The problem of experts originates in the “social distribution of knowledge” that Berger and Luckmann (1966) emphasized. Thus, an analysis of expert failure should be grounded in an appreciation of the division of knowledge in society. We need some understanding of where expertise comes from. The social division of knowledge was not designed and imposed from above by a knowledge elite. It emerged spontaneously as the unintended consequence of many individual choices aiming at local ends and not any overall design for the system. The division of knowledge is a spontaneous order. It is bottom up and not top down. The theory of expert failure may go wrong if it does not begin with this bottom-up understanding of the social division of knowledge.
In earlier chapters we saw that some authors build on hierarchical notions of knowledge. As I will suggest in this chapter, the division of knowledge in society is quite rich and complex. If we employ a grossly simplified hierarchical model of knowledge, it becomes possible to imagine the growth of knowledge can be planned or that theorists can construct causal models to explain the “ideology” of other humans. Mannheim (1936) is an example. We saw earlier that Mannheim (1936) exempted science and his own ideas from the category of “ideology,” placing his theory above the system and not in it. Cole (2010) does not make the gross simplifications of Mannheim. He is, nevertheless, another example of a theorist adopting a relatively hierarchical view of knowledge. He explicitly calls for more “hierarchy” and wants to empower a “knowledge elite.” Cole’s hierarchical view of knowledge has implications for practice and for the theory of expert failure.
Cole criticizes an important document published by America’s National Academy of Sciences. The “NAS Report” (NAS 2009) reviews forensic science in the United States and makes recommendations for reform. Cole (2010) supports the Report’s central reform measure, which is federal oversight of forensic science through a regulatory body. But he criticizes the Report for adopting “obsolete models of science” found in the works of Karl Popper and Robert Merton that can only “impede the achievement” of the Report’s “purported goal: the ‘adoption,’ by forensic science, of a vaguely articulated ‘scientific culture’” (Cole 2010, p. 452). One “obvious problem” is Popper’s idea of scientific “boldness,” which is “tellingly missing from the NAS Report’s discussion of the ‘scientific method’” (p. 452).
Popper said scientists should make bold conjectures. As Cole deftly puts it, in the Popperian view, “only by thinking big, taking risks, and making ‘bold conjectures’ would scientists advance knowledge” (Cole 2010, p. 452). Cole considers it a great error to apply such Popperian sensibilities to daily work in forensic science. “Popper’s theory,” Cole insists, “applies to the sort of theory-generating scientist who works at the apex of the academic establishment.” It probably does not apply to forensic scientists, who are usually not research scientists. Forensic scientists are mostly “technoscientific workers from whom we would probably not desire boldness” (pp. 452–3).
Cole emphasizes the supposed difference between “discovery science” and “mundane science.” He wants Popperian boldness only in discovery science. But in Chapter 3 we saw a telling example of boldness in “mundane science.” There we saw Odling (1860) defend “the conflict of testimony” (p. 167) with an example. It is worth quoting Odling at greater length than I did earlier. Odling (1860, p. 168) says:
Three or four scientific men of eminence, retained by the patentee in an action for infringement, declared that a particular chemical reaction could not take place. They were supported by works of authority, and the reaction most certainly did not take place when the ordinary modes of experimenting were adopted. But it was most important for the defendant to show that this reaction was practicable, and his witnesses, after various attempts, succeeded in devising a method by which it was effected with facility. Accordingly they contradicted the witnesses for the plaintiff, and declared positively that the reaction could take place. And this illustration shows the importance of having a subject investigated by men desirous of establishing different conclusions. Had this case been referred to an independent commission, they would probably have decided it by the mere knowledge of the day, and this new reaction, with its important consequences, would have been altogether overlooked.
Applying the hierarchical model of knowledge to this case would probably have resulted in false victory for the plaintiff and an injustice to the defendant.
Cole (2010) adopts a hierarchical view of knowledge. In this chapter and the next I will develop a nonhierarchical view of knowledge that owes much to F. A. Hayek and Bernard Mandeville.
We have seen other examples of theorists adopting a simplified and relatively hierarchical view of knowledge. Among them, Alfred Schutz is unusually interesting. Schutz (1946) recognized the Hayekian division of knowledge and gave it great importance: “Knowledge is socially distributed and the mechanism of this distribution can be made the subject matter of a sociological discipline” (1946, p. 464). And yet Schutz gives the well-informed citizen the job of adjudicating expert opinions coming from different domains. Schutz says, “There is a stock of knowledge theoretically available to everyone, built up by practical experience, science, and technology as warranted insights” (1946, p. 463). Thus, even though Schutz fully acknowledges Hayek’s insight that knowledge is dispersed, he sees the social “stock of knowledge” as consisting in “warranted insights” rather than, mostly, practices that have emerged over time as aspects of the division of labor. These “warranted insights” are, apparently, conscious thoughts, and they are “theoretically available to everyone.” In spite of his emphasis on division of knowledge, his theory of experts builds on a theory of knowledge that is much more hierarchical and disembodied than Hayek allows.
