The Possibility of Practical Reason, 2nd Edition

11  The Way of the Wanton[1]

In his 1970 paper “How is Weakness of the Will Possible?” Donald Davidson introduced his solution to the philosophical problem of akrasia by framing it as a problem about the self: [2]

Davidson did not refer to the self as such, but when he spoke of “the agent’s representative”, he was invoking the same idea—the idea that a person’s psyche hosts a conversation among several voices, one of which has the distinction of speaking for the person himself.

Davidson’s solution to the problem of akrasia did not live up to its picturesque introduction. It did not identify “the agent’s representative” but merely divided the conative attitudes into prima facie, all-things-considered, and “all-out” judgments. The last of these judgments was cast in the role of The Will, insofar as it constituted the agent’s immediate intention to act, but Davidson never explained why this attitude should be conceived as representing the agent rather than simply bringing up the rear in a parade of attitudes passing through the agent’s mind.

A more successful attempt at identifying “the agent’s representative” appeared in the following year, when Harry Frankfurt published “Freedom of the Will and the Concept of a Person”.[3] Like Davidson, Frankfurt discussed cases of conflicting motives, but he offered an explanation of how one of those motives gains the authority to represent the agent himself. A motive gains this authority, Frankfurt explained, from the agent’s desire that it predominate, which Frankfurt called a second-order volition. Frankfurt illustrated this phenomenon with the example of a drug-addict who has a second-order desire to resist his craving for the drug: [4]

Frankfurt is usually interpreted as saying that human agency stems from the occurrence of higher-order volitions. What has not been widely noted, I think, is that Frankfurt traced the roots of agency further, to a particular interest shared by all human agents. In “Identification and Wholeheartedness”, he wrote: [5]

Here Frankfurt spoke, not of particular higher-order desires to be motivated in one way or another on a particular occasion, but of a standing desire to be motivated as we want, and to be so motivated because we want to be. This is, in fact, a third-order desire, for there to be correspondence and causal influence between our second- and first-order desires.

The importance of this general interest in our own motivational integrity was reflected in Frankfurt’s portrait of the paradigm nonagent, a figure that he dubbed the “wanton”:[6]

The characteristic feature of the wanton is not that he merely lacks higher-order desires as to how he is motivated. Merely to lack such desires might just amount to indifference, which is not sufficient for wantonness. What characterizes the wanton is “his mindless indifference to the enterprise of evaluating his own desires and motives”, which is the lack of a third-order desire to engage in forming second-order desires.[7]

Frankfurt envisioned this “enterprise” as requiring some motive of its own: [8]

The wanton lacks this motive for reflective evaluation, and so he is like someone who loses interest in checking his arithmetic: [9]

What’s analogous to wantonness is not the lack of an opinion as to whether the last calculation is right or wrong but rather a lack of interest in the enterprise of forming such an opinion. The wanton thus lacks a third-order motive for forming second-order evaluations of his first-order attitudes.[10]

Frankfurt thus posited a single motive whose operation lies behind all human agency. Our agency arises from our concern over “whether the desires by which we are moved to act as we do motivate us because we want them to be effective in moving us or whether they move us regardless of ourselves and even despite ourselves.”[11]

But how can “the desires by which we are moved to act as we do motivate us because we want them to be effective in moving us”? How, that is, can our wanting a desire to be effective in moving us make any difference as to whether it actually is effective? And even if a first-order desire can be made effective by our wanting it to be so, how do we thereby become identified with it, so that it becomes more truly our own, attaining the status of what Davidson called “the agent’s representative”?

I am going to offer answers to both questions. Although I will trace these answers to various passages in Frankfurt’s papers, I cannot claim that they are his. They certainly do not belong to the standard interpretation of Frankfurt’s analysis of agency. The most I can claim is that they belong to an interpretation under which Frankfurt’s analysis is correct—correct as the analysis of something, that is, though not necessarily of agency.

