Sokal Revisited
Contributed by: Brad Kelley
Yesterday 1gm linked to an article in the NY Sun by Michael Shermer reviewing Alan Sokal’s new book, Beyond the Hoax, which describes in more detail than he (Sokal) has done so far his views of the wonderful hoax he pulled and it’s massive, voluminous fallout. I don’t think I’ll read it, for reasons I give below, but I agree with most of what Shermer has to say. My field is Science Studies, the academic home of the very movements he decries (as do I)—postmodernism, social constructionism, and relativism (Shermer uses a slightly different list, and we could quibble, but same idea). This hoax is famous in our field, inciting eruptions of vitriol and intense emotion on both sides.
Still, the hoax wasn’t exactly what Shermer and many other scientific rationalists make it out to be. Lots of heat, no light, I think is the saying. The journal wasn’t peer-reviewed, and though it might or might not have been a major source that pointy-headed academics turned to, the editors of the journal weren’t, as I understand it, turning down many articles. If you sent something in, it pretty much got published. I don’t plan on reading Sokal’s book, for one because there are so many other good things out there, and from reading Sokal’s writings on this and other things before, well, let’s say that as a philosopher, he’s a very good scientist. “Ham-fisted” in making and keeping to subtle distinctions, one of my colleagues correctly said.
And my point here is that this issue –the nature of science, its alleged rationality, objectivity, and/or reliability--turns on extremely subtle distinctions. I’m in Shermer’s (and Dawkins’s, and Weinberg’s, and Gould’s, etc.) camp on this, but often Shermer (and quite especially the others) make it out like this is a knock-down drag out victory—the scientific methodology trumping the alternative at every turn, racking up all the argument victories, and the other side being pretty much the losers at every turn. This view has a derogatory name, “scientism,” and it is can be as dogmatic as any counterpart on the other side, sometimes equally blindly. I wish it were that easy, but that’s just not true, historically or philosophically. (Weinberg is the worst offender in this regard. Pick up anything he writes on the topic, and the attitude will be obvious: We are brilliant and rational; the other side is composed of morons.” Consider Shermer’s last paragraph in this light:
“There is progress in science, and some views really are superior to others, regardless of the color, gender, or country of origin of the scientist holding that view. Despite the fact that scientific data are "theory laden," science is truly different than art, music, religion, and other forms of human expression because it has a self-correcting mechanism built into it. If you don't catch the flaws in your theory, the slant in your bias, or the distortion in your preferences, someone else will, usually with great glee and in a public forum — for example, a competing journal! Scientists may be biased, but science itself, for all its flaws, is still the best system ever devised for understanding how the world works.”
Agreed. But there are problems with this view—my view, Shermer’s view –like there are with any theories. Take the one he mentions, the “fact” that the simplest, most basic of observations are ”theory-laden.” Ouch. First of all, many on the side of science fought the battle against this now commonly accepted view, and often used similar arguments against it as are used against the science critics now. And it is all well and good to say, “despite the fact that scientific data are theory-laden…,” but that theory-ladenness is a bedeviling problem for us! Add to that the under-determination problem and various skeptical conundrums that have built up over the years (since Phyrro and Sextus Empiricus, at least), and we have some real issues to solve. (In another parallel to mathematics, someone said that God exists because mathematics is consistent, and the Devil exists because we can’t prove it. That’s the boat we are in.)
To keep with the boat metaphor, Shermer points out that we are in Neurath’s boat—we are constantly pulling out planks while in the water, and putting in new and better ones; science indeed has self-correcting mechanisms. This is a nice feature of our rational methodology, but it is at best a necessary not a sufficient condition for a rational methodology. There are bigger problems we face on other scores; it isn’t our ablility to make minor (if important) corrections that is at issue—it’s generally our lack of arguments that we get on the right track in the first place.
In my view, what distinguishes our side isn’t (just) that we have a superior methodology, for that is notoriously difficult to demonstrate, but that when we have a problem, we recognize it and set to solving it. When Newton offered his proof that the calculus was giving correct answers, he made a bumbling error, one that, when pointed out, obviously needs to be rectified. Newton’s proof used the notion of an infinitesimal (though that wasn’t his term for it), and when his variable for it occurred as the numerator of a fraction, he let it be zero. When the same symbol appeared as the denominator, in which case it mathematically couldn’t be zero, he let it be a small but positive quantity. As the philosopher Berkeley, not otherwise much known for his mathematical acumen, pointed out, you just can’t do that. (He referred to infinitesimals as “the ghosts of departed quantities.”) But it was recognized as a problem after that. A big enough of one to give up using calculus? Of course not, but any educated scientist and mathematician would have recognized it as a real problem. Finally, in the 1800s, it was solved by Dedekind.
