Classification of Verbal Charateristics Based on their Implication Force by Means of Cluster‐ and Factor Analysis

Abstract

One of the first linguists who carried out systematic analysis of the use of multivariate methods for linguistic data was J. Tuldava, whose works are devoted both to working out a general approach and methods of how to use this kind of analysis for linguistic data, and to possible ways of interpreting the results ‐ groups, or clusters, of objects and charateristics (Andreev & Tuldava, 1989; Liiv & Tuldava, 1987; Tuldava, 1976,1981). Multivariate statistical analysis is aimed at “compressing”; the information, reflecting the structure of an object, when this information is “dispersed”; in correlation tables, being presented by a large number of correlations which are too numerous to observe. Such lack of explicitness often creates serious difficulties in attempts to reveal systematic tendencies in language. This article deals with the problem of classifying a number of formal characteristics of the English verb, based on their implication (“diagnostic”) value. The general approach and some concrete methods proposed by J. Tuldava in his works on cluster and factor analysis of linguistic data, are used. The investigation is based on 8375 verbs ‐ a complete list of English verbs included in the seventh edition of the Concise Oxford Dictionary of Current English, (1987). The characteristics under investigation are qualitative and describe phonetic, morphemic, derivational and syntactic aspects of the verbs. All these characteristics can be divided into those which are manifested within the word base (“internal”) and those whose manifestation takes place outside the word base (“external”). The former include phonetic and morphemic, the latter derivational and syntactic characteristics.

Notes

Address correspondence to: S. Andreev, 13–3, ul. Bakunina, 214000 Smolensk, Russia.