Artificial Intelligence and the Modern Condition of Expertise

Abstract

The following essay offers a political theory of the expert. I draw connections between the authority derived from expertise to the emerging role of artificial intelligence in a political condition marked by an impersonal sovereign. I begin with a detailed description of the role of the expert and expertise with specific attention to the function of knowledge as a source of authority. Through the work of Max Weber and Carl Schmitt, I connect the political role of the expert to the rise and prominence of the administrative or total state. The essay focuses on the connection between the expert as an authority in the modern state and the emergence of a political logic that invites artificial intelligence into the realm of authority.

[. . .] it may be worth our while a little to reflect on FOUR SORTS OF ARGUMENTS, that men, in their reasonings with others, do ordinarily make use of to prevail on their assent; or at least so to awe them as to silence their opposition. The first is, to allege the opinions of men, whose parts, learning, eminency, power, or some other cause has gained a name, and settled their reputation in the common esteem with some kind of authority. When men are established in any kind of dignity, it is thought a breach of modesty for others to derogate any way from it, and question the authority of men who are in possession of it. This is apt to be censured, as carrying with it too much pride, when a man does not readily yield to the determination of approved authors, which is wont to be received with respect and submission by others: and it is looked upon as insolence, for a man to set up and adhere to his own opinion against the current stream of antiquity; or to put it in the balance against that of some learned doctor, or otherwise approved writer. Whoever backs his tenets with such authorities, thinks he ought thereby to carry the cause, and is ready to style it impudence in any one who shall stand out against them. This I think may be called ARGUMENTUM AD VERECUNDIAM.

John Locke, An Essay Concerning Human Understanding,book IV, chat XVII, section 19

Ted Chiang’s (2000) very short story “Catching crumbs from the table” is written as a report on the conditions of science after a point when scientific inquiry will have “moved beyond the comprehensibility of humans.” In the story, written as a short scientific report, a genetic therapy has been developed that is administered during fetal brain development, allowing some biological humans to receive direct neural transfer (DNT) of scientific learning. Having this ability effectively alienates these “metahumans” from other humans as they acquire the knowledge associated with it. Those without this genetic ability to perform DNT are left outside of the frontiers of science; those with the genetic modification, the metahumans, perform the work that sustains scientific discovery. What human researchers are left to do is interpret the scientific work of metahumans, to perform the work of hermeneutics as scientific inquiry. This activity gives rise to artifact hermeneutics: human researchers seek to understand how something works by reverse engineering the artifact itself, but this yields limited insight and primarily provides human researchers with just enough information to grasp what has already been done. Chiang’s report from the future ends in a sanguine tone, noting that the technologies that gave rise to metahumans were invented by humans, so it is conceivable that further inventions by humans can enable humans to “upgrade” their minds so that they match the intellectual power of metahumans. Even in the future the hope for a solution to the problems of technology lie in the technological power of humans.

The allegorical message of Chiang’s fictional report has affinities with the present conditions of expertise and artificial intelligence. Obviously, there is not a genetic treatment (called Sugimoto gene therapy in the story) that prepares a human fetus to develop into a metahuman with an ability to receive information and data directly into the brain. But I think that the conditions of modern expertise in the administrative state can be understood as a condition in which complex knowledge generation is difficult to access and creates a bifurcation in which a small group has “access” to information and data that are grasped primarily through interpretation. Over the next six sections, I will offer a political theory of the expert that explores the conditions of authority based in the administrative state, conditions that prefigure the impersonal reasoning of algorithm and artificial intelligence. The role played by the expert is based on a fabricated experience and an interpretation of it. Experts are not simply “the ones who know,” but rather the ones whose knowing is deployed for action. They are the interpreters of a knowledge system that is increasingly removed from conventional understanding. Experts are practitioners of a hermeneutics of fabricated experience, and their importance grows the more deeply systems become embedded in complex generators of knowledge and information such as artificial intelligence. A key facet of the expert is as an agent in an impersonal system of authority. Expertise presages artificial intelligence through this impersonality. Recent developments in AI, such as ChatGPT, are designed and built specifically to be impersonal, a feature that limits how interesting it is but protects its authority by avoiding “opinion.” ¹

The essay begins with two primary political concepts, authority and sovereignty, approaching them through early modern observers of the administrative state, specifically Weber and Schmitt. I then introduce how the expert fits into this authority structure. Sections three through six elaborate on the qualities I use to describe and understand modern expertise. In section three, I discuss how matters of fact supplant matters of truth in the political role of the expert and follow this section with a commentary on how the specific knowledge of the expert, metis, is associated with the modern notion of policy in sections four and five. Finally, I conclude with a consideration of how artificial intelligence culminates many of the themes developed through the systems generated by the expert and the administrative state, such that expertise might simply be a bridge concept to a world much like what Chiang described but with the flair of the political, a situation where scientific discovery and knowledge are outside our grasp, leaving us to collect the crumbs of understanding.

1

Democracy as a political institution confounds authority. In a state where the social is the source of authority, expressed in the abstraction of the public, authority itself seems to dissolve. The problem of sovereignty concerns who decides upon the actions that will shape social conditions. Law is insufficient in this respect, since in a proactive sense it is at best a prescription that cannot answer the question of its own application. What authority does law truly have when alone it functions as a norm? To answer this question of authority by appeal to the rule of law is to respond in abstraction, to appeal to a norm that seeks out a leveling in the treatment of all members of society but fails to address the inherent inequality of authority. If we can agree upon one thing about power under the rule of law, so critical to modern expressions of democracy, it is that it is a power unyielding and without exception, except when it is not.

