1. Paul

The Practical Self begins with Descartes’s withdrawal into isolation. And it ends where the Meditations end: in a world of others, offering objections and replies, holding one another to account. But its route from self-consciousness to objectivity is decidedly not Cartesian—it denies that we can establish a connection between self-consciousness and the world on theoretical grounds alone. Instead, it argues that we have practical grounds to accept that we are the agents of our thinking and that our assent to this claim is sustained by practices which relate us to an objective world. Elliot Samuel Paul’s thoughtful comments identify two gaps in this argument and argue that Descartes’s account of agency and objectivity is needed to bridge them. The route from self-consciousness and objectivity needs supplementation with Cartesian resources.

Paul’s first gap concerns the claim that we have practical grounds to accept that we are the agents of our thinking. As I introduce the notion of a practical ground in The Practical Self, we have practical grounds to accept a claim only when two conditions are met: it is theoretically undecidable, and it is practically required. A claim is theoretically undecidable when there can be no theoretical grounds to accept or deny it. And a claim is practically required when assenting to the claim is a rational requirement on the pursuit of ends that we are required to set. Paul’s worries about bad faith target the first condition: he claims that, unless more is said, we have theoretical grounds sufficient to deny that we are the agents of our thinking.

What are these grounds? Paul claims that arguments for hard determinism give us theoretical grounds to deny that we are the agents of our thinking. For if hard determinism is true, then we cannot do anything which deviates from that which is determined by the past states of the world and the laws of nature. This is supposed to undermine the claim that it is theoretically undecidable whether we are the agents of our thinking because Paul takes it that it is theoretically undecidable whether p only when the theoretical grounds to accept p and not-p are equal. I argue in The Practical Self that we lack theoretical grounds to accept that we are the agents of our thinking (PS, pp.88–99, pp.103–110). So if arguments for hard determinism give us theoretical grounds to deny that we are the agents of our thinking, and if there are no theoretical grounds that support the claim that we are the agents of our thinking, then the question can be decided: we are not the agents of our thinking.

Paul thinks this gap can be bridged because he thinks there are resources in Descartes that show we have theoretical grounds to accept that we are the agents of our thinking. I come to Descartes below. For now, the question is whether we need theoretical grounds to support the claim that we are the agents of our thinking in order to counter the theoretical grounds adduced by the hard determinist. Absent such additional grounds, is the question of our intellectual agency decided?

Something is decidable only when it is knowable. And it does not suffice for something to be knowable that there are theoretical grounds in its favor. The grounds have to settle the question—they must be objectively sufficient.Footnote ¹ The invocation of knowledge here is neither ad hoc nor a placeholder for some eventual replacement in terms of degrees of belief: practical commitment is constrained by what we know. Consider the way in which our decisions about what to do are shaped by our knowledge of which options are closed. Perhaps there is some reason to think that the train to London has been cancelled—so long as I do not take its cancellation to be settled, I may still commit to catching it. The link between practical grounds and undecidability is explained by the way practical commitment is constrained by what can be known.Footnote ²

Paul appeals to Kant’s Antinomies as a contrasting alternative on which undecidability is established by showing equally good reasons to endorse both a thesis and its antithesis. But Kant does not take an antinomy to be necessary to show the undecidability of the traditional questions of rationalist metaphysics—that would elide the treatment of the ideas of Soul and God in the Paralogisms and the Ideal. Rather, the role of the Antinomies is to show the conflict that arises within reason when we seek to know the unconditioned and, in doing this, to undermine a certain use of the cosmological idea. It does not support an account of undecidability that requires equal and opposing grounds.

The mere presence of an argument for hard determinism is not enough, then, to show that it is theoretically decidable whether we are the agents of our thinking—the argument must also be good. Paul notes that I do not discuss arguments for determinism, and if any of them are sound, then there can be no practical ground to accept that we are the agents of our thinking. But if that is the case, it is not that we need some theoretical grounds in favor of agency to balance the scales. We need to live with determinism.

Could there be an argument for determinism that shows that we are not the agents of our thinking? That depends on delicate issues about the interaction between the two conditions on practical assent. For one might think that if we are practically required to accept that we are the agents of our thinking, then there cannot be theoretical grounds to deny it, simply because it must be possible to do what we are required to do. That is, theoretical grounds which determined that p was false would preclude us from accepting p on practical grounds—so the very fact that we must have faith that we are the agents of our thinking would suffice to exclude theoretical grounds which show it to be false. That would not tell us where the argument for hard determinism goes wrong, but it may give us confidence that it is unsound.Footnote ³

Paul is right to situate the discussion of our intellectual agency in the context of debates about the relation between freedom and determinism. These were central to early modern discussions of self-consciousness and objectivity. But if the claim that we are the agents of our thinking is theoretically undecidable, that is not because the reasons for and against it somehow balance out. It is because the reasons on either side do not settle the question.Footnote ⁴

Paul’s second gap concerns the relation between faith and objectivity. In The Practical Self, I argue that our faith in ourselves as the agents of our thinking is sustained through practices that relate us to others and that this provides a connection to objectivity of the sort pursued in the Cartesian and Kantian projects. Paul objects that this proposal cannot work unless we have theoretical grounds for believing in an objective world. The thought seems to be that unless we have independent grounds for thinking that there is an objective world, the practices that relate us to others cannot sustain our intellectual agency.

