Experts have knowledge not possessed by others. Those others, the laity, must decide when to trust experts and how much power to give them. We hope for a “healer” but fear the “quack,” and it is hard to know which is which. Experts may play a strictly advisory role or they may choose for others. Psychiatrists, for example, may have the legal power to imprison persons by declaring them mentally unfit. There is, then, a “problem with experts,” as Turner (2001) has noted. When do we trust them? How much power do we give them? What can be done to ensure good outcomes from experts? What invites bad outcomes? And so on.
Different people know different things, and no one can acquire a reasonable command of the many different fields of knowledge required to make good decisions in all the various domains of ordinary life, including career, health, nutrition, preparing for the afterlife, voting, buying a car, and deciding which movie to watch this evening. Each of us feels the need of others to tell us things such as which medical therapy will best promote long-term health, which religious practices will produce a happy existence after death, which dietary practices are most likely to postpone death, which schools are most likely to launch our children into good careers, what hairstyles are the most fashionable, and which car will be the most fun to drive. In our work lives, too, we rely on the expertise of others. Corporate managers require the opinions of experts in many areas, including engineering, accounting, and finance. Investigating police officers require forensic scientists to tell them whether there were “latent” fingerprints at the crime scene and whether any such prints “match” those of the police suspect. A civil attorney may seek expert opinion on the risks of using a product that harmed their client. And so on. We rely, in other words, on the opinions of experts. We rely on experts even though we are conscious of the risk that experts may give bad advice, whether from “honest error,” inattention, conflict of interest, or other reasons. The “problem of experts” is the problem that we must rely on experts even though experts may not be completely reliable and trustworthy sources of the advice we require from them.
The problem of experts has been recognized in some degree and in some form, however vaguely, from an early point in the history of Western thought. But it has emerged as a distinct problem only quite recently. The earliest explicit treatment of a general problem of experts may be Berger and Luckmann (1966), who link the problem to the division of knowledge in society. “I require not only the advice of experts, but the prior advice of experts on experts. The social distribution of knowledge thus begins with the simple fact that I do not know everything known to my fellowmen, and vice versa, and culminates in exceedingly complex and esoteric systems of expertise” (p. 46).
The problem of experts originates in the “social distribution of knowledge” that Berger and Luckmann emphasize. I will examine the social distribution of knowledge in Chapters 6 and 7. The basic idea, given earlier, is simple enough, however: Different people know different things. The division of labor entails a division of knowledge. No one would deny this humble truth. As we shall see, however, there seems to have been only limited clear awareness of the division of knowledge in Western thought before F. A. Hayek (1937) elevated it to a central theme of political economy. The idea has even been denied or decried by skilled scholars. And few thinkers have been willing and able to recognize the implications of dispersed knowledge as understood by Hayek. As we shall see, Hayek drew our attention to the ways in which dispersed knowledge is embedded in practice rather than in books, formulas, and rational thought.
In Chapter 6 I will discuss the surprising fact that the modern literatures on the problem of experts to be found in sociology and in science and technology studies can be traced to a source only rarely cited in those literatures, namely, Hayek’s 1937 article “Economics and Knowledge.”
It would be difficult or impossible to recognize a general problem of experts without a clear and explicit recognition of both the existence and the importance of the division of knowledge in society. We should not be surprised, therefore, that it was only after Hayek’s articulation of the problem of dispersed knowledge that we had a clear and explicit recognition of a general problem of experts.
Nor is it likely that one could clearly see a problem of experts in a society lacking a relatively high degree of Weberian rationality. Weber’s meaning is contested. I will therefore define “Weberian rationality” as the practice of applying procedures of counting, numerical measurement, and conscious deliberation to a large number and variety of choices in human life, the tendency to articulate ends pursued, the further tendency to apply an efficiency criterion to the choice of means to achieve those ends, and a disposition to seek constancy, harmony, and order among the ends pursued. (See, generally, Weber 1927 and Weber 1956) Modern corporate enterprises, of course, exhibit a high degree of Weberian rationality. This sort of rationality contrasts with Marco Polo’s representation of thirteenth-century business practices. (See any standard edition of his memoir, which is usually entitled The Travels of Marco Polo in English.) The particulars of his story cannot all be trusted, of course, and it is disputed whether even the broadest outlines are factual. It is nevertheless notable how little careful calculation seemed to have entered his business dealings as described in his famous narrative. His business venture was an adventure. Presumably, the uncertainties of the venture would have overwhelmed any attempt at careful calculation. And, importantly, double-entry bookkeeping was not yet available to him. (Marco Polo’s adventure is dated 1271–95, which precedes any plausible date for the “first” example of double-entry bookkeeping. See Yamey 1949, p. 101; de Roover 1955, pp. 406–8 and 411; Edwards 1960, p. 453.) Without double-entry bookkeeping, good calculations would be difficult or impossible.
In a world of low Weberian rationality there will be priests, soothsayers, physicians, jurists, and other experts. But it seems unlikely that any thinker in such a world would very clearly anticipate the “problem of experts” considered in this book. Only when a large variety of choices are made at least in part through Weber-rational processes is it reasonably possible for a social observer to recognize a general problem of experts.
