2.1. Function, correctness, and excellence

Let me start with a few general remarks about capacities. Epistemic capacities are those we exercise or employ in forming and revising beliefs (i.e., in making judgments).Footnote ³ To be a believer just is to possess epistemic capacities. As a foil for epistemic capacities, I will use the capacity to play chess. This capacity is exercised in making chess moves, and to be a chess player just is to possess such a chess-playing capacity. Capacities in general are defined by their function (point, aim)—roughly: by what they enable their possessors to do when exercised successfully. Plausibly, the function of a chess-playing capacity is to generate moves that promote checkmating one’s opponent. This is what such a capacity enables one to do when employed successfully. Identifying the function of an epistemic capacity is bound to be more controversial. For present purposes, I will assume that the function of an epistemic capacity is to generate true beliefs. But not much hangs on this assumption. The main argument of this paper should go through even if one assumes a different function (e.g., knowledge or understanding).Footnote ⁴

What is important is the idea that a capacity’s function fixes a fundamental standard of success or correctness, and that we can evaluate possessors of a capacity (as such) in terms of how well they exercise their capacities relative to this fundamental standard. Thus, the function of an epistemic capacity fixes believing truly as a standard of correctness, such that we can evaluate possessors of this capacity (i.e., believers) in terms of how well they form and revise beliefs relative to this standard. The standard of correctness for chess moves fixed by the capacity to play chess is to make moves that promote winning. Accordingly, as I use the term, a “correct chess move” is not just a legal move—i.e., one permitted by the rules of chess—but rather one that advances the game’s defining objective of checkmating one’s opponent.

Moreover, it is important to emphasize that capacities come in degrees. Both Magnus Carlsen and I possess the capacity to play chess, but his capacity is vastly more developed than mine. He is far more skilled or competent at playing chess. As I understand them, terms such as “skill,” “competence,” or “virtue” refer to certain stages in the development of a capacity. I will use the term “excellence” as an umbrella notion encompassing them all.

To a first approximation, we can say that to excel as a possessor of a certain capacity is, at least in part, a matter of being disposed to succeed reliably enough (i.e., to meet the relevant standard of correctness) when one exercises that capacity under appropriately normal conditions.Footnote ⁵ An excellent chess-player, then, is one who has developed their chess-playing capacity to the point that, when they exercise this capacity (i.e., play chess), they tend to do so successfully (i.e., make correct moves). Similarly, an excellent believer is, in part, someone who has sufficiently developed their epistemic capacity so that, when they exercise this capacity (i.e., make a judgment), they tend to do so successfully (i.e., make a correct judgment). (The qualifier “in part” will become important later.)

One final clarification: as I use the term, “excellence” does not necessarily denote rarity or an exceptional achievement. Achieving some degree of excellence is often a natural stage in the life cycle of a capacity. For instance, I take it that all mature believers possess at least some degree of epistemic excellence.

2.2. What is it to exercise a capacity well?

With these brief clarifications in place, we can now turn to the question of what it is to exercise a capacity well. As I see it, this is not merely a matter of conforming to the relevant standard of correctness. The manner in which agents attain—or fail to attain—correctness also matters for the quality of their exercises. A more illuminating approach is to understand the quality of an exercise in terms of how well the agent responds to considerations that bear on how to meet the relevant standard of correctness. Let me explain.

For any given capacity, its standard of correctness is associated with a range of correctness-relevant considerations—considerations that bear, positively or negatively, on whether an exercise of that capacity would meet its standard. In chess, such considerations are those that bear on whether a certain move would promote winning. For example, the fact that advancing one’s rook to c4 would capture an opponent’s bishop bears positively on whether performing this move would promote winning the game. In the epistemic case, correctness-relevant considerations are those that bear on whether believing P would be true (i.e., evidence for or against P). For example, the fact that a reliable source has testified that P bears positively on whether P is true.

