework – even if it is differently worked out by cultures and subjects); c. semiotic space, whose overall perception bears immediately upon social practices and cultural knowledge. Cognitive linguistics favor conception (b), with a very little touch of (c). Talmy speaks of fictive contours and fictive motions, as specific linguistic creations which might reshape the perceptive field itself. Talmy again, and Vandeloise, talk of forces and of control/containment relations which are at the very heart of the cognitive import of such basic words as against in English or dans in French. Actually, both tend to consider perception (with or without language), not as a 2-level activity (crude topology/geometry first, and only then force, control, access), but as an activity immediately integrating all these aspects. We are of course of the same opinion, but we think more radically that –
this approach of perception should be extended to include a broader repertoire of dimensions, which are unavoidably shaped by the social and cultural context;
Instability and the theory of semantic forms
–
perception, for what concerns its ‘continuity’ with semantics, is less a matter of encountering concrete, external things or places, than a matter of establishing qualified relations with things, space, and other perceiving agents; therefore another conception of subjective experience, as well as a more intersubjective perspective, are here fundamental. They put forth immediately intertwined attentional, modal, behavioral, axiological values, which cognitive linguistics treats only as secondary or derived, and at best in a very parsimonious way.
Precisely, the Gestalt and phenomenological tradition doesn’t dissociate the grasping of forms and values; as we have said, perception, action, and expression are here more tightly intertwined than in any other approach. ‘Forms’ in this sense: –
–
–
–
are to be simultaneously defined in all modalities (visual, auditive, tactile, motor and kinesthetic. . .), cf. the very important concept of synesthesy (objects, moves, changes that appear explicitly in one sensorial modality, are ‘felt’ in other sensorial and kinesthetic modalities as well); have immediate functional and agentive values (degree of spontaneity, distinction active/passive, differentiation of roles). Cf. Gibson’s affordances (1979), which have been directly inspired by Lewin’s Aufforderungscharakter and Köhler’s requiredness (1938), e.g. artifacts like a hammer, a chair, are perceived immediately with their gestual, postural, functional values; seeing a mailbox immediately sketches, depending upon our attitude, parts of an integrated social scenario; have also immediate esthetic and ‘behavioral’ values, with emotional resonance. Recall the examples of Köhler (1929, 1938): a wave, a musical crescendo. Cf. also Michotte’s work (1946) on the perception of movements as behavioral styles (walking, running [away, after], swimming, flying); include an immediate perception of forces or causes, of intentional moves (intersubjectivity, animacy, agency), and of expressive values (joy, fear, demand).
Perception in this sense has to be considered as instantiating a general structure of cognition, and not only as resorting to a purely sensorial and peripheral organization. As a slogan, we could say that ‘to perceive is from a single move to act and to express’. Perception already gives access to, and sketches, a meaning. It implies not only the presence of things, but a perspective of the subject, and a suggestion of acting. Perception in space is not grasping pure configurations or shapes, nor only a basis for other, subsequent ‘associative’ or ‘metaphorical’
Yves-Marie Visetti and Pierre Cadiot
interpretations: it is from the outset a dynamic encounter of ‘figures’ with no necessary dissociation between forms and values, apprehended in the course of actions, and deeply qualified by a specific mode of access or attitude. It is this notion of a qualified relation (which is a way of ‘accessing’, of ‘giving’, of ‘apprehending’) that we want to transpose into semantics, in order to view it as a kind of perception and/or construction of forms. At this level, any distinction between abstract or concrete, or between interior or exterior perception, is irrelevant. Therefore ‘figures’ are objective counterparts, phenomenological manifestations of the relations we have with them. Needless to say, the perceived relations are not prescribed by some kind of pre-existent exterior world: they are conditioned by a global perspective or purpose, which constitutes subjects and objects simultaneously. Any perceptive relation can thus be modulated towards its subjective side, or towards its objective one, in a way which is constitutive of the act of perceiving. As a relation, it can be transposed to multiple situations or referents. Language only radicalizes this: at its deepest level, it defines, differentiates, and records primarily the relations – not the referents, which depend upon another, more thematic, linguistic and cognitive level (e.g. the contrast such as house/home: the two words possibly have the same referent, but not the same relation to it). And as soon as language comes into play, relations are definitely socially reconstructed. On the whole – and this is called polysemy – they are intrinsically transposable to a diversity of ‘themes’, in a variety of semantic domains correlated to a variety of social and cultural practices. Language activity appears, up to a certain degree, as a ‘new’ layer of social perception, made of intrinsically transposable, highly unstable seeds of ‘forms’ (forms of relations), to be stabilized in a variety of domains: experiential (qualia and their evaluations), practical (actions and their domains), theoretical, mythical, etc. . An insight into the semantics of nouns In several recent works, we have applied to a set of strongly polysemic nouns of ‘Basic French’ a description principle, which takes into account all their uses on an equal footing.6 We were thus moving away from the dominant lexicological approach, which promotes a certain notion of ‘literal’ meaning, supposedly combining tangible, concrete, reference and denominative function in a primary layer. As for us, on the contrary, the meaning of the most frequent nouns can and must be devised long ‘before’ any logic of classification or of categorization of referents. As a matter of fact, nouns – at least the most frequent ones – are ‘ways of access’, or ‘ways of establishing relationships’, prior to be-
Instability and the theory of semantic forms
ing labels in a game of entities categorization and denomination. Their prior function is to be interpreted in terms of analogical generative potentials (or seed of forms), which we call motifs. These motifs may be intuitively presented as generic ‘experiential bundles’, and described, in the phenomenological and Gestalt style, according to different intertwined modalities: perception, action, qualia and evaluation. Of course, we do not intend to give a full description of them: such an enterprise would be endless. The only thing to do is simply to put forward some of their principal dimensions, which are already very enlightening for the question of polysemy and of the so called ‘figurative meanings’. We shall only give here a few examples, trying to choose them in such a way that their polysemic distribution in French would be similar to the one of their usual translation into English. Let us start with some motifs which seem to provoke a perception and/or a construction of forms of visual type. The words which correspond to them seem indeed to have as a basic signification a ‘schematic’ form, which is easily, almost mechanically, transposable from one domain to another. –
–
ARBRE (‘tree’): arbre fruitier (‘fruit tree’), arbre généalogique (‘family tree’),
arbre syntaxique (‘syntactical tree’); also some uses considered as more figurative: arbre de la Vie (‘Tree of Life’), arbre de la Connaissance (‘Tree of Knowledge’); VAGUE (‘wave’): vague d’enthousiasme (‘wave of enthusiasm’), vague de chaleur (‘heat wave’), Nouvelle Vague (‘New Wave’).
These examples already show that motifs are not generally limited to configurational values (like a dynamical shape). Indeed, as in the gestalt theory of visual perception, motifs unify a bundle of synesthetic values going far beyond purely morphological determinations. For example, the motif of ARBRE unifies a branching process with a specific coherence stemming from the root, and giving rise to a perspective of growth, generativity, support. Depending upon the specific use, some of these dimensions are salient, others are pushed into the background, or even vanish. The important point is that language offers the possibility to grasp simultaneously all these aspects, because they are put into transaction with each other, and blend together, giving rise to a kind of coalescence. At the same time, language offers the possibility of dissociating this same unity (up to a certain point), and of enriching it (if needed), in order to give rise to a variety of profiles. Beyond the synesthetic values just exemplified, other nouns give direct access in their motif to dynamical-functional and practical (action-oriented) dimensions of meaning. Of course, this immediate relation to praxis makes in-
Yves-Marie Visetti and Pierre Cadiot
creasingly more problematic the attribution of an original ‘material’ meaning! Thus, for instance: –
–
–
–
BOUCHE (‘mouth’): can be used in French as in English for a river (‘mouths
of the Ganges’), a volcano, etc. French also uses it for the subway’s entrance (bouche du métro). One can see that the motif of BOUCHE includes dynamical-functional aspects, roughly evoking ‘entry and exit’; CLEF (‘key’): clef anglaise (‘adjustable spanner’), clef de voûte (‘keystone’), clef du succès (‘key of success’), clef du mystère (‘key to the mystery’), pointclef (‘keypoint’), mot-clef (‘keyword’). One can propose that the motif of CLEF unifies ‘exclusive access, (un)locking, and accuracy’. One can also see that the word clef can evolve according to a mainly perceptual and functional model (clef anglaise, clef de voûte), or according to a more explicitly intentional and practical model (point-clef, mot-clef, clef du mystère); TABLE (‘table’): table à jouer (‘gaming table’), table des matières (‘table of contents’), table de multiplication (‘table of multiplication’), tables de la loi (‘tables of the laws’). As a first approach, the motif of TABLE includes ‘visibility, availability, stand for an activity’; MUR (‘wall’): mur de briques (‘brick wall’), mur de Berlin (‘Berlin Wall’), se cogner la tête à un mur (‘to hit one’s head against a wall’), se heurter à un mur d’incomprehension (‘to come up against a wall of incomprehension’). These examples show that MUR integrates in its motif ‘to separate, to stand up, to surround, to protect, to hit’. It is to be stressed that an agonistic dimension is already immediately present in this motif, and not subsequently inferred (but of course it is neutralized in many denominative uses).
In this search for the motifs, the lexicalized figurative meanings play a very important role. Indeed, they do not function as heavily analogical mechanisms, but on the basis of an immediate promotion of the corresponding motif, which therefore appears as a general access principle, a qualitative relational index, immediately available in a variety of domains. Beyond the functional and practical aspects just evoked, some words record intersubjective relationships, giving them the status of a general relational principle. With the increasing importance of the intersubjectivity in action, the notion of Gestalt is expanded further, towards a social field of application. CLIENT (‘customer’) is here a very good example. Usually, in the lexical presentations of the word, one starts from the commercial meaning (client habituel ‘regular customer’), and one treats all the other meanings as derived or figurative. But in so doing, one does not see a unifying central motif, which could roughly be defined by ‘customarily related, who is taken care of, among
Instability and the theory of semantic forms
a series’. For example, one can say of a somewhat unpredictable horse: c’est un client vicieux (it is a bad-tempered customer); of the next opponent of a sport team: leur prochain client n’est pas facile (their next customer is not so easy); of a somewhat odd celebrity interviewed by a journalist: c’est un drôle de client (he is an odd customer), etc. Our view is that all these uses, whose list is still open, are not understood by going through the commercial isotopy (even if sometimes this analogy may fugitively appear in the mind); they are directly elaborated on the basis of the preceding central motif and of the ongoing context. Other words yet give access through their motif to a certain general ‘quality of sensation’, or to a certain ‘norm of evaluation’, which can be applied to an open set of entities, situations, states, etc., impossible to be determined a priori. These linguistic qualia have of course very important perceptual and emotional correlates, which are like their emblems; but being linguistic, these qualia are something else than these perceptible emblems: they are transposable to many kinds of experiences. Here are some examples: – –
–
–
BOUE (‘mud’): traîner quelqu’un dans la boue (‘to drag someone’s name in the mud’); NUIT (‘night’): the motif here tends to split into two sub-motifs, which nevertheless remain linked; the first evokes darkness: la nuit tombe (‘night is falling’), la nuit de l’ignorance (‘darkness of ignorance’), la nuit des temps (‘the mists of time’); the second evokes a period of rest: passer une bonne nuit (‘to have a good night’); FOUILLIS (‘mess’): we shall not try to explicit a motif, but only underline that it is neither concrete nor abstract, being totally entangled, as a generic qualia, between physical, psychological, and axiological aspects; NUAGE (‘cloud’): here again, a tentative definition of a motif would not be appropriate; it is better to delineate it through the specific phraseology of the word (colloquialisms), of which the motif is a unifying principle. For example: les nuages s’accumulent (‘clouds are gathering’: in French, it applies to many kinds of situations where a threat is looming, as in English ‘to be under a cloud’); être dans les nuages (‘to be in the clouds’); un nuage de tristesse passa sur son visage (‘his face was clouded with sadness’); and inversely, one can talk of un bonheur sans nuages (a happiness without clouds: ‘a perfect bliss’).
All these examples show that the notion of Gestalt can only be recast in semantics if it is taken in its widest diversity. Even less of course than for grammatical units, configurational or morpho-dynamical aspects do not suffice, since the motifs absorb many other dimensions. As testified by polysemy, by the (so
Yves-Marie Visetti and Pierre Cadiot
called) figurative meanings, and by their surrounding phraseology, nouns, at least the most frequent ones, register in their most internal kernel the coalescence of all these dimensions, much more than their dissimilation. This is why it is necessary to introduce motifs as unifying principles for the lexical diversity. On the other hand, this kind of unity does not define an invariant; on the contrary, motifs can be dissociated, and sorted out at the lexical level of profiling. Therefore, profiling does not consist of a ‘simple’ instantiation, but of a recapture of the motifs through more global dynamics. We contend that this process must be understood as a stabilization process, applied to unstable seeds, and this leads us to the global theory sketched in the next section.
. Motifs, profiles, themes: Instabilities and stabilizations Our global theoretical perspective presents language activity as a construction, and/or a perception of semantic forms. That does not mean that we intend to reduce it to the perception or construction of simple ‘external’ entities. On the contrary, it means that we aim at describing the more specifically linguisticsemantic part of a global process giving rise to ‘thematic forms’, which are inextricably both linguistic and semiotic. These forms can be sensible, imaginary, or ideal; and their construction depends upon the subject’s activity as well as upon the semiotic (social, cultural) ambient medium. This is why we have taken up concepts and principles inherited from the gestalt and phenomenological traditions: indeed, they put into place, at least at the level of individual subjects, the appropriate setting for this kind of widening of perspective, and at the same time for its focalization on language. In support of their approach of psychology, the gestaltists from the Berlin School (principally Köhler 1920) laid the basis of a general theory of Forms and organizations. Drawing upon their hypothesis of an isomorphism between the structures of the subjective immediate experience, on the one hand, and the functional dynamical organization of the brain, on the other, they devised a theory both phenomenological and physical, inspired by field theory, statistical physics, and dynamical systems. But they considered it at that time a speculative theory, or a building metaphor, and not a genuine model for the phenomenological mind and/or for the brain, hoping that future progress in neurosciences, in physics, in mathematics, and in the methodology of phenomenological descriptions, would confirm their insight. Since then, many works in various areas have pursued in the same direction, and actually gone far beyond, towards multiple theories of complex dynamical systems. Although we
Instability and the theory of semantic forms
do not offer here any precise project to establish a model, we think that calling upon the most general principles of the gestalt theory of Forms can help to stabilize our own theory, and to prepare its integration into the important interdisciplinary field just mentioned. As a reminder, here are some of the most fundamental features of this theory:7 – – – –
–
–
relations between parts and wholes: synthesis by reciprocal determination of all dimensions of the field of forms continuous substrates, continuous modulations of forms, and at the same time delineation of forms by means of discontinuities figure/ground and trajector/landmark organization no form without an ‘internal’ time of constitution: time of integration and/or differentiation, identification of forms through the dynamical chaining of different profiles forms are intrinsically ‘transposable’ (transposition does not mean a twostep process, going from a field A to another field B, but it refers to the immediate availability of an organizing ‘scheme’ in an open variety of domains) ‘schemes’ are not formal types, as in logical approaches, but ‘potentials’ to be actualized, evolving through practice.
Last, but not least, there appears in gestaltist writings, notably those by the so-called ‘microgenetic’ schools (Werner 1956; Flavel & Draguns 1957; cf. also Kanizsa 1991: 118), that forms are to be considered as the result of dynamical stabilization processes, i.e. as units in an ongoing continuous flow, comprising more or less stable ‘phases’, depending on the moment and on the part of the flow. Of course, for lack of mathematics and physics, it was only possible to develop these concepts of stability/instability as from the 1960s. This more recent aspect of the theory is essential for the theory of semantic forms we want to build. Modern mathematical and physical concepts of instability, and recent advances in the theory of complex systems, allow us, not to build a model for the moment, but at least to conceive and to formulate a unified setting for language activity seen as a construction of forms.8 Without taking into account such a notion of instability at the very heart of the linguistic theory, we would be obliged, either to drop the immediate link between language and action-perception (as logical approaches do), or to consider concrete, externally stabilized, referential uses as a first building layer (as cognitive linguistics mostly do). In all cases, this would imply the isolation of literal meanings, and the processing of all other uses by means of metaphor and metonymy (which
Yves-Marie Visetti and Pierre Cadiot
strangely enough would admit at a later stage transformations such as mixing, deformations, etc. excluded from the first stage). Let us see now how the dynamical principles we favor are redistributed in our theory. . Motifs Let us first recall that we view linguistic motifs as unstable seeds of semantic forms, which can be stabilized only by, and with: (i) the other constituents of surrounding syntagms, (ii) more distant elements of the co-text, and (iii) an ongoing context-and-topic. This stabilization process is not a ‘simple’ instantiation, but a recapture of the motifs by non immanent profiling dynamics, partly linked to specific semantic domains, partly constituted by generic grammatical means. This entire process gives rise to the variety of lexical profiles (uses) of the words. Each motif blends, and intertwines, different dimensions that can be dissociated only later (if ever) in the stabilization process, by inscription into a more specific semantic domain. Therefore, a motif does not belong to a specific domain: on the contrary, it encompasses several ones (to the extent that ‘semantic domains’ can always be sharply distinguished from one another). In a sense, motifs define the functional kernel of many linguistic units, whether monomorphemic or polymorphemic. Most importantly, these unstable ‘seeds’ do not entirely control their own stabilization parameters from the inside, nor are they by themselves generative of the lexical values they motivate. Language activity has a polysystemic, multi-level organization, with strongly interacting and at the same time possibly uncoupled ‘levels’. As a physical (thermodynamic) metaphor, this organization is not that of a homogeneous system, made of uniformly individuated and stabilized entities. It is that of a heterogeneous medium, with several coexisting more or less differentiated ‘phases’, ongoing phase transition, and diffusion/reaction processes. More precisely, for our dynamical approach of the semantic reconstruction up to the level of text and discourse, we need –
–
coalescence and/or transaction between dimensions of meaning, the dissociation of which could only happen ‘downstream’ in a stabilization process in the co-text and context: this implies to introduce ‘upstream’, and constitutionally, a structural instability at the level of motifs (see a little lower in the text); openness and immediate susceptibility of the linguistic motifs within the thematic and situational frame, allowing for a generalized form of indexi-
Instability and the theory of semantic forms
–
cality (rooted in the themes of the discourse); because of this plasticity of the motif, and unpredictability of the exact part which is taken up at each occurrence, its internal organization has to be a complex, chaotic one (see lower in the text); permanence of this type of organization through the traditional layers of integration (morphemes, words, phrases, texts).
In order to build such a theoretical linguistic concept, it is quite relevant to draw upon the various mathematical notions of instability. It even appears that we must go further than the Elementary Catastrophe Theory of R. Thom and E.C. Zeeman, from which the very few existing semantic models derive.9 Pursuing the same lines, we can represent the participation of a given unit in the global construction of meaning (e.g. the contribution of this unit to the construction of an ongoing scene or scenario, or to a network of ‘mental spaces’) through a dynamical system operating in a certain semantic space, each state of which corresponds to a particular contribution of the unit. This dynamical system is coupled to certain parameters to be found in the co-text and in the context, and it controls in a reciprocal way some (other or the same) parameters in its semantic environment. If the analysis is situated at a microgenetic temporal scale, it is possible to postulate that the essential result of the construction is directed by the ‘asymptotic’, stabilized states of this dynamical system. In the right cases, the set of all these asymptotic states constitutes an attractor set, i.e. a region of the semantic space (a point, a cycle, or a more complex set, once called a ‘strange attractor’), towards which converge all the trajectories, whatever their initial position in a wider region, called a basin of attraction. This attractor set represents a more or less complex state of the unit concerned, which may change according to the contextual parameters influencing the dynamics (and which also reciprocally influences these parameters). Thus, depending upon the contextual variations, a given attractor set can slightly move in the semantic space, without changing qualitatively its internal ‘geometry’ (structural stability). It can also change qualitatively, or even split up into several other different attractors (‘structural instability’, ‘bifurcation’). In this way, a linguistic unit appears as a more or less unstable dynamical system, engaged in a reciprocal determination process with a certain part of the context. This ‘deformation’ process generally results in a more stabilized version of the initial dynamics, which drives the system into a certain attractor set, concentrating, so to speak, on the resulting value, or use, of the unit. Hopefully, then, the modelizing process would consist in defining a motif as an unstable dynamical system, and in studying it relatively to a family of possible deformations (i.e. according to the
Yves-Marie Visetti and Pierre Cadiot
different semantic fields and phrase constructions where the word appears in a corpus), so as to describe exhaustively the different cases of stabilization, as had once been done in other areas by the Elementary Catastrophe Theory. Structural instability is one of the key concepts of the dynamical system theory. But there is another one, coming from the seminal work of D. Ruelle and F. Takens (1971), and which we have just alluded to (cf . Bergé et al. 1984; Dahan-Dolmenico et al. 1992; Ruelle 1993, 1996). Even if the ambient dynamic is stable, its asymptotic states can be very complex, because the corresponding attractor set itself has a very intricate topological structure, constituted by a bundle of dense, entangled trajectories, going through it in an unpredictable way (‘strange attractor’). The attractor then represents a chaotic state, i.e. a global envelope of stabilization, which is accurately defined from an ideal geometrical point of view, but the trajectories of which cannot be known in their exact asymptotic evolution (unless the initial conditions are perfectly determined, and the computation ‘infinite’). This important property, called ‘sensitivity to the initial conditions’, defines a kind of ‘stable turbulence’, which is of very considerable interest for our concept of motif, to the extent that a motif can be promoted as such by certain uses, in particular the ‘figurative’ or ‘metaphoric’ ones. In this model, the promotion of a motif corresponds to a chaotically organized state, which results in trajectories inside the semantic space whose asymptotic evolution remains unpredictable. We, therefore, see in which various meanings we need here to take up the mathematical concepts of instability in semantics: –
–
–
first within the framework of a stable dynamics, comprising a chaotic attractor set, and consequently a kind of regional instability (allowing the promotion and the contextual elaboration of a motif, with fluctuating trajectories, and unpredictability of what is asymptotically integrated in this kind of use); secondly, in the framework of light global fluctuations of the global dynamical landscape, which do not imply important qualitative transformations (but only amplifications, or a kind of smoothing, inducing more simple or generic variants); thirdly, in the case of genuine structural deformations (structural instability), which modify the topology, and/or the number, of the attractor sets and their basins, and so reveal new principal contrasting dimensions, allowing a whole polysemic diversity of uses.
Let us emphasize that these phenomena can be simultaneously observed, depending on the dimensions on which the analysis is directed. Moreover, and
Instability and the theory of semantic forms
this is most important, two dynamics can be topologically very similar, and even have exactly the same attractor sets, while strongly differing in the degree of their structural instability. When this dynamical setting is combined with a ‘morphemic’ conception of motifs (coalescence, transaction, between dimensions not yet dissociated at this level), several aspects of the construction of meaning, which are ordinarily presented as very distant ones, can be brought together without incoherence. Strange as it may appear, ‘figurative’ meanings appear very akin to the generic ‘definitions’ devised by lexical studies, and also to the generic ‘potentials of meaning’ brought out by linguistic theories. As a matter of fact, a generic definition of a motif promotes it through a global description of the topology of its attractor set, which reveals the intertwining of other linguistic motifs on its ground, while a figurative meaning also promotes the motif, not in a synoptic way, but rather by collecting some of its aspects along a largely unpredictable trajectory.10 In both cases, the motif, as a dynamical chaotically organized unit, is perceived as such in the discourse – though in a more or less synoptic and global manner. What can be said, then, about the ‘meaning potentials’, which various linguistic theories postulate in order to introduce some kind of unity and generativity at the heart of a lexical unit? In a dynamical setting like ours, such a ‘meaning potential’ is only another structurally unstable form of the motif, topologically very close to its chaotically stable ones (promoted by definition and/or figurative meanings). This structurally unstable form represents the generative potential of the corresponding linguistic unit, in as far as it is immediately available in an indefinite number of semantic lexical fields, through recapture and re-stabilization within their own dynamical frameworks. Each use then corresponds to a certain stabilization path. In this way, polysemy becomes a central and constitutive phenomenon in language organization and activity. It may, therefore, be seen that it is possible to bring together in a unified setting deeply entrenched aspects of language activity, as well as more innovative ones: the key being to recognize at a theoretical level, and from the very beginning, a certain dynamical state, or semantic ‘phase’ (let us use here again the thermodynamic metaphor), which potentially combines the different forms of instability we have just mentioned. There now remains to see how this primordial instability is most of the time recaptured, and re-stabilized (‘profiled’) in order to construct the variety of semantic forms.
Yves-Marie Visetti and Pierre Cadiot
. Profiles What do we then call profiles, or profiling?11 Roughly speaking, profiling, which is of course context-and-situation dependent, occurs –
– –
by stabilization in lexical organizations (e.g. domains like music, cooking, sailing, architecture, business, law, mathematics; fields articulating several experiential domains and practices; denominative paradigms); correlatively through grammatical units and constructions; also through discourse organization (e.g. anaphors, comparisons).
From the point of view of the present theory of semantic forms, profiling implies – – –
figure/ground repartition of the lexical content in semantic fields;12 possible dissociation of the involved motifs, through stabilization in cotext and context; enrichment by new aspects, or on the contrary impoverishment of the involved motifs.
By this process, words (initially considered with all the ‘morphemic openness’ of their motif) become lexical units indexed on lexical classes, with more stabilized and individualized meanings.13 Plasticity of the motifs through profiling is a key point in our theory. Depending on the reciprocal determination of the co-text and the context, some features can be completely neutralized, or on the contrary made salient. In many cases, some features are so to speak virtualized; they remain as a possible ‘aspect’ inside the dynamics of construction, without being explicitly integrated in the constructed forms. Nevertheless, they are as it were reserved, and may come back to the foreground if the discourse needs it afterwards.14 One of the reasons of these virtualization processes is that, by entering into a specific semantic domain to contribute to the formation of a lexical unit, a motif functions as a simple motivation; its proper contribution can be superseded by other afferent features, which are more important in this context. These features are either recorded in the lexicon, as a particular use of the word, or indexically integrated on the spot. But let us emphasize that even if these modulations of meaning are already registered in the lexicon, it is always the global stabilization dynamics in the current phrase, or in a larger cotext, and the peculiarities of the ongoing topic, which determine what exactly will be taken up from the lexical registration. Let us also point out that profiling is a differential process, which happens through contrasts and coordination
Instability and the theory of semantic forms
between several inter-defining lexical units, which are the results of reciprocal stabilization paths. At the level of a clause, lexical profiles stabilize through grammatical units and constructions, whose meanings stabilize correlatively at the same time. In this way, each statement appears as a view on the ongoing thematic organization, offering individuation, hierarchical structure, chaining, and grounding in the situation. In particular, a lexical profile can offer a certain view, or aspect, of a thematic unit. But this view is only a characterization of the unit: it cannot by itself decide what constitutes the thematic identity of the unit throughout the discourse.15 The determination of a profile is not in the first place a matter of type instantiation, even if pre-recorded types can come into play. Types, in our view, are anticipations which pertain to the thematic level of language organization (like scenarios or ‘actors’). The determination of profiles is performed, more fundamentally, by the mobilization of multiple frameworks which open the way to the thematization process. Among the most current frameworks are the –
– – – –
– – – – –
modulation of specific differences of a lexical unit on the generic ground of a class: a lexical class appears as an area in a semantic space, where features, depending on the specific unit, circulate from the foreground to the background (allowing, for example, metonymical shifts, e.g. school considered as a building, or as an institution); elaboration of functions and mereology, through lexicalization of parts and functions (a gaming table has legs, but a table of contents does not have any); exploration of the semantic neighborhood (synonyms, antonyms); fixation of a hyperonym, i.e. choice of a lexical unit bringing to the foreground some generic features of a given semantic class; introduction of a scalar structure into a class (e.g. few, many, too many; icy, cold, tepid, warm, hot); more generally, introduction of a global ‘geometrical’ structure into a class (putting for example a week, generic value at the center, and a dense, emblematic parangon on the periphery16 ); dissociation between processes, and roles or participants; choice of a part of speech (nouns, verbs, etc.); quantification, determination; aspects, tenses, modalities; constructions and grammatical functions.
As one can see, the problem of the construction of the lexicon, in its relation to the functional kernel of language (motifs and grammar), pertains indeed to the problematic of complex systems. First, the systemic variation is orga-
Yves-Marie Visetti and Pierre Cadiot
nized around unstable dynamics (here called motifs), which produces by being stabilized the diversity of profiles, whether new or registered. Secondly, there is a permanent adjustment of the system’s categorization networks. The lexicon is not a set of labels, nor a nomenclature of concepts processed as such by the arrows of reference. It is the historical and heterogeneous result of a multitude of accesses to themes; these accesses are never registered alone, but in clusters, and at different depths of unification, stabilization, and exteriorization. The lexicon can only function because it is liable to establish in its own formats, and to register immediately, distinctions which have been up to then original – which implies the weakening or the ‘virtualizing’ of other already established distinctions, without losing them. Lastly, the mobilization of motifs and profiles is aimed at the construction of thematic targets, which have their own structure. But language activity is not to be seen as a complete resorption of these semantic phases into a completely stabilized and/or externalized thematic level. It rather rests on the permanent co-existence of these different phases of meaning through the discourse. . Themes In order to complete the presentation of our theory of semantic forms, we must say at least a few words concerning the level of thematic forms and thematic spaces (recall that we take here ‘thematic’ in the full, literary sense of the word). At this level, the aggregation of profiles into thematic forms distributed throughout the text or the flow of speech (referring for example to narrative entities like actors, actions, and their transformations) is performed. In the same way as we have recalled in Section 2 some principles from the phenomenological and gestalt theory of perception and action, in order to transpose them into semantics, we should now come back to the phenomenological and semiotic theory of the thematic field (notably in Gurwitsch’s work), in order to connect our theory of semantic forms, to contemporary research on discourse, narratives, text semantics, etc. It would allow a criticism of the objectivist approaches, often correlated to referential semantics, and to the primacy of denominative uses. And it would also open the way to describing the new, original, motifs created by the discourse which elaborates, in a more or less innovative way, the pre-given linguistic motifs. In the framework of the present paper, we shall limit ourselves to the following fundamental points:17
Instability and the theory of semantic forms
–
–
–
–
In a situation of spontaneous speech, profiles are not perceived separately from the themes to which they give access, being nothing else than the transitory presentation of these accesses. The profiling dynamics cannot really enter into a stable state without a minimal thematic positioning, including the grasping of an ongoing topic. Profiling, therefore, depends constitutionally upon the movement of thematization. But it is not necessarily completely absorbed in the ongoing thematic units, since it can evoke lexically registered features not immediately relevant to the topic. At the thematic level a global dynamics of construction and access to themes is carried out, set as common objects of interest in the intersubjective field. Themes, in this sense, are partly externalized in our perceptive and practical world as concrete objects or as effective actions. But this is only a partial aspect of their identity which is made, as already said, of an organized history of profiled accesses (e.g. an history developed in the structure of a scenario). Language opens on an exteriority which can be simultaneously sensible, imaginary, and ideal. We are here in a complete opposition to certain approaches to referential semantics, which claim to favor concrete denominative uses, but actually have a very limited conception of what ‘reference’ means. To refer to a theme is not only to refer to its concrete facets, nor only to refer to its abstract, ideal, ones. Such conceptions do not allow one to understand that language by nature addresses fiction as well as reality. Think for example of a chess game, and its pawns; the theme of the game is a synthesis of many different aspects; and necessary as they may be, pawns are a simple material substrate, invested by this whole thematic organization. More precisely, their visible and tangible configuration only defines a crucial perceptive facet of the ongoing theme – i.e. of the game. A thematic unit builds up its identity through a synthesis of successive profiles; an actor, for example, is identified by the open set of the participant profiles, which compose it from one clause to another, and in this way define its transformations and interactions with the other actors in an ongoing narrative (once schematized by the scripts and frames of the psychological semantics). The thematic organization implies an elaboration of relevance (e.g. assessing the lines of continuity of topics, which are called isotopies in the semiotic tradition). Its tactics of implementation in the discourse still rests upon grammatical routines, but even more upon rhetorical norms set by the different domains of discourse and types of texts (e.g. the medical do-
Yves-Marie Visetti and Pierre Cadiot
–
–
main, with types of texts like a medical report, a letter to a colleague, a scientific paper, etc.). The logic of categories or classifications, the denominative uses, or also the transformations of referents, are to be understood at that level, starting from generic thematization frameworks (e.g. time-space organization, practical world of everyday life, a-temporal systems like logical or conceptual systems), and also from more specialized frameworks, upon which technical terminologies depend. To sum up, the thematization activity can and must be understood at the semantic level of its linguistic accesses and effects, and without contradiction, as a global access to other textual, pragmatic, imaginary, conceptual, perceptive, and practical layers, which are less directly linguistic, but still semiotic (therefore cognitive and social at the same time).
The concept of motif, as we have seen, has allowed us to describe the functional kernel of language, and its unfolding in a permanently adjustable lexicon. But speech does not only stabilize, it also renews linguistic and lexical instability. Existing motifs are modulated, and new ones are sketched (even deeply elaborated) through discourses, some of which are the starting point of an instituted modification, effectively registered in diachrony. It is therefore crucial that the dynamical structure of motifs (which is, so to speak, the most internal ‘phase state’ of language) allows an immediate interaction with the ongoing thematics. This kind of susceptibility makes it possible to index a renewed motif on an existing lexical unit, which condenses some essential dimensions of a new original theme, after having cut out part of its structure (e.g. its precise event structure). Of prime importance are the metaphoric innovations, and in a more commonplace manner, the uses mixing metonymic shifts and figurative operations.18 This is a complete reversal relative to other theories, which start from an ontology conceived independently from language, or give primacy to the reference to a practical, concrete world, without asking what perception or practice consist of. Ontologies are complex thematic constructions, they are a very peculiar result of text, discourse, and other social practices, and not a universal starting point for semantics. Quite differently, figurative meanings may be considered a very important clue to the study of motifs in lexical semantics, and these in effect transgress ontological divisions. As we have said, we postulate that in many cases this kind of use promotes a linguistic motif, i.e. elaborates and puts it forward without absorbing it completely in a conventionalized lexical profile. We gave examples concerning nouns in Section 2.
Instability and the theory of semantic forms
. Conclusion In this paper, we have systematically analyzed the principles according to which it is possible to build an analogy, and even a continuity, between language and perception. Starting from the case of prepositions, we identified several obstacles, or misleading choices made by cognitive linguistics. In particular –
– –
an erroneous model of perception, strangely disconnected from action, expression, and other essential dimensions of anticipation, leading to an inadequate separation between grammar and lexicon; the non-consideration of polysemy as a fundamental property of language; an inability of the theory to allow immediate interactions between the thematic developments in discourse, and the presumably most ‘interior’ level of language (the level of ‘schemes’ in cognitive linguistics).
In order to remedy all these deficiencies, we have introduced a more radically dynamical setting, which gives a fundamental role to the mathematical concepts of instability. On this basis, the construction of semantic forms can be distributed between three ‘phases’ called motifs, profiles, and themes. Indeed we claim that a theory of forms, suited to linguistics and likely to offer a coherent and global view on language activity, is possible only by introducing a diversity of concurrent semantic ‘phase states’, in a process made of structurally unstable or chaotic resources, and of partial stabilization dynamics. In this way, we rejoin a Humboldtian conception of language, which considers it as an energeia, i.e. not as a finished product, but as a self-organized activity. This implies that we consider languages, not only as means to build (re)presentations, but also as capabilities of being immediately modulated, transformed, by their own activity. In order to better support this conception from a cognitive point of view, it appears necessary to come back to phenomenological and Gestalt theories of perception and action. In this way, the discussion really opens on what can rightly be taken up again from them for semantics, while not forgetting the historical, social, and ‘transactional’ nature of what we have called linguistic motifs and lexical profiles. How is, then, the alleged unity of a word constituted? Our description makes it a compromise between three concurrent dynamic integration formats. At this level of the word (and even beyond in the case of compound lexical units), our theory puts in the center a ‘phase state’ of meaning, the instability of which (structural instability, instability in the sense of chaotic structures) can be described as morphemic. It makes possible the coalescence of dimensions which can be dissociated only further in a stabilization process, and thus
Yves-Marie Visetti and Pierre Cadiot
radically differentiates motifs from what other theories call types. Motifs are generic in a specific sense, since they allow homogeneous thematic developments, as well as heterogeneous thematic dissociations or blendings, as in figurative meanings. From this point on, the question of polysemy can be redistributed in a new way among the three postulated meaning ‘phases’. It is also possible then to take into account the immediate interaction between the ongoing discourse, and the linguistic anticipations registered at these three levels, whether in the time of a conversation, or of a literary work, or also in the general evolution of language in diachrony.
Notes . Under the French heading linguistique de l’énonciation, we mean a linguistic current which can be traced back to K. Bühler (1934), through the work of E. Benveniste (1966/1974), and more recently, through the important contributions of A. Culioli (1990/1999), O. Ducrot (1984), J.C. Anscombre and O. Ducrot (1983). . For a full presentation, see our book: Pour une théorie des formes sémantique – Motifs, Profils, Thèmes (Cadiot and Visetti 2001a). See also Visetti and Cadiot (2000), and Cadiot and Visetti (2001b). . It draws on several recent works on prepositions: cf. Cadiot (1991, 1997, 1999b). . Some among the most important references to the authors quoted in this paragraph are given in the References section. An excellent introduction to Husserl’s phenomenology can be found in Salanskis (1998). For an introduction of some of Vygotsky’s ideas in cognitive linguistics, cf. Sinha and Jensen de Lopez (2000). . Readers interested in having more details on phenomenology and perception, in the perspective of a transposition in the field of semantics, may refer to our book (Cadiot and Visetti 2001a, particularly Chap. 2). . Cf. Cadiot (1999a); Cadiot and Nemo (1997a, b, c); Nemo and Cadiot (1997); Cadiot and Tracy (1997); Visetti and Cadiot (2000); Cadiot & Visetti (2001a, b: Ch. 3, Section 3.1); see also Tracy (1997); Lebas (1999). . For a reconstruction of Gestalt theory, and its assessment in the contemporary field of cognitive sciences, cf. Rosenthal and Visetti (1999). For a presentation and illustration of a general dynamical paradigm in cognitive sciences, see Port and Van Gelder (1995), and most of all, J. Petitot’s works quoted in the References section. See also Petitot, J., Varela, F., Pachoud, B. and Roy, J.-M. (Eds.) 1999. . Far beyond the remarkable insights of the historical gestaltists, we see now mathematicians, physicists, biologists, computer scientists, modelizers in cognitive and social sciences, lay the foundations of a framework crossing their particular domains, and in which questions of stability and instability, invariant and variation, regulation and viability, can be profoundly re-thought, and sometimes modelized. The following terms make it somewhat
Instability and the theory of semantic forms
explicit: multiple spatial and temporal scales (at least two, micro and macro); importance of the topological, dynamical, and statistical characteristics; reciprocal determinations of local and global aspects; multiple dynamics for the formation of unities (births, deaths, coalitions, etc.); co-existence of several dynamical ‘phases’; adaptation, and active preservation of the internal and external viability domains; natural drift by coupling with a proper environment; behavioral repertoire organized around unstable dynamical processes, which constitute the system’s functional kernel. On the whole, all the system’s characteristics are historically determined. Given the great variety of the fields and the models involved, we cannot do better than referring the interested reader to the site of the Santa Fe Institute (www.santafe.edu), and to the entire series of the SFI’s Studies in the Sciences of Complexity. See also Weisbuch (1991). For a philosophical analysis of this paradigm shift, cf. Cilliers (1998). . Examples of semantic models based upon Elementary Catastrophe Theory can be found in Thom (1974) or Zeeman (1977); in Brandt (1986), Petitot (1985, 1992, 1995), or Wildgen (1982); and more recently, with different theoretical orientations, in Piotrowski (1997), or Victorri and Fuchs (1996). . This being said, the event of a figurative meaning does not only involve the level of motifs; it also implies processes at the thematic level, e.g. blendings, according to G. Fauconnier (1997) or Fauconnier and Turner (1999). . We use the same term as Langacker (1987), but in a different theoretical framework. There is no theory of instability in Langacker’s cognitive grammar. Furthermore, we have already criticized the strictly ‘configurational’ schematism he makes use of at the level of grammar. Lastly, we do not have the same conception of the ‘thematic’ level, nor of the alleged primacy or typicality of physical uses. . The ground of a semantic field corresponds to its most generic features, and also to some more specific, but less relevant or salient ones, when the field is dynamically stabilized by the occurrence of a specific lexical profile (playing here the role of a figure). . Not all words, however, possess a specific motif. Numerous technical terms are actually words indexed in a unique specific domain, which, furthermore, are very rarely used in a figurative meaning (examples chosen at random in a dictionary: galvanoscope, gastritis, gasoline). Of course, speech can always unlock the semantic game, and invent new meanings, which imply the creation of new (most of the time transitory) motifs. As an exercise, try for instance to say to your best friend: You are a real gastritis, or You are my favorite gasoline, and see what happens. . Cf. for instance the analysis of the polysemic word cut by D. Touretzky (1994). . Take for example a cooking recipe: the identity of the chicken (the central actor of the ongoing scenario) remains the same throughout. Nevertheless, its profiles change constantly, from the market up to the plate. . A same word can possibly be placed in either position, e.g. the word street which functions according to the case as the generic term of the paradigm of urban ways (avenue, boulevard, lane, etc.), and as a kind of ‘parangon’ in denser (at the same time metonymical and somewhat figurative) meanings, like to run about the streets, to find oneself out on the street, to come down into the street, man in the street . . .
Yves-Marie Visetti and Pierre Cadiot . For more details, see Cadiot and Visetti (2001a: Ch. 3, Section 3.2.3). . Cf. for example Fauconnier and Turner (1999), Coulson and Fauconnier (1999).
References Anscombre, J. C., & Ducrot, O. (1983). L’argumentation dans la langue. Liège/paris: Mardaga. Benveniste, E. (1966/1974). Problèmes de Linguistique générale. (2 Vol). Paris: Gallimard. Partial Transltion 1973. Problems in General Linguistics (Miami Linguistics, No 8). Miami: Miami University Press. Bergé, P., Pomeau, Y., & Vidal, C. (1984). Order within Chaos: Toward a Deterministic Approach to Turbulence. NY: Wiley. Brandt, P. A. (1992 [1986]). La charpente modale du sens: Pour une Sémiolinguistique Morphogénétique et Dynamique. Amsterdam: John Benjamins. Bühler, K. (1990 [1934]). Theory of Language; The Representational Function of language. Translated by D. F. Goodwin. Amsterdam: John Benjamins. Cadiot, P. (1991). De la grammaire à la cognition, la préposition POUR. Paris: Editions du CNRS. Cadiot, P. (1997). Les prépositions abstraites en français. Paris: Armand Colin. Cadiot, P. (1999a). Principe de conformité et génération analogique en sémantique nominale. Verbum, 31, 383–407. Cadiot, P. (1999b). Espaces et prépositions. Revue de Sémantique et pragmatique, 6, 43–70. Cadiot, P., & Nemo, F. (1997a). Propriétés extrinsèques en sémantique lexicale. Journal of French Language Studies, 7, 1–19. Cadiot, P., & Nemo, F. (1997b). Pour une sémiogenèse du nom. Langue Française, 113, 24– 34. Cadiot, P., & Nemo, F. (1997c). Analytique des doubles caractérisations. Sémiotiques, 13, 123–145. Cadiot, P., & Tracy, L. (1997). On n’a pas tous les jours sa tête sur les épaules. Sémiotiques, 13, 105–122. Cadiot, P., & Visetti, Y. M. (2001a). Pour une théorie des formes sémantiques: motifs, profils, thèmes. Paris: Presses Universitaires de France. Cadiot, P., & Visetti, Y. M. (2001b). Motifs, Profils, Thèmes: une approche globale de la polysémie. To appear in Cahiers de Lexicologie. Cassirer, E. (1946 [1924]). Language and Myth. Translated by S. Langer. Reprint by Dower Publications. Cassirer, E. (1953 [1923]). The Philosophy of Symbolic Forms. Vol 1: Language. Translated by R. Manheim. New Haven: Yale University Press. Cilliers, P. (1998). Complexity and Postmodernism. Understanding Complex Systems. London: Routledge. Coulson, S., & Fauconnier, G. (1999). Fake Guns and Stone Lions: Conceptual blendings and Privative Adjectives. In B. A. Fox, D. Jurafsky & L. Michaelis (Eds.), Cognition and Function in Language (pp. 143–158). Stanford, California: Publications of the CSLI.
Instability and the theory of semantic forms
Culioli, A. (1990/1999). Pour une linguistique de l’énonciation (3 Vols). Paris: Ophrys. Dahan-Dolmenico, A., Chabert, J.-L., & Chemla K. (Eds.). (1992). Chaos et déterminisme. Paris: Seuil. Ducrot, O. (1984). Le dire et le dit. Paris: Minuit. Fauconnier, G. (1997). Mappings in Thought and Language. Cambridge, MA: Cambridge University Press. Fauconnier, G., & Turner, M. (1999). Metonymy and conceptual integration. In K. U. Panther & G. Radden (Eds.),Metonymy in Language and Thought (pp. 77–90). Amsterdam: John Benjamins. Flavel, J., & Draguns, J. (1957). A Microgenetic approach to perception and thought. Psychological Bulletin, 54 (3), 197–216. Gibson, J. J. (1979). The Ecological Approach to Visual Perception. Boston: Houghton-Mifflin. Reprint Hillsdale, N.J.: Lawrence Erlbaum Associates, 1986. Gurwitsch, A. (1964). The Field of Consciousness. Pittsburgh: Duquesnes University Press. Gurwitsch, A. (1979). Human Encounters in the Social World. Pittsburgh: Duquesnes University Press. Husserl, E. (1982 [1913]). Ideas pertaining to a Pure Phenomenology and to a Phenomenological Philosophy. First Book: General Introduction to Pure Phenomenology. Translated by F. Kersten. The Hague: M. Nijhoff. Husserl, E. (1989 [1912–1928]). Ideas pertaining to a Pure Phenomenology and to a Phenomenological Philosophy. Second Book: Studies in the Phenomenology of Constitution. Translated by R. Rojcewicz and A. Schuwer. Dordrecht: Kluwer Academic Publishers. Husserl, E. (1973 [1939]). Experience and Judgement. Translated by J. S. Churchill and K. Ameriks. London: Routledge and Kegan Paul. Kanizsa, G. (1991). Vedere e pensare. Milan: Il Mulino. Kant, I. (1999 [1781/1787]). Critique of Pure Reason. P. Guyer & A.W. Wood (Eds.). Cambridge: Cambridge University Press. Koffka, K. (1935). Principles of Gestalt Psychology. New-York: Harcourt Brace. Köhler, W. (1920). Die physischen Gestalten in Ruhe und im stationären Zustand. Berlin: Braunschweig. Excerpts in A Source Book of Gestalt Psychology 1938. W. D. Ellis (Ed.). London: Kegan Paul. Köhler, W. (1964 [1929]). Gestalt psychology. New York: Liveright. Köhler, W. (1938). The Place of Value in a World of Facts. New-York: Liveright Pub. Köhler, W. (1971). Selected Papers. Ed. by Mary Henle. New-York: Liveright. Lakoff, G. (1987). Women, Fire and Dangerous Things. Chicago: University of Chicago Press. Langacker, R. (1987, 1991). Foundations of Cognitive Grammar (2 Vols). Stanford: Stanford University Press. Langacker, R. (1999). Grammar and Conceptualization. Berlin/New-York: Mouton de Gruyter. Lebas F. (1999). L’indexicalité du sens et l’opposition ‘en intension’/‘en extension’. PhD Thesis. Paris: Université Paris 8. Merleau-Ponty, M. (1962 [1945]). Phenomenology of Perception. Translated by C. Smith. London: Routledge and Kegan Paul. Merleau-Ponty, M. (1973 [1969]). Prose of the World. Translated by J. O’Neill. Northwestern University Press.
Yves-Marie Visetti and Pierre Cadiot
Michotte, A. (1946). La perception de la causalité. Louvain: Études de Psychologie, Vol. 8. Nemo, F., & Cadiot, P. (1997). Un problème insoluble? Revue de Sémantique et Pragmatique, 2, 9–40. Petitot, J. (1985). Morphogenèse du sens. Paris: Presses Universitaires de France. English version to appear 2001. Bern: Peter Lang. Petitot, J. (1992). Physique du sens. Paris: Editions du CNRS. Petitot, J. (1995). Morphodynamics and Attractor Syntax. Dynamical and Morphological Models for Constituency in Visual Perception and Cognitive Grammar. In R. Port & T. Van Gelder (Eds.), Mind as Motion: Explorations in the Dynamics of Cognition (pp. 227–281). Cambridge: MIT Press. Petitot, J., Varela, F., Pachoud, B., & Roy, J.-M. (Eds.). (1999). Naturalizing Phenomenology. Issues in Contemporary Phenomenology and Cognitive Science. Stanford: Stanford University Press. Piotrowski, D. (1997). Dynamiques et structures en langue. Paris: Editions du CNRS. Port, R., & Van Gelder, T. (Eds.). (1995). Mind as Motion: Explorations in the Dynamics of Cognition. Rastier, F. (1987). Sémantique interprétative. Paris: Presses Universitaires de France. Rastier, F. (1997 [1989]). Meaning and Textuality. Translated by P. Perrone and F. Collins. Toronto: University of Toronto Press. Rastier, F. (2000). On Signs and Texts: Cognitive Science and Interpretation. In P. Perron et al. (Eds.), Semiotics as a Bridge between the Humanities and the Sciences (pp. 409–450). New-York/Toronto: Legas Press. Rastier, F. (2001 [1994]). Semantics for Descriptions. To appear. Chicago: University of Chicago Press. Rosenthal, V., & Visetti, Y. M. (1999). Sens et temps de la Gestalt. Intellectica, 28, 147–227. Ruelle, D., & Takens, F. (1971). On the nature of turbulence. Communications in Mathematical Physics, 20, 167–192. Ruelle, D. (1993). Chance and Chaos. Princeton: Princeton University Press. Ruelle, D. (1996). Turbulence, Strange Attractors, and Chaos. World Scientific Series on Nonlinear Science, Series A, Vol. 16. World Scientific Publishing Co. Salanskis, J.-M. (1994). Continuity, Cognition, and Linguistics. In C. Fuchs & B. Victorri (Eds.), Continuity in Semantics (pp. 127–153). Amsterdam: John Benjamins. Salanskis, J.-M. (1998). Husserl. Paris: Les Belles Lettres. Sinha, C., & Jensen de Lopez, K. (2000). Language, Culture and the Embodiment of Spatial Cognition. Cognitive Linguistics, 11 (1/2), 17–41. Talmy, L. (2000). Towards a Cognitive Semantics (2 Vols). Boston: MIT Press. Thom, R. (1980 [1974]). Modèles mathématiques de la morphogenèse. Paris: Christian Bourgois. Touretzky, D. (1994). Continuity, Polysemy and Representation: understanding the verb cut. In C. Fuchs & B. Victorri (Eds.), Continuity in Semantics (pp. 231–240). Amsterdam: John Benjamins. Tracy, L. (1997). La clé du mystère: mettre le référent à sa place. Langue Française, 113, 66–78. Vandeloise, C. (1991 [1986]). Spatial prepositions: A Case Study for French. Translated by A. Bosch. Chicago: University of Chicago press.
Instability and the theory of semantic forms
Vandeloise, C. (1999). Quand dans quitte l’espace pour le temps. Revue de Sémantique et pragmatique, 6, 145–162. Victorri, B., & Fuchs, C. (1996). La polysémie – construction dynamique du sens. Paris: Hermès. Visetti, Y.-M. (1994). Les modèles connexionnistes entre perception et sémantique. Sémiotiques, 6–7, 15–48. Visetti, Y.-M. (2001). Constructivismes, émergences: une analyse sémantique et thématique. To appear in Des lois de la pensée au constructivisme, M.J. Durand (Ed.). Paris: Editions de la Maison des Sciences de l’Homme. Visetti, Y.-M., & Cadiot, P. (2000). Instabilité et théorie des formes en sémantique – pour une notion de motif linguistique. TLE (Théorie, Littérature, Enseignement), 18, 137–169. Vincennes/Paris: Presses Universitaires de Vincennes. Vygotsky, L. (1978 [1930]). Mind in Society: The Development of Higher Psychological Processes. Cambridge, MA: Harvard University Press. Vygotsky, L. (1986 [1934]). Thought and Language. Cambridge, MA: Harvard University Press. Weisbuch, G. (1991). Complex Systems Dynamics. Santa Fe Institute Studies on the Sciences of Complexity. Addison Wesley. Werner, H. (1956). Microgenesis and aphasia. Journal of Abnormal Social Psychology, 52, 236–248. Wertheimer, M. (1959 [1945]). Productive thinking. New York: Harper. Wildgen, W. (1982). Catastrophe Theoretic Semantics. An Elaboration and Application of René Thom’s Theory. Amsterdam: John Benjamins. Zeeman, E. C. (1977). Catastrophe Theory: Selected Papers 1972–1977. Mass: AddisonWeslay.
Schematics and motifs in the semantics of prepositions Pierre Cadiot Paris 8, Lattice
.
Introduction
My aim is to show that it is, at least at some deep level, pointless to approach the semantics of prepositions by their spatial meaning or use, which is commonly identified with its literal or prototypical meaning, as opposed to its secondary or extended metaphorical meaning(s) (see e.g. Lindstromberg 1997). Such an assumption is based on the hypothesis of an a priori isomorphism between the properties of spatial language, and especially prepositions, and those of externally perceived space. This view can be challenged in many ways. On a cognitive side, it should be underlined that, inherent to this current conceptualization of language (more or less construed as a reputedly faithful copy of the world), there are assumptions which overlook or forget both the teachings of phenomenology and of psychology (gestalt theory and developmental psychology) dealing with autocentric spaces bound to the exploration and discovery of the body. Speech should firstly be considered as a reinforcement of the egocentric activity, which accompanies the experience, rather than representing it. Presentation of spatial descriptions in terms of connecting to one another detached entities (landmark/trajectory) is bound to a displacement of scenes towards the point of view of the observer (for example the experience of a child who is to be “picked up” only has an element of verticality when perceived from a certain distance in a schematic framework). In other words it is the result of an inference, bound to a projection in a naturalized space with strong euclidization of its frame. However, what seems to influence language at its most profound level is the Lebenswelt, the experience of the body in its relationship to the environment and to others. This insistence
Pierre Cadiot
on the self-centered physical experience is exemplified by experiential, qualitative terms: resistance/yielding, holding tight, rupture, softness, roughness, bury, block, insert, get rid of, drown, touch, etc. For example the motif of containment is much richer than the relationship between the container and the contained. English up is conceptually an aspectual marker for completion, and not essentially an indication of verticality etc. Meaning is thus firstly anchored in anticipation, qualitative, often synesthesic feelings, and not in a directional grasping of objects. On the linguistic side, it is well known that in many languages localization or local meaning is not expressed by prepositions, but holistically, and more specifically by adverbs, classifiers, nouns, or typically by predicates (position and/or movement verbs). For example in Tzeltal, a Mayan language, there is only one so-called preposition which introduces instrumental, purpose, manner, time or place adverbials after the verb, namely the word ta. Thus, /ta + place/ can mean in/to/at/from a place (see Brown 1994). In a similar fashion, the Chinese preposition zai has both a very generic locative-functional and an existential meaning combined with adverbials or classifiers which specify space positions or localizations. Partly like in many other languages (e.g. Spanish and Serbo-Croatian), in Palikur, an Arawak language of Northern Brazil and F. Guyana, the so-called classifiers do not show the difference between on and in (support and containment). But more specifically in Palikur (and possibly in many other languages of the Amazonian area), one does not keep distinct object and place, what and where. Space is not conceived of as a uniform and homogeneous pre-existing receptacle for all experiences and events occurring in it. For example, the classifiers inuewu, ukuni and inugik designate what may happen in what we call ‘sky’. Each of these three markers can be considered a specific indication for accessing what is up in the natural environment from the viewpoint of a human being: inuewu is used to refer to natural, more or less cosmological, movements in the sky: sun, rain, stars, swarms of birds or insects; ukuni is more concerned with what is seen as being most probably closer and at a rather low altitude, such as dark clouds; inugik is used, it seems, to avoid the distinction between objects and events, such as disappearances, planes, birds flying away, or God! But such a non-distinction between what and where can be evidenced more marginally in every language. Think of the English ‘the sky is low’ or of Korean, where there is a different word for ‘top’ depending on whether or not it is considered the result of an ascension, whether you see it, or whether you attempt to attain it.
Schematics and motifs in the semantics of prepositions
. Cognition and typology As far as Indo-European languages are concerned, it is well known that we tend to fancy a systematic isomorphism between the system of spatial prepositions and a topological, systemic vision of some kind of neutralized space. This space is commonly conceived of as a mapping of geometric entities (plan, point, line, volume or positions) and idealized in some schematic imagery, for instance according to Lakoff (1987), who includes prepositions among words that have “image schematic meaning”. Space is also established with kinesthetic schemas or principles, such as orientation, motion, path, shape, landmark/trajectory distribution, points of view (deictic, intrinsic) and so on. It is based as well, contrary to Tzeltal or Palikur, on a clear distinction between objects and places or positions. A second aspect is the tendency to assimilate the core meanings of individual prepositions with some basic, more or less case notions like agent, purpose, manner, means, instrument or cause, and with the different Euclidian locations or localizations. For example, as was recalled by V. Brøndal (1950), Leibniz considered that prepositions and particles give an insight into the main categories of intuition or cognition. Another aspect of this oversimplification, which we have already mentioned, is bound to the pervading idea of a literal, more or less concrete meaning of linguistic expressions. Too closely identified with an idealized notion of reference, it underestimates the qualitative and aspectual dimensions of spatial indications themselves. This alleged first meaning is referentially construed as an extraction from the exteriorized representational layer, notwithstanding the qualitative/aspectual dimension of spatial expressions. Furthermore, more generally, it ignores the essential fact that words, before being referential labels are the mediators in the apprehension and qualitative constitution of objects. These objects themselves emerge from the semiotic process. Moreover, simplification underevaluates and relegates the immediate requalification of words in other domains or registers, be they temporal and/or qualitative, which can hardly be separated from space. This requalification, which is wrongly disguised when it is considered a matter of transposition, of transfer to separated domains, or as a metaphor (i.e. an extension of the meaning), is genetically founded in the continuity of the immediate donation of the world in the experience. This continuous requalification of (spatial) language is a logical consequence of a necessarily phenomenological access to the world. Another point I would like to make is the fact that English is often presented as a satellite-framed language (Talmy 2000, among many others). This means that bordering and vectorization of positions and movements are to be
Pierre Cadiot
expressed by satellite morphemes: affixes, adverbs, prepositional phrases. Such a typological feature – contrasting partly with French being more or less verbframed, because its verbs of movement integrate in their core meaning vectorizing and bordering cues (aller, entrer, partir, passer, sortir, venir) – implies that English verbs are strongly qualitative: they do not encode a movement, but a way of moving. As such, their aspect could be said to be massive, or imperfective, although, as is well known,1 not as clearly as for example in Russian or in German. So, the propensity to delineate spatial expressions in terms of landmark/trajectory might be reinforced by this typological feature, i.e. by some necessity for the processes or eventualities to be externally delimitated, a tendency which is probably more pregnant in English than in other languages. Anyway our point is to show that it should not be projectively confused with a universal way of construing space. There are several linguistic arguments, which could be borrowed from different languages, against this generalization. It would be confusing not to relate to the view that prepositions are relational words. First of all, as is well known, they are very akin to adverbs and particles. Taking into consideration that in Old French, adverbial (or intransitive, absolute) uses of prepositions were more frequent, it might even be appropriate to speak of prepositional and adverbial uses of these morphemes (good examples are après,2 avec, contre, derrière, devant, pour). Another aspect is the evident fact that the difference between close grammatical classes and open lexical classes is overestimated in contemporary linguistics. Partly for specific linguistic reasons, in French, the direct identification of movement with a schematic path (landmark/trajectory) is not so evident. This leads to the well-known gap between grammar and space cognition (landmark/trajectory) when comparing French and English verbs, or particle constructions with their Latin derived counter parts in English: (1) Paul part. Paul go PATH away ‘Paul is going away.’ (2) Paul sort. Paul go PATH out ‘Paul is going out.’ (3) Paul entre. Paul go PATH in ‘Paul is going in.’
Schematics and motifs in the semantics of prepositions
(4) Paul entre dans la maison. Paul go PATH to PLACE in the house. ‘He goes into the house’. (5) Paul traverse la rivière à la nage. Paul PATH across the river swims ‘He swims across the river.’ (6) Paul revient à cinq heures. Paul come PATH back at five ‘He will return at five!’ (7) Où est-il arrivé? Where has he arrived PLACE? ‘Where has he got to?’
Contrary to the English verb, which frames the scene by use of the satellite (prepositions or prepositional phrases), the French verb contains inherent indications for framing the scene. In the grammatical dimension of the problem (satellite-framed vs. verb-framed), we are more concerned with general characteristics than with specific spatial orientations: The English verbs are often more imperfective or even massive as in French (compare il sort/*he goes. He is going is possible but only if we already know where or when!).
. A new theoretical framework I now wish to summarize some of the views we have arrived at in recent work dealing with French prepositions (see Cadiot & Visetti, this volume; Lebas 1997). a. Classifying prepositions and their meanings in terms of an ontological categorization can be considered a theoretical ‘mistake’ in approximately the same sense that it is a mistake to confuse phonetics and phonemics (I will thus use the terms emic/etic, borrowed more generally from Pike 1971). b. Space is far from being in any sense a primitive: in a Kantian version, it is the very generic foundation of human experience. It can be vectorized through human perception and action, thus becoming discontinuous, it can be indexical and biased by the human body or by attraction, it may be deictical or object-internal. Moreover, it can be functionalized, subjectified, metaphorized, and so on.
Pierre Cadiot
c. The semantics of prepositions emphasizes the problem of the relation to space and to the physical world in a particularly striking form. Most often there is a trend towards relying on a very general psychological prototype, according to which language, at its most fundamental level, encodes tangible and/or physical structures. Therefore, in order to describe prepositions, as well as other categories of words, linguistics should favor spatial and/or concrete uses, and even take them as a primary basis for all the other ones. This idea leads, in cognitive semantics and also in grammaticalization theories, to a hierarchy of meanings, which starts from spatial or physical values, taken as literal meanings, up to temporal or abstract meanings, which are supposed to be derived from the previous ones by some kind of metaphorical transfer process. However, authors like Lakoff (1987), Langacker (1987), Talmy (2000) or Vandeloise (1986) underline that these primary values proceed from specifically linguistic schemes, which should not be confused with perceptive ‘external’ structures: indeed they are far more schematic, and at the same time genuinely linguistic, since for example they shape space by introducing ‘fictive’ contours or ‘fictive’ motions (Talmy 2000). But in spite of this important adding, the primacy (and/or the prototypical status) of the spatial and/or physical meanings is not questioned. d. As for the type of those linguistic schemes, and their relation to our external, everyday perception, we have seen that two main attitudes can be distinguished: –
–
Sometimes (Langacker 1987, particularly) the realm of dimensions prescribed by the historical Kantian framework is centered on purely abstract ‘configurational’ dimensions (abstract topology, abstract dynamics); those dimensions are supposed to be a permanent and obligatory basis of language in all semantic domains. On the contrary, dimensions like forces (and a fortiori dimensions like interiority, animacy, agency,) are considered as less grammatical, secondary dimensions, coming only from more or less prototypical uses (e.g. referring to the external perceived space); they can only add themselves to the configurational dimensions, and never neutralize them. Sometimes the realm of dimensions is not reduced (Talmy 2000; Vandeloise 1999), but this realm is considered primarily as part of our experience of the external physical world; spatial uses are more than typical, they are the primary ones; and all other uses are considered to be derived by a kind of metaphorical process.
Schematics and motifs in the semantics of prepositions
The approach we advocate is different.3 It aims at going beyond this kind of schematism, while keeping some of its ‘good’ properties. The exact abstraction level as well as the interior diversity of each scheme are a first key matter: abstract topological and/or cinematic characterizations are too poor, and schemes weighted from the beginning by spatial or physical values are too specific, and furthermore rely on a very peculiar conception of spatial and physical experience. Actually, more intentional or praxeologic dimensions, intuitively related to interiority, animacy, expressiveness, appropriation, control, dependence or anticipation are needed. By entering in the process of discourse, these dimensions – configurational or not – can be neatly put forward by speech, or alternately kept inside the dynamics of the construction of meaning as a more or less virtual aspect of what is thematized. In particular, configurational or morphological values are not a systematic basis: they may be pushed in the background, or even disappear, superseded by others, which are quite equally fundamental and grammatical. There is, therefore, no a priori reason for sorting out and eliminating some of these dimensions from the original linguistic schemes we are seeking. But we have to consider these dimensions at a very generic level, so as to assume that they are systematically put into play, and worked out by each use. For these reasons, and for others, which will appear later, we shall drop the designation of scheme, and adopt the word motif to express the kind of ‘seed of meaning’ we attribute to each preposition.
. Some sketchy considerations on French prepositions There are great differences in the systems of prepositions in French and English, especially concerning so-called colorless or only weakly depictable space prepositions. In describing some of them, I will show that – –
there is no clear-cut distinction between spatial and non-spatial uses or senses and that the emic meaning of it should be accessed in an immediate combination of schematic and intentional dimensions.
