Optimizing Structure in Context: Scrambling and Information Structure (review)

A revised version of the author’s 1996 doctoral dissertation, this book is concerned with scrambling phenomena in German and Korean and the architecture of the grammar. The analysis combines the frameworks of lexical-functional grammar (LFG) and optimality theory (OT). The main claim of the book is that scrambling relates to different components of the grammar and happens as a result of the interplay of these modules rather than as an outcome of an operation by any one particular module. The LFG framework is chosen because it represents grammatical information in a set of distinct but, crucially, parallel structures. According to the author, ‘each component of the grammar has its own interests, represented in some form of principles or constraints, and scrambling happens when the signals from different components mismatch’. OT is used as a concrete tool to demonstrate how principles from different modules of the grammar interact and how conflicts between them are resolved.

The book opens with a useful introduction to the issues discussed in later chapters. In Ch. 1, the author outlines previous approaches to scrambling phenomena and their theoretical and empirical shortcomings. Then follows a description of main tenets and mechanisms of OT and LFG theories uses throughout the book.

Ch. 2 investigates the basic phrase structure of German and Korean with the goal of determining the unmarked phrase structure in these languages. In particular, Choi argues that while both German and Korean exhibit evidence for a VP constituent, the structure of matrix clauses in these languages differs: In German matrix clauses are CPs, whereas in Korean matrix clauses are IPs. In addition, C proposes a set of structural constraints that derive the canonical phrase structure in the spirit of OT.

In Ch. 3, C shows that scrambled and nonscrambled structures differ slightly in their interpretations and classifies these meaning-related effects of scrambling into two distinct categories: semantic (specificity effect) and discourse (antifocality effect). Based on these interpretational differences, C develops a new model of information structure based on two information features, ‘newness’ and ‘prominence’, and then proposes a set of information structuring constraints as the main sources of the focus-related effects of scrambling. Then, C argues that the specificity effect can also be subsumed under the general information-based approach.

Chs. 4 and 5 are concerned with accounting for scrambling phenomena in German and Korean, respectively, using the constraints developed in the previous chapters. It is shown that each alternative structure is the ‘optimal’ output encoding the syntactic and discourse-pragmatic information provided in the input in the best possible way by means of phrase structure and/or prosodic structure. The semantic and discourse effects discussed in Ch. 3 are shown to follow naturally from the constraint competition from different modules of the grammar.

The discussion of Korean in Ch. 5 adds an interesting dimension to the discussion of scrambling since Korean, unlike German, has a special morphological topic marker, nun. The author examines the difference in information encoding between the scrambling of the regularly case-marked phrases and that of the phrases marked with nun and shows that the difference results from the interaction between constituent order and morphological marking.

Finally, Ch. 6 summarizes the main arguments of the book and concludes the discussion.