The influence of reference acceptance and reuse on conversational memory traces.

Abstract

Not all pieces of information mentioned during an interaction are equally accessible in speakers’ conversational memory. The current study sought to test whether 2 basic features of dialogue management (reference acceptance and reuse) affect reference recognition. Dyads of speakers were asked to discuss a route for an imaginary person, thus referring to the landmarks to be encountered. The results revealed that the participants’ conversational memory for the references produced during the interaction depended on whether these had been reused during the interaction and by whom, along with landmark visibility during the interaction. These findings have implications for partner adaptation in dialogue, which depends in part on what speakers remember of past interactions.

Conceptual match as a determinant of reference reuse in dialogue.

“…The results corroborated this idea, as the participants recalled the content words they had generated themselves better than the content words which had been generated by their partner. This confirms and extends previous work on how production and generation affect collaborative dialogue (Knutsen & Le Bigot, 2014;Knutsen et al, in press), shedding light on how lowlevel processes affect speakers' memory for past interactions (see also Knutsen & Le Bigot, 2015).…”

Explicit feedback from users attenuates memory biases in human-system dialogue

“…To answer this question, the current study examined participants' recognition of references to landmarks presented and accepted during an interaction with a simulated dialogue system. First, the results confirmed that the participants remembered the references they had presented themselves better than the references that had been presented by the system, confirming that the self-presentation bias also occurs in human-system dialogue (Knutsen & Le Bigot, 2015;Knutsen et al, in press). This is also in line with the idea that similar psychological processes are at play when human beings interact with other humans or with automated dialogue systems (Brennan, 1991;Powers et al, 2005).…”

“…When two human partners interact, they remember better self-presented information and/or information accepted through verbatim repetition (Knutsen & Le Bigot, 2015;Knutsen et al, in press). Is this also the case when a human user interacts with a system?…”

“…Importantly, studies on human dialogue suggest that dialogue memory is more or less accurate depending on whether one needs to retrieve initially self-or partner-produced information from memory (Hjelmquist, 1984;Jarvella & Collas, 1974;Knutsen & Le Bigot, 2015;Stafford, Burggraf, & Sharkey, 1987;Stafford & Daly, 1984). For instance, Knutsen and Le Bigot (2015; see also Knutsen et al, in press) have recently shown that the distinction between self-and partner-production at the time of common ground construction directly affects dialogue memory.…”

“…To do so, they rely on their memory for past interactions, or dialogue memory (e.g., Gibbs, 1986;Keenan, MacWhinney, & Mayhew, 1977;Le Bigot et al, 2013;Pasupathi & Hoyt, 2010). However, such memory is subject to a number of biases which cause some of the encoded pieces of information to Memory biases in human-system dialogue 5 become less readily accessible than others (Knutsen & Le Bigot, 2015;Knutsen, Ros, & Le Bigot, in press). The first goal of the current study is to determine whether these biases are also observed when a human user interacts with a dialogue system.…”