. The preposition en Let us first have a look at following phrases concerning the preposition en:
Pierre Cadiot
(8) pommier en fleurs ‘apple tree in bloom’ (9) chien en chaleur ‘dog in heat’ (10) femme en cheveux ‘hair-dressed woman’ (11) propos en l’air ‘words up in the air’
The sense of these expressions can be paraphrased by the following intuitive formulations or characterizations: “globally saturated physical image” (8), “invasion” (9), “emblematic access” (10), “taken over from the inside/outside” (11). They tend to show that space is only involved at an etic level, and in some sort of continuous variation. The characterizations can be resumed in a unique notion, or motif, of coalescence, with no emic limits or bounding and asymmetrically oriented toward the referent of the second NP. The image of the first NP is, so to say, absorbed in the image of the second (fleurs, chaleur, cheveux, air). But this motif is not only schematic or perceptual. It coalesces with a more instructional dimension: one has to associate the resulting image with its perspective, and with the intention through which or by which it was brought about. The scene is necessarily animated by the process that has generated it. Otherwise other prepositions like dans (with its bounding instruction) or even avec would be more appropriate. A more direct evidence for this rather intuitive interpretation of en can be drawn from other data where space is not involved: (12) a.
Max est en tort. Max is prep in (the) wrong ‘Max is wrong.’ b. *Max est en raison. Max is prep in (the) right ‘Max is right.’
(13) a.
Max est en beauté. Max is prep in (the) beauty ‘Max is handsome.’ b. *Max est en laideur. Max is prep in (the) ugliness ‘Max is ugly.’
Schematics and motifs in the semantics of prepositions
(14) a.
Max est en vie. Max is prep in (the) life ‘Max is alive.’ b. *Max est en mort. Max is prep in (the) death ‘Max is dead.’
There seems to be a rather regular paradigm of such cases, where only the resulting states exemplified by the examples (a) can be correctly introduced by en, whereas those in (b) are not compatible with en. All the acceptable expressions can be associated with an intentional, subjective object-oriented path or a purpose that brought them about. For example Max est en vie (14a) is pragmatically possible only in as much as one has reasons to believe that he could be dead (after some accident, presumably); Max est en tort (12a) is acceptable because he has done or said something which happened to be wrong or inappropriate; to be en beauté ‘in beauty’ (13a) means more than to be beau ‘beautiful’: that he tried or at least, wished to be handsome. . The preposition par Even more evidently, it is impossible to differentiate spatial and not spatial interpretation in the case of par. (15) être emporté par le courant ‘to get carried away by the current’ (16) passer par le jardin ‘to go through the garden’ (17) prendre par la gauche ‘to take a left hand turn’ (18) regarder par le trou de la serrure ‘to look through the key-hole’ (19) attraper par la cravate ‘to grab by the tie’ (20) tuer par balle ‘to kill by bullets’ (21) convaincre par son comportement ‘to convince by one’s behaviour’
Pierre Cadiot
(22) impressionner par son intelligence ‘to impress by/with one’s intelligence’ (23) passer par des moments difficiles ‘to come through hard times’ (24) renoncer par lassitude ‘to give up from/because of lassitude’
In English, by seems to work better with concrete referents and tends to internalize them in the scope of the schema, while with more abstract complements, through or even because of are better, and with seems at least to initiate a motion of externalization, or “parallelization”. As is well known, par is typically used to express agentivity in passive constructions or in any type of constructions where a process is described from the point of view of its activation. So it expresses an inner activation principle. Being “inner” corresponds to the schematic dimension, being “agentive” to the intentional one. But both are intimately correlated and coactive in every instance, even when it corresponds to no specific local thematic or referential intuition. . The case of sur (“on”) I shall briefly distinguish the following configurations. The first deals with the predicate to be on, creating a region where the localization of the subject is enabled by the predicate. The second consists in a frame for the events taking place in this region. .. The region sur What I shall call the region sur is constructed at the level of the predication être sur ‘to be on’, i.e. it is a construction of a site based on the connection between the preposition and the assigned NP, on the localization of the noun subject and the contact enabled by the predicate. A typical example is: (25) Le livre est sur la table. ‘The book is on the table.’
In other cases, this very same region is established by the context of the sentence, which allows for an adjustment or requalification of lexical and syntactic expectations, as on can gather here: (26) Max s’est effondré dans le fauteuil. ‘Max collapsed in the (arm) chair.’
Schematics and motifs in the semantics of prepositions
(27) Max a posé timidement une fesse sur le fauteuil. ‘Max timidly sat on the (arm) chair.
The motif contact is permitted and enabled by the predicate. As opposed to a table or a sidewalk, an armchair is not a priori an acceptable object for the predicate être sur ‘to be on’. The requalification is facilitated by the specific reference. .. The zone sur There is a zone established as a frame for what happens in the region sur. In the zone, the possible fluctuations between contact and localization increase: (28) Les enfants jouent sur le trottoir. ‘The children are playing on the sidewalk.’
Still, there is a simple correlation between a topological notion and a univocal localization in the thematic space. However, this correlation is nullified, or made more complex, by many other uses with spatial implications. It may happen that the prepositional phrase does not localize the subject of the sentence. (29) Pierre joue avec sa poupée sur la table. ‘Pierre plays with his doll on the table.’ (30) Pierre a vu un chat sur le balcon. ‘Pierre saw a cat on the balcony.’
Nothing indicates that the referent of Pierre is localized by the region sur (on the table, on the balcony). In fact, the contrary is noticeably more likely. The region sur no longer has determined spatial limits at the etic level (in the sense offered by Pike, that is, from the exterior, or in a linguistically marked fashion). In the following examples, quite particular to French, we can hypothesize that the motif is further developed: (31) Pierre travaille sur Paris. ‘Pierre works *on/in Paris.’ (32) Pierre est représentant sur la région Nord. ‘Pierre is a representative *on/for/in the north.’
Here, the preposition sur is used in the construction of “functional spaces” (zones specified only in the domain of the predication) and not of physical spaces – where it is more evident – but the topological instruction of contact is entirely preserved. From the preceding examples we might believe the
Pierre Cadiot
motif of contact to be simply topological. In fact, it can easily be requalified with new interpretative effects for which the spatial inferences are decreasingly concrete, proving themselves to be inseparable from temporal and qualitative modulations (Dendale & De Mulder 1997, whence the following examples): – Support (weight or imminence) (33) Une menace planait sur la ville. ‘A threat hovered on?/over the town.’
– Foundation (assessment) (34) Juger les gens sur l’apparence. ‘To judge people on?/by their appearance.’ (35) Il fut condamné sur de faux témoignages. ‘He was convicted on false testimony.’
– Covering (36) La couverture est sur la table. ‘The tablecloth is on the table.’
– Objective (goal) (37) Marche sur Rome ‘The March on Rome’ (38) Fixer un oeil sur quelque chose ‘*to pose/*to fix /*to leave/to fast one’s eyes on something’
–
Visibility, immediate access (as opposed to inclusion which would signify dependence, interposition of a border or a screen). (39) Il y a un trou sur ta manche. ‘There is a hole *on/in your sleeve.’
Furthermore, the semantic cues support and/or foundation can be extended easily to uses that are definitively non-spatial, as in: (40) Impôt sur le revenu ‘Tax on income’ (41) Agir sur ordre ‘Act on orders ’ (42) Pierre a travaillé sur cette question depuis longtemps. ‘Pierre has been working on this question for a long time.’
Schematics and motifs in the semantics of prepositions
(43) Sur cette question, Pierre n’a rien à dire. ‘On this issue, Pierre has had nothing to say.’
Here the motif of contact is invested in a thematic zoning, which can be specified only in the domain opened by the predication or the introducing nominal argument. Let us also remember the temporal uses differentially specifiable, which emerge from the motif of contact.4 (44) Sur ce, il disparut à jamais *‘On/after this he disappeared for ever’ (45) Pierre est sur le départ ‘Pierre is about to leave’ (46) Il y a eu des gelées sur le matin ‘There was a frost this morning/on the morn (archaic)’ (47) Il faut agir sur le champ ‘One must act at once’
Finally, in compter sur ses amis ‘to count on ones friends’ and in miser sur le bon cheval ‘to bet on the right horse’, without entirely abandoning a certain value of ‘to lean on’, which may be considered a modulation of the original motif, the preposition sur is requalified as a rectional marker by grammaticalization.5 To resume our discussion on sur, all these examples not only invalidate purely spatial and physical explanations, they also weaken explanations based on abstract topological schemas. These often seem artificial and demand further qualifications, which call into doubt their validity. Above all, this type of schematics does not provide operable explanations, and as a result doesn’t explain why only certain values and not others are called upon (by interaction with the surrounding lexical material, as we say). What is missing here is the possibility of recognizing the affinity and interrelation of these different values, which we would like to stabilize by way of lexico-grammatical motifs. In this way, the topological instruction, even when purely configurational and despatialized (i.e. conceived independently of the perceived space) seems to flag behind a richer, more open delimitation of two segments or phases, as they are construed during any type of contact. Compared to the image of surface often invoked (as a geometrical notion), or to that of height (Weinrich 1989), this motif of contact would have the same statue as that of coalescence for the preposition en, or of means in the case of the preposition par. Beyond its dynamic value, the motif contact also offers a static characteristic, which provides a border or a stabilized variation (localization, support). However, it is fundamentally an aspectual motif, intentional in aim and in practice. At
Pierre Cadiot
once a motif of exploitation and of valorization of this contact by a type of immediate interaction (leaning, rebounding, perlaboration), it gives the values of objective imminence, achievement, effect, transition, cause and effect. Its configurational expression, once fully deployed, includes the following parameters: (a) an axial orientation of momentum, (b) a transversal orientation for the contact zone, and (c) an exteriority which is delimiting phase (a) and (b). Notice that even if the contact zone is in fact the topological frontier of the access zone, it is still not appropriated as its border, but remains exterior. To conclude, localization can certainly be explained in Euclidean terms like surface, height, width, etc. But the diversity of possible instances of localization (the rich variety of contributing elements) calls for dimensions that are more dynamic (force, figure/background) compared to the more configurational ones. In the phrase “cup on the table”, we might emphasize the importance of the features bearing-weight. As in French, the English expressions “bandage on the arm”, “drawing on the wall”, “handle on the door”, “apple on the branch”, show us that the preposition on constitutes the sight as a background, which guarantees a detachability for the figure, regardless of any more objective relations with the object/surface. . The case of sous One can uncover five experiential types, which bear a family resemblance in the Wittgensteinian sense: a. Low position: sous la table ‘under the table’, sous les nuages ‘under the clouds’ b. Covering/protection: sous la couette ‘under the covers’, (objet enfoui) sous la neige ‘under the snow’, sous une même rubrique ‘under/in the same rubric’ c. Exposition: sous la pluie ‘under*/in the rain’, (marcher) sous la neige ‘(walk) in the snow’, sous les regards ‘under the eyes of x’, sous les bombes ‘under fire’, sous la menace ‘under the gun’ d. Inaccessibility: sous terre ‘underground’, sous le sceau du secret ‘under heavy guard’ e. Depending from the external: sous surveillance ‘under surveillance’, sous influence ‘under the influence’, sous la contrainte ‘under pressure’, sous garantie ‘under warranty’, sous arrestation ‘under arrest’. These uses involve a co-adjustement of the values selected from the NPs assigned by the preposition, and in some cases by the introductory element (see the example of the snow). Together they evoke family resemblances of covering,
Schematics and motifs in the semantics of prepositions
protection, inaccessability, exposure to, and dependence upon, in varying degrees of explicitness. Among the notions evoked above, some of them seem more oriented towards a topological schematic pole (a surface constructed by the PP, which establishes an interior space based on that boundary). The others are closer to a more instructional pole (Cadiot 1999), which consists of the more dynamic values, aspectualized by a quasi praxeological perspective (i.e. there are no exit dynamics and the opening is blocked), and indexed on the ambivalence of the situation (covering vs. exposed). Articulating these two poles of the boundary, which remains separate from the interior space, is just the configurational expression of this blocking and ambivalence. As in the case of sur, this complex motif is diversely profiled and stabilized: by valorization, specification, or on the contrary by inhibition, retreat and aspectualization of the different values it unifies. . The case of contre Let us note the following four experiential types: a. Proximity with contact: s’appuyer contre le mur ‘leaning against a wall’ b. Opposition (conflict): être contre le mur de Berlin ‘to be against the Berlin wall’, contre toute attente ‘against all expectations’ c. Exchange: échanger sa vieille voiture contre un scooter ‘to trade one’s old car for a scooter’ d. Proportion/comparison: vingt mauvais films contre un bon ‘twenty bad films *against/for one good one’. For contre ‘against’, we propose a motif instituting the affinity of opposition and reconciliation (force/counter-force, posing/opposing). This motif is up to a certain point sustainable within a Kantian schematic framework, capable of reflecting relational categories like force in a plurality of not necessarily physical spaces. But we insist again on the necessity to modulate this motif/schema and to specify it in accordance with plausible profiles. As a result, values such as counter-force or dynamics coming together can disappear almost completely from the profile. Even when they are virtualized, as in “a sofa against the wall”, they remain as a motivation for the internal perspective or aspect of the dynamics.
Pierre Cadiot
. By way of conclusion We hope that these short considerations, which evidently call for considerable developments and should be in a systematic mood confronted with other languages, will be understood as a way of challenging the routine frozen expression “spatial preposition”.
Notes . Think of a contrast like to enter (evidently borrowed from French) vs. to get in. . Après is a nice example of the more qualitative origin of spatio-temporal indications. The preposition was originally an adverb inherited from Latin ad-pressum/ad-premere (English: to press). Many modern uses of the French preposition, either spatial, partly spatial or not spatial at all are bound to this etymology: grimper après un mur ‘to climb behind/over a wall’; la clé est après la porte ‘the key is on the door’; il me court après ‘he is pursuing me’; il est toujours après moi ‘he is always hovering over me’; et après, qu’est ce que tu veux que ça me fasse? ‘and finally, after all, what should I care?’. . It draws on several recent works on prepositions: cf. Cadiot (1997, 1999). See also the work of Berthonneau (1998, 1999), Leeman (1997a, 1997b) and Vandeloise (1986, 1999). . In Palikur, amin is present in all cases in which there is a notion of open-dynamic space, of expected departure, of going out (which makes it nearly a cognate of sur). As in French sur, amin also indicates something like “to be interested in”, “to prepare for”, and is also a basis for deontic “must”. . Grammaticalization is a way to refer to the very fact that, especially in rectional contexts, the preposition loses the greatest part of its semantic impact because the corresponding paradigm is closed or limited. Nevertheless, rather than being thought of as a categorial process, it should be relativized as an intensionalization and not as the vanishing of any meaning. The need for another explanation is evident when one seeks to find some reason for the specific choice of so-called grammaticalized prepositions: why, for example, consister en (consist in), insister sur (insist on), compter avec (count with)? The core meaning is bleached by the strong limitation of the paradigm, but is maintained.
References Berthonneau, A. M. (1998). Espace et Temps: quelle place pour la métaphore? Verbum, 20 (4), 352–382. Berthonneau, A. M. (1999). A propos de dedans et de ses relations avec dans. Revue de Sémantique et de Pragmatique, 6, 13–41. Bowerman, M. (1996). The origins of children’s spatial semantic categories: cognitive versus linguistic determinants. In Gumperz, J.-J., Levinson, S. C. (Eds.), (pp. 145–176).
Schematics and motifs in the semantics of prepositions
Brøndal, V. (1950). Théorie des prépositions. Introduction à une sémantique relationnelle. Copenhague: Munksgaard. Brown, P. (1994). The INs and ONs of Tzeltal locative expressions: the semantics of static descriptions of location. Linguistics, 32, 743–790. Buridant, C. (2000). Grammaire nouvelle de l’ancien français. Paris: Sedes. Cadiot, P. (1997). Les prépositions abstraites en français. Paris: Colin. Cadiot, P. (1999). Espaces et Prépositions. Revue de Sémantique et de Pragmatique, 6, 43–70. Dendale, P., & De Mulder, W. (1997). Les traits et les emplois de la préposition spatiale sur. Faits de Langue, 9, 211–220. Denis, M. (Ed.). (1997). Langage et cognition spatiale. Paris: Masson. Franckel, J.-J., & Lebaud, D. (1991). Diversité des valeurs et invariance du fonctionnement de en préposition et pré-verbe. Langue Française, 91, 56–79. Groussier, M. L. (1996). Prépositions et primarité du spatial: de l’expression de relations dans l’espace à l’expression de relations non spatiales. Faits de Langue, 9, 221–234. Guimier, C. (1978). EN et DANS en français moderne: Etude sémantique et syntaxique. Revue des langues romanes LXXXIII, Montpellier III, 277–306. Gumperz, J.-J., & Levinson, S. C. (Eds.). (1996). Rethinking Linguistic Relativity. Cambridge: Cambridge University Press. Klein, W., & Nüse, R. (1997). La complexité du simple: l’expression de la spatialité dans le langage humain. In Denis, M. (Ed.), (pp. 1–24). Kwon-Pak, S. N. (1997). Les prépositions spatiales: sur quelques emplois de la préposition par. Thèse de Doctorat en Sciences du Langage. Strasbourg: Université Strasbourg 2. Lakoff, G. (1987). Women, Fire, and Dangerous Things, What Categories reveal about the Mind. Chicago: Chicago University Press. Lang, J. (1991). Die französischen Präpositionen, Funktion und Bedeutung. Heidelberg: C. Winter Universitätsverlag. Lebas, F. (1997). L’ indexicalité du sens et l’opposition intension/extension. Thèse de Doctorat en Sciences du Langage. Paris- Saint-Denis: Université Paris 8. Leeman, D. (1997a). Définir une préposition: hypothèses et perplexités. Revue de Sémantique et Pragmatique, 2, 183–200. Leeman, D. (1997b). Sur la préposition EN. Faits de Langue, 9, 125–144. Lindstromberg, S. (1997). English Prepositions explained. Amsterdam: Benjamins. Talmy, L. (2000). Toward a Cognitive Semantics. Cambridge, Mass.: MIT Press. Vandeloise, C. (1986). L’Espace en français. Paris: Le Seuil. Vandeloise, C. (1999). Quand DANS quitte l’espace pour le temps. Revue de Sémantique et Pragmatique, 6, 145–162. Weinrich, H. (1989). Grammaire textuelle du français. Paris: Didier/Hatier.
The theoretical status of prepositions The case of the “prospective use” of in Franck Lebas Université Blaise Pascal – Clermont-Ferrand II, UFR Lettres, Langues & Sciences Humaines
.
Introduction
In studies of theoretical aspects of prepositions, it appears that the subset named ‘spatial prepositions’ is favoured by the linguistic community. Some of them always have their place in studies of prepositions, whether in virtue of being called ‘spatial’ or simply because they are very common. The preposition in is certainly one of the most emblematic examples of this recurrent focalisation. Here I will leave aside the central problems that are generally discussed concerning the preposition in and turn to the two extremes of the theoretical scale. First, I will consider the theoretical status of prepositions in general, and use in only as a non-specific instance of the category. Second, I will propose a rapid description for the meaning of the French preposition dans – my arguments will be explained with examples of in – and narrow to a very specific use named here “prospective use”, and exemplified by He will arrive in three hours, which is generally forgotten or, when studied, considered peripherical and non-problematic. The reason why this use is generally left aside is that it is in fact problematic for most approaches of prepositions. The reason of its being central in this study is that it supports the approach presented here. In turn, this approach supports a general theory of language, promoted by a small (but growing!) set of French researchers, whose principal aspects I will present. Meanwhile, I will argue that the well-known debate of the ‘spatial’ status of some prepositions (among which in) diverges from the main issue of the status of the prepositional category itself. Of course, this debate will be given an answer here: no, there is no such thing as a ‘spatial preposition’. But this is
Franck Lebas
only a consequence of the characteristics of the prepositional category on one hand, and of more general hypotheses on language on the other hand.
. Guillaume and the prepositions If one is interested in building a theory of prepositions, the name of Gustave Guillaume (1964, 1974) is one of the very first references to arise, because he showed that prepositions are characterised by mental operations that are fundamental for language activity. Here are the main principles of his theory that I will relate to in the discussion. First, linguistic interpretation is considered through its genetic process and the linguistic signs bring material and constraints to this process. This theory can be thought of as opposing structuralist theories: linguistic signs do not bring conceptual or structural elements that would then be put together during interpretation, but they influence directly the construction of semantic (ideogenesis) and syntactic forms (morphogenesis). Second, there are two classes of words. The predicative class is characterised by its incidence to a base. Interpretation being a construction, every phase can be thought of as a base to which some words, typically nouns, adjectives and verbs, have an incidence. This is a technical and dynamic way of characterising the semantic words, as opposed to the words of the non-predicative class. Prepositions are examples of such non-predicative words, but they have an even more specific property: prepositions play a role towards the interval between two incidences, they are not themselves incidental. In less technical terms, prepositions are said not to contribute semantically to a thought in construction, but to act in the interval between two successive thoughts. Third, Guillaume introduces a mechanism that can (roughly) “stop”, at some specific points, the genetic process brought by a word during interpretation. This is called a seizure (Fr. “saisie”) and can explain some aspects of polysemy acting at an abstract level. This mechanism is an impressive way of explaining, for instance, the different values of the French determiners un ‘a’ and le ‘the’. The generic vs. specific meanings of each of these determiners are correlated to “early” or “late” seizures in the genetic process. Theoretically, this mechanism applies to the entire linguistic system. In particular, some disciples of Guillaume (Cervoni 1991; Moignet 1981; Roccheti 1982) have tried to relate some aspects of prepositions to different seizures in their genetic process, but the level of abstraction is such that it becomes very difficult to correlate these hypotheses to linguistic facts.
The theoretical status of prepositions
The attempts to examine prepositions inside Guillaume’s framework encounter serious difficulties, even if we leave aside the problem of the seizures as an explanation of some aspects of their polysemy. Maybe the most serious difficulty is to explain the possibility for prepositions to be more or less semantic, or colourful as many linguists have said since Spang-Hanssen’s book in 1963 (see Vandeloise 1993, for a review). In fact, it is difficult to build a theory of prepositions that explains the very abstract role of some of them (to or of for instance) and the very semantically specific role of others. The supporters of Guillaume’s theory have to face the paradox of a prepositional class defined as non-predicative – hence with no semantic contribution – and examples of prepositions, such as in or during, which have to be considered semantic. To date, they propose to solve this problem by postulating a two-stage process, one being characteristic of the prepositional status and the other being a semantic contribution, a real incidence (Cervoni 1991: 168–169). I have argued (Lebas 1999) that this answer is contradictory and I will relate only to the three main principles mentioned above.
. The Extrinsic Properties Theory As described above, the theoretical challenge is to integrate the semantic nature of morphemes that are precisely pointed out as having no semantic contribution to interpretation. I will argue here that this puzzle can be solved inside a new theoretical framework, based on a concept of “extrinsic property” that Pierre Cadiot and François Nemo introduced in their work on lexical semantics (Cadiot & Nemo 1997a, b, c; Cadiot & Visetti 2000; Lebas 1999). This solution has also a lot to do with the studies carried out by Pierre Cadiot on several French prepositions (Cadiot 1997). In fact we could consider that several hypotheses derived from local aspects of language now come to shape a coherent theory that helps solve the problems these hypotheses encountered locally. I hope this will become clear after I have explained briefly what these hypotheses are. . The concept of “extrinsic property” The expression “extrinsic property” was introduced by Pierre Cadiot and François Nemo to reveal the paradoxical nature of the properties that language is supposed to indicate about the world. If we take the canonical example of the noun chair (I will use nouns to explain the theory, then come back to prepo-
Franck Lebas
sitions), we can see that nearly all theories of language have carried on the intuition that its core meaning derives from properties of the objects that we usually call chairs: Chairs have specific shapes, specific parts, specific dimensions, etc. These properties are intrinsic because they are supposed to depend on no other thing than the object itself. Using intrinsic properties is not the only way for characterising an object. All chairs also have in common that they are perfectly suited to a particular way of sitting. People do not sit on a chair, in a sofa or on a stool in the same way or for the same reasons, because these different kinds of furniture are adapted to different sitting gestures, to different uses, habits and functions. These characteristics are extrinsic to the objects, since they apply only when people interact with these objects. The central claim of the Extrinsic Properties Theory is that the core meaning of nouns is composed of such properties that are derived from human uses, usages, habits, behaviours, objectives, etc. These properties are extrinsic to the objects they point at, but it is very important to note that they are nevertheless thought of as being properties of these objects. For instance, we claim that people tend to name chairs the sitting devices that have a particular back because these devices directly seem to be particularly suited to the specific way of sitting indicated by the word chair: extrinsic properties are more fundamental than intrinsic ones, at least as far as language is concerned (see Cadiot & Lebas, to appear, for a discussion). There have been numerous linguistic studies, for example Cadiot and Nemo (1997a–c) or Lebas (1997a, b, 1999), showing that the main semantic properties of nouns can be explained by this theory, especially their polysemy, their referential behaviour and their metaphorical uses. These successes help to accept the necessary generalisation of the “extrinsic properties” hypothesis (exemplified by nouns of artefacts such as chair) to every kind of noun (natural kinds, abstract nouns, etc.), and also the possible generalisation to other linguistic categories, such as prepositions. . Prepositions don’t have a “semantic aspect” Once we have accepted that semantic contents of nouns are made of – and only of – extrinsic properties, the fact that these properties can be of very different types could lead to the idea that the meaning of prepositions is made of one special and characteristic type of these properties. For instance, we could suppose that in indicates a “relation between objects” described in terms of extrinsic properties. These properties would surely constitute the very essence of
The theoretical status of prepositions
the ‘inside’ concept, describing what it is in a “subjective” approach for an object to be in another one: the physical control of the containee by the container (this is detailed below), the process of inserting an object into another one, the structuring capability brought by the container, all these conceptual abilities and “readings of the world” being integrated and structured as one complex form out of the practice of objects in early childhood. No purely geometrical aspect of this concept is at stake here, as some linguists have already argued before (cf. Vandeloise 1986, 1993, 1999). To sum up, the inclusion relation we have in mind when conceiving of the word in is derived from the usages that constitute the concept of inclusion, not from the memory of prototypical relations (whatever that might be!). It should be clear now that the semantics of relations can be thought of as fully compatible with the semantics of objects, and it takes little effort to conceive of a very general meaning of the terms object and relation in this claim. But our concern here is to know to what extent these hypotheses can point out not only the semantics of a particular preposition but also the characteristic of the prepositional category. For example, whatever the theory, we have to show the difference between in, inserted and to insert. This is the point where we have to come back to Guillaume. I propose that what makes the semantic material of prepositions particular is not to be found in the way this material is arranged nor in some syntactic information that would be added to these words, but very simply in the linguistic use that this material allows for. It is the possibility of interpreting their semantic material in a special way that motivates the intuition of a special category for prepositions. This special way of interpreting is the non-incidental way, an interpretative attitude that prevents the semantic material from contributing to the process of building up mental objects (I will neglect here the difference between “non-incidental” and “non-predicative” (see Cervoni 1991: 68–73 for details). Instead, the semantic influence of prepositions is directed to the process of interpretation itself. This proposal is new for three reasons. First, what is usually considered as the semantic aspect of colourful prepositions is viewed here as the whole meaning but directed towards grammatical processes. Colourful prepositions, as opposed to more abstract or colourless prepositions, don’t have a semantic aspect added to their meaning; they simply have different uses that give rise to an intuitive semantic particularity. In fact, diachrony tells us to read this difference the opposite way: colourless prepositions have the privilege of some very special (so called abstract) uses that other colourful do not have (or do not seem to have at a first glance).
Franck Lebas
Second, the characteristic of prepositions is not to indicate a special type of relation between objects, nor to indicate a relation in a special way. Prepositions simply do not indicate a relation between objects! This is their particularity, even if the outcome of this interpretation is constituted, among other things, of a relation to which the preposition has contributed. To phrase this in a perhaps less puzzling way, a preposition happens to contribute to an element of interpretation that we tend to call a relation, but that relation is not the sense of – or represented by – the preposition in this interpretation. The role of prepositions is far too indirect and grammatical to say that their sense in context is a relation between two elements of the constructed sense, even in examples as simple as the flower in the pot. Third, even if prepositions can contribute to a relational element of the constructed sense, they have many uses that have nothing to do with constituted relations. The prospective use of in, which will be dealt with later, is a very clear example, but we can also consider verbal uses, like to consist in, and many different examples like the person in charge, In his hurry he forgot his keys, where the concept of relation could be employed but would be far from grasping the precise meaning. Not to mention more abstract prepositions. . Departing intuition from semantics For methodological reasons, the theoretical propositions made inside Extrinsic Properties Theory were made for, and from the analysis of, non canonical, peripherical uses of linguistic terms. As for prepositions, what guided us was the hypothesis that only one meaning should apply to all uses, especially the uses that are traditionally left aside by linguists and those that are invariably analysed through meaning transfers, figurative uses, and so on. This being so, what becomes the main problem is to come back to the canonical examples and to remember that basic, unsophisticated sentences like the flower is in the pot have to be accounted for! For those static uses, we have to decide how to handle direct intuition. The theory presented here proposes an analysis of intuitions that other theories present as explanations, namely that prepositions indicate primarily relations, or at least that they can sometimes indicate relations. In fact, when I say that prepositions do not indicate relations, I am implying consequently that other words do. Certainly the adjectival class can be supposed to have this property. But what I mean in saying so is that adjectives, when interpreted, keep a kind of connection with the relation or quality they have semantically contributed to. I don’t mean to say that the semantic contribution of adjectives is to bring a re-
The theoretical status of prepositions
lation or a quality “as is”. Adjectives, too, are postulated to bring only extrinsic properties that can constitute what is called later a semantic relation or quality, in the same way prepositions do. The difference lies only in the domain of incidence: adjectives have a direct incidence to mental objects that can be referred to, whereas prepositions have an incidence to the interpretation of sequences of words. I have shown (Lebas 1999) that the semantics of French prepositions à and de can be described as two particular ways of interpreting a sequence of two words. What is characteristic of à is to impose a conceptual distance between the terms it puts in relation, and X-de-Y states that Y is a context for interpreting X. This kind of theoretical proposition is easy to accept in the case of very abstract and grammatical prepositions. As a general, grammatical structuring of semantic elements is a satisfactory way of describing in technical terms the concept of linguistic abstractness. But when it comes to more colourful prepositions, it appears strange and meaningless to say that what is to constitute a semantic relation (relation of inclusion in the case of in) is to be viewed as a grammatical structuring. Indeed, the semantics of these colourful prepositions is not transformed in grammatical structuring; it is only interpreted as such. When applied to contexts where prepositions finally contribute to static relations between real objects, the semantic difference could be considered minimal, but in fact the conceptual implications of this syntactic construal are essential. Among other things, it is well known that prepositions cannot be modified by adjectives, the relation to which they contribute can not be referred to, prepositions can appear alone only in certain circumstances, etc. These particularities are not only syntactic; they correspond to a special semantic status for the relations involved. In a word, these particularities constitute the contingence of the semantic relation, as opposed to other linguistic ways of construing that mark relations as incorporated to the objects they are applied to. Conversely, adjectives like tall and late, for instance, are relational insofar as someone has to evaluate something with respect to a relevant scale, but the tallness or lateness will always be “attached” to the object they qualify, they are not contingent to it. Before turning to the preposition in in greater detail, we should keep in mind that what is proposed here for characterising the prepositional category is in fact the description of a specific way of interpreting linguistic signs. It can be considered as a description of words that we label as prepositions only to the extent that the semantics of these words are particularly suited to this interpreting way. As a result, we can expect all so-called prepositions to behave in a non-prepositional way in some contexts, and conversely we can expect other categories to behave sometimes in a prepositional way.
Franck Lebas
Subcategories are treated in the same manner. For example, some prepositions are called “spatial” only because of their being particularly suited to contexts where they contribute to spatial relations.
. The meaning of French dans The theoretical scheme that has been introduced will now be applied to the case of the French preposition dans ‘in’. As indicated above, I propose to define the core meaning of dans with the kinetic concept of insertion. The term “concept” is there only to refer to the synthetic aspect of this semantic content. I deliberately do not use the term “representation” here, since I believe linguistic signs to be directly associated to conceptual abilities that form a coherent structure. In the case of in, this structure consists in every aspect of the process of inserting as a meaningful process. When saying “insertion”, I don’t refer to the technical ways of inserting objects, but to the functions and implications of this process. Claude Vandeloise (1993, 1999) has already made attempts in this direction and argued in favour of a control of the containee by the container as one of the main dimensions of in. I fully agree with this proposal, but at the same time I wonder whether this statement is explicit enough. What does it mean for something to “control” something else? As I said, I would not use the term “control” as a representation, from which we could make some inferences. Instead, I would say that the concept of insertion and the concept of container are dependent, and that they generate together a particular way of structuring things which is highly connected to the practice of the physical world. Let us take a simple example. When a child plays at putting a small toy in a box, he observes that generally when he holds the box and moves or walks, then the toy stays inside the box. He understands gradually why it is so, and makes the synthesis of the “walls” of the box as a unique structure that prevents the toy from falling in every horizontal direction. Things are completely different when using a plate instead of a box, because you have to behave in a very different way to be able to maintain a toy on a plate. This constitutes one of the main differences between in and on. This is an example of the practice of the physical world, which is, of course, supported by – not reduced to – visual practice. We can suppose that these practices are articulated, though supported by very different cognitive regions, and the result is the ability to use vision and gestures with great efficiency. Linguistics is indeed in quite a different domain of practice, but I make the assumption that it can nevertheless be articulated
The theoretical status of prepositions
to this global synthesis (see Lebas 1999; Cadiot & Lebas to appear, for the psycholinguistic aspects). And in the same way as vision develops new capabilities towards the insertion process, the linguistic practice of this concept opens to “inserting” things that cannot be touched nor seen, and of course opens to the communicational aspects. In this framework, we should not expect the linguistic analysis to extract a few descriptive elements that would grasp all that is characteristic of such a semantic content as “insertion process”. We cannot sum up in a few terms all that it is to insert something, even if we decide to stick to one modality. We can refer to this synthesis because of our common practice of it, but we cannot describe it (yet). Instead, we can try to make assumptions about what synthesis is correlated to what linguistic unit. For instance, I would not say that in is associated to a holding process, a spatial stabilizing process, a static inclusion state, etc. All these syntheses I could not describe fully either, but they nonetheless have a differential value inside a phenomenological theory. And what we can observe through the linguistic aspects not only tells us which of these possibilities to select, but also gives hints at what exactly constitutes the semantic content that we try to determine. I will try here to show the main lines that lead to the description of the meaning of dans. The kinetic aspect will be demonstrated first, and then the prospective use will reveal the choice for a definition based on the insertion concept. . The kinetic aspect of ‘dans’ Some studies of dans have revealed a possible connexion of some uses with the concept of causation (Leeman 1999). Here are two clear examples: (1) Dans la bataille, Paul a été blessé. ‘In the fight, Paul was hurt.’ (2) Dans sa grande sagesse, Pierre décida de ne pas combattre. ‘In his great wisdom, Peter decided not to fight.’
If we tried to phrase these examples in terms of static inclusion, we would obviously lose what is essential to the global meaning. It would not only be imprecise to say that Paul was hurt during the fight; it would be a real misinterpretation. And saying that Peter took his decision during a moment of particular wisdom, or inside a state of mind of wisdom would be patently absurd. In fact,
Franck Lebas
the causative aspect of these uses are central to the global meaning, and the theoretical hypothesis I posited concerning the prepositional category implies that the meaning of dans helps build these particular causative relations. At this point, we have to be very careful not to overspecify this core meaning. If we say (3) Pendant la bataille, Paul a été blessé. ‘During the fight, Paul was hurt.’
there is a strong tendency to build the very same causative relation. Hence, the meaning of dans in (1) only brings another (strong) motive for a causative reading, not the causation relation itself. As for example (2), we can observe that there has to be a modifier of the noun. Without it, the sentence is very awkward, if acceptable at all (cf. 4): (4) Dans sa sagesse . . . ‘In his wisdom . . . ’
We could suspect the reason to be that wisdom cannot refer to a property of Peter but only to the local event of him being wise in such situations. Whether his being wise at that moment is due to a permanent feature of his personality is not directly implied in this sentence. I propose that this observation as well as the previous ones is to be linked with the kinetic core meaning of dans in these uses. It is the process of engaging into something that can be considered as a source for the intuition of a causation, even if we prudently retain only the kinetic aspect of this contribution for the moment. In fact, it should become clear, by comparing the two different causative relations in these examples, that causation is the only abstract word we have for pointing out what is in fact reduced to this conceptual process of engaging into. For example, it would be very artificial to say that the preposition dans ‘in’ in example (1) is equivalent to /because+during/, and I can think of no equivalent at all for example (2)! That is to say that we do not have to end up with a wording of a particular use, and that an indirect reference to the meaning through its genetic process is not only theoretically justified but also more precise. To sum up, I propose that the kind of uses exemplified above reveals the kinetic nature of the core meaning of dans, and that this kinetic nature sometimes develops into a static relation in other uses (the opposite hypothesis is impossible for obvious reasons). The English preposition in, though being different, would probably share this first description, but we would have to examine other uses to verify this. If we suppose very roughly that English in is
The theoretical status of prepositions
equivalent to French dans as well as French en, the kinetic natures of the two French prepositions (I shall not go into details for en) correspond together to the kinetic nature of in. What is specific to dans, and to some uses of in only, is the phrase “engage into” that I used above, which is the first sketch of the following description. . The prospective use of ‘dans’ What I called the “prospective use” of dans ‘in’ is exemplified by (5): (5) Le dîner sera prêt dans vingt minutes. ‘Diner will be ready in twenty minutes (from now)’
In English, we usually have to add from now to specify one of two possible interpretations (the other one being in twenty minutes’ time, which corresponds to en vingt minutes in French). As mentioned in the introduction, this use is most often left aside because it manifests a clear difference with all the other uses of dans. Indeed, this use doesn’t fit into any classical description simply because the referent of what follows dans corresponds to the (immediate) outside of what is said to be in it. A dinner that is said to be ready dans vingt minutes is programmed to be ready immediately after a period of twenty minutes from now, definitely not at a time “inside” this period. I cannot even think of an explanation of this observation that would simply “continue” some hypothesis made from the canonical uses, and I never heard of any convincing one. This problem is even one of the few that resist so patently the “domain specific” approaches. Indeed, it is not enough to propose, as does for instance Claude Vandeloise (1999) with Il a plu dans la journée ‘It rained in the day’, that the uses of dans (or in) applying to time are specific because of the particular structure of time. Whatever can be thought of this argument, the localist approach – stating that spatial structures and operations shape the foundations of semantics – cannot explain how the concept of inclusion (associated to in) can give birth to a use that is best expressed in terms of exclusion. I will draw the conclusion from these observations that something new has to be brought into the theoretical framework so as to account for this particularity of dans (and the corresponding uses of in). It happens that this case is not isolated, and that elsewhere in semantics there is a need for the very same theoretical tool. For instance, I have shown elsewhere (Lebas 1997a, b, 1999) that a specific use of French ils ‘they’ – the collective use we see in They have raised taxes again for instance – demonstrates the same kind of break, and that only a particular attitude towards semantic interpretation, which I have named
Franck Lebas
intensionalisation (Fr. mise en intension), could explain this observation. Intensionalisation is at stake each time a new use appears through the attitude of leaving some elements of meaning in an intensional state. By definition, these semantic elements are in an intensional state (vs. extensional state) because they happen not to have actually a role in the context, a semantic incidence, but are nevertheless conceived of as if they had. In fact, it quickly appeared that intensionalisation was a very general, virtually universal phenomenon, but I cannot go into detail here. In the present case, intensionalisation simply refers to the fact that what normally specifies the destination of what is said to be dans ‘in’ is supposed to be “settled” and that, as a consequence, specifies – we could say measures – the process itself instead of its result. Saying that the “normal” object is thought of as semantically “settled” is identical to saying that the same core meaning is still used within a different syntactic arrangement that does not allow for a specification of what is normally specified. In this use, we do not say “where” something is said to be dans ‘in’ but “how far” we have to go to conceive of this position. In fact, we can consider that the object of dans is indeed determined, but only in an indirect way: determining “how” incidentally determines “what” in that case. The way I propose to account for this use is coherent with the fact that it is very particular and associated with specific constraints. In French, this is revealed by the absence of any determination: (6) a.
Le dîner sera prêt dans vingt minutes. ‘Dinner will be ready in twenty minutes from now.’ b. Le dîner sera prêt dans les vingt minutes. dinner will be ready in twenty minutes ‘Dinner will be ready within (the next) twenty minutes’
(6a) represents the prospective use whereas (6b) the temporal inclusion use. These are very clear-cut interpretations, which means that the presence/absence of determiners is essential to the two different meanings that we observe. My suggestion is that the object of dans in the prospective use, because it acts as a measure (of a kinetic process), has to be clearly non-referential. The absence of the determiner is a strong, local indication for interpreting the syntactic object as not referring to any “real” (mental) object (a particular and identified period of time in that case), and in fact it is a direct indication of the intensionalisation of the usual syntactic object. Of course, the main conclusion of this analysis is that this very specific combination of semantic attitudes and constraints is directly and exclusively
The theoretical status of prepositions
compatible with the kinetic insertion definition of dans. This is the only use of dans that clearly and compulsorily leads to the insertion part of the definition, but since we can conceive of every other use as developed from this core meaning, there is no need for other evidences. . The prospective use of ‘dans’ is proximal Moreover, it is possible to correlate this hypothesis to the syntactic constraints exemplified by (7a–c): (7) a.
Je le ferai dans cinq minutes. ‘I’ll do it in five minutes from now.’ b. *Je l’ai fait dans cinq minutes. I did it in five minutes *‘I did it in five minutes from then.’ c. Demain, je le ferai dans une heure. tomorrow, I’ll do it in one hour *‘Tomorrow, I’ll do it in one hour from then.’
These constraints are indeed very difficult to account for if one thinks of this use in continuation of the other uses of dans. On the contrary, they appear quite natural if intensionalisation is postulated to be characteristic of this prospective use. The reason is simply that the attitude of specifying the interpretative process itself is attached to the moment of utterance. This attitude cannot be conceived of with a distance – the prospective use of dans is proximal, not distal – because it applies to the interpretation process and not to some stable semantic stage of it. Leaving the domain of prepositions, we can note that this is not the case, for instance, of vingt minutes plus tard ‘twenty minutes later’: (8) a.
Je le ferai vingt minutes plus tard. I’ll do it twenty minutes later. b. Je l’ai fait vingt minutes plus tard. I did it twenty minutes later. c. Demain, je le ferai vingt minutes plus tard. Tomorrow, I’ll do it twenty minutes later.
. Systemic implications At this stage it is necessary to insist on several theoretical points. First, determination in French being very different from that in English, it is not surpris-
Franck Lebas
ing that the temporal inclusion use corresponds to a different balance in English, involving the word within (I’ll do it within twenty minutes) instead of determination in French (Je le ferai dans les vingt minutes). Second, the prospective use corresponds to a specific semantic attitude indeed, but does not have to correspond to a syntactic structure that is different from that of the other uses. In French the absence of determination after dans is so coherent with intensionalisation in that case that it is a strong and obligatory syntactic indication. Alternatively, the global system of English happens to be such that the same syntactic structure (in with no determiner) can be interpreted in two ways, one being the prospective use, the other one the more “classical” inclusion use. Third, the fact that French dans and en correspond (roughly) to only one English preposition does not preclude these analyses. I believe the nature of the prepositional category to be very similar in French and in English. I strongly believe that intensionalisation is a universal linguistic phenomenon, but still English and French are different in their structure and we would have to give a different description to the core meaning of in. Indeed, a description of French en has been proposed (Cadiot 1999) which is very different from the description of dans presented here, and both are certainly different from that of in, which (to me!) has yet to be given.
. Conclusion The prospective use of dans (which we also see in corresponding uses of English in) has linguistic properties that are so peculiar that they are very difficult to account for in most theoretical frameworks. I tried to demonstrate that the Extrinsic Properties Theory is perfectly suited for building a real explanation – not only a description – of these constraints, based on a description of dans in terms of kinetic insertion. The hypotheses that we have to put forward are so specific to the phenomenological theory that I have used here that we can consider this study to provide a strong support for it. In addition, the prepositional category itself has been characterized, inside of the Extrinsic Properties Theory and with the help of Guillaume’s heritage. It appears that what makes prepositions specific is their incidence to interpretation itself, not directly to the semantic objects built through interpretation. This theoretical status for prepositions is coherent with the very abstract role of the majority of them, and with the variations that we observe on the colour scale. The challenge of integrating colourful and colourless preposi-
The theoretical status of prepositions
tions in one theory is given a response here, as I tried to demonstrate with the example of in.
References Cadiot, P. (1997). Les prépositions abstraites en français. Paris: Armand Colin. Cadiot, P. (1999). Espaces et prépositions. Revue de Sémantique et Pragmatique, 6, 43–70. Cadiot, P., & Lebas, F. (Eds.) (To appear). La constitution extrinsèque du référent. Langages. Cadiot, P., & Nemo, F. (1997a). Propriétés extrinsèques en sémantique lexicale. Journal of French Language Studies, 7, 1–19. Cadiot, P., & Nemo, F. (1997b). Pour une sémiogénèse du nom. Langue Française, 113, 24– 34. Cadiot, P., & Nemo, F. (1997c). Analytique des doubles caractérisations. Sémiotiques, 13, 123–144. Cadiot, P., & Visetti, Y.-M. (2000). Instabilité et théorie des formes en sémantique. TLE 18 [Frontières instables]: 137–169. Cervoni, J. (1991). La préposition, étude sémantique et pragmatique. Paris – Louvain-laNeuve: Duculot. Guillaume, G. (1964). Langage et science du langage. Paris: Nizet; Québec: Presses de l’Université de Laval. Guillaume, G. (1974). Leçons de linguistique 1949–1950, Série A. Paris: Klincksieck; Québec: Presses de l’Université de Laval. Lebas, F. (1997a). Conséquences théoriques des frontières de la polysémie. Application au pronom il. Langue Française, 113, 35–48. Lebas, F. (1997b). L’indexicalité du sens et l’opposition “en extension”/“en intension.” Sémiotiques, 13, 163–176. Lebas, F. (1999). L’indexicalité du sens et l’opposition en intension/en extension. Thèse de doctorat de l’Université Paris 8, Saint-Denis. Leeman, D. (1999). Dans un juron, il sauta sur ses pistolets. Aspects de la polysémie de dans. Revue de Sémantique et Pragmatique, 6, 71–88. Moignet, G. (1981). Systématique de la langue française. Paris: Klincksieck. Roccheti, A. (1982). Sens et forme en linguistique italienne: Etudes de psycho-systématique dans la perspective romane. Thèse d’Etat de l’Universite de Paris III, 2 vol. Spang-Hanssen, E. (1963). Les Prépositions incolores en français moderne. Copenhague: Gads Forlag. Vandeloise, C. (1986). L’espace en français. Paris: Seuil. Vandeloise, C. (1993). Les analyses de la préposition dans: faits linguistiques et effets méthodologiques. Lexique, 11, 15–40. Vandeloise, C. (1999). Quand dans quitte l’espace pour le temps. Revue de Sémantique et Pragmatique, 6, 145–162. Vandeloise, C. (Ed.). (1993). La couleur des prépositions. Langages, 110.
Temporal semantics of prepositions in context David S. Brée and Ian E. Pratt-Hartmann Computer Science Department, University of Manchester
We examine how the temporal semantics of prepositions and verbs interact to determine the interval in which the main proposition in a sentence holds. The central question is whether these two sources of temporal semantics are simply additive or rather interactive. Only preposition phrases specifying a temporal interval, such as 1992, are considered, not those specifying a duration, such as two years. Also, the present tense, modals and generics are not considered. It is shown that the interaction is usually additive except that: – quantification is generally determined by the aspect, when present, and by the type of verb otherwise, except that throughout and until require universal and after and before require existential quantification; – temporal since must have the perfective aspect, as was known; – by can only take the perfective aspect with event verbs; – for most prepositions the progressive aspect requires the time of reference to be first updated to within the interval determined by the preposition.
We have developed a temporal logic suitable for capturing the semantics of temporal prepositional phrases (Pratt & Brée 1995). In this logic it is also possible to record the temporal semantics provided by the tense and aspect of the verb in the sentence. The purpose of this paper is to ascertain whether these two sources of temporal semantics can simply be summed, or whether there is an interaction between them. There is clearly some interaction: since is only given a temporal interpretation if the sentence in which it occurs has a perfect aspect, otherwise it is inferential (Brée 1985). However, it is quite possible that there is no other interaction; the semantics of tense and aspect on the one hand and that of temporal prepositions on the other may be simply additive. From the point of view
David S. Brée and Ian E. Pratt-Hartmann
of cognitive simplicity, this would be what one would expect. Hence the issue addressed in this paper.
.
Preliminaries
We will consider only declarative sentences having two parts: a matrix and an optional temporal prepositional phrase. The matrix has a tensed verb and an optional aspect, given by an auxiliary verb. (Modals are not considered.) The temporal prepositional phrase has a preposition and a temporal noun phrase. When processed, the sentence yields a timeless proposition and a temporal interval over which this quantified timeless proposition holds. The context and the matrix verb (and auxiliary, if present) contribute to the fixing of this interval globally. The temporal prepositional phrase, if present, refines this interval. We assume each sentence is spoken (or written) at a particular time, the time of utterance, or tou. It describes a situation that holds at some time, often called the time of event or toe (even though the situation described might not be an event at all but some state). The toe is usually different from the tou; they are only the same when a speaker is describing ongoing situations, such as in sports reporting on the radio. Each sentence is interpreted within a temporal context which is set in two ways: –
–
by the time of reference (tor), in the Reichenbachian sense. Any narrative or description assumes that the writer and reader have agreed on a tor that may be different from both the tou and the toe. The tor may be earlier, at the same time as or later than the tou, as reflected by the tense of the matrix verb (Section 2.2). The toe is set relative to the tor rather than to the tou (but recall that the tor may be the same as the tou). In English the aspect of the matrix verb partially determines this relationship (Section 2.3).1 by the situation, for example: (1) Gunner had staked out the joint.
must clearly mean that the staking out occurred within a reasonable period before the tor, not several years earlier, let alone a millennium earlier. We indicate this context by stating that the proposition holds within an interval which we will simple call Context.
Temporal semantics of prepositions in context
A temporal prepositional phrase may introduce either a time duration, i.e. a length of time not tied to the time line, e.g. two months, or a temporal interval, i.e. a length of time located on the time line, e.g. 1992. We will only consider temporal intervals. Some matrix clauses are generic, e.g. they may indicate a habit. Generics can generally be distinguished from non-generics as they accept frequency adverbials, such as frequently or always, whereas non-generics do not, cf. *Ridley (2) Gunner killed at dawn, frequently. rats Generics are not covered in this paper. We now introduce a simple framework for representing the temporal information conveyed by verbal tense and aspect before turning to the main topic of this paper: the semantics of temporal prepositions (Section 3) and how this interacts with tense and aspect, for both past and future of the matrix verb (Section 4). The present tense has several uses, in particular for generics, and is not covered in this paper.
. The matrix The tense and aspect of the verb in a sentence together set temporal constraints on the possible interval in which the main event in the sentence occurs. We will now summarise the ways they do this, as presented in Quirk et al. (1985). Our purpose is to provide a background against which the semantics of temporal prepositions can be investigated, not to add to the knowledge of the semantics of tense and aspect. But first we need to consider the type of verb, because this too plays a role in determining the temporal semantics. . Verb types The acceptability of some temporal prepositions depends on the type of verb used in the matrix, cf. in (3) Gunner killed Ridley ? five minutes. for *in (4) Gunner liked Ridley five years. for
David S. Brée and Ian E. Pratt-Hartmann
Therefore we need to distinguish different types of verb. The most common classification is that due to Vendler (1957) in which a distinction is made between states, achievements, accomplishments and activities. Note that many verbs can be classified into more than one of these semantic types, depending on the context in which they are used, for example transitively or intransitively, cf. the tower. [transitive–achievement] (5) Gunner climbed for pleasure. [intransitive–state] We will not consider these differences due to context here, but rather we assume that the classification into semantic type is made by some other device which has access not only to the verb but also to all the matrix (other than the verb’s tense and aspect) and its context. When we refer to the type of a verb, this is shorthand for the type as determined by the entire matrix plus context – see (Verkuyl 1999) for how this may be done. Quirk et al. (1985) go so far as to classify verbs into a hierarchy resulting in 11 semantic types. We wish only to make distinctions between different types of verb if the distinction leads to different temporal semantics. We will see that we need to distinguish verbs that indicate an event, e.g. kill, from those that indicate a non-event. We will also need to divide non-event verbs into those that are states, e.g. like, from those that are activities, e.g. follow. The reasons for our choices follow. Events are not true over all sub-intervals of the interval in which they hold, whereas non-events are, cf. by (6) [Event] Gunner killed Ridley noon. *until *by (7) [Non-event] Gunner liked Ridley noon. until In (6) the killing took place at some interval, most likely a matter of minutes or even hours. However, of none of the sub-intervals within this interval could it be said that the timeless proposition Gunner kill Ridley holds. This contrasts starkly with the sub-intervals in (7); in all of them it is the case that Gunner like Ridley holds. A test for event versus non-event verbs is the acceptability or otherwise of the combination with in or for governing a duration, cf. (3) with (4). In choosing examples to illustrate the different semantics for events versus nonevents, we must be careful to choose verbs, such as kill and like, which virtually require one or the other reading, rather than verbs such as climb, which
Temporal semantics of prepositions in context
can readily be used to indicate either an event (the achievement of having climbed) or a non-event (the act of climbing). We will use the above examples repeatedly. We need to divide non-event verbs into those that are states and those that are activities, e.g.: State
Gunner liked Ridley for two months. ⇒ ?Gunner was liking Ridley for two months.
Activity Gunner followed Ridley for two months. ⇒ Gunner was following Ridley for two months.
The reason is that, while a progressive aspect can be used with an activity verb with only a small change in meaning (the activity is not terminated), when combined with a state verb, the result is only marginally acceptable; the meaning is changed to an activity, sometimes introducing a sense of deceit on the part of the agent. There is no temporal truth functional difference here. Rather the difference seems to be due to the need for the progressive aspect to be applied to some action, even if the action is continuous throughout the interval; this is a property that state verbs do not possess. These three types of verb, namely event, state and activity, will be sufficient distinctions for our purpose. For the sake of comparison, Vendler’s achievements and accomplishments are our events. The difference between them, for Vendler, was that accomplishments are ‘extended in time’, whereas achievements are not, cf. Accomplishment Gunner wrote his report. Achievement Gunner finished his report.
We have found no need to make this distinction as our temporal intervals can be of any duration. Vendler’s scheme has been extended by several authors. Steedman (1999) provides a good overview and a more extensive classification. It should be noted that Steedman classifies Vendler’s activities as events whereas our equivalent activities we classify as non-events, since they have the sub-interval property. Also for comparison, our events are Quirk et al.’s (1985) punctual dynamic verbs, our state verbs are their stative verbs and our activity verbs are their durative dynamic verbs. These three top level categories include all of their 11 semantic types as special cases.
David S. Brée and Ian E. Pratt-Hartmann
. Tense In English there are strictly only two verbal tenses: the past and the non-past or present. The past tense generally sets the tor to be earlier than the tou. The present tense has several uses. For instance, with an event it can indicate an unfolding action, with a non-event it is more likely to indicate a property that holds at the tor and possibly for the whole of the context interval, cf. Event Gunner kills Ridley. Non-event Gunner likes Ridley.
While the present can be used when the toe is later than the tou, especially with a temporal adverbial, it is the modal will that usually indicates the future, e.g. Present Gunner leaves tomorrow. Modal Gunner will leave soon.
To avoid having to talk about a past tense and a future modal, we will refer to this use of the modal will as the “future tense”, even though in English it is not strictly a tense (Enç 1996) and even though will does not always indicate that the toe is earlier than the tou. All three tenses (past, present and future) contribute to setting the time of the main event or non-event. If there is no temporal adverbial in the sentence, then this is done by reference to the tor at the point where the sentence is being interpreted, as set up by the previous context, i.e. anaphorically. If there is an explicit temporal adverbial in the sentence, e.g. yesterday, at dawn, then the tense will pick out a unique referent for the adverbial, i.e. cataphorically, cf. (8) Ridley arrived at dawn. [Cataphoric] (9) Gunner killed him. [i.e. just after dawn; Anaphoric]
In this section we will be concerned with anaphoric determination of the time of the event using tense and aspect. In the following section we look at the use of prepositions to set the toe cataphorically. .. Past and future tenses Even though the future with will does not always behave in the same way as the past tense (Enç 1996), e.g. in embedded clauses, we will find that the past and future tenses have the same interaction with temporal prepositions.
Temporal semantics of prepositions in context
Events are presumed to occur just a little later than the tor as set by the previous sentence. As a result of interpreting the sentence, the tor is updated to the toe. We will represent a simple past or future tensed sentence describing an event, such as killed (10) Gunner Ridley. will kill by Past |= ←•(tou, •(Context, •(tor + ε, G kill R))). Future |= •→(tou, •(Context, •(tor + ε, G kill R))). where, following Pratt and Brée (1995): –
–
–
– – –
•(t, α) indicates that the quantified timeless proposition α holds of the time interval t, which is a unique time interval within its temporal context, e.g. Context; ←•(t, α) indicates that the quantified timeless proposition α holds of the unique time interval ending at the start of t (with the beginning of the interval set by the context); •→(t, α) indicates that the quantified timeless proposition α holds of the unique time interval beginning at the end of t (with the end of the interval set by the context); φ is the existential quantification of a timeless proposition φ over some sub-interval of the temporal context; ε is a small duration; . . . is a timeless unquantified proposition.
For example, the formula for the past tense should be read as: there exists a unique tou, sometime before which there is a unique interval Context, within which there is a unique interval tor + ε (which is a little later than the current tor), during which there is an interval in which Gunner kill Ridley is true. For non-events we have universal rather than existential quantification. A simple tensed sentence describing an activity, such as: followed (11) Gunner Ridley. will follow will be represented similarly by: Past ←• |= (tou, •(Context, •(tor + ε, G follow R))). Future •→
David S. Brée and Ian E. Pratt-Hartmann
where φ is the universal quantification of φ over all sub-intervals of its temporal context, here tor + ε. Similarly, a simple tensed sentence describing a state, rather than an event, such as: liked (12) Gunner Ridley. will like will be represented by: Past ←• |= (tou, •(Context, •(tor, G like R))). Future •→ Note that we assume that ε = 0 for state verbs.2 There are some well-known exceptions to these general rules which we will be ignoring. For example: –
Will may also be used to indicate a deduction: (13) Ridley will be at home by now.
–
A modal in the past does not necessarily indicate a past tor. It may be a conditional: (14) Gunner should lie low for a week now.
–
The past can indicate a counterfactual: (15) Control wishes that Gunner had been more careful.
–
When the past tense is used in a clause that is part of a matrix which uses the future tense, then the past tense can indicate an event in the present, or even in the future: (16) “Control will be delighted when he hears that I killed you.”
.. Summary Since the present tense has several uses (Quirk et al. 1985) and these differ somewhat from the way the past and future tenses are used, its interaction with temporal prepositional phrases will need to be considered separately at a later time. The semantics of the past and future tense in sentences without any aspect, temporal preposition or temporal adverbial will be represented by:
Temporal semantics of prepositions in context
Past ←• |= (tou, •(Context, •(tor + ε, Qφ))). Future •→
where: – – –
Q is for event verbs and for non-event verbs, φ is the timeless proposition conveyed by the sentence, and ε = 0 for state verbs.
Notation: In general, tenses interact with aspects in the same way so the relation between the tor and the tou, given by the tense, will only be specified if there is ambiguity. . Aspect The function of aspect is to set the time of an event with respect to the tor. Without any aspect, an event or an activity, as it is introduced, is given a time just after the tor (Section 2.2) and the tor is updated (Section 1). To put the next event or activity earlier than the tor is the function of the perfective aspect; to put it at the tor is one function of the progressive aspect. Surprisingly, the two aspects may be used in combination. .. The perfective aspect It is sometimes argued (e.g. by Comrie 1985), that the perfect in English indicates a tense relative to the absolute tense set by the main verb, rather than an aspect. We prefer to follow the more traditionally grammarians, e.g. Quirk et al. (1985), who use the term perfective aspect (rather than the more common perfect aspect) to emphasize the contrast with the progressive aspect. We represent the perfective aspect, which sets the time of an event to be earlier than the tor, by embedding the existential quantification of the proposition representing the timeless meaning of the clause in ←•(tor, . . .), e.g. had Event Gunner killed Ridley:←•(tor, G kill R), will have had State Gunner liked Ridley; ←•(tor, G like R). will have had (17) Activity Gunner followed Ridley; ←•(tor, G follow R). will have Note that:
David S. Brée and Ian E. Pratt-Hartmann
– – –
the future perfective assumes a tor later than the tou. The addition of, e.g. by tomorrow, makes this explicit and the sentences in isolation clearer; with non-event verbs the perfective alters the quantifier from universal to existential; with an activity verb, the perfective implies that the activity is over. We have not represented this implicature.
.. The progressive aspect The progressive aspect is used to indicate that something has a duration or is not complete. As Quirk et al. (1985: 209) put it: “. . . there is some point of orientation from which the temporary event or non-event described by the verb can be seen to stretch into the future and into the past.” The point of orientation is within the tor. As the extent of stretch into the future and the past is not specified, we will note only that the event or non-event holds over the whole of the tor interval. With an event verb the progressive is used to ensure that what would otherwise be an event is interpreted as an activity holding in an interval which includes at least the tor, e.g. Event James opened the door and looked in. Gunner was killing Ridley; •(tor, prog(G kill R).
where prog(φ) is a function that turns an event into an activity, i.e. enables the proposition to hold for all subintervals of the temporal context (here the tor). Recall that the tor is an interval rather than a point in time. Here the killing is presumably by some slow activity such as suffocation. As a non-event already has duration, the progressive aspect, when applied to a non-event verb, might seem to be unnecessary. But this is not so. With an activity verb, the progressive signals that the activity takes place throughout the tor, rather than at some small time later than the tor, e.g. was Activity Gunner following Ridley; • (tor, prog(G follow R)), will be where prog applied to a proposition with an activity verb indicates that the activity is not complete. Note that this difference reflects the so-called ‘imperfective paradox’: from the progressive of an activity the activity itself holds, but from the progressive of an event, the event itself does not: Activity Gunner was following Ridley. ⇒ Gunner followed Ridley. Event Gunner was killing Ridley. ¬ ⇒ Gunner killed Ridley.
Temporal semantics of prepositions in context
With a state verb there is, indeed, no difference between the progressive aspect and the representation of the non-aspectual, as ε = 0 already (see Section 2.2). However a progressive can be used with state verbs; not surprisingly it then usually signals some peculiarity, e.g. that there is a pretence or effort on the part of the agent: ? knowing was know State Gunner Ridley; •(tor, prog(G R)). will be ?liking like where prog applied to a proposition with a state verb indicates that the verb is not to be taken literally. Note that the progressive can turn some state verbs into activities, those to which growing to could be added. These verbs have both a state meaning and the meaning of the activity undergone to reach the state, e.g. resembled (18) Gunner was resembling Bond. was growing to resemble These normally state verbs, when combined with a progressive, need to be treated as activities. What is common to all three types of verb is that the progressive sets the toe to be at the tor, rather than just after the tor, i.e. ε = 0. Certain special uses of the progressive (Quirk et al. 1985) will not be covered: –
with going to, to indicate a future (in the past): (19) Gunner was going to kill Ridley tomorrow.
–
with attitude verbs, to indicate tentativeness: (20) I was hoping to borrow some money. (21) I was wondering if you could help me.
.. The perfect progressive aspect The simple combination of the temporal features of the perfective (setting the toe < tor) and progressive (setting toe = tor) aspects would be contradictory. However, these two aspects are used together without contradiction, so their temporal semantics cannot be the sum of their two parts. What happens is that the progressive still has the effect of giving something a duration and being incomplete, while the perfective places this non-event earlier than the tor. Universal quantification at the tor is retained from the simple progressive.
David S. Brée and Ian E. Pratt-Hartmann
With an event, a progressive turns the event into an activity, while the perfective puts the resulting activity earlier than the tor, e.g. Event Gunner had been killing Ridley; ←•(tor, prog(G kill R)).
We have chosen universal rather than existential quantification here, even though an event with no aspect is given existential quantification, on the grounds that the progressive turns an event into an activity (for a while can be added). In this example the killing would have had to be by some slow activity! As with the simple progressive, the perfect progressive with a state verb again indicates an activity and possibly deceit on the part of the agent, e.g. had ? State Gunner been liking Ridley; ←•(tor, prog(G like R)). will have With an activity verb, the perfect progressive is quite natural. It is very similar to the simple perfective in that it postulates that an activity occurred earlier than the tor, cf. (17), so it will be represented in the same way, e.g. had (22) Activity Gunner been following R; will have ←•(tor, prog(G follow R)). Also the activity is not complete, which can be most easily seen in activities that can lead to accomplishments, cf. cleaned (23) Gunner had cleaned his rifle. had been cleaning With a simple perfective the activity has lead to the desired accomplishment, with the perfect progressive there is an implicature that it has not yet done so. With an activity that does not lead to an accomplishment, the simple perfective implies that the activity is over by the tor, whereas the perfect progressive implies that it still holds, cf. followed (24) Gunner had Ridley. been following Capturing this subtlety would take us beyond our present purpose of providing a simple representation of tense and aspect in order to investigate their interaction with the semantics of temporal prepositions.
Temporal semantics of prepositions in context
Table 1. Temporal semantics for the aspects Aspect
–
Progressive
– Perfective
•(tor + ε, µ(σ))† ←•(tor, µ(σ))‡
•(tor, prog(µ(σ))) ←•(tor, prog(µ(σ)))
µ(σ) •(tor, α) ←•(tor, α) prog(φ)
†
‡
The timeless unquantified proposition given by the sentence σ. Proposition α holds in time interval tor. Proposition α holds from the beginning of the context up to the unique interval which is the tor. A function that turns an event φ into an activity, that with a state verb gives an implicature of deceit on the part of the agent and with an activity verb an implicature that there is no accomplishment. If the verb is a non-event, then instead of ; if the verb is a state, then ε = 0. If the verb is an activity that can lead to an accomplishment, then there is an implicature that there is an accomplishment.
