Wednesday, 9 July 2014

Sen's merisms as senility's mereology

Edited- in view of a comment

A merism is a linguistic attempt to define a universal domain, for example by saying 'high and low' or 'by sea or land' or things like 'Will and Testament' where there were two types of jurisdiction regulating acts of an essentially similar kind.

Amartya Sen's 'Idea of Justice' posits the following merism as providing a 'covering set' for the subject of its discourse- viz. 'niti' and 'nyaya'  which, I suppose, is meant to replace the quaint and faintly vulgar distinction between Theory & Practice'- as in 'Theoretically everybody is innocent till proven guilty. In practice, the Bihari servant did it and we'll beat him till he confesses.
Sen says '‘Neeti’ is about rules and institutions and ‘nyaya’ is about their realisation' - i.e. the former is theoretical or transcendental whereas the latter is about real world outcomes in which 'we realise all the different aspects of human life and take into account all the relevant factors.'

Why would the outcomes of properly designed rules and institutions diverge from ad hoc actions to the same end which genuinely 'take all the different aspects of human life into account'?

If Theory and Practice are different it must be because either the Theory is crap or the Practice hypocritical or because of some informational or cognitive constraint for which a workaround is being sought.
However, if we have rational expectations, then provided 'regret minimizing' learning is implemented then we know in advance that the two either converge or one or both are abandoned- i.e. they cease to be part of whatever 'Concrete Universal' they are supposed to be concerned with.
But, this is not the approach Sen takes.
Rather, there seems to be some mysterious sort of mereological claim behind Sen's book,  'The idea of Justice' which appears as misconceived as his much vaunted but meaningless work on Arrow's theorem and which, perhaps, arises out of a misprison, not as with Harsanyi, of a Methodical, but a Mathematical sort.
Thus, Sen says, '...incompleteness of rankings would not prevent making comparative judgments of justice in a great many cases, where there might be fair agreement on particular pairwise rankings, about how to enhance justice and reduce injustice'. 
Aumann agreement- where we can't agree to disagree- is much much stronger than 'fair agreement' which in turn must be much weaker than Baumol 'superfair' (i.e. envy free and therefore 'zero regret learning' based) agreement.  Yet, mathematics say that both arise through 'regret minimizing algorithms' while everything else is sublated away.
In good faith discussion and collective choice, re. any given pairwise ranking  either there is a coordination or a discoordination game. In either case, 'fairness' militates for your changing your mind as you learn rather than be bound once and for all by stare decisis. Thus, 'fair agreement' would be provisional and defeasible w.r.t. the Hannan consistent, regret minimizing, 'reflective' equilibrium without any mereologically hiatus occurring.

