Sundar Pichai, epistemic ethics & Coase's Antidosis

Epistemic ethics is the notion that people who do a particular type of rigorous research should find common 'action guiding' principles on the basis of what they study. So a Geneticist who knows that there is no difference, at the DNA level, between an Iyer and an Iyengar should not behave as though Iyengars are superior to Iyers. Also she should let me take her out on a date. Who knows? Maybe I really am a handsome 24 year old who has been nominated for the Fields Medal. Why not take a chance?

Speaking of Iyengars, I notice that Google is now headed by one such youngster. Since Google does research, it follows that Pichai may be bound by the epistemic ethics relevant to the sort of research his company is involved in. Just recently, some guy wrote a memo about how women and minorities are getting special treatment in that Company and boo hoo it isn't fair.

Now, all right thinking people acknowledge that when you run a big company what you should do is to listen to each and every employee and gently lactate the milk of human kindness as you do so.  Pichai, of course, is not denying that he listened very carefully and lactated very profusely but the fact is he also fired the dude who was causing him a headache.

Did Pichai violate epistemic ethics?

Scott Aaronson, at Shtetl optimised, has lactated profusely and cogitated profoundly on this question.

He says-

'if James Damore deserves to be fired from Google, for treating evolutionary psychology as potentially relevant to social issues, then Steven Pinker deserves to be fired from Harvard for the same offense.

Yes, I realize that an employee of a private company is different from a tenured professor. But I don’t see why it’s relevant here. For if someone really believes that mooting the hypothesis of an evolutionary reason for average differences in cognitive styles between men and women, is enough by itself to create a hostile environment for women—well then, why should tenure be a bar to firing, any more than it is in cases of sexual harassment?

Was James Damore employed for the specific purpose of investigating evolutionary psychology? Is he an eminent man in that field? Would a janitor at Harvard, who made similar points to Damore, perhaps not as cogently, and not by means of an email memo, been sacked? I think so. The janitor is employed to mop the floors. Airing his views on evolutionary psychology might constitute a nuisance and it might alarm some students and faculty.

Scott might reply 'But, Pichai & Damore are both Science Nerds'

'They share an epistemic ethics.'

Why should Pichai get to fire Damore?

Ronald Coase developed a Theory of the Firm based on the notion that where transaction costs are high or markets are missing, it makes sense to aggregate certain sorts of activities under a distinct corporate personality rather than to just contract on the open market for specific goods and services as required.

According to this theory, Corporate decision making has a different information set and pay-off matrix from that of any individual. Further, the Chief Executive may be aware of particular cost schedules and other strategic considerations which are not 'common knowledge'. In this case, though all employees may have the same duty to the Corporation, they do not necessarily have the same duties to each other by virtue of their ex officio capacity. As a janitor I have a duty to the Company. However it is not a duty I can properly discharge by barging into the Board Room and explaining my theory of option pricing. There are exceptions to this rule. If I genuinely have important information then a Court may find that I had a duty to share that information even if that was not the specific purpose for which I was hired.

In this particular case, the CEO is empowered to say what is or is not the Company's interpretation of a specific course of conduct in relation to a possible breach of existing policy regarding safety in the workplace. Pichai was legally entitled to sack someone whom he thought had infringed that policy because he had the right to interpret his own Company's rule in this respect. The employee concerned had no such right. Because the law is clear on this point, Damore will face an uphill legal battle if he sues for damages.

However, even he had no legal right to do so, there may have been good commercial reasons for Pichai to sack the fellow pour encourager les autres. His Board would probably have approved. However, in that case Damore may be entitled to damages.

Of course, we all understand that a thing can be permitted by the law and advisable on commercial grounds and yet be unethical. In particular, it may be legal and profitable for a Scientist to endorse some pseudo-scientific rubbish so as to make money, but we would still feel this this to be a violation of epistemic ethics. We would say- 'This man may be doing good Science but he is a bad Scientist because he is undermining Scientific method and the hard won reputation for integrity of other toilers in this field.'

Perhaps, there is some genuine philosophical puzzle or aporia here?

Scott now presents a bizarre argument which staves off this possibility-

But the reductio needn’t stop there. It seems to me that, if Damore deserves to be fired, then so do the 56% of Googlers who said in a poll that they opposed his firing. For isn’t that 56% just as responsible for maintaining a hostile environment as Damore himself was? (And how would Google find out which employees opposed the firing? Well, if there’s any company on earth that could…) Furthermore, after those 56% of Googlers are fired, any of the remaining 44% who think the 56% shouldn’t have been fired should be fired as well! And so on iteratively, until only an ideologically reliable core remains, which might or might not be the empty set.