The basic idea, then, is that we can evaluate chess-players and believers in terms of how well they respond to such correctness-relevant considerations when they exercise their respective capacities—by selecting chess-moves and forming or revising beliefs. More precisely, we can evaluate their exercises along at least three dimensions, asking:Footnote ⁶

1. A. Whether they are responding correctly—i.e., whether their response results in a correct move or judgment.

2. B. Whether they are responding excellently—i.e., whether their response (chess-move or judgment) manifests a sufficient degree of excellence, such that not too easily would they respond incorrectly.

3. C. Whether they are responding correctly through excellence—i.e., whether they attain correctness as a manifestation of their excellence at chess or judgment.

Correspondingly, one can fall short of performing fully well in at least four distinct ways:

1. i. One responds neither correctly nor excellently (neither A nor B)—e.g., through lack of excellence, one makes an incorrect move or judgment.

2. ii. One responds correctly but not excellently (A without B)—e.g., one happens to make a correct move or judgment by mere (good) luck.

3. iii. One responds excellently but not correctly (B without A)—e.g., one carefully considers the correctness-relevant factors of one’s situation but is (faultlessly) deprived of crucial information about what the correct move or judgment would be (bad luck).

4. iv. One responds both correctly and excellently, but not correctly because excellently (A and B without C)—i.e., Gettier-style cases, where correctness and excellence combine accidentally, through a combination of bad and good luck.

These evaluations of a chess-player’s and a believer’s performances run in parallel. In the next section, I will begin highlighting an important structural difference between the evaluative structure of epistemic capacities and that of capacities in sports, arts, or games. The best way to get this difference into view is to shift our focus slightly. Up to this point, we have considered only how capacity-bearers are evaluated when they exercise their capacities—i.e., when they play chess or make judgments. To highlight the distinctive character of epistemic capacities, we should now consider cases in which agents do not exercise their capacities. My suggestion will be that epistemic capacities differ from capacities in sports, arts, and games in how such non-exercises bear on their possessor’s level of excellence.

3.1. Excellence at chess

Suppose you are playing chess and have an excellent opportunity to capture your opponent’s queen. Suppose further, however, that instead of making the move, you simply abandon the game. Does this refusal to make the correct move reveal a lack of skillfulness at chess on your part? Not necessarily. You may simply have lost interest in the game and prefer to do something else. A lack of motivation to continue playing, by itself, does not count against your skillfulness at chess; you might still be an exceptionally skilled chess player.

This suggests that not exercising one’s chess-playing capacity—even on occasions where success would be easily attainable—is fully compatible with being a skilled or competent chess player. What we expect of a skilled chess player, simply qua skilled player, is only that they succeed—i.e., make correct moves—when they exercise their chess-playing capacity. But we do not necessarily expect them to exercise that capacity in the first place.Footnote ⁷

Moreover, this feature is hardly unique to chess; it plausibly characterizes capacities in sports, arts, and games more generally. Imagine, for example, a pianist who refrains from exercising their capacity to play the piano, even though they are in an excellent position to do so well. This alone need not demonstrate any shortcoming of musical skill; they may still be among the finest pianists alive. Consider the case of Glenn Gould: he famously withdrew from public concerts at the height of his career, declaring that he “detest[ed] audiences.” (Gould, Reference Gould1966) Yet his refusal to perform is surely fully compatible with him being one of the most skilled pianists of his time (or ever).

To be clear, a chess player’s or a pianist’s refusal to perform may, of course, reveal other shortcomings. For instance, a concert pianist who refuses to perform out of caprice may fall short of professional standards. But the crucial point is that such non-exercises of their respective capacities do not, by themselves, imply any deficit in skillfulness at chess or piano playing. At any given occasion, a chess player or musician may always refrain from employing their capacities without thereby demonstrating a deficiency in the relevant form of excellence.

