Anaphora and language design by Eric Reuland (review)

Anaphora and language design presents a novel approach to the theory of binding and anaphoric relations that govern pronouns—pronominals such as her, and anaphors such as herself. The book has two goals. The first is to substitute Canonical Binding Theory (CBT) with a derivational theory of anaphora, based on how chain formation and lexical properties of anaphors interact. The second goal is to offer a truly minimalist derivational theory of anaphora. The book achieves these goals in an elegant and empirically precise way. Therefore, it stands as a must-read for scholars interested in anaphora, regardless of their theoretical background. Below I summarize the contents of the book.

Chapter 1 offers an introduction to the key topics and notions discussed in the book, followed by an outline of the following chapters. The chapter begins by discussing how CBT fails to capture pronoun data in English, Frisian, and Icelandic. For instance, condition A ("anaphors must be bound in the local domain" [Chomsky 1981:188]) fails to distinguish between simple, SE-type, and complex, SELF-type, anaphors in Frisian (e.g., zich vs. zichzelf), and their licensing patterns. This calls for an alternative solution, whose key notions and desiderata are sketched out in the remainder of the chapter. The pre-theoretic intuition behind this solution is simple. Anaphors carry features that "match" or, more accurately, agree with those of the noun phrases that precede them. This agreement mechanism allows us to establish a syntactic "connection" or chain, which in turn licenses an anaphoric (semantic) relation between the pronominal element and the noun phrase.

Chapter 2 discusses the basic syntactic principles according to which anaphoric relations ("binding") are established. First, a distinction between Narrow Syntax and Logical Syntax is introduced. Narrow Syntax amounts to iterated applications of the Merge and Agree operations and Logical Syntax to computing and establishing structural relations (such as binding) among constituents. Anaphoric relations are licensed in A(rgument)-binding contexts, syntactic contexts in which the related noun phrases are arguments of a predicate.

Chapter 3 analyses the factors that play a role in binding: the lexical makeup of pronouns and their role in Chain formation. The morpho-syntactic properties of anaphors are discussed in detail. SE anaphors usually lack grammatical features (e.g., gender, number), unlike the typically feature-rich SELF anaphors. When distinct anaphors enter into syntactic computations, their features (or lack thereof) may license the formation of "syntactic chains", structural relations among noun phrases. Hence, SELF anaphors (e.g., herself ) establish a connection with previous noun phrases by "repeating" grammatical information (e.g., gender). Chain formation and specific features of lexical items, as general principles of grammar, are shown to capture the anaphor data, via their principled interaction.

Chapter 4 offers a brief analysis of economy conditions in processing. Two key principles are put forward. First, anaphoric relations should override logophoric (i.e., discourse-bound) relations. Anaphoric relations locally resolve the interpretive status of pronouns, unlike the open logophoric readings that must probe "beyond" the sentence. Second, when an ambiguity is resolved, the grammar cannot re-introduce discarded interpretations. So, if herself is anaphorically related to the girl, no logophoric interpretation of this anaphora can be "snuck in", in further discourse.

Chapters 5 and 6 discuss how chain formation, combined with the Agree operation, can account for the anaphoric relations that emerge between anaphors and binders. These chapters analyse a rich set of data, ranging from various types of pronominals, to anaphors embedded in subjunctive contexts. The chapters also offer an elaborate discussion on how conditions A and B can be derived as the result of SE and SELF anaphors denoting operations on argument structure. Anaphors, in their role as reflexive pronouns, grant that the identity of the two underlying variables does not render themindistinguishable, for binding purposes. In other words, in a sentence such as The girl washes herself, the morpheme -self in herself grants that the verb washes can distinguish between binder and bindee. Even if they denote the same individual...