Book review

Modularity in Language

Yuasa, Etsuyo. 2005. Modularity in Language. Constructional and Categorial Mismatch in Syntax and Semantics. Berlin: Mouton de Gruyter.

Reviewed by Vanja Kljajevic, NewHeights Software.

Modularity in Language proposes a modular theory of grammar in which prototype analysis and other basic principles of categorization are applied to grammatical categories and constructions. In this approach, a construction is a pairing of form and meaning. The main idea of the book is that syntax, semantics and pragmatics are autonomous modules of grammar, each of which has a set of distinctive primitives irreducible to other levels, as well as its own combinatoric rules and principles. The alignment of representations across levels results in prototypes. Conversely, the lack of alignment results in idiosyncrasies. In case studies from English and Japanese, Yuasa analyzes several types of clauses and grammatical categories as examples of representational mismatch in grammar. These cases are presented as evidence for the hypothesis of autonomous levels in grammar.

The book consists of seven chapters. Chapter 1 introduces the basic concepts of the multimodular approach to grammar, such as constructions, levels, and the differences between the concepts of level and strata in Chomsky's theories (e.g. the D- and S-structures). In this chapter, Yuasa also defines three goals of the book: (1) to propose an original analysis of Japanese and English instances of mismatch in complex clauses, (2) to determine the conditions for mismatch cases that support the hypothesis of autonomous levels, and (3) to prove that mismatch cases support the motivation hypothesis, i.e. the claim that language generally tends to create mismatch cases in order to maximally exploit the existing resources.

Chapter 2 explains correspondence and mismatch among the representations of different levels. Constructions and categories convey different types of information at different levels of grammar. Yuasa adopts McCawley's idea of fuzzy grammatical categories and applies it to the analysis of clausal constructions. Linguistic properties of categories and constructions are viewed as distributed across different levels of grammar. Default alignment of these properties constitutes a prototype, which can be either categorial or constructional. These prototypes are regulated by distinct sets of principles. On the one hand, constructional prototypes are regulated by the General Homomorphism Constraint, which requires that properties belonging to different levels have identical hierarchical relations (e.g. c-command relations) and the same linear order at different levels. On the other hand, the alignment of properties in categorial prototypes is regulated by restrictions that match semantic classes with pragmatic functions, following Croft's suggestion of universal typological prototypes. However, language tolerates violations of default alignment of properties in both constructional and categorial prototypes, i.e. combinations of mismatched representations belonging to different prototypes.

Chapters 3, 4 and 5 discuss examples of mismatch in the following order: the coordination-subordination mismatch, nonrestrictive relative clauses, and performative adverbial clauses. As examples of the coordination-subordination mismatch, Yuasa analyzes the pseudocoordinate conditional in English and Japanese te-coordination. Unlike prototypical cases of coordination, in which both the syntactic and the semantic representations are coordinate and thus congruous, English coordinate conditional clauses combine two incongruous representations: the syntactic representation of a prototypical coordinate with the semantic representation of a prototypical subordinate clause. She finds the reverse pattern in the case of Japanese te-coordination: the syntactic representation of a subordinate is combined with the semantic representation of a prototypical coordinate clausal construction. Yuasa builds her argument for the coordinate-subordinate mismatch by showing that coordinate and subordinate structures behave differently with regard to backward pronominalization, adherence to the Coordinate Structure Constraint, deranking, and endocentricity. Then she lists all the characteristics of coordinate and subordinate clauses in both analyzed cases and examines previous analyses of these clausal constructions, showing that none of the previous accounts is adequate. Finally, she explains that previous accounts failed because, unlike her model, their theoretical apparatus could not capture the dual 'mismatched' nature of these constructions.

Similarly, in Chapter 4 Yuasa argues that none of the existing hypotheses on nonrestrictive relative clauses is adequate. These hypotheses include the Main Clause Hypothesis, the Subordinate Clause Hypothesis, or the combination of the two, which suggests that nonrestrictive relative clauses are subordinate to their antecedents at D- and S-structures, but they are raised at LF. Yuasa claims that nonrestrictive relative clauses are actually a case of constructional mismatch of syntax and semantics. The constructions are said to involve incongruous associations of the semantics of an independent clause and the syntax of a relative clause. Again, she lists properties of subordinate and independent clauses such as grammatically signaled incorporation, capacity to express an assertion and bear an independent speech act, behavior with regard to mood, tense and aspect, scope relations to elements of the main clause, adverb preposing, etc., determining which of these properties are present in English and Japanese nonrestrictive relative clauses. From the perspective of he hypothesis of autonomous levels, Yuasa concludes that the 'curiously mixed' properties of non-restrictive relative clauses can be explained only as an instance of constructional mismatch: the syntactic representation of a prototypical relative clause is associated with the semantic/pragmatic representation of a prototypical independent clause. This, Yuasa claims, not only explains the 'curiously mixed' properties of this construction type, but it also further supports her hypothesis, creating a coherent explanation of a set of distinct grammatical constructions.

