Whereas other branches of linguistics and sociology have undergone a significant computational and qualitative transformation over the last decades, Conversation Analysis (CA) and Multimodal Interaction Analysis remain predominantly qualitative fields of research to this day. In part, this is due to core tenets of the field, such as the emphasis on using emic categories and adhering to the sequentially of an interaction. In addition, the local, contextually embedded accomplishment of actions in interaction also makes quantitative multimodal analyses of interactions a tricky undertaking. In part, it may also be due to the fact that the extensive transcription and annotation of large video corpora is extremely time consuming. Nonetheless, there have been a few voices within CA that have raised the questions if quantitative analyses could also be used in CA research (Stivers 2015). These calls coincide with an increase of computational models that can be applied to video recordings of human interaction and beyond that can largely facilitate or even fully automate the process of annotation.

Computer Vision tools can be used to recognize and categorize embodied elements of communication, such as gestures or facial expressions, as well as for demarcating environmental features, such as the background or furniture, distances between participants to an interaction and much more. Automatic speech recognition tools have become increasingly precise and reliable, even in dealing with challenges of spoken or non-standard language. There has been a rich variety of sophisticated Natural Language Processing tools, that can label grammar, moods, topics, narrative sequences, etc. Furthermore, fields such as Corpus Linguistics have developed elaborate methods to process, query and analyse linguistic data quantitatively to derive data-driven trends and insights. This is why we would like to raise the question if and how quantitative methods could also be used effectively in CA and interactional linguistics in order to investigate human interaction from a multimodal perspective.

In the workshop, we would like to support an exchange of ideas and approaches that could expand our understanding of how computational methods could complement qualitative analyses, and also how computational approaches could benefit from theoretical insights. We therefore invite empirical as well as theoretical contributions that describe or reflect the use of quantitative or computational tools in the context of multimodal analyses of interaction. These can cover human interactions, including speech, embodied conduct and sign language, including film and documentary recordings.

The workshop will consist of two presentation sessions and a panel discussion between invited experts representing theoretical and computational approaches.

Please send your abstract of maximum 500 words (excluding bibliography, figures and tables). The deadline for submission is December 11, 2022. Notifications of acceptance will be sent out by January 15, 2023.

In order to guarantee diversity, one person may be the first author of only one submission and co-author of one other submission.

Please, submit your abstracts via EasyChair: https://easychair.org/conferences/?conf=camva2023