No other methodology does that. There are, I suppose, rival methodologies, but none is committed to solving their own difficulties. Of any competing methodology that invokes faith, I simply ask, point me to one interesting sentence that has ever been written about how, precisely, faith leads to something that is epistemologically worthy or reliable. One won’t find it. (Of people, like Gould and most likely Francis Collins, who think that there are two epistemologies that lead to different kinds of knowledge, what you won’t find—ever—is any account of how that second methodology works. But scientists, with the help of scientifically-minded philosophers, explicitly and willingly take on the tasks involved in specifying every step of their methodology.)
It bothers me that some on our side just don’t seem to understand that we have our problems, too. Hume’s problem of induction is a good place to start, but Goodman’s “New Riddle of Induction” is pretty good, too. The experimenters’ regress, and the Quine-Duhem problem (and various versions of that problem), work as well. Still, the basic problem is that we don’t really have an adequate account of how our methodology works, at least one that solves all or most of these problems. Logical positivism was an early, valiant and quite brilliant attempt to account for science, but, as is well known, it didn’t work. Since then, no other methodology has come to the fore, or become dominant. Bayesianism was what I cut my philosophical teeth on, and its beauty is that by just one simple little theorem (and a little interpretation), one might generate a complete theory of the rationality and reliability of science! But it doesn’t work
I say often that I am glad that fundamentalists are idiots, because if they understood Hume’s problem of induction, they would use it against us unmercifully. We have no answer. None. We’re working on it, though. By the way, how are you guys doing that Problem of Evil? What? You say, what problem? (In less obstinate moods I think there are some interesting things to say about Hume’s problem. Other times I have trouble recalling what they are.)
An underlying distinction I allude to here is between the critics on science from the left—the postmodernists, social constructionists, radical sociologists of science, call them what you will—and the critics from the right—the fundamentalists, creationists, and such. The latter are morons. They don’t deserve a second look, powerful as they can be politically in the US. The former aren’t, I argue. They are confused, but what they are getting at taps into something vitally important. They deserve a hearing, and they are not stupid.
The problems that are raised against scientific methodology, the good ones anyway, are epistemological problems. The radical sociologists of science who we rail against take the same problems, problems raised initially by Hume, Kant, Descartes and others in philosophical history, and bungle them, sometimes drawing unwarranted metaphysical conclusions about the nature of reality. (Harry Collins’s account of the problem of grue and the experimenter’s regress, which if I recall correctly occur in his book, Changing Order, is a case in point.) Here’s a bold and provocative claim, but one I will defend: the radical sociologists of science have added nothing to the arguments critiquing scientific methodology that wasn’t there before (often long before) but the intensity of their critiques has speeded up our progress by many years. They have forced us—some of us—to try to meet the challenges to scientific methodology head on, and not put it off until later.
(By the way, if you want arguments for anit-realism, don’t go through the radical sociologists of science/ postmodernists! There’s plenty of that in our own tradition, from Berkeley to Nelson Goodman and the later incarnations of Hilary Putnam in our time.)
The “Science Wars,” as they are called, are worth getting to know. Sokal’s attack was a minor skirmish, funny—a great horse-laugh-- but ultimately not too enlightening. The best book on the broader issues is probably Zammito’s A Nice Derangement of Epistemes, but it is heavy going. Haack is always good to read. Mayo isn’t as explanatory to the beginner as Haack, but her account of Kuhn, Lakatos, and Popper is alone worth the price of the book. It, too, is heavy going for the novice. Alan Chalmer’s What is This Thing Called Science is a must read, as is Ian Hacking’s Representing and Intervening, both very accessible. Philip Kitcher, too, has numerous books and articles that are excellent on these issues.
One last point. You hear people on our side condemn those on the other side because they aren’t scientists. Yes, you who are regulars at 1gm already see it—this is an ad hominem of the worst sort. First, the movement to critique scientific methodology, the one that led directly to the radical sociology of science, can be said to have begun with Ludwig Fleck and Pierre Duhem, a biologist and physicist, respectively. They are worth reading. Kuhn, who inspired so much critique of the rationality of science (eschewing that influence in later years), was a physicist. But science is, in the end, too important to be left to the scientists. All it takes to learn about these debates, crucial ones in our times, is read this literature and apply the canons of rational inquiry, broadly conceived, clearly and fairly.
Comments
Hi Brad. Good to see another post from you here, and I think this one was very thoughtful. I have some papers to go edit for colleagues, so I'll keep this quick (and a little dirty), and offer some suggestions.
