People segment the stream of experience into events, or temporal segments that have a beginning and an ending. But how are such event boundaries defined? Linguistic theories of event encoding draw a distinction between bounded events that include an inherent endpoint ("eat a pretzel") and unbounded events that lack such an endpoint ("eat cheerios"). Even though the literature on event cognition has not focused on such abstract aspects of event structure, we hypothesize that sensitivity to boundedness could shape the way events are processed. In the present study, we show that viewers are sensitive to event boundedness in a category identification task and distinguish it from event completion; furthermore, viewers identify bounded events more easily than unbounded events. Sensitivity to boundedness emerges even when viewers are prevented from encoding the events linguistically and thus does not depend on the online use of linguistic distinctions. We conclude that event cognition relies on highly abstract properties of events and their boundaries, and sketch implications of these findings for the way events are described, processed, and used to interact with the world.

中文翻译：

---

有尽头吗？观众对抽象事件结构的敏感性。

人们将经验流分割为事件或具有开始和结束的时间段。但是这样的事件边界是如何定义的呢？事件编码的语言理论区分了包括固有端点的有界事件（“吃椒盐卷饼”）和缺乏这种端点的无界事件（“吃cheerios”）。尽管关于事件认知的文献并未关注事件结构的这些抽象方面，但我们假设对有界性的敏感性可能会影响事件的处理方式。在本研究中，我们表明观众对类别识别任务中的事件有界性很敏感，并将其与事件完成区分开来；此外，与无界事件相比，查看者更容易识别有界事件。即使在阻止观众对事件进行语言编码时，也会出现对有界性的敏感性，因此不依赖于语言差异的在线使用。我们得出结论，事件认知依赖于事件及其边界的高度抽象属性，并勾勒出这些发现对事件被描述、处理和用于与世界互动的方式的影响。