The example of Schutz (1946) suggests that Hayek’s insights on dispersed knowledge are not as simple or trivial as summary statements may seem to suggest. We had better spend some time on the topic if we hope to avoid the sort of errors, as I imagine them, that I have attributed to Mannheim, Cole, and Schutz.
The term “division of knowledge” often refers to how we divide education or science into topics or disciplines. We typically distinguish, for example, science from the humanities. McKeon (2005) seems to have this sort of “division” in mind. Our knowledge may be divided into that concerning the public sphere and that concerning the private sphere, with different rules or principles belonging to each sphere. (Given the “entanglement” of the nominally public and private spheres explained in Wagner 2010, this particular division of knowledge may be doubted.) In this sense, the “division of knowledge” may be perfectly understood by an individual mind. Or the “division of knowledge” may refer to a perhaps philosophical distinction between types of knowledge. The difference between a priori and a posteriori knowledge, for example, may be described as a “division of knowledge.” These are, of course, perfectly legitimate uses of the term, but distinct from the meaning of interest here. In this volume the term refers to the fact that different people know different things.
F. A. Hayek (1937, 1945) is generally credited with the insight that knowledge is dispersed. This attribution seems to be about right. But as we will see, others anticipated Hayek in various degrees. In particular, Mandeville (1729) gives dispersed knowledge a treatment that rewards study today. Both George (1898) and Mises (1920) had previously given the idea a clear and explicit scientific statement and drawn from it the inference that comprehensive central planning of production cannot be as productive and efficient as its advocates had imagined.
It is difficult to trace the history of Hayek’s notion of dispersed knowledge. The problem is that the basic idea that different people know different things is trivial and obvious. The implications of this humble insight, however, are neither trivial nor obvious, especially when it is combined with a broader understanding of the nature of the knowledge that is thus divided among participants in the division of labor.
It is a truism to say that different people know different things. Keil et al. (2008) note, “As adults we all believe that knowledge is not distributed smoothly and homogeneously in the minds of others” (p. 259). We recognize that “bits of knowledge and understanding cluster together in ways that reflect different areas of expertise” (p. 259). Lutz and Keil (2003, p. 1081) find that “Children as young as 3 years of age already have a sense of the division of cognitive labor. They understand that adults are not omniscient and that they do have different areas of expertise.” Older children display more sophistication. The researchers found that “Children as young as 4 years were able not only to make attributions about stereotypical roles but also to make judgments about quite general and seemingly abstract domains such as biology and mechanics” (p. 1081). It is to be expected, then, that past writers will have often noted the division of knowledge without thinking themselves to have made some sort of discovery. For this very reason, perhaps, no one prior to Hayek succeeded in bringing this foundational fact forward as an explicit and central theme of social science. An exception should perhaps be made for Mandeville. But his treatment of dispersed knowledge seems to have left inadequate traces on subsequent thinkers. Even Adam Smith and David Hume, who were greatly influenced by Mandeville, adopted a less radical view of knowledge dispersion than we find in Mandeville. It was only with Hayek that the theme stuck in economics.
Hayek’s insight that knowledge is dispersed is important because we cannot somehow aggregate divided knowledge, thereby overcoming or eliminating dispersion of knowledge. Our understanding of dispersed knowledge is therefore linked to our understanding of the nature of the knowledge thus dispersed. There are large literatures on themes such as tacit knowledge and extended cognition that relate to the picture of knowledge I develop in this chapter. I will refer to this literature only incidentally, however, and direct my attention to the theme relevant to a theory of expert failure, the division of knowledge. As I develop in greater detail presently, the knowledge corresponding to the division of labor is “constitutive” and evolutionary. It may also be exosomatic, tacit, and synecological. When put in the right order these labels give us the acronym SELECT, which represents the idea that knowledge may be Synecological, EvoLutionary, Exosomatic, Constitutive, and Tacit.
