Аннотация:Abstract This study, using the matched‐guise technique, was designed (a) to determine how West Welsh pre‐adolescents would react to Welsh speakers reading a passage of prose in one or other of three language varieties (R.P. English [EE] vs. Welsh‐accented English [EW] vs. Welsh [WW]), and (b) to examine what effect language of testing (English vs. Welsh) might have on children's social evaluations of speech styles. The results showed that EW was judged less good and more snobbish than WW while the former was rated as less strong than EE; no differences accrued between EE and WW. Two interaction effects emerged showing that differences arose in the way subjects evaluated EW depending on the language of testing. He was rated as more selfish than either EE or WW when the language of instructions and the scales were in Welsh and less intelligent than both of them when the testing situation was in English. The results are discussed in relation to previous matched‐guise research in Wales and their methodological implications more generally underscored.