When Frankfurt describes desires as moving us “because we want them to be effective in moving us”, he certainly seems to have a causal process in mind. Speaking of the so-called willing addict, who wants to act on his addiction, Frankfurt writes: [12]

“Overdetermination” sounds like a causal process. Frankfurt appears to be saying that the addict’s first-order desire to take the drug is made effective by two independently sufficient causes: it is effective both “because he is psychologically addicted” and “because he wants it to be.” The latter of these causes is the addict’s second-order volition to be actuated by his first-order desire to take the drug. The question is how this cause operates: how can a second-order desire cause a first-order desire to become motivationally effective?[13] A related question is why Frankfurt describes this process as the overdetermination of the addict’s desire rather than the overdetermination of his behavior.

I think that an answer to the first question lies in the follow passage introducing Frankfurt’s notion of a second-order volition: [14]

The end of this passage is subject to at least two readings. The second way of wanting to want to X is a desire not just for some desire to X but for an effective desire to X, which is what Frankfurt would call a second-order volition. On the one hand, the passage could mean that this second-order volition must be a desire for the effectiveness of a particular first-order desire that one already has and, happily, wants to have. The second-order desire must refer to an already existing desire and must wish, of that desire, that it be effective. On the other hand, the reason why wanting an effective desire to X entails having a desire to X may be, not that it requires an already existing desire to X that can serve as its object, but rather that it just is a desire to X. Wanting an effective desire to X is, by definition, wanting that one actually do X as the result of a desire to do so. It therefore involves wanting oneself to X, which is a reflexive way of wanting to X.

I think this second reading is preferable, because it doesn’t require that one’s second-order desire be a desire for a first-order desire that one already has. According to this reading, both ways of wanting to want to X are desires for a first-order desire that one does not already have. What Frankfurt is saying, on this reading, is that in wanting an effective first-order desire to X, which one does not already have, one already does have a desire to X, after all — namely, this very second-order desire.[15]

This reading helps to explain the following passage from a different paper:[16] “Suppose a man wants to be motivated in what he does by the desire to concentrate on his work. It is necessarily true, if this supposition is correct, that he already wants to concentrate on his work. This desire is now among his desires.” Why “now”? The answer, according to this interpretation, is that the desire to concentrate on his work just consists in his second-order desire for such a desire to motivate what he does.[17]

This interpretation also enables us to explain how a second-order volition can cause a previously existing first-order desire to become motivationally effective though it was not effective before. A second-order volition has a first-order motivational force matching that of the first-order desire whose effectiveness it endorses. A pre-existing desire to X can thus be reinforced by a second-order desire that it be effective, because the latter desire is also a desire (for oneself) to X. Hence, a second-order volition can cause a first order desire to become motivationally effective by reinforcing it.

Thus, when Frankfurt says that the case of the willing addict involves “overdetermination of his first-order desire to take the drug”, he means that the addict has two desires to take the drug: his addictive urge to take the drug and a desire that he take the drug because of that urge, which is a desire for himself to take the drug and hence a reflexive desire to take it.[18] Because the addict has two desires to take the drug, the fact that he wants to take it is overdetermined—which, I think, is what Frankfurt means by the “overdetermination of his first-order desire to take the drug.”

This explanation necessitates a subtle clarification about the content of second-order volitions. A second-order volition that one be effectively moved by a first-order desire cannot have the content that one be effectively moved by the first-order desire alone. The content of a second-order volition must be that one be effectively moved by the first-order desire as reinforced by this very volition. Otherwise, the volition would tend to be self-frustrating.

For suppose that the willing addict formed a second-order volition to be effectively moved by his addiction alone. If he ended up taking the drug because his addictive urge was reinforced by his second-order volition, then he would not have taken it because of the addictive urge alone, and so he would not have satisfied the volition, after all. His volition would be necessarily self-frustrating in the sense that its motivational effectiveness would tend to cause it to go unfulfilled: in order to be fulfilled, it would have to be motivationally inert.

