Nassim Nicholas Taleb's The Black Swan: The Impact of the Highly Improbable is one long polemic against views that are taken for granted by professional economists and philosophers (both of whom are repeatedly and fairly criticized throughout the book--historians are also criticized, but these have largely abandoned the views attributed to them since their so-called 'linguistic turn'). Luckily, philosophy has no Nobel, so we are spared his worst ire. Leaving aside the economists, philosophers have responded by ignoring the book. This is a shame because not only is Taleb sometimes eerily spot on in his observations, Taleb is also extremely erudite about philosophy and its history (Nelson Goodman’s Grue and Averroes can be found in the same chapter), and full of pointed criticisms of philosophical practice and how it is implicated in supposedly ‘scientific’ ideas that have had disastrous afterlife in social policy [I call this the ‘Socratic Problem.’].) In some ways Taleb takes the impact of philosophy more seriously than professional philosophers are inclined to do.
The lack of attention to The Black Swan within the discipline can be explained, in part, by the admission that most philosophers don't read books (anymore), but I have no such excuse; more important because I have been discussing ideas anticipated by Taleb in the life of this weekly blog-entry (including a modest criticism of one of Taleb’s articles), I am obliged to engage with it. The book is full of philosophical ideas that are treated in a very lively manner—I would happily teach it to an introductory class in philosophy of social science (except, perhaps, that I dislike teaching books with which I agree so much). Today I discuss two aspects of Taleb’s book. First, I articulate his main criticism of philosophers; second, I criticize the core distinction in the book (I do so in spirit of constructive engagement—rather than as attempt at refutation).
First, Taleb's two core (related) criticisms of the philosophers (especially epistemologists and decision theorists) is that i) we are obsessed with confirmation (which prevents us from real learning and gives us false confidence) [Popper is one of Taleb's heroes] and that ii) we engage in what he calls the 'ludic fallacy'--we model decision (or even probability) on what he calls the "sterilized" uncertainty "we encounter" in (casino) "games". The second is really four related problems: a) we pretend as if we can assign probabilities to events that are fundamentally (epistemically) opaque to us; b) we pretend that there is commensurability between our credence(s) and the probable structure of the world; c) when we create our toy-models, we fail to ask about the unknown, unknowns--Taleb's Black Swans--that fall outside our models of reality. (In the artificial environment of the model of well-oiled turn of a roulette table this may be legitimate, but not in the environment of human decision, including the vendetta's among casino owners and their employees.) There is a fourth (which I have discussed on this blog): when we mathematically model unknown, unknowns (that is, genuine Knightian uncrtainty) as "randomness" and put it inside the model, we create false confidence in our numbers (recall how I took Samuelson and Arrow to task for this move against Knight [somewhat unfairly, Taleb fails to realize that Knight should be one of his good guys]).
Now, Taleb admits there are social circumstances in which our toy-models will do reasonably well (even outside a casino). In (reasonably) closed environments where unknown, unknowns can be safely ignored and where we have independent knowledge about the nature of underlying distributions or where we can safely rely on the law of large numbers, we can bring our mathematical tools (including the Gaussian, which is one of Taleb’s bete noirs) to bear on problems. Such circumstances are what he calls ‘Mediocristan’ (circumstances where Black Swans can be ignored). These are opposed to what he calls ‘Extremistan,’ where a singleton (event/observation, etc) can impact the total outcome, that is, places where unknown, unknowns can be disastrous or extremely profitable. (I am going to resist the urge to call such unknown unknowns, ‘highly improbable,’ or ‘fat tails,’ because often our very models hide the black swans from us, what Taleb calls the ‘masquerade problem’.)
Now, my simple criticism is that Taleb has a tendency to treat Mediocristan and Extemistan as given rather than as, in part, manufactured or constructed (by, say, policy-makers and their consultants). Or, at least, many social Extremistan’s are the (perhaps unintentional, but foreseeable even if current social scientific orthodoxy refuses to admit it) consequence of policies that are designed to privilege some well connected political group. (This is as true of right-wing deregulation pushed by well connected managerial class as it is of, say, left-wing development projects.) Now implicitly Taleb realizes this (he attacks, for example, the culture of ‘bonus’ incentives to managers of nuclear plants and financial assets), but too often he treats Extremistan as given (some of his core examples concern wealth distribution). And this leads him to discount the possibility that we (concerned, informed citizens) can influence politicians and their experts to think about (social) decision making in terms of robustness to error (the underlying idea is that one cannot avoid, say, stock market crashes, but can do quite a bit to avoid financial meltdown or the socialization of their costs). That is, in its mistrust of the experts (and the way we sell ourselves to highest bidders) the book is anti-political in a way that reinforces the status quo (even if Taleb teaches his readers how to think profitably about black swans).
Recent Comments