Testing theories of working memory: On the relevance of semantic distractor features in the irrelevant speech effect on serial recall

工作记忆模型对干扰物特征对当前行为目标的追求是有害的做出了不同的假设。计划中的研究项目旨在调查语义干扰因素在与任务无关的背景声音干扰连续回忆中的作用。本文提出的实验是基于一项初步研究，该研究比较了两类干扰句子：语义匹配的最后一个词和语义不匹配的最后一个词。从工作记忆模型中可以得出相反的预测，即这些句子应该产生的干扰程度。根据假设不相关的声音效果是由自动干扰引起的模型，干扰序列的破坏潜力仅由其声学特征而不是其语义特征决定。根据允许听觉干扰物捕获注意力的模型，违反先前构建的语义期望的单词应该产生特别大的不相关的声音效果。根据这一假设，与语义匹配的句子相比，语义不匹配的句子对序列回忆的破坏更大。这是一个新奇而有趣的发现，扩展了之前已知的语义效应。然而，在得出明确的结论之前，有必要对语义干扰物特征如何影响听觉干扰物对序列回忆的干扰进行系统的研究。本研究项目提出了这样的调查。设计了10个实验来考察语义干扰因素对工作记忆表现的影响、影响的轨迹和相关过程。对于当前工作记忆模型充分性的经验评估，期望有重要的见解。

Working memory models make contrasting assumptions on which distractor features should be harmful to the pursuit of current behavioral goals. The planned research project seeks to investigate the role that semantic distractor features play in the disruption of serial recall by task-irrelevant background sound. The proposed experiments are based on a pilot study with promising results, in which two categories of distractor sentences were compared: sentences with semantically matching last words and sentences with semantically mismatching last words. Opposing predictions can be derived from working memory models as to the amount of disruption these sentences should produce. According to models that assume that the irrelevant sound effect is caused by automatic interference, the disruptive potential of a distractor sequence is only determined by its acoustic and not its semantic features. According to models that allow for auditory distractors to capture attention, words that violate a previously built-up semantic expectation should produce a particularly large irrelevant sound effect. In line with this assumption, sentences with semantically mismatching last words caused greater disruption of serial recall than sentences with semantically matching last words. This is a novel and interesting finding that extends previously known semantic effects. However, before definite conclusions can be drawn, a systematic investigation seems necessary on how semantic distractor features affect the disruption of serial recall by auditory distractors. This research project proposes such an investigation. Ten experiments are designed to examine the conditions, under which semantic distractor features are harmful to working memory performance, the locus of this effect, and the processes involved. Important insights are expected as to the empirical assessment of the adequacy of current working memory models.