Knowledge is synecological if the knowing unit is not an individual, but a collection of interacting individuals. Smith (2009) makes a similar point in contrasting “constructivist” and “ecological” rationality in economics. He links ecological rationality with “adaptive human decision and with group processes of discovery in natural social systems” (2009, p. 25). Hutchins (1991, 1995) famously explained how no one on a modern ship knows personally all the information that nevertheless enters into the decision-making process guiding the ship. The knowing entity is the ship and its crew, not any one person. There is an evolved “division of cognitive labor” (1991, p. 34) that was not fully designed by anyone. Each person in the crew interacts with others and with the ship to generate choices that may be more justly attributed to the system – the ship and its crew as a whole – than to any one person on the ship. Some readers may prefer to say that only individuals know and only individuals choose. But if we are not privy to the details of the ship’s division of cognitive labor, then we cannot specify which persons knew which things and which persons made which choices. We do not need a map of the ship’s division of cognitive labor to recognize that their interactions are generating potentially adaptive outcomes that depend on new information coming from both outside and inside the ship. In other words, we do not need a map of the ship’s division of cognitive labor to see that it is thinking, learning, and acting in much the way individual humans think, learn, and act.
As we saw in the previous chapter, Leonard Read (1958) taught us that no one person knows how to make a pencil. Probably no one person knows all that must happen in the pencil factory to produce a pencil. Here too there is a division of cognitive labor. Moreover, no one knows how to make all the inputs to pencil making and inputs to the inputs to pencil making, and so on. It is the overall division of labor that “knows” how to make a pencil. Pencil-making knowledge is distributed across all participants in the social division of labor; it exists in the system. Pencil-making knowledge is synecological.
I borrow the term “synecological” from ecology. The Oxford English Dictionary defines “synecology” as “The study of the relationships between the environment and a community of organisms occupying it. Also: the relationships themselves.” Etymologically, “syn” means “same.” Thus, etymologically, the word means “same ecology.” The interacting elements are in the same ecology. I use the term “synecological” to suggest that such knowledge is generated by the interactions of elements in an environment and is not separable from these elements, their interactions, or their environment.
There are a variety of views that represent knowledge as evolutionary. The idea of evolutionary knowledge can be found in Mandeville (1729). As we will see later in this section, Vasari (1568) also takes an evolutionary view of knowledge, though without offering a philosophically grounded theory of evolutionary knowledge. More recently, Radnitzsky and Bartley (1987) provide something of a canonical statement of the modern evolutionary epistemology associated with names such as Karl Popper (1959) and Imre Lakatos (1970). Donald Campbell (1987, pp. 73–9) reviews some history of the idea of evolutionary epistemology, but traces it back no further than Herbert Spencer.
Generally, knowledge is “evolutionary” if it changes over time in a process of variation, selection, and retention. Longo et al. (2012) have given us a theory of what we might call “creative evolution.” This theory has been imported to the social sciences and humanities by Felin et al. (2014), Koppl et al. (2015a, 2015b), and Devins et al. (2015, 2016). If Koppl et al. (2015a) are right, then the theory of creative evolution describes the coevolution of the division of knowledge and the division of labor.
Knowledge is exosomatic if it is embodied in objects existing outside the organism that uses such knowledge. As Ingold notes, Lotka (1945, p. 188) seems to be the first to use the term “exosomatic” for “the products of human knowledge” (Ingold 1986, p. 347). Lotka says: “In place of slow adaptation of an atomical structure and physiological function in successive generations by selective survival, increased adaptation has been achieved by the incomparably more rapid development of ‘artificial’ aids to our native receptor-effector apparatus, in a process that might be termed exosomatic evolution.” He links exosomatic evolution to the idea that, for humans, “[k]nowledge breeds knowledge” (1945, p. 192).
Karl Popper (1979) emphasized the exosomatic nature of much human knowledge, using the book as a principal example. A more suitable example for this discussion might be an egg timer. The knowledge of when to remove the egg from the boiling pot is embodied in the egg timer, which exists apart from the cook.
Knowledge is “constitutive” if it constitutes a part of the phenomenon. It is “speculative” if it explains the phenomenon. As we shall see presently, these are overlapping categories. The fisherman’s knowledge is constitutive of fishing, for example, no matter how much or little of it can be found in theories of fishing.
Finally, knowledge is tacit if it exists in our skills, habits, and practices rather than in an explicit form that could be written down. The knowledge of how to ride a bicycle is tacit. Ryle (1949) and Polanyi (1958) are standard sources on tacit knowledge, though many other figures, such as George (1898, pp. 39–41), have noted the phenomenon in one way or another.
The knowledge that coevolves with the division of labor is synecological, evolutionary, exosomatic, constitutive, and tacit. SELECT knowledge seems to be necessarily evolutionary. It is characteristically, but not necessarily, synecological, exosomatic, constitutive, and tacit. What Prendergast (2014) says of Mandeville applies to the characterization of knowledge given here. In social evolution there is an “accumulation of knowledge derived in the course of economic activity and embodied in practices, procedures, goods and technologies” (p. 105).
