I recently read a paper by the anthropologist David Graeber entitled ‘The Sword, The Sponge, and the Paradox of Performativity: Some Observations on Fate, Luck, Financial Chicanery, and the Limits of Human Knowledge‘. Graeber sent it to me because we are hoping to write an article on the emergence of probability theory and its application in the financial markets.
The working title of our paper is ‘The Betrayal of Freedom and the Rise of the Future Machines’. The basic idea is to show that the predictive powers of social sciences — including economics and finance — were shown to be fairly vacuous in the 1960s from a variety of different directions. The response by the horrified professions was to bury the evidence and double down on probabilistic prediction. This coincided with the rise of finance and the whole thing produced the weird world of meaningless numbers and extreme instability that we face today.
Anyway, here I will provide a gloss on Graeber’s excellent paper as an accompaniment to my recent piece on the anthropology of the economics profession. It is not available online except behind a paywall but I urge readers to seek it out. It is one of the best psychological/anthropological descriptions of how and why people — from village elders to econometricians — try to use arcane and difficult methods to predict the future and dictate how people should organise themselves socially and economically. In order to discuss the paper I must first introduce to non-anthropologists two key terms.
The first is ‘mana‘. This is a difficult term to pin down as it has any different dimensions in many different cultures. But Graeber’s main angle of attack in the paper is that mana is a power that people believe they can gain control over to predict and influence the direction of future events.
The second is ‘performativity‘. Performativity is a sort of ‘thinking/doing makes it so phenomenon’. For example, the Queen of England is the Queen of England because everyone believes her to be the Queen of England. If everyone in the world stopped believing tomorrow that she was the Queen of England she would cease to be the Queen of England. Her social position is literally only real insofar as we believe it. Her royal actions and symbolism are thus a way to ‘perform’ this belief and reinforce it.
Now, onto the essay. Graeber thinks that many of the phenomena that anthropologists know as mana are actually very similar to concepts that we in the West employ such as fate, luck, chance and probability. Graeber notes that we as a society have been taught to think of events in terms of probabilities. But this is not present in non-Westernised cultures. He cites the following conversation he had with an educated Malagsy while he was in Madagascar doing fieldwork:
David: What do you think the chance is that a bus will come in the next
David: I was thinking of running up the hill to get some cigarettes. I figure
it’ll take maybe five minutes. What do you think the chance is that
a bus will come before I’m back?
Zaka: I don’t know. A bus might come.
David: But is it likely to?
Zaka: What do you mean?
David: You know, what’s the chance? Is there a very large chance it will
come? Or just a small chance?
Zaka: A chance can be big or small?
David: Well, is it more like 1 in 10? Or more like 50–50?
Zaka: How would I know? I don’t know when the bus is going to arrive. (p32)
It is clear that Zaka the Malagasy thinks David the American’s questions to be absolutely bizarre. It simply does not make any sense to him. Despite being educated he does not even think to ascribe to chance a numerical estimate. Graeber concludes:
Even when my Malagasy did become fluent, I never heard people employing language in the way that people would do so in America, for example, “I’d say 3 to 1 the cops won’t even notice that I’m parked here.” In fact, I discovered not only that such a way of thinking was unknown to most Malagasy, but also that, once explained, it seemed just as peculiar, exotic, and ultimately unfathomable as any of those classic anthropological concepts, such as mana, baraka, or s´akti, regularly employed in other parts of the world to put a name on the play of chance or to explain otherwise inexplicable conjunctures or events. Once I began to think about it, I realized that this puzzlement was a pretty reasonable response. Chance actually is a very peculiar concept. Zaka was right: the main thing is that we do not know when the bus is going to arrive. This is the only thing that we can say for certain. Anything could happen—the bus might break down, there might be a strike, an earthquake might hit the city. Of course, all these things are, from a statistical perspective, very unlikely, million-to-one chances, really. But it is that very application of numbers to the unknowable that struck my Malagasy interlocutors as bizarre—and not without reason. What a statistical perspective proposes is that we can make a precise quantification based on our lack of knowledge, that is, we can specify the precise degree to which we do not know what is going to happen. (pp32-33)
I am entirely in agreement with Graeber here. We can assume that the bus was not on a strict timetable because we are not dealing with an advanced and well-organised society. Thus giving the chance that the bus would arrive while Graeber went to get cigarettes a numerical estimate was a pretty mystical thing to do altogether. (I do find it amusing that Graeber said that the chances of a strike, the bus breaking down or the earthquake hitting were ‘a million-to-one’ though… it seems like even he finds it difficult to get away from his pre-conceived cultural way of thinking about such things!).