When it is not, the question of authority arises. Schmitt framed this problem in terms of the sovereign as the one who decides on the exception, through which the application of the law is realized. It is clear from his framing of the issue that the question of the exception is ultimately unavoidable, and there will inevitably be a person whose role it must be to decide when the law does not apply. In theory, this is the only way to have functional law: Law is defined by its boundaries, and under Schmitt’s theory of sovereignty, law is determined by the decision of the exception, a decision he views as ultimately falling to a personal authority.

But another ingredient in this complication of authority is the governance produced by and through modern mass democracy. Among the insights about modernity in Weber’s (2013) Economy and Society is his observation that modern democracy comes with, arguably produces the need for, bureaucracy:

Bureaucracy inevitably accompanies modern mass democracy, in contrast to democratic self-government of small homogeneous units. This results from its characteristic principle: the abstract regularity of the exercise of authority, which is a result of the demand for “equality before the law” in the personal and functional sense–hence the horror of ‘privilege, and the principled rejection of doing business “from case to case.” (XI,8,A, emphasis original)

Modern mass democracy transforms the state according to the interest of the demos, but the demos does not govern but rather is governed. Weber distinguished between democratization understood as a process by which the demos acquires power and gains a greater share in government (deemed impractical) and democratization as the process of universal accessibility of office and the expansion of “the sphere of influence of ‘public opinion’ as far as practicable” (XI.8.B). The result of this inevitable coexistence between modern democracy and bureaucracy is a tension between the necessary production of impersonal authority in the interest of a leveling of the governed, so that all are treated the same under the law without privilege, and the potential for autocratic rule in and by this impersonal system. Mass democracy produces bureaucracy, which holds the potential to contravene democratic values, but the means for democratic society to address this potential is also bureaucratic. We might call this the iatrogenic problem of mass democracy: the treatment of the pathologies of bureaucracy produces more bureaucracy.

In terms of authority, bureaucracy is an organization type based upon rationality and law down to the point that authority itself, being fully impersonal, derives from the constitutive features of bureaucracy. To make sense of this, Weber had to define a difference between power and authority. Power is a discrete agent’s ability to control or direct; authority he defined as the likelihood that commands would be obeyed. When those subject to the commands consider the commands just and appropriate, then you have legitimate authority and power.

These definitions of authority and power suffice to Weber’s objective and give some theoretical clarity to the idea of bureaucracy. Democracy, though, confounds authority specifically because it cultivates a culture of questioning authority, of questioning the power of anyone to command, of questioning the ability to control, of questioning to the extent that legitimacy is an open question. Authority in a modern democracy is placed in a bureaucracy because bureaucracy has no specific personality (following Weber, we could say it has no potential for power), and its authority is legitimate only to the extent that it is not placed in any person. This also happens to be the most frustrating and dangerous feature of bureaucracy: that it is an arrangement of impersonal operations.

What Weber described as the modern state brought forth by democracy Schmitt (2004) would later call the economic state or the administrative state. He distinguished this state type from the legislative state and the governmental state, both of which retain a semblance of the personal. Importantly, Schmitt attached the potential for total power in primarily in the administrative state. Where Weber was somewhat elusive in his connection of democracy to the formation of modern bureaucracy, Schmitt was more direct:

Of much greater significance is the realization that the cause of the contemporary “total state,” more precisely the total politicization of all human existence, is to be sought in democracy, [. . .] the total state needs a stable authority in order to move ahead with the necessary depoliticizations and to reestablish free spheres and living spaces from within itself. (p. 90)

Schmitt the legalist maintained that “the state is law in statutory form,” but that “When the dualism of state and society was set aside [. . .] the state’s will and the people’s will became identical, as one would expect under the democratic logic, designating every expression of the people’s will as ‘statute’ and giving it the entire dignity and majesty due this concept by virtue of its connection with law and justice” (pp. 18 and 23). The rise of democratic logic in governing, especially the use of plebiscites in matters of decision, conflated society and state and degraded the authority of the legislature, the “one lawmaker” as Schmitt described it. This arrangement presented a different state from the legislative or the governmental, a third type, the administrative state, in which “command and will do not appear authoritarian and personal, [in which] men do not rule, nor are norms valid as something higher. Instead, the famous formula ‘things administer themselves’ holds true” (p. 5).

Schmitt’s use of the phrase “things administer themselves” seems intended to give the reader pause; it hints at the absence of human agency at the heart of the state, of the res publica (public thing) itself. The formula underscores the impersonality of the state, such that the administrative state seems to eschew power, command, and authority because of the apparent absence of personality from the deployment of such governing features. But it is not a clean formula. The administrative state achieves order by satisfying the demos, similar to how a household order is the satisfaction of the members of it (hence economics deriving from nomos of the oikos, customs of the household). Schmitt observed the continuing need for agency in a state that is intentionally impersonal, an agency that is a thing to administer the interconnected things that administer the state. In an abstract sense, bureaucracy is a rule by nobody, an observation made later by Arendt (2009) in her eulogy for politics (p. 97). This type of rule comes about alongside the total politicization of human existence, as Schmitt observed. But the abstraction is insufficient to the practical. There comes a point when the situation demands practical agency. Schmitt found in this realism the rebirth of the sovereign, an ever-present necessity for state order. But modern conditions continuously degrade this sense of realism and create a new one, a realism around rule, a rule of the rule-set, of the algorithm itself, with the potential to replace the very spark of human distinction with what humans themselves create, namely, an artifice of intelligence.

