|
|
International Workshop
on Cross-cultural and Culture-specific Aspects of Conversational Backchannels
and Feedback
December 5-7, 2006 Data for analysis
available at the workshop
1.
Japanese free conversations: audio 2.
English free conversations: audio 3.
Spanish free conversations: audio 4.
English direction-giving dialogs: audio 5.
Spanish direction-giving dialogs: audio 6.
English quiz/tutorial dialogs: audio 7.
Iraqi free conversation: video 8.
Iraqi direction-giving dialogs: video 9.
Swedish free conversation: vidoe 10.
English courtroom talk: video 11.
Swedish courtroom talk: audio Abstracts An Overview of English Backchannels John
Heritage, UCLA English
language backchannels acknowledge prior utterances or utterance components in
terms of three basic dimensions: (i) locutionary (dealing with the fact
that another has spoken and with the prospects for that speaker's
continuation as a speaker), (ii) epistemic (dealing with the knowledge
relations between speaker and addressee concerning the matter at hand), and (iii)
affiliative (dealing with issues of agreement the content of the speaker's
talk and alignment with its project). These
three dimensions and their interweaving will be summarized across a range of English
language backchannels: mm hm,
uh huh, mm, yes (and its tokens: yeah, yep etc.), oh, really, oh really,
declarative partial repeats (with and without oh prefaces), interrogative
partial repeats (with and without oh prefaces), okay, right and alright. Some
cross-linguistic contrasts and comparisons will be indicated together with
some suggestions about nods and head shakes. Linguistic feedback: some
cultural and linguistic differences Jens Allwood, Goteborg University 1. The role of feedback in communication 2. A simple model of the system of linguistic and
communicative feedback 3. Possible and actual sites of linguistic and cultural
differeces 4. Some corpus based facts concerning feedback - data
from Swedish, German, English and Danish. 5.
Conclusions Ten Questions about Back-Channels,
and Some Incomplete Answers Nigel Ward, The University of Texas at El Paso This talk will briefly discuss some
questions that have guided my research over the past ten years: What scientific significance is there to the
study of back-channeling? What are the practical applications of models of
back-channeling? How do listeners know how to time their
back-channels? How do the phonetic components of a non-lexical
back-channel convey specific meanings? How does the prosody of a back-channel convey specific
meanings? How do spoken backchannels relate to gaze, head
nods, and other gestures? How can dialog systems be designed to enable
back-channelling behavior? How can we teach second-language learners to be
better back-channelers? What methods are most useful for the study of
back-channeling?What tools are needed to support the study of
back-channeling? NEGOTIATION, FEEDBACK AND CULTURE Melvin F. Shakun, New York University We introduce our formal framework for
multiplayer negotiation, Evolutionary Systems Design (ESD). Referring to this
framework, we consider feedback and culture in negotiation with regard to
human and computer agents. Negotiation and
Response Strategies in Japanese Multi-Party Conversations Yasu Katagiri I will describe our initial effort in
collecting and analyzing multi-party conversation data in our MEXT
funded project. The objective of
the project is to seek for a theoretical framework for describing the dynamics
of conversational interactions, from low level turn/floor management,
collaborative topic transition, to the display and management of evaluative attitudes toward proposals and
other participants, in the domain of consensus building. Perception of Virtual Non-verbal Conversational Behaviors Across CulturesDavid Traum*, David Novick+, Bilyana Martinovski*, David Herrera+ and Dusan Jan**Institute for Creative Technologies, Los Angeles, CA +The University of Texas at El Paso, El Paso, TX Research in sociolinguistics and cultural anthropology has shown differences across cultures with respect to non-verbal behaviors in conversation. These non-verbal behaviors include proxemics, gaze, and pauses between turns. Can virtual worlds with multiple conversants saliently represent these differences?ŽÊ Will culturally appropriate parameters for proxemics, gaze and pause be perceived as realistic by native speakers? To answer these questions, we conducted a cross-cultural study of the perceptions of non-verbal behaviors in a virtual world. Native speakers of American English, Mexican Spanish, and Levantine Arabic observed six two-minute animations representing multi-party conversation. The animations were silent.