Taleb and Feyerabend

By Felix Salmon
July 21, 2009

Back in March, Scott Locklin’s second-ever blog entry was a takedown of my Wired article which ended with this:

I’ll be fine until someone sends me another piece of nonsense by Nassim Taleb, at which point, as my favorite wrestler, the Iron Sheik likes to say, “I will make him humble; old country way.”

Happily, that day has now come, and Locklin’s screed against Taleb makes for delicious reading. Among his vocabulary: clowns; impostures; mush headed absurdity; Berkeley tapwater; careen; chicken entrails and pixie dust; utile; dyspepsia; serfs; mountebanks; and “heretical alpha monkey of the quants”.

Substantively, Locklin makes a very astute comparison:

Taleb, on the other hand, is sort of the Paul Feyerabend of quantitative finance. Like Feyerabend, he is well read, a good writer and quite charming. Like Feyerabend, Taleb seems to earn his daily bread by showing up and being kind of witty. Finally, both Feyerabend and Taleb are very much Against Method. This means, effectively, they’re both intellectual nihilists. Feyerabend thought we couldn’t know anything for various reasons too silly to get into right now. Taleb thinks all of quantitative finance is nonsense and we should do away with quants.

I’m a fan of Feyerabend, which is maybe why I’m a fan of Taleb as well: both of them force us to carefully examine things we believe mainly because clever and successful experts tell us that they’re true. And I haven’t asked Taleb about this, but I suspect that he wouldn’t be very upset to be called “the Paul Feyerabend of quantitative finance”. Certainly the two share an interest in thinking big.

Locklin ends his blog entry by saying that a successful quant like Jim Simons is “making money more or less proving people like Taleb wrong”. To use Locklin’s own analogy, that’s like saying that a successful physics experiment more or less proves Feyerabend wrong. Which given that Feyerabend was a successful physicist before he became a philosopher, is a bit weird.

It’s perfectly possible to spend a lifetime in science without ever having to grapple with or worry about Feyerabend’s philosophy — just as it’s perfectly possible to spend a lifetime in finance without expending any time on Taleb’s arguments. Many of those scientists and financiers will be very successful. But for those of us who want to take a step back and re-examine the foundations of science, or finance, then grappling with Feyerabend and Taleb is I think a useful and illuminating thing to do.

Update: Taleb emails to say he considers himself closer to Hayek than to Feyerabend.


We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/

Taleb is correct – the economic system could not withstand the amount of debt that was created (as Locklin acknowledged), and the growth in that debt was non-linear. I don’t understand why Locklin has a hard time understanding the non-linearity in the growth of debt, it’s pretty obvious, it dramatically deviated from a linear pattern. And I would bet that many models that the financial quants (which is what Locklin has been unsuccessfully aspiring to become) developed cannot account for non-linearities like the step function in debt the financial system could not absorb.

If you believe that the financial system of the world can be modeled, then anybody who says that models are flawed at some point (we obviously reached that point last year) presents a threat to your belief system (which in the case of aspiring quants, is an insult to your religion). The weather system of the world is easier to model than the financial system, but we can’t even do that with any degree of accuracy. There are almost 7 billion black boxes in the world economy, and their responses to changes in the system can not be modeled, let alone approximated.

Locklin then disparges the concept of converting debt into equity. He doesn’t understand the concept (creditors and lenders would reduce principal in exchnage for a share of the assets that are collateral for the debt – what is hard to understand about that?), and so launches into a personal attack on Taleb.

What I don’t understand, Felix, is why you bring attention to this blog? You posted enough for the day, this wasn’t necessary, go enjoy Shanghai.

Posted by KenG | Report as abusive

Taleeb is on the bestseller list for a very good reason. While the crowd was mesmerized by the “high-fiving” on CNBC-duced, Taleeb, like a select few, were disturbed by the fact that never in history, were the financial markets so leveraged and so vulnerable to a shock. I never let a day go by without remembering the words of the famous UK Economist, David Ricardo…”Finance is a conspriacy against the laity.”

See all of the Bullish Rubes, soon, at the bottom !!!

Posted by John P. Crowley | Report as abusive

Locklin is as angry, indignant, and articulate as Taleb.

Therefore, he is fun to read and will get attention.

Posted by Dave | Report as abusive

One of Taleb’s flaws is that he paints a narrow picture of modeling. One criticism I’ve had of financial models is they are top-down instead of bottom-up. You take a group of 1000 mortgages, securitize, and get xyz performance & risk. That is really an idiot way to do it. What you should do is build a model for an individual mortgage, income, credit score, collateral. Aggregate the 1000 individual models and see how they react to a given change.

This goes to the next problem with models. Most MBS models had no provision for either default or a fall in home prices, and ignored the relationship between the two. Really, is that modeling, or hoping?