.. Summary This completes the preliminaries concerning aspect. The results are summarised in Table 1. An aside: as the perfective aspect with an activity verb indicates that the activity is over at the tor, it might be thought that the perfective aspect turns an activity into an event, complementing the function of the progressive aspect in turning an event into an activity. While this is true of some activity verbs, in particular those that lead to some accomplishment, e.g. in (25) Gunner had cleaned his rifle five minutes. for it is not true of activity verbs in general, e.g. cry, let alone of state verbs. So the nice symmetry between the functions of the progressive and perfective that is suggested does not hold. . Summary In order to incorporate the semantics of tense and aspect into a natural language understanding program, these need to be expressed as a procedure. The procedure in Table 2 will generate the semantics of aspect given in Table 1 together with tense semantics given in Section 2.2. Note that the double quotes around strings indicate that the string is to be written out. There are three phases:
David S. Brée and Ian E. Pratt-Hartmann
Table 2. Procedure for determining the temporal semantics of a sentence without a temporal preposition µ(Σ) = “|=”; case tense(σ): % Set the tor relative to the tou past “←•(tou,”; future “•→(tou,”; esac; “•(Context,”; if perfective?(σ) then “←•” else “•”; % Set the toe relative to the tor “(tor”; if ¬(perfective?(σ) ∨ progressive?(σ) ∨ state?(σ)) then “ + ε”; “,”; if (¬event?(σ) ∨ progressive?(σ)) ∧ ¬perfective?(σ) then “ else “”; if progressive?(σ) then “prog(”; % Now the progressive operator “µ(σ)”]; % and finish off with the matrix. LEGEND Inputs: Σ is a sentence σ is the matrix of Σ
Functions: µ(Σ) returns the proposition found in Σ tense(σ) returns the tense in σ
Boolean functions: perfective?(σ) is true iff σ has a perfective aspect progressive?(σ) is true iff σ has a progressive aspect event?(σ) is true iff σ indicates an event state?(σ) is true iff σ indicates a state verb as opposed to an activity ]
– – – –
output as many right parentheses as necessary.
a past or future tense shifts the temporal context, and thus also the tor away from the tou; if there is no aspect then the toe is moved to slightly after the tor; a perfective, if present, shifts the toe to be earlier than the tor; if the verb (plus context) is a non-event, or if the progressive aspect is used, then there is a universal quantification, otherwise existential quantification, of the main proposition.
This procedure does not pretend to capture all the subtleties of tense and aspect (Steedman 1999). Our aim has been to provide a framework that is sufficient to embed the semantics of temporal prepositional phrases, to which we now turn.
Temporal semantics of prepositions in context
. Temporal prepositions We turn now to the semantics of various prepositions that can govern a definite temporal interval, which we call the ‘prep interval.’ We will examine three groups: – – –
those that fix the time of the main event to be the same as that of the interval given in the prepositional phrase: at, on, in, during, throughout; those that fix the time to be earlier or later than the prep interval: after, before; those that fix the time to be in a range with the prep interval at one end and usually the tor at the other: since, until, till, by.
Certain temporal prepositions have not been covered as they occur infrequently with intervals and have special uses: for which is used when making plans: (26) Gunner invited Ridley for the weekend;
over which is an alternative to throughout, generally one in which some activity is developing: (27) Gunner had been growing suspicious of Ridley over the year;
through, also an alternative to throughout, used with a constant activity: (28) Gunner will be recording Ridley’s movements through the summer;
within which is generally used with a definite duration linked to the tor: (29) Gunner killed Ridley within the hour.
between . . . and (30) Gunner killed Ridley between dawn and dusk.
from . . . to (31) Gunner followed Ridley from dawn to dusk.
We have also ignored any pre-modifiers, e.g. 5 minutes after, and so have not dealt at all with ago which only appears in this form. In the absence of any aspect, we have chosen to place the prep time later than the tor. The reason for this is similar to the reason for using of tor + ε when there is no preposition. The normal expectation in a story is that time
David S. Brée and Ian E. Pratt-Hartmann
advances with each event that is reported. To place an event earlier than the tor it is usual to use the perfective aspect. Consider the short sequence: (32) Gunner the hotel in a hurry. He found Ridley in the station at 5 o’clock. left * Ridley arrived at 4 o’clock. He was about to leave. had
After the first sentence the tor is the time of Gunner’s leaving the hotel. It is clear that the prep time of 5 o’clock is later than this. In the third sentence the prep time is earlier than the updated tor=0500 and so a perfective aspect is required. After this third sentence the tor remains at 0500: it is at 0500 that Ridley was about to leave. Of course, if there is no tor, as in a story that begins: (33) Gunner left Moscow in 1992.
then there is no restriction on the prep time. We now consider the semantics of each of the above three groups of prepositions in the absence of any aspect in the matrix. The first group is subdivided into those which have an existential reading and throughout, which indicates universal quantification. . At, on, in and during The prepositions at, on, in and during indicate that the main proposition takes place in the interval that they govern, i.e. the prep interval. The difference between them (Brée 1992) is of little import here, but can be summarised as: at is used with prep intervals that are very short, i.e. clock times, natural events, e.g. dawn, temporal boundaries, e.g. beginning, and even religious festivals, e.g. Easter. on is used for a complete day or an indefinite but qualified part of a day, e.g. on a fine morning; in is used for prep intervals longer than a day and for unqualified parts of a day, e.g. in the morning; during is used for prep intervals of any length, except clock times; in particular it is available to introduce prep intervals that are described, not by an explicit interval, but by an event, e.g. during the Second World War, but it is not confined to non-temporal descriptions of prep intervals, e.g. during 1998. The matrix proposition holds at some interval within the prep interval, which itself is later than the tor, e.g.
Temporal semantics of prepositions in context
at dawn dawn killed on Sunday Sunday Event G R ; •→(tor, •( , G kill R)). will kill in 1992 1992 during the War the War was State Gunner sad on Sunday; •→(tor, •(Sunday, G sad)). will be followed Activity Gunner R on Sunday; •→(tor, •(Sunday, G follow R)). will follow
The quantification is determined by the verb type, as when there is no preposition. It might be objected that even for non-event verbs the quantifier should be existential since, for instance, there might be no claim that Gunner was sad every moment of Sunday, only that he was sad some of the time. However, non-events have the implicature, if not the truth functionality, that they hold over the whole of the prep interval, e.g. Food was in short supply in the winter of 1945. So we have chosen to retain universal quantification here. . Throughout Whereas with at, on, in and during quantification over the prep interval is determined by the verb type, throughout requires universal quantification over this interval. Universal quantification cannot be applied to events (unless the event is generic), so the progressive aspect is then required. Throughout cannot be used with short intervals that take at, but religious festivals are acceptable. shot *Ridley Event Gunner throughout March. [Generic] will shoot rats was State G worried throughout March; •→(tor, •(March, G worried)). will be followed Activity G R throughout March; •→(tor, •(March, G follow R)). will follow . After and before After and before place the time of the matrix event (or non-event) later or earlier than, rather than in, the prep interval, respectively. Frequently the matrix event is assumed to be close to the prep interval, but this is not necessarily the case, cf.
David S. Brée and Ian E. Pratt-Hartmann
(34) Gunner ran off after shooting Ridley. (35) Gunpowder was re-invented in the West after it had been invented in China.
So after and before indicate a temporal range open at one end, or, more precisely, bounded by the context interval. Once again the prep interval is later than the tor, as illustrated by the following example: after (36) Gunner found Ridley at dawn. He killed him 5 a.m. before in which the 5 a.m. must be later than the tor=dawn. If the prep time is to be earlier than the tor, then the perfective aspect is required, e.g. * after (37) Gunner killed Ridley at 5 a.m. He found him 4 a.m. had before The range for the main proposition is that half of the time line within the Context interval which is on the appropriate side of the prep interval given by the after /before phrase, e.g. killed after •→ Event G R 5 a.m.; •→(tor, (0500, G kill R)). will kill before ←• was after State Gunner sad the shooting; will be before •→ •→(tor, (t(shooting), G sad)). ←• followed after Activity Gunner Ridley the shooting; will follow before •→ •→(tor, (t(shooting), G follow R)): ←• where t(shooting) indicates the interval during which the shooting occurred. The quantifier is determined by the verb type. Note that activity verbs are given an existential rather than the universal quantification that they would have had with no preposition. . Since, until and by While after and before indicate a range of time whose beginning or ending is set by the context, since, until, till and by indicate an interval beginning
Temporal semantics of prepositions in context
(until, till, by) or ending (since) at the tor. This is always the case for since; as a consequence it requires a perfective aspect. With until, till and by, on the other hand, the beginning of the interval is the tor only by default; it may be some other time, sometimes made explicit using from, e.g. until (38) Gunner had been on tenterhooks then. [Beginning at the begintill ning of the Context] (39) Ridley had been killed by then. [Beginning at the beginning of the Context] until (40) Ridley will be in Moscow from the end of March June. [Beginning till at end of March]
As there is no discernible semantic difference between until and till, and till is less usual, it will not be used in future examples. Temporal since requires a perfective and so is unacceptable with no aspect. Quantification over the interval must be universal with until, so with an event until must be in the context of a progressive. There are no simple restrictions on the use of by. However, with an event in the past tense, by may or may not be acceptable, cf.: found •→(tor, ←•(dawn, G find R)) (41) Gunner ? Ridley by dawn; ? . killed Both find and kill are event verbs. This suggests that a distinction may be needed between different types of event verbs, such as that used by Vendler, viz. achievements and accomplishments. But both achievements and accomplishments are acceptable: finished (42) Gunner the report by dawn. wrote The verb itself is not sufficient to make by unacceptable. For example, embedding kill can make by (at least more) acceptable: (43) The evidence suggests that Gunner killed Ridley by dawn.
With a future tense there is no difficulty: (44) Gunner will kill Ridley by dawn.
We leave the task of finding the restrictions on the use of by with past tense events for another paper.
David S. Brée and Ian E. Pratt-Hartmann
With a non-event verb, either state or activity, the non-event is true of all sub-intervals, so not only until but also by can be used. With until the universal quantification is over the interval starting at the tor (by default), but with by the interval begins at some unspecified time later than the tor but earlier than the prep interval: liked until State G R 1990; •→(tor, ←•(1990, G like R]. will like by •→(t, followed until Activity Gunner Ridley dawn; will follow by •→(tor, ←•(dawn, G follow R]. •→(t, where ‘]’ stands for as many right parentheses as required. With until (but not with by) the tor marks the beginning of the interval only by default. This default may be overridden using from: until (45) Gunner followed Ridley from dusk dawn; *by •→(tor, •→(dusk, ←•(dawn, G follow R))) . * . Summary The temporal semantics of these four classes of preposition are summarised in Table 3. Note: – –
The preposition interval, τ(φ), is later than the tor set up by the previous context. The interval in which the matrix proposition, µ(σ), holds is set: – – –
– –
in the prep interval by at, on, in, during and throughout; later than the prep interval by after and since; earlier than the prep interval by all the other prepositions: before, until, by.
Since is not permitted without a perfective aspect. Throughout and until are the only two prepositions that require universal quantification of the matrix proposition and so are unacceptable with event verbs. At, on, in, during and by accept the same quantification as when there is no preposition, i.e. existential quantification for events and universal quantification for non-events. The remaining prepositions, viz:
Temporal semantics of prepositions in context
Table 3. Temporal semantics for temporal preposition phrases φ in a matrix σ with an event, state or activity verb in the past or future tense, having no aspect Preposition
Event
State
Activity
none at /on /in / during throughout after before since until by
•(tor + ε, µ(σ))
•(tor, µ(σ))
•(tor + ε, µ(σ))
→ • (tor, •(τ(φ), µ(σ)))
→ • (tor, •(τ(φ), µ(σ)))
→ • (tor, •(τ(φ), µ(σ)))
*
→ • (tor, •(τ(φ), µ(σ)))
→ • (tor, •(τ(φ), µ(σ)))
→ • (tor, → • (τ(φ), µ(σ)))
→ • (tor, → • (τ(φ), µ(σ)))
→ • (tor, → • (τ(φ), µ(σ)))
→ • (tor, ← •(τ(φ), µ(σ)))
→ • (tor, ← •(τ(φ), µ(σ)))
→ • (tor, ← •(τ(φ), µ(σ)))
* *
*
*
→ • (tor, ← •(τ(φ), µ(σ)))†
→ • (tor, ← •(τ(φ), µ(σ)))†
→ • (tor, ← •(τ(φ), µ(σ)))‡
→ • (tor, ← •(τ(φ), → • (t, µ(σ)))) → • (tor, ← •(τ(φ), → • (t, µ(σ))))
µ(σ) τ(φ) •(t, α) •→(t, α) ←•(t, α) t * † ‡
–
The proposition given by the matrix σ. The interval indicated by the temporal prepositional phrase φ. Proposition α holds in the unique time interval t. Proposition α holds from the unique time t up the interval marking the end of the context. Proposition α holds from the beginning of the context up to the unique time t. is any interval within the context, here between tor and τ(φ). Unacceptable. The tor is by default. In the past tense, some event verbs are unacceptable in some contexts.
after and before, allow an existential quantification not only for events but also for activities. By may not occur with some event verbs in the past tense.
The stage is now set for us to address the question: is the temporal semantics of a sentence determined by a simple summation of the semantics given by the temporal prepositions to those already given by tense and aspect, or is there an interaction?
. Temporal prepositions in a past or future matrix We have now established a semantics both for tense and aspect on the one hand and for the different temporal prepositions, when found in a matrix having a verb with no aspect, on the other. We now explore their combination. We look first at the use of temporal prepositions in sentences with a verb having a
David S. Brée and Ian E. Pratt-Hartmann
perfective aspect. In the next sections we turn to sentences with a verb having progressive and perfect progressive aspects. . Perfective aspect Recall that the perfective, with no temporal preposition, forces an existential quantification of the matrix proposition, µ(σ), over the interval earlier than the tor: ←•(tor, µ(σ)). In a matrix with no aspect, the prep interval itself, τ(φ), is later than the tor: •→(tor, *(τ(φ), Qµ(σ))), where * is a temporal relation (•, ←•, •→), depending on the preposition, and Q is a quantifier (, ), depending on the verb type as well as the preposition (Table 3). In all cases we will see that the perfective takes precedence over the preposition in that the prep interval is set before the tor rather than later, thus: ←•(tor, *(τ(φ), Qµ(σ))). What is of interest is the selection of the quantifier: is it determined by the verb type, the aspect or the preposition? .. At, on, in and during With at, on, in and during the event or non-event is situated in the prep interval (Section 3.1): had Event Gunner killed Ridley on Sunday; ←•(tor, •(Sunday, will have G kill R)). had State Gunner been sad on Sunday; ←•(tor, •(Sunday, G sad)). will have had Activity G followed R on Sunday; ←•(tor, •(Sunday, G follow R)). will have The combination with the perfective aspect is purely additive, with existential quantification determined by the perfective aspect.3 .. Throughout Throughout indicates that the matrix event or non-event occurs in every sub-interval of the prep interval (Section 3.2). With events the progressive aspect is required, so events are not acceptable with just the perfective aspect. With non-events a simple perfective puts the prep interval earlier than the tor: had State Gunner liked R throughout 1992; ←•(tor, •(1992, will have G like R)).
Temporal semantics of prepositions in context
had Activity G followed R throughout 1992; ←•(tor, •(1992, will have G follow R)).
Note that throughout requires universal quantification. .. After and before After and before situate the matrix event or non-event later or earlier, respectively, than the prep interval (Section 3.3): had after Event Gunner killed Ridley 5 a.m.; will have before •→ ←•(tor, (0500, G kill R)). ←• had after State Gunner been sad the shooting; will have before •→ ←•(tor, (t(shooting), G sad)). ←• had after Activity Gunner followed Ridley the shooting; will have before •→ ←•(tor, (t(shooting), G follow R)). ←• Note that the quantification is determined by the perfective aspect. .. Since, until and by Recall that since places the matrix (non-)event after the prep interval but before the tor, while until and by place it after the tor but before the prep interval (Section 3.4). As the perfective aspect places the event earlier than the tor, and both until and by with no aspect indicate an interval beginning at the tor, we would expect neither to be acceptable. However, until, just with non-events, and by, just with events, can mark the end of an interval whose beginning is set by the Context: since had Event Gunner killed Ridley by dawn; will have *until •→ ←•(tor, (dawn, G kill R)). ←• *
David S. Brée and Ian E. Pratt-Hartmann
since had State Gunner liked Ridley until 1992; will have *by •→ ←•(tor, ←• (1992, G like R)). * since had Activity Gunner followed Ridley until dawn; will have *by •→ ←•(tor, ←• (dawn, G follow R)). *
Note that the quantification is determined by the verb type. With the future tense the beginning of the Context is later than the tou, |= •→(tou, •(Context, •→(tor, . . .))), and so with by an event that may have occurred earlier than the tou is generally indicated by making the matrix conditional using if . . . not, e.g. (46) Gunner will have killed Ridley by dawn, if he hasn’t done so already.
There is an alternative reading for by. Rather than the tor being updated after the matrix is processed, the tor is first updated to the prep interval and then the matrix event is placed earlier: (47) By dawn, Gunner had killed Ridley; tor ⇐ dawn; ←•(tor, G kill R).
This phenomenon takes us beyond the scope of this paper. Note that: – –
With since, the quantification of the matrix proposition is existential for events but universal for non-events. For until and by, when there is no aspect the beginning of the interval in which the matrix event may occur is normally the tor. However, the perfective sets the prep interval to be earlier than the tor, so the beginning of the interval in which the matrix proposition holds is set instead by the Context, but only for non-events with until and only for events with by.
Temporal semantics of prepositions in context
.. Summary The interaction of the perfective aspect with different prepositions depends on the type of the verb, as shown in Table 4, which summarises the details given above. Note: –
The prep interval is set earlier than the tor, following the perfective aspect (Table 1). The temporal relation between the matrix interval and the prep interval is identical to that found with no aspect (Table 3). For most prepositions, quantification is existential, irrespective of the type of verb, which is consistent with the perfective aspect (Table 1). However:
– –
–
– –
with since quantification is determined by the type of verb, as was the case with the other prepositions when there was no aspect. Maybe this is because temporal since requires a perfective. throughout and until require universal quantification as usual.
By is acceptable with events but not with non-events (which were acceptable with no aspect).
Table 4. Temporal semantics for preposition phrases φ in a matrix σ with an event or non-event verbs in the past or future tense, having a perfective aspect Preposition
Event
Non-event
none at /on /in /during throughout after before since until by
←•(tor, µ(σ)) ←•(tor, •(τ(φ), µ(σ))) * ←•(tor, •→(τ(φ), µ(σ))) ←•(tor, ←•(τ(φ), µ(σ))) ←•(tor, •→(τ(φ), µ(σ))) * ←•(tor, ←•(τ(φ), µ(σ)))
←•(tor, µ(σ)) ←•(tor, •(τ(φ), µ(σ))) ←•(tou, •(τ(φ), µ(σ))) ←•(tor, •→(τ(φ), µ(σ))) ←•(tor, ←•(τ(φ), µ(σ))) ←•(tou, •→(τ(φ), µ(σ))) ←•(tor, ←•(τ(φ), µ(σ))) *
µ(σ) τ(φ) •(t, α) •→(t, α) ←•(t, α) *
The proposition given by the matrix σ. The interval indicated by the temporal prepositional phrase φ. Proposition α holds in the unique time interval t. Proposition α holds from the unique time t up the interval marking the end of the context. Proposition α holds from the beginning of the context up to the unique time t. Unacceptable.
David S. Brée and Ian E. Pratt-Hartmann
. Progressive aspect We turn now to considering the interaction between the progressive aspect and temporal prepositions. Recall that one function of the progressive aspect is to indicate that a proposition holds at the tor rather than some time later (Section 2.3.2): (48) •(tor, prog(µ(σ))).
On the other hand, a temporal preposition with no aspect introduces a prep interval that is after the tor (Section 3). The progressive aspect does not affect this, as illustrated by the following sequence in which 1205 is acceptable, as it is later than the previous tor of 1200, but 1130 is not acceptable: 1205 (49) Gunner arrived in Moscow at noon. At he was looking for Ridley *1130 when he suffered a heart attack. Note also that the tor is updated to a time within the matrix state, e.g. the heart attack took place during the looking. Compare this with no aspect: (50) *At 1205, Gunner looked for Ridley when he suffered a heart attack.
which is unacceptable because the at phrase and the when clause have the same function: to fix the toe. The progressive of a state verb is unnecessary; it is frequently interpreted as deceit on the part of the agent (Section 2.3.2). This is also the case when a temporal prepositional phrase is introduced. .. At, on, in and during At /on /in /during in a progressive matrix set the time of the non-event resulting from applying the progressive to within the prep interval (cf Section 3.1). An issue arises as to whether the prep interval is after the tor or the tor is within the prep interval, cf.: (51) Gunner was killing R on Sunday; •→(tor, •(Sunday, prog(G kill R))). •(Sunday, •(tor, prog(G kill R))).
Recall that (52) Gunner was killing Ridley; •(tor, prog(G kill R)).
Note also that if a when clause is added to (51) then the when, which normally refers to the tor, is the interval in which the killing took place rather than the
Temporal semantics of prepositions in context
whole of the Sunday. This suggests that we choose the tor within the prep interval, i.e. •(Sunday, •(tor, prog(G kill R))). However, under a normal flow of time Sunday should be later than the previous tor, cf.: 6 (53) Gunner arrived at 5 a.m. He was killing Ridley at , when . . . *4 This suggests that we need to note both the old and the new tor: (54) •→(tor, •(Sunday, •(tor , prog(G kill R)))),
where tor is the new updated tor. So in general we have: was Event G killing R on Sunday; will be •→(tor, •(Sunday, •(tor , prog(G kill R))). was State ?G being sad on Sunday; will be •→(tor, •(Sunday, •(tor , prog(G sad))). was Activity Gunner following Ridley on Sunday; will be •→(tor, •(Sunday, •(tor , prog(G follow R))). Note how the progressive aspect applied to a state verb introduces a note of deceit and, applied to an activity verb, prevents completion (Section 2.3.2). .. Throughout Throughout gives a universal quantification over the entire prep interval, which again is later than the tor (cf Section 3.2). There is no need to include an updated tor: was Event G killing R throughout March; will be •→(tor, •(March, prog(G kill R))). was State ?G liking R throughout March; will be •→(tor, •(March, prog(G like R))). was Activity Gunner following Ridley throughout March; will be •→(tor, •(March, prog(G follow R))).
David S. Brée and Ian E. Pratt-Hartmann
.. After and before After and before place the matrix (non-)event later or earlier than the prep interval (cf Section 3.3). Again the tor is updated to a new tor that is after/before the prep interval, respectively: was after Event Gunner killing Ridley the tip-off; will be before •→ •→(tor, (t(tip-off), •(tor , prog(G kill R)))). ←• was after ? State Gunner liking Ridley that; will be before •→ •→(tor, (that time, •(tor , prog(G ‘like’ R)))). ←• was after Activity Gunner crying the shooting; will be before •→ •→(tor, (t(shooting), •(tor , prog(G cry)))). ←• The past tense sentences can also be given a future intent reading, in which the tor is not updated to the time of the main event. The sentence then conveys that the actor, here Gunner, has an intention to carry out an act, e.g. killing, at a later time. This reading is the same as in: after (55) Gunner was going to kill Ridley the tip-off. before Also a case could be made that the prep interval may occur earlier than, rather than after, the tor, e.g. (56) Ridley found Gunner. He was crying after the shooting.
We leave these two possibilities aside for a future investigation. .. Since, until and by Recall that since requires a perfective aspect for a temporal interpretation and so is not acceptable with a progressive aspect (Section 3.4). Until is a universal quantifier and so normally requires a non-event verb and cannot be used in the context of an event verb. Here, however, the progressive aspect turns an event into an activity and so until can be used. Just as with throughout there is no need for an updated tor to be included:
Temporal semantics of prepositions in context
was until killing Ridley dawn; will be by •→(tor, ←•(dawn, prog(G kill R.]. •→(t, was until ? liking Ridley dawn; State Gunner will be by •→(tor, ←•(dawn, prog(G like R.]. •→(t, was until Activity Gunner following R dawn; will be by •→(tor, ←•(dawn, prog(G follow R]. •→(t, Event Gunner
where t is an unbound time in the context, and ‘]’ indicates that all open left parentheses should be closed. .. Summary These observations on the interaction of prepositions and the progressive aspect are drawn together in Table 5. Note: –
– –
The progressive aspect turns an event into an activity (Table 1), so that both throughout and until, which require universal quantification, are acceptable with events. The temporal relation between the tor and the prep interval is identical to that found with no aspect (Table 3). Quantification of the matrix proposition is determined by the progressive (Table 1) but over an updated tor.
. Perfect progressive aspect Finally we turn to the interaction between temporal prepositions and the perfect progressive aspect. Recall that the combination of the progressive and perfective aspects combines their effects, giving an event a duration and placing this earlier than the tor (Section 2.3.3): (57) ←•(tor, prog(µ(σ)))
As the perfect progressive of a state verb is unnecessary, it is frequently interpreted as deceit on the part of the agent (Section 2.3.3). This is still so when a temporal prepositional phrase is introduced.
David S. Brée and Ian E. Pratt-Hartmann
Table 5. Temporal semantics for preposition phrases φ in a matrix σ with any verb in the past or future tense, having progressive aspect Preposition
Any verb
none at/on/in/during throughout after before since until by
•(tor, prog(µ(σ))) •→(tor, •(τ(φ), •(tor , prog(µ(σ)))) •→(tor, •(τ(φ), prog(µ(σ)))) •→(tor, •→(τ(φ), •(tor , prog(µ(σ))))) •→(tor, ←•(τ(φ), •(tor , prog(µ(σ))))) * •→(tor, ←•(τ(φ), prog(µ(σ)))) •→(tor, ←•(τ(φ), •→(t, prog(µ(σ)))))
µ(σ) τ(φ) •(t, α) •→(t, α) ←•(t, α) prog(φ)
tor t *
The proposition given by the matrix σ. The interval indicated by the temporal prepositional phrase φ. Proposition α holds in the unique time interval t. Proposition α holds from the unique time t up the interval marking the end of the context. Proposition α holds from the beginning of the context up to the unique time t. A function that turns an event φ into an activity, that with a state verb gives an implicature of deceit on the part of the agent and with an activity verb an implicature that there is no accomplishment. The updated tor. is an unbound time in the context. Unacceptable.
Mostly the perfect progressive is the simple addition of the perfective and the progressive aspects. So when the progressive sets the new tor it is to a time earlier than the present tor. ..
At, on, in and during
had been killing Ridley on Sunday; will have ←•(tor, •(Sunday, •(tor , prog(G kill R)))). had State ?Gunner been being sad on Sunday; will have ←•(tor, •(Sunday, •(tor , prog(G sad)))). had Activity Gunner been following Ridley on Sunday; will have ←•(tor, •(Sunday, •(tor , prog(G follow R)))). Event Gunner
Temporal semantics of prepositions in context
..
Throughout
..
After and before
had Event Gunner been killing Ridley throughout March; will have ←•(tor, •(March, prog(G kill R))). had ? State Gunner been liking Ridley throughout March; will have ←•(tor, •(March, prog(G like R))). had Activity Gunner been following Ridley throughout March; will have ←•(tor, •(March, prog(G follow R))). had after been killing Ridley 5 a.m.; will have before •→ ←•(tor, (0500, •(tor , prog(G kill R)))). ←• had after State ?Gunner been liking Ridley the shooting; will have before •→ ←•(tor, (shooting, •(tor , prog(G like R)))). ←• had after Activity Gunner been following Ridley the shooting; will have before •→ ←•(tor, (t(shooting), •(tor , prog(G follow R)))). ←• Event Gunner
.. Since, until and by The perfect progressive aspect can be used with since, not surprisingly as since indicates an interval ending at the tor (required by the perfective). Until, which, by default, indicates an interval starting at the tor, can also be used. The interval ends at the prep interval introduced by until, which is set earlier (by the perfective aspect) rather than later than the tor. The start of the interval is determined by the Context. On the other hand, the use of the perfect progressive with by is unacceptable. This is not surprising as by is unacceptable with perfective non-event verbs and the progressive ensures that even events are to be considered as activities.
David S. Brée and Ian E. Pratt-Hartmann
since had Event Gunner been killing Ridley until dawn; will have *by •→ ←•(tor, (dawn, prog(G kill R))). ←• * since had State Gunner been liking Ridley until 1992; will have by* ←•(tor, •→ (1992, G ‘like’ R)). ←• * since had Activity Gunner been following Ridley until dawn; will have *by ←•(tor, •→ (dawn, G follow R)). ←• *
.. Summary These observations on the interaction between temporal prepositions and the perfect progressive aspect are drawn together in Table 6. Note: –
–
– – –
The temporal relation between the interval in which the matrix occurs and the tor is identical to that found with prepositions with no aspect (Table 3). The perfect part of the perfect progressive puts the prep interval earlier than the tor (Table 1), overriding the preposition in a sentence with no aspect (which puts the prep interval later than the tor). The progressive part of the perfect progressive turns an event into an activity (Table 1). The quantification of the matrix proposition is universal as with the simple progressive aspect (Table 3). By is not acceptable because the perfective itself is unacceptable with nonevents and the progressive turns even events into non-events.
Temporal semantics of prepositions in context
Table 6. Temporal semantics for preposition phrases φ in a matrix σ with any verb in the past or future tense, having perfect progressive aspect Preposition
Any verb
none at /on /in /during throughout after before since until by
←•(tor, prog(µ(σ))) ←•(tor, •(τ(φ), •(tor , prog(µ(σ))))) ←•(tor, •(τ(φ), prog(µ(σ)))) ←•(tor, •→(τ(φ), •(tor , prog(µ(σ))))) ←•(tor, ←•(τ(φ), •(tor , prog(µ(σ))))) ←•(tor, •→(τ(φ), prog(µ(σ)))) ←•(tor, ←•(τ(φ), prog(µ(σ)))) *
µ(σ) τ(φ) •(t, α) •→(t, α) ←•(t, α) prog(φ)
tor *
The proposition given by the matrix σ. The interval indicated by the temporal prepositional phrase φ. Proposition α holds in the unique time interval t. Proposition α holds from the unique time t up the interval marking the end of the context. Proposition α holds from the beginning of the context up to the unique time t. A function that turns an event φ into an activity, that with a state verb gives an implicature of deceit on the part of the agent and with an activity verb an implicature that there is no accomplishment. The updated tor. Unacceptable.
. Summary The above analysis is brought together in Table 7. Each column of this table shows the interaction of the preposition with aspect. For each group of prepositions, the semantics for event and non-event verbs are both given. In a few cases there is a difference between state and activity verbs which is shown in the footnotes. When there is no aspect and no temporal preposition, the interval is set close to the tor. The introduction of a temporal preposition moves the interval in which the main proposition holds to a time set by the prep interval given in the prepositional phrase, which is itself after the tor. Most prepositions have different effects; those that have the same effect, e.g. at, on, in and during, generally take different classes of noun phrase; until and till, however, are indistinguishable in use:
E E˜
E E˜
E E˜
E E˜
E E˜
E E˜
E E˜
at/on/in/ during
throughout
after
before
since
until/ till
by
tor
‡
t †
← •(tor, → • (τ(φ), µ(σ))) ← •(tor, → • (τ(φ), µ(σ))) ← •(tor, ← •(τ(φ), µ(σ))) ← •(tor, ← •(τ(φ), µ(σ)))
→ • (tor, → • (τ(φ), µ(σ))) → • (tor, → • (τ(φ), µ(σ)))‡
→ • (tor, ← •(τ(φ), µ(σ)))
→ • (tor, ← •(τ(φ), µ(σ)))‡
* ← •(tor, ← •(τ(φ), µ(σ))) ← •(tor, ← •(τ(φ), µ(σ)))
*
*
→ • (tor, ← •(τ(φ), µ(σ)))††
→ • (tor, ← •(τ(φ), µ(σ)))‡‡
→ • (tor, ← •(τ(φ), → • (t, µ(σ)))
← •(tor, → • (τ(φ), µ(σ)))
← •(tor, → • (τ(φ), µ(σ)))
← •(tor, •(τ(φ), µ(σ)))
* *
*
*
← •(tor, •(τ(φ), µ(σ))) ← •(tor, •(τ(φ), µ(σ)))
→ • (tor, •(τ(φ), µ(σ)))
→ • (tor, •(τ(φ), µ(σ)))
→ • (tor, •(τ(φ), µ(σ)))
← •(tor, µ(σ)) ← •(tor, µ(σ))
•(tor + ε, µ(σ))
•(tor + ε, µ(σ))†
Aspect: Perfective
None
← •(tor, ← •(τ(φ), prog(µ(σ))))
→ • (tor, ← •(τ(φ), → • (t, prog(µ(σ)))) → • (tor, ← •(τ(φ), → • (t, prog(µ(σ))))
* *
← •(tor, ← •(τ(φ), prog(µ(σ))))
→ • (tor, ← •(τ(φ), prog(µ(σ)))) → • (tor, ← •(τ(φ), prog(µ(σ))))
← •(tor, → • (τ(φ), prog(µ(σ))))
← •(tor, → • (τ(φ), prog(µ(σ))))
← •(tor, ← •(τ(φ), •(tor , prog(µ(σ)))))) ← •(tor, ← •(τ(φ), •(tor , prog(µ(σ)))))
* *
→ • (tor, ← •(τ(φ), •(tor , prog(µ(σ)))))
→ • (tor, ← •(τ(φ), •(tor , prog(µ(σ)))))
→ • (tor, → • (τ(φ), •(tor , prog(µ(σ)))))
← •(tor, → • (τ(φ), •(tor , prog(µ(σ))))) ← •(tor, → • (τ(φ), •(tor , prog(µ(σ))))
→ • (tor, •(τ(φ), prog(µ(σ)))) → • (tor, → • (τ(φ), •(tor , prog(µ(σ)))))
← •(tor, •(τ(φ), prog(µ(σ))))
→ • (tor, •(τ(φ), •(tor , prog(µ(σ)))))
← •(tor, •(τ(φ), prog(µ(σ))))
← •(tor, •(τ(φ), •(tor , prog(µ(σ)))))) ← •(tor, •(τ(φ), •(tor , prog(µ(σ)))))
→ • (tor, •(τ(φ), •(tor , prog(µ(σ))))) → • (tor, •(τ(φ), prog(µ(σ))))
← •(tor, prog(µ(σ)))
← •(tor, prog(µ(σ)))
Perfect progressive
•(tor, prog(µ(σ)))
•(tor, prog(µ(σ)))
Progressive
The timeless unquantified proportion from the matrix σ. The interval indicated by the temporal prepositional phrase φ. Proposition α holds in the unique time interval t . Proposition α holds from the unique time t up the interval marking the end of the context. Proposition α holds from the beginning of the context up to the unique time t . A function that turns an event φ into an activity, that with a state verb gives an implicature of deceit on the part of the agent and with an activity verb an implicature that there is no accomplishment. the updated value of tor. is an unbound time in the context. * Unacceptable. †† The tor is by default. If the verb is a state, then ε = 0. ‡‡ In the past tense some event verbs are unacceptable in some contexts. If the verb is an activity, then is replaced by .
E E˜
none
µ(σ) τ(φ) •(t, α) → • (t, α) ← •(t, α) prog(φ)
V
Preposition
Table 7. Temporal semantics for Event (E) or Non-event (E˜ ) verbs, within |= [←•or•→](tou, •(Context, . . .))
David S. Brée and Ian E. Pratt-Hartmann
Temporal semantics of prepositions in context
At/on/in/during/throughout fix the main interval to be the same as the prep interval τ(φ), with either universal quantification (throughout) or quantification determined by the type and aspect of the verb (at, on, in, during). After/before fix the main interval to be later/earlier than τ(φ). Since fixes the interval to start at the end of τ(φ) and end at tor; it requires a perfective aspect to do so. Until/till/by fix the interval to end at τ(φ) and to start either at the tor, by default (until), or at some sub-interval between the tor and the prep interval (by), with universal quantification. A perfective aspect moves the interval in question to be earlier than the tor. With since, this shift is required, so the perfective aspect is always required with since. Until (till), which by default indicates an interval beginning with the tor when there is no aspect, and by which indicates an interval after the tor, might be expected not to take a perfect(-progress)ive aspect. In fact until just with non-events and by just with events then indicate an interval, with its beginning point set to the start of the Context rather than to the tor. The progressive aspect turns events into activities, i.e. there is no completion of the event. When combined with a state verb, there is a sense of action on the part of the agent, transforming the state into an activity and generally giving a sense of deceit on the part of the agent. The progressive with an activity verb only has the effect of indicating that the activity still holds at the updated tor, i.e. the tor after the sentence has been processed. It retains this function when combined with temporal prepositions. With at, on, in, during, after and before the tor is first updated to an interval set by the preposition, and then the progressive requires universal quantification over this updated tor. With the prepositions that require universal quantification, viz: throughout and until, and with by, this updating is unnecessary. The combination of the perfective and progressive aspects both sets the interval to be earlier than the tor and forces events to be interpreted as activities. In general, the combination also does this when used with temporal prepositions. By, which is only acceptable with the perfective if the verb is an event, is unacceptable with the prefect progressive as the progressive turns even events into non-events. With past and future tenses, the way the semantics of aspect and of the temporal prepositional phrase combine is:
David S. Brée and Ian E. Pratt-Hartmann
Table 8. Procedure for the temporal semantics of all sentences µ(Σ) = if
(π(φ) = since ∧ ¬perfective?(σ))∨ % Exceptions (π(φ) = by ∧ perfective?(σ) ∧ (state?(σ) ∨ progressive?(σ))) then “* else “ |= ; case tense(σ): % Set the tor relative to the tou past “←•(tou,”; future “•→(tou,”; esac; “•(Context,”; if perfective?(σ) then “←•” % Set the toe relative to the tor else if φ =‘’ then “ • else “•→ ; “(tor”; if φ = ‘’ ∧ ¬perfective?(σ) ∧ ¬progressive?(σ) then “ + ε ; “,”; case π(φ) : % Set up the prepositional constraint at|on|in|during|throughout “•”; after|since “•→”; before|until|by “←•”; esac; “(τ(φ),”; % and add the prep interval if ¬(π(φ) = throughout|since|until|by) ∧ progressive?(σ) % update tor first then “ • (tor , ; % t any time between tor and prep interval if π(φ) = by then “•→(t, ; case π(φ) % Select the quantifier throughout|until if progressive?(σ) ∨ ¬event?(σ) then “ else “* ; after|before if progressive?(σ) ∨ state?(σ) ∧ ¬perfective?(σ)) then “ else “ ; T if progressive?(σ) ∨ (¬event?(σ) ∧ ¬perfective?(σ)) then “ else “ ; esac; if progressive?(σ) then “prog(”; “µ(σ)”]; % and finish off with the matrix. LEGEND Inputs: Functions: Σ is a sentence µ(σ) returns the proposition found in σ σ is the matrix of Σ π(φ) returns the temporal preposition in φ φ is the temporal phrase in Σ τ(φ) returns the time found in φ tense(σ) returns the tense in σ Boolean functions: perfective?(σ) is true iff σ has a perfective aspect progressive?(σ) is true iff σ has a progressive aspect event?(σ) is true iff σ indicates an event state?(σ) is true iff σ indicates a state verb ] output as many right parentheses as necessary.
Temporal semantics of prepositions in context
1. The tense sets the context and thus the tor to being before or after the tou: ←• |= (tou, •(Context, . . . •→ The toe is then set in relation to the tor: – – –
if there is a perfective, then to be earlier (whether or not there is a temporal preposition): ←•(tor, . . .; if no temporal preposition nor aspect is present, then to be at or just after: •(tor + ε, . . .; otherwise, to being after: •→(tor, . . .;
2. If there is a temporal preposition, then its temporal constraint is now introduced, together with the interval signalled by its noun phrase. 3. In addition: –
–
With progressives, for all prepositions except since, until, by and throughout, the quantification is to be over the tor updated to be within the context set by the preposition: •(tor , . . . For by an unbounded interval t later than the tor is introduced so that quantification is over the interval between t and the prep interval: •→(t, . . ..4
4. Then the appropriate quantifier is added:5 – – – – – –
if there is a progressive aspect then ; if the temporal preposition is throughout or until, then ; if there is a perfective aspect then ; if the temporal preposition is after or before and the verb is an activity, then ; if the verb is a non-event then ; otherwise .
5. If there is a progressive aspect, its effect is now set to operate on the matrix proposition. 6. Finally the matrix proposition is inserted. This procedure is shown in Table 8. The combination of the effect of tense and aspect on the one hand (1, 2, 5, 6) and the temporal prepositional phrase on the other (3, 4, 5) are mostly additive. In setting the toe relative to the tor (2), prepositions place the toe
David S. Brée and Ian E. Pratt-Hartmann
to be later than the tor rather than at or just after the tor, but this is hardly an interaction. However, the main interactions are: –
Quantification is generally determined by the aspect, when present, and by the type of verb otherwise, except that: – throughout and until require universal quantification, – after and before require existential quantification with activities.
–
–
–
Temporal since can only occur with a perfective aspect, as was known. The reason is that quantification with since is always over an interval ending at the tor. Temporal by can only take a perfective aspect with (some) events. As the progressive turns an event into an activity, perfect progressives do not take by. The reason is that quantification is always over an interval that is later than the tor. With the prepositions at, on, in, during, after and before, but not with throughout, until and by, the progressive requires that the tor be updated to an interval as indicated by the preposition. The universal quantification is then over this updated tor. The reason is that throughout, until and by force universal quantification.
Notes . Note that the tor and the toe may be on opposite sides of the tou. This overcomes Comrie’s (1985) objection to Reichenbach’s system. . This is in line with ter Meulen (1995), but not with Steedman (1999) who suggests that ‘context’ should determine ε. We incorporate context into the decision of whether a verb is being used to describe an event or a non-event. . Actually universal quantification might be called for with state verbs. This is an open question. . When quantification is existential this step makes no semantic difference and may be omitted. . Note that these rules are ordered, the first one that is acceptable is taken. The format is slightly different from that given in Table 8, but the effect is the same.
Temporal semantics of prepositions in context
References Brée, D. S. (1985). The durative temporal subordinating conjunctions: since and until. Journal of Semantics, 4, 1–46. Brée, D. S. (1992). Words for time. In F. Macar, V. Pouthas, and W. Friedman (Eds.), Time, Action and Cognition (pp. 337–348). Dordrecht, NL: Kluwer. Comrie, B. (1985). Tense. Cambridge, UK: Cambridge University Press. Enç, M. (1996). Tense and modality. In S. Lappin (Ed.), The Handbook of Contemporary Semantic Theory (pp. 345–358). Oxford, UK: Blackwell. Meulen, A. G. B. ter (1995). Representing Time in Natural Language. Cambridge, MA: MIT Press. Pratt, I. E., & Brée, D. S. (1995). An approach to the semantics of some English temporal constructions. In Proceedings of the Seventeenth Cognitive Science Conference (pp. 118– 123). New Jersey: Erlbaum. Quirk, R., Greenbaum, S., Leech, G., & Svartvik, J. (1985). A Comprehensive Grammar of the English Language. London, UK: Longmans. Steedman, M. (1999). Temporality. In J. van Bentham and A. ter Meulen (Eds.), Handbook of Logic and Language (pp. 895–938). Amsterdam, NL: Elsevier. Vendler, Z. (1957). Verbs and times. The Philosophical Review, 66, 143–160. Verkuyl, H. J. (1999). Aspectual Issues: studies on time and quantity. Stanford, CA: Center for the Study of Language and Information.
TSL[v.20020404] Prn:7/05/2002; 13:22
F: TSL5011.tex / p.1 (115)
Prepositions and context Ian Pratt-Hartmann and Nissim Francez Manchester University, UK / Technion, Haifa, Israel
This paper examines the semantics of spatial preposition phrases in sentences which report the occurrence of events, states or activities. The category into which a sentences falls – event-, state- or activity-reporting – is called its aspectual class; and it is well-known that aspectual class distinctions must be taken into account in order to understand the semantics of most temporal constructions in natural language. The goal of this paper is to establish that aspectual class distinctions are also relevant to the interpretation of spatial adverbials. This is a surprising fact, because standard explanations of aspectual class are so intimately bound up with distinctively temporal phenomena in language. The phenomena reported in this paper thus lead us to view the issue of aspectual class in a new light.
.
Background
This paper examines the semantics of spatial preposition phrases in sentences which report the occurrence of events or the holding of states, for example: (1) Mary kissed John in the garden (2) Mary is asleep in the garden.
In the sequel we shall focus particularly on the different ways spatial prepositions behave with respect to events and states, and on how these behaviours are related. The background to our analysis is a theory of the semantics of English sentences originally proposed by Davidson (1967) and extended and elaborated by Parsons (1990), which has by now attained the status of orthodoxy. According to this theory, which we may refer to loosely as the Davidson–Parsons account, sentences such as (3) Mary kissed John
TSL[v.20020404] Prn:7/05/2002; 13:22
F: TSL5011.tex / p.2 (116)
Ian Pratt-Hartmann and Nissim Francez
(4) Mary is asleep
are implicitly existentially quantified. Specifically, 3 and 4 have the respective logical forms: (5) ∃e[kiss(Mary, John, e) ∧ time(e) < ] (6) ∃s[sleep(Mary, s) ∧ time(s) = ],
where e ranges over events and s over states. In addition, adverbial modifiers are taken to contribute extra conjuncts to the quantified formula, so that sentences 1 and 2 have the respective logical forms: (7) ∃e[kiss(Mary, John, e) ∧ place(e) ⊆ place(garden) ∧ time(e) < ] (8) ∃s[sleep(Mary, s) ∧ place(s) ⊆ place(garden) ∧ time(s) = ].
An important advantage claimed for this account is that the proposed logical forms explain the fact that sentence 1 implies sentence 3 and sentence 2 implies sentence 4. In quantifying over both events and states in the above examples, we are alluding to a familiar classification of sentences introduced by Vendler (1967). For present purposes, we can describe this classification as follows. We call the sentence (9) Oswald shot Kennedy
event-reporting, because it does not admit of qualification by a range of temporal adverbials including for, until and throughout: (10) *Oswald shot Kennedy for 3 minutes.
By contrast, the sentence (11) Mary was asleep
is non-event-reporting, because it can be modified by such adverbials. Among non-event-reporting sentences, some have the further property that they resist the progressive form: (12) *Mary is being asleep/happy/a doctor (13) *Mary is loving John (14) *Mary is believing in homeopathy.
Such sentences will be called state-reporting. By contrast, the sentences (15) John followed Mary
TSL[v.20020404] Prn:7/05/2002; 13:22
F: TSL5011.tex / p.3 (117)
Prepositions and context
(16) Mary worked on the paper (17) Mary thought about Vienna
allow both qualification by for-adverbials and the progressive form. Hence they can be neither event-reporting nor state-reporting. These sentences will be called activity-reporting. The category into which a sentence falls – event-reporting, state-reporting or activity-reporting – is known as its aspectual class. The topic of aspectual class has been extensively discussed in the literature (e.g. Verkuyl 1993; ter Meulen 1996; Steedman 1997). It is obvious from the way the aspectual classes are defined that we can expect considerable interaction between aspectual class on the one hand, and the semantics of verb-tense and -aspect and temporal adverbials on the other. Again, the semantics of these temporal constructions has been extensively studied in the literature (see, e.g. Dowty 1979; Stump 1985; Crouch & Pullman 1993; Hwang & Schubert 1994; Ogihara 1996; Pratt & Francez 2001), and, as expected, the issue of aspectual class features in these accounts. The goal of this paper is to establish that aspectual class distinctions are also relevant to the interpretation of spatial adverbials. This is a surprising fact, because aspectual class is so intimately bound up with distinctively temporal phenomena in language. The observations reported in this paper thus lead us to view the issue of aspectual class in a new light.
. The locations of events What does it mean to say that an event is located somewhere in space? One might guess that any event is located in some rather ill-defined region containing all its participants: agent, patient, instrument and so on. But in fact the Davidson–Parsons theory often implies the contrary. Consider the sentences (18) Oswald shot Kennedy in his limousine (19) Oswald shot Kennedy in the book repository.
Given the standard account of the assassination, most speakers would – perhaps with some hesitation – accept sentence 18; by contrast, sentence 19 seems definitely false. Yet it is interesting that the spatial preposition phrases in both sentences exclude one of the participants in the shooting event. Our intuitions are clearer still for the sentences
TSL[v.20020404] Prn:7/05/2002; 13:22
F: TSL5011.tex / p.4 (118)
Ian Pratt-Hartmann and Nissim Francez
(20) Kennedy was shot in the presidential limousine (21) Kennedy was shot in the book repository,
where we easily judge 20 true and 21 false. Note that, if we judge 20 to be true in virtue of the existence of an appropriately located shooting event, and if Oswald was indeed the perpetrator of the only shooting of Kennedy, then it follows, on the Davidson–Parsons account, that sentence 18 is true as well. (Parsons lays great emphasis on the fact that his event-semantics accounts for the logical relationships between active sentences and their corresponding passivizations.) Similar considerations suggest that it must be common for events to be located where their patients (or ‘themes’) are. Thus we say that John telephones Mary in her (= Mary’s) office or that John writes a letter on his desk without any suggestion that John himself needs to be in Mary’s office or on (rather than merely seated at) his desk. Common, but not universal: if John proves a theorem at his desk, it is John and not the theorem that is at his desk. And if Archimedes discovered the principle of flotation in a bathtub then it was again he and not the principle that was getting wet. Cases where the theme of the event is a physical object and where the event is still not taken to be the location of the theme are hard, but not impossible, to think of. If we are told that John hit Mary on the table, then the most natural reading is that they were both sitting or standing on the table, though geometrical constraints may play a special role here. Indeed, some events seem to have locations much smaller than that of their participants. Imagine two shapes, a triangle and a circle, moving around in an animated computer display containing a large, fixed rectangle. At some point, the triangle and circle appear to collide, and then immediately rebound in opposite directions. Most speakers would judge the sentence (22) The circle hit the triangle in(side) the rectangle
to be true if and only if the unique point of impact lies inside the rectangle, even if only a portion of the circle and triangle do so. According to the Davidson–Parsons account, adverbials contribute to the meanings of the sentences they modify by imposing extra conditions on the events over which the sentences in question implicitly existentially quantify. But the above examples suggest that where spatial prepositions are concerned, the extra conditions in question are ones of which we have no grasp, except by reference to the way the event is classified in the unmodified sentence. In particular, it is unclear that we can know where an event is located except with specific reference to the type of event it is.
TSL[v.20020404] Prn:7/05/2002; 13:22
F: TSL5011.tex / p.5 (119)
Prepositions and context
Our conclusions from the above discussion are as follows. The Davidson– Parsons account of adverbial modification succeeds in replicating speakers’ pre-theoretic judgements concerning certain implication relations between event-reporting sentences with and without modification by spatial preposition phrases. In doing so, however, it forces some specific and quite surprising conclusions about the spatial location of events. In particular, the functionsymbol place(e) occurring in the logical forms proposed above seems to have no obvious sense which we can grasp in the absence of a specific classification of the events to which it is applied. Whether this fact constitutes a genuine difficulty for the Davidson–Parsons account, however, is unclear.
. The location of activities and states In the previous section, we concentrated on sentences which report the occurrence of events, and we examined their modification by spatial preposition phrases. In this section, we turn our attention to sentences which report the holding of states. As we shall see, spatial prepositions exhibit rather different behaviour in these cases. Consider the state-reporting sentences: (23) Mary is asleep (24) Mary is happy.
At first sight, 23 and 24 appear to be logically parallel, and indeed are most naturally analysed on Parson’s theory as: (25) ∃s[asleep(Mary, s) ∧ time(s) = ] (26) ∃s[happy(Mary, s) ∧ time(s) = ],
where s is a variable ranging over states. However, adding spatial preposition phrases produces very different results: (27) Mary is asleep in the garden (28) #Mary is happy in the garden.
The symbol # indicates a reading which is best regarded as the result of a re-interpretation, or as being otherwise ‘secondary’. That there is something odd about the interpretation of 28 can be seen by comparing it with 27. Sentence 27 makes a claim about the state Mary is presently in, and, moreover, where that state is located. (We may reasonably assume that any state of Mary’s
TSL[v.20020404] Prn:7/05/2002; 13:22
F: TSL5011.tex / p.6 (120)
Ian Pratt-Hartmann and Nissim Francez
being asleep is roughly coextensive with Mary herself.) Its meaning is therefore accounted for by assigning it the logical form (29) ∃s[asleep(Mary, s) ∧ place(s) ⊆ place(garden) ∧ time(s) = ].
In particular, it is immediate from the assigned logical forms that 27 implies 23. By contrast, sentence 28 cannot easily be interpreted as a claim about Mary’s present state and that state’s location. And its meaning cannot be captured by a logical form analogous to 29. If 28 means anything, it is presumably roughly synonymous with (30) Mary is generally happy when she is in the garden,
a kind of interpretation which appears unavailable for 27. In particular, it is clear that 28, understood in this way, does not imply 24. In fact, most, but not all, state-reporting sentences behave more like 24 than 23: (31) #Mary loves John in the garden (32) #Mary believes in homeopathy in the garden.
By contrast, activity-reporting sentences, just like event-reporting sentences, can be modified by spatial preposition phrases in this way: (33) John followed Mary in the park (34) Mary worked on the paper in the garden (35) Mary thought about Vienna in her office.
In view of these observations, we can now see sentence 27 as an oddity. On the one hand, it passes the usual tests for classification as a state (rather than a process or event); on the other, it admits of localization by a spatial preposition phrase in a way that seems often to be resisted by state-reporting sentences. The explanation for this behaviour may be the close semantic relation between the sentences: (36) Mary was asleep (37) Mary slept
the second of which is unproblematically activity-reporting, and thus capable of spatial location. Other examples of locatable states also appear to be ‘activity’-related, in a relaxed sense:
TSL[v.20020404] Prn:7/05/2002; 13:22
F: TSL5011.tex / p.7 (121)
Prepositions and context
(38) Alice is a banker in London (39) Bob is a down-and-out in Manchester.
In view of the strong semantic links to activities in these exceptional cases, we conclude that state-reporting sentences ordinarily do not admit of location by spatial prepositions in the way that event- and activity-reporting sentences do.
. Spatial restriction Yet it would be incorrect to conclude that state-reporting sentences cannot be modified by spatial prepositions. For example, consider the sentences (40) The road is muddy (41) The road is muddy near the church.
Sentence 40 is a clear case of a state-reporting sentence, and sentence 41 shows that it can be unproblematically modified by a spatial preposition. However, when we come to look at the interpretation of such pairs, analogous treatment to that of event- or activity-reporting sentences would be incorrect. For example, 40 and 41 would be assigned the respective logical forms: (42) ∃s[muddy(road, s) ∧ time(s) = ] (43) ∃s[muddy(road, s) ∧ near(place(s), place(church)) ∧ time(s) = ].
To see that these analyses are wrong, observe that formula 43 implies formula 42, but sentence 41 does not seem to imply sentence 40: a muddy road is a road that is almost everywhere muddy, not one that is muddy in just one place, and on this understanding of muddiness, neither one of the sentences 40 or 41 implies the other. Spatial preposition phrases often modify state-reporting sentences in this way: the unmodified state-reporting sentence claims something about the entire region occupied by the theme of the state; the modified sentence restricts that claim to the region occupied by the specified part of that theme. Here are some typical examples: (44) The field is flooded (45) The lawn is diseased (46) The field is flooded in the corner (47) The lawn is diseased at the end of the garden.
TSL[v.20020404] Prn:7/05/2002; 13:22
F: TSL5011.tex / p.8 (122)
Ian Pratt-Hartmann and Nissim Francez
Sentences 46 and 47 are best paraphrased as (48) The part of the field in the corner (of the field) is flooded (49) The part of the lawn at the end of the garden is diseased.
Hence, a theory which takes unmodified sentences to quantify existentially over states, and spatial preposition phrases to contribute extra conjuncts to the quantified formulas will give incorrect semantics in such cases. With these new examples behind us, it is worth taking some time to reconsider those state-reporting sentences, such as 24, which cannot straightforwardly be modified by spatial preposition phrases. The spatial preposition phrase we considered in connection with this sentence, namely in the garden, picks out a region which we expect to include the space occupied by the relevant participant, namely Mary. By contrast, the spatial preposition phrases considered in connection with sentences 44 or 45 pick out regions which we expect to be included in the space occupied by the relevant participants. And it is easy to see that the former sort of spatial preposition phrases make just as little sense with most state-reporting sentences. To illustrate, the A1 is a road that stretches from London to Edinburgh, and thus lies wholly within Great Britain, but not wholly within England. Consider, then the sentences: (50) The A1 is bumpy (51) The A1 is bumpy in England (52) *The A1 is bumpy in Great Britain.
Following our earlier analysis, we see that 50 states that the A1 is (generally) bumpy over the whole of its extent; 51 states that the A1 is (generally) bumpy over the intersection of its extent with the territory of England; and 52 just sounds strange. The reason for this strangeness is now clear: the job of the spatial preposition phrase in connection with state-reporting sentences is to restrict the predicate supplied by the unmodified sentence to a part of the object to which it would be applied if the spatial preposition phrase were absent; and a spatial preposition phrase picking out too big a region results in a trivial restriction, and therefore in a sentence which can be rejected on general (Gricean) grounds. Thus, the real difference between the two sorts of state-reporting sentences exemplified by 24 and 50 is a matter of spatial homogeneity: being happy is a property possessed at best by Mary as a whole – not by any part of her. Hence, spatial preposition phrases which restrict that happiness to something larger than Mary are redundant, and spatial preposition phrases which restrict it to
TSL[v.20020404] Prn:7/05/2002; 13:22
F: TSL5011.tex / p.9 (123)
Prepositions and context
something smaller than Mary are absurd. Being bumpy, by contrast, is a state that can be possessed by bits of the A1 as well as the whole of it: thus, sentences such as 52, which restrict that bumpiness to something larger than the A1, are still redundant, but sentences such as 51, which restrict it to some part of the A1, make good sense. We might be tempted to summarize our explanation by saying that the behaviour of spatial preposition phrases is conjunctive for event- and activityreporting sentences but mereological for state-reporting sentences. In the former case, a spatial preposition phrase adds an extra conjunct concerning the events or activities over which the sentence existentially quantifies. In the latter case, it replaces an individual involved in some predication with a part of that individual.
. Spatial preposition phrases and context It transpires that the ‘mereological’ behaviour identified in the previous section is a bit more general than the above examples would have us think; for spatial preposition phrases operate in analogous ways on sentences which quantify over objects. For example, in the context of discussing a matrix of numbers, consider the sentences: (53) Every element is negative (54) Every element is negative in the third row.
Of course, being less than zero is not a thing which a number can do in one place rather than another: this state cannot sensibly be said to have a location at all. Rather, what is happening here is that the spatial preposition phrase is limiting the universal quantification in the subject to a smaller context than it would otherwise have: not the whole matrix, but just the third row thereof. Moreover, example 54 illustrates an important connection between what we called the conjunctive and mereological behaviours of spatial preposition phrases. For, to restrict universal quantification to a proper sub-region of its normal context can be seen as a matter of conjoining a suitable condition to some predicate guarding that quantification. Here is another example, this time again involving the state of being happy: (55) The prime minister is happy in Israel, but unhappy in Great Britain.
TSL[v.20020404] Prn:7/05/2002; 13:22
F: TSL5011.tex / p.10 (124)
Ian Pratt-Hartmann and Nissim Francez
This sentence can be interpreted in two ways. On the one hand, it can be used to make a statement about a single prime minister (of some contextually provided political entity), namely, that he is happy whenever he is in Israel and unhappy whenever he is in Great Britain. But it can equally be used to make a statement about two individuals, the Prime Minister of Israel and the Prime Minister of Great Britain, namely, that one is happy and the other unhappy. The location of the happiness, if happiness can be said to have a location, is immaterial here: the job of the spatial preposition is simply to limit the context of quantification within which the definite description is interpreted. These observations suggest that the conjunctive and mereological behaviours contrasted above may not be so different after all. For it may be that states such as the state of being muddy are best thought of in terms of universal quantification over parts: a muddy road is a road that is muddy (almost) everywhere. Thus, ignoring tense-information, sentences 40 and 41 would be assigned the respective logical forms: (56) ∀x[part(x, road) → muddy(x)] (57) ∀x[part(x, road) ∧ near(place(x), place(church)) → muddy(x)]
so that again, the spatial preposition phrase functions so as to conjoin a condition to part of the sentence it modifies. What remains unclear at this point is whether this analysis can be carried out for state-reporting sentences in general. On this question depends the possibility of an account of spatial preposition phrases that works uniformly across all aspectual categories. Our conclusions regarding the modification of state-reporting sentences by spatial preposition phrases, then, are as follows. When spatial preposition phrases modify state-reporting sentences, they do not report the location of any state; rather, they substitute parts for wholes. The parts in question may be parts of some of the individuals mentioned in the unmodified sentence; or they may be parts of an implicit context over which quantification in the unmodified sentence would otherwise range. It is tempting, in the light of the examples we have considered, to attempt a unification of the two ways in which spatial preposition phrases appear to behave, by taking state-reporting sentences to be implicitly universally quantified, where the quantification ranges over parts of the participants in those states. Whether such a unification can be carried through, however, is unclear.
TSL[v.20020404] Prn:7/05/2002; 13:22
F: TSL5011.tex / p.11 (125)
Prepositions and context
. Summary In this paper, we examined the role of spatial preposition phrases as applied to sentences belonging to different aspectual classes, with special reference to the Davidson–Parsons account of adverbial modification. According to this account, simple unmodified sentences existentially quantify over events and states, and adverbial modifiers add conjuncts to the meaning of the unmodified sentence. Our main findings were as follows. For those sentences which we called event-reporting and activity-reporting, the Davidson–Parsons account, though not unproblematical, nevertheless correctly reproduces speakers’ intuitions about the logical relations between modified and unmodified sentences. For those sentences which we called state-reporting, we found that the Davidson–Parsons account failed to capture the effect of modification by spatial preposition phrases. When applied to state-reporting sentences, spatial preposition phrases function so as to replace the participating individuals by those parts of them intersecting the region specified in the complement of the preposition phrase. Finally, we drew some loose parallels between the two ways in which spatial preposition phrases function, noting that spatial preposition phrases often function so as to restrict the context over which quantification would normally range to a part of that context. Whether these loose parallels are manifestations of a unified account, however, remains to be seen.
References Crouch, R. S., & Pullman, S. G. (1993). Time and modality in a natural language planning system. Artificial Intelligence, 63, 265–304. Davidson, D. (1967). The logical form of action sentences. In The Logic of Decision and Action, N. Rescher (Ed.), Pittsburgh: University of Pittsburgh Press. Dowty, D. (1979). Word Meaning and Montague Grammar. Dordrecht: D. Reidel. Hwang, C. H., & Schubert, L. K. (1994). Interpreting tense, aspect and time adverbials. In D. M. Gabbay & H.-J. Ohlbach, (Eds.), Proceedings of the First International Conference on Temporal Logic [Lecture Notes in Computer Science 827] (pp. 238–264). Berlin: Springer Verlag. Ogihara, T. (1996). Tense, Attitudes and Scope. [Studies in Linguistics and Philosophy 58]. Dordrecht: Kluwer. Parsons, T. (1990). Events in the Semantics of English: a Study in Subatomic Semantics. Cambridge, MA: MIT Press. Pratt, I. and Francez, N. (2001). Temporal prepositions and temporal generalized quantifiers. Linguistics and Philosophy, 24 (2), 187–222.
TSL[v.20020404] Prn:7/05/2002; 13:22
F: TSL5011.tex / p.12 (126)
Ian Pratt-Hartmann and Nissim Francez
Steedman, M. (1997). Temporality. In J. van Benthem & A. G. B. ter Meulen (Eds.), Handbook of Logic and Language (pp. 895–935). Amsterdam: Elsevier. Stump, G. (1985). The Semantic Variability of Absolute Constructions. Dordrecht: Kluwer. ter Meulen, A. G. B. (1996). Representing Time in Natural Language. Cambridge, MA: MIT Press. Vendler, Z. (1957). Verbs and times. The Philosophical Review, 66, 143–160. Verkuyl, H. J. (1993). A Theory of Aspectuality: the interaction between temporal and atemporal structure. Cambridge: Cambridge University Press.
Prepositional phrases as noun modifiers in contemporary Hebrew Grammatical, semantic and pragmatic motivations Esther Borochovsky-Bar Aba and Hava Reppen Department of Hebrew, Tel-Aviv University
Introduction In this paper we wish to discuss prepositions in prepositional phrases modifying nouns, such as mumxe le’irak “specialist to (in) Iraq”, mumxe me’irak “specialist from Iraq” etc. Our study will be limited to those phrases the head of which is originally nominal, i.e. it is not the result of a transformation of a verb or another linguistic item. This study may contribute to the study of the noun phrase in Israeli Hebrew, which, until now, mainly described the various phrases (various construct states, adjectival phrases and appositional phrases),1 whereas prepositional phrases as noun modifiers have not been subject to extensive descriptions. The sporadic studies of prepositional phrases (like Ornan 1964; Rosen 1977; Schlesinger 1994) mention some factors motivating their use:2 style,3 syntactic constraints4 and a syntactic-semantic distinction.5 Thus, the uniqueness of prepositional phrases versus other phrases has been mentioned, yet most of the attention was paid to the characterisation of the alternative phrases. We wish to present here preliminary findings which point to the uniqueness of the prepositional phrase as a noun modifier: the prepositional phrase serves more transparently than other phrases to determine the semantic relationship between the head of the nominal phrase and its modifier, and enables us to make more precise distinctions.
