Philosophical Studies

, Volume 168, Issue 2, pp 545–568 | Cite as

Rationalizing flow: agency in skilled unreflective action



In recent work, Peter Railton, Julia Annas, and David Velleman aim to reconcile the phenomenon of “flow”—broadly understood as describing the “unreflective” aspect of skilled action—with one or another familiar conception of agency. While there are important differences between their arguments, Railton, Annas, and Velleman all make, or are committed to, at least one similar pivotal claim. Each argues, directly or indirectly, that agents who perform skilled unreflective actions can, in principle, accurately answer “Anscombean” questions—”what” and “why” questions— about what they do. I argue against this claim and explore the ramifications for theories of skilled action and agency.


Agency Skilled action Flow Expertise Nondeliberative action Self-awareness 

1 Introduction

Branch Rickey—a member of the Baseball Hall of Fame and the manager who helped break the color barrier by signing Jackie Robinson—purportedly once said, “an empty head means a full bat.”1 When trying to understand the genius of top athletes, essayist and novelist David Foster Wallace (2006, p. 155) echoed a similar sentiment:

The real secret behind top athletes’ genius, then, may be as esoteric and obvious and dull and profound as silence itself. The real, many-veiled answer to the question of just what goes through a great player’s mind as he stands at the center of hostile crowdnoise and lines up the free-throw that will decide the game might well be: nothing at all.

Rickey and Wallace do not mean that successful baseball and basketball players are stupid. Far from it.2 Rather, they mean to draw attention to the way in which athletic expertise is “unreflective,” that is, the way in which expert action unfolds in the absence of self-reflective thought or conscious self-awareness. Sometimes this phenomenon seems to be an artifact of the fast speed at which sports are played, but it is not unique to fast action or sports. A skilled pianist, for example, might find the rhythm of a sonata only once she stops thinking about whether she is playing well, or an expert poker player might make the best decisions when she trusts her gut instincts and resists the temptation to scrutinize her actions.
In a related vein, Mihaly Csikszentmihalyi suggests that actions in “flow” are unreflective. Flow characterizes a state of “optimum experience” in which “a person’s body or mind is stretched to its limits” and one becomes “so involved in an activity that nothing else seems to matter” (Csikszentmihalyi 1990, pp. 3–4). Csikszentmihalyi elaborates (1990, p. 155):

In normal life, we keep interrupting what we do with doubts and questions. ‘Why am I doing this? Should I perhaps be doing something else?’ Repeatedly we question the necessity of our actions, and evaluate critically the reasons for carrying them out. But in flow there is no need to reflect, because the action carries us forward as if by magic.

The phenomenon of flow—of being carried forward unreflectively in the performance of a difficult action—is hard to understand from the perspective of common philosophical views about agency and action. It is hard to understand how an agent can actively choose what to do, or engage in practical reasoning about what to do, or endorse what she is doing, for example, if she is not thinking about herself, or even thinking about anything at all, while she acts.

In recent work, Peter Railton (2009), Julia Annas (2011), and David Velleman (2008) aim to reconcile the phenomenon of flow—broadly understood as describing the unreflective aspect of skilled action—with one or another familiar conception of agency. While there are important differences between their arguments, Railton, Annas, and Velleman all make, or are committed to, at least one similar pivotal claim. Each argues, directly or indirectly, that agents who perform skilled unreflective actions can, in principle, accurately answer “Anscombean” questions—“what” and “why” questions— about what they do.3 When an individual accurately answers such questions, she demonstrates that her action, while unreflective, is nevertheless agential, and not rather an unintentional or mindless behavior (Railton), a subrational knack (Annas), or an expression of Frankfurtian “wantonness” (Velleman).4

While there is much to admire in their accounts of skilled unreflective action, Railton, Annas, and Velleman are wrong to think that all agents who perform skilled unreflective actions can, in principle, accurately answer “Anscombean” questions. This is not just to say that Railton, Annas, and Velleman simply misrepresent one incidental characteristic of skilled unreflective action (i.e. that agents who perform skilled unreflective actions can accurately answer Anscombean questions). Rather, these authors take the capacity (in principle) to accurately answer Anscombean questions to be central evidence for skilled unreflective actions being agential. But the cases I will discuss and analyze show that an individual’s capacity to accurately answer Anscombean questions is not good evidence for skilled unreflective actions being agential.

First, I will make some brief stage-setting remarks about skill and unreflective action (Sect. 2). Then I will summarize Railton, Annas, and Velleman’s arguments (Sect. 3). Next, I will argue against the pivotal claim by discussing several cases of agential skilled unreflective actions with respect to which the acting individual cannot, even in principle, accurately answer Anscombean questions (Sect. 4). Finally, in the conclusion, I will advance the speculative suggestion that these cases may in fact be paradigmatic of skilled action (Sect. 5).

2 Preliminary remarks

“Skilled unreflective actions” are actions that (a) involve mastery of a given skill, such as tennis, ballet, or poker; and (b) unfold without the individual who performs them occurrently thinking about what she is doing. I will offer arguments in Sect. 4 that some skilled actions are in fact unreflective in this sense. Furthermore, while I will refer to skilled unreflective actions “in the flow,” I will not focus on the details of Csikszentmihalyi’s account of flow. Rather, I will focus on whether Railton, Annas, and Velleman’s claim about agency in the flow is right. I will draw from empirical literature as well as from the reports of skilled individuals themselves.

My examples will be top-level athletes, musicians, etc. However, examples like these ultimately serve as a device for presentation of a broader set of cases of skilled unreflective action, which are ubiquitous. Consider a banal example, like tying one’s shoes. Shoe-tying involves skill, and it is typically unreflective. This becomes clear by observing children who are clumsily learning to tie their shoes; they have to think hard about what to do because they do not know how to do it.5 Watching a child struggle to tie her shoes can make one feel like an expert in the daily tasks of life. Indeed, expertise is exercised in pedestrian ways like this, and ultimately, the questions Railton, Annas, Velleman, and I are asking about agency in unreflective action apply to banal cases like shoe-tying. But more familiar and dramatic examples of expertise bring the key questions into focus. This is because when athletes, musicians, etc. perform well, what they do represents, in some sense, action at its best. But this is puzzling from the perspective of familiar theories of action and agency, which prioritize agents’ self-reflective capacities. The hope, then, is that by making sense of unreflective expert action, we will make headway in understanding the relationship of action to reflective thought in broader contexts as well.

Throughout, I will refer to “action” and “behavior” interchangeably, unless otherwise noted. Furthermore, when I refer to agential skilled unreflective action, I will mean that the action in question is an expression of agency in some “full-fledged” sense. What it means for an action to express agency in a full-fledged way is different for each of Railton, Annas, and Velleman, as I will discuss in Sect. 3. However, the essential point is that each aims to show that skilled unreflective actions can be agential (in some full-fledged sense) in some cases but not in others, and that evidence for a given skilled unreflective action being agential is the agent’s capacity, in principle, to accurately answer Anscombean questions. Hereafter, I will refer to such actions—for which the individual who performed them can accurately answer “what” and “why” questions about that action—simply as “Anscombean.”6

3 Anscombean questions


‘Automatized’ behavior can be found across the entire span of human activity. The resulting actions, while not anticipated by conscious intention, may nonetheless be intentional, and done for reasons. The jazz saxophonist’s solo riff, the basketball guard’s well-timed jump, the experienced driver’s smooth downshift, and the wit’s lightning riposte aren’t unintentional or mindless behavior, like absent-mindedly tapping one’s foot while writing or succumbing without realizing it to the emotional contagion of a crowd. Rather, they are complex, structured, purposeful activities done mindfully but fluently, without deliberation or intention-formation. Yet were we foolish enough to interrupt these individuals in mid-stream, they could typically answer the Anscombean question, ‘What are you doing?,’ without any further observation or inference (Railton 2009, p. 97).7

The expert pianist plays in a way not dependent on conscious input, but the result is not mindless routine but rather playing infused with and expressing the pianist’s thoughts about the piece… [The expert pianist] can ‘give an account’ of what he does, which involves being able to explain why he is doing what he is doing. Such a person understands what he is doing, unlike the person who can pick up a knack in a purely unintellectual way, without understanding what it is he is doing and why. (Annas 2011, pp. 14, 20)

The reflective capacity required for [the skilled agent’s] 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 merely lies 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 (Velleman 2008, p. 188).