Those two criticisms, top-down vs. bottom-up and pollyanish assumptions do not taint the principle of modeling, but rather how models were built and used.

Posted by winstongator | Report as abusive

I thought about Feyerabend a lot yesterday. First, this Richard Posner post brought him to mind:

“One begins to wonder whether the current crop of economists has made or will make an original contribution to fighting the economic crisis. It is not as if such weapons as easy money (the Federal Reserve’s expansion of the money supply to drive down interest rates), bailouts of banks and other financial institutions, and deficit spending were inventions of modern economists. They were well known in times past (for financial crises are nothing new) to bankers and other businessmen and to economists such as Walter Bagehot (a nineteenth-century journalist rather than a professional economist), Irving Fisher, and John Maynard Keynes. Keynes in particular seems the best guide to digging our way out of our current economic crisis, even though he has been dead for 63 years and his great work The General Theory of Employment, Interest and Money (1936) has been neglected rather completely for the last 30 years or so.”

Now, Feyerabend:

“The withdrawal of philosophy into a “professional” shell of its own has had disastrous consequences. The younger generation of physicists, the Feynmans, the Schwingers, etc., may be very bright; they may be more intelligent than their predecessors, than Bohr, Einstein, Schrodinger, Boltzmann, Mach and so on. But they are uncivilized savages, they lack in philosophical depth — and this is the fault of the very same idea of professionalism which you are now defending. [6]”

I too find the earlier writers more relevant. Perhaps Economics has fallen into the same trap as Philosophy and Physics.

Then I read this Justin Fox post:

“The same goes for Matt Taibbi and his already legendary Goldman Sachs rant. If all financial journalists shared Taibbi’s disregard for fairness and his juvenile glee in name-calling, financial journalism would be unbearable. But since most financial journalists try so hard to be fair that they often miss the truth, and write with such sobriety as to be unreadable, Taibbi offers a welcome change of pace and point of view.

This country has spent 50 years in the grips of a pretty narrow consensus on what journalism is supposed to look like. That’s unraveling. And while a lot of things about the future of this business worry me, the increasing diversity of journalistic methods and standards isn’t one of them.”

And Feyerabend:

“An essential part of a general education of this kind is acquaintance with the most outstanding propagandists in all fields, so that the pupil can build up his resistance against all propaganda, including the propaganda called ‘argument’. It is only after such a hardening procedure that he will be called upon to make up his mind on the issue rationalism-irrationalism, science-myth, science-religion, and so on. His decision in favour of science – assuming he chooses science – will then be much more ‘rational’ than any decision in favour of science is today.”

I agree that Taleb is reminiscent of Feyerabend. But let me end with a quote from Bagehot, and two from Burke:

“The great difficulty which history records is not that of the first step, but that of the second step. What is most evident is not the difficulty of getting a fixed law, but getting out of a fixed law; not of cementing (as upon a former occasion I phrased it) a cake of custom, but of breaking the cake of custom; not of making the first preservative habit, but of breaking through it, and reaching something better.”

“History is a preceptor of prudence, not of principles.”

“No rational man ever did govern himself, by abstractions and universals.”

I’m pretty sure that Feyerabend would have given me grief about using Bagehot and Burke. But then, he pretty much gave me grief for every view I proposed to him.

Locklin’s comment on Feyerabend shows only that Locklin knows nothing about Feyerabend. The message of “Against method”, for example, is not intellectual nihilism (which I suppose means something like “no distinction can be made between good and bad reasons”). Feyerabend argued at length against people like C. G. Hempel, K. Popper, I. Lakatos, and L. Laudan that there is no one set of rules or procedures adequately characterizes those episodes in the history of science that everyone would agree constituted progress.

His ironic way of putting his conclusion is that if you, the true believer in Scientific Method, think that science must have a method if it is to be rational, then by your own lights — given that in actual science there are counterexamples to every attempt at a philosophical characterization of scientific reason — science is not rational. It’s an ad hominem destruction of mainstream philosophy of science. There is no overarching essence of reason; there are only local “rationalities”. That’s nihilism only if you’re a True Believer.

Feyerabend was not averse to following out the relativizing implications of his view. If there’s no God’s Own Method by which to adjudicate differences between local methods, then adjudication becomes more difficult than mainstream philosophers of science (and especially Popper, F’s teacher and prime target) have thought. It may even be that to adjudicate, e.g. between Western and Chinese medicine, one ends up always to some degree begging the question.

That’s still not nihilism. As long as you remain within Western medicine, there are good and bad arguments, there is evidence and non-evidence, and so on.

It’s too bad that comments like Locklin’s aren’t biodegradable.

