The study also highlights the constructs of current linguistic theory, arguing for distinctive features and the notion 'onset' and against some of the claims of Optimality Theory and Usage-based accounts.
The importance of Henk Zeevat's new monograph cannot be overstated. [...] I recommend it to anyone who combines interests in language, logic, and computation [...]. David Beaver, University of Texas at Austin
Date: Fri, 10 Oct 2003 11:01:37 +0200 From: Bert Remijsen <A.C.L.Remijsen@let.leidenuniv.nl> Subject: Lexical Categories: Verbs, Nouns and Adjectives
Baker, Mark C. (2003) Lexical Categories: Verbs, Nouns, and Adjectives, Cambridge University Press, Cambridge Studies in Linguistics 102.
Bert Remijsen, Leiden University
DESCRIPTION OF THE BOOK'S PURPOSE AND CONTENTS
This book presents a hypothesis on the nature of the lexical categories within the Generative Grammar research program (Principles and Parameters framework). This is a topic on which Generative Grammar has had little to say so far. The lexical categories are either assumed or ill defined in terms of the vague features [N] and [V]. The three core chapters of Baker's book present and discuss definitions of verbs, nouns and adjectives, respectively. These categories are defined in syntactic terms, and Baker hypothesizes that they are present universally. Verbs are defined as lexical categories that take a specifier, and nouns as bearers of a referential index. The third lexical category, adjectives, is distinguished negatively, having neither of these characteristics. These hypotheses are evaluated on the basis of a wide range of languages and syntactic processes. Problematic cases, such as languages in which nominal and adjectival predicates are indistinguishable from verbs, are accounted for in terms of (null) functional heads. These can change the valence of one lexical category in terms of the above-mentioned definitions to that of another. But even for languages where some of the categories have been claimed to be collapsed together, Baker maintains that there is evidence for the same three-way contrast. His heuristic is to look for category-critical behavior in contexts where the interference of obscuring functional heads is minimal.
This is a book that brings generative and functional / descriptive linguistics closer together. As Dixon (1997:132) writes, adequate language description requires that "each analytic decision has to be approached as an open question". One of the most basic decisions is undoubtedly that of the lexical categories: which categories can justifiably be distinguished in this language? An answer to this question is likely to figure at the beginning of any morphosyntactic description (Payne 1997). As Baker recognizes, this issue has received little attention in generative theory. Nouns, verbs and adjectives were distinguished in Chomsky (1970) in terms of the binary features [N] and [V], which were not fleshed out theoretically. The lack of coverage of this topic in generative grammar limits the relevance of this and comparable frameworks for the descriptive linguist. This is the gap Baker intends to fill with this large-scale study of the lexical categories in the generative framework, as it aims "[to redeem] the long-standing promissory note known as + / - N and + / - V. Such a theory should provide a unified account of the range of grammatical environments in which one lexical category can be used but not another, and of differences in the internal structure of words and phrases headed by the various lexical categories." (p.17) Baker is not satisfied with vague semantic characterizations of the lexical categories, such as that nouns prototypically refer to things, verbs to events and adjectives to properties. Such an account is unsatisfactory because of the great number of apparent contradictions. Baker uses the English minimal-set example of 'hunger' (noun) vs. 'to hunger' (verb) vs. 'hungry' (adjective), where each form evidently refers to the same event, irrespective the difference in lexical category. Baker's alternative is to define the lexical categories in terms of their syntactic behaviour: - A verb is a lexical category that takes a specifier - A noun is a lexical category that bears a referential index - An adjective has neither That is, the theoretically empty binary features [V] and [N] are replaced by privative syntactic features. And while these definitions bear obvious relations to the above- mentioned functional characterizations in terms of prototypes, they afford a clean cut between the categories. In general, there are three possible strategies to distinguish the lexical categories: semantically, morphologically, and syntactically. The semantic approach is that of the semantic prototypes and fuzzy boundaries between them, which has obvious drawbacks. The approach via morphology, well known from the grammarians of Latin and Greek, is useless for languages of the isolating type. Baker has explored in detail the potential of the third possibility. His syntactic definitions are harder to check than the alternative semantic and morphological ones. That is, one can establish at a glance whether a Latin noun takes inflection characteristic of nouns, but it takes a little more work to look for evidence of a referential index. Still, if this can give us clear-cut distinctions between the lexical categories, the exercise is well worth the effort. It is also worth noting that Baker's approach crucially depends on the assumption of Principles & Parameters principles such as the ones mentioned below. As a consequence, this syntactic strategy for distinguishing the lexical categories is relatively theory-dependent. But I would be inclined to turn this the way around, and say that this important work makes the generative framework more attractive. The reader will decide.
Baker brings his definitions of the categories to bear on a wide variety of languages, with special attention for languages that have been claimed not to have the category under consideration, and for languages claimed to have more than these three categories. Data from Mohawk, Edo and Chichewa, partly collected by the author himself, constitute the backbone of the cross-linguistic underpinning of the hypotheses. Many other languages have been taken into account on the basis of primary descriptions and secondary analyses. It is clear that the author has taken seriously the challenge to support his hypotheses with cross-linguistic evidence.