Railton, Annas, and Velleman each recognize that skilled behavior is often unreflective. And each suggests, if indirectly, that skilled unreflective behavior, in some cases at least, is agential. This means that Railton, Annas, and Velleman reject an occurrent property view of agency, according to which skilled action is agential only in case the acting individual occurrently thinks about or consciously monitors what she is doing, when she is doing it. But each ties agency in skilled unreflective action to an individual’s self-reflective thought nevertheless. Railton, Annas, and Velleman do this by appeal to a particular dispositional capacity that agents who perform skilled unreflective actions have, compared with individuals who behave skillfully and unreflectively in some non- or sub-agential way. Railton explicitly describes this capacity as the ability to answer “the Anscombean question.” When a person accurately answers such a question, we are to believe, according to Railton, that the behavior in question was not mindless or unintentional (i.e. that the action was agential).8

Annas and Velleman do not refer to Anscombean questions explicitly, but their claims are closely related, and can be cashed out in these terms (as I show in more detail in Sects. 3.13.3). Annas claims that the expert pianist can explain what she is doing and can give an account of it. This presumably includes the ability to say what she is doing and why, and more importantly, it distinguishes the pianist’s action from the exercise of an unintellectual knack, which Annas elsewhere describes as the product of mere technical abilities which are closed to the guidance of reasons (2011, pp. 19, 25). And Velleman’s claim about the skilled agent’s ability to re-engage with her self-awareness, self-criticism, and self-correction also implicitly encompasses the capacity to accurately answer Anscombean questions, as well as crucially serves as evidence for the skilled individual’s agency. Thus, while there are important differences between their views, Railton, Annas, and Velleman all appear to be committed to the claim that skilled unreflective actions that are agential are Anscombean. Provisionally:

(ANS): all agential skilled unreflective actions are Anscombean.

As I will show in what follows, ANS is entailed by Railton, Annas, and Velleman’s accounts of agency in skilled unreflective action. To be clear, though, ANS is not the claim that all behavioral expressions of agency—actions properly so-called—are Anscombean. ANS pertains to skilled unreflective actions. Nor is ANS the claim that all skilled actions are Anscombean. As I will discuss, Velleman emphasizes that behaviors which are not agential can be skilled and even desirable; and Railton and Annas at least claim that such behaviors can be complex. Nor, finally, is ANS the claim that some agential skilled unreflective actions are Anscombean. Surely this is true. What is untrue, I will argue, is that all skilled unreflective actions that are agential are Anscombean.

3.1 Fluent agents

One aim of Railton’s (2009) account of unreflective skilled action is to stop a vicious—and familiar9—regress in theories of practical reason. The problem is, in short, that “choosing one’s reasons,” or “endorsing certain reasons,” or “identifying a certain reason,” or “throwing one’s weight behind one reason rather than another”—all of which are candidate acts that philosophers have thought to be distinctive of agency—are themselves all full-fledged acts (Railton 2009, p. 103). This regress problem pertains to “any model of action that seeks to understand the distinctive operations of autonomous or rational agency in terms of some special sort of action on the part of the agent” (2009, p. 103). For the regress to be avoided, Railton argues, some psychological process must be identified that enables agents to respond to reasons for action, but not via a further act of choosing, endorsing, etc. He writes (2009, p. 104), “there must exist non-deliberative causal psychic processes ‘of the right kind’ to be aptly responsive to a given consideration as such, and aptly expressive of one’s identity or values, even in the face of competing interests.”

Railton names this non-deliberative causal psychic process “practical competence,” and beings that have practical competence are what he calls “fluent agents.” Athletes, craftsmen and other individuals who act masterfully yet unreflectively exemplify fluent agency. Railton identifies two necessary conditions for having practical competence, and thus being a fluent agent. The first is a set of basic coping skills—such as the ability to focus attention, to make and set goals, to interpret what’s at stake in situations—which Railton calls developmental precursors to and conditions for mature practical rationality (2009, p. 86).10 These skills comprise what Railton calls “agent-competence.” Competent agents know how to do things, although they might not know which things are worth doing. Second, fluent agents must be able to respond to incoming information with the right sort of “gut feelings.” Drawing upon dual-process models in psychology, Railton describes the mind’s automatic and largely affective “triage system,” which, when properly refined by experience, offers individuals a rapid, intuitive sense of what’s happening in a situation and how to respond (2009, p. 93). This affective sense of things is what Railton calls “practical intelligence.”11

Together, agent-competence and practical intelligence enable fluent agents to respond to good reasons for acting without having to deliberate (or think reflectively) about what to do. This responsiveness to good reasons for acting marks fluent agents as “full-fledged” agents.12 Railton writes (2009, p. 98), “why see deliberative agency as the truest form of acting for the sake of reasons? Why not see deliberative agency as one more domain—distinctive and important, but by no means predominant—in which humans can develop greater or lesser skill at responding aptly to reasons?”

But how can we tell the difference between an exercise of fluent agency and “unintentional or mindless behavior?” For one, fluent agents can accurately answer Anscombean questions, as Railton suggests in the quotation above. But this is one of many characteristics Railton uses to describe fluent agents. He also appeals to a number of phenomenological characterizations (2009, p. 105): the fluent agent will not simply “find herself” having done something and wondering why (a phenomenological condition we might call INTELLIGIBILITY); she will have a sense of having selected from various possible actions (without needing to have made a conscious choice between them, which we might call NONDELIBERATIVE FREEDOM); her action will feel that it came from “within” (SELF-DIRECTEDNESS without, as Railton says, “self-centeredness”); and what she does will be consistent with her scheme of values (VALUE-EXPRESSIVENESS).13 On my interpretation, these phenomenological characteristics are what an action being Anscombean ostensibly expresses. When the fluent agent accurately reports what she is doing and why, she indicates having had experiences that plausibly represent her place in the action. For Railton, then, we can say:

ANSR: (1) all agential skilled unreflective actions are Anscombean; and (2) agential skilled unreflective actions are shown to be Anscombean when the report of the agent that performs them represents (some preponderance of) INTELLIGIBILITY, NONDELIBERATIVE FREEDOM, SELF-DIRECTEDNESS, and VALUE-EXPRESSIVENESS.

3.2 Virtue and virtuosos

Annas (2011) argues that virtue is a certain kind of reliable disposition that has been acquired through habituation. My interest here is not in her account of virtue per se (though it is interesting and important), but in her understanding of skill, which forms the basis of her account of virtue. Like Railton, Annas emphasizes that habituation sharpens, rather than blunts, one’s responses to certain situations; she discusses the fast, flexible reactions of expert athletes as well as the courageous acts of skilled soldiers (2011, p. 14). Also, Annas recognizes that skilled action is often unreflective. For the skilled pianist, she writes (2011, p. 29), conscious thoughts “have, in a useful philosophical term, effaced themselves. The pianist is not aware of them at the time, and it would interfere with her performance if she were (as thinking about how to skate and ride a bicycle interferes with actual skating and riding).” Recognizing this puts Annas in a similar quandary as Railton, for she recognizes (2011, p. 19) that only certain unreflective skills provide a model for virtuous actions.14 She must find a way to distinguish unreflective skills that can serve as the basis for a theory of virtue from their reflex-like, unintentional cousins.

To do so, she distinguishes “skills” from mere “subrational knacks.” The skilled person, she argues, as quoted above, can “give an account” of what she does, whereas a person with a mere subrational knack cannot. “If asked how she produced a certain effect,” the pianist whose conscious thoughts have effaced themselves “would have something to say about how it was done; just as she was taught, she can go on to teach others” (2011, p. 29). Giving an account of what one is doing marks the skilled person as having what Annas calls the “need to learn” and the “drive to aspire.” Together these represent the desire to understand oneself, to direct one’s own actions, and to improve what one does; and effectuating these desires requires the ability to articulately convey one’s reasons for action (2011, p. 20). Putative skilled individuals who fail to articulate their reasons for doing what they do are most likely driven by “technical” skills or “natural gifts,” Annas argues, which are on par with subrational knacks. Individuals who perform skilled unreflective actions and who have the need to learn and the drive to aspire are good models for virtue, then, because they are full-fledged agents, engaged in efforts to understand and improve themselves. Annas’ version of ANS, therefore, is:

ANSA: (1) all agential skilled unreflective actions are Anscombean; and (2) agential skilled unreflective actions are shown to be Anscombean when the report of the agent that performs them represents the agent as having the need to learn and the drive to aspire.

3.3 Higher wantons

Velleman (2008) proposes a phenomenological interpretation of Frankfurt’s account of identification. On Velleman’s interpretation of Frankfurt, full-fledged actions are characterized by the agent having a sense of being “in touch” with the mechanisms guiding her behavior (2008, p. 180). Velleman characterizes this sense of being in touch as a solution to a problem that being conscious of our motives creates. While non-human animals might act on the basis of their desires without reflective hesitation, the problem for human beings is that “consciousness just seems to open a gulf between subject and object, even when its object is the subject himself” (2008, p. 179). On this view, non-human animals are by default in touch their motives. A person, by contrast, “is continually falling out of touch with his motives, by becoming reflectively conscious of them” (2008, p. 180). Identification is the hard-won solution to the problem of continually falling out of touch with our motives. But, Velleman asks, why not simply aim to become so engrossed in our activities that we eschew the problem of self-reflective consciousness altogether? When we “lose ourselves” in an activity—losing consciousness of ourselves, at least temporarily, in a craft or a sport—the phenomenological gulf between subject and “object” (i.e. oneself) disappears. The problem, however, is that in taking this shortcut around the problem of self-reflective consciousness we risk becoming what Frankfurt calls “wantons,” or beings indifferent to which of our first-order desires effectively guides our behavior. When losing ourselves in an activity, we don’t want to become like the dog that sees a passing car and just has to chase it.15 So Velleman’s question is: what distinguishes the wanton, who can surely do complex, skilled and even intelligent things, from the expert in flow, who dispenses with reflective self-awareness but not in such a way that threatens her status as an agent?

