Gregory L. Murphy

The role of theories in conceptual coherence

The question of what makes a concept coherent (what makes its members form a comprehensible class) has received a variety of answers. In this article we review accounts based on similarity, feature correlations, and various theories of categorization. We find that each theory provides an inadequate account of conceptual coherence (or no account at all) because none provides enough constraints on possible concepts. We propose that concepts are coherent to the extent that they fit peoples background knowledge or naive theories about the world. These theories help to relate the concepts in a domain and to structure the attributes that are internal to a concept. Evidence of the influence of theories on various conceptual tasks is presented, and the possible importance of theories in cognitive development is discussed.

Comprehending Complex Concepts

Abstract Recent theories of concepts have raised the issue of how people combine simple concepts (like engine and repair ) to form complex concepts (like engine repair ). This article approaches this issue by asking how people comprehend modified noun phrases of this sort. One explanation of how complex concepts are understood (the feature weighting model) provides a simple mechanism in which the primary feature of the modifying concept is made more salient in the modified concept. nother explanation focuses on how world knowledge directs the combination process. The two explanations are compared in their ability to account for the interpretation of various kinds of noun phrases. Two experiments are reported which evaluate the feature weighting models predictions for adjective-noun phrases. These contrasts suggest that the combination process does require reference to world knowledge. The consequences of accepting such an account are discussed.

On metaphoric representation

The article discusses claims that conceptual structure is in some part metaphorical, as identified by verbal metaphors like LOVE IS A JOURNEY. Two main interpretations of this view are discussed. In the first, a target domain is not explicitly represented but is instead understood through reference to a different domain. For example, rather than a detailed concept of love per se, one could make reference to the concept of a journey. In the second interpretation, there is a separate representation of love, but the content of that representation is influenced by the metaphor such that the love concept takes on the same structure as the journey concept. It is argued that the first interpretation is not fully coherent. The second interpretation is a possible theory of mental representation, but the article raises a number of empirical and theoretical problems for it. It is concluded that many of the data cited as evidence for metaphoric representations can be accounted for by structural similarity between domains.

Food for Thought: Cross-Classification and Category Organization in a Complex Real-World Domain

Seven studies examined how people represent, access, and make inferences about a rich real-world category domain, foods. The representation of the category was assessed by category generation, category ratings, and item sortings. The first results indicated that the high-level category of foods was organized simultaneously by taxonomic categories for the kind of food (e.g., vegetables, meats) and script categories for the situations in which foods are eaten (e.g., breakfast foods, snacks). Sortings were dominated by the taxonomic categories, but the script categories also had an influence. The access of the categories was examined both by a similarity rating task, with and without the category labels, and by a speeded priming experiment. In both studies, the script categories showed less access than the taxonomic categories, but more than novel ad hoc categories, suggesting some intermediate level of access. Two studies on induction found that both types of categories could be used to make a wide range of inferences about food properties, but that they were differentially useful for different kinds of inferences. The results give a detailed picture of the use of cross-classification in a complex domain, demonstrating that multiple categories and ways of categorizing can be used in a single domain at one time.

Audience Design in Meaning and Reference

We argue that the speaker designs each utterance for specific listeners, and they, in turn, make essential use of this fact in understanding that utterance. We call this property of utterances audience design. Often listeners can come to a unique interpretation for an utterance only if they assume that the speaker designed it just so that they could come to that interpretation uniquely. We illustrate reasoning from audience design in the understanding of definite reference, anaphora, and word meaning, and we offer evidence that listeners actually reason this way. We conclude that audience design must play a central role in any adequate theory of understanding.

Basic level superiority in picture categorization

In a seminal paper, E. Rosch, C. B. Mervis, W. D. Gray, D. M. Johnson, and P. Boyes-Braem ( Cognitive Psychology , 1976, 8 , 382–439 ) found that an object can be categorized faster at the basic level (e.g., hammer ) than at either a subordinate ( club hammer ) or a superordinate level ( tool ); they attributed this result to basic categories having more distinctive attributes. But numerous factors other than the number of distinctive attributes might have caused this result; for example, basic categories routinely have shorter and more frequent names than do subordinates, and are typically learned earlier and occur more often than either subordinate or superordinate categories. In this paper, we report three experiments, all of which used artificial subordinate, basic, and superordinate categories, and all of which either held constant or systematically varied several of these “other” factors. All three studies replicated the finding that objects can be categorized fastest at the basic level (but the relative speeds of subordinate and superordinate categorizations differed from past results); and all three strongly supported the claim that distinctive attributes are the factor underlying the results, though it appears that only perceptual attributes are critical.

