The logic of pronominal resumption by Ash Asudeh (review)

Ash Asudeh’s The logic of pronominal resumption is devoted to the description and formal representation of resumption across natural languages, a general phenomenon by which a pronoun occupies the base position of a syntactic dependency. Building empirically on a representative sample of languages (Irish, Hebrew, Swedish, Vata, and English) and two types of syntactic dependency (unbounded dependencies and raising constructions, although only one chapter is specifically devoted to the latter), this book, composed of thirteen chapters, is on the one hand clearly inspired by traditional generalizations or distinctions made in the literature, but on the other hand brings a quite novel approach to resumption, coming mainly from the specific framework defended by the author—lexical-functional grammar (LFG) associated with glue semantics (based on linear logic proofs). My overall impression of the book is that it is a very valuable reading for anyone interested in that phenomenon and especially for those interested in how resumption can be formalized.

Building on traditional literature on the topic (McCloskey 2002, 2005, Sells 1984), the author uses three fundamental empirical generalizations as guiding principles for his own theory, which is clearly expressed in Chs. 1 and 2.

The first one is the distinction originating from Sells (1984) between true and intrusive resumptives, which the author restates as grammatically licensed versus processor resumptives (i.e. not fully grammatical, whose production would be related to processing).

The author also builds on a second well-established generalization in the literature based on a distinction between two lines of approach to resumption (distinguishing resumptive strategies across languages or within the same language) regardless of the specific model adopted (head-driven phrase structure grammar (HPSG), LFG, generative grammar): either a syntactic ‘base-generation’ of both the resumptive pronoun and the detached (wh-) constituent, and a binding relation between the two, or a ‘movement’ approach to resumption that equates the two elements to one syntactic function, the resumptive being more or less like a gap. These two approaches just follow from the duality of a resumptive construction, which interacts with both pronominal anaphora (binding processes) and movement or unbounded dependencies (see McCloskey 2005:96, Sharvit 1999, and also Rouveret 2011 for an extended discussion of the issue).

Building on these first two distinctions, the author ends up distinguishing between three kinds of resumption in unbounded dependencies: anaphora-like (true/grammatically licensed) resumptives, which he calls syntactically active resumptives (SARs); gap-like (true/grammatically licensed) resumptives—syntactically inactive resumptives (SIRs) in A’s terminology; and intrusive or processor resumptives. Ch. 2 restates traditional arguments to distinguish between these three uses of resumption, such as island sensitivity, weak crossover, reconstruction, or binding by a quantified antecedent. A very precise analysis is developed for each type of resumption throughout the book: Irish (Ch. 7) and Hebrew (Ch. 8) display SARs, whereas cases of resumption in Swedish (Ch. 9) and Vata (Ch. 10) are used to illustrate the analysis of SIRs. Processor resumptives are discussed further in Ch. 11, with cases of resumption in English.

The last fundamental generalization that A uses (borrowed from McCloskey 2002) relies on the observation that resumptive pronouns are just ordinary pronouns, and that both should make strictly equivalent contributions. A uses this generalization as the starting point of his theory of resumption. But the main originality of his theory of resumption is undoubtedly the correlation with resource sensitivity, and more precisely the assumption that resumptive pronouns constitute a resource surplus (compared to gaps) in semantic composition, thus requiring some managing or consuming device. The general intuition is that the difference between each kind of resumption (SARs, SIRs, and processor resumptives) resides in whether and how it can manage/consume the resource surplus created by the resumptive, through specific properties given to the complementizer system of the language. The intuition is first developed in Chs. 5 and 6, where the author presents an instructive discussion of different types of logic and of the resource sensitivity of natural language, before...