The consequence of the degradation of Schmittian realism in favor of a realism of the artificial is that intelligence itself rises to the status of rulership. This is a distinctly modern condition of legitimacy. Rule by intelligence differs from rule by divine edict or tradition or rule by charisma. This type of rule demands acknowledgement of the changed face of knowledge itself, of the very instrumentalization of the process of knowledge, the idea of the novum organum. ² Intelligence is conceived as coming from experiment, and so it is in the one who experiments that we place authority. In other words, the modern disposition towards impersonal rule is incomplete, leaving a remnant of agency to hold the elements of authority necessary to state order. This remnant is endowed to the one who experiments to know, the expert.

2

What is expert authority? If we subscribe to Weber’s distinctions, then we might say that it is domination made legitimate by an appeal to knowledge premised on experience. Because experiences are limited by opportunity, time, context, and a variety of other factors often outside of individual control, expert authority must be translatable to others who lack the experience. The authority claimed by the expert is based upon an appeal to an experience and the willingness of others to accept that experience to such a degree that they are willing to adopt reasonable inferences from an experience that they did not have. This, I propose, is the enigma of expert authority. Let’s consider its constituents, which together will give an account of the role of the modern expert.

• 1. Expert knowledge is metis. The distinction is not always maintained in plain language, but the expert is one who applies knowledge, an instrumentation of knowledge towards manipulation.

• 2. An appeal to an expert in an argument is still an appeal to authority, but it is a different appeal than to an authority premised on status, tradition, or charisma. The argumentative tactic of the appeal to authority, what Locke called the argumentum ad verecundiam in the quotation above, is fallacious to the extent that the authority subject in the appeal is not legitimate. But it is just as likely to be considered a valid argument, as it is in Aristotle’s Rhetoric. Locke himself described this form of appeal in mostly a favorable sense as a valid argument approach. The success of the argument depends upon the modesty of those to whom the appeal is made. Verecundia can be translated as modesty, shyness, or bashfulness, and as Locke described in the passage, to question the appeal is impudence. Interestingly, Locke’s framing of the argument is not to appeal to the authority figure but rather to the modesty of the one hearing the argument.

An appeal to an expert corresponds to an appeal to the modesty of the listener. The expert is the one who knows as a result of being the one who has “experienced,” and it would be impudent for one who lacks experience to deny the expert. But what are we to make of the actual claim of the expert? What is the reasoning process behind the expert claim?

• 3. The expert does not deign to argument. The expert reasons through evidence. The modern commonplace to “let the evidence speak” is the appeal of the expert: argumentum ad indicium. The expert rejects rhetorical appeal and brings an argument to its completion, and thus to its end. Syllogisms and rhetoric are relics of a time when people made cases, but the case is always now plainly presented: “There is a problem. We require the problem to be solved. What will solve the problem is such and such, and I know this because of my evidence.” This is the rhetoric, simple as it is, of the expert. The presentation of facts is the culmination of the agency of experts. Evidence is what is plain and clear (ex-video), what does not need the support of a convincing argument. The expert speaks not from opinion (doxa), thus requiring convincing; rather, the expert speaks plainly (evidenter). But the expert does not speak from the standpoint of pure objectivity. The expert’s knowledge, metis, is a kind of trained subjectivity.

“Expert opinion” is an oxymoron to the extent that we comply with the sense that opinion is not necessarily based upon fact or knowledge. Qualifying the opinion as being of an expert necessitates that the statement is based upon fact or knowledge, or upon evidence as herein discussed. The Greek idea of doxa, which is frequently translated “opinion,” derives from the verb dokein meaning to appear or seem. In other words, an expert does not have an opinion, but those who are not experts who believe what the expert has stated have opinions because their views are not based upon facts or knowledge but rather upon appeal to the expert. Thus, an expert opinion is a view by non-experts of how things seem based on what the expert has presented. Only non-experts can have “expert opinion,” which would be better phrased “expert-based opinion.”

• 4. What happens when you argue with the expert? The expert’s evidence is of such type that the non-expert is alienated from it. The expert is the one who experiences the making of the evidence. What we the non-experts get is the product, the artifact of the evidentiary process. Arguing with an expert is futile. When it is done, it frequently begins with the proviso, “I am not an expert.” This is an admission that I am without evidence, but in violation of Locke’s argumentum ad verecundiam, I embrace my impudence irrespective of my lacking the experience of the expert when I argue with the expert.

Because the expert is not engaged in an argument, the dispute is one-sided. Always, the expert can simply appeal to the evidence. Likewise, the expert will demand of any dispute that it be presented as an appeal to the evidence. One does not argue with the expert. Instead, one must look to fabricate evidence out of experience that contradicts the claim of the expert, thus playing the game of the expert. One has options here. The simplest approach is an appeal to a different expert with a different experience that produced different evidence. Experts can disagree based upon their experiences and the facts and evidence concomitant to their experiences, but this is fundamentally different from a non-expert disagreeing with an expert; a non-expert and an expert speak from different grounds.

