There are a number of issues that arise from Tom Reeves provocative and interesting paper. However, I would like, currently, to address a response from Rod Sims, given below:
[quoting Sims, 26 Feb 96] In reviewing the query and response about whether or not to use CAI, it seems that this is at the crux of what Tom is trying to say--or get us to understand. As far as we "down under" think (if I don't speak for all my colleagues here, they'll quickly let you know), the notion of CAI disappeared shortly after the meteorites removed the dinosaurs!
CAI is a manifestation of behaviorism, and it seems we have not only moved beyond that S-R world to cognitive approaches, but even further to the world of constructivism and building models of the world using the best tools.
In using various terminologies, we have to take care not to misrepresent and misuse learning theories or constructs. There is already enough confusion around in this area.
Firstly, Computer Assisted Instruction (CAI) has, as a descriptive term, different connotations for different cultures. Rod Sims, in his denigration of CAI, is referring to a narrow view of CAI and one which originated in the UK and became widely adopted there and in Australia but not in the US. In one sense, as with CAL, CAI (particularly in the US) should be seen as a blanket descriptor, one that refers to a large range of computer-use strategies (as CAL does in Australia and in the UK)--not simply "behaviorist" ones.
Secondly, we should also be careful in representing cognitivist theories of learning as "recent"--or even constructivism as a single and cohesive theory of learning. Perhaps more importantly, we should avoid a rather simplistic view of past, where developments are represented as progressive and the most recent adoptions of learning theory as the best.
Cognitivist theories and behaviorist theories have existed side by side and for some time--and we should not forget the psychology of, for example, Tolman's work, where cognitive theories are represented in a behavioral tradition. There will always be learning situations which can be profitably interpreted and analyzed using behavioral constructs. In other situations, cognitivist approaches to analyzing conditions of learning provide more appropriate tools. A useful reference in this regard, is Hill (1985), where it is suggested that criteria for an "ideal" theory of learning would have to include both connectionist (behavioral) and cognitivist constructs.
For most of us, learning theories have two main values. One is in providing us with a vocabulary and a conceptual framework for interpreting the examples and conditions of learning that we observe. The other is in suggesting where to look for solutions to practical problems. The theories do not give us solutions but they do direct our attention to those variables or conditions that are crucial in finding solutions.
Tom's paper, recently published in this forum, makes a coherent and well-argued plea for researchers in IT to address socially relevant questions and to do so using methods that are relevant to current needs in IT education. He argues the need to develop new theories (of teaching and learning) based upon a sound body of systemic research. I would like to extend this to include a plea not too misrepresent or misuse those theories we already have to guide us in the directions Tom advocates.
Hill, W. F. (1985). Learning. A survey of psychological interpretations (4th ed.). New York: Harper and Row.