Earlier thinkers may not have been in a good position to identify a general problem of experts. They did produce, however, some analyses that addressed questions of expertise. Sometimes experts are viewed favorably, especially by writers putting themselves forward as experts. We can also find complaints against experts. Aristophanes, for example, mocked Athenian oracle mongers (Smith 1989). In Henry VI, Shakespeare has Dick the butcher exclaim, “The first thing we do, let’s kill all the lawyers.” The line may have alluded to Wat Tyler’s Peasant Revolt of 1381, whose partisans “expressed a particular animosity against the lawyers and attornies” (Hume 1778, vol. 3, p. 291), perhaps because “they had deliberately advised the barons to reinstitute servile labor as the just and traditional due” (Schlauch 1940). The physicians in Molière’s L’Amour médecin are incompetent fops more likely to kill than heal. DeFoe (1722) contrasts skilled physicians with the “Quack Conjurers” (p. 33) brought forward by the plague. He laments “the foolish Humour of the People, in running after Quacks, and Mountebanks, Wizards, and Fortune tellers” (p. 42).
Only very recently has Western thought produced any clear and coherent statements of a general problem with experts. Nor have we yet had a common discussion of the problem of experts. Scholars producing similar and complementary analyses do not seem to be aware of one another. We can nevertheless identify a literature on experts stretching back to antiquity with the aid of a simple classification. This simple classificatory scheme helps to us identify works that are relevant to a theory of experts even when they do not include a clear and coherent statement of a general problem of experts. Indeed, as we shall see, the literature includes theoretical treatments of experts in specific domains. In some cases the only problem with experts seems to be that nonexperts are not sufficiently obedient.
In this and the next two chapters, I will discuss many past thinkers. But I will not attempt to provide anything approaching a complete survey of the literature on experts. There are important figures I will not mention at all, such as Scheler (1926), Szasz (1960), Tetlock and Gardner (2015), and Conrad (2007). There are others, such as Jurgen Habermas (1985) and James B. Conant, who will be mentioned, but given, perhaps, insufficient attention. Rather than attempting a survey of the literature on experts, I have tried to show that there is a literature on experts. This literature is spread across several fields, including philosophy, science and technology studies, economics, political science, eugenics, sociology, and law. I believe the simple taxonomy provided in the next section creates a common framework for comparing different theoretical treatments of experts. It makes them commensurable.
Theories will differ according to how they model experts and how they model nonexperts. Sometimes experts are modeled as disinterested, neutral, objective, “scientific,” incorruptible, and so on. Call such experts “reliable.” They are reliable because nonexperts can rely on them to provide good guidance. Sometimes experts are modeled as interested parties who may act on local, partisan, or selfish motives, or as subject to unconscious biases, or as corruptible, and so on. Call such experts “unreliable.” They are unreliable because nonexperts cannot always rely on them to provide good guidance. Nonexperts may be modeled as rational, reasonable, active, skillful, or otherwise competent. Call such nonexperts “empowered.” They have, potentially at least, the power to form a reasonable estimate of the reliability of experts or, perhaps, to override or ignore the advice of experts or decisions made for them by experts. Nonexperts may be modeled instead as unable to reason well, subject to the supposed dictates of “culture” or “ideology,” or otherwise passive and causally inert. They lack the potential to competently judge or to choose among expert opinions. Call such nonexperts “powerless.” They do not have the power to form a reasonable estimate of the reliability of experts or to competently override or ignore decisions made for them by experts. This taxonomy gives us four categories for a theory of experts: reliable-empowered, reliable-powerless, unreliable-empowered, and unreliable-powerless. Any theory of experts can be fitted into this humble taxonomy.
Technical fields such as psychometrics give the word “reliable” a technical meaning from statistics that I am not invoking here. Kaye and Freedman (2011, p. 227) say, “In statistics, reliability refers to reproducibility of results. A reliable measuring instrument returns consistent measurements.” Thus defined, reliability is distinct from validity. “A valid measuring instrument measures what it is supposed to” (Kaye and Freedman 2011, p. 228). These scientific definitions are important in many contexts, including assessments of the quality of forensic-science evidence. The word “reliable” may also have, however, the broader meaning I invoke in this chapter.
My taxonomy does not touch all dimensions. It does not address how or whether the theorist models themself. Does the theorist model themself as an “expert”? Does the theorist express sympathy for some parties over others? And so on. And it may be difficult to classify a theory in which the reliability of experts or the power of nonexperts is endogenous. Nevertheless, any theory will contain a model of the expert and a model of the nonexpert, and we can usually classify the experts as generally “reliable” or “unreliable” and the nonexperts as generally “empowered” or “powerless.” Table 2.1 represents my simple taxonomy.
A taxonomy for theories of experts
| Nonexperts are powerless | Nonexperts are empowered | |
|---|---|---|
| Experts are reliable | Theories in this box tend to view experts favorably. Some theorists in this group have argued in favor of experts deciding for nonexperts in at least some areas such as reproduction. | Theories in this box tend to view nonexperts as dependent on experts, but reasonably able to ensure that experts serve the interests of nonexperts. |
| Experts are unreliable | Theories in this box tend to view the thinking of both experts and nonexperts as determined by historical or material circumstances. These circumstances may be oppressive to both experts and nonexperts, but experts tend to be in a position of dominance over nonexperts. | Theories in this box tend to view experts as posing a risk to nonexperts in part because experts may place their own interests above those of nonexperts. Mechanisms may exist, however, to protect nonexperts from the errors and abuses of experts. |
American progressives such as Wilson (1887) tended to view experts as reliable and nonexperts as powerless. Nonexperts were expected to follow the dictates of the experts and little weight was given to the idea of resistance or of feedback from the nonexpert to the expert.
