Semantic memory and the generation effect: Some tests of the lexical activation hypothesis.
Semantic memory and the generation effect: Some tests of the lexical activation hypothesis.
复制标题
语义记忆和生成效应：词汇激活假设的一些测试。
DOI：
--
复制
发表时间：
1985
期刊：
影响因子：
--
通讯作者：
J. Hampton
中科院分区：
文献类型：
--
作者：
J. Gardiner;J. Hampton
英国伦敦城市大学的约翰·M·加德纳和詹姆斯·A·汉普顿如果一个单词是自己创造的，而不是读出来的，那么它更有可能被回忆起来。这一发现--生成效应--被归因于语义记忆，更具体地说，归因于组成词在主观词典中的表征的语义特征的增强激活。三个实验表明，对于有意义但没有意义的字母二元语法(例如，E-T与E-C)；对于单位化的两位数字(例如，28比2、8)；对于熟悉但不熟悉的名词复合词(例如，芝士蛋糕与奶酪番茄酱)，同样的生成效应发生在有意义而不是无意义的字母二元组(例如，E-T与E-C)上。这些结果表明，语义记忆参与对生成效应是关键的，因为被回忆的项目必须形成一个完整的功能单位。主观词汇中的表征既不是效果产生的必要条件，也不是充分条件。近年来，人们对一个人的记忆是如何通过自我生成要记住的项目来影响的重新产生了兴趣。这种兴趣很大程度上源于并集中于对生成中呈现的单词的卓越记忆，而不是阅读任务--Slamecka和Graf(1978)所描述的生成效应；另见Jacoby，1978)。在SLA-Mecka和Graf的生成任务中，受试者被给予一个刺激词，一个规则，说明它与目标反应之间的某种关系，以及反应词的首字母。阅读任务的不同之处只在于回答词被完整显示。他们的研究表明，在许多不同类型的规则下，都会出现大而相似的生成效应；这种效应发生在再认记忆、线索回忆和自由回忆学习中；而且这种效应不会延伸到刺激单词--它仅限于对应答单词的记忆。随后的研究表明，生成效应对记忆项目的词汇状态很敏感，对于McElroy来说，两位作者在本文的研究中做出了同样的贡献。我们非常感谢弗朗西斯·盖伊的实验援助。如需重印，请发送至伦敦北安普顿广场城市大学社会科学与人文学系心理科John M.Gardiner，EC1V OHBV England。如果作者要求，也可以提供实验材料的副本。Slamecka(1982)的研究表明，以字母移位规则产生(或阅读)的可发音非词在记忆中不存在代发效应。根据这一发现，他们反对对生成效应的简单唤醒或努力解释，而是提出语义记忆参与是生成效应出现的必要条件。更具体地说，他们从词汇激活的角度提出了一种解释，根据这一解释，生成任务导致语义特征的激活，从而增加了获得痕迹的可能性。我们称之为词汇激活假说。Slamecka和Fevreiski(1983)进一步完善了这一观点，表明即使在自我生成尝试失败的情况下，也会发生生成效应。因此，就项目的表面特征和语义特征的区别而言，表面特征的自生并不是效果所必需的。事实上，Jacoby(1983)已经证明，在一项感知认同的测试中，表面特征比语义特征更重要，自我生成实际上是不利的。他发现，与阅读过的单词相比，自我生成的单词更不容易被识别。Slamecka和Fevreiski(1983)对规则激活假说的含义进行了有益的总结。他们说732
John M. Gardiner and James A. Hampton The City University, London, England A word from a list is more likely to be recalled if it was self-generated rather than read. This finding--the generation effect--has been attributed to semantic memory and, more particularly, to the enhanced activation of the semantic features comprising a word's representation in the subjective lexicon. Three experiments showed that a similar generation effect occurs for meaningful but not for meaningless letter bigrams (e.g., E T vs. E C); for unitized but not for nonunitized 2-digit numbers (e.g., 28 vs. 2, 8); and for familiar but not for unfamiliar noun compounds (e.g., cheesecake vs. cheese ketchup). These results indicate that semantic memory involvement is critical to the generation effect only in that the item to .be recalled must form some integrated functional unit. Representation in the subjective lexicon is neither a necessary nor a sufficient condition for the effect to emerge. During recent years there has been renewed interest in how a person's memory is influ- enced by having self-generated the items to be remembered. Much of this interest has stemmed from, and focused on, the finding of superior memory for words presented in a generate as opposed to a read task--the gen- eration effect as delineated by Slamecka and Graf (1978; see also Jacoby, 1978). In Sla- mecka and Graf's generate task, subjects were given a stimulus word, a rule specifying some relation between it and the target response, and the initial letter of the response word. The read task differed only in that the response word was shown in full. Their study showed that a large and similar generation effect occurred with a number of different sorts of rule; that the effect occurred in recognition memory, cued-recall, and free-recall learning; and that the effect did not extend to the stimulus word--it was restricted to memory for the re- sponse word. Subsequent research indicated that the gen- eration effect is sensitive to the lexical status of the item to be remembered, for McElroy The authors contributed equally to the research reported in this article. We are greatly indebted to Frances Guy for experimental assistance. Requests for reprints should be sent to John M. Gardiner, Psychology Division, Department of Social Science and Humanities, The City University, Northampton Square, London, EC1V OHBv England. Copies of the experimental materials are also available from the authors on request. and Slamecka (1982) showed that there was no generation effect in memory for pronounc- able nonwords generated (or read) with a letter transposition rule. On the strength of this finding they argued against simple arousal or effort interpretations of the generation effect, and instead proposed that semantic memory involvement was a necessary condition for the generation effect to emerge. More particularly, they suggested an interpretation in terms of lexical activation, according to which the gen- erate task leads to an enhanced activation of semantic features, and a consequent increase in the likelihood of gaining access to the trace. We refer to this as the lexical activation hy- pothesis. Slamecka and Fevreiski (1983) lent a further refinement to this view by showing that a gen- eration effect occurs even when the self-gen- eration attempt fails. Thus, with respect to the distinction between the surface features and the semantic features of the item, self-gener- ation of the surface features is not essential to the effect. Indeed, Jacoby (1983) has shown that in a test of perceptual identification, where surface rather than semantic features are at a premium, self-generation is actually disadvan- tageous. He found that self-generated words were less likely to be identified than words which had been read. Slamecka and Fevreiski (1983) provided a useful summary of the implications of the lex- ical activation hypothesis. They remarked that 732
人类的学习和记忆。
影响因子：
24.8
作者：
Johnson,MK;Hasher,L
通讯作者：
Hasher,L