Leaving aside the Maths, suppose Sen's operationalized 'fair agreement' can be expressed by a maxim- which is called a 'nyaya' in the Indic tradition- then either that maxim is universalisable or it is not. If it is universalisable then a categorical imperative exists and we are on our way to a transcendental identification of a fully just society. If it is not universalisable then either the point is moot whether it is a specific instance of a more general maxim which is universalisable or, alternatively, we can't be certain that our 'fair agreement' is purely conventional and arises out of shared prejudices or else is self-serving merely and requires 'ideological' window-dressing so that we have 'false consciousness' or  'plausible deniability' re. our true and ugly motives. In this case, we are deluding ourselves that we have 'fair agreement' regarding Justice as opposed to what is cheaper, or more convenient, or more profitable to believe or pretend to believe.
By the same token- suppose the 'fair agreement' can't be expressed by a maxim, then two possibilities arise
1) there is some algorithm or heuristic which generates the same set of 'fair agreements' re. pair-wise comparisons of Justice.  Either this can be expressed in some future Language which evolves on that problematics' fitness landscape or it is essentially apophatic. Either way, what can't be denied is that this view implicitly endorses the transcendental identification of a fully just society even if that lies beyond human ken.
2) no algorithm can generate the same set of 'fair agreements'. In this case overlapping Brouwerian 'choice sequence' can still do the job- so, at least for Intuitionism, the transcendental identification still exists.
From the philosophical point of view- and Sen is writing here as a philosopher- it makes mereological sense to say that which is just in a particular instance is part of Justice. We don't need 'well ordering'- i.e. the ability to stipulate the 'best'- in order to construct a complete pair-wise ordering. Thus, Sen goes on to say, in the passage quoted above- 'A partial ordering can be very useful without being able to lead to any transcendental identification of a fully just society. The approach of the human development is a special application of this general strategy of making do with what can be very widely accepted, without expecting that this strategy will solve every decisional problem we face.'
What's wrong with that?
Surely Sen is just saying 'guys, let's not postpone piecemeal Social Engineering (except Sen would fight shy of this Popperian term) because we're all so caught up in working out the Mathematics of the Perfect Society's Voting rules or Preference Revelation Mechanisms or whatever. More especially, for very poor countries like India and Bangladesh- could we just agree to feed the hungry already because they'll be dead by the time we've fine tuned the sermon we want to read them by way of Grace?'
If this was is what Sen is saying, then why are Bhagwati and Panagariya, but also low I.Q shitheads like me, so angry with Sen?
The answer has to with his use of the collocation 'human development', more particularly human beings in 'Developing Societies'- i.e. people living in places where, by definition, though Institutions are weak and Baoulding 'Pscyhic Capital' is low, there are a heck of a lot of 'low lying fruit' (of the sort which G.D.P does capture) arising from the implementation of up to date technology and the breaking of sort of Klepto or Kakistocratic rent-seeking coalitions which flourish best under the umbrella of Identity Politics- i.e. the notion that human beings differ radically from one another on the basis of things like Colour or being Circumcised or Church attendance or whatever.
Sen may be on the side of the angels but, the fact is, his notion of 'Human Development' is highly manipulable by Kakistocracies such that he himself becomes a sort of P.R trophy or mascot for the worst of regimes.
What is shocking about Sen- a bright guy serving at the heart of Western Academia at precisely the point when, at least for the mathematically literate, Statistics and Game Theory solved all the Philosophical aporias of the Social Sciences and, with extraordinary success, united Economics to the Life Sciences in a manner that 'keeps on giving'. Just recently, there was a headline in a National newspaper showing how 'regret minimization' in portfolio selection is also the algorithm which gives rise to Mother Nature's prodigal diversity.  Sen, as a Bengali, would have been aware of the 'Hannan consistency' result when it appeared. Hannan was taught by a Bengali. He says that back then Bengali students had a better grounding in Stats than Americans.
Sen, as a philosopher, must know David Lewis famous book on Conventions as Schelling focal point solutions to co-ordination problems.
In the Seventies, when he taught at the L.S.E, he would certainly have known all about the Price equation and Maynard Smith and Evolutionary Stable Strategies and Ken Binmore's work and thus Aumann and Peyton Young and Schelling and so forth.
Yet, he refuses to say, well, of course any given 'fair agreement' is defeasible such that it informs a Brouwerian choice sequence terminating at an incomputable Schelling focal point, which is too transcendental- selah!- indeed it reconciles the Rahim of Rawls with Dworkin's Ram- or Judge Hercules- such that 'harmonious construction' is univocal and thus Habermas was not actually an utter shithead all along and as for the late Richard Rorty or the recent Hilary Putnam, they too are redeemed because, though failing to recognise their own felicity, they met the Unicorn, tarried in Peach Blossom Valley, but then moved on for Love is a viva, in articulo mortis,  only its Savants fail, thinking Thought's Quest its only Grail.

Is there any way of looking at Sen's intellectual trajectory and not coming to the conclusion that he was simply engaging in strategic gesture politics?
I can't see how.
Still, if only so I get to finish the bottle of Bacardi I've just broached, let us continue to argue the toss.
Suppose Sen's 'fair agreement' is homologous to Rawlsian 'overlapping consensus'- then, there is no 'strain of commitment' arising from a common Research Project re. the ergodics of what obtains. (Earlier, Rawls had assumed that Samuelson's Economics (which is ergodic or else empty) is a free 'plug-in' for all agents in the 'original position- i.e. Muth rationality was 'enabled'.)
In this case, even if we have no strategic behaviour by reason of costless and ubiquitous  'zero knowledge proofs' - or else everything is 'common knowledge' - still nothing mereological can be predicated of 'fair agreement' no matter what protocols re. 'Public Justification' are observed.  No 'fair agreement' re. Justice-as-Fairness can itself be a part of either Justice or Fairness save coincidentally or by virtue of backward induction. If the former, then Justice is Stochastic, whereas, if the latter, it is Occasionalist.
Is there some other way of conceptualizing 'fair agreement'?
One obvious answer is that 'fair agreement' arises where there is an underlying co-ordination problem. We all want to use the word 'Justice' from time to time in an inter-subjective context and it makes sense for us to do so in terms of a Schelling focal point, provided we aren't really interested in Justice but simply in improving our own lives- or those of people we feel sympathy for.
Provided there is a pay-off for a decisive counter-party, with different interests. to pretend that the improvement we desire is one they only concede because of their greater nobility and attachment to transcendental ideals (which by itself has a reputational effect and lowers their transaction costs while also making possible the extraction of Manorial Tiebout rents for their principals) then there is a sustainable preference falsification availability cascade here of the sort Sen invokes when he writes-
'...Parisians would not have stormed the Bastille, Gandhi would not have challenged the empire on which the sun used not to set, Martin Luther King would not have fought white supremacy in ‘the land of the free and the home of the brave’, without their sense of manifest injustices that could be overcome. They were not trying to achieve a perfectly just world (even if there were any agreement on what that would be like), but they did want to remove clear injustices to the extent they could...'