Again, Paul thinks that this gap can be bridged, and again, he thinks that Descartes provides the resources. He finds in Descartes theoretical grounds, sufficient at least for ordinary life, to accept that the objects of perception are independent of our perception of them. We will come to Descartes below. For now, the question is whether we need theoretical grounds to think that there is an objective social world in order for our relation to others to sustain practical assent in the claim that we are the agents of our thinking.

Paul’s framing of the issue can make it seem as if practical assent in our own agency is sustained by our acceptance of an objective social world. And that suggests a picture on which one sustains the other if and only if it has independent support. But it is central to the account set out in Chapter 5 of The Practical Self that faith is sustained not by belief but by practice. And although these complex forms of normatively governed cooperative behavior involve beliefs and experiences, it is not through relations of evidential support that they sustain items of practical assent. They do so by situating practical assent in a form of life, making it easier to accept that which we must accept on practical grounds alone.

How does this work for the case of intellectual agency? The Practical Self argues that our faith in ourselves as agents is sustained through discursive practices that relate us to others. The broad idea is that our practices of holding one another accountable for our thinking presuppose that we are the agents of our thinking and that this makes it easier for us to accept that which we are required to accept on purely practical grounds (PS, pp. 147–156). These practices are centered on other people—we would not continue our conversations and prolong our disagreements without accepting the existence of others, their good faith, and our capacity to change their minds (PS, p. 161; cf. Theaetetus 210b–c). Nor would our involvement carry such high stakes without the emotional and affective responses occasioned by such interaction.Footnote ⁵ But we do not need independent grounds for the existence of others in order for these practices to support our sense of ourselves as agents: participation in the practice is enough.

For these reasons, I do not see that we need theoretical reasons to accept that we are related to an objective world in order for social practice to sustain practical assent. Nor do we need theoretical reasons to accept that we are the agents of our thinking in order to show that the claim is theoretically undecidable. Still, even if we do not need such reasons, might we have them anyway? Paul argues that Descartes provides us with good theoretical grounds to accept both claims. His illuminating presentation of Descartes’s account of clarity shows how it is supposed to underwrite perfect certainty that we are the agents of our thinking and moral certainty—certainty sufficient for the conduct of life (Med. 7:149)—that we are related to an objective world. If Descartes is right, then we have good theoretical grounds for recognizing our intellectual agency and, at least for ordinary purposes, taking ourselves to be situated in an objective world. Cartesian clarity shortcuts faith.

How does clarity provide theoretical grounds for these claims? As Paul presents the story, it is broadly experiential: clarity is the appearance of truth such that to perceive something clearly is to have it presented to you as true. It comes in degrees: something can be presented more or less clearly. And when a claim is presented with perfect clarity, there can be no reason to doubt that it is true. But when a claim is presented with imperfect clarity, the reason to accept it is compatible with there being reason to doubt it. If there is also a reason to doubt, then we cannot accept it with the perfect certainty of the sort at stake during contemplation of truth. But imperfect clarity is often enough for the moral certainty of everyday life. It is the clarity of our experience that gives us reason to accept that we are the agents of our thinking, situated in an objective world.Footnote ⁶

Paul suggests that the clarity of a perception is an aspect of its phenomenology. On this way of presenting Descartes’s views, it is the phenomenology of experience which provides us with (perfect) reason to accept that we are the agents of our thinking and (imperfect) reason to accept that we are situated in an objective world. Should we accept that the phenomenology of experience is decisive in this way? Consider our status as thinking agents. Paul agrees with Descartes that experience provides us with evidence that we are the agents of our thinking. He suggests that there is an obvious phenomenological difference between jumping into the pool and being pushed into the pool, and that this shows up in deliberation about what to do. I am less confident that the phenomenology is so decisive. Claims about how things seem are sensitive to context, including what we know and what we accept—they are not a source of neutral data.Footnote ⁷

For this reason, rather than focus on the phenomenology of experience, I tried to give structural reasons for thinking that experience is ill-suited to explain our grasp of ourselves as the agents of our thinking: that it is the passive receptivity of experience which makes it ill-suited to explain our grasp and knowledge of ourselves as intellectual agents (PS, pp. 88–95). Paul suggests that this conflates the passivity of experience with the passivity of its object. But someone who shares the structural concern need not deny that we can experience and thereby come to know that others are acting—in this sense, there is no general assumption that the perception of agency requires agential perceiving. Rather, the concern is that the passivity of experience cannot account for the distinctive way in which we know our own agency.Footnote ⁸ It is these structural concerns, not the phenomenology, which will settle whether our status as the agents of our thinking can be decided by appeal to experience.

