Automatic semantic similarity priming
Automatic semantic similarity priming
复制标题
DOI：
10.1037/0278-7393.24.3.558
复制
发表时间：
1998-05-01
影响因子：
2.6
通讯作者：
Boisvert, S
中科院分区：
文献类型：
--
作者：
McRae, K;Boisvert, S
含义相似但不相关的单词可能不会自动相互启动（J.R. Shelton & R.C. Martin，1992）。这与分布式模型的中心预测相矛盾，需要对传播激活和复合线索理论进行可疑的修订，并导致自动启动不会挖掘单词含义的说法。假设项目相似度低是造成先前零效应的主要因素。实验 1 使用词汇和语义决策任务的单个和配对呈现，获得了高度相似项目的启动。实验 2 复制了 Shelton 和 Martin，以表明实验 1 的结果不是由于方法论因素造成的。实验 3 显示了对于短 SOA 和长 SOA 高度相似的项目的启动，而对于长 SOA 则仅启动不太相似的项目。因此，语义相似性是语义相关性的一个重要方面，自动启动确实挖掘了词义。
Words similar in meaning but not associated may not automatically prime one another (J.R. Shelton & R.C. Martin, 1992). This contradicts a central prediction of distributed models, necessitates questionable revisions of spreading activation and compound cue theories, and has led to the claim that automatic priming does not tap word meaning. Low item similarity was hypothesized to be the primary factor responsible for the previous null effects. Experiment 1 obtained priming with highly similar items using single and paired presentation with lexical and semantic decision tasks. Experiment 2 replicated Shelton and Martin to show that the results of Experiment 1 were not due to methodological factors. Experiment 3 showed priming for highly similar items for short and long SOAs, but for less similar items for a long SOA only. Thus, semantic similarity is an important aspect of semantic relatedness, and automatic priming does tap word meaning.