Page 80 - Handbooks of Applied Linguistics Communication Competence Language and Communication Problems Practical Solutions

58 Madeleine Brabant, Bernadette Watson and Cindy Gallois

ever, he noted that the vast majority of French speakers did accommodate to an
English-speaking stranger. Work like this has highlighted the importance of
tailoring the method to the research question. Questionnaires are excellent de-
vices for studying overt attitudes, norms, and social rules, but they may be poor
predictors of communication in context.

1.1.2.2. Simulated intercultural communication: matched guise technique

Lambert and colleagues (1960) pioneered a uniquely social-psychological tech-
nique for the study of intercultural communication. The Matched Guise Tech-
nique (MGT) in its earliest form involved a single bilingual or multi-lingual
speaker who recorded two or more monologues of relatively neutral content
(e.g. giving directions in a city) in each of his languages (speakers were always
male). The passages were exact translations of each other, so that speaker char-
acteristics and content were controlled as much as possible. Later work (see
Giles and Powesland 1975) used the same technique to measure different ac-
cents within a language (e.g. RP, London, Scottish, Welsh accents in English).
The idea was that differences in evaluations of the speaker could be attributed to
the impact of the language or accent.
The MGT had the great advantage of being a fairly unobtrusive measure of
reception, and thus arguably less susceptible to social desirability and other
biases than direct techniques like questionnaires. On the other hand, it came
under fairly stringent critiques from within and outside psychology. First, there
had been an almost exclusive reliance on male speakers. Gallois and Callan
(1981) replicated some of the early work in Australia for male speakers, but re-
sults were nearly reversed when the speakers were female. As it turns out, other
social group memberships (gender, age, professional role) and other features of
the context are extremely important influences on evaluations of speakers, so
that the decontextualized aspect of the MGT was called into question. On an-
other front, Nolan (1983) pointed to the danger in assuming that the same
speaker across guises had the same characteristics. He noted that speaker stereo-
types about language (and even more, about accent) have a significant impact on
vocal style. In addition, the diglossic nature of many bilingual environments
means that speakers are unlikely to be equally at home in both or all their lan-
guages for any particular content; once again, the impact of contextual features
cannot be ignored. In recent years, researchers have tended to return to more di-
rect measures of evaluation, but the MGT is still seen regularly in the literature.