Velleman labels skilled agents—in particular those who achieve the Daoist ideal of wu wei, or “effortless action”—“higher wantons.” Higher wantons dispense with deliberative self-regulation, but not in such a way as to threaten their status as full-fledged agents. Velleman writes (2008, p. 188):

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.16

Higher wantons are “beyond technique” (2008, p. 183), that is, beyond the “techniques” of agency—such as “self-scrutiny, self-criticism, and self-correction” (2008, p. 187). But these techniques merely lay “dormant,” ready to be activated when needed. When the higher wanton reactivates her techniques of agency, and thus returns to self-awareness, she stands poised, ceteris paribus, to answer Anscombean questions. Of course, Velleman does not discuss Anscombean questions as such. But the techniques of agency Velleman describes are precisely those cognitive capacities for which the ability to accurately answer Anscombean questions is thought to be evidence.17 Furthermore, it is not just that the higher wanton, but not the wanton simpliciter, could report what she is doing in some generic way (e.g. “I am going to the bar,” “because I want to get drunk,” etc.), but that she could do so under a certain description that reflects her antecedent acts of self-scrutiny, self-criticism, etc. Therefore:

ANSV: (1) all agential skilled unreflective actions are Anscombean; and (2) agential skilled unreflective actions are shown to be Anscombean when the report of the agent that performs them represents her as having re-engaged with the techniques of agency which she antecedently used in the acquisition and development of her unreflective skills.

4 Against ANS

Is ANS (and its variants) true? Athletes and craftsmen often say that when they are performing well they are not “thinking.”18 Here is one example, from cricket player Ken Barrington: “when you’re playing well you don’t think about anything and run-making comes naturally. When you’re out of form you’re conscious of needing to do things right, so you have to think first and act second. To make runs under those conditions is mighty difficult.”19 There are many analogous examples.20 This apparent lack of occurrent thought in skilled action is, of course, the very phenomenon that Railton, Annas, and Velleman are out to explain. Does this putative mindlessness also mean that some agential skilled unreflective actions are not Anscombean?

At the very least, experts frequently say that they don’t know what they’re doing or why. Hall of Fame NFL running back Walter Payton said, “people ask me about this move or that move, but I don’t know why I did something. I just did it.”21 Kimberly Kim, the youngest person ever to win the US Women’s Amateur Golf Tournament, said, “I don’t know how I did it. I just hit the ball and it went good.”22 Larry Bird, the great Boston Celtic, purported said, “[A lot of the] things I do on the court are just reactions to situations… A lot of times, I’ve passed the basketball and not realized I’ve passed it until a moment or so later.”23 Payton, Kim, and Bird are reporting what Beilock and Carr (2001) call “expertise-induced amnesia.” These experts don’t seem to have episodic memories of what led them to perform well, even immediately after acting.24

Do statements like those from Payton, Kim and Bird count as counterexamples to ANS? One reason to think so follows from a principle of charity: presumably athletes like Payton, Kim, and Bird are being honest about their experiences. If they are, then the explanatory burden is on those who wish to show that these individuals are wrong about what they say they don’t know. Furthermore, there is no reason to think that ANS must be true in order for athletes to succeed in the midst of the game or for artists to excel while creating art. Payton’s posture (e.g. a slight shift in his hips, a lowering of his shoulder) might tell the defender trying to tackle him what Payton is going to do next, even if Payton himself doesn’t know it and the defender can’t consciously identify the postural cues to which he’s responding. Neither player’s action needs to be Anscombean in order for them to do it well. Of course, sometimes it is valuable for players to articulate why they do what they do, for example when they are trying to teach others how to play. But this is a claim about coaching, not playing. Indeed, there is a general assumption within professional sports that the best players make bad coaches.25

Perhaps, however, ANS is vindicated by the ostensible fact that individuals like Payton, Kim, and Bird can say that they are “trying to get a first down” or even just that they are “playing basketball,” even if they can’t say exactly how or why they are doing these things. This idea should be resisted. Course-grained statements like “I was trying to get a first down” do not vindicate ANS. Such statements show that an action was directed at (excuse the pun) a goal, but ANSR, for starters, requires that an individual’s report count as evidence for experiences like INTELLIGIBILITY, etc. Course-grained reports do not count as evidence of such experiences; for example, they do not indicate that the individual had any sense of having selected from various possible actions. In fact, they might not count as evidence for any particular experiences. Payton might answer a “what” question by saying “I was trying to get a first down” just because he’s making a reasonable inference about his intentions after the fact. Course-grained statements do not vindicate ANSA either. A statement like “I was trying to get a first down” is no “account” of one’s action, if, as Annas argues, giving an account of one’s action represents one as having the need to learn and the drive to aspire. For example, what could course-grained statements like these teach to others? Finally, such statements do not vindicate ANSV. They don’t demonstrate anything about the individual’s recall of her self-scrutinizing acts of training. The skilled individual’s training dealt with when and how and why to speed up or slow down or juke and fake or just lower your shoulder and push. A total novice could say “I was trying to get a first down.”26

Taking stock, when asked Anscombean questions, some individuals fail to answer them. Further, what these individuals actually say—course-grained statements about what they were doing—fail to vindicate ANS. What could vindicate ANS, given these considerations? In what follows, I’ll analyze two kinds of arguments. The first has to do with the role psychologists believe self-awareness plays in skilled action (Sect. 4.1). Perhaps a form of self-awareness is thought to be necessary for skilled action, and perhaps this suggests that all exercises of agential skilled unreflective action are Anscombean. The second has to do with the “self-talk” in which skilled agents are sometimes found to be engaged while performing (Sect. 4.2). Perhaps the things agents say to themselves—like “watch the ball!’’—render ANS true.

Unfortunately (for ANS), neither of these possibilities bears out. After showing why, I’ll complete my discussion of the evidence against ANS by explaining why some skilled unreflective actions that are not Anscombean are nevertheless agential.

4.1 Skill and self-awareness

What is the relationship between self-awareness and skilled action? Is there reason to think that some kind of self-awareness is necessary for the performance of skilled action, such that agents who act unreflectively would nevertheless be able to report the contents of this awareness, thus vindicating ANS?

Put simply, no. The literature in sports psychology suggests that self-awareness is often detrimental to expert performance. This suggests that some expert athletes, at least, will not be sufficiently self-aware to accurately answer Anscombean questions. Were they sufficiently self-aware, they wouldn’t be capable of expert performances.

The way in which self-awareness can interfere with expert performance is known anecdotally as “Steve Blass Disease.”27 Steve Blass was a hugely successful Major League Baseball player who inexplicably lost his ability to pitch. Within the span of 3 years he went from winning 20 games in a season and pitching Game 7 of the World Series to playing in the minor leagues, and shortly after that to being out of professional baseball altogether. It wasn’t just that Blass lost his edge. Suddenly and inexplicably, he couldn’t even hit the catcher’s mitt from the mound. There was nothing physically wrong with Blass, so far as anyone could tell. The common interpretation is that Blass’ problem stemmed from something like chronic over-thinking. This is a common enough “disease” in profession sports to earn its own name.28

Controlled studies suggest that Steve Blass Disease is an extreme form of a common problem. The problem is that performing skilled actions well stands in an inverse relationship to self-focused awareness. For example, in an article tellingly titled, “Overthinking skilled motor performance: Or why those who teach can’t do,” Flegal and Anderson (2008) showed that highly skilled golfers needed twice as many attempts to sink a putt after describing their previous putts than controls who had simply spent time on an unrelated task between putts. Beilock and Gray (2012) similarly show that experts’ performance is largely unaffected by distraction, but is harmed by skill-focused attention, whereas novices perform worse when they are distracted and perform better when they attend to what they are doing. As Beilock et al. (2003, p. 310) put it: “It is as if experts cannot pay enough attention to remember as well as novices, at least when performing under routine conditions so close to what they have practiced in the past.” On this view, experts ordinarily utilize proceduralized or “chunked” “action programs,” that is, action-control processes that automatize the step-by-step details regarding what to do. When an expert attends to those ordinarily automated action programs, her performance suffers. In general, the more proceduralized a skill-set, the less decomposable it will be from the agent’s perspective.29

However, as Montero (2010, pp. 106–108) points out, there are different ways in which one can be self-aware while acting. One can be aware of one’s body through one’s senses (“sensory bodily awareness”), an awareness which can in turn be driven by either “top-down” endogenous attention—e.g. when a painter intentionally focuses on the feel of the brush in her hand—or by “bottom-up” exogenous attention—e.g. when a painter’s attention suddenly turns to her fingers because the handle of the brush slips in her hand.30 By contrast, one can be aware of one’s body in a more “cognitive” way (“cognitive bodily awareness”), by thinking propositional thoughts about where your body should be or what it should be doing. Furthermore, one can be self-aware in all of these ways either peripherally or focally.

