Labeling Bias and Categorical Induction: Generative Aspects of Category Information.

Abstract

When a person is characterized categorically with a label (e.g., Linda is a feminist), people tend to think that the attributes associated with that person are central and long lasting (S. Gelman & G. D. Heyman, 1999). This bias, which is related to category-based induction and stereotyping, has been thought to arise because a category label (e.g., feminist) activates the dominant properties associated with the representation of the category. This explanation implies that categorical information influences inferential processes mainly by conjuring up main attributes or instances represented in the category. However, the present experiments reveal that this attribute-based explanation of induction does not provide a complete picture of inferential processes. The results from 3 experiments suggest that category information can affect inferences of attributes that are not directly related to the category, suggesting that categories not only activate likely attributes but also help integrate unlikely or even unrelated attributes.One of the core purposes of categories is to integrate information and apply it to future predictions (Anderson, 1990;Rosch, 1978;E. E. Smith, 1994). For example, one forms the category "liberal" by classifying people who have similar political opinions and then uses the category to make an inference, such as predicting the policy that a member of the category advocates. In this regard, categories help bind information and serve as a medium for inferences.How do categories integrate diverse information and license inductive inferences? Main psychological accounts of inductive judgments suggest that two factors-(a) matching attributes between entities and (b) the coverage of premises over a conclusion (hereafter I call these two factors attribute-based similarity)-are the guiding forces of inferences (see Heit, 2000, for a review; Kunda & Thagard, 1996;Osherson, Smith, Wilkie, Lopez, & Shafir, 1990;Osherson, Stern, Wilke, Stob, & Smith, 1991;Rips, 1975;Sloman, 1993;E. R. Smith & Zarate, 1992;Tversky, 1977). According to this view, for example, the conclusion Lions have Disease X, given a premise, Zebras have Disease X, is credible proportional to the extent to which the two entities-zebras and lions-have features in common and the magnitude of an entity in a premise (i.e., zebras) exceeds that of a conclusion (lions; i.e., coverage).This similarity-based account of categorical induction has inspired many areas of cognitive psychology and social cognition and has yielded a rich array of empirical findings that illuminate the relationship between feature-based similarity and category membership

The Semantic Web and Human Inference: A Lesson from Cognitive Science

“…One source of this bias seems to come from the need to maintain cognitive economy [5]. One of the most compelling findings in the cognitive psychology of inductive inference is that people make a predictive inference on the basis of the categorical information that is immediately recognizable [6]- [10]. By arranging concepts categorically, we treat individual objects as a group and deal with the characteristics of the group as a whole, rather than individual objects separately.…”

Linking Syntax and Inductive Reasoning: Categorical Labeling and Generic Noun Phrases

“…Previous studies showed that categorical labeling such as in (9) bolsters the estimation of highly likely attributes (e.g., "KINATE contains no fat." see Gelman & Heyman, 1999;Walton & Banaji, 2004;Yamauchi, 2005). The main explanation suggested for this "labeling effect" is that categorical statements activate the underlying "content" of the category.…”

“…The following examples help illustrate the design of the experiment (see Gelman & Heyman, 1999;Walton & Banaji, 2004;Yamauchi, 2005; for a similar procedure).…”

Category labels versus feature labels: Category labels polarize inferential predictions

“…We think that category labels play a guiding role in feature inference (E. M. Markman, 1989;Yamauchi, 2005;Yamauchi et al, 2007;Yamauchi & Markman, 2000) and promote a reasoning strategy using abstract rules (see Sloman, 1996, andSloman, 1994, for the distinction between rule-based and similarity-based reasoning strategies). First, participants' inferential projections should depend primarily on the matching/mismatching status of the labels when the two labels are characterized as the names of categories.…”

“…These sentences roughly mean the same thing, but they evoke different kinds of implications. Gelman and Heyman (1999) suggested that categorical noun labels (e.g., "poem" in Sentence 1), unlike feature labels (e.g., "poetical" in Sentence 2), evoke a sense of immutability, endurance, and centrality linked to the identity of an object, even in 5-year-old children (see also Walton & Banaji, 2004;Yamauchi, 2005).…”