Esther Borochovsky-Bar Aba and Hava Reppen
Let us compare, for example the noun phrases: mumxe me Áirak “a specialist from Iraq” and mumxe ’iraki “an Iraqi specialist”. The prepositional phrase (in the first example) defines most transparently the origin of the modified noun, whereas the adjectival phrase (the second one) does not necessarily determine an origin, but denotes a feature or a characteristic connected to this place of origin.6 Such a functional distinction is named by Benveniste (1967) “actual” versus “virtual”. To strengthen this supposition we would like to mention that it is the fact that the adjective denotes a feature or characteristic, and not necessarily an objective actual fact, which makes it possible to use it, and not the parallel prepositional phrase, to denote an evaluating property. For example, the phrase ’ima polaniya “Polish mother” serves to denote particular features of mothers attributed to mothers of Polish descent (such as being preoccupied with feeding their children, being full of expectations from their children, like expecting them to become doctors etc.) and does not mean that this mother actually comes from Poland. (Therefore, the phrase ’ima polaniya “Polish mother” is commonly used to describe mothers who are not from Poland.) The prepositional phrase cannot fill this role because it denotes, unequivocally, some actual relationship. In our current study we will not discuss the differences between prepositional noun phrases and other phrases, but will focus on the differences between the various prepositional phrases differing in their preposition. This raises the question what determines the occurrence of a particular preposition in a modifying prepositional phrase. First, we will discuss cases where the preposition is obligatory and not interchangeable, and afterwards, cases in which the preposition is chosen in accordance with the speaker’s intention to express the various relationships between the head and its modifier in the prepositional phrase. In the cases in which there is an optional choice, the relationship reflected through the preposition may be either semantic (see Sec. 2) or pragmatic (see Sec. 3).
.
The obligatory preposition
Sometimes we find in a prepositional phrase an obligatory preposition, which cannot be changed, and whose appearance is dictated by a number of factors. We will present cases where the preposition is determined by the head of the prepositional phrase (1.1) and cases where it is determined by the modifier (1.2).
Prepositional phrases as noun modifiers in contemporary Hebrew
. The obligatory preposition determined by the head Let us start with cases where the preposition is dictated by the head: (1) hayi-ti xelek me-ha-xavura hazot I-was part from-the-group this ‘I was part of this group.’
Being part of something is by definition tied to the whole, and this tie is normally realized by the preposition min/me/mi7 = from “of ” which is true for many other languages in partitive constructions, and not only for Hebrew. This is true also when in the position of the head we find words denoting quantity as part of a whole: (2) ’exad me-’ede ha-re’iya. one from-witnesses the-sight ‘One of the eye-witnesses.’ (3) ’asarot mi-korbanot ha-’ason . scores from-victims the-disaster ‘Scores of the disaster’s victims.’ (4) shemets me –ha-mevuxa she-’axaza bi. some from-the-embarrasment which-overwhelmed me ‘Some of the embarrasment which overwhelmed me.’ (5) ’arba’im ’axuz mi-yexolet tifkud-am. Forty percents from-ability functioning-their ‘Forty percents of their ability to function.’
Another example of obligatory preposition is one dictated by an abstract noun which by its definition denotes certain semantic relations. Let us examine sentences (6) and (7): (6) ha-pa’ar ben kesem ha-xazon le-ven mimush-o. the-discerpancy between magic the-vision to-between realization-its ‘The discrepancy between the magic of the vision and between its realization.’ (7) ha-keren tishmor ’al shivion ben kol kvutsot The-fund will-maintain on equality between all groups ha’oxlusiya. the-population ‘The fund maintain equality between all groups in the population.’
Esther Borochovsky-Bar Aba and Hava Reppen
In these sentences we find in the head of the prepositional noun phrase an abstract noun which denotes a constant relation between (at least) two entities, and these two entities are realized in two prepositional phrases with the prepositions: ben . . . ’uven . . . = between . . . and between . . . In contrast, in sentences as: (8) ha-’or ben ha-trisim na’asa yoter ve-yoter bahir. the-light between the-shutters became more and-more bright ‘The light between the shutters turned brighter and brighter.’ (9) tenaki bevakasha gam ’et ha-lixlux ben ha-kis’ot. You-clean please also acc the-dirt between the-chairs. ‘Please clean also the dirt between the chairs.’
the prepositional phrase opening with the preposition ben = between denotes the “place” of the light/dirt “between” the shutters/chairs, and the preposition is not dictated by the head, and is not obligatory; it can be replaced by other prepositions such as under/above/behind (the shutters/chairs), thus denoting various locative relations. . The obligatory preposition determined by the modifier Sometimes the obligatory preposition is connected with a particular semantic characteristic of the noun that follows it. For example, a noun denoting classification or categorization of the head, will necessarily appear with the preposition mi = from: see examples (10)–(12): (10) tsmaxim mi-zan kaze nitan le-gadel raq be-xamamot. plants from-species such possible to-raise only in-greenhouses ‘Plants of such species can only be raised in greenhouses.’ (11) ron, ’itonay mi-motsa kurdi, ’azav ’et misrade ha-’iton. Ron journalist from-origin Kurdic left acc offices the-newspaper ‘Ron, a journalist of Kurdic origin, left the newspaper’s offices.’ (12) klavim me-ha-sug she-molix ’ivrim. dogs from-the-kind that-leads blinds ‘Dogs of the kind that leads the blind.’
The nouns zan = species, motsa = origin, sug = kind ets. will only be joined to another noun by mediation of the preposition min = from. It should be mentioned that the entity represented by the noun in the head position is not described by the nouns following the preposition (kind, species, origin), but by
Prepositional phrases as noun modifiers in contemporary Hebrew
the element joining these nouns (kaze = such, kurdi = Kurdic, shemolix ’ivrim = that leads the blind). Contrary to the above nouns which dictate the use of the preposition mi = from, nouns which denote an inherent characteristic8 of the head, appear with the obligatory preposition be = in. Here too the nouns following the preposition do not carry the main content, but are carriers for the adjectives which follow them, presenting the main message:9 (13) xotex daf le-pisot ktanot be-godel zehe. cuts page to-pieces small in-size identical ‘(He) cuts the page into small pieces of identical size.’ (14) brexa be-roxav kaze lo tikanes la-xatser shelaxem. swimming.pool in width this not will-enter to-the-yard yours A swimming pool of this width will not fit into your yard. (15) ’erom lo mushlam shel ’isha be-gil-i nudity not complete of woman in-age-my ‘Incomplete nudity of a woman my age.’ (16) ve-kol na’im be-mivta ’anglosaksy kaved. and-voice pleasant in-accent English heavy ‘And a pleasant voice in a heavy English accent.’ (17) hu ’ana hallo muzar be-signon pariza’i he answered hallo strange in-style parisian ‘He answered saying a strange “hallo” in an affected parisian style.’ (18) rak sorgim dakim be-tsurat lev mafridim bene-nu only bars thin in-shape heart separate between-us le-ven-am. to-between-them ‘Only thin heartshaped bars separate between us and them.’ (19) shalaf-ti mi-sham ’et simlat ha-ktifa be-tseva yayin pulled-I from-there acc dress the-velvet in-color wine. ‘I pulled out the wine colored velvet dress.’
As mentioned before, these nouns which denote an inherent feature of the head’s meaning, appear only with the preposition be = in. In both groups mentioned above, the preposition heading the prepositional phrase is an obligatory non-interchangeable preposition. However, it is worth mentioning that the occurrence of either preposition is not arbitrary and semantic connections can be found to their concrete (spatial) occurrences, having their own meaning: An inherent feature of cohesion (examples 13–19) is indicated by the prepositional
Esther Borochovsky-Bar Aba and Hava Reppen
phrase with be = in, while characterizing an object by virtue of its belonging to some group (examples 10–12) is indicated by mi = from, which basically denotes the place of origin. In other words, even though the preposition is an obligatory non-interchangeable one, and as such does not denote any meaning by itself, the fact that it is one preposition which appears, and not another one, is not arbitrary but semantically motivated.10
. Semantic factors Denoting semantic distictions is the foremost factor which dictates the occurrence of specific preposition in a prepositional noun phrase. These distinctions are best presented in minimal pairs differing only in the preposition, that is to say, the burden of denoting the semantic relationship between the head and the modifier lies on the preposition alone. This relates to those cases where there might be various relationships between the head and the modifier, and it is the preposition that denotes these relationships. Let us look, first, at several examples of known, even trivial differences, denoted by the various prepositions, i.e., mostly in temporal and locative relations:11 (20) hu shalax dgima shel ha-ra’al le-ma’abada meyuxedet he sent sample of the-poison to-laboratory special be-london in-London. ‘He sent a sample of the poison to a special laboratory in London.’
The prepositional phrase with be = in denotes a cohesion between the entity indicated in the head and the location denoted by the modifier. Other locative relations are denoted through other prepositions, such as ma’abada leyad/mixuts lelondon = laboratory near/outside London etc. In the next example the prepositional phrase with be = in denotes a similar relation of cohesion between the head and the modifier, but in the dimension of time, and not of place: (21) ha-hatsagot be-sheva12 meyu’adot li-yladim. The-shows in-seven destined to-children ‘The shows at seven o’clock are for children.’
Other relations of time can be denoted with other prepositions: hahatsagot lifne/’ad sheva. = The shows before/until seven etc.
Prepositional phrases as noun modifiers in contemporary Hebrew
Following is a detailed discussion of a more abstract semantic difference between the prepositions le = to and mi = from, denoting respectively field of expertise and origin: (22) ’amatsia bar’am, mumxe le-’iraq Amatsia Baram, specialist to-Iraq. ‘Amazia Bar’am, specialist in Iraq.’ (23) mumxe me-’iraq specialist from-Iraq ‘A specialist from Iraq.’
in the phrase mumxe le’irak = specialist to (in) Iraq, the first noun denotes a person’s particular occupation, and the prepositional phrase denotes the domain of occupation.13 In the phrase mumxe me’irak = specialist from Iraq, the prepositional phrase denotes the origin of that person. We can conclude that le = to denotes Iraq as the field of expertise of the person denoted in the head of the phrase, namely, a person who continuously deals with Iraq, whereas the mi = from denotes Iraq as the place of origin of that person.14 The different semantic relations in the pairs mumxe le’irak = specialist to = (in) Iraq versus mumxe me’irak = specialist from Iraq, are indicated through the preposition alone. However, the same meaning of prepositions can be identified in cases where only one of them appears, excluding the other preposition. For instance, the relation “the person being occupied in”: “the field of expertise” is denoted in the following examples: (24) nish’al mumxe le-kafe ’ex mezahim kafe tov. we-will-ask specialist to-coffee how identify coffee good. ‘We will ask a coffee specialist how good coffee can be identified.’ (25) she-hi mora le-misxak that-she teacher to-drama. ‘That she is a drama teacher.’ (26) she-hu more le-te’atron that-he teacher to-theater. ‘That he is a theater teacher.’ (27) ha-profesor le-filosofya mitpa’el me’od me-ha-sefer. the-Professor to-Philosophy impressed very.much from-the-book ‘The philosophy Professor is very impressed by the book.’
In the phrases above, the preposition mi is blocked (there are no phrases such as *more mimisxak = a teacher from drama etc.), because the noun cannot be
Esther Borochovsky-Bar Aba and Hava Reppen
used as an origin. However, when the noun can be used as an origin, it can, of course, join the head nouns with the mediation of mi, as in example (28): (28) ’ani makir more mi-natsrat I know teacher from-Nazareth.15 ‘I know a teacher from Nazareth.’
On the other hand, it is possible to see many phrases (see examples 29–30) in which the noun in the head position may accept a prepositional phrase modifier beginning with mi = from,16 but not one beginning with le = to, because the head noun does not denote an occupation: (29) ’aval ha-’emet hi she-’anaxnu yedidim tovim me-’ota ha-branja but the-truth is that-we friends good from-same the-clique ‘But the truth is that we are good friends sharing the same clique’ (30) shne yeladim mi-yerushalayim niftse’u ’etmol. two children from-Jerusalem were-injured yesterday ‘Two children from Jerusalem were injured yesterday.’
So far an emphasis was put on the use of the preposition to denote diverse semantic relations, as, for instance, in the phrase mumxe le’irak = specialist to (for) Iraq – the preposition le = to (for) is necessary in order to denote the “field of expertise”, and if we omit the preposition, this meaning can not be implied. However, in many cases we can decode the semantic relations by way of the lexical meaning of each component of the phrase, without the preposition le = to. We could, for instance, combine the nouns mumxe = specialist and ginun = gardening, or more = teacher and tsiyur = drawing, in a construct states such as: *mumxe ginun = gardening specialist or *more tsiyur = drawing teacher. And yet, we do not find such phrases.17 The appearance of the preposition signals therefore the existence of some semantic relations, which could be decoded even without it.18
. Pragmatic factors Prepositions very often serve to denote differences in the pragmatic dimension. Here we wish to present some pragmatic issues indicated by the preposition, such as nuances in the points of view of the addresser, some emphases made by the addresser or presuppositions shared by the addresser and the addressee.
Prepositional phrases as noun modifiers in contemporary Hebrew
. be = in versus mi = from – “integration” versus “detachment” Let us look at examples (31) and (32): ha-tsva’it hexlit kanir’e lekatser (31) mishehu ba-rabanut19 someone in-the-rabbinate the-military decided probably to-shorten halixim. procedures ‘Someone in the military rabbinate probably decided to shorten procedures.’ (32) mishehu me-ha-rabanut hatsva’it hexlit kanir’e Someone from-the-rabbinate the-military decided probably lekatser halixim. to-shorten procedures ‘Someone from the military rabbinate probably decided to shorten procedures.’
Both sentences are identical in their meaning and their truth value, and the difference between them is probably pragmatic. They both speak of the violation of an order by some person in the military rabbinate. In the first sentence there is an emphasis on the fact that this person belongs to the system, is an integral part of it (this may imply that this person’s misdeeds reflect on the entire system). It seems that the preposition be = in, which in its basic temporal and local aspects denotes cohesion, coherence, synchronization, enhances this nuance of cohesion between the single person and the system. The preposition mi = from, that basically denotes origin presents here the person as a single individual being set apart from the system (and that is why in the sentence with mi = from it is not implied that the whole system is involved). In the following example in which the prepositional phrase opens with be = in, we can sense the same cohesion between a person and the system he belongs to: (33) shelbi mekave she-yatsliax lehasir ’et hitnagdutam Shelbi hopes that-he-will-succeed to-remove acc objection-their shel bexirim ba-pentagon of high-officials in-the-pentagon. ‘Shelbi hopes to overcome the objection of high Pentagon officials.’
In this sentence too, the replacement of be = in by mi = from will bring about the change mentioned above.
Esther Borochovsky-Bar Aba and Hava Reppen
However, sometimes such a replacement is prevented, because in a certain context the cohesion of a person to the system it belongs to is built in. (34) hu ’avad ke-rofe be-kupat xolim.20 he worked as-physician in-fund-patients ‘He worked as a physician in a health fund.’
The verb ’avad = worked + the preposition ke = as (’avad ke = worked as) denotes doing some work within an organization, which makes it impossible to detach the person from that system. In other contexts the preposition be = in can be replaced by the preposition mi = from: (35) rofe klali be/mi-kupat xolim hursha. physician general in/from-fund patients was-convicted. ‘A general physician in/from the health fund was convicted.’
Similar examples are (36) and (37) below: (36) betor katsin zutar ba-tsava hu haya yoter mutslax. as officer petty in-the-army he was more successful ‘As a petty officer of the army he was more successful.’
In (36), the replacement by mi = from is prevented, probably due to the occurrence of the preposition betor = as, the function of which is to present the noun following it as having a function in a system. But in a different context mi = from is possible: (37) ’emesh hofi’a ba-televizya katsin zutar be/mi-tsva ’artsot last.night appeared in-the-T.V. officer petty in/from-army states ha-brit the-union. ‘A petty officer of the U.S. Army was on television last night.’
In other cases, however, the prepositional phrase starts with mi = from and excludes be = in: (38) ha-dover lo hitir le-mishehu me-ha-rabanut the-spokesman not allowed to-someone from-the-rabbinate ha-tsva’it lehishtatef. the-military to-participate. ‘The spokesman did not allow someone from the military rabbinate to participate.’
Prepositional phrases as noun modifiers in contemporary Hebrew
In this sentence it would sound odd to have the prepositional phrase with be = in, because in this context there is a focus on one individual, while the system is only presented as a background. The preposition be = in is also blocked when the head noun is a proper noun: (39) ’amatsya bar’am me-’universitat xayfa . . . Amatsya Bar’am from-university Haifa ‘Amatsya Baram from Haifa University . . . ’
The proper noun denotes the person as an individual and not as a part of a system; therefore it cannot be specified by a complement opening with the preposition be = in even when the complement denotes the system in which the person has a task. . be = in versus le = to – belonging to a system versus collaboration with others in that system The nouns in the head position studied here are such as [xaver] = friend/colleague, [shutaf] = partner. These nouns, by definition, imply the contact to another person. A friend has to be a friend of someone,21 etc. The interpersonal contact belongs to a certain domain, and this domain could be realized by the preposition le = to, as in examples (40)–(43): (40) nosea ’im xaverim laklafim22 shel ha-horim la-yam. going with friends to-the-cards of the-parents to-the-beach ‘Going to the beach with his parents cards companions.’ (41) ’ani hayiti bat.hazug shela lelevayot. I was companion her to-funerals ‘I was her companion for funerals.’ (42) ’ulay ’afilu lo ’efxad she-yesh lanu shutaf lasod. maybe even not I-will-dread that-have us partner to-the-secret. ‘Maybe I will not dread the fact that we have someone to share the secret with.’ (43) takir, zot xavera sheli la’avoda. meet, this friend mine to-the-work ‘May I introduce my colleague from work.’
The prepositional phrases with le = to define the domain in which the “partnership” is being realized. However, the domain of occupation and the system
Esther Borochovsky-Bar Aba and Hava Reppen
that a person belongs to are very often realized by the prepositional phrase with be = in (as shown also in 3.1). Let us look at examples (44)–(47): (44) hu shutaf be-’asakim rabim. he partner in-businesses many ‘He is a partner in many businesses.’ (45) hem xaverim be-’amuta lema’an barak. they members in-group for Barak ‘They are members of a group supporting Barak.’ (46) hayi-ti xaver be-mo’adon ski shel ’itona’im was-I member in-club ski of journalists. ‘I was a member of a journalists’ ski club.’ (47) xaverim be-direktoryon dor le-dor members in-board “Dor le-Dor” ‘ “Dor le Dor” board members.’
It seems that the speaker’s choice whether to denote the domain by be = in or by le = to is not arbitrary, but is pragmatically motivated: in the phrases with le = to such as xaverim la’avoda = “colleagues to the work” = “work colleagues”, it is stated that the persons are part of a certain system, but the emphasis lies on the interpersonal contact between them, while phrases with be = in emphasize the belonging of the person to the system, although there are other members in a similar position in this group.23 . mi = from denoting a previous source of knowledge In the following examples the prepositional phrase with mi = from does not denote the origin of the entity (be it a person or an object) of the head noun, as may seem at first glance, but points to some previous piece of information concerning this entity; information known both to the addresser and the addressee and thus constitutes a kind of origin. Yet it is not the origin of the entity itself, but it concerns the previous knowledge shared by the addresser and addressee. In order to explain our claim, each of the following examples will be accompanied by such an alleged ‘piece of information’ (in parenthesis), which can be implied from the specific context. Clearly, it is only one possible implication. (48) rina me-ha-pa’am ha-kodemet lo tavo hayom. Rina from-the-time the-former not will-come today
Prepositional phrases as noun modifiers in contemporary Hebrew
‘Rina from the former meeting will not come today.’ (Rina who was here the other day, and we both know it). (49) u-leyad-o ’omed arye . . . gavoa ve-xagigi ba-xalifa and-near-him stands Arye . . . tall and-solemn in-the-suit me-’etmol. from-yesterday ‘And next to him stands Arye. . . tall and solemn in the suit from yesterday.’ (he wears the same suit he wore yesterday; I the speaker have seen him in it; and maybe the addressee saw it or heard about it as well). (50) yavo’u ’el-ay ha-xaverim me-ha-klafim. will-come to-me the-friends from-the cards. ‘My card companions will come to me.’ (those friends with whom the addresser normally plays cards with, and the addressee knows about it). (51) ha-mexabel me-’apropo. the-terrorist from-Apropos. ‘The terrorist from Apropos.’ (the terrorist who attacked the Apropos restaurant and we remember him from that event). (52) ’ata zoxer ’et ha-bayit mi-kfar saba? you remember acc the-house from-Kfar Saba? ‘Do you remember the house from Kfar Saba?’ (the house both the addresser and the addressee saw together in Kfar Saba and wanted to purchase).24
. Conclusion We have shown the unique use of prepositional phrases as noun modifiers in Hebrew. We have also attempted to show different factors which dictate the use of a particular preposition, excluding others. There are cases where the speaker has no option of choosing the preposition, as it is dictated by rigid grammatical rules, while in other cases the preposition is chosen in order to express certain semantic relations, in accordance with the various meanings it denotes. Finally we have shown how the choice of prepositions can be determined by pragmatic motivations of the speaker. All these factors do not necessarily work separately, but are very often combined to affect the use of a particular preposition in a prepositional phrase: an
Esther Borochovsky-Bar Aba and Hava Reppen
obligatory preposition is not always an arbitrary one; having a meaning does not prevent it from being chosen for pragmatic uses and so forth.
Notes . We will mention several papers that deal with various kinds of noun phrases: Ornan (1964), Berman (1978), Borochovsky (1986), Berman-Ravid (1987), Taube (1991), Rosenhois (1989), Schlesinger et al. (1996), Shatil (1997). . Ornan’s research presents prepositional phrases as products of two different structural transformations: transformation of a verb into an action noun and ommission of the verb, but Ornan does not deal with the functions of prepositional phrases in comparison to the function of other noun phrases. . Schlesinger (1994: 56) points at uses where, in his opinion, prepositional phrases are stylistic variants of the other noun phrases, particularly of construct states. . Rosen (1977: 175–176) discusses a syntactical constraint which motivates the use of a prepositional phrase with le = to instead of using the construct state. . Rosen (1977: 180–181) points at cases where there is a syntactic and semantic difference between the prepositional phrase and the construct state. . See the distinction of Rosen in previous footnote no. 5, as may also be studied in the works of his students Taube (1991) and Shatil (1997). . These three versions are allomorphs of the same Hebrew preposition equivalent to the English “from”. From now on we’ll use mi to represent all three. But see Tobin’s article in this book for a different opinion. . We have two ways of describing an object: one by describing it as belonging to a group of similar objects (by the preposition mi = from), the other by characterizing it through an inherent feature (by the preposition be = in). On this distinction and other means of characterization see Borochovsky and Trommer (1993). . For a similar phenomenon see Taube (1986: 78). . These findings agree in a way with approaches concerning the semantics of prepositions which give a theoretical priority to the “spatial” meaning of the prepositions (sometimes identified as their prototypical meaning) and relate their more abstract or extended meaning to the former (see for instance: Anderson (1971), Bennett (1975), Lakoff (1987), Kemmer & Bat-Zeev Schyldkrot (1995), Lindstromberg (1998)). In Hebrew we can see in Reppen (1995) a semantic motivation for an obligatory preposition. See particularly pp. 13–16. This approach is being challenged, however, by Cadiot (2000) who claims that it is pointless to make such a distinction between the “spatial” literary meaning and the extended one. For a similar view see Visetti (2000). . These uses of the prepositions have been widely dealt with in the literature (for example, see Leech 1969; Bennett 1975; Talmy 1983).
Prepositional phrases as noun modifiers in contemporary Hebrew . It can also be said in the same meaning: hahatsagot shel sheva = The shows of seven o’clock. . It is interesting to mention here that there are also phrases in which the modifier denoting the domain of occupation opens with the preposition be = in and there is a slight semantic difference between these and the above mentioned opening with the preposition le: The modifier opening with le presents the occupation as having some nuance of an official approval. For example the expression mumxe leginun = specialist to (in) gardening – is used to denote a person who seems to have an official title for this profession while the expression mumxe beginun = specialist in gardenning is used to denote someone who works in gardening as a hobby, he has some knowledge about it, but it’s not his official title. This difference accounts for the fact that conventional professions such as more letsiyur = teacher to drawing (drawing teacher); profesor lebalshanut = professor to (of) linguistics – occur always with the preposition le and not be. . It should be mentioned that these two expressions differ from each other in another aspect, in the order the immidiate constituents of the lexeme mumxe = specialist join the prepositional phrases le’irak = to Iraq and me’irak = from Iraq. The lexem mumxe = specialist contains two constituents: ‘person’ and ‘expertise’. In the expression mumxe le’irak = specialist to = (for) Iraq the constituents’ order is: ‘person’ + (‘expertise’ + ‘to Iraq’) while in the expression mumxe me’irak = specialist from Iraq the order is: (‘person’ + ‘expertise’) + ‘from Iraq’. This difference is formally expressed by the fact that the prepositional phrase le’iraq = for Iraq cannot function as a predicate, while the prepositional phrase me’irak = from Iraq can be placed as a predicate: the specialist is from Iraq. For the use of this criterion to analyze various expressions see Rosen (1977a: 89–91). . In addition to the fact that not every lexeme can denote an origin, it seems that the origin, contrary to the occupation, has to be realized by a definite noun, because the origin is something definite while an occupation is something general. . In fact, quite a number of nouns may be complemented by a prepositional phrase beginning with mi = from, denoting origin, belonging etc. . Although in these examples there are no syntactic or other constraints on such phrases. . Reppen (1999) who deals with various phrases used as names of organizations, shows that whenever the noun modifier denoting a domain of occupation is an action noun, it joins the head by mediation of the preposition le e.g. ’aguda lehitnadvut = society to volunteering (voluntary society) and never directly to create a construct state, such as *’agudat hitnadvut = voluntary society. . ba is the portmanteau form of the preposition be when it includes the definite article ha = the, i.e. whenever the noun following the preposition is definite. . In this sentence the phrase be-kupat xolim (in a health fund) can be considered either as a noun modifier (as it is interpreted here) or as a complement of the verb or of the whole sentence. . Rosen (1977a: 149–154) calls such a relation “inalienable belonging” to differentiate it from “possession” or “ownership. Bartning (1992), who discusses various possibilities of decoding the semantic relations between nouns connected by the French preposition de, claims (p. 169) that when the first noun of the noun phrase denotes a family relationship
Esther Borochovsky-Bar Aba and Hava Reppen
e.g., l’epouse de Jean = Jean’s wife, or interpersonal relations such as l’ami de Paul = Paul’s friend, there is only one possibility of decoding: a relation to another person. . la is the form of the preposition le when it includes the definite article ha = the, exactly like ba/be, which was mentioned above. . Phrases such as ’axim le-tsara = brothers to trouble = people sharing the same problem – present a metaphoric use, where there is no conventional social “system” but some other common issue between them. Thus, the meaning of contact between the persons is retained and the preposition is le = to as expected. . Bartning (1992: 184) discusses similar cases in which the relation between the nouns in the phrase is decoded by the context and not by the meaning of the nouns themselves. For example: la jeune fille du bac = The girl from the ferryboat. She explains that relying on the context or on the story one can give the paraphrase: The girl whom the Chinese had met on the ferryboat. It is important to note here that Bartning after Cadiot (1989) thinks that “de” belongs to the so called “empty prepositions” which do not have a meaning of their own. Cadiot (1989: 27) claims that these prepositions do not encode the relation between the nouns, but transfer it either from the context or from the meaning of the nouns. In our discussion we claim that the Hebrew preposition mi = from, which in its most concrete occurrence denotes a place of origin, also denotes the origin of our acquaintance with an object.
References Anderson, J. M. (1971). The Grammar of Case: Towards a Localistic Theory. London: Cambridge University Press. Bartning, I. (1992). La préposition de et les interprétations possibles des syntagmes nominaux complexes Essai d’approche cognitive. LEXIQUE 11, 163–191 P.U.L. Bennett, D. C. (1975). Spatial and Temporal Uses of English Prepositions: An Essay in Stratificational Semantics. London: Longman. Benveniste, E. (1967). Fondements syntaxiques de la composition nominale. B.S.L.62.1: 15– 31 and in Problemes de linguistique générale, Tome II: 145–162. Berman, R. A. (1978). Modern Hebrew Structure. Tel-Aviv: University Publishing Projects. Berman, R. A., & D. Ravid (1987). ’al dargat hamilun shel tserufe smixut (On the degree of lexicalization of construct state phrases). Balshanut ’ivrit xofshit, 24, 5–22 (Hebrew). Borochovsky, E. (1986). Herarxiat haleva’im haba’im ’axare shem ha’etsem (The Heirarchy in Post Nominal modifiers). Leshonenu, 50, 103–118 (Hebrew). Borochovsky, E., & Trommer P. (1993). ’iyun semanti-pragmati bayaxas hasemanti hamexune hefex (Semantic-Pragmatic Study of the Sense Relation Called Opposition). Leshonenu, 57, 215–250 (Hebrew). Cadiot, P. (1989). Dimensions de la préposition. Travaux de linguistique et de philologie, XXVII, 57–74. Cadiot, P. (2000). Schematics and Intentionality. A paper given at “Prepositions” conference, at The University of Haifa, June 2000.
Prepositional phrases as noun modifiers in contemporary Hebrew
Kemmer, S. E., & Bat-Zeev Shyldkrot, H. (1995). A Reexamination of the ‘empty prepositions’ à and de in French. In E. Casad (Ed.), Cognitive Linguistics in the Redwoods (pp. 347–388). Berlin: Mouton de Gruyter. Lakoff, G. (1987). Women, Fire and Dangerous Things: What categories reveal about the life of the mind. Chicago and London: University of Chicago Press. Leech, G. N. (1969). Towards a Semantic Description of English. London: Longmans. Lindstromberg, S. (1998). English Prepositions Explained. Amsterdam: John Benjamins. Ornan, U. (1964). Hatserufim hashemaniyim bilshon hasifrut ha’ivrit haxadasha, beyixud ’al pi ha proza shel X. N. Bialik (The nominal phrase in Modern Hebrew with particular reference to the prose of H. N. Bialik). Ph.D. dissertation, Jerusalem: Hebrew University (Hebrew). Reppen, H. (1995). Hatnayot lehofa’ato shel tsayan yaxas mutsrax ba’ivrit bat yamenu (Conditions for the Occurrences of a Governed Case-Relation Marker in Contemporary Hebrew) Ph.D. dissertation. Tel-Aviv: Tel-Aviv University (Hebrew). Reppen, H. (1999). Ma ben ‘histadrut rof ’e hashinayim’ uvein ‘histadrut lirfu’at shinayim’: ’iyun taxbiri bishmot ’irgunim ba’ivrit hayisre’elit (‘Dentists Association’ versus ‘Association for Dentistry’: A syntactical Study of Organizations’ names in Israeli Hebrew). Leshonenu, 62, 317–332 (Hebrew). Rosen, H. B. (1977). ’ivrit Tova (Good Hebrew). Jerusalem: Kiryat sefer (Hebrew). Rosen, H. B. (1977a). Contemporary Hebrew. The Hague: Mouton. Rosenhois, J. (1989). Sharsheret smixuyot ba’ivrit haxadasha (a Chain of Construct states in Modern Hebrew). Leshonenu, 53, 93–106 (Hebrew). Shatil, N. (1997). Kvutsot Shmot-’etsem Ba’ivrit Hameduberet (Noun Phrases in Spoken Hebrew). Ph.D. dissertation. Tel-Aviv: Tel-Aviv University (Hebrew). Schlesinger, Y. (1994). Prakim betoldot halashon ha’ivrit (History of the Hebrew Language 11). Tel-Aviv: Ha’universita Haptuxa (Hebrew). Schlesinger, Y., Ravid, D., & Sarel, Z. (1996). Tipuax hahavxana hamishlavit baxinux haleshoni behaxsharat morim (Stimulating a Distinction between Registers in Teachers’ Language Education). Tel-Aviv: Mofet (Hebrew). Talmy, L. (1983). How language structures space. In H. Pick & L. Acredolo (Eds.), Spatial Orientation: Theory, Research, and Application (pp. 225–282). New York: Plenum Press. Taube, D. (1986). Tserufe shem ’etsem1 + shem ’etsem2 – i uvxinat hamaratam betserufe smixut makbilim (Nounphrases consist of noun1 + (noun2 + suffix i), and their substitution with parallel construct states phrases), MA Thesis. Jerusalem: The Hebrew University (Hebrew). Taube, D. (1991). ’al tserufe shem ’etsem + shem to’ar gzur shem ’etsem ba’al hasofit –i ba’ivrit bat yamenu” (“The Phrases Consist of Noun + Adjective Derived from Noun by the Derivational Suffix –i in Modern Hebrew”). In M. Goshen-Gotshtein, S. Morag, & S. Kogot (Eds.), Shay Lexayim Rabin (pp. 117–132). Jerusalem: Akademon (Hebrew). Visetti, Y. M. (2000). Three spaces and one semantic theory. A paper given at “Prepositions” conference, at The University of Haifa, June 2000.
The Hebrew prepositions mi-/min ‘from, of ’* Same or different? A sign-oriented approach Yishai Tobin Department of Foreign Literature and Languages Ben Gurion University, Be’er Sheva
The thesis I am going to sustain is not one that would surprise the man on the street. Tell him that if two ways of saying something differ in their words or arrangements they will also differ in their meaning, and he will show as much surprise as if you told him that walking in the rain is conducive to getting wet. Only a scientist can wrap himself up in enough sophistication to keep dry under these circumstances. (Bolinger 1977: 1)
.
Introduction
This paper presents a sign-oriented analysis (based on the concepts of invariance, markedness, distinctive feature theory, and iconicity) that explains the semantic distinction between the Hebrew prepositions mi-(me-)/min ‘from, of, more than; since’ that traditionally have been considered to be: (a) synonyms, or: (b) phonetically determined abbreviations, contracted forms, or prefixal alternates (mi-/(me-)) versus a full, separate, or independent morpheme (min), because: (c) they all perform the same syntactic or grammatical function(s) (e.g., Goshen Gotestein et al. 1977; Reppen 1995; A. Rosén 1965; Wallenrod & Aaroni 1949; Wittenberg 1977; Yalin 1942) as illustrated in example (1):1 (1) hu barax min ha-bayit / me-ha-bayit. he fled from the-house / from-the-house ‘He ran away from home.’
Yishai Tobin
Furthermore, the prefix mi-, unlike the free form min, can also be: i.
inflected for person number and gender: (mimeni ‘from-me’; mimxa/mimex ‘from you’ (masc./sg., fem./sg.); mimenu ‘from him’; mimena ‘from her’; mimenu/me-itanu ‘from us’; mikem/miken ‘from you’ (masc./pl., fem./pl.); mehem/mehen ‘from them’ (masc./pl., fem./pl.);2 and ii. combined with other independent prepositions or other words to form compound prepositions: me-al ‘from-on’ = ‘above’; mi-taxat ‘from-under’ = ‘beneath’; me-ever ‘from-across’ = ‘beyond’, or other compound forms: me-asher ‘from that’ = ‘than’; mi-ma-she ‘from-what-that’ = ‘than’ used in comparative constructions. . Traditional and neo-traditional approaches
The dictionary (Alcalay 1963) lists mi-(me-) and min separately but refers to mi- as (“abbreviated from min”) with the following meanings: ‘from, of, more than; since’ without stipulating a part of speech (p. 1152). The dictionary entry for min, on the other hand, lists min as a “preposition” but does not mention mi- (although it includes idioms and expressions containing both inflected and uninflected mi-) with the following meanings: ‘from, of; out of; since; because; than’ (pp. 1367–1368). Another dictionary (Wittenberg 1977: 575) discusses the “letter mem” which performs several functions among them: (1) locative mem: mi-yaar ‘from a forest’; (2) the mem of cause: hismik mi-busha ‘blushed with shame’; (3) mem of agency: muksam me-ha-noem ‘he is spellbound by the speaker’; (4) the mem of comparative and superlative degrees: leah yafa mi-neomi ‘Leah is more beautiful than Naomi’. Wittenberg (1977: 644) defines min as “a preposition which contracts into “mem”, see entry for mem;” followed by a list of idioms and expressions containing both min and mi-. Berman (1978: 122–124) goes beyond the traditional analyses by establishing three morphological classes of prepositions expressing case relations: (a) prefixes, (b) separate words, and (c) either prefixes or separate words (including mi-/min). She then presents the “complicated restrictions on when the ‘separate word’ form of min can be used,” claiming that it occurs freely mainly in “adverbial” types of expressions, where min has the literal sense of ‘from’ in the sense of direction ‘away from’. She further states that in this sense it can alternate with its prefixed counterpart mi-, but only the enclitic form is possible in instances of “government” (where the verb taking a Patient NP governs a specific preposition). Berman provides three examples (ex. 1 above) and:
The Hebrew prepositions mi-/min ‘from, of ’
(2) hu lakax et ze ?min ha-yeled / me-ha-yeled he took OM this from the-boy / from-the-boy (OM=Object Marker) ‘He took it from the boy.’ (3) hu paxad *min ha-raayon / me-ha-raayon he feared from the-idea / from-the-idea ‘He was afraid of the idea.’
The main disadvantage of this claim concerning “instances of government”, however, is that it can easily be contradicted by counter examples such as example (4) taken from a best selling novel/murder mystery: (4) . . . ve-nehena min ha-maxshava she-yuxal lirot et . . . and-enjoyed from the-thought that-will be able to-see OM ha-shinuim ha-kalim she-taaseh doktor neidorf the-differences the-light that-will do doctor Neidorf ba-maamar she-yitparsem in-the-article that-will-be-published (Gur 1991: 7) “. . . and he enjoyed the thought that he will be able to see the slight differences that Dr. Neidorf will make in the article that will be published.”
Glinert (1989) discusses mi- (without mentioning min) for the following grammatical functions: i. ii.
iii. iv.
v.
vi.
as a preposition linking quantifiers to their nouns (shisha me-ha-nerot ‘six of the candles’) (Glinert 1989: 72–73); or as a partitive in general, particularly in its inflected form, in a wide variety of constructions introducing relative or main clauses containing superlative, predicate, subject, oppositional, parenthetical, and attributive functions (all of which follow the general pattern above and can appear either with mi- or with min) (Glinert 1989: 75–77); as a ‘case preposition’ for indirect objects that cannot undergo passivization (Glinert 1989: 144); or as a governed preposition for verbs indicating “a psychological effect on someone” (zuaza mi- ‘shocked at’; hushpa mi- ‘influenced by’; neelav mi‘offended by’; hufta mi- ‘surprised at’) (Glinert 1989: 145); as part of a list of prepositions that “introduce objects and adverbials” (Glinert 1989: 150) giving an example of an adverbial use of min (ratsnu min he-xatser el ha-masait ‘We ran from the yard to the truck’) (Glinert 1989: 154) (for which mi-/me- is also possible (Y.T.)); as a preposition governing indirect objects governed by verbs which: (a) denote “sources of experience” (paxad ‘fear’; salad ‘dislike’; nehena ‘en-
Yishai Tobin
joy’ (cf. (counter)example (4) with min above); hitpael ‘be impressed with’; mabsut ‘be glad at’; hityaesh ‘despair with’; or (b) “with most verbs of distancing” (nizhar ‘beware of ’; niftar ‘be rid of ’; histayeg ‘have reservations about’; histir ‘conceal from’; raxok ‘distant from’); or (c) denote “sources of things or information” (kibel ‘receive’; lamad ‘learn’; shama ‘hear’) (Glinert 1989: 161); or (d) “miscellaneous verbs” (bikesh ‘ask (for)’, darash ‘require (of someone)’); vii. as a “governed preposition” which is either omitted before finite clauses or retained (or not) before infinitives (Glinert 1989: 205); viii. or as a comparative marker in various kinds of comparative constructions (Glinert 1989: 216–217) alluded to above in compound constructions with mi-: yoter X mi-Y ‘more X than Y’; X yoter me-asher Y ‘X more than Y’; X yoter mimeni ‘X more than-I’; yoter X mi-she Y ‘more X than Y’; X mi-y ‘X-er/more X than Y’. Reppen (1995, 1996) presents the most detailed study of min as a “governed case-relation marker” and its “prefixal alternants mi-/me-” which tend to overlap in her study. Her most basic claim is that: “The preposition has three alternants: min as a single word is restricted to environment (sic) immediately preceding the definite article ha- where it is optional. It also appears as an enclitically prefixed element mi or me, the choice of which is determined by phonological conditions.” (Reppen 1995: iv–v). The use of min (or alternatively mi-/me-) is then described for various message types denoted as a temporal adjunct, as a marker of locality, as a governed preposition as part of verb valency; as a denoter of emotional states, as a point of departure for verbs denoting states or coming into states with animate subjects, and other uses or message types in contrast with other prepositions or prepositional phrases such as le ‘to’ or al yedei ‘by, by means of ’. Reppen takes a Jakobsonian view of invariant or core meaning for min as a ‘source’ or ‘beginning point’ similar to our own, but, like in all the other studies mentioned above, she does not postulate the possibility that there may be a semantic distinction between min versus mi-/me- which might explain the choice of one over the other in all of their shared message types, functions, and uses. . The sign-oriented approach The present semantic analysis will be based on the following sign-oriented hypotheses:
The Hebrew prepositions mi-/min ‘from, of ’
a. the Hebrew prepositions mi-/min ‘from, of, more than; since’ differ in their monosemic, invariant meanings, or signifiés (e.g., Bolinger 1977; Saussure 1913/1959; Tobin 1990; and Waugh & Rudy 1991); b. this difference in meaning is expressed by a marked distinctive semantic feature in the sense originally postulated by the Prague School: e.g., Andrews and Tobin (1996), Jakobson (1975), and Tobin (1988); c. this marked distinctive feature (Semantic Integrality) (e.g., Tobin 1990, 1994) is based on the perception of entities in continuous versus discontinuous space, time, or existence in the sense originally associated with the school of the Psychomechanics of Language (Guillaume 1945, 1971; Hirtle 1982; Wickens 1991); d. according to this analysis: the independent morpheme min is unmarked in Hebrew and indicates a relationship between the source of an entity or entities and the entity or entities themselves viewed either in continuous or discontinuous space, time, or existence, while the dependent morpheme mi- is marked for Semantic Integrality and therefore indicates a relationship between the source of an entity or entities and the entity or entities themselves which can be potentially perceived individually in discontinuous space, time or existence, but must now be viewed in continuous space, time or existence; e. these signs are part of a semantic system in the sense of the Columbia School (e.g., Contini-Morava 1989; García 1975; Gorup 1987; Huffman 1996; Kirsner 1979; Reid 1991). The data will be analyzed on the microlevel: i.e., sentences in context as well as on the macrolevel: i.e., in the form of semiotic-oriented discourse analyses originally inspired by Diver (1969) and further refined in the concepts of word systems and communicative strategies (Aphek & Tobin 1988, 1989), and in the ‘from sign to text’ and ‘from text to sign’ approaches developed in Tobin (1989, 1990, 1993, 1994).
. Semantic Integrality One of the most fundamental problems in linguistics in general and in the study of language universals in particular is the distinction between what is extralinguistic versus what is linguistic, or, in other words: which universals are part of language per se and which are reflections of universal principles of cognition, perception, or communication reflected in diverse ways in different
Yishai Tobin
Alternative ways of perceiving entities in a spatio-temporal-existential cline a + b = a + b (discontinuous space, time, or existence] or a + b = [ab] (continuous space, time, or existence] semantic integrality: the perception of potentially discontinuous entities viewed in a continuous set Figure 1. Semantic Integrality.
languages (García, van Putte & Tobin 1987). Indeed, this question of language universals is one of the key issues dividing traditional and neo-traditional sentence-oriented as well as Saussurian, semiotic, or sign-oriented approaches to language (Contini-Morava & Sussman Goldberg 1995; Contini-Morava & Tobin 2000; Klein-Andreu 1983; Reid & Otheguy in press; Tobin 1989, 1990). The semantic feature relevant to the present analysis, Semantic Integrality, has been previously postulated as a semiotic feature of human cognition and perception (Tobin 1990). The marked feature Semantic Integrality is based on the assumption that there are two alternative ways of perceiving a plurality of entities in space, time, or existence either as discrete entities (a + b = a + b) or as potentially discrete entities perceived as part of a continuous set (a + b = [ab]) as illustrated in Figure 1. . The spatio-temporal-existential cline Semantic Integrality may be viewed as a marked distinctive feature that can be used to explain linguistic concepts both in the grammar and the lexicon in a unified and systematic way that reflects the human perception of entities in space, time, and existence. Indeed, there may very well be a universal semantic development of linguistic forms on a spatio-temporal-existential cline going from the most concrete spatial messages to the more abstract temporal, to the most abstract existential kinds of messages: (a) existence in space or in a place (e.g., in the room), (b) existence in time (e.g., in the morning), (c) abstract existence (e.g., to be in trouble, to be in the in-group) (Tobin 1990: Ch. 3). This spatio-temporal-existential cline has been considered a cognitive universal (e.g., Givón 1979; Traugott 1978; Wierzbicka 1972, 1980). It has been applied to the historical development of language in general and in the development of pidgins and creoles in particular (e.g., Traugott 1975); various
The Hebrew prepositions mi-/min ‘from, of ’
word classes such as copulas, prepositions, and particles (Aphek & Tobin 1988; García, van Putte & Tobin 1987; Traugott 1975, 1978); first language acquisition (Traugott 1974; Clark 1973); the origin and development of idioms in general (Pike 1976) and the development of phrasal verb idioms in particular (Makkai 1972). These studies provide additional evidence for Kronasser’s Law (Kronasser 1952; Kova´cs 1961) which claims that semantic change tends overwhelmingly to move from the concrete to the abstract. . The analysis: Invariance, markedness and iconicity The distinctive feature of Semantic Integrality may provide a means to better understand the connection between human cognition and perception and language and how this connection may be realized differently in various language systems.3 These analyses were based on the Jakobsonian notion of markedness as it is used to distinguish between opposed members of a lexical or grammatical system (Andrews 1990; Battistella 1990, 1996) (the last reviewed in Tobin 1998). The asymmetric value relationship of the concept of binary markedness developed by the Jakobsonian-van Schooneveldian School of linguistics may be summarized as follows for the distinctive feature of Semantic Integrality: a. The unmarked (U) member of an opposition is neutral. It makes no claim with regard to Semantic Integrality and is less complex cognitively. It means: perceive entities in continuous or discontinuous space, time, or existence. In the present system the Hebrew prepositions mi-/min indicate a relationship between the source of entities and the entities themselves. In this system the independent morpheme, min is unmarked. It allows for both potential readings, (a + b = a + b) or (a + b = [ab]) but favors the first discrete or discontinuous perception of the source of entities and the link with the entities themselves. b. The marked (M) member of the opposition makes a specific claim regarding Semantic Integrality. It is more complex cognitively. It means: perceive potentially discontinuous entities in continuous space, time, or existence. In the present system the Hebrew prepostion, the bound morpheme mi- is marked: the only reading possible is the continuous reading: (a + b = [ab]). Therefore the marked form mi- is the most efficient choice for indicating a continuous collective perception linking a source of entities to those entities themselves viewed within a cognitive set.
Yishai Tobin
The unmarked member of an opposition which makes no specific claim concerning the absence or the presence of a distinctive semantic feature is usually more flexible and open-ended than the marked member of an opposition. The unmarked member of an opposition allows for any and all possibilities and is very often the first to be acquired and the one more frequently used although this does not have to be an absolute rule. The marked member of an opposition which does make a specific claim for the presence of a distinctive semantic feature is less flexible and usually more limited in its distribution because of its greater precision in meaning. It is often acquired later and generally may be used less frequently than the unmarked member of an opposition although this is not necessarily an absolute rule. Therefore it should always be remembered that neither order of acquisition nor the rate of frequency is the sole criterion for choosing the markedness value. However the asymmetric markedness relationship implies that the marked members usually may be replaced by the unmarked members while the opposite is not necessarily true. The feature of Semantic Integrality which makes a claim for the perception of potentially discrete entities as a single continuous set has been chosen as the marked feature for the following reasons: i.
The process of first perceiving concrete or abstract entities as they appear discretely in discontinuous space, time, or existence is the fundamental perception of identification. The later perception of these potentially discrete entities as forming part of an integral set occupying continuous space, time, or existence is a more complex cognitive task. This greater cognitive complexity underlies and justifies the choice of Semantic Integrality as being the marked feature synergetically. ii. If the notion of isomorphism is taken seriously, it becomes evident that the notion of integrality or fusion functions on all the levels of language as well. iii. In phonology there is an opposition between vowels and diphthongs, nongeminated and geminated consonant phonemes, and stops/fricatives and affricates forming an obstruent consonant class. In all of these cases the second or marked member of the opposition represents a more complex unit or sound nucleus composed of the merger, fusion, or integration of two vowels or consonants and may be marked for Integrality (Tobin 1997). iv. In morphology or syntax there are also simple, analytic, independent units versus more synthetic, agglutinative, or compound units composed of the fusion, merger, or integration of different bound or unbound morphemes
The Hebrew prepositions mi-/min ‘from, of ’
or clause types, etc., both in word formation and in clause and sentence formation. In all of these cases, the fused, merged, or integrated units are perceived as being more complex or marked as opposed to the simple or more neutral counterparts. v. There is usually some sort of iconic connection between the signals themselves and their markedness values (Tobin 1990, 1994). The signs marked for the more complex or specific meanings are usually larger or more complex in their signal or form while the signs which are unmarked and are simpler and less complex in their meanings are usually simpler in their signals and forms as well. In the mi-/min system there is an iconic connection between the forms and the meanings of the members of the system and their markedness values. The unmarked morpheme min, which allows for a discrete, discontinuous reading, is an independent and autonomous word, while the marked member of the system, mi-, which only allows for a continuous reading, is a bound morpheme which connects and is suffused with the noun to which it is related.
. Microlevel analysis . Idioms and expressions If one briefly examines the idioms and expressions that appear in the dictionary for mi- and min several semantic and structural generalizations may be reached. There are fewer idioms with the marked form mi- with a tendency (but not a hard and fast rule) for collocations with indefinite phrases or definites with proper names: (5) a.
me- ei- matai? from- where- when ‘From what time? Since when?’ b. mi- zman from- time ‘long since’ c. kibalti mixtav mi-tel-aviv received-I letter from-tel-aviv ‘I received a letter from Tel-Aviv.’
Yishai Tobin
d. matok mi-dvash sweet from-honey ‘sweeter than honey.’
The definitions presented for the unmarked form, as previously listed, include ‘out of; because’; and ‘than’ which was ‘more than’ for the marked form. These additional dictionary definitions for the unmarked form min are compatible with our analysis. If only the unmarked form allows for discontinuous perceptions on the spatio-temporal-existential cline, it fits that min would have the additional message of ‘out of ’ or the possible causation related to ‘because’ without stipulating a continuous or discontinuous perception of the factors involved in the causation process. By the same token, the marked continuous meaning of mi- with an obligatory continuous reading makes it the more suitable candidate to be preferred for comparative constructions such as: yoter X mi-Y ‘more X than Y’; X yoter me-asher Y ‘X more than Y’; X yoter mimeni ‘X more than-I’; yoter X mi-she Y ‘more X than Y’; X mi-y ‘X-er/more X than Y’. Not unsurprisingly, there are many more idioms and expressions attributed to the unmarked form min with its neutral meaning. Many of these idioms and expressions follow the structural pattern of min ha-X ‘from the-X’ with the definite article ha- ‘the’ (again a generalization rather than a hard and fast rule!): (6) a.
min ha-din from the-rule/judgement ‘it is worthwhile; you should; it is incumbent upon . . . to’ b. min ha-minyan from the-number/quorum ‘registered, regular, ordinary’ c. min ha-nimna from the-avoided/prevented/refrained ‘It is impossible, unlikely’ d. min ha-stam from the ordinary/vague ‘probably’ e. min ha-raui from the worthy ‘It is worthwhile, it is proper’ f. yotse min ha-klal departs from the-rule ‘exceptional, extraordinary; excellent’
The Hebrew prepositions mi-/min ‘from, of ’
The neutral unmarked meaning of min allowing for both discontinuous and continuous perceptions on the spatio-temporal-existential cline makes it particularly suitable to collocate with the definite article. The definite article hahas been analyzed elsewhere to mean differentiation /focus /semantic integrality for all of its uses (definite article, deictic /vocative marker, question or interrogative marker, relative clause marker) (Tobin 1997). The so-called definite article differentiates, separates, and focuses on entities in discourse which can then be integrated into the larger message. The unmarked min allows for both integral (continuous) and non-integral (discontinuous), perceptions of entities rendering its frequent collocation with ha- both frequent and appropriate. The marked form mi- also collocates with the definite article but only in messages which have integral kinds of messages. I have found sentences containing the idiom yotse min ha-klal ‘exceptional, extraordinary; excellent’ with both forms and my claim is that the use of the marked mi- will trigger or is motivated by a more integral message. Example (7) is taken from a psychology text (Forward 1990). It represents a typical use of the idiom yotse min ha-klal where a list of exceptions is removed from a general class or rule, this “separation” explains why this idiom contains the unmarked form min which allows for discontinuous perceptions rather than the marked form. In this particular passage traumatic sexual and physical abuse are being emphasized as extreme examples of causes of psychological damage and the focus is on these particular exceptions: (7) yetsuyan ki-be-asher le-aspektim shel ‘hemshexiyut’ Will-be-noted that-in-which to-aspects of ‘continuity’ ve-’xazaratiyut’ yesh yotsim min ha klal. pgiya and ‘reversion’ there-are departures (min) from the-rule. damage minit o fisit yexola le-hiyot traumatit gam im hi sexual or physical can to-be traumatic also if she mitraxeshet paam axat, ve-hi alula li-grom le-nezek nafshi happens time one, and-she liable to-cause to-damage soul kaved. heavy. ‘It should be noted that with regard to aspects of ‘continuity’ and ‘reversion’ there are exceptions to (min) the-rule. Sexual or physical damage is liable to be traumatic even if it only occurs once, and is liable to cause serious psychological damage.’
Example (8), on the other hand, presents the usual idiom yotse min ha-klal with the marked form yotse me-ha-klal. This example is taken from a vegetarian
Yishai Tobin
cookbook (Brown 1984: 200) from a passage dealing with root vegetables and bulbs. The example explains why this class of vegetables in general should be cooked with their peels in order to preserve the vitamins found in the peel. The author then remarks in an almost parenthetical statement, that this does not include celery which should be peeled before cooking while “others” within the group of root vegetable and bulbs (not denoted by the definite article) should be peeled afterwards. In this passage the group itself as a set of a particular kind of vegetable requiring specific treatment in cooking is being stressed, thus motivating the unusual use of the marked form mi- in this established idiom. The emphasis is not on the celery but on the general class of vegetables which is the focus of this particular passage: (8) yerakot shoresh u-pkaot yesh le-kartsef be-mivreshet vegetables root and-bulbs there-is to-scrape with-a-brush ve-le-vashel be-klipatam. klalit lo kolfim lifnei ha-bishul: and-to-cook in-peel-their. generally no peel before the-cooking: . . . shoresh karpas hu ha-yotse me-ha-klal, she-oto . . . root celery he the departure from(me-)-the-rule, the-him yesh liklof lifnei ha-bishul. axerim kolfim axarei . . . there-is to-peel before the-cooking. Others are peeled afterwards. ‘Root vegetables and bulbs are to be scraped with a brush and cooked with their peels. Generally one doesn’t peel before cooking: . . . celery is theexception to (mi-)-the-rule, that has to be peeled before cooking. Others are peeled afterwards.’
. Utterances in context A similar communication strategy of preserving the marked form mi- for entities viewed as an integral set can be found in all styles and registers of both spoken and written Hebrew. Example (9) is taken from a written request of a student to be given a room in the university dormitories. The reason the student gives in the application is ‘from his actually being’ (me-estem heyoti) ‘from a rural settlement’ (mi-moshav) (both indicated by the appropriate marked form emphasizing the perception of continuous space in an indefinite phrase which is very far away ‘from the university’ (min ha-universita) (indicated by the appropriate unmarked form emphasizing the distance and the perception of discontinuous space in a definite phrase)). The student then uses the inflected forms of the marked form to further strengthen the argument of the distances involved from one place to the other:
The Hebrew prepositions mi-/min ‘from, of ’
(9) makom megurai meruxak min ha-universita kol she-kax place living-my remote from (min) the-university all the-thus me-etsem heyoti mi-moshav, muetet from (me-) actually-being-mine from (mi-)-moshav, reduced mimeni ha-taxbura ha-tsiburit ve-ha-efsharut from (mi-)-me The-transportation the-public and-the-possibility li-nsoa mimenu la-universita ve-le-hefex eina to-travel from (mi-)-it to-the-university and-the-opposite is-not realit. realistic. ‘My place of residence is far away from (min) the university so my being from (mi-) a moshav, reduces from (mi-) me the possibility to use public transportation from (mi-)-it (the moshav) to the university and the opposite which makes it (the possibility to use public transportation) unrealistic.’
The following pair of examples (10a, b) taken from the murder mystery (Gur 1991) show two alternative ways of viewing two men who come from a kibbutz in the south: first they are shown in contrast with the unmarked form (afilu ‘even’ . . . min ha-kibbutz ‘from the-kibbutz) to people from cities in the north or the center of Israel (mi-haifa/mi-tel-aviv) and then later the two kibbutzniks ‘from the south’ (me-ha-darom) are viewed as being part of (mi-tox ‘from-within’) a larger audience with the integral form: (10) a.
kshe-neidorf hirtseta be-shabatot ba-boker when-neidorf lectured on-saturdays in-the-morning ba-maxon hayu kulam baim. hayu baim in-the-institute were all-they come were come mi-haifa u-mitel-aviv ve-afilu shnei from (mi-)-haifa and-from (mi-)-tel-aviv and-even two ha-kibutznikim min ha-darom hayu megiim. (11) the-kibbutznikim from (min) the south were arrive. ‘When Neidorf lectured on Saturday morning at the Institute everybody would come. They would come from (me-)-Haifa and from mi--Tel-Aviv even the two kibbutzniks from (min) the south would be there.’ b. mi-tox ha-kahal ha-gadol hitgalu lo From (mi-)-within the-audience the-large revealed to-him le-shavrir shniya shnei ha-kibutznikin me-ha-darom. (19) to-split second two the-kibbutzniks from (me-)-the-south.
Yishai Tobin
‘From within the large audience the two kibbutniks from (me-)-thesouth appeared before him for a split second.’
The next set of examples comes from the vegetarian cookbook (Brown 1984). Examples (11a–c) show how leaves are ‘removed from’ a sauce or a cabbage or how the crust is ‘cut off ’ from the bread. The common denominator of these examples is the emphasis on the integrality of the sauce, cabbage, or bread: the “whole” having its “parts” removed from it. Not unsurprisingly, the marked form mi- is used in these utterances: (11) a.
mesalkim et alei ha-dafna me-ha-rotev remove OM leaves the-bay from (me-)-the-sauce ‘The bay leaves are removed from (me-)-the sauce.’ b. mesirim shmone alim xitsoniim me-ha-kruv remove eight leaves external from (me-)-the-cabbage ‘Take off eight outside leaves from (me-)-the-cabbage.’ c. xotxim et ha-krum me-ha-lexem (64) cut OM the-crust from (me-)-the-bread ‘Cut the crust off from (me-)-the bread.’
On the other hand, when one is trying to get rid of the carbon gases from legumes (i.e. to remove the gas from the plant) the unmarked form is used to emphasize the physical separation of the undesirable part from the whole: (12) derex nosefet le-hipater min ha-gazim she-be-kitniyot hi way another to-get-rid of (min) the gases that-in-legumes she le-hanbit otan lifnei ha-bishul. (193) to-make-sprout them before the-cook ‘Another way to get rid of the gas from (min) legumes is to cause them to sprout before cooking them.’
Another example of the use of the unmarked form to emphasize a physical removal or separation of entities as an act of prevention of a negative outcome can be found in Gur (1991: 13) where the removal of an electric plug from the socket is being focused upon as a means to prevent the burning of the coil a kettle or samovar:
The Hebrew prepositions mi-/min ‘from, of ’
(13) . . . ve-le-hakpid al hotsaat ha-teka min ha-xashmal . . . and-to-be-strict on removal the-plug from (min) the electricity le-axar ha-shimush. Rak lifnei xodesh nisraf ha-meixam to-after the-use. only before month was-burned the-samovar ha-saruf (13) the-burned . . . ‘Be strict about removing the plug from (min) the electricity after using it. It Was only a month ago that the burnt kettle burned up.’
Other examples emphasizing physical separation with the unmarked form come from spoken Hebrew. Example (14a) was overheard in a conversation between two students. One of them was telling how her five year old brother only listens to classical music. When his kindergarten teacher started playing children’s songs in his class, he left the room in anger saying “yuk”. Example (14b) comes from a television news program about various diseases that were spreading in Ruwanda after the civil war there. The reporter was describing how a woman dragged her dead husband’s body far away from her house to keep the plague out. In both instances where the message of a total physical separation is being emphasized, the unmarked form is being used: (14) a.
ha-yeled yatsa min ha-gan the-child left from (min) the-kindergarten ‘The child walked out of (min) the kindergarten.’ b. . . . ve-garera oto harxek min ha-bayit . . . and-dragged him far away from (min) the house ‘. . . and she (the wife) dragged him far away from the house.’
The next two examples from Gur (1991) contain the same verb yatsa ‘leave, depart’ in two different spatial perspectives reflected by the choice of preposition. Example (15a) describes how a police detective observes two suspects going out of one room and entering another as two distinct acts with the unmarked min reflecting a discontinuous perception. Example (15b), on the other hand, describes how a very large man is being helped out of a car that seems too small (comparative mi-) for him (or possibly both he and the man helping him) viewing both the men and the car as participants in the act using the marked form reflecting a continuous spatial perception: (15) a.
be-zavit ayin ra’ah gold et in-corner eye saw Gold OM min ha-xeder ha-katan from (min) The-room the-small
shnei ha-xashudim yotsim two the-suspects leave ve-ponim el ha-mitbax. (20) and-turn to the kitchen
Yishai Tobin
‘In the corner of his eye Gold saw the two suspects go out of min the small room and turn to the kitchen.’ b. hu patax et ha-delet le-hildesheimer ve-azar lo he opened OM the-door to-Hildesheimer and-helped to-him la-tset me-ha-mexonit, she-hayta ktana mi-day to-get out from (me-)-the-car, that was small from (mi-)-enough gam le-gufo shelo. (51) also to-body-his his ‘He opened the door for Hildesheimer and helped him get out of (mi-)-the-car, that was too (mi-) small for his body as well.’
. Macrolevel analysis . From sign to text As was seen in examples (7)–(15) the choice of a marked versus an unmarked form is not arbitrary but motivated by a subtle semantic distinction which can be traced to the invariant meanings of the forms in question. In other words, one opposed meaning may be seen to be more suitable to express a specific kind of message rather than the other. The repeated exploitation of this meaning to communicate that specific kind of message can be related to or associated with a particular leitmotif in both spoken and written discourse. If a specific sign or a set of signs containing the same or a similar semantic feature consistently occurs in contexts where a particular theme or leitmotif is being discussed, this will give the discourse or text a sense of coherence. The ability to trace the consistent exploitations of particular signs in specific contexts within a text is called the “from sign to text” approach. It allows one to view the text in the hierarchical ascending order of sign and system to context and text. In the present paper, the “from sign to text” approach implies: i.
The distribution of the marked versus unmarked forms is not random, but is skewed along thematic lines within the structure of a specific text. ii. The choice and preference of the marked or unmarked forms can be related directly to particular themes, specific characters, and recurring or related events in the plot or subplots within the structure of the text. iii. Therefore the choice and preference of the marked or unmarked forms can serve as part of a larger system of textual coherence and cohesion.
The Hebrew prepositions mi-/min ‘from, of ’
.. From sign to text: Noah and the ark (Genesis 6–7) In the story of Noah and the ark, Noah is commanded to collect all the animals of the earth and bring them to the ark (Genesis 6: 19–20). The first time this task is introduced, all the various kinds of species appear with mi- save for the cattle that appears with the unmarked min – this collocation, on the surface, at least – appears to be arbitrary: i.e., mi- and min are synonyms: (16) u-mi-kol ha-xai mi-kol basar shnayim mi-kol and-from-all the-living from-all flesh two From-all tavi el ha-teva le-haxayot itxa zaxar u-nekeyva you-will-bring to the-ark to-keep-alive with-you male and-female yehiyu. me-ha-of le-minehu u-min ha-behema they-will-be. From-the-fowl to-kind-his and-from the cattle le-mina u-mi-kol remes ha-adama le-minehu, shnayim to-kind-her, And-from-all creep the-earth to-kind-his, two mi-kol yavou eleixa le-haxayot. from-all they-will-come to-you to-keep-alive. ‘And of (mi-) every living thing of all flesh, two of (mi-) every sort shalt thou bring into the ark, to keep them alive with thee; they shall be male and female. Of (mi-) the fowl after their kind, and of (min) the cattle after their kind, of (mi-) every creeping thing after its kind, two of (mi-) every sort, shall come unto thee, to keep them alive.’