Socrates and the philosophers of the Academy belong in the “reliable-powerless” category. I discuss this group in Chapter 3, where I attempt to defend at some length the claim that Socrates, Plato, and Aristotle put themselves forward as experts who should be obeyed. In other words, they wanted the rule of experts. As we will also see in Chapter 3, expert witnesses in law have often adopted views that place them as well in the reliable-powerless box.
Eugenicists such as Francis Galton (1904), Karl Pearson (1911), and J. M. Keynes (1927) tended to fall into the category reliable-powerless. The “innate quality” of the population (Keynes 1926, p. 319) was to be preserved and enhanced. Eugenicists often sought to achieve this end in part by denying the benefits of child rearing to some individual members of that community, with eugenic experts deciding who would be denied.
Keynes’s letter to Margaret Sanger of 23 June 1936 indicates a “shifting” of his views. Keynes said, “In most countries we have now passed definitely out of the phase of increasing population into that of declining population, and I feel that the emphasis on policy should be considerably changed, – much more with the emphasis on eugenics and much less on restriction as such” (Keynes 1936). This “shifting” may seem ominous. Singerman (2016, p. 541) says, however, “Eugenics contained many different ideologies and policies, and Keynes was among many enthusiasts who never endorsed forcible sterilization or other state violence.” Presumably, Keynes favored a population policy that was in some way interventionist, neither laissez faire nor dictatorial. Singerman notes that Keynes “never advocated any specific policies in print” (2016, p. 541). In private conversation, he has wryly commented that Keynes seems to have taken the attitude that “eugenic policies will be extremely important … tomorrow.” Singerman’s researches seem to carry us only up to about 1930. The potentially ominous letter to Sanger is consequently beyond the scope of his inquiries. Overall, however, current evidence does not seem to support the claim that Keynes specifically favored or advocated forced sterilization. And yet the number and “innate quality” of the population was in some unarticulated way to be a matter of state policy.
A great variety of policy positions in economics may be described as “Keynesian.” And Keynes’ own views on economic policy are contested. It seems fair to interpret Keynes, however, as an advocate of the rule of experts in both economic policy and population policy. Singerman (2016) links Keynes’ policy preferences in these two areas, which might at first seem to be unrelated:
After the catastrophe of the Great War, Keynes continued to link ethics and eugenics as he sought to construct a moral society. A crippled continent faced unemployment, starvation, and revolution, so achieving a world of good things and good states of mind was possible only through technocratic management of population and economy. At the same time, this very mechanism for building an ethical society itself required attention to the nature of population as well as the number. The creation of both the caste of technocratic managers and the educated democratic citizenry who would follow them demanded active – if always ambiguous – measures to address the biological characteristics of those citizens. By the end of the 1920s, Keynes judged the immediate pressure of overpopulation to have receded. He could begin to foresee how material progress could positively combine with biological science, to produce individuals both taught and shaped to make good use of the absence of want.
For Keynes, economy, population, and morality are all subject in some degree to state planning. And state planning in these three areas is linked: The success of planning in any one area depends on successful planning in the other two. This vision of planning in three interlocking domains seems to limit the scope for human agency. Our choices must be right and rational. Eugenic experts will act to produce a future population that makes right and rational choices as judged by those experts today. The real authors of future moral choices are today’s eugenic, moral, and economic experts, especially Keynes. This mechanistic vision exalts the knowledge of one man and one mind, while mortifying the freedom and creativity of future generations.
It is, unfortunately, necessary to make a kind of disclaimer on Keynes and Keynesianism. Some readers may mistakenly believe (whether angrily or gleefully) that I think Keynes’ views on eugenics somehow taint or disprove Keynesian economics. I repudiate such callow logic. Although I reject the sort of economic policies usually associated with Keynes (Koppl 2014, pp. 129–39), I have profited from his economic theory and drawn on it in my own work (Koppl 2002). It is not my opinion that Keynesian economists of any sort secretly wish for eugenic policies or otherwise fail some moral test. Nor do I think that Keynes’ eugenic opinions somehow prove that policies of economic “stimulus” are bad or ineffective.
Mannheim (1936, 1952) is an important figure in the unreliable-powerless box. As Goldman explains (2009), Mannheim “extended Marx’s theory of ideology into a sociology of knowledge.” Marx lacked a clear understanding of the division of knowledge in society and tended to view the cultural and ideological “superstructure” as caused by the base without causally influencing the base. The material forces of production cause the bourgeoisie to think in one way and the proletariat in another. True freedom of thought outside the revolutionary vanguard will not exist until the workers’ paradise has arrived. Mannheim was subtler than Marx on these points, but in most cases the thinking of a social class or of a “generational unit” is “determined in its direction by social factors” (1952, p. 292).