The inverse relationship between skilled action and self-awareness is not found in all of these forms of bodily awareness equally. The empirical literature demonstrating the detrimental effects of self-awareness on skilled action deals with top-down cognitive bodily awareness that occupies focal attention. Tellingly, this is precisely the form of self-awareness to which a self-report must testify in order to vindicate at least ANSA and ANSV. A report of sensory bodily awareness (e.g. “I shifted left because a sudden feeling made me look that way”) does not make either of these variants of ANS true.31 On a more general level, this is because a report of sensory bodily awareness simply describes how performing a particular action felt; it does not indicate what the agent was doing or why. More specifically, Annas’ skilled agent must report something that amounts to an account of what one does. Only a report of what Montero calls cognitive bodily awareness can comprise the kind of account of one’s actions that Annas needs, because this is the only kind of report that is sufficiently abstract and teachable to others. Second, a report of sensory bodily awareness does not vindicate ANSV because ANSV requires that the agent say something indicative of having self-scrutinizing, self-criticizing or self-correcting thoughts or attitudes. Only a report of cognitive bodily awareness contains the right kind of representation of the self, of the one who is doing the scrutinizing, etc., needed to make ANSV true.

4.2 Instructional nudges

Experts often talk to themselves while performing—giving themselves encouragement or instructions—and some might take this self-talk as evidence that ANS is true. Consider the example of what Sutton (2007) calls “instructional nudges.” These are usually short phrases like “watch the ball!” or “keep your feet moving!” Instructional nudges are interesting because they appear to link the agent’s more reflective acts of training to their performance when in the flow. Sutton, for example, argues that instructional nudges act like “cognitive short-cuts,” reminding the agent of what she learned in practice and giving her a little added flexibility in the flow of activity. He claims (Sutton et al. 2011, p. 93) that instructional nudges are evidence that expert athletes are “thoughtful” even in moments of flow. “Expertise,” he writes, “is in part the training up of the right indirect links between thought and action, not the evacuation of thought from action.”32

But how exactly do instructional nudges work? Sutton isn’t sure, but he suggests that perhaps they are tactics for automatizing the links between memory (e.g. of what one practiced during training) and action (2007, p. 776). This seems right, but to which memories do instructional nudges automatize the links? Are they memories of antecedently formed intentions, like “I will watch the ball?” Or are they memories of particular feelings in the body, e.g. the way one’s body feels when hitting the ball the right way? Or another possibility: are the automatized links between memory and action created by instructional nudges mere associations between a cue (e.g. a pitch tailing into the strike zone) and a response (swing!)?

Empirical evidence speaks against at least one of these possibilities, namely that instructional nudges automatize links between antecedently formed intentions and action. In the game of cricket, at least, it turns out that the instructional nudge “watch the ball” doesn’t actually cause players to watch the ball. Instead, it leads batters to look ahead to where they predict the ball will be. So the instruction “watch the ball” doesn’t cause the batter to act in virtue of the literal meaning of the words watch-the-ball. It is not like taking a deep breath when the doctor says “take a deep breath.” The instruction works instrumentally, causing the batter to do something that promotes the attainment of her goal (i.e. to hit the ball), but not necessarily by doing what she thinks she’s doing. Even if the batter feels that she has successfully fulfilled her intention to watch the ball, she didn’t actually watch the ball!33 If asked, she might say, “I hit that one well because I was really watching the ball.” Here we would have a case where the individual’s answer to an Anscombean question isn’t too coarse-grained, but is simply false. Not that this would make her action deficient in any way; it’s possible that instructional nudges work just because they distract the individual from the kind of detrimental self-awareness that Beilock describes.

This example is not anomalous. It is an instance of a broader and common pattern of confabulation in skilled action. Confabulation is common when skilled agents are asked to report the techniques they use “on the field.” For instance, Jeannerod (2006) shows that baseball players who successfully catch objects falling at an accelerating rate report that those objects are falling at a constant speed. These individuals are confabulating reasons—based in naïve physics—when they are asked why they are moving to the spot where the accelerating object will fall.34

It is also noteworthy that pretty much anything can serve as an instructional nudge. Dancers’ practice of “marking” their routines and tennis players’ grunting when they hit the ball are ways of nudging oneself to act in a particular way.35 In each of these cases, the nudge does seem to provide a link between what the agent learned during practice and what the agent does while performing. But there is no reason to think that in all cases the nature of the link is to remind one of one’s reflective intentions regarding what to do or of one’s reasons for doing something one way or the other. That is, there is no reason to think that the link instructional nudges represent between practice and performance vindicates ANS. It is an open and interesting question how instructional nudges do in fact work, in particular whether they automatize links between bodily feelings and actions or whether they create associations between cues and responses, or through some other mechanism. Note that the success of either of these alternate explanations would not bode well for ANS. If instructional nudges guide behavior by instantiating associations between cues and responses, it becomes hard to maintain the claim that those responses count as full-fledged actions.36 And as I argued in (Sect. 4.1), reports of bodily feelings (i.e. “sensory bodily awareness”) don’t vindicate ANS.37

4.3 Agency without Anscombean answers

For ANS to be shown false, there must be at least some cases of agential skilled unreflective action which are not Anscombean. ANS would not be threatened, in other words, if the cases I have put forward are not really cases of action at all, but manifestations of some sort of sub-agential processes. Velleman raises this possibility in suggesting that the actions of the wanton simpliciter can be complex and skillful. More strikingly, Velleman suggests that it is sometimes desirable to leave off being an agent and cede control of one’s behavior to the guidance of one’s “motivational nature” (2007, p. 213). Drawing upon his claim that the constitutive aim of action is to do what “makes sense,” Velleman writes (2009, 138fn), “there are occasions when trying to do what makes sense, makes no sense at all. On such occasions, the rational thing to do is to leave off being an agent for a while.” Is this what happens in the flow?

I do not think so. The cases of skilled unreflective action I have discussed are not instances of individuals leaving off being an agent for a while, regardless whether doing so is desirable. One reason I say this is because of the way in which we typically describe skilled unreflective action. We describe athletes’ play, for example, with person-level adjectives, such as courageous or timid, creative or mechanical, joyful or uninspired. This suggests that there is a prima facie sense in which skilled unreflective actions really are actions, and not expressions of an agent’s mere motivational nature (or of some equivalent sub-agential features of agents). In what follows, I will offer two additional reasons for thinking that some non-Anscombean skilled unreflective actions are agential. In (Sect. 4.3.1) I will discuss recent neuroscientific research on improvisational musicians which suggests that some skilled unreflective actions exemplify important characteristics of agency without being consciously self-monitored in the way that one would expect of Anscombean actions. In (Sect. 4.3.2) I will argue that some non-Anscombean skilled unreflective actions are “self-expressive,” which is to say, they disclose elements of who an agent is.

4.3.1 Improvisational music and neural architecture

It has not been my aim in this paper to give an account of skilled unreflective action (i.e. a scientific explanation of the mechanisms underlying skilled unreflective action). However, it is important that my arguments about agency in skilled unreflective action be empirically plausible and consistent with contemporary science. And indeed, neuroscientific evidence lends support to there being genuine counterexamples to ANS.38

Especially striking are the results of several recent neuroimaging studies of expert improvisational musicians [Limb and Braun (2008); Bengtsson et al. (2007); Berkowitz and Ansari (2008)]. Limb and Braun (2008), for example, used fMRI to compare the neural activity of professional jazz pianists while they improvised to the pianists’ neural activity while performing over-learned sequences of notes. Limb and Braun found a fascinating pattern. Improvisational performance appears to be characterized by widespread deactivation of the lateral portions of the prefrontal cortex (LOFC and DLPFC) and focal activation of the medial prefrontal cortex (MPFC). This is significant, Limb and Braun argue (2008, p. 4), because the MPFC “plays a role in the neural instantiation of self, organizing internally motivated, self-generated, and stimulus-independent behaviors;” the MPFC is associated with higher behavioral goals and “maintaining an overriding set of intentions while executing a series of diverse behavioral subroutines.” These are qualities associated with agential action. However, the areas which Limb and Braun found significantly deactivated during improvisation—the lateral portions of the PFC—are “thought to provide a cognitive framework within which goal-directed behaviors are consciously monitored, evaluated and corrected… [and which supports] self-monitoring and focused attention” (4). From these results, the authors conclude:

Musical creativity vis-à-vis improvisation may be a result of the combination of intentional, internally generated self-expression (MPFC-mediated) with the suspension of self-monitoring and related processes (LOFC- and DLPFC-mediated) that typically regulate conscious control of goal-directed, predictable, or planned actions (4–5).

