The similarity and diversity of semantic relations

Abstract

There is a rich variety of semantic relations in natural languages. Subjects' perceptions of similarities among relations were studied for a wider variety of relations than had been used in previous studies. Forty subjects sorted 31 cards bearing five example pairs of each of 31 semantic relations. Subjects were able both to distinguish the relations and to perceive their similarities. A hierarchical clustering analysis of the sorting data indicated that the subjects perceived five families of semantic relations (contrasts, class inclusion, similars, case relations, and part-wholes). The five families were distinguished in terms of three properties of semantic relations: contrasting/noncontrasting, logical/pragmatic, and inclusionlnoninclusion. Within each family, relations also were sorted in ways consistent with their defining properties. Relations were therefore viewed not as unanalyzable primitives, but in terms of the relational properties that distinguished them.Semantic relations between concepts are basic components of language and thought (Bierwisch, 1970; H. H. Clark, 1970;Evanechko & Maguire, 1972;Evens, Litowitz, Markowitz, Smith, & Werner, 1980;Katz, 1972;Leech, 1974;Miller, 1969; Miller & JohnsonLaird, 1976;Osgood & Richards, 1973;Palermo, 1978).Relying on associative and introspective data, Wundt (1893) identified and classified a wide range of relations. Wundt recognized two general classes: outer associations (part-wholes, word sequences) and inner associations (categorical relationships, similars, coordination, causal relations). At least 13 other classification systems were proposed prior to 1911 (Warren, 1921).More recently, researchers have been directly concerned with the study of semantic relations in two related areas: the development of general models of semantic memory (Anderson, 1976;Norman & Rumelhart, 1975; Schank, 1973) and the study of semantic decisions (Cohen, 1977;Danks & Glucksberg, 1980;Kintsch, 1980;Smith, 1977). In both areas, research has focused on a limited number of relations, and the similarities between relationships have not been explored. General models of semantic memory have been developed to account for the comprehension and representation of propositional knowledge (Anderson, 1976;Norman & Rumelhart, 1975; Schank, 1973). In these models, information is represented by a network oflabeled relaPreparation of this article was supported in part by a faculty research grant from Trenton State College to the first author. We thank Mary Crawford for comments on an earlier version of this paper. Requests for reprints should be sent to Roger Chaffin, Department of Psychology, Trenton State College, Trenton, NJ 08625. tions between nodes that stand for concepts. Relations serve as unanalyzed, primitive terms; consequently, their number has been restricted in the interest of economy. As a result, only a fraction of the total number of relations in the English language have been considered. The scope and power of current models would be enhanced by encompassing the wid...

Focaused search of semantic cases: The effects of question form and case status

“…Adults can make comparisons among semantic cases in a manner similar to the way they compare ordinary taxonomic categories (Chaffin & Herrmann, 1984;Shafto, 1973). The language of young children reflects their growing awareness of different semantic cases (Bowerman, 1973).…”

The Subjectivity of Lexical Cohesion in Text

“…In psychology, the focus has been mostly on classical relations; however, there have been recent calls to broaden the focus and include non-classical relations as well (McRae and Boisvert, 1998;Hodgson, 1991). Some researchers have always included some non-classical relations, such as Evens et al (1983), Chaffin and Herrmann (1984), and researchers in Library and Information Science. However, as stated earlier, the research on lexical semantic relations has been done out of the context of text, and then assumed to be relevant within it, and in lexical cohesion research, the analysis of lexical semantic relations was done by experienced linguists with particular points of view.…”

Relation similarity as a function of agreement between relation elements

“…The purpose of the present research was to ascertain whether perceptions of relation similarity originate in a comparison of relational elements similar to those used in linguistics to define relations (Leech, 1974;Lyons, 1968Lyons, , 1977. If it could be shown that perceptions of relation similarity vary with the overlap in relation definitions, then the theory that the comprehension of relations involves processing definitional elements of relations would gain further support (Chaffin & Herrmann, 1984a, 1984bJohnson-Laird, Herrmann, & Chaffin, 1984; see also Klix, Hoffman, & van der Meer, 1983).…”

“…A recent study (Chaffin & Herrmann, 1984b) showed that relation similarity varies systematically across relations. The study required subjects to sort word pairs that represented 31 relations into groups of similar relations.…”

“…The three ways that relation similarity varied in Chaffin and Herrmann's (1984b) study were explained as being due to agreement of elements that define relations. For example, subjects frequently sorted pairs such as alive-dead with those such as hot-cold because both pairs shared the antonym elements of (1) a common dimension on which word meanings are represented (2) bilaterally and (3) symmetrically about the dimension's midpoint; however, neither pair was sorted with synonyms (carauto), since synonymity is not defined by a common dimension with symmetric opposition.…”