Thomas Kuhn "Structure of Scientific Revolutions"

It occurred to me that perhaps it would be best to answer my own questions up front to get reactions, as opposed to just leaving them here. So here's my take on Kuhn.

1) What did you take the term "paradigm" to mean?

The two things that defined a paradigm most definitively (Kuhn does seem to be terribly vague on this point, IMO) were the problem-solving solutions taught in the process of teaching new scientists, and the particular historical perspective adopted for pedagogical means. A paradigm can more clearly identified by these two criteria than by simply referencing "the world-view which the scientific community shares" (though that certainly has it's place and deserves to be discussed).

I think this is the best approach because science is learned, mostly, by solving problems. The ideas, on the face, aren't the most complex and difficult things to get at with respect to other subjects, IMO. They aren't meant to be. The complexity comes in the application of these ideas to modeling problems or finding solutions to problems. This helps later in the generation of hypothesis and in approaching problems/puzzles/questions within the lab (as you do find problems), as you have a basic rule-of-thumb by which to start finding a solution. As such, a new paradigm, if exists, will teach problems differently at some level. This doesn't occur at every level, as can be witnessed by the still-in-use problem solving techniques of Newtonian mechanics. They're still-in-use because they still work for their set of situations, and they're a hell of a lot easier to use than the problem solving techniques used to model other situations. (Further, relativity is pretty elegant in that it reduces to Newtonian mechanics in the conditions Newtonian mechanics was generated to model. More on this later in "incommensurability", though)

The historicism of the scientific pedagogy is another central point to Kuhn that serves as a historical looking-back-upon for scientific revolutions. Those mentioned in the history most often are those who revolutionize. There are other figures as well, mentioned for this or that law/hypothesis/interpretation (as Kuhn would put it, "normal science", albeit significant contributions to normal science none-the-less), but the history of blindingly brilliant individuals who broke the conservatism inherent to the scientific process is a main sticking point in demarcating paradigms as these individuals, through history, come to be less accurate descriptions of their biography and more heroic figures enshrined in the values and assumptions that the scientific community should adopt. The story is often told in the frame of "He observed, couldn't explain, postulated, proposed experiment, and since then..." etc, etc. In short, the story follows the general rule-of-thumb for scientific progress currently being taught. I know a handful of these stories myself, but I haven't checked to see if they're true. Further, I don't care if they're true. They're useful to understanding science.

I dislike "world-view", not because it doesn't seem to encapsulate what a paradigm is, according to Kuhn, but rather because it lacks anything resembling a decent empirical point of demarcation in which we can identify paradigms.


2) Did you read his work as prescriptive, or descriptive?

Mostly descriptive, for myself. There are some points (such as the scientific communities need to rely upon conservative values in order to progress at all, etc) that are prescriptive, but I think Kuhn was more attempting to explain why paradigms change and challenge the assumption that science = progress at all times.

3) How did you react to the concept of incommensurability?

I'm still thinking on this one, myself. In some sense, such as the point Kuhn makes between Einstein's concepts of matter and Newton's concepts of matter, I can see as somewhat incommensurable. But, simultaneously, this might be too metaphysical to be counted as incommensurable -- even under the prescriptions of Kuhn, scientists within the paradigm that had a positive prediction in solving an outstanding puzzle whilst attempting to not destroy all puzzles so far solved in that solution. That doesn't indicate that new paradigms are incommensurable to each other. Primary point that I noticed he left out: The periodic table of the elements has survived several models of the atom. Also, the ideal gas law is intentionally built on "falsity" and the Newtonian paradigm, yet it is still in wide use today. So, in part, I could see things being incommensurable, but there are some obvious arguments I don't think Kuhn addresses very well.

4) How did you react to SSR?

I thought it was well thought out and presented a decent argument for a novel position. I was initially turned-off to Kuhn due to the nature in which it was presented to me by others, but reading it myself I don't think those others had the right of it. In fact, after reading it, I sort of thought it was so specialized that I still wonder why it's so widely read. This isn't to say that it's bad, mind, but usually books on the sociology and philosophy of an academic discipline in conjunction with a hypothesis about their internal changes just doesn't attract this much interest.

What did you take the overall thesis of the book to mean?