This is Romantic hogwash unworthy of even a 16 year old Howrah schoolgirl.

Parisians did not storm the Bastille because they suddenly developed an interest in Justice but because they thought they'd be better off under a different type of regime. Similarly, Gandhi did not condemn Western Education, Western Medicine, Western Technology, Western Democracy and, after 1918, the British Raj which was bringing more and more of these horrible things to India, because he suddenly developed strong feelings about Justice but because he himself and the class he belonged to stood to gain, and indeed did greatly gain, by such strategically simulated preference falsification but for which the Indian National Congress would have had to show that it was better at administering India than the I.C.S. Similarly, Martin Luther King may have appealed to the notion of Justice- indeed, he was a bit of a Rawlsian avant la lettre-  but everything he campaigned for was of direct benefit not just to himself and the Church he represented but to other 'coloured' people like himself in America at that time. What was the upshot? The Daley dynasty's Chicago (satirized in the Kelsey Grammer starrer 'Boss') produces an Obama who was financed by, among others, the heiress to the Playboy Empire. These are Tiebout Manorial rents with a vengeance and explain the disappearance, since 1970, of the American middle class and the rise of the 'one percent'.

Is Sen simply a victim of a senile Sehnsucht for the romantic certainties of Undergraduate prattle about the sanctity of 'Struggle' (pronounced Ees-tragal, as in Bengali)?
Perhaps. But he is making a mereological claim- indeed, he has done so pretty consistently over the last forty five years- viz. that part of the idea of Justice, or Social Welfare, or what have you, is the commission of fatally prejudicial actions involving throwing away information or arbitrarily changing the rules of the language game or plain, out and out, lying.
In particular, Sen makes great play of the fact that 'well-ordering' isn't essential and so no 'best' outcome or 'ideal' situation might exist and so deontics must be concerned with pair-wise comparison.
This begs the question- is Economics not, as Samuelson said, essentially concerned with ergodic processes? Leaving that question aside- is Life itself not the product of Evolution, that is a dynamics that is mathematically describable, rather than the one off creation of an arbitrary God?
If so, Schelling focal points that aren't strategic but Muth rational and Euvoluntary exist (and, probably Baumol 'superfair' w.r.t Sen's 'fair agreement' stipulation) for 'Justice' or 'Social Welfare' though not effectively computable still generate a 'well ordering' and, furthermore, function is a manner which obviates pair-wise comparison and the invidious manner in which that procedure throws away information about dynamics.

Ultimately Sen's merisms are those of a sophomore- ''OMG rich people have nice stuff, while poor people don't got shite to eat'- but such merisms deny mereology a dynamic- i.e. they forbid Development, more especially Human Development..


  1. Was there a specific article or paper which prompted this burst of spleen? If so, you've failed to link to it.
    I haven't come across 'merism' in the literature before- who has time to read everything?- but, imagine it arises in 'Law & Econ' with which I'm not that familiar.
    I suppose a merism for the Market would be 'buyers & sellers'. So long as one always has to quote both one's buy and sell price, like an old fashioned stock jobber, then no 'uncorrelated asymmetry' (Maynard Smith) arises and so the Evolutonary Stable Strategy is mixed Nash rather than a pure conditional.
    To quote Wikipedia- 'The usual applied example of an uncorrelated asymmetry is territory ownership in the hawk-dove game. Even if the two players ("owner" and "intruder") have the same payoffs (i.e., the game is payoff symmetric), the territory owner will play Hawk, and the intruder Dove, in what is known as the 'Bourgeois strategy''.
    I think the charitable way to look at Sen's merism 'niti and nyaya'- for Justice Public Discourse- is to separate out the 'Bourgeois strategies' of 'Niti' (where the Institution or 'rule bound' individual has well defined 'territory ownership' and the other party feels they don't) from 'Revolutionary' Nyaya which takes nothing for granted regarding 'territory ownership'.
    I must admit, it's some years now since I read Sen's book which, in any case, was meant for a popular audience. Still, my question is, why not give the man the benefit of the doubt? Surely we can all accept that 'uncorrelated asymmetries' can be distortionary and that Sen is the leading Public Intellectual calling attention to this?

    1. Many thanks for your valuable comment. I've posted about it here-
      Re. your opening question, I'm sure I read something somewhere, not about Sen, but on the topic of merism and mereology- though perhaps that wasn't the language used.
      Will get back to you here if I can manage to think of what it was.

  2. What great misunderstanding of a mathematical sort is Sen guilty of with respect to Arrow's theorem? Surely Sen's treatment of the topic is a masterpiece of lucidity.