Developmental Psychology Copyright 1990 by the American Psycholol~ical Asscx~tien, Inc. 1990, Vol. 26, No. 3, 488-493 0012-1649/90/$00.75 Semantic Familiarity, Relevance, and the Development of Deductive Reasoning Shawn L. Ward and Willis E Overton Temple University This study examined the developmental progression of deductive reasoning between Grades 6 and 12 and the role of semantic content relevance in the solution oftbe Wason selection task. A rating procedure was used to establish the degree of relevance between antecedent and consequent clauses of conditional (if .... then .... ) propositions. Results were consistent with the general position that formal deductive reasoning becomes available in adolescence and that relevant propositional content is not a sufficient condition for adequate deductive reasoning performance. High-relevant content enhanced level of performance for those having deductive reasoning competence, and low- relevant content failed to support adequate performance. Findings are interpreted within the distinc- tion between the use of the inference rule competence and the facilitation of relevant semantic con- tent. A considerable body of contemporary research has demon- strated that the ability to reason deductively undergoes a trans- formation and a well-defined developmental progression be- tween the ages of approximately 10-11 years and 17-18 years (Bady, 1979; Bucci, 1978; Byrnes & Overton, 1986, 1988; Moshman, 1979; Moshman & Franks, 1986; O'Brien & Over- ton, 1980, 1982; Overton, Byrnes, & O'Brien, 1985; Overton, Ward, Noveck, Black, & O'Brien, 1987; Pollack, Ward, & Over- ton, 1988). This same body of research, as well as additional studies (Clement & Falmagne, 1986; Franco & Overton, 1984; Markovits, 1986; Overton, Yaure, & Ward, 1986), supports the position that prior to adolescence, formal deductive reasoning competence is largely unavailable, whereas by late adolescence, individuals demonstrate a high level of competence in solving deductive reasoning problems. These findings do not, however, suggest that available deductive reasoning competence is mani- fested in all contexts. In fact, earlier research (Gilhooly & Fal- coner, 1974; Manktelow & Evans, 1979; Wason, 1968), which reported poor reasoning in adults, provided the starting point for empirical explorations of possible factors that moderate the underlying competence that is available to adults (Wason, 1983). One factor that has proven to be a moderator of successful performance is the semantic content of the reasoning problems. In early research (Bracewell & Hidi, 1974; Van Duyne, 1974), this variable was referred to as thematic or realistic content. Portions of this article were presented at the annual meetings of the Eastern Psychological Association, Alexandria, Virginia, April 1987. The research was based on a portion of a dissertation submitted by Shawn L. Ward to Temple University in partial fulfillment of the re- quirements for the PhD degree. Correspondence concerning this article should be addressed to either Shawn L. Ward, who is now at the Department of Psychology, Le Moyne College, Syracuse, New York 13214, or Willis E Overton, De- partment of Psychology, Temple University, Philadelphia, Pennsylvania 19122. 488 Early findings in this area (Wason & Shapiro, 1971; Yachanin & Tweney, 1982) have suggested that when thematic content was used, performance by adults was enhanced. However, later research failed to replicate---or demonstrated only weak sup- port for--this effect (Wason, 1983). Following these efforts, sev- eral explanations have been offered to account for this elusive thematic content effect (Cbeng & Holyoak, 1985; Griggs, 1983; Overton et al., 1987). Griggs (1983) suggested that it is the familiar rather than the realistic nature of the content that enhances performance, and this suggestion has been supported in several studies (Griggs & Cox, 1982, 1983). Cbeng and Holyoak (1985; Cbeng, Holyoak, Nisbett, & Oliver, 1986) have claimed that direct real-world ex- periences account for the findings. Experiences are represented as inductions from pragmatic activities such as permissions or causations. The inductions form generalized knowledge struc- tures called pragmatic reasoningschema. When similarly struc- tured material is presented at a later time, the schema is elicited, and a richer source of examples is provided from which to draw inferences. Recently, Overton et al. (1987) reexamined the role of seman- tic content in deductive reasoning in a series of developmental experiments that assessed performance on a modified version of the Wason selection task. In this task, individuals are given a conditional rule and are required to select conditions that would establish the validity or nonvalidity oftbe rule. The Ov- erton et al. (1987) experiments demonstrated that regardless of the specific nature of the content, a developmental progression occurs between 4th and 12th grades and deductive competence is largely unavailable before 8th grade. However, familiar se- mantic content also significantly enhanced performance. The conclusion drawn by these authors was that both logical knowl- edge (i.e., formal reasoning competence) and world knowledge (i.e., familiar as distinct from abstract semantic content) are necessary but distinct features of adequate logical reasoning performance. Despite the fact that Griggs (1983), Cheng and Holyoak