Kuhn's 'decisive transformation in the image of science'?

greenspun.com : LUSENET : History & Theory of Psychology : One Thread

Kuhn begins the Introduction to his book - "The Structure of Scientific Revolutions" (P1, 1962) - with the words; "History, if viewed as a repository for more than anecdote or chronology, could produce a decisive transformation in the image of science by which we are now possessed". Why did Khun believe that we need 'a decisive transformation in the image of science by which we are now possessed'?

-- Jeff Ryan (thephilosoph@hotmail.com), July 06, 2001

Answers

At the time Kuhn was writing, Logical Positivism (or actually, very early post-positivism) was still the dominant mode of philosophical thought about science. LP generally viewed (good) science as a strictly rational and steadily progressive enterprise. Kuhn challenged all that, arguing that paradigm shifts in the history of science did not appear to be based on strictly rational grounds. In addition, they were examples of decisive breaks ("ruptures," as Bachelard had called them long before Kuhn) in the steady culumlative progress it was assumed science should follow. In short, Kuhn legitimized scientific revolution, whereas the logical positivists had believed that revolution was a sign that the science that had gone before was dysfunctional. (Recall, logical positivism originally grew out of the crisis casued by relativity theory and quantum theory overthrowing Newtonian mechanics at the beginning of the 20th century.)

-- Christopher Green (christo@yorku.ca), July 07, 2001.

The best work of Kuhn as far as I'm concerned is his development of the incommensurability thesis. The logical positivists believed that theories could be compared within a common language (i.e., logic), and their view of scientific progress was one in which old theories became integrated within new, larger theories, which became integrated within even larger theories etc. For example, Galilean physics was integrated into Newtonian physics was integrated into relativity theory. Kuhn critically examined some instances in which this was alledgedly the case, for example he looked at the meaning of the concept of 'energy' in classical physics and quantum mechanics, and at the meaning of 'motion' in Aristotle's and Galileo's theories. He came to the interesting conclusion that these terms referred to fundamentally different things. Therefore, he concluded that these theories could not be expressed in a common language; they were in his words incommensurable (not to be confused with incomparable). The conclusion was that scientific progress was not gradual, with earlier theories being eaten up by later ones, but that it was characterized by abrubt changes including quite fundamental changes of the meaning of concepts. Hence his idea of the scientific revolution - to be contrasted with evolution and not to be interpreted in the completely radical and relativist way some of Kuhn's 'followers' did. Kuhn first seems to have come to his conclusion by trying to accomodate for the obvious falsity of Aristotle's theory of motion and the - at first sight - childish mistakes in it, while at the same time trying to retain the idea that Aristotle was not a complete idiot. Through this historical analysis he 'saw' that Aristotle's world was a completely different one from Galileo's because for Aristotle 'motion' referred to a qualitative change we would probably call something like 'growth', and not to a change in position (which was Galileo's meaning of the term). I think this is why he said that history has more to offer philosophy of science than just description and anecdote. And because his analysis did change our idea of science and scientific progress quite radically, and in passing destroyed what was left of logical positivism after Popper all but killed it, I would say his judgement was accurate.

-- Denny Borsboom (ml_borsboom.d@macmail.psy.uva.nl), July 24, 2001.

It seems that you refer to a "Whiggish" view of history, clearly present in Hegel but available in many historians, in which the later development is always an advance on the previous development. Positivism congratulated science, and itself, as advances on previous "metaphysical" ideas.

Kuhn would seem to reject this telic view; yet, he too would admit that today's science is indeed more powerful than yesterdays, wouldn't he? So, the result of scientific revolutions is still a kind of progress, isn't it? We can abandon realism as a criterion for progress, but, we still can't abandon progress. I refer you to the technological history of warfare; for example:

http://www.mhs.ox.ac.uk/geometry/essay.htm

What do you think? Am I off base here?

Yours, P. Fettner Temple University

-- Peter Fettner (pfettner@astro.temple.edu), November 26, 2001.


You're not entirely off-base; for instance, Kuhn saw sciences getting increasingly varied and specialized and increasingly precise over time, and he didn't think this process was likely to be reversed. But what he was arguing against was the idea that (1) a science progresses inevitably towards a particular outcome, and (2) that the outcome can be used to explain why the science arrives at the conclusions it does. This last point is the most crucial. It might seem obvious - an effect can't be used to explain its cause - but it runs against some powerful intuitions, for instance the idea that physics came up with atomic theory because atoms are really there. I have trouble shaking such intuitions myself. But the fact is, the idea that atoms are really there came as the result of an intellectual process and is worthless as an explanation of how Bohr et al. came up with that result - worse than worthless, because it produces bad history, a story in which the younger Bohr is confusedly groping around with a hodge-podge of right- and-wrong ideas towards a new theory, rather than what he was actually doing (according to Kuhn), which was modifying earlier theories in a way that kept the evolving ideas as coherent as possible. So Kuhn the historian was demanding that we tell history from past to present rather than try to make the present cause the past. (And this is one reason why Kuhn thought we needed to transform our image of science: so that we could tell its history.)

