Information structure: The syntax-discourse interface (review)

The objective of this book is to critically present the big questions related to the mapping between information structure (IS), syntax, and intonation across different theoretical perspectives, including generative grammar, head-driven phrase structure grammar, lexical-functional grammar, dynamic syntax, construction grammar, representation theory, and role and reference grammar.

The second chapter, which takes up more than a third of the book, is an introduction to various relevant notions (some of them very briefly treated, others more extensively) from the formal and functionalist perspectives, and it sets the stage for the issues covered in the rest of the book. Among them are the dichotomy of topic/focus, which is defended in lieu of the traditional focus/presupposition put forth in Chomsky 1971, Jackendoff 1972, Rooth 1985, Reinhart 2006, and others, and the question of how these notions relate to the syntax and the prosody. The weakest part of the chapter is the discussion of the relation between IS and prosody. It is cursory and a bit superficial, with the notions of pitch accent and stress, in particular, confused. This topic by itself would merit a substantial chapter.

In the second and subsequent chapters, Erteschik-Shir defends (44) a discourse-based definition of ‘topic’ and ‘focus’ within file-card semantics (Heim 1982), given in 1 below.

• 1.

  1. a. TOPIC instructs the hearer to locate on the top of his file an existing card with the appropriate reference.

  2. b. FOCUS instructs the hearer to either (i) open a new card and put it on top of the file, and assign it a new label (for an indefinite), or (ii) locate an existing card and put it on top of the file (for a definite).

It is argued that an advantage of such definitions of focus and topic is that they allow for nonbinary divisions of the sentence (unlike the Chomsky/Jackendoff dichotomy of focus vs. presupposition) and can thus accommodate Q&A cases like the following.

• 2.

  • Q: Where is the book?

  • A: I gave it to Mary. (where I is assumed to be the topic, Mary the focus, and the rest of the sentence the background)

While Q&A is the most widely used operational test across different views of focus to determine the ‘focus’ of the sentence, it is also well known that Q&As have limitations in that they allow for accommodation. Indeed, the Q above can be interpreted immediately, in which case it is appropriate to answer it with a sentence like The book is on the shelf. Or alternatively, some tacit discourse presuppositions can be accommodated, that is, the Q can entail that the book is not here, therefore Someone TRANSFER the book to y (where TRANSFER ranges over the set of verbs that contain that notion, such as give, take, send, ship, mail, etc.). Such presupposition defines the PP, to y, as the focus variable and the focus value provided by the answer is Mary. The answer furthermore provides extra (unsolicited) information, which is that the subject of the transfer is I. Indeed, a more minimal answer to the Q is Someone gave it to Mary. Thus, we end up with the Chomsky/Jackendoff view of focus, whereby focus is the part of the sentence that is asserted or nonpresupposed.

A quite different issue is the means that a language employs to identify the focus. As argued by Chomsky (1971) and many others, in English, as in many other languages, it is identified intonationally: the focused part of the sentence contains the ‘intonational nucleus’ or nuclear pitch accent, which is the last pitch accent in the intonational phrase (Pierrehumbert 1980), and is perceived as rhythmically most prominent (i.e. as bearing the nuclear stress). This by no means entails that all languages should identify the focus in the same way. Other languages might use other prosodic means, like prosodic boundaries or morphosyntactic markers.

Another case claimed to favor the IS view in terms of the noncomplementary notions of topic and focus is...