Mannheim seems oblivious to the division of knowledge and its coevolution with the division of labor. For example, he describes the “mental climate” of “the Catholic Middle Ages” as “rigorously uniform,” notwithstanding some differences between clerics, knights, and monks (1952, p. 291). But the “mental climates” of cobblers, courtiers, and concubines were all very different from one another, as were those of sailors, saints, and sinners; pages and popes; blacksmiths, beggars, and bandits; goldsmiths; maidens; widows; weavers; tillers; jugglers; jewelers; minstrels; masons; hunters; fishers; and tillers. The “mental climate” of “the Catholic Middle Ages” varied within in each such social role, between places, and over time. It varied, importantly, with one’s place in the division of labor.
In another passage Mannheim says: “The real seat of the class ideology remains the class itself … even when the author of the ideology, as it may happen, belongs to a different class, or when the ideology expands and becomes influential beyond the limits of the class location.” By exempting science and his own ideas from the category of “ideology,” Mannheim placed his theory, in a sense, above the system and not in the system. (He did not “put the theorist in the model,” to use language developed in Chapter 4.) From his seat above the system Mannheim proclaimed boldly that “man’s thought … now perceives the possibility of determining itself” (Mannheim 1940, p. 213).
Foucault (1980, 1982) also belongs in the “unreliable-powerless” category. The individual is oppressed by a “discipline,” which is (1) a knowledge system (“régime du savoir”) such as medicine or psychology combined with (2) a communication system such as academic journals and the popular press and (3) some elements of more direct power, such as the ability to write prescriptions or to imprison someone declared “insane.” The victims of such “disciplines” may resist, and some generally do. But they resist only from a position of subjugation that they cannot easily escape. They are trapped in part because the knowledge system defines reality for the oppressed person, which makes it hard for the oppressed to recognize that they are indeed oppressed. Some high school students may be rebellious, while others see nothing to rebel against. In any event, the school’s “discipline” gives power to the teachers, the administrators, and, importantly, the overall system, but not to the students putatively served by the system.
Foucault (1982) has expressed a special interest in “the effects of power which are linked with knowledge, competence, and qualification: struggles against the privileges of knowledge.” For Foucault such struggles “are also an opposition against secrecy, deformation, and mystifying representations imposed on people” (1982, p. 781). Foucault’s linking of power and knowledge is suggestive and points to the possibility that experts may have pernicious power. His refusal to view power as extrinsic to the social fabric or exclusive to a small oppressing minority is also suggestive. Power is, somehow, in the daily fabric of social life and inextricable from it.
Jürgen Habermas probably also belongs in the “unreliable-powerless” category of Table 2.1. Turner says that for Habermas, “There is an unbridgeable cultural gap … between the world of illusions under which the ordinary member of the public operates and worlds of ‘expert cultures’” (2001, p. 128). (Turner cites Habermas [1985] 1987, p. 397.) Schiemann (2000) challenges Habermas’s distinction between communicative and strategic action.
Merton (1945) and Goldman (2001) are probably best placed in the “reliable-empowered” category. For Merton, experts are indispensable and mostly reliable, notwithstanding their potential to place special interests above the general interest. Because “intellectuals in the public bureaucracy” are in a position of “dependency” to the democratically elected policy maker, there is a relatively high risk of the “technician’s flight from social responsibility” through excess deference to the policy maker’s goals (p. 409). There is a relatively low risk of expert abuses such as imposing the expert’s values on others.
Adopting the perspective of veritisitic social epistemology, Goldman (2001) examines the problem of “how laypersons should evaluate the testimony of experts and decide which of two or more rival experts is most credible” (p. 85). He reviews some strategies for nonexperts to evaluate experts and places his discussion in the context of the philosophical literature on testimony. Unlike many other philosophical analyses of the issue, however, Goldman gives some attention to the social context of testimony. He notes, for example, that novices may be able to make reliable judgments of interests and biases (pp. 104–5). Goldman might have reached a more optimistic position if he had given even greater attention to social structure. I will note presently, for example, the importance of competition in the market for expert opinion.
Goldman neglects ways that details of context can matter. When two competing experts present alternative interpretations of a given issue, the truth may hinge on factors a novice is capable of judging. In DNA profiling, for example, the judgment about whether the genetic information in the crime-scene sample matches that of the suspect’s sample depends in part on the interpretation of an “electropherogram,” which is nothing more than a squiggly line. In some cases, novices will be able to see for themselves whether the two squiggly lines have peaks at the same locations. In such cases, modularity in expert knowledge allows novices to make reasonable judgments without acquiring all the specialized knowledge of the contending experts.
Goldman notes that the power of nonexperts is generally enhanced by the existence of competing experts whose exchanges the nonexpert can observe (2001, pp. 93–7). His overall conclusions are “decidedly mixed, a cause for neither elation nor gloom” (2001, p. 109). He thus views laics as neither utterly powerless nor fully empowered.
Finally, Berger and Luckmann (1966) are best placed in the unreliable-empowered category, as are Stephen Turner (2001) and Levy and Peart (2017). These thinkers recognize that experts may be unreliable and generally oppose giving experts an unchecked power to decide for others. And in one way or another they recognize the ability of nonexperts to challenge the authority of experts. Turner, for example, insists that “We, the non-experts, decide whether claims to cognitive authority … are to be honoured” (2001, p. 144).
I noted that the problem of experts originates in the “social distribution of knowledge” that Berger and Luckmann (1966) emphasize. In Chapter 7 I will explain how the idea of the division of knowledge in society migrated from the “Austrian economics” of Ludwig von Mises and F. A. Hayek to the phenomenological sociology of Alfred Schutz, from whence it entered the sociology of knowledge via Berger and Luckmann (1966).