I do agree with those that think Kuhn divorced truth from science. However, I think Kuhn totally rejected the idea of truth in the first place. So, whilst against scientific realism, his particular rejection of it has been less emphasized than that rejection itself by those who previously introduced me to Kuhn. I, myself, am a scientific realist. But because of Kuhn I'm more aware of the difficulties on my position, and for that appreciate the book.

I am interested in what you mean by 'truth' in this case. Given that particular measurements will still be accurate, and particular theories will be true within their domain of applicability. So how does Kuhn divorce truth from science?





I understand the discomfort here, but how is it possible to provide an empirical demarcation for something as vague and all-inclusive as worldview? The medieval worldview, with the 'superlunary spheres' and the ptolmaic cosmology was brought down by the scientific revolution of Newton and Galileo. But the Newtonian worldview of 'matter in motion' has been subsequently undermined by QM and relativistic physics. In fact I don't even know if there is 'a worldview' in modern science any more. But whether there is or not, Kuhn's point is well illustrated by these changes, particularly the transition from the medieval to the modern view. (I will await with interest the view that emerges from the rubble of modernity.)

I think what Kuhn means by truth are statements such that is always true and always will be true. For example, 2 + 2 = 4, or "The blue book is blue", but much more dramatic than these examples. (though perhaps not restricted to these... I'm still a little uncertain on this myself) The conception Kuhn is against is that, in changing Paradigms, like the change from Newtonian to QM conceptions of momentum, for example, that the reason for the change of Paradigms is that the new Paradigm more accurately represents the universe around us. Kuhn disagrees, stating that the reason one changes Paradigms is because it solves an outstanding and irritating puzzle that the scientific community has yet to solve.

For the specific line where Kuhn talks about truth, I'll just quote him at the end of SSR. It sort of steals the thunder from the passage to read it before reading the preceding 170 pages, but what the hey:



So he divorces science from truth by the following rough outline, I'd say:

1) There are paradigms adopted by the scientific community, which include the ways that they interpret the world, the values they share, the measuring devices they construct, and the problems they create for themselves to solve, amongst other things (roughly -- seriously, I wish he were more clear)

2) These paradigms are incommensurable -- the positions held by scientists throughout the ages contradict each other in a fundamental way.

3) A paradigm change is not characterized by a truth-value justification, as is commonly assumed. Rather, the function of a paradigm change is to solve an outstanding problem, or "Resolve a crisis" (as Kuhn puts it). Science does not evolve towards truth, but evolves from present webs of knowledge.

4) There is no reason to think that the resolution of a crisis actually brings one closer to truth. Instead, it satisfies a set of values held by the scientific community.

So... that is the way in which he divorces science from truth. He doesn't think that truth is a tenable concept for evaluating scientific theories in the least. Instead, he thinks there is another function that selects for theories, like "the fittest theory wins" (He brings up allusions to Darwin, though stresses that these can become out of hand). He doesn't explicate in this book what that function is, but more states that truth isn't it in any sense of the word truth, whether it be capital-"T" or little "t" type. To him, I think, it's just an untenable concept for evaluating scientific theories.

That's my exact problem with using the idea of "world-view": It's too vague for a decent identification of what a paradigm actually is, though Paradigm can be interpreted to mean "world-view". I would prefer, instead, the problem solving methods present in text-books, and the historical tales of scientific heroes told. So, the training process might best reflect a paradigm.

You might be able to make head-way with a "Fundamental concept" approach, but I think there are difficulties with it. For example, It may seem that our conceptions of matter are fundamental, but that itself may just be because we think that the behavior of matter would effect the behavior of, say, animals in an important way. Aristotle certainly did not think this was so, but that teleology was the fundamental metaphysic of the universe, matter be damned. So, maybe there is some traction here, but it could present some metaphysical and interpretation difficulties, I think

Yikes, this new forum is going to take some getting used to. Anyway...

Nice post! I haven't read the SSR in quite a while, but it remains one of my mental touchstones when I'm thinking about the philosophy of science. I read it for the first time as a student in a course about the phi of sci, along with some examples of Popper, Feyerabend and a few others. To address your questions (out of order):

4.) I honestly don't remember having a particularly strong reaction to the book in either a positive or negative sense. The ideas it contained did not seem disarmingly revolutionary to me at the time, doubtless because so many of them had already filtered down to me by indirect means due to the popularity of the book. My specialization within the philosophy dept was in the history of philosophy (rather than in epistemology, aesthetics, or logic, say), which can hardly be counted as a specialization at all. And the phenomenon that Kuhn described seemed like a very common occurrence within a variety of intellectual pursuits; often similar transitions took place within a single generation. So the theory that science was not an exception to this did not challenge any entrenched positions I had nor seem all that shocking to me, a non-scientist. Re-reading it a few years later, but still a while back, the book actually seemed a little more fresh than on the first reading. No doubt in part because I was not reading it as part of a course, but also because I had learned a little bit more about the history of science and scientists rather than the historiography of the same.