The collocation of behema ‘cattle, beast’ with the unmarked form min, however, turns out not to be arbitrary. Later in the passage (Genesis 7: 2–3) one learns that the cattle or beasts alone are further subdivided into two subcategories: clean (those that are permitted to be eaten like the fowl) versus unclean (those that are forbidden to be eaten). All the clean beasts and the fowl, therefore, form one category for which seven are to be put in the ark, on the one hand, but the unclean beasts, on the contrary, form a separate and distinct category of which only two are placed in the ark, on the other hand. Therefore the unique collocation of the behema (‘cattle’) in the previous example (16) with the unmarked form min is semantically motivated. By the same token, in the next example (17), the collocation of ha-behema ha-tehora (‘the clean beast’) with the marked form mi- together with the fowl – all of which are to enter the ark “seven-seven” and form one integral group or set – is semantically motivated as is the collocation of ha-behema ha-lo tehora (‘the unclean beast’) with the unmarked form min as an “exception” (or as the idiom goes: yotse min ha klal):
Yishai Tobin
(17) mi-kol ha-behema ha-tehora tikax lexa shiva ish rom-all the-cattle the-pure you-will-take to-you seven man ve-ishto u-min ha-behema ha-lo tehora hu shnayim ish and-wife-his and-from the-cattle the-no pure he two man ve-ishto. gam me-of ha-shamayim shiva shiva zaxar and-wife-his. also of-fowl the-sky seven seven male u-nekeyva le-haxayot zera al-pnei kol ha-arets. and-female to-keep-alive seed on-the-face-of all the-land. ‘Of (mi-) every clean beast thou shall take to thee seven and seven, each with his mate; and of (min) the beasts that are not clean two [and two], each with his mate; of (mi-) the fowl also of the air, seven and seven; male and female; to keep seed alive upon the face of all the earth.’
As the text continues (Genesis 7: 8–9) and the loading of the ark takes place before the flood, Noah, his sons, his wife, and his sons’ wives, enter the ark individually, and then there is a list of all the classes of creatures that enter the ark. In this antediluvian part of the text, each class of creature enters separately, as if it is being “counted off ”, not unsurprisingly, this is done exclusively with the unmarked form: (18) min ha-behema ha-tehora u-min ha-behema asher from (min) the-beast the-pure and from (min) the-beast that einena tehora is-not pure u- min ha-of ve-xol asher romes al ha-adama: and from (min) the-fowl and-all that creep on the-earth shnayim shnayim two two bau el noax el ha-teva zaxar u-nekeyva kaasher tsiva came to Noah to the ark male and-female as-that commanded elohim et noax. God OM Noah. ‘Of min clean beasts, and of min beasts that are not clean, and of min fowls, And of everything that creepeth upon the ground, there went in two and two unto Noah into the ark, male and female, as God commanded Noah.’
After “the windows of heavens were opened and the rain was upon the earth forty days and forty nights” the text reiterates the roster of who had entered the ark and which classes of creatures (all listed in the previous examples) without using the prepositions. This is then followed by a general summary (Genesis
The Hebrew prepositions mi-/min ‘from, of ’
7: 15–17) of what occurred where all the creatures are grouped into a single set or category referred to as ‘of all flesh’ with, not unsurprisingly, the marked form mi-kol basar: (19) . . . mi-kol basar asher bo ruax xayim . . . mi-kol . . . from (mi-)-all flesh that in-it spirit life . . . from (mi-)-all basar bau . . . flesh came . . . ‘. . . of (mi-)-all flesh wherein is the breath of life . . . went in . . . of (mi-) of all flesh . . . ’
The text continues with the story of the flood that destroyed all the living creatures on earth of all sorts and types, which, once again, are listed without the use of these prepositions. At the close of the text, when the devastation is again summarized we have the use of the marked form (translated as ‘whatsoever’ and ‘both’) when listing the creatures that ‘were blotted out from the earth’ where their physical separation from the earth is emphasized by the use of the unmarked form: ‘all (kol asher )in whose nostrils was the breath of life, whatsoever (mi-kol asher) was in the dry land died. And he blotted out every living substance which was upon the face of the ground, both man, (mi-adam) and cattle, and creeping thing, and fowl of the heaven; and they were blotted out from the earth (min ha-arets); and Noah only was left and they that were with him in the ark.’ (Genesis 7: 22–24) . From text to sign The “from sign to text” approach is only one way to analyze a text semiotically. It is also possible to view an entire discourse or a text itself as a unified whole composed of a signal and a meaning: i.e., a sign in its own right where the message of the text represents its meaning or signifié, and the non-random distribution of the language in the text represents its form, signal, or signifiant. When it is possible to view a text as a sign in its own right, it then becomes possible to analyze that text and the non-random distribution of the signs within that text in a method called the “from text to sign” approach. This approach attempts to show that the consistent choice of a marked or an unmarked form may be directly determined and motivated by the message of the text itself as it functions as the meaning half of the text as a sign: i.e., the larger textual message will motivate the choice of the marked or the unmarked form throughout the text. The message of the text, functioning as the meaning of the “text as sign” involves a conflation of elements including the characters,
Yishai Tobin
the plot, the themes and leitmotifs that all function together within the larger system of the text. This approach involves the hierarchical descending order from text and context to system and sign. The “from text to sign” approach has explained the use of unusual marked forms in classic and contemporary texts (e.g., curiouser and curiouser in Alice’s Adventures in Wonderland, darlingest in Social Disease (Rudnick 1986) and worse/worser/worst in the Jerusalem Post) and in other texts of literary and journalistic prose as well as poetry in English and Hebrew (Tobin 1990, 1993, 1994/1995). In the present study the text to be analyzed is horim marilim (toxic parents) (Forward 1990). This text suggests a therapeutic method to handle the consequences of parent abuse which maintains the wholeness (or the ‘integrality’) of the individual. The author does not isolate elements, but suggests that any experience in the past affects future experiences, emotions, or cognitions. This concept that the damage caused by abuse spreads into every area of the child’s experience is the integral message of the text. One would then expect that the marked form mi- would appear more frequently in the text than the unmarked form min. And this hypothesis is true: there are 135 instances of the marked form as opposed to only 10 instances of the unmarked form. The marked form appears in those contexts relating the past to the present and the future, presenting the child as an integral whole which includes his fears and distrust based on the past, and presenting the ‘toxic parents’ as integral entities as well, albeit negative ones. It is not surprising that in a text whose message is the impossibility of separating the past from one’s life that the marked form would predominate. By the same token, those instances with the unmarked form either appear in established idioms (viz. ex. (7) yotse min ha-klal ‘exceptional’, or lehotsi et ha-martsea min ha-sak ‘to let the cat out of the bag’) or in contexts emphasizing separation (emotional or otherwise) from the parents presenting them in a particularly negative light. An example of the opposite from text to sign strategy can be found in the Hebrew version of Sophocles’s Oedipus Rex. One can say that the message of this classic play is that one cannot escape his destiny or fate. Those characters who actively attempt to avoid (try to be separated) from their fate will appear more with the unmarked form rather than the marked form. This hypothesis is supported by the data. Oedipus, who escapes his parents only to kill his real father, (an active attempt to escape his fate), represents the context where 70% of the unmarked form appear. This is followed by Jocasta, Oedipus’s mother, who ordered Oedipus to be killed as a child (in order to have him escape his fate) represents the context where 12% of the unmarked form appear. The dis-
The Hebrew prepositions mi-/min ‘from, of ’
tribution of the remaining 18% of the unmarked form appear sporadically in the contexts of the other characters and events. An example of both strategies appearing in the same text can be found in Amos Oz’s story ahava meuxeret (‘late love’). This text deals with Shraga Ungar, an elderly man who lectures in kibbutzim about the danger of a new Holocaust starting in Russia. The text deals with loneliness and the great distances that separate people and the need for love to unite people. This message about the danger of the lack of contact between human beings motivates the distribution of the marked and unmarked forms in this text. Shraga feels alienated from the people in the kibbutz to whom he lectures. He complains that it is as if he does not exist there: he does not listen to what the kibbutzniks say to him and viceversa. He feels lonely and dispensable there. In this context the unmarked form predominates. On the other hand, when Shraga meets an old woman friend who used to accompany him on his lectures and with whom a relationship develops, in this context the marked form predominates. Shraga’s view of himself is a negative one. He feels that his body is disintegrating and recurrent leitmotif in the text focuses on his bad physical condition. In this context he discusses his obesity, his high blood pressure, his weak gums which give him bad breath, etc. His view that his body is falling apart is further supported by the prevalent use of the unmarked form in the description of his various ailments. Furthermore, he feels that people are obviously avoiding him because of his physical condition in which, again, the unmarked form predominates. In other words there is a non-random distribution of their marked and unmarked forms so as in contexts of loneliness and alienation, the unmarked form predominates and in the context of unity and affection, the marked form predominates.
. Summary and conclusions In this paper I have shown how supposed synonymous or full versus abbreviated forms sharing a semantic domain actually differ in meaning based on the theoretical tenets of invariance, markedness and distinctive feature theory. The methodology employed includes both micro-sentence and macro-discourse level analyses following semiotic approaches to text analysis. In every case it was shown that the choice of one linguistic sign over another was not arbitrary but motivated by a subtle semantic distinction. Despite the fact that this semantic distinction is, indeed, a subtle one, in the words of Dwight Bolinger (1977: 17): “True, these are subtle differences, but who says semantic distinctions have to be gross? ”
Yishai Tobin
Notes * Much of the data presented here were taken from seminar papers written by my students Tal Berebi and Nurit Tobali from Haifa University and the Ben-Gurion University of the Negev, respectively. . The historical distribution of the allomorphs mi- versus me- is phonetically determined: me- precedes the so-called “guttural” consonants aleph /‘/ (a glottal stop), heh /h/ (a glottal fricative), ayin/het /’h/ (voiced/voiceless pharyngeal fricatives) and a resh />/ (a velar /r/). Although most of these consonants are not pronounced at all, or not uniformly pronounced by native speakers of Israeli Hebrew, the actual alternation of these allomorphs is purely prescriptive today. We will use the form mi- in our discussion as representing both regardless of whether an individual speaker follows the prescriptive rule or not. The highly prescriptive attitudes towards language prevalent in Israel is discussed in Tobin (1977: Ch. 5). . The inflection of mi- is irregular and Glinert (1994: 43) states: “. . . the inflection of mi- is so odd as to defy simple explanation. Of the two forms for ‘from us’, mimenu is the more colloquial. It is in fact identical with the ‘him’ form.” I disagree with Glinert’s stylistic statement and have found the form me-itanu to be more informal and colloquial. . Semantic Integrality has served as the basis for analyzing the following linguistic phenomena: systems of grammatical number in general and so-called “irregular” plurals, mass/count nouns, agreement in English; the quantifiers many vs. much; each vs. every; and some vs. any in English; the adverbs of comparison like vs. as in English and their so-called Hebrew equivalents kemo/ke-; the connectives also vs. too in English; the restrictives only vs. just in English; the deletives without vs. -less in English and their so-called Hebrew equivalents beli/le-lo; the synthetic vs. analytic systems of comparatives and superlatives (the) more/most X/-er/-est X in English; the diminishers few vs. less in English and their so-called Hebrew equivalents me’at/ketsat; the conditionals if vs. whether in English; possessive constructions in English and Hebrew; and contracted vs. non-contracted forms in English and Hebrew in Tobin (1990, 1994) as well as a means to distinguish between the lexical pairs: little/big versus small/large (Tobin 1999a). A semantic distinction between other so-called full versus abbreviated forms (until/till) based on another set of semantic features may be found in Tobin (1999b).
References Alcalay, Reuben (1963). The Complete Hebrew/English Dictionary. Ramat-Gan/Jerusalem: Massada. Andrews, Edna (1990). Markedness Theory: The Union of Asymmetry and Semiosis in Language. Durham, NC/London: Duke University Press. Andrews, Edna, & Tobin, Yishai Tobin (1996). Towards a Calculus of Meaning: Studies in Markedness, Distinctive Features and Deixis. Amsterdam/Philadelphia: John Benjamins. Aphek, Edna, & Tobin, Yishai (1988). Word Systems in Modern Hebrew: Implications and Applications. Leiden/New York: E.J. Brill.
The Hebrew prepositions mi-/min ‘from, of ’
Aphek, Edna, & Tobin, Yishai (1989). The Semiotics of Fortune-Telling. Amsterdam/Philadelphia: John Benjamins. Battistella, Edwin (1990). Markedness: The Evaluative Superstructure of Language. Albany, NY: SUNY Press. Battistella, Edwin (1996). The Logic of Markedness. Oxford: Oxford University Press. Berman, Ruth A. (1978). Modern Hebrew Structure. Tel-Aviv: Tel-Aviv University. Bolinger, Dwight (1977). Meaning and Form. London/New York: Longman. Brown, S. (1984). sefer ha-bishul ha-tsimxoni (The Vegetarian Cookbook). Tel-Aviv: Kineret. Carrol, Lewis (1865/1970). Alice’s Adventures in Wonderland. London: Penguin. Clark, Herbert (1973). Space, time, semantics and the child. In T.F. Moore (Ed.), Cognitive Development and the Acquisition of Language (pp. 18–33). New York: Academic Press. Contini-Morava, Ellen (1989). Discourse Pragmatics and Semantic Organization: The Case of Negation and Tense-Aspect with Special Reference to Swahili. Berlin/New York: Mouton de Gruyter. Contini-Morava, Ellen, & Sussman Goldberg, Barbara (1995). Meaning as Explanation: Advances in Linguistic Sign Theory. Berlin/New York: Mouton de Gruyter. Contini-Morava, Ellen, & Tobin, Yishai (Eds.). (2000). Between Grammar and Lexicon. Amsterdam/Philadelphia: John Benjamins. Diver, William (1969). The System of relevance in the Homeric verb. Acta Linguistica Hafniensia, 12, 45–68. Forward, S. (1990). horim marilim (Toxic Parents). Tel-Aviv: Matar. García, Erica (1975). The Role of Theory in Linguistic Analysis: The Spanish Pronoun System. Amsterdam: North Holland. García, Erica, van Putte, Florimon, & Tobin, Yishai (1987). Cross linguistic equivalence, translatability, and contrastive analysis. Folia Linguistica, 21, 373–405. Givón, Talmy (1979). On Understanding Grammar. New York: Academic Press. Glinert, Lewis (1989). The Grammar of Modern Hebrew. Cambridge: Cambridge University Press. Glinert, Lewis (1994). Modern Hebrew: An Essential Grammar. London/New York: Routledge. Gorup, Radmila (1987). Semantic Organization of the Serbo-Croatian Verb: The System of Concentration of Attention. Munich: Verlag Otto Sagner. Goshen Gotestein, M., Livne, Z., & Span, S. (1977). ha-dikduk ha-ivri ha-shimushi. Tel-Aviv: Schocken. Guillaume, Gustave (1945). L’architectonique du temps dans les langues classiques. Copenhagen: Munksgaard. Guillaume, Gustave (1971). Leçons de linguistique, 1948–1949, Série B: psycho-systématique du langage: principes, méthodes et applications I. Québec/Paris: Presses de l’Université Laval/Klinsieck. Gur, Batia (1991). retsax be-shabat ba-boker/A Saturday Morning Murder. Jerusalem: Keter. Hirtle, Walter (1982). Number and Inner Space: A Study of Grammatical Number in English. Québec: Presses de l’Université Laval. Huffman, Alan (1996). The Categories of Grammar: French lui and le. Amsterdam/Philadelphia: John Benjamins.
Yishai Tobin
Jakobson, Roman (1957). Shifters, Verbal Categories and the Russian verb. Cambridge, MA: Harvard University Press. Kirsner, Robert (1979). The Problem of Presentative Sentences in Modern Dutch. Amsterdam: North Holland. Klein-Andreu, Flora (Ed.). (1983). Discourse Perspectives in Syntax. New York: Academic Press. Kova´cs, Ferenc (1961). À propos d’une loi sémantique. Acta Linguistica Hungarica, II (3–4), 405–411. Kronasser, Heinz (1952). Handbuch der Semasiologie. Heidelberg: Carl Winter Universitätsverlag. Makkai, Adam (1972). Idiom Structure in English. The Hague: Mouton. Pike, Kenneth (1967). Language in Relation to a Unified Theory of the Structure of Human Behavior. The Hague: Mouton. Reid, Wallis (1991). Verb Number in English: A Functional Explanation. London/New York: Longman. Reid, Wallis, & Otheguy, Ricardo (Eds.). (In press). Issues in Neo-Saussurean Linguistics: Advances in Sign-Based Approaches to the Study of Language. Amsterdam/Philadelphia: John Benjamins. Reppen, Hava (1995). Conditions for the Occurrence of a Governed Case-Relation Marker in Contemporary Hebrew: Especially min and Cognate Case-Relation Markers. Ph.D. dissertation. Tel-Aviv University. (in Hebrew). Reppen, Hava (1996). “The agent of an Action”, “the stimulation of a state” and “the instrument” that are realized by the prepositional phrases “al-jede + NP, “min” + NP and “be” + NP.” In Esther Borochovsky and Yishai Tobin (Eds.), Proceedings of the Eleventh and Twelfth Annual Meetings of the Societatis Linguisticae Europaeae Sodalicium Israelënse (pp. 99–106). (in Hebrew). Rosen, A. (1965). Hebrew. Jerusalem: Niv. Rosén, Haiim (1977). Contemporary Hebrew. The Hague: Mouton. Rudnick, Paul (1986). Social Disease. New York: Ballantine. Saussure, Ferdinand de (1913/1959). A Course in General Linguistics. New York: Philosophical Library. Sophocles (1994). Oedipus the King (trans. by Aharon Shabtai). Tel-Aviv: Schocken. Tobin, Yishai (Ed.). (1988). The Prague School and its Legacy. Amsterdam/Philadelphia: John Benjamins. Tobin, Yishai (Ed.). (1989). From Sign to Text: A Semiotic View of Communication. Amsterdam/Philadelphia: John Benjamins. Tobin, Yishai (1990). Semiotics and Linguistics. London/New York: Longman. Tobin, Yishai (1993). Aspect in the English Verb: Process and Result in Language. London/New York: Longman. Tobin, Yishai (1994/1995). Invariance, Markedness and Distinctive Feature Analysis: A Contrastive Study of Sign Systems in English and Hebrew. Amsterdam/Philadelphia: John Benjamins (hardcover)/Be’er Sheva: Ben-Gurion University Press (paperback). Tobin, Yishai (1997). Phonology as Human Behavior: Theoretical Implications and Clinical Applications. Durham, NC/London: Duke University Press.
The Hebrew prepositions mi-/min ‘from, of ’
Tobin, Yishai (1997). Same versus different crosslinguistically: The articles in English, Spanish and Hebrew. In R. Hickey & S. Puppel (Eds.), Language History and Language modelling: A Festschrift for Jacek Fisiak on his Sixtieth Birthday (pp. 1831–1858). Berlin/New York: Mouton-de Gruyter. Tobin, Yishai (1998). Review: The Logic of Markedness by Edwin Battistella. Language, 74 (4), 832–834. Tobin, Yishai (1999a). One size does not fit all: A semantic analysis of ‘small’ and ‘large’ vs. ‘little’ and ‘big’. In Christopher Beedham (Ed.), Langue and Parole in Synchronic and Diachronic Perspective (pp. 169–180). Amsterdam: Elsevier. Tobin, Yishai (1999b). Till vs. Until: A sign-oriented approach. Sémiotique Appliquée/ Applied Semiotics, 8, 426–436. Traugott, Elizabeth C. (1974). Explorations in linguistic elaboration: Language change, language acquisition, and the genesis of spatio-temporal terms. In J. Andersen & C. Jones (Eds.), Historical linguistics I (pp. 263–314). Amsterdam: North Holland. Traugott, Elizabeth C. (1975). Spatial expressions of tense and temporal sequencing: A contribution to the study of semantic fields. Semiotica, 15 (3), 205–230. Traugott, Elizabeth C. (1978). On the expression of spatio-temporal relations in language. In Joseph Greenberg (Ed.), Universals of human language, Volume 3, Word structure (pp. 369–400). Stanford: Stanford University Press. Wallenrod, R., & Aaroni, A. (1949). Fundamentals of Hebrew Grammar. Tel-Aviv: Shilo. Waugh, Linda, & Rudy, Stephen (1991). New Vistas in Grammar: Invariance vs. Variation. Amsterdam/Philadelphia: John Benjamins. Wickens, Mark (1991). Grammatical Number in English. Amsterdam/Philadelphia: John Benjamins. Wierzbicka, Anna (1972). Semantic Primitives. Frankfurt: Athenaum. Wierzbicka, Anna (1980). Lingua Mentalis. New York: Academic Press. Wittenberg, Morris (1977). Comprehensive Modern Hebrew-English Dictionary. Jerusalem: Yeda. Yalin, David. dikduk ha-lashon ha-ivrit. Jerusalem: Dvir. ——— (1958). The Holy Scriptures (according to the Masoretic text). Philadelphia: The Jewish Publication Society of America.
A contrastive analysis of French and Hebrew prepositions The case of sans, bli-belo and lelo Susanne Feigenbaum Department of French, University of Haifa
.
Introduction
Prepositions are characterized by a high level of polysemy. One finds that for French the most problematic areas concern the local and the temporal prepositions, because they have a semantic content as well as a colorless, case assigning function. Moreover, as a semantic group, they possess multiple lexical valencies, which have to be acquired separately (see Ben Rafael & Lebas, this volume). Other prepositions, like sans (without), do not behave as functors and therefore, they may be defined as lexical, colored, circumstantial, concrete or expressive, as in Cadiot (1997: 37). They are usually studied for their semantic contents and are not necessarily counted among the items that are problematic from the point of view of their lexical valency. This article deals precisely with the issue of sans and its valencies. Let us start with the assumption, present in most of the grammars, that sans has one meaning, namely that of a deletive, like in the example (1) a.
une maison sans jardin ‘a house without a garden’ b. une maison avec un jardin ‘a house with a garden’
Deletion can be defined as the relationship between a positive and a negative referent, where the negative referent presupposes a contradictory relationship with its affirmative counterpart. Thus, statement (1a) presupposes that houses have gardens and can therefore considered to be antonymic with (1b). More-
Susanne Feigenbaum
over, statements with sans may be described as a composed assertion (2a) according to the semiotics of Charles Sanders Peirce (1931–1958: 2.345), or as an inverse negation (2b) in syntax (Muller 1991): (2) a. (+A) & (–B)]. b. /X while not Y/
Yet, this relationship (henceforth /without/) is not always present with sans, as when it acts as a negative marker of a predicate or when it replaces an adverb. For instance, the group sans retard in the sentence (3a) may be paraphrased by a negative clause (3b) or by a lexical equivalent (3c): (3) a.
Je vous apporterai les esquisses sans retard. ‘I shall bring you the sketches without delay.’ b. Je ne tarderai pas de vous apporter les esquisses. ‘I shall not be long in bringing you the sketches.’ c. Je vous apporterai les esquisses immédiatement. ‘I shall bring you the sketches immediately.’
In this case, we conclude that the manner adverb sans retard “without delay” is not the contradiction of avec retard “with delay,” because the whole sentence does not presuppose that there is usually a delay in bringing the sketches. A similar lack of presupposition concerns adjectives: (4) a.
Les nuages sont sans couleur. ‘The clouds are colorless.’ b. Les nuages sont blancs. ‘The clouds are white.’
The statements (4a) and (4b) may be considered to be referentially equivalent, white being taken as the absence of color. Since clouds are mostly experienced as white, sans couleur does not presuppose its opposite. In this article, we shall speak of plurivalency, which is another way to define polysemy. We claim that semantic distinctions like the ones described above may be treated within semiotic plurivalency and that they have their implications for the linguistic system of the equivalent prepositions of sans in Hebrew. The arguments will be presented in the following way: The second section presents the theoretical frame in which the semiotic plurivalency of sans can be summarized. The third section presents the link between the two languages. In it one will find the list of the Hebrew lexemes that constitute the equivalent paradigm for the uses of sans. In the fourth section, I shall revise the relevant points and conclusions of an earlier paper (1997b), concerning the expressive-
A contrastive analysis of French and Hebrew prepositions
ness of some of these prepositions. This discussion will be based on extensive data and other linguistic research, as in Tobin (1994). Finally, I shall summarize my findings with regard to the linguistic changes going on in Hebrew and the dynamic status of these prepositions.
. The semiotic valency of sans Peirce considers that a sign may be apprehended in three ways: a) by itself, b) in contrast with another object and c) as a medium between two objects. In the first case, the sign is monovalent, in the second it is divalent and in the third, trivalent. . Trivalent sans As I have said before, presupposition is an eminent factor in the relationship /without/. From the semiotic point of view, the preposition /without/ is trivalent. It has a mediating valency, presenting the link between /+A/ and /–B/. This link exists because /–B/ indicates a negative polarity of a presupposed affirmative pole which is a lexical component of /+A/, or which is implied by /+A/. It is furthermore important to remember that in Peircean semiotics, the triadic sign includes a dyadic sign. Thus, for (c), we have first the contradiction of <+pleuvoir> and then the triadic conjunction of <+tonner> with <–pleuvoir> (cf. Feigenbaum 1996: 294). Where the various presuppositions are concerned, they are illustrated in (5a–c). (5) a.
Il prend son café sans lait. ‘He drinks his coffee without milk.’ b. Elle sort sans se retourner; elle entre sans que les autres la remarquent. ‘She leaves without turning back; she enters without being seen by the others. ’ c. Il tonne sans pleuvoir. ‘It thunders without raining.’
The context on which the presupposition is based is either cultural (5a), habitual (5b) or experimental (5c).
Susanne Feigenbaum
. Divalent sans The dichotomic valency is given when /without/ contradicts an element that is not a lexical component of /A/, but is to be found in the co-text or in the universe of discourse: (6) Marcus attend en bas, mais cette fois sans les billets pour le cinéma. ‘Marcus is waiting downstairs, but this time without the tickets for the cinema.’
This statement can be contrasted with (5c), where it is expected that process /A/ is accompanied by /B/ and where the absence of /B/ puts /A/ into a new isotopy. On the contrary, the sentence (6) above does not imply that the nature of “waiting” has changed. The discourse referent in the sense of Karttunen (1971) is based solely on /B/ and as a consequence, the dichotomic link of sans billets with the affirmative /with/ statement is not of a lexical but of a discursive nature. . Monovalent sans Sans is semiotically monovalent in lexically complex units like idioms and composites, because the term introduced by sans is not presupposed, neither lexically by /A/ nor by the discourse. Let us illustrate the case of the adverbial composite: (7) Tu n’as pas très faim aujourd’hui, sinon tu aurais mangé sans attendre. ‘You are not very hungry today, otherwise you would have eaten immediately.’
The group sans attendre represents the same case as sans retard in example (3). It has no semantic affinity with the preceding verb, but functions as a manner adverbial, the meaning of which can be glossed by ’immediately’. Monovalency is also present with adjectival composites: (8) a.
C’était une ville sans nom. this was a town without a name ‘This was an anonymous town.’ b. un animal sans défense an animal without defense ‘a good-natured animal’
A contrastive analysis of French and Hebrew prepositions
According to Corbin (1997) lexically complex units are characterized by the fact that they occupy the position of a syntactic constituent. Very often, for reasons of coherence the group /sans + noun/ becomes a composite because it cannot to be taken in the transparent sense. If one would take the expression sans nom literally, it would imply that there were a semantic affinity between /name/ and /town/. However, the componential features of /town/ do not include /name/. The fact that there are no towns without names belongs to our knowledge of the world and not to the lexicon. Therefore, (8a) is read in the sense that one does not know the name of the town, or that its name is of no importance. Another semiotic characteristic of lexically complex units mentioned by Corbin (ibid: 54) is their referential stability. Thus, sans nom as such is often found, like une île sans nom (an unknown island), un pays sans nom (an unknown land), or une angoisse sans nom (un undescribable fear). Similarly, sans défense (8b) means that defense is not required facing this animal, but “defense” itself is not a lexical feature of “animal”. Again, sans défense is a unit with a stable reference, as with un oiseau sans défense (a harmless bird), une jeune fille sans défense (a helpless young girl), or un homme simple, crédule et sans défense (a simple, trusting and harmless man). Finally, the affixal status of sans is confirmed by the fact that it exists in a number of nominal composites, like sans-fil (wireless), boire-sans-soif (drunkard), or sans-culotte (revolutionary). What we have, then, with monovalent sans, is a preposition devoid of its relational reference. As a result, it relies solely on the negative seme. Where negativity is concerned, the monovalent sign resembles the dyadic sign. However, contrary to the latter, monovalent sans does not contradict any other assertion. A first consequence of this independence both from the preceding term /A/ and from a /+B/ in the discourse is the morphological bond with the following term /B/. This is coherent with Riegel’s (1985) theory that groups like the above are attributive composites, where sans acts as a negative affix. There is a second referential consequence. As one knows, negative words produce not only contrariety, but some of them are negative semantic primitives in the sense of Wierzbicka (1972). Deletives play a role in this field when it comes to make up for lexical gaps. As an example, “bald” is restricted to /+A head – B hair/, whereas for other body parts “hairless” is used. Since there are no deletive prefixes in French (Riegel 1985; Feigenbaum 1997c), one uses /sans + noun/: (9) Il avait des mains sans poils ‘His hands were hairless.’
Susanne Feigenbaum
To conclude, these composites with sans occupy the syntactic position of simple adjectives or nouns. The next examples show that among other things they may figure in attributive positions in the same way as adjectival or nominal attributes do: (10) Malades et sans ressources, pauvres de nous!1 Ill and adj attr without ressources, poor us ‘Ill and resourceless, poor us!’ (11) Un sommet baptisé “sans cravate” a summit baptized nom attr “without tie” ‘A so called “tie less” summit meeting’
. From deletive to negative In Table 1 I have indicated how the three valencies differ with respect to the presupposition of sans. First, when trivalent sans establishes a tight lexical relation between the terms preceding and following it (ex. 1), the deletive seme is preponderant. Second, when divalent sans stands in discursive relation to an affirmative expression, we do not speak of deletion but of contradiction (ex. 6). Third, when monovalent sans figures in collocations, it represents a negative primitive (ex. 8), which does not presuppose any affirmative contextual or lexical element. Table 1 might also be imagined as scissors, with trivalency at the open end, closing down on monovalency, negativity being the stable element. It need not surprise us that attributes with sans often have the value of a negative sentence, as in “Vous croyez, docteur, que c’est sans gravité de changer de marque de lait?” (Do you think, doctor, that it is not serious to use another brand of milk?). Here, sans is not a deletive but the transformation of a contradicted adjective: (12) ce n’est pas grave = c’est sans gravité it is not serious = it is without seriousness
Table 1. The plurivalency of sans Trivalent
Divalent
Monovalent
¬B→+B ¬B→+A
¬B→+B
¬B
A contrastive analysis of French and Hebrew prepositions
Thus, when describing sans, we have to account for its negative value as well. A sound confirmation of this aspect can be found in the fact that in some dialects sans is equal to the negative adverb non (not).2 As a result of findings like the above, I have expressed the following hypotheses in Feigenbaum (1997c), to which I shall refer during the rest of this paper: i.
Trivalent and divalent sans belong to syntax whereas monovalent sans belongs to the lexicon. ii. Lexical sans is more expressive than syntactic sans.
. The Hebrew equivalents of sans In Hebrew, there is more than one equivalent for sans, which we shall divide into two groups. In the first group we find several prepositions and in the second group a number of verbal forms. In order to provide a complete picture, we shall briefly recall the facts concerning the second group from Feigenbaum (1997b), but then we shall go into details concerning the prepositions, which are the main theme of this article. . Deprived of The designation of deletion and of something missing can, of course, be expressed by other lexical means than prepositions. As an example, take ‘deprived of ’ or ‘lacking’ and the French verb dépourvu de. Hebrew, too, has such semantically transparent verbs for the deletive relation. Some of them are infinite forms of the following verbs illustrated in (13): xaser (to miss) – occurring in the form xasar (‘missing’ – active) and mexusar (‘missing’ – passive) – and lintol (to deprive of) – in the form netul (deprived of). (13) a.
xasar.busha, xasar.ereKh act missing shame, missing value ‘shameless, without value’ b. mexusar avoda pass missing work ‘unemployed’ c. netul.hakarah deprived of consciousness ‘unconscious’
Susanne Feigenbaum
The link between the above examples and our preposition sans is that they are equivalent to nominal and adjectival composites in French. Thus, xasar busha means sans-gêne, xasar ereKh means sans valeur, mexusar avoda means sanstravail, and netul hakarah stands for sans connaissance. At the same time, these verbal forms are equivalent to the French prefix dé-, as in (14) zejtim netul ei galin = olives dénoyautées olives deprived of pits = olives pref dis pitted
Because these deletive verb forms have not been classified as prepositions in Hebrew, a more specific discussion will not be pursued here. . The prepositions The Hebrew equivalent of sans is first of all a set of “negative complex particles” (Glinert 1989: 293). Part of them have an iconic morphological structure resembling the English word ‘without’. This is most obvious with the word belo, consisting of the preposition be ‘with’ and the negative adverb lo ‘not’. Thus, belo may be glossed as /with – no/. This reading is evident in the following example, rooted in Mishnaic Hebrew (Azar 1995: 183): (15) Kama hen jafot beperot vekama hen jafin belo how they beautiful with fruit and.how they beautiful with.no perot! fruit ‘How beautiful they are with fruit and how beautiful they are without fruit!’
Let us add, however, that the reading ‘with’ for the preposition be in (15) has changed in modern Hebrew. Nowadays, the preposition corresponding to ‘with’ is the word im: (16) Im o bli sukar? With or without sugar?
We shall also discuss a variant of belo, the form lelo, based on the preposition le and the negative word lo. Furthermore, there is a group based on the ancient negative adverb belah. This root is either used as an independent morpheme spelled beli, or it is combined with the prepositions me and le, resulting in mibeli and libeli.3 Since all the particles combined with the negative bases are semantically opaque
A contrastive analysis of French and Hebrew prepositions
(cf. also Borochovsky and Reppen, this volume), we shall consider them as morphologically complex words, according to the following description: (17) a.
beli, mibeli, libeli [[(me) (le)] PREP [beli] ADV] PREP b. belo, lelo [[be, le] PREP [lo] ADV] PREP
To conclude, all the prepositions are transparent for negativity, because they contain the morphemes beli and lo. . The distribution of beli, mibeli, lelo and belo The syntactic distribution of the Hebrew prepositions corresponds largely to the one of sans (Feigenbaum 1997b), as they precede NPs, infinitives, or clauses. Some remarks will be made for each one of them, based on Even-Shoshan (1985). .. Beli It is a negative word, which can be glossed as ‘with not’, ‘in a manner that not’ or ‘with lack of ’. It precedes an NP (18a), an infinitive (18b) or a clause (18c): (18) a.
beli levush with not np clothes ‘without clothes’ b. beli leharim kol with not inf make sound ‘without saying a word’ c. beli she tabit meaxoreha in a manner that not P’ she look back ‘without her looking back’
Biblical Hebrew also knew bebeli (be + beli), as well as libeli (le + beli). In modern Hebrew, however, these forms have become archaic.4 .. Mibeli In Biblical Hebrew, this negative word has the same syntactic distribution as beli, but its semantics presents a more intensive aspect. In Modern Hebrew, the use of mibeli is restricted to infinitives (19a) and clauses (19b):
Susanne Feigenbaum
(19) a.
mibeli lifgoa without inf offend ‘without offending’ b. Haaxot xisnah et hatinok mibli she hi badka et ha kartis. the nurse vaccinated the baby without P’ that she checked the card. ‘The nurse vaccinated the baby without checking the card.’
.. This negative word can be read as ‘with no’, ‘lacking’ and as ‘without’. Even if the first interpretation is more adapt for traditional Hebrew, ‘with no’ is theoretically always possible before an NP, for instance in the following example, cited by Even-Shoshan (ibid): (20) Bena’alaim, belo na’alaim, haein hakol axat? ‘With shoes, with no shoes, is not everything the same?
However, in most cases, the reading (b) ‘without’ seems more natural: (21) a.
nitan lomar belo hafrasa ‘it may be said with no exaggeration’ b. ‘it may be said without exaggeration’
The choice between the two readings above is restricted to NPs. In modern Hebrew, belo precedes also infinitives and clauses (22), in which case it can only mean ‘without’: (22) a.
belo laxshos without inf suspect ‘without suspecting’ b. belo she haia xashad without P’ that there was suspicion ‘without there being any suspicion’
.. It has the same definition like the others: ‘with no’, ‘lacking’. Out of context, both readings are acceptable for (a), whereas ‘without’ is more coherent with (b): (23) a.
erets lelo neft ‘a country with no/without oil’ b. medina lelo misuI ‘a state with no/without taxes’
A contrastive analysis of French and Hebrew prepositions
c.
halvaot lelo ribit ‘loans without interest’
Let us underline that in modern Hebrew, lelo figures before NPs alone. . The valency of beli, mibeli, lelo and belo Is there a semiotic distinction in Hebrew between the three valencies elaborated for French sans? Concerning trivalency (cf. 19b) and divalency (cf. 15, 16), this question is redundant. As we have seen before, the syntactical rules of Hebrew predict the use of one or the other preposition and therefore, they have preference over the semantic rules. Let us now verify the equivalents adhering to monovalency. We have shown that from the morphological point of view, monovalent sans is an affix. Recall that sans barbe ‘beardless’ in the phrase sa figure grasse, molle et sans barbe (his fat, lumpy and beardless face) is a lexically complex unit and for that reason, it admits the following notation (cf. Feigenbaum 1997c: 31): (24) [[sans]PREP [barbe]N] ADJ
We shall adopt the same notation for the corresponding collocations in Hebrew. Table 2 states some of the French examples from Quémada (1992) and their Hebrew counterparts from Achiasaf (1999), Bahat and Mishor (1995), Bannitt (1983), Even Shoshan (ibid), Inbal (1990) and Scharfstein (1939). It shows that the most frequent prepositions are beli and lelo. Table 2. A comparison of collocations Hebrew
French
English glosses
beli akifim beli bajot beli kaxal ve srak beli/belo neder beli paxad, libeli xat beli sforot, libeli sforot beli xejk beli xesed bli shemi lelo hafsakah, beli heref lelo pegam, lelo revav lelo sjag
sans détour sans peine sans fard sans promesse sans peur, sans crainte sans nombre sans bornes sans merci sans nom sans cesse sans tache sans fin, sans limites
directly easily bluntly without promise fearless countless without limits merciless unknown endless flawless, immaculate endless
i Zilkha
(1972: 19) mentions bnei beli shem (Job 30.8)
Susanne Feigenbaum
To conclude, these Hebrew expressions are lexically complex units, since they satisfy the features mentioned before: they occupy one syntactic function and they possess a stable reference. Moreover, they are also characterized by referential opacity, especially in idiomatic items, e.g. beli (belo) neder or lelo revav.5 Therefore, the prepositions in these lexically complex units are inserted on the morphological level and the notation for lelo sajag ‘endless’ corresponds to the one given below: (25) [[[le]PREP [lo]ADV] PREP [sajag]N] ADJ
. The value of beli: Two opinions When establishing the list for the Hebrew equivalents, one inevitably asks oneself whether one of the items is more representative than the other. A first comparison concerns syntax. From the description above it appears that beli and belo are the words with the greatest syntactic valency, since they figure in all environments. On the contrary, the two other prepositions are each restricted to one environment alone, for lelo precedes only NPs whereas mibeli precedes only clauses or infinitives. We resume this to be as follows: (26) /lelo [NP], mibeli [INF, P’], beli [NP, INF, P’], belo [NP, INF, P’]/
There are, however, two criteria that distinguish beli from belo. First, belo can be read ‘with no’ before NPs, whereas beli is always read ‘without’. Second, only beli functions anaphorically, in the same way as does sans: (27) a.
Il vaut mieux avec que sans. it is better PREP with than PREP without ‘You are better off with than without.’ b. Adif im measher beli/*belo preferably with than beli/*belo ‘You are better off with than without.’
As one can see in (27), the prepositions sans and avec ‘with’ can be stranded and so can beli and im, whereas belo is excluded in this context. All these are indications that the value of beli as an independent word is solidly established whereas belo appears to conserve its status as a negative clitic. A second means of evaluating the equivalents of sans concerns the lexical level. One of the points that confirms the previous syntactical analysis concerns lexically complex units. Although the twelve idiomatic types we have extracted from the dictionaries represent a relatively small corpus, they show that col-
A contrastive analysis of French and Hebrew prepositions
locations with beli are more common than those with lelo (cf. Table 2). To resume, the structural analysis indicates that beli represents the word with the highest value compared to sans. Based on this assumption and the fact that beli seems idiomatic, I suggested in Feigenbaum (1997b) that one would prefer beli to lelo, if one were to be expressive. The choice between beli and lelo cannot, however, rely on structural considerations alone. Tobin (1994: 97–103) studies the two prepositions in the /NP DELETIVE NP/ context, showing that English and Hebrew have similar systems where the feature Semantic Integrality is concerned. Thus, statements using the preposition without as in “I am without a jeep today.” are unmarked for Semantic Integrality, whereas those using the suffix -less like “I am jeepless today” are marked (ibid: 99). According to Tobin, the same distinction can be made in Hebrew between the marked lelo (corresponding to the suffix “-less”) and the unmarked beli (corresponding to the preposition “without”). As one can see from his example cited below, a set is marked for semantic integrality if it is “a single cognitive set perceived in continuous space” (ibid: 100). (28) a.
xultsa beli kaftorim shirt without buttons ‘a shirt without buttons’ b. xultsa le-lo kaftorim shirt to-no buttons ‘a buttonless shirt’
In (28a) the message is that the buttons are missing from the shirt, whereas in (28b), one deals with “a pullover, a shirt made without buttons.” By the way, the distinction between the marked and the unmarked groups corresponds to our analysis of semiotic valencies. In Feigenbaum (1997c) we maintained that trivalency designs a state whereas monovalency is either a quality or a lexical subcategorization.6 Indeed, (28a) is in accordance with our definition of trivalency in French and (28b) with monovalency. This is illustrated by the following comparison of monovalent groups: (29) a.
les sans-logis; ‘the homeless’ b. une chemise sans col ‘a collarless shirt’
Relevant to our discussion is that in the /NP DELETIVE NP/ environment, lelo is the preposition that best corresponds to monovalent sans. Thus, sans logis corresponds to lelo korat gag and sans col to lelo zavaron. Since this seems to be
Susanne Feigenbaum
in contradiction with our initial belief that beli has the greatest value compared to sans, we decided to reexamine our analysis with the help of a larger corpus, to which the following sections will relate.
. Verification of the data For this paper, approximately 8000 examples with the Hebrew prepositions in question were analyzed from the daily Israeli newspaper Haaretz, published during the years 1990 and 1991. Thus, the former findings could be specified and modified. . Quantitative specifications As we can see from Figure 1, lelo is by far the most frequent preposition (ca. 4000 lelo = 50%, 1700 beli = 22%, 1300 mibeli = 16%, 1000 belo = 12%). It is also interesting to note that the Hebrew prepositions are most typically found before an NP (81%), whereas the quantitative analysis made for sans in Feigenbaum (1996) for the same type shows that the ratio is only 43%. In this respect, the high ratio of lelo gains additional weight. As we have seen, lelo is restricted to NPs and one may therefore say that lelo is the most typical representative of the deletive prepositions in its most natural environment. The quantitative verification for prepositions followed by NPs confirms this hypothesis, for the results show that the general proportion for lelo holds for this environment, too (3997 lelo = 67%, 1098 beli = 19% and 836 belo = 14%).7 These facts give some indication concerning markedness. The common assumption is that markedness is correlated with low frequency. We therefore assume that in the /NP deletive NP/ environment, beli is marked with respect to lelo, and belo is marked with respect to beli and lelo. BELO 12 % MIBELI 16 %
LELO 50 %
BELI 22 %
Figure 1. Haaretz 1990/1991 – Quantitative distribution.
A contrastive analysis of French and Hebrew prepositions
. Another look at lexically complex units Our original hypothesis concerning the expressiveness of beli is based on its being the largest set in the list of idioms and collocations. In this analysis, we checked whether an idiom registered for beli also figures with lelo and vice versa. It appears that the dominance of beli does not withstand reality. In fact, more and more idioms accept all the prepositions capable of figuring before a NP. As an example, we have found examples where the idiom beli kaxal ve srak ‘without ornaments’ is also preceded by lelo and belo. Such findings indicate that the use of lelo is more productive than is normally shown by dictionaries. Still, there is something to be said for beli, as long as one distinguishes tokens and types. In order to verify this difference, we have extracted more than sixty types that can be defined as lexically complex units and that qualify for monovalency. It is true, on the one hand, lelo is in the majority where the tokens of these groups are concerned. This may be seen in the following list, where the ten most frequent lexically complex units are arranged according to their rate of recurrence. Here, lelo takes over: (30) lelo safek ‘doubtless’, beli kesher ‘without any relationship’, lelo heref ‘incessantly’, lelo dixui ‘immediately’, lelo pega ‘flawless’, lelo heker ‘unconsciously’, lelo ho’il ‘without use’, lelo ma’ane ‘without solution’, lelo sjag ‘endless’, lelo dofi ‘faultless’8
Indeed, the result of the token analysis summed up in the total ratio of 60 to 40 in favor of lelo compared to /beli–lelo/.9 On the other hand, with the focus on the analysis of types, lelo remains in the minority. We have divided the monovalent units extracted from the Haaretz study into three groups, named ‘any’, ‘beli or belo’, and ‘lelo’ (cf. Figure 2): a. Any: those units that behave like beli kaxal ve srak, admitting any preposition, including lelo and belo. To this group belong half of the types listed.
15 % lelo
46 % any 36 % beli_belo
Figure 2. Distribution of Hebrew prepositions with lexically complex units.
Susanne Feigenbaum
b. Beli or belo: those that go only with beli or belo, like beli mesim ‘unintentionally’.10 This group represents more of a third in the list. c. Lelo: those that go only with lelo, like lelo revav ‘spotless’.11 They make up 15% of the list. Taken together, the types with beli or belo from the first two groups constitute 85%. This means that the majority in this type class is higher than the one of the lelo unit in the token analysis.
. Evaluation I think that an empirical analysis, like the one used with this corpus, is helpful and necessary in order to verify a wide range of factors that influence the use of one or the other preposition. Thus it appears that the hypotheses concerning the correlation of lelo with monovalency, on the one hand, and that of beli with expressiveness, on the other, is not a clear-cut case. Let us look again at the role of lelo in groups like ‘pitiless’ or sans merci, which are monovalent. The question we asked ourselves was whether the use of lelo in this respect is systematic and whether the Hebrew equivalent for such an evidently idiomatic expression belongs to group c) in the type division made above. The answer is not entirely consistent with our expectations, as stated in (31): (31) a. lelo raxam = 11 items b. lelo raxamim = 1 item ‘pitiless’ c. belo raxam = 2 items ‘pitiless’ d. beli raxamim = 3 items ‘without pity’
The relationship of /12 : 3/ is in favor of the idiom with lelo compared to the analytic expression with beli. Still, the same noun raxam admits belo, although much less often. Therefore, one must admit that the use of lelo is not totally systematic. Another reflection concerns the role of lelo in lexical subcategorization, illustrated in (29). Indeed, when speaking of a masok lelo tajas ‘pilotless aircraft’, one refers to a type of plane. On the contrary, if one were to use masok beli tajas, one would refer to a plane without a pilot. (Clearly, one would rather
A contrastive analysis of French and Hebrew prepositions
see the former in the air than the latter!) It is true that this distinction is productive, yet it is not entirely systematic either. First, one finds cases with lexical subcategorization or qualities, where beli is used: (32) a.
benzin beli oferet ‘leadless petrol’ b. bnei-beli-dat ‘persons without religion’ c. bnei-beli-gil ‘ageless persons’
Second, the overwhelming use of lelo before NPs embraces also appositions which indicate states: (33) Ha mifgash mijuad le bogrim lelo bnei zug. ‘The reunion is intended for veterans without their spouses.’
Third, in some cases the difference in meaning between the two is not as firm as in (28). At least this is so in (34), where the creditor will be disappointed in any case: (34) check beli kisui/check lelo kisui ‘a check without security’
Finally, a word concerning the rhetoric effect due to its Semantic Integrality. Tobin (1994: 101) claims that marked expressions with lelo have a stronger rhetorical effect than the unmarked ones with beli. This idea is confirmed by a wellknown slogan used by the City of Tel Aviv these days that states that Tel Aviv is a place working around the clock: (35) Tel Aviv, ir lelo hafsakah ‘Tel Aviv, town without a pause’
Still, when citing this slogan, people might replace lelo with beli (for instance referring to Tel Aviv as ha ir beli hafsaka ‘the town without a pause’). With regard to the use of beli in idioms, we have seen that the quantitative analysis is ambiguous. Beli is regressing as a token, yet it is dominant as a type. From this point of view, the examples below with beli are significant because they represent idiomatic language, connotating familiarity (the word taxles, for instance, comes from Yiddish). (36) a.
beli rosh u’beli raglaim without head and without feet ‘without rhyme or reason’12
Susanne Feigenbaum
b. Tafsu otanu beli mixnasaim. ‘They caught us without pants.’ c. Ele devarim beli taxles ‘These are words without foundation.’
Thus, if one would like to preserve a popular connotation, one is likely to choose beli. Another example may illustrate this. Take the French un(e) sans cervelle ‘a mindless person’. Normally, one prefers to translate it into Hebrew with a synonym, as in ‘a naïve woman’ (petiah) or a ‘hollow minded person’ (nevuvat moax). Yet, there exists a literal translation with beli, which has a popular ring: (37) Hu xai im beli-moax. he lives with without-brains ‘He lives with a mindless person.’
Finally, the role of belo should also be discussed, especially when it comes to register. We have seen that in the written corpus of the Haaretz newspaper, lelo is the largest set, whereas belo is used much less than any other prepositions (cf. Fig. 1). At this point, it will be interesting to consider that the text we took is dated 1990–1991, a period where lelo was considered to be “bad style”.13 From this perspective, the use of belo becomes marked not only because of its low frequency but also because of its high style. To begin with, it is associated with legal language, as the following examples show: (38) a.
mahalakh belo takdim ‘a move without precedence’ b. belo nohal maavar gvul ‘without a procedure for frontiers’ c. lehofia b’bet mishpat belo orekh din ‘to show up in court without a lawyer’
Next, we often find belo instead of lelo for lexical subcategorization. Thus, the expression lelo ribit (23c) is synonym to (39a), lelo kisui (34) to (39b): (39) a.
malveh belo ribit ‘a loan without interest’ b. belo kisui u’belo sikui ‘without security and without chance’
A similar association exists between (40a) and (40b):
A contrastive analysis of French and Hebrew prepositions
(40) a.
manhig lelo xazon ‘a leader without foresight’ b. ha’ish belo t’khunot ‘the man without talents’
The same may be said for lexical complex units in adjectival positions. The equivalent for ‘without response’ is mostly lelo maneh, but one also finds belo maneh (the rate is 15 : 1): (41) Hem notru belo maaneh u’beli hitjaxasut. ‘They remained without response and without consideration.’
And for sans nom (cf. 8a), beli and belo appear: (42) a.
benei beli shem people without a name ‘anonymous people’ b. sfinot belo shem u’belo kol ships without a name and without a sound ‘unknown and silent ships’
In short, belo transmits a literary or learned style, so that sayings like the following obtain a nearly proverbial status: (43) Aruxa belo kinuax kmuah ke isha balat ain axat ‘A meal without desert equals a woman with one eye.’
. Concluding remarks The main purpose of this paper was to establish a correspondence between the plurivalency of French sans and that of its Hebrew equivalents. The research that had been previously carried out on French has shown that sans is not always a deletive, but that it is also used in a reduced way, that is, like a negative morpheme. It has also indicated a correlation between this negative aspect and deletive or negative affixes in other languages. Since Hebrew possesses several words that have the meaning of sans, we wanted to confirm the hypothesis that this language, too, differentiates valencies by using different morphemes. Our findings are, albeit, not uniform. We believe that this is so because Modern Hebrew, although it is a well-established language, is still undergoing dynamic changes. For one, there is a tendency to connect specific words to specific syntactic surroundings, like mibeli to clauses and infinitives and lelo to NPs. These
Susanne Feigenbaum
words, then, become more and more productive, especially where lelo is concerned.14 But there is also the fact that the words beli and belo, which are etymologically well established, compete with these productive uses. We have concentrated on NPs in order to substantiate a distinction between beli, belo and lelo. It may be summarized as follows: – – –
In the case of monovalency related to subcategorization, there exists a tendency to use lelo. In the case of monovalency related to idioms, beli is preponderant. For a chosen register with literary effects, or in legal language, one tends to prefer belo.
Notes . From the newspaper Le Matin (11.02.97). . Renacle (1956: 368–371) observes a specific use of the group /SANS + infinitive/, where the meaning is equivalent to /NON + participe passé/. Accordingly, the expression in the Wallon dialect “I magnèt do boûre sins saler” corresponds to the standard French “ils mangent du beurre non salé”. . For instance, zemax beli jaaseh kemax = ‘a plant which does not produce flower’ (Joshua, 8; 7). This is an example of its Biblical use as a negative adverb. . We find it in the idiom libeli arokh ‘without limits’. . The word neder means “vow” while in beli neder it means ‘without a binding promise’. Similarly, the word revav means “fat” while in lelo revav it means “flaw”. . In Feigenbaum (1997c) we checked English and German affixation related to sans. . A separate analysis of articles published on the Internet (Snunit 1996) shows that all NPs are preceded by lelo. . Also on TV, 10/12/2000: “lelo heref ” by B. Nethaniahu. . I think it is also interesting to notice that when a preposition takes over, its counterpart is far behind. Thus, on the one hand, lelo safek takes up 78% of all the safek expressions but on the other hand, beli (belo) kesher takes up 92% of all the ones with kesher. . There exists also mibeli mesim. This idiom is an exception to the rule that mibeli figures only before phrastic complements. . Here is a new example: haim tsnuim lelo revav ‘a modest, immaculate life’ (by Prof. Kazir, on the radio station Kol Israel, 29/5/2000). . In German = ohne Hand und Fuss; in French = sans queue ni tête. . I would like to thank Asher Stern for the information that on November 14, 1994, the Academy of the Hebrew Language decided not to forbid the use of lelo.
A contrastive analysis of French and Hebrew prepositions . It has no separate entry in Even-Shoshan (1985). It figures, however, in Achiasaf (1999) and in the Hebrew dictionary of Word2000.
References Achiasaf, O. (1999). Dictionnaire pratique bilingue français-hébreu. Rosh Hayain: Prolog. Azar, M. (1995). The Syntax of Mishnaic Hebrew. Jerusalem: The Academy of the Hebrew Language. Bahat, Sh., & Mishor, M. (1995). Dictionary of Contemporary Hebrew. Jerusalem: Maariv. Banitt, M. (1983). Dictionnaire Français-Hébreu. Jerusalem: Massada. Cadiot, P. (1997). Les prépositions abstraites en français. Paris: Colin. Corbin, D. (1997). Locutions, composés, unités polylexématiques: lexicalisation et mode de construction. In M. Martins-Baltar (Ed.), La locution entre langue et usages (pp. 53– 101). Paris: Ophrys. Even-Shoshan, A. (1985). Hamilon Haxadash. Jerusalem: Kiryat Sefer. Feigenbaum, S. (1996). Le connecteur sans. In Cl. Muller (Ed.), Dépendance et intégration syntaxique (pp. 292–298). Tübingen: Niemeyer. ——— (1997a). Plurivalence sémiotique et polysémie: le cas de sans. Semiotica, 115 (3/4), 361–380. ——— (1997b). Une comparaison entre sans et ses équivalents en hébreu. Journée d’Etude de Linguistique contrastive Français-Hébreu; 5/1/97, Tel Aviv University, oral communication. ——— (1997c). Le mot composé sans + SN: ses rapports syntaxiques, sémantiques et morphologiques. Travaux de Linguistique, 34, 21–47. Glinert, L. (1989). The Grammar of Modern Hebrew. Cambridge: C.U.P. Inbal, Ch. (1990). Dictionnaire Français – Hébreu; Hébreu – Français. Jerusalem: Zack. Karttunen, L. (1971). Discourse Referents. Bloomington: Indiana University Linguistics Club. Mainguenau, D. (1994). Précis de Grammaire pour les Concours. Paris: Dunod. Muller, Cl. (1991). La négation. Genève: Droz. Peirce, Ch. S. (1931–1958). Collected Papers of Charles Sanders Peirce. Ch. Hartshorne, P. Weiss, & A. W. Burks (Eds.), Vols 1–8. Cambridge, MA: Harvard University Press. Quémada, B. (1992). Trésor de la langue française. Paris: Gallimard. Renacle, L. (1956). Syntaxe du parler wallon de la gleize. Paris: Les Belles Lettres (Bibliothèque de la Faculté de Philosophie et Lettres de l’Université de Liège, CXXXIX). Riegel, M. (1985). L’adjectif attribut. Paris: Presses Universitaires de France. Scharfstein, Z. (1939). Ozar Hamilim vhanivim. Tel Aviv: Shilah. Snunit. (1996). Education Information System. Jerusalem: Hebrew University. Tobin, Y. (1994). Invariance, Markedness and Distinctive Feature Analysis. Amsterdam: John Benjamins. Wierzbicka, A. (1972). Semantic Primitives. Frankfurt: Athenäum. Zilkha, A. (1972). Negation in Hebrew. Michigan: Ann Arbor University Microfilms.
A language in change Declined prepositions in spoken Modern Hebrew as a case study Inbar Kimchi-Angert The Department of Hebrew Language, The Hebrew University of Jerusalem
In this fieldwork based research, 20 declined prepositions of spoken Modern Hebrew were analyzed. 57 native speakers from the middle socioeconomic class were recorded while reading and creating declined prepositions. The declensions were divided into normative and non-normative declensions. I checked their frequency and tried to find the motives for using non-normative declensions. The analysis of the declensions by means of gender, age, ethnicity, education, residence and family relations showed that in many cases there is a connection between the form of the declined preposition and the demographics of the speaker. I have also found that not only are there differences between subjects, but there are also cases in which the same subject declined a preposition in two different forms, showing confusion and a lack of differentiation between the forms. Dissimilation and analogies of different kinds influence the form of declined prepositions in the spoken language. Some of the non-normative forms are known from earlier periods of the language, others are a creation of Modern Hebrew. Based on the findings from the fieldwork and on other data collected from the media, I showed that declining prepositions in a non-normative form is a phenomenon related neither to a lack of knowledge of the language nor to a lack of education. On the contrary, the changes become popular amongst the educated people. We witness nowadays a use of new forms that though they differ from the normative declension they are logical and coherent and are probably not about to disappear from the language.
Inbar Kimchi-Angert
Introduction The morphological aspect of declined prepositions in Modern Hebrew is rather unexplored territory. Many studies of prepositions deal with the syntax or the semantics of the prepositions. In this paper the issue is different. The normative declensions of Hebrew prepositions were contrasted with linguistic processes that cause shifts and changes in some of those declensions in the spoken language. Several new declensions can be found in the non-formal written language as well. However, they have not been accepted yet by the Academy of the Hebrew Language. Hebrew prepositions decline in two ways, as shown in Table 1:1 Table 1. The distribution of prepositions according to the declension system A: declension like nouns in singular identical to the noun
similar to the noun
etsel, biglal, be‘ad, biwvil, mul, min, ‘al-yad
et, be, le, ‘im, wel
B: declension like nouns in plural
C: special prepositions
axarey, el, bli, lifney, mittaxat, ‘al
bein, kmo
Group A is of prepositions that decline like the noun in its singular form, while group B is of prepositions that decline like the noun in the plural form. Group C presents special prepositions that have two normative declensions. Some examples will demonstrate this division: Group A, singular: ‘a friend’ – xaver, ‘my friend’ – xaver-i, and in much the same way: ‘chez’ – etsel, ‘chez moi’ – etsl-i, ‘to’ – le, ‘to me’ – l-i. Group B, plural: ‘friends’ – xaver-im, ‘my friends’ – xaver-ay, and in much the same way: ‘after’ – axarey, ‘after me’ – axar-ay, ‘to’ – el, ‘to me’ – el-ay. Group C: ‘between’ – bein, ‘between them’ masc. – bein-am (singular) as well as bein-ey-hem (plural); ‘like’ – kmo, ‘like you’ pl. masc. – kmo-xem (biblical) as well as kmot-xem (mishnaic). As if the situation is not complicated enough, there is also a sub-division in the category of the singular (group A): prepositions that decline exactly like the noun versus those that decline similarly to it. For example, in the second person singular feminine:
Declined prepositions in spoken Modern Hebrew
‘your friend’ – xaver-ex, and in much the same way: ‘chez toi’ – etsl-ex, while ‘to you’ – l-ax. In the nominal system, which is seemingly logical, it is possible to see when the noun is singular and when it is plural. But as far as prepositions are concerned, the division into categories and sub-categories is arbitrary and illogical for the average speaker. The situation invites shifts from one declension to another, thus creating non-normative declensions. This phenomenon is not new, and is known from earlier periods of the language.2 The variety of declensions that can be found in earlier texts was reviewed elsewhere (Kimchi-Angert 1999: 11–26). The spoken language is the focus of this study, because linguistic changes start in the spoken language, and it is only later that they enter the written language (Blanc 1989). The morphological aspects of prepositions in Hebrew have been neglected, especially in Modern Hebrew. I have found only two studies that focused on this issue. One is Glinert (1989: 497–500), which discusses it in the last chapter of his book on Modern Hebrew. However, he bases his discussion on findings gathered from only 6 subjects, a sample size that is too small for a statistical analysis. The other is Ravid (1996), who studied only 5 prepositions, all of them in the 3rd person singular masculine. Many researchers who have mentioned the morphological aspect treated all declensions that divert from the normative way of declension as mistakes, and therefore as not worthy of discussion.3 Since Modern Hebrew is relatively young and vulnerable to change and since immigrants are a large portion of the population, the importance of the norm in Israel is enormous. My starting point was synchronic and descriptive, not judgmental. I wanted to see which normative and non-normative declensions are in use, and to what extent the use is related to the demographics of the speaker. When there is a norm, and there is a norm in Hebrew, subjects will not report accurately on their way of speaking. However, when a subject uses a non-normative form, this use constitutes a support to the existence of a non-normative form. When there is a consistency between informants, there is more support to the existence of the non-normative form (Aissen 1992: 9–10). According to Rosen (1953: 4–6), norm is what the linguists see as correct, while standard is what people who know the language use. My goal is to confront the norm with the standard use.
Inbar Kimchi-Angert
Method Participants The sample was comprised of 57 native speakers. The demographics of the sample was as follows: Gender: 28 men, 29 women. Age: 5 between 18–21 (∼army service), 27 between 22–30 (∼students), 19 between 31–50 (∼parents), 6 between 51–70 (∼elderly). Education: 3 did not finish high school, 11 with high school diploma, 43 studied in an institution of higher education. Ethnicity: 19 of eastern origin (sepharadi), 38 of western origin (ashkenazi). Residence: 42 in the city, 15 in kibbutz or moshav. As such, the sample is not representative of the Israeli population at large. For example, 75% of the subjects have continued their studies beyond highschool, and of course this is not the general situation in Israel. Since the socioeconomic class of the sample is higher than average, it is reasonable to assume that shifts that were made by the subjects exist also in lower socio-economic parts of the society.
Fieldwork The subjects were requested to read aloud paragraphs that were written in a conversational style and had declined prepositions in them. Written Modern Hebrew does not signal all vowels, therefore this was an opportunity to see shifts of vowels.4 Shifts of this kind appear when a preposition in singular (group A) is moved between sub-categories, as explained before. This is also the case when there are analogies and dissimilations, as in the prepositions el – ‘to’: eley-xem v. aley-xem (=‘to you’ pl. masc.), min – ‘from’: mimmenn-o v. mimmenn-u (=‘from him’), kmo – ‘like’: kamo-xem v. kmo-xem (=‘like you’ pl. masc.) etc. The subjects also had to rephrase sentences, and by doing that, to decline prepositions. For example, rephrase a as b: (1) a.
yawav-ti mul ha-mera’ayn-im. Sat.I in.front.of the.interviewers ‘I sat in front of the interviewers.’ → b. yawav-ti mul-am. Sat.I in.front.of.them ‘I sat in front of them.’
Declined prepositions in spoken Modern Hebrew
The whole procedure was recorded on tape, and analyzed later. Declined prepositions were also randomly collected from the radio and television in order to expand the database. Most of the speakers were reporters and politicians.
Results and discussion5 Shifts between and within the two main categories Shifts between the two main categories of declension exist, as well as shifts between the two sub-categories of the singular declension. Both kinds of shifts were not found to be very common, and some of them are known from earlier periods of the language. The shift within the sub-categories of group A from similar to identical, like the non-normative itt-ex instead of the normative itt-ax (‘with you’ sing. fem.), is related to a low socio-economic class (Ravid 1996: 189). The socioeconomic class of the sample used in this research was higher than average, hence the almost complete absence of this phenomenon from the result of this study. The shift from group B to group A, i.e. from plural suffixes to singular suffixes was found to be subject-dependent. Three men in their twenties were relatively consistent in making that switch in the case of the 2nd person singular feminine. Two prepositions, bli and mittaxat, had in earlier periods a mixed declension of both singular and plural suffixes in some of the persons, therefore they had a higher tendency to be found in the non-normative form when declined by other subjects as well. The shift from group A to group B, i.e. from singular suffixes to plural ones has not been mentioned before in studies of Modern Hebrew. This shift is relatively rare, though it is attested already in medieval Piyyut (Yeivin 1996: 106–107). However, evidence for its existence was found in the case of non-normative ba‘ad-av instead of the normative ba‘ad-o (‘for him’ sing. masc.), as well as in several other cases.
Analogy to the paradigm I have examined the tendency of speakers to make an analogy to the paradigm of a declined preposition. In other words, to use the first syllable that is used
Inbar Kimchi-Angert
in most declensions of a preposition, while declining the irregular form. Let us take for example the preposition min (=‘from’). The normative declensions of this preposition are presented in Table 2.6 Table 2. The normative declensions of min Person
Singular
Plural
1st 2nd masculine 2nd feminine 3rd masculine 3rd feminine
mimmen-ni mimm-xa mimm-ex mimmenn-u mimmenn-ah
mimmen-nu mikk-em mikk-en me-hem me-hen
In the 2nd person plural the normative form is mikk-em (=‘from you’). This form is different from most of the others in the paradigm, because it does not start with the syllable mim like mimmenn-i, mimm-xa etc. This is the cause of the development of the non-normative form mim-xem that starts with mim.7 The distribution of the declensions is as seen in Figure 1. About one third of the subjects created the normative declension, almost half created the non-normative declension, and about one fifth created both declensions. A specific analysis of the distribution by demographic variables has been done, and the following results have emerged: The distribution by age is presented in Figure 2. 50
40
30
non-normative
20
both normative
%
10
0
Figure 1. The use of different declensions of min 2nd person plural.