Hayek’s discovery of dispersed knowledge was not the truism that different persons know different things. He discovered that this humble insight is of central importance to social theory. Hayek seems to have credited himself with this discovery when he described his 1937 essay “Economics and Knowledge,” which identified the division of knowledge as the central issue in economics, as “the most original contribution I have made to the theory of economics” (Hayek 1994, p. 79). Earlier writers gave differing degrees of attention to the division of knowledge and they adopted different attitudes to it.
Mandeville’s treatment of knowledge so thoroughly anticipated Hayek that we may wonder if Hayek was not guilty of unconscious borrowing. His article on Mandeville (Hayek 1978) is generous toward Mandeville and credits Mandeville with the “twin ideas” of evolution and spontaneous order (p. 250). But Hayek does not seem to recognize Mandeville’s theory of dispersed knowledge. It seems unlikely that this snub was intentional. Even if we credit Mandeville fully, however, it was Hayek and not Mandeville, George, or Mises who caused the idea to stick. After Hayek, there is widespread recognition of a problem of dispersed knowledge; before Hayek, there was not. Often even the very persons who invoke it do not adequately understand the concept. But the idea is nevertheless widely recognized as real and important. It is because of Hayek that the idea has acquired a permanent place in the lexicon of social science. Thus, it seems fair to speak of “Hayekian” dispersed knowledge rather than “Mandevillean” dispersed knowledge.
The distinction between constitutive and speculative knowledge is important for a good understanding of Hayekian knowledge dispersion. This distinction modifies slightly Hayek’s distinction between “constitutive” and “speculative” ideas (1952b, pp. 36–7). Knowledge is “constitutive” if its possession becomes one of the causes of a social phenomenon. Such knowledge is constitutive of the phenomenon. Knowledge is “speculative” if it explains something, be it a social phenomenon, a natural phenomenon, or something else.
This distinction is often between two aspects of knowledge, but some constitutive knowledge has no evident speculative dimension. The specialized knowledges associated with each position in the social division of labor enable, and in this sense “cause,” that division of labor. It is true, of course, that an earlier and perhaps less refined division of labor gave rise to the constitutive knowledge that then enabled the subsequent, perhaps more refined division of labor. We have here yet another example of feedback in an evolutionary system. The logic here is close to that of Young (1928), who, however, took a more hierarchical view of knowledge.
Sailors had a constitutive knowledge of their craft long before scientists acquired a speculative knowledge of the mathematical principles of sailing (Mandeville 1729, p. 143). Wild chimpanzees have been seen to make tools (Goodall 1964). They have a constitutive knowledge of tool making. Presumably, however, they do not have any speculative knowledge of tool making since they do not possess human language. Whatever we might be able to teach chimpanzees to do, the gestural language of wild chimpanzees as currently understood (Hobaiter and Byrne 2014) would not seem to allow the explanatory function defining speculative knowledge. For humans, too, constitutive knowledge need not correspond in any way to speculative knowledge. But constitutive knowledge for humans might also be speculative. Classical mechanics is speculative knowledge because it explains the motions of bodies celestial and terrestrial. It is also constitutive knowledge because it guides us in the construction of bridges and buildings.
Speculative knowledge is theoretical and explicit. Speculative knowledge does not necessarily guide action. When it does, the knowledge precedes the action and is separate from it. Constitutive knowledge may be tacit and practical. By definition, constitutive knowledge guides action. It need not exist prior to or separately from the actions so guided. I did not know how to ride a bicycle prior to the action and my bicycle knowhow is not independent of my bicycle riding. If I found that I could no longer ride a bicycle I might lament that I had “forgotten” how. And I would mean by that statement only that I had lost the skill, not that my memory had failed me.
Hayek’s notion of dispersed knowledge refers mostly to constitutive knowledge. It refers to the knowledge that enables the division of labor. Expertise is generally constitutive knowledge derived from the expert’s place in the social division of labor.
The social division of knowledge is featured prominently in Plato’s Apology. In it, Socrates explains how he became a gadfly to Athens. Chaerephon had asked the oracle at Delphi whether there was anyone wiser than Socrates and, Socrates reports, “the Pythian prophetess answered that there was no man wiser.” Like Captain Renault in Casablanca, Socrates was shocked, shocked by this report. He said he was thus driven to question his fellow Athenians in an effort to prove the oracle wrong. His first stop was a politician. Their exchange left Socrates thinking “he knows nothing,” a plausible report given the man’s profession. Other politicians were no better. Socrates moved from politicians to poets, whom he found incapable of explaining their own works. Much like the politicians, they thought themselves wise when they were not. Finally, he arrived at the artisans, who, Socrates says, “knew many fine things” and were thus wiser than he. But, like the poets and politicians, “they thought that they also knew all sorts of high matters” of which they knew nothing. From this experience he draws his famous conclusion: “He, O men, is the wisest, who, like Socrates, knows that his wisdom is in truth worth nothing.”
