Political Epistemology Beyond Democratic Theory: Introduction to Symposium on Power Without Knowledge

ABSTRACT

Jeffrey Friedman’s Power Without Knowledge builds a critical epistemology of technocracy, rather than a democratic argument against it. For its democratic critics, technocracy is illegitimate because it amounts to the rule of cognitive elites, violating principles of mutual respect and collective self-determination. For its proponents, technocracy’s legitimacy depends on its ability to use reliable knowledge to solve social and economic problems. But Friedman demonstrates that to meet the proponents' “internal,” epistemic standard of legitimacy, technocrats would have to reckon with the heterogeneity of people’s ideas, which he presents as one of two aspects of a political anthropology of ideational beings. The other aspect is ideational determinism: the shaping of our conscious actions by our interpretations, and of our interpretations by “ideational exposures” (which are, to some extent, heterogeneous). For the most part, our symposiasts agree with this anthropology or leave it uncontested, but they fall back on democratic theory to point toward alternatives to technocracy. This raises the question, which Friedman does not ask, of whether his political anthropology undermines a certain brand of democratic theory: the liberal brand that attaches respect to people’s opinions as products of “free reflection,” i.e., as underdetermined.

Keywords:

• democratic theory
• ideational determinism
• ideational heterogeneity
• intellectual charity
• political anthropology
• political epistemology
• political ignorance
• social science
• populism
• Jeffrey Friedman

Until now, the primary critique of technocracy has been supplied by democratic theorists, who have equated technocracy with epistocracy. As an inherently undemocratic project, they argue, rule by epistemic elites is illegitimate on its face—regardless of whether it accomplishes its objectives. Jeffrey Friedman’s Power Without Knowledge: A Critique of Technocracy (Oxford University Press, 2019) poses an entirely different challenge to technocracy. Friedman asks if technocracy may be illegitimate even on its own, negative-utilitarian terms, regardless of its democratic illegitimacy, because it is ineffective at achieving its goal: the relief of human suffering.

The aim of this symposium is to confront some of the many implications of Friedman’s subtle and nuanced critique of technocracy. However, a key aspect of Friedman’s argument, which passes largely without comment in the symposium, is that it is not intended as a contribution to democratic theory so much as an attempt to extend political theory beyond democracy. This may have been missed because, as a critique of technocracy, Power Without Knowledge can be leveraged to support what many theorists take to be their raison d'être: defending the people’s democratic authority against impingement by technocratic experts, among others. Yet Friedman’s critique of technocracy is not grounded in a principle of democratic authority, but in a political epistemology that suggests the possible unreliability of technocratic knowledge claims. This epistemology, in turn, is built on a political anthropology grounded in our status as ideational creatures: creatures who, unlike other biological entities, consciously interpret the world before acting in it. This is an anthropology that may (it seems to me) have disturbing ramifications for regime types other than technocracy, including democracy (in certain conceptions of it), although it is particularly troubling for technocracy.

Because of the sweep and intricacy of the book’s progression from anthropology to epistemology to technocratic critique, my first order of business will be to provide a blow-by-blow account of how the argument proceeds. Section II will then introduce the symposium’s ten responses to the book, followed in section III by brief questions about its implications for democracy and democratic theory.

I. A PRÉCIS OF POWER WITHOUT KNOWLEDGE

The book’s Introduction addresses what Friedman sees as confusion created by the standard view of technocracy as rule by “experts.” By assuming, rather than demonstrating, the validity of experts’ knowledge claims, theorists such as Isaiah Berlin and Jürgen Habermas long ago elided the question of whether technocrats do in fact possess the knowledge they require: knowledge of how to efficiently manipulate individual behavior for utilitarian ends. Following their example, political theorists have persistently bracketed this question.

In Friedman’s view, the standard equation of technocracy with epistocracy contributes to this bracketing of whether technocratic knowledge claims are unreliable, as it simply stipulates the adequacy of the knowledge of technocratic social scientists or other “cognitive elites.” The Introduction additionally points out that the type of knowledge claimed by epistocrats is also claimed, in less sophisticated form, by many ordinary citizens in contemporary democracies. Friedman maintains that in the United States (and, I would add, in Europe), utilitarian knowledge claims are the bread and butter of mass politics. If we understand technocracy as a polity’s attempt to “solve, mitigate, or prevent social and economic problems among its people,” using broadly utilitarian criteria of what is socially and economically problematic, we will be led to acknowledge that ordinary citizens often think that they can engage in this project and, indeed, that they might be right (5; parenthetical page references without citations, here and throughout the volume, refer to Power Without Knowledge.)

Thus, Friedman redefines technocracy as a polity dedicated to solving, mitigating, and preventing utilitarian social and economic problems, regardless of whether these efforts are made by epistocrats or by ordinary citizens (“citizen-technocrats”). This definition suggests that rather than being an inherently elitist project, technocracy is an inherently epistemic one. This is because the technocratic project requires answers to epistemic questions—how widespread are social and economic problems, what are their causes, and how can they be solved?—that are asked, and answered, by elites and masses alike. Thus, an empirical political epistemology of technocracy would examine the actual basis on which both citizen-technocrats and epistocrats make their knowledge claims in any given time and place. Friedman makes a wide-ranging start in this direction in Part II (Chapters 4-6). A normative political epistemology of technocracy would try to determine the likely adequacy of these knowledge claims. This is the goal of Part I (Chapters 1-3).

Chapter 1: Naïve Technocratic Realism

Chapter 1 attempts to pre-empt an objection to both normative and empirical epistemology derived from an epistemic disposition that Friedman, adapting psychologists’ usage, labels “naïve technocratic realism.” The naïve realist holds that some aspect of reality is self-evident, such that our ideas about (for example) how to solve a social or economic problem are not, in fact, “ideas” so much as theoretically unmediated and accurate reflections, or intuitions, of a transparent reality. If this were true of technocratic realities, an empirical political epistemology of technocracy would collapse into the political epistemologist’s own unerring perceptions about how to solve social problems, which would then be projected into the heads of technocrats in the real world (whether epistocrats or citizen-technocrats), who would be seen as perceiving these realities as unproblematically as the theorist does. And normative political epistemology would have no point, as all political actors’ knowledge would be just as reliable as the political epistemologist’s own knowledge, and would require no critical evaluation.

The most obvious objection to naïve technocratic realism is what Friedman calls “the fact of technocratic disagreement.” Insofar as technocrats disagree about what Friedman delineates as the four types of technocratic knowledge—knowledge of (1) the scope of social and economic problems and (2) their causes, and knowledge of (3) the efficacy of solutions to them and (4) the solutions’ costs—then technocratic knowledge cannot be self-evident: disagreement entails the non-transparency of the realities being perceived. The naïve realist, however, might respond to the fact of technocratic disagreement by challenging whether it is really a fact or whether it is, instead, a ruse created by special interests, or class interests, to mask the self-evident truth. Friedman responds to this challenge by pointing out that in actual technocratic debate, it is often argued that technocratic policies cause counterintuitive, unintended consequences. If such consequences are possible, he argues, honest technocratic disagreement would be warranted, since an unintended consequence is, by definition, not self-evident, and thus might legitimately be disagreed about.