And Kuhn was basically saying (I don't think he used these exact words) that there is no explanatory value in claiming that physics evolves towards the truth, or towards a true description of reality. You can - if you want - say that physics has evolved towards the truth, but that tells us no more than that you believe that e.g. atomic theory is true. It doesn't tell us why it's true. For that you have to go to Bohr et al.

By the way, a conclusion that I draw, though Kuhn didn't, is that philosophy has nothing to say about when or why ideas are true, or when we're justified in believing them. In other words, epistemology just isn't a subject. Philosophy much more than science is undergoing the image crisis.

Kuhn's opposition to teleology fit right in with his idea of incommensurability. To elaborate on the example Denny gave above: for Aristotle, motion meant a change in quality or a change in state with a definite endpoint (Aristotle believed that effects could cause their causes), so that motion not only included a rock's moving towards its place in the center of the universe, and fire reaching outwards to its place on the periphery, but also an acorn growing into a tree, a man returning from sickness to health, and so forth. Newton's idea of momentum - that an object in motion stays in motion unless acted on by an outside force - would have made no sense in an Aristotelian system, since it wouldn't have involved a change in quality and therefore wouldn't have been motion.

The point that Kuhn made was that you could see motion as a change in quality or you could see it as a change in place, but that there was no third thing, "the datum" or "what is really there," for you to look at, to compare the ideas to, in order to see which was right. What would such a third thing be? You can compare the two ideas to each other, to see which is better, but not to an independent measure. And the belief that there would be such an independent something, "what is really there," is teleology dressed as "empiricism." To claim that Newton chose his idea of motion over Aristotle's on the ground that he saw that it was really there would be to claim that Newton's conclusion - his laws of motion - caused him to come to it. (This doesn't mean that nothing's really there, just that "really there" has no explanatory value. It's a conclusion, not a cause.)

I agree with everything that Denny said above except the phrase "to be contrasted with evolution." What Kuhn describes is an evolutionary process, just not a gradualistic one. New paradigms are modifications of old, just as new biological species are variations on old. But the new paradigms don't simply contain the characteristics of the old. They destroy many of those characteristics. (In later years, Kuhn said that he'd probably overstated the suddenness of paradigm shifts, but this doesn't change his fundamental ideas of incommensurability and of scientific change being noncumulative. And another crucial idea of his is that paradigms are models of how to solve problems, not rules that scientists follow; so science is an analogic activity not a methodical one, and each field has its own paradigms rather than following some universal "scientific method.")

I agree somewhat with Christopher, except I very much disagree with the statement "paradigm shifts in the history of science did not appear to be based on strictly rational grounds." Kuhn said no such thing, and when people claimed that he'd said it he told them they were projecting, and when they said, well, by your account the shifts can't be rationally based, he vehemently disagreed with them, and told them why he thought the grounds were rational. Having said this as emphatically as I can, I'll add that the phrase "not rational" does appear occasionally in Kuhn's work - very occasionally - and I haven't in all cases figured out what it's doing there. But if you compare Kuhn's account of Kepler in The Copernican Revolution to, say, Arthur Koestler's in The Encyclopedia of Philosophy, in Kuhn's account Kepler's ideas hold together, while for Koestler they're an odd mixture of modern science and old superstition. Rationality wasn't an issue for Kuhn. He simply assumed that if you're not self-defeating and if you try to make your beliefs and values as non-self-contradictory as possible, then you're rational.

-- Frank Kogan (edcasual@earthlink.net), March 09, 2002.


Frank, for what it's worth, I think you're reading an awful lot of 1980s and 1990s "science studies" into 1960s Kuhn. As for the question of "rationality," Kuhn says, "as in political revolution, so in paradigm choice -- ther is no standard higher than the assent of the relevant community" (p. 94). Now one is at liberty the "deflate" the idea of "rationality" to that of "community assent" if one wishes, but to do so blurs an important distinction in the history of thought, even if not in reality. My point (and Kuhn's, I would hold) is that "community assent," rather than "canons of reason" is often the final arbiter in whether a new paradigm wins out over the old one. And scientists often make their move to the new paradigm (or fail to) with their scientific community, not with infallible (or even terribly compelling) *reasons* for doing so.

-- Christopher Green (cgreen@chass.utoronto.ca), March 10, 2002.


I'd have had no trouble with your original statement if you'd said "paradigm shifts in the history of science were not based on deductive reasoning derived from shared premises and shared sense impressions," though this says no more than that we can't use deductive logic to derive our premises. It doesn't follow that our premises can't be tested and that our choice of premises can't be rational, and that some people and some communities, for that matter, can't be more rational than others in their choices. To argue otherwise would be to make the word "rational" inapplicable to most human behavior, and for practical purposes would drive it out of the language. The way to test your premises is to try them out and see what happens, which I think is a good description of what Kuhn says is being done when new paradigms are introduced into science.