Socrates’ conversations with artisans revealed a social division of knowledge. Each knew his separate art and was, in this regard, “wise.” Socrates blasts their “philosophical pretensions” and explicitly esteems practical, humble, workmanly knowledge over theoretical knowledge, at least among the humble artisans: “I found that the men most in repute were all but the most foolish; and that some inferior men were really wiser and better.” In his conversations with the poets, moreover, he discovered that some knowledge is tacit. He says: “not by wisdom do poets write poetry, but by a sort of genius and inspiration.” They know how to write poems, but can explain nothing of it to others.
In Plato’s Apology, the knowledge divided among the people is constitutive knowledge. Much of it exists in the tacit forms of habit and knowhow. The poets could not say how to write a good poem, but they could do it. Finally, the division of knowledge corresponds to the division of labor. Plato’s Socrates does not consider whether this division of knowledge and labor was planned or emergent.
Ancient writers do not seem to have written much that very clearly anticipates Hayek’s notion of spontaneous order according to which a seemingly planned social order may be the unintended consequence of human action. This point matters for us because in my interpretation of the division of knowledge it is itself a spontaneous order that emerges together with the division of labor.
Hayek (1978) says the ancient Greeks saw the problem of unplanned order, “of course,” but to discuss the problem they used the distinction between “natural (physei)” and “artificial or conventional (thesei or nomõ)” (pp. 253–4). This vocabulary, Hayek says, “produced endless confusion” because of its “ambiguity” (p. 253). If the ancient Greeks were as familiar with the problem as Hayek seems to suggest, it seems puzzling that they could not overcome their supposed problem of vocabulary. It seems hard to suppress the suspicion that the confusion Hayek refers to was in their understanding and not only in their exposition. In any event, it seems that ancient writers did not produce any very clear anticipations of the idea of spontaneous order. And if that is right, they would not likely have produced a clear statement of the division of knowledge in anything like the broadly Hayekian terms I articulated previously.
Hayek does provide a few suggestive quotes. In “all free countries,” he says, there was the “belief that a special providence watched over their affairs which turned their unsystematic efforts to their benefit” (p. 254). He quotes Aristophanes to illustrate this belief:
But the notion of divine intervention to overrule the chaos and misery of choices made higgledy-piggledy seems far from Adam Smith’s invisible hand.
Hayek’s quote from “the Attic orator Antiphon” seems closer to our themes. Antiphon says great age “is the surest token of good laws, as time and experience shows mankind what is imperfect.” This remark hints at both the limits of human knowledge and the tendency of accumulated experience to produce results superior to what design can produce.
Finally, Hayek (1978, p. 255) quotes a passage from Cato that expresses esteem for tradition as embodying the accumulated wisdom of many minds, each of which is weak and fallible. Roman law was exemplary because it
was based upon the genius, not of one man, but of many: it was founded, not in one generation, but in a long period of several centuries and many ages of men. For, said he, there never has lived a man possessed of so great a genius that nothing could escape him, nor could the combined powers of all men living at one time possibly make all the provisions for the future without the aid of actual experience and the test of time.
This remark represents, for Hayek, a natural law tradition that kept alive some notion of spontaneous order. In Hayek’s view, the tradition reached a zenith with the “Spanish Jesuits of the sixteenth century.” These “Spanish schoolmen of the sixteenth century … emphasized that what they called pretium mathematicum, the mathematical price, depended on so many particular circumstances that it could never be known to man but was known only to God” (1989, p. 5). They seem to have had, therefore, a relatively clear recognition of the Hayekian problem of dispersed knowledge. They reached, Hayek says, very “modern” results before being “submerged by the rationalist tide of the following century” (1978, p. 255).
Another sixteenth-century thinker, Georgio Vasari, also reached results compatible with the general view adopted here. His Lives of the Painters, Sculptors and Architects chronicles the gradual accumulation of knowledge in these arts. He represents progress in the arts as a recovery of ancient practice and knowledge. There is, therefore, an element of teleology in his story. The ancient perfection to which he appealed, however, had no precedent and was achieved by the same piecemeal discovery process that Vasari chronicled for the modern figures covered in his history:
Having pondered over these things intently in my own mind, I judge that it is the peculiar and particular nature of these arts to go on improving little by little from a humble beginning, and finally to arrive at the height of perfection; and of this I am persuaded by seeing that almost the same thing came to pass in other faculties, which is no small argument in favor of its truth, seeing that there is a certain degree of kinship between all the liberal arts. Now this must have happened to painting and sculpture in former times in such similar fashion, that, if the names were changed round, their histories would be exactly the same.
