The Processing of Prosodic Cues to Rhetorical Question Interpretation: Psycholinguistic and Neurolinguistics Evidence

Kharaman, Mariya; Xu, Manluolan; Eulitz, Carsten; Braun, Bettina

doi:10.21437/interspeech.2019-2528

Cited by 11 publications

(17 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This result clearly goes in the same direction and has the same magnitude as [10]. Previous perception data have shown that listeners make use of prosodic cues to RQ vs. ISQ meaning [25,26] in an experimental setting. Here, we show that speakers use these cues also in spontaneous settings to help their listeners deduce the intended meaning.…”

Section: Discussionsupporting

confidence: 65%

The prosodic realization of rhetorical and information-questions in German spontaneous speech

Braun¹,

Einfeldt²,

Esposito³

et al. 2020

Speech Prosody 2020

Self Cite

View full text Add to dashboard Cite

Previous results from laboratory experiments show that German speakers use prosody to distinguish between informationseeking questions (ISQs) and rhetorical questions (RQs). In the current paper we investigate whether pitch accents and edge tones (i.e., those that were typical for RQs and ISQs in the experimental data) are also used in spontaneous speech.As compared to laboratory data, spontaneous speech data are syntactically and lexically more varied. However, notwithstanding more variation in their prosodic realization, RQs and ISQs in spontaneous speech essentially exhibit the same prosodic characteristics as RQs and ISQs in lab speech. Specifically, RQs were most often realized with an L*+H nuclear accent in both polar and wh-questions. Edge tones differed across question types. For polar ISQs, the most frequent edge tone was a high-rise H-^H%, while polar RQs were mostly realized with a high plateau (H-%) or a low-rise (L-H%). Wh-ISQs equally often ended in a low edge tone (L-%) and H-^H%, while wh-RQs most frequently terminated in L-%. RQs were furthermore produced with a slower speaking rate in both settings.Given the similarities between the results for spontaneous vs. lab speech, the use of experimental data to investigate the prosodic realization of different illocution types is validated.

show abstract

Section: Discussionsupporting

confidence: 65%

The prosodic realization of rhetorical and information-questions in German spontaneous speech

Braun¹,

Einfeldt²,

Esposito³

et al. 2020

Speech Prosody 2020

Self Cite

View full text Add to dashboard Cite

show abstract

“…The set of stimuli consisted of 32 German wh-questions (e.g., Who likes lemon? ), which were manipulated by fully crossing four factors: (a) nuclear accent type (late-peak vs. early-peak accent), (b) voice quality (breathy vs. modal voice on the final noun of the wh-question), and (c) duration of the utterance (lengthening or shortening of the utterance duration by 10%), resulting 64 test trails in eight test conditions (cue combinations) (see [14]) in (d) two different experimental settings (Experiment I vs. Experiment II) summing up to 16 conditions in total.…”

Section: Methodsmentioning

confidence: 99%

“…We used the probabilities in Figure 1 for the 16 test conditions. The proportions in Experiment I were taken from the outcome of a psycholinguistic experiment [14,16,23] while the proportions of Experiment II were entirely hypothetical in nature:…”

Section: Present Studymentioning

confidence: 99%

“…Most experiments so far were designed to focus on a detailed picture of the weighting and interplay of the tested cues and went with a high number of conditions or cues, accepting the lower generalisability in terms of lexicalisations [8][9][10][11][12][13]. Fewer studies chose to focus on the lexical generalisability and tested a high number of lexicalisations in fewer conditions, in particular in research at the prosody-pragmatics interface [14,15]. We have previously shown that AL systems predict outcomes of a classical 2x2x2 study design (three factors with two levels each, eight test conditions) reliably and fast [16].…”

Section: Introduction 1backgroundmentioning

confidence: 99%

See 1 more Smart Citation

Reliable Estimates of Interpretable Cue Effects with Active Learning in Psycholinguistic Research