First, in the end, if there are to be 'sides' on this issue, I come out on the one defending science as an ultimately rational but fallibilist enterprise; my problem with pieces like Shermer's (and those from the other 'side') is that they are mainly political, and have no intellectual content whatsoever; I would go further and say that Shermer's piece laughably naive and intellectually irresponsible, because it is obvious from his ignorant remarks that he has not read anyone he claims to be criticizing, all the while accusing them of cognitive bias (I made a comment along these lines yesterday). It strikes me that one might expect an empirically-minded person to go see for him or herself what claims are actual being made rather than reliance on prejudice and hearsay.
But at any rate, I appreciate the effort you've made to spell out in more responsible fashion some of the difficulties, philosophical and theoretical, encountered in practice of science. That said, here are a few problems I have with how you set out the issues. First, you offer a sweeping overview of the many problems with Shermer's (face-value) position. But, from a strictly stylistic point of view, I think it might be bewildering for a non-specialist to see all the problems, mostly because many of them you only mention, and sometimes you are not even consistent with the term you use. (For instance, you allude to the 'underdetermination problem', then, later, the 'Quine-Duhem problem'. So far as I know, these are the same problem, namely that individual statements are not verified or impugned one by one but only as cohesive parts of a larger theory. Therefore, the truth and falsity of any individual claim, taken by itself, is underdetermined by the empirical evidence which could be claimed to support them. Conversely, a given experimental result may support a number of different individual statements, taken by themselves). Perhaps it would have been more helpful to focus on one or two issues in detail, rather than mentioning (rather than explaining) a myriad of them. Another example:
In the first instance, Newton didn't use fractions in the Principia, he used ratios, drawn (mostly) from Euclid and Apollonius. (Fractions are a specific application of ratios, but not equivalent to them; in particular, they are typically used in expressions of relationships of proportion to one another, not equivalence. You will not see equals signs anywhere in the Principia). The point: you don't use ratios to express the division of one number by another but to show a relationship of one proportion to another (a:b :: c:d), so there is no problem about dividing something by zero and coming up with an undefined number. So long as the variable zeroed out (or not) on one side of the ratio is consistently used on the other side of a proportion, I don't see the problem). More to the point, since the use of Newton's infintesimals is to approximate quantities for approaching a limit--rather than ending up in paradoxes about how the limit is ever reached if there are in principle infinitely many points leading up to it--then it is not obvious why this procedure is illegitimate. Some more explanation would have been helpful, I think.
Second, I think some of your claims are too broad, as are also some of your targets. Such as:
Who is the target here? David Bloor? Foucault? Latour? Ludwig Fleck? Kuhn? (All of them? They are very different in style, methodology, tradition, inspirations, etc.) On the one hand, if the claim they would have added is that there are sceptical problems about induction, underdetermination, and so on, then you're surely right, but that's pretty anodyne. If the claim is that the more specific problems that they posed added nothing, then I think this is plainly false. Foucault, for instance, cites many predecessors in his historical analyses of the human sciences. A central problem of those investigations is the very status of certain human traits (like 'neurosis', 'insanity', 'criminal tendencies', etc.) within particular domains of putatively scientific inquiry, and the issues that this raises turn largely on questions about historigraphy, institutions, and social practices. (Ian Hacking has a nice, 10 page essay on some of these issues called Making People Up). I do not recall Descartes, Hume, or Kant discussing these issues; indeed, the very idea that there is such a discipline as 'the human sciences' esp. as an experimental (rather than introspective) mode of investigation, is not a central preoccupation of their work, except perhaps in implicit form.
Zzzzzzzzzzzzzzzzzzzzz
Thanks for your long and helpful response Adam. Though it surely will provoke lots of response of jpaul's type, too--Zzzzz. I do want to say that between my Word version and the posting on 1gm, a great many of the paragraph markers were lost, making it more confusing than it already was.
Several responses. I learned the Newton point in a grad class in Phil of Mathematics many years ago, and have forgotten some of the details. I know that was how it was presented by a wonderful professor, but need to look again at the literature on this. Berkeley's criticism is easy to find--if I remember correctly, it's in The Analyst. I'm pretty sure I'm right on the basics, even if my terminology is wrong. The mistake, which Berkeley is historically given credit for pointing out, is one of a pretty elemental fallacy of equivocation, which Dedekind explicitly sought to correct. I look forward to reading the Analyst again, and some of the secondary literature, but if you think my original post was a snooze, jpaul, this one would be excruciatingly boring.
