======== WDL Dis. #22 (Bcc: all WDL participants) ====== From bruemd@sainc.com Mon Mar 6 12:10:33 2000 From: "David Bruemmer" To: wdl@cse.msu.edu John, Your five levels [(5)knowledge, (4) inborn behavior, (3) representation, (2) architecture, (1) timing] provide much needed differentiation. As you have pointed out, vague terms such as 'prior knowledge' and 'learning' can be misleading. Desultory use of such terms certainly hampers the penetrating, systematic investigation you call for. However, I feel you have misunderstood the modus operandi which drives much of the field of psychology. I am not a psychologist. Currently, I am, among other things, trying to develop a robotic landmine detection platform. I am perhaps a dilettante (hopefully in the old sense of the word), but through excursions into a number of fields, I have come to understand that different fields use different tools. Dust mite research conducted with a SEM microscope isn't necessarily any more correct than the clinical research which studies people's allergic reactions to mites. Different "lenses" allow entirely different dimensions to be observed. Each dimension must be examined in fundamentally different ways. For neuroscience to tell pyschology that it is more rigorous and systematic, would be analagous to chemistry declaring itself a more correct, precise science than physics. They both see different dimensions and although, certainly, these dimensions overlap, each science is systematic (and ad hoc) in its own way. You mention Physics and Astronomy as examples of mature sciences. Certainly, these are old endeavors. I think it is important to point out that the 'systematic' research which has brought us the concept of dark matter is very different than the systematic research which studies how to build bridges. Rather than assert that fields necessarily become less ad hoc and more 'scientific' as they mature, I venture the following model as another viewpoint of how all fields cycle, roughly, in the following manner: Visionary minds use, in addition to empirical study, intuition, imagination, and perhaps even a bit of dumb luck to birth new, revolutionary insight. Funding flows and positivist optimism emerges that new rules will capture a universal truth essential to the field. A host of researchers jump on the band wagon and, for a time, systematic study seems to support the new rules. Further systematic research finds fundamental flaws in the original theory and the hapless visionary (hopefully lucky enough to be dead) becomes discredited or falls from favor. Researchers become dissollusioned, funding slows, other fields look down on you. Back to 1. This is but a pet theory of mine and hopefully no one will waste their time pointing out any number of readily available contradictions. Nonetheless, I believe this cycle can be seen in the history of the computer model of mind and yet again in the neural network craze that usurped it. It is true for the field of AI in general as it has also been true on a grander scale for Physics (Newton, Einstein and now, once again, some scientists believe that they are on the brink of perfect, universal equations). How can we avoid this cycle (or at least make it more profitable)? I believe we must remain distributed across dimensions as well as across fields. You have said that psychology should stop looking at level 5 (knowledge) and start to investigate (1) - (4). You wrote: The same paradigm seems applicable to psychology. Psychology will become a much beautiful science if the field shifts its current emphasis on (5) to (1) - (4) instead. (5) is very ad hoc by nature. A lot questions of long debates in psychology could be answered systematically if psychology searches answers through (1) - (4). If you mean that we must examine development at all five levels, then I wholeheartedly agree. This conference should be an important step in the right direction. However, what I also hear implied is that (5) is a misdirected, unscientific dimension we would do well do avoid. We should not view one toolbag/lense as more appropriate than another. As Susan points out, we must maintain an interest in the effect as well as the cause. This caveat is especially important when talking about brain/mind. We must not assume that we fully understand the effect (a developing mind), sufficiently to turn our backs on level (5) and begin modeling the cause (1) -(4). If we do, I fear we are all the more likely to spiral through the steps in my pet theory. Instead research in each area should progress in tandem, with understandings from each dimension informing the others.