2.) I read the work as primarily descriptive. Although I was aware of some of the prescriptive qualities that you mentioned, I tended to interpret them as indicative of a commonly held position within the scientific community rather than as directives.

1.) and 3.) are very related questions for me. Please keep in mind that my interpretation of "paradigm" is by this time several degrees removed from the book. (An interpretation based on multiple readings years apart, years ago. And I can't refresh my views with a brief skim, as I can't find my copy. Did I sell it at some point?!) So what I have to say about it is less an exegesis, and at this point more of a personal notion. I remember that when I first read it a paradigm seemed like a relatively precise body of data: the pedagogical directives (restrictions on what would be considered an effective experiment, etc.) forming the core, this branching out into the general historical perspective. Overall, I think I envisioned it to be something like a combination of Foucault's "archive", an abstract database, and a metaphysical card catalog. However, as time has gone on the idea of a scientific paradigm has become much more diluted. The relationship between theory and practice seems less one-sided to me at this point, and it also seems possible that all sorts of "nonscientific" data might also belong to a given paradigm. For example, how funding is distributed for research is determined to some degree by personal, political, and aesthetic tastes at the time. And so how and why experimentation takes place is effected by not only the value of the inquiry to scientific curiosity, but also to its immediate practical implications and simply how fashionable it is in the popular imagination. To me, the rhetoric used to sell a theory is a piece of the paradigm, even if it is not an important piece of scientific knowledge, because it shapes the scientific community's historical perspective. Unfortunately, I can't recall whether or not Kuhn addressed any of this.

If a paradigm is simply a restricted body of knowledge, then the idea of incommensurability seems largely tenable; although I liked your counterexamples and of course there are others. However, if "paradigm" is given a much more vague definition, which is where I lean these days, the incommensurability between paradigms seems less total than it might be. Rather than it being a case of black and white it allows for shades of gray, and the opportunity for a hardline stance might even be the exception.

On a related, but slightly tangential note, and in contrast to the above reservations: I wonder to what degree might incommensurability exist within a single paradigm? At this point, it is taken for granted that QM and relativity theory do not contradict one another, but we are a far cry from a unified theory nor is one method reducible to the other. Nonetheless, they seem to me to belong to a common paradigm. Of course, relativity is not a problem for quantum mechanics, nor is QM a problem for relativity theory. Rather their non-identity is a problem for science in general, and this might be considered one of the exemplary problems that incites a paradigm shift. Still, this disjunction seems to me to partake of incommensurability as much as, if nor more than, it is an example of an unexplained phenomenon.

Overall, my impression of Kuhn's book continues to be positive. It provides a general framework around which to organize a larger sense of science's history. I think regular exceptions to its theory are easily found when studying this history in detail, but it is useful as an approximate, large-scale, x-ray image of the history of science. If it does not provide the solution to all of the historian's quandaries, it is at least a spur to some fruitful questions. So long as it is not adhered to too rigidly, and is used in combination with a few other models, it yields valuable insights into how science can work as an intellectual discipline.

Question 2 interests me the most. "Structure of Scientific Revolutions" seems to be more descriptive than prescriptive. It is a good survey of the history of science that provides insight into the "sociology" of science. According to Freeman Dyson in "The Scientist as Rebel" Kuhn was unhappy with dogmatic interpretations of his book. Dyson met Thomas Kuhn at a meeting of historians and quoted Kuhn as proclaiming: "One thing you people need to understand: I am not a Kuhnian!"

The idea of 'capital-T' truth is difficult. I have raised it before on the Forum, and the consensus always seems to be that Truth is a meaningless idea. There are only true propositions, or to put it another way, truth is a matter of correspondence between a proposition, and that which it seeks to represent. But the correspondence theory of truth is subject to many criticisms, not least that it always assumes that we can ascertain the truth from the outset, otherwise there would be no possibility of correspondence. So it is an a posteriori explanation which doesn't seem to allow for a priori truths.