Surely, a volition of this form is inappropriate for creatures like us, to whom “[i]t matters greatly . . . whether the desires by which we are moved to act as we do motivate us because we want them to be effective in moving us or whether they move us regardless of ourselves and even despite ourselves.”[19] If we want it to be the case that our desires are effective because we want them to be, then we had better not want them to be effective all on their own; our desire for them to be effective had better leave open the possibility of their being effective because of our hereby wanting them to be so.

Indeed, given our general desire that our motives for action be effective because we want them to be, our second-order volitions may not only leave open that possibility but positively favor it. Our desire that some first-order motive be effective may be, more specifically, the desire that it be effective partly because of being reinforced by this very desire. Second-order volitions would then incorporate the concern that is fundamental to agency, in Frankfurt’s view. Wanting our first-order motive to be effective when reinforced by this very volition would be an instance of our concern that motives actuate us because we want them to, rather than regardless of or despite ourselves.

The resulting model of agency can be summarized as follows. Our concern that motives generally actuate us because we want them to leads us to reflect on our desires and to form a second-order volition that one of them actuate us because of our hereby wanting it to. This volition is a desire for us to do that which the first-order desire is a motive for doing, and so it reinforces that desire, which consequently actuates us because we want it to.

How can this process constitute our identifying with the first-order desire? How does it authorize the desire to represent us in the contest among our motives? I do not believe that an answer to these questions can be extracted from Frankfurt’s own words. Frankfurt’s conception of identification remains opaque throughout his writings on agency. I am going to propose an answer, partly on the basis of hints in the texts but largely on the strength of its own plausibility. I think of it as the answer that might have been in Frankfurt’s mind as he wrote or may be in ours as we read, insofar as we find him persuasive.

Frankfurt states “The Problem of Action” as follows: [20]

Frankfurt points out that the contrast between “instances in which purposive behavior is attributable to a creature as agent and instances in which this is not the case” can be applied to the behavior of lower animals, such as spiders.[21] The appearance of the contrast in the behavior of spiders constrains possible solutions to the problem of agency: “The conditions for attributing the guidance of bodily movement to a whole creature, rather than only to some local mechanism within a creature, evidently obtain outside of human life. Hence they cannot be satisfactorily understood by relying upon concepts which are inapplicable to spiders and their ilk.”[22]

An analysis of action in terms of reflective attitudes, which are presumably unavailable to spiders, would appear to violate this constraint. Frankfurt therefore suggests that there must be general conditions of agency whose satisfaction in humans requires these sophisticated attitudes but whose satisfaction in spiders does not: “While the general conditions of agency are unclear, it may well be that the satisfaction of these conditions by human beings depends upon the occurrence of events or states which do not occur in the histories of other creatures.”[23]

To my knowledge, Frankfurt never attempts to clarify what he here calls “the general conditions of agency”, which pick out the behaviors that are attributable to spiders rather than to sub-arachnoid mechanisms, just as they pick out the behaviors that are attributable to people rather than sub-personal mechanisms. But here is a passage that I find suggestive:[24]

Here Frankfurt makes explicit what is elsewhere only implicit—and yet, I think, pervasively implicit—in his writings on agency: his approach to the topic is phenomenological, informed primarily by observations about “our sense of our own agency”, of what it is like to be an agent. And what it is like to be an agent, Frankfurt says, is to be somehow “in touch” with the mechanisms guiding our behavior.[25]

Suppose that the phenomenological notion of being “in touch” with what guides one’s behavior expresses “the general condition of agency”, as Frankfurt conceives it. If this general condition of agency requires for its satisfaction in humans the occurrence of events or states that do not occur in other creatures, as Frankfurt suggests, the reason should be that those events or states are necessary in order for humans, but not for other creatures, to be in touch with the mechanisms guiding their behavior. Maybe spiders achieve agency with less sophisticated mental equipment, in other words, because they need less equipment to be in touch with their own mechanisms of behavioral guidance. So, at least, we might hypothesize on the basis of these few passages.