Friedman is not, at this early point in the book, in a position to argue that counterintuitive unintended consequences are real. But since plausible mechanisms are routinely proposed that would explain why such consequences might be real, he contends, we can conclude that technocratic debate might always be warranted, as the intuitions that naïve realists treat as reflections of self-evident truth may actually be misleading. Thus, we cannot avoid the need to investigate whether technocratic knowledge claims are, in fact, reliable, nor should we avoid asking how, empirically, people’s technocratic knowledge claims are created—since they are apparently not caused by technocratic realities in themselves. If they were, then there would be no fact of technocratic disagreement.

Chapter 2: The Lippmann-Dewey Debate

In Chapter 1, Friedman did not argue that technocracy is hobbled by irresolvable disagreement, nor that it is plagued by unintended consequences. This is because, as he emphasizes in his reply to the symposium, in Chapter 1 he was merely noting, in effect, that all human action, being the action of fallible beings, is imperfect, such that it might be the subject of legitimate disagreement. This almost truistic fallibilism grounds the need for political epistemology in general, in opposition to the infallibilism of the naïve realist. But no critique of a specific form of human action, such as technocracy, can be built on such a broad foundation. In Chapter 2, then, Friedman begins to produce a deeper foundation by reconstructing Walter Lippmann’s political epistemology, and the responses to it that John Dewey published in the 1920s and beyond.

Friedman portrays Lippmann as not just a fallibilist but an interpretivist, according to whom our ideas about society and government are “stereotypes” based on partial and possibly unrepresentative information samples. Lippmann held that society presents us with too much information for anyone to process. To make this blooming, buzzing confusion intelligible, we must impose interpretations on it that screen out most of it, and screen in only information that our interpretations process, by virtue of inferences about society as a whole, as plausible. In the Lippmann-Dewey debate, Friedman contends, Dewey never recognized this interpretive dimension of Lippmann’s argument. Thus, he treated journalists and social scientists as able to give ordinary citizens the technocratic knowledge they need, but without any explanation of what would ensure that the interpretations that screen in journalists’ and social scientists’ “information samples” will tend accurately to represent the whole of the societies whose social and economic problems they are trying to solve.

Chapter 3: Interpretation, Ideational Determination, and Heterogeneity

In Chapter 3, which is the theoretical heart of the book, Friedman contends that the particular interpretive problem facing technocrats is a function of a further complicating factor, beyond the problem of too much information and, thus, the need for interpretation. This is the problem of ideational heterogeneity.

Friedman brings this factor into view by noting that the problem of interpretation afflicts not only those whom Lippmann studied, namely the subjects of technocratic policy making—citizen-technocrats and epistocrats—but also the objects of their decisions, the agents whose behavior they seek to shape through their policy interventions. Even in areas of technocracy that border on the hard sciences, such as epidemiology or climate policy, technocratic initiatives are legal decrees that are intended to change the behavior of human beings. This entails predicting how human beings’ actions will be affected by a given law or regulation. Thus, technocratic subjects must produce subjective predictive interpretations of the actions of other people, actions that are, in turn, based on those people's subjective interpretations of how they should respond to a given technocratic policy.

Friedman judges this double hermeneutic problematic in light of his political anthropology. Interpretations stem from the organization of our beliefs into structured webs. The nodes of these webs are, in culturally and experientially variegated modern societies, likely to be somewhat heterogeneous across individuals. Thus, the interpretations that guide the conscious actions of those whom the technocrat would predict and control may vary among the population to be predicted and controlled. However, this variability is not statistical noise that, as such, could be predicted by means of probability theory. It is strictly determined, Friedman holds, by nonrandom processes of individual belief formation and of the interpersonal communication of ideas. In addition, interpretations may differ between the subjects and the objects of technocratic prediction and control. The subject may impute to the objects (agents) a certain interpretation of the social environment, but the objects’ interpretations of the social environment may differ from those of the subject (the technocrat). Both of these forms of ideational heterogeneity—within a population and between the population’s members and the technocrat—may cause the technocrat to make predictive errors, leading, in turn, to unintended consequences.

However, Friedman also points out the limits of this cause for concern about technocracy. If everyone’s ideas were completely heterogeneous, there would be no hope for technocratic knowledge; technocrats would be completely at sea. But our ideas are only presumptively and defeasibly heterogeneous, and only to some extent. A defeasible presumption of heterogeneity is grounded in the fact that there is no “unitary human idea generator” (147), but mass media and educational systems are partly unitary idea generators (171). Thus, it would seem, the same cultural forces that give rise to a degree of heterogeneity also give rise to a degree of homogeneity, leading to shared behavioral norms and shared interpretations of the social environment (169).

Part I closes, then, by suggesting that social scientists should, case by case, “judiciously” research the empirical balance between ideational heterogeneity and homogenizing counterforces in the particular populations whose behavior technocrats would attempt to predict and control. Friedman proposes, as the model for this kind of social science, intellectual history, which tackles the webs of belief of the individuals who are its objects of inquiry in all their idiosyncracy, yet does not foreclose ideational homogeneities between interpreter and interpreted that may allow the interpretation to be accurate. A judicious technocracy would embrace the intellectual historian’s defeasible presumption of ideational heterogeneity and would apply it forwards to those who are the object of technocratic prediction and control (not just “intellectuals”), with the aim of identifying whether the targets of specific future interventions will likely be subject to particular ideational homogeneities on the basis of which reliable policy forecasts can be made. This approach requires, above all, the intellectual charity necessary to understand others on their own terms.

In effect, Part II (Chapters 4-6) will take this approach to understanding contemporary technocrats, asking whether the ideational homogeneities they display are likely to allow them to recognize ideational heterogeneity in others. Friedman’s answer is a resounding no, because a lack of intellectual charity is the sine qua non of contemporary social science and popular politics alike. But Friedman leaves no doubt that this could change—if social scientists’ and ordinary citizens’ ideas changed. Part II, then, is both an empirical description of contemporary modes of technocratic thinking and an attempt to trigger, through this description, a “cultural revolution” (342) against these uncharitable patterns of thought. This normative thrust against technocratic culture emerges from Friedman’s claim that contemporary modes of thought are so pathologically uncharitable that it is as if both epistocrats and citizen-technocrats are under a malign pressure to make behavioral predictions for their own sake, regardless of their accuracy. Part II makes this argument over the course of two chapters (Chapters 4–5) that deal with contemporary social scientists and one (Chapter 6) that deals with contemporary citizen-technocrats.

Chapter 4: Economic Theory, Positivism, and Uncharitability

Friedman observes that social scientists tend to hook their claims to predictive knowledge of individual behavior on various objective variables, such as the incentives treated as causal by neoclassical economists, the internal (putatively psychological) mechanisms treated as causal by behavioral economists, and the vast array of other causes, from the trivial to the grandiose, for which positivist empiricists seek correlative observational or experimental evidence. In each case, the price of the social scientist’s claim to predictive knowledge is to foreclose the possibility that those being predicted have idiosyncratic beliefs (or, in many cases, any beliefs at all), since beliefs are subjective.

Thus, neoclassical economic theorists insist, a priori, that “incentives matter,” meaning that people mechanically, and therefore predictably, respond to objective incentives, sans interpretation. This allows the economist-technocrat to assume that she can manipulate people so as to solve social and economic problems by changing the objective incentives they face. Behavioral economists, such as Daniel Kahneman and Amos Tversky, tweak this assumption by adding an ever-growing list of “irrational” (psychological) exceptions to the rule of “rational” (incentives-based) behavior. These exceptions are established through experimentation that treats artificially induced laboratory behavior as the mechanical expression of psychological heuristics and biases, not the ideational product of subjects’ interpretations of the laboratory situation and the tasks given to them during the experiment. More generally, both laboratory and field experiments are predicated on the assumption of human uniformity, which is the only thing that could ground the external validity of their findings across all humankind.