Right, Hume, Kant and Descartes certainly did not carry on their discussions in the context of anything like what became “Science Studies” in the 20th century. Still, much of the radical soc of science seems to be bastardized, confused versions of their skeptical arguments (of course, in some cases, the arguments were intended to be AGAINST skepticism). There is extensive literature on Kant's influence on 20th century philosophy and postmodernism.
The end of your post makes me think you think the issue has something involving the human sciences per se. Though those might raise interesting side issues, the main issue, as I see it, is about ordinary science and its defense. So your claim that the human sciences didn’t exist in the time of Descartes, et al., is confusing at best.
I didn't have Foucault in mind at all. I frankly have to admit I have never gotten much out of reading Foucault, though I did try valiantly. I know lots of smart people disagree with me about his ‘worth,’ but when I ask these people why he matters, they either give me something mundane and platitudinuous, or something bold and provocative which, frankly, seems wacky to me. I will admit, the problem could be me, bad indigestion when I have read him or something. (Fallibilism starts at home, to paraphrase Quine.) Barnes and Bloor, Collins, Pickering, Pinch, Shapin and Schaeffer (sp?), are some of the names I had in mind. Yes, they are all very different, but drawing those distinctions would have resulted in something WAY too long--Norm was nice enough to give me the space he did (much to jpaul's dismay as it is). Latour I put aside a bit (he's a philosopher, not a sociologist for one thing); he's a more complicated case.
I certainly didn't mean Fleck--he had a lot of influence, but way before those I am talking about. I am talking about more recent work, post-Kuhn. In my attempt at brevity, I didn't make that clear enough.
Let me rephrase my bold claim (I TOLD you it was provocative--it provoked you!): Many if not all the arguments that have been used by the post-Kuhnian radical sociologists I allege to have been bastardizations of and based on confused interpretations of debates that already existed in 20th century philosophy, many of which existed in some form or other since Descartes, Hume, and Kant. While he might not agree with my claim, Hacking provides some evidence along these lines in The Social Construction of What?, which I should add to my list of must-reads in this area.
I didn't mean to seem to multiply problems by calling underdetermination by two names. I do think that, whatever you call it, it is a whole set of somewhat related problems rather than just one problem. Calling it Duhem-Quine is odd, historically, because Quine means something quite different by it than was intended by Duhem. There's some good literature on this, which I could provide references for on request.
The end of your post makes me think you think the issue has something involving the human sciences per se. Though those might raise interesting side issues, the main issue, as I see it, is about ordinary science and its defense. So your claim that the human sciences didn’t exist in the time of Descartes, et al., is confusing at best.
Yes, the my point is that a great, great many “defenders” of science completely ignore many of the great skeptical arguments that do indeed pose hurdles for presenting a rational scientific methodology. I WISH, Adam, that this point were anodyne. My point is that it SHOULD be anodyne. Why then don’t we find a lot more agreement on it among the otherwise smart people that are on our side in this?
Sorry about that Brad. I hadn't noticed that some of the paragraph breaks had been stripped out. I think I've found them all and fixed them.
Thanks for posting this.
This is because I didn't know whom you included in the "radical sociologists", and since the context was Sokal, I assumed French theorists (such as Foucault, among others) might be included. Was Foucault a 'sociologist'? Or Latour? Well, their influence on that discipline as well as anthropology and many other areas is immense. Indeed, most of the French theorists Sokal is criticizing are responding to currents in the French human scientific setting (linguistics, psycho-analysis, historiography, etc.), which he does not understand. But the list you provide is sufficient to see, now, what you mean.
Here I guess my point is how much emphasis you want to put on the 'one form or another': if you mean it as a very general claim about a certain general set of problems, I don't see how anyone could disagree. If you mean more specific problems, say, about the influence of social factors on the status and promulgation of scientific claims, then I think it's pretty controversial. Notoriously, it is easy to find continuity or discontinuity in history depending on one's own preferences, which is why a more exact statement might help.
But let me stop nitpicking and ask a simpler question, my "real" question. Why is it, given the many slight of hands and oversimplification you diagnosis in the so-called "defenders of science", do you insist on defending them? My own view is that one should simply, in one way or another, reject the terms of the debate. This is actually, kinda of at least, what you might be read as doing by problematizing some of their oversimplifications. For while I think the issues you raise are deep, difficult, and interesting, I think too often in these so-called 'cultural war' "debates" on the status of science, neither side is really talking to the other; it's more about sloganeering, oversimplification, and being already convinced (which is what I meant by "political"). One side will be full of moralisitic self-righteousness about being "defenders of science" or defenders of "truth" (whatever that means) and insist that there is a "real" world "out there" (whatever that means); I take it for granted that this is just empty sloganeering because, as I explained in an earlier comment, no one has ever denied, not seriously anyway, that the world exists, or seriously claimed that all claims are equally "true". The other side, likewise, pretends as if all scientists and its supposed "defenders" are somehow deadpan positivists, and then proceeds to attack the strawman thus constructed.