It is not yet clear whether other sorts of skilled agents would demonstrate similar patterns. Pragmatic considerations might prevent putting tennis players in fMRI machines! But I strongly suspect that Limb and Braun’s findings are common to other cases of creative improvisational action. The key point, in any case, is not the finding that the areas of the brain associated with conscious monitoring and self-focused attention are deactivated during improvisation. That supports what Railton, Annas, and Velleman say (namely, that skilled behavior can be unreflective). Rather, the key point is that Limb and Braun did not report increased activation in the reflexive habit control centers of the brain, but rather in the areas of the brain associated with internally-motivated, stimulus-independent, intentional behaviors—in short, in the areas of the brain associated with agential action. This is presumably why they call the MPFC the “neural instantiation of the self.” Perhaps this phrase should not be taken too literally. Nevertheless, these findings lend support to the claim that some skilled unreflective actions are neither reflexively habitual nor Anscombean but are agential nevertheless.

4.3.2 Self-expression

Perhaps there are reasons to be skeptical about strong claims from functional neural localization. Imagine, then, a tennis player who plays fearlessly but genuinely thinks of herself as a defensive player. What should we say about when and how her behavior expresses something about who she is? Is she really a fearless person or really a defensive person? In a case of conflict like this between her behavior and her self-understanding, which should we take to represent her?39

There are a number of interpretive options: (a) a report-taking view, according to which she is defensive because what she avows expresses who she is, in which case she will have left the guidance of her behavior to (something like) her “motivational nature,” which is not self-expressive; (b) an action-taking view, according to which she is fearless because her actions express who she is, in which case what she avows represents some kind of reflective illusion; (c) a neither/nor view, according to which she is neither fearless nor defensive, perhaps because she is in a state of contradiction, or because she is somewhere in-between fearlessness and defensiveness; and finally, (d) a context-dependent view, according to which she is both fearless and defensive, but in different contexts.40

ANS requires that (b), the action-taking view, and (d), the context-dependent view, be false. If (a), the report-taking view, is correct, then the player’s reflectively disavowed fearless play will count as an instance of skilled unreflective behavior which is not Anscombean (presuming her disavowal counts as a failure to accurately answer Anscombean questions) and not self-expressive (i.e. agential). ANS countenances this. ANS also countenances (c), the neither/nor view, because according to this view the player is neither fearless nor defensive, so there is no possibility for the player’s behavior to be self-expressive in spite of what she avows. According to (d), the context-dependent view, however, her behavior could be self-expressive when she is on the court, regardless whether she can accurately answer Anscombean questions about that particular bit of behavior. And according to (b), the action-taking view, her behavior is self-expressive even if it conflicts with what she reports. Both (b) and (d) allow for cases of agential skilled unreflective action that are not Anscombean.

Why think that (b) or (d) is correct? Above I noted that we typically describe athletes’ play with person-level adjectives. This gives prima facie support to (b) and (d). It is also worth noting that the particular adjectives we tend to use to describe skilled unreflective actions are terms associated with the evaluation of character and even virtue (e.g. courageous, creative, inspired), rather than with mere “grading” evaluations, in Smart’s (1961) sense.41 Why do we do this? For one, how an individual behaves when in the flow is generally not one-off; agents exhibit patterns of skilled unreflective action. These patterns manifest in perception and attention, as well as in action. The creative tennis player will reliably see opportunities to hit an unexpected shot, for example; and the inspired woodworker will predictably attend to an interesting grain in a piece of wood.42 These patterns of perception, attention, and action can be thought of as representing what an agent cares about.43 Caring in this sense is not the same as explicit belief about what one ought to do or evaluative judgment of what one has reasons to do. But caring in this unreflective sense is self-expressive nevertheless. It represents just those elements of who we are that persist regardless whether we endorse them, like being forgetful, or generous to a fault, or even being implicitly biased.44 These are all examples of what one cares about, in the sense that being forgetful, etc. are manifestations of patterns of perception, attention, and action. Such patterns are self-expressive.

Another reason to think that non-Anscombean skilled unreflective actions are self-expressive is that agents in these contexts are not “alienated” from what they do. The tennis player who plays one way but describes herself in another, for example, may feel no alienation from her behavior. Alienation involves more than a simple conflict between what one says and what one does. For one, alienation involves negative affect.45 But it is plausible that the tennis player in this case experiences no negative affect when her play is discordant with her self-understanding. This may be because she is simply unaware of the discord between her avowal and her play, or because she does not find it disturbing or problematic.

These considerations may not add up to the claim that how one behaves “on the field” has the final say about who one is. But it is a problem for ANS if skilled unreflective actions that are not Anscombean are self-expressive in any meaningful sense. For if ANS is true, skilled unreflective actions that are not Anscombean are akin to mindless or unintentional behaviors, subrational knacks, or wanton behaviors that, while complex and maybe even desirable, nevertheless stem from a person’s sub-agential motivational nature. Such mere behaviors are in no way self-expressive. But ANS is false, and many skilled unreflective actions that are not Anscombean are in fact self-expressive.

5 Conclusion

Railton, Annas, and Velleman recognize that skillful unreflective action does not fit neatly with traditional conceptions of agency, and each sheds distinct and important light on this phenomenon. I have argued, however, that each subscribes to the view that all agential skilled unreflective actions are Anscombean (ANS), and that this view is false.46 I have presented counterexamples and have rejected interpretations of these counterexamples that would make variants of ANS true for Railton, Annas, and Velleman. I then offered arguments for thinking that some skilled unreflective actions that are not Anscombean are nevertheless agential. Briefly now, in conclusion, I would like to make a more speculative suggestion, which is that paradigmatic agential skilled unreflective actions are non-Anscombean.

In the essay from which I quoted earlier, David Foster Wallace expresses consternation over how “great athletes usually turn out to be stunningly inarticulate about just those qualities and experiences that constitute their fascination” (2006, p. 152). He settles on the following explanation: great athletes are great because they can “shut off the Iago-like voice of the self.” Wallace’s claim is stronger that Railton, Annas, and Velleman’s. While they agree that expert action is in many cases unreflective, Wallace’s claim is that expertise is achieved at least in part because the expert acts unreflectively. In this sense, Wallace’s claim parallels the empirical literature in sports psychology that I discussed earlier. Both suggest that reflective awareness is detrimental to skilled action. However, Wallace goes further still. The last sentence of his essay reads,

…those who receive and act out of the gift of athletic genius must, perforce, be blind and dumb about it—and not because blindness and dumbness are the price of the gift, but because they are its essence (2006, p. 155, emphasis added).

“Blindness and dumbness” are not stupidity, as Wallace notes.47 They are, rather, a “gift” with benefits in some contexts—e.g. when trying to achieve flow on the court or the field—and consequences in others—e.g. when penning an autobiography about what one did when one was in the flow. This gift, Wallace argues, represents the “essence” of athletic expertise. Athletic geniuses must be blind to what they are doing, in some way that ordinary agents are not. It is their persistent inability to hear the “Iago-like voice of the self”—that is, to switch into a self-aware, self-critical mode—that enables them to perform as they do. While Wallace suggests this as an explanation for the vacuity of many sports autobiographies, it equally suggests that skilled agents might paradigmatically fail to accurately answer Anscombean questions.

Wallace’s claim is, of course, speculative. It is a question for future research whether it can be vindicated. One avenue for such future research is to consider why any skilled unreflective actions are Anscombean. What makes those skilled agents distinct who can accurately report what they are doing and why? Is it a capacity that stems from their character? Is it a characteristic of particular kinds of skilled unreflective actions? Is it an effect of context? This line of questioning inverts the philosophical strategy most philosophers of action employ. Rather than take self-reflective action as the norm and unreflective action as the exception, future research should consider the possibility that the truly mysterious cases are those skilled actions in which the agent knows what she’s doing and why.48


  1. 1.

    Reported in Hagen (2011).

  2. 2.

    Wallace writes (2006, p. 153), “Anyone who buys the idea that great athletes are dim should have a close look at an NFL playbook, or at a basketball coach’s diagram of a 3–2 zone trap…”.

  3. 3.

    Railton (2009), Annas (2011), and Velleman (2008) are not concerned with the actual views of Elizabeth Anscombe. Nor should anything I say be interpreted as commentary on her views.

  4. 4.

    In arguing this way, Railton, Annas, and Velleman join a broader debate about the role of self-awareness in skilled action which has been ongoing in phenomenology and related fields. See, for instance, Brownstein (2010); Dow (ms); Dreyfus (2002a, b, 2005, 2007a, b); Gallagher (2005); Gottlieb (2012); McDowell (2007a, b); Montero (2010); Rietveld (2010); Schear (2013); and Sutton (2007). Note also that Railton (2009), Annas (2011), and Velleman’s (2008) goals are each broader than I have space here to address. I will thus be honing in on one particular claim each makes. Finally, note that the larger goal of understanding agency in unreflective action is also Frankfurt’s, who writes (2002, p. 90): “I am not so sure that deliberation and practical reasoning do play fundamental roles in our agency, if this means that they are essential to our capacity to function as agents. No doubt they figure prominently in our selection of ends and in our design of plans for reaching our goals. It is far from clear, however, that they are indispensable conditions of action.”

  5. 5.