Progress in all the liberal arts proceeds by the same piecemeal process of accumulation that Vasari chronicled for Tuscan art from Cimabue to Michelangelo. It was, says Vasari, the same process in the ancient world and in the modern world. The knowledge acquired in this process is largely tacit. In both poetry and “the arts of design,” works made in the “fire” of inspiration are better than those made “with effort and fatigue” (p. 274). And the process is not necessarily cumulative. Knowledge can both arise and disappear. Luca della Robbia invented glazed terra cotta sculpture, which was a “new form of sculpture” that “the ancient Romans did not have” (p. 280). His techniques were family secrets and when his family became “extinct,” knowledge of the technique was largely lost and “art was deprived of the true method of making glazed work” (p. 280).
Vasari relates at least one story that reveals a clear appreciation for the distribution of knowledge. Lorenzo Ghiberti won a competition to make a pair of bronze doors for the Florence Baptistery. These are now the north doors of Baptistery and depict scenes from the Old Testament. Ghiberti’s rivals for the commission kept their work “hidden and most secret, lest they should copy each other’s ideas.” Ghiberti, by contrast, “was ever inviting the citizens, and sometimes any passing stranger who had some knowledge of the art, to see his work, in order to hear what they thought and these opinions enabled him to execute a model very well wrought and without one defect” (p. 292). By crowdsourcing criticism of his work, Ghiberti was able to improve it enough to prevail in the competition.
Vasari chronicles the emergence of a tradition from its earliest beginnings with Cimabue to its apotheosis in Michelangelo. (It does not matter for our purposes here that Vasari’s choice of Cimabue as the one who gave “the first light to the art of painting” seems more political than factual.) In the process painters learned how to represent the human body realistically, how to foreshorten figures, how to represent a figure “shivering with cold” (p. 323), perspective, how to give figures “grandeur and majesty” (p. 723), and so on. Competition and emulation among members of the community gave rise to a set of practices that, in turn, produced a series of objects and innovations. And yet this history seems to have had little or no influence on subsequent social thought. Mandeville makes no reference to it in the Fable. Adam Smith’s library included a copy of Vasari’s Vite (Bonar 1894, p. 116). No reference is made to it, however, in the Wealth of Nations, The Theory of Moral Sentiments, or Lectures on Justice, Police, Revenue and Arms. Whether because it was a victim of the “rationalist tide,” or because it was considered a thing apart from social theory, or for another reason, Vasari’s history seems to have left no discernible trace on social theory, notwithstanding the detailed and vivid chronicle he gave us of the evolution of a spontaneous order.
Hayek says that Mandeville’s “speculations … mark the definite breakthrough in modern thought of the twin ideas of evolution and of the spontaneous formation of an order” (1978, p. 250). Mandeville recognized that knowledge is divided and gave importance to the fact. He recognized the tacit dimension of knowledge. He distinguished constitutive from speculative knowledge, elevating the former relative to the latter. He saw constitutive knowledge as emergent from the division of labor. And he viewed the division of labor and both constitutive and speculative knowledge as products of a slow evolution driven by individual action, but not any human design. Hayek’s assessment of Mandeville’s role in creating the “twin ideas” of evolution and spontaneous order supports the impression that Mandeville may have been the first (or at least the first modern writer) to achieve anything like the Hayekian conception of human knowledge. As we shall see, later writers seem to have fallen short of Mandeville’s radical vision until Hayek resuscitated it beginning with his classic essay of 1937, “Economics and Knowledge.” Even since Hayek’s much cited work on this theme, most scholars take a less radical view of knowledge – one that overestimates the power of rational thought and speculative knowledge, at least if Mandeville and Hayek were right. Prendergast (2014, p. 87) nicely summarizes Mandeville’s views on these topics: “[F]or Mandeville, innovators were people of ordinary capacity who were alert to the opportunities and challenges of their environment. As a result of specialisation, they possessed tacit knowledge which was actualised in what they did rather than in theoretical propositions.”
Bernard Mandeville noted the division of knowledge in society. Because “Men differ” in “Inclination, Knowledge, and Circumstances,” Mandeville explained, “they are differently influenced and wrought upon by all the Passions” (1729, vol. II, p. 90). He recognized this divided knowledge as an emergent phenomenon having more to do with experience than “reasoning a Priori” (1729, vol. II, p. 145). We “often ascribe to the Excellency of Man’s Genius, and the Depth of his Penetration, what is in Reality owing to length of Time, and the Experience of many Generations, all of them very little differing from one another in natural Parts and Sagacity” (1729, vol. II, p. 142). The art of sailing, for example, has been explained mathematically. But it is practiced “without the least scrap of Mathematics” (1729, vol. ii, p. 143). Ignorant sailors impressed into service soon learn to sail “much better than the greatest Mathematician could have done in all his Life-time, if he had never been at Sea” (1729, vol. ii, p. 143). Through practice, the knowledge becomes “habitual” (1729, vol. ii, p. 140).