Of course, the idea of capital T -Truth is rather romantic or spiritual. Perhaps this is why it is rejected in our secular age. I associate it with Herman Hesse novels and spiritual philosophy, in both of which truth is often spelt with a capital T. And I feel quite at home with this usage.

So I think that Kuhn's rejection of the idea of science 'gradually disclosing the truth' is very interesting in its own right. I think this is one of the key motifs of post-modernism, and maybe Kuhn's book is actually a key text in the origin of post-modernism.

Modernism still believed in the idea of Progress, and I am sure the idea of Progress was itself adapted from the Christian anticipation of the Second Coming and the eventual redemption of the world. Science, after all, started out with a commitment to Truth, because it grew from a tradition wherein Truth still had a meaning. But the further it strayed from its spiritual roots, the less meaningful it appeared to be.

Anyway, the book I read after the SSR was The Tao of Physics by Frithjof Capra, and I really could not understand the former until I read the latter. I recommend it.

I haven't read Kuhn, but I'm vaguely familiar with what he says, and besides, the notion of paradigms/forms of life/conceptual schemes crop up elsewhere, so commenting on them isn't a complete trek into the wilderness.

I would characterize a paradigm, to generalise it, as essentially being practice; the shared customs and habits exhibited in our day to day activities. Thus we might imagine a paradigm shift involving some individual or group breaking with traditional practices, and doing something else. If their behaviour finds a niche and catches on, perhaps because it allows us to do the usual things more easily, or even do new and exciting things, eventually supplanting the old practices, then we will say there has been a paradigm shift.

Some notions of incommensurability, I think, would be susceptible to the same arguments employed against private languages. For example, if we mean a Newtonian physicist could not, in principle, comprehend certain theorems of quantum mechanics, then clearly we would be wrong. All that would be required is for him to learn quantum theory. However, if all that we are saying is that it is impossible to formulate certain theorems of quantum mechanics in a Newtonian vocabulary, then that would seem to be true. We can say things of the same objects under discussion by Newtonians, cats, for example, in quantum talk, that we cannot say in Newton talk. In this case, 'incommensurable' might be too strong a word. Are we willing to say that QM contradicts Newtonian mechanics? That for 200 years all our physical statements were false, but now they are true? I think it would be better to say that they are two different vocabularies, and unsurprisingly, we can say and do things with one, that cannot be done with the other.

There's an article by Davidson called "On the very idea of a conceptual scheme" that I've been intending to read for a while now that deals with this subject. You might want to give it a read.


Surely 2+2=4 is just as dependent on our paradigms as anything else? 2+2=4 is merely a linguistic rule, we might imagine a linguistic community that had a different rule, along with a complicated system of other rules governing its use. I think the best way to reconcile truth with paradigms would be to pull a Rorty, so to speak. Say that Truth is a property of sentences/vocabularies, and the world does not speak, only people do; it may cause us to adopt certain beliefs, but it cannot propose a language for us to speak. This is not on the face of it to adopt idealism; we can still say that something is out there that is not caused by human mental states, but we must say that it has no intrinsic nature. We may re-describe the world as we please with new vocabularies. The world is out there, but truth is not.

How come, then, that the actual result of the calculation will be same in any conceivable language domain, and regardless of the terminology used to express it? I am afraid if you're a mathematical realist, and there are still a large number of them, this proposition will not be accepted.

Perhaps it's best to just drop what kind of truth Kuhn is arguing against -- I'm not sure which or what, and besides, he is stating that it's a clunky notion. Perhaps it is more clear to say: Kuhn is arguing against scientific progress towards a universal end-goal description often associated with Truth. I'm not sure I agree with this, but I think this better communicates what Kuhn is stating. I state this in response to your post, because I always took "capital-T" type truth to follow the correspondence theory of truth. I'm not sure how to think of truth in any other way, actually.

