The Bactra Review: Occasional and eclectic book reviews by Cosma Shalizi   110

Against Relativism

Philosophy of Science, Deconstruction and Critical Theory

by Christopher Norris

Oxford: Blackwell, 1997

On Deconstructionist Realism

Christopher Norris is a professor of philosophy at the University of Wales-Cardiff. This is the latest in a series of books and collections of his essays whose thrust is telegraphed by their titles --- Uncritical Theory, The Truth about Postmodernism, Reclaiming the Truth, New Idols of the Cave, etc. Norris, as might be imagined, is in favor of the Truth, and he likes Reality and Reason too, though generally without the capital letters. In fact Norris is a flat-out scientific realist, and thinks that quarks and leptons (for example) really exist and are more or less as advertised in the Particle Properties Data Booklet. He entertains grave suspicions about instrumentalists, who regard scientific notions as more or less convenient fictions, more or less useful conceptual tools --- for how could they be so convenient if they weren't, pretty much, true? For cognitive relativists, who are (these days) claiming that scientific ideas are convenient fictions alright, convenient for capitalism, patriarchy, the military-industrial complex, and the other usual suspects, he has unmitigated contempt. I share all these convictions, of course, but I'm less serious about them than Norris is, since I don't rub shoulders with those who disagree on a daily basis.

What sets Norris apart from your garden variety anti-relativist is his passionate commitment to the deconstruction of Jacques Derrida and Paul de Man, to whom he has devoted a book each (plus several books on deconstruction-in-general). In fact, he goes so far as to claim that Derrida and de Man, properly understood (by Norris, for one) are actually on the side of the angels, or at any rate that of the decent materialists and realists. There is a certain piquancy to seeing the whole apparatus of trace, differance, aporia, etc., etc. being turned against various and sundry Foucauldians, social constructivists, "strong" sociologists of knowledge, Richard Rorty, post-modernists, Heideggarians, practical manglers, etc., and it's amusing to watch Norris flail at, essentially, everyone else who cites Derrida as an authority for getting him so flagrantly and ridiculously wrong. Far be it from me to take a side in this quarrel.

The book under review is a collection of essays, mostly previously published in various learned journals. These are written to explode relativism by showing that arguments advanced in its favor are tosh, or to castigate writers (Quine, Kuhn, van Frassen) who, while not relativists themselves, have propounded influential arguments which relativists have used, or to expound the thoughts of Derrida and de Man on realism. There is a great deal of overlap and repetition between the essays, entirely natural in their original setting, but not corrected here.

There is also a vast amount of repetition within each essay. When Norris wants to argue some point, he begins by stating some premise, and then re-stating it, and then re-stating it again, tweaking it slightly each time, and giving more or less (generally less) explicit justification for the change, until he finally drifts into his conclusion. Not especially concise to begin with, this diffusive procedure gives him one of the lowest densities of argumentation per page of any philosophical writer who does, in fact, argue. That initial premise is often taken from some other writer; in these exegetical cases it is sometimes very hard to tell when Norris is speaking on behalf of the author he is interpreting, and when he is speaking to us in his own voice. I sympathize, since I have (as constant readers have pointed out) the same problem, but sympathy doesn't inhibit annoyance.

One needn't write obscurely about obscure writers: vide Kolakowski on Husserl, or Norris himself on Heidegger. Thus I get suspicious when the glass darkens, which it does whenever Derrida and de Man are invoked, which is also when Norris goes into full-throttle exegesis. "These statements [about Pascal's style and his contrast between "real" and "nominal" definitions] are characteristic of `late' de Man in the impression they give of raising fundamental issues in a style of extreme elliptical precision which somehow omits --- or disdains to make explicit --- the most crucial argumentative moves" [p. 51]. This is extremely charitable, both in conceding precision, and in supposing that those key moves were made. Far be it from me to dispute that de Man is elliptical, disdainful of his readers, and talking about a very important problem --- here, the relation between empirical knowledge, and the kind we have in math, logic, and other pure deductive systems. That he gets us any further in understanding this nasty mess is altogether a separate question. The best evidence I can see that he does is that Norris, a smart and well-read man, says so with every sign of sincere conviction. I note that many people who have read these authors have concluded that they're relativists at best, but that Norris does not take them to task for this, as (in the analogous situations) he chides Quine and Kuhn.

Maybe the deconstructionist masters really are as much in favor of truth and reality as Norris wants us to believe; maybe they support motherhood and apple pie, too; maybe not; howsoever, to paraphrase Leszek Kolakowski's remark about Marx and democracy, there are better arguments in favor of reason and realism than the fact that Derrida is not quite so hostile to them as he at first appears. The one which Norris hammers home in this book is the one compressed into Richard Dawkins's quip that "there are no social constructivist at 30,000 feet", and Ian Hacking's dictum about positrons: "if you can spritz something with them, then they are real". That is: if science really was just so much ideological baffle-gab, why would it work, which it very, very plainly does? And it doesn't just work in technological applications, convincing though those are; even very recondite and conceptual aspects of sciences like special relativity and quantum mechanics can have observational, empirical consequences. (Curiously, for all his interest in quantum mechanics, Norris seems unaware of recent relevant work by physicists on topics such as decoherence and the measurement problem; nor that we never talk about "complementarity" any more, thank the gods.) Of course this doesn't mean that the sciences haven't also been tools of capitalism, imperialism, and all the rest: but the reason chemistry is a much better tool of imperialist domination than alchemy is that it's much more true. (He doesn't put it quite that way.)

