Geförderte Projekte

Towards a prosodic grammar for rhetorical questions

  • Teilprojekt P6 Innerhalb der Forschungsgruppe FOR 2111 "Questions at the Interfaces", 2.  Runde
  • PIs: Bettina Braun und Nicole Dehé
  • Projektmitarbeiterinnen:  Marieke  Einfeld, Daniela Wochner, Katharina Zahner

Kurzusammenfassung:

Rhetorische Fragen (RQs) unterscheiden sich von  informationssuchenden Fragen ( ISQs) in ihrer phonologischen und phonetischen Realisierung. Phonologisch werden rhetorische Polarfragen im Deutschen mit einem hohen Plateau realisiert (H-%), informationssuchende Polarfragen typischerweise mit einem steigenden Grenzton (H-^H%). Rhetorische Ergänzungsfragen endeten fast ausschließlich in einem tiefen Grenzton (L-%), informationssuchende Ergänzungsfragen erlaubten mehr tonale Variation (L-%, L-H%, H-^H%). Entgegen der semantischen Literatur ist die Unterscheidung zwischen tiefen und hohen Grenztönen unzureichend. Unabhängig vom Fragetyp wurden RQs hauptsächlich mit einer besonderen Art von L*+H Akzent realisiert, in dem L und H mit der betonten Silbe aligniert waren, fortan (L+H)*. Aus phonetischer Perspektive wurden RQs am Beginn der Äußerung häufiger mit behauchter Stimmqualität produziert als ISQs. Darüber hinaus wiesen ihre Konstituenten eine längere Dauer auf, insbesondere das Objekt. 

Das wesentliche Ziel der zweiten Phase ist die Entwicklung einer prosodischen Grammatik für RQs. Darunter verstehen wir zu überprüfen ob (a) eine neue Akzentkategorie (L+H)* gerechtfertigt ist und zumindest zum tonalen System des Deutschen hinzugefügt werden muss, (b) welche Kombination aus phonologischen und phonetischen Merkmalen aus der ersten Phase (Tonakzente und Grenztöne sowie Dauer, Stimmqualität und Tonhöhenumfang) zu wohlgeformten RQs führen, (c) welche der Merkmale und ihre Ausprägungen sprachspezifisch sind, und welche der Grammatik von mehr als einer Sprache zugehörig sind. 

In diesem Projekt untersuchen wir die prosodischen Merkmale in zwei weiteren deutschen Varietäten (Norddeutsch in Kiel und Schweizerdeutsch in Bern), mit einem Fokus auf dem (L+H)* Akzent, ebenso wie nicht-tonale prosodische Unterschiede im Isländischen (Stimmqualität und Dauer). In vier großangelegten Perzeptionsstudien, flankiert von neurolinguistischer Evidenz, testen wir die Wohlgeformtheit von verschiedenen prosodischen Realisierungen von RQs im Deutschen und Isländischen. Darüber hinaus untersuchen wir zwei weitere Sprachen, deren prosodische Systeme sich vom Deutschen unterscheiden (zwei italienische Varietäten als silbenzählende Sprachen und Mandarin Chinesisch als Tonsprache). Die Eigenschaften dieser Sprachen führen zu anderen Beschränkungen bezüglich der relevanten prosodischen Merkmale von RQs (Dauer und f0). 

Link zur Webseite

The Production and Perception of Rhetorical Questions in German

  • Teilprojekt P6 innerhalb der Forschungsgruppe 2111 "Questions at the Interfaces"
  • PIs: Bettina Braun und Nicole Dehé
  • Mitarbeiterinnen: Jana Neitsch, Daniela  Wochner, Katharina Zahner

Kurzzusammenfassung:

Dieses  Projekt  untersuchte die prosodische Realisierung von  rhetorischen Fragen im Vergleich zu informationssuchenden Fragen mit dem gleichen Wortlaut, sowohl in Hinblick auf die tonale Realisierung (Akzente, Grenzentöne) als auch in Hinblick auf die Dauerstruktur und die Stimmqualität. Weiterhin untersuchte das Projekt, ob Fragen aufgrund von prosodischen Markern als rhetorisch oder informationationssuchend klassifiziert werden können, zum einen von automatischen Klassifizieren, zum anderen von Hörern und welche akustischen Hinweise am relevantesten für die Kategorisierung sind.  

Link zur Webseite

Processing prosody across languages, varieties, and nativeness

  • DAAD Project
  • PIs: Bettina Braun, Anne Cutler, Adriana Hanulíková, Andrea Weber, Sabine Zerbian
  • Mitarbeiterinnen: Yuki Asano, Ann-Katrin Grohe, Claudia Friedrich, Helena Levy, Nadja Schaufler, Katharina Zahner

Kurzzusammenfassung:

The overall scientific aim is to investigate the processing of variability in the prosodic domain (intonation, rhythm) in two closely related languages (English, German). Although both languages have broadly similar prosody (e.g., are stress languages), precise timing patterns differ both across these languages and across varieties of each (Northern vs. Southern German; US, British, and Australian English). These subtle differences can signal semantic contrasts and are hence vital for communication. Language learners transfer their native language (L1) timing patterns to second languages (L2), however, leading to potential comprehension difficulty. Although research on comprehension of L2 or accented speech has expanded recently, prosodic factors have hardly been addressed. Our comparisons across languages (English, German), varieties of each, and nativeness (L1, 2) will comprehensively address the case of prosodic processing differences for the first time, enable new developments in variability research on prosodic variability and on such classic factors as frequency (familiar vs. unfamiliar varieties of both L1 and L2), and enable separation of the relative contribution of the principal research factors to variation and its resultant processing effects (e.g., by comparing the present results from related languages against ongoing studies comparing English with African languages, by the German team, and Asian languages, by the Australian team). A powerful special feature of the proposed research is that the expertise of various internationally known psycholinguists will be pooled. Members of the SpeechNet BaWü (http://ifla.unistuttgart. de/index.php?article_id=218&clang=0), a consortium of psycholinguists at the universities of Tübingen, Freiburg, Konstanz, and Stuttgart will collaborate with colleagues from the University of Sydney in Australia.

