Abstract

Analyzing concepts expressed by words has long been a method favored by philosophers for gaining insight into the structure of thought. Recently it has been suggested that the prevalence of seemingly unavoidable counterexamples to almost all of the conceptual analyses so far proposed indicates that the concepts selected for analysis have been too general and that concepts as specific as the psychologically real senses of words should be selected. Unfortunately there is no consensus among linguists as to the specific criteria and tests that should be used for determining which word senses are psychologically real. ;The methods I have followed in seeking to identify appropriate criteria and tests for individuating psychologically real word senses include: examining psycholinguistic studies of semantic aspects of the mental lexicon; evaluating the individuation criteria and tests proposed by linguists and examples of their application to semantic analyses of particular words in the light of the results of ; analyzing the principles and practices of word-sense individuation in an unabridged English dictionary, on the assumption that those principles and practices can shed light on word-sense individuation in the mental lexicon inasmuch as they are end products of a centuries-long evolution toward optimal means of enabling people to find what they want to know about the senses of particular words. ;I found that the psycholinguistic experiments purporting to show which psychologically real senses a particular word has or even how many such senses it has are too seriously flawed to shed much light on either question. However, psycholinguistic experimentation on other semantic aspects of the mental lexicon has yielded reliable results which, in conjunction with the results of , indicate that none of the tests so far proposed for determining which senses of a word are psychologically real is wholly reliable. Applying a collection of tests to a given word often yields an intuitively satisfying individuation. Testing in this way the dictionary analyzed in suggests that its individuations often approximate psychologically real ones