Norris has an excellent dissection of the relativist position which Larry Laudan has called "skepticism about everything except the social sciences." This is the idea that, while reliable scientific knowledge is unobtainable, reliable knowledge of the social, political and psychological motivations of scientists is so easily obtained it can be laid on with a shovel. Whether (say) an air-pump actually creates a partial vacuum or not is unknowable; what led some centuries-dead people to agree to or dissent from this proposition can be determined with precision and certainty and set forth in a book which can be read over a weekend. In a related vein, Norris has re-analyzed some case-studies beloved of cognitive relativists in some depth, coming, naturally, to different conclusions. Another very interesting study is on the development of jet engines, which relativists don't talk about, but which should worry them a good deal.

I will now pick a number of nits.

Norris talks about Gödel's Theorem, of course. His basic point is right: it does not spell the end of mathematical knowledge, much less complete epistemic free-for-all. The theorem (in the strengthened version of Quine) asserts that, in any consistent axiomatic system in which we can construct basic arithmetic, there will be at least one true but unprovable statement. (The condition can fail in two ways: either the system is too dumb to be able to count, or it's inconsistent.) As Jaako Hintikka notes, this is strictly equivalent to asserting that there is no computer program which, if we let it run forever, would spit out all and only the true statements of that system, but the first way of putting it sounds much sexier. Norris gets this right (pretty much) in two of the three essays where it comes up. On p. 22, however, he gives the following gloss on the theorem: "there will always exist at least one indispensable axiom whose validity cannot be proved in terms of that same system." This is bizarre, since, in the first place, the true theorem (which, I repeat, Norris gets right elsewhere) is about what can be proved starting from the axioms, not whether the axioms themselves can be proved, and, in the second place, if we could prove an axiom using the others, we'd call it a theorem and remove it from our list of axioms! Note that the analogy Norris draws (a loose one, by his own admission) to Derrida's idea that every philosophical system relies on at least one ineliminable metaphor would collapse if we substituted the true theorem for the mutilated one.

Norris objects to the semantic notion of truth (the notion that the sentence "Ripe cherries are red" is true if and only if ripe cherries are red, and similarly for other sentences and their "disquoted" counterparts), regarding it as one step away from dissolving, in the hands of relativists, into no sort of truth at all. So far as I could tell, the only support Norris offers for this peculiar idea is that there is at least one relativist who agrees with him. Matters are not helped by Norris's attempt to link this question to that of whether there is anything to causality over and above "constant conjunction," an invariant pattern of association or succession. If there is an argument here, on either score, it is spread so thinly over his pages as to be entirely indiscernible. I suspect there has been some massive failure of communication here, and only become more suspicious when I find Norris calling disquotation an "equivalence relation" (p. 140), since an equivalence relation (as he must know) is one which is reflexive, symmetric and transitive, and disquotation is plainly none of these things.

Some of Norris's statements about geometry lead me to suspect that he is ignorant of Hilbert's great work on the formal nature of Euclidean geometry. Generally, he could benefit from some exposure to model theory.

Norris spends part of an essay castigating Bertrand Russell's 1912 paper "On the Notion of Cause," comparing it unfavorably to Russell's 1948 book Human Knowledge. The only problem with this contrast is that Russell expounded basically the same theory of causation on both occasions. Rather than actually reading the paper, Norris has evidently relied on two or three often-quoted sentences --- most notably "The law of causality, I believe, like much that passes muster among philosophers, is a relic of a bygone age, surviving, like the monarchy, only because it is erroneously supposed to do no harm" --- which do not convey the distinction Russell drew the crude A-then-B statements beloved of "philosophers and Papuans" like "fire burns cotton" or "thrown rocks break windows," and the functional relationships actually found in developed scientific theories. The major refinement in the intervening years had to do with the very useful notion of "causal line," which Russell was using by 1927. --- This'd be a minor point, were Norris not so given to chastising others for mis-reading authors considerably easier to get wrong than Russell is.

Overall, it's good to have someone like Norris on the side of the Truth, Reason and Reality, but this isn't really a book for anyone not already a hardened addict of the science-versus-relativism genre. People coming to this book without some acquaintance with both analytical philosophy of science and post-structuralism are in for rough going. The main problem with it is not so much the presence of Uncle Willard, or even that of Uncle Jacques, as the awful style: lose that and we might have something here.


Errata: p. 226, for "irrational numbers", read "imaginary numbers" (an error of roughly the same magnitude as confusing allegory and alliteration); p. 233, for "thorems", read "theorems"; p. 252, for "Mil's", read "Mill's"; p. 301, makes it sound as though the Michelson-Morley experiments happened after Einstein proposed special relativity.
xii + 330 pp., end-notes to each chapter (no unified bibliography), index of names
The French Disease / Philosophy of Science
Currently in print as a hardback, ISBN 0631198644, US$72.95 [Buy from Powell's], and as a paperback, ISBN 0631198652, US$28.95 [Buy from Powell's]. LoC Q175 N594
15 November 1999