Link to Website

A new window on intonational form and function


Project summary

This research project applies new psycholinguistic methods to the study of intonational form and function. Only recently, some of these methods (e.g., eye tracking, cross-modal priming) have been successfully employed to investigate intonational processing on-line. Therefore, intonational comprehension research is not restricted anymore to probe for utterance interpretation at the end of utterances when all information is available, but can finally access the time course of integrating intonational information as the utterance unfolds over time. This not only provides valuable information on how listeners actually process different parts of an intonation contour, but allows us to tackle and resolve a number of theoretical issues in intonational phonology and intonational meaning that could not be addressed earlier. These include (a) the question of whether an intonation contour is better represented as a holistic tune or as a sequence of pitch accents which can freely combine with each other and (b) the question on whether the semantic contribution of an utterance is computed compositionally from the meaning of its parts or linked to the overall tune. We will investigate these questions for a semantically fascinating intonation contour in German, the hat pattern, as an example in place for other contours with more than one pitch accent.

Relevant publications

  • Braun, B., Asano, Y. & Dehé, N. (2019). When (not) to Look for Contrastive Alternatives: The Role of Pitch Accent Type and Additive Particles. Language and Speech. Advance Online Publication.
  • Braun, B., & Tagliapietra, L. (2009). The role of contrastive intonation contours in the retrieval of contextual alternatives. Language and Cognitive Processes.
  • You can find a selection of the experimental sentences here.
  • Braun, B. & Asano, Y. (2011). How freely can German pitch accents be combined? 7. Tagung zu Phonetik und Phonologie im deutschsprachigen Raum. Osnabrück, Germany. (talk)
  • Braun, B. & Asano, Y. (2012). Eye-tracking data on the immediate contribution of prenuclear accents and f0-interpolations to utterance interpretation in German. 13th Conference in Laboratory phonology. (poster)
  • Soundfiles for a Submission to  Frontiers can be  found here (Experiment 1, Experiment 2: contrastive topics, Experiment 2: control condition)

Bias in polar questions

  • Research project funded by the German Reserach Council (DFG) within the priority program XPrag.de: New pragmatic theories based on experimental evidence. Funding period: May 2014-April 2017.
  • PIs: Bettina Braun, Maribel Romero
  • MitarbeiterInnen: Filippo Domaneschi, Anja Arnhold


Project summary

Beyond their (arguably common) truth-conditional contribution, different types of polar questions (e.g. Is Jane coming?, Is Jane not coming?, Isn't Jane coming (too/either)?, Is Jane really coming?) are often claimed to have different use-conditional content, most notably pertaining to two kinds of bias: original speaker bias and contextual evidence bias. The current empirical generalizations on polar questions and their bias are partial - existing approaches describe some but not all polar question types - and at times contradictory. Three main lines of analysis have been developed: the first line exploits the notion of "usefulness" of the proposition expressed by the sentence radical (line A), the second line relies on the contribution of the verum operator (line B), and the third line models the differences between the questions at issue in terms of speech acts (line C). Each of these lines accounts for a different set of data and assumes a different pragmatic architecture of discourse and conversational moves and goals. The novelty of the present project lies in the use of experimental production and perception data that will allow us to decide between these analyses or to develop one of them further. Our first goal is to arrive at an empirically founded characterization of the empirical data. Specifically, we will start out with a large-scale, semi-spontaneous production study in German and English, in which we manipulate the two kinds of bias and analyse the linguistic realization of polar questions (positive vs. negation questions, high vs. low negation, intonation). Crucially, pitch accent placement and type as well as boundary tones, which so far have received little attention in the polar questions at issue, will be analysed in detail. The gathered information will be used to design and conduct perception experiments that tackle three subtle and controversial, but theoretically crucial issues, namely: the split between low and high negation questions, the nature of Ladd's (1981) ambiguity, and the acceptability of high negation questions with either. The second goal is to evaluate, modify or develop the existing analyses further, informed by the new experimental results. Specifically, we plan to develop a unified and comprehensive account of polar questions (including the contribution of intonation) that can explain the cross-linguistic differences between English and German (and possibly among other languages later on) and that can be extended to similar pragmatic effects in other question types (e.g. rhetorical effects wh-questions).

The mental representation of non-lexical tone: A Bantu case study

  • This project was funded by the Konstanz-Essex-Development fund (KEDF)
  • PIs: Bettina Braun, Nancy Kula (University of Essex)


Project summary

We use AX tasks and production experiments with novel-words to investigate how Bantu speakers (a language with high and low tones that may spread to the right) mentally represent lexical and non-lexical tones and how productive they can use tone spreading rules when producing novel verbs.