There is some irony, perhaps, in the fact that “social constructionism” would owe so much to followers of the “Austrian” school of economics. The first of the Austrians, Carl Menger, defended “abstract theory” in the methodenstreit of the 1880s. Later, Mises would say, “In the course of social events there prevails a regularity of phenomena to which man must adjust his actions if he wishes to succeed.” The sovereign (whether literal monarch or a democratic electorate) knows that it cannot suspend the law of gravity by fiat. But the sovereign rarely understands that the laws of economics are just as binding and just as independent of human will. “One must study the laws of human action and social cooperation,” Mises averred, “as the physicist studies the laws of nature” (1966, p. 2).
There may be little or no incongruity between the “social constructionism” of Berger & Luckmann themselves and the “naturalism” of Mises. The term “social constructionism,” however, often refers to an anti-realist view of the world that puts any notion of truth under a cloud. One textbook treatment says, for example, “the notion of ‘truth’ becomes problematic. Within social constructionism there can be no such thing as an objective fact. All knowledge is derived from looking at the world from some perspective or other, and is in the service of some interests rather than others” (Burr 1995, p.4). Lest the reader mistake his comments for nothing more than healthy skepticism, he continues, “The search for truth … has been at the foundation of social science from the start. Social constructionism therefore heralds a radically different model of what it could mean to do social science” (p. 4). This antirealist position is traced back to Berger and Luckmann’s classic work. The author says “the major social constructionist contribution from sociology is usually taken to be Berger and Luckmann’s (1966) book The Social Construction of Reality” (p. 7). Unfortunately, Burr gives us no hint of the important fact that Berger and Luckmann explicitly invite the reader to “put quotation marks around” the terms “knowledge” and “reality” (p. 2).
The term “social construction of reality” has come to identify the view that all social reality – and in some versions all reality – is socially constructed and can, therefore, be made and unmade at will. In other words, there can be no scientific theory of society. Such a view, however, is in tension with the view that knowledge is dispersed widely across the population. To reformulate society at will, we would have to predict the consequences of our designed institutions. Those consequences depend on the interaction between dispersed bits of knowledge and future contingencies, which cannot be predicted without those very bits of knowledge, which, however, are unavailable to the planner if knowledge is dispersed.
Berger and Luckmann are probably not best classified as constructivists in the modern sense, although the term has an elastic meaning (Berger 2016). Paul Lewis (2010, p. 207) says: “Berger and his associates such as Thomas Luckmann see themselves as integrating the key insights of [Max] Weber and [Emil] Durkheim, namely that society is produced by subjectively meaningful human activity and yet appears also to possess objective, things-like status.” If this attractive interpretation is correct, it puts further distance between Berger and Luckmann and the sort of constructivism Burr attributes to them. They do say at one point, however, “the social world was made by men – and, therefore, can be remade by them” (p. 89). This remark seems to neglect the economic idea of spontaneous order, which is often identified with Adam Smith’s “invisible hand.” I refer to the idea that there are systematic but unintended consequences of human action. Standard economic examples include money, the macroeconomic determination of the price level, and the microeconomic determination of relative prices. Importantly for this volume, both the division of labor and its correlated, corresponding, coevolved division of knowledge are spontaneous orders. They are emergent unintended consequences of human action.
In my view, Berger and Luckmann do not give adequate weight to the economic idea of spontaneous order. This criticism is supported by at least one scholar who has made a direct comparison of Hayek (1967b) with Berger and Luckmann (1966). Against Berger and Luckmann, Martin (1968, p. 341) rails: “Surely society is in part ‘a non-human facticity’ and the attempt to eliminate this self-deluding and self-destructive.” On just this point, Martin (1968, p. 341) says, Hayek gives us a “critical codicil” with his “understanding of the non-human complexities and ‘arbitrary’ mechanisms derived from the intended and unintended consequences of each man’s choice, now as well as in the past.” Advances in ethology and psychology since 1966 have shown us just how much of human social life is derived from our prehuman ancestors and is, therefore, a product of biological evolution rather than free human invention (Cosmides, Tooby, and Barkow 1992).
Berger and Luckmann introduce experts with the problem of knowing which expert to trust. As we saw above, they say, “I require not only the advice of experts, but the prior advice of experts on experts. The social distribution of knowledge thus begins with the simple fact that I do not know everything known to my fellowmen, and vice versa, and culminates in exceedingly complex and esoteric systems of expertise” (p. 46).
How do you judge expertise without having the expertise you are judging? Berger and Luckmann identify the strategy we generally use in the “natural attitude.” I assume my everyday reality conforms to that of other members of my society. “I know that there is an ongoing correspondence between my meanings and their meanings” (p. 23). I apply this strategy to experts as well. I assume, for example, that forensic science is “science” and that the forensic scientist shares my view of science as objective and reliable. We rely on the label since we don’t have the expert knowledge itself.
Whether the default strategy of the natural attitude is effective in coping with expertise depends on the details of the social processes governing the sort of expertise in question. If there is a tight feedback loop between the experts’ claims or actions on the one hand and the public’s experiences on the other hand, then a kind of conformity is likely between the expectations of the public and the competencies of the experts.
