Monday, 3 April 2017

Pragmatism vs pragmatics

Dave Maier has a well written post at 3Quarks in which he attempts to 'coordinate' Wittgenstein and Pragmatism.

He says-
In my view, a minority one to be sure, what properly characterizes Wittgenstein’s “hinge” propositions – those propositions “exempt from doubt” and thus making inquiry conceptually possible – is surprisingly simple. It is obscured by Peirce’s unfortunate formulation of our properly anti-skeptical attitude as “fallibilism,” in which no empirical beliefs, given their corrigibility in principle, are held with probability 1. Following Isaac Levi, my teacher at Columbia, I prefer “infallibilism,” which states simply that a proposition is believed only if it is held without doubt. As even some Peirce interpreters concede (e.g. Christopher Hookway), fallibililsm gives back to skepticism with one hand what it takes away with the other. If, as Moore famously pointed out, “I believe it but it isn’t true” makes no sense, neither does “I believe it but it might not be true.” Instead, I say: only if you believe it – that is, regard it as true, without doubt – do you believe it. With this in place, Wittgensteinian hinges are easily seen simply as: beliefs tout court.

Maier is saying there is a coordination game involving 2 different textual availability cascades. This game presents particular difficulties because-  Wittgenstein('s) ...thought is particularly difficult to coordinate with anything else, due to his abnormally strong (and peculiar) philosophical personality, as well as the obscurity and multiple reasonably valid interpretations of his various mostly incomplete writings. We certainly don’t simply want to say: oh look, Wittgenstein is a pragmatist; but on the other hand it is difficult to see how his thought can be used for pragmatist ends outside the explicitly Wittgensteinian context. Most pragmatists don’t want to bother, and most Wittgensteinians tend to resent the effort.

Maier adds that his solution is a minority one. What is it precisely? 

Wittgenstein had said- 'Moore’s paradox can be put like this: the expression “I believe that this is the case” is used like the assertion “This is the case”; and yet the hypothesis that I believe that this is the case is not used like the hypothesis that this is the case. (PI p. 190)

I think Maier is saying that to really believe something is to regard it as more than a hypothesis because you have no doubts on the matter.
The problem here is, on this interpretation, Moore's paradox returns with redoubled force. 'I believe x but know x isn't true' becomes true for every x, where x is an empirical statement, for anybody who believes they haven't yet gained omniscience or that further progress in the empirical sciences is still possible.

Maier is a student of Isaac Levi- probably best known to Economists from a paper of his on the Tony vs George decision problem re. the Iraq War.  In that case, his argument was vitiated by

1) assuming Tony and George could agree that their decision problem can be represented as a state functional form.
In reality, the weaker party would never agree that such a representation is possible. Why? The stronger party, by definition, has a bigger choice menu- including the option of compensating the weaker party or otherwise taking on a bigger share of the cost or the risk. Sadly, this also means being able to fuck up the weaker big time. A full 'state functional form' does not militate to consensus. It militates to gangsterism and punitive anal rape and... but let us allow Tony and George some little privacy.

2) assuming Tony and George can only achieve consensus if they both agree to shift to a state of full belief so that a particular sort of calculus can be applied. This is quite mad. Nobody in the history of the World has ever had full beliefs. If they did, consensus would be impossible, unless it were inevitable. In other words 'full belief' wouldn't matter even a tiny bit.

Why? Well if I have full beliefs about every possible state of the world, I already believe we will reach such and such consensus. You also have a belief that a particular consensus will be reached. If these beliefs match- by the Revelation Principle- the thing goes ahead. If they don't, consensus is impossible no matter what calculus is used. What's more, we know this in advance. That's why we aren't going to do something silly- viz first formulate a hypothesis about every state of the world and then decide, just for the sake of getting to a consensus, to say we have full belief, without a shadow of doubt, in each and every one of these hypotheses.

Suppose you and I are trying to reach a consensus on what type of Pizza to order. Do you really need to have total belief that Lord Krishna won't turn me into a pig in my next life if we order pepperoni? Do I really need to have total belief that your wife will throw you out of bed for farting, and thus precipitate the collapse of your marriage and your eventual descent into alcoholism, if we order the quattro formaggi?
Nope. Not at all. That would be silly, though I'm sure that's pretty much what would happen.

3) Assuming that Tony and George can form a group and that a 'principal' can aggregate information for this group in a transparent manner such that a good decision is made.
Unfortunately, Economic theory predicts that under transparency, the principal will do 'stupid shit' and there is some 'laboratory work' to confirm this notion, which everybody already instinctively grasps.

Levi's paper- featuring Tony & George fucking up globally- appeared in a volume called, hilariously, 'Arguments for a Better World'. It showed that math-savvy Ivy League Professors do have something to contribute to fucking up, not just countries- like Russia in the Nineties- but everybody, everywhere.

What was Levi's mistake? The answer is that 'consensus' is a Coordination problem, not a Decision problem. What people do is choose what appears Schelling focal- though it is not likely to be effectively computable- and hedge using discoordination games. That's what Tony did- and made a lot of money out of doing- and also why Gordon Brown stopped doing that particular type of stupid shit as soon as he took over.

Returning to the Wittgenstein's approach to Moore's problem, which Maier- Levi's student- so brusquely shouldered aside, we see that the former's remark rises above the inane if recast as a story about coordination games getting hedged through discoordination games. That's how Pragmatics works. Not Pragmatism- the stupid Academic Availability Cascade- but pragmatics as semantics' no goodnik twin brother wot got thrown outta shul for being a complete tsedreyt but, look at him now!, a big shot he is on Wall Street yet!
The odd thing is Chomsky & Semantics turned out to be the true Madoff.
Robert Aumann the true Rabbi.


  1. I think Levi was making a point of a technical kind in the Tony/George decision problem. More generally, consider the Paris Harrington theorem. We may know something is true without being able to prove it. In particular computable functions which grow very fast and thus cant be proven in Peano Arithmetic to be everywhere defined.

  2. Thank you for your comment. I'm sorry, I should have responded sooner. As you say, Levi has a technical bone to pick with Sen. Factorisation of weak preference into a strong preference plus an indifference relation is not unique under his type of'joint deliberation' because admissibility varies.
    However, this would be true of any deliberative process for which admissibility is an input. I like the notion that the latter's domain might grow exponentially faster than the former can handle! How would it work?