How was Feyeraband a “successful physicist”? Also, it seems quite incorrect to refer to Feynman as an uncivilized savage. Any arguments referring to Feyeraband’s science acumen look like they should be taken with a grain of salt.

Posted by Phil | Report as abusive

I don’t think Feyerabend was a successful physicist or really a physicist at all – he never got a doctorate in physics and I don’t think he ever published non-philosophical papers.

The description of Feynman as a savage is right on the money though – Feynman was the first of the generation of quantum theorists who specifically said that they didn’t care about the interpretation of their results in terms of underlying reality.

Posted by dsquared | Report as abusive

For Feynman, try reading “What Do You Care What Other People Think?: Further Adventures of a Curious Character”
& “Surely You’re Joking, Mr. Feynman!”

In both books, he explains that he doesn’t have any use for philosophy, and, at the same time, shows, as far as I’m concerned, that he doesn’t understand anything about philosophy.

Here’s an Amazon Review with more detail:

http://www.amazon.com/review/R2AWMP7NEA3 VHL

However, I did enjoy reading both books way back when.

Thanks for the kind words and link, Felix; in particular the good natured link to my skewering of your outrageous Wired article.

I liked both Feyerabend and Taleb when I first read them. I certainly think criticism is necessary and helpful or I wouldn’t do it (OK, maybe I would; I am a sick bastard). However, if you’re going to kick out the props and say nothing is understandable, you’re not being helpful. You’re giving up on your craft and participation in society. At some point in life I realized this; somewhere around the time I put aside my leather jacket for the Brothers Brooks, and I have disliked this sort of adolescent “nobody knows everything” bong-hit pose since then. Nihilism is an intellectual pose which dates back to the ancient Greeks; if Taleb wants to roll that way, I suggest he takes up residence in a barrel like Diogenes, and stays out of the financial times and policy decisions.

Mr. Locklin,

I think that you are confusing nihilism with Zen. The problem is not that “nothing is understandable”, but that “the words that can be spoken are not true words”.

Your obsession with nihilism (like, perhaps, your Brooks Borthers suit) is neither more nor less than a representation of your own state of mind.

Posted by Anon | Report as abusive

I view Quants as Civil Engineers of past, where many bridges used to collapse but the science of bridge building got progressively better, so will Quant Finance too. It is true that in Finance there are some endogenous risk factors unlike many other sciences where saying that a price will rise will actually cause the price to rise(when u bid on it). So with this kind of self fulfilling prophecies the Quant models are in need of meta knowledge, but all things said it is no reason considering history of science to throw the Quant babies with their bathwater!

Posted by Arun | Report as abusive

Is there truth, is there fact? Or just some absurdly human interpretation of our experiences?

When the mass of people believe in something, then it generally becomes true. The value of gold being my favorite example (good electrical conductor, shiny, rare, difficult to convert into something else). That we believe that Gold represents value is the only reason it HAS value… get a grip. Everything in life is belief.

Taleb’s contribution is that belief has financial limiations because leverage can’t go on forever – there are underlying risks and costs.

Fayerabend says similar things about philosophy of science, that if we end up in a cycle of belief then we can create wonderfully complex models for life, but you need to question the statistical underpinnings and whether we have observed enough to turn belief to fact.

Bayes is the unsung hero of experiential-reality… But that’s a longer essay.

Posted by Nic | Report as abusive

I wrote the following comment on Scott Locklin’s critique of your post on the Gaussian copula, and it’s currently awaiting moderation.
I found your blog post because Felix Salmon, who certainly lives up to his first name, cheerfully provided a link to it in his own blog post of July 21, 2009. Here, you said, “The problem is, houses cost too much because money was too cheap. That’s the financial crisis in one line. It’s not any more complicated than that….” Did I not already have other information and some critical acumen of my own, I might have read your hand-wavy explanation and gone away flattered that I’m among the enlightened.

Assuming that we know what it means for money to be too cheap, that we know when it is too cheap, and that on the given occasion it really was, we’re still faced with the circumstance that housing suffered a bubble this time, while tulip bulbs and companies trading in South America were more or less spared the insane run-up. We must also try to account for the circumstance that house prices didn’t rise at all uniformally across the United States, as one learns from even a little time spent examining the histories of the twenty cities in the Case-Shiller index. Moreover, in order to accept your opinion that “cheap money” was the cause of the housing bubble, we have to assume that hand-wavy quantitative financial analysts, complacent risk managers, authors of legal restrictions on land use, and supposedly powerful members of Congress such as Barney Frank and Chris Dodd are all actually non-causal. So I do think the matter is more complicated than that. Oddly, for one who is so sensitive to others’ simplifications, it appears you may even have made a habit of simplifying. Consider that in this same blog post, you gave us the formula for popular science-writing, then immediately allowed that Isaac Asimov had a much different way of going about it.