Indeed, it is not clear why Descartes takes his own appeal to an experience of freedom to have explanatory force. As Paul emphasizes, this experience is supposed to be most clear when we exercise our capacity to make and withhold judgments (see, e.g. Med. 7:57; Prin. 8a:7, 19–20). But what is gained by insisting that it is the experience that explains the knowledge that we are exercising this capacity? In the Fifth Set of Objections, Gassendi denies a claim about the will which Descartes takes to be self-evident through experience, and Descartes replies that either Gassendi does not “attend to the actions the mind performs within itself” or is unfree (Med. 7:377). The irony is supposed to be that in denying that he has an experience of freedom, Gassendi must have had such an experience, since denial is a paradigmatic exercise of freedom. So either he has failed to attend to his free act of denial, or he is not actually free. The power of this explanation seems to rest on the thought that we know we are free in exercising our capacity for judgment rather than some supposedly independent phenomenology of experience.Footnote ⁹

What about our relation to the objective world? Paul insists that objects are presented as independent of our perception of them—in this, they contrast with after-images. And he claims that this gives us a theoretical reason to hold that we are related to an objective world, albeit one which is defeated by other reasons in the context of alethic contemplation. But it does not follow from the fact that experience does not present objects as dependent on our perception that it thereby presents them as independent of our perception—experience might be neutral on the matter.Footnote ¹⁰ Nor does it follow from the fact that experience presents us with trees, tables, and the like that it thereby presents us with objects as independent of our perception, for it might be that we can use the concepts of ordinary objects to characterize the content of our experience without any commitment to the mind-independence of those objects.Footnote ¹¹

For these reasons, I suggested in The Practical Self that the sense in which experience presents its objects as independent is better understood in terms of its structural features: it presents its environment as something on which we have a partial perspective. Consider the difference between an object of visual awareness moving out of view and its receding from existence. This is experienced as a difference between the diminishment of the experiential relation we stand in to the object and the diminution of the object of experience itself (PS, pp. 64–65, cf. Textor, Reference Textor2019). It is in virtue of these structural features that experience presents its objects as independent of our perception of them.

Does this have implications for the Cartesian project? Paul accepts that the grounds provided by experience are not sufficient for knowledge that we are related to an objective world, at least so long as we are engaged in the project of the Meditations. But he thinks that this does not matter: perfect certainty is not needed for the conduct of life. I worry that this undersells what is at issue. If the Cartesian project were motivated by a quixotic craving for perfect certainty, then Paul is right that we could admire its intent whilst refusing to abandon anything which fell short. But if, as Paul suggests, Descartes endorses a conception of our perceptual capacities on which their exercise alone is insufficient for the knowledge that we are related to an objective world, then our grounds for thinking that we are so related can only come from outside the perspective of self-consciousness—with the reassurance, for instance, that God does not deceive. This raises a question about whether Descartes can provide an account of our relation to an objective world that is intelligible from the point of view of the self-conscious subject. In The Practical Self, I identified this as Kant’s complaint (see PS, pp. 19–20, pp. 26–31). For those who take it seriously, it is little reassurance to be told that it does not trouble us in ordinary life.Footnote ¹²

2. Brennan

The Practical Self claims that we are required to have faith in ourselves as the agents of our thinking and that this faith is sustained by a set of practices that relate us to a world of others. Hayley Brennan’s rich comments target each of these claims. She argues that there are alternative ways to support the claim that we are the agents of our thinking. And that the importance of discursive practices for self-consciousness can be extracted from their supposed role in sustaining faith. I take these in turn.

First, faith. In ch.4 of The Practical Self, I argue that we have practical grounds to accept a claim when it is theoretically undecidable and practically required. A claim is theoretically undecidable when there are no theoretical grounds sufficient to accept or deny it. And a claim is practically required when assenting to the claim is a rational requirement on the pursuit of ends that we are required to set. Thus, if we cannot settle that we are the agents of our thinking on theoretical grounds and if it is a rational requirement that we accept this claim in order to pursue ends that we are required to set, then we have practical grounds to accept that we are the agents of our thinking.

Brennan raises two challenges to this argument. First, she identifies theoretical grounds for assent which have not been ruled out. Second, she identifies alternative practical grounds for assent which provide something distinct from, and stronger than, faith. These challenges raise important issues about the relation between self-consciousness and agency and the nature and variety of practical assent.

Are there theoretical grounds to accept that we are the agents of our thinking? Brennan’s case in support focuses on the nature of self-consciousness. This is introduced in The Practical Self as the capacity to recognize some aspect of one’s perspective as one’s own. Does the exercise of this capacity involve thinking of oneself as the agent of one’s thinking? Not necessarily. For there could be creatures who recognize themselves solely as the receptive subject of their experiences—without, that is, thinking of themselves as the agents of their thinking. It is for this reason that I distinguished self-consciousness in its more general form from deliberative self-consciousness—that form of self-consciousness which also involves thinking of oneself as the agent of some aspect of one’s mental life. The argument of The Practical Self aims to show that deliberative self-consciousness requires faith that one is the agent of one’s thinking (PS, pp. 97–100).