Turner (2001) provides the useful example of the massage therapist. “The massage therapist is paid for knowledge, or for its exercise, but payment depends on the judgements of the beneficiaries of that knowledge to the effect that the therapy worked. The testimony of the beneficiaries allows a claim of expertise to be established for a wider audience” (p. 131). This type of expertise, Turner explains, “shades off into the category of persons who are paid for the successful performance of services” (p. 131).
Competition among experts may not be sufficient to keep the expectations of novices aligned with the competencies of experts if the novices cannot independently judge the results of expert advice or practice. This can be the case with experts in the area of used cars, as Akerlof (1970) pointed out. But even when autonomous learning by consumers is difficult, entrepreneurial market processes may create effective feedback loops anyway. Used car dealers in the United States, for example, now offer reliable warranties on their used cars, thus substantially reducing the risk of getting a lemon. Private rating agencies may be an effective surrogate for consumer judgment of outcomes. The Insurance Institute for Highway Safety, for example, probably does a good job of supplying naïve consumers with expert knowledge of automobile safety.
Berger and Luckmann focus most of their attention on universal experts. The fact that such experts claim special knowledge that applies to all things in the universe does not seem to be the driving factor in Berger and Luckmann’s analysis. The driving factor seems to be the relative independence of their knowledge from either empirical control or reasoned judgment by nonexperts. In the extreme, these two characteristics create a “[h]ermetic corpus of secret lore” (p. 87).
In such cases, the division of knowledge is not easily overcome. In the market for universal experts, competition may not improve veracity. Berger and Luckmann note the “practical difficulties that may arise in certain societies” when, for example, “there are competing coteries of experts, or when specialization has become so complicated that the layman gets confused” (p. 78). Pluralism in systems of expert knowledge may create “socially segregated subuniverses of meaning” in which “role-specific knowledge becomes altogether esoteric” (p. 85).
Universal experts are “the officially accredited definers of reality” in general (p. 97). But experts in modern society can acquire the power to define reality in a particular realm. “All legitimations, from the simplest pretheoretical legitimations of discrete institutionalized meanings to the cosmic establishments of symbolic universes may, in turn, be described as machines of universe maintenance” (p. 105). Thus, experts may be engaged in subuniverse maintenance.
When either the universe or a subuniverse is at stake, experts are eager to claim a monopoly and reluctant to allow competition. Who is to be considered an expert is very important. “The outsiders have to kept out” and the insiders “have to be kept in” (p. 87). Berger and Luckmann note: “there is the problem of keeping out the outsiders at the same time having them acknowledge the legitimacy of this procedure” (p. 87).
In addition to pumping up your own expertise, you disparage the expertise of others. With nihilation, “The threat to the social definitions of reality is neutralized by assigning an inferior ontological status, as thereby a not-to-be-taken-seriously cognitive status, to all definitions existing outside the symbolic universe” (p. 115). As a part of this strategy, the disparaged group may be taken to know who the “real” experts are. They do not own the truth, we do. And “Deep down within themselves, they know that this is so” (p. 116).
A body of experts may defend its monopoly with “intimidation, rational and irrational propaganda … mystification,” and “manipulation of prestige symbols” (p. 87). Physicians, Berger and Luckmann note, manipulate prestige symbols, mystify, and propagandize on the power and mystery of modern medicine (p. 88). The “general population is intimidated by images of the physical doom that follows” from rejecting a doctor’s advice. “To underline its authority the medical profession shrouds itself in … symbols of power and mystery, from outlandish costume to incomprehensible language.” Medical insiders are kept in – that is, kept from “quackery” – “not only by the powerful external controls available to the profession, but by a whole body of professional knowledge that offers them ‘scientific proof’ of the folly and even wickedness of deviance” (p. 88).
The theory of experts addresses the issues of epistemic authority and epistemic autonomy. To the extent that epistemic authority is problematic or epistemic autonomy desirable, there is a problem with experts. In this book I will try to show that there is a market for expert opinion whose structure determines the reliability of experts and the power of nonexperts. Most earlier treatments of experts and expertise have taken the reliability of experts and the power of nonexperts to be exogenous. There can be ambiguity, however, about what is endogenous and what is exogenous in a model. We might imagine, for example, that workers will achieve epistemic autonomy in the workers’ paradise and only in the workers’ paradise. It seems ambiguous whether the power of nonexperts is exogenous or endogenous in such a model. Nevertheless, it seems fair to say that expert reliability and nonexpert power are fully endogenous in only a relatively small fraction of discussions of experts and expertise.
Having sketched out a crude map of the literature on experts, I should probably indicate where I can be found on that map. I view experts as posing a risk to nonexperts in part because experts may place their own interests above those of nonexperts. I will identify mechanisms, however, that protect nonexperts from the errors and abuses of experts. Thus, I may fit best in the unreliable-empowered category of my simple taxonomy. In the theory I develop in this book, I have tried to make expert reliability and nonexpert power fully endogenous. Experts have great potential to be unreliable, and the wrong institutions may leave nonexperts more or less powerless. But there is a perfect moral, epistemic, and cognitive parity between experts and nonexperts. Thus, the overall institutional structure in society or the institutions dominant in a particular area of human thought and action may produce reliable experts, an empowered laity, or both.
