Maki William S, Buchanan Erin
Department of Psychology, Texas Tech University, Lubbock, Texas 79409, USA.
Psychon Bull Rev. 2008 Jun;15(3):598-603. doi: 10.3758/pbr.15.3.598.
There has been much debate about the relation between knowledge for meaning (semantic memory) and knowledge for words in context (associative memory). Many measures of that knowledge exist, but do they all measure the same thing? In this study, scaling, clustering, and factor-analytic techniques were used to reveal the structure underlying 13 variables. Semantic similarity determined from lexicographic measures is shown to be separable from the associative strength determined from word association norms, and these semantic and associative measures are in turn separable from abstract representations derived from computational analyses of large bodies of text. The three-factor structure is at odds with traditional views of word knowledge. The expression of long-term knowledge about words and the concepts they represent may be better viewed in terms of associative, semantic, and thematic information.
关于意义知识（语义记忆）与语境中单词知识（联想记忆）之间的关系，一直存在诸多争论。存在许多衡量该知识的方法，但它们衡量的都是同一事物吗？在本研究中，运用了标度、聚类和因子分析技术来揭示13个变量背后的结构。由词典学方法确定的语义相似性被证明与由单词联想规范确定的联想强度是可分离的，并且这些语义和联想测量又与从大量文本的计算分析中得出的抽象表征是可分离的。这种三因素结构与传统的单词知识观点不一致。关于单词及其所代表概念的长期知识的表达，或许从联想、语义和主题信息的角度来理解会更好。
