Theories and Similarity: Categorization under Speeded Conditions

Theories and Similarity: Categorization under Speeded Conditions Christian C. Luhmann (christian.luhmann@vanderbilt.edu) Department of Psychology, Vanderbilt University 301 Wilson Hall, Nashville, TN 37203 USA Woo-kyoung Ahn (woo-kyoung.ahn@vanderbilt.edu) Department of Psychology, Vanderbilt University 301 Wilson Hall, Nashville, TN 37203 USA Thomas J. Palmeri (thomas.j.palmeri@vanderbilt.edu) Department of Psychology, Vanderbilt University 301 Wilson Hall, Nashville, TN 37203 USA Abstract A largely accepted view in the categorization literature is that similarity-based reasoning is faster than theory-based reasoning. In the current study, we explored whether theory- based categorization behavior would continue to be observed when people are forced to make category decisions under time pressure. As a specific test of the theory-based view to category representation we examined the causal status hypothesis, which states that properties acting as causes are more important than properties acting as effects when categorizing an item. Subjects learned four categories of items composed of three features and learned causal relations between those features. In two experiments we found that participants gave more weight to cause features than to effect features even under rapid response conditions. We discuss implications of these findings for categorization. Introduction When posed with categorization tasks in everyday life people recruit information from a variety of sources. In general, previous work on categorization has focused on two sources of information: similarity and theories. One family of categorization theories has centered on the notion of similarity (e.g., Kruschke, 1992; Nosofsky, 1986; Smith & Medin, 1981; Rosch & Mervis, 1975). On this view concept learning and use is based on computing the similarity between an object to be categorized and a stored representation of a category (e.g., exemplars, Nosofsky, 1986; or prototypes, Hampton, 1995). An alternative view assumes that people have theories that embody relations between properties and influence categorization behavior (Carey, 1985; Keil, 1989; Murphy & Medin, 1985; Rips, 1989). An illustrative example comes from Keil’s (1989) discovery experiment. When presented with an animal that had the appearance and behavior of a horse but the insides and lineage of a cow, adults would categorize the animal as a cow. This behavior suggests that lineage has a special status above and beyond perceptual features, presumably reflecting the importance of lineage in our lay theory of biology. Similarly, Medin and Shoben (1988) showed that people would rather accept a square cantaloupe than a square basketball, presumably because “being round” is more central in naive theories of physics (i.e., the domain in which basketballs are grounded) than in naive theories of biology (i.e., the domain in which cantaloupes are grounded). The similarity-based and theory-based views are not necessarily incompatible (e.g., Sloman & Rips, 1998). In fact, many proponents of either view allow for, or even advocate, the operation of both kinds of processes (e.g. Sloman, 1996; Smith & Sloman, 1994). However, these proposals typically put the two views on unequal footing. A persistent bias present in these ‘hybrid’ models is that similarity-based categorization is primary. For instance, in the developmental literature, it has been argued that theory- based mechanisms cannot precede similarity-based mechanisms in development because theories must be acquired through similarity-based mechanisms (Quine, 1977; Vygotsky, 1962; but see Keil, Smith, Simons, & Levin, 1998). Thus, only after sufficient experience has been obtained may theories be developed and used, amending (or supplanting) similarity-based information. In addition to the idea that similarity-based categorization is developmentally primary, there is a notion that similarity- based information is accessed more rapidly, and perhaps more automatically, than theory-based information. This assumption may be motivated by the observation that novices (e.g., children) use similarity-based reasoning and thus it is a somehow simpler mode of reasoning (cf. Keil et al, 1998). Smith and Sloman (1994) make this argument explicit by assuming that theory-based reasoning is a type of rule-based reasoning, arguing that rule-based reasoning is, “more analytic and reflective than similarity-based categorization” (pp. 377-378).