What would account for such a difference between humans and spiders? Two of the papers under consideration—“Freedom of the Will” and “Identification and Wholeheartedness”—begin with the distinction between human persons and lesser creatures, including the lower animals. In the first paper, the difference is attributed to the presence in persons of hierarchically ordered attitudes. But in the second, it is attributed to the presence of reflective consciousness, and Frankfurt remarks in a footnote, “The notion of reflexivity seems to me much more fundamental and indispensable, in dealing with the phenomena at hand, than that of a hierarchy.”[26] Perhaps, then, reflective consciousness is what makes additional events and states necessary to put persons in touch with the mechanisms guiding their behavior, thus enabling them to satisfy the general condition of agency.

I have now inched my way along a branch of Frankfurtian texts that can no doubt be lopped off behind me at any one of several junctures. Rather than trust my weight to the texts any longer, I am going to leap off into thin air. What follows cannot be supported by textual evidence at all; it is a speculative reconstruction.

According to the standard interpretation of Frankfurt, the difference between the motives with which one is “identified” and the motives to which one stands as a “helpless bystander” is that one wants the former to be effective in determining one’s behavior but wants the latter not to be effective.[27] What produces the opposite of identification with one’s motives (a condition for which Frankfurt has no negative term correlative to “identification”) is the presence of a negative higher-order attitude.

I want to suggest an alternative to this interpretation.[28] In the phenomenological sense of “identification” that I attribute to Frankfurt, the opposite of identification—call it dis-identification for now—is produced by reflective consciousness itself.

Being the subject of a desire usually entails being the subject of various thoughts symptomatic or expressive of the desire. Being thirsty, for example, entails thinking thirsty thoughts: looking around for quantities of liquid; wondering if they are potable; considering ways of reaching them, avidly imagining their taste; and so on. All of these thoughts are framed from the perspective of a potential drinker, but none explicitly represents the occupant of that perspective. They are framed from the point of view of a potential drinker who remains out of the picture, at the unrepresented origin of that point of view. Of course, the thoughts symptomatic of thirst may include the first-personal thought “I’m thirsty,” but that thought is in the first instance an atomic expression of thirst, like smacking one’s lips or crying “Water!” rather than a compositionally analyzable attribution of thirst to oneself.

The difference between that expression of thirst and the attribution of thirst to oneself defines a continuum of possible thoughts that include awareness of one’s thirst in various degrees of explicitness. Sometimes one looks for a drink without yet knowing that one is thirsty; sometimes one looks for a drink while knowing about one’s thirst but not attending to it at all; sometimes one attends equally to the possible drink and the dryness of one’s throat or the urgency of one’s craving; sometimes one focuses on the thirst to the exclusion of the prospects for slaking it.

Across this continuum, one becomes progressively less engrossed in the activities motivated by thirst. At the former end are the cases in which one “loses oneself” in gazing at the cool drink being served at the next table, or in peeling an orange, or in assaulting the shell of a coconut. In the middle of the continuum are the cases in which one undertakes such activities with cool self-possession. At the latter end are cases in which one is distracted by one’s thirst from the very activities that it would motivate. Cases of the first kind can end with the thought “Oh, I must be thirsty”: noticing that one’s attention has become engrossed in the pitcher of water carried by a waiter, one belatedly becomes aware of one’s thirst. Cases of the last kind can end with the thought “Stop thinking about how thirsty you are and get a drink!”