Brennan is skeptical that these two forms of self-consciousness can be separated. And she thinks this has implications for whether there are theoretical grounds to accept that we are the agents of our thinking. In broad terms, her thought is that since we cannot separate the two forms of self-consciousness, the theoretical grounds we have for recognizing that we are the receptive subjects of mental episodes thereby provide theoretical grounds for recognizing that we are the agents of our thinking.

Why think that we cannot distinguish self-consciousness from deliberative self-consciousness? The two come apart if it is possible to recognize some aspect of one’s perspective as one’s own without thereby thinking of oneself as the agent of one’s thinking. I motivated this claim by appeal to the possibility of creatures who recognize only that they are the receptive subject of their experiences. Consider an animal that shows a sophisticated grasp of the disabling conditions for different sensory modalities, displaying not only sensitivity to items in the world but also some awareness of the perceptual relation they stand in to their environment.Footnote ¹³ Such a creature would be self-conscious in virtue of recognizing itself as the subject of its experiences. But it would not think of itself as anything other than the receptive location of certain perceptual episodes.

Brennan agrees that I need not think of myself as an agent when recognizing that I am the receptive subject of perceptual experience. But she points out that if this recognition involves self-ascribing episodes of perceptual experience, then even the recognition that I am the subject of perceptual experiences involves a self-conscious judgment. And if making such a judgment requires me to recognize that judgment as my own, then all self-consciousness involves recognition of myself as an active thinker. A creature who self-ascribes her experiences does not, then, have to think of itself as the agent of its experiences. But it does need to think of itself as the agent of that very self-ascription. So, if all exercises of self-consciousness involve self-ascription, and if self-ascription requires thinking of oneself as the agent of one’s thinking, then there is no gap to be drawn between self-consciousness and deliberative self-consciousness. All exercises of self-consciousness involve thinking of oneself as the agent of one’s thinking.

This line of reasoning assumes that any instance of self-ascription involves recognition of self-ascription. For then even the self-ascription of experiences requires recognition that one is self-ascribing and, thus, recognition that one is the agent of one’s thinking. In The Practical Self, I was wary of making this assumption—see PS, p. 99—but Brennan is right that the distinction between self-consciousness and deliberative self-consciousness wanes without it. And this might explain some of the resistance we feel in imagining a creature which can do no more than recognize itself as a receptive subject of experience.Footnote ¹⁴

Say that this is right. What are the implications for the claim that we are required to have faith in ourselves as the agents of our thinking? I noted above that there are practical grounds to accept a claim if and only if there cannot be theoretical grounds that establish it. And this is where the impossibility of distinguishing self-consciousness from discursive self-consciousness is supposed to have bite. For if we can have theoretical grounds to recognize that we are the receptive subject of our experiences, and if we cannot recognize this without also recognizing that we are the agents of our thinking, then we have theoretical grounds to accept that we are the agents of our thinking. There is no need for faith.

But these connections cut both ways. Mike Inwood, my much-loved and much-missed predecessor at Trinity College, once commented on the line that sleep is the brother of death, saying: People usually take it to show that the Greeks were not afraid of death, but maybe they were just really scared of falling asleep. Brennan takes it that if self-consciousness itself already involves a commitment to the claim that we are the agents of our thinking, then there are unproblematic theoretical grounds to accept the claim. Perhaps we should conclude instead that all exercises of self-consciousness involve an article of faith.

The strength of Brennan’s case here will depend on the supposed theoretical grounds we have to recognize ourselves as the receptive patient of mental episodes.Footnote ¹⁵ In The Practical Self, I consider three sets of theoretical grounds—those made available by experience, those made available through conceptual mastery, and those made available through transcendental argument—and suggest that none can explain our sense of ourselves as the agents of our thinking. More generally, no theoretical stance can capture the distinctive ways in which we relate to ourselves when engaged in episodes of conscious thinking (PS, pp.125–129). If these arguments are right, and if Brennan is right that we cannot think of ourselves as the receptive patient of mental episodes without also thinking of ourselves as the agents of our thinking, the result is an extension of our puzzle rather than its solution.

Are there alternative practical grounds to accept that we are the agents of our thinking? The argument set out in The Practical Self is modelled on Kant’s account of the practical postulates. Brennan suggests that Kant has other resources that offer a more satisfying account of the practical basis for accepting that we are the agents of our thinking. In particular, she thinks that Kant’s account of freedom offers a more promising model for accepting that we are the agents of our thinking. On Brennan’s reading, Kant holds that we have practical cognition that we are free. And she thinks this is stronger than the mere faith we have that God exists. She argues that our status as thinking agents is closer in both topic and justificatory structure to freedom than God: we have not faith but practical knowledge.

