The importance of being coherent: Category coherence, cross-classification, and reasoning☆
Section snippets
Cross-classification
Past research has identified three influences on category preference when more than one category is available. First, people are more inclined to use the category with the greatest relevance to the property in question (Heit and Rubinstein, 1994, Kalish and Gelman, 1992, Murphy and Ross, 1999, Ross and Murphy, 1999). Second, inferences are more often made from categories with increased mental activation relative to others (Macrae et al., 1995, Sinclair and Kunda, 1999, Smith et al., 1996). And
Category coherence
Category coherence refers to the extent to which category features go together in light of prior theoretical, causal, and teleological knowledge (Medin, 1989, Murphy and Medin, 1985; see Murphy, 2002, for a review) rather than being just incidentally co-occurring. “Lives in water, eats fish, has many offspring, is small” describes a more coherent category than “lives in water, eat wheat, has a flat end, is used for stabbing bugs” (Murphy & Wisniewski, 1989). It is well documented that most
Current research
The present research was guided by two major goals. The first goal was to consider the extent to which category coherence influences category use in reasoning from multiple categories. In particular, when high and low coherence categories are placed in direct competition with one another, are higher coherence categories favored over less coherent ones? The second goal was to begin to explore cognitive processes underlying differential use of high versus low coherence categories, looking at how
Pretesting of categories
In using natural categories, we must specify how the degree of coherence for particular categories is established. For example, consider the categories of “ministers” and “county clerks.” Our intuition is that, consistent with a common cause notion of coherence, ministers are associated with deep underlying traits such as belief in God, compassion for others, and satisfaction in attending to the spiritual needs of a community. These features give rise to surface behaviors such as being on a
Experiment 1: Basic inference
With materials from the pretest in hand, we were now able to develop a set of inference problems involving entities belonging to multiple categories. Specifically, problems similar in structure to those used by Nelson and Miller (1995) were developed. These problems asked participants to make property inferences about individuals belonging to two categories, with the difficulty being that the categories provided conflicting information. For six coherence problems, high and low coherence
Experiment 2: Information selection
In Experiment 2, we developed problems in which participants could select categories about which they desired inference-related information. For example, one could choose the category “waiter” and find out that “20% of waiters prefer Coca-Cola to Pepsi.” For each problem, there were two high and two low coherence categories; participants could choose as few as one category or as many as all four. The dependent measure was the number of high versus low coherence categories about which
Experiments 3a and b: Single-category explanations
One possible cognitive explanation for the results of the first two experiments is that individuals engage in explanation-based reasoning about new properties. Heit and Rubinstein, 1994, Sloman, 1994 found that people were more inclined to transfer a property from one category to another when they could generate a single coherent explanation for its presence in both categories. Sloman (1994) offered the example that “Many ex-cons are hired as bodyguards. Therefore many war veterans are hired as
Experiment 4
The materials used in this experiment were similar to those of Experiment 3a except that each problem described people who were members of two categories (one high and one low coherence category) rather than one. Participants were asked to generate three different explanations for the stated preference. At the end of the task, they were asked to go back and circle the most plausible explanation (from among the three) for each problem. We hypothesized that explanations would make reference to
Summary of results
The primary purpose of the experiments was to better understand the role of category coherence in reasoning about cross-classified entities. In pretests, we identified social categories that varied in similarity, one marker for coherence, and then provided confirmation that these categories also differed on other measures associated with coherence including entitativity (Haslam et al., 2000) and the presence of deep features (Ahn, 1998, Keil, 1989). We also found that distinctiveness and
Conclusions
Category-based induction involves not only assigning an entity to one or more categories but also deciding which of these categories to use to inform inference. Based on the experiments presented in this paper, we conclude that natural social categories vary in coherence, the coherence of social categories is an important determinant of which one or more categories are selected and used to make an inference, and category-based explanation may serve as an important mechanism for linking novel
References (52)
Why are different features central for natural kinds and artifacts? The role of causal status in determining feature centrality
Cognition
(1998)- et al.
