Representation, Memory, and Development

Essays in Honor of Jean Mandler

Nancy L. Stein, Patricia J. Bauer, Mitchell Rabinowitz, George Mandler

• Publisher: Psychology Press
• Year: 2014
• ISBN:
  9781135635893
• Edition: 1
• Topic: Psychology
• Subtopic: History & Theory in Psychology
• Index: Psychology

A festschrift to honor Jean Mandler, this volume contains contributions from leading scholars focusing on the child's development of memory, visual representation, and language. It is appropriate for students and researchers in cognitive psychology, language acquisition, and memory.

1

Being There Conceptually: Simulating Categories in Preparation for Situated Action

Lawrence W. Barsalou

Emory University

A constant theme in Jean Mandler’s work is that a child’s developing knowledge is grounded in sensory—motor experiences of events (e.g., Mandler, 1987, 1992). Rather than being detached from events, knowledge remains grounded in them. Rather than being amodal, knowledge retains its sensory—motor origins. The essay to follow arises in the tradition of this work and reflects its influence.

THE SITUATED VIEW OF CONCEPTS

According to the view developed here, people conceptualize a category differently across situations, with each conceptualization embedded in a background situation. A single situation—independent concept does not represent the category; the concept does not represent the category in isolation, independently of the situations in which it occurs. Consider the category of chairs. According to the situated view, different conceptualizations of chairs are represented in their respective situations. Thus, one situated conceptualization might represent office chairs in business environments, another might represent easy chairs in homes, another might represent theater chairs in theaters, another might represent airline chairs in jets, and so forth. A single situation—independent concept does not represent chairs across situations, and the conceptualizations do not represent isolated chairs. As the category is encountered in different situations, a situated conceptualization develops for each, linked together in a radial concept, as described later.

EVIDENCE FOR THE IMPORTANCE OF SITUATIONS

Findings across diverse areas demonstrate the importance of situations in intelligence and behavior. In developmental psychology, the Vygotskian approach has stressed the importance of situations in acquiring cognitive and social skills (e.g., Vygotsky, 1991). From this perspective, Jean Mandler illustrated the importance of situations in children’s ability to remember stories and events (e.g., Mandler & Johnson, 1977; Mandler, 1987). In the social and personality literatures, situations predict behavior at least as well as traits (e.g., Mischel, 1968). In perception, situations greatly facilitate object recognition when an object occurs in a predicted context (e.g., Biederman, 1981), with Jean Mandler and her collaborators providing some of the earliest demonstrations (e.g., Mandler & Parker, 1976; Mandler & Ritchey, 1977; Mandler & Stein, 1974). In memory, situations play a central role in elaborating perceived scenes (e.g., Intraub, Gottesman, & Bills, 1998) and in retrieving information from memory (e.g., Tulving & Thomson, 1973). In language comprehension, texts can be incomprehensible when the relevant situation is not known (e.g., Bransford & Johnson, 1973). Indeed, language comprehension can be viewed as preparation for situated action (Barsalou, 1999a). In pragmatics, situations are central to establishing common ground between communicators, both human (e.g., Clark, 1992) and nonhuman (e.g., Smith, 1977). In problem solving and reasoning, it may be difficult to draw valid and useful conclusions without the support of a concrete situation (e.g., Cheng & Holyoak, 1985; Gick & Holyoak, 1980; Johnson-Laird, 1983). In linguistics, the importance of situations has motivated the theory of construction grammar, where syntactic structures evolve out of familiar situations (e.g., Goldberg, 1995). In philosophy, the importance of situations has motivated the theory of situation semantics, where logical inference is optimized when performed in the context of specific situations (e.g., Barwise & Perry, 1983). At a more general level, arguments about the central role of situations in cognition can be found in Barsalou, Yeh, Luka, Olseth, Mix, and Wu (1993), Clark (1997), Glenberg (1997), and Greeno (1998).

Across these diverse areas, the common theme is that situations are fundamental to cognition. By incorporating situations into a cognitive task, processing becomes more tractable than when situations are ignored. Because specific entities and events tend to occur in some situations more than others, capitalizing on these correlations constrains and facilitates processing. Knowing the current situation constrains the entities and events likely to occur. Conversely, knowing the current entities and events constrains the situation likely to be unfolding.

By focusing on situations, the cognitive system simplifies many tasks. Rather than having to search everything in memory across all situations, the cognitive system focuses on the knowledge and skills relevant in the current situation. As a result, it becomes easier to recognize objects and events that may be present; it becomes easier to remember relevant information and skills; it becomes easier to resolve the ambiguities of language; it becomes easier to solve problems and perform reasoning; it becomes easier to predict the actions of other agents. For all these reasons, it would not be surprising if situations turned out to be central for concepts.

CURRENT THEORIES OF CONCEPTS

Most current views implicitly view concepts as unsituated, assuming that concepts have been abstracted from the situations in which they occur. Although these theories could be readily extended to represent situated concepts, they typically do not.