How do the cases of freedom and God compare? There is a way of presenting Kant’s views on which they share a structure. In broad terms, we have a distinction between theoretical and practical grounds for accepting a claim (CPrR 5:4–5). Theoretical grounds arise from reason’s interest in the cognition of objects: they are evidential grounds which bear on the truth of the claim (5:120). Practical grounds arise from reason’s interest in “the determination of the will” (5:120). There is thus a distinction between endorsing some claim on grounds which are broadly evidential and assenting to some claim on grounds which are broadly related to the determination of our will.Footnote ¹⁶

At this level of description, the cases of God and freedom are on a par: our assent to each is not determined by reason’s interest in the cognition of objects—it cannot be, since there can be no theoretical grounds to determine whether we are free and whether God exists. We are nevertheless permitted to accept each claim because it relates to a determination of the will. In the case of freedom, it is the fact of reason—that we are subject to the moral law—together with the recognition that “freedom and unconditional practical law reciprocally imply each other” (CPrR 5:29) which gives us practical grounds to accept that we are free. In the case of God’s existence, it is the fact that we are required to set the highest good as our end, together with a set of claims about the necessary conditions for setting that end, which gives us practical grounds to accept that God exists. In both cases, practical grounds are assured by the connection between the postulated claim and some moral requirement.

There are also differences. Brennan emphasizes that freedom follows immediately from the moral law, whereas God’s existence is a result of unpacking further conditions and inferences. This is an important difference and explains, perhaps, why Kant thinks our awareness of freedom is somehow more secure than our commitment to God’s existence.Footnote ¹⁷ But the difference should not be overstated: in both cases, our grounds for accepting the claim trace back to a determination of the will, and it is this difference in foundation which underwrites the legitimacy of assent. Appropriately so. There is no question for Kant that we can base legitimate cognitive attitudes solely on theoretical grounds. The puzzle is whether we can form legitimate cognitive attitudes that outstrip the availability of theoretical grounds. His foundation of practical assent in the will resolves the problem. This applies equally in the cases of freedom and God.

There are certainly passages where Kant seems to accord a different status to freedom—it is cognized (CPrR 5:30), it is proved (CPrR 5:47), it is, according to the third Critique, among the things that can be known (CPJ 5:468). But Kant also counts freedom as one of the postulates (CPrR 5:132) and does so even in contexts where our relation to the postulates is precisely contrasted with attitudes of cognition and knowledge (Bxxx; JL 9:68n). These contrasting claims raise a set of puzzling interpretative questions.Footnote ¹⁸ Brennan places weight on a passage in which Kant says that we cognize freedom and that without the moral law it would have remained unknown (CPrR 5:30). But cognizing or knowing freedom is not the same as cognizing or knowing the proposition that I am free. And it is compatible with cognizing freedom in a practical sense that the appropriate cognitive attitude to take toward the proposition that I am free is one of practical assent.Footnote ¹⁹ I do not see, then, that Kant’s claims about cognizing freedom detract from a shared structure of assent for both God and freedom. I use the term “faith” to pick out this mode of assent, but its substantive nature is a mode of assent determined by practical grounds. In this way of carving up the terrain, freedom and God go together.

Brennan is right that there remain differences between the cases of freedom and God—in particular, that our grounds for assenting to our own freedom do not turn on anything distinctive about the nature of setting ends. But this seems to me a point where the case for accepting that we are the agents of our thinking is more usefully modelled on God’s existence than on freedom. For I argue in The Practical Self that our commitment to our status as intellectual agents shows up most clearly in extended stretches of cognitive activities structured by decisions and the setting of ends. When we reason about what we will do, we treat the constraints imposed by our decisions as revisable in a way which allows us to retract the decisions without thereby counteracting the available evidence. That requires us to take a stance toward our decisions, which contrasts with the stance we take toward the constraints imposed by the environment (PS, pp. 125–128; cf. Soteriou, Reference Soteriou2013, p. 287 f.). It is the practical commitment involved in setting ends, rather than the fact of reason, which underwrites our sense of ourselves as intellectual agents.

This is not to deny that there are ways in which our relation to our own intellectual agency more closely mimics Kant’s account of our relation to our own freedom. And there may be projects for which it is useful to distinguish between the varieties of practical assent: not just practical cognition and faith, but perhaps also hope, acceptance, and trust. But it is the general distinction between theoretical and practical grounds for accepting a claim that is needed to explain our grasp of ourselves as the agents of our thinking. Lichtenberg’s puzzle is resolved when we see that we must accept the claim on practical grounds.

Brennan’s final set of considerations concerns the relation between practical assent and discursive practice. In The Practical Self, I argue that our faith in ourselves as the agents of our thinking is sustained by a set of practices that relate us to a world of others. Brennan is skeptical that we must have faith in ourselves as the agents of our thinking. But she thinks that the importance of discursive practices for self-consciousness can be secured independently of their supposed role in sustaining faith.

Brennan’s proposal draws on the Fichtean claim that a demand from another rational being is a condition on my own self-consciousness.Footnote ²⁰ She suggests that deliberative self-consciousness—that form of self-consciousness in which we recognize ourselves as the agents of our thinking—is made possible by practices of holding one another accountable. It is because we hold and are held accountable for our thinking that we come to have the capacity to recognize ourselves as the agents of our thinking. Deliberative self-consciousness thus requires discursive practice—and without any appeal to faith.