However, Friedman contends that all of these bad ideas are only figuratively a function of a perverse technocratic system intent on prediction at the cost of accuracy. The bad ideas are not caused by an “injudicious” technocracy; they cause it. Thus, for example, the lack of intellectual charity that we see among social scientists whose findings have technocratic relevance is equally hegemonic in the branch of social psychology dedicated to studying the authoritarian personality, an area of research that has few technocratic applications or ambitions—indicating that there is a broader culture of intellectual uncharitability that sometimes happens to lead to “injudicious,” ideationally insensitive forms of technocratic social science. In the authoritarian-personality literature, even arch-enemies of a certain conception of positivism, such as Theodor Adorno (in The Authoritarian Personality), show themselves to be masters of uncharitability, mangling and ignoring evidence of ideational determination and ideational heterogeneity to maintain that what might otherwise seem to be hideous misinterpretations of the social world (such as anti-Semitism) are the products of noninterpretive unconscious psychosocial determinants. Adorno’s complaint about positivism, of course, is that it universalizes for all time the particularities of our present, oppressive era (Adorno [1969] 1976). Friedman would undoubtedly agree, but he goes farther and denies that, even within our era, human beings can be reduced to noninterpretive products of their psychic (rather than ideational) or social (structural) environments.¹

Chapter 5: The Spiral of Conviction

Friedman maintains that an important obstacle to a cultural revolution against positivism and other technocratically pathological patterns of thinking is a tendency toward dogmatism among experts. Here again he builds upon Lippmann, who contended that interpretations confer intelligibility on the world in a manner that encourages partiality. An interpretation judges incoming information with a bias toward confirming what the interpretation illuminates. Only this information is intelligible, plausible, and interesting—in light of the interpretation; the overabundance of other information is screened out as misinformation, or irrelevant information, or craziness. Thus, the more information an interpretation screens in, the more the interpretation will seem to be confirmed, even though it is possible that what is really happening is that a false picture of the world is being unfairly reinforced through the selective perception and retention of information.

This produces a tendency towards interpretive dogmatism among those who have assimilated a great deal of information, a tendency Friedman labels the “spiral of conviction.” Spirals would seem likelier to occur among those whose interpretations have screened in a large amount of confirmatory information, which builds confidence in the interpretation, and Friedman argues that this is what may really be indicated by studies that show “motivated reasoning” to be more intense and widespread among the relatively well informed than the relatively uninformed. Friedman illustrates the operation of spirals of conviction by delving into several technocratic economists’ analyses of what caused the financial crisis of 2008. These highly credentialed, politically influential social scientists were, Friedman contends, aware of information that undermined their mainstream neoclassical view of the crisis, but they were able to dismiss this information as utterly implausible—although doing so required intellectual somersaults that are classic markers of dogmatism.

Chapter 6: Public Ignorance, the Intentions Heuristics, and Populism

The asymmetrical tendency of the well informed to spiral into dogmatism suggests, to Friedman, that epistocrats may be at a cognitive disadvantage relative to uninformed citizen-technocrats.

In his discussion of the Lippmann-Dewey debate, he pointed out that if one accepts (as Dewey did) that social-scientific knowledge is reliable, one starts down a slippery slope toward epistocracy—so long as one also accepts that solving social problems is the overriding goal of the polity. If the goal is to solve social problems, and if social scientists are likelier to know how to do it than ordinary citizens, why should the latter be involved in crafting the solutions? Even if the people are involved in defining the problems, i.e., determining the ends to be pursued, “the fundamental dilemma of democratic technocracy” is, in Friedman’s view, that if technocratic truths are not self-evident, and if we grant that social scientists are likelier than everyone else to gain access to them, the people have no legitimate role in determining the policy means for meeting those ends (108-13). However, if social science, being injudicious, does not provide reliable predictive knowledge, as Chapter 4 suggests may be true, then we have no reason to start down the slippery slope. And while citizen-technocrats, being non-expert by definition, lack social-scientific “knowledge,” their ignorance should also leave them relatively free of the spirals of conviction that would follow from having too much of this type of knowledge, as suggested by Chapter 5. In Chapter 6, then, having delineated the epistemic pathologies of epistocrats, Friedman turns hopefully to citizen-technocrats. Unfortunately, he finds that, at present, they seem to be as injudicious as epistocrats tend to be.

Friedman begins with an exacting demolition of the manner in which political scientists have analyzed survey research on the political ignorance of ordinary citizens. According to these writers, whose analysis reproduces that of the economist Anthony Downs (1957), it is only to be expected that the public will be politically ignorant, as each individual voter has but a minuscule chance of affecting the outcome of a mass election. Voters, knowing these odds, correctly estimate that it is not worth their time to “study up” on politics and government, as it is highly unlikely that their votes will affect public policy or the conduct of government. This is a classic “incentives matter” story that, as usual with neoclassical economics, imputes to agents accurate subjective knowledge of the objective incentives they face.

Friedman contends, however, that in fact there is no evidence that voters are aware of the odds against their individual votes mattering, and that there is much evidence against it. There is, he points out, survey evidence indicating that they think their votes do matter; there is the massive fact that they continue to vote, by the hundreds of millions, every year; and there is the fact that when they vote, they do not cast random ballots. Friedman exploits this last fact to argue that voters must think that their vote choices are adequately informed: if they did not, they would be unable to decide whom they should vote for, and would have no way of proceeding other than by flipping a coin. The real question, then, is not why they vote despite the odds, but why they think themselves capable of rendering sound political judgments on the slender informational foundation that survey research shows that they often have.

To answer this question, Friedman assembles survey and focus-group data suggesting that in the United States in recent decades, voters have often tended to treat the solutions to social and economic problems as so straightforward that they do not require very much knowledge or even very much thought. Rather, they seem to think, social and economic problem solving requires leadership that is forceful enough to overcome the resistance of special interests to doing what “obviously” needs to be done. In short, Friedman suggests, American voters have been naïve technocratic realists. Accordingly, they have viewed politics as a contest that pits “strong” politicians with good intentions against either “weak” but good politicians or evil ones—those with bad intentions, who have been bought out by the dark powers of the establishment. Such voters, Friedman infers, will tend to assume that all they need to know about politics is which politicians are well intentioned and strong.

In this analysis, naïve technocratic realism takes shape as a real empirical phenomenon (and a volatile, populist, Manichean one)—not merely, as in Chapter 1, a theoretical position that would militate against political epistemology. However, the connection between naïve realism in Chapters 1 and 6 is more than merely fortuitous. A naïvely realistic electorate, fixated on the question of politicians’ intentions, will be as bereft of an awareness of unintended consequences as the naïvely realistic theorist constructed in Chapter 1. Thus, the true concern, for a normative political epistemologist of technocracy, is not the fact that voters may be ignorant of the often-trivial political information tested for by survey researchers, but that they will be ignorant of the very possibility that public policy may backfire. Moreover, such voters will tend to reject the possibility of genuine technocratic disagreement, like the cynical naïve realist of Chapter 1, for they will be unaware of any epistemic difficulties that might justify such disagreement—such as the difficulty we face, as technocrats, in putting ourselves into the shoes, or rather the heads, of those whose behavior we intend to manipulate.