Experts are usually defined by their expertise. An expert is anyone with expertise. This is the norm no matter what box one occupies in Table 2.1. But this way of defining “expert” may have some tendency to suggest that experts are reliable. It underlines the nonexperts’ relative ignorance, which may have some tendency to suggest that they are powerless in the sense of Table 2.1. I define an “expert” as anyone paid to give an opinion. This definition leaves it an open question whether experts are reliable or unreliable and might be less likely to suggest that nonexperts are powerless.
The literature on expert witnesses at law is an exception to the generalization that experts are usually defined by their expertise. Hand (1901) notes the distinction between experts and other witnesses. Unlike experts, other witnesses “shall testify only to facts and not to inferences” (p. 44). In other words, an expert witness expresses an opinion. Hand views nonexperts as unable to judge the relative merits of competing expert claims, and wishes to put things in the hands of an expert “tribunal.” He thus views experts as relatively “reliable” and nonexperts as “powerless.”
Woodward (1902) quotes (without citation) St. Clair McKelway’s definition. “An expert must be regarded as any specialist giving evidence in the form of opinion, no matter what his real or reputed standing in his specialty or in the community” (Woodward 1902, p. 488). McKelway’s definition has the advantage of precision and neutrality. It tells us what an expert witness is, not what an expert witness should be. Unfortunately, Woodward proposed an amended version that added a normative element to the definition of expert. “An expert is a specialist, the value of whose evidence, given in the form of opinion, is proportioned to his character, to his reputation for honesty in the community, and to his standing in his specialty or profession. It is not sufficient that he be thought wise, he must also be accounted honest” (p. 488).
Idealizing experts may encourage a moralistic approach to the problem of experts. Such moralizing cannot replace a positive theory of experts, in which we try to examine cause and effect objectively. In the theory of experts, as in economic theory, we need positive theory to ground our normative prescriptions.
The definition of “expert” should leave it an open question whether experts so defined are “reliable” or “unreliable” in the sense of Table 2.1. It should leave it open whether they are wise or foolish, saints or knaves, respected or derided, heeded or ignored. Much of the discussion of experts, however, includes moralizing over the expert’s knowledge and character. This moralizing may have been more salient in the nineteenth century, but it has never disappeared. The infinitely repeated phrase “junk science,” which seems to have been coined by Huber (1993), suggests that errors of scientific testimony are at least partly attributable to some form of ethical failure by the expert. Huber derides the “scientific cranks and iconoclasts who peddle their strange diagnostics and quack cures not at country fairs but in courtrooms across the land” (pp. ix–x). He thus views experts as unreliable and nonexperts as often powerless to distinguish “science” from “junk science.” Although moralizing is likely justified in all too many cases, the term “junk science” spuriously dismisses the notion that questions of applied science are often difficult and ambiguous. Its use often seems to reflect the view that if two experts give inconsistent opinions, at least one of them is personally corrupt. In any event, moralizing does not help us to produce a scientific theory of experts and expert failure.
Peart and Levy (2005) say that the “key feature” of the concept of an “expert” is “someone who makes recommendations about how others might achieve human happiness” (p. 4, n. 1). This characterization links experts to expert advice. Unfortunately, it considers only one sort of advice – how to achieve human happiness – and it does not define experts by their advisory role. Later in the same work, the authors “provide a more restrictive, formal definition of an expert” (p. 4, n. 1). They “define an expert as someone who uses all the evidence in a transparent manner to estimate a model” (2005, p. 241, emphasis in original). This definition seems to elevate the expert’s expertise to something like the best available opinion. Peart and Levy recognize that “[r]eal-world experts sometimes fall remarkably short” of their “idealized expert” who “has no interests other than truth and so estimates a model with all the data in the most efficient manner” (p. 243). Indeed, their research on experts began in discussions on eugenics (2005, p. vii; Levy and Peart 2017, p. xv), and they do not stint in criticism of eugenics experts. But taking this idealized vision as definitive of experts may make it harder to identify the conditions tending to render experts more reliable and those tending toward less reliability.
By the definition in Peart and Levy (2005), neither wizards nor forensic scientists are experts. The wizard mystifies and is thus not transparent. (See the discussion of alchemy in Mackay 1852.) The forensic scientist typically makes a subjective judgment of similarity (Nichols 2007; NAS 2009; Koppl 2010c), which, as such, cannot be transparent. There is the further point that transparency may be hard to define precisely. Levy and Peart (2017) do offer a formal definition of transparency. It applies, however, only in the narrow context of statistical estimation. And it seems all too consistent with their recognition that “full transparency … may be a pipedream” (2017, p. 183). They say, “The estimate is transparent if for an arbitrary reader t
In the economic literature on “credence goods,” mechanics, doctors, and others bundle diagnosis and treatment (Darby and Karni 1973; Wolinsky 1993, 1995; Emons 1997, 2001; Dulleck and Kerschbamer 2006; Sanford 2010). In this literature, “experts” are able to perform a diagnosis that the customer cannot. Here too, experts are defined by their expertise. Sanford (2010) says, “[E]xperts are able to, with effort, understand and interpret data relevant to their customers’ situations in a way customers themselves cannot. That is, true expertise is acquired, on a customer-by-customer basis, at a cost to its purveyors” (p. 199). Experts in this literature may provide incorrect diagnoses because they cheat or because they shirk. Sanford’s model distinguishes “true experts” from “quacks,” who “lack the ability to effectively carry out their duties” (2010, p. 199). The question then becomes the proportion of each type in the market.
