Reviewed by Andrew Carnie, University of Arizona

Focus and topic are among the most useful terms in information structure and the syntax, pragmatics, and semantics of discourse, and yet at the same time they are among the most poorly or inconsistently defined technical notions we work with. Part of the problem is that no clear correspondence holds between these semantico-pragmatic concepts and the syntactic constructions that are supposed to realize them. Luis López provides us with an innovative way to think about the relationship between information structure and the syntactic and semantic interfaces. L asks us to abandon primitive notions of topic/focus and instead characterize various syntactic constructions in terms of features: [+a] for strongly anaphoric items, [+c] for contrastive items, and [+spec] for specifics. This innovation has important theoretical implications: e.g. pragmatic interpretation is interleaved with syntactic phase generation, and  the information structure status of syntactic terms is determined by their feature-triggered structural positions.

Primarily using data from Catalan, L argues that there is no one-to-one link between particular discourse-related constructions—such as clitic left dislocation (CLLD), clitic right dislocation (CLRD), focus fronting—and their discourse function. He argues instead that there is a clear correspondence between the construction, the [±a, ±c] features, syntactic positions, and the precise meaning of each expression. The [±a] and [±c] features are assigned by special pragmatic rules to items in the edges of phases: [+a] is assigned in the specifier of vP and [+c] is assigned in the specifier of FinP. L gives clever arguments from scope and bound variable interpretation for the syntactic position and value associated with each feature: CLLD constructions are [+a, +c], CLRD is [+a, -c], focus fronting is [-a, +c], and default rheme status is [-a, -c]. [±a, ±c] features are assigned derivationally through pragmatic rules at phase boundaries rather than being base generated.

L observes a pattern whereby the feature valuation of a constituent and any subconstituent extracted from it must be identical, but only when the movement crosses a phase boundary. Movement within phases does not obey this constraint. From this L concludes that the feature must have been assigned during the derivation (at the phase boundary) and not in the lexicon/numeration; otherwise we would expect no contrast in behavior as the critical feature would be present throughout the entire derivation. L extends these results to various phenomena using object movement.

This book is not for the layperson; it is highly technically sophisticated. It covers significant ground in syntactic and pragmatic theory. It is also highly innovative and challenges some very foundational assumptions people make both about information structure and syntactic structure. L’s book is lush with interesting data, and the argumentation effectively uses the clever interplay and interaction of grammatical processes. This results a work that belongs on the reading list of every linguist interested in the intersections of syntax, semantics, and pragmatics.