How strong is this connection? Fichte’s version of the claim is standardly understood as involving full-blown necessity, established—at least in the opening theorems of Foundations of Natural Right—by means of transcendental argument. In The Practical Self, I allowed that the discursive practices of holding one another accountable were central but denied that they were necessary: they sustain faith for most people, but there may be some among us who can hold onto the idea that they are the agents of their thinking without all the conversational back and forth. Brennan is likewise wary of the full-fat necessity claim. But she thinks that there can be a tight and quasi-generative connection between self-consciousness and practice, even if the relation is not strictly universal and necessary. Perhaps. But once we give up on a properly necessary connection, we need some alternative way to understand what it means to say that deliberative self-consciousness is made possible by the practices of holding one another accountable.

The more substantive issue concerns the argument for Brennan’s Fichtean proposal. If it is supposed to avoid notions of faith and practical assent, then it must provide us with theoretical grounds for thinking that holding and being held accountable make possible deliberative self-consciousness. Perhaps this is to be established by transcendental argument. Or perhaps through empirical reflection on the role of socialization in enabling deliberative thought.Footnote ²¹ But if the conditional is going to be discharged, we also need theoretical grounds to endorse its antecedent: that we are deliberatively self-conscious. The Practical Self claims there can be no such grounds. And this looks in line with Fichte’s own understanding of his argument. As he puts it in the Wissenschaftslehre nova methodo, “the scope of Kant’s practical postulate is too narrow, for he limits it entirely to belief in God and immortality; but we will see that consciousness in its entirety is included within this postulate” (Fichte, Reference Fichte and Breazeale1992, p. 298). Fichte’s starting point, the principles which ground the Wissenschaftslehre, are postulates—items of practical assent.Footnote ²²

Brennan is right, then, that one need not think of practices as sustaining faith—perhaps the practices of holding one another accountable are the very basis for intellectual agency, a social foundation for deliberative thought. But if the aim is to bypass any appeal to practical assent, then both the connection between deliberative self-consciousness and discursive practice and our status as deliberatively self-conscious must be established on theoretical grounds alone. Brennan’s Fichtean proposal can avoid scaffolding; I am less confident that it can avoid faith.

3. O’Brien

The Practical Self is shaped by engagements with Descartes, Kant, and what is sometimes described, misleadingly, as the Strawsonian tradition.Footnote ²³ Paul and Brennan have represented Descartes and Kant in their contributions. Lucy O’Brien’s perceptive comments complete the troika. They center on two broad issues. First, the nature and status of the starting points for the arguments presented in The Practical Self. Second, the resources provided by the consciousness that we are acting and their relation to experience. They raise questions about the structure of practical justification and its relation to agency. I take them in turn.

Start with the argument for the claim that we are required to have faith that we are the agents of our thinking. In broad terms, this holds that we are required to set the end of evaluating our perspective on the world and that it is a condition on doing this that we take ourselves to be the agents of our thinking. What grounds do we have to accept the starting point of this argument? O’Brien claims that if we have theoretical grounds, then the starting point assumes that which the argument claims is impossible—knowledge that one is the agent of one’s thinking. And if we lack grounds, why accept the conclusion of the argument?

O’Brien notes that these questions can be avoided if the starting point for the argument is not a descriptive claim established on theoretical grounds but a practical imperative: evaluate your perspective on the world! Indeed, it must be so if the argument is to provide a distinctively practical form of justification. For if there were theoretical grounds for the starting point, and assuming that the second claim is established on broadly theoretical grounds, then we would have theoretical grounds for the conclusion. And that is exactly what is denied. So any argument which seeks to show that we have practical grounds to accept a claim must trace our assent back to some determination of the will—the imperative that we evaluate our perspective on the world.

This does not preclude questions about the starting point of the argument—even if we do not need theoretical grounds to accept that we must settle questions about our perspective, we still need some reason to think that it is a genuine imperative and not an “expression[…] of decline, of the final exhaustion of life” as Nietzsche describes Kantian morality (AC §11, see PS, pp.124–125). O’Brien does not go this far. But she worries that if the starting point of the argument is the imperative to evaluate our perspective on the world, then it is too close to the claim that we are the agents of our thinking for the argument to have suasive force. For if this starting imperative is simply the practical demand to think for ourselves—to make use of our own understanding, as Kant puts it (E 8:35)—then the argument shows only that we must take ourselves to be capable of thinking because we are required to do so.

This in itself need not be a problem—as Brennan points out, a more direct connection would still underwrite practical grounds for accepting that we are the agents of our thinking. But it would make much of the argument in Chapter 4 otiose. O’Brien argues that if the requirement to evaluate our perspective goes no further than the requirement to think, then we do not need to appeal to considerations about our capacity to set ends in order to show that we have must have faith that we are the agents of our thinking.

O’Brien’s charge here connects with Brennan’s claim that the argument for practical assent would have been more perspicuously modelled on Kant’s argument for the postulate of freedom. For Kant, the connection between the moral law and freedom is much closer and more direct than the circuitous connection between setting the highest good as our end and God’s existence. Similarly, according to O’Brien, there is a much closer and more direct connection between the requirement that we think for ourselves and our status as the agents of our thinking than there is between our capacity to set certain ends and our agential status. Why take the longer path?

