This volume consists of chapters contributed by a group of SLA experts whose work has had a significant influence on the field of contemporary SLA research. Its goal is to provide readers with accessible synopses of various approaches to the study of later language development. Twelve chapters define language either as a computational system that exists beyond general human cognition or as a dynamic system that links the two inextricably. All of them nonetheless presuppose that adult SLA is distinct in that its initial state is constitutive of a completed grammatical system. Each chapter, except for the introduction and the conclusion, follows the same format: introduction of the theory or approach, verification of its validity, discussion of its limitations, an illustrative study, and discussion of the theory's relevance to some of the SLA issues identified by VanPatten and Williams in the introduction. Each chapter ends with discussion questions and further suggested reading.

VanPatten and Williams begin by clarifying concepts central for the understanding of SLA theories, including the notions of model, theory, and hypothesis. They then present a brief account of antecedent research approaches, such as Krashen's monitor theory (chapter 2). Beyond the introduction, three chapters demonstrate how and why generative grammar, as formulated by Chomsky (1981), can help shed light on important aspects of SLA. White (chapter 3) describes a now quite familiar research program, in which Universal Grammar (UG) principles are conceived of as constraints on learners' interlanguage grammars. She shows that not all second language (L2) syntactic knowledge can be ascribed to language input and the learner's generalized statistical learning mechanisms. Carroll (chapter 9) delineates the autonomous induction theory, in which UG is assumed to operate via the language acquisition device, which is controlled by language-specific processing mechanisms. Carroll's framework echoes the methodological assumption that UG should be studied in parallel with other elements (see Chomsky, 1972; Grimshaw, 1981). VanPatten (chapter 7) describes a two-layered processing model whereby observable data are parsed by processing principles, the result of which is then verified by UG (see Townsend & Bever, 2001, for a similar comprehension model). The underlying assumption—that UG is directly accessed in real-time processing—is, however, one of the most hotly debated topics in linguistics (see Gahl & Garnsey, 2004; Newmeyer, 2003). In chapter 8, Pienemann proposes a different processing model that is constraint-based.

Bardovi-Harlig and Ellis (chapters 4 and 5) jointly argue that SLA is meaning-driven, such that grammatical knowledge arises as learners are encouraged to communicate using the newly acquired language, and therefore their grammars should be investigated either as a holistic system (Bardovi-Harlig) or a system of constructions (Ellis). Gass and Mackey (chapter 10) present the interaction hypothesis, which stresses the essential functions of input, interaction, and employment of the target language use in SLA research. From a rather different angle, DeKeyser (chapter 6) contends that SLA can be studied as a mode of linguistic skill acquisition, whereas Lantolf and Thorne (chapter 11) argue that sociocultural variables (e.g., private speech) mediate SLA processes in a way that is indispensable. The volume closes with Ortega's examination of the preceding chapters against five SLA problems she identifies.

All of these chapters are written in such a succinct and precise manner that they will certainly be a welcome addition to current scholarship. A careful perusal of the volume will provide its audiences not only with important background information but also with ways to address SLA issues innovatively. The collection exemplifies a much appreciated effort to bring together the researchers whose work is typically viewed as disparate, so that readers can see how differing approaches might supplement each other.