On the boundaries of conversation By John Wilson (review)

Abstract

212 LANGUAGE, VOLUME 66, NUMBER 1 (1990) attribute to the French speakers are mirror images of each other, while the actual responses of the French speakers are in between, is interpreted by Persoons as proof that the Dutch speakers try to justify their aggressive attitude by projecting it onto the other community. Persoons further shows that the Flemish students' attitudes are situationally determined: the more the interviewer is perceived to be a member of the other group (from monolingual Dutch via accommodating and gallicized to monolingual French), the more radical the attributed responses are. Van Bezooijen's paper stands out by its sophisticated way offerreting out the contribution that different objective speech characteristics make to a number of attributions by different linguistic groups. In a first experiment it was shown that the subjects who had listened to three versions of a set of recordings (integral, with lowpass filtering, and with random splicing ) rated speakers more reliably on personality when the recording left prosody intact, and more reliably on intelligence and SES when pronunciation was left intact. A second experiment showed that voice quality was relatively more important for foreigners than for native speakers , and that pronunciation (regional accent) played no role for foreign listeners but played an important role in the attribution of intelligence and SES by all native speakers, and in the attribution of negative personality traits by native speakers of the same dialect as in the recording. The first paper of the second half of the book, by A. Kerkhoff, R. van Hout & T. Vallen, echoes some of these findings: prestige and attractiveness are two factors that play a separate role in attitudes toward standard Dutch. The originality of their contribution consists of simultaneous comparisons between Mediterranean minorities' attitudes toward both the standard majority language and the dialect, and between attitudes toward the standard majority language of both dialect speakers and Mediterranean minorities. Unfortunately, there seems to be some confusion in the paper between statistical prediction and real-world causation in the discussion of the role that shame for the native language plays in the second-language acquisition process. Moreover, some of the tables and figures in this contribution are impossible to interpret without constant reference to the text. The originality of A. Vermeer's paper, which also deals with the acquisition of Dutch as a second language by Mediterranean minorities, resides in the style of presentation rather than in the research methodology itself. The relationship between a host of variables at six different points in time is visualized in one schema, which is difficult to absorb, especially as the structure of this schema is different from the structure of the corresponding text. The last two contributions provide an interesting confrontation of two different explanations for why the correlations found between language attitude and language use are often so low: K. Jaspaert & S. Kroon argue that attitude is not a mental construct, but an intermediate variable between social factors and behavior . Unfortunately, their argument is based on path analyses ofdata from rather crudely operationalized attitude measurements (as the authors admit). If their findings could be corroborated by other data, however, they would be a strong argument for replacing attitude data by direct measures of social factors (which would probably have a much higher reliability in the first place). H. Münstermann and R. van Hout provide a more mundane explanation of the discrepancy between attitude and use: statistical problems, viz. heteroscedasticity and false assumptions oflinearity in regression analyses . It is somewhat ironical that this obvious point should only be raised in the last paper, while other contributions do not mention, for instance, whether their data were visually inspected for linearity. After reading this volume one can only conclude that the problem of the discrepancy between attitude and behavior and between different attitude measures is far from resolved. The book provides an impressive array of suggestions for methodological improvement, however , which makes it stimulating reading, even for researchers not particularly interested in the Dutch language area. While the book does not provide a global picture of language attitudes in this area (even contributors dealing with related topics hardly refer to each other), the individual papers are clearly of a more homogeneous...