ABSTRACT

Working memory (WM) uses knowledge and relations to organize and store multiple items in fewer structured units, or chunks. We investigated: (a) whether a crowd that exceeds the WM capacity is retained better if individuals can be grouped in social chunks; and (b) what counts as a social chunk: two individuals involved in a meaningful interaction or just spatially close and face-to-face. In a delayed change-detection task, participants were more accurate in reporting changes in arrays involving facing (vs. non-facing) dyads whether they depicted meaningful interactions or not (Experiments 1, 2 and 4). This advantage survived a secondary task that increased WM load, only when facing dyads formed meaningful interactions (Experiment 3). Thus, WM uses representation of interaction to chunk crowds in social groups. The mere face-to-face positioning is sufficient to trigger social chunking, although without a semantic anchor this process is fainter and more susceptible to interference.

摘要

工作记忆 (WM) 使用知识和关系将多个项目组织和存储在更少的结构化单元或块中。我们调查了：（a）如果可以将个人分组到社交块中，是否可以更好地保留超过 WM 容量的人群; (b) 什么才算是一个社交块：两个人参与有意义的互动，或者只是空间上接近和面对面。在延迟变化检测任务中，参与者更准确地报告涉及面对（与非面对）二元组的阵列变化，无论他们是否描绘了有意义的交互（实验 1、2 和 4）。这种优势在增加 WM 负载的次要任务中幸存下来，只有当面对 dyads 形成有意义的交互时（实验 3）。因此，WM 使用交互表示来划分社会群体中的人群。仅仅面对面的定位就足以触发社交分块，尽管没有语义锚，这个过程更微弱并且更容易受到干扰。