What happens when experts disagree? This is indeed a strange predicament, for on what can experts disagree? They cannot disagree on the facts, though they can differ on what the facts entail. What the facts entail is the concern of policy. Sometimes this distinction is muddled, as it was in the discussion of the United States Science Court in the late 1970s. This elaborately designed process was intended to resolve questions of scientific fact but to leave matters of social value aside. It would consist of a panel of scientist judges before whom experts with opposing scientific positions would argue their cases, a disagreement that would be resolved by a report from the judges. ³ It was proposed as a limited institution; one proponent called it a set of procedures, rather than an established institution, to be used as needed. The process itself is a testament to the role of the expert: it is explicitly called an “experiment” and centers the dispute on experts being judged by other experts, not a citizen jury but a panel of experts. If one goes just slightly deeper than the surface of this idea, the science court is an idea for working out the facts because it is assumed that the facts will, once made clear, guide policy decisions. The proponents of the science court define scientific fact as “a result, or more frequently the anticipated result, of an experiment or an observation of nature” (Kantrowicz, 1977, p. 333). From the standpoint of the public, there are two interesting components. First, the public is not a party to the argument itself, which raises the question of whether it could be reasonably called an argument, especially since one of the stated reasons for the science court is to break free of public arguments where irrelevant technical claims are posed, which “generates enormous confusion in the minds of the public” (p. 333). Second, the meaning of fact is tied to an understanding of the results of experiments, even something called “anticipated results,” which sounds like it would be outside of the boundaries of the experiment itself. But it is not, for this is precisely the role of the expert: to take the experience (experiment) and translate it into actionable intelligence. A primary goal of the court is to insulate experts from the public, where confusion enters through argument. The goal of using knowledge for the design of policy has been a consistent theme of expertise. More recently, a committee called the Committee on the Use of Social Science Knowledge in Public Policy under the National Research Council (2012) released a report “to lay a foundation for continuous improvement in the conduct of social science research and its application to public policy”.

Disagreement is possible among experts, but it is settled by the determination of fact, not values, which makes it an entirely technical disagreement, thus not a matter of argument. That expertise uninhibited by argument can still suffer contestation highlights nothing more than the specific quality of expertise and of the expert in the modern state, for the expert is not merely a scientist concerned with knowledge but a social concept of agency for converting knowledge into action.

• 5. The expert of the experiment is not one who has experienced but rather one who has fabricated experience, yet another way in which the knowledge of the modern expert corresponds with metis. It is a cunning knowledge derived from the trick of the experiment. The expertise derived from the experiment is not from having done anything but rather from manipulating conditions and sometimes tricking people into doing what they are not aware they are doing, which is one feature of the social science experiment. From this trickery comes a knowledge that the expert translates to the active arena. Herein lies much of the impersonal quality of expert authority. It derives from the alienation of the non-expert from the production of the knowledge. An appeal to the evidence is mediated by the expert, who was involved in the production of the evidence. The fabrication of experience also gives the modern expert the ability to make use of others’ experiences, to interpret the experiments that others fabricate, and thus to expand the inherent authority provided by knowledge.

• 6. The expert, who reasons through evidence produced in fabricated experiences, eschews certainty. The famous adage from Bacon (1898), “If a man will begin with certainties, he shall end in doubts; but if he will be content to begin with doubts, he shall end in certainties” (p. 38), is true in its metaphorical sense. The work of the expert is never complete; beginning with doubts, the expert smooths the way to understanding. Bacon equates contemplation (contemplativa) to a similar process of action itself: “For the two ways of contemplation are not unlike the two ways of action, commonly spoken of by the ancients: the one plain and smooth in the beginning, and in the end impassable; the other rough and troublesome in the entrance, but after a while fair and even; so it is in contemplation . . .”(ibid). ⁴ The experiences of the expert are marked by the patience of doubt and the rigor of smoothing out an understanding about the world. Bacon’s metaphor is a commentary on certainty above all. To begin in certainty is an error because it is so fragile: any contrary experience will send a person of strong certainty into a condition of irreconcilable doubt. To begin in doubt is a protection against the fragility of certainty. The advancement of learning never reaches the point of certainty but rather of decreased doubt (Bacon’s fair and even path).

Certainty is categorically opposed to doubt; only faith permits certainty. In expertise, doubt forever lingers. In the early twentieth century, John Dewey took up what Bacon started. Dewey critiqued the philosophical quest for certainty and called for the integration of theory and practice to the point of raising actual experience above the idealistic systems of contemplation. Dewey (1929) called the scientific method and empirical analysis, his version of Bacon’s novum organum, “the most powerful tool we possess for developing other modes of knowledge” (chpt. 9). He proposed that ideas be “tested” in experience, subjected to experiment and measurement. Knowledge and ideas are by necessity integrated into action so that their truth is ascertainable by whether effects align with principles. We must not start with conviction (certainty) but be open to testing our convictions; truth is made by our experiences and our determinations of how well aligned our beliefs are with the grand experiment of living in the world. In other words, truth is an artifice insofar as we find it useful to living, insofar as it makes our paths fair and even. Beginning in doubt, we move to the horizon of certainty. This movement towards the horizon is the action made possible by the expert, the action of resolving a problem. As Dewey (1929) put it: “Action is the means by which a problematic situation is resolved. Such is the net outcome of the method of science.” There can never be certainty, so we settle instead for resolution.

• 7. Dewey’s pragmatism made contemplation an instrument for action. In philosophical terms, Dewey collapsed the distinction between a vita contemplativa and a vita activa, making the active life (action) priority to and determinant of contemplation. Thought and ideas are true to the extent that they are useful to the actions required of being in the world. Dewey stated plainly that the idea of a spectator theory of knowledge—his phrase for a collection of philosophical dispositions ranging from the ascetic contemplative to the academic—was anachronistic. The experimental method, being centrally about experience itself, merged theory and action insofar as all theories were ultimately hypotheses about the world. We “know” the world by experiencing it.