This last thought is naturally couched in the reflexive second person, because it occurs when one has put a distance between oneself and one’s thirst—that is, between one’s reflecting self and one’s thirsting self. Attentively reflecting on one’s thirst entails standing back from it, for several reasons. First, the content of one’s reflective thoughts is not especially expressive of the motive on which one is reflecting: “I am thirsty” is not an especially thirsty thought, not necessarily the thought of someone thinking thirstily. Second, attentive reflection is itself an activity—a mental activity—and as such it requires a motive, which of course is not thirst. Reflecting on one’s thirst is therefore a distraction from acting on one’s thirst, and in that respect it is even a distraction from being thirsty. Most importantly, though, consciousness just seems to open a gulf between subject and object, even when its object is the subject himself. Consciousness seems to have the structure of vision, requiring its object to stand across from the viewer—to occupy the position of Gegenstand.

The distancing effect of reflective consciousness can be represented in terms of perspectives or points of view. The perspective one occupies as the subject of thirst is not the same as the perspective one occupies as the subject of reflective thinking about being thirsty. In the former perspective, one’s thirst is out of view, at the unrepresented point of origin, from which it issues in thirsty thoughts and actions directed at possible sources of drink, which dominate the field of view. In the latter perspective, one’s thirst is placed within the field of view, whose unrepresented origin may now contain reflective curiosity or concern but is unlikely to contain thirst, since thirst is not a motive for reflecting. The reason why becoming reflectively aware of one’s thirst tends to make one less engrossed in thirsty activities is that such awareness draws one away from the perspective in which thirst occupies the governing point of origin rather than the passive field of view.

These remarks return me to my current suggestion, that we are distanced—or, as I put it, dis-identified—from our motives by reflective consciousness itself. The more conscious we become of a motive, the more it becomes the object of our thought; and the more it becomes the object of our thought, the less we think from the perspective of its subject; and the less we think from the perspective of the motive’s subject, the less engrossed we are in the activities that it motivates.

This suggestion bears on what I have interpreted Frankfurt as believing about the general condition of agency. My interpretation is that what makes for agency in general, according to Frankfurt, is the phenomenological condition of being “in touch” with the mechanisms guiding one’s behavior. And the feature that distinguishes a person from a lower animal now turns out to take the person out of touch with those mechanisms, since what distinguishes him as a person is reflective consciousness, which opens a gulf between him as the subject of reflection and his motives as the object. A lower animal is already in touch with its motives in a way that a person is not, because the animal lacks the capacity for stepping back to a reflective distance. A person is continually falling out of touch with his motives, by becoming reflectively conscious of them.

That is why a person’s ability to satisfy the general condition of agency “depends upon the occurrence of events or states which do not occur in the histories of other creatures.”[29] In order to satisfy the condition of agency, a person must overcome a disadvantage that other creatures do not face—namely, the distancing effect of reflective consciousness. It is in order to bridge this reflective gap that human agency requires events and states not present in other creatures. Their function is to return the person and his motives to a self-rapport of a kind that lesser creatures are not capable of losing. How can second-order volitions perform this function? My earlier account of their motivational efficacy explains how.

As I explained earlier, a second-order volition to be actuated by a desire to drink is a desire for oneself to act on that lower-order desire, possibly as reinforced by this very higher one. It is therefore a desire for oneself to drink, which is a reflexive way of wanting to drink. That is why the volition can reinforce the lower-order desire to drink, thereby enabling it to take effect in one’s behavior.

Because the second-order volition is a reflexive desire to drink—that is, a desire for oneself to drink—it enables one to combine perspectives that were previously disjoint. In wanting oneself to drink, one occupies a perspective in which one is simultaneously the subject of a desire to drink and the subject of reflective consciousness of oneself as wanting to drink. Reflection on one’s first-order desire to drink may have drawn one out of one’s perspective as subject of that desire, but reflexively wanting oneself to drink introduces a desire to drink into one’s new perspective as subject of reflection. The higher-order desire enables one’s reflecting self and one’s thirsting self to share a perspective, because one now thirsts reflectively. Therein lies the identification, which closes the reflective gap, putting one in touch with the mechanisms guiding one’s behavior and hence fulfilling the general condition of agency.