The Arts of Brewing, and making Bread, have by slow degrees been brought to the Perfection they now are in, but to have invented them at once, and à priori, would have required more Knowledge and a deeper Insight into the Nature of Fermentation, than the greatest Philosopher has hitherto been endowed with; yet the Fruits of both are now enjoy’d by the meanest of our Species, and a starving Wretch knows not how to make a more humble, or a more modest Petition, than by asking for a Bit of Bread, or a Draught of Small Beer.
“It is,” Mandeville says,
not only that the raw Beginners, who made the first Essays in either Art, good Manners as well as Sailing, were ignorant of the true Cause, the real Foundation those Arts are built upon in Nature; but likewise that, even now both Arts are brought to great Perfection, the greatest Part of those that are most expert, and daily making Improvements in them, know as little of the Rationale of them, as their Predecessors did at first.
This passage expresses very strongly the priority of practical, experiential knowledge over theoretical knowledge.
Mandeville values constitutive knowledge above speculative knowledge and seems to think the overlap between them is small. Though he denied it (I 292), Mandeville sometimes seems to disparage education altogether, as when he opines that “the Knowledge of the Working Poor should be confin’d within the Verge of their Occupations” (1729, vol. I, p. 288). Prendergast (2010) explains that Mandeville placed great importance on education, but attacked “the view that the poor were poor because of their lack of education” (p. 415, n.1). Moreover, the education he favored was often conducted outside of schools and when “The Knowledge of Parents is communicated to their Off-spring, and every one’s Experiences in Life, being added to what he learned in his Youth, every Generation after this must be better taught than the preceding; by which Means, in two or three Centuries, good Manners must be brought to great Perfection” (1729, vol. ii, pp. 145–6). Even those receiving a university education should concentrate on matters trade-related. “No Man ever bound his Son ’Prentice to a Goldsmith to make him a Linen-draper; then why should he have a Divine for his Tutor to become a Lawyer or a Physician?” (1729, vol. I, p. 293).
Mandeville’s attitude to the poor is a contested issue. On the one hand, he advises legislators that “the surest Wealth” of a nation “consists in a Multitude of laborious Poor” (1729, vol. I, p. 287). He says: “To make the Society happy and People easy under the meanest Circumstances, it is requisite that great Numbers of them should be Ignorant and Poor” (pp. 287–8). It may be that Mandeville had as harsh and exploitative a view of the poor as Kaye believes. Kaye holds to this view even while noting that “here, as elsewhere,” Mandeville was able “to make a current creed obnoxious by the mere act of stating it with complete candour” (see p. lxxi of Kaye’s introduction to Mandeville 1729). This skill, however, may suggest a more Swiftian interpretation of Mandeville’s attack on charity schools. He rather clearly says that the poor can abide their bad condition only because they are ignorant and have never known anything else. He seems, then, to attribute any differences between prince and pauper to their different positions in the division of labor, rather than any innate differences. “Human Nature is every where the same,” he says; “Genius, Wit and Natural Parts are always sharpened by Application, and may be as much improv’d in the Practice of the meanest Villany, as they can in the Exercise of Industry or the most Heroic Virtue” (1729, vol. I, p. 275). He says wryly that “A Servant can have no unfeighn’d Respect for his Master, as soon as he has Sense enough to find out that he serves a Fool” (p. 289), and “No Creatures submit contentedly to their Equals, and should a Horse know as much as a Man, I should not desire to be his Rider” (p. 290).
Mandeville’s attack on the charity schools was an attack on the clergy, not the poor. Mandeville adhered to Peart and Levy’s (2005, p. 3) analytical egalitarianism. Before coming to the charity schools, Mandeville developed at length the theme that “the Clergy are not possess’d of more intrinsick Virtue than any other Profession” (1729, vol. I, p. 173). The clergy are just like you and me, which makes them despicable, immoral, and corrupt. With “brutish Appetite” they “indulge their Lust.” We have “reason to believe,” Mandeville tells us, that what the clergy say “is full of Hypocrisy and Falshood, and that Concupiscence is not the only Appetite they want to gratify; that the haughty Airs and quick Sense of Injuries, the curious Elegance in Dress, and Niceness of Palate, to be observ’d in most of them that are able to shew them, are the Results of Pride and Luxury in them” (1729, vol. I, p. 173).