    I do not have space to address the relationship between “knowing-how” and “knowing-that” in this paper, although what I say here may have ramifications for recent debates between “neo-Ryleans” and “Intellectualists.”

  6. 6.

    I am grateful to an anonymous reviewer for Philosophical Studies for suggesting this term.

  7. 7.

    Note that Railton includes “the wit’s lightning riposte” as an example of a skilled unreflective action (as I’m calling it). On Railton’s account, all action—including deliberative “mental” action—has nondeliberative skill (as he calls it) at its core. I discuss this aspect of Railton’s view in Brownstein (ms). See also Arpaly and Schroeder (2012).

  8. 8.

    While Railton says that agents who perform skilled unreflective actions can “typically” say what they are doing, he clearly intends an action being Anscombean to count as evidence for the action being agential. So cases of non-Anscombean yet agential skilled unreflective action will not just be exceptions to a claim about what’s typical, but will call into question what Railton takes to be evidence for showing that skilled unreflective actions are agential.

  9. 9.

    See Ryle (1949/2009).

  10. 10.

    Railton likens agent-competence to the skills needed for being an effective administrator. Just as an effective administrator can guide her organization to irrational ends, “rational driving is a distinctive way of deploying one’s driving competence and rational discourse and social conduct are distinctive ways of deploying one’s linguistic and social competences” (2009, p. 86). Rational action is one way of deploying agent-competence (2009, p. 86). While deciding to steal may indicate a failure in one’s rational competence, stealing ineptly indicates a failure in one’s agent-competence (2009, p. 87). Railton attributes weakness of will, indecision, failures to adjust means to ends, and other ordinary irrationalities to failures in agent-competence. (The reasons there are to be a competent agent might then count for what reasons there are to be rational. On the debate over whether there are reasons to be rational, see Railton (1997); Schroeder (2004); Kolodny (2005); Raz (2005); and Broome (2005, 2007).)

  11. 11.

    Practical intelligence is not excellence in practical reasoning (Railton 2009, p. 90). In describing practical intelligence, Railton stresses the primacy of affect in our immediate responses to the world. He writes, “Those with highly developed practical skills—in sports, seamanship, teaching, etc.—appear to draw upon this evaluation system, as refined by experience, in forming a quick intuitive ‘sense’ of what is happening in a given situation, and how to respond appropriately” (2009, p. 93). See also Railton (2012).

  12. 12.

    There is some ambiguity in Railton’s claim that fluent agency is “full-fledged.” He writes (2009, p. 103) that the regress problem can only be stopped if there are “ways in which individuals could come to embrace one reason over others autonomously, but not via a further ‘full fledged’ act.” I think what Railton means is that exercises of fluent agency are not “full-fledged” in the sense that fluent agents don’t choose, endorse, etc. their reasons for action reflectively; it seems that Railton identifies full-fledged acts, properly so-called, with reflective acts, in other words. However, exercises of fluent agency are indeed full-fledged in the sense that fluent agents act on the basis of reasons and do so autonomously.

  13. 13.

    Railton does not present these as necessary or sufficient conditions for fluent agency. Rather, he presents them as considerations that collectively distinguish agential skilled unreflective action from unintentional or mindless behavior. This is why I add “some preponderance of” in the formulation of ANSR (see below).

  14. 14.

    This recognition cuts against one criticism of her virtue-as-skill account (see Jacobson 2005), for Annas need not claim that all skilled actions are good models for virtuous actions. All Annas needs is a claim about some instances of skilled action. However, she endorses ANS insofar as she claims that all individuals whose skilled unreflective actions can serve as a model for virtue can give an account of what they do.

  15. 15.

    I am indebted to Taylor Carman for this metaphor.

  16. 16.

    While Velleman says that higher wantons dispense with self-regulation, I think he means that they dispense with deliberate or deliberative self-regulation. Formally speaking, in psychology, self-regulation is just the general process of managing cognition, emotion, and behavior. That said, it is usually (and often tacitly) assumed in the psychological literature that self-regulation is for managing emotion and behavior in the service of one’s reflective ideals or avowed goals. For discussion, see Fujita (2011).

  17. 17.

    Perhaps there is an interpretation of Velleman on which the higher wanton might still not have the capacity in principle to accurately answer Anscombean questions even after her techniques of agency have been reactivated. This strikes me as a coherent interpretation, though not a very convincing one, for what would self-awareness, self-criticism, etc. get you if you still couldn’t report what you were doing? If true, though, this interpretation would get Velleman off the hook for ANS.

  18. 18.

    What athletes et al. mean when they use the word “thinking” in this context is, of course, not clear. See footnote 32.

  19. 19.

    Quoted in Sutton (2007, p. 767).

  20. 20.

    For example, Yogi Berra purportedly said, “Think? How can you hit and think at the same time?” (quoted in Sutton et al. 2011, p. 80). Barbara Montero (2010, p. 106) quotes choreographer George Balanchine as saying, “don’t think, dear; just do.” See Sutton (2007) and Montero (2010) for illuminating discussions of statements like these.

  21. 21.

    Quoted in Beilock (2010, p. 224). Here is another example from football which exemplifies what John McDowell might call a “demonstrative concept.” In ‘The Art of the Pass,’ Sports Illustrated reporter Tim Layden (2010) has the following exchange with NFL quarterback Phillip Rivers: “Rivers wiggles the ball in his right hand, fingers across the laces as if ready to throw. He purses his lips, because this isn’t easy to articulate. ‘You always want to pick a target,’ he says. ‘Like the chin [of the receiver]. But on some routes I’m throwing at the back of the helmet. A lot of it is just a natural feel.’ Rivers strides forward with his left leg, brings the ball up to his right ear and then pauses in midthrow. ‘There are times,’ he says, ‘when I’m seeing how I’m going to throw it as I’m moving my arm. There’s a lot happening at the time. Exactly where you’re going to put it is still being determined.’ Even as it’s leaving the fingertips? More head-shaking and silence. Finally: ‘I don’t know,’ says Rivers. ‘Like I said, there’s a lot going on.’”

  22. 22.

    Quoted in Beilock (2010, p. 231), who follows the quote by saying, “[Athletes like Kim] can’t tell you what they did because they don’t know themselves and end up thanking God or their mothers instead. Because these athletes operate at their best when they are not thinking about every step of performance, they find it difficult to get back inside their own heads to reflect on what they just did.”

  23. 23.

    Quoted in Levine (1988).

  24. 24.

    Perhaps Payton/Kim/Bird-type reports are non-Anscombean because the report has come well after the agent has acted. Perhaps richer reports would be given if agents were really interrupted in mid-performance. I am doubtful about this. Research on memory in skilled action suggests that time-lags between performance and report are relatively unimportant. See Beilock and Carr (2001) and Sutton (2007) for a review of research on memory in skilled action, and also Anderson (2003) and Stone et al. (2010) on “retrieval-induced forgetting.” Note also: it is possible that individuals form episodic memories while performing skilled unreflective actions, but that the content of these memories is, for some reason, unavailable for report. See Sect. 4.1 for related discussion.

  25. 25.

    Pace Annas (2011, p. 29), who assumes that it is in the nature of skilled agents to be able to teach their skills to others. Professional hockey player Thérèse Brisson, for example, said, “Recently retired hockey players who played at high levels rarely make the ideal coaches for youth hockey. They know what to do, but they can’t communicate how to do it!” Quoted in Beilock (2010, p. 225).

  26. 26.

    Thanks to Alex Madva for making this point about novices.

  27. 27.

    For an eloquent discussion and interview with Steve Blass, see the radio program “This American Life.”

  28. 28.

    Dreyfus (2007a) and Montero (2010) both discuss the case of Chuck Knoblauch, an All-Star baseball player who suffered a similar fate as Blass. Montero is skeptical of Dreyfus’ interpretation of the Knobuauch case. She makes a distinction between the putative detrimental effects of attention on expert action and the (more plausible, in her opinion) detrimental effects of attention on highly automatized, everyday skills.

  29. 29.

    See Beilock and Gray (2012); Brown and Carr (1989); and Keele and Summers (1976).

  30. 30.

    Montero (2010) argues that proprioception is a form of sensory bodily awareness. She argues against claims made by O’Shaughnessy (1998) and Gallagher (2003) that agents are usually unaware of proprioceptive information from their bodies. Montero argues (2010, p. 113), for example, that ballet dancer Britt Juleen experiences awareness of proprioception when she describes the aim of performance to be “feeling totally immersed [sic] the feeling of my body moving.”

  31. 31.

    I am unsure whether a report of sensory bodily awareness could vindicate ANSR. Instead of choosing or endorsing a reason for action, the fluent agent, Railton argues, is “attuned to the reasons he faces” (2009, p. 106). Attunement is not quite a technical term for Railton, but it is close. Perhaps a report of a bodily feeling could count as evidence of attunement to one’s reasons for action.

  32. 32.

