DOUBLE
REFRACTION
Looking twice at the history of science

Tuesday, October 7, 2014

Barry Barnes' Scientific Knowledge and Sociological Theory, 40 years on

2012 was the fiftieth anniversary of Thomas Kuhns’ influential and controversial book The Structure of Scientific Revolutions. 2014 is the anniversary of a book that was nearly as influential and nearly as controversial as Kuhn's, at least among historians and sociologists of science. Barry Barnes’ Scientific Knowledge and Sociological Theory was the first full-length exposition of what soon became known as the Strong Programme in the sociology of knowledge. The programme was ‘strong’ in the sense that it used sociology to explain established scientific theories, as opposed to explaining scientific institutions or explaining discredited beliefs. When I read the book last week I found it surprisingly radical and surprisingly prescient. I also found what I think are the roots of a gross error that persists to this day. Expand post.

10 comments:

  1. This intricate issue may get simpler, when we start with principles that scientists actually do follow rather than with one that historians ought to follow.

    For example, take Quine's "natural tendency to disturb our system of beliefs as little as possible." It has also been called the maxim of minimum mutilation. Taken individualistically, it could simply mean that scientists try to conserve their own previous work, views etc. as valid as possible. They stick to their guns, they try to be consistent, whatever you want to call it. (There may be other such rules scientists tend to follow and I neglect the rare exception of a complete reorientation here.)

    If scientist X and Y meet with a new fact or problem, they will both try to explain it in ways that preserve their respective earlier views, publications etc. without doing too much harm to them. As neither is clairvoyant, the fact that, say, X had the better views, publications etc. than Y emerges only in retrospect. Hence both scientists followed the same (social) rule, and the good reason/bad reason distinction is a retrospective one.

    From this, the symmetry principle seems to be no more than a caution against Whig retrospective to me.

    ReplyDelete
    Replies
    1. Dear Joachim,

      [This comment may appear twice due to technical glitch. If it does, please note that the two comments are not identical and that you are now reading the right one.]

      Thanks for your comment, and apologies for my long delay in replying.

      What do you mean when you say that 'the good/bad reason is a retrospective one'?

      Do you mean that Xs reasons were in fact better than Ys reasons (that is, no more conducive to true beliefs than Ys reasons)? Or do you mean that Xs reasons were no better than Ys reasons, and that historians are making things up when they say otherwise?

      If the latter, and if the point is generalised, then we are forced to conclude that there was no point in the past when the reasons in favour of our current beliefs were stronger than the reasons against those beliefs. If that is the case, then why should we continue to hold those beliefs? Too much symmetry about past beliefs leads to skepticism about present beliefs.

      If the former, then what's wrong with the historian reporting the fact that X's reasons were better than Y? Granted, neither X nor Y were aware of this fact. But historians routinely describe events that were beyond the epistemic reach of some or all of the actors who were involved in those events. Consider descriptions of the posthumous influence of the writings of an author, or descriptions of how a living person is perceived by people they are not acquainted with, or descriptions of the subconscious causes of a person's actions.

      This whole question may be moot, however, because I am not the only one who insists on distinguishing between the goodness and badness of the reasons for past beliefs. The same distinction is made by anyone who accuses a historian of violating the symmetry principle. Making that accusation means saying a historian has explained a true belief in terms of good reasons and a false belief in terms of bad reasons.

      Consider that hoary classic, Leviathan and the Air Pump. There the authors accuse historians of taking an asymmetric approach to the dispute between Thomas Hobbes and Robert Boyle regarding the possibility of a vacuum. They say that historians have explained Boyle's view as a consequence of his experiments, and explained Hobbes' view as a consequence of his misinterpretation of Boyle's experiments. Why do Shapin and Schaffer say that this is an asymmetric approach to the issue? Presumably, it is because they believe that 'making inferences from experiments' is a good reason for believing something, whereas 'misinterpreting someone else's experiments' is a bad reason for believing something.