Kuhn writing about Kepler: "Kepler was an ardent Neoplatonist. He believed that mathematically simple laws were the basis of all natural phenomenon, and that the sun is the physical cause of all celestial motions." Now I don't find this irrational, even though I don't believe either of the two things. What would have been irrational is if Kepler had not tried to make his ideas work. He succeeded brilliantly with planetary orbits, he failed with motions, but he tried. And if we can't use "rational" to distinguish between a Kepler, who tried to make his ideas work and to hang together, and, say, a Hitler, who just believed whatever he felt, then I'm not sure what good the word is.

I haven't defined "rationality," and I don't think I need to, other than to mouth platitudes about "trying to make one's beliefs and values as consistent and non-self-contradictory as possible, and not being self-defeating." It's not a word that needs a universal definition. It's certainly not "community assent" - I know of no community that would assent to that definition of "rationality." But this doesn't mean that you can't be rational in assenting to the community. It depends on how the community goes about attaining assent, right? Really, the term "rational" is impossibly vague if we don't give it a place and a time of day and a conversation to be part of.

"There is no standard higher than the assent of the relevant community." If Kuhn had left his statement at that, he'd have been doing exactly what he shouldn't have: using the outcome ("community assent") to explain the revolution that preceded it. But he didn't leave it at that. He followed with "To discover how scientific revolutions are effected, we shall therefore have to examine not only the impact of nature and of logic, but also the techniques of persuasive argumentation effective within the quite special groups that constitute the community of scientists." And this is what a lot of the book is about, how scientific communities make the shift, and Kuhn doesn't see these communities as making nonrational choices, and he doesn't assume that persuasion is necessarily nonrational (and neither should we, unless again we're going to insist that only deductive statements based on shared premises can be "rational").

People aren't being absurd, though, in raising the issue of nonrationality, considering Kuhn's choice of vocabulary. Notice "logic" in the sentence I just quoted (maybe he should have said "deductive logic"), and his use of "faith" and "conversion" in later chapters. But I want to reiterate that "rationality" is not Kuhn's issue - he doesn't think it's under threat, and he doesn't want to liberate us from it - and in Structure he doesn't address it. That doesn't mean we can't address it, but honestly, speaking for myself, I think it's a wrong issue, a nonissue, the wrong conversation, and simply detracts from seeing what is interesting in Kuhn. He's interested in understanding past modes of thought, and in understanding the process by which modes of thought change, and he thinks that to do so you need to get a handle on incommensurability. Rationality can take care of itself.

If you're interested, Kuhn did meet the issue of "rationality" head on, when it was forced on him; e.g., the three articles "Reflections on My Critics," "Rationality and Theory Choice," and "Afterwords," the first from 1969 and all three in the posthumous collection The Road Since Structure. I don't own the book so I can't quote directly, but in the first piece he says flat out (in 1969) that current theories of rationality don't apply to science and therefore need to be changed, which is to say that if the theory tells you that science is not rational, then there's something wrong with your theory.

Sorry I went so long in this post, but what I find frustrating about a lot of these conversations is that people allude to what seem like provocative ideas without actually stating the ideas, as if the real discussion were happening elsewhere and we were only allowed to cite such discussions rather than engage in them ourselves. I've barely read '80s-'90s "science studies," and I don't think I'd trust anyone who used the word "deflate" to describe what is, among other things, supposed to be an attack on reductionism. And I'll say further - admitting that this is prejudice, since I haven't read them - that my guess is that people who'd deflate (or conflate, whatever) "rationality" into "community assent" are addicted to the empiricism that they're trying to overthrow, as if they believed that logical empiricist "canons of reason" (buzz word) and "community assent" were the only two choices, so that if you don't have the first (which in their gut they think is the only legitimate one), you're thrown back on the second.

I see Kuhn as part of the ongoing rebellion against the philosophical "foundationalism," a rebellion that goes back to the late 19th century, if not earlier. (An instance of "foundationalism" would be the belief that you can only test an activity [e.g., quantum physics, evolutionary biology] using criteria that are so absolutely independent of the activity that the criteria would exist even if the activity didn't. Anti-foundationalists say that they can't make sense of such "independence.") Specific inspirations for Structure were Quine's "Two Dogmas of Empiricism" and Wittgenstein's Philosophical Investigations. Where I risk misreading Kuhn is in wanting him to be more Wittgensteinian than he is, which is to say that Kuhn thinks that theories of meaning and (once it was forced on him) theories of rationality have value and that he really ought to have such theories (though he doesn't), and I keep wanting him not to, to accept that "rational" is a term for making specific comparative judgments, rather than some gold standard.

An interesting question would be whether we can use Kuhnian ideas of paradigms and incommensurability to understand nonsciences, where there is no analogue to what he calls "normal science" and where paradigms and meanings are never fully shared across a relevant community, hence no paradigm in the Kuhnian sense to be shifted, but where, nonetheless, there are people more or less in the community and others not, people with greater and lesser understanding of each other, shifts in the discussion, and so forth. (And the brute fact here isn't that paradigm choice is something other than rational, but that - whether it's rational or not - these communities as a whole don't arrive at a choice, and the arguments don't necessarily get resolved and may not always be resolvable.)



-- Frank Kogan (edcasual@earthlink.net), March 10, 2002.


Moderation questions? read the FAQ