Von Neumann and the Raiders of Rationality

While I think that we might pursue the Contrast as something directed more toward Game Theory itself than Poundstone's book, I think a short take on the book's overall logic may yield some insights as to how exactly Game Theory functions in relationship to real-life situation.

That I'm even having to make this distinction points to Poundstone's fundamental take on Game Theory: as something that one can describe in mathematical terms but also as something that, as it relates to real-world and interpersonal situations, constantly bleeds over into personal conflicts which its requirements of pure reason cannot account for.

Poundstone makes it a point several times in the book to emphasize that criticisms of game theory as not accounting for an effect of a real world situation is problematic because game theory doesn't assume imperfect human players (see the discussion on guilt in the dilemma on 223 or the problem of measuring human investment on 171-172), but yet the entire logic of his book rests upon the indeterminacy within game theory between pure reason and human indeterminacy.

So, we frequently get two different kinds of chapters (the division more stated at the beginning of the book and merged a bit by the end): there are numerous chapters simply detailing the material premises and results of the games (chapters 1, 6, and 11 are good examples) while others detail von Neumann's personality and biography or historical accounts (see chapters 2, 7, and the rather heartbreaking 9).

This division in chapters and logic seem counterproductive. If we wanted a strict account of game theory or a straight biography of von Neumann, we might find ourselves disappointed by what seems like info that distracts from one or the other books' provisional main point.

But it seems as though the kind of logic we find here can also be seen abundantly in game theory. For example, one of the dilemmas the book opens with asks us to imagine that we are faced with a clock ticking down an hour, and at the end of the hour, both our mother and our spouse will be killed. We have the option of saving one of them by pressing a button. If we save one, the other dies. If we refuse to press a button, both die. (it just occurred to me that one could probably see a lot of these prisoners' dilemmas in the Saw series).

Poundstone says that one can only decide which of the two he/she likes better and choose accordingly, but the woman/man of reason this advice presumes simply does not exist. The decision is, in most cases I'd presume, impossible to make (a great film might be made with Norman Bates put in this situation...one would assume he'd choose Mother, I guess).

While Poundstone stone-faces it and pretends as though we should accept this precept without a second thought, it indicates what he will later call the ultimate failure of game theory or the prisoner's dilemma to dictate political or economic affairs because rationality has nothing to do with the way human beings actually live. The model is incredibly useful for providing certain explanations regarding biological evolution, but as Lacan tells us, human desire has very little to do with our unthinking biology. To use an example, our body feels hunger, but what we ultimately eat has nothing to do with nutritional value (if it did, nobody would eat McDonald's).

Additionally, the widespread appeal and draw of game theory plays on irrational motivations in order to make decisions. One might map out reasons to save Mom or spouse, but the decision, in the end, would likely not be a rational one (if one's mom or spouse was continually abusive or otherwise unpleasant, that situation might be different). In the end, game theory can't explain these things. The chess example used in chapter 1 is a good one. Von Neumann held that chess isn't really a game to be played at if it were played rationally. One could mathematically produce strategies that would dictate every move, reaction, surprise move, etc. But no one can actually play the game in such a way, so it, like poker, does inevitably depend on intuition and the kinds of things that reason and quantification can't dictate. We can say the same with economics. Von Neumann initially assumed that game theory could dictate military strategy and economic decisions, but game theory remains unable to do simply because these systems have human motivations.

So, I guess if we're dealing with game theory as a contrast, perhaps that's the way we can approach it. Game theory is great at modeling certain relationships between humans, economic forces, etc. but ultimately fails to predict or provide models for human behavior because it's logic and ontology is based on a purely rational, quantifiable regime whereas the kinds of policy formations we seek are based in human irrationality, flesh, etc.

The dichotomy ultimately takes place in the von Neumann biography as well. Poundstone regularly depicts von Neumann as able to perform incredibly complex mathematical equations without effort, like a human calculator (34, 95, 179) but at the same time his quarrels with his wife, his personal affability, and his love of humor seem to play against it. We can also characterize our relationship to computing technologies in a similar fashion. I might produce a digital animation that makes you laugh. It relies on a series of manipulations of binary code but moves beyond it to work on the human register.