Realism versus constructivism: beyond Popper and Kuhn

Before turning to some new readings, we had an intriguing discussion of the implications of Popper and Kuhn respectively for the teaching of science ... and people's current positions on the ideas that each emphasized. Among the thoughts that came from that was that Popper would endorse a principle of teaching not content but process, ie that science classes should not teach people what "is", but rather help people acquire the perspective and tools of falsification. Some content might be needed to provide a foundation for falsification. Kuhn, on the other hand, might endorse the teaching of science more as is currently done, to give students an appreciation of contemporary "reality", ie the currently accepted paradigm, so they could participate effectively in "normal" science. This picture is, of course, a little ironic, since one tends to think of Popper, the "realist", in the context of rigorous science education, and Kuhn, the "constructivist", in terms of challenges to the validity of existing paradigms.

In this context, there emerged a sense that some blending of Popper's and Kuhn's approaches was needed, in particular that an appreciation of the tentativeness of knowledge and associated commitment to falsification is important but that one also needs some form of normal science, ie an accumulating set of problem-solving observations that provide a more or less agreed upon sense of "reality" for practical purposes as well as to provide a basis and grist for subsequent falsification. The questions that follow from this are (1) is there any relation between paradigm shifts and falsification? (2) is there any "progress" across paradigm shifts? (3) what role does "reality" play in paradigm shifts?

Again, there is irony here. Popper was concerned with "reality" and "progress", and yet it is Kuhn's "normal science" that is in some ways most easily yields some understanding of those terms. Kuhn is most typically associated with a challenge to science as "Truth" and yet it is Popper who seems to most completely do away with that concept. Perhaps a way to blend the two would be to conceive a way to incorporate some of Popper's falsification commitment into Kuhn's paradigm shifts, in a way that adds some cumulative character to them but without invoking anew a "mirror of nature" approach to science (Rorty), ie a measure of progress based on a correspondence theory of truth. Could it be that, contra Kuhn, a choice between incomensurables actually involves some component of "falsification", not in the narrow sense of there existing falsifying observations but rather in a broader sense? Perhaps the incomensurables themselves create a new paradigm within which it is possible to see one as a subset of the other and this, rather than arbitrary social forces, is in fact that basis for choices between them? A continually widening set of observations accounted, rather than a correspondence theory, for would then be the criterion for progress. And "falsification" would be broadened to include "subsuming". And "incomensurable" would be understood not as a permanent state but rather as the incentive that drives still broader paradigms within which existing ones can be compared (without, however, ever mentioning "reality" in any absolute sense). We'll be exploring this set of possibilities later in the course (Grobstein, From Complexity to Emergence and Beyond).

The Goodman and Hemple readings, as summarized by Sunyoung, moved us in interesting ways in this direction. Goodman argues that there isn't A way that the world is, speakable or otherwise (contra mystics). Instead there are multiple speakable ways the world is, ie that the world is nothing more (and nothing less) than the collection of meaningful (useful?) statements about it. That's clearly a construvist position, perhaps an extreme one, and raises some significant questions. Is there "actually" one world with a number of different descriptions? or simply a number of different worlds? How is one to account for the fact that some some things are recognized by us as being the same even when they change (Descartes' ball of wax intact and melted)? Is there an "objective" world or only subjective ones? Could the "objective" be simply that which everyone agrees on? Could things be agreed on BECAUSE there is an objective world out there?

Hemple represents these issues in terms of the "stubborness of facts", and with a reminder that contemporary constructivism was anticipated by the Vienna Circle's Otto Neurath and his metaphor of science as "like seafarers condemned to repair and rebuild their ship forever on the open ocean, without any possibility of taking it into a dry-dock and rebuilding it there on a firm basis". Facts, for Hemple, are things that are there whether we are or not. They are characteristics of what is out there, of an objective world that can in fact be "carved at the joints", as McKenna describes it in relation to common-sense or "Plain Jane" metaphysics. Are there in fact such facts? Or only commonalities in our individual descriptions? Is the law of falling bodies a "fact" or only a convenient way to make sense of shared experiences and understandings (most people have never in fact seen a body fall according to the law since it would do so only in an ideal case with no other influences, in a perfect vacumn). We'll return to this issue this coming week in the particular case of relativistic and quantum physics.

Davidson offers a somewhat different challenge to a relativistic constructivism, asserting that the very idea of incommensurable understandings, in the sense of understandings that cannot be inter-translated, is itself incoherent. If there wasn't a common framework within which understandings could be compared, he argues, then one wouldn't be able to notice incommensurabiliites. Can one notice difference between two understandings without having a third more general one in which they can be compared? That too is a question we'll be returning to this week in connection with the history or physics. And perhaps one tht will help us conceive ways that paradigm shifts can be progressive without resorting to "mirror of nature" characterizations of science.