Power Without Knowledge: a partial critique of a Critique of Technocracy

As I noted (recall) earlier in the week, a blog post (here) by the French scholar, Cyril Hédoin, got me to read Jeffrey Friedman’s (2019) Power without Knowledge: A Critique of Technocracy (OUP). I may have met Friedman back in 2006, when our mutual friend, Fonna Forman invited us to participate in a panel at APSA. In the process, I wrote Fonna “I looked up Friedman. Interesting character! Is he unemployed now?” Apparently, he was except for his Critical Review editorial-ship. I don’t actually recall the panel itself, and much to my embarrassment, I doubt I eever read Critical Review.

A few years ago, Jeffrey Friedman asked me to review Neil Levy’s Bad Beliefs for Critical Review. He gave me excellent editorial feedback, but I didn’t like the polemical direction he was pushing me. So, I published my review essay elsewhere (here thank you Vasso Kindi). Then I learned he died at the end of 2022, quite young still.

Power without Knowledge: a Critique of Technocracy (OUP, 2019) is a very uneven book. But it is ambitious, and it is written with real urgency; and it repeatedly reminds us of important insights that are recurrently overlooked or pushed aside in the face of each ‘crisis’ or new method or technology that promises huge payoffs. In what follows, I first want to give you a sense of some of the commitments and strengths of the book. (I will also say something about its weaknesses.) I won’t assume you will have read my earlier piece on Friedman, but it can’t hurt if you do. I then want to close with differentiating my own views from Friedman’s.

First, the book reminds us that there are at least three under-determination problems in policy salient social science and human life more general: (i) the very same data can generate different theories/perspectives; he uses this point to help explain phenomena like polarization. (Friedman himself is quite clearly an epistemic holist.) Friedman calls the process a ‘spiral-of-conviction’ and, drawing on Lippmann, he anticipates elements of Kevin Dorst’ account of rational polarization (see especially pp. 241-3). (ii) There is always more data that gets ignored (for all kinds of respectable reasons) that may make existing theories more fragile than they appear. He quite rightly connects this to Hume’s problem of induction (see pp. 167-168) (iii) Even quite robust theories that are successful in their intended domain, may generate serious unintended consequences and we are in a terrible position to know the magnitude of these effects (and if they are, on balance, better or worse than whatever baseline we’re projecting from) (pp. 46-49). All three of these are developed in very subtle ways throughout the book.

As regular readers know, I was once quite infatuated with memes as a tool in human sciences. And the reason for this was that it would be a way to get access to the role of ideas in social change. I eventually soured on it because I couldn’t see how one could avoid Post hoc ergo propter hoc fallacies.

When it comes to policy salient social sciences, I embrace what I have been calling methodological analytic egalitarianism (MAE) (which I plausibly trace back to Hobbes and Mandeville). This approach posits homogeneous agents — or natural equality — and uses observed difference to figure out the social/environmental causes that produce differences (including in embryonic and child development). It’s a neat way to block empirical social science from using, say, a hierarchical or morally flawed status quo as evidence for the naturalness of hierarchy.

For much of the second half of the twentieth century, MAE was, in fact, deployed by people in mainstream economics (and arguably in Rawlsian political philosophy). Most folk that reject MAE (and related views) are either people that flirt with defending hierarchical political projects (eugenics, racial hierarchy, imperialism, etc.) or people that really see difference as emancipatory. Jeffrey Friedman rejection of MAE is quite different in character.

Without, I think, fully appreciating the moral and evidential benefits of MAE to his own politics, Friedman quite rightly notes that, in practice (think of revealed preference, the reliance on incentives/heuristics in explanations), MAE screens off human heterogeneity due to our beliefs/ideas. And building on insights we find in Quine and Lippmann, Friedman correctly notes that even when ideas are tracked somehow in social science, how these are interpreted by individual agents is incredibly hard to capture (radical translation may really begin at home). What social scientists often hope is that the differences average out so that we can make somewhat decent claims about how populations and groups behave. Friedman has a whole battery of arguments that suggest that this faith is somewhat misplaced. And hence this converges with his arguments that emphasize the fragility of policy relevant social science due to other reasons of underdetermination. Keep that in mind.

Now, most people that emphasize unintended consequences and the fragility of social science are often categorized as reactionary, conservative, or populist. And because left populism has been so dormant under Pax America, emphasis on such fragility tends to be coded ‘right.’ Friedman himself tries to forestall this by emphasizing his commitment to a kind of ‘negative consequentialism’* that he associates with Popper’s vision of a politics aimed “the least amount of avoidable suffering for all,” (p. 12) Friedman couples this with commitment to Rawls’ Difference Principle (and redistribution (p. 322)). There is, thus, what we may call a humane social-democratic sensibility at work in Friedman, but without the commitment to progressivism’s embrace of technocracy. I use ‘sensibility’ because the normative side of his commitments is not worked out in any detail and there is no way of knowing how it all hangs together.