Under perfect information, no missing markets, no convexities etc; it would be true that people who oppose Damore's firing might be judged to have the same opprobrious quality. But in that case, there would be no Google because the Coasian conditions for the firm to exist would not be met.

It is rational for a person who does not want to be fired- perhaps because his transfer earnings are low- to not want anyone to be fired. Of course, in a perfect information world, nobody earns any economic rent and so it is not rational to object to being fired oneself or care about anyone else being fired.

Obviously, a perfect Arrow Debreu world with perfect forward markets for everything would also be one where there would be no Epistemics or Ethics or Epistemic Ethics. All knowledge would be available everywhere. No Choice would bear an externality or be other than regret minimizing ( which fulfills virtue ethics.)

Scott's foray into epistemic ethics turns out to be utterly worthless.

This raises the question-

Can Epistemic ethics exist in an imperfect Coasian world?

Sure.

Think of a epistemic duty as being like a liturgical requirement in ancient Athens.

Scott thinks Pichai had an epistemic duty to retain Damore.

By Antidosis, Pichai can say 'I'll swap places with a CEO ready to retain Damore provided I get a company equal to Google to manage.' Scott can put together such a Company and do the deal. 

Actually, Financial Markets allow something like Athenian Antidosis already. It may be that Damore has the right business model. Some VCs get together and make a hostile bid. Or maybe there is a Board Room coup. Or a Management buy out.

It seems the Coasian firm already has a way to do epistemic ethics.
Scott doesn't, but he has lactated very profusely on his blog.
That's something nobody can take away from him.

Scott Aaronson's Naked Emperor Equilibrium

Scott Aaronson has a post at Shtetl Optimized on Kolmogorov's epistemic ethics.Scott writes

 If it means anything to be a lover of truth, it means that anytime society finds itself stuck in one of these naked-emperor equilibriums—i.e., an equilibrium with certain facts known to nearly everyone, but severe punishments for anyone who tries to make those facts common knowledge—you hope that eventually society climbs its way out. But crucially, you can hope this while also realising that, if you tried singlehandedly to change the equilibrium, it wouldn’t achieve anything good for the cause of truth.

This is stupid. Clearly, in a situation where people are too frightened to say 'The Emperor is naked' the very last thing you'd want is for this to become explicit 'common knowledge' because then the task of the Secret Police just gets very much easier. Essentially they'd just need to get an independent truth telling savant to say 'All subversives will be put to a very painful death along with their near and dear on the morrow of the xth night subsequent to this announcement- where x stands for the percentage of potential subversives we have targeted- unless, of course, these potential subversives have the good sense to top themselves first.'

 Imagine an economy with homogeneous labour input (or else assume hatred of the regime is randomly distributed across occupations). A one off technological shock which raises productivity by ten percent might mean that the ten percent of the population who by a Pareto Law, represent most of the Subversion potential can be got rid off.

In this case, the alethic Savant who makes it explicitly common knowledge that the regime is murderously efficient only in propagating a stupid lie- e.g that the Emperor aint naked- would also be the best, the most economical, instrument of the Secret Police.  His announcement that such and such is the dastardly plan of those sociopaths would cause around ten percent of the population to commit suicide on the tenth night.

Actually, that isn't quite true. All rational agents with this 'common knowledge' would be committing to 'Newcombe' or 'Kavka' type mechanisms to guard against being 'potentially subversive' on the xth night. Thus any random decimation on the morrow of night x would be sufficient to not just fatally deplete the class of potential subversives but would set off an endogenous 'arms race' of increasingly costly signals of commitment to the 'naked Emperor equilibrium'. Indeed, to adapt the language of St. Paul, we would have a particularly robust, for 'mysterious', moral Economy.

In this context, can there be an action guiding epistemic ethics? In other words, is there something about studying a subject properly which also tells you what to concentrate on within that subject and how to deal with colleagues who may have different views or be more vulnerable to official displeasure?

Scott, believes in

‘the Kolmogorov option. This is where you build up fortresses of truth in places the ideological authorities don’t particularly understand or care about,’

However this wouldn’t work in either Stalin’s or Mao’s Utopia because doing stuff the authorities don’t care about, more especially if it requires brain power, would be considered ‘bourgeois idealism' punishable by a spell in a Labour Camp.

People like Kolmogorov & Hua Luogeng showed the State that Mathematicians were willing to roll up their sleeves and do applied work so as to over-fulfil the 5 year plan or whatever. Luogeng was even able to save the lives of one or two 'bourgeois idealist' Pure Mathematicians.

