Watch your language, Pt. 2: Galison vs. Staley

In Pt. 1, I discussed the historiographical problem of under what circumstances it is useful to criticize someone else’s characterization of history, highlighting Peter Galison’s rebuke in Image and Logic to Andy Pickering’s account of the discovery of the J/ψ particle from Constructing Quarks.  I noted that Galison took exception to Pickering’s idea of “tuning” experiment to theory on the count of its adherence to an antipositivist understanding of the history of experiment as proceeding in some sort of theoretical relationship to theory rather than on its own terms.  This independence of experimental tradition from theoretical concerns is part of a useful view of history Galison calls “intercalation”.  I noted that the issue of theory-dependence can have political overtones, but that the issue is also important to understanding how knowledge-production works, and to constructing coherent and accurately-worded historical accounts.

But just how important is accuracy in wording?  When is one making a point and when is one just nitpicking?  To address this question I want to skip ahead a couple of years to a special issue of Perspectives on Science dedicated to Image and Logic in which philosopher of science Kent Staley disputed Galison’s division of modern particle detection into epistemologically distinct “image” and “logic” traditions.  Galison responded in the same issue entirely confident that he was being visited by some easily vanquished ghost out of the historiographical past.  Yet, to my mind, this is a dispute that Staley won.  I’ll explain why, and then get on to the ultimate question of whether it matters.

First off, I should say that I’m predisposed to Staley’s argument.  When I first thoroughly read Image and Logic in 2003, as impressed as I was (and still am) with the historiographical craftsmanship, I felt that Galison was making too much of the distinction between the two traditions, primarily since those working in the logic tradition always seemed so eager to move toward the tracking of distinct particles by densely packing counter detectors together.  That they did not work in that tradition from the start always seemed to be more of a nod to the practical advantages of using their brand of detector (which Galison details superbly) rather than any fundamental adherence to the epistemology of logic detection.  But I never thought too much about it.  However, when I recently stumbled across this issue of Perspectives on Science, I was interested that Staley had had the same impression, and I was eager to see what he had to say.

Staley, it turned out—technical philosopher of science that he is—came at things from the opposite perspective.  To his mind it was statistical argumentation—the epistemological hallmark of the logic tradition—that permeated through all particle detection to the image tradition.  His paper details how experimenters in the image tradition often used statistical arguments in order to argue that their vast collections of images depicted the things that they said they did.  Similarly he shows how the obviousness of evidence to be found in so-called “golden events” could also be found in some experiments in the logic tradition.  In a stretch, Staley goes so far as to say that even obvious evidence, as in a golden event, could be seen as using the same statistical standards of evidence, except that the clarity with which the experiment met the statistical standard made it possible to leave that standard implicit and unaddressed.

The status of Staley’s claims is very important.  Staley was very careful not to deny that there actually were image and logic traditions—what he is doing is attacking Galison’s characterizations of the traditions as operating in distinctly binary epistemological modes.  Where Galison sees a marker of the logic tradition’s epistemology in the phrase “anything can happen once”, Staley suggests it “has to be taken with a grain of salt”.  It is a “slogan” (p. 225), i.e. something that magnifies the differences between two sides for the sake of defining them as sides.

Staley, in the end, might have overplayed his hand a little bit, primarily because he considers the statistical and epistemological affinities between image and logic traditions to be indicative of a “unity” of scientific method—a phrase he repeats and emphasizes, and that was sure to draw comparisons to older generations of positivistic philosophies.  In any event, Galison had then recently finished co-editing a book called The Disunity of Science (1996), emphasizing just what a fractured and discordant—and thus wonderful and powerful!—thing “science” historically has been.  The point might have been better taken if Staley had said something along the lines of  the traditions being epistemologically “contiguous” rather than “unified”, but, in any event, he had made his claims clear enough.

Where Staley insists, “My finding is modest,” Galison sees a fundamental attack on his historiographical methodology, and will have none of it: “Kent Staley wants to re-hoist the venerable banner of the Unity of Science movement,” he begins.    A few lines later: “Where I take Staley’s project to derail is in reducing exprimental argumentation to the application of statistical reasoning” (Galison’s emphasis).  Galison goes on to reiterate the fact that the object of detection in the image tradition was to produce indisputable visual evidence of phenomena, which was something that was impossible in the logic tradition, which had to use statistical argument.  Bubble chamber physicists might use statistical arguments (as in “bump hunting”, which Staley describes in some detail), but it was not at the root of their work.