In fact, one of the more interesting arguments in Friedman’s book is that in our society a kind of soft-progressivism is endemic. In fact, He calls this a ‘naive technocratic worldview’ (and associated with a naive realism) in which social ‘problems’ are treated as relatively easily amenable to particular interventions or solutions and that doubts or disagreement about these get moralized (pp. 307-315). Again this echoes familiar Conservative arguments, but in Friedman there is never an attempt to suggest that the underlying phenomena that give rise to the sense that there is a problem ought to be treated as inevitable and part of the human condition. His own preference is to provide people with considerable resources so that they can engage robustly in practices of ‘exit.’

Okay, I think this gives a sense of what Power without Knowledge: is about. I have not mentioned some of its virtues yet. First, it has the best account of the so-called Lippmann-Dewey debate. (It’s a bit of a shame that Friedman is unfamiliar with some other excellent treatments of it, and he does not know more recent attempts to renew Dewey’s program — by Kitcher —in democratic oversight of science, and unfamiliar with Nancy Cartwright and Anna Alexandrova’s attempts to make social science be more effective and congruent with values. ) Second, while he is a partisan in the voter ignorance literature, his is the very best account I know of the more general epistemic weaknesses of all the competing approaches in it. Third, he is very alert to the many ways how intellectual libertarian fellow-travelers may instantiate technocratic commitments.

Let me close this post with a general reflection. Friedman’s book builds on the recognition of the many mechanisms that also drive my own so-called ‘Platonic skepticism’ about the ability of truth to dominate mere opinion in a democratic context. As regular readers know, I think such ‘Platonic skepticism’ about the role of opinion is endemic in the history of liberalism (and also one of the reasons why socialists and Marxists too easily conflate left and right liberals with conservatives). I don’t mean to suggest such ‘Platonic skepticism’ is intrinsically liberal. For example, Arendt also tends to embrace it as a feature of contemporary mass society without, thereby, becoming liberal.

In fact, Arendt’s essays on these matters helped me see the following: the public production of truth is rather time-consuming and costly. And, alongside strategic agents trying to shape our perceptions, this is why, for her, we should not expect truth in the public domain. It’s rather the province of rather expensive and slow-moving social institutions of scientific discovery and the justice system. Both of these are fallible, of course, but much less so than public opinion. (Lippmann himself was enmeshed in the world of public opinion on the side of the victors usually, so his critique is really as immanent as it gets.)

In fact, what Arendt’s position entails — and the best treatment of it is 1984 — is that those that fancy themselves powerful, smart, and elite often participate in (if not generate) the very delusions that are common to public opinion formation. (My favorite contemporary example is Elon Musk.)+ But what she tends to downplay — and with her libertarians, Marxists and anarchists —, and Jeffrey Friedman ignores entirely, is that the state itself with its bureaucracies and record-keeping is (to use Nick Cowen’s phrase) a massive machinery of record to help constitute public truth.

I use the language of ‘constitution’ because the state’s role here is not just the provision of a public good that facilitate private transactions (through deeds, registries, administrative rules, etc.) The state is also (recall this post inspired by Tom Pink) witness to truth and generates a massive infrastructure of authoritated beliefs. This is one reason why the state also showers resources on systems of justice and science.

Friedman misses this constitutive role in his (otherwise quite generous) treatment of Lippmann. He sees Lippmann as offering a set of proposals that he (correctly) characterizes as very close to a Dewey style civil service epistocracy. (p. 124; Lippmann is an original advocate of revolving doors among bureaucracy, science, and industry!) And Friedman correctly discerns that these will not suffice for the technocratic and policy goals of handling epistemic complexity. But this assumes — and Friedman is a fantastic critic of this naive epistemic worldview — that the liberal State’s main epistemic goal is to solve problems.

But that’s a mistake, the state’s main epistemic role is to help constitute public truth: this is why, historically, weights, measures, mints, public records, are so important to state building. The state is capable of a long-term and costly commitment in witnessing truth through such practices. Obviously, some of the payoffs of this are to facilitate private market and non-market interactions. But among the main payoffs is the creation of a shared enough public sphere.

Friedman is not wholly unaware of this. He cites James C. Scott’s work generously. But he only does so to critically discuss the ‘homogenizing impulse’ in ‘technocratic manipulation’ and the ‘legibility’ of populations (e.g., p. 136). Obviously, there are many real political dangers when public truth is unaccountable and monopolizes our attention or resources. The criticisms and worries are apt; but they end up missing how without states routinely and massively witnessing truths, we’re left diminished without sufficient tools to handle the complexity of mass society and the very advanced division of epistemic labor. I agree with Friedman that technocracy is not the answer, but the question of how to turn knowledge in responsible and epistemically efficacious power is left unanswered.

---

*In the book he uses ‘negative utilitarianism,’ but since Friedman rejects the epistemic reliability of minimizing/maximizing social quantities I prefer not using ‘utilitarianism.’'

+Friedman writes “January 20, 2017, it does seem…as if a randomly selected voter of the sort long feared by empirical scholars of American public opinion—a voter with virtually no knowledge of public affairs and no awareness that he needed it—had been parachuted into the Oval Office.” (p. 304)