Think about this. If I have a coin that is fairly balanced I can give an objective probability estimate. The chance of the coin landing on either heads or tails is 0.5 or 50%. This number is not mystical. It is objective. But when Graeber asks what the chance of the bus turning up is or whether the cops will notice that he is parked illegally, these estimates are not objective. They are entirely mystical. They are, in fact, as Graeber rightly points out, the same sorts of magical thinking that many so-called primitive cultures use to try to grapple with the future.
The amusing thing, however, is that we actually employ people and give them social prestige to engage in these mana-like numerical concoctions. This is just like in supposedly more primitive societies where soothsayers and astrologers are given status as village elders and power to decide how the society should be organised. In our Western societies we hire economists and financial experts. Graeber writes:
Almost invariably, too, there are certain specialists who claim privileged, exclusive knowledge. In very hierarchical societies, elites will either try to monopolize such matters themselves (e.g., Azande princes maintain exclusive rights to officiate over the most important oracles) or attempt to forbid them as forms of impiety (both Catholic and Sunni authorities have been known to do this at one time or another). There is also a frequent, although not universal, tendency for these techniques to draw on forms of knowledge seen as foreign and exotic: the Arabic lunar calendar in Madagascar, Chinese numerology in Cuba, Babylonian zodiacs in China, and so on. (It is not the past, perhaps, but the future that really is a foreign country.)
From this perspective, it is quite easy to see that economic science has become, in contemporary North America above all, but in most of the industrialized world (or, perhaps better said, financialized world), exactly this sort of popular ‘technology of the future’. There are specialists who try to keep a monopoly on certain forms of arcane knowledge that allow them to predict what is to come, although in a way that, insofar as the situation becomes political, inevitably slips into performativity. At the same time, fluctuations in the financial markets, speculation on stocks, investments, and the machinations of commodities traders or central bankers, all these have become the stuff of everyday arguments over coffee or beer or around water coolers everywhere—just as they have become the veritable obsessions of certain cable watchers and denizens of Internet chat pages. There is also a tendency—quite typical of such popular technologies of the future as well—for idiosyncratic (‘crackpot’) theories to proliferate on the popular level. (pp39-40)
Graeber is absolutely correct here. This is where we come to the notion of ‘performativity’. If we examined the situation objectively we would quickly see that these people are mostly engaged in soothsaying but we do not. Why? Because the performance buttresses our hierarchical social and economic structures and lends them credibility and weight. We do not want to know the reality: namely, that our social structure is determined in line with political and social power. And thus we create fictions that ground it as being somehow ‘objective’.
The statistical estimations are mostly about performance in this regard. They are similar to a symbolic ceremony by the Queen of England. And in the financial markets this performance generates dynamics of its own. For example, when everybody is convinced that the markets will be calm they will get statistical read-outs saying that the markets will be calm precisely because they are acting as if the markets will be calm. But when they start to panic because something unseen happens, their statistical read-outs will suddenly change.
