Sunday, August 5, 2012

Christopher Norris On Formalized Language and Set Theory

-An Excerpt from Christopher Norris' Badiou’s Being and Event, pgs. 52-55. 

In effect, the very confidence initially displayed by Frege and Russell in the power of their logical language (technically speaking: that of the first-order quantified predicate calculus) to offer a complete and perfectly consistent formalization of the set-theoretical domain was itself a sure sign of the project’s being headed for just such an obstacle somewhere along the way. It is precisely this recurrent gesture of containment—this move to control and delimit the scope of enquiry through various techniques of always premature ‘totalization’—that Badiou regards as having posed a chief obstacle to progress by evading the radical or, indeed, socio-political order. Hence his objection not only to the claim that intuition might yield valid insights or conceptual progress (since intuition is most often just the name applied to preconceived belief) but also to that narrowly logicist idea that the exploratory scope of set theory might be circumscribed by a purely formal programme whose terms are specifically in advance and which therefore pre-emptively restricts any future developments to what is conceivable at present. So it was, Badiou thinks, that the Frege-Russell project was predestined to run aground on those paradoxes that sprang to view as soon as it encountered its limit-point condition in that realm of the ‘pure multiple’—or that formally unrestricted set-theoretical domain—which required that statements of the relevant class be open to the test of self-reflexive application.

I should offer some further detail at this stage since the episode in question is among the most crucial for Badiou’s understanding of set theory and of the complex relationship between genesis and structure that has characterized its history to date. The great promise of set-theory as envisaged by Cantor, Frege, Russell and its other early proponents was that of reducing mathematics to a purely logical or axiomatic-deductive structure of entailment relations that would leave no room for anomaly or paradox. That claim encountered its first major setback when Russell showed—by purely logical means—that set theory was intrinsically prone to generate just such problem, namely the kinds of self-reflexive, self-predicative or auto-referential paradox that resulted from its dealing with formulas such as ‘the set of all sets that are not members of themselves’ or ‘he who shaves the barber in a town where the barber shaves every man except those who shave themselves’. Yet, as Badiou points out, despite their somewhat contrived appearance such paradoxes all derive from a basic formula (that of the set which is not a member of itself) which, so far from being forced or extraordinary, in fact turns up—and quite acceptably so—in each and every possible specification of a set. Thus ‘it is obvious that the set of whole numbers is not itself a whole number’, and so on for any range of similar instances (p. 40). To this extent it is an inbuilt feature of set-theoretical thought, one that arises whenever it is a question of asserting ‘the constitutive power of language over being-multiple’, and which therefore cannot be regarded as something pathological or (as Russell and Frege supposed) in need of surgical excision. However, it does take on such a negative, subversive or system-threatening aspect when its implications are followed through in the context of an ultra-logical programme which identifies truth with formal validity, in turn, with the classical ideals of consistency and total closure under logical entailment. In that context the acceptable face of self-reference—its ubiquitous and therefore unobjectionable presence—undergoes a distinct change of expression and becomes, in effect, the un-doer of that whole optimistic logicist project.

Russell’s answer was to make it a stipulative rule that statements in formal languages such as those of mathematics or the logical calculus should not be self-referring in a way that gave rise to difficulties of this sort. Rather they could best be averted by a ‘Theory or Types’ which distinguished clearly between various orders or levels of statement, that is, those belonging to the first-order language of direct or mater-mode assertion, those that referred to such first-order statements from a higher logical level, and so on up through successive stage of increasingly abstract formal (i.e. meta-linguistic) specification. Only thus, Russell thought, could set theory—as a crucial component of present-day developments in logic and mathematics—be kept on its path towards an ever more secure, since ever more precisely codified conception of validity or truth. Still his purported ‘solution’ to these problems struck many, then and now, as objectionably ad hoc and as having more to do with interest of pragmatic or methodological convenience than with principles self-evident to reason. Indeed, the set-theoretical paradoxes have remained a spur to philosophic thought and a potent source of speculative ideas both within mathematics and across a range of other disciplines ever since Russell first discovered them. Their impact was intensified by various related developments, including—most notably—Gödel’s undecidability-proof to the effect that any formal system sufficiently complex to generate the axioms of elementary arithmetic or first-order logic would necessarily include or entail at least one statement the truth or validity of which could not be proved within the system itself. In other words, one could have either truth as matter of rigorous logical procedure or consistency (‘completeness’) as a matter of intra-systemic coherence but surely not both unless by some maneuver, like Russell’s, that looked suspiciously like a mere device for saving logico-mathematical appearances. Nevertheless set theory survived these and other challenges through the effort of various thinkers to provide some method of formal restatement in axiomatic terms that would keep the paradoxes safely out of view or at least prevent them from doing real harm. During the past century it has become absolutely central to every branch of pure and applied mathematics, as well as to every mathematics-based development in the physical and even (in certain contexts) the social and human sciences.

-Christopher Norris, Badiou’s Being and Event, pgs. 52-55.

No comments:

Post a Comment