Stress Effects in Vowel Perception as a Function of Language-Specific Vocabulary Patterns

Abstract

Background/Aims: Evidence from spoken word recognition suggests that for English listeners, distinguishing full versus reduced vowels is important, but discerning stress differences involving the same full vowel (as in mu- from music or museum) is not. In Dutch, in contrast, the latter distinction is important. This difference arises from the relative frequency of unstressed full vowels in the two vocabularies. The goal of this paper is to determine how this difference in the lexicon influences the perception of stressed versus unstressed vowels. Methods: All possible sequences of two segments (diphones) in Dutch and in English were presented to native listeners in gated fragments. We recorded identification performance over time throughout the speech signal. The data were here analysed specifically for patterns in perception of stressed versus unstressed vowels. Results: The data reveal significantly larger stress effects (whereby unstressed vowels are harder to identify than stressed vowels) in English than in Dutch. Both language-specific and shared patterns appear regarding which vowels show stress effects. Conclusion: We explain the larger stress effect in English as reflecting the processing demands caused by the difference in use of unstressed vowels in the lexicon. The larger stress effect in English is due to relative inexperience with processing unstressed full vowels.

Perception of local and non-local vowels by adults and children in the South

“…These differences suggest that children were comparatively more uncertain about vowel identity, which corresponds to their greater reliance on stress cues in making identification decisions. This interpretation is consistent with the finding that English has a large perceptual benefit of stress, and the perception of English unstressed vowels, especially short unstressed vowels, can be quite poor (Warner and Cutler, 2017). The children's repetitions seemed unrelated to the dialect and age of the talker because they were distributed with almost equal proportions across the dialects and age groups.…”

A Cross‐linguistic Examination of Toddlers’ Interpretation of Vowel Duration

“…As mentioned above, vowel duration has many different linguistic functions. The languages that concern us here, English and Dutch, are phonologically similar in several respects: Both allow consonant clusters as onsets and as codas; both mark lexical stress with elevated pitch and elongated duration; both commonly reduce vowels to schwa in unstresssed syllables (though English more so than Dutch; Warner & Cutler, ); and in both languages the child‐directed vocabulary is dominated by monosyllables and trochees (e.g., Finegan, ; Kooij, ). However, Dutch listeners seem to weight vowel duration more heavily than English listeners do when judging vowel identity, as described below.…”

Managing Speech Perception Data Sets

“…One major comparative finding is that unstressed vowels are recognized far more poorly than stressed ones in English, while this effect is small and limited to a few vowels in Dutch. In Warner and Cutler (2017), we argue that this difference comes not from acoustic differences in the unstressed vowel space, but rather from listeners' differing need to distinguish among unstressed vowel qualities. Dutch has more unstressed vowels with full vowel quality (not schwalike quality), while unstressed English vowels are usually schwa.…”