As I have said, I do not attribute this conception of identification to Frankfurt; I attribute it primarily to myself as a reader of Frankfurt: it is what I have in mind when I find myself persuaded by Frankfurt’s talk of identification. Maybe it is what other readers have in mind, too. For those whose reading of Frankfurt is informed by some such conception of identification, I want to indicate three significant consequences.

First, this conception of identification insulates Frankfurt’s view from a well-known objection that was raised by Gary Watson and addressed by Frankfurt on several occasions.[30] The objection is that second-order volitions cannot succeed in identifying the agent with a first-order desire unless the agent is already identified with the second-order volition. If he can make the “analytically puzzling statement” that a first-order desire is “a force other than his own”, then he should be able to make a similar statement about a second-order volition.[31] And if the second-order volition is a force other than his own, then its endorsement of a first-order desire can hardly make the latter “truly his own”. How, then, does the agent become identified with his second-order volition? If the answer is that the second-order volition becomes his own by being endorsed at a higher order, a vicious regress of volitions will ensue.

This regress disappears under the conception of identification that I have introduced. For one thing, the agent is identified with his desires by default, until he steps back to reflect on them. And for each reflective step back, the agent can bridge the resulting distance by forming a reflective desire, with which he will be identified by default. No third-order desire is required unless and until the agent takes a second step back, to reflect on his second-order volition. When the agent stops stepping back, the regress stops with him.

Secondly, the agent’s second-order volition can endorse itself, at least implicitly, since it may be a desire for his first-order motive to be efficacious, not by itself, but with reinforcement from this very desire. Reflective desires represent a general solution to the problem of reflective detachment—a way of being reflectively conscious of a desire without stepping back from it. They provide this solution because they provide a perspective in which the agent is simultaneously the subject of the desire and the subject of reflection upon it. Of course, an agent can step back from a reflective desire, to a perspective of higher-order reflection. But then he can close the resulting gap, with another reflective desire, for which there is no problem of reflective detachment, unless he chooses to create one by reflecting further. So my conception of identification has some good news for readers of Frankfurt: the vicious regress is gone. Now for the bad news.

The bad news is that Frankfurt’s theory, as I have now interpreted it, does not solve the problem stated by Davidson—the problem of picking out “the agent’s representative” among the competing motivational forces within him. The agent’s representative is supposed to speak with the authority of the agent, articulating what will count as his best judgment and executing what will count as his decision. Representing the agent is thus a normative matter of being authorized, as his proxy, to make commitments for which he will be responsible.

But I have now interpreted higher-order volitions as identifying the agent with his motives, not in the normative sense of authorizing them to act as his proxies, but in the phenomenological sense of putting him “in touch” with them, by bridging the reflective gap. Under my interpretation, Frankfurt’s theory becomes a theory of how to stay engaged or even engrossed in one’s activities, despite the distancing effects of reflective consciousness. As such, it may no longer pick out a proper part of the psyche that (in Davidson’s words) “can execute the decision and take the rap.”[32]

Indeed, Frankfurt’s theory may no longer be a theory of the self, under my interpretation. Any motive, even the most fleeting whim, can be the object of reflective identification; and what an agent identifies with at one moment, he can step back from at the next. Surely, being the object of the agent’s identification is too ephemeral a status to carry the title of self.

Of course, Frankfurt himself has moved beyond the theory of the self expounded in the papers that I am interpreting here. His new theory, while still invoking the notion of identification, focuses mainly on the fixity of the self, its role as a stubborn bulwark against ambivalence and vacillation.[33] Perhaps, then, Frankfurt himself believes that what his earlier papers offered was not, after all, a theory of the self.