The answer turns on a distinction which O’Brien notes: between the habitual claim that I am the agent of my thinking and the progressive claim that I am agentially thinking. The latter is the topic of Descartes’s Meditations and Lichtenberg’s remarks on the cogito. And if the argument for practical assent is to connect with those discussions, it needs to establish more than the habitual. It needs to show that we have practical grounds to accept that we are agentially thinking. This is where the capacity to set ends is important. It is only because our status as intellectual agents is a condition on setting particular ends that we can establish the claim that we are agentially thinking and not just that we are a habitual agent of thought.

How so? In broad terms, the argument claims that self-conscious judgments—judgments in which I recognize some aspect of my perspective as my own—are individuated in part by their position in extended stretches of cognitive activity (PS, pp. 119–123; cf. Ryle, Reference Ryle1958). These cognitive activities are structured by the pursuit of ends, including among them the end of evaluating the propriety of those perspectives. Our capacity to set these ends constrains what we will do: when we decide on a course of action, we treat it as a constraint on our reasoning. But when we reason in this way, we treat the constraints imposed by our decisions as revisable in a way which allows us to retract the decisions without thereby counteracting the available evidence (PS, pp. 125–128; cf. Soteriou, Reference Soteriou2013, p. 287 f.). That is to say, we treat them as self-imposed. And when we do this, we take ourselves not just to be an agent of thinking in the habitual sense, but the agent of the thinking involved in this particular project. It is the requirement to set a particular end which secures practical assent in the progressive.

Consider now the argument for the claim that faith in ourselves as the agents of our thinking is sustained by practices that relate us to an objective world. Again, O’Brien worries about the starting point in ways that connect with the concerns raised by Paul. She notes that the discursive practices of holding and being held to account are ones that involve talking to others. But if talking is a way of thinking with words—a way of voicing our thoughts, as Jennifer Hornsby puts it (Reference Hornsby2005, p. 112)—then the existence of discursive practices already entails the ability to exercise agency in thinking. And if these practices necessarily involve people who are independent of me, then the existence of discursive practices also already entails the existence of an objective social world. Thus, if we have grounds to believe in the existence of discursive practices, we already have grounds for thinking that we are intellectual agents capable of exercising agency in collaboration with objective others. No argument is needed.

Note that this worry cannot be avoided by reconstruing the argument to objectivity as grounded in an imperative. The discursive practices which relate us to others are not necessary conditions on practical assent. If they were, we could perhaps argue that our entitlement to recognize ourselves as related to others is ultimately grounded in a determination of the will—by arguing, for instance, that it is a necessary condition on setting the end of evaluating our perspective that we accept both that we are agents and that we are related to others. That would be a way to establish practical grounds for accepting both claims (at least, modulo further arguments for their theoretical undecidability).Footnote ²⁴ But it is a central part of the discussion of practice in chapter 5 of The Practical Self that this is not the case. Conversation with others is one of the ways in which we scaffold our faith—it is not a necessary condition for its existence.

The question, then, is whether we need independent grounds to accept the existence of discursive practices in order to accept that these practices sustain our faith in the claim that we are the agents of our thinking. O’Brien’s thought is that the argument for faith precludes such grounds—if talking with others entails being the agents of our thinking, there can be no theoretical grounds to accept that discursive practices exist on pain of her previous objection resurfacing. And she claims that if we lack such grounds, then we have no reason to accept that this practice sustains our faith. The connection to objectivity is defunct.

This is a powerful line of criticism. But it assumes that we need independent grounds to believe in the existence of discursive practices in order to understand that they sustain our faith. And this can be resisted. Practices sustain practical assent by situating it in a form of life. They give structure rather than justify. Someone involved in these practices can reflect on them and come to understand the way in which they provide that support. They can recognize, for instance, that baptism celebrates the initiation of a new member into the community or that repeated communion signals one’s continued membership. When this happens, the participant does not acquire evidential warrant for the claims accepted as part of the practice—from the point of theoretical reason, there may be nothing to be said in their favor. Nor do they gain practical support, at least so long as they do not take the practice to be a necessary condition on practical assent. Rather, they make explicit that which is already contained in the practice, even though there may be no grounds independent of participation in the practice to accept its claims.

Still, even if we do not need independent grounds to accept the existence of practices, do not we anyway have such grounds? The deeper challenge posed by O’Brien’s comments is not that we need such grounds to recognize that practices sustain our faith, but simply that we need such grounds. Otherwise, the argument in Chapters 3 and 4 of The Practical Self looks to face a problem of other minds. If I lack theoretical grounds to recognize that I am the agent of my thinking, then I lack theoretical grounds to recognize that you are the agent of your thinking. And if the practical grounds to recognize that I am the agent of my thinking relate to a determination of my will alone, then they give me no grounds to accept that you are the agent of your thinking. The result is a structural problem of other minds: I have no grounds to recognize that you are a thinking agent—and a fortiori, no grounds to think that there is a practice involving other thinking agents. This looks like a problem, irrespective of whether we need such grounds in order for participation in the practice to sustain our faith.

