Header image  
NSF IRI-9311496  
  HOME ::
Computational demonstrations of cognitive processes that transcend the boundaries of individual agents

This project employs computational models to explore aspects of cultural process. In particular, the project focusses on cognitive processes that are inexplicable in terms of the properties of individuals. The cultural process operates in the dynamics of a system composed of learning agents together with the environment in which the agents live, including the products of agent activity such as tools and behaviors. A central fact about human cognition is that it proceeds in a cultural process. The simulation models consist of communities of connectionist networks. Connectionist networks were chosen not because they have any special neural character, but because they provide a good vehicle for modelling learning as the internalization of structure in the environment through action in the environment -- a crucial feature of cultural process. The networks stand for simplified agents which must generate material structure in the external world in order to effect the behavior of others in the "community" of agents. We do this by simulating simple worlds in which learning agents create (in interaction with each other and with the objects in their world) structures which mediate (or come to mediate) the organization of their behavior.

The goal of this work is not to replicate the existance of any particular cultural process (one sense often implied by "modelling"). Instead, the simulations are intended to aid in the articulation of what cultural process is, and what it affords cognitive agents who are embedded in it. The intent is that the work supplement empirical investigation of cultural process (as it occurs in the real world) by refining the theoretical constructs which we use to look at the world. To this end, we have produced the following papers.

Learning in the cultural process

Edwin Hutchins and Brian Hazlehurst

In this paper, we introduce a framework for simulating cultural process. The general idea is to simulate a world in which learning agents create external structures which mediate their behavior. This simulation demonstrates the simple (but very important) concept that such a system is capable, through generations of time, of producing agents endowed with cognitive powers that are not attainable in the lifetime of any individual agent. We argue that such an outcome is possible without effecting the genetic organization of individuals. However, we also argue that the cultural process is capable of guiding both learning and phylogenetic evolution, which leaves open the possiblity of genetic organization tracking cultural process.

Appears in C. Langton, C. Taylor, J.D. Farmer, & S. Rasmussen (Eds.) Artificial life II. SFI studies in the sciences of complexity (Vol. 10).  New York: Addison Wesley, pp 689-706, 1991.

How to invent a lexicon: The development of shared symbols in interaction

Edwin Hutchins and Brian Hazlehurst

In this paper, we elaborate upon the framework by considering more explicitly the problem of creating shared symbolic structure. A lexicon is (among other things) a set of public structures for denoting or implicating shared meanings. In other words, the existence of a lexicon requires the sharing of forms and meanings -- and mappings between these -- among members of an interacting population of agents. Leaving aside cosmic and theological events which could create such an outcome, how could a lexicon come to be? The solution provided here is based upon a convergence of agents' schemes for classifying visual phenomena. All agents have a capacity for such classification, but convergence upon a singular scheme is shaped by the constraints for consensus when employing the scheme in interaction with other agents.

Appears in N. Gilbert & R. Conte (Eds.) Artificial Societies: The computer simulation of social life, London: UCL Press, pp 157-189, 1995.

A shorter version appears in E. Goody (Ed.) Social intelligence and interaction, Cambridge University Press, pp. 53-67, 1995.

The emergence of propositions from the coordination of talk and action in a shared world

Brian Hazlehurst and Edwin Hutchins

In this paper, we draw upon earlier work in order to address the question: Where do propositions come from? We claim that propositions can emerge from a cultural process in which there is a need to share the sequential organization of shared symbolic structures. The idea is that the management of joint attention in the world creates a context in which a certain organization of language-like structures can emerge. This language-like system entails referential properties (see How to invent a lexicon), but now also has syntactic properties because the management of attention requires the negotiation of action between interlocutors in time. We show how such negotiation in the service of coordinated action is sufficient for the development of a systematic, rule-describable "language."

Appears i n: Language and Cognitive Process. (Special issue on Connectionist Approaches to Language Development), K Plunkett (ed).  13(2/3):373-424, 1998.

An overview of this line of work is available in:

Hutchins E. & Hazlehurst B. Auto-organization and Emergence of Shared Language Structure, In Cangelosi, A. and Parisi, D. (Eds) Simulating the Evolution of Language, London: Springer-Verlag, pp 279-305, 2002.