Declined prepositions in spoken Modern Hebrew
80
60 non-normative %
both
40
normative 20
0 18–21
22–30
31–50
51–70
Figure 2. The use of different declensions of min 2nd person plural by age group. 100 80 60
non-normative
%
both
40
normative
20 0
high school
more than high school
Figure 3. The use of different declensions of min 2nd person plural by level of education.
The younger the speaker, the greater is the use of the non-normative declension, and vice versa. The distribution by education is presented in Figure 3. Subjects with a higher level of education tend to use the normative declension more than those with lower level of education.
Inbar Kimchi-Angert 80
60
%
non-normative
40
both normative 20
0 women
men
Figure 4. The use of different declensions of min 2nd person plural by gender.
The distribution by gender is presented in Figure 4. We can detect two interesting differences. The first is that men decline the preposition in the non-normative way more than women do. What is the cause of this phenomenon? Studies about the differences between the language of men and the language of women have found that women tend to use a more conservative language (Crawford 1995). There is a clear relation between norm of language and conservative language. The second difference is that women randomly switched between normative and non-normative declensions of the same person more than men did. Again, if we look at studies about gender differences in language, we see that women tend to be unassertive in their way of speaking (Lakoff 1975). We see that in syntactical and semantic uses. For example, women use uncertainty verbs like ‘it seems to be’, and hedges such as ‘possibly’ and ‘somewhat’. They use more tag questions in order to get approval, etc. (Mulac & Lundell 1994: 303). These switches between forms can be explained as a morphological aspect of women’s unassertiveness in speech. In the 3rd person plural we have a similar situation. The normative declension is me-hem (=‘from them’), which is different from all the other declensions of this preposition.8 It starts with me and not with mi like mimmenni, mimm-xa etc. 10% of the subjects that were asked to use the declension of the 3rd person plural created the non-normative form mi-hem. It is my belief
Declined prepositions in spoken Modern Hebrew
that the use of this non-normative declension will become more common in the future. Analogy to the paradigm was found to be quite powerful in other prepositions as well. Non-normative forms are common in the 2nd and 3rd person plural of the prepositions el – ‘to’ and kmo – ‘like’. In the other persons of these two prepositions all declensions starts with el- and ka- respectively, while the normative declensions of the 2nd and 3rd persons are aley-xem, aley-hem, kmo-xem and kmo-hem. The analogy to the paradigm caused the formation of the non-normative forms eley-xem, eley-hem, kamo-xem and kamo-hem.9 Non-normative forms in the 2nd person plural of the preposition et – ‘nota accusativi’ are also being used. In all other persons the declensions start with ot-, while the normative declension of the 2nd person plural is et-xem. Due to the analogy to the paradigm the form ot-xem is widely used.
Dissimilation I have also examined the power of dissimilation. In the preposition min we have the declining form mimmennu that serves for two persons: 3rd person singular masculine and 1st person plural. This situation is very uncomfortable for the speaker, because it invites ambiguity in its use. There have been suggestions in earlier periods to change the vocalization in order to solve the problem, but they were not accepted (Avinery 1964: 354). In the 3rd person singular masculine two problems exist. The first is the ambiguity that is mentioned above, and the second is the fact that other declined prepositions in this person end with o like b-o (=‘in him’), etsl-o (=‘chez lui’), and not with u like mimmenn-u.10 80% of the subjects in the sample read and created the non-normative form mimmenn-o, thus solving both problems of the normative form. In the 1st person plural all the subjects (100%) created the form me’itta-nu, which is a combination of two prepositions, the first of them is min, and the second is et, which is the ‘nota accusativi’ in Hebrew. The unique non-normative declension me’itta-nu is the only one that had been accepted by the normative linguists. It seems that the majority of the speakers preferred to solve the ambiguity problem by changing both declensions.
Analogy to the preposition ‘le’ I have examined the influence that the preposition le (=‘to’) has on other prepositions that end in the same consonant, i.e. ‘l’. All those prepositions are
Inbar Kimchi-Angert 80 70 60 50 %
non-normative 40
both
30
normative
20 10 0 chez vous
chez eux
for you (pl.)
for them
Figure 5. Etsel and biwvil as declined in 2nd and 3rd person plural.
part of the singular category, and include the following: biglal – ‘because of ’, mul – ‘in front of ’, etsel – ‘chez’ and biwvil – ‘for’. In the 2nd and 3rd person plural, the declension of le – ‘to’ is la-xem and la-hem, respectively. The normative declension of the other group of prepositions is biwvil-xem and biwvil-am. The identical suffixes of the declensions in other persons, as well as the final consonant of the prepositions, influence the suffixes of some of those prepositions, thus creating the non-normative declensions biwvila-xem and biwvilahem.11 This shift is a creation of Modern Hebrew. I have found this influence in two of the prepositions – etsel and biwvil. The use of different declensions of etsel and biwvil in the 2nd and 3rd person as found in the sample is as seen in Figure 5. The shift is more common in etsel than in biwvil, but even in the latter at least 25% used the non-normative declension. Usually, the higher the age and the education of the subject, the higher was the use of the normative declensions. Many subjects used both normative and non-normative declensions. This can be explained in two ways: First, a great confusion and uncertainty exist when trying to determine which is the “correct” form. Second, a viewing of the two forms as equivalently acceptable takes place. What characterizes the subjects that used both declensions? There were more women than men and more subjects of eastern origin than of western origin. I have mentioned above the possible relation between women’s unassertiveness in speech and the use of both declensions. McMillan et al. (1977: 558) explain women’s lack of as-
Declined prepositions in spoken Modern Hebrew
sertiveness as a reflection of their lack of self esteem, which could be the result of considering themselves as an “unequal” sector of the society. It may be that these factors prevail in Hebrew speakers of eastern origin, and hence their use of both normative and non-normative declensions.
The plural suffixes ‘m’ and ‘n’ Another issue I have examined was the use of ‘n’ for 2nd and 3rd person plural feminine. In order to clarify this point, here are the normative declensions of the preposition biglal in these persons: ‘because of you’ (pl. masc.) – biglal-xem, while ‘because of you’ (pl. fem.) – biglal-xen, and in much the same way ‘because of them’ (pl. masc.) – biglal-am, while ‘because of them’ (pl. fem.) – biglal-an. The morphological differentiation between masculine and feminine is one of the main characteristics of Hebrew, in personal pronouns, in nouns, in verb conjugation as well as in declined prepositions, and it has other suffixes besides ‘m’ and ‘n’. In cases of mixed groups the masculine form is used. In our case, the ending ‘m’ for prepositions in plural masculine versus the ending ‘n’ for those in plural feminine differs only in the point of articulation, so the switch is easy. Shifts between the two suffixes are known from the Bible as well as from the Mishna (Yalon 1964: 77; Epstein 1964: 1228–1230). Two demographic variables were found to be significantly related to the use or non-use of ‘n’ for plural feminine (p < 0.05 in T test). The first is gender. Women created the feminine suffix significantly more than men did. Although there are no substantive data on this phenomenon, one possible explanation of the gender difference is by the particular significance this ‘n’ has for women. It is a linguistic marker that is a part of their identity as women. The second variable that is related to the use or non-use of ‘n’ for plural feminine is origin. Subjects of western origin created the plural feminine suffix significantly more than those of eastern origin. Since there was no relation of the kind with other demographic variables, the difference can be explained by variables that were not checked in this study, like the education and the socio-economic class of the subjects’ parents.
The most common changes in declined prepositions After analyzing the results of this study, I am able to present the most common changes in declined prepositions. The processes that take place in the creation and use of non-normative declensions are presented in Figure 6. It is clear that the dissimilation is the most influencing process on morphological changes.
Inbar Kimchi-Angert 100 90 80 70 60 % 50 40 30 20 10 0
etsel-chez
oxtem-nota accus. you (pl.) mimxemfrom you (pl.) kamoxemlike you (pl.) mimmenofrom him eleyxem-to you (pl.) me’ittanufrom us
Biwvil-for
dissimilation both dissimilation and analogy analogy to the paradign analogy to the final consonant
Figure 6. Linguistic processes found to be most productive.
Next in its importance is the force of an analogy to the paradigm of a declined preposition. A combination of dissimilation and analogy to the paradigm is stronger than the analogy alone. This phenomenon can be explained by the vast power of the dissimilation. Finally, another influencing process is an analogy to another preposition. This analogy takes place by the final identical consonant of the prepositions involved. No relation was found between the way parents declined prepositions and the way their children declined them.
Conclusions In this study I have identified and analyzed morphological changes in declined preposition in spoken Modern Hebrew. Based on fieldwork, I have shown that speakers of different demographics use non-normative declensions, and that generally it is not possible to attribute this use to a low socio-economic class. The average socio-economic class of the sample is higher than the average in the Israeli society. I believe that in the general population this tendency for the use of non-normative declensions is more pervasive than the results I have shown here. Several demographically related tendencies were found. A relation was found between higher age and education and the tendency to use normative declensions. Subjects of western origin were found to be more consistent than
Declined prepositions in spoken Modern Hebrew
those of eastern origin. Men were found to be more consistent than women on one hand, and tended to use more analogy-based declensions on the other. Both women and subjects of western origin were significantly more strict in the use of the plural feminine suffix ‘n’ in declined prepositions. Not only were there differences between speakers when they declined prepositions in speech, but there were many cases in which a particular speaker used both normative and non-normative declensions. This phenomenon indicates that we are in the middle of a transition in which prepositions, like other components of a language, are going through morphological changes. A living language is in a constant state of change, and these changes have a linguistic logic behind them. It seems to me that they can be an integral part of Modern Hebrew. Thorough fieldwork on the subject has not been conducted until now. Hopefully, this research will contribute to the understanding of the linguistic forces that cause morphological changes in the language.
Acknowledgements This paper is based on part of my MA thesis: Inbar Kimchi-Angert, A Morphological Analysis of Declined Prepositions in Spoken Modern Hebrew, The Hebrew University of Jerusalem, 1999, written under the guidance of Dr. S.E. Fassberg. I would like to thank Dr. S.E. Fassberg, Alon Angert and Sigal Bekker for their much-appreciated help.
Notes . There are a few prepositions that do not decline, like lamrot (=despite), bewel (=owing to) and terem (=before), but most prepositions do. . See Gesenius (1910: 300–305) and Joüon (1993: 336–347, 686–687) for the biblical period, Qimron (1976: 229–248, 284–285) for Dead Sea Scrolls, Segal (1936: 162–165), Yalon (1964: 13–15, 25) and Kutscher (1972: 16–21) for mishnaic period, and Yahalom (1978, 1985) and Yeivin (1996: 106–107) for the payyetanic period. . Such as Alcalay (1951), Sivan (1969) and Rabi (1994). . However, some subjects had changed not only the vowels but added a consonant in the case of normative biwvil-am changed into non-normative biwvila-hem. For a possible explanation of this change see results and discussion.
Inbar Kimchi-Angert . The results of the declensions in the plural persons are presented as masculine persons only in order to simplify the presentation. The issue of masculine and feminine suffixes (m and n respectively) is discussed below. . See Gesenius (1910: 303) and Joüon (1993: 342) for a review of the creation of each declension. . Due to the lack of dagew, i.e. the gemination of a consonant, the tenuis form of the consonant is used, in this case x instead of k. . The reason for the different vowel, e instead of i, is a compensatory lengthening, a linguistic process that is not very strong nowadays due to the phonetic loss of dagew in most consonants of Modern Hebrew. . The normative forms aley-xem and aley-hem are phonetically identical to the 2nd and 3rd person plural of ‘al (=‘on’). Therefore, dissimilation is another process that causes the shift in el, from normative to non-normative eley-xem and eley-hem. . The reason for the unique ending is an assimilation to the 3rd person hu, when ‘from him’ = min+hu. As for the supposed reduplication of the preposition, see Gesenius (1910: 303). . Identical in 1st and 3rd person singular, masculine and feminine.
References Aissen, J. L. (1992). Fieldwork and linguistic theory. In W. Bright (Ed.), International Encyclopedia of Linguistics Vol. 2 (pp. 9–10). New York: Oxford University Press. Alcalay, R. (1951). Linguistic corrections: Comments on correcting common linguistic mistakes. Leshonenu La‘am, 2 (11–12), 5–46 (in Hebrew). Avineri, I. (1964). Yad Hallaschon. Tel-Aviv: The Izreel Publishing House (in Hebrew). Blanc, H. (1989). Human Language. Jerusalem: The Bialik Institute (in Hebrew). Crawford, M. (1995). Talking Differences: On Gender and Language. London: Sage. Epstein, J. N. (1964). Introduction to the Text of the Mishna. Jerusalem/Tel-Aviv: Magnes Press, Vol. 2 (in Hebrew). Gesenius, W. (1910). Hebrew Grammar. E. Kautzsch (Ed.), A. E. Cowley (Trans.). Oxford: Clarendon Press. Joüon, P. (1993). A Grammar of Biblical Hebrew. T. Muraoka (Trans. and revised). Rome: Editricie Pontificio Istituto Biblico. Kimchi-Angert, I. (1999). A Morphological Analysis of Declined Prepositions in Spoken Modern Hebrew, MA thesis. Jerusalem (in Hebrew). Kutscher, E. Y. (1972). Rabbinic Hebrew. In M. Bar-Asher (Ed.), A Collection of Articles on Rabbinic Hebrew, Vol. 1 (pp. 1–35). Jerusalem: The Hebrew University of Jerusalem (in Hebrew). Lakoff, R. (1975). Language and Woman’s Place. New York: Harper and Row. McMillan, J. R., Clifton, A. K., McGrath, D., & Gale, W. S. (1977). Women’s language: Uncertainty or interpersonal sensitivity and emotionality? Sex Roles, 3 (6), 545–559.
Declined prepositions in spoken Modern Hebrew
Mulac, A., & Lundell, T. L. (1994). Effects of gender-linked language differences in adults’ written discourse: Multivariate tests of language effects. Language and Communication, 14 (3), 299–309. Qimron, E. (1976). A Grammar of the Hebrew Language of the Dead Sea Scrolls, Ph.D. thesis. Jerusalem (in Hebrew). Rabi, J. (1994). Improving language without pedantry. In C. Cohen (Ed.), Sound Language (pp. 337–341). Jerusalem: The Bialik Institute (in Hebrew). Ravid, D. (1996). The acquisition of declined prepositions: Psycholinguistic principals. In O. Schwarzwald (Rodrigue) & Y. Shlesinger (Eds.), The Hadassah Kantor Jubilee Book (pp. 184–195). Ramat-Gan: Bar-Ilan University (in Hebrew). Rosen, H. B. (1953). On standard and norm, on processes and mistakes. Leshonenu La‘am, 4 (6), 3–8 (in Hebrew). Segal, M. H. (1936). The Grammar of Misnaic Hebrew. Tel-Aviv: Dvir Publishing House (in Hebrew). Sivan, R. (1969). Better Hebrew Usage. Tel-Aviv: Karni (in Hebrew). Yahalom, J. (1978). A Collection of Geniza Fragments of Piyyute Yannai. Jerusalem: Makor (in Hebrew). Yahalom, J. (1985). Poetic Language in the Early Piyyut. Jerusalem: Magnes Press (in Hebrew). Yalon, H. (1964). Introduction to the Vocalization of the Mishna. Jerusalem: The Bialik Institute (in Hebrew). Yeivin, I. (1996). Characteristic linguistic features of Piyyut. In M. Bar-Asher (Ed.), Studies in Hebrew and Jewish Languages (pp. 105–118). Jerusalem: The Bialik Institute (in Hebrew).
The French preposition in contact with Hebrew Miriam Ben-Rafael University of Tel Aviv
.
Introduction
It is widely agreed that languages change in contact with other languages. The changes they undergo, whether the errors that are introduced in the process of acquiring the second language (L2) or the modifications that occur in the first language (L1) in contact with L2, are systematic and unavoidable. One speaks, in this perspective, of L2 acquisition and of L1 attrition; errors, variations, changes and even attrition are then understood as phases of the same general process, that is, of L1/L2 interaction. Errors accrued in L2 acquisition are often explained as L1 interferences. Other explanations view these errors as the product of learning strategies and as manifestations of a temporary linguistic system – that is, of an interlanguage or IL (Selinker 1992). It is also argued that L2 structures may present specific difficulties which lead to IL errors, unrelatedly to L1 (Frei 1971; Bouton 1969). Moreover, the literature speaks of a tendency of reduction of the LI lexicon in contact situations and of the emergence of codeswitching patterns (Romaine 1989; Jacobson 1998; Meyers-Scotton 1992). As for syntax, some scholars believe that variations from the standard language represent mostly characteristic tendencies of LI itself (Blanche-Benveniste 1990). However, it is recognized that L1 syntactic variations are associated with a propensity to simplification and remain far less numerous than lexical changes. It is in the context of these considerations that this chapter focuses on French/Hebrew contact, and discusses the specific case of French prepositions. Long neglected, the study of the preposition has undergone remarkable development in recent years, especially in French linguistics. Numerous studies
Miriam Ben-Rafael
have offered new classifications and have re-defined the status of prepositions (Cadiot 1997; Borillo 1998; Gaatone 2000; Picabia 2000). Arguing that previous positional and functional definitions of prepositions are inadequate, they see prepositions as relateurs – linkers (Hagège 1996; Cadiot 1997: 18). Scholars question issues such as the morphology and assumed invariability that have long been attributed to French prepositions. They re-analyze the prepositions associated with verbs and their possible alternations in different verbal contexts. Cases in point are the prepositions de and à often found in French with infinitive verbs. According to Cadiot (1997: 60–85), the alternation of à/de is ordered by a specific “aspect principle”, and governed by a number of norms which determine the distribution of these prepositions. Cadiot shares, for instance, a dynamic perception of à vs a static perception of de; moreover, preceding an infinitive, à leans towards the subject, and de towards the object (ibid: 75); the alternation à/de could express continuity vs discontinuity (ibid: 81); de would express a concrete configuration of the object, while à would be a means of abstract or dative configuration (ibid: 77). Gaatone (1998: 189–193) studies prepositional alternation in relation to the passive form. He shows the complexity of the alternation of par and de as well as of à, avec, entre, par les soins de, and dans. The use of de remains relatively rare in the passive mode while par is definitely the privileged preposition here. Gaatone underlines, however, that cognitive verbs (comprendre, connaître, savoir) accept de as well as par. The same, he says, applies to the verbs of perception, companionship and assistance. Among the verbs of acceptance, de remains a possibility, but it is par that is preferred. Hence, there is a whole series of verbs which accept both de and à as possibilities, though one or the other may be preferred (Gaatone 1998: 198–210). Studies have also been carried out on French prepositions in language contacts. For the most part, these studies show a process of simplification. Aub-Busher (1993: 210) points out that the use of French prepositions in a Caribbean French-based Creole is more restricted than in standard French. In this Creole, prepositions are not generally used in adjectival phrases, neither to mark an indirect object complement, nor in locative expressions with certain nouns (l’école/“the school”, instead of à l’école “at school”). Kihm (2000) similarly notes the disappearance of de in genitive constructions in Haitian Creole (manch chimiy la/“sleeve blouse the”, instead of manche de la chemise). The studies also suggest interferential phenomena. In Canadian French, for instance, one finds prepositional constructions called “preposition stranding”, where the preposition is set at the end of the sentence. Durand (1993: 259) of-
The French preposition in contact with Hebrew
fers here several examples: “Quelle heure qu’il est arrivé à?” and “Quoi-ce-que tu as parlé à Jean de?”. These examples may be attributed to English interferences. Scholars caution, however, that similar processes are found in numerous varieties of French in France itself (Durand 1993: 260). Roberge and Vinet (1989: 71) quote, among other instances, “parler avec a toujours été difficile” and “Mitterand j’ai voté pour”. The prepositional system has also been studied in the process of L2 acquisition. Contrastive analysis (CA) of this system in different languages shows that there is often a lack of equivalence, which may create difficulties for learners (Feigenbaum 2000a, 2000b). Duškova (1969: 18) points out the prepositional errors of Czech learners of English, Csecsy (1971) of Hungarian students of French, and Ben-Rafael (1994) of Hebrew speaking students of French. Bouton (1969) emphasizes here that the French prepositional system presents great difficulties for learners, whatever their L1. While in the formation of Creoles the prepositional system tends to be simplified, in other situations of French acquisition, it rather allows the construction of space, localization and movement in L2. Giacobe’s (1992) three-stage research on narratives is eloquent in this latter respect. In the first stage, a micro-system of three prepositions (de, à, en) establishes the relation between the theme and its “relata”. These prepositions are used, whether after a verb of movement (à, de) and localization (en), or in verbless sentences indicating motion. In the second stage, when the IL has already developed up to a given level and includes in its lexicon the L2 prepositions à, de, and en without verbs, two new forms appear – por (= pour) and sur. In the third stage, when the learner already controls the schemes of verbs like sortir and aller, s/he is able to distinguish the relative mobility of elements vis-à-vis each other, and dans, then, appears. As a whole, these considerations show that prepositions are an interesting aspect of linguistic variation in situations of language contact. It is in this context that our own research focuses on the prepositional system when French comes into contact with Hebrew.
. The research setting The population of Israel, a country of immigrants, is of many linguistic origins. The dominant language is Hebrew and as a rule immigrants, including the French speaking immigrants, become bilingual by acquiring this language. This situation accounts for numerous L1 variations, which go into making an
Miriam Ben-Rafael
“Israeli French” or “Franbreu”. Previous studies of “franbréophones” – Israelis of French speaking origin – and of “hébréophones” – Hebrew speaking Israeli learners of French – have found several features common to “Franbreu” and the hébréophone interlanguage (IL) (Ben-Rafael 1993a, 1994). These common features mainly concern syntactic elements such as the verbal system (simplification), verbal valency, gender choice and the omission of determinants. These features often reflect Hebrew structures. Yet, lexical borrowings and codeswitchings, which are most frequent in Franbreu, are almost nonexistent in IL. The present study considers two Hebrew/French contact situations: (1) a formal acquisition situation, involving native Hebrew-speaking learners of French as a second language at school; and (2) a post-immigration situation involving former native French-speaking immigrants who have lived in Israel for about thirty years and have acquired Hebrew as a second language. The study begins with an analysis of the prepositional errors of learners of French. The data of the study have been collected from a large set of syntactic errors in IL produced in writing and orally by high-school students (three classes in a senior high-school in a Tel-Aviv suburb). In a second stage, prepositional occurrences have been collected in the Franbreu of a sample of Frenchspeaking immigrants by means of interviews and by recording spontaneous conversations among them. The aim of this study is to answer the following two questions: (a) to what extent are IL errors and Franbreu variations due to the contact of French with Hebrew, and (b) to what extent can the errors found in IL and in Franbreu be meaningfully compared? In a general manner, both in French and in Hebrew, prepositions connect words or groups of words and set them in a dependency relation, serving as markers of their linguistic functions. For example, in (1a) the preposition à indicates time, en brings in means, and devant signals location; in Hebrew, (1b), these functions are carried out by the two prepositions be and mul: (1) a. à cinq heures son père venait la chercher en voiture devant le lycée b. beša`a xameš ‘avi’a halakh laqaxat ‘ota bemekhonit mul hatikhon ‘At five her father went to fetch her by car in front of the high school.’
At the same time, the meanings of prepositions in the two languages are not always equivalent. In French, as in Hebrew, the same preposition can be used to indicate different functions, depending on the words to which they are linked. For example, avec in French marks three different functions – “how”, “by what means” and “with whom”. On the other hand, the same Hebrew functions would be introduced by be and im:
The French preposition in contact with Hebrew
(2) a. il marche avec difficulté avec un baton avec son père b. hu holekh beqoši im maqel ‘im ‘aviv ‘He walks with difficulty with a stick with his father.’
Similarly, à in French indicates “where”, “how” and “where to”, while in Hebrew these functions are expressed by two different prepositions, be or its variant ba (the contraction of be and the determinant ha) and im: (3) a. à la maison il parlait à voix basse à son ami b. babayit h’u diber beqol namukh ‘im haxaver šelo ‘At home he spoke quietly with his friend.’
Hence, prepositions may have different functions, be the same in one linguistic system and become diversified in the other. The lack of congruence may cause confusion among Hebrew learners of French and among French speakers who are in contact with Hebrew. The following presentation will consider the French prepositions when they are in contact with Hebrew in those two different linguistic situations. . The IL of French learners To first consider IL, prepositional choice in French L2 is often a function of Hebrew L1. The fact that Hebrew and French prepositions are not always equivalent is a most common cause of interferential errors (Ben Rafael 1994), especially regarding the uses of à, avant, dans, pour and de. All this is shown in the following brief contrastive analysis of the French and Hebrew prepositional systems. . A contrastive analysis (CA) of French and Hebrew prepositional systems Table 1 shows the problematics of the use of à, avant, après and derrière. In French, à, its variant au (contraction of à and the determinant le) and en, all express both static and dynamic spatial functions; in Hebrew, two prepositions, be and le, are necessary to distinguish between these two functions. Moreover, in French one distinguishes between avant and après to express a temporal function, and devant and derrière to express a spatial function; in Hebrew, two prepositions are sufficient to fulfill these two functions – lifne and axare. The expression il y a, which some scholars regard today as a preposition (Grevisse 1986: 1504; Picabia 2000), makes the French temporal prepositional
Miriam Ben-Rafael
Table 1. French and Hebrew prepositional systems Fr prepositions
H prepositions (English translation)
j’ habite à Paris je vais à Paris j’ habite au Canada je vais au Canada j’ habite en France je vais en France je suis arrivé avant Kèrèn je suis assise devant Kèrèn je suis arrivé il y a deux heures je suis arrivé après Kèrèn je suis assise derrière Kèrèn
‘ani gar beparis (I live in Paris) ‘ani nose ‘a leparis (I travel to Paris) ‘ani gar beqanada (I live in Canada) ‘ani nose ‘a leqanada (I travel to Canada) ‘ani gar betsarfat (I live in France) ‘ani nose ‘a letsarfat (I travel to France) hig ‘ati lifne qeren (I arrived before Karen) ‘ani yoševet lifne qeren (I am sitting in front of Keren) hig ‘ati lifne ša ‘atayim (I arrived two hours ago) hig ‘ati ‘axare qeren (I arrived after Keren ) ‘ani yoševet ‘axare q (I am sitting behind K)
system even more complex in comparison to Hebrew, which employs here the single preposition lifne. The great majority of the IL prepositional errors involve à, which covers a wide range of meanings in the IL. From the very beginning of French L2 acquisition, à is the preposition that is the most frequent in the learning material at school. The learner encounters à in the most varied contexts and confronts its numerous uses and meanings. In Sans Frontières I (1989),1 for instance, the learner is exposed to sentences such as: J.M. habite à Paris, il est né à Marseille “JM lives in P, he was born in M.”, where he deduces that à is equivalent to the Hebrew be. With other sentences, like Il dit bonjour à M. Lenoir “he says hello to Mr L.”, he assumes that à corresponds to le. Yet we know that à has many other meanings as well. In addition, à may be joined by other prepositions, such as dans, en, chez, which still diversify its meanings. Hence we may expect errors to multiply. . The French preposition à in IL Our data show that à assumes a variety of specific meanings in IL. .. à IL = dans L2 (static) In the following IL occurrences, à and its variants au and aux express a static relation to location, and their counterparts would be be in Hebrew, and dans in French.2 (4) a.
à cette monde nous devons gagner d’argent ‘In this world we need to earn money.’
The French preposition in contact with Hebrew
b. la grève aux hôpitaux est finie ‘The strike in the hospitals is over.’ c. le PLO a envoyé des lettres aux territoires ‘The PLO sent letters to the territories.’ d. j’ai à ma famille trois fils ‘I have three sons in my family.’
The use of à in these examples represents an interference from L1 doubled by an overgeneralization. The preposition à, having been learnt from the beginning as equivalent to be, is transferred to the target language French, in inappropriate contexts. .. à IL = dans L2 (dynamic) In the following occurrences the IL à is equivalent to the French dans. However, this time it indicates a dynamic relation. Its HL1 equivalent is thus le or its variant la: (5) a.
elle n’ a pas retourné à son pays original ‘She didn’t go back to her native country.’ b. ils sont allés à Midi ‘They went to the South.’ c. elles m’ont pris à une boite de nuit ‘They took me to a night club.’
.. à IL = à L2 temporal preposition In French, à may also express a temporal relation, such as: “je vais à l’opéra à midi, à 4 heures, à minuit (from Sans Frontières 1). It is similar to the Hebrew be: ‘ani holekh la’opera batsohorayim, be’arb`a, bexatsot (“I go to the opera at noon, at four, at midnight”). Yet , while both prepositions express a time relation, the French à and the Hebrew be do not completely overlap, again leading to errors. In IL, one thus finds occurrences such as: (6) a.
à première septembre (for: le premier septembre) berišon beseptember ‘September 1st’ b. à vendredi (for: le vendredi) beyom šiši ‘on Fridays’ c. à nuit (for: la nuit) balayla ‘at night’
Miriam Ben-Rafael
d. à le passé (for: dans le passé) ba`avar ‘in the past’ e. au l’après midi (for: l’après-midi) axare hatsohorayim ‘in the afternoon’
À and its variant au are incorrect in the above uses in French. The timefunction that they mark in IL should be expressed in French by the determinants le, la or l’ followed by a noun or by dans. In Hebrew, in contrast, this time relation would be expressed by the preposition be and its variant ba or by ’axare (= après). .. à IL = chez L2 Another category of errors involving à in IL consists of its use in place of the French preposition chez: (7) a.
les enfants sont allés à leur oncle (for: chez) hayeladim halkhu ladod šelaem ‘The children went to their uncle.’ b. il est arrivé à l’ami de son père (for: chez) hu higy‘a laxaver šel ‘ab’a šelo ‘He arrived to the friend of his father.’
In these examples, à expresses a change of location and movement. The IL à is an equivalent of the Hebrew le. In (6) and (7), we see that the IL preposition does not change with the destination of the movement, which in this case concerns persons. À, then, is the equivalent of Hebrew le. In FL2, in contrast, it is chez which is used, in the case of a human destination. Moreover, IL à is again a substitute of French chez in (8a) and (b), which describe static situations but in which the locations indicated by the prepositions are still of a human nature. In Hebrew, this type of relation is signaled by ’etsel (= chez) as in French, and this means that the use of à here cannot be viewed as a Hebrew interference: (8) a.
il vit à ses parents (instead of: chez) hu xay ‘etsel hahorim šelo ‘He lives with his parents.’ b. il mange à sa sœur (instead of: chez) hu ‘okhel ‘etsel ha’axot šelo ‘He eats at his sister’s place.’
The French preposition in contact with Hebrew
.. à IL = pour L2 À IL is also equivalent to the French pour: (9) a.
c’était un homme qui avait un télégramme à Daniel (instead of: pour Daniel) ze haya ‘iš šehaya lo mivraq bišvil Daniel ‘This was a man who had a telegram for Daniel.’ b. c’est un cadeau à mon ami (instead of: pour mon ami) zot matana bišvil haxaver šeli ‘It’s a gift for my friend.’
À may then express a function of destination in IL, similar to pour in French and bišvil in Hebrew. It is often understood as the counterpart of le or la whether movement is involved – as in je vais à l’opéra (“I go to the opera”) and il va à Tel-Aviv (“he goes to T.A”) – or when it is spoken of a human destination – as in il dit bonjour à la concierge (“he says hello to the housekeeper”) or il donne la lettre à Pierre (“he gives the letter to Pierre”). .. à IL = sur, de L2 À IL may also be understood as the equivalent of the French preposition sur (“on”) or au sujet de and de (“about”) or of the Hebrew preposition al (“on”): (10) a.
le premier film est à une fille haseret harišon ‘al bat ‘The first movie is about a girl.’ b. ils sont en colère à Israël hem qo ‘asim ‘al yisra’el ‘They are angry with Israel.’ c. ils parlent à le livre hem medabrim ‘al hasefer ‘They speak about the book.’
.. à IL = be L1 + abstract noun Moreover, à may also be used as a substitute for the Hebrew L1 be when it introduces an abstract noun: (11) a.
il a fait à façon bizarre (instead of: de façon bizarre) hu Áasa be’ofen muzar ‘He did oddly.’
Miriam Ben-Rafael
Table 2. IL functions of -à H
IL
Fr
be le be le/’etsel le/bišvil ‘al be
à à à à à à à
dans (static) dans (dynamic) à (time) chez pour sur, de, au sujet de de façon, avec + abstract noun
b. il parle à patience (instead of: avec patience) hu medaber besavlanut ‘He speaks with patience.’
We see, in these examples, a lexical Hebrew L1 interference. These occurrences are the translation of Hebrew. However, while the Hebrew preposition be followed by an abstract noun is correct in Hebrew L1, its IL equivalent à is unacceptable in French. In short, à plays a major role in the IL of Hebrew L1 learners of French. It is a pluralistic preposition that has many meanings and functions. Table 2 summarizes these functions.
. Verb valency Another major source of French L2 prepositional errors are prepositional verbs. Here again we find that most errors are explained by the lack of equivalence between French and Hebrew. A short contrastive analysis (Table 3) shows that Hebrew L1 and French L2 verbs often differ in valence. Some French L2 verbs imply direct or indirect objects, while their Hebrew L1 counterparts do not. Two examples are aider (“to help”) and regarder (“to look”) (see Table 3, ex 1, 2). Even when both the Hebrew and the French verbs require prepositions, these rarely overlap (ex 3). Moreover, Hebrew verbs that are followed by a verb in the infinitive do not usually require a preposition, where many French verbs do (ex 4, 5). All these differences again provide occasions for error. . Interferential errors We find interferential errors in a variety of patterns.
English translation
I help my mother I look at the nice girl he speaks about his family he decides to come he begins to speak
Fr
1. j’ aide ma mère 2. je regarde la belle fille 3. il parle de sa famille 4. il décide de venir 5. il commence à parler
Table 3. Contrastive Analysis: verbs and their valency
‘ani ‘ozeret le’im’a ‘ani mistakel ‘al habaxura hayafa hu medaber ‘al mišpaxto hu maxlit () lavo hu matxil () ledaber
H j’ aide à ma mère je regarde sur la belle fille il parle sur sa famille il décide () venir il commence () parler
Fr literal translation of H The French preposition in contact with Hebrew
Miriam Ben-Rafael
.. Omission of de De is omitted and the verb is followed by a direct infinitive object as in Hebrew; the IL sentences follow the Hebrew equivalent pattern (a) (cf. Table 4): Table 4. Pattern a. HL1
IL
FL2
V + inf
v + inf
V + de + inf
(12) a.
elle a essayé () skier hi nista la‘asot ski ‘She tried to ski.’ b. il a décidé () voir hu hexlit lir’ot ‘He decided to see.’ c. je n’ oublierai pas () offrander (= offrir) ani lo ‘eškax latet ‘I won’ t forget to give.’
With pattern (b), de is omitted after a predicative adjective following the copula être, as in H (cf. Table 5): Table 5. Pattern b. HL1
IL
FL2
v être + adj (predic) + v inf v être + adj (predic) + v inf v être + adj (predic) + de + v inf
(13) a.
il est possible () rester en Europe efšar lehiša’er ‘It’s possible to remain in Europe.’ b. c’est interdit () retourner asur laxzor ‘It’s forbidden to go back.’
.. omission of à Similar processes of omission appear with à. In Table 6, à is omitted before a direct infinitive object, as in H (pattern c): Table 6. Pattern c. HL1
IL
FL2
v + inf
V + inf
V + à + inf
The French preposition in contact with Hebrew
(14) a.
ils ont commencer () balayer hem hitxilu let’at’e ‘They began to sweep.’ b. nous allons aider () cueillir les pommes anaxnu na ‘azor liqtof ‘et hatapuxim ‘We will help to pick the apples.’
In Table 7, à is omitted in structures of the type /avoir + infinitive/, as in Hebrew (pattern d): Table 7. Pattern d. HL1
IL
FL2
yeš (avoir) + inf
v avoir + inf
v avoir + à + inf
(15) a.
j’ai beaucoup () dire yeš li harbe lehagid ‘I have a lot to say.’ b. j’ai très dur () vivre ici qaše li lixyot po ‘It’s very difficult for me to live here.’
. Non-interferential errors We also find non-interferential errors, caused by inappropriate uses, superfluous elements, or overgeneralizations. In the following examples, de and à have been added where they are not needed: (16) a.
j’aime *de voir beaucoup de TV (instead of: j’aime beaucoup voir la télévision) ‘I like very much to watch TV.’ b. je préfère *d’aller en étranger (instead of: je préfère aller à l’étranger ‘I prefer to go abroad.’ c. les grandes parties ne laissent pas les autobus *à rouler le samedi (instead of: les grands partis ne laissent pas rouler le samedi) ‘The big parties don’t let buses drive on Saturdays.’ d. nous sommes dans un pays qui se déclare *à être démocratique (instead of: nous sommes dans un pays qui se déclare être démocratique) ‘We are in a country which declares itself to be democratic.’
Miriam Ben-Rafael
The next sentences illustrate the incorrect use of pour, in place of de or aux (contraction of à + les): (17) a.
il m’a demandé *pour sortir (for: il m’a demandé de sortir) ‘He asked me to go out.’ b. le professeur pose des questions *pour les élèves (for: le professeur pose des questions aux élèves) ‘The teacher questions the students.’
. The French preposition in Franbreu In spite of the lack of equivalence between the French and Hebrew systems, prepositional changes in Franbreu are not too frequent – and much less so than in IL. Our data, for instance, indicate that à is used in Franbreu much as it is in French. Certain errors occur only among individuals who settled in Israel a long time ago (see examples 18a, b) or who arrived in early childhood (19a, b); they are by no means characteristic of the general Franbreu trend. . à (Franbreu) = à (IL) vs en, à, dans (French) (18) a.
j’ ai immigré *à Israël en 1951 (for: en Israël) ‘I immigrated to Israel in 1951.’ b. je suis venu en Israël *à 56 (for: je suis venu en Israël en 56) ‘I came to Israel in 56.’
(19) a.
je suis né *à la France *à la ville de Paris *à 1974 (for: en, dans, en) ‘I was born in France in the city of Paris in 1974.’ b. l’histoire se passe *en Japon (for: l’histoire se passe au Japon) ‘The story takes place in Japan.’
. Franbreu preposition = Hebrew preposition One also finds interferences which are calques from Hebrew: (20) a.
y a moyen de demander chez qui lifnot (instead of: à qui s’adresser) ‘etsel mi lifnot ‘It’s possible to ask whom to address.’ b. on voit bien là dans cette photo (for: on voit bien là sur cette photo) batmuna haz’ot ‘It is quite clear here in this picture’.
The French preposition in contact with Hebrew
c.
j’ai été sur la rue Sokolov (for: j’ai été rue Sokolov) halaxti ‘al Sokolov ‘I went to Sokolov street’.
. Quantitative adverb + de + French noun vs quantitative adverb + () Franbreu noun In some cases, French constraints are not respected and de is omitted after French quantitative adverbs, as it is the rule in Hebrew. Thus assez, beaucoup and peu behave like their Hebrew equivalents maspiq, harbe and me`at. This is shown in the following examples where des is not a preposition but marks incorrectly the regular French indefinite determinant. (21) a.
y a pas assez *des limites (for: assez de limites) eyn maspiq gvulot ‘There are not enough limits.’ b. j’ai beaucoup *des amis (for: beaucoup d’amis) yeš li harbe xaverim ‘I have a lot of friends.’
Nonetheless, de is still omitted much more often in IL than in Franbreu. . Confusion and attrition In our Franbreu data, we also find occurrences that indicate confusion or attrition: (22) a.
sur la rue il y a un magasin (for: dans la rue) ‘In the street there is a store.’ b. à la fin de compte il fait tomber le nid (for: en fin de compte) ‘At the end he drops down the nest.’
. Verb valency in Franbreu Most of the prepositional variations in Franbreu, which, it must be emphasized, are less numerous than in IL, involve verbal valencies. We find interferential prepositions that are Hebrew calques. In example (23), the second speaker relies on the Hebrew al “on”, instead of using pour “for”:
Miriam Ben-Rafael
(23) Locutor 1. j’ai pas envie de me battre avec elle (I don’ t want to fight with her) Locutor 2. te battre *sur quoi (to fight on what).
A similar interference occurs in (24), where the Hebrew verb is followed by al, whereas the French one is directly transitive: (24) regarde *sur ton matériel (instead of: regarde ton matériel) ‘Look at your material.’
In (25), the error source is the Hebrew me “from”, instead of using à son père: (25) ce roman elle l’a caché *de son père (this novel she has hidden it from her father).
Errors are numerous when a verb is followed by an infinitive. In (26), for instance, the preposition is omitted, as in Hebrew. In (27), de is incorrect and replaces erroneously à: (26) si je peux pas trouver () lesader ‘et ze (instead of: si je peux pas trouver à arranger ça) limtso lesader ‘et ze (= trouver arranger ça) ‘If I can’t figure out how to fix it’. (27) on m’a aidé *d’intégrer (instead of: à m’intégrer) ‘They helped me to integrate.’
. The resistance of the French preposition Yet, despite all these variations in Franbreu, French prepositions often appear to resist change, as in chérie sois pas en laxats “my dear don’t be pressured”, where en corresponds to the Hebrew be “in”, but also to the French en, such as en colère “angry” or en sueur “sweating”. The same codeswitching with the Hebrew be occurs in j’ ai été en mil’uim “I was on reserve duty”, which again is similar to expressions like en classe “in class”, en vadrouille “roving about” or en vacances “on holiday”. Despite the changes influenced by Hebrew, the French prepositional model remains relatively stable. Even in codeswitching, the prepositions that are used before Hebrew borrowings in Franbreu are the same ones that would be used in French.
The French preposition in contact with Hebrew
(28) a.
ils sont prêts à lehanhig des šinuyim de misgeret ‘They are ready to introduce formal structural changes.’ b. si on me demande de letaqen . . . ‘If they ask me to repair.’ c. haxug a décidé de litmokh ba ‘The department has decided to support her.’
In (28a), the assigned preposition is à, in (b) and (c), it is de. French constraints are respected even with regard to the gender of borrowed nouns that follow à. When the borrowed noun is masculine in Hebrew, the contractive au (à + le) is used, even when the equivalent French noun is feminine: (29) a.
c’est un martse au tekhnyon ‘He is lecturer at the Technion [university]’ b. il travaille au beyt sefer ‘He works in the school.’
Similarly, if a Hebrew borrowing does not keep its Hebrew gender but rather adopts the gender of its French equivalent, French prepositional constraints are still respected. In the following occurrence, du, the contraction of de + le, is used because the French determinant accords with the French masculine gender of the French word restaurant rather than with the feminine gender of the Hebrew equivalent, mis‘ada, which is used: (30) de serveuse elle est devenue la patronne du mis‘ada (restaurant) ‘From a waitress she has become the boss of the restaurant.’
Hence, even in the case of Hebrew borrowings, one observes the strong resistance (31) of French prepositional forms: (31) a.
ça veut dire qu’ il y a quelque chose de mamaš qui se passe (cf: quelque chose de réel) ‘It means there is something real that happens.’ b. je vais faire un peu d’hašqaya (cf: un peu d’arrosage) ‘I’m going to do some watering.’
For the most part, the insertion of Hebrew borrowings does not lead to incorrect usage of French prepositions. However, concomitantly with the correct use of French forms, one also finds some digressions. In examples (32a, b), de and à, which bring in, respectively, the verbal borrowings faire mimuner and lehit’apeq, are correct by French criteria; but in its second occurrence, in the phrase lehit’apeq à manger (32b), à incorrectly replaces the French preposition de:
Miriam Ben-Rafael
(32) a.
je vais essayer de faire mimuner (cf: je vais essayer de faire financer) ‘I will try to get some financial help.’ b. j’arrive pas à lehit’apeq *à manger beaucoup de cornflakes (as: j’arrive pas à me retenir/but unlike: de manger beaucoup de cornflakes) ‘I can’t manage to keep myself from eating a lot of cornflakes.’
. Conclusion This chapter has examined the French prepositional system in two different situations of French/Hebrew contact. More specifically, these pages have dealt with: (a) French L2 acquisition by Hebrew L1 learners and (b) Franbreu production by bilingual French and Hebrew locutors, in order to learn and compare how the French prepositional system operates in both IL and Franbreu. We may summarize our conclusions as follows: 1. our findings have revealed prepositional errors and variants in IL as well as in Franbreu; 2. these errors are much more numerous in IL than in Franbreu; 3. in IL, errors also stem from overgeneralization and the formation of specific rules; 4. in Franbreu, errors and variants may stem from confusion and some attrition of French L1. In a general manner, the findings indicate that if, on the one hand, IL is characterized by the development of a new prepositional system, on the other hand, Franbreu illustrates the relative resistance of French to change. While a contrastive analysis has shown significant differences between the French and Hebrew prepositional systems, it appears that the two forms of French/Hebrew contact do not converge. Why this difference? How can we explain it? As we noted in the introduction, the literature of language contact emphasizes that change is principally lexical and concerns codeswitching or borrowings. Syntactic change remains minimal, and when it occurs, it mainly relates to the spoken language, and reflects tendencies inherent to L1 itself or a propensity to simplification (Blanche-Benveniste 1990; Gadet 1989; Walter 1988; Ager 1990; Mougeon & Beniak 1991). Moreover, even when researchers speak of syntactic changes, either in general or in language contact situations, they underline the fact that such changes are slow, and that elements which seem to disappear from the language tend rather to confine themselves to specific contexts of the spoken language (Blanche-Benveniste 1990: 213). The syntax and
The French preposition in contact with Hebrew
the grammar are widely considered to be the “hard core” of the language and the reflection of its specificity. This core, it is contended, remains, in contrast to the lexicon, relatively resistant to external influences (Hagège 1987: 26–27). The general process of language erosion is variable and broadly depends on the specific circumstances of every particular community. However: When an attempt is made to outline, beyond the wide variety of particular cases, a general profile of the process of erosion and of the manner that the different components of the language are affected, it will be recalled that in most cases, the core elements resist change longer than the lexicon. (Hagège 2000: 110)
It is when they, too, are affected, that one may speak of the beginning of the death of a language. In the case of Franbreu, the French/Hebrew contact seems to sometimes precipitate the emergence of certain tendencies of variation that are natural to French, while, on the other hand, the influence of Hebrew syntactic structures is also noticeable. However, when substantial differences exist between the two linguistic systems, it is French that prevails. This is the case of the French negative system, for instance, which withstands the influence of the Hebrew negative system (Ben-Rafael 1993b). This is also the case of the French prepositional system in spite of the fact that Hebrew prepositional features may have some influence, and that some typical French forms do not appear in Franbreu (such as je vais sur Paris). The French prepositional system has remained, in fact, broadly unchanged. This resistance to change of the French prepositional system in French/Hebrew contact confirms that the grammar, as indicated in the literature, constitutes the hard core of the language. Regarding IL, we know that it varies according to the level of learners and the conditions of learning but that, at the same time, it consists of an organized structure (Corder 1981: 71). IL is a simplified language marked by formal features of both, the target language and L1. Researchers have emphasized the importance of L1 transfers and the re-emergence of L1 at all levels – lexical, structural and syntactic (Selinker 1992). This seems to relate, as far as syntax and grammar in general are concerned, to what has been said above. That they both are viewed as parts of the hard core of the language may explain, among other things, why a large number of errors appear in these domains when L2, the target-language, confronts L1, the source-language. As for the Hebrew-speaking learners of French, errors regarding the different fields of syntax – gender, determinants, possessive forms, word order, the verbal system and, at last, the prepositional system – are numerous in-
Miriam Ben-Rafael
deed (Ben-Rafael 1989, 1994). The factors that explain these errors are varied, although among all factors considered, it is mainly the influence of Hebrew L1 structures that stands behind the majority of syntactic Hebrew/French IL errors. In conclusion, in order to answer the question concerning the differences found in the prepositional system of Franbreu and Hebrew/French IL, one may suggest that it is the link of the prepositional system to the hard core of both French and Hebrew, which brings about that in French/Hebrew contact: 1. Franbreu witnesses the resistance of the French prepositional system to Hebrew influence; 2. French prepositions meet difficulties to be adopted in IL. Hence, the same theoretical principle actually explains these two contrasting phenomena in different contact situations involving the same languages.
Notes . Method of French second language acquisition for beginners. . In all IL examples, we consider and discuss only the errors which involve prepositions.
References Ager, D. E. (1990). Sociolinguistics and Contemporary French. Cambridge: Cambridge University Press. Aub-Busher, G. (1993). French and French-based Creoles: the Case of the French Carribean. In C. Sanders (Ed.), French Today (pp. 129–214). Cambridge: Cambridge University Press. Ben-Rafael, M. (1989). Interférences syntaxiques et stratégies d’apprentissage: le français, langue étrangère en Israël. Mémoire de maîtrise. Tel-Aviv University, French Department. Ben-Rafael, M. (1993a). Contact de langues, acquisition et attrition: Profils franbréophones. In J.C. Pochard (Ed.), Profils d’apprenants – Actes du 9e Colloque international: Acquisition d’une langue étrangère, perspectives et recherche (pp. 91–101). St. Etienne: Publ. Université St. Etienne. Ben-Rafael, M. (1993b). Attrition and Acquisition Compared: The French Negative in Language Contact. Communication presented at the 10th World Congress of Applied Linguistics AILA. Amsterdam: University of Amsterdam.
The French preposition in contact with Hebrew
Ben-Rafael, M. (1994). Interférences syntaxiques et stratégies d’apprentissage: le français langue étrangère en Israël. In Firmin Sierra Martinez et al. (Eds.), Las lenguas en la Europa Comunitaria (pp. 75–89). Amsterdam: Editions Rodopi. Blanche-Benveniste, C. (1990). Le français parlé, études grammaticales. Paris: CNRS. Borillo, A. (1998). L’espace et son expression en français. Paris: Armand Collin. Bouton, C. P. (1969). Les mécanismes de l’acquisition du français langue étrangère chez l’adulte. Paris: Klincksieck. Cadiot, P. (1997). Les prépositions abstraites. Paris: Editions Ophrys. Corder, J. P. (1981). Error Analysis and Interlanguage. Oxford: Oxford University Press. Csecsy, M. (1971). Les prépositions: Interférences franco-hongroises. Le Français dans le Monde, 81, 40–43. Durand, J. (1993). Sociolinguistic Variation and the Linguist. In C. Sanders (Ed.), op.cit (pp. 257–286). Duškova, L. (1969). On Sources of Errors in Foreign Language Learning. IRAL, 7(1), 11–36. Feigenbaum, S. (2000a). A contrastive analysis of French and Hebrew prepositions: The case of sans, bli, and lelo. Paper presented at the Conference Prepositions, grammatical, semantic and pragmatic aspects. University of Haifa, June 2000. Feigenbaum, S. (2000b). Selon et ses synonymes: une analyse contrastive. Paper presented at the Conference Prep an 2000, la préposition dans tous ses états. Tel-Aviv: TAU, 3–9/9. Frei, H. (1971). La grammaire des fautes. Genève: Slatkine. Gaatone, D. (1998). Le passif en français. Bruxelles: Duculot. Gaatone, D. (2000). Les prépositions: une classe aux contours flous. Paper presented at the Conference Prep an 2000. Tel-Aviv: TAU, 3–9/9. Gadet, F. (1989). Le français ordinaire. Paris: Armand Colin. Giacobe, J. (1992). Acquisition d’une langue étrangère. Paris: CNRS éditions. Grevisse, M. (1986). Le bon usage. Edited by A. Goosse. Paris/Louvain-la-Neuve: Editions Duculot. Hagège, C. (1987). Le français et les siècles. Paris: Odile Jacob. Hagège, C. (1996). Les relateurs comme catégorie accessoire et la grammaire comme composante nécessaire. Faits de Langue, 9, 6–19. Hagège, C. (2000). Le souffle de la langue. Paris: Odile Jacob. Jacobson, R. (1998). Codeswitching Worldwide. Berlin: Mouton de Gruyter. Kihm, A. (2000). Quand ‘de’ disparaît: les constructions génitives en créole haïtien. Paper presented at the Conference Prep an 2000. Tel-Aviv: TAU, 3–9/9. Meyers-Scotton, C. (1992). Social Motivations for Codeswitching. Oxford: Clarendon. Mougeon, R. & Beniak, E. (1991). Linguistic Consequences of Language Contact and Restriction. Oxford: Oxford University Press. Picabia, L. (2000). ‘Il y a’ est-il une préposition? Paper presented at the Conference Prep an 2000. Tel-Aviv: TAU, 3–9/9. Roberge, Y. & Vinet, M. T. (1989). La variation dialectale en grammaire universelle. Montréal: Presses universitaires de Montréal. Romaine, S. (1989). Bilingualism. Oxford: Blackwell. Sans Frontières 1, Méthode de français. (1989). Paris: Clé International. Selinker, L. (1992). Rediscovering Interlanguage. London: Longman. Walter, H. (1988). Le français dans tous les sens. Paris: Robert Laffont.
‘Preposition’ as functor The case of long in Bislama Dennis Kurzon Department of English, University of Haifa
.
Prepositions as word class
Prepositions seem to be taken for granted. Let us take two features of this word class which show what I mean, and which therefore need some clarification at the outset. Such a clarification would be essential before one sets out on a detailed analysis of any set of such words in one or more languages. Firstly, the term ‘preposition’ (and its equivalent in SOV languages – ‘postposition’1 ) refers only to the position of the word class in a structure, and not to its function or to some general semantic category. Terms for some of the other word classes tend to be more helpful. For example, ‘noun’ is the English equivalent of Latin nomen, meaning “name”; hence, nouns have traditionally been defined as the names of things. The term ‘adverb’ means “attached to the verb”, again reflecting how the word class has been traditionally defined as a “modifier of the verb”, although this does ignore the fact that in many languages, including English, adverbs modify adjectives and other adverbs, too. Secondly, prepositions constitute a small set of items in any language (words, or morphemes in, say, Semitic languages), sharing this feature with conjunctions and pronouns, in contrast to the large sets of nouns, verbs, adjectives and adverbs. Whereas there is no doubt that these four large groups2 are contentives in that members of these groups can be semantically defined, there may be a question concerning the semantic status of the smaller groups. If antonymy is considered a criterion by which contentives are distinguished from functors (Radford 1997: 45), then many prepositions would be classified as contentives despite their major syntactic function within the phrase, as we
Dennis Kurzon
shall see. Such contentive prepositions include the pairs inside/outside, on/off as in the following two pairs of sentences: (1) a. She put the packet inside the shed. b. She put the packet outside the shed. (2) a. They got on the bus b. They got off the bus.
But prepositions such as for, at, of do not seem to have antonyms, as we may see in the following: (3) I bought it for/? five dollars. (4) They are coming at/? five o’clock.3 (5) a.
It consists of five parts.
For the verb consist in the last example (5a), we also find “it consists in NP”, but there is no strictly semantic difference between of and in; the difference is usually grammatical between a countable complement, as in example (5a) above, and a non-countable complement as in (5) b. The happiness of a country consists in the freedom of its citizens (Oxford Dictionary)
A counterexample to this pair, where semantic meanings of the sentences do differ, may seem to be (6) a.
I am afraid of him
and (6) b. I am afraid for him
but this difference seems to be a function, on the one hand, of the grammatically required of -insertion (see 5. below) in the adjectival phrase afraid of him, since afraid cannot be followed by a NP complement (unlike the equivalent verb to fear), as against a semantically defined PP for him (“on his account”). The difference is, hence, not due to any semantic distinction in se between of and for in this context, but between the zero meaning of the of -insertion and the semantic meaning of for (for which there is no semantically defined antonym). What this seems to indicate is that the separate word class of prepositions does not reflect a homogenous set of grammatical and semantic distinctions. There should be at least two subclasses of prepositions – those that have full
‘Preposition’ as functor
semantic meaning, i.e. those words with antonyms, e.g. after/before, and those that, we may say, are highly polysemic and act syntactically more as functors. The net result seems to be that it serves no purpose to speak of the meaning of any member of this sub-class of prepositions, for their meaning is highly dependent on the context. Several scholars who have studied prepositions do not posit two classes only, as we have done here, but have established three classes or subclasses of prepositions. Emonds (1985: 254), for example, argues for the following three in English: + ___ NP, e.g. at, for, of, with + ___ (NP), e.g. across, down, in over, up + ___ e.g. apart, ahead, away
The first subclass of prepositions requires a complement, with those of the second subclass a complement is optional, while prepositions of the third subclass cannot select a complement. This is supported by Foskett (1991), who includes together under prepositions, but as a member of the third type, i.e. those that occur without a complement. He shows that prepositions of the first subclass may be found with prepositional verbs, those of the second type occur in phrasal verbs where the verbs themselves have idiomatic meaning, while the third type of preposition may be found in phrasal verbs which do not have idiomatic meaning. Compare (examples based on Foskett 1991: 32): (7) The teacher always ran the students’ work down
and (8) She put the puzzle together.
where the phrasal verb run NP down has an idiomatic meaning (in contrast to run down meaning, among other things, “collide with”), hence down is a member of the second subgroup, while the phrasal verb put together is not idiomatic; its meaning is a composite of its constituent parts. Rauh (1993) initially distinguishes between two types of prepositions – lexical and non-lexical prepositions, setting up a set of criteria which characterize lexical prepositions.4 But he makes a further distinction among the non-lexical prepositions: one subgroup – prepositions in fixed phrases (or idiomatic) – fulfil some of the criteria, while prepositions of the second subgroup – case prepositions – fulfil other criteria. Phrases containing a member of the first subgroup have the following characteristics: they act more like adjectives or adverbs than like PPs, the preposition is not head of the phrase, and there is no proform. As for the second subgroup (the case preposition), he argues that the
Dennis Kurzon
structure of the second type is NP, not PP, and coordination with other PPs, furthermore, is ungrammatical. Likewise, König and Kortmann (1991) distinguish three subclasses of prepositions. The first subclass consists of very frequent, monosyllabic prepositions, all of which have “a broad range of meanings that include very abstract, ‘grammatical’ meanings and uses” (p. 112); they also usually allow stranding. The second subclass contains less frequent, typically disyllabic prepositions with a limited range of meanings, while the third subclass consists of infrequent phrasal or complex prepositions, which are composed of relational nouns and prepositions of the first class. They also note that most members of the first subclass have very versatile meanings, their dictionary entries being far longer than other prepositions (due, they claim, to Zipf’s Law of Abbreviation; 1991: 111). This is not only true for English, but for many other languages. If we take a look at frequent prepositions in several languages, we would notice that on the basis of meaning, they may be divided into at least two groups. Let us take three dictionaries, with the caveat, of course, that dictionary entries reflect the opinion of the compiler, and the set of criteria used to distinguish meanings is certainly not uniform among dictionaries, even dictionaries of the same language. The meanings given may often be a matter of opinion, and are not based on some objective test. In the 1969 edition of the American Heritage Dictionary of the English Language, we find highly polysemic prepositions such as for (28 meanings),
with (26), on (26), to (21), of (18)
compared to: at (13), in (13), by (12)
For French, we can cite the number of meanings given in the Lexis dictionnaire de la langue française of 1975: de (15 meanings),
à (13), pour (11),
compared to sur (9), en (8), par (7) and
avec (5).
The same may be seen in Hebrew. According to the Even Shushan Hebrew Dictionary (1991), we can again divide prepositions into two groups: Áal (10), be- (8), le- (8), min (8)
compared to bishvil (6), kmo (5), el (4), ke- (4), Áim (4)
‘Preposition’ as functor
Although there is no non-arbitrary (and motivated) way we can draw a line and say that a preposition that has more than x number of meanings is highly polysemic, with its meaning being relegated to a secondary feature as against its syntactic function,5 there seems to be here a continuum which does make a distinction between highly polysemic prepositions, on the one hand, and prepositions with only a handful of meanings, on the other. It is possible to hypothesize that the more meanings a preposition purportedly has, the less important its semantics becomes, and its syntactic function becomes more central. An illustration of this is the Hebrew particle et. It functions like a preposition, and it takes personal endings like most prepositions (e.g. oti “me”, otah “her”, cf. bi “in me”, bah “in her/it”), but it is meaningless. The sole function of et without personal endings is to mark definite direct objects, and the form with personal endings marks direct object pronouns (e.g. hi ro’ah oti “she sees me”).6 Integrating the two features of prepositions discussed here, we may say that there are languages in which the term for ‘preposition’ much more accurately reflects the syntactic function of this group of words. In German, for example, as well as the term “Präposition” we also find “Verhältniswort”, which may be translated “relation word”. In Hebrew, the word for ‘preposition’ is “relation word” (milat-yahas), too. (See Brøndal 1950 for a brief discussion of the history of the term “preposition”.) The term “relation” in this context has been defined as involving “two or more participant entities” in an asymmetrical relation; in terms of cognitive grammar, “one participant is selected for foregrounding, while the other participant serves as a background” (Taylor 1993: 153). It is my argument that the grammatical function of prepositions, or at least of the subgroup of highly polysemic prepositions, is more central than their semantic meaning. These prepositions are probably undergoing a process of grammaticalization. Diachronically, they related to spatial and temporal dimensions of the message, but have gradually been taking on grammatical functions, losing much of their semantic meaning.7
. Bislama prepositions Evidence of this internal classification of prepositions does not come only from English and other well-known languages. In this paper, the function and meaning of one preposition (long) in the English-based creole, Bislama, will be analyzed, with the purpose of showing that it is in effect meaningless, or is so polysemic that we cannot give it a basic meaning in a synchronic study.8 Bis-
Dennis Kurzon
lama is one of three Melanesian creoles; this one is spoken, and written, in the Republic of Vanuatu (formerly the New Hebrides) in the South Pacific Ocean. The best known of these creoles is Tok Pisin, spoken in Papua New Guinea, while the third creole is Pijin, spoken in the Solomon Islands. At one stage, these were not separate languages due to the (forced) mobility of Melanesian labourers, and the intermingling of the labourers from different parts of the region, as a constituent of a ‘divide and rule’ policy of the colonialists. However, already by 1885, Tok Pisin had become separated “when labourers from the New Hebrides and Solomons stopped going to Samoa and labourers from New Guinea stopped going to Queensland” (Siegel 1998: 354). Although Tok Pisin has a far longer history than Bislama, today it is not held with the same regard in Papua as Bislama is held in Vanuatu (Crowley 1990a: 3). There, Bislama shares with English and French the status of official languages, but Bislama is creeping ahead in strength. Because of the competition between the two dominant colonial powers, Britain and France, the respective languages have become “tools of division at the national level”, so Bislama functions as “the effective unifying language” (ibid. p. 2). The examples of Bislama in this paper are from Vanuatu publications in various fields, and from personal communication with more locally based scholars of the language (Terry Crowley, Robert Early, John Lynch, Miriam Meyerhoff). The basic assumption concerning the apparently meaningless preposition long is as follows: This preposition is one of a set of five prepositions (the others being blong, from, wetem [=with], olsem [=as]) which differ from a second set of prepositions which derive from verbs (Crowley 1990b: 76), e.g. kasem as a verb means “to catch”, while as a preposition it may be glossed “till”, “as far as”, as in (9) a.
Mi no kasem mane long bang yestede “I didn’t get any money from the bank yesterday” b. Bae mi slip long 6.00 kasem 8.00. “I’ll sleep from 6.00 to 8.00.”
and bitim as a verb means “to beat”, and as a preposition “more than”, e.g. (10) a.
Ajentina i bin bitim Jemani long Wol Kap long 1986 “Argentine defeated Germany in the World Cup in 1986” b. Mi save dring ti bitim yu. “I can drink tea more than you” (examples from Crowley 1990b: 76, 79)
‘Preposition’ as functor
(but see example (17) below on the use of long meaning “than” in the correlative moa . . . long). Furthermore, in the first group we may distinguish long and blong from the other three, for these two, which Crowley calls full prepositions, cannot be stranded, while from, wetem and olsem may be (ibid. p. 82). Compare: (11) ples ia nao, mi stap long hem this is the place I live at ()
with the obligatory pronoun complement following the preposition, and (12) Mi stap tekem wan kos we mi kam long Vila from I’m taking a course that I came to Vila for.
in which from is stranded. Therefore, according to at least one syntactic criterion, the subclassification set up has been justified.
