Wednesday, 28 October 2026

Kaushik Basu's ignorant knowledge curse

 Can knowledge reduce human welfare? That depends on how we define knowledge & welfare. If knowledge means useful information and welfare means utility, then the answer is no.

Kaushik Basu disagrees. 

The knowledge curse: how knowledge can reduce human welfare- Kaushik Basu and Jörgen Weibull

Greater knowledge is always an advantage for a rational individual.

No. Adequate knowledge is useful. There are diminishing or negative returns to acquiring, storing & accessing information.  

However, this article shows that for a group of rational individuals greater knowledge can backfire, leading to a worse outcome for all.

It is possible that a shared delusion turns out to be useful. But knowledge that such is the case means people can keep the useful behaviour while discarding the delusion- e.g. being good, so Santa doesn't put you on his naughty list, can lead to good habits which are advantageous even though Santa Claus was locked up long ago for being a creepy pedophile.  

Surprisingly, this can happen even when new knowledge does not mean the discovery of a new action but simply provides a deeper understanding of the interaction at stake.

Understanding (Verstehen) is not Knowledge (Eklaren). Verstehen ist nicht Erklären" (Understanding is not explaining). Wilhelm Dilthey,  wrote: "Die Natur erklären wir, das Seelenleben verstehen wir" ("We explain nature, we understand the life of the soul").

 I say- 'I empathise with you'- it does not mean I can explain you or, indeed, know you at all. Indeed, understanding you may mean rejecting any and every type of objective knowledge about you. Theodicy is a verstehen of God such that no erklaren is conceivable. The Faith this gives rise to is founded upon a mystery. But, this is also the case for the objects of our intimate, concrete, love or, indeed, distant, wholly abstract, affection.

More specifically, enhanced knowledge about the current state of nature may hinder cooperation among purely self-interested individuals.

Anything at all can hinder cooperation. More generally, if there is a coordination game, or pooling equilibriums featuring 'cheap talk'. there is likely to be a discoordination game, or separating eqbilibrium featuring some costly (or costly to disguise) signal. 

The paper describes this paradoxical possibility—a ‘knowledge curse’—and analyses the evolutionary process that occurs if, initially, only a few people have access to the greater knowledge. It concludes with a tentative comment on ways to avert this potential knowledge backlash

Basu was cursed with stupidity. Fuck does he know about anything useful?  

Consider a two-person world, or a world of many who are matched pairwise and play the Base Game where one player chooses between the rows and the other columns. 
 A B A 53, 53 51, 52 B 52, 51 50, 50

Both go for 53,53.  

The numbers represent expected payoffs that players obtain, with the row player’s payoff listed first and the column player’s listed second, for each of the four strategy combinations. Thus, if player 1 chooses A and 2 chooses B, 1 gets an expected payoff of 51 and 2 an expected payoff of 52. There may be variations in payoffs, but since they do not know when and why the variations occur there is not much they can do about this. It is natural that players will settle into using the strictly dominant strategy A, which gives a higher expected payoff than strategy B, irrespective of what strategy the other player uses. As a consequence, they get an expected payoff of 53 each. The assertion in §1, that the discovery of a new action can cause everybody’s welfare to drop can be illustrated with a simple example. Suppose that, in addition to actions A and B, there is discovery of a new action, C. Here is the Expanded Base Game, with action C for each player added in: A B C A 53, 53 51, 52 0, 55 B 52, 51 50, 50 0, 54 C 55, 0 54, 0 1, 1

Knowledge has decreased because there is greater uncertainty. The Muth rational solution (i.e. one which assumes people use the correct economic theory) remains the same. 

New knowledge of one type may reduce knowledge of another type. In this case, knowledge about what the other will do falls while some other sort of knowledge rises.  

What happens after this scientific advance? Both players will now choose their dominant strategy C and they will end up with a payoff of 1 each. There is, however, no surprise in this. The new action has converted the game into an extended Prisoner’s Dilemma.

It may do. It may not. What matters is whether C has changed the likelihood of some D, E, F etc.  

What would be surprising is welfare collapsing with the advance of science even without the discovery of any new action.

Not really.  Ecclesiastes 1:18 states, "For with much wisdom comes much sorrow; the more knowledge, the more grief". The idiot is perfectly happy. The savant is a miserable fuck. 

This counter-intuitive possibility is the focus of the present study. To examine it, consider a scientific advance. A Galileo Galilei comes along and discovers that if the moon is in phase 1, the game that humans play is the Phase-1 Game given below A B A 4, 4 101, 1 B 1, 101 98, 98

If so, there is no impredicativity or interdependence. But, in that case, there is no game.  

That is, if the moon is in phase 1 and the two players choose action A, they will get payoff 4 each. And so on. Strategy A is still strictly dominant. Galilei also discovers that if the moon is in phase 2, then the game that humans play is the following Phase-2 Game : A B A 102, 102 1, 103 B 103, 1 2, 2 In this phase, strategy A is strictly dominated by strategy B, and if both players use this strategy, they each receive payoff 2. Now note that the expected payoffs, when both phases of the moon are equally likely, are the ones given in the Base Game payoff bimatrix. That is, if, for instance, both choose A, each will half the time (when the moon is in phase 1) get 4 and half the time (phase 2) get 102, which makes for an expected payoff of 53. Once the knowledge of the scientific breakthrough spreads through society, human income drops to an expected level of 3, the average of payoff 4 in phase 1 and payoff 2 in phase 2. This calamitous collapse, from 53 to 3, is brought about simply by the advance of knowledge.

No. It is brought about by the manner in which the game has been rigged. The fact is, Galileo's discover stuff which has endless applications in all sorts of fields. True, a particular 'coordination game' may be adversely affected but there is a bigger payoff for some other reason. Thus, discovering that the earth goes round the Sun may reduce Religious Faith which in turn causes more people to tell lies, but it may also lead to greater astronomical knowledge permitting mariners to cross great oceans and discover new continents.  

It is not our claim that this will always happen, but the fact that it can is alarming enough.

It really isn't. We get that Science can undermine religious fate- if people think they are descended from monkeys, they might rob and rape and grab my bananas- but science can make our lives less 'nasty, brutish & short'.  

There is a lot of writing about the existential risks that human society faces. However, we tend to associate such risks with human folly, irrationality and lack of knowledge.

Religious people may say it is because people no longer fear that the Devil will fuck them in the ass 

What is summed up in the above parable is that in some interactions, perverse outcomes can be the result of science and rationality.

Only by an arbitrary ipse dixit assertion. 

Game theory is all very well if the payoff matrix is accurately known but, because of Knightian Uncertainty and impredicativity, it never is.  

No comments: