A syntax of substance by David Adger (review)

This book by David Adger results from the investigation of two separate, yet overlapping, issues in the theory of syntax in the generative tradition. The first one is the hypothesis, which goes back to Chomsky 1970, that there is a tight structural parallelism between the verbal and the nominal domains. This builds on the assumption that nouns can take complements as verbs do, so the core part of the NP would be the noun plus its complement, much like the core part of the VP would be the verb plus its complement. Other modifiers to the noun or to the verb (adjuncts) would be peripheral to this core nucleus.

The second main question A’s book deals with is what determines the label of a phrase. For example, how is that a verb phrase receives its label from the verb and not from the category the verb combines with (say, its direct object)? There must be general mechanisms (or labeling algorithms, as they have been recently called) that govern label determination, and a significant amount of work in recent years in the theory of syntax has been devoted to their identification.

When tackling these two issues, A decides to challenge the received views. On the one hand, he denies that there is a close structural parallelism between nominal and clausal structures. On the other hand, he deals with the labeling issue by rejecting the view endorsed by many authors that labeling is connected to structure-building operations. What links the two parts of A’s book is the fact that the analysis of the nominal domain where the NP/VP parallelism is abandoned becomes the main empirical motivation for the new approach to his theory of labeling. In particular, Chs. 2 and 3 discuss labeling and phrase structure theory, while Chs. 4–6 focus on NP structure.

Before turning to a short description of the content of the book and to a general evaluation, a preliminary consideration is in order. A’s book is a typical work in the generative tradition, which may raise some eyebrows. This is so because this book, as is not unusual in the generative tradition, seems to put into question fundamental tenets of the received theory, instead of building on them. Therefore, it might appear that generative linguists need to reinvent the wheel each time they write a new book. In turn, this may be taken as an indication of the weakness of this research tradition, since it might seem that, after decades of research, generative linguists have not been able to reach a stable consensus on the vocabulary of their own theory (Daniel Everett’s (2014) review of A’s book is representative of this reaction). Understandable as this reaction may initially seem, it is not justified. First, a high level of innovation is characteristic of hard sciences, and being suspicious of innovation is more typical of premodern science (although, admittedly, it is always very difficult to compare methodologies, styles of argumentation, and ways the knowledge progresses across different fields). However, the main reason why I think that this reaction is not justified, at least as far as A’s book is concerned, is that such a reaction would reveal a cursory reading and a superficial understanding of its contents. Indeed, if one reads the book carefully, it is easy enough to see the ‘standing on the shoulders of giants’ situation and even a strong continuity with the preceding tradition in some fundamental aspects. Let us take a representative example. As I said, A denies that NPs and VPs have a parallel organization. He does not deny, however, that both have a hierarchical organization. In fact, that sentences, clauses, and phrases are organized hierarchically has been a main tenet of the generative tradition from its beginning sixty years ago, and no generative linguist has ever denied this. Even more telling, it is instructive to see how A motivates the differences between NPs and VPs. One classical diagnostic to probe...