I think the deeper reason for this is that the exercise of capacities in sports, arts, and games is ultimately conditional on the presence of a separate motivation (e.g., a desire) to engage in the relevant domain. This motivation is “separate” in the sense that excellence in these domains does not require possessing it, no matter how skilled or competent one is. Being skilled in the exercise of a capacity and being disposed to exercise it are separate matters and can therefore come apart. Some players may be highly skilled at chess yet often lack any inclination to play; other players may be highly motivated to play but lack skill.

Note, moreover, that nothing said here conflicts with the familiar thought that the acquisition and retention of skill typically require extensive practice.Footnote ⁸ If a pianist were to stop practicing altogether, they might eventually lose their musical skill. This, however, does not show that a disposition to practice is itself constitutive of such skill. Rather, it shows only that skill is temporally fragile: absent continued exercise, it may fade. An agent may therefore count as highly skilled or competent at a time despite lacking any inclination to engage in the relevant activity at that time. A loss of motivation may explain a subsequent decline in skillfulness, but it does not itself constitute such a decline.

3.2. Epistemic excellence

Let us now turn to epistemic capacities and consider how they differ from the capacities to play chess or the piano. What I want to suggest is that, in contrast to skilled chess players or musician, what we expect from an excellent believer is not only that they tend to succeed when they exercise their epistemic capacities, but also that they exercise those capacities in response to certain opportunities for epistemic success (e.g., opportunities to believe truly). In other words: there are occasions for epistemic success such that, if one did not exercise one’s epistemic capacities (i.e., make a judgment) on that occasion, this would imply a shortcoming or malfunction of these capacities—independently of whether one had a desire to know.

The clearest examples of such occasions are ones in which one has excellent (sufficient, undefeated) evidence for P readily available. Intuitively, if one were in such a situation but did not respond by making a judgment on whether P, this would constitute a shortcoming of one’s epistemic capacities—a deficit of epistemic excellence. Consider a couple of examples for illustration:

Plausibly, in these cases, Mary and John exhibit a shortcoming of epistemic excellence precisely by failing to believe what their evidence supports—namely, that Mary’s husband is guilty and that John’s proof is flawed. Their epistemic failings consist not (or not only) in exercising their epistemic capacities poorly (e.g., by forming a false belief), but in failing to exercise those capacities at all—by failing to form a belief for which they possess excellent evidence. Moreover, it seems clear that both subjects are epistemically criticizable even though they may lack any motivation to comply with the epistemic standard of correctness on these occasions. In Mary’s case especially, what explains her resistance to the evidence is plausibly a desire to avoid knowing the truth about her husband. Thus, we can easily imagine that, on some level, she fears that her husband is indeed guilty but accepting this would cause her great psychological discomfort. By not believing what her evidence supports, she protects herself against such discomfort. But this lack of motivation to find out the truth obviously does not absolve her from a negative assessment of her epistemic capacities. The ignorance of her husband’s guilt is epistemically objectionable.

The lesson I want to draw from examples like these is that there is a deep and relevant difference between epistemic capacities and capacities in domains of sports, arts, and games. In general, we do not expect skilled agents to engage their chess-playing capacities unless they possess some independent motivation to do so (e.g., a desire to play and win), but, simply qua skilled player, they need not possess any such motivation. It is only when, and for as long as, the relevant motivation is present that skilled players engage in responding to considerations that bear on whether a given chess-move would be correct (“correct” in the sense of promoting winning).