Einfeldt¹,

Sevastjanova²,

Kazak³

et al. 2021

Interspeech 2021

Self Cite

View full text Add to dashboard Cite

Studying the relative weighting of different cues for the interpretation of a linguistic phenomenon is a core element in psycholinguistic research. This research needs to strike a balance between two things: generalisability to diverse lexical settings, which requires a high number of different lexicalisations and the investigation of a large number of different cues, which requires a high number of different test conditions. Optimizing both is impossible with classical psycholinguistic designs as this would leave the participants with too many experimental trials. Previously we showed that Active Learning (AL) systems allow to test numerous conditions (eight) and items (32) within the same experiment. As stimulus selection was informed by the system's learning mechanism, AL sped-up the labelling process. In the present study, we extend the use case to an experiment with 16 conditions, manipulated through four binary factors (the experimental setting and three prosodic cues; two levels each). Our findings show that the AL system correctly predicted the intended result pattern after twelve trials only. Hence, AL further confirmed previous findings and proved to be an efficient tool, which offers a promising solution to complex study designs in psycholinguistic research.

show abstract

“…In rhetorical questions and exclamatives, the (LH)* accent is accompanied by further prosodic modification, in particular lengthening and non-modal voice quality Wochner, Forthcoming), not necessarily co-occurring with the accented word but occurring across the utterance. Listeners, in turn, do not only use information on the pitch accent type when identifying rhetorical questions, but additionally use durational and voice quality cues (Kharaman et al, 2019). Intensity and voice quality could not be analyzed with the present data set because of remote data collection; the fact that 13 The words are not even listed in the CELEX corpus (Baayen et al, 1993); dlexDB (Heister et al, 2011) reveals a very low frequency, ranging from 1 for "Mandalas" (0.43 occurrences per million) to 116 for "Melanie" (50.43 o.p.m.).…”

Section: Crosstalk Between Tone and Intonation And Cross-linguistic I...mentioning

confidence: 99%

How experience with tone in the native language affects the L2 acquisition of pitch accents

2022

Self Cite

View full text Add to dashboard Cite

This paper tested the ability of Mandarin learners of German, whose native language has lexical tone, to imitate pitch accent contrasts in German, an intonation language. In intonation languages, pitch accents do not convey lexical information; also, pitch accents are sparser than lexical tones as they only associate with prominent words in the utterance. We compared two kinds of German pitch-accent contrasts: (1) a “non-merger” contrast, which Mandarin listeners perceive as different and (2) a “merger” contrast, which sounds more similar to Mandarin listeners. Speakers of a tone language are generally very sensitive to pitch. Hypothesis 1 (H1) therefore stated that Mandarin learners produce the two kinds of contrasts similarly to native German speakers. However, the documented sensitivity to tonal contrasts, at the expense of processing phrase-level intonational contrasts, may generally hinder target-like production of intonational pitch accents in the L2 (Hypothesis 2, H2). Finally, cross-linguistic influence (CLI) predicts a difference in the realization of these two contrasts as well as improvement with higher proficiency (Hypothesis 3, H3). We used a delayed imitation paradigm, which is well-suited for assessing L2-phonetics and -phonology because it does not necessitate access to intonational meaning. We investigated the imitation of three kinds of accents, which were associated with the sentence-final noun in short wh-questions (e.g., Wer malt denn Mandalas, lit: “Who draws PRT mandalas?” “Who likes drawing mandalas?”). In Experiment 1, 28 native speakers of Mandarin participated (14 low- and 14 high-proficient). The learners’ productions of the two kinds of contrasts were analyzed using General Additive Mixed Models to evaluate differences in pitch accent contrasts over time, in comparison to the productions of native German participants from an earlier study in our lab. Results showed a more pronounced realization of the non-merger contrast compared to German natives and a less distinct realization of the merger contrast, with beneficial effects of proficiency, lending support to H3. Experiment 2 tested low-proficient Italian learners of German (whose L1 is an intonation language) to contextualize the Mandarin data and further investigate CLI. Italian learners realized the non-merger contrast more target-like than Mandarin learners, lending additional support to CLI (H3).

show abstract

The Processing of Prosodic Cues to Rhetorical Question Interpretation: Psycholinguistic and Neurolinguistics Evidence

Cited by 11 publications

References 15 publications

The prosodic realization of rhetorical and information-questions in German spontaneous speech

The prosodic realization of rhetorical and information-questions in German spontaneous speech

Reliable Estimates of Interpretable Cue Effects with Active Learning in Psycholinguistic Research

How experience with tone in the native language affects the L2 acquisition of pitch accents

Contact Info

Product

Resources

About