ŽÊ While all of the animations had Afghani characters in a Central Asian setting, the parameters of the characters' non-verbal behaviors were set to values reported in the literature for Americans, Mexicans, and Levantine Arabs; as presented, the animations differed mainly with respect to proxemics. Two different animations for each culture were presented to each observer, and the order of presentations was balanced across observer groups. The observers were asked to rate the realism of the overall animation, the characters' proxemics, the characters' gaze behaviors, and the characters' pauses in turn-taking. The results contained both expected and unexpected elements. As expected, Arab subjects judged the Arab proxemics to be more realistic than both American and Mexican proxemics (p<.01). Arab subjects also judged the Arab animation more realistic overall than the American animation (p<.01). Arab subjects did not judge American proxemics to differ from Mexican proxemics. And judgments of Arab subjects about gaze and pause did not show significant differences across cultures, which was expected because these parameters did not significantly differ across the animations. The judgments of the Mexican and American subjects did not show differences between any of the cultures with respect to proxemics or overall realism. Surprisingly, in the aggregate the subjects saw significant differences between some of the individual animations, even if they did not see significant differences between the sets of animations representing the different cultural parameters. For example, the aggregated subjects judged the proxemics of animation "Arab 1" to differ significantly from those of both "American 1" and "Mexican 2" (p<.001). There was suggestive evidence (p<0.5) that American subjects distinguished the proxemics of "Arab 1" from "American 1", but Mexican subjects apparently did not perceive these differences (p>.59).ŽÊ There is suggestive evidence that Mexican subjects distinguished "Arab 1" from "Mexican 2" (p<.13), but Mexican subjects did not distinguish the pairs of Arab and Mexican animations. What accounts for Americans and Mexicans not perceiving differences in proxemics and overall realism among the different cultures represented in the animations, particularly when they judged individual animations to differ significantly? The significant differences in perceptions of the individual animations suggest that the animations differed from each other along dimensions other than proxemics, gaze and inter-turn pause length. Possible factors include gesture, coincidental coordination of gesture among the characters, and limitations of the virtual world, which may have affected the representations of the different cultural models in different ways. Developing a Cross-Cultural Corpus of Non-verbal Behaviors in ConversationDavid Novick+, Bilyana Martinovski*, David Traum*, David Herrera+ and Dusan Jan**Institute for Creative Technologies, Los Angeles, CA +The University of Texas at El Paso, El Paso, TX To further the study of cross-cultural differences in conversation, it would be helpful to have an audiovisual corpus for analyzing non-verbal behaviors such as proxemics, gaze and inter-turn pause, particularly in multi-party interaction. We report here on the factors associated with the design and collection of such a corpus. To study proxemics in multi-party, it would appear counterproductive to seat the conversants. While the The UTEP Corpus of Iraqi Arabic, for example, provides an excellent resources for analyzing gesture and gaze, the conversants were seated as a way of assuring that their gestures and gaze would be recorded on camera. But asking conversants to stand poses corresponding problems, including those of scenario, context, and audiovisual recording. A cocktail partly might be a straightforward scenario for standing conversation, but most cocktail parties do not involve strangers in a laboratory. The physical context of the interaction might affect proxemics, so interpersonal distances observed in the corpus might differ from those that occur "in the wild." And if one permits the conversants to move around, which seems necessary to establishing natural proxemics, then obtaining reliable video records of the interaction becomes chancy. To pilot a cross-cultural corpus of multi-party interactions for analyis of non-verbal behaviors, we recorded naturally occuring conversations in Los Angeles, CA, and Ciudad Juarez, MX. With approval from UTEP's Institutional Review Board, these conversations were recorded in public places. Some of the conversants were aware of being recorded, and others were not. Informal assessment of the pilot corpus indicates the advantanges and limitations of this approach. The advantages include the n aturalness of the interactions and their settings; these are "feral" conversations taking place because the conversants have real motivations and the actual context in which they placed themselves. The disadvantages include the lack of a meaningful audio track, the lack of multiple camera angles, and a lack of close-ups to track gaze. We propose to provide excerpts of the pilot corpus for group analysis at the workshop and to solicit discussion of design factors. |