By contrast, what we expect of an excellent believer is that they respond to their evidence simply because it is the evidence—not just when and because they satisfy some further motivational condition, such as possessing a desire to find out the truth. This difference in expectation, I suggest, reflects a distinctive feature of epistemic capacities. To excel as a possessor of epistemic capacities requires not only a disposition to succeed—i.e., to judge correctly—when one exercises these capacities, but also a disposition to exercise those capacities in the first place when one’s evidential situation calls for it. An excellent believer is therefore not merely someone who tends to get it right when they respond to their evidence, but someone who is reliably responsive to their evidence to begin with. They possess a readiness to engage in judgment when their evidence calls for it.Footnote ⁹

Of course, it is difficult to specify precisely when one’s evidential situation calls for a judgment. Any such specification will depend on substantive and controversial commitments in epistemology. Different accounts of evidence and evidence-possession—such as internalist and externalist views—will yield different verdicts about when not making judgments reflects a genuine deficiency of epistemic excellence.Footnote ¹⁰ For present purposes, I want to remain largely neutral on these issues. What matters for the contrast between epistemic capacities and capacities in domains of sports, arts, and games is that possessors of the former can display a shortcoming of excellence simply by not exercising their epistemic capacities, regardless of whether they have a desire to know the truth.Footnote ¹¹

3.3. Summary of the difference

We can now summarize the relevant contrast between epistemic capacities and capacities in domains of sports, arts, and games by focusing on the kind of responsiveness to standards of correctness that excellence in each domain requires.

Consider first the capacity to play chess. Not exercising one’s chess-playing capacity in response to opportunities for success—such as opportunities to make a winning move or to outplay an opponent—does not, by itself, constitute a shortcoming or malfunction of that capacity. An agent may possess the relevant skill yet refrain from responding to such opportunities simply because they lack an independent motivation to perform. In the absence of such motivation, not performing a correct move does not reflect poorly on the agent’s chess-playing excellence. Accordingly, excellence in chess requires only conditional responsiveness to the considerations that bear on correctness in that domain—conditional, that is, on the agent’s being motivated to engage their chess-playing capacity in the first place.

Epistemic capacities differ in this respect. As we have seen, not exercising one’s epistemic capacities in response to certain opportunities for epistemic success can itself constitute a shortcoming of epistemic excellence. Crucially, this is so even in the absence of any independent motivation to find out the truth. This difference is reflected in the kind of responsiveness that epistemic excellence requires. Unlike excellence in chess, epistemic excellence demands unconditional responsiveness to the considerations that bear on the relevant standard of correctness—i.e., one’s evidence. An excellent believer is not only disposed to judge correctly when they judge but also reliably disposed to engage their epistemic capacities when their evidential situation calls for it. In this sense, epistemic excellence involves a standing readiness to respond to evidence as such, rather than a responsiveness that is activated only given the presence of some further motivational condition.

4.1. The challenge

Let us begin by recalling the core thesis:

As note before, CC-E is a thesis about the sources of epistemic normativity—i.e., about what makes it the case that one epistemically ought to believe something, or that some consideration is an epistemic reason for believing it. There are various ways of developing this form of epistemic constitutivism. A relatively straightforward—albeit somewhat simplified—way of spelling it out as an account of epistemic reasons is the one introduced in the opening section:

On this view, epistemic reasons are precisely those considerations to which believers respond when they exercise their epistemic capacities well. Since well-functioning epistemic capacities are responsive to correctness-relevant considerations, and since correctness in the epistemic domain is truth, the view delivers the familiar and plausible result that epistemic reasons are considerations bearing on whether a belief is true—i.e., evidence for and against P.Footnote ¹⁴ In this way, CC-E promises both a grounding story of epistemic reasons and a vindication of the evidentialist intuition that epistemic reasons are truth-directed.Footnote ¹⁵

However, constitutivist accounts of normativity face a well-known challenge. Critics have argued that constitutive functions, by themselves, are insufficient to ground genuine normativity, including the normativity of epistemic reasons (e.g., Côté-Bouchard, Reference Côté-Bouchard2016; Enoch, Reference Enoch2006, Reference Enoch and Brady2011).Footnote ¹⁶ Applied to CC-E, the challenge can be put as follows. Many capacities have constitutive functions yet do not seem to generate genuine normative reasons—e.g., the capacities to play chess, knit, or throw darts. Thus, for example, the fact that moving one’s knight would promote checkmating does not, by itself, appear to provide one with a genuine normative reason to perform that move. By contrast, CC-E maintains that epistemic capacities do precisely that—give rise to genuine normative reasons. What explains this asymmetry? What is special about epistemic capacities, such that their constitutive function grounds normative reasons, while the constitutive functions of these other capacities do not? Proponents of CC-E owe us an explanation here.

