Situated Cognition:
On Human Knowledge and Computer Representations
William J. Clancey
© Cambridge University Press 1997
CONTENTS
Introduction: What is situated cognition?
- Descriptions and coordinations
- On comparing human knowledge and computer representations
- Reader's guide
PART I REPRESENTATIONS AND MEMORY
- Plans, drawings, and interpretations
- Mechanisms vs. descriptions of behavior
- Aaron's program vs. what an artist knows
- The experience of an artist participating in a community
- The structural and functional aspects of "situatedness"
2. Mycin's Map
- A simple introduction to knowledge representations
- A knowledge base is a kind of map
- General vs. situation-specific models
- Searching a descriptive model
- The map isn't the territory
- Explaining rules
- Syntactic and semantic interpretation
3. Remembering Controversies
- Arguments for memory as dynamic process, against structure
storage
- How descriptive modelers talked about symbols and representations
- The psychological view of a physical symbol system
- The stored-schema view takes hold
- "Actively doing something all the time"-From Associationism to
Contextualism
- Classical memory: Stored traces and isolated modules
- What's wrong with simple connectionism?
4. Sensorimotor Maps vs. Encodings
- von Foerster and Bateson: Descriptions of information
- The owl monkey's map
- Dynamic, systemic processes of representing
- Maturana: In-formation vs. In-struction
- Dewey: Coordination memory
PART II SITUATED ROBOTS
5. Navigating Without Reading Maps
- Parallel, layered machines
- Cooperating, self-organizing robots
- Toto: Recognizing landmarks, learning paths
- How Toto creates and uses maps
- Appraisal of Toto
- Pengi: Indexical representations
- How Pengi uses labels
- Talking about representations
- Classical conditioning architecture
- Computational neuroethology
- Comparison and assessment of minimalist robots
6. Perceiving Without Describing
- What mediates behavior?
- A feature-learning robot
- A chaotic model of perception
- How a rabbit discriminates odors
- Implications for a theory of perception
7. Remembering Without Matching
- Neural Darwinism: An alternative to the encoding view of learning
- Neuronal groups and classification
- Coordinating categorizations by global maps: sequences and
concepts
- Design of Darwin III: Synthetic neural systems
- A situated cognition interpretation of Neural Darwinism
- Prometheus: Coupled recognition-action machines
- Probabilistic topological maps
- Coupling of perception and action
- Summary of neuropsychological claims
8. Engineering Transactional Systems
- Frames of reference for describing knowledge
- The transactional perspective
- Comparison of mechanisms of representing and coordinating motions
- Behavior systems, development, and emergence
- Methods for developing new behaviors
- An evolutionary perspective on adaptation
- Does AI need artificial life first?
- The conceptualization problem
PART III ECOLOGICAL THEORIES
9. Transactional Experience
- A message in Nice
- Inventing a synthetic paintbrush
- Contextualism revisited: No elementary properties
- Conceptual composition over time: Consciousness, feedback, and
error
10. Dialectic Mechanism
- The philosophy of change
- "Both-and" logic
- Examples of dialectic relations
- Conceptual dependency hierarchies
- Scientific levels as dialectic
11. The Ecological Approach to Perception
- Towards the reconcilation of the "situated" and the "symbolic"
views
- The reciprocal view of knowledge
- Towards a theory of "knowing about"
- Niches, affordances, and invariants
- Energy and information
- Gibson on information and perception
- Different views of "construction" and "contained in"
- Relating perception and conception
PART IV SYMBOLS RECONSIDERED
12. Coupling vs. Inference
- Putting inference in its place
- Examples of direct perception in people
- The debate: What theorists misunderstood or poorly explained
- Direct = without inference, not without processing at all
- Information = invariant (stable) dynamic relation, not an
isolated representation
- Algorithmic theory = description, not a mechanism
- Gibson: Perception = resonance mental = symbolic;
- Ullman: Perception = symbolic mental = subjective
- Perceptual = categorical, but not conceptual
- Gibson's interpreters respond
- Shaw and Todd: Abstract machine theory
- Grossberg: Adaptive resonance
- Reed: Perceiving is an act
- Prazdny: The symbolic account is more eccentric
- Bickhard: Functional indicators
- Fodor and Pylyshyn: Perceiving is knowing what you know
about
- Correlation as a semantic relation
- Relation of perception to beliefs, inference, and judgment
13. The Varieties of Symbol Systems
- Reformulating the physical symbol system hypothesis
- Symbolic meaning vs. "distal access"
- How to develop a broader view of symbol systems
- Diverse examples of "symbols"
- Reconsidering human reasoning
- The varieties of conceptual relations
- Heuristic coordination
14. Reformulated Dilemmas
- The procedural-declarative controversy
- Elementary deliberation: Serial vs. parallel, Thinking vs.
reflexes
- The frame problem
- Symbol grounding
- Searle's Chinese room
- What transfers?
Conclusions: Lessons for Cognitive Science
- Clarifications about situated cognition
- How to participate in a scientific controversy
- Beware an either-or mentality
- Try both narrow and broad interpretations of terms
- Given a dichotomy, ask what both assume
- Beware imposing spatial metaphors
- Beware locating relations
- Try viewing "independent" levels as co-determined
- Don't equate a descriptive model with the causal process
being described
- Recognize that first approximations are often
overstatements
- Beware that words can sometimes mean their opposites
- Enduring dilemmas are possibly important clues
- Periodically revisit what you have chosen to ignore
- Beware of building your theory into the data
- Locate your work within historical debates and trends
- "It's not new" does not refute a hypothesis
- Beware errors of logical typing
- Recognize conceptual barriers to change
- To understand an incomprehensible position, start with
what the person is against
- Recognize that the "born again" mentality conceives sharp
contrasts
- Recognize how different disciplines study and use as tools
different aspects of intelligence
- Recognize the different mental styles of your colleagues
- A proper treatment of descriptive modeling
Notes
References
Author Index
Subject Index
Back to William J. Clancey
Home Page