The first thing to note about the argument of this chapter is that it is intellectually charitable to ordinary citizens in their technocratic capacity—despite Friedman’s contention that they are populists in the making, ready to follow their tribune as he tramples institutional and normative constraints in the fight against special interests. It might be useful to contrast this intellectual charity to the stance taken by epistocratic critics of democracy, such as Bryan Caplan (2007) and Jason Brennan (2016). These writers follow Downs in crediting voters with knowing the crucial fact that their (individual) votes do not really matter—a flattering concession to public knowledge, given that no theorist of democracy recognized this mathematical truth until Anthony Downs did in 1957. Yet intellectual charity is not a matter of projecting flattering thoughts into the minds of others; this is what analytic philosophers do under the guise of “interpretive charity.” Rather, intellectual charity is a matter of trying to grasp what is actually in the real minds of real people, and Friedman contends that there is no reason to think that very many contemporary voters have the knowledge that Downs, Caplan, and Brennan ascribe to them.

Moreover, when such theorists use voters’ “recognition” that their votes don’t matter to explain both public ignorance and the policy views of ordinary citizens, their conclusion is that while citizens are rational in deciding to be ignorant, this decision licenses—in their minds—irrational opinions, since voters “know” that these opinions will not be translated into law. However, as Friedman argued in a long review essay on Caplan’s Myth of the Rational Voter (Bennett and Friedman 2008), the assertion that voters’ opinions are “rationally irrational” tells us nothing about the specific opinions voters hold. The theory of rational irrationality licenses voters to believe anything they might fancy, but cannot tell us why voters fancy the particular ideas they hold except by saying, tautologically, that they must “prefer” these “irrational” opinions over others or they would not hold them. By contrast, Friedman’s approach is to ask what it would be rational for voters to believe if they were raised in a culture that, like ours, is hospitable to democratic technocracy. In a culture in which it is endlessly reiterated that voters’ individual votes do matter and, moreover, that this is a good thing, a citizen might be expected to infer that each voter’s opinions are valuable regardless of how well-informed the voter is: we do not, after all, impose knowledge requirements on voters of the sort that epistocratic theorists favor. Thus, they might conclude, the issues about which they have opinions are readily deciphered; anyone can do it, and everyone is encouraged to do it in the voting booth.

This line of reasoning, in the mind of a citizen-technocrat interpreted with intellectual charity, produces what Friedman calls the ontology of a simple society, in which social and economic problems are easily solved—since, epistemically, the solutions are readily deciphered. On this foundation, Friedman builds up the ideal type of a “naïve technocratic worldview” that seems to fit actual public opinion and actual voter behavior, while creating a compelling picture of the inner life of the citizen-technocrat in cultures like ours (283-303). One key feature of this picture is the voter’s equation of good intentions with good policy outcomes and bad intentions with the creation or perpetuation of social and economic problems. Friedman points to a great deal of evidence suggesting that voters bypass any consideration of unintended policy consequences, such that good or bad intentions are all that matter. This collapse of consequences into intentions would make sense, given the simple-society ontology.

Another useful contrast here might be with defenses of voter knowledgeability produced by democratic theorists such as Habermas (2006) and Robert Goodin and Kai Spiekermann (2018), according to which underinformed citizens can get by with knowledge proxies such as cues and heuristics. Logically, such proxies work only if voters are sufficiently knowledgeable to be able to assess their accuracy—in which case, Friedman points out, knowledge proxies would not be needed. Moreover, it seems likely that voters who fit the profile Friedman develops would be uninterested in deliberately choosing knowledge proxies to begin with, as they would not think they need to know much beyond the intentions of candidates (and parties, and bills). In effect, theorists such as Habermas attribute to voters the same awareness of their need for more-than-intuitive knowledge that neoclassical economists such as Caplan ascribe to them, producing a deliberate search for heuristics. By contrast, Friedman sees the use of intentions as “heuristics” as an as-if process: citizen-technocrats act as if intentions are valid proxies for knowledge of technocratic consequences, but they actually believe that good intentions ontically cause good consequences and bad intentions cause bad consequences. This is intellectually charitable to citizen-technocrats because, given the premise of a simple society, it is quite reasonable for them to assume that intentions will not face complications on the road to outcomes.

Part III: Third-Order Uncertainty about Second-Order Heterogeneity

Chapter 6 concludes Part II on a doubly tentative note. First, Friedman maintains, it is not clear whether it is better to be ruled by relatively open-minded citizen-technocrats who are, however, innocent of the very possibility of unintended consequences; or by relatively well-informed epistocrats (well-informed, at any rate, about the possibility of unintended consequences, if not about the heterogeneous ideas that may cause them) who are, however, as experts, likely to be relatively dogmatic. Friedman calls this the Hobson’s choice of technocracy. More importantly, though, from the perspective of a theorist of technocracy rather than a theorist of democracy, Parts I and II do not establish the magnitude of the errors that might be committed by injudicious technocrats or even by judicious ones. This uncertainty is expressed in a three-page section (317-19) that leads off Part III.

Given our inevitable ignorance of future history, Friedman points out, we cannot know if the unintended consequences of the mistakes made either by judicious technocrats or by clueless citizen-technocrats, doctrinaire neoclassical economists, or positivist empirical researchers will outweigh the social and economic problems their efforts might solve. Yet if we are to be truly internal critics of technocracy, our most natural standard of legitimacy will be the utilitarian standard of doing more good than harm, overall. Since we (theorists) cannot, however, know the magnitude of the mistakes that judicious technocrats or injudicious technocrats will make in the face of ideational heterogeneity, this internal standard of legitimacy does not allow us to render normative judgment on either form of technocracy—judicious (ideationally sensitive) or injudicious—because we cannot know if the unintended consequences of the mistakes outweigh the social and economic problems that particular technocrats will mitigate in the unknowable future.

In his reply to the symposium, Friedman emphasizes that this type of uncertainty is distinct from the problem of ignorance facing injudicious technocrats with which we are left at the end of Part I: ignorance of the ideas of the agents manipulated by technocratic policies. The problem with which we are left at the end of Part II is theorists’ third-order ignorance about the performance of technocrats, who in turn will encounter second-order ignorance of the ideas of the human objects of their efforts at prediction and control when they (technocrats) come up with first-order policy analyses and recommendations. Any political theorist (not just a theorist of technocracy) who wants to render a normative judgment about a political system that will produce emprical tendencies, but not certainties, will face third-order ignorance. No political theorist can divine the future, so no political theorist can forecast the magnitude of effects produced by any political system—unless by means of a philosophy of history like Marx’s.

Lacking such a philosophy of history (which I suspect Friedman would accuse of insensitivity to the contingencies of people’s ideas), Friedman suggests that all we can do is make logical comparisons among alternatives that allow judgments of their relative, not absolute, consequences. Just as it is logical that a judicious (ideationally sensitive) technocracy is preferable to a technocracy that, like our own, ignores the ideational determination of conscious human action, Friedman proposes that to evaluate the legitimacy of a judicious technocracy, we need some other point of comparison. He finds this in exitocracy, which he sketches in Chapter 7.

Chapter 7: Capitalism, Socialism, and Exitocracy

Exitocracy is the ideal type of a technocracy that takes advantage of a more reliable epistemic device than even judicious technocrats’ attempts to understand and predict the ideas of the objects of their actions. This device is what Albert O. Hirschman (1970) called “exit,” in contrast to “voice.” In applying this dichotomy to technocracy, Friedman includes the ratiocinations of epistocrats as well as those of citizen-technocrats under “voice.”