Finally, if Frankfurt’s theory is intended to solve the problem of reflective awareness, as I suggest, then it may turn out to be a half measure, stopping short of a complete solution. Although we can bridge the reflective gap and get “in touch” with our motives by means of higher-order volitions, we can eliminate the gap entirely by becoming so engrossed in an activity that we stop reflecting and lose ourselves. There is at least one philosophical tradition that recommends transcending reflective awareness in this manner. It is the Daoist tradition, especially as represented in the Zhuangzi.[34] In my interpretation of Frankfurt, his theory of agency becomes a prolegomenon to that work.[35]

The spiritual ideal expressed in the Zhuangzi is one of effortless action, as described by the phrase wu wei. The word wei means “action”, and wu wei is its negation—literally, “nonaction”. But “nonaction” does not mean doing nothing at all; it means acting without deliberate intention or effort—spontaneous activity: “It describes a state of personal harmony in which actions flow freely and instantly from one’s spontaneous inclinations—without the need for extended deliberation or inner struggle—and yet nonetheless accord perfectly with the dictates of the situation at hand, [and] display an almost supernatural efficacy.” [36]

A prime example of spontaneous activity is the exercise of a “knack”, which is exemplified throughout the Zhuangzi by stories such as this: [37]

Although we would regard carving oxen as the purposeful application of a skill, Butcher Ding has gone “beyond technique” and his movements are due to “spiritual desires” that “proceed” as if on their own. This process requires a shedding of ordinary purposes and precepts, as another artisan explains: [38]

The performance of artisans like Butcher Ding and Woodworker Qing is guided by an inexpressible knack. Wheelwright Pien says: “You can’t put it into words, and yet there’s a knack to it somehow. I can’t teach it to my son, and he can’t learn it from me.”[39]

The way to exercise such a knack is not to keep one’s eye on an ultimate goal, or to follow the precepts of a method, or even to focus on one’s actions themselves. On the contrary, Woodworker Qing must forget external goals (“congratulations and rewards, titles or stipends”), forget evaluative judgment (“blame or praise . . . skill or clumsiness”), and indeed forget himself: “I forget I have four limbs and a form and body.”

Such forgetfulness is necessary because spontaneous action is inhibited by distinctions between good and bad, right and wrong—or shi and fei, often translated as “That’s it” and “That’s not”:[40]

To “roam that free and easy take-any-turn-you-please path” is not a matter of taking a random walk through life; it is rather a way of attaining effective activity (“nonaction”). It can even be a solution to the problem of “how to rule the world”: “Let your mind wander in simplicity, blend your spirit with the vastness, follow along with things the way they are, and make no room for personal views—then the world will be governed.”[41]

A “wandering mind” is conducive to effective activity because of its responsiveness to the objective demands of one’s materials and circumstances—its capacity to “follow what is inherently so.” Thus, Butcher Ding “rel[ies] on the Heavenly patterns” and Woodworker Qing “examine[s] the heavenly nature of the trees.” This responsiveness to “things the way they are” is what requires the mind to be emptied of external goals and evaluative judgments.

Ultimately, emptying the mind of goals and judgments leads to forgetfulness of the self. Here is another description of the same phenomenon: [42]

The Zhuangzi’s conception of spontaneous activity has been compared to the “flow” experience described by the psychologist Mihaly Csikszentmihalyi.[43] Csikszentmihalyi conducted research in which subjects were prompted to record their activities, and their feelings about them, at regular intervals during the day. He then identified a category of “optimal experiences” that occur in the course of highly challenging activities in which the subject exercises appropriate skills.

According to Csikszentmihalyi, “flow” begins as follows: [44]

As in the “knack” stories of the Zhuangzi, evaluative judgment is suspended: [45]

Also as in the “knack” stories, awareness of the self disappears: “[O]ne of the most universal and distinctive features of optimal experience [is that] people become so involved in what they are doing that the activity becomes spontaneous, almost automatic; they stop being aware of themselves as separate from the actions they are performing.”[46] Csikszentmihalyi goes on to explain that this loss of self-consciousness “does not involve a loss of self, and certainly not a loss of consciousness, but rather, only a loss of consciousness of the self.”[47]

As we have seen, Frankfurt regards reflective awareness as the distinctive characteristic of humanity. A spiritual ideal of transcending reflective awareness would thus be, in Frankfurt’s terms, an ideal of transcending what makes us human. But transcending what makes us human is just what the Zhuangzi and Csikszentmihalyi recommend.