It is important to distinguish this challenge from the familiar “shmagency” objection. As the objection is typically understood, it grants that agents or believers may have constitutive functions, but disputes that such functions can ground genuine reasons.Footnote ¹⁷ Applied to CC-E, the complaint is that even if well-functioning believers respond to truth-relevant considerations, those considerations do not thereby constitute genuine reasons unless we already have a reason to be well-functioning believers in the first place. In effect, then, the claim is that proponents of CC-E—and constitutivists more generally—face a dilemma: either there is a prior reason to be a well-functioning believer, or there is not. If there is no such reason, then there are likewise no reasons to exercise one’s epistemic capacities well by forming beliefs in ways that respond to considerations bearing on the standard of correctness set by the function of those capacities. If there is such a prior reason, however, then well-functioning epistemic capacities cannot be explanatorily fundamental for epistemic normativity. Any normative force possessed by correctness-relevant considerations would instead be derived from the antecedent reason to be a believer. Either way, CC-E would be in serious trouble.

I think constitutivists should respond by rejecting the presupposition underlying this dilemma—that grounding reasons in constitutive functions requires a prior reason to be an agent or believer. As others have argued, absent further assumptions it is far from obvious why such a reason is needed for the constitutivist grounding story to work.Footnote ¹⁸ The constitutivist’s question is why some considerations, but not others, constitute reasons for particular actions or beliefs. Their answer, very roughly, is that these are the considerations to which well-functioning agents and believers respond. One might, of course, raise various questions about the details of this story. But it is unclear why its success should necessarily depend on whether, in addition to these particular reasons, there is also a further, more general reason for being an agent or believer.Footnote ¹⁹ Perhaps there is no special reason to be an agent or believer—why should this imply that we cannot have any particular reasons to perform specific actions or believe particular propositions?Footnote ²⁰ Of course, one might also harbor more general doubts about whether grounding reasons in non-normative facts is possible at all. But that is a different objection, and not one that specifically targets constitutivism.Footnote ²¹

The present challenge, however, is distinct from the “shmagency” objection. It does not turn on the assumption that the constitutivist grounding story requires a further reason to be a well-functioning believer. Rather, it begins with the observation that many capacities with constitutive functions do not give rise to genuine reasons. If so, it is fair to ask what makes epistemic capacities special: why do they give rise to genuine normativity while other capacities presumably do not? Answering this question need not involve positing any special reason to be a well-functioning believer. Instead, proponent of CC-E should identify an internal difference between these capacities and argue that this difference plausibly carries normative significance. This is the challenge I have in mind.Footnote ²²

One difficulty in addressing the challenge is the lack of clear and widely accepted criteria for what it takes for something to count as genuinely normative—a difficulty reflected in the frequent but elusive appeal to “normative oomphiness” in discussions of normative authority. Still, one prominent—though not uncontroversial—feature often taken to mark epistemic reasons as genuinely normative, and to align them with moral reasons in this respect, is their categorical force (e.g., Kelly, Reference Kelly2003):

I focus on Categoricity as the relevant mark of genuine normativity. The central question, then, is whether CC-E can explain why considerations that bear on the standard of correctness fixed by our epistemic capacities possess this kind of categorical force—i.e., a force that does not depend on an agent’s contingent aims, interests, or desires. I will argue that the preceding reflections on the distinctive nature of epistemic capacities place CC-E in a position to vindicate the categorical force of these considerations. In this way, the account of epistemic capacities helps address the challenge of explaining why epistemic capacities can ground genuine normativity whereas other capacities cannot.

