Speech Science Forum - Leendert Plug (University of Leeds)

Abstract:

I will report on an experiment aimed to test the hypothesis that English listeners orient to canonical forms in judging the tempo of reduced speech. To date, few studies have assessed how listeners estimate the tempo of speech that features deletions. Orientation to canonical forms should mean that perceived tempo is high relative to articulation rate calculated on the basis of surface phone strings. Findings of previous studies (Koreman 2006, Reinisch 2016) may be taken to provide some evidence for listeners’ orientation to canonical forms, but both Koreman and Reinisch point out that listeners may perceive speech with deletions as relatively fast simply because they have learnt to associate phonetic reduction with fast speech, where it is most pervasive. ​

In the experiment I will report on here we attempted to test the hypothesis that listeners orient to canonical forms using highly controlled stimuli featuring exactly one instance of deletion each, and no other instances of phonetic reduction. We capitalized on the fact that in English, the non-realization of schwa in an unstressed syllable (e.g. support, terrain, derive) may result in a surface consonant cluster associated with a different word than the intended one (sport, train, drive). After constructing maximally ambiguous surface realizations for relevant lexical pairs, we presented listeners with sentences containing these realizations. For each sentence form, we convinced some listeners that the sentence contained a disyllabic word (support, terrain, derive etc.), and others that it contained a monosyllabic word (sport, train, drive etc.). Asking listeners to judge the tempo of the sentences allowed us to assess whether the difference in imposed interpretation ― a difference of one canonical phone and syllable ― had an impact on perceived tempo. ​

The results revealed the predicted effect of the imposed word interpretation: sentences with an imposed ‘disyllabic’ interpretation for the ambiguous word form were judged faster than (the same) sentences with an imposed ‘monosyllabic’ interpretation. However, the results were complicated by effects associated with our experimental design, so further research is warranted.