In truth their ‘future machines’ are just feeding back to them their own activity. It is like an animal looking in a mirror and thinking that another animal is staring back at them. In reality, it is just their own reflection. The whole situation would be hilarious but these dynamics are wreaking havoc on our societies. They are also trapping us as political actors because they give us a sense of fatalism about the future. They encourage us to think that the ‘experts’ have the whole thing figured out and that ‘politics’ should be structured in line with this. This is the true poison of modern economics and it is what makes modern economists such dangerous clowns. The unfortunate thing is that almost every single one of them, wrapped in their socially-sanctioned delusion of scientificity, have absolutely no idea what they are doing.
Thanks. The Graber paper sounds very good. Your last two paragraphs are particularly interesting and relate to my book about reflexivity: http://emergenteconomics.com/book/ . One of the things i’ve since wrestled with is whether or not reflexivity is a special case which only occurs in specific areas at specific times — which is what Soros suggests in stock markets — or whether it’s a general phenomenon which characterises most things. I’m beginning to suspect the latter.
Every sphere that involves self-other relationships.
The key references in this regard are Part III of Sartre’s Being and Nothingness and Part II Chapter 4 of Merleau-Ponty’s Phenomenology of Perception. Both available online and linked to there.
I wasn’t a fan of Being and Nothingness and thought Sartre better expressed his philosophical ideas in his plays and novels. Nausea, in my view, is a more convincing platform for conveying his thoughts about the importance of subjectivity than B&N. In Camera/Hell is other people also in some ways better describes the difficulty of intersubjective relations than does a straight philosophical text. Similarly The Outsider is one of the better expositions of existentialism.
There are lots of uses of reflexivity in the social science and anthropology literature: Keynes, Soros, Mannheim, Giddens, Beck, Geertz, Ruby, Bloor, Bourdieu…
In the novels and plays Sartre is describing things in terms of allegory and metaphor. The philosophical work actually provides analytical tools that I believe can be used in analysing how people make decisions in the face of uncertainty. I’m going to write a paper on Sartre, temporality and decision-making at some point.
Dan, I would be glad to know more about your work on reflexivity. Your suspicion is quite right. But it is interesting how von Neumann and Morgenstern got around this problem in the theory of games. What I mean is that they were able to circumvent the problem of the content of other minds. Probably only a genius like von Neumann could have successfully engineered this maneuver.
But Soros does not appear to want to do this, that is, to avoid confronting the content of other minds. In so doing, he is facing potentially formidable logical problems, which technically, involve matters of epistemic logic should he wish to drill deeply into the epistemic structure that reflexivity inevitably entails.
“Think about this. If I have a coin that is fairly balanced I can give an objective probability estimate. The chance of the coin landing on either heads or tails is 0.5 or 50%. This number is not mystical. It is objective. But when Graeber asks what the chance of the bus turning up is or whether the cops will notice that he is parked illegally, these estimates are not objective. They are entirely mystical.”
I am not sure “mystical” is the right word here, though.
If you can’t get an objective, numerical probability, then you might get an epistemic probability, on the basis of empirical evidence and inductive argument. If someone tells you that they have seen a bus come along some road every day at 3 pm for years, you have the basis for a reasonable belief that it is a posteriori probable a bus will come on that road today at 3 pm. This probability is not an objective, numerical probability, of course — and you could be wrong, but it is still a reasonable epistemic probability.
Where there is very little or no evidence, you may not be able to make any reasonable epistemic probability too, of course. Where no evidence exists at all to make any judgement, you face what Keynes would have called “fundamental uncertainty.”
I would say that a good way to classify probabilities in line with Keynes’ thinking is as follows:
There is not sufficient evidence and even if there were the events must be assumed to be non-ergodic unless you can prove otherwise (e.g. you know that the cops are scheduled to check that place at a given time every day and how often they miss their schedules or that the bus has a timetable and you know all the prior instances when that bus was late etc.). When a numerical estimate is given in the circumstances that we are taking about it is largely mystical. It is very similar to the ‘mana language’ that supposedly primitive societies use to deal with unknown events.
Frankly, I don’t think Keynes went nearly far enough on this. GLS Shackle properly studied how these issues relate to the type of examples that Graeber is giving.