Friedman argues that exit enables people, to some extent, to take advantage of their knowledge of their own unhappiness in a given situation versus their satisfaction after they leave this situation behind. Even as he details how inaccurate this type of personal knowledge can be, he maintains that, for all its imperfections, it is more reliable than the four types of technocratic knowledge, because the latter require even judicious technocrats to grasp the conditions of society as a whole and the presumptively heterogeneous ideas of the millions of anonymous agents within it. Recurring to Dewey’s famous metaphor, Friedman writes that it is easier to judge whether one’s own shoes pinch than to decide which shoes will avoid pinching the feet of everyone in a mass society (321). Both injudicious and judicious technocrats must “judge the significance of social problems for anonymous others . . . speculate about their causes, and . . . speculate about the efficacy of various solutions and the side effects they may cause. . . . Inasmuch as it is inherently difficult for anyone to have such knowledge—even those who are judiciously attentive to ideational heterogeneity—the exitocratic alternative would appear to be the better one,” epistemically speaking (330).

Friedman constructs exitocracy as a society somewhat like our own, in principle. It has a capitalist element, in that competition among owners of the means of production can provide workers and consumers choices of employment, goods, and services from which they may exit if they are dissatisfied. There is also a robust private sphere in which non-economic exit and personal voice play a significant role. However, economic exit is a dead letter without the economic means to take advantage of it, so Friedman sketches a neo-Rawlsian Difference Principle with “entry goods” replacing “primary goods,” grounding a potentially radical wealth redistribution (335).

Plainly, both the capitalist and socialist elements of this ideal type require technocratic judgments. Thus, exitocracy is a type of technocracy, not an alternative to it. It is an “extraordinary” type of technocracy, though, in that it aims, wherever possible, to devolve decision-making authority to exit users, rather than using the authority of voice to block exit (322).

As a form of technocracy, exitocracy faces the same second-order epistemic problem—ideational heterogeneity—that faces “ordinary” judicious and injudicious technocracy. As Friedman puts it,

The administration of an exitocratic Difference Principle would require the administrators to make behavioral predictions that nobody—no matter how judiciously attentive to ideational heterogeneity—is well positioned to make. . . . A judicious technocrat attempting to decide which level of taxation is “excessively” high . . . would have to answer the empirical question of how the ideationally heterogeneous capitalists of her time and place will respond to various possible rates of taxation. This question is no less difficult to answer than the question of how they might respond to various levels of a minimum wage or to other “ordinary” technocratic initiatives. Thus, the epistemological critique of technocracy presented in Part I would apply to redistribution in the service of enhancing people’s entry opportunities. The same critique would apply, as well, to the provision of other public goods, including the good of legally establishing and defending the private sphere. All of these exitocratic activities would entail highly fallible, speculative behavioral predictions about anonymous, presumptively heterogeneous others. (340-41)

Nevertheless, the fact that exitocracy is a type of technocracy is, in Friedman’s view, a feature, not a bug. The technocratic status of exitocracy is what allows it to serve the comparative function that is the rationale for Part III. It is true that the “ordinary” technocratic aspects of an exitocracy are epistemically problematic, but just as a judicious technocracy would address this problem by attempting to charitably understand the ideas of the populations whose behavior is being predicted, an exitocracy would address the problem by devolving decision making authority, wherever possible, from collective voice users to individual exit users. This would reduce the scope of ordinary technocratic decision making, producing a comparative advantage over judicious technocracy, but without completely eliminating ordinary technocracy, which is impossible. Thus, we have a “hierarchy of technocratic legitimacy” (343), with exitocracy (or rather, as Friedman clarifies in his reply, judicious exitocracy) at the top, followed by judicious ordinary technocracy, followed by injudicious ordinary technocracy.

Of these three options, injudicious ordinary technocracy is the least equipped to deal with the problem of ideational heterogeneity. But since there is nothing below injudicious technocracy in the hierarchy, we have nothing worse to compare it to. Therefore, Friedman holds, we cannot pass judgment on its stand-alone or absolute legitimacy. We know that it is relatively illegitimate compared to either a judicious exitocracy or a judicious ordinary technocracy, but if pressed to say whether our own technocracies, which are primarily injudicious, are illegitimate in the abstract, Friedman maintains that we must be silent. However, this does not entail silence about the cultural status quo. In noticing the pathological epistemic consequences of contemporary ideas, we are enlisting comparisons “upward” in the hierarchy, contrasting the injudicious (ideationally insensitive) status quo against a culture in which we came to grips with our ideational determination and heterogeneity.

Cultural critique of this sort is what Friedman envisions as the future of the political epistemology of technocracy. This would entail a shift of theorists’ attention from institutions and policies to ideas. Just as Friedman did in Part II, future theorists would attempt to critically assess (and, unlike Friedman, historically contextualize) the culture that generates injudicious technocracy—not just the culture of positivism and neoclassical economics, but of intellectual uncharitability, “of moralizing, and opinionating, and arguing; the culture of the inhuman humanism that effaces our ideational nature” (345). Thus, we return to political anthropology.

The political anthropology of ideational heterogeneity explains our disagreements without recourse to dehumanizing psychological or structural explanations (or, as I think Friedman would say, pseudo-explanations). And the political anthropology of ideational determinism entails a culture of tolerance, for nobody deliberately chooses to hold bad ideas: therefore, they cannot be blamed for holding them. Our task, then, as political actors and theorists alike, is to understand and resist bad ideas, not demonize as willfully evil those who hold them. This includes understanding and resisting the dehumanizing methods that currently blind us to their existence as ideas, and which occlude the interpretive and thus fallible nature of our own ideas. Thus, cultural critique includes an auto-critique of the antihumanist ideas that have invaded our own webs of belief. Power Without Knowledge, then, is ultimately a book that is less about technocracy than false consciousness.

Afterword: Technocracy and the Left

In a brief afterword, Friedman follows Foucault in comparing technocracy to early modern Cameralism, about which he makes two points. First, Cameralism was, as Foucault saw, a project with a potentially infinite remit. Its goal was to increase state revenue by tending to the welfare of the society that generated it, and this opened absolutely everything to state control. Friedman (349) quotes the historian of Cameralism, Keith Tribe (1988, 75): “‘The range of matters that eventually became the object of regulation . . . is quite bewildering in its variety,’ in that anything and everything, including not just the processes of industry and agriculture but the conduct of what we would call personal life, might bear on the people’s well-being and therefore be a fit object of gute Polizei. (Polizei—usually translated as ‘police,’ sometimes as ‘policy’—was derived from polis.)” Second, Cameralist writers recognized that this very wide remit entailed great epistemic demands, but they had a quite naïve view of how to meet these demands: namely, by writing ever-more-massive encyclopedias to advise Cameralist bureaucrats about “everything from silver mining, viticulture, fruit growing, vegetable gardening, and animal husbandry to medicinal herbs, flowers, ‘brewing, bleaching, advice on how to get spots off clothes, how to get rid of headaches’” (350, quoting Tribe 1988, 25-26). The needed knowledge was vast but readily available. A recognition that knowledge of something so difficult to understand as people’s ideas might be required was very far from the Cameralists’ imagination, as they were naïvely ignorant of ideational complexities that might cause unintended consequences.

