good@baviki.enet.dec.com (08/21/90)
This was inspired by Matthew Diamond's recent article on treating virtual reality scientifically, but is more general than a followup to that specific article. One of the many things I find attractive about virtual worlds is that it brings many philosophical assumptions out from the background to explicit consideration. One of these is what it means to do science. In terms of Kuhn's analysis of science in The Structure of Scientific Revolutions, I think that human- computer interaction is in a pre-paradigmatic phase. That is, there are no commonly accepted ways of doing science in this field. One school adopts a philosophy often found in the physical sciences where there are facts out there for us to discover. Another school adopts a philosophy where there is no dualistic objective/subjective distinction, but takes an interpretive or constructivist approach where we construct meaning for ourselves. The first approach, based on the physical sciences, is more prevalent in human-computer interaction, but I don't think it is established as a scientific paradigm, especially over the past few years. It seems to me that virtual worlds, by its nature, implies an interpretive/constructivist approach. Since this approach is counter to more prevalent philosophies in science, I think we can expect some discussion of what it means to "treat virtual reality scientifically." Three books which are related to this interpretive, constructive, contextual, or phenomenological approach to human-computer interaction include: Work-Oriented Design of Computer Artifacts - Pelle Ehn (Lawrence Erlbaum Associates, 1988) Understanding Computers and Cognition - Terry Winograd and Fernando Flores (Ablex, 1986) Mind Over Machine - Hubert Dreyfus and Stuart Dreyfus (Free Press, 1986) Ehn's book actually works to synthesize the "Heideggerian" approach of the latter two books with approaches based on the philosophies of Wittgenstein and Marx. There's also a short article on our group's approach to contextual design by Wixon, Holtzblatt, and Knox in the CHI '90 proceedings (pp. 329-336). Michael Good Good@Baviki.Enet.Dec.Com