I agree that it's very interesting, though I'm not sure how you link it to modernism/post-modernism. Do you mind expanding on that? To me, it just seems like a sociological tract on the scientific community which happens to disagree with truth as an end-process for anything. People may value truth, still, but according to Kuhn, I think he's stating that no process we currently follow can legitimately claim that it's getting closer to truth, not that it doesn't want to or try to find truth. (EDIT: So, it doesn't get to truth in a teleological relationship -- truth doesn't pull it through some necessary string of events that some day will equate to, TADAH, pure objective truth, and science will be done. Instead, it jumps from where it currently resides to something else, and this something else is no better or worse than the previous paradigm, just different. I'm just putting this in here for clarifications sake. I don't know if I've communicated it clearly, and I think I've jumbled together a lot of ideas in interpreting Kuhn)

But this is very significant in its own right, isn't it? This is where we see the abandonment of the Quest for Truth, in the romantic sense of some all-encompassing Truth that all of us will bow to, and the emergence of the idea of truth as being something embodied in one or another cultural narrative: a religious narrative, or a scientific narrative, or a political narrative. This is how I see SSR as part of the emergence of post-modernism.

I know I am speaking in very broad terms here. My interepretation is largely historical. In the view of the Enlightenment, the enterprise of science was the pursuit of scientific truth, liberated from the smothering embrace of religious dogma. If you study the history of the Enlightenment and the Scientific Revolution, the key breakthrough was, very briefly, the combination of Cartesian mathematics, Newtonian physics, and Galileo's cosmology. In doing this the accretion of thousands of years of scholastic metaphysics and Aristotlean science was swept aside. Baron Holbach famously said 'all I see are bodies in motion'. LaPlace developed his statistical determinism. And so on - this was the Scientific Revolution. You may recall the statement of Lord Kelvin that by the middle of the 19th C, the physical description of nature was very nearly complete, save for the 'dark clouds' of the Michelson Morley experiments. And perhaps this is the nearest Science ever came to believing that they were nearing the Truth.

The great difficulty was, of course, that resolution of the results of the Michelson Morley experiments was one of the key drivers for the development of relativity. This of course was very soon thereafter followed by the discoveries of sub-atomic physics and QM. So in the space of a couple of hundred years, the picture had changed from the Medieval universe of Ptolmaic cosmology and 'the changeless superlunary spheres' to a mechanical universe governed by Newton's laws, to the current 'scientific cosmology' which is actually completely incoherent in any philosophical sense (what with dark matter, string theory and multi-verses).

So I personally think we are actually in the middle of another scientific revolution. Heaven knows what the end result will be, or how far off it is. But I think the model that was created for, and by, the Scientific Revolution and the European Enlightenment is well and truly broken, as broken as the crystal spheres that came before it. Hopefully the next one that develops will be intelligible, and user-friendly, but we really don't know yet.

So that is a very quick sketch of the 'story so far', but as I said in my previous post, it is greatly indebted to The Tao of Physics, which is not a mainstream publication, but which has helped me a great deal in making sense of what the hell is going on.

That's what I always thought, too, but Michael Polanyi in section 1.3 of Personal Knowledge (1958) quotes Einstein himself as denying it! Not only that, but the M.-M. experiment actually showed a positive result, which was confirmed in subsequent more accurate experiments of the same kind by D. C. Miller, as late as 1926.

From the detailed footnote on page 13, it would appear that it was not until 1952 that Synge explained the apparent ether drift as an effect of the circular rather than linear motion of Miller's (and presumably also M.-M.'s) interferometer. What appears to be a quite inconsistent explanation was given by some other authors in 1955, in terms of "statistical fluctuations and temperature effects". At any rate, other experiments of a different kind, including one by Michelson, Pease and Pearson (1929), have shown the absence of any ether drift, in line with the predictions of special relativity.

I can't remember if Kuhn commented on this (it's decades since I read the book under discussion), but I doubt if he would have been surprised that people believed in relativity in spite of the Michelson-Morley experiment, and not because of it!

The Wikipedia article http://en.wikipedia.org/wiki/Michelson%E2%80%93Morley_experiment says:

hmmm. I read Isaacson's recent bio of Einstein a couple of years back. I seem to recall it said that Einstein downplayed the importance of the Michelson Morley results in terms of their overall importance for his development. But I am sure they were one of the factors. Be that as it may, the point remains that Kelvin's assessment of the 'physical description of the Universe nearing completion' is nowadays a stock illustration of the dangers of scientific hubris.

In saying this I don't want to belittle the scientific outlook in the least. But I do think it ought to maintain a 'metaphysical modesty' which I am sure is characteristic of many of the great physicists, unlike the hordes of acolytes who draw unwarranted assumptions from it as a matter of routine.