Functional prosody transfer in disambiguating questions and statements in an unknown language

The ability to distinguish questions from statements presents an important linguistic competence that is crucial for successful communication. Languages vary in how they mark this distinction, with intonation playing a crucial role in many languages. Portuguese or Basque use solely intonation to distinguish between polar questions and statements, while in English, Polish or German morphosyntactic markers and intonation are used. Mandarin, for instance, relies solely on intonation when there is no question particle. Still other languages, such as Finnish, do not use intonation but question particles alone [1, 2]. Such cross-linguistic differences in the marking of interrogativity might pose challenges for adult language learners when encountering an unfamiliar language. Therefore, in this study, we test to what extent the interpretation of utterances as statements or questions in an unknown language – here Polish – is modulated by the role of intonation in cueing the distinction between statements and questions in listener's native language (L1) – here L1 listeners of Finnish, German, and Mandarin Chinese.

To this end, we pitch-manipulated Polish utterances (e.g., macie kawę; do you have a coffee / you have a coffee) at two positions relevant for question interpretation, i.e., the nuclear pitch accent and the final boundary tone (last two tonal targets in the utterance), and asked L1 speakers of Finnish, German and Mandarin (N=24 per group, no knowledge of Polish, age range between 23 and 39), along with Polish controls (N=25) to judge whether the utterance was a statement or question in a simple forced-choice perception experiment (1 indicating statement, 100 indicating question). For manipulation, we created a continuum of 5 pitch accents (PA) by 5 boundary tones (BT, see [3] for a similar procedure). Modifications were done at the syllable nuclei, with PA modifications (first CV- syllable) ranging from PA1=160 Hz to PA5=208 Hz (with 12 Hz intervals) and BT modifications (second CV syllable) ranging from BT1=146 to BT5=234 (with 22 Hz intervals). The prenuclear region was equalized to an average value of 185 Hz to compensate for the fluctuations of pitch that may influence the judgment before the target manipulation at the utterance end.

Results show that overall pitch manipulations at the boundary tone affected question interpretation more than manipulations at the nuclear pitch accent. Moreover, non-native listeners as a group differed in their sensitivity to pitch manipulations from native Polish listeners: While Polish listeners are unaffected by increasing f0 at the pitch accent in general, non-native speakers are more likely to opt for question interpretation (esp. for level 4 and 5); similar tendencies are found for boundary tone manipulations. Within non-native listeners, interesting differences across groups emerge: Finnish speakers, for instance, are more likely than German and Mandarin speakers to perceive higher boundary tones as indicating a question (esp. level 1 and 2), which might reflect typological differences in functional prosody in question marking. Our findings shed light on functional prosody transfer in disambiguating statements and questions based on intonational cues, which has implications for foreign-language prosody teaching and learning.