Let us now consider Baker's definitions of verbs and nouns, which distinguish the three lexical categories. Verbs are defined as lexical categories that have a specifier. As Baker writes, "[t]he most challenging aspect of defending [this definition] is not to show that all verbs have specifiers but that the other lexical categories cannot have them." He deals with nominal and adjectival predicates by postulating a functional head (Pred) above the noun / adjective predicate, so that the specifer can be related to this functional head rather than to the lexical category below it. Baker attributes problematic cases of languages in which nominal and adjectival predicates are indistinguishable from verbs to a null Pred, which at face value suggests that nouns and adjectives themselves can take specifiers. Baker provides compelling evidence for the reality of such a potentially null functional category, among others from causative verb formation. In many languages, genuine verbs can be raised into a dominating verb head to form a causative verb, whereas nouns and adjectives cannot, even if they are indistinguishable from verbs in simple predicates. The ungrammaticality of causatives that involve the raising of a noun or an adjective can be accounted for in terms of the 'Proper Head Movement Generalization' axiom, which vetoes movement of a lexical head to a lexical head movement via a functional head. This supports the hypothesis of a null Pred blocking the movement of predicate nouns and adjectives. Baker duly reports that in one case, the Imbabura dialect of Quechua, causatives can be formed by means of the same morphological derivation from either verbs, nouns or adjectives, in violation of the prediction that Pred would bar causative formation for nouns and adjectives via the same mechanism as for verbs. This appears to be the only phenomenon that is difficult to reconcile with Baker's hypotheses. Pred is also evident from unaccusativity diagnostics. The subject of a nominal or an adjectival predicate behaves like the subject of transitive verbs but unlike the subject of an unaccusative verb or the object of a transitive verb, with respect to a number of syntactic processes that involve movement. These regularities are to be expected if the former are directly dominated by a functional head, rather than by a lexical one. The independently-motivated 'Empty Category Principle' (ECP) states that is more difficult to extract from a functional phrase (i.e., Pred) than from a lexical one (V). These examples illustrate how Baker's account of the distinction between the lexical categories uses principles that were already available in the generative tradition.
Whereas the feature that distinguishes verbs is purely syntactic, the defining characteristic of nouns has more of a semantic basis. Nouns are defined as lexical heads that have a referential index, or in semantic terms, a criterion of identity - i.e., a noun can be evaluated with respect to sameness in relation to another noun. Baker makes this feature syntactically relevant by postulating that the referential index has to be licensed by matching with a structurally adjacent index, such as a theta role of a sister node. Just as in the case of the definition of verbs, this definition of what a noun is ties in with syntactic behavior that is typical of this category. Examples include the ungrammaticality of nouns on their own, their ability to bind pronominal elements, and their suitability to fulfill theta roles. The association between nouns and both quantifiers and determiners can also be analyzed in a more explanatory adequate way on the basis of the above-mentioned definition of what a noun is. In this context and elsewhere in the book, Baker relates the syntactic regularities expressed by generative principles to the nature of the lexical categories. This provides a stronger, more explanatory motivation of why those principles should be the way they are. This is an important quality of the book.
As he does in relation to the other categories, Baker discusses in considerable detail languages that have been claimed not have nouns. He uses examples from, among others, Nahuatl to show that in some languages adjectives can take on nominal inflection to appear in the same position as the contentious nouns. As with other problematic cases, Baker contends that the noun category is there in Nahuatl, but that the distinction with adjectives is obscured by functional heads that lend adjective projections a referential index. "The critical task, then, is to find ways of isolating the lexical heads from their functional support systems" (p. 177). He goes on to demonstrate that in Nahuatl and other languages like it, noun incorporation offers a heuristic to distinguish nouns from non-nouns. In an appendix, Baker argues that adpositions constitute a functional rather than a lexical category, on a par with the Pred in that they can alter the valence of a noun - the adposition can bind the referential index, so that the resulting phrase is relative free in its syntactic distribution.
I would conclude that the author fulfills the ambition to present a unified account of the lexical categories without fuzzy boundaries between them. As such this book has a lot to offer to students of language, whatever their theoretical perspective. Within the generative tradition, the author widens the scope to include an essential but hitherto neglected topic. For those outside the generative tradition, the book offers an overview of syntactic heuristics critical to the distinction between the lexical categories. As such the book may be a valuable source for descriptive linguists confronted with an unclear lexical category distinction. There certainly is a threshold to overcome for the reader who is not familiar with the generative framework, although the author has made an effort to briefly introduce relevant principles before using them. Still, I expect that many may feel it is worth the effort, because this work is directly relevant to linguistic description and comparison. The book is both scholarly and written with enthusiasm.
Chomsky, N. (1970) Remarks on nominalization. In Jacobs, R. & Rosenbaum, P. (eds.) Readings in English transformational grammar. Waltham (Mass.): Ginn, pp. 184-221.
Dixon, R. M. W. (1997). The rise and fall of languages. Cambridge University Press.
Payne, T. E. (1997) Describing morphosyntax. Cambridge University Press.
ABOUT THE REVIEWER:
BIOGRAPHICAL SKETCH Bert Remijsen's main research interest is prosodic typology. Fieldwork research on this topic in the Raja Ampat archipelago (west of New Guinea) inspired him to start working on a language description of Matbat, an undocumented Austronesian language of the island Misool. The most vexing problem so far in this project is the distinction between the grammatical categories.