Through action, we encounter the world. This encounter is the experience from which we learn as well as the means through which we resolve “problematic situations.” A quest for certainty is quixotic because the conditions of our encounter are in flux. Our quest is guided by a series of attempts; it is a quest not for certainty but for expertise. Authority is not based on contemplative truth but on active truth, not produced by certainty but produced by the attempt itself. The attempt is the most human of actions, where actions are the means, as Dewey said, of resolving problematic situations.

Herein we discover how the expert possesses the unique authority of the modern state, an authority that admits its limits and its uncertainty but is no less forceful in its ambitions to resolve problems. In the experiment there is ambition, for the attempt itself is a justification of the expert’s authority: the experiment conditions (smooths) expectation by admitting beforehand that a matter may not turn out as planned. Consequently, we may need another attempt, and another, and another. And at each attempt, we will look to the expert, the one who fabricates experience (experiments), each time confident in having learned from the previous experiment (attempt).

3

As noted above, the method of science does not produce certainty. The new instrument that Bacon discovered was not an instrument to settle our metaphysical problems but rather an instrument that reshaped how we witness and act in the world. The novum organum focused attention upon the world in its physical and material sense. This limited perspective of science, in which our new instrument conditioned our sense of knowledge, forges claims and facts that are at best contingent or hypothetical. What we moderns have inherited is the condition that we exist in a world where we do not know for certain but must still act; we cannot know with certainty what our actions will produce or fully justify them, but in our attempts, we continuously test what we believe will happen and, perhaps, validate our ideas about what should happen. The expert is the agent who attends to these attempts, and the experiment is an activity that helps the expert to better specify hypotheses that will then guide social action. The mid-20th century philosopher of science Karl Popper (1957) saw clearly that science is not an operation of verification but rather an operation of falsification. ⁵ Where this leaves us in the world of action is in that hazy place of pragmatism, the place where we try. We attempt. As we build systems of authority in a world where traditional authorities have been negated by science itself, in the sense that we have rendered false our testament to the divine; we turn to the world of facts. This is a very different world from that of truth. Facts are simply that which has occurred or been done, and based upon that which has been done, we do again. The pragmatist position of deriving truth from what works reinforces the modern transformation of truth to the extent that facts—that which is done—are the condition for our conception of truth. Such is the condition that gives birth to the modern expert.

Properly speaking, an expert identity is impersonal. The expert acting as expert does not promote personal views. It is a conduit identity; the expert reveals the evidence and gives at most an interpretation of what the evidence says. Evidence presented by one expert contradicting evidence presented by another does not undermine the impersonal quality of both. The experience of the expert that gives him or her the authority attributed by that label is not an experience of transcendence or mysticism; it is a measured experience. The evidence produced by this experience can always be compared to other experiences (even if the conflict between the two is unresolved). The moment an expert speaks in the grammar of “I feel” or “I think” in a way that is not conditioned on the statement being “in my expert judgment,” the aura of expertise collapses. The best attack on the expert is to expose them personally, to expose their personal views or predilections in a way that connects them to the expert claim.

The expert speaks but not from opinion; the language of the expert is, by definition, factual. The words of the expert are not of the expert. They are the words of facts, and the expert brings them forth. In this respect, it is tempting to connect the modern expert to the traditional priest or the classical oracle. Each of these social positions is protected in their words by something compelling. The priest is guided by the thusness of tradition, the oracle by the signs of the world. But the impersonality of the modern expert is distinct because its authority derives from observable fact. The expert is beyond the oracle and the priest in the sense that the authority derives from the physical world that all can (at least potentially) share in experiencing. It is even reasonable to say that the roles of priest or oracle precede the social role of the expert, which functions as the perfection of them, for the expert, speaking in the grammar of facts, speaks in a way that approaches the actual incontestable claim and does so plainly, evidenter. But the social role of the expert is also like the priest or the oracle because the fullness of the expert’s claims leaves little space to disagree without fundamentally challenging the existing order; the expert of the modern state speaks from the perspective of the public from which the modern expert gains its authority, in the language of generality and impersonality. This quality is apparent in how experts describe the claim. They assiduously avoid the agency of the first person and rely instead upon claims that remove them from responsibility: “research has shown” and “according to research/literature/evidence.” A disagreement with the expert is not a disagreement with the person but a disagreement with the fundamental process of knowledge production.

Thus, to challenge, to contest the expert, is to question the normative foundation behind the administrative state and to its constituent public. Despite the intrinsic admission of uncertainty that underlies expertise, the expert acts in a way that seems insulated from contest. It is, in many respects, this admission of attempting that protects the expert from warranted challenge. Any challenge of the modern expert must face the response that the expert claim is at best hypothetical and contingent and that new information will only improve expertise, a logic that parallels the logic of technology: all problems, even those of expertise, are resolvable through expertise. In this respect it is accurate to claim that a challenge to the modern expert is a new form of heresy. We see this reinforcement of authority emerging in artificial intelligence, which is in a constant state of improvement as it is “used.”

4

Before getting into the matter of artificial intelligence, let us consider more closely the claim that “expert knowledge is metis.” Considered as a distinct type of knowledge, metis involves acting on the world in a deliberate manner, meaning it involves a cunning application of practical knowledge. Practical knowledge in a modern sense can be thought of as causal knowledge in the general sense that my actions will yield or produce some specific outcome. It is not necessary that the fullness of causation be known in the sense that we can attribute any result to a final cause; rather, it is sufficient for a reasoned connection to be made between a preceding event and an outcome. At a most basic level, causal knowledge is a framework for connecting desired ends with known actions, where the latter might also be action that we deem practical in the sense of doable. Causal knowledge is primarily an orientation towards answering “why” in a sense that allows for control over the answer.