Mandeville’s Letter to Dion (1732) may give support to the view that Fable had firmer moral foundations than Kaye seems to allow. Mandeville says that “The Fable of the Bees was a Book of exalted Morality” (p. 24) aimed at Christian hypocrisy. Thus attacked, his “Adversaries were obliged to dissemble the Cause of their Anger” (p. 25) by imputing immoralist ideas to Mandeville.
Edwards (1964) and Harth (1969) both view the Fable as satirical. Harth (1969) castigates Kaye’s attempt to “impose an artificial unity” on the Fable, which only served to “flatten his satire into an insipid exercise in literary paradox” (pp. 325–6). He notes the passage in Fable in which Mandeville has Cleomenes saying, “There is, generally speaking, less Truth in Pangyricks than there is Satyrs” (Harth 1969, p. 322, and Mandeville, 1729, vol. II, p. 59). Edwards compares Mandeville to Swift (1964, pp. 198, 203, and 204) and emphasizes the “complexity of Mandeville’s tone,” which includes, he avers, heavy doses of irony (p. 204). In Edwards’s plausible interpretation, Mandeville did have a low opinion of the charity schools, but not because he had a low opinion of the poor. Mandeville asks:
[W]hy must our concern for Religion be eternally made a Cloke to hide our real Drifts and worldly Intentions? Would both Parties agree to pull off the Masque, we should soon discover that whatever they pretend to, they aim at nothing so much in Charity-Schools, as to strengthen their Party, and that the great Sticklers for the Church, by Educating Children in the Principles of Religion, mean inspiring them with a Superlative Veneration for the Clergy of the Church of England, and a strong Aversion and immortal Animosity against all that dissent from it.
The use I am making of Mandeville’s work, however, does not require me to sort out the important interpretive question of his attitude to the poor. My attention here is focused solely on his importance for our understanding of the division of knowledge in society.
As his use of the word “habitual” seems to suggest, Mandeville recognized the existence of tacit knowledge. The word “knowing,” he explains, has a “double Meaning” (1729, vol. ii, p. 171): “There is a great Difference between knowing a Violin when you see it, and knowing how to play upon it” (1729, vol. ii, p. 171). This is “the Difference between Knowledge, as it signifies the Treasure of Images receiv’d, and Knowledge, or rather Skill, to find out those Images when we want them, and work them readily to our Purpose” (1729, vol. ii, p. 171). Here, interestingly, part of tacit knowledge is the ability to skillfully call up and deploy explicit knowledge.
Prendergast (2014, p. 105) says: “Mandeville appears to have been the first to develop a theory of social evolution based on the accumulation of knowledge derived in the course of economic activity and embodied in practices, procedures, goods and technologies.”
Mandeville also recognized what we might call the “division of opinion” in society. Speaking of the “Judges of Painting,” he says, “There are Parties among Connoisseurs, and few of them agree in their Esteem as to Ages and Countries, and the best Pictures bear not always the best Prices” (1729, vol. I, p. 326). Such judges are, of course, experts, and it is interesting to note Mandeville’s skeptical view of them.
We saw Mandeville explain how the imparting of knowledge from one generation to the next improves “Manners.” The “Precepts of good Manners” for Mandeville “are no more than the various Methods of making ourselves acceptable to others, with as little Prejudice to ourselves as possible” (1729, vol. ii, p. 147). In keeping with the overall spirit of his “licentious system” (Smith 1759, VII.II.104), Mandeville gives a dark cast to this seemingly happy idea. “Manners and Good-breeding,” he says wryly, “consists in a Fashionable Habit, acquir’d by Precept and Example, of flattering the Pride and Selfishness of others, and concealing our own with Judgment and Dexterity” (1729, vol. I, p. 69).
Good manners are the art of getting along. Mandeville does not think manners so conceived are easily acquired. It takes “two or three Centuries” of accumulated experience to bring manners to “great Perfection” (1729, vol. ii, p. 146). It is a slow evolution of prudent, agreeable, and sociable behavior shaped by commerce. Mandeville thus articulates the doux commerce thesis discussed by Hirschman (1977, pp. 56–63). This term, Hirschman explains, “denoted politeness, polished manners, and socially useful behavior in general” (p. 62). Recently, Henrich et al. (2005) and Pinker (2011) have given empirical support to this view. There is an epistemic dimension to the doux commerce thesis, at least in the form Mandeville gives it. The knowledge of how to behave in a prosocial manner emerges slowly from the accumulated experience of generations. It exists, I would add, mostly in the tacit form of accumulated habit.