Scott mentions Lysenko’s idiocy but doesn't get that it was useful to both Stalin and Mao. It gave them an excuse to starve the peasantry thus bolstering the ‘dictatorship of the proletariat’.

In the West we are fortunate to have autonomous universities competing with each other, more especially in ‘games against nature’. Does this circumstance permit the deduction of a categorical imperative for Scientists of the sort given in Scott's post?

Certainly, under a specific type of Dialetheia which features ‘scientific truth’ discoverable in one way and ‘deontic truth’ discoverable in another way. However, this is inelegant and not robust.

If we forbid Dialethia we either have some sort of evidentiary decision theory- which is equally fraught with problems; we might end up ‘managing the news’ or else get caught up in backward causation type paradoxes.

Alternatively we could embrace Gibbard type ‘semantic normativism’ but this too has problems.
Finally, we could think of ‘Truth’ as the solution of a coordination game. However, if Knightian uncertainty obtains, we ought to be hedging through ‘discoordination games’ so similar problems arise.

Kolmogorov himself gives us a demarcation criteria for ‘alethic’ research programs separating them from ‘Preference falsification’ based ‘availability cascades’. Clearly, only a ‘bandwagon’ would instrumentalise complexity classes far beyond our computational capability for prescriptive purposes.
Sadly, this approach doesn’t take us very far either because it fails for co-evolved process

What is the Economic theory of ‘naked-emperor equilibriums—i.e., an equilibrium with certain facts known to nearly everyone, but severe punishments for anyone who tries to make those facts common knowledge’?

This sounds like a ‘pooling equilibrium’ on the basis of ‘cheap talk’. ‘A costly signal’ (e.g. something heavily punished) gives rise to a separating equilibrium- which means there is an arbitrate opportunity between coordination and discoordination games.

However General Equilibrium theory predicts that the moment this is exploited the whole becomes, at worst, ‘anything goes’, or at best ironic in a Hegelian manner. The ‘martyr’ becomes the bedrock of the everlasting Ecclesia founded upon the lie he exposed.

I think this question and its ironic outcome lies at the very origin of Western Philosophy in the duel between Isocrates and Aristotle. The former wrote a letter to the boy Alexander urging him to get shot of his tutor who was having him grapple with difficult and abstract subjects like Pure Mathematics. It is better, Isocrates says, to concentrate on perfecting rhetorical style and the arete proper to pre-eminence.

Isocrates is referred to in the Phaedrus where Socrates comes to the conclusion that the palinode (i.e. being able to change your mind) is paradigmatic of philosophy as something which gives birth to the new. Ironically, Socrates- a descendant of Daedalus- finds an Ariadne’s thread which leads him to become a pharmakos- a scapegoat- for his beloved Polis.

Isocrates’ ‘Antidosis’ (a Coasian mechanism design quirk in Athenian law whereby, if you were selected to discharge a public function then you could challenge some richer guy to do it or else swap estates) is the opposite of Socrates’ own defence against the charge of misleading the youth.

Essentially, in Scott’s terms, this ‘good heretic’ is saying he’d be better at the job of clothing the Naked Emperor in a seamless robe of diaphanous opulence.

This is an idea which the Russians were familiar with from Doestoevsky. If you want to get a job with the Secret Police, you pretend to be a dissident till they pick you up for interrogation at which point you turn the tables on them and explain how you are an even better agent provocateur than those they set upon you. Of course, the thing works the other way as well. A true dissident who really wants to change the system will have disguised himself as a secret policeman straight off the bat.

Perhaps this was the game everyone was playing. People like Luzin, by reason of their mystical leanings, were actually safer if denounced by their own. Those in real danger were the Old Bolsheviks.

The Soviets and the Americans develop O.R to a high standard at around the same time. Two very different systems begin to mirror each other in one respect- their Political Paideia can have a purely mathematical description. The Soviets had Nobel prize worthy mathematical economists just as the Americans did, but it was the latter which developed systematic schools attracting adherents across the globe. Ultimately, in the ’90’s we had the ultimate ‘antidosis’ competition. The Harvard Econ Dept was hired to give Russia an American type economy. We all know how that turned out.

Perhaps we became complacent at some point in the Eighties. It is sad that shrill complaints re. ‘trait based’ discrimination have vitiated proper statistical research into the genuine article as illumined by Sowell, Loury, Fryer & c. There are a lot of bright and good people doing Junk Social Science in order to show they are ‘engaged’.

Misology, too, is a type of ‘costly signal’. If a smart and erudite guy makes it a point to utter ignorant non sequiturs it must be the case that he really really cares, right? Either that or he’s just lazy and likes publicity. But, that’s the nature of our current ‘naked Emperor equilibrium’.