      Delete
  2. I am unclear about exactly what you think Barnes's error is. Sure, philosophers have been interested in specifying the causes of, or constant conjunctions associated with, justified true belief. But presumably Barnes's point is that philosophers are committed to the position that that species of causation is ultimately internal-- it has to do with the causal consequences of rational minds engaging with the world. That's not at all the same thing as the external, social causality Barnes has in mind. Now you can, in a Mertonian spirit, specify social conditions of possibility for mind being able to engage with the world in such a fashion that it is caused to come to justified true beliefs. But as you know, the Mertonian paradigm is exactly what Barnes is arguing against.

    ReplyDelete
    Replies
    1. Hi Stephen,

      Thanks for your comment. If Barnes wrote only what you just wrote, then I would have no real complaint. But that is not what he wrote, and it is not how he has been read by most later writers on the symmetry principle.

      In the passages that I quoted in the above post, Barnes does not say that philosophers have focused on one kind of cause of belief at the expense of another kind of cause of belief. Instead he says that philosophers' accounts of belief are not *causal* at all, or perhaps that those accounts only invoke a trivial cause of belief, namely the natural world. I find the former view in the second Barnes quote ('...this particular perspective...' and 'Science is conceived...') and the latter view in the first Barnes quote ('Many academic theories...).

      I admit that Barnes and his followers have sometimes written roughly what you wrote in your comment. For example, in 1976 Bloor defined the symmetry principle as the view that all beliefs should be explained with 'the same type of cause.' There are two caveats, however.

      Firstly, sociologists have rarely disentangled this reading of the symmetry principle from the view that Barnes presented in the passages I quoted above.

      Secondly, the Barnes reading is the dominant reading in the literature. It dominates Bloor's 1976 book, despite the 'types of cause' definition that I just mentioned; it dominates an important 1982 article on the symmetry principle by Barnes and Bloor; it dominates Collins' writings on the symmetry principle; and it is the reading that motivates Golinksi's claim that historians of science had not studied science as a 'human product' prior to the late twentieth century.

      Perhaps Barnes was using the word 'cause' in an unusual way in his 1974 book. Perhaps he really did have in mind only the external, social causes you mentioned in your comment. I can kind of see why he might have used the word in this way, since it is natural to think that our mental processes are autonomous in a way that our socially-determined behaviour is not autonomous (although later sociologists of science, such as Shapin, have rejected this distinction and accused philosophers of science of insisting upon it!).

      However this terminological peculiarity does not let Barnes off the hook. He accuses philosophers and sociologists of explaining true beliefs as the result of 'a direct apprehension of reality' or as beliefs that 'derive directly from awareness of reality.' This is an unfair characterisation of epistemologists as diverse as Francis Bacon and Karl Popper. There is nothing 'direct' about the method that Bacon proposes in his New Organon or that Popper put forward in his Conjectures and Refutations. Both authors describe long, complex, fallible processes by which scientists finally arrive at true beliefs.

      Delete
    2. The argument about Barnes making a big error, in (apparently) arguing that historians and philosophers have claimed true beliefs are uncaused, is flat wrong. Barnes never claimed as such, even in the quotes you cite. Barnes said some philosophers write as if true beliefs are their own cause. Not uncaused, but their own cause. Your claim of Barnes' error relies on an elision. Of course, you could just read Bloor's 1981 article on the strengths of the strong programme, where Bloor notes that Laudan makes just the same elision you do.

      Delete
  3. Hi Darrin. Thanks very much for reading the post, and for the comment. Thanks also for drawing my attention to Bloor's 1981 reply to Laudan, which I had not read when I wrote this post. You are right that Bloor says in that article that he wrote about beliefs causing themselves, not about beliefs being uncaused. But I don't think that makes much difference to Laudan's case or to mine, for two reasons.