Following historians such as Daniel T. Rodgers (1998), Friedman then speculates that this naïveté was passed down to the contemporary left by Progressives, who were frequently educated by such students of Cameralism as Gustav von Schmoller and Adolf Wagner. Unlike the Cameralists, however, the Progressives were democrats. Thus, their epistemic naïveté concerned the knowledge of the people, not the bureaucrats. (Newspapers, not encyclopedias, were to be the transparent medium of “the facts.”) We inherit, then, two left traditions: the democratic, reformist, technocratic-utilitarian, Progressive, post-Cameralist tradition and the democratic, revolutionary, non-utilitarian, non-technocratic, emancipatory Marxist tradition.

In the former tradition, the naïve and simplistic Cameralist approach remains unchallenged, awareness of unintended consequences remains untheorized at best, and ideational humanism remains unimaginable. This tradition should also be hostile to the message of Power Without Knowledge on the grounds that, in challenging technocratic legitimacy, it challenges the legitimacy of democracy—conceived, narrowly, as democratic technocracy. (I will return to this point in section III.) At the same time, the capitalist element of exitocracy should prompt hostility from those in the Marxist tradition, despite the fact that the redistributive element would enable people to exit from social problems more effectively (arguably) than would even judicious popular attempts to centrally plan solutions to them. Anticipating these hostile responses to the book, Friedman suggests that both traditions have backed themselves into an embrace of technocracy not because they, any more than the Cameralists, have come up with an epistemic justification for technocracy; but because both traditions, like Cameralism, have simply failed to think about the political epistemology of technocracy.

II. CRITICAL RESPONSES TO POWER WITHOUT KNOWLEDGE

I would like to thank Jonathan Benson, Étienne Brown and Zoe Phillips Williams, Kevin Elliott, Oscar Larsson, Adam Lerner, Eric MacGilvray, Alfred Moore, Zeynep Pamuk, Robert Reamer, and Julian Reiss for their thought-provoking contributions to the symposium. Each of their papers is preceded by an abstract that summarizes their arguments in some detail, so in this section I will restrict myself to saying a few words about conceptual relationships among them. I will delay until section III any mention of Friedman’s extensive responses to each of these contributions. Friedman’s reply is a major work in its own right that puts each of the symposium papers, which fasten on one or at most two chapters of the book, into the context of the book’s argument as a whole.

For example, Reamer’s paper, and especially Benson’s, focus on the “ordinary” technocratic features of an exitocracy (Chapter 7). Contrary to Friedman, they consider these features to be flaws that undermine the legitimacy of exitocracy, given (for the sake of argument) that ideational heterogeneity poses a serious problem for behavioral prediction. Insofar as ordinary injudicious technocrats are unable to make reliable predictions of the behavioral effects of their policy proposals, Reamer and Benson argue, exitocrats should likewise be unable to predict the effects of the various institutions, such as property rights and redistributive regimes, that they might implement. Therefore, Reamer turns to anarchism as an alternative to exitocracy, while Benson alludes to an alternative regime of “pure procedural legitimacy.”

Benson’s paper consists of a detailed list of the ordinary technocratic projects in which an exitocracy would have to engage, and the unintended consequences that might flow from them. In addition to providing an abbreviated list of the same kind, however, Reamer’s paper minimizes the magnitude of ideational heterogeneity, on the Wittgensteinian grounds that language homogenizes the ideas of the native speakers of a given language. Reamer goes so far as to suggest that technocrats in our day capitalize on this homogeneity, allowing them to make accurate inferences from the behavior of population subsamples (so long as the larger population speaks the same language as the subsample). Thus, Reamer is defending the judiciousness of ordinary technocracy while also proposing anarchism as an even more extraordinary type of alternative to it than exitocracy is, as anarchism would more comprehensively institutionalize the exit mechanism by doing away with the state’s monopoly legal system. Anarchism is intended, however, to supplant exitocracy as a superior form of technocracy, while Benson’s pure procedural legitimacy would abandon technocracy altogether—assuming (for the sake of argument) that ideational heterogeneity is so troublesome that it undermines all forms of technocracy.

Larsson, like Reamer, attempts to minimize ideational heterogeneity, but he does so by attributing homogenizing effects not to language, but social structures. He then proposes that to some extent, some technocracies already seize on this homogeneity in order to predict human behavior reliably (just as Reamer contends that technocrats already seize on linguistic homogeneity to predict behavior reliably). The fact that these technocracies are injudiciously inattentive to ideational heterogeneity does not matter to Larsson, because he, like Reamer, minimizes the impact of that problem. That Larsson does so by citing Foucault on the predictability of human behavior—but in order to endorse technocracy, rather than condemn it—makes his contribution particularly fascinating. He attempts to square this apparent circle by appealing to the democratic use of structural homogeneities, whereby stakeholders “attend to public problems, often with a technocratic outlook, in an effort to resolve social and political problems through the pragmatic utilization of existing knowledge” of the homogeneities. Larsson thus merges a version of Foucault with a version of Habermas.

Lerner takes a different path than Larsson’s and Reamer’s. He thinks that Friedman underestimates heterogeneity because he notices only heterogeneous ideas, not heterogeneous mental qualia. Even though accounting for diverse qualia would put even more pressure on judicious technocrats, Lerner proposes no alternative forms of technocracy or alternatives to technocracy, as he is interested in ideational heterogeneity as a pure researcher, not as an institutional reformer. However, in his field of research—international relations—he finds grounds for bracketing the individualist ontology that Friedman connects to ideational determinism. Friedman locates the site of both heterogeneous and homogeneous ideas in the webs of belief characteristic of individual minds, while Lerner wants to allow for the possible existence of collective ideational entities, such as states.

Larsson’s and Lerner’s contributions engage with Friedman’s political anthropology more directly than do any of the other papers. Brown and Williams, for example, defend the appropriateness of naïve technocratic realism, at least in some cases, which has the effect of bypassing the question of ideational heterogeneity. As I said in summarizing Chapter 1, a rejection of naïve realism, while a necessary precondition for a critical epistemology of technocratic knowledge claims (or any others), is insufficient as a critique of any particular type of knowledge claim (such as technocratic knowledge claims). Such a critique requires something specific, such as the presumption of ideational heterogeneity, that would raise doubts about whether this particular type of knowledge claim is more dubious than human knowledge in general—which is fallible, yes, but not necessarily so fallible as to be unreliable. By focusing on naïve realism to the exclusion of ideational heterogeneity, then, Brown and Williams are foregoing a critique of Friedman’s critique of technocracy. This may be linked to concerns, expressed at the end of their paper, about the democratic legitimacy of technocracy. If technocratic truths are self-evident, as suggested by a defense of naïve realism, then surely they can be put to work by ordinary citizen-technocrats, securing democratic legitimacy for technocracy but ignoring the challenge to technocracy—ideational heterogeneity—that stems from Friedman’s political anthropology. The same approach seems to undergird Brown and Williams’s suggestion that, at least in extremis, we should take “epistemic gambles”—even in the absence of judicious ideational knowledge. Without the need for such knowledge, ordinary citizen-technocrats can govern by intuiting that, at least in otherwise-hopeless situations, some technocratic policy would be worth a shot in the dark, once again taking advantage of a type of self-evident truth.

MacGilvray, I think, might be said to make a similar end run around the problem of ideational heterogeneity. Drawing on the classical tradition and on Weber, he defends the possibility of an open-minded technocracy that is judicious—but only in the colloquial sense of the term. For MacGilvray, exposure to conflicting perspectives can attune citizen-technocrats to the complexity of political issues, as Aristotle and Polybius maintained. And responsible leaders, of the sort envisioned by Weber, can sort through the claims of dueling, doctrinaire epistocrats. MacGilvray’s focus, then, is on the “complexity” of policy consequences in a general sense, and on the spiral of conviction, but not on the predictive problem posed by ideational heterogeneity.