    Sutton is critiquing Dreyfus, who argues that expert athletes do not think at all during performance. Much of Sutton’s claim against Dreyfus turns on what one means by “thought.” Dreyfus claims (2007a, b) that the content of experience in flow is nothing more than attractive and repellent forces (attracting one into certain courses of action and repelling one from others). When he says this, I think his aim is to show that experts in flow need not have explicit or implicit propositional thoughts about what they’re doing. This is what counts as “thought” for Dreyfus. See Dreyfus (2002b).

  33. 33.

    Abernathy (1981). I suspect this is the case in other ball sports as well. An anonymous reviewer for Philosophical Studies suggests that a batter who successfully predicts where the ball will be, based on having watched the ball during some part of its flight path, could count as having watched the ball. This seems right, if the question is whether the batter has watched the ball at any point in time. But when a batter tells herself to watch the ball, I think she is telling herself to watch the ball at and through the point of contact. (In tennis at least, this is the way I was taught.) She is telling herself to look where the ball is, in other words, not elsewhere, even if elsewhere turns out to be where the ball will be. In this case, when the batter successfully predicts where the ball will be, she does not do so in virtue of having watched the ball in the sense that she meant it.

  34. 34.

    See also Fourneret and Jeannerod (1998); Jones (1988); and Marcel (2003a, b).

  35. 35.

    Dancers often “mark” their routines by modeling moves in a less-than-complete way, often by sequences of hand gestures that represent the individual moves of the routine. The interesting question about marking is what it adds to a dancer’s training and practice that merely imagining oneself performing the moves cannot do. See Kirsch (2010) for discussion. He argues that marking is likely either a way of anchoring projection to the targeted behavior or a form of self-imposed priming that increases the vividness of the dancer’s imagination.

  36. 36.

    Although I think there is an important difference between mere associative links and the sorts of associative mental states that might underlie skilled action. See Brownstein (ms) for discussion.

  37. 37.

    Although see footnote 31 for a qualification regarding ANSR.

  38. 38.

    For an overview of the neuroscience of flow, see Bruya (2010). For example, in Bruya (2010), Dietrich and Stroll (2010) describe a process of what they call “phenomenological subtraction,” in which there is a reduction of the specific contents of conscious experience when an agent is in flow due to down-regulation of the functional networks in the prefrontal cortex during strenuous activity. On neural localizations of skilled vs. unskilled action, see also Milton et al. (2007) and Milner and Goodale (1995, 2008).

  39. 39.

    This question could be cashed out in terms of character. Does her style of play express her character or does her self-understanding? I hesitate to use the term “character,” however, given its pride of place in recent debates about the moral psychology of virtue ethics.

  40. 40.

    I am drawing from interpretive options for belief-attribution in cases of apparent belief-behavior discord. See Schwitzgebel (2010) and Brownstein and Madva (2012a, b) for discussion.

  41. 41.

    Although I am trepidatious about using the term “character” (see footnote 39), it is worth noting that John Wooden, the most successful coach in the history of college basketball, purportedly said, “sports do not build character. They reveal it.” See also Lieberman and Eisenberger (2004) for an empirically oriented proposal that habits inform an “intuitive self” which is distinct from one’s “self-conception.” The intuitive self, they argue, is not reliant on the behavioral memories provided by episodic memory.

  42. 42.

    See, for instance, Gabbett and Abernathy (2013) for evidence of “perceptual expertise” in rugby players.

  43. 43.

    In calling patterns of perception, attention, and action representative of what an agent cares about, I am drawing upon Shoemaker (2011), who argues that agents can be “attribution-responsible” for actions even if they cannot answer for those actions (i.e. even if they are not “answerability-responsible”). I think Shoemaker’s account of responsibility for actions that conflict with one’s evaluative judgments could be made to show that individuals are also (attribution-) responsible for their skilled unreflective actions, but I do not have space to explore this question here, For discussion, see Brownstein (under review) Note also: Shoemaker’s sense of “care” is not Frankfurt’s. Another project for future work is to explore the relationship between these two different senses of what it is to care about something.

  44. 44.

    See Brownstein (under review) for my account of self-expression and responsibility in the context of implicit bias.

  45. 45.

    See Christman (2011, pp. 195–196) for argument.

  46. 46.

    Although see footnotes 17 and 31 for caveats.

  47. 47.

    See footnote 2.

  48. 48.

    Thanks to Alex Madva for pushing me to see this conclusion, in particular these questions for future research.



For much needed help and feedback on this paper, many thanks to Nobel Ang, Michael Brent, John Christman, James Dow, Katie Gasdaglis, Tamar Gendler, Daniel Kelly, Eliot Michaelson, James Rocha and Steven Weimer. Special thanks to an anonymous reviewer for Philosophical Studies for cogent, detailed, and invaluable suggestions, and to Alex Madva for reading multiple drafts of this paper and for continuing to be an exemplary interlocutor and friend. I am also deeply grateful for all the helpful comments and questions I received after presenting parts of this paper at the 2012 Midsouth Philosophy Conference, the 2012 LSU Philosophy Conference, and the 2012 Eastern Division meeting of the American Philosophical Association.