I don’t buy it.  In arguing for unity (or continuity) between traditions, Staley was making a pointed critique of Galison’s characterization of the tradition—his adjectives, essentially—he was not suggesting that the image tradition be reduced to some fundamental statistical structure.  By framing Staley as an old-fashioned reductionist, Galison turns the tables and puts Staley on trial thereby reducing the burden on his own work.  Galison helpfully offers Staley some fallback positions to his ill-fated program: if statistics is to be taken as funademental, even in golden event experiments, perhaps it is at the level of cognition rather than articulation.  Or, perhaps the philosophically trained historian can go back and, à la Imre Lakatos, rationally reconstruct the logical progression of science regardless of the actual path of its historical development.

It’s a false choice—Galison wants Staley’s head for his crime of adhering to an obsolete program.  The only question is the means of execution.  The first is contrary to experimental results in cognitive psychology; the second (which Galison feels Staley would prefer) is contrary to good historical method: “This would be neither science, nor philosophy, nor history—it would be a festival of just-so stories.”  Galison appeals to the standard of historiographical coherence at the very end of his attack, too: “Why inflict such reconstructive violence on the historical record?”

What Galison appears to have missed is that it is Staley who is accusing Galison of historiographical incoherence.  No petty attack on adjectives—Galison did, after all, name his book after the divide—Staley insisted that these adjectives are central to Galison’s understanding of the emergence of “hybrid” detectors in the 1970s, and his central mechanism of historical change, the creation of “pidgin” and “creole” languages in scientific “trading zones”.  For Staley, though, no such contrivances are, in this case, necessary, because the lack of a profound (one might say incommensurate) epistemological difference allowed image and logic physicists to understand and appreciate each others’ work perfectly well.

As Staley points out (p. 225), it’s certainly not as though bubble chamber physicists refused to allow the arguments for the detection of particles in spark chambers.  Indeed, though their work habits and interpretive methods differed, there was enough argumentative overlap already to allow experimenters to assess the strengths and weaknesses and complementary qualities of either approach.  Thus when the practical means became available to join the strengths of the traditions, they were seized upon—not inevitably, certainly, but logically enough.

Staley characterizes physicists’ adherence to one or the other tradition as based upon an evaluation of the prospects of either—which one might, practically speaking, yield more and better results (p. 225)?  I would go further and suggest that the primary difference was one of personal preference.  As Galison details, different chambers entailed abiding by the different constraints to be found in different work environments, and with making different kinds of arguments.  Thus, the choice to work in one or another tradition would be based on personal background and habitual preferences for certain styles of work and argument, but almost certainly not on the fundamental cognitive differences between looking at images and doing statistics that Galison seems to rely upon in formulating his attack on Staley.

But does it, in the end, matter?  I think so.  Galison’s history still stands perfectly well—Staley never challenged that, and I also don’t see the narrative of  Image and Logic as altered in any significant way by his argument.  Nevertheless, if Galison can go after Pickering on the details, I don’t see why Staley can’t go after Galison.  If it matters whether the experimental process of bump hunting took place according to practices specific to experimental tradition or according to a practice undertaken in view of higher theory, surely it matters whether physical experimentation developed in divergent or convergent epistemological traditions.  Whether Staley was right or wrong, that’s a case that needs to be taken more seriously than Galison took it.

Technical point: One of the arguments Galison used against Staley was the assembly of “atlases” of images from experiments in the image tradition (an important point in his new book with Raine Daston Objectivity as well), which he takes as a distinguishing feature of that tradition that could never be shared by the logic tradition.  He quotes Manchester experimenter Patrick Blackett to the effect that it is only through the study of such atlases that an experimental eye can be developed and the “hitherto unknown be detected”.  Galison neglects the fact that the interpretations of these images were not born obvious, and, as he well knows, Blackett lost out on winning a prewar Nobel Prize for the discovery of the positron to Carl Anderson because he, unlike Anderson, was unsatisfied with apparently golden events.  As he wrote in 1934: “The experiments were … continued until about 4000 tracks of electrons and about 400 tracks of positrons had been obtained, giving a body of evidence sufficient to justify quantitative conclusions.” (This is quoted in my own, “The Heuristics of War: Scientific Method and the Founders of Operations Research,” BJHS 40 (2007): 251-274; also note that Blackett’s images were not as clear or dramatic as Anderson’s.)  Blackett did not want to be accused of “tuning”.  Clearly, then, at least Blackett appreciated that there was a contiguousness between the interpretation of visual and statistical forms of evidence.  He, of course, would go on to win the 1948 Nobel Prize for his innovation of attaching counters to a cloud chamber to signify the passage of a charged particle through it thereby triggering a camera to record an image of the passing.