1st Semester 2025/26: Measuring alignment in conversation

Instructors

Anna Palmann & James Trujillo

ECTS

6

Description

In social interaction, humans frequently engage in conversation through a complex dynamic of communicative behaviors (e.g. speech, hand gestures, facial expressions, turn-taking, etc.) unfolding over time. Often speakers adapt aspects of their communicative behaviors throughout the course of a conversation. Such alignment can happen both, within a single speaker (e.g., speech-gesture alignment) or between two or more speakers and on different levels (e.g., alignment in gesture behavior, accent, lexical choice, sentence structure).

 

In this project, we will quantify alignment on different levels by measuring overlap within and between speakers in a conversation corpus consisting of free and task-based dialogues. We will then compare the strength of alignment across different conditions (e.g., with or without background noise, different tasks) in order to draw conclusions about communicative alignment in social interaction.

Organisation

In the first week, we will have an introductory session covering important theories of alignment and approaches to measure it. After that, students will delve more into the literature and work through a tutorial for measuring alignment on a toy dataset. At the end of week 1 we will have a brainstorming session, where students form groups and decide on one level of alignment they would like to investigate in depth and come up with a concrete research question.

 

In the second and third week, students will work largely independently. They will come up with a plan on how to quantify their type of alignment given the provided dataset, create an analysis pipeline, and execute the analysis. In week 3 we will have a group discussion with the possibility for feedback.

 

The fourth week is dedicated to finalizing the projects, writing a short report including the code used for analysis, and giving presentations in the last session.

 

Potential projects could focus on gesture-speech alignment within one speaker, lexical alignment between speakers, gestural alignment between speakers, or other forms of alignment.

Prerequisites

Basic programming skills (R or Python). Methodological tutorials will be available, and a group’s specific approach can be based on their collective skills.

Assessment

• final presentation of the project
• short research report (per group)
• analysis pipeline code (as jupyter notebook or R file)

References

• Duran, N. D., Paxton, A., & Fusaroli, R. (2019). ALIGN: Analyzing Linguistic Interactions with Generalizable tech Niques -A python library. Psychological Methods, 24(4), 419–438.
• Fusaroli, R., Rączaszek-Leonardi, J., & Tylén, K. (2014). Dialog as interpersonal synergy. New Ideas in Psychology, 32, 147-157.
• Louwerse, M. M., Dale, R., Bard, E. G., & Jeuniaux, P. (2012). Behavior matching in multimodal communication is synchronized. Cognitive science, 36(8), 1404-1426.
• Pickering & Garrod, Towards a mechanistic psychology of dialogue. Behavioural and Brain Sciences, 27:169-225, 2004.
• Pouw, W., Trujillo, J. P., & Dixon, J. A. (2020). The quantification of gesture–speech synchrony: A tutorial and validation of multimodal data acquisition using device-based and video-based motion tracking. Behavior Research Methods, 52(2), 723–740.
• Rasenberg, M., Özyürek, A., & Dingemanse, M. (2020). Alignment in multimodal interaction: An integrative framework. Cognitive science, 44(11), e12911.