    Firstly, I don't agree with your reading of the passages I quoted above (from Barnes rather than Bloor). Consider the accusation that philosophers have been 'treating truth as unproblematic and falsehood as needing causal explanation.' Surely what Barnes means here is that (according to said philosophers) truth *does not* need causal explanation, not even a causal explanation that appeals to the truth itself.

    Secondly, Bloor's claim that philosophers have treated true beliefs as 'self-explanatory' is little better than the claim that Laudan and I attribute to him. What Bloor means by this, to judge from the examples he uses in the 1981 paper, is that philosophers have explained true beliefs by appeal to the truths they refer to, eg. they have explained Galileo's belief that the moon is rocky by appeal to the rockiness of the moon. Now, many philosophers think that the rockiness of the moon is *part* of the explanation of Galileo's belief. But I do not know any philosophers or historians of science who think that it is a *complete* explanation. A complete explanation would include references to Galileo's telescope, is beliefs about how telescopes work, his drawings of the moon, etc.

    ReplyDelete
  4. Michael, I think you either fail to adequately grasp the Barnes-Bloor position you critique, or this is a case of refusing to see the overall argumentative structure. For charity's sake, I'll eliminate the former option. Let's take your first claim, where you write that "surely" Barnes intends to mean that some philosophers think truth does not need causal explanation. The problem here is that this was the claim in dispute, so it helps little to plead that surely Barnes must mean what you misinterpret him to mean. No, Barnes means what Bloor means and Bloor, in that 1981 article, directly quotes Laudan saying that if a belief is rational one needs to enquire no further into causes. Which implies a self-explanatory model of causation. I think in this instance you would be better accepting you misinterpreted and deal with the bigger explanatory story at stake.
    To wit, your second claim. "What Bloor means by this" is followed by not explicating what Bloor clearly says he means by this, which is that philosophers often divide causes into two species, thus ending up with an asymmetric explanatory enterprise. That the rockiness of the moon is part of the cause of Galileo thinking the moon is rocky is tagged to a point about philosophers not assuming that one chain in the causal elements is sufficient. But the insufficiency of one causal factor is irrelevant to an argument that points to some types of causal factors being privileged, like some great hierarchy of being of types of causal species. You could perform your same lack of sufficiency defence on all the examples Bloor cites and skim over the crucial point about species of causes in each case. This, to me, suggests an unwillingness to keep the overall argumentative structure of the symmetry versus asymmetry debate, and the old rationality debate, in clear view. Once it is in clear view, your charge against Barnes fails and your attempt to evade Bloor's point fails too.
    It's good that philosophically inclined historians still find these issues important, but it would be even better if we kept the structure of the arguments in view too.

    ReplyDelete
    Replies
    1. Hi Darrin, and apologies for the very late reply.

      On what Barnes means: the best way to find out what Barnes meant in his book is to examine passages from it. In my last post I quoted a passage that says - nearly in so many words - that only falsehoods require causal explanations. You wrote in your last comment that this is 'the claim in dispute.' Precisely! And that is why I chose that passage to resolve the dispute!

      On the 'two species' reading of Barnes: as you say in your last comment, Bloor and Barnes accuse philosophers of dividing scientific beliefs into two 'species' and of engaging in a different 'explanatory enterprise' for beliefs of each species.

      But what is the difference between these two explanatory enterprises, according to Barnes/Bloor? Is it the difference between explaining a belief and not explaining a belief, or between explaining a belief in terms of itself and explaining it in terms of something else (the 'self-explanatory' view you mentioned in your earlier comment), or between explaining a belief in terms of epistemic factors (experiments, observations, and suchlike) and explaining it in terms of social factors (self-interest, professional interests, self-fashioning, and suchlike)?

      My view is that Barnes and Bloor slide between these three different versions of the symmetry principle (and between other versions as well). In different places they state each of these versions, and in other places their statement of the principle is vague enough to accommodate all three.