According to the Zhuangzi, the Daoist sage “has the physical form of a human being but lacks the human essence.”[48] He explains: [49]

The “likes” and “dislikes” mentioned here appear to be based on judgments of right and wrong, or “That’s it” and “That’s not”, and so they appear to be reflective attitudes. They no doubt include motives of the sort that result from thinking about “Goodwill and Duty”, or “skill or clumsiness”—self-critical instances of “That’s it” and “That’s not”. Hence, lacking a human essence, in the Zhuangzi’s sense, must entail lacking that “concern with our own motives” that makes us “care about what we are.”[50] It therefore entails lacking what Frankfurt identifies as the source of human agency, according to my interpretation.

Creatures who lack human reflectiveness are at an advantage in attaining effortless action: [51]

The millipede and snake are like Frankfurt’s spider, who attains agency without all of the complications that are necessary for human beings. In all of these instances, agency is more readily attained because there is no reflectiveness that needs to be overcome.

Yet we humans cannot avoid exercising our reflective essence on the way to transcending it. Butcher Ding has gone “beyond technique”: that is, he cultivated a skill and then left it behind. So it must be with all of the artisans. They must have undertaken many years of self-critical practice in order to attain their capacity for self-forgetful spontaneity. Hence, the spiritual ideal of the Zhuangzi is not an alternative to that of reflective agency: it is the next step.

When a human being “finds flow” in the exercise of a skill, does he instantiate agency, as Frankfurt conceives it, or does he instantiate wantonness instead? Or is this case rather a challenge to the categories of agent and wanton altogether?

When we are in flow, we are not actuated by the self-concern that is central to agency in Frankfurt’s conception. We are not “concerned with our own motives”, and certainly not “ceaselessly alert to the danger that there may be discrepancies between what we wish to be (or what we wish to seem to be) and how we actually appear to others and to ourselves.”[52] Such self-concern is precisely what must be left behind if flow is to be attained. When a person is in flow, “it never occurs to him to consider whether he wants the relations among his desires to result in his having the will he has”—which is how Frankfurt describes the wanton.[53] The spontaneous activity celebrated by the Zhuangzi and Csikszentmihalyi thus seems to resemble a manifestation of wantonness.

But surely the actors described by the Zhuangzi and Csikszentmihalyi are “in touch” with the mechanisms guiding their behavior; they are anything but “helpless bystanders” to their activities.[54] And having set aside their human essence, they would seem entitled to be judged by the conditions of agency as they apply to animals, in whom reflection and higher-order attitudes aren’t required.[55] So can these actors be wantons and agents simultaneously?

The way to answer this question, I think, is to consider the sense in which these actors are “beyond technique”. They have acquired their skills through training that involved self-scrutiny, self-criticism, and self-correction. The reflective capacity required for this training—the capacity to monitor their own performance, to consider how it falls short of an ideal, and to correct it accordingly—is no longer exercised after they have perfected their skill; but it is merely dormant, ready to be reactivated by the first misstep. Though Woodworker Qing loses himself in his work, forgetting even that he has four limbs, a single false stroke will recall him to self-awareness and re-engage his capacity for self-criticism and self-correction. He is beyond technique not because he has lost it but because his spontaneous activity now accords so well with the technique that it can lie dormant.

Actors in flow have thus achieved a higher wantonness. They act wantonly in the sense that they have dispensed with self-regulation. But they have dispensed with self-regulation only because it has been so effective as to render itself unnecessary. And their capacity for self-regulation remains in reserve in case it is needed. Hence, their wantonness is also a consummate example of agency.