. Semantics of long In this paper I will focus both on the syntactic features and on the semantics of long. In the two sentences given above (11) and (12), it may be seen that long has a locational/directional meaning; in the first long means “at” , while in the second “to”. This may be its basic meaning, historically speaking, but as I have already said (see note 8), a diachronic approach is not relevant to this study. What is of interest is the use and meanings of the preposition in present-day Bislama, as reflected in the written texts under investigation. Further examples of long with the same locational meaning are: (13) sidaon long hem Sit down on it. (14) Man i bon long wol ia blong kasem gutfala laef. Man is born in this world to achieve a good life.
and with directional meaning, often in collocation with go/ko,9 as in (15) Minista blong Edukesen, Yut mo Spot, Honorabol The education, youth and sport minister, the honourable Joe Natuman i lidim wan delikesen long wan ofisol visit i ko long J.N. leads a delegation on an official visit to Morises Mauritius
Dennis Kurzon
Of course, with the gamut of locational/directional meanings we also find the related temporal meaning, as in (16) bae man ia i glad blong kambak long wan narafala dei This man will be glad to return on another day. [lit. FUTURE man this is glad . . . ] (17) tumora long hem on the next day [lit. “tomorrow on it”]
If we move away from locational and temporal meanings, we find other semantic meanings of long such as “from”, as in (18) Credit union hemi difren long ol narafala institution or The Credit Union is different from/to other institutions or organesen. organizations
“than” (but see the use of bitim in (10b) above), as in (19) Pikinini hemi gat moa janis long yumi . . . blong lanem . . . Children have more chance than us to learn
and “about” as in (20) yu save contactem mifala long hem you can contact us about it
Example (18) shows an interesting phenomenon, which reflects standard as against non-standard English. The adjective different has as complement a PP with head to, but in more colloquial English, the preposition from is head. I offer both translations. In such cases, Bislama, like many other creoles and pidgins, chooses one of its zero-content prepositions. As for long with the meaning of “about”, a more frequent way of expressing this may be by the compound preposition long saed blong (see ex. 23 below), but even the title of a tourist guide to the Bislama language (Tryon n.d.) uses long on its own with this meaning: (21) Evri samting yu wantem save long Bislama be yu everything you want to know about Bislama but you (were) fraet tumas blong askem too afraid to ask. [lit. afraid too/very]
‘Preposition’ as functor
. Syntactic functions of long . We find, however, that the preposition long has two further uses, but this time we are not in the field of semantics, but in those of lexis and grammar. That is to say, we are not concerned with an extension of the meaning of the preposition, but with an extension of its function. The first non-semantic function is that of the formation of compound prepositions. We find long in combination with adverbs or other particles to form compound prepositions, a function similar to that of of in English, as in on account of, by means of, in front of,10 e.g. (22) ol pikinini oli mas sidaon kwaet raon long hem The children must sit down quietly around ( ) him. (23) Hemi kontinu long Yunivesiti blong Papua Niu Gini blong stadi She goes on to the Papua New Guinea University to study long saed blong lao. (about) law.
In the first example, raon is primarily the adverb “round”, as in (24) bae yumi tanen raon FUTURE we turn round
but in combination with long it functions as a preposition. In the second example, long, on its own, as we have seen (ex. 21), may carry the meaning of “about”, but the compound preposition long saed blong, as in our example, is used more frequently. . The second non-semantic function is syntactic in essence. What is to be discussed here is, as far as it can be, theory-neutral, although the general approach as far as the structure is concerned is the X-bar model. For example, the prepositional phrase may be said to have the structure: PP → P P → P NP, while a verb phrase with two complements may have the following structure: V → V NP PP.
Dennis Kurzon
Readers would find no difficulty, however, in translating the structures discussed here into forms they are familiar with. The preposition long has a function within the sentence (or verb phrase), indicating – for want of a better term – case. Apart from the nominative case (or subject), we can identify two other cases, given in their traditional terminology – accusative and dative. In the context of Bislama, we are talking about its underlying structure; these cases are not marked morphologically by a bound morpheme in the way case is marked in, say, Russian or Latin. Firstly, long may be a marker of the dative. According to Robert Early (p.c.), this is not necessarily equivalent to the indirect object; the distinction found in English, for example, between direct and indirect object does not seem appropriate to the situation in Bislama. The data based on written texts, however, suggest otherwise. While in (25) oli stap lisen gud long olgeta storian ia They were listening well to these stories.
lisen long seems to be a direct gloss of “listen to”, we find similar structures with long in collocation with verbs that in English are ditransitive, for example ask, tell. Let us look at the following examples, firstly, verb phrases – with potential ditransitive heads – with one complement: (26) yu save askem long olfala You can ask () the old guy. (27) yu mas talem long olgeta You must tell () them. (28) oli talem long Meri They told () Mary
and secondly, ditransitive verbs with two complements: (29) oli talem toktok ya long Meri They told this message to Mary (30) Yufala i mas givim plante wok long olgeta You must give a lot of work to them
Our first reaction is that the preposition seems to function as a marker of the dative case. However, we may be looking at long as a marker not of the dative case only but as a marker of any verb complement. We can illustrate long with this general function in the following sentence:
‘Preposition’ as functor
(31) Yu save singsing long wanwan laen fastaem You can sing ( ) one line first time.
However, in this example we are not dealing with the Bislama equivalent of English ditransitive verbs, but with a verb that could be either intransitive or transitive. It is, therefore, important to distinguish between two classes of verbs in Bislama: transitive (including what would be ditransitive in English) verbs and intransitive verbs. The distinction between these two classes, moreover, is morphologically marked. Transitive verbs end in a suffix, which has at least three allomorphs, depending on vowel harmony: -im, -em, and -um, as in bitim (‘to hit’), askem (‘to ask’) and yusum (‘to use’).11 We have seen in the above examples of ditransitive verbs that not only do they have this suffix, but one of the complements is a prepositional phrase with long as its head. However, there are intransitive verbs – verbs which are of course not morphologically marked by -em, etc. – that may nevertheless be followed by an object; Terry Crowley has called them ‘cognate object verbs’. In such cases, the object is preceded by long. In the above example (31), we have the verb singsing, which means “sing”, functioning as a transitive verb with the object in the form of a prepositional phrase headed by long, while in (32) the same verb functions intransitively: (32) hemi stap singsing she is singing
In the following, we meet another cognate object verb givhan, “to help” (lit. “give a hand”): (33) bae oli givhan long ol Provisal Gavman blong faenem ol betta they will help () provincial governments to find better way blong developem ways to develop
with the object “ol Provisal Gavman” marked by a preceding long. If the situation were as described above, there would be no problem, for we could then say that complements of cognate object verbs and one of the two complements of ditransitive verbs are marked by long. However, let us look again at ditransitive verbs when followed by two complements, as in example (29), repeated here: (29) oli talem toktok ya long Meri They told this message to Mary
Dennis Kurzon
One of the complements is preceded by long. It may be concluded from this example that the PP is the equivalent of the indirect object, but the following is an obvious counterexample: (34) yumi mas lanem hem long olsem wanem blong sevem mane we must teach him/her how to save money
The pronoun object of the ditransitive verb lanem is not preceded by long, as would be expected on the basis of example (29) above, but the complement clause, “olsem wanem blong sevem mane”, is marked as complement by the preceding long. In a grammaticality test among experts (but not native speakers) in the Bislama language, there was general agreement that the following holds with regard to what have been called ditransitive verbs. Generally speaking, if a verb allows two complements (objects), then the second one in sequence is a PP, with long as the head, while the first complement is a NP. It is the second complement that is closely related to the indirect object. We therefore find patterns such as (35) a.
V + NP + PP [long + NP]
but not (35) b. *V + PP [long + NP] + NP
nor (35) c. *V + NP + NP
Hence, we find a sentence such as (36) a.
Central Bank blong Russia i bin kivim woning long The Russian Central Bank gave a warning to kafman ia. the government
in contrast to the ungrammatical: (36) b. *Central Bank blong Russia i bin kivim long kafman ia The Russian Central Bank gave to the government woning. a warning
and
‘Preposition’ as functor
(36) c. *Central Bank blong Russia i bin kivim kafman ia woning. The Russian Central Bank gave the government a warning
That is to say, one cannot move the PP to first position after the verb, and one cannot have two NPs which are complements of the verb, with the first complement as indirect object as in English. So, Early’s comment above (p.c.) that a distinction in Bislama between the direct and the indirect object does not seem to hold is not supported by the evidence. On the contrary, the indirect object does occur second in sequence after the verb and is in the form of a PP with long as head. To make things more complicated, example (34) above, with the complement clause marked by long, occurs in a pamphlet issued by a Vanuatu credit union, but was judged ungrammatical by all the scholars, while the occurrence of long before the first NP (still as ‘indirect’ object, by the way) with a following clause complement was judged grammatical, e.g. (37) Yu save soem long ol pikinini olsem wanem nao oli save You must show () the children how now they can sevem mane. save money
It is obvious that more work is required in this area, including eliciting responses from native speakers. One may rely to a certain extent on published material, since Bislama does have the status of an official language in Vanuatu. There are discussions going on concerning the standardization of the language, initially concerning its orthography, but in the future other features of the language will be discussed. The amount of published material, however, seems fairly limited. It includes two weekly newspapers, the Bible, various commercial and financial publications, language and other textbooks; but there is very little literature in Bislama. Native speaker responses, therefore, will help to solve the problem concerning ditransitive verbs and the role of long in this general context. It has also been pointed out that there may be dialectal differences between a northern variety and a central and southern variety, which does not allow a definite solution to the problem (Miriam Meyerhoff, p.c.). This seems to indicate, however, that in the case of a language such as Bislama where the amount of written data is limited, an analysis of a corpus has fairly obvious limitations. The elicitation of responses from one or more native speaker would provide data which can solve the problem, either as an issue of grammaticality or as a matter of dialect.
Dennis Kurzon
Long as a complementizer may also occur alongside blong, which in its function as a complementizer12 is equivalent to English to, examples of which occur in (33) and (34). The only example I have found in the corpus of long with this function is: (38) ol bigman/woman oli interest long joinem credit union ia. men and women are interested in joining this credit union
But this may be an example of a different aspect of the phenomenon that has just been discussed. This sentence may be either of non-native usage, or of native usage which is influenced by the superstrate language (English), in that long is normally the equivalent of “in”. In fact, some of the inconsistencies noted in the data may be due to the same sociolinguistic situation. Authors of some of the publications may not be native Bislama speakers, i.e. native ni-Vanuatu, as they are called, but outsiders, especially Australians, who are competent speakers of the language, but cannot be regarded as native speakers. A much larger group of authors may be native speakers who are nevertheless influenced by English, since it has the role of the superstrate language. Clark’s comment concerning Bislama vocabulary newspapers, radio broadcasts, and political speeches are typically larded with lexical and syntactic Anglicisms (1983, cited by Thomas 1990: 250)
(mentioned in the context of urban and rural differences) may still be valid today.
. Conclusion Creoles and pidgins typically have one or two prepositions which are polysemic to such an extent that, apart from clear locational/directional/temporal meanings, they may be said to acquire their meaning through context. In AngloNigerian Pidgin, for example, we find the preposition fò. In answer to the ques˙ tion whether fò is polysemic, Mann (1993: 64) says that it may encompass “a ˙ set of possible meanings – the appropriate one being evoked in the right context”, or it may have zero sense, a ‘grammatical dummy’, evoking one meaning: ‘preposition’. Its meaning is based “on the speakers’ deictic interpretation of the utterance”. This is supported by Jones’ study of Sierra Leone Krio (1990) in which the equivalent preposition, ff, “functions as a multipurpose preposition with a core directional or locational meaning” (p. 852), similar to na in Krio, and long in Tok Pisin (and in Bislama, too). He then lists some of the meanings
‘Preposition’ as functor
that ff can carry: ‘in’, ‘to’, ‘for’, ‘against’, ‘on’, ‘of ’, ‘at’, ‘into’, ‘by’, ‘from’, ‘with’ and ‘during’. From this list, it may be seen that any statement concerning the core meaning of the preposition is diachronic in nature, for prepositions in English such as against, of, with are neither temporal nor locative, while by is agentive, e.g. (39) This damage was done by alley cats
as often as it is temporal, as in (40) By the time he comes . . . .
or locative, as in (41) She walked by his house.
The relation set up by long is, according to the data, semantically indeterminate in that the meaning it has depends entirely on the context. But we can go even further, for at times the meaning of the preposition long is quite irrelevant; what is of importance is its syntactic role as a complement marker. Its occurrence between two phrases to form one phrase may seem similar to of -insertion in English, where nouns or adjectives are followed by a complement, e.g. Both (42) a. *Poirot is envious Miss Marple.
and (43) a. *Poirot’s envy Miss Marple
need of -insertion to make them grammatical: (42) b. Poirot is envious of Miss Marple
and (43) b. Poirot’s envy of Miss Marple.
respectively (examples from Haugeman 1994: 173–176). The compounding of nouns, for example, without a relational word, viz. preposition, does exist in Bislama but it is far less extensive than the case in English, e.g. (44) tumora bae i gat wan turis-bot i go long Anejum tomorrow there will be a tourist boat to Aneityum Island
Compounding, moreover, occurs more at the phrase level than at the word level (Arends et al. 1995: 326–327). As an equivalent of of -insertion, the Bislama
Dennis Kurzon
preposition long seems to share this function with blong (see note 12). This may indicate two default possibilities, not a happy situation to have, but if we consider the frequency of occurrence with regard to this function, blong seems to occur far more in such circumstances than long. However, this is a topic for another paper. It has been shown that Bislama provides at least one example of a preposition which supports the stand taken at the opening of this paper. The principal function of the preposition long is to indicate a syntactic relationship between the verb and its complement, and sometimes between the verb and one of its modifiers. Although more research has to be carried out in this area, especially in the elicitation of spoken language from native speakers, it is clear from the examples presented here that long tends to be so polysemic that it eventually becomes semantically meaningless. Its meaning in any particular context depends almost entirely on the context and not on a previous semantic definition found in the dictionary. Moreover, this preposition has provided additional evidence that prepositions cannot be regarded as one, almost homogeneous class of functors, but must be further subdivided into subclasses based on syntactic criteria.
Notes . There is the term ‘adposition’, which would adequately cover both preposition and postposition, but apart from individual researchers (e.g. Hawkins 1993), the term has not been universally accepted. . Strictly speaking, one speaks of three major classes, since adverbs are in most cases derived from adjectives. . “They are coming from five o’clock”, for example, is not antonymous, for this sentence does not rule out that they will come at five o’clock; “from five o’clock” means at any time from five onwards. Neither is “before 5 o’clock” antonymous. In fact, the only phrase opposite in meaning would be “not at 5 o’clock” . This could mean “before”, “after” 5 o’clock, or even not at all. . Briefly, these criteria relate to syntactic and semantic projection properties, proforms, category-specific movement and coordination, case properties, argument structure and theta grid and the definition of their properties (Rauh 1993:121). . Hence, the division I make here is into two groups and not three, which would seem to be the case if a more extensive set of criteria, including syntactic criteria, were used. The figures are rough and ready ones, and should not be taken as fundamental to this study. Perhaps, in any case, a continuum would represent the situation more realistically if all possible criteria are taken into account.
‘Preposition’ as functor . Not to be confused with et, meaning “with”. Its use as a semantically defined preposition is limited, in modern Hebrew, to names of partners of a firm, e.g. Cohen et Levi “Cohen and Levi”, and to the inflected form, e.g. iti “with me”, as in hi ratsa iti “she ran with me”. . Of course, from a diachronic point of view, prepositions probably started their existence as grammatical units, derived from case endings, e.g. in Sanskrit there is a very small number of adpositions (see note 1) which were a subclass of indeclinables in Panini’s grammar. . Obviously, diachronically, if there is enough evidence, we may discover a basic historical meaning. I have used the term ‘creole’ to describe Bislama. The three Melanesian dialects are often referred to as pidgins. The distinction between these two terms is that while a pidgin has no native speakers, though its speakers are often proficient in it as a second language, creoles do have native speakers. It seems that children in certain areas are acquiring these languages as their first language (e.g. Bislama; Thomas 1990: 237). . There is usually no distinction between voiced and voiceless consonants. Hence, both spelling variants may be found. This collocation is an example of serial construction – the occurrence of another verb with the predicate marker i, following the main verb of the sentence, and usually indicating “some kind of inherent directionality in the action” (Crowley 1990b: 68). Crowley claims that only five Bislama verbs are involved in serial constructions: go, gobak, kam, kambak and stap; e.g. hem i lukluk i kam long me he looked towards me . Other prepositions are used in compounding, such as to as in due to, according to, from as in away from. But of seems to be the productive means of forming new compound prepositions, since it is probably a case of of -insertion, for the of is followed by a NP complement. . -um is found far less frequently. This latter verb often appears as usem or even usim. . The most frequent function of blong is as preposition, meaning “of ” (see Section 2); it derives from English belong (to).
References Arends, J., P. Muysken, & N. Smith (1995). Conclusions. In Arends, Muysken & Smith (Eds.), Pidgins and Creoles: An Introduction (pp. 319–330). Amsterdam: John Benjamins. Brøndal, V. (1950). Théorie des Prépositions. trans. Pierre Naert. Copenhagen: Ejnar Munksgaard. Clark, R. (1983). Review of Le Pidgin Bislama(n) et le Multilingualisme aux NouvellesHébrides by J.-M. Charpentier (1979)). Language in Society, 12, 539–542. Crowley, Terry. (1990a). The position of Melanesian Pidgin in Vanuatu and Papua New Guinea. In Verhaar (Ed.) (pp. 1–18). ——— (1990b). Verb serialization and prepositions in Bislama. In Verhaar (Ed.) (pp. 57–89).
Dennis Kurzon
Foskett, Colin. (1991). Together – a preposition? In Rauh (Ed.) (pp. 29–52). Haugeman, L. (1994). Introduction to Government and Binding Theory (2nd ed.). Oxford: Blackwell. Hawkins, Bruce W. (1993). On universality and variability in the semantics of spatial adpositions. In Zelinsky-Wibbelt (Ed.) (pp. 327–349). Jones, Frederick C. V. (1990). The grammatical items bin, ff and mcs in Sierra Leone Krio. Linguistics, 28, 845–866. König, Ekkehard and Bernd Kortmann. (1991). On the reanalysis of verbs as prepositions. In Rauh (Ed.) (pp. 109–125). Mann, Charles C. (1993). Polysemic functionality of prepositions in pidgins and creoles: The case of ‘fò’ in Anglo-Nigerian Pidgin. In F. Byrne & J. Holm (Eds.), Atlantic meets ˙ Pacific: A Global View of Pidginization and Creolization (pp. 57–67). Amsterdam: John Benjamins. Radford, Andrew. (1997). Syntactic Theory and the Structure of English. Cambridge UP. Rauh, Gisa. (1993). On the grammar of lexical and non-lexical prepositions in English. In Zelinsky-Wibbelt (Ed.) (pp. 99–150). Rauh, Gisa (Ed.) (1991). Approaches to Prepositions. Tübingen: Gunter Narr Verlag. Taylor, John, R. (1993). Prepositions: Patterns of polysemization and strategies of disambiguation. In Zelinsky-Wibbelt (Ed.) (pp. 151–175). Thomas, A. (1990). Language planning in Vanuatu. In Richard B. Baldauf & Allan Luke (Eds.), Language Planning and Education in Australasia and the South Pacific (pp. 234– 258). Clevedon: Multilingual Matters. Tryon, Darrell. (1987). Bislama: the national language of Vanuatu, Canberra: Pacific Linguistics. ——— (n.d). Evri samting yu wantem save long Bislama be yu fraet tumas blong askem. South Pacific: Media Masters. Verhaar, John W. M. (Ed.). (1990). Melanesian Pidgin and Tok Pisin. Amsterdam: John Benjamins. Zelinsky-Wibbelt, C. (Ed.). (1993). The Semantics of Prepositions: From Mental Processing to Natural Language Processing. Berlin: Mouton de Gruyter.
Prepositions in modern Judeo-Greek (JG) Biblical translations* Julia G. Krivoruchko Department of Foreign Languages, University of Haifa
.
Introductory definitions
For the purposes of this examination, by modern JG we mean Judeo-Greek from the sixteenth to the twentieth centuries. From sociolinguistic point of view, modern JG may be defined as a sociolect of Greek spoken by Romaniote Jews,1 or rather a group of such sociolects of different regions and communities. The exact relationship between them remains to be examined. The history of Jewish translations of the Bible into Greek extends over more than two thousand years. The continuity in this uniquely long tradition, argued as long ago as 1924 by D. S. Blondheim, was recently validated on many occasions by N. de Lange (e.g., 1982, 1993; see also Weinreich [1973] 1980: 78–79). Further on, we shall call the sub-language of these translations Biblical Judeo-Greek (=BJG). This somewhat lengthy abbreviation aims to emphasize the distinction between spoken JG and BJG, so that the results obtained in the research of the latter should not be mechanically extended to the former. A description of BJG prepositions must consider numerous heterogeneous factors. These include the dialect features of the text, its inner consistency, the stylistic inclinations of the translator/s,2 the relationship of the text to its original, the influence of rabbinical scholarship on the choice of syntactic constructions and lexical items, etc. Since this is the first attempt to approach this complicated subject matter, at times we limit ourselves to formulating questions rather than answering them.
Julia G. Krivoruchko
. Prepositions in earlier stages of BJG . Ancient BJG The use of prepositions in ancient BJG has been an object of extensive research. Biblical scholars long ago discovered that prepositions, however small and subtle, provide an excellent touchstone for understanding the historical and stylistic background of the text. A dissertation has been dedicated to almost every particular preposition, not to mention numerous articles and a mass of comparative scholarship.3 This extraordinary attention derives, however, not from the interest in the prepositions as such, but largely from the significance of ancient JG texts for western culture: the Septuagint is one of the most studied texts, and certainly the most studied translation in the world’s philological history, and the New Testament is a text of primary importance for Christian doctrine. Along with clarifying theological and semantic intricacies, philological investigations of the Septuagint have done much service for textology and linguistics. Inter alia, the study of prepositions helped to distinguish translations from Hebrew/Aramaic and original Greek compositions within the Biblical corpus. R. Martin (1960: 295) has demonstrated on statistical grounds that “the relative infrequency of certain prepositions is especially characteristic of translation Greek and stands in sharp contrast to normal prose Greek style”, cf. Martin (1974). To paraphrase his conclusion, certain Greek prepositions belonged to the periphery of active vocabulary of Biblical translators, and their functions were redistributed among the other prepositions. The prepositional system that resulted from such an attitude may be described as extremely “compressed”, i.e., quantitatively reduced and highly polysemic in its core parts. The rabbinical authorities, in all probability, did not recognize the Septuagint translation as precise enough. As it was rapidly embraced by the Church, the Hebrews needed a different one that would distance them from their ideological opponents.4 Besides, many Hellenized Jews no longer understood Hebrew.5 For this reason or another, new translations were produced for use in the Jewish environment. The most famous of them was that of Aquila. Composed before 140 AD, it drew on rabbinical scholarship with its persistent inclination to etymological reasoning. Aquila’s rendering of the original was judged by later Christian authorities as extremely literal (e.g., δουλεύων τῃ ‘Εβραικῃ λέξει Orig. Epist. ad African. 3).6 This approach inevitably influenced the translation of major parts of speech, and also the prepositions. A Paradenbeispiel is Aquila’s propensity to
Prepositions in Judeo-Greek Biblical translations
translate Heb. ét ‘direct object marker’ through σύν ‘with’, probably reinforced by the homography of this marker with the root of declinable preposition ét ‘with’ (e.g., bara ˘elohim ét haššamayim weét ha’arez Gen. 1 : 1 ἔκτισεν ὁ θεὸς σὺν τὸν οὐρανὸν καὶ σὺν τὴν γην). Modern scholars tend to explain this odd usage either as a homerism or as an influence of rabbinical exegesis.7 Whatever factor influenced the final choice, it was surely not linguistic normativity or the acceptability of the translation for the Greek speakers, since normally σύν was not constructed with the accusative.8 Summarizing, the system of prepositions in ancient BJG did not coincide with that of contemporary normative Greek. Variation in the number of prepositions and in their syntactic government may be observed. . Medieval and later BJG Jews did not abandon the practice of translating the Scriptures into Greek after Aquila.9 Unfortunately, we possess very few data about the language of Byzantine Greek-speaking communities, as the quantity of surviving texts is small. The most significant of them are presented in Table 1. We have included continuous texts, and also glossaries since they frequently contain translations of prepositional phrases. Manuscripts whose Jewish provenance is disputable, such as Graecus Venetus, Vaticanus Gr. 343, were excluded. As one may see, only one full Pentateuch text survived, the famous Humaš mi Qušt’a. All the rest are fragments chosen on liturgical or linguistic grounds: the fragments of Jonah and Isaiah are haftarot, those of Daniel and Esdra are in Aramaic. The dating of many texts is amazingly imprecise. In all the above editions only a few sentences are dedicated to prepositions; neither systematic presentation nor survey has ever been attempted.10 A huge chronological gap between antiquity and modern times in this research field still has to be filled.
. Prepositions in modern BJG During our work on an edition of a JG manuscript YBZ 3519 (Krivoruchko, to appear) it became clear that the prepositional system of this manuscript is unique and differs greatly from the corresponding systems of the texts presented in Table 1, as well as from less exotic forms of Greek. In all probability, the manuscript originates from the 19th century Epirus (Giannina?), while the composition itself may be slightly older.11 It seems to belong to the tradition that flourished in northwestern Greece and the neighboring Ionic islands, as it
Julia G. Krivoruchko
Table 1. Selected sources of medieval and later BJG Date of composition*
Content
Editor
Date of edition
Script of edition
600–1200, ms
Eccles. (fragm.) idem + gloss. to I Kings Jon. gloss. to Mal., Job cet. Genizah fragments
D. Blondheim N. de Lange
1924 1982
Gr Heb+Gr
D. C. Hesseling N. de Lange
1901 1980 1996
Gr Heb+Gr Heb+Gr
marginal gloss. Fb in Codex Ambrosianus interlinear gloss. to Prov. Constantinopolitan Pentateuch
A. E. Brooke, N. McLean H. P. Rüger
1906– 1940 1959 1547 1890 1897 1576 1914 1957– 1958 1971– 1981
Gr
a 1100–1200, ms 900–1100, ms
“Byzantine”, ms 700–1100, ms a 1547, pr
a 1576, pr a 1627, ms 1500–1700, ms a 1839, ms
Job Jer., Dan., Ezra (fragm.) gloss. to Ps., Lam., Eccl. (fragm.) gloss. to Megilloth
L. Belléli D. C. Hesseling M. E. Fabian** A. Danon D. Goldschmidt M. Altbauer, Y. Shiby
Gr Heb Gr Gr Heb+Gr Heb+Gr Heb+Gr
*For the manuscripts (ms) we cite the editor’s opinion; for the printed sources (pr) the date of publication. **This edition was inaccessible to us.
exhibits some features of the local dialects. The manuscript contains the translation of four haftarot from Isaiah.12 As it is one of the most extensive coherent texts in modern BJG, it may be considered a representative sample for the period under investigation. The inquiry into modern BJG prepositions is complicated by the absence of a description of the prepositional systems of the Modern Greek dialects that seem to have formed the basis (or at least, the closest comparandum) for the corresponding JG sociolects. The dialectologic description of Epirus dialects, in many respects praiseworthy, is not satisfactory for prepositions. To our knowledge, no special inquiry has gone into this matter. In the grammar of the northern dialects by A. A. Papadopulos (1927) sufficient attention has not been paid to prepositions. The prominent dialectologists E. A. Bongas and N. B. Kosmas were given awards by the Athens Academy for their research on Epirus linguistics, but the major part of their work was dedicated to the lexicon and not to the description of the grammar. All these scholars commonly omitted the phe-
Prepositions in Judeo-Greek Biblical translations
nomena that were identical to standard Greek (SG), only occasionally noting that certain SG forms were not used. Therefore, we often have to draw our conclusions e silentio, hardly the best point of departure. We proceed now to review the prepositions of modern JG Biblical texts: first as a system, secondly from a syntactic viewpoint, and finally from a semantic and functional perspective. Phonetic details irrelevant for the present discussion, such as the exact degree of consonant palatalization or the prenasalization of plosives are ignored. For the convenience of the reader only, the JG accents are marked as identical with Epirus Greek (EG).13 . The system of modern BJG prepositions: their quantity and quality Let us start with simple prepositions, otherwise known as “true” or “proper prepositions”, and proceed to so-called “improper” ones, which are essentially collocations of adverbs of place with true prepositions.14 Historically, improper prepositions continue the prepositional adverbs of ancient Greek and the combinations of adverbs with proper prepositions. Grammatical authorities often treat them in like manner, as they are followed by weak personal pronouns. As we shall see, they also correspond to proper prepositions in Hebrew. Table 2 presents a comparison between the SG system of proper prepositions, the comparable system of the Giannina region of Epirus, and that of the JG Biblical translation we have studied. As a basis of comparison, we have chosen the checklist of SG prepositions by Holton et al. (1999: 407–408) as the most comprehensive. Taking into account the nature of our text, we have excluded the prepositions used only in mathematical discourse. The “prepositions of katharevousa origin used in formal discourse” (ibid.: 405–407), are marked with shading. For reasons already mentioned, the Epirus forms are given only if they differ from the standard. To the best of our knowledge, there is no evidence that Gianniotes avoided any particular preposition. However, the plus sign is used to mark the usage explicitly confirmed by dialectologists. As is seen from Table 2, the number of prepositions used by our translator is much smaller than in the standard and the dialect. Only eight proper prepositions occur in our material. As the number of the published BJG texts increases, some other prepositions are likely to appear. One wonders whether dialect and JG systems included the almost fully synonymous items μέχρι/ίσαμε/έως/ως ‘until’ (of time)/‘as far as’ (of place), as we found only ως. In other case of near synonyms δίχως/χωρίς ‘without’, we found only δίχως.
Julia G. Krivoruchko
Table 2. BJG proper prepositions
Another prominent feature is the total absence of katharevousa prepositions in our material. Even the katharevousa prepositions recorded by the dialectologists in the area are absent. As the katharevousa prepositions form an integral part of formal and elevated SG styles, one possible conclusion is that the translators avoided such styles or perhaps were only partly familiar with them. After all, learned discourse in Greece was predominantly a Christian discourse, mediated, supported, and developed by the Church. Its highest form was intimately connected with the liturgy. The Jewish communities might have wanted to distance themselves linguistically from Christians. They also might have felt no need to develop special liturgical styles, as they already had a special liturgical language, Hebrew. How likely is it that the liturgical text was composed of entirely colloquial, therefore wholly profane, elements?15 An archaic impression, if required, could be achieved only within the limited repertoire of demotic prepositions. A case to consider in this connection is the usage of εις/σε. Historically, σε developed from εις and became a basic preposition, with the central meaning “in”. In SG, εις persisted in phraseological expressions. In JG Biblical translations we ob-
Prepositions in Judeo-Greek Biblical translations
serve εις/σε/σ’ functioning as variants of the same preposition with marked preference for the full form, appearing about nine times more often than the two shorter forms together. The same relationship exists between διά and για. In this case, as in the previous one, the latter developed from the former. In the BJG text we observe them functioning synonymously with twice more διά than για, not to mention διά as a part of the frequent connector διά ότι ‘because’ (= Heb. ki). The preponderance of εις and διά may be interpreted as an objective phonetic fact (i.e., as a stylistically neutral feature) or as a result of conscious application of certain poetic principles. A statistical comparison with demotic texts from the same region and period would help to evaluate the true nature of this phenomenon.16 Note also that the members of pairs εις vs. σ’, διά vs. για have a different metrical value: two syllables vs. one. As the JG Biblical translations were read during the liturgy, this quality might be very welcome. Metrical variations could be also achieved by interchanging the almost synonymic από (in causal sense) and εξιτίας. In SG, prepositions such as από, για, με, and σε may be optionally elided to απ’, γι’, μ’, and σ’ (Tr 1012 β, γ;17 Holton et al. 1999: 371). Interestingly, the elisions are not popular with our Jewish author. Of all the above variants only σ’ is found, and even that is used very rarely. Several reasons may have contributed to this situation. First, the hiatus must have occurred in EG less than in SG, and local JG might have shared this feature with the wider linguistic environment.18 Second, elision might be judged too colloquial to appear in a liturgical text. A telling detail is that our translator avoided the dialect form οχ’ which is found in other modern JG texts of Epirus provenance, and preferred από. He thereby rejected regional limitations and aspired to the newly emerging national standard. However, only the dialect features that the author was aware of were eliminated: slight phonetic variations creep in unnoticed. For instance, in the majority of cases we found the reduced form εξιτίας ‘because of ’ (50:2 bis, 54:14, 51:13) as opposed to a single use of normative εξαιτίας (49:19). To sum up, the principles of Jewish Biblical translation in ancient and in modern times seem to have produced the no-surplus prepositional systems. Yet the “compression” is not total, as the system still offers the translator several options to construct the texts of a complex rhetorical and metrical structure.
Julia G. Krivoruchko
. Syntax of modern BJG prepositions Let us now consider the above checklist from a syntactic point of view. In SG, the demotic prepositions are normally constructed with the accusative, except for εξαιτίας ‘because of ’, εναντίον ‘against’, and μεταξύ ‘between’, which take the genitive, and certain syntactic structures which permit the nominative. In a few phraseological expressions, the relict forms of dative may be found (Holton et al. 1999: 371–405). By contrast, the majority of katharevousa prepositions take genitive, while some of them may adjoin also accusative, nominative and the relict forms of dative (ibid.: 405–407). The system of JG prepositions under analysis includes only one preposition that theoretically should be followed by genitive, namely εξιτίας. In practice, in half of the cases it is not. We find εξιτίας τη δίψα ‘because of thirst’19 (50:2), εξιτίας το πολ κατοίκων ‘because of (too) many inhabitants’ (49:19), constructed with the accusative just like the basic demotic prepositions. The generalization of the accusative as the sole prepositional case is a well-known tendency, particularly widespread in northern dialects (Kontosopoulos 1994: 101; Horrocks 1997: 216–217). As can be seen from Table 3, a similar syntactic peculiarity is observed with “improper” prepositions.
Table 3. BJG improper prepositions SG ανάμεσα (σε)* απέναντι (από/σε) γύρω (από/σε) δίπλα (σε) έξω (από) κάτω (από) κοντά (σε) μαζί (με) μακριά (από) μέσα (από) (σε) μπροστά (από/σε) πάνω (από) (σε) πίσω (από) πλάι (σε)
Modern BJG καταγναδίς/καταγναδίν
EG options Gen
αγνάντια +
σιμά
Gen
ομπροστά απάνω/απάνου
Gen Acc
σ’μά
* Από and σε are placed in separate brackets when in combination with the preposition they give different meanings, and together if synonymous.
Prepositions in Judeo-Greek Biblical translations
In SG, adverbs of place such as “under”, “between”, “around” may be followed by the genitive of a weak personal pronoun or by one of the basic prepositions από or σε, which, in turn, take accusative (Holton et al. 1999: 365–369). In JG the accusative appears directly after the adverbs without any από or σε interceding: απάνου το βονό ψηλό ‘on the high mountain’ (40:9); απάνου απαλάμες ‘on the palms’ (49:16); απάνω την πλάτη ‘on the back’ (49:22), etc. Secondly, modern JG translations exhibit an unexpected but perfectly Greek phenomenon of substantivation of prepositions. To δίχως ‘without’ was used to translate Heb. efes ‘nothing’ (40:17). Non-standard, but well explicable is the treatment of Heb. k˘e ‘as, like’. Classified by certain grammarians as a conjunction despite its ability to add personal suffixes, it betrays its contradictory nature in the translation process.20 The expected EG equivalents σαν + acc. and ως + nom. converge into σαν + nom.: k˘ero‘eh σαν ο πιστικός ‘as a shepherd’ (40:11). In certain contexts we cannot determine whether the following form is nominative or accusative: Heb. k˘emar midli σα σταλαματιά από σίκλο ‘as a drop from a pail’ (40:15). A remarkable feature of EG, fairly widespread typologically, is the usage of από and sometimes of σε instead of genitive (Kosmas 1997: 57, cf. Tr 523 γ). Undoubtedly, it reflects the general tendency towards the attrition of genitive in Modern Greek. This phenomenon is also discernible in BJG texts, where από + gen. is one of the options to translate the Hebrew status constructus. This construction, however, may also be translated by pure genitive, and the translator’s choice is hard to predict. Awkward structures with three successive από are found, e.g., in 49:19. Consequently, the syntax of modern BJG prepositions can be characterized as not wholly identical, but very close to that of EG; neither does it contradict the general tendencies of syntactic development in the region. . Modern BJG prepositions vs. their Hebrew originals Our understanding of modern BJG prepositions would be incomplete without analysis of the translation technique, which is the ultimate raison d’être of this peculiar system. First, note that certain passages of our text fully harmonize with EG, as they tend more to dynamic-equivalence translation than to formal equivalence. At times the translator was less concerned to “reproduce [. . . ] formal elements, including: (1) grammatical units, (2) consistency in word usage, and (3) meanings in terms of the source context” (Nida 1964: 165), but aimed at “the closest natural equivalent to the source-language message” (ibid.: 166). Different
Julia G. Krivoruchko
meanings of Hebrew prepositions came to be translated by different Greek prepositions. The principal equivalencies may be stated as follows: Table 4. Hebrew-BJG translational equivalents Hebrew preposition
Semantic/syntactic function
BJG equivalent
al b˘e
‘on’ (1)* locative (‘in’) (2) instrumental (3) price (‘with’, ‘by’) (4) final/ causal (5)
απάνω/απάνου εις (Tr 1013 β) με (Tr 1013 β)
el ét ét (et) k˘e l˘e
min
neged
adverbial (6) directional (‘towards’) (7) comitative (‘with’) (8) DIR OBJ (9) comparative (‘as’) (10) INDIR OBJ/commodi (11) final (‘for’) (12) temporal (‘until’) (13) separative (‘from’) (14) comparative (‘then’) (15) causal (‘because of ’) (16) ‘against’ (17)
*See examples.
Examples: (1) al har-gavoah απάνου βονό. ψηλό on mountain. high ‘on the high mountain’ (40:9) (2) bammidbar εις. τον. έρημο ‘in the desert’
(40:3)
(3) b˘ehazaq με. χέρα. δυνατή with hand strong ‘with strong hand’ (40:10)
διά/για (Tr 1016 α, β) εξιτίας, από (Tr 1028 η) με (Tr 1013 γ) εις (Tr 1014 α) με (Tr 1013 α) acc. /με σαν (Tr 1013 β) gen. διά/για (Tr 1013 α, β) ως (Tr 1017) από (Tr 1028 α) από (Tr 1028 κ) εξιτίας καταγναδίς/καταγναδίν
Prepositions in Judeo-Greek Biblical translations
(4) b˘ekesef με. ασήμι ‘with silver’
(52:3)
(5) b˘ekol
hatto’teha διά. όλα τα. φταιξίματά. της because of all the sins her ‘because of all her sins’ (40:2)
(6) [harimi] bakoah [qolék] με. δύναμη with strength ‘mightily’ (40:9) (7) el Avraham avikem εις τον. Αβραάμ τον. πατέρα. σας to the. Avraham the. father. yours ‘to your (sc. masc. pl.) father Avraham’ (51:2) (8) ito με. ταύτον ‘with him’ (40:10) (9) weet-banayik και. με. τα. πιδιά σου and. with. the. children. yours ‘and your (sc. fem. sg.) children’ (49:25) (10) k˘ero‘eh σαν. ο. πιστικός as. the. sheperd ‘as a shepherd’ (40:11) (11) l˘e’ummim [natattiw] εις. τις. φυλές ‘to the peoples’ (55:5) (12) l˘ema‘˘a´séhu διά τη. δουλειά. του for. the. work. his ‘for his work’ (54:16)
Julia G. Krivoruchko
(13) l˘e‘olam ως. πάντα for. always ‘forever’ (40:8) (14) [laqahti] miyyadék από. το. χέρι. σου from. the. hand. yours ‘from your (sc. fem. sg.) hand’
(51:22)
(15) mé’efes [watohu nehš˘evu-lo] κατώτιρα από. το. δίχως lower of. the. without ‘less than nothing’ (40:17) (16) [tézri] miyyošév εξιτίας. το. πολ. κατοίκων because. of. the. many. inhabitants ‘because of (too) many inhabitants’ (49:19) (17) negdo καταγναδίν. του ‘against him’
(40:17)
The translator does not attempt to reproduce the etymology of Hebrew compound prepositions, but their semantics only.21 Unlike Aquila, he never inserts πρόσωπον or its local dialect variants, such as προυσώπατα (Bongas 1964: 19) to translate the compounds with panim: mipnéy hamat εξιτίας τη σκάνια ‘because of the fury’ (51:13), lefanaw ομπροστά του ‘in front of Him’ (40:10), cf. Aquila Isa. 7:2 ἀπό (ἐκ Q) προσώπου ἀνέμου for mipnéy-ruah ‘from the wind’. The author is eager to distinguish different prepositions explicitly and accurately. Alongside έρχεται εσένα for ba’u-lak (49:18), we find ρτει εις εσένα ‘will enter you (fem. sg.)’ for yavo’-bak (52:1). Quite often, his desire to produce an intelligible and semantically precise text results in Greek having more prepositions than the original. In ´s’u-marom éynéykem ‘lift your (masc. pl.) eyes’ (40:26) the translator inserts εις to express direction: σηκώστε εις το ψηλάδι τα μάτια σας. Direct object after verbs of motion, a characteristic of Hebrew, is regularly avoided: Mizrayim yarad-ammi εις του <>22 ‘My people went down to Egypt’ (52:4); halak h˘ašékim ιπήγαινε εις τα σκοτεινά ‘went into darkness’ (50:10).
Prepositions in Judeo-Greek Biblical translations
In coordinate constituents or constructions understood as such, prepositions are inserted to clarify the syntactic function and to underline parallelism. In 52:2 hitna‘˘ari mé‘afar ‘shake off the dust’, the preposition m˘e is translated through από. In the immediately following clause hitpathi mosréi zawwa’rék (q˘eré reading pro hitpathu) ‘solve the bonds of your (fem. sg.) neck’, mosréi zawwa’rék was also supplied with an από. In 40:19 w˘ezoréf bazahav y˘eraqq˘e‘ennu ur˘etuqot kesef zoréf, ‘the jeweller decorates it with gold and silver chains . . .’ με το χρυσάφι translates bazahav, and to άλυσους ασημένιους another με is added. A substantial amount of preposition usage seems to have been inspired by demands of metric balance and rhetorical symmetry. Occasionally the translator respects the valencies of Greek verb even in cases where a more Hebrew-like version might cause no misunderstanding. W˘e’el-mi t˘edammyun él ‘and to whom you will compare God’ (40:18) he translates through με ποιόν. The use of εις ποιόν would not have created any particular misapprehension. But unlike his ancient predecessors, our interpreter was not inclined to confine himself to the already used (and more frequent) equivalent of el. Of course, it is not always possible to satisfy the demands of precision in the traditional sense and those of Greek grammar. Fortunate instances are those like w˘e’et-y˘erivék anoki ariv ‘and I will fight your (fem.sg.) enemy’, where the direct object is replaced by a prepositional one: με το μαλωτή σου εγώ θέλα μαλώσω (49:25). The rendering resulted in a natural Greek construction, at the same time echoing the ancient tradition of translating ét through σύν.23 Quite often, through the resourceful use of a preposition an otherwise complex translational problem may be solved. For example, the literal translation of the phraseological expression w˘e’ekr˘etah lakem b˘erit (55:3) ‘I cut you (masc. pl.) an alliance’ would be almost incomprehensible, but the insertion of the preposition με it makes it somewhat clearer: κόψω με σας συνθήκη ‘I will conclude with you an alliance’. In 54:12 w˘e´samti kadkod šimšotayk uš˘e‘arayik l˘e’avnéy eqdah wekol-g˘evulék l˘e’avnéy héfez ‘and I will put (> change) your (fem. sg.) windows into rubies, your (fem. sg.) gates into valuable stones and all your (fem. sg.) boundary into precious stones’, l˘e is replaced by με ‘with’. The promise to transform all the gates and all the boundaries of Jerusalem into jewels has shrunk into the far less sumptuous but more realistic (and therefore more understandable) plan to decorate it with precious stones. However, finding the middle way between Greek and Hebrew is not always the case, and the compromise between exactness and linguistic normativity may be detrimental to the latter. Attempts to follow Hebrew closely lead now and then to peculiar and nowadays obsolete forms of Greek. Such a case is με
Julia G. Krivoruchko
δίχως ασήμι = b˘elo’-kesef (55:1), where b˘e = με, lo’ = δίχως, lit. ‘with without money’= ‘gratis’.24 This expression may be compared to informal με χωρίς, attested in modern SG sources (LKNE 1998: 387, s.v. δίχως 1) but rejected by grammarians (Tr 1996: 1029). Constructions such as ποιός εις εσάς for mi bakem ‘who of you (masc. pl.)’ (50:10) and λυπηθεί με σένα for yanud lak ‘will console you (fem. sg.)’ (51:19) exhibit an orientation to formal equivalence, since nothing seems to rule out the normal ποιος από σας and λυπηθεί εσένα. As there are no prefixed verbs in Hebrew, the Greek prefixed verbs are almost never activated in a close translation like ours. Prepositions compensate for this deficiency, appearing in combinations where normal Greek would not use any. Instead of [prefixed verb + direct object] we observe [unprefixed verb + preposition + object]. For instance, in the above example hitna‘˘ari mé‘afar ‘shake off (imperat. fem.) the dust’ (52:2) we encounter τινάξου από το χώμα and not αποτινάξου το χώμα. Literalism sometimes ends in total disregard of the syntactic structure of the target language. The text contains a number of examples where, to the best of our knowledge, the use of prepositions is hard to justify in any variety of Greek. Such are dab˘eru al-lév ‘speak to the heart’ (40:2) = ομηλάτε απάνω την καρδία instead of ομηλάτε στην καρδία. The Hebrew phraseological expression šomé‘a b˘eqol meaning “obey” is reproduced as κούει εις τη φωνή (50:10), šim‘u élay as κούστε . . . εις εμένα (51:1), cf. 55:2. The translator ignored the literary equivalent υπακούω and seemed unmoved by the fact that ακούω normally takes pure accusative. Translating y˘eré’ yhwh ‘God-fearer’ (50:10) as φοβάται από τον A (= Αδοναι), our author adds από to a standard direct object, just as his predecessors did in the Septuagint (Deut., Ps., IV Kings, etc.). This anomalous usage makes evident the translational nature of the text, as it can only be explained as a calque from the Targum and/or Metsudat David ad locum or some similar source. In antiquity the preposition l˘e in the constructions like hayah l˘e ‘to become something’ was translated through εἰς: ἐγένετο . . . εἰς ψυχὴν ζωσαν ‘became a living soul’ (Gen. 2:7). In ancient and modern Greek, auxiliaries of that kind do not require prepositions at all.25 Despite this, our translator made an effort to preserve at least some preposition in the corresponding sentences. In 40:4 w˘ehayah he‘aqov l˘emišor ‘a detour will become a plain’ he used για: θα είναι ο στραβόδρομος για σάδι. In every particular case, it is difficult to foresee where the line between the normal Greek and the Hebrew-like syntax will be drawn, and by which cri-
Prepositions in Judeo-Greek Biblical translations
teria the final version will be shaped. The translator’s decisions are not easily predictable. However, he certainly tries to take into consideration the intelligibility of the text for the monolingual Judeo-Greek-speaking audience and does not ignore language standards without weighty traditional reasons.
. Conclusions: Modern BJG in a wider perspective The analysis of the BJG prepositional system provides valuable aid for understanding the entire range of Greek Biblical translations, from the oldest we have mentioned to those still to be published. A change of preposition often signals not only historical grammar development, but also the transformation of cultural and pragmatic attitudes and last but not least changes in the understanding of the text. Biblical translations originating from traditional Jewish environments form a highly specific class of linguistic corpora (res. sub-languages). The primary goal of these translations was to preserve the sense and also the form of the original: they would mold the target language into the forms of Hebrew/Aramaic.26 In so doing they constantly challenged the norms of the target language, which is an intriguing theoretical problem in itself. The phenomena resulting from the imitation of syntax and semantics of Hebrew, including Hebrew prepositions, will have much in common in all Jewish calque languages, such as Ladino, Ivretaytsh (Yiddish-Taytsch, Lerntaytsh), Sharh, etc. Comparative research in this field will be very worthwhile.
Notes * I would like to thank Prof. Theodosia-Soula Pavlidou for helpful criticism and the staff of Yad Ben Zvi Institute, Jerusalem, for their kind permission to use their archives and library. . Romaniotes are a Jewish minority in Greek-speaking lands that predated the immigration waves from Italy, Spain and Portugal and almost ceased to exist during World War II. . Although the translation is unquestionably traditional and might well be produced by a group, further on we will use the singular for convenience. . Classic works are Johannessohn (1910) and Johannessohn (1926). Particularily rich is the bibliography on New Testament prepositions that makes extensive use of the Septuagint studies and papyrological research; see Perschbacher (1995: 233–234). . Cf. Swete 1900: 30. . See Orig. Ep. ad African. 2.
Julia G. Krivoruchko . Similar wording is often found in modern works, e.g., “a painfully literal translation” (Nida 1964: 12), “crudities and absurdities which beset the excessive fidelity of Aquila” (Swete 1900: 45). See also Fernández Marcos (1979: 98–100), Soisalon-Soininen (1972: 177– 184), Field (1875: XXI–XXIV). . This translation is also found in Septuagint, e.g., Eccles. 2: 17. For discussion see Hyvärinen (1977: 26–29), Barthélemy (1963: 10–27), Swete (1900: 39, n. 2). . See e.g., the reaction of Hieronimus in Epist. ad Pammach. 11. . On Theodotion, Symmachus, Hexapla, and other later versions of probable Jewish origin see Fernández Marcos (1979: 107–159) (the last edition of this work was unavailable to us), Swete (1900: 42–86). For Byzantine translations see De Lange (1992), De Lange (1993), Bádenas de la Peña (1996). . Hesseling (1897: XLVII–XLVIII), Swete (1900: 41, 48), De Lange (1982: 66–67, 71). The reviewers were equally brief: Belléli (1897: 136, 143). . For the paleographical description of the manuscript, see Sznol (1996–1997). Selected fragments of the text and discussion of its provenance may be found in Krivoruchko (1999: 647–653). . Further on we refer to this text by chapter and verse only. . We discuss the BJG phonetics in detail in Krivoruchko (to appear). . Cf. eigentlichen Praepositionen vs. uneigentlichen Praepositionen in Thumb (1910: 93, 101–102). . On using the JG Biblical translations during the liturgy see Matsas (1953: 11, n. 1). . Appropriate material would be a narrative text, like General Makriyannis’s memoires, and a text of non-Christian religious content such as the mi‘radj poem from Epirus published by F. Kotzageorgis (1997). . In the Triandafillidis’s grammar (=Tr) references are to paragraphs. . See Bongas (1964: 15) and Kosmas (1997: 57). The data about the relevant facts are deficient and contradictory, but the phenomenon itself is undoubted. . All the examples are given in my translations, which aim to show the exact way the text was understood by the JG interpreter. . E.g., in Bauer and Leander (1922: 647, 650–651); but Gesenius and Kautzsch (1909: 313). . We use this term as an equivalent of “zusammengesetzte Praepositionen”, see Bauer and Leander (1922: 646–647). . The end of the line is missing, but the preposition is clearly visible. . Not that the translator was consciously imitating his ancient predecessors: he was hardly aware of their existence. . The same combination serves to translate mibli (Eccles. 3:11) (Altbauer & Shiby 1971– 1981: 408). . “The use of εἰς with the accusative after εἶναι and γενέσθαι may safely be regarded as a Hebraism” (Conybeare and Stock 1905: 81); cf. Hyvärinen (1977: 46–47), Belléli (1890: 290). . E.g., for the Ladino translations see Sephiha (1973), Bunis (1996).
Prepositions in Judeo-Greek Biblical translations
References Altbauer, Moshe, & Shiby, Yakov. (1971–1981). Glosar yavani-yehudi le Hamesh ha-Megillot. (A Judeo-Greek glossary of the Hamesh Megillot). Sefunot, 15, 367–421. Bádenas de la Peña, Pedro. (1996). Aproximación a la historia de las versiones de las escrituras al griego vulgar. Trans. Revista de traductología, 1, 8–37. Barthélemy, Dominique. (1963). Les devanciers d’Aquila. Leiden: Brill. Bauer, Hans and Pontius Leander. (1922). Historische Grammatik der hebräischen Sprache des Alten Testaments. [Halle (Reprint Ed.). (1965)]. Hildesheim: Olms. Belléli, L. (1890). Une version grecque du Pentateuque du seizème siècle. Revue des Études Grecques, 3, 289–308. Belléli, L. (1897). Rec. on D. C. Hessling (Ed.). Les cinq livres de la loi (la Pentateuque). Revue des Études Juives, 35, 132–155. Blondheim, D. S. (1924). Échos du Judéo-Hellénisme. Étude sur l’influence de la Septante et d’ Aquila sur les versions néo-grecques des Juifs. Revue des Études Juives, 78, 1–14. Bongas, E. A. (1964). Ta Glossika Idiomata tis Ipeirou (Voreiou, Kentrikis kai Notiou). The Dialects of Epirus (North, Central and South). v. 1. Gianniotiko kai Alla Leksilogia. (The Gianniote and Other Vocabularies). Ioannina: Society for Epirus Studies. Brooke, Alan E., & McLean, Norman. (1906–1940). The Old Testament in Greek. Cambridge: Cambridge University Press. Bunis, David M. (1996). Translating from the head and from the heart: the essentially oral nature of the Ladino Bible-translation tradition. In W. Busse & M.-C. Varol-Bornes (Eds.), Hommage à Haim Vidal Sephiha (pp. 337–357). Berne: Peter Long. Conybeare, F. C., & Stock, S. G. (1905). Grammar of Septuagint Greek. Boston: Ginn. Danon, Abraham. (1914). Meïrath Énaïm, version en néo-grec et en caractères hébraïques. Journal Asiatique. 11e serie. 4, 5–65. De Lange, N. R. M. (1980). Some new fragments of Aquila on Malachi and Job. Vetus Testamentum, 30, 291–294. De Lange, N. R. M. (1982). Two Genizah fragments in Hebrew and Greek. In J. A. Emerton & S. C. Reif (Eds.), Interpreting the Hebrew Bible. Essays in honour of E. I. J. Rosental (pp. 61–83). Cambridge: Cambridge University Press. De Lange, N. R. M. (1992). Byzantium in the Cairo Genizah. Byzantine and Modern Greek Studies, 16, 34–47. De Lange, N. R. M. (1993). Jews of Byzantium and the Greek Bible: Outline of the problems and suggestions for future research. In G. Sed-Rajna (Ed.), RASHI 1040–1990: Hommage à Ephraïm E. Urbach (pp. 203–210). Paris: Éditions du Serf. De Lange, N. R. M. (1996). Greek Jewish Texts from Cairo Genizah. Tübingen: Mohr. Fernández Marcos, N. (1979). Introducción a las versiones griegas de la Biblia. Madrid: CSIC. Field, F. (1875). Origenis Hexaplorum quae supersunt sive veterum interpretum Graecorum in totum Vetus Testamentum fragmenta. Prolegomena. v. 1. Oxford: Clarendon. Gesenius, Wilhelm. (1909). Hebräische Grammatik völlig umgearbeitet von E. Kautzsch. Leipzig [Reprint (1962). Hildesheim: Olms.] Goldschmidt, Daniel. (1958). Tirgumei ha-mikra leyivanit meet yehudim bnei ha-mea hatav-zain. (The Bible translations by the 16th century Jews). Kiriat Sefer, 33, 131–134.
Julia G. Krivoruchko
Hesseling, Dirk S. (Ed.). (1897). Les cinq livres de la loi (Le Pentateuque). Leiden: von Doesburgh; Leipzig: Harrasowitz. Hesseling, Dirk S. (1901). Le livre de Jonas. Byzantinische Zeitschrift, 10, 208–217. Holton, David, Mackridge, Peter, & Philippaki-Warburton, Irene. (1999). Greek: A Comprehensive Grammar of the Modern Language. London: Routledge. Horrocks, Geoffrey. (1997). Greek: A History of the Language and its Speakers. London: Longman. Hyvärinen, Kyösti. (1977). Die Übersetzung von Aquila. Uppsala: Almqvist and Wiksell. Johannessohn, M. (1910). Der Gebrauch der Kasus und der Präpositionen in der Septuaginta I. Kirchhein N.-L.: Max Schmersow. Johannessohn, M. (1926). Der Gebrauch der Präpositionen in der LXX. Goettingen: Vandenhoeck and Ruprecht (Mitteilungen des Septuaginta-Unternehmens. Bd. 3, 3). Kontosopoulos, N. G. (1994). Dialektoi kai Idiomata tis Neas Ellinikis. (The Major and Minor Dialects of Modern Greek). Athens: n.p. Kosmas, N. V. (1997). To Glossiko idioma ton Ioanninon. (The Dialect of Ioannina). Athens/Giannina: Dodoni. Kotzageorgis, Fokion. (1997). To islam sta Valkania: Ena ellinofono mousoulmaniko xeirografo apo tin Ipeiro tou 18ou aiona. (Islam in the Balkans: A Greek Muslim Manuscript from the 18th Century Epirus). Athens: Sindesmos ton en Athinais Megalosxoliton. Krivoruchko, J. (1999). Ta ellinika xeirografa se Evraiki grafi: I proeleusi kai ta kiria xaraktiristika tous glossas tous. (The Greek manuscripts in Hebrew script: their provenance and main linguistic features). In A. Argyriou, K. Dimadis and A. D. Lazaridou (Eds.), The Greek World between East and West 1453–1981. Acts of the 1st European Congress of Modern Greek Studies, Berlin, 2–4 October 1998 v. 1 (pp. 647–660). Athens: Ellinika Grammata. Krivoruchko, J. (To appear). Haftaroth from Isaiah in Epirus Romaniote Translation. LKNE = Aristotle University of Thessaloniki. Institute of Modern Greek Studies (Manolis Triandafillidis Foundation). (1998). Leksiko tis Koinis Neoellinikis (Dictionary of Standard Modern Greek). Thessaloniki: Institute of Modern Greek Studies. Martin, R. (1960). Some syntactical criteria of translation Greek. Vetus Testamentum, 10, 295–310. Martin, R. (1974). Syntactical Evidence of Semitic Sources in Greek Documents. Cambridge, MA: Society of Biblical Literature. Matsa, Josef. (1971–1981). Ha-shira ha-yehudit bivanit. (Jewish poetry in Greek). Sefunot, 15, 235–365. Matsas, I. (1953). Gianniotika evraika tragoudia. (Gianniote Jewish Songs). Ioannina: Ipeirotiki Estia. Nida, Eugene A. (1964). Toward a Science of Translating. With Special Reference to Principles and Procedures Involved in Bible Translating. Leiden: Brill. Papadopoulos, A. A. (1927). Grammatiki ton voreion idiomaton tis neas ellinikis glossis. (Grammar of the Northern Dialects of Modern Greek). Athens: n.p. Perschbacher, W. J. (1995). New Testament Greek Syntax: An Illustrated Manual. Chicago: Moody Press. Rüger, H. P. (1959). Vier Aquila-Glossen in einem hebräischen Proverben-Fragment aus der Kairo-Geniza. Zeitschrift für die Neutestamentlische Wissenschaft, 50, 275–277.
Prepositions in Judeo-Greek Biblical translations
Sephiha, H. V. (1973). Le Ladino, Judéo-espagnol calque: Deutéronome, versions de Constantinople (1547) et de Ferrare (1553). Paris: Centre de Recherches Hispaniques, Institut d’Études Hispaniques. Soisalon-Soininen, I. (1972). Einige Merkmale der Übersetzungweise von Aquila. In J. Schreiner (Ed.), Wort, Lied und Gottesspruch: Festschrift für Joseph Ziegler, V. 1 (pp. 177–184). Würzburg: Echter. Swete, Henry Barclay. (1900). An Introduction to the Old Testament in Greek. Cambridge: Cambridge University Press. Sznol, Shifra. (1996–1997). Haftara Nehamu (Yeshayahu 40, 1–26) batargum leyehudityevanit. (The haftara “Nehamu” (Isaiah 40, 1–26) in the Judeo-Greek translation). Beit Mikra, 3, 332–342. Thumb, Albert. (1910). Handbuch der Neugriechischen Volkssprache. Strassburg: Trübner. Tr = Triandafillidis, M. (Ed.). (1996). Neoelliniki grammatiki tis dimotikis. (Modern Greek Grammar of Demotic). Rev. ed. Thessaloniki: Aristotle University of Thessaloniki. Institute of Modern Greek Studies (Manolis Triandafillidis Foundation). Weinreich, Max. [1973] (1980). History of the Yiddish Language. (Shlomo Noble, Trans.). Chicago/London: University of Chicago Press.
Quddiem and some remarks on grammatical aspects of Maltese prepositions Rami Saari Department of Linguistics The Hebrew University of Jerusalem, Israel
Introduction Maltese, a modern vernacular form of Arabic with its own linguistic standards and peculiarities, is one of the only Semitic languages written in the Latin alphabet.1 It is the national language of the Republic of Malta, where it is spoken by about 350,000 inhabitants (Vanhove 1993: 1) and where it shares with English the status of an official language. Though Maltese linguists used to believe from the eighteenth century onwards that the Semitic structure of their language was due to the fact that it is a remnant of Phoenician or Punic, in fact Maltese has its origins in colloquial Arabic. The Arabic language was brought to the Maltese islands (Malta, Gozo and Comino), situated in the middle of the Mediterranean Sea, by Moslem conquerors in the ninth century. Since the Arabs were driven out from the islands by Count Roger the Norman in 1090 and Malta became part of the Kingdom of Sicily (1090–1194), a large number of European loan words, mostly Romance, were gradually introduced into Maltese vocabulary. Being the language of an entirely Christian people since the thirteenth century, when the German Emperor Frederick II, King of Sicily, expelled the Moslems (Kinder & Hilgemann 1978: 173; Laspina 1971: 43; Prevaes 1993: 96), Maltese was highly influenced for many centuries by European languages spoken in the Maltese islands, mainly Italian and English, but Spanish and French too (Farrugia 1998: 287). Modern standard Maltese reflects these clear foreign influences in its phonology, morphology, syntax and vocabulary. Nevertheless, from a linguistic point of view Maltese is an isolated Arabic dialect, which gained an official status. The first text written in Maltese is Pietro
Rami Saari
Caxaro’s Cantilena, a poem presumably composed in the middle of the fifteenth century and discovered only in 1966. The first Maltese literary novel, Ine˙z Farru˙g, was written by Anton Manwel Caruana and published in 1889.
The goal of the paper The goal of this paper is to present several grammatical aspects of Maltese prepositions and to expose some of the main problems to be dealt with in research into the complicated definition of the Maltese parts of speech in general and the Maltese prepositions in particular. Though this paper concentrates on only one preposition, quddiem, the examples which are used in it help to refer to some of the essential questions about Maltese prepositions in a more general way. Since Maltese lacks a sizeable corpus of written texts attesting the different stages of its development (Borg 1978: 1), the examples brought in the present paper are all taken from Maltese literary works and dictionaries written and published between the end of the 1880’s and the first half of the 1990’s. Therefore they follow the standard Maltese orthography. Some of the letters of this Latin script which have diacritic marks or deviate notably from the rules ˙ of pronunciation used in Italian are: c˙ [ˇc], g˙ [ˇg], A[h], gA[originally ä or g, ˙ nowadays obsolete, reflected in the lengthening of the previous vowel or pronounced h when in the end of the word], x [š], z˙ [z], and z [c, dz]. For more ˙ details: Cohen (1970).
The etymology of quddiem The preposition the paper deals with, quddiem, has its origin in Arabic wordstock and it means ‘in front of ’. It is worthwhile to mention that its antonym wara means both ‘behind’ and ‘after’. However, quddiem marks only the local meaning, ‘in front of ’, whereas another Maltese preposition, qabel, Arabic qabla, contains the temporal meaning of ‘before’. Quddiem is both phonologically and etymologically the Maltese version of qudd¯ama, a word which exists in Classical Arabic and appears in many colloquial forms of contemporary Arabic dialects as qudd¯am after the apocope of the case endings of Classical Arabic and the disappearance of the short final vowels in general (Procházka 1993: 190). Qudd¯ama may be traced back even to pre-Islamic Arabic poetry, where it appears for instance in the line q¯aãidan qudd¯ama hayyin “a chief in ˙ front of a tribe” (Tarafa bin al-äabd as quoted in Arazi and Masalha 1999: 892.) ˙
Quddiem and some remarks on grammatical aspects of Maltese prepositions
Both Classical Arabic and Modern Literary Arabic sometimes use the preposition ’am¯ama as its variant, but this Arabic word cannot be found in Maltese at all. The origin of qudd¯ama, like the origin of a whole subgroup of Arabic prepositions (Wright 1991: 280–282; Voigt 1999: 24–25, 30), is a triliteral noun used in the accusative singular form, determined by the following genitive. This form is the Arabic expression of the adverbial origin of prepositions in general and it explains why prepositions function as heads of adverbial phrases. Classical Arabic presents also several nouns in the adverbial -u case (qudd¯amu, ’am¯amu, ’awwalu, haytu etc. Wright 1991: ii 288). These words ˙ are not followed by the genitive and they function like adverbs. In the case of qudd¯ama the noun in question is qudd¯amun “fore part, front part” (Aquilina 1990: 1164). The phonological peculiarities of the Maltese form derive firstly from the disappearance of the final short vowel a, secondly from the so called ‘im¯ala’ sound shift, in which a long, stressed a¯ , in this case in the second syllable of the word, became a long stressed ¯ı, through the diphtongisation ie which occurred in all Maltese dialects, and thirdly from the pronunciation of the emphatic uvular q in the beginning of the word as a voiceless glottal stop or ‘hamza’, the way it is usually pronounced in the urban Arabic dialects spoken on the Mediterranean coast.
The declension of Maltese prepositions As a preposition quddiem may come before a noun: (1) Quddiem Omm Sidna in-front Mother Lord-1.pl.poss.suff ‘In front of the Mother of our Lord.’
(Caruana 1978: 74)
However the preposition may precede a possessive suffix, which is to be considered representative of the personal pronoun: (2) Ried jibla’ want-3.m.sg.perf swallow-3.m.sg.imperf def.artl-platt ta’ quddiemu plate of in-front-3.m.sg.poss.suff ‘He wanted to swallow the plate in front of him.’ (Sammut 1991: 28)
Besides quddiem there are many other Maltese prepositions which may behave in the same way, i.e. they may form prepositional phrases and they may receive possessive suffixes as well, for example daqs ‘like, as much as’, dwar ‘about, con-
Rami Saari
cerning’, flok ‘instead of ’ and kontra ‘against’. Some Maltese grammars do not mention Maltese prepositions of European origin as declining ones or do not mention them as Maltese prepositions at all (Brother Henry 1980: 192; Chetcuti 1986: 35). The declension of Maltese prepositions of European origin with Arabic possessive suffixes shows that these foreign elements have been fully adapted to the Semitic morphology of Maltese. This is the case regarding flok (originally a prepositional phrase which went through a semantic extension; nowadays the word functions mainly as a compound preposition made of Arabic f¯ı ‘in’ and Italian luogo ‘place’, which has its origin in Latin locus) and the Italian loan words contra ‘against’, Maltese kontra, and favore ‘favor’, Maltese favur ‘in favor of ’ (Vanhove 1993: 350).