This quality of answering why in an instrumental way ties causal knowledge to the scientific method. Knowing the causal relationship is adequate to the method of science, but it is inadequate to its implication, which is that the knowledge is practical, meaning it entails practice, or in a haughtier sense it entails praxis. Knowing the causal structure is a minor component of scientific or causal knowledge. The idea is to act on the world, as Freire would put it, to act on the world in order to transform it. Moreover, this word praxis connotes in Greek a meaning of proof or demonstration by doing. Put differently, the idea of praxis entails an attempt to show that one’s knowledge works to achieve some purpose.

The knowledge associated with praxis and the practical is made and not discovered, which is its key distinction from theory or theoria. This remark will strike some as obvious and others as ridiculous. But for those who would dispute it, I will simply state that few practical concerns bother to go to the length of even asking the question of ultimate cause but are satisfied with the causal structure that is sufficient to allow for doing or acting. In this sense, the knowledge that the situation demands is constructed out of a preconceived intent. If one were to seek to mitigate damage to the environment or climate, one does not challenge the premise of civilization itself but is instead satisfied with the changes to social and economic systems that will have practical results while retaining the pleasures of modern life. This is how we get policies such as carbon capture and the promotion of new lightbulbs and battery technology, policies that do not address the problem of climate change in the sense of ultimate cause. If a policymaker is concerned with addressing poverty, it is unlikely that one will explore a Christian understanding of poverty based on humility or the ancient idea of askesis; rather they will focus upon the material wants and needs of those who seem to suffer poverty. This is how we get policies such as wealth transfers or social benefits, which do more to substantiate the materialism that makes poverty a problem in a modern capitalist state than they do to see how modern poverty emerges from a macro-social condition based on growth and accumulation. In other words, the knowledge is shaped by the problem. In some cases, the knowledge shapes the problem, but this is not because the knowledge is independent of the problem but rather because the knowledge is shaped in such a way that it opens new ways of seeing problems.

To bring this matter back to the expert, the knowledge undergirding expertise is made from the experiment (out of the fabricated experience), and if done to an appropriate level, the knowledge offers practical implications. Experts may gain a better understanding of the way things work through their experiments, but the extent of their expertise is determined by how that understanding is translated in practical knowledge. How often does an interesting finding, even the most obscure astrophysical discovery, when reported to the public get attached to some “implication” that answers why this matters. Such a connection is frequently framed as what we can do with this new understanding, which is the question behind the making of knowledge and to the idea of metis: that the knowledge is concerned with manipulation, that an understanding is fine to an extent, but the fullness of knowledge is achieved in its use. The expert is an expert to the extent that he or she takes knowledge and attempts to accomplish something with it. That the expert does so from a position of impersonality and from the standpoint of incontestability testifies to the connection between metis and sovereignty, for the expert possesses the remnants of sovereignty in a world where authority is impersonal.

Incontestability also happens to be the distinctive feature of the artificial activity conducted by the expert in the administrative state, namely, policy. It is a feature that stems from the impersonality of the expert’s activity, a feature often connected with a sense of neutrality or impartiality. A commonplace understanding of the relationship between the legislative and the administrative maintains that legislative norms are realized in the administrative neutral implementation. In other words, according to the Weberian sense of the technical-rational features of bureaucracy, it is thought that a functional government requires a norm-setting arena (the legislative) over the practical execution (the bureaucracy), but the practical execution is confined to a neutral, non-normative, non-judgmental carrying-out of the requisite activities to fulfill the normative guidance. A bureaucrat acts neutrally (Weber’s sine ira ac studio).

Schmitt explained this distinction between the legislative and the administrative as that between statutes and measures. Both the legislative state type and the administrative type share in eschewing personal will as the basis of command. The expression of command characteristic of the legislative state is the statute, which maintains the distinction between the norm-setting arena of law and the legal application appropriate to the executive. Contrastingly, the administrative state “finds its existential justification in the purposefulness, the usefulness, and, in contrast to the conformity to norms and the normatively grounded legislative state, the directly concrete, factual appropriateness of its administrative measures, orders, and commands” (Schmitt, 2004, p. 9). Whereas statute is based on normative grounds, on the acceptance of a higher view of norms to justify action, the administrative state finds justification in “factual necessity, the condition of things, the force of circumstances, the necessity of the moment, and other non-normative, situation specific justifications” (p. 8). Schmitt acknowledged that the legislative state suffered from “ a certain abstractness” (p. 10) but is nevertheless based on a “several-thousand-year-old distinction . . . of the nomos against the mere thesmos, the ratio against the mere voluntas, intelligence against the blind will without a basis in norms” (p. 11). This idea, which Schmitt suggested predated even the Greek culture from which he initially derived it, distinguishes between the higher laws governing human behavior (and which presumably should govern states) and the “mere” laws of human creation, the rules and precepts. According to Schmitt, the latter have no justification other than practicality and are consequently subject to particular conditions.