Pamuk and Reiss, in contrast, do grapple with ideational heterogeneity, but indirectly. Both authors attend exclusively to Friedman’s critique of positivist epistocracy. While neither is persuaded that positivist social science is unable to cope adequately with ideational heterogeneity, both are open to that possibility. Pamuk, however, would like to see evidence of the magnitude of the difficulties caused for positivist technocracy by ideational heterogeneity, while Reiss, who endorses the claim that “pernicious” ideational heterogeneity is a serious problem, nonetheless believes that there can be positivist progress toward an understanding of the laws of human behavior.

Another similarity between these two papers is that insofar as technocracy cannot cope adequately with pernicious heterogeneity, both papers depict non-technocratic democracy as the obvious normative alternative. Pamuk sees popular sovereignty as an alternative to technocracy simply because, as a form of sovereignty that is not grounded in epistemic claims, its legitimacy is immune from epistemic doubts. Reiss conceptualizes technocracy in the standard way, as epistocracy, such that “technocracy” becomes a foreign entity inside the democratic body politic. He therefore sees ideational heterogeneity and other forms of pernicious heterogeneity—among many problems for positivist social science that he beautifully schematizes—as plaguing experts, not ordinary citizens. In this view, what Friedman calls the “fact of epistocratic disagreement” (69) becomes the central problem for technocratic legitimacy, because all of the problems discussed by Reiss are bound to cause experts to disagree about how to solve social and economic problems; yet the demos is, by definition, unqualified to adjudicate these disagreements. Therefore, technocracy (epistocracy encased in democracy) is, for Reiss, “a bad idea”—which seems to leave a democracy purged of technocracy as the alternative.

Like Benson, then (if we take democracy to be a pure procedure of the sort he has in mind), Pamuk and Reiss think of democracy as beyond technocracy: a form of decision making that need not confront the epistemic quandaries explored in Power Without Knowledge. None of these authors, however, explain why democracy should serve this role. To use Pamuk’s example, Hobbes viewed absolute sovereignty as an alternative to epistemic uncertainty and conflict; yet Hobbes’s sovereign was hardly democratic. An absolute democracy would remove us from the epistemic vagaries of technocracy, but so would an absolute monarchy or any other regime in which power is vested unconditionally in some individual or group, regardless of the consequences of their actions. Of course, democratic theory abounds in reasons for preferring democracy to other regime types, but from Friedman’s deliberately adopted internal normative perspective, technocratic legitimacy stands or falls on its consequences—its ability to deliver the (negative-utilitarian) goods—not on whether it allows for popular participation or consent. In this sense, technocracy is beyond democracy, even when the former is conceptualized as including a democratic element, democratic technocracy.

Elliott’s and Moore’s contributions, unlike most of the others, touch on the legitimacy of democratic technocracy. In line with Friedman’s remarks about populism, Elliott contends that democratic technocracy has supplanted democracy simpliciter in the minds of a significant body of citizens, who have unwittingly picked up Progressive normative assumptions. These assumptions undermine democracy not in the fashion of Friedman’s “fundamental dilemma of democratic technocracy” (the slippery slope from democratic technocracy to epistocracy, given certain epistemic assumptions), but because if technocracy does not deliver the goods, post-Progressive citizen-technocrats are perfectly willing to endorse nondemocratic alternatives. Elliott, then, can be read as envisioning non-technocratic democracy as an alternative to democratic technocracy, just as Pamuk and Reiss, and possibly Benson, envision non-technocratic democracy as an alternative to epistocracy.

In a sense, the same is true of Moore. He objects to Friedman’s redefinition of technocracy to include the category of democratic technocracy, for while this definition illuminates the lawless tendencies of technocratic populists, who want social and economic problems solved at any cost, it obscures non-technocratic aspects of modern democracy, such as the pluralism that Moore sees as being opposed by both populists and epistocrats. If we follow Friedman too far, according to Moore, “the critique of technocracy becomes an undifferentiated critique of more or less the whole of modern democratic politics,” such that even pluralism is subsumed by technocracy: “parties compete to offer solutions to social problems, epistocrats claim to be able to solve social problems without input from voters, and populists claim to know how to solve social problems if only they could win the acclaim of the majority and execute their program without obstruction or compromise.” This picture of modern politics is precisely what Elliott finds illuminating about Friedman’s redefinition of technocracy: it allows us to see how far modern democracy has strayed into technocratic problem solving. I should note that Friedman does not actually claim that democratic technocracy has completely subsumed democracy. In the book’s Preface, he recognizes the rise of a decidedly non-technocratic emancipatory politics alongside democratic technocracy (ix), and in its Afterword, he alludes to Marxism as the historical source of this type of politics (351). But Elliott and Moore are referring to Chapter 6, where Friedman paints such a compelling picture of the ideal-typical dynamics of naïve democratic technocracy that one comes away seeing it almost everywhere one looks in contemporary mass politics.

I hope to have suggested that all the papers but two (Lerner’s and MacGilvray’s) attempt to escape the problems identified in Power Without Knowledge by turning to democracy (although this is far from the only thing the papers attempt to do). Brown and Williams, like Larsson, attempt to defend democratic technocracy by pushing back against Friedman’s critique of naïve technocratic realism—which Friedman depicts as the backbone of democratic technocracy as currently practiced. Pamuk, Reiss, and perhaps Benson turn to non-technocratic democracy as an alternative to epistocracy. Elliott turns, by implication, to non-technocratic democracy as an alternative to the populist tendencies of democratic technocracy. Finally, Moore is concerned to defend non-technocratic democracy against both populist antipluralism and what he sees as the antidemocratic implications of exitocracy. Thus, he worries about the prospect that the “architects” of exitocratic policies and institutions would want to shield their creations from popular interference. Of course, from the internal perspective of technocracy, that is irrelevant. Legitimacy stems from getting the job done, regardless of who does it; this is why either citizen-technocrats or epistocrats, or both, might be the architects, depending on who is thought to know what they are doing. Like each of the other appeals to democratic legitimacy, then, except Larsson’s and Brown and Williams’s defenses of naïve technocratic realism, Moore’s paper reaches beyond technocracy to democracy.

But it is not really a critique of Power Without Knowledge, which is, after all, not defending technocracy in any of its forms (including exitocracy) so much as exploring the ramifications of technocracy from within, bracketing the legitimacy of its utilitarian orientation. I read the book to be saying that if one attaches normative significance to solving, mitigating, or preventing social and economic problems, then one must face up to the attendant epistemic demands, which should drive one up the hierarchy of legitimacy, from injudicious technocracy to judicious ordinary technocracy to judicious exitocracy. If one does not want to follow this logic, however, as appears to be true of many of the contributors, one needs either to deny the epistemic demands by defending naïve realism or challenging the importance of ideational heterogeneity; or else to renounce, or at least subordinate, the goal of social and economic problem solving. Only then can democratic legitimacy be allowed to prevail over a putative technocratic legitimacy that may be, but often is not, democratic.

