Following previous work in L1 acquisition, SLA has witnessed a veritable explosion of research activity in the domain of temporality since the mid 1980s. In the hefty book under review here, Kathleen Bardovi-Harlig, for several years one of the most active protagonists in the field, sets out to organize the dislocated array of findings, methods, and approaches to date.

SECOND LANGUAGE SPEECH PERCEPTION AND PRODUCTION: ACQUISITION OF PHONOLOGICAL CONTRASTS IN JAPANESE. Takako Toda. Landam, MD: University Press of America, 2003. Pp. 213. $63.00 cloth.

This volume is an investigation into second language (L2) phonology. The subject of Toda's study is the development of the syllable timing system of Japanese by learners who are native speakers of English. Toda points out that acquiring a L2 phonology involves more than producing the language without an accent. It is also necessary to develop a system of representations and rules that will govern language use. English and Japanese differ in that while English relies on stress to determine syllable weight, Japanese uses time. This study investigates the extent to which learners of Japanese as a L2 have acquired that aspect of the Japanese syllable and makes use of both perception and production tasks. One of the assumptions of this study—an assumption founded on previous studies in L2 phonology—is that perception and production are interrelated, but that they are not mirror images of one another.

WASHBACK IN LANGUAGE TESTING: RESEARCH CONTEXTS AND METHODS. Liying Cheng, Yoshinori Watanabe, and Andy Curtis (Eds.). Mahwah, NJ: Erlbaum, 2004. Pp. xxi + 233. $27.50 paper.

Since Alderson and Wall (1993) posited a seemingly rhetorical question, “Does washback exist?” language testing professionals have directed attention to the impact of testing on teaching, learning, educational outcomes, and individuals. This edited volume is the culmination of language testing professionals' conscious efforts to better understand washback. It bears rich testimony to the complex nature of washback phenomena in every corner of educational systems. It is a useful resource for those seeking guidance on their own washback studies—for teachers and school administrators who are haunted by the power of tests and for those who need empirical evidence for and against test washback in specific testing situations.

The growing interest in error analysis shows a concern with errors as evidence of learner language systems. A study of errors becomes a means of revealing what the learner has yet to know. However, the analyst takes a lenient view of errors as such. What has to be remedied is the system, not the symptoms. Nevertheless, the errors produced may be more or less serious. As the analysis will normally reveal a host of deficiencies in the learner system, it may be difficult to know where to start and what to emphasize in applying the results. This is where error evaluation fits in, i.e. the study of error gravity (see Fig. 1.).

This paper responds to David Birdsong's discussion of the positions and the data presented in Felix (1988) and Zobl (1992). We attempt to clarify a number of misconceptions that we feel characterize his discussion and elaborate further on the performance and competence issues contained in the earlier papers.