Consider classical theories, which typically assume that rules describe the objects in a category independently of situations (e.g., Bruner, Goodnow, & Austin, 1956). For example, a rule might attempt to capture the physical properties of chairs that are necessary and sufficient for membership. Although such rules could also attempt to capture situational properties, they typically do not. Instead, classical theories abstract across situations, rather than establishing rules for subsets of chairs within particular situations. Classical theories to date represent the extreme view of unsituated concepts.

Prototype theories similarly tend to assume that unsituated abstractions represent categories (e.g., Hampton, 1979; Rosch & Mervis, 1975). Rather than being definitional, however, these abstractions are statistical, representing the most frequent properties across situations, with situation-specific properties canceling themselves out. Although subprototypes could develop for concepts in particular situations, this possibility has not been explored.

The view that categories are embedded in intuitive theories similarly tends to ignore situations (e.g., Murphy & Medin, 1985). Although intuitive theories constrain the form that concepts take, situations have not played a central role in these accounts. Even when concepts are constrained by intuitive theories, they are nevertheless assumed to remain constant across situations (but see Gelman & Diesendruck, 1999, for an account of intuitive theories that is compatible with situated concepts).

Exemplar models have much potential for implementing situated concepts but typically have not. Many exemplar models assume that the entire set of exemplars stored in memory for a category represents it on each occasion (e.g., Lamberts, 1994; Nosofsky, 1984). Thus, a fixed representation stands for the category on all occasions, not a situation-specific one. Furthermore, exemplar representations typically only include physical properties of exemplars, not properties of associated situations. Notably, some exemplar models have more of a situated character. For example, Nosofsky and Palermi’s (1997) random-walk model assumes that only a subset of exemplars is retrieved in the current context, such that the category representation changes across contexts (also see Barsalou, Huttenlocher, & Lamberts, 1998). Similarly, Medin and Schaffer’s (1978) context model assumes that context is important in categorization, although context is typically implemented as other properties in the objects being categorized, not as their situational properties. As these theories illustrate, exemplar models can implement situated conceptualization if two conditions are met: (a) situational information is represented in exemplars along with object properties, (b) situation-specific subsets of exemplars are retrieved during categorization.

In contrast to the previous four classes of models, connectionist models clearly implement situated conceptualization. Not only do they represent a category differently across situations, they include situational information in these representations. Consider Rumelhart, Smolensky, McClelland, and Hinton’s (1986) account of the room schema. In an auto-associative net, subsets of object properties are linked to subsets of room properties, such that correlated sets of object and situational properties form attractors. When a subset of object properties is activated, related situational properties become active, thereby situating the object. Conversely, when situational properties become active, relevant properties of the object become active, resulting in a situation-specific representation of it. Connectionist models have been explicitly formulated to implement situated conceptualization, and they do so elegantly.

CONCEPTS AS GROUNDED IN PERCEPTUAL SIMULATION

The importance of situations for concepts follows from the proposal that people represent concepts with perceptual simulations (Barsalou, 1999b). This next section briefly outlines the theoretical assumptions of perceptual symbol systems, and then reviews some of the empirical evidence for this approach. The following section then illustrates how viewing concepts as grounded in perceptual simulation predicts the importance of situations in concepts.

The first assumption of this view is that selective attention focuses on components of experience. During perception of sensory events, people focus on shapes, colors, sounds, smells, etc.; during perception of proprioceptive events, people focus on movements, facial expressions, vocalizations, etc.; during perception of introspective states, people focus on emotions, cognitive operations, beliefs, etc. Once attention selects a perceived aspect of experience, associative areas in the brain capture the respective pattern of activation in the relevant perceptual, proprioceptive, or introspective area. Later, these associative areas partially reactivate these perceptual representations in the absence of perceptual input, thereby simulating the experience of what an external or internal event was like. Using such simulations, people conceptualize objects, external events, and internal events in their absence.

Barsalou (1999b) illustrated how these simulation mechanisms implement a fully-functional conceptual system, including the type-token distinction, categorical inference, the productive construction of novel simulations, the representation of propositions, and the representation of abstract concepts. Also illustrated are how these simulation mechanisms could underlie the knowledge that supports basic cognitive processes, including perception, categorization, memory, language, and thought. Additional articles extend this theory (Barsalou, 1993, 1999a; Barsalou & Prinz, 1997; Prinz & Barsalou, 2000; Prinz & Barsalou, in press). In one of the earli...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Table of Contents
• Dedication
• Preface
• 1 Being There Conceptually: Simulating Categories in Preparation for Situated Action
• 2 Building Toward a Past: Construction of a Reliable Long-Term Recall Memory System
• 3 The Origin of Concepts: Continuing the Conversation
• 4 Scripts, Schemas, and Memory of Trauma
• 5 On Animates and Other Worldly Things
• 6 How to Build a Baby That Develops Atypically
• 7 Pretense and Representation Revisited
• 8 Early Concepts and Early Language Acquisition: What Does Similarity Have to do With Either?
• 9 A Stitch in Time: The Fabric and Context of Events
• 10 The Reemergence of Function
• 11 The Procedural-Procedural Knowledge Distinction
• 12 Spatial Language: Perceptual Constraints and Linguistic Variation
• 13 Conceptual Development in Infancy: The Case of Containment
• 14 Memories for Emotional, Stressful, and Traumatic Events
• Author Index
• Subject Index