III. POLITICAL EPISTEMOLOGY BEYOND TECHNOCRACY

In assessing the symposium, then, I think we must begin by reiterating that Power Without Knowledge is not intended as a contribution to democratic theory. As Friedman says in the book’s Introduction, only one relatively minor strand of democratic theory, the Utilitarian-Progressive strand—which finds few (if any) defenders among contemporary theorists—equates political legitimacy with technocratic problem solving (18-19). This is why Power Without Knowledge is best seen as an attempt to expand political epistemology beyond democracy. Nonetheless, I wonder if Friedman’s political anthropology does not pose a challenge to at least one mainstream justification of democracy—not democratic technocracy, but democracy as a form of egalitarianism predicated on mutual respect, or what Niko Kolodny (2014b) describes as “social equality.”² This type of justification may stand behind some contributors' desire to return from technocracy to democracy.

As we have seen, Friedman’s political anthropology consists of two elements, ideational heterogeneity and ideational determinism. The first of these does most of the heavy lifting in his critique of technocracy. Thus, his response to the symposium, which discusses the anthropology even more extensively than does the book itself (where the term political anthropology does not appear), fixes attention on the positivist leanings of the majority of his interlocutors (Benson, Larsson, Lerner, MacGilvray, Pamuk, Reamer, and Reiss) precisely because positivism, in Friedman’s view, has no anthropology—officially. The positivist is willing to consider any independent variable as a possible determinant of human behavior, such that nothing is ruled out as constitutive of the human. Yet unofficially, Friedman argues, all independent variables are required, by the positivist, to be capable of imposing themselves on all human beings, grounding external validity and thus the aspiration to produce lawlike generalizations. Thus, positivist anthropological openness, or nullity, is shadowed by a refusal to contemplate causes of human behavior that might be perniciously heterogeneous, including human ideas. For this reason, Friedman argues, the commitment to universalism that makes positivism an ideal device for generating empirically grounded technocratic behavioral predictions also guarantees that these predictions will be insensitive to the empirical diversity of human ideas, making positivism unscientific and eroding the reliability of its predictions.

Ideational heterogeneity, then, receives the lion’s share of Friedman’s attention, both in his reply and in his book. It is true that ideational determinism links ideational heterogeneity with pernicious behavioral heterogeneity: one aspect of ideational determinism is the determination of conscious behavior by subjective interpretations. This is the aspect that is most salient to the critique of technocracy, as the epistemic goal of the technocrat is behavioral prediction. But the other aspect of ideational determinism is the mental dimension: the creation of interpretations from beliefs and of beliefs from mental operations performed on ideational inputs that are, for the most part, communicated to each of us (somewhat heterogeneously) from other people. The inputs that are communicated to us determine our webs of belief, and these webs determine our conscious actions (including communicative actions). To be sure, at each step we may consciously work on the materials given to us, but Friedman assumes that this work, the work of thinking, is nomological, determined by neurological architecture that is not itself under our control (137). Thus, the whole process is determined. The communicative starting point of the process distinguishes Friedman’s view from atomistic individualism, as he notes in his reply to Lerner, but it also raises the prospect that we are robots programmed by ideational inputs, each marching to the slightly discordant beat of the ideas to which we happen to have been exposed over the course of our lives. Thus, the opinions that, according to an important strand of democratic theory, deserve respect, and entitle us each to an equal share of political power, would seem to have no value in themselves, as they are in fact (in Friedman’s view) the strictly determined products of whatever newspaper articles, blog posts, Twitter screeds, movies, novels, and courses in political theory we happen to have read or watched. This is particularly true for our political opinions, as we know (or think we know) very little about politics, government, or society that is not communicated to us from others.

What may be problematic here, for democratic theory, is not the heterogeneity or homogeneity of these ideational exposures, nor their fallibility, but the fact that there is no escaping them: we cannot resist our beliefs, which are coercive over us or, perhaps better put, constitutive of us. Friedman contends that in this sense, free will is a myth. To believe in it “is to believe in uncaused causes: determinants of action that come from nowhere to close the putative gap between reasons and actions.” Friedman denies that there is such a gap: “an agent whose action stemmed from her free (underdetermined) will would lack sufficient reasons for action: that is to say, ideational determinants. Human minds, however, are saturated with reasons for action, and we know that these are sufficient to determine their actions because otherwise (lacking sufficient reasons for an action) a human being would be unable to act. Like Buridan’s Ass, she would be unable to ‘choose’” (142).

Setting aside the age-old philosophical questions, this is a compelling picture of what political ideas—not just technocratic ideas—look like. What predictability we do find in them stems from the cultural streams (journalism, high culture, mass culture, pedagogy) to which we and our fellow citizens are exposed; both the heterogeneities and the homogeneities across individuals are traceable to each individual’s ideational exposures. Seen in this light, one must wonder why we are each entitled to equal respect for our political beliefs rather than, say, equal concern for our welfare—the starting point of technocracy. Pursuing the deterministic side of Friedman’s ideational anthropology, then, widens the object of political epistemology to encompass all the political thoughts, and conscious actions, of human beings; but having left technocracy behind, it may also return us to technocracy because, arguably, it makes more sense to care for one another as sensual, suffering creatures than to respect one another as transcendent minds.

The last section of Friedman’s reply reiterates a theme taken up in his chapters on Lippmann and on democratic technocracy: the scandal that neither political scientists nor political theorists pay much attention to journalism, despite the fact that virtually all knowledge of political issues must flow to us through journalistic hands. But the same scandal attends political scientists’ and theorists’ tendency to neglect cultural influences more widely, beyond mere journalism. Does it not matter that our ideas about justice, equality, emancipation (and democracy itself)—not merely our ideas about the scope, causes of, and solutions to social and economic problems—are mediated to us by human interpreters who are themselves ideationally determined?

Our mutual determination by the streams of self-confirming ideas we communicate to one another would seem to defeat any case for respecting our beliefs as independent or free, even as it dictates that we must treat one another with intellectual charity if we want to understand the idiosyncracies that separate us from one another. Intellectual charity acknowledges the empirical reality of ideational determination even as it undermines the notion that any idea deserves respect apart from its objective truth value. Democratic theorists should find much to think about in Friedman’s warning (in his reply) that any of us—the people at large, technocratic elites or masses, journalists, or political theorists—can get caught in spirals of conviction even more insidious than the shadows on the wall of the cave.

Notes

1 An important part of Chapter 3 is likewise a critique of Charles Taylor’s social holism, which, like Freudianism and Marxism, effaces ideational determination. Although Taylor, as a cultural theorist, does not deny people’s determination by ideas, he does (Friedman argues) portray the ideas expressed in cultural wholes as functional responses to self-evident social truths, not contingent products of autonomous streams of ideas (autonomous from both human will and objective realities).

2 Kolodny’s epic examination of the justification of democracy argues that it amounts to “equal opportunity not only for informed influence, but also for autonomous influence: influence knowingly in accord with judgments that are themselves reached by free reflection on what one takes to be relevant reasons” (Kolodny 2014a, 310, emph. added). Though Kolodny leaves the resemblance unremarked, this principle is similar to Rawls’s “criterion of reciprocity,” according to which contributors to public reasoning over the form of social institutions must “think it at least reasonable for others to accept them, as free and equal citizens. . . . Citizens will of course differ as to which conceptions of political justice they think the most reasonable, but they will agree that all are reasonable, even if barely so” (Rawls 2005 [1997], 446). Similarly, Amy Gutmann and Dennis Thompson maintain that democracy is “a conception of government that accords equal respect to the moral claims of each citizen, and is therefore morally justifiable from the perspective of each citizen” (1996, 26).