Category differentiation in object recognition: Typicality constraints on the basic category advantage.

When people are asked to decide whether an object is in a given category, they generally respond faster when the category is at the basic level (e.g., car) than when it is at the superordinate level (e.g., vehicle) or the subordinate level (e.g., sedan). Basic categories have shorter and more frequent names, are learned earlier, and are usually more highly differentiated than other categories (they are both specific and distinctive), but it is not clear which of these factors is responsible for the faster response to basic categories. In three experiments with natural language categories, we found evidence that objects can be identified fastest as members of differentiated categories, even when such categories have longer names and are learned later than less differentiated categories. Specifically, we argued that atypical subordinate categories (e.g., racing car) are highly differentiated and should therefore be responded to as fast as basic categories in object recognition. The results supported this view and also ruled out the hypothesis that objects are necessarily identified as members of basic categories before further identification. We discuss the implications of these findings for the use of category names as definite descriptions in discourse.

Noun phrase interpretation and conceptual combination

Abstract Four experiments investigated the process by which people understand adjective-noun and noun-noun phrases in order to evaluate competing models of concept representation and conceptual combination. In three experiments, subjects judged whether noun phrases (NPs) were sensible. The results showed that when modifiers were conceptually complex (nouns and nonpredicating adjectives), the NPs took longer to interpret than when simpler modifiers (predicating adjectives) were used. Also, when an adjective modified part of a nouns schema, it was understood more quickly than when it modified nonschematic aspects of the noun. The results were interpreted as supporting a schema-modification view of comprehending NPs. A final experiment investigated this view by measuring the reading times of sentences containing NPs. The results showed that when the context activated relevant conceptual structures, all NPs were equally easy to comprehend, as predicted by the schema modification view. Without a helpful context, noun-noun phrases were considerably slower than adjective-noun phrases.

The locus of knowledge effects in concept learning.

Three experiments investigated how knowledge influences concept formation and representation in a standard concept acquisition task. The primary comparison was among arbitrary concepts, which had meaningless features; meaningful concepts, which had meaningful features from different domains; and integrated concepts, which had meaningful features interconnected by common knowledge. Experiment 1 found that learning was superior for the integrated concepts but that there was little difference as a function of feature meaningfulness. Experiment 2 suggested that the integrated Ss were learning to form a knowledge-based schema as their concept representation because they did not distinguish the typicality of features that differed in frequency. Experiment 3 introduced a category whose features were from the same domain but were not otherwise related. This concept was as difficult to learn and use as the meaningful concepts were. These comparisons help specify the ways in which knowledge does and does not influence concept formation.

The Representation of Polysemy: MEG Evidence

Most words in natural language are polysemous, that is, they can be used in more than one way. For example, paper can be used to refer to a substance made out of wood pulp or to a daily publication printed on that substance. Although virtually every sentence contains polysemy, there is little agreement as to how polysemy is represented in the mental lexicon. Do different uses of polysemous words involve access to a single representation or do our minds store distinct representations for each different sense? Here we investigated priming between senses with a combination of behavioral and magnetoencephalographic measures in order to test whether different senses of the same word involve identity or mere formal and semantic similarity. Our results show that polysemy effects are clearly distinct from similarity effects bilaterally. In the left hemisphere, sense-relatedness elicited shorter latencies of the M350 source, which has been hypothesized to index lexical activation. Concurrent activity in the right hemisphere, on the other hand, peaked later for sense-related than for unrelated target stimuli, suggesting competition between related senses. The obtained pattern of results supports models in which the representation of polysemy involves both representational identity and difference: Related senses connect to same abstract lexical representation, but are distinctly listed within that representation.