In many contexts from casual everyday conversations to formal discussions, people tend to repeat their interlocutors, and themselves. This phenomenon not only yields random repetitions one might expect from a natural Zipfian distribution of linguistic forms, but also projects underlying discourse mechanisms and rhythms that researchers have suggested establishes conversational involvement and may support communicative progress towards mutual understanding. In this paper, advances in an automated method for assessing interlocutor synchrony in taskbased Human-to-Human interactions are reported. The method focuses on dialogue structure, rather than temporal distance, measuring repetition between speakers and their interlocutors last n-turns (n = 1, however far back in the conversation that might have been) rather than utterances during a prior window fixed by duration. The significance of distinct linguistic levels of repetition are assessed by observing contrasts between actual and randomized dialogues, in order to provide a quantifying measure of communicative success. Definite patterns of repetitions where identified, notably in contrasting the role of participants (as information giver or follower). The extent to which those interacted sometime surprisingly with gender, eyecontact and familiarity is the principal contribution of this work.

No abstract

Abstract-The way dialogue partners collaborate to achieve a joint task is dependent on the way they construct a common ground of knowledge. Diverse conversational mechanisms are involved in developing a common ground, and repetition phenomena appear to be strongly connected to these processes. This article describes the use of an automatic method to detect, within dialogue transcripts, linguistic cues of engagement and synchrony, by observing repetitions at different linguistic levels. We focus on the relationship between repetition patterns and task-based success in interaction with task-based experience and partner familiarity. We conduct our analysis on the data of the HCRC Map Task corpus. Results suggests that, among other patterns, significant amounts of repetitions play a role for unfamiliar participants, with greater success, in particular, at first attempts.

The Human Communication Research Centre Map Task Corpus is a landmark data source for analysis of multimodal dialogue corpora. The task in this dialogue is the communication by an information giver of a path on a map to an information follower. The corpus includes a measure of task success for each dialogue: the deviation from path scores (the information follower's path from the information giver's path). The original HCRC Map Task works gives information on how the scores for each map were calculated. However, upon analysis and inspection of the corpus materials, we find that it is not possible to reproduce the recorded scores exactly. We report four candidate reconstructions that all appear rational, and note the correlations these lead to with the deviation scores recorded with the corpus. This is important to anyone wishing to reproduce the HCRC method on map-task corpora.

Collaborative dialogue is an important category of human interaction and is widely studied in the literature, especially in fields that attempt to develop new technologies that enable wider varieties of collaborative dialogues. The ingredients of collaboration in dialogue are less thoroughly addressed. We describe the theoretical framework within which we are working and our approach to the construction of a theory of what may make dialogue collaborative. We study a multimodal dialogue corpus (MULTISIMO) testing for positive and negative correlations between dialogue content features and interaction features that one might reasonably imagine are related to assessments of degrees of collaboration. The duration before the second speaker’s first turn and degree of imbalance in the number of words produced by speakers negatively correlate with collaboration assessments (that is, imbalances of content and a delay in the first speaker yielding the floor lead to diminished perceptions of collaboration), while a monotonically increasing duration of focus in successive dialogue sections (rather than overall dialogue duration) correlates positively (that is, when participants are deemed to be extending the duration of the task rather than increasing speed with experience, this is perceived as collaborative).