“When a numerical estimate is given in the circumstances that we are taking about it is largely mystical.”
OK, I have no problem with that.
Though the point above I made is different: about epistemic probabilities.
Regarding Shackle, I would be a bit careful about his views on probability and uncertainty, since he takes a rather extreme view:
See Crocco, M. 2002. “The Concept of Degrees of Uncertainty in Keynes, Shackle, and Davidson,” Nova Economia 12.2: 11–28.
I don’t think his views are nearly as extreme as certain Post-Keynesians make them out to be. Yes, he is skeptical of a very large amount of economic work. He even goes so far as to say that economists should limit how much policy advice they should give. But I think he’s right on both counts.
I tend to agree with you LK.
I think that when Graeber is asking for a numerical estimate, he’s asking for a degree of belief. Or at least it seems to do so since he thinks about probability alongside fate, luck, and chance in its article (which I can’t read, unfortunately).
So it seems that we need to make a difference between the economist/financial expert on the one hand, and another individual on the other. When the first is talking about and giving numerical estimates, he may believe he’s doing something scientific and objective while the second is mainly sharing his degree of belief about what, according to him, is the most probable given what he knows (or believe to know) at a certain time and place. For instance, when Graeber talked about the earthquake, he could have used another huge fraction instead of ‘a million-to-one’ to express the same message : he did not expect to see that kind of event to happen when he was talking. The number in itself does not count, only its scale. And I’m not sure Graeber believed he was doing something very accurate. If we understand the usage of numbers in that way, that is, not as an objective truth but as a subjective belief, do we still have “to get away from [t]his pre-conceived cultural way of thinking about such things”? The problem of course, as you emphasized perfectly, is when others “think that the ‘experts’ have the whole thing figured out and that ‘politics’ should be structured in line with this.”
Also, in that perspective, Graeber said :
“But it is that very application of numbers to the unknowable that struck my Malagasy interlocutors as bizarre—and not without reason.”
What if he asked the same question about the arrival of the bus, but NOT in a numerical manner? If numerical values are used in the west to “scale” our degree of beliefs, maybe Malagasy use another procedure to do so. Also, before the rise of probability, how do we dealt with these kind of questions? If we asked them tout court.
And that doesn’t include the use of these delusions (and the deluded) as a means to gain power by those who either are true believers or understand the delusions and use them for their own purposes. Visions of the Wizard of Oz come to mind.
Well, under the Bayesian interpretation of probability, all probabilities are literally degrees of belief, though I’m not sure how many economists, or people who design complex financial derivatives, interpret probability this way. Lars P. Syll and Noah Smith have posted some rants against Bayesian probability some time ago.
Bayesianism is at the root of this nonsense. Their so-called ‘priors’ are exactly what I am talking about when I refer to mystical numbers.
I think that frequentism has its fair share of subjectivity as well, although I should mention that I’m not a statistician by training. The following article illustrates well what I’m taking about:
As an economics noob,
“…that the predictive powers of social sciences — including economics and finance — were shown to be fairly vacuous in the 1960s from a variety of different directions. The response by the horrified professions was to bury the evidence and double down on probabilistic prediction. This coincided with the rise of finance and the whole thing produced the weird world of meaningless numbers and extreme instability that we face today.”
Doesn’t that sound like something Adam Curtis would say in the beginning of his documentary? Just imagine it in your head. It sounds so apt! 😀
AS I have argued, economic phenomenon are generated by nonergodic stochastic processes — accordingly applying probabilities derived from existing data is of no use in predicting future outcomes.
Clearly, patently, though, human beings APPEAR to have and/or develop minds that tend to do exactly that, i.e. to assign and apply probabilities to future events that “are generated by nonergodic stochastic processes.” Not to mention to treat ‘the very coinage(s) of their brain(s)” (individual or inter-subjective) as “really there” or “really real” and (gasp!) as having independent causal powers and effects. Alienation/estangement and reification are not truly eradicable, it appears.