The causal status effect in categorization: An overview
- et al.
Causal status as a determinant of feature centrality
Cognitive Psychology
(2000) The cognitive structure of social categories
Cognitive Science
(1985)- et al.
Knowledge selection in category learning
- et al.
Correlated properties in natural categories
Journal of Verbal Learning and Verbal Behavior
(1984) - et al.
Feature inference and the causal structure of categories
Cognitive Psychology
(2005) - et al.
Category coherence and category-based property induction
Cognition
(2004) - et al.
Food for thought: Cross-classification and category organization in a complex real-world domain
Cognitive Psychology
(1999) When explanations compete: The role of explanatory coherence judgments on likelihood
Cognition
(1994)
Feature centrality and conceptual coherence
Cognitive Science
Knowledge structures and linear separability: Integrating information in object and social categorization
Cognitive Psychology
Linear separability and concept learning: Context, relational properties, and concept naturalness
Cognitive Psychology
Cognitive foundations of natural history
Ad hoc categories
Memory & Cognition
Ideals, central tendency, and frequency of instantiation as a determinant of graded structure in categories
Journal of Experimental Psychology: Learning, Memory, and Cognition
Inference and classification learning of abstract coherent categories
Journal of Experimental Psychology: Learning, Memory, and Cognition
Essentialist beliefs about social categories
British Journal of Social Psychology
Properties of inductive reasoning
Psychonomic Bulletin & Review
Similarity and property effects in inductive reasoning
Journal of Experimental Psychology: Learning, Memory, and Cognition
Is the acquisition of social categories based on domain-specific competence or knowledge transfer?
Race in the making: Cognition, culture, and child’s construction of human kinds
On wooden pillows: Multiple classification and children’s category-based inductions
Child Development
Category learning with minimal prior knowledge
Journal of Experimental Psychology: Learning, Memory, and Cognition
Concepts, kinds, and cognitive development
Structural alignment in induction and similarity
Journal of Experimental Psychology: Learning, Memory, and Cognition
Cited by (37)
Revising deference: Intuitive beliefs about category structure constrain expert deference
2017, Journal of Memory and LanguageInfluence of category coherence and type of base-rate acquisition on property generalization
2017, Acta PsychologicaCitation Excerpt :These findings are consistent with Patalano et al.'s (2006) study, demonstrating high validity of the present data. Based on the pretest results, 12 category pairs were selected according to criteria of distinguishable coherence levels and equal-estimated frequency (Haslam et al., 2000, 2006; Patalano et al., 2006). Table 2 lists the experimental materials.
Category coherence in children's inductive inferences with cross-classified entities
2015, Cognitive DevelopmentCitation Excerpt :The results of this research extend previous developmental research that has found that coherence supports children's learning of single categories (e.g., Booth, 2014; Krascum & Andrews, 1998; Murphy, 2002; Nguyen et al., 2011). Our findings are consistent with the adult concepts literature, and add a developmental aspect to this body of research, lending further support for the claim that coherence affects beliefs about what makes the categories of a cross-classified concept informative for induction (Hayes et al., 2011; Patalano et al., 2006, 2009). Thus far, we have interpreted the findings of the current investigation as suggesting that coherence influences children's inductive reasoning with cross-classified entities.
Up to standard? Market positioning and performance of Hong Kong films, 1975-1997
2015, Journal of Business VenturingLatent scope bias in categorization
2014, Journal of Experimental Social PsychologyCitation Excerpt :On the surface, this finding may appear to contradict previous findings in the literature. For example, Patalano, Chin-Parker, and Ross (2006) demonstrated that people prefer to use more coherent categories in making predictions about others, and one feature of coherent categories is that they have more features. While this could seem a contradiction to the present predictions, there is a striking methodological difference in the two approaches.
- ☆
This work was supported by National Science Foundation Grant SBR 97-20304. We thank Nick Haslam and Gregory Murphy for comments on an earlier draft of this paper. Thanks also to Melissa Paulson and Jane Erickson for their assistance in data collection.