      This vagueness helps to explain why leading historians now make the absurd claim that we have only recently discovered that science is (to quote Golinski) 'a human product.' We have always known that science is a human product, that is to say that scientific beliefs are caused by human thoughts and actions. We have (perhaps) not always known that many true beliefs are caused by human thoughts and actions that are *social* in nature. Barnes and Bloor conflate these two points, and as a result they lend credibility to an absurd claim by coupling it with a plausible one.

      Delete
    2. Michael, you suggest if we go back to Barnes and quote him then all can be resolved. But your actual method is at variance with your claimed method. You quote Barnes from SKST in order to show that Barnes is saying historians believe true beliefs cannot be explained causally. You critique Barnes for supposing historians or philosophers really think true beliefs do not have causal explanations. Now, I replied that you misinterpreted Barnes, collapsing 'their own cause' into 'uncaused'. I suggested Bloor's discussion shows the error of your ways. You then sought to evade Bloor's refutation of your misinterpretation of Barnes, but I do not think your reply escapes my critique that you are not grasping the shape of the argument (species of causes).
      So let's return to your assertion that if you just go to Barnes, your critique of Barnes stands. No, it does not. And it is easy to show. All one has to do is actually just keep reading. On p. 4 of SKST Barnes says "it should be noted at this point that it is only causal elucidation by reference to bias, or interference with normal faculties of reason and cognition, which is held to be inapplicable to true beliefs. Other kinds of causal account remain possible". I think, then, if you follow your method and go to Barnes, your reading of Barnes does not hold water. Note also that your original framing of Barnes' argument had Barnes answering YOUR question of whether true beliefs can be explained causally or not. But that is not what Barnes was discussing, instead Barnes was discussing the "shape of explanation" point I claimed you were not grasping. So in fact your apparent close reading of Barnes is actually the kind of whiggish transposition of question that you critique. Thus, Barnes explicitly frames his discussion of truth and causes with the idea that academic theories follow common sense theories (p. 2), where what we take to be true is treated as unproblematic and what we take to be false is treated as needing some elaborate explanation to account for divergence from the obvious. You can invent multiple versions of the symmetry principle all you like, but the asymmetry involved in that common sense explanatory scheme is all you actually need. Barnes then notes on p. 4 that what is shared by many explanatory approaches is a "condemnation" of causal explanations of beliefs, signalled by thinking we need causes when there is distortion from the taken for granted. On. p. 5 we then see Barnes saying "the idea of truth as a normal, straightforward product of human experience" is the claim he is interested in dismantling. So on p. 7 when you cite Barnes again referring to "not in need of causal explanation", Barnes has spent several pages talking about what counts as causes and what does not in that phrasing of the issue, but your argument has to overlook all of it. So where really is your vaunted method of "going back to Barnes"? I think, instead, you are telescoping your reading to support a very tendentious misinterpretation. Ultimately, two possible readings of your misinterpretation of Barnes arise. One, you never really grasped the point of symmetry, indicated by an unwillingness to really work with the 'shape of the argument'. That is, you fail to understand the point about two species of causes. A second reading might be that you really do understand the point about two species of causes, but for some reason want to evade its consequences, and so you invent some critique of Barnes to muddy the waters. I have no idea which reading is true, but textually, I don't see much evidence that you have got inside the two species of causes idea. I worry about the implications for historical practice that implies!

      Delete
    3. Darrin, Thanks for pushing me on Barnes. Yes, he does make the 'shape of explanation' point that you refer to, ie. that academics have tended to give one kind of explanation for true beliefs and another kind of explanation for false beliefs. But this is consistent with the above post and with my comments in this thread. In my post I argued that Barnes attributed a different view to academics, ie. the view that true theories have no causal explanations. The point I made in my last comment was precisely that Barnes attributed *both* of these views to academics, ie. that he was unclear and/or inconsistent. After reading your comment I re-read the first chapter of his 1974 book, and this has not changed my view that Barnes conflated the two attributions. I will lay out the evidence for this reading in a new post in the next few days.

      Delete