There is a genuine issue here. It is a version of a problem which arises for Kant’s account of freedom—if my grounds for accepting that I am free rest in my own consciousness of the moral law, then it looks as though I can have no grounds for recognizing that you are free. For Fichte, this was the “gross deficiency” of the Kantian system. “How should we draw the boundary of all rational beings?,” he asks in a letter to Reinhold from 1795, “To which of these appearances should I assign the concept of rationality and to which should I not?… I ride a horse without asking its permission and without wishing to have it ride me in turn. Why do I have more qualms when it comes to the man who lends me the horse?” (Fichte, Reference Fichte1988, p. 408). If there is an answer to Fichte’s question, it will mimic the story about our practical grounds for accepting that we are the agents of our thinking: just as my grounds for accepting that I am the agent of my thinking reside in my capacity to set ends, so too will my grounds for accepting that you are thinking reside in your capacity to set ends and the relation of that capacity to the first-person plural. But this is a promissory note for a further discussion: O’Brien’s charge has force.Footnote ²⁵

I finally turn to our consciousness that we are acting and its relation to experience. Chapter 3 of The Practical Self argues that there are no experiential grounds to accept that we are the agents of our thinking (PS, pp. 88–95). O’Brien disagrees. There is a form of experiential awareness, she claims, which is both active and non-observational: consciousness of action is not confined to the receptive. This, she suggests, offers us a much simpler response to Lichtenberg’s puzzle: we know that we are the agents of our thinking on the basis of self-consciously experiencing thinking going on.Footnote ²⁶

O’Brien quotes Gassendi’s comments on the cogito in the Fifth Set of Objections as precedent for this approach: “You could have made the same inference from any one of your actions,” he complains to Descartes, “since it is known by the natural light that whatever acts exists” (Med. 7:259). But in appealing to an experience of thinking as the explanation for our knowledge that we are thinking, O’Brien more closely aligns with Descartes than his objector. Gassendi objects to Descartes’s method of radical doubt because he denies that the will is free to judge when there are no reasons that compel its assent. Descartes, in response, insists that “the will has the freedom to direct itself, without the determination of the intellect, toward one side or the other” (Med. 7:378). And, crucially, he insists that if Gassendi fails to see this, it is because he either fails to attend to or lacks the experience of freedom which demonstrates it: “although these propositions are self-evident, I am not prepared to set about proving them here. These are the sorts of things that each of us ought to know by experience in his own case… I affirm what I have experienced and what anyone can experience for himself, whereas your denial seems merely to be based on your own apparent failure to have the appropriate experience” (Med. 7:377). O’Brien could say the same.

I noted above in response to Paul that I am skeptical about whether these phenomenological reports settle the question of whether we experience ourselves acting. O’Brien notes that the structural considerations appealed to in Chapter 3 are similarly indecisive. For why confine the experiential to the receptive? Rather than trying to justify that restriction here, I want to ask what active, non-observational experiential awareness would look like. In her wonderful Self-Knowing Agents, O’Brien defends the idea that we know that we are acting on the basis of agent’s awareness: “a form of awareness had by creatures capable of controlling their actions, mental and physical, that is independent of any capacity of the creature to understand the term or concept “I,” that is both non-conceptual and non-perceptual in nature and yet that is capable of immediately warranting the self-ascription of the action that the creature is aware of in this way” (O’Brien, Reference O’Brien2007, p. 76). This form of awareness is not receptive because it is grounded in or constituted by the rational control we have over our actions: “we are agent-aware of our actions in virtue of carrying them out as a direct result of an active consideration of ways we might act” (2007, p. 120). Thus, for O’Brien, one form of active, non-observational experiential awareness is the non-observational awareness involved in a subject acting on the basis of her active considerations of the options available to her.

Now, there is a way of presenting this view on which it looks steadfastly opposed to the view defended in The Practical Self: the knowledge that we are the agents of our thinking is grounded in conscious experience because “there is a general entitlement immediately to self-ascribe those states… which are conscious” (2007, p. 119). This is a theoretical reason for assent, one which bears on the truth of the claim in question. But there is another way of presenting the views on which they look less far apart. O’Brien and I agree that we have grounds to accept that we are the agents of our thinking. We agree that these grounds stem from a process of deliberation which involves active consideration of the options available to us. And we agree that this marks an important point of contrast with our receptive knowledge of the environment as made available in perception (see, e.g., 2007, pp. 96–106, pp. 168–169). We differ as to whether this process should be classed as a form of experience, and thus whether it provides a kind of theoretical ground for the recognition that we are thinking. But that seems like a disagreement about classification rather than a difference in structure.

In a way, this should come as no surprise. O’Brien and I swim in the same waters. The Practical Self explains our grasp and warrant for the claim that we are the agents of our thinking in terms of the distinctive structure of practical justification. Self-Knowing Agents grounds it in a supposedly independent phenomenology of experience. This marks an important difference. But our disagreement is organized around a shared commitment to the insight which structures Self-Knowing Agents (2007, p. vii): that the first-personal aspects of self-consciousness cannot be properly understood independently of our capacity for action. The question is whether this insight can be captured by drawing a distinction within the class of experience or whether it requires, as Kant thought, a distinction within the structure of justification.Footnote ²⁷