The adapting of non-Semitic elements into Semitic patterns Non-Semitic nouns in Maltese may have the broken plural forms (Aquilina 1973: 310–311) and these internal plurals have exerted considerable structural pressure on the subsequent historical development of Romance Maltese morphology (Borg 1978: 279), for instance serp ‘serpent’ pl. sriep, storja ‘story, history’, pl. stejjer, ˙gurnata ‘day’ pl. ˙granet and pjazza ‘square’, pl. pjazez are based on analogy with the inflectional patterns of Maltese Semitic nouns such as kelb ‘dog’, pl. klieb, ftira ‘flat and round bakery product’, pl. ftajjar, ˙gurdien ‘mouse’, pl. ˙grieden and ˙garra ‘jar’ pl. ˙garar. Much in the same way, we may find in Maltese, besides declining Semitic prepositions such as daqsna ‘like us, as much as we (do)’, fuqkom ‘on you (pl.)’ and dwarhom ‘about them’ also the suffixation of the possessive pronouns to non-Semitic prepositions, for instance flokna ‘instead of us’, favurikom ‘in favor of you (pl.)’ and kontrihom ‘against them’. In my opinion, the origin of this phenomenon is the great Maltese flexibility in adapting non-Semitic elements into Semitic word structure.
The restrictions of the declension Unlike the above-mentioned examples and unlike the vast majority of Maltese prepositions, there are also several Maltese prepositions that cannot decline with possessive endings. Such prepositions are the Semitic bla ‘without’, ˙go ∼ ˙gew ∼ ˙gewwa ‘inside’ and sa ‘until’ as well as the Italian permezz ‘by, by means of ’ and rigward ‘concerning, with regard to’. In the case of Semitic
Quddiem and some remarks on grammatical aspects of Maltese prepositions
prepositions, the reason for this behavior is mainly morphological, whereas the above mentioned Italian prepositions do not decline with possessive suffixes mainly because of semantic reasons, just like some Semitic prepositions which were originally nouns and still function solely as nouns whenever the possessive suffix is attached to them, for example tul ‘along, during; length, height’, but tulna ‘our length, height’ and waqt ‘at, during; time’, but waqthom ‘their time’. I consider the preposition in its purest shape as a non-lexical element, whose grammatical function is most evident in those cases which have no other meaning or usage. Naturally, these words have minimal precision and extremely wide semantic fields. Therefore, their contamination in Maltese is quite a widespread phenomenon. Thus we find in this language many contaminated forms such as minnflokok (lit. ‘from in your place’) ‘instead of you’, ˙go fina (lit. ‘inside in us’) ‘inside us’ and sa minn meta konna (lit. ‘till from when we were’) ‘ever since we were’. Sometimes a compound form of adverb and preposition is semantically equal to a simple preposition, e.g. barra minn, Alief and gAajr, all meaning ‘except’. Besides contaminated forms the usage of synonyms may also solve the problem of prepositions which cannot decline with possessive endings. For example the prepositional phrase bla t-tfal ‘without the children’ can be expressed by the phrase mingAajr it-tfal which includes the compound preposition mingAajr ‘without’ (< minn ‘from’ + gAajr ‘except’) and is semantically identical. However, since bla cannot decline with possessive endings, the only way to say ‘without them’ in Maltese is mingAajrhom.
The usage of prepositional phrases in the creation of new word stock Maltese prepositional phrases were and are still used rather widely in the creation of new adverbs, nouns and adjectives. The presentation of the adverbs in this paper precedes the presentation of the nouns and the adjectives, because, in my opinion, the main task of the prepositional phrase in Maltese is to express the extension or the complementation focused on the adverb. Therefore, the linkage between Maltese prepositional phrases and adverbs is a very strong one. As a matter of fact, besides a few Maltese adverbs such as hekk ‘this way’, dritt (dritt) ‘directly’, sew(wa) ‘correctly, honestly; very much’ and qajl(a) ‘slowly’, most of the Maltese adverbs are compound prepositional phrases, for instance bilfors (lit. ‘with the power’) ‘necessarily’ flimkien (lit. ‘in the place’) ‘together’, malajr (lit. ‘with the air’) ‘soon, quickly’, mill-˙gdid (lit. ‘from the new’) ‘again’
Rami Saari
and tassew (lit. ‘of the truth’) ‘really’. Non-compound prepositional phrases may function just as well like adverbs: (3) Kemm hu liebes ta˙c-˙cajt! How he dress-3.m.sg.perf of-def.art-joke! ‘He’s dressed so funnily!’
(Calleja 1995: 4)
(4) WieAed minnhom, kontra qalbu, One from-3.pl.poss.suff, against heart-3.sg.poss.suff, ˙ daAal gewwa enter-3.m.sg.perf inside ‘One of them went inside against his own will.’ (Friggieri 1991: 35) (5) U baqgAu jgAixu dejjem gAal qalbhom and stay-3.pl.perf live-3.pl.imperf always for heart-3.pl.poss.suff ‘And they lived for ever and ever happily.’ (Magri 1994: 86)
In this kind of prepositional phrases the complement of the preposition is not necessarily a noun, but possibly another preposition: (6) GAal min irid ikun for who want-3.m.sg.imperf be-3.m.sg.imperf jafha ta’ fuq fuq know-3.m.sg.imperf-3.f.acc.suff of up up ‘For whom wants to know it superficially.’ (7) Kollox inqaleb ta’ taAt fuq everything turn-3.m.sg.perf of down up ‘Everything turned upside down.’
(Magri 1994: xxv)
˙ (Zahra 1994: 43)
The usage of quddiem in adverbs The preposition quddiem may be used alone as an adverb meaning ‘forward’: (8) Pawlu qabe˙z quddiem ji˙gri Paul jump-3.m.sg.perf forward run-3.m.sg.imperf ‘Paul jumped forward beginning to run.’ (Caruana 1978: 8)
Some other Maltese prepositions may be used as well just like regular adverbs, ˙ for example gewwa ‘inside’, qabel ‘before’ and wara ‘after or afterwards’. Unlike these, there are some Maltese prepositions like bAal ‘like, as’, fost ‘among’ and skond ‘according to’, which require a noun and cannot be used alone as adverbs. Besides very few exceptions which are due to ellipsis, e.g.
Quddiem and some remarks on grammatical aspects of Maltese prepositions
(9) Barra l-kotba gAandi moAAi wkoll out def.art.-book-pl at-1.sg.poss.suff brain-1.sg.poss.suff also ‘Besides books I have got my brains too,’ (Friggieri 1991: 90)
where ‘barra l-kotba’ is used instead of ‘barra mill-kotba’, adverbs cannot function like prepositions, i.e. they cannot form phrases with nouns. But quddiem can also form a part of a prepositional phrase, which may function as an adverb. It is to be considered the nominal complement of the preposition bi ‘with, in’ in the prepositional phrase bil-quddiem ‘in advance, beforehand’. In the sentence Irrid il-flus bil-quddiem ‘I want the money in advance’, the Semitic prepositional phrase bil-quddiem may perfectly well be replaced by the European adverb anti˙cipatament. Other examples of quddiem used as a nominal complement of a preposition in a prepositional phrase are (10) QiegAed fir-rewwixta fuq quddiem stand in-def.art-revolt on front-side ‘He led the revolt.’ (11) Dan isservik lezzjoni this-m serve-3.m.sg.perf-2.acc.suff lesson for-def.artgAall-quddiem front-side ‘This will teach you a lesson for the future.’ (Aquilina 1990: 1163)
In this last sentence the prepositional phrase is to be considered as an adverb. But it may be analyzed just as well as a preposition accompanied by a definite noun, because, due to syncopation of the case markers, Maltese cannot morphologically differentiate the Arabic preposition qudd¯ama from the Arabic noun qudd¯amun.
The usage of quddiem in other syntactic categories The common Maltese substitution of expressions, which have different origins and similar semantic fields, has not yet been investigated thoroughly. However, it exists not only in adverbs, but also in nouns (for instance kelmtejn qabel, literally “two words before”, means ‘introduction’, though introduzzjoni may be used just as well). It exists in adjectives, where sempli˙cjott ‘simple-minded’ may be expressed by nieqes ftit minn moAAu, literally “he’s lacking something of his brains”, and it exists also in verbs, so that one may use Semitic tama or European spera in order to say ‘he hoped’.
Rami Saari
The relationship between prepositions and nouns in Maltese Quddiem alone functions not only as a preposition and as an adverb, but also in its original usage as a noun which means ‘the front side’ or ‘the front part’. This noun may appear as the nomen rectum in a construct state or as a definite noun accompanied by ta’, the Maltese preposition which indicates possession, for instance (12) Quddiem il-libsa (∼ il-quddiem tal-libsa) front-side def.art-dress (∼ def.art-front-side of-def.art-dress) kien maAmu˙g be-3.m.sg.perf dirt-pass.part ‘The front part of the dress was dirty’.
This noun has its own plural form, quddimini, which means ‘the two front parts of a bodice’. However, in a sentence such as (13) Il-quddiem Alla biss jafu def.art-front-side God only know-3.m.sg.imperf-3.m.acc.suff ‘Only God knows the future,’ (Aquilina 1990: 1164)
the noun indicates a different meaning. Il-quddiem meaning ‘the future’ may appear only as a definite noun, because both the definite article and the possessive suffixes in Maltese constrain the semantic and the syntactic role of nouns which may function as prepositions. In the case of an indefinite form of the word ‘future’, a Maltese speaker would use the non-Semitic futur, which may be definite as well by the attachment of the definite article: il-futur, ‘the future’. Besides quddiem there are many other Maltese prepositions, both Semitic and non-Semitic, which are originally nouns and most of them still function as such, for instance favur ‘favor’ or ‘in favor of ’, qalb ‘heart’ or ‘in the middle of ’, tul ‘length’ or ‘along’ and waqt ‘time’ or ‘at, during’. Unlike these noun-prepositions we find in Maltese very few words which are nothing but prepositions, for instance the compound bla ‘without’ and the “basic” or “primary” prepositions fi ‘in’, minn ‘from’ and sa ‘till, until’. According to Voigt (1999: 36) even this type of preposition emerged from nouns in a period previous to the separation of the Semitic languages. Most Maltese prepositions still show a close relationship with the noun category, and therefore the Maltese prepositional phrase shall be considered as basically the same kind of morpho-syntactic pattern as the Maltese construct state.
Quddiem and some remarks on grammatical aspects of Maltese prepositions
The nominalization of prepositional phrases Maltese extensively nominalizes prepositional phrases: (14) It-tabib daAal gAand ta’ quddiemna. def.art-doctor enter-3.m.sg.perf at of in-front-1.pl.poss.suff. ‘The doctor entered the house of the people living in front of us.’ (15) Kull filgAodu malli every in-def.art-morning with-def.art-rel.pron tqum get-up-2.m.sg.imperf ‘Every morning when you get up’
is the title of a popular program on Maltese radio. The prepositional phrase mentioned in this title, filgAodu ‘in the morning’, substitutes here for the noun ‘morning’, as can be clearly seen also in the following examples: (16) U jorqdu sa fil-gAodu and sleep-3.pl.imperf till in-def.art-morning ‘And they will sleep till the morning.’
(Friggieri 1991: 31)
(17) Minn fil-gAodu sa filgAaxija from in-def.art-morning till in-def.art-evening ‘From the morning till the evening.’ (Friggieri 1991: 144)
The same happens with some other temporal expressions, for instance: (18) GAall-Aabta tas-sitta ta’ fil-gAaxija for-def.art-hit of-def.art-six of in-def.art-evening ‘At about six o’clock in the evening.’ (Friggieri 1991: 69) (19) Sa madwar il-Adax ta’ bil-lejl till about def.art-eleven of with-def.art-night ‘Till about eleven o’clock at night’ (id.).
The lack of consistency in the usage of the hyphen in Maltese orthography of such words as filgAaxija (∼ fil-gAaxija) ‘evening’ and filgAodu (∼ fil-gAodu) ‘morning’ may suggest that these words are not perceived by Maltese authors either as pure nouns or as pure prepositional phrases, though their syntactic usage and semantic functions in this language show clearly that they are both prepositional phrases and nouns. Moreover, nominalized prepositional phrases may function in Maltese as the subjects of the sentences in which they appear, though several special morpho-syntactic rules concerning definiteness and plural formation must be applied to them:
Rami Saari
(20) Tal-Aaxix waqaf quddiem id-dar of-def.art-vegetable stand-3.m.sg.perf in-front def.art-house ‘The greengrocer stood in front of the house.’ (Born 1991: 34) (21) Mhux ta’ b’xejn igAiru gAalik neg-3.m.sg-neg of with-nothing envy-3.pl.imperf for-2.poss.suff tar-raAal l-ieAor of-def.art-village def.art-other ‘It’s not for nothing the people of the other village envy you.’ (Friggieri 1991: 56) (22) Ta’ quddiem qab˙zu of front-side jump-3.pl.perf ‘The ones who were in the front side jumped.’
˙ (Zammit 1987: 24)
The subject of the first sentence, Tal-Aaxix “the greengrocer”, is, as a matter of fact, the nominalized prepositional phrase ‘of the vegetable’. Maltese forms many nouns related to professions according to the same pattern, for example tal-Aalib (‘of the milk’) ‘the dairyman’, tal-˙gelati (‘of the ice-cream’) ‘the icecream seller’, tal-laAam (‘of the meat’) ‘the butcher’, tal-Aob˙z (‘of the bread’) ‘the baker’, tal-Aanut (‘of the shop’) ‘the seller’ and even tal-grocer (‘of the grocer’) meaning actually ‘the grocer’. Procházka (1993) does not indeed deal with ta’, but this word, which can be declined, precedes nouns and expressing the possession in Maltese it substitutes for the Classical status constructus. Therefore, in my opinion, it should be considered a preposition just like English of and Italian di, in spite of some differences in usage in comparison with other prepositions.
Prepositions and prepositional phrases which form Maltese adjectives The adjective quddiemi (f. quddiemija, pl. quddiemin) means ‘situated in the front or in a front position’ and it is derived from the first noun in question by the suffixation of the Maltese adjectival ending -i (originally Semitic -¯ıy; Pennacchietti 1974: 30). However, there are some Maltese prepositions which take instead the ending -an or a suffixation of two semantically identical but morphologically different endings: -an + -i > -ani, for example fuqan(i) ∼ fuqâni ∼ fuqqani ∼ faqqani ∼ fawqani, ‘high, elevated’ (Aquilina 1987: 369) and fostan(i) ∼ fustani ∼ wistani ‘occupying a middle position’ (Aquilina 1987: 358). Also adverbs may take the same endings in order to form adjectives or they may be used as complements in a prepositional phrase. “The outer sea”, for instance, appears both as il-baAar il-barrani and as il-baAar ta’ barra (Magri
Quddiem and some remarks on grammatical aspects of Maltese prepositions
1994: 6). There are even more Maltese prepositions, both Semitic and nonSemitic, which cannot form an adjective in either of these ways, for example bejn ‘between’, ma’ ‘with’ and skond ∼ skont ‘according to’. Many Maltese adjectives are expressed by prepositional phrases, for instance the sentence M’humiex ta’ quddiem in-nies “They are not presentable” contains the adjective ta’ quddiem in-nies ‘presentable’. This Maltese adjective is actually both a relative and a prepositional clause. An adjective of European origin, pre˙zentabbli, may be substituted for it. Other examples of Maltese adjectives formed by prepositional phrases are presented in the following sentences: (23) Dak li kien l-aktar ’il quddiem that-m rel.pron be-3.m.sg.perf def.art-more towards front-side fosthom among-3.pl.poss.suff ‘The one which was the most advanced among them.’ (Born 1991: 27) (24) Saru i˙zjed bla Aniena u ibsin become-3.pl.perf more without mercy and harsh-pl ‘They became more merciless and harsh.’ (Caruana 1978: 72) (25) “GAidli, fittex “tell-2.sg.imp-1.sg.indir.obj.suff, try-2.sg.imp gAidli,” tell-2.sg.imp-1.sg.indir.obj.suff,” ordnalu l-isqof, nervu˙z order-3.m.sg.perf-3.m.sg.indir.obj.suff def.art-bishop, nervous u fuq ix-xwiek. and on def.art-thorn-pl “Tell me, try to tell me,” the bishop ordered him, nervous and tense.” (Friggieri 1991: 21) (26) Dik art mill-aAjar gAar-raba’ that-f land from-def.art-best for-def.art-field ‘That is an excellent soil for cultivating.’ (Friggieri 1991: 122) (27) Dak ta’ snienu tad-deheb that-m of tooth-pl-3.m.sg.poss.suff of-def.art-gold ‘That one with the golden teeth.’ (Magri 1994: 58) ˙ ta’ temp = ragel ˙ xiA, avvanzat (28) Ragel fi˙z-˙zmien man of time = man old, advance-pass.part in-def.art-time “ ‘A man of time’ = an old man, ‘advanced in time’.” (Magri 1994: 89)
Rami Saari
In these examples, the prepositional phrase functioning as an adjective doesn’t require number concord and agreement by gender form. The attributive clause is therefore marked by syntactic means, i.e. the usage of the preposition ta’ “of ” or the adjunction of the noun to the prepositional phrase. However, morphological agreement is essential in Maltese in nominalized, or rather ‘adjectivized’, prepositional phrases, which contain possessive suffixes relating to the preceding noun whose qualities are described; for instance, ‘a robust, healthy man’ is ra˙gel b’saAAtu (Magri 1994: 25), actually ‘a man with his health’, whereas ‘a robust, healthy woman’ is mara b’saAAitha, actually ‘a woman with her health’; tifel rasu fuq il-kotba (lit. ‘a boy his head on the books’) is ‘a diligent boy’, whereas ‘a diligent girl’ is tifla rasha fuq il-kotba. The usage of this kind of phrase is based on a pattern of asyndetic sentences which include a predicative phrase with no prepositions at all such as ra˙gel qalbu tajba (lit. ‘a man his heart is good’) ‘a good-hearted man’ and mara qalbha tajba (lit. ‘a woman her heart is good’) ‘a good-hearted woman’.
Some examples of prepositional phrases used in Maltese phraseology Besides all the above-mentioned functions, prepositions may also determine the semantics of Maltese verbs, because there are in this language quite a few verbs the meaning of which is changed by the prepositions they govern and not only by the effect of their different collocations in sentences (Aquilina 1976: 67). Prepositional phrases play an important role also in the rich Maltese phraseology. This can be demonstrated clearly by the following examples: (29) Dejjem minn ta’ quddiem biex igAin lil min always from of front-side in-order-to help-3.m.sg.imperf to who he hu fil-b˙zonn in-def.art-need ‘He is always too ready to help those in need.’ (Aquilina 1990: 1163) ˙ (30) “B’wi˙cc˙ u minn quddiem” jigifieri bla mistAija “with-face-3.m.sg.poss.suff from front-side” i.e. without shame ‘ “With his face in the front” means shamelessly.’ (Fenech 1984: 289) (31) ’Il quddiem inkunu nistgAu towards front-side be-1.pl.imperf can-1.pl.imperf nagAmlu l-affarijiet aktar legali make-1.pl.imperf def.art-matter-pl more legal ‘From now on we shall be able to make things more legal.’ ˙ (Zahra 1994: 47–48)
Quddiem and some remarks on grammatical aspects of Maltese prepositions
Conclusion The investigation of constraints, which give rise to the differences in the syntactic behavior of Maltese prepositions and prepositional phrases, is still a topic which needs a lot of research. These prepositions and prepositional phrases reflect both the basic Arabic structure of Maltese and the considerable nonArabic component which was widely adapted into the Semitic patterns of this language. Thus, this paper clarifies some of the prepositional mechanisms of Maltese, taking into account the fast rhythm of dynamic changes occurring in this language, which has been exposed to strong and prolonged European influences.
Notes . This paper focuses mainly on morphology and syntax. Therefore it will not deal with the sociolinguistic status of Maltese within Arabic.
References Aquilina, J. (1973). The Structure of Maltese. A Study in Mixed Grammar and Vocabulary. Malta: The Royal University of Malta. ——— (1976). Maltese Linguistic Surveys. Malta: The University of Malta. ——— (1987–1990). Maltese-English Dictionary, 2 Vols. Malta: Midsea Books. Arazi, A., & Masalha, S. (1999). Al-äaqd al-tham¯ın. Six Early Arab Poets: New Edition and Concordance. Based on W. Ahlwardt’s The Divans of the Six Ancient Arabic Poets. Jerusalem: The Hebrew University of Jerusalem, Institute of Asian and African Studies, The Max Schloessinger Memorial Series. Borg, A. (1978). A Historical and Comparative Phonology and Morphology of Maltese. Thesis Presented to the Senate of the Hebrew University. Jerusalem. Born, W. (1991). Is-salib tal-fidda. Blata l-Bajda: Merlin Library. ˙ zwiti Maltin. Brother Henry F. S. C. (1980). Grammatika Maltija. Malta: Gi˙ ˙ Calleja, O. (1995). 4 drammi: Satira, Anestesija, I˙gsma iAirsa, Cens perpetwu. Malta: Gutenberg Press. Caruana, A. M. (1978). Ine˙z Farru˙g. Rumanz storiku. Edizzjoni annotata mill-professur ˙ Aquilina. Malta: A.C. Aquilina & Co. Union Press. G. ˙ (1986). TgAallem il-Malti f ’20 lezzjoni: TaAri˙g prattiku ta’ ortografija u ta’ Chetcuti, G. grammatika fil-livell tal-matrikola u tal-matrikola avanzata. Malta: Merlin Library. Cohen, D. (1970). Le système phonologique du maltais, Aspects synchroniques et diachroniques. In D. Cohen (Ed.), Études de linguistique sémitique et arabe, [Janua Linguarum, Series Practica 81], (pp. 127–150). Paris: Mouton.
Rami Saari ˙ (1998). Il-Malti Lsien ImAallat. In Karl Borg (Ed.), Lingwa u Lingwistika Farrugia, G. (pp. 277–291). Valletta: Klabb Kotba Maltin. Fenech, K. (1984). Idjomi Maltin. Malta: Interprint. Friggieri, O. (1991). Fil-g z˙ ira Taparsi jikbru l-fjuri. Malta: Grima. Goldenberg, G. The Contribution of Semitic Languages to Linguistic Thinking. In G. Goldenberg (Ed.), Studies in Semitic Linguistics: Selected Writings by Gideon Goldenberg (pp. 1–9). Jerusalem: The Magnes Press, The Hebrew University. ˙ Laspina, S. (1971). Grajjet il-g z˙ ejjer Maltin. Malta: A.C. Aquilina & Co. Kinder, H., & Hilgemann, W. (1978). The Penguin Atlas of World History, Vol. 1. London: Penguin Books. ˙ ˙ Mifsud Chircop, San Gwann: ˙ Magri, M. (1994). Grejjef missirijietna. Edizzjoni kritika ta’ G. Publishers Enterprises Group. Pennacchietti, F. A. (1974). La classe degli aggettivi denotativi nelle lingue semitiche e nelle lingue berbere. In A. Caquot & D. Cohen (Eds.), Actes du premier congrès international de linguistique sémitique et chamito-sémitique (pp. 30–39). The Hague: Mouton. Prevaes, M. H. (1994). The emergence of standard Maltese: the Arabic factor. Thesis. The Netherlands: The Catholic University of Nijmegen. Procházka, S. (1993). Die Präpositionen in den neuarabischen Dialekten. Dissertationen der Universität Wien. Thesis. Wien: Verband der wissenschaftlichen Gesellschaften Österreichs. Saari, R. The Maltese Prepositions. Thesis in preparation. The Hebrew University. Jerusalem. Sammut, F. (1994). Paceville. Blata l-Bajda: Merlin Library. Vanhove, M. (1993). La langue maltaise: études syntaxiques d’un dialecte arabe “périphérique”. Wiesbaden: Harrassowitz Verlag. Voigt, R. (1999). Die Präpositionen im Semitischen – Über Morphologisierungsprozesse im Semitischen. In Tradition and Innovation, Norm and Deviation in Arabic and Semitic Linguistics, L. Edzard and M. Nekroumi (Eds.), 22–43. Wiesbaden: Harrassowitz Verlag. Wright, W. (1991). A Grammar of the Arabic Language. Cambridge: Cambridge University Press. ˙ ˙ Zammit, T. (1987). Nies bla sabar u stejjer oAra. L-istejjer kollha migburin minn T. Cortis bi studju ta’ O. Friggieri. Blata l-Bajda: Merlin Library. ˙ Zahra, T. (1994). Is-seba’ tron˙giet mewwija. Blata l-Bajda: Merlin Library.
Locative prepositions in language acquisition and aphasia Mark Leikin Faculty of Education, University of Haifa
Prepositions have attracted a great deal of attention in psycholinguistics and neurolinguistics. This has to do with the ambiguous status of prepositions in language. They perform auxiliary syntactic functions, and in this respect they are like other functors. In terms of syntactic theory, however, the class of prepositions is treated as a ‘lexical category,’ especially when they determine the character of spatial and temporal relationships (e.g., Grodzinsky 1990). In this case, prepositions may be treated like content words that carry a definite semantic load. Such an approach has been used in studies of acquisition of Russian locative prepositions in childhood (Leikin 1998) and their handling by adult aphasic patients (Leikin 1996). Language acquisition and aphasic language breakdown represent two types of notable deviation from ‘normal’ language behavior. Prepositions present a considerable challenge to children in the course of language acquisition and cause great difficulties for aphasic patients. It is usual to hold that prepositions tend to be omitted in Broca’s aphasia, while in Wernicke’s aphasia they are mostly substituted (Caplan 1987; Grodzinsky 1990). However, prepositional substitutions have been shown to appear in the case of agrammatic (Broca’s) aphasia too (Friederici 1985; Miceli, Silveri, Pomani, and Caramazza 1989). Moreover, in certain contexts and in certain tasks (e.g., structured tests) this error pattern is even more evident (Heeschen and Kolk 1988; Leikin 1996). To date, the prepositional substitutions have not been studied like substitutions (paraphasias) of content words (e.g., semantic and phonetic paraphasias). As regards children, prepositions appear in the child’s speech later than all major types of content words. As a rule in the period 1; 10–2; 4 (years; months) the prepositions are omitted in obligatory contexts (Brown 1973) including, in
Mark Leikin
Russian (Gvozdev 1961), constructions with governed oblique cases (e.g., . . . stole . . . , ‘(on) the table’). At age 2; 2–3; 5, prepositions consequently appear in children’s speech in a fairly invariant order in different languages: in, on, behind, under, above, in front of, and so on (Johnston 1988). First and most frequently, prepositions express locative relations but thereafter the number of prepositions in the child’s language rapidly increases and the variety of semantic and syntactic functions expressed by them expand. The acquisition order indicates that the cognitive and semantic complexity level of locatives determines it (Johnston 1988; Leikin 1998). Initial prepositional meanings are concrete, have deictic features, and they are similar in this respect to adverbs, including the possibility of using them in an adverbial function (Leikin 1998). In Russian locative adverbs to some extent duplicate the spatial prepositions (e.g., the preposition v, ‘in’ and the adverb vnutry, ‘in/within’). However, adverbs have greater figurativeness (deictic features) than prepositions, as well as high specificity and simplicity. Children prefer adverbs at this stage, which suggests that they first acquire more physically concrete meanings and then more abstract ones. This fact is in line with the well-known hypothesis (Piaget 1963; Vygotsky 1962) that cognitive development precedes language development. That is, children first acquire spatial relations cognitively, and only later are these relations verbalized. In turn, the influence of cognitive development on language development seems to be reversed (Bowerman 1989; Vygotsky 1962). Further development of prepositional meanings is affected by the development of cognitive skills, and these in turn exert an inverse influence on the cognitive development of locatives. In this regard, note that omissions and substitutions of prepositions are found in the speech of older children too, when (approximately by age 5) most of the prepositions have been acquired (Cromer 1991; Leikin 1998). Hence one more stage is likely, namely the acquisition of a concept in the sense of full adult meaning. The acquisition of prepositions seems not to be completed at the moment of their appearance in speech but to continue for a considerable period of childhood (Leikin 1998). It was also hypothesized (Jackendoff 1987; Leikin 1998) that this stage is characterized by the formation of the semantic system of locatives. One of the most important indications of the acquisition of the prepositional system is the ability to treat prepositions functionally, when the preposition immediately executes its semantic function in any context.
Prepositions in language acquisition and aphasia
.
The semantic system of prepositions
It was suggested (e.g., van Schooneveld 1978) that locative meaning is the core meaning of prepositions that have multiple meanings, and that prepositions form a definite system. The systematic character of locatives is connected with the systematic character of spatial relations: the semantic system of locatives somewhat correlates with perceptual representation of the visual world existing in the ‘visual brain’ (Jackendoff 1987; Leikin 1996; Talmy 1983). The structures of spatial expressions across languages have been described as a certain system of semantic distinctive features (hereafter called SDF) or mentally represented primitives (cf. Jackendoff 1987; van Schooneveld 1978; Talmy 1983). In this study, a model of the Russian locative prepositional system is used (Leikin 1996). This pattern is fairly similar to van Schooneveld’s (1978) model as well as to Jackendoff ’s (1987). In particular, this concerns the principle of hyponymy and the assumption that a limited number of semantic features constitute a hierarchy of inclusion relationships. These similarities can very likely be traced to the theory of Jakobson (Jakobson 1990). In our model (Leikin 1996) the system of locatives is represented as a hierarchical system of a finite stock of SDFs and the means for generation of the meanings. That is, the system is considered to be in the process of expanding (sort of tree structure). Four object-centered features distinguish the basic type of SDF: (1) Coexistence of two objects in time and space (e.g., the Russian preposition s, ‘with’); (2) Functionality, i.e., the functional asymmetry between objects so that one is the located object (hereafter called the Figure) and the other one is the reference object (hereafter called the Landmark) (e.g., pri, ‘attached to’); (3) Location, specifying location of the figure with additional meaning of stasis over a period of time; (4) Path, specifying motion along a path such as Trajectory of motion. The second type of SDF appears in the hierarchical structure after incorporation into the system of two new measures: the viewer’s position and properties of the landmark (e.g., objects’ intrinsic coordinate axes). These are such SDFs as Vertical Orientation, Horizontal Orientation, Surface Contact, Front, Back, and Container. The meaning of the preposition may be described by a single feature, as in the preposition s, ‘with’ (the feature Co-existence+) or as just any one of realization of a SDF. For example, the meaning of the preposition v, ‘in’ (Enclosure+) is only a particular realization of the SDF Location (Location+). Note that a realization of any feature appears here as a certain initial meaning that is not necessarily designated by a concrete preposition. Finally, a combina-
Mark Leikin
tion of features and their realizations may describe a meaning. For example, the meaning ot, ‘from/away from’ is described by the combination of several SDFs, ‘inclusion of the Figure in the neighborhood of the Landmark’ (Location+, Source +) and ‘motion ex loco’ (Path+, Target–). The model presents the system of locative meanings in the process of expanding. Two means for the derivation of new meanings are suggested. First, the determination of the feature-operator in the field of generic meaning (one of the realizations of a SDF). In this case, the features Functionality, Location, and Path play the role of such an operator. For example, the determination of the feature-operator Functionality in the field of the nearest generic meaning (realization of the SDF Co-existence) leads to the formation of the new meaning, ‘definition of the figure by way of the landmark’ (SDF Functionality+). The second means of derivation is the negation of known (marked) specific meanings within the limits of the nearest generic meaning. For example, the meaning u, ‘by/at’, appears as a result of a negation of the meaning v, ‘in’, within the limits of the meaning ‘definition of the figure by way of the landmark’. The result of this negation (‘non-inclusion of the figure in the volume of the landmark’ – the new marked specific meaning) is reformulated with the addition of a marked specific meaning to the nearest generic meaning as ‘inclusion of the figure in the neighborhood of the landmark’ (u, ‘by/at’). As a result of these derivational processes it is possible to pattern the system of locative meanings (Fig. 1) (Leikin 1996). This pattern is presented in terms of lexical units, and represents the suggested hierarchy of meanings and trends of derivation of the system. This model was applied for the analysis of prepositional substitutions obtained from adult aphasic patients and normal preschool children (Leikin 1996, 1998). away from with out of
attached to to as far as
in across/via
by/at
on
behind
in front of
under
above
Figure 1. Part of the Russian system of locative meanings.
Prepositions in language acquisition and aphasia
. Description of the previous studies . Method Eighty-eight normal children and 46 adult aphasic patients (all Russianspeaking) participated in the study. The children (see Leikin 1998) were selected from five age groups: (1) 17 children aged 3:0–3:9; (2) 15 children aged 4:0–4:11; (3) 17 children aged 5:0–5:11; (4) 20 children aged 6:0–6:6; and (5) 19 children aged 6:7–7:2. The aphasic group (see Leikin 1996) ranged in age from 47 to 66 years (M=55.3 SD=6.2). Educational level ranged from 10th grade through master’s degree. Etiology was vascular in all cases. All subjects were right-handed patients in stable neurological condition without spatial disturbances. Patient’s language abilities were assessed by the Western Aphasia Battery (Kertesz 1982) and Luria’s Neuropsychological Assessment (Luria 1969). These subjects represented mild to moderate degrees of aphasia. Of these, 16 were considered to be Broca’s Aphasia, 16 Wernicke’s Aphasia, and 14 Transcortical Sensory Aphasia. Several tasks were administered to participants, whereby each subject had to manipulate two objects (a small sphere and a large cube open on one side) in non-verbal and verbal situations. In the Non-verbal imitation task, a participant was required to establish spatial relationships between the sphere and the cube in real space according to the pattern demonstrated by examiner. The subjects were asked to place the objects according to the pattern. In the Verbal tasks the subjects carried out three types of linguistic operations with locative prepositions: (1) oral repetition of prepositional phrases such as ‘the sphere is on the cube’; (2) comprehension of such phrases; and (3) naming of a demonstrated patterns of spatial relationships. In the comprehension tasks, the phrases were read aloud by the examiner and the subject was required to place the objects in the appropriate relationships. In the Naming task, the examiner demonstrated each pattern of the relationship and subjects were asked to state where the sphere was placed. Six types of locative relationships and six prepositions (the Russian equivalents of the prepositions in, on, under, above, in front of, and behind) were studied in constructions such as ‘the sphere is on the cube’. . Brief description of the results Children (Leikin 1998). None of the children had difficulties in the Non-verbal tasks and Repetition. By contrast, verbal tasks were found to be relatively dif-
Mark Leikin
ficult for the children. Errors in the usage of prepositions occurred in all age groups, but performance steady improved with older age. Analysis of variance with three factors (age, task, and prepositions) revealed strong age effect, that is, almost each following age group was significantly better than the previous one. The effect of the task was also significant in all age groups in favor of the Comprehension task. Also, the variance in the results correlated with specific prepositions. The most difficult one was above, the prepositions in front of and under were somewhat easier, while the prepositions on, behind, and in were easiest. This pattern was most distinct in the oldest age groups and largely conformed to the acquisition order. In the Comprehension task two errors patterns were noted. The first was non-response (5% of all errors). The second was miscomprehension (substitution) of prepositional meaning (95%). In this case, children placed the objects in inappropriate relationships, for example, in front of instead of behind. The substitutions were mostly among the six studied meanings (96%). In the Naming task the children demonstrated the tendency to state first where the sphere was (the pragmatic end-result). However, in most cases the children specified spatial relation and landmark. Different language means for expression of relationships were shown: (1) deictic terms (e.g., tut, ‘there’); (2) different words and grammatical forms that specified the relationship (e.g., doma, ‘at home’); (3) familiar prepositional constructions (pod kryshaj, ‘under the roof ’); (4) locative adverbs (e.g., vnutri, ‘in/within’); (5) locative prepositions – 70% of all responses (e.g., na, ‘on’). However, only those responses in which the spatial relationship was not expressed (4% of errors) or was expressed incorrectly were considered errors. Mostly these were substitutions (96% of errors), for example, above instead of in front of. The children also showed two special strategies in naming spatial relations. The first included the search for an adequate expression for a given relationship, in the course of which several prepositions and/or adverbs appeared. For example, attempting to express the relationship above, the child said: ‘The sphere is on . . . in . . . under, above the cube’. It was interesting that a similar strategy was shown in the Comprehension task, too. The second strategy was apparent in an attempt to intensify the semantic content of the expressions with the help of additional lexical and/or syntactic means (e.g., ‘The sphere is under, the cube is above’). Thus, although as early as age 3:4–4:0 children’s speech is characterized by the almost complete possession of syntax, including prepositions (Johnston 1988; Cromer 1991), the findings showed that the children experienced specific difficulties concerning prepositions. However, they displayed linguistic
Prepositions in language acquisition and aphasia
care for the prepositions. For example, the children were not instructed to use prepositions only, but they used prepositions (correctly or not) in most cases.
Aphasic patients (Leikin 1996). The aphasic patients had no problem with the Non-verbal tasks, whereas in Verbal tasks (including Repetition) they showed significant difficulties. However, the in-groups results differed according to specific features of particular aphasic syndromes. For example, while patients with Wernicke’s Aphasia demonstrated worse results in the Repetition task, patients with Broca’s Aphasia were poor in the Naming task. Between-groups analysis of variance revealed significant differences in all tasks. The one exception was noted in the Naming task, where the results of the Wernicke’s Aphasia and Transcortical Sensory Aphasia groups were not significantly distinct. The effect of the task was substantial, too. The Repetition task was found to be easier than the Comprehension and Naming tasks, whereas production of prepositions (Naming task) was generally inferior to their comprehension. The variance in the results with particular prepositions was noted only in the Comprehension task. The comprehension deficit was more marked in Transcortical Sensory Aphasia. However, in this task, regardless of aphasia type, the patients demonstrated the same error patterns as those shown by the children: non-response (1.7% of all errors) and miscomprehension of prepositional meanings (98.3%). Also like the children, the aphasic adults ‘searched’ for an adequate relationship concerning the given preposition, but they used this strategy more extensively than the children. The second special strategy was repetition, and sometimes a detailed analysis of the presented phrase. For example, for the phrase “The sphere is in front of the cube’, the patient said, ‘The sphere is from of the cube . . . It means that if the sphere is in front of the cube, the cube must be behind the sphere. Then . . . then the sphere is behind the cube”. Three error types were noted in the Naming task and partly in the Repetition task: non-responses (4.8% and 20% of errors, respectively), omissions (1.6% and 3%), and substitutions (95.6% and 77%). Substitutions were of three various types: (1) prepositional substitutions (e.g., under for in front of ); (2) adverbial substitution (e.g., from behind for under); (3) ‘description’ (e.g., ‘The sphere. The cube is underneath’ for above). The aphasic patients, like the children, used two special strategies in Production tasks. The first involved the search for an adequate expression, for example, ‘The sphere is above . . . on the cube’ for the relationship on. These searches were more frequent and lengthy than those made by the children. The second strategy was the attempt to intensify the semantic content of expres-
Mark Leikin
sion (e.g., ‘The sphere stands on the top part of the cube’). The patients used a great number of various words with locative meanings. Even agrammatic patients (Broca’s Aphasia) used different prepositions. Thus, the results showed that the prepositional system was not completely destroyed in aphasic patients. In certain situations, even the patients with Broca’s aphasia were able to handle prepositions and demonstrated linguistic care for them. At the same time, not only similarities but also the distinctions between different aphasic groups were found. This indicates that there may be general and separate mechanisms for the inconsistent handling of prepositions by patients with different aphasic syndromes (Leikin 1996). Taken together, the results displayed significant quantitative and qualitative differences between the children and the aphasic patients. The differences were also shown in comprehension-production ratio. The Comprehension task was easier than the Naming task with the children, whereas with the aphasic patients this relation varied according to the form of aphasia. The same was true for the relations among the ability of subjects to comprehend prepositions, to use them for the expression of spatial relationships, and to use other language means along with the prepositions to this effect. Simultaneously, the same error-types as well as a few similar strategies of performance were shown in both groups of the subjects. . Substitutions Substitution was the major error type in both children and aphasics. The aphasia type as well as children’s age had no significant influence on the character of these substitutions. However, the patterns of substitutions and the verbal behavior related to them were different in aphasic patients and children. With the children, substations existed in parallel to specific types of verbal behavior, for example, extensive use of adverbs. Note, however, that up to 40% of adverbial substitutions in the younger groups and approximately 15% of such substitutions in older groups were also semantic substations (e.g., within instead of on). By contrast, the aphasic patients most commonly substituted one preposition for another. These substitutions also demonstrated a wide repertoire of different prepositions (e.g., with, to, to/up to, between) whereas the children used only a few prepositions other than the six studied (from, at/near, at/by, through/across, with, and opposite). The children comprehended the prepositions better than they produced them. The Naming task, however, proved more successful than the Comprehension task, since the children were able to use any suitable linguistic terms for
Prepositions in language acquisition and aphasia
expression of locative relationships. This correlation was stable in all age groups as well as the order of the prepositions’ complexity. In the aphasic groups this was not as evident. The variance in the results significantly correlated with certain prepositions mainly in the children’s group. At the same time, both in children and aphasic patients some patterns of substitutions occurred significantly more frequently than others (e.g., above instead of on). As a result various prepositions had different ‘strengths’ from the viewpoint of the substitution pattern, that is, some locative meanings appeared as substitutors more frequently, and the opposite. Consequently, results of substitutions may be transformed by an algorithm (Leikin 1996) as follows. We will consider the relation A/B (e.g., the substitution above→on) as belonging to the binary relation if (a) there is a also pair B/A (e.g., the substitution on→above) and the frequency of appearance of element A divided by the number of all considered cases > 0.3; or if (b) there is no pair B/A, but this frequency > 0.6. With the results of this transformation we obtain the following patterns (Figs 2 and 3). In these, the significant trends of substitutions are marked as arrows, so that a continuous line indicates the most frequent occurrence and a dotted line depicts less but valid frequency of occurrence. The arrow shows the direction of the substitutions from the substitute to the preposition that has been replaced. Figures 2 and 3 display only more frequent patterns of substitutions, although all substitutions types were utilized for further linguistic analysis. In addition, the data from different children’s age groups were pooled for this effect since no significant influence of age factor on the character of the substitutions was found. The aphasic patients’ data were assembled for similar considerations. The substitutions revealed were analyzed with the frequency of the prepositions and their prosodic and phonetic properties. Likewise, using the proposed model of the system of the locative meanings (see above), further analysis of the substitutions was performed in terms of semantic content (sets of SDF) of the substitute and the prepositions that have been substituted. Figures 2 and 3 show that the structures of substitutions were different in the children and the aphasic patients. This fact seems to reflect the different patterns of semantic complexity of the prepositions and the different character of the semantic relationships among them in children and in adult aphasic patients. Also, the structures of substitutions varied depending on the task. These differences reflect not only the task-dependent character of the results but seemingly also the differences between comprehension and production of the prepositions.
Mark Leikin Comprehension
in
Naming
behind
in front of
under
above
in
on
under
behind
above
on
in front of
Figure 2. Directions of substitutions of children in Comprehension and Naming tasks (most frequent types of substitutions). Comprehension in
under
behind
on
above
in front of
Naming
behind
in
behind
in front of
above
under
Figure 3. Directions of substitutions of aphasic patients in Comprehension and Naming tasks (most frequent types of substitutions).
The analysis of the substitutions did not reveal any connection between the nature of errors and such features of the prepositions as frequency, prosodic and phonetic complexity, or similarity. Rather, the results reflected the semantic properties of the prepositions and the relationships among them. An analysis of the substitutions from the viewpoint of a proposed pattern of locative meanings showed that the nature of the substitutions to some extent correlated with SDFs and their structure within the system. Note, however, that even in the aphasic group the structures of more frequent substitutions reproduced the proposed model (Fig. 1) only partly. Even so, the linguistic analysis of substitutions makes it possible to classify the substitutions as follows: 1. Substitutions by common feature (the neutralization of one or more SDF): a. Neutralization of the opposition by the feature Contact in the presence of the common feature Vertical Orientation (e.g., on→above,1 beyond→on). b. The substitution of a specific meaning by a more common generic meaning in the presence of the common feature ‘inclusion of the figure in the neighborhood of the landmark’ (neutralization of the features
Prepositions in language acquisition and aphasia
Orientation, and Way of Orientation) (e.g., at→in front of, beside→in front of, by/about→above, toward→in front of). c. Neutralization of the feature Way of Orientation in the presence of the common feature Orientation (e.g., to the left→under, sideways→ behind). 2. Substitutions by contiguity of meanings: a. Substitutions by similarity of meanings (e.g., inside→in). b. Substitutions by contrast of meanings (e.g., outside→in, from→in). The main opposition is by the feature Enclosure (‘. . . is not outside, but in the cube’). If this meaning is disclaimed a new generic meaning ‘inclusion in the neighborhood’ acquired different concrete realizations (e.g., under→above, in front of→behind, under→on). In these cases, there is a change of the position inside the meanings Vertical Orientation or Horizontal orientation. 3. Substitution by nature of orientation (e.g., above→in front of, above→ behind). 4. Reformulating of relationships (e.g., through → in front of, between→ under). 5. Introduction of supplementary orientator (e.g., on the height→above, from the rear→behind). 6. Substitutions of ‘Amnestic’ type – a definition of lost words: a. Definition of relationships by way of negation (e.g., the sphere is not outside→in). b. Double definition (e.g., at the back of above→behind, at the top of above→on). The searches noted in all subjects’ groups included all types of observed substitutions in different combinations. These partly reveal relations between certain spatial meanings and most likely certain stages of internal derivational processes (Leikin 1996).
. Discussion The children were found to demonstrate linguistic care for the prepositions. At the same time, the developmental and linguistic features of the children’s difficulties were recognizable. In production, they demonstrated consecutive
Mark Leikin
(age-dependent) and simultaneous (within age-group) functioning of different means of expression of spatial relationships (e.g., deictic terms). Many of these means are typical of younger children and cannot be considered errors. However, these cases can explain the course of acquisition of linguistic means for locative expressions. For example, the locative adverbs appear in children’s speech before prepositions and reflect such features of children’s language as dependence on situation (Leikin 1998). In Russian, adverbs are also more illustrative and substantive than prepositions, and their functional usage is acquired in early stages of language development (Gvozdev 1961). Simultaneously, the children demonstrated the synonymy of prepositions and adverbs. Accordingly, we speculate that the acquisition of locative prepositions in Russian is based not only on the adaptation of cases, as proposed by Gvozdev (1961), but also on the acquisition of adverbs (Leikin 1998). The results showed that the use of adverbs remained more natural for the children even after the prepositions had been acquired. However, in the situation of compulsory naming, adverbs as well as prepositions were involved in semantic substitutions. In this case, an attempt to use prepositions only caused a large number of such substitutions. The findings seem to reflect the formation of the system of spatial meanings including not only prepositional meanings but also the meanings of the adverbs. This process was seen also both in the decline of the number of errors with age increase and in the differences between the prepositions in terms of their difficulty for the children. The order of complexity for the studied prepositions was the same for all age groups, and largely conformed to the acquisition order of them. No such sort of correlation was found in aphasic patients. Significant differences were found between the results of the children and the aphasic subjects. However, the aphasic patients evidently did not form an integral group. The results revealed differences in verbal behavior and similarities in the error patterns in different aphasic groups. The differences are easily accounted for by mechanisms of specific aphasic syndromes (e.g., Caplan 1987). By contrast, the similarities in error patterns revealed by the aphasic patients, as well as by the children, seemed to relate directly to the properties of the prepositions as a language class. The findings showed that neither the patterns of the most frequent substitutions of children and aphasic patients nor the repertoire and the character of the substitutions were similar. The substitutions of the children provided far less evidence for the proposed features of the prepositional semantic system (e.g., Trajectory of Motion, Contact) than that found in the aphasic group. The distinctive features evident in the children’s groups closely correlated with
Prepositions in language acquisition and aphasia
the spatial relations of non-linguistic reality. That is, the children’s SDF system primarily reflected properties of their spatial cognition. In consequence, the semantic features had different ‘strengths’. For example, the features Enclosure, Common Orientation, and Proximity were more significant than others (e.g., Contact). The children also revealed some semantic features that were not typical of Russian language. For example, the feature Proximity within the meanings in front of, under, and above was revealed (e.g., two separate meanings within the meaning in front of : pered-okolo, ‘in front of-by/at’ and peredvozle, ‘in front of-by/near). Such opposition by the feature Proximity has explicit expression, for example, in certain Caucasian languages, but not in Russian. But it correlates well with properties of spatial cognition. At the same time the children’s SDF system had already been affected by interconnections proper to Russian language. Accordingly, the children’s errors can be attributed to insufficient acquisition of the prepositional semantic system and linguistic operations for the search and selection of prepositional meanings from this paradigm (see Leikin 1998, for more details). By contrast, the adults’ SDF system seemed to reflect not only real spatial relations but also primarily the relationships specific in present language (Leikin 1996). The substitutions of the aphasic patients revealed a wide range of distinctive features and their connections, which closely resembled the proposed system of locative meanings. However, they revealed two more groups of relations. First, the analysis showed the presence of relations that were not typical of Russian (e.g., on-above, in front of-by/at). Seemingly, these findings demonstrate a specific return of aphasic patients to cognitive reflection of the real world. In this case, however, a more complex description of space than that of the children was found. Secondly, the analysis of substitutions showed that the actual relations of locative meanings were often more varied than would be expected from the analysis of prepositional meanings in literary texts (Leikin 1996). For example, derivation of the meaning in front of from the meaning by/at/near was more natural than the use of a long chain of oppositions (Fig. 1). Such a link between two meanings seems to be more a reflection of language experience than a spatial relationship (Leikin 1998). In this context, the semantic substitutions of aphasic patients may be viewed as the result of an erroneous selection of meanings from a paradigm. The results suggest that in the case of aphasia the system of distinctive features itself is not destroyed, but that processing operations with this system are disordered.
Mark Leikin
. Conclusion The findings showed that both normal preschool children and aphasic adults demonstrated difficulties in handling locative prepositions. These difficulties were dissimilar, and this was expressed in the terms of quantitative and qualitative differences between the results of the two groups. At the same time, there were also definite similarities between two groups, which related to the semantic properties of prepositions. However, the aphasic patients’ errors did not reflect regression back to children’s language competence. On the contrary, aphasic patients preserved a certain degree of their previous well-formed linguistic competence. As for the children, the acquisition of prepositions seems to be not completed at the moment of their appearance in speech but this process continues for a considerable period of childhood. The results suggest that both preschool children and adult aphasic patients were able to operate with prepositions as with content words. In this case, the main error type in the usage of prepositions was substitution (both expressive and receptive). The analysis of these substitutions showed that they were of semantic type, and in this respect did not differ from semantic substitutions of content words (semantic paraphasias). Semantic substitutions of the locative prepositions were at least partly in line with the proposed model of a semantic system of locatives.
Notes . Hereafter X→Y means X instead of Y.
References Bowerman, M. (1989). Learning a semantic system: What role do cognitive predispositions play? In M. Rice & R. L. Schiefelbush (Eds.), The Teachability of Language (pp. 133–169). Baltimore: Paul H. Brooks. Brown, R. (1973). A First Language: The Early Stages. Cambridge: Harvard University Press. Caplan, D. (1987). Neurolinguistics and Linguistic Aphasiology. Cambridge: Cambridge University Press. Cromer, R. F. (1991). Language and Thought in Normal and Handicapped Children. Oxford: Basil Blackwell. Friederici, A. D. (1985). Levels of processing and vocabulary types: Evidence from on-line comprehension in normals and agrammatics. Cognition, 19 (2), 133–166.
Prepositions in language acquisition and aphasia
Grodzinsky, Y. (1990). Theoretical Perspectives on Language Deficits. Cambridge: The MIT Press. Gvozdev, A. D. (1961). Voprosy Izutchenia Detskoj Rechy [Issues of Studying Children’s Speech]. Moscow. Heeschen, C., & Kolk, H. (1988). Agrammatism and paragrammetism. Aphasiology, 2 (2), 299–302. Jackendoff, R. (1987). On Beyond Zebra: The Relation of Linguistic and Visual Information. Cognition, 26 (2), 89–114. Jakobson, R. (1990). On Language. Cambridge: Harvard University Press. Johnston, J. J. (1988). Children’s verbal representation of spatial location. In J. Stiles-Davis, M. Kritchevsky, & U. Bellugi (Eds.), Spatial Cognition: Brain Bases and Development (pp. 195–205). Hillsdale, NJ: Lawrence Erlbaum. Kertesz, A. (1982). Western Aphasia Battery. New York: Grune and Stratton. Leikin, M. (1996). The application of distinctive semantic features to the production and comprehension of prepositions in different forms of aphasia. In E. Andrews & Y. Tobin (Eds.), Towards a Calculus of Meaning: Studies in Markedness, Distinctive Features and Deixis (pp. 381–414). Amsterdam, Philadelphia: John Benjamins. Leikin, M. (1998). Acquisition of locative preposition in Russian. Journal of Psycholinguistic Research, 27 (1), 91–108. Luria, A. R. (1970). Traumatic Aphasia. The Hague: Mouton. Miceli, G., Silveri, M., Pomani, C., & Caramazza, A. (1989). Variation in the pattern of omissions and substitutions of grammatical morphemes in the spontaneous speech of so-called agrammatic patients. Brain and Language, 36 (3), 447–492. Piaget, J. (1963). The Origins of Intelligence in Children. New York: W.W. Norton. van Schooneveld, C. H. (1978). Semantic Transformations, Prolegomena to a Calculus of Meaning. Physsardt. Talmy, L. (1983). How language structures space. In H. Pick & L. Acredolo (Eds.), Spatial Orientation: Theory, Research and Application (pp. 225–282). New York: Plenum. Vygotsky, L. S. (1962). Thought and Language. Cambridge, Mass: MIT Press.
Index
A accusative 240 activity 117, 120, 121, 123, 125 adverbials 147 after 91, 97, 102, 105, 109 analogy 197 Anglo-Nigerian Pidgin 244 aphasia 283, 287, 289, 290, 295 Arabic 6, 269–271, 275, 281 Aramaic 250, 251, 263 artificial intelligence 2 aspect 75, 83, 87 aspectual class 117, 125 at 90, 96, 100, 104, 109 attractor set 25 attributive 147 B before 91, 97, 102, 105, 109 beli 179 belo 180 Bislama 5, 231, 235–246 by 92, 97, 102, 105, 109 C Canadian French 210 chaotic state 26 Chinese 42, 142 code-switching 5 cognate object verbs 241 cognitive linguistics 9 colorful semantic PPs 1 colorless case PPs 1 Columbia School 149 comparative constructions 148
complex systems 15, 23, 29 composites 174 content 14 contentives 231 context 123–125 control 63, 66 D dative 240 declension 195, 271, 272 direct 240, 243 dissimilation 201 distinctive feature 165 distinctive feature theory 145 ditransitive verbs 240–242 divalent 173 during 90, 96, 100, 104, 109 E Elementary Catastrophe Theory 25 English 4, 42–45, 50, 54, 68–72, 76, 80, 83, 115 event 115–121, 123, 125 event versus non-event verbs 78 existential 154 extrinsic property 61 F fieldwork 193 figurative meaning 19, 27 forms 17 French 5, 18–21, 44–45, 47, 54, 66, 69–72 French-based Carribean Creoles 210 from sign to text 160, 161
Index
from text to sign 163 functors 231, 233
G generic definition 27 German 235 Gestalt 9, 21, 41 Gestalt psychology 16 governed case-relation marker 148 government 146, 147 grammar 13 Greek 249–253, 257–258, 260–263 demotic 254 Epirus dialects 251–253, 264 katharevousa 254
H Hebrew 4, 6, 193, 234, 235 homogeneity 122
I iconic 153 iconicity 145 idioms 174 “improper” prepositions 256 in 90, 96, 100, 104, 109 incidence 60, 65, 70, 72 inclusion 63, 65, 67, 69, 70, 72 indirect object 240, 242, 243 insertion 66, 67, 71 instabilities and stabilizations 22 integrality 149, 152 intensionalisation 70–72 intransitive 241 invariance 145, 165 isomorphism 152 isotopies 31 Italian 269–270, 272–273, 278
J Jakobsonian 148, 151 Judeo-Greek 5, 6, 249–267
K Kantian schematism 9 kinetic 66–70, 72 Korean 42
L language acquisition 2, 6, 283 Latin 240 lelo 180 lexical valency 171 lexicon 13, 29 linguistic motivation 13 linguistic qualia 21 linguistic schematism 9 logical form 116, 119–121, 124
M macrolevel analysis 160 Maltese 5, 269–281 marked 152, 155, 159, 165 markedness 145, 152, 153, 165 meaning potentials 27 mereological 123, 124 microgenetic 23 model of perception 33 monosemic 149 monovalent 173 motif 12, 22, 42, 47, 48, 51–55 motifs 10, 12, 19, 24 profiles 10, 22 themes 22
N negation 172 nominalization of prepositional phrases 277 norm 195 noun modifiers 127, 139, 141
O object 240 of -insertion 245
Index
on 90, 96, 104, 109 ontology 32 oppositional 147 P Palikur 42, 43, 56 paradigm 197 parenthetical 147 partitive 147 parts of speech 270 passivization 147 perceptive model 11 perfect progressive aspect 85, 103 perfective aspect 83, 96 phenomenological 9 phrasal verb idioms 151 phraseology 280 Pijin (Solomon Islands) 236 plurivalency 172 polysemic 10 polysemy 21, 33, 172 Prague School 149 praxis 19 predicate 147 preposition, spatial 115, 117–125 prepositional phrase 127, 128, 130–141 presupposition 173 profiles 28 profiling 22, 28 progressive aspect 84, 100 proper prepositions 253 prospective use 59, 67, 69–72 Q quantifiers 147 R Russian 5, 240 S schema 41, 43, 44, 46–48, 50, 53, 55 seizure 60, 61
semantic 152 semantic forms 9 Semantic Integrality 5, 150 semantics 283–285, 288–292, 294–296 semantics of nouns 18 semantics of temporal prepositional phrases 75 semiotic valency 5 semiotics 172 Semitic word structure 272 Serbo-Croat 42 Sierra Leone Krio 244 sign-oriented 150 sign-oriented approach 148 signifié 163 signifiés 149 since 92, 97, 102, 105 Sophocles 164 space 41–48, 51–53, 55, 56 Spanish 42, 269 spatio-temporal-existential 150 speech acts 2 spoken language 193 stabilization of semantic forms 10 state 115–117, 119–125 static 64, 65, 67, 68 structure 14 structural stability 25 subject 147 superlative 147 synesthetic 19 syntactic categories 275 syntax 256 T temporal interval 77 temporal prepositions 89 tense 75, 80, 87 thematic identity 29 thematic forms 22, 30 themes 10, 30 theory of catastrophe 4 theory of dynamics 3 theory of forms 22
Index
throughout 91, 96, 101, 105, 109 till 109 time duration 77 time of reference 76 Tok Pisin 236, 244 transitive 241 translation technique 257 trivalent 173 Tzeltal 42, 43
until 92, 97, 102, 105, 109
U unmarked 154, 155, 159, 165 unstable ‘forms’ 13
X X-bar 239 X-bar syntax 1
V verb types 77 verb valency 148 verbal aspect 4 vowel harmony 241
In the series TYPOLOGICAL STUDIES IN LANGUAGE (TSL) the following titles have been published thus far: 1. HOPPER, Paul J. (ed.): Tense-Aspect: Between semantics & pragmatics. 1982. 2. HAIMAN, John & Pamela MUNRO (eds): Switch Reference and Universal Grammar. Proceedings of a symposium on switch reference and universal grammar, Winnipeg, May 1981. 1983. 3. GIVÓN, T.: Topic Continuity in Discourse. A quantitative cross-language study. 1983. 4. CHISHOLM, William, Louis T. MILIC & John A.C. GREPPIN (eds): Interrogativity: A colloquium on the grammar, typology and pragmatics of questions in seven diverse languages, Cleveland, Ohio, October 5th 1981-May 3rd 1982. 1984. 5. RUTHERFORD, William E. (ed.): Language Universals and Second Language Acquisition. 1984 (2nd ed. 1987). 6. HAIMAN, John (Ed.): Iconicity in Syntax. Proceedings of a symposium on iconicity in syntax, Stanford, June 24-26, 1983. 1985. 7. CRAIG, Colette (ed.): Noun Classes and Categorization. Proceedings of a symposium on categorization and noun classification, Eugene, Oregon, October 1983. 1986. 8. SLOBIN, Dan I. & Karl ZIMMER (eds): Studies in Turkish Linguistics. 1986. 9. BYBEE, Joan L.: Morphology. A Study of the Relation between Meaning and Form. 1985. 10. RANSOM, Evelyn: Complementation: its Meaning and Forms. 1986. 11. TOMLIN, Russel S.: Coherence and Grounding in Discourse. Outcome of a Symposium, Eugene, Oregon, June 1984. 1987. 12. NEDJALKOV, Vladimir (ed.): Typology of Resultative Constructions. Translated from the original Russian edition (1983). English translation edited by Bernard Comrie. 1988. 14. HINDS, John, Shoichi IWASAKI & Senko K. MAYNARD (eds): Perspectives on Topicalization. The case of Japanese WA. 1987. 15. AUSTIN, Peter (ed.): Complex Sentence Constructions in Australian Languages. 1988. 16. SHIBATANI, Masayoshi (ed.): Passive and Voice. 1988. 17. HAMMOND, Michael, Edith A. MORAVCSIK and Jessica WIRTH (eds): Studies in Syntactic Typology. 1988. 18. HAIMAN, John & Sandra A. THOMPSON (eds): Clause Combining in Grammar and Discourse. 1988. 19. TRAUGOTT, Elizabeth C. and Bernd HEINE (eds): Approaches to Grammaticalization, 2 volumes (set) 1991 20. CROFT, William, Suzanne KEMMER and Keith DENNING (eds): Studies in Typology and Diachrony. Papers presented to Joseph H. Greenberg on his 75th birthday. 1990. 21. DOWNING, Pamela, Susan D. LIMA and Michael NOONAN (eds): The Linguistics of Literacy. 1992. 22. PAYNE, Doris (ed.): Pragmatics of Word Order Flexibility. 1992. 23. KEMMER, Suzanne: The Middle Voice. 1993. 24. PERKINS, Revere D.: Deixis, Grammar, and Culture. 1992. 25. SVOROU, Soteria: The Grammar of Space. 1994. 26. LORD, Carol: Historical Change in Serial Verb Constructions. 1993. 27. FOX, Barbara and Paul J. Hopper (eds): Voice: Form and Function. 1994. 28. GIVÓN, T. (ed.) : Voice and Inversion. 1994. 29. KAHREL, Peter and René van den BERG (eds): Typological Studies in Negation. 1994.
30. DOWNING, Pamela and Michael NOONAN: Word Order in Discourse. 1995. 31. GERNSBACHER, M. A. and T. GIVÓN (eds): Coherence in Spontaneous Text. 1995. 32. BYBEE, Joan and Suzanne FLEISCHMAN (eds): Modality in Grammar and Discourse. 1995. 33. FOX, Barbara (ed.): Studies in Anaphora. 1996. 34. GIVÓN, T. (ed.): Conversation. Cognitive, communicative and social perspectives. 1997. 35. GIVÓN, T. (ed.): Grammatical Relations. A functionalist perspective. 1997. 36. NEWMAN, John (ed.): The Linguistics of Giving. 1998. 37. RAMAT, Anna Giacalone and Paul J. HOPPER (eds): The Limits of Grammaticalization. 1998. 38. SIEWIERSKA, Anna and Jae Jung SONG (eds): Case, Typology and Grammar. In honor of Barry J. Blake. 1998. 39. PAYNE, Doris L. and Immanuel BARSHI (eds.): External Possession. 1999. 40. FRAJZYNGIER, Zygmunt and Traci S. CURL (eds.): Reflexives. Forms and functions. 2000. 41. FRAJZYNGIER, Zygmunt and Traci S. CURL (eds): Reciprocals. Forms and functions. 2000. 42. DIESSEL, Holger: Demonstratives. Form, function and grammaticalization. 1999. 43. GILDEA, Spike (ed.): Reconstructing Grammar. Comparative Linguistics and Grammaticalization. 2000. 44. VOELTZ, F.K. Erhard and Christa KILIAN-HATZ (eds.): Ideophones. 2001. 45. BYBEE, Joan and Paul HOPPER (eds.): Frequency and the Emergence of Linguistic Structure. 2001. 46. AIKHENVALD, Alexandra Y., R.M.W. DIXON and Masayuki ONISHI (eds.): Noncanonical Marking of Subjects and Objects. 2001. 47. BARON, Irene, Michael HERSLUND and Finn SORENSEN (eds.): Dimensions of Possession. 2001. 48. SHIBATANI, Masayoshi (ed.): The Grammar of Causation and Interpersonal Manipulation. 2002. 49. WISCHER, Ilse and Gabriele DIEWALD (eds.): New Reflections on Grammaticalization. 2002. 50. FEIGENBAUM, Susanne and Dennis KURZON (eds.): Prepositions in their Syntactic, Semantic and Pragmatic Context. 2002. 51. NEWMAN, John (ed.): The Linguistics of Sitting, Standing and Lying. n.y.p. 52. GÜLDEMANN, Tom and Manfred von RONCADOR (eds.): Reported Discourse. A meeting ground for different linguistic domains. n.y.p. 53. GIVÓN, T. and Bertram MALLE (eds.): The Evolution of Language from Pre-language. n.y.p.