The 64 Trillion Dollar Question: Is money, like Queen Elizabeth’s title, and, indeed, like all titles and entitlements, also a “performativity”?
Reblogged this on Sindelo The Genius.
I’m just a dumb economist so I may not understand your argument but how are economists maintaining a monopoly on their arcane information? We publish our research in outlets no less accessible than you, and train a thousand or so new economists every year, not even counting all the people without PhDs who claim that title- without the ill effect that wold attend someone claiming the title medical doctor or electrician, I might add.
Um, by process of training them you are maintaining a monopoly. It’s very simple really. A single producer can sell a lot of goods to a lot of people but its still a monopoly. This is pretty basic stuff. Come on.
Talk about simple stuff, our research is non rival and as I explained, the title economist is non excludable, hence not a private good. I think you don’t know much about economics.
No such thing as a public or semi-puiblic monopoly? Get out of here… Really! Get serious. You think those students are protesting over nothing. Wake up, buddy…
Doesnt that also raise the question of whether those research outlets you mention, even though accessible, are open to heterodox point of views?
* that is open to receiving and publishing heterodox point of views
Oh, come on. Do you think a non-economist can get published in an economics journal or get grants to do economics-type work? Even dissident economists (heterodox economists) barely have access to mainstream journals. And economists rarely (only~20% of the time) publish their data sets, making what empirical work they do do largely immune from review. The economics profession is hermetic.
Oh, you don’t think that almost bring the world banking sector down and would have done so had not certain governments bailed them out, systems run by your graduates, isn’t a serious ill effect?
Have you seen Donald MacKenzie’s An Engine, Not a Camera? It’s a performative analysis of the financial market, but I have to say I’m interested in this paper of David Graeber. I read his book on Debt and I’m doing a research on performativity (my dissertation was about performativity and econometrics), but I’ll see if I can find it.
Pingback: Econometricians, Financial Markets and Uncertainty
Pingback: Econometricians, Financial Markets and Uncertainty: An Anthropological View | BRANDAKA
Phil, since you are going at this from an anthropological perspective, I have some references/papers I can email you if you wish. You and David have predecessors, but I am sure you know this. Besides, this does not take away from your own individual approach to the subject. David will no doubt be aware of Evans-Pritchard’s Witchcraft, Oracles and Magic Among the Azande and the studies by Levy-Bruhl and the somewhat controversial ones by Levi-Strauss. But a more recent treatment of risk assessment is Rebenato’s Plight of the Fortune Tellers.
Picking up on Evans-Pritchard if I may, my fellow grad students and I were one day trying to understand what it was that so-called primitive people didn’t understand about causality (just so it is clear, we rejected outright the notion of the “primitive” other than as related to technological development).
A chap goes down a forest path and is mauled by a tiger. Now, his explanation is that someone is out to get him for something he might or might not have done. Now, does he believe that had he been a little earlier or later, the tiger would have missed him? He agrees. Does he believe that had he taken another path, the tiger would have missed him? He agrees. Does he not now believe that what happened to him is a coincidence and not the consequence of human malevolence? No, he doesn’t. And this is because he rejects something we take virtually for granted, the concept of mind-independent causal chains.
Put another way: instead of the rejection of mind-independent causal chains, we have the rejection of the accidental concurrence of independently manifested causal chains. In either case, the chap in question entertains and inhabits a somewhat different conceptual world than we do, though there is considerable overlap at certain points. And this is so even though we both inhabit, in a quite strict sense, the same physical world.
The same could be said about neoclassical economists compared to heterodox economists, that the members of these two somewhat heterogeneous communities entertain and inhabit rather distinct conceptual worlds concerning the economy, while the economic system itself is the same for both. In putting the issue this way, I am, implicitly, ignoring one sort of relativism and rejecting another.
Reblogged this on synthetic_zero and commented:
review essay on David Graeber on the “paradox of performativity”