Introduction Anthropologists synthesize cultural systems formulated at a symbolic level based on how behavior or productions of behavior relate to contexts of use. Using a range of systematic methods as well as literary and expository techniques and conventions, they communicate a sense of the content of cultural systems, possible ranges of variation, and how these change over time. They produce detailed descriptions of human behavior, though of necessity these are often synthetic, drawn from

more »

... rtial experiences or individuals' accounts of experience. Knowledge is one of the most complex of these behaviors and is the most common focus for research in Cultural Anthropology: what people know and how knowledge is created, transmitted, distributed and instantiated. Much of the past century of anthropology was dominated by fundamental issues relating to knowledge, with anthropologists variously exercised by instances of apparently irrational knowledge (Sperber 1986); e.g. magic, spirits and weather control, while others were overwhelmed by the complexity of everyday life (Geertz 1973). In particular this has given rise to questions about the prospect of representing knowledge systematically. Some, such as Richards (1993) dispute this is possible. Others, including the contributors to this collection, argue that knowledge can be represented systematically. The issue, in part, is what constitutes a knowledge system. Richards appears to be arguing against a single deterministic systemic representation of knowledge, and argues instead for an 'improvisational' characterization, based on 1 interactive decision-making within a constantly changing historical context, idiosyncratic for each farmer and where that historical context constrains or directs the appropriate applications. There is no disagreement on any aspect of what Richards proposes. However, Richards refusal to see the embedding of knowledge and context as a system does constitute a conflict. One of the points of an inferencing system for humans is to deduce what kind of system one is in. A person embedded in a cultural system is interlocked in potentially many knowledge systems (see articles by Leaf and Lyon, this number), and each system is non-deterministic for the most part. To interact with a system, the person has not only to decide what system it is, but which variant of the system it is, and to some extent must reconstruct a history for events whose cause is not directly apparent. This was one of the ideas of Austin in "How to do things with words". It is this notion of getting in synch with the world that intelligent systems, especially social intelligent systems, require. We can further clarify this process of 'getting in synch' by examining the relationship between applied scientific and cultural knowledge and their applications. Specifically, applications (or instantiations) of scientific knowledge are not themselves science, and such applied knowledge has properties not unlike those described by Ellen and Harris (2000) for 'Indigenous Knowledge' (IK). In particular, in applying domain related knowledge, additional knowledge is used that is not directly related to the domain knowledge, but nevertheless is necessary to apply domain knowledge in a contingent world. This inclusion of knowledge that must be instantiated in order to enable domain knowledge to be instantiated builds on Ellen's concept of prehension (Ellen 1986 (Ellen , 1993 , which in part includes the anticipatory knowledge a person brings to a situation where they must instantiate knowledge. Describing or formalizing enabling knowledge permits us to more formally describe what Ellen and Harris suggest is 'tacit, intuitive, experiential, informal,