Why not just reject the terms of this non-debate? Why not just say, "defender of science" is itself a massive oversimplification that turns out not to be very helpful for thinking seriously about either science or its influence?
It seems to me your most immediate response might be--but correct me if I'm wrong--that you want to defend the rational credentials of science as a self-correcting enterprise. On this point, let me make one final distinction that I think might help. On the one hand, there is what we might call the cultural authority of science, its role in public life and claim vis-a-vis other modes of inquiry. On the other hand, there is its cognitive authority, which is what would answer such questions as: is scientific investigation the best way of settling such and such factual question? Granted, these are complexly interrelated: part of science's cultural authority, the prestige it is accord, is based on its supposed cognitive authority to get the facts right. But more generally it is not quite that simple, and it is simply a mistake to conflate these two issues (I don't think you do; I think the so-called cultural-war 'debate' positively invites us to). For instance, I think some people (legitimately) troubled by the overwhelming prestiged that is accorded to science as a cultural institution end up attacking its cognitive authority in order to deflate its cultural pretensions. But it seems to me one can have serious concerns about the role of science in public life and its status vis-a-vis other modes of investigation--for example, one might be skeptical that there is some magical method that will settle all our problems, or that science is somehow "more rational" than other modes of inquiry, or that other disciplines like literature are somehow 'second-rate' compared to science. One can be skeptical about all of this, I say, without denying, in the least, that science provides a useful set of conceptual tools for manipulating and controlling bits of the physical world (which is what I call it's 'cognitive authority' in the end amounts to for me).
Oh, and Brad, I meant to thank you for your generous and detailed response to my original comment.
Great discussion, folks. I will offer my own two cents (considerably less grounded and intelligent than yours, I have little doubt):
I think in certain respects some parsing is required about the nature of the disagreement. While I think in certain instances it's fair to claim that the latter sociological thinkers have "bastardized" earlier scholarship, it's also more than abundantly fair to say that the "postmodern" positions taken up in challenge to positivist science in the public domain (i.e. debates over "truth" "out there" particularly with respect to things like climate science) are bastardizations of postmodernism (to a degree--for the moment let's leave Derrida, Baudrillard and the like back in semiotics and linguistics where they belong).
From my own toe-dipping into this subject matter (not my expertise, though I have passing familiarity), it sure seems as if the problem with not having "more agreement on it among the otherwise smart people" to some extent comes down to either (or both): the difference between analytical and dialectical reasoning, and the fundamental hurdle to consensus that I see all the time in my field, which is the "war" if you will between qualitative and quantitative empiricism. I myself am more qualitatively oriented, so I am often in the position of saying, "Well, great you have numbers: but the way you established your data categories is just as subjective as what I do, only you're not remotely reflexive about it." For my part I can talk about "what's happening" and be extraordinarily interpretive and reflexive, but I don't have "firm numbers" to back up these claims. It's possible to integrate the two approaches, but it's never perfect.
Anyway, what I'm trying to point out is the different "regimes of knowledge" if you will that are acceptable means of getting at the "truth" of the matter, but which will always be subject to revision, are always conditional, and always (and this is where Foucault comes in, Brad) subject at least in part to the "historicity" of Knowledge or Truth in its time. The endpoint of this Foucauldian line of reasoning can be "bastardized" to infer that all empiricism is inherently corrupt; whereas the actual point, as I understand it, is to recognize the degree to which even the very best science is conducted in a socio-cultural milieu/paradigm, and we should watch for these inflections.
So: some of the divisions we see here among "smart people" are folks who cannot, as Brad says, "digest" the real thrust of the other paradigm. And here's the real kicker: I get the sense (although by no means is this anything other than a hunch) that SOME of this may simply come down to brain wiring.
As a parting shot: "postmodernism" gets a bit of a snarky dismissal among most scientists, when it's clear they don't understand any of the stuff there, if they've even bothered to read it. Dawkins is exhibit A. The same is true on the other side, of course: I've met many people in the Social Sciences/Humanities who don't know some very basics about what happens in "real" science. Sometimes being smart and educated just isn't enough once you start to specialize and/or have to deal with the grind that is modern scholarship, particularly in universities.
Post a comment