The conditions of modern expertise expose how factual necessity achieved justification when an idea of “the public” replaced the general, preestablished norms that shaped the legislative statute. This occurs once the notion of the public is established as a calculable idea independent of the people, who in their plebiscitary power can only provide assent or dissent (as Schmitt put it, “the people can only respond yes or no . . . They cannot advise, deliberate, or discuss” (p. 89)). The higher norm that supported the nomos of the legislative state is replaced by the notion of the public good, which elevates fact, necessity, general will, and utility. As the legislative function is gradually delegated to the administrative, the administrative function is itself in a position to bring order, and it does so in the only way that its condition allows: through organization. Schmitt saw this happening alongside the transformation of the state “into the empty functionalism of momentary majority decisions” (p. 12), but its development proved deeper; the administrative state adopted a sense of intelligence that resulted in the fusing of statute and measure: policy.

5

Statute is the province of the legislature over the administrative function of execution, which Schmitt called measure. Something must always “back-up” the measure. Statute, on the other hand, stands on its own. Justification of statute derives from its source, whether it be divine or legislative (as Schmitt put it, the preestablished norm). The implication of the word measure, as translated into English from the German, is that of an act with a prescribed extent, hence the belief that the bureaucracy would function in a neutral manner according to legislative statute. The emergence of policy as a governing concept in the twentieth century coincided with the rise of what Schmitt called the total state (notably different from the authoritarian state, which depends upon a personal will) in which the administrative (acting as the bureaucracy) “transforms the law of the parliamentary legislative state into the measures of the administrative state” (p. 13). I believe that this transformation is best called policy.

Policy is a warped merging of measure and statute, warped by the sense that evidence is itself a sufficient norm, which is one step through which fact becomes truth. This is how the expert can speak as both administrator and legislator: the fact as observed by the expert through the experience of the expert carries the influence of truth itself; evidence is enough to justify action, and it gives more meaning to the idea of the expert’s measured experience. From the standpoint of command and authority, policy is the substance of the rule by expert that shapes the administrative state and transforms sovereignty into an idea that can be impersonal. Policy—the merging of statute and measure—is the arena of the expert, of the man-made justifier of man-made order (where higher norms are absent and even considered irrelevant). Policy within the state is by design general (i.e. it is public policy), so its effects, which determine its truth, are determined by the general benefits produced (which is what often makes public policy appear to be based on an ethos of utilitarianism). ⁶

This kind of rule–a transformation of sovereignty—is a new version of an old idea. Where it is old, we find words available such as noocracy or epistocracy, both of which speak to a power derived from intellect (nous) or knowledge (episteme). Such ideas in Western thought can be traced to Pythagoras and his mathematikoi or to Plato’s Republic (Politeia) or Laws (Nomoi). Both forms of rule, noocracy and epistocracy, are basically conditions on aristocracy, conditions through which the aristoi is more specifically defined in terms of knowledge. In recent debates these old ideas have taken a new flourish, primarily as a way of questioning the justice of democracy itself. Jason Brennan (2017) is a prominent voice promoting epistocracy in his Against Democracy, a lengthy argument to defend the idea that democracy, insofar as it subjects all people to the decisions of its process, potentially places them also in the unjust position of being subject to incompetent governance. Instead, Brennan argues, a more just governance would give decision making authority to those of demonstrated competence, thereby reducing the potential for harm brought by incompetent political decisions. Brennan’s position is that democracy is more likely than epistocracy to produce bad decisions, which makes democracy more unjust. This position has had its own critics who have pointed out that there is no way to guarantee either that the epistocrats would not themselves be systematically biased (the process for gaining office, such as through education, might be systematically biased) (Estlund, 2007, pp. 221–223) ⁷ or that they would be trustworthy (Moraro, 2018).

These concerns are not those of the modern expert, though. The political conditions of modern expertise elude the questions of justice and morality that lay behind positions for and against democracy or epistocracy. In place of these concerns is the practical. Aristotle’s derivation of moral expertise through the phronimos is one ingredient in the eventual elevation of practicality achieved through the novum organum, an elevation made plain in the pragmatist conception of truth. The expert of the administrative state takes its authority from the discovery of fact, and where fact is regarded level with moral truth and the good, the authority of the expert is firmly placed, for the expert does not derive authority in any personal sense but does so from the communicable discovery of what works, of the fact through the novum organum. We are led to a world not of epistocracy, not of democracy, not of aristocracy, but of pragmatocracy, a world in which authority is placed in the things done (pragmata) discovered through experiments, in which the affairs of the state are the state. ⁸ The expert is not a leader of political or moral life but is an instrument in the operation of the total state functioning according to practical necessity. The remnants of agency in the administrative state accumulate in the expert, the expert who acts for the public, for Nobody, a thing administering itself, pragmata with agency.

The question of “who decides” has occupied much of the debate around democracy, and rightfully so since it is a foremost question of politics. That Nobody decides, that in effect the decision is a product of planning, design and manipulation (of metis), is rarely admitted even if it is often experienced, as when a person encounters the non-decisions of bureaucracy. Still, it is commonly postulated that some decision is necessary; at some point down the line a decision about the design must have been made. Somewhere and sometime a decision about what is to be done was proffered, and the question then turns directly back to that who in that time and place. This is especially evident when applied to the burgeoning applications of artificial intelligence and algorithm-based implementation. But the debate is strictly speaking fallacious; it is based on a version of a “turtles all the way down” reductio ad absurdum. It is similar also to the Aristotelian unmoved mover, or the Neoplatonic One, a problem in politics remedied by the event of the Constitution, a document that is created in a previous generation and held to bind future ones. Contemporaries face conditions because of historical factors outside of their control; certainly there is some debate about how to go about changing those conditions, how to alter preconception and predilection, but such concerns are outside the practical ones that are the matter of policy. Policy proceeds from, but it also avoids the key political concern of sovereignty, namely, who decides. Policy avoids this question in its relationship to measure, which is the necessity of action in circumstance, but it still retains the constitutive force that follows from statute following the fact as norm. To answer, then, that Nobody decides is not an avoidance of the problem but a precise answer to the question of decision in a world where sovereignty has been dissolved, where authority requires modesty in facing fact. Agency remains in the form of Nobody through the expert. Modern expertise is the remnant of agency in a society that turns progressively toward Nobody, and sovereignty remains in an altered form of the impersonal.