Alfred Schutz (1946) defines experts by their expertise. He says, “The expert’s … opinions are based upon warranted assertions; his judgments are not mere guesswork or loose suppositions” (1946, p. 465). For Schutz, the expert’s knowledge is valid by definition. Schutz comes close to recognizing a problem with experts. But his definition of “expert” prevents him from recognizing adequately the full scope of the problem. Wierzchosławski (2016, p. 248) describes Schutz (1946) as “a classical phenomenological approach” to “expert studies.”
Schutz says that the outside world imposes “situations and events” on us that we did not choose. The extensive cooperation with strangers characteristic of the modern world has make these “imposed relevancies” ubiquitous and potentially oppressive. “We are, so to speak, potentially subject to everybody’s remote control” (p. 473). This “remote control,” together with the social distribution of knowledge, obliges us to rely on others for the knowledge that guides our choices. At this point, Schutz is very close to Berger and Luckmann’s (1966) statement of the problem of experts quoted above. In the end, however, Schutz seems to fear the man in the street and to place his hopes on experts and, as I discuss below, well-informed citizens. He thus places himself squarely in the top row of Table 2.1, viewing experts as fundamentally reliable. Schutz does not quite recognize a fully general problem of experts because he views expert knowledge as unproblematic.
Following their common teacher Alfred Schutz (1946), Berger and Luckmann (1966) define experts by their expertise. But they self-consciously set aside the question of the validity of the expert’s knowledge (p. 2). By explicitly setting aside any question of the reliability or validity of the expert’s knowledge, they were empowered to take a fairly skeptical view of experts. Their skeptical attitude, however, did not correspond to any theory indicating when experts are likely to be more reliable and when less. Bloor (1976) went further than Berger and Luckmann by postulating the methodological norm of “symmetry,” whereby the sociologist of science does not appeal to “truth” to help explain why one scientific theory prevails over another. Berger and Luckmann (1966) offer a phenomenological analysis of expertise, but not a causal theory of the ideas of scientists or other experts.
Bloor’s symmetry principle has lost ground in science and technology studies. Lynch and Cole (2005, p. 298) say that “Proposals for critical, normative, activist, reconstructivist and other modes of ‘engaged’ STS research are abundant in the literature.” They call such engagement the “normative turn” in science and technology studies (205, p. 270). Such “normativity” has become “central” to their discipline (Lynch and Cole 2005, p. 298). Turner (2010, p. 241) says, “Major figures such as Harry Collins and Bruno Latour have modified their views in order to avoid being seen as skeptics, particularly about global warming.” Forensic science provides the immediate context for the discussion in Lynch and Cole (2005).
Turner attributes the normative turn in STS at least in part to an extrinsic motive: I don’t want to look like a global warming skeptic. The desire to signal goodness on global warming and the desire to intervene in political outcomes are both extrinsic motives for the normative turn in science and technology studies. I do not know how important such extrinsic motives were in taking the normative turn. Presumably, intrinsic motives had at least some role to play as well. After all, some of the practices in, for example, forensic science might well seem shockingly “unscientific” to even the most ardent Bloorian. The relative importance of extrinsic and intrinsic motives in the normative turn of science and technology studies is a matter of guesswork and judgment, not scientific measurement. In any event, good scholarly practice usually requires us to consider the arguments and evidence of other scholars without imputing inappropriate motives to them.
Like Schutz (1946), Collins and Evans (2002) focus on “expertise,” which they view as “not truth-like” because expertise is “essentially imprecise” and “does not provide certainty” (Collins and Evans 2003, pp. 436, 448). They say, “Since our answer turns on expertise instead of truth, we will have to treat expertise in the same way as truth was once treated – as something more than the judgement of history, or the outcome of the play of competing attributions. We will have to treat expertise as ‘real’, and develop a ‘normative theory of expertise’” (2002, pp. 236–7). They regard the “area of prescription” they wish to enter as “upstream” from “downstream” descriptive analysis, wherein Bloor’s “symmetry” postulate is “central” (2002, p. 286, n. 28). The “prescriptions” they favor “depend on a preference for a form of life that gives special value to scientific reasoning” even though they explicitly “do not provide an argument for why [they] value it specially.”
If we define experts by expertise, we must either, like Berger and Luckmann (1966), set aside questions of expert reliability or, like Schutz (1946) or Collins and Evans (2002), mostly defer to the experts. Whichever choice we make, it becomes difficult to ask when experts might be more reliable and when less reliable. If, instead, we define “experts” by their place in the social order, we can ask about their economic incentives and discover, perhaps, when they are more likely to provide reliable advice and when they are less likely to do so. In other words, we need a theory of experts and not a theory of expertise. As I noted previously, I define “expert” as anyone paid to give an opinion. This economic definition skirts the necessity of somehow deciding ex ante whether the expert’s knowledge is reliable, unreliable, real, spurious, or something else. It allows us to ask which market structures tend to generate more reliable expert opinions and which market structures tend to generate less reliable expert opinions. And it allows us to ask which market structures tend to generate more empowered nonexperts and which market structures tend to generate less empowered nonexperts.