School of Philosophy, Psychology and Language Sciences, University of Edinburgh.
Cognitive and Information Sciences, University of California, Merced.
Top Cogn Sci. 2018 Jan;10(1):55-79. doi: 10.1111/tops.12300. Epub 2017 Nov 13.
When people communicate, they coordinate a wide range of linguistic and non-linguistic behaviors. This process of coordination is called alignment, and it is assumed to be fundamental to successful communication. In this paper, we question this assumption and investigate whether disalignment is a more successful strategy in some cases. More specifically, we hypothesize that alignment correlates with task success only when communication is interactive. We present results from a spot-the-difference task in which dyads of interlocutors have to decide whether they are viewing the same scene or not. Interactivity was manipulated in three conditions by increasing the amount of information shared between interlocutors (no exchange of feedback, minimal feedback, full dialogue). We use recurrence quantification analysis to measure the alignment between the scan-patterns of the interlocutors. We found that interlocutors who could not exchange feedback aligned their gaze more, and that increased gaze alignment correlated with decreased task success in this case. When feedback was possible, in contrast, interlocutors utilized it to better organize their joint search strategy by diversifying visual attention. This is evidenced by reduced overall alignment in the minimal feedback and full dialogue conditions. However, only the dyads engaged in a full dialogue increased their gaze alignment over time to achieve successful performances. These results suggest that alignment per se does not imply communicative success, as most models of dialogue assume. Rather, the effect of alignment depends on the type of alignment, on the goals of the task, and on the presence of feedback.
当人们进行交流时，他们会协调各种语言和非语言行为。这种协调过程被称为对齐，人们认为这是成功交流的基础。在本文中，我们质疑这一假设，并研究在某些情况下，不对齐是否是一种更成功的策略。更具体地说，我们假设只有在交流是互动的时候，对齐才与任务的成功相关。我们提出了一个在差异识别任务中的结果，其中一对对话者必须决定他们是否看到了相同的场景。在三种条件下，通过增加对话者之间共享的信息量来操纵互动性（无反馈交换、最小反馈、完整对话）。我们使用递归定量分析来测量对话者的扫描模式之间的对齐程度。我们发现，在无法进行反馈交换的情况下，对话者会更加注视对方，并且在这种情况下，增加的注视对齐与任务成功率降低相关。相比之下，当反馈是可能的时，对话者通过多样化视觉注意力来更好地组织他们的联合搜索策略。这可以从最小反馈和完整对话条件下整体对齐减少中得到证明。然而，只有进行完整对话的对话者随着时间的推移增加了他们的注视对齐，以实现成功的表现。这些结果表明，对齐本身并不意味着像大多数对话模型所假设的那样具有交际上的成功。相反，对齐的效果取决于对齐的类型、任务的目标以及反馈的存在。
