The English Noun Phrase
The Nature of Linguistic Categorization
Cambridge: University Press, 2007
Hardback. xiii-379 p. ISBN 978-0-521-84961-6. £85.00*
Reviewed by Evelyne Chabert
Université Stendhal – Grenoble 3
The book under review is one in a series (Studies in English Language) meant to shed new light on a variety of linguistic topics. Its aim is to show that many English noun phrases defy classification, as is partly indicated by its subtitle, The Nature of Linguistic Categorization. The author holds that structural attempts at categorization are most often invalidated by a number of in-between cases which make up as many counter-examples, even if, in the case of NPs, syntactic, semantic as well as pragmatic criteria are taken together. To capture the internal structure of problematic nominal constructions then, it is necessary to give up any hope of arriving at a neat, strict, straightforward typology, and to adopt a different, cognitive-pragmatic, approach. Only notions borrowed from this theoretical framework are liable to explain some of the thorny formal features of NPs.
The two parts of the book correspond to the two main steps in the reasoning. The first (The structural approach, possibilities and limitations) consists in an examination of five problematic nominal constructions, namely close appositions (the poet Burns), appositions with of (the role of president), binominals (her nitwit of a husband), pseudo-partitive constructions (a box of chocolate, a piece of cake), and sort/kind/type-constructions (a sort of kitchen). These being notoriously problematic, and existing structural accounts of them being unsatisfactory, Keizer develops her own contribution, which is to examine whether, by trying to improve on existing accounts, one might arrive at a satisfactory classification. Improving on existing accounts consists in providing a more unified account, bringing together criteria of a different nature: syntactic, semantic, pragmatic and cognitive (e.g. syntactic and semantic headedness, referentiality and predication, definiteness, determination, quantification, the information status of the nominal elements). Still, she shows that even by combining such diverse factors, no discrete, invariable categories can be established. The only solution is to opt for a compromise: one can at best adopt a formal representation of a prototype for each class. Some constructions will be regarded as prototypes, i.e. best examples in a category, while others will be considered as deviating from the prototype. The necessary existence of such prototype effects reflects the extreme complexity of linguistic reality, characterized by gradience (category membership is a matter of degree) and fuzziness (linguistic categories are not always hermetic).
In the second phase of her demonstration (i.e. the second part of the book, The cognitive- pragmatic approach: some applications), Keizer shifts her attention to three cases of information packaging, i.e. in which the speaker seems to have a choice between two or more constructions: complements vs modifiers within the NP (yielding such possibilities or impossibilities as the student of physics with long hair / *the student with long hair of physics, the fear of a terrorist attack in Britain / the fear in Britain of a terrorist attack), discontinuous NPs (a review of his new book came out yesterday / a review came out yesterday of his new book), and possessive constructions (the author's opinion / the opinion of the author). Her point is to show that the choice is not arbitrary, but governed by a combination of pragmatic and cognitive factors, which, again, impedes the possibility of strict classification. And here again, only prototypes can provide a proper account of the structures: some uses of a construction are determined by a cluster of prototypical factors, while others are generated by different criteria and thus deviate from the prototype.
The scope of this review is not to expose the treatment of each and every construction that the book attempts to account for, but, by way of illustration, it might be useful to take a closer look at Keizer's treatment of two structures, one from part 1, the other from part 2: close appositions (the poet Burns, the name Kennedy, my friend John), and possessive constructions (the author's opinion / the opinion of the author). These epitomize her approach.
Where close appositions are concerned, the aim of Keizer's study is to establish the fact that, instead of forming a unified subset of appositions, close appositions constitute an aggregate of different subtypes, which requires both that the very notion of close apposition be redefined and that the existence of prototype effects be admitted. After reviewing at length the existing treatments of the issue, she adds her own criteria. The combination of all these factors leads her to a threefold conclusion:
— A new definition of close appositions should be proposed, in which neither nominal element has independent reference; only the structure as a whole is referential. A consequence of this is that both nouns, and not only the head, fall within the scope of the determiner.
— Close appositions should not be thought of as one homogeneous category, but as a set of different subtypes. The model yields a certain pattern with four subtypes which can be formally represented in the following (simplified) way:
type 1 a det+N+NP the poet Burns
type 1 b det+N+N the word recession
type 2 NP+det+N Burns the poet
type 3 N+NP Doctor Smith
type 4a poss+N+NP my friend John
type 4b poss+N+det+N my friend the poet
— Some constructions, within each subtype, will always be better examples than others. These are prototypical cases, fulfilling all the conditions for category membership. But there are also less prototypical cases, lacking certain properties of the category. For instance, in earlier treatments, an indefinite NN structure (a value K, I have a friend John who is in Linguistics with me) could not qualify for membership in the category of close appositions (because the NP does not refer to an unequivocal entity). Nor could a combination of two common nouns (my friend the poet). In Keizer's treatment, the prototypical structure is type 1a; all other types do qualify for category membership but as less prototypical constructions. The crucial criterion for membership is that the NN construction should enable the speaker to produce "a referring expression which is felicitous in a given context" . In this view then, category membership is dependent on the discourse situation, and is therefore pragmatic in essence.
In the case of the possessive constructions (i.e. the genitive and the of-construction), Keizer holds that ten (mostly cognitive and pragmatic) factors are at play. In her view, the prototypical use of the genitive construction is triggered by the interaction of a cluster of them, namely: a topical, active possessor and a salient possessee; the relation between the two concepts is intrinsic and (semi-)active; the possessor functions as a reference point for the identification of the possessee; the possessor has human reference, is singular and structurally simple (factors (i), (iii), (iv), (v), (vi) and (vii)). On the other hand, the use of the of-construction is determined by a different cluster. In cases in which the genitive construction is used while other, competing factors than those above-mentioned are at work, then the construction is less prototypical.
Though comprehensive and challenging, the book has a few weaknesses. The first is formal: the book may appear a bit too stodgy at times, in particular because of its internal structure, with too many subparts and the very repetitive pattern of the eight chapters devoted to the eight nominal constructions. The second weakness is in the content itself: it is not very clear how the two main parts are logically articulated. It seems that the study of the three constructions in part 2 does not really tie in with the examination of the first five in part 1. In other words, how the two parts concur to the demonstration is no obvious matter.
However, the main interest of the book is not lost, it gives one food for thought: not only does it seek to incorporate cognitive-pragmatic notions into the discussion of the structure of noun phrases, but it also challenges our temptation to discard gradience and fuzziness from our accounts of linguistic facts, and to presuppose that linguistic phenomena are orderly objects always qualifying for neat categorization.
*Paperback reissue, 2010. ISBN 978-0-521-18395-6. £28.99.
Cercles © 2011