Perception of child-produced Polish sibilants: a comparison of native English speakers and Polish Heritage speakers

The Polish language has a complex sibilant structure when compared to languages like English. Of particular interest here are the alveolo-palatal and retroflex sibilants. There have been some previous studies on Polish sibilants examining production and perception of children (under 5 years). However, there is a greater need for understanding adult perception of children’s productions and the perception of different populations listening to children’s productions. Contributing to perception studies would, therefore, allow for a more in-depth analysis of this field of research. This paper builds on the findings of a production-perception study of Polish sibilants in typical children (Zygis et al., 2023) and expands the results by examining English and Heritage Polish population perceptions of Polish children’s productions. The Zygis et al. study examined Polish children and their production and perception of the contrasting sibilants. The study looked at the perception of the children for their own production and adults’ productions. Their study acquired recordings of 80 Polish children aged 35–106 months producing words with /s, ʂ, ɕ/. One of their tasks involved the child participants hearing their own productions of word-medial sibilants: /kasa/, /kaʂa/ and, /kaɕa/ at random. They then had to choose between three images (corresponding to Polish words, e.g.: kasa for cash register) to indicate the stimuli they heard. Their study found that there were a number of acoustic parameters that children used to identify sibilants. They observed that especially the younger children, “appear [to] pay more attention to formants independent of the sibilant and [that] the cue weighting [for these young children] changes during the acquisition process” (Zygis et al., 2023). For the present study, we wanted to explore the perception of these word-medial sibilants for different phonetic environments and for non-native listener populations. The three phonetic conditions included: the whole word as in the original study, the isolated sibilant, and the (isolated) sibilant together with the preceding vowel. The audio files (taken from Zygis et al, 2023) were edited and played to both native English and Polish Heritage listeners at McMaster University in Hamilton, to determine the perception of the three-way Polish sibilant distinction. This distinction is non-existent in English for English listeners or influenced by both Heritage and English phonetics/phonology for Heritage speakers. The sibilant distinction in English lies between /s/ and /ʃ/, therefore the task for the English native participants was to choose between buttons that indicated “kasa | as | s” (for the /s/ sibilant) or “kasha | ash | sh” (for /ʃ/) to indicate which sibilant they perceived. The Heritage speakers of Polish were English participants with varying levels of Polish fluency residing in the Southern Ontario area. They used the same design (three-way sibilant distinction) as the original study. A total of 41 English and 13 Heritage listeners participated in the study. It was hypothesized that the English native listeners would categorize all Polish alveolars as (English) alveolars, but it was not clear how retroflex and alveolo-palatal contrasts from the children’s complex productions would be resolved by the English listeners. It was further assumed that the perception of stimuli with vowel transitions (e.g., /kasa/ and /as/ in contrast to isolated /s/) would significantly differ comparing English listeners and Polish Heritage listeners. In our results, English participants increasingly categorized all manipulations of /s/ as /s/, and /ɕ/ as the /ʃ/ sibilant, especially for the older children’s productions. Their perceptions for the retroflex /ʂ/ was split, half as /s/ perceptions, across conditions. Phonetic information in the form of formants (on top of the spectral noise of the isolated sibilant) did not significantly improve distinction for the English participants. The Polish Heritage speakers showed difficulty in correctly identifying /ʂ/ variations especially in the older children. Phonetic environment and age had varying effects depending on the sibilant. As Polish Heritage participants are familiar with three-way sibilant contrasts, it was interesting to see how these Heritage speakers’ classification differed from that of English participants, especially for stimuli from children who are in the very initial stages of speech development (i.e., decreased articulatory and acoustic accuracy).

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.