Constructing a Lexicon of English Verbs (review)

The purpose of this ambitious book is to revise Simon Dik’s theory of functional grammar by developing the authors’ lexical component so that the syntactic and semantic components of the overall grammar are more explicitly integrated with the lexical system of verbs. Accordingly, of particular interest is Farber and Usón’s proposal that the syntax of predicates is largely predictable from the lexical semantics of verbs, a familiar notion but argued for with rare perspicacity in the present volume. This is only one of numerous controversial issues addressed with much discussion of a wide variety of supportive and opposing scholarly opinions.

The brief preface (vii–ix) acknowledges the authors’ debt to their colleague, the late Spanish linguist, Leocadio Martín Mingorance, to whom the book is dedicated, and then provides a sketch of the contents and connections throughout the book. Most usefully preceding the six chapters of the book is a list of abbreviations (xix–xxi). Ch. 1, ‘The lexicon and linguistic theory’ (1–36), positions the authors’ approach among others in which the lexicon, particularly the verb as central to argument structure, plays a central role in determining the syntactic structure of sentences. The chapter prepares for the elaboration in the subsequent chapters of a functional-lexematic model (FLM), incorporating ideas on lexical organization from Dik and Coseriu. Like subsequent chapters, this one ends with a summary, but unlike the others, it outlines the content.

Ch. 2, ‘Simon Dik’s functional grammar’ (37–66), narrows discussion in order to describe how Dik’s approach will be revised and adapted to FLM. As singled out in the final summary to the chapter, the model seeks to organize the verb lexicon into a set of domains and subdomains that will allow the syntactic behavior of predicates to be predicted by the semantic content of lexical verbs.

Ch. 3, ‘Lexical organization and the FLM’ (67–142), moves on to semantic fields. F&U reject the description of semantic fields in terms of binary features in favor of a prototype and network analysis. In the course of discussion, they approach the problem of relating more and less specific (or ‘complex’, in their terms) verbs, e.g. foretell, which they define as ‘say correctly that something will happen in the future,’ predict ‘foretell something accurately, based on previous knowledge/experience’ and forecast ‘predict something (of weather and events difficult to foretell)’ (83). They end up endorsing monolingual dictionary definitions of items to the extent that they define more complex words in terms of simpler words. Thus, they use existing agreement among definitions in various monolingual English dictionaries as evidence for the hierarchical relationships they posit among verbs of different degrees of specificity. I note here that dictionaries divide most items into numerous senses before they define them; indeed, as F&U abstract elements from various senses in devising a flow-chart network of relationships among various verbs, the same verb often appears in different positions in the network according to the particular sense at issue. As explicitly mentioned at several points in the book, F&U pitch the current benefits of this approach for computational linguistics and machine modelling of language. Nevertheless, the approach will be controversial to linguists concerned with reducing rampant polysemy in the description of lexical items in favor of core even if prototypical meanings.

Ch. 4, ‘Towards a semantic syntax’ (143–202), discusses the issue that I have most focused on here. It contains a basic observation of great interest (anticipated in the previous chapter, e.g. 138–39), that semantically more complex verbs are often regularly more restricted syntactically than the more basic verbs to which they are related in a semantic hierarchy. This is an excellent point in favor of the relation between the semantics of a verb and its syntax, and it is generously demonstrated for numerous sets of verbs. Generally, the more specific verbs lose syntactic frames (privileges...