4.2. Toward answering the challenge

To see how this works, return once more to CC-E’s account of epistemic reasons:

Whether this account can vindicate Categoricity depends crucially on how we understand the relevant notion of capacity at work here. If the capacity invoked in Epistemic Reasons is modeled on capacities familiar from sports, arts, and games, it is difficult to see how CC-E could secure categorical force. To bring this out, consider again the case of chess. Whether a chess player’s lack of responsiveness to correctness-relevant considerations—such as opportunities to make a winning move—counts as a shortcoming of her chess-playing capacity depends entirely on whether she has an independent motivation to engage in playing chess. Absent such motivation, her lack of responsiveness does not reflect a deficiency of chess-playing excellence. Consequently, if the notion of capacity at work in Epistemic Reasons were the one familiar from sports, arts, and games, the resulting account would render epistemic reasons similarly conditional on an agent’s desires, ends, or goals. A believer’s failure to respond to evidence would count as a shortcoming only if she had a desire to find out the truth. Thus, the account would fail to capture the categorical force of epistemic reasons.

By contrast, once we recognize the distinctive nature of epistemic capacities, CC-E’s prospects of vindicating Categoricity look considerably better. The key insight here is that, as argued in the previous sections, excelling as a possessor of epistemic capacities requires unconditional responsiveness to correctness-relevant considerations (i.e., evidence). Hence, lack of such responsiveness can by itself constitute a shortcoming or malfunction of epistemic capacities, independently of whether the subject has any desire to know the truth. Put differently, both chess players and believers are assessable in terms of how well they respond to considerations that bear on correctness in their respective domains. But in the epistemic case, the authority of these considerations is categorical or inescapable, in that a believer cannot avoid a negative assessment of her excellence simply because she lacks a desire to heed them.

It is important to be clear about the sense in which epistemic standards are inescapable. What is inescapable for believers—but escapable for chess players—are assessments of excellence, not assessments of correctness. Arguably, assessments of correctness are equally inescapable in both domains: just as an incorrect belief does not cease to be incorrect simply because the believer is indifferent to the truth, an incorrect chess move does not cease to be incorrect simply because the player is indifferent to winning. The crucial difference lies elsewhere: in how such indifference bears on the agent’s relevant excellence. An indifferent chess player’s incorrect move need not count against her skillfulness. By contrast, a believer’s failure to respond to decisive evidence does count against her epistemic excellence—even if she lacks any desire to conform to the epistemic standard of correctness.

Interestingly, this distinction also helps to defuse a well-known argument by Philippa Foot (Reference Foot1972), intended to challenge the idea that moral standards differ from standards of games or etiquette in virtue of their categorical force—an argument that has since been extended to epistemic standards as well (e.g., Côté-Bouchard, Reference Côté-Bouchard2016). Foot argues that the standards governing games, etiquette, and similar practices are no less inescapable than moral—and, by extension, epistemic—standards:

Foot’s claim seems plausible if “standards” are understood merely as standards of correctness. A breach of etiquette does not cease to be a breach simply because the agent is indifferent to complying with standards of etiquette. In this sense, the application of correctness standards is indeed equally insensitive to an agent’s contingent desires across domains of etiquette, morality, and epistemology. However, once we distinguish between assessments of correctness and assessments of excellence, Foot’s strict analogy between these domains breaks down. While assessments of correctness may be equally inescapable across these domains, assessments of excellence are not. In the case of games and etiquette, an agent’s indifference can exempt them from a negative assessment of their excellence. Again, as noted, if a chess player is indifferent to winning, then passing up an opportunity to play a winning move need not reflect poorly on their chess skill. Likewise, if an agent does not care about following rules of etiquette, then their violation of these rules need not reveal a lack of mastery of them.