  1. Abernethy, B. (1981). Mechanisms of skill in cricket batting. Australian Journal of Sports Medicine, 13(1), 3–10.Google Scholar
  2. Anderson, M. (2003). Rethinking interference theory: Executive control and the mechanisms of forgetting. Journal of Memory and Language, 49, 415–445.CrossRefGoogle Scholar
  3. Annas, J. (2011). Intelligent virtue. New York: Oxford University Press.CrossRefGoogle Scholar
  4. Arpaly, N., & Schroeder, T. (2012). Deliberation and acting for reasons. Philosophical Review, 121(2), 209–239.CrossRefGoogle Scholar
  5. Beilock, S. (2010). Choke: What the secrets of the brain reveal about getting right when you have to. New York: Free Press.Google Scholar
  6. Beilock, S., & Carr, T. (2001). On the fragility of skilled performance: What governs choking under pressure? Journal of Experimental Social Psychology: General, 130(4), 701–725.Google Scholar
  7. Beilock, S., & Gray, R. (2012). From attentional control to attentional spillover: A skill-level investigation of attention, kinematic, and performance outcomes. Human Movement Science, 31, 1473–1499.Google Scholar
  8. Beilock, S. L., Wierenga, S. A., & Carr, T. H. (2003). Memory and expertise: What do experienced athletes remember? In J. L. Starkes & K. A. Ericsson (Eds.), Expert performance in sports: Advances in research on sport expertise (pp. 295–320). Champaign: Human Kinetics.Google Scholar
  9. Bengtsson, S., Csikszentmihalyi, M., & Ullén, F. (2007). Cortical regions involved in the generation of musical structures during improvisation in pianists. Journal of Cognitive Neuroscience, 19(5), 830–842.CrossRefGoogle Scholar
  10. Berkowitz, A., & Ansari, D. (2008). Generation of novel motor sequences: The neural correlates of musical improvisation. Neuroimage, 41, 535–543.CrossRefGoogle Scholar
  11. Broome, J. (2005). Does rationality give us reasons? Philosophical Issues, 15, 321–337.CrossRefGoogle Scholar
  12. Broome, J. (2007). Wide or narrow scope? Mind, 116, 359–370.CrossRefGoogle Scholar
  13. Brown, T., & Carr, T. (1989). Automaticity in skill acquisition: Mechanisms for reducing interference in concurrent performance. Journal of Experimental Psychology: Human Perception and Performance, 15, 686–700.Google Scholar
  14. Brownstein, M. (2010). Conceptuality and Practical Action: A Critique of Charles Taylor’s Verstehen Social Theory. Philosophy of the Social Sciences, 40(1), 59–83.CrossRefGoogle Scholar
  15. Brownstein, M., & Madva, A. (2012a). Ethical Automaticity. Philosophy of the Social Sciences, 42(1), 67–97.CrossRefGoogle Scholar
  16. Brownstein, M., & Madva, A. (2012b). The Normativity of Automaticity. Mind and Language, 27(4), 410–434.CrossRefGoogle Scholar
  17. Brownstein, M. (manuscript). Nondeliberative action.Google Scholar
  18. Brownstein, M. (under review). Responsibilit(ies) and implicit bias.Google Scholar
  19. Bruya, B. (Ed.). (2010). Effortless attention: A new perspective in the cognitive science of attention and action. Cambridge, MA: MIT Press.Google Scholar
  20. Christman, J. (2011). The politics of persons: Individual autonomy and socio-historical selves. Cambridge: Cambridge University Press.Google Scholar
  21. Csikszentmihalyi, M. (1990). Flow: The psychology of optimal experience. New York: Harper and Row.Google Scholar
  22. Dietrich, A., & Stroll, O. (2010). Effortless attention, hypofrontality, and perfectionism. In B. Bruya (Ed.), Effortless attention: A new perspective in the cognitive science of attention and action (pp. 159–179). Cambridge, MA: MIT Press.CrossRefGoogle Scholar
  23. Dow, J. Manuscript. Just doing what I do: Expert action, reflection, and self-ascription.Google Scholar
  24. Dreyfus, H. (2002a). Intelligence without representation: The relevance of phenomenology to scientific explanation. Phenomenology and the Cognitive Sciences, 1(4), 367–383.CrossRefGoogle Scholar
  25. Dreyfus, H. (2002b). Refocusing the question: Can there be skillful coping without propositional representations or brain representations? Phenomenology and the Cognitive Sciences, 1, 413–425.CrossRefGoogle Scholar
  26. Dreyfus, H. (2005). Overcoming the myth of the mental: How philosophers can profit from the phenomenology of everyday experience. Proceedings and Addresses of the American Philosophical Association, 79(2), 43–49.Google Scholar
  27. Dreyfus, H. (2007a). The return of the myth of the mental. Inquiry, 50(4), 352–365.CrossRefGoogle Scholar
  28. Dreyfus, H. (2007b). Response to McDowell. Inquiry, 50(4), 371–377.CrossRefGoogle Scholar
  29. Flegal, K., & Anderson, M. (2008). Overthinking skilled motor performance: Or why those who teach can’t do. Psychonomic Bulletin & Review, 15(5), 927–932.CrossRefGoogle Scholar
  30. Fourneret, P., & Jeannerod, M. (1998). Limited conscious monitoring of motor performance in normal subjects. Neuropsychologia, 36, 1133–1140.CrossRefGoogle Scholar
  31. Frankfurt, H. (2002). Reply to Michael E. Bratman. In S. Buss (Ed.), Contours of agency: Essays on themes from Harry Frankfurt, 85–90. Cambridge, MA: MIT Press.Google Scholar
  32. Fujita, K. (2011). On conceptualizing self-control as more than the effortful inhibition of impulses. Personality and Social Psychology Review, 15(4), 352–366.CrossRefGoogle Scholar
  33. Gabbett, T. & Abernathy, B. (2013). Expert-novice differences in anticipatory skill of rugby league players. Sport, Exercise, and Performance Psychology. doi:10.1037/a0031221.
  34. Gallagher, S. (2003). Bodily self-awareness and object perception. Theoria et Historia Scientiarum, 7, 53–68.Google Scholar
  35. Gallagher, S. (2005). How the body shapes the mind. Oxford: Oxford University Press.CrossRefGoogle Scholar
  36. Gottlieb, G. (2012). Unreflective action and the argument from speed. Pacific Philosophical Quarterly, 92(3), 338–362.CrossRefGoogle Scholar
  37. Hagen, P. (2011). Manuel’s lineups varied and successful. Retrieved September 29 2011 from
  38. Jacobson, D. (2005). Seeing by feeling: Virtues, skills, and moral perception. Ethical Theory and Moral Practice, 8, 387–409.CrossRefGoogle Scholar
  39. Jeannerod, M. (2006). Motor cognition: What actions tell to the self. Oxford: Oxford University Press.CrossRefGoogle Scholar
  40. Jones, L. (1988). Motor illusions: What do they reveal about proprioception? Psychological Bulletin, 103, 72–86.CrossRefGoogle Scholar
  41. Keele, S., & Summers, J. (1976). The structure of motor programs. In G. E. Stelmach (Ed.), Motor control: Issues and trends (pp. 109–142). New York: Academic Press.Google Scholar
  42. Kirsch, D. 2010. Thinking with the body. The annual meeting of the cognitive science society. Accessed 13 May 2013.
  43. Kolodny, N. (2005). Why be rational? Mind, 114, 509–563.CrossRefGoogle Scholar
  44. Layden, T. (2010). The art of the pass. Sports Illustrated. Accessed 13 May 2013.
  45. Levine, L. (1988). Bird: The making of an American sports legend. New York: McGraw Hill.Google Scholar
  46. Lieberman, M. D., & Eisenberger, N. I. (2004). Conflict and habit: A social cognitive neuroscience approach to the self. In A. Tesser, J. V. Wood, & D. A. Stapel (Eds.), On building, defending and regulating the self: A psychological perspective (pp. 77–102). New York: Psychology Press.Google Scholar
  47. Limb, C., & Braun, A. (2008). Neural substrates of spontaneous musical performance: An fMRI study of jazz improvisation. PLoS One, 3(2), e1679.CrossRefGoogle Scholar
  48. Marcel, A. (2003a). The sense of agency: Awareness and ownership of action. In J. Roessler & N. Eilan (Eds.), Agency and self-awareness (pp. 48–93). New York: Oxford University Press.Google Scholar
  49. Marcel, A. (2003b). Introspective report: Trust, self-knowledge and science. Journal of Consciousness Studies, 10(9–10), 167–186.Google Scholar
  50. McDowell, J. (2007a). What myth? Inquiry, 50(4), 338–351.CrossRefGoogle Scholar
  51. McDowell, J. (2007b). Response to Dreyfus. Inquiry, 50(4), 366–370.CrossRefGoogle Scholar
  52. Milner, A. D., & Goodale, M. A. (1995). The visual brain in action. Oxford: Oxford University Press.Google Scholar
  53. Milner, A. D., & Goodale, M. A. (2008). Two vision systems re-viewed. Neuropsychologia, 46, 774–785.CrossRefGoogle Scholar
  54. Milton, J., Solodkin, A., Hluštík, P., & Small, S. (2007). The mind of expert motor performance is cool and focused. Neuroimage, 35, 804–813.CrossRefGoogle Scholar
  55. Montero, B. (2010). Does bodily awareness interfere with highly skilled movement? Inquiry, 53(2), 105–122.CrossRefGoogle Scholar
  56. O’Shaughnessy, B. (1998). Proprioception in the body image. In J. Bermúdez, A. Marcel, & N. Eilan (Eds.), The body and the self (pp. 175–204). Cambridge, MA: MIT Press.Google Scholar
  57. Railton, P. (1997). On the hypothetical and non-hypothetical in reasoning about belief and action. In G. Cullity & B. Gaut (Eds.), Ethics and practical reason (pp. 293–321). Oxford: Clarendon Press.Google Scholar
  58. Railton, P. (2009). Practical competence and fluent agency. In D. Sobel & S. Wall (Eds.), Reasons for action (pp. 81–115). Cambridge: Cambridge University Press.CrossRefGoogle Scholar
  59. Railton, P. (2012). Two cheers for virtue, or, might virtue be habit forming? In M. Timmons (Ed.), Oxford studies in normative ethics volume 1. Oxford: Oxford University Press.Google Scholar
  60. Raz, J. (2005). The myth of instrumental rationality. Journal of Ethics and Social Philosophy, 1(1), 1–28.Google Scholar
  61. Rietveld, E. (2010). McDowell and Dreyfus on unreflective action. Inquiry, 53(2), 183–207.CrossRefGoogle Scholar
  62. Ryle, G. (1949/2009). The concept of mind. New York: Routledge.Google Scholar
  63. Schear, J. (Ed.). (2013). Mind, reason and being-in-the-world: The McDowell–Dreyfus debate. New York: Routledge.Google Scholar
  64. Schroeder, M. (2004). The scope of instrumental reason. Philosophical Perspectives, 18, 337–364.CrossRefGoogle Scholar
  65. Schwitzgebel, E. (2010). Acting contrary to our professed beliefs, or the gulf between occurrent judgment and dispositional belief. Pacific Philosophical Quarterly, 91, 531–553.CrossRefGoogle Scholar
  66. Shoemaker, D. (2011). Attributability, answerability, and accountability: Towards a wider theory of moral responsibility. Ethics, 121, 602–632.CrossRefGoogle Scholar
  67. Smart, J. J. C. (1961). Free-will, praise, and blame. Mind, 70(279), 291–306.CrossRefGoogle Scholar
  68. Stone, C., Barnier, A., Sutton, J., & Hirst, W. (2010). Building consensus about the past: Schema consistency and convergence in socially-shared retrieval induced forgetting. Memory, 18(2), 170–184.CrossRefGoogle Scholar
  69. Sutton, J. (2007). Batting, habit and memory: The embodied mind and the nature of skill. Sport in Society, 10(5), 763–786.CrossRefGoogle Scholar
  70. Sutton, J., McIlwain, D., Christensen, W., & Geeves, A. (2011). Applying intelligence to the reflexes: Embodied skill and habits between Dreyfus and Descartes. Journal of the British Society for Phenomenology, 42(1), 78–103.Google Scholar
  71. Velleman, J. D. (2007). What good is a will? In A. Leist & H. Baumann (Eds.), Action in context. Berlin: De Gruyter/Mouton.Google Scholar
  72. Velleman, J. D. (2008). The way of the wanton. In C. Mackenzie & K. Atkins (Eds.), Practical identity and narrative agency. New York: Routledge.Google Scholar
  73. Velleman, J. D. (2009). How we get along. Cambridge: Cambridge University Press.Google Scholar
  74. Wallace, D. F. (2006). Consider the lobster and other essays. New York: Little, Brown and Co.Google Scholar

Copyright information

© Springer Science+Business Media Dordrecht 2013

Authors and Affiliations

  1. 1.Department of HumanitiesNew Jersey Institute of TechnologyNewarkUSA

Personalised recommendations