When reading polysyllabic words, assignment of lexical stress is a challenge for readers, especially in languages, such as English or Italian, in which stress position is not strictly determined even though words as well as nonwords typically contain several sublexical cues to stress that readers might use. Here, we attempted to identify such cues using a corpus analysis and to examine their impact on human performance in a megastudy in which participants ( = 45) assigned stress to nonwords ( = 800), stimuli particularly revealing of stress cue use because they have no predefined stress pattern. Hierarchical regression results confirmed an impact of sublexical cues examined in former studies and revealed a role for cues not previously examined, including similarity to real words. These results are informative for computational models of reading as they indicate that readers assign stress to nonwords based on not only sublexical but also lexical information.

中文翻译：

---

阅读意大利语时词汇重音分配的线索：多音节非单词的大型研究

在阅读多音节单词时，词汇重音的分配对读者来说是一个挑战，特别是在英语或意大利语等语言中，尽管单词和非单词通常包含几个词汇下提示来强调读者可能会遇到的重音位置，但重音位置并没有严格确定。使用。在这里，我们尝试使用语料库分析来识别此类线索，并在一项大型研究中检查它们对人类表现的影响，其中参与者（= 45）将压力分配给非单词（= 800），刺激特别揭示了压力线索的使用，因为他们没有预定义的压力模式。分层回归结果证实了之前研究中检查的词汇下线索的影响，并揭示了以前未检查过的线索的作用，包括与真实单词的相似性。这些结果为阅读的计算模型提供了信息，因为它们表明读者不仅基于词汇信息，还基于词汇信息将重音分配给非单词。