As a remnant agency, the expert is destined to disappear. For centuries people have been preparing for the rise of the impersonal algorithm, a condition prefigured by bureaucracy itself. People are more inclined to rely on the advice of algorithms than of other people when issues are difficult or complex, a preference that challenges the popular idea of the wisdom of crowds as a defense of democracy and highlights how instead the crowd produces bureaucracy, as Weber and Schmitt both suggested. ⁹ The sovereign as Nobody is realized through the order of organization achieved through the artifice of decision, or artificial intelligence.

6

A way to think about pragmata with agency is through artificial intelligence, which can be viewed as a conceptual extension of the expert and a realization of the impersonal features that shaped this specifically modern social role. What troubles many about the operation of AI is its opacity, a problem that differs only slightly from the visceral reaction people have to experts who rely on evidence that is inaccessible to most people. But its opacity is tamed by the sense that artificial intelligence operates impersonally by design rather than by will. A person can expect that the decision of an artificial intelligence follows the algorithm as closely as possible without corruption, whereas an expert must be trusted insofar as the expert’s bias and prejudice are willfully suppressed. Concerns about bias in artificial intelligence are simply design problems, whereas bias and prejudice in human beings are problems of the psyche, that elusive, human-distinguishing quality. An artificial intelligence is under constant states of improvement by design, similar to a social system. This meliorist quality of AI is its most attractive feature to the modern state.

Improvement is not always obvious, though. The administrative values of accountability and transparency, which we derive from social attempts to reconcile the power of the administrative state in bureaucracy and the expert, are also applied to AI. But just like accountability of bureaucracy begets bureaucracy, just like transparency demands new bureaucratic practices to make bureaucracy transparent, to hold AI accountable or to make it transparent will require artificial intelligence. No human intelligence can grasp the reasoning behind AI. In fact, it may even be difficult for AI to explain its reasoning in a linear, understandable way to a human intelligence. The point is that accountability and transparency are social values of bureaucratic society, reflective of the tendency of mass democracy to produce complex systems for governing. We might regard this as a social law: problems in technology demand technological solutions; translated to politics to mean problems in administration demand administrative solutions. The only way out is to obey the ancient advice of Lao Tzu: do nothing, take actionless action, and relinquish control. But we live in a world where it is always demanded that we do something, and then something else, and then something more, until we have done enough to say that we did it. We act to produce fact because doing something solves problems.

This is the idea behind the commonplace preposition, used so often in contemporary political debates, “In an increasingly complex society . . .,” and in such a society, administration bears the burden of simplification. The simplification is not, however, of the sort that the people, the demos, can access it. On the contrary, it is a simplification to serve the administrative function, to serve the objective of policy, the merging of measure and statute. Artificial intelligence is a product of that simplification in administrative work, but its own working is phenomenally complex. And the only response seems to be to use the tools of administration (accountability and transparency) to guarantee that it works “appropriately.” This is little different than how we conceive of experts, except we can personalize the experts despite their impersonal role.

The fabricated experience of the experiment is an essential part in the political function of the expert. Like Chiang’s human scientists, the expert of the modern state is charged with interpreting the otherwise indecipherable knowledge produced from the experiment into a language of authority. The interpretation is an expression of policy. In this sense, the expert plays a role that veers toward the theological in the scientific. Although the modern sensibility draws a strong distinction between these arenas of reason, the hermeneutic demands of knowledge generation into a framework of action and practice cannot happen without the interpretive ability of the expert. This function of the expert underscores the argumentative justification of expertise: the expert does not argue but speaks plainly, much like a priest does not argue but testifies on matters of faith and an oracle does not argue but divines through the ethereal and ineffable. But arguments are made based on the words of humans in these social roles. The argumentative structure is an appeal to the modesty of the listener, and a dispute against their words is an act of impudence. There is very little difference between the request that we trust the experts and the formulation that we have faith in experts, especially insofar as humans are increasingly alienated from the knowledge process that justifies authority in the administrative state. But the impudence is no less justified insofar as the systems themselves have become so complex as to elude human understanding, a problem that is both exacerbated by and solved through the rapidly growing functionality of artificial intelligence. Chiang’s story has more than a hint of irony in its message that science has advanced beyond the grasp of humans. Bacon's novum organum, the instrument of science itself, was seen throughout the Enlightenment as an instrument that would liberate humanity from the ignorance that supported faith-based traditional authority. That same instrument of discovery—commonly called scientific method - we now confront as being on the precipice of exceeding our own grasp, so that we now find ourselves in a similar position as the scholastic hermeneutic of interpreting what we witness has been done (fact) to facilitate our actions. The quality of the total state that Schmitt alluded to, the condition of things administering themselves, is realized through artificial intelligence, a situation in which we subject ourselves to the rule of nobody and seek to understand that rule in the artifacts produced through it, strengthening its position as we use the tools that empower it to try to control it. But the public is ready for this. It has been conditioned to this behavior through experts.