Reviewed by Dan Michel, University of Florida

Gunther Kaltenböck’s It-extraposition and non-extraposition in English presents a functionalist account of the alternation between it-extraposition and nonextraposition in English. The book proposes that the early traditional extraposition analysis, where the CP subject is moved to object position, does not adequately account for this alternation. While the transformational analysis that K argues against is no longer in widespread use, the remainder of the book is not adversely affected by this. K’s proposal is that a functionalist account is needed in addition to any transformational theory.

Ch. 1, ‘Introduction’ (1–9), briefly states the work’s intention and describes the International Corpus of English (ICE-GB), a computerized million-word database composed of 300 spoken and 200 written texts in various categories. Drawing from this corpus, K demonstrates that it-extraposition is statistically the unmarked form (1,701 occurrences) and that nonextraposition is substantially rarer (217 occurrences). Ch. 2, ‘Previous studies’ (10–26), criticizes previous work on extraposition for the lack of corpora, or for insufficiently representative corpora. Ch. 3, ‘Defining the class’ (27–64), proposes working definitions for the investigated structures, while acknowledging that some constructions seem to rest in fuzzy areas. Helpful summary charts are provided.

Ch. 4, ‘Formal properties of it-extraposition and non-extraposition’ (65–153), is a detailed report, including forty-six tables, of the environments in which the structures under scrutiny appear in the corpus. Preliminary explanations regarding distribution of the constructions are presented, but detailed support is not provided until Ch. 5.

Ch. 5, ‘Functional properties of it-extraposition and non-extraposition’ (154–278), constitutes the heart of the book. First K presents a multitude of theoretical frameworks for information structure, finally settling on a topic-comment structure neatly summarized in two tables (180). K groups it-extraposition into two informational types: Type 1 (given complement extraposition), which is backward-looking to comment on an old topic, and Type 2 (new complement extraposition), which is forward-looking to present a new topic (181). Of these, Type 1 is statistically marked (28.5% of it-extrapositions). Nonextraposition is also divided into two types: Type 1 (given subject clause) is also backward-looking but is here unmarked (80.2% of nonextraposition), while Type 2 (new subject clause) is marked (251). Collapsing the two types of referential duties, it-extraposition clearly favors the introduction of new content while the less common nonextraposition favors reference to prior information.

Ch. 6, ‘Conspectus: Factors influencing the choice of non-extraposition’ (279–93), collects factors from Ch. 5 that influence the use of nonextraposition, including formality and weight (word count) and information status. Ch. 7, ‘Summary and conclusion’ (294–98), extols the value of utilizing a corpus and the need for functional explanation to account for the earlier observation that the two constructs ‘are not usually interchangeable in a given context’ (279). Overall, this book provides a wealth of extraposition data and commentary. In addition, this book is an example of how a large corpus-based study can force one to reconsider what is marked, and the potential reasons for that markedness.