Figure 75: What represents and/ or describes what?

 

 

Cambridge, Mass.: MIT Press 1986; paperback edition 1989; 5^th print 1996.

 

content

The cognitive system constructs models

Models/ default hierarchies represent situations

Cognitive systems represent (portions of) the world

The model describes (aspects of) the environment – and the system’s actions

Empirical rules model the world/ represent the situation

 

 

 

The cognitive system constructs models

 

p.12

In Common with many recent theoretical treatments, we believe that cognitive systems construct models of the problem space that are then mentally "run" or manipulated to produce expectations about the environment (Craik 1943; Gentner and Stevens 1983; Johnson-Laird 1983). Induction consists of generating and revising the units from which mental models are constructed. For a variety of reasons we will take condition-action rules as the most important of these units.

… Mental models are best introduced by contrast with the notion of a schema, which has been extremely influential in cognitive psychology and artificial intelligence, under that name as well as under the labels “script”, “frame” and “concept” (Bartlett 1932, Piaget 1936, Minsky 1975, Schank and Abelson 1977, Rumelhart 1980). The premise underlying the schema notion is that information about the likely properties of the environment are stored in memory in clusters that can be accessed as large units and that can serve to generate plausible inferences and problem solutions.

… Although schemas are valuable for chunking information together, they exact a toll in inflexibility. What is to be done with situations or things that are not neatly matched by any existing schemas?

 

Bibliography

 

Sir Frederic C. Bartlett: Remembering. A study in experimental and social psychology. Cambridge: Cambridge University Press 1932; various eds till 1995.

Jean Piaget: La naissance de l’intelligence chez l’infant. Neuchâtel: Delachaux et Niestle 1936;
English: The Origins of Intelligence in Children, New York: International Universities Press 1952; 6^th print 1975;
u. d. T.: The Origin of Intelligence in the Child. London: Routledge & Kegan Paul 1953; various eds till 1997;
German: Das Erwachen der Intelligenz beim Kinde. Stuttgart: Klett 1969; 5. ed. 2003; further München: Deutscher Taschenbuch Verlag 1992.

Kenneth James William Craik: The nature of explanation. Cambridge [Eng.]: University Press 1943; again 1952, as paperback 1967.

Marvin Minsky, A Framework for Representing Knowledge. In Patrick Henry Winston (Hrsg.): The Psychology of Computer Vision. New York: McGraw-Hill 1975.

Roger C. Schank, Robert P. Abelson: Script, Plans, Goals and Understanding. An Inquiry into Human Knowledge Structures. Hillsdale, N. J.: Erlbaum 1977.

David E. Rumelhart: Schemata: The Building Blocks of Cognition. In Rand J. Spiro, Bertram C. Bruce, William F. Brewer (Hrsg.): Theoretical Issues in Reading Comprehension. Hillsdale, N. J.: Erlbaum 1980.

Dedre Gentner, Albert L. Stevens (Ed.): Mental Models. Hillsdale N. J.: Lawrence Erlbaum 1983.

Philip N. Johnson-Laird: Mental Models. Towards a Cognitive Science of Language, Inference, and Consciousness. Cambridge: Cambridge University Press 1983, 6. ed. 1995.

 

 

Models/ default hierarchies represent situations

 

p. 14

… Although mental models are based in part on static prior knowledge, they are themselves transient, dynamic representations of particular unique situations. They exist only implicitly, corresponding to the organized, multifaceted description of the current situation and the expectations that flow from it.

… Models must consist of components that can be flexibly constructed and interrelated. Our most basic epistemic building block is a condition-action rule, which has the form ”IF such-and-such, THEN so-and-so”, where the IF part is the condition and the THE part is the action.

 

p. 16

… To represent environments fraught with novelty, mental models cannot rely exclusively on precompiled structures. Flexibility can come only through the use of combinations of existing knowledge structures.

 

p. 19

… Default hierarchies are capable of representing both the uniformities and the variability that exist in the environment. This representation serves to guide the kinds of inductive change that systems are allowed to make in the face of unexpected events.

 

 

Cognitive systems represent (portions of) the world

 

p. 29

In our framework a cognitive system represents the world with which it interacts using mental models, constructed from rules and organized into default hierarchies. In assembling a model of the current situation (often, in fact, a range of models, which are allowed to compete for the right to represent the environment), the system combines existing rules … The assembly of a model, then, is just the simultaneous activation of a relevant set of rules.

 

p. 30-31

… Because a mental model is the cognitive system’s representation of some portion of the environment, and because the relation between the model and the environment is critical to understanding the model’s role in the cognitive system, our first step must be a more precise treatment of the environment … As we will see, learning a representation of the transitions function is the critical goal in the construction of a mental model.

 

 

The model describes (aspects of) the environment – and the system’s actions

 

p. 33

… A model may or may not be a valid description of the environment.

 

p. 36

… Appendix 2B demonstrates that a hierarchy of default rules with exceptions can represent knowledge more concisely that is, with fewer total rules than a system restricted to “exceptionless” rules.

The representation of events and event sequences at different levels of a default hierarchy is the most basic way for a system to deal within variability within a class of events.

 

p. 39

… The model need only describe aspects of the environment and of the system’s actions that are relevant to the attainment of goal-satisfying states.

 

 

Empirical rules model the world/ represent the situation

 

p. 43

… Whereas the function of empirical rules is to model the world, the primary function of inferential rules is to produce better empirical rules.

 

p. 205

… Rules, organized in a default hierarchy, compete with one another for the right to represent the situation and to predict its successor situations.

… People model the physical and social worlds using empirical rules organized in default hierarchies.