In Chapter 5, Yuasa takes another example of dependency mismatch - performative adverbial clauses. She examines the mixed properties of these adverbial clauses by dividing them into properties they share with independent clauses: (1) they can perform an independent speech act; (2) quantified NPs and negation in the main clause do not extend their scope to these clauses; (3) their tense is independent of the tense in the main clause; (4) they are poor antecedents. Performative adverbial clauses also exhibit some features of subordinate clauses (e.g. subordinating morphemes, special verb forms, and word order). Yuasa explains the mixed properties in terms of a syntax-semantics/pragamatics constructional mismatch. While 'normal' adverbial clauses instantiate the prototypical adverbial clause construction, performative adverbial clause constructions match the syntax of a prototypical adverbial clause with the semantics/pragmatics of a prototypical independent clause. Yuasa claims that this incongruous correspondence of representations is the 'result of the autonomy of different levels of grammar' (p. 112). Again, she examines which properties of this clausal construction also characterize independent clauses, and which properties rather characterize subordinate clauses, first in English and then in Japanese.

At this point, the book begins to resemble an intellectual exercise in exemplification of the proposed theory, without further adding to or refining it, and Yuasa switches her focus from the instances of clause construction mismatch to mismatch cases within grammatical categories. So Chapter 6 is a multi-modular analysis of Japanese subordinating conjunction markers that exhibit properties of both prototypical subordinating conjunction markers and of nouns. More specifically, Yuasa examines previous accounts of the subordinating conjunction markers totan 'instance', toori 'to go through', kuse 'habit' and ageku 'the last phrase of renga or linked verse', all of which have mixed properties. Yuasa proposes an explanation in terms of categorial mismatches at the levels of syntax and semantics. The markers behave syntactically as typical nouns, while semantically they function as typical subordinating conjunction markers. Yuasa suggests grammaticalization as a possible source for the categorial mismatch. She explains that semantic information can be affected by historical change independently of syntactic information, which leads to the violation of default correspondences in prototypical grammatical categories, resulting in categorial mismatches. In addition to the studied Japanese subordinating conjunction markers, Yuasa presents several examples of grammaticalization as categorial mismatches, such as the English subordinate conjunction marker while, and the complementizer kuti in Chichewa.

Chapter 7 is a conclusion which briefly summarizes the main points from the previous chapters and emphasizes the role of semantics in grammar. Yuasa's claim that mismatch across levels of grammar occurs in order to make the most of limited resources is perfectly in accord with the principle of least effort, which extends to the realm of cognition. What is missing is a section to tie up her theorizing on multi-level grammar with the ideas on modularity in language, which would then better support the book's title Modularity in Language. As it stands modularity within grammar is really all that the book covers. Also, a better incorporation of the concepts of categorization and conceptualization into modularity of language as a cognitive system would make her theorizing on categorial mismatch in chapter 6 more coherent. It would also strengthen her argument on the motivation for mismatch, especially regarding language acquisition. Thus, the main objection to the book pertains to the concept of module or 'the system of level in Yuasa's proposal: it lacks the theoretical context that would render it more plausible. First of all, it is not clear which conception of modularity Yuasa's multi-modular approach is based on: e.g. Fodorian, Chomskyan, Jackendovian? Second, the author does not determine how the proposed modules function within human cognition, how they interact, and how they process the information they contain. If addressed and answered, these questions could have strengthened Yuasa's book and made it more interesting to a wider readership. That said, Modularity in Language is a well written book that meets the goals set in the introduction. The amount of collected evidence is impressive and will keep the book a valuable source of data from English and Japanese even if and when a better proposal comes along.

References

Bates, E. 2001. Modularity, Domain Specificity and the Development of Language. Philosophy and the Neurosciences. A Reader. Edited by William Bechtel, Pete Mandik, Jennifer Mundale, and Robert S. Stufflebeam. Oxford: Blackwell Publishers. 134-152.

Chomsky, N. 1986. Knowledge of Language: Its Nature, Origin and Use. New York: Praeger.

Chomsky, N. 1995. The Minimalist Program. Cambridge, Mass.: The MIT Press. 3

Jackendoff, R. 2000. Fodorian Modularity and Representational Modularity. Language and the Brain. Representation and Processing. Edited by Yosef Grodzinsky, Lew Shapiro and david Swinney. San Diego: Academic Press. 4-30.

Jackendoff, R. 2002. Foundations of Language. Brain, Meaning, Grammar, Evolution. Oxford: Oxford University Press.

Kempen, G. 1999. Sentence Parsing. Language Comprehension: A Biological Perspective. Verlagi, Springer.

Platzack, C. 2000. Multiple Interfaces. Cognitive Interfaces. Constraints on Linking Cognitive Information. Edited by van der Zee & Urpo Nikanne. Oxford: Oxford University Press.

Rosch, E. (1977). Human Categorization. Advances in Cross-Cultural Psychology. Edited by Neil Warren. London: Academic Press.

Links

© 2002-present ICLA; all rights reserved.
Site admin information and contact