By contrast, standards of excellence apply to possessors of epistemic capacities as such. A believer cannot avoid assessments in terms of epistemic excellence simply because they do not care about the truth. If they fail to revise a belief in the face of clear counterevidence, this reflects poorly on their epistemic excellence—even if they prefer to remain ignorant on the matter. Thus, unlike assessments of excellence in sports, arts, or games, assessments of epistemic excellence are categorical: they do not depend on the agent’s contingent aims, interests, or desires.

Recognizing this distinction enables CC-E to resist the Foot-inspired challenge to the distinctive authority of epistemic standards. Even if the standards of games and etiquette are inescapable in the sense that assessments of correctness do not vary with an agent’s contingent desires, they lack the categorical authority characteristic of assessments in terms of epistemic excellence. And it is precisely this feature—this inescapable dimension of epistemic excellence—that allows CC-E to explain why epistemic capacities give rise to genuinely normative notions.

4.3. Explanatory direction

Before concluding, it is worth considering another worry about CC-E’s project of explaining epistemic reasons in terms of well-functioning epistemic capacities.

One might agree with CC-E that there is indeed a systematic connection between epistemic reasons and well-functioning epistemic capacities in that epistemic reasons align closely with the considerations to which epistemically excellent believers are responsive. One might also grant that such a tight correspondence calls for explanation. It is surely not a mere coincidence that epistemic reasons systematically align with the considerations to which believers respond when exercising their epistemic capacities well. But one might object to the way in which CC-E seeks to explain this systematic alignment.

There are two natural ways to explain why this alignment is no coincidence: either by explaining epistemic reasons in terms of well-functioning epistemic capacities, or by explaining well-functioning epistemic capacities in terms of epistemic reasons. As we have seen, CC-E treats capacities as explanatorily fundamental and seeks to account for epistemic reasons in terms of them. According to Epistemic Reasons, a consideration counts as an epistemic reason just in case, and because, failure to respond to it would constitute a malfunction or shortcoming of one’s epistemic capacities. But why privilege this direction of explanation? Why not instead reverse the order of explanation and maintain that a well-functioning epistemic capacity is one that reliably tracks independently existing reasons for belief? On this alternative, reasons are explanatorily prior, and the notion of well-functioning is understood in terms of responsiveness to them.

This reasons-first strategy fits naturally with the familiar primitivist view of reasons, according to which reasons are simply considerations that count in favor of something and admit of no further analysis or explanation (see, e.g., Scanlon, Reference Scanlon1998). By adopting this strategy, one could explain the systematic connection between epistemic reasons and well-functioning epistemic capacities while retaining a commitment to primitivism about reasons. The question, then, is whether there is any good reason to prefer CC-E’s capacity-first explanation over this reasons-first alternative.

I think there is. Note that there are many considerations that count in favor of holding certain beliefs, not by bearing on their correctness, but by bearing on the practical advantages of holding them. Familiar examples include the fact that believing in God would make one happy, or the fact that one has been promised a large sum of money for believing that the earth is flat. Such considerations plausibly count in favor of forming the relevant beliefs. A primitivist about reasons therefore appears committed to treating them as genuine reasons for holding these beliefs.Footnote ²⁴

But this creates trouble for the reasons-first strategy. If well-functioning is explained in terms of responsiveness to a primitive notion of reasons, then well-functioning epistemic capacities should be responsive to these practical considerations as well. But that is highly implausible. Someone who forms the belief that God exists on the ground that doing so would make them happy is not exercising their epistemic capacities well. Responsiveness to practical incentives is plausibly not a mark of epistemic excellence.

The reasons-first strategy thus threatens to yield a highly revisionary account of epistemic excellence, one on which responsiveness to evidence is not constitutive of well-functioning belief formation. By contrast, the capacity-first strategy preserves the ordinary idea that epistemic capacities function well precisely when and to the extent that they respond to truth-relevant considerations, and that failures of such responsiveness constitute shortcomings of epistemic excellence. This strikes me as a clear advantage of CC-E’s capacity-first strategy over the reasons-first alternative.