1. Introduction
Early on in Transparency and Reflection Boyle identifies a common theme among philosophers in the Western tradition that informs his project.Footnote 1 It is the idea that
[there is] something distinctive of human mindedness, some essential openness to self-awareness, which informs our capacities for knowledge and action and thereby sets our lives apart from those of animals that lack this fundamental capacity for self- reflection… human beings [are] distinctively self-knowing animals. (2; first two emphases added)
Boyle seeks to return to “the classical idea that the topic of self-knowledge must be approached via a consideration of the distinctive nature of human minds,” as opposed to treating it “as a special problem in epistemology, on a par with the problems of perceptual knowledge, knowledge by induction, knowledge by testimony, and so on” (4f.). The primary concern of an account of self-knowledge, Boyle thinks, should be a form of self-awareness that “belongs intrinsically to certain of our first-order mental states themselves” (214).
Boyle identifies Kant as the historical progenitor of this “metaphysics first” explanatory direction that he favors.Footnote 2 In Kant, he finds the claim that there is a necessary connection between “rationality, self-awareness, and self-determination” (5, 19)—a claim that preoccupied Boyle in earlier work. But Boyle now thinks that “many purported arguments linking rationality with self-awareness are question-begging”; for, they build “an assumption of self-awareness into their characterization of the cognitive task that a rational subject must perform” (18). The “Kantian claim,” Boyle has come to think, “overintellectualizes rationality and sets the bar for possessing it implausibly high” (19). His present aim is to “vindicate the broader Kantian intuition” in a way that avoids this charge of overintellectualizing and question begging. This, he thinks, can be accomplished by recognizing that the self-consciousness distinctive of minds like ours is only implicit. Crucially, however, a rational being like us “must be capable of framing a first-person notion of herself, and concepts of the basic modes of presentation in which she can engage, by merely reflecting on an implicit self-awareness already involved in her first-order thinking” (19; first emphases added). The idea is that explicit propositional self-knowledge requires only making explicit what is already implicit in being in first-order mental states. To explicate this idea, Boyle calls on a “basic distinction” due to Sartre, “between pre-reflective and reflective self-awareness,” which Sartre presents as the distinction between “nonpositional self-consciousness” and “positional self-knowledge” (20).
In Section 2, I present Boyle’s diagnosis of the shortcomings of contemporary epistemic accounts of self-knowledge in terms of their failure to explain the key “transparency fact” regarding self-knowledge highlighted by Gareth Evans (Reference Evans and McDowell1982). In Section 3, I briefly explain Boyle’s use of the Sartrean notion of nonpositional self-awareness in his account. This use is designed to provide a metaphysical explanation of the transparency fact that is faithful to the spirit of the aforementioned Kantian claim and yet avoids the charge of overintellectualizing our first-order mentality. I then argue that Boyle’s explanation suffers a certain instability. To the extent that we take seriously Boyle’s ambition to provide an account of the nature of first-order mental states in terms of a built-in implicit form of (“nonpositional”) awareness that accommodates Evans’ transparency, we lose the idea that ordinary (“positional”) self-knowledge is achieved by “merely” making explicit what is already implicit in being in first-order mental states. We also risk losing the Kantian claim of the distinctiveness of our own mentality. On the other hand, to the extent that we regard nonpositional awareness as a form of genuine self-awareness, apt to give rise to self-knowledge through mere reflection, we risk overintellectualizing the nature of first-order mental states and their transparency. I conclude, in Section 4, by raising some questions concerning Boyle’s success in providing a viable alternative to contemporary (epistemic) accounts of basic self-knowledge (see Appendix).Footnote 3
2. Transparency and Self-Knowledge
In Chapter 2 (“Transparency and Other Problems”), Boyle offers a diagnosis of the failure of existing “epistemic” accounts of privileged self-knowledge. Such accounts all fail to provide an explanation of what he takes to be the central explanandum of the topic of self-knowledge, as epitomized by Evans’ “transparency fact.” In a much-cited passage, Evans says:
[I]n making a self-ascription of belief, one’s eyes are, so to speak, or occasionally literally, directed outward—upon the world. If someone asks me “Do you think there is going to be a third world war?” I must attend, in answering him, to precisely the same outward phenomena as I would attend to if I were answering the question, “Will there be a third world war?” I get myself in a position to answer the question whether I believe that P by putting into operation whatever procedure I have for answering the question whether P. (Reference Evans and McDowell1982, 225; cited in Boyle [Reference Boyle2024, 37])
As I read Evans, he offers his observation as part of developing what he describes as “the most powerful antidote to a Cartesian conception of the self.” He highlights the transparency fact as part of rejecting as misguided the Cartesian idea of privileged introspective access to mental states that are—for their subjects—“luminous.” As he remarks, to determine whether one is seeing a tree, one “does not in any sense gaze at, or concentrate upon, his internal state” but rather “gazes again at the world (thereby producing, or reproducing, an informational state in himself)” (Reference Evans and McDowell1982, 227). Evans’ transparency is thus not a matter of a special first-person perspective or point of view on our own mental states. His is transparency to the world, which is intended to replace altogether the Cartesian idea of transparency to oneself. Although he accepts that subjective experience is a matter of the world presenting itself to a subject who occupies a specific (first-person) vantage point, he denies that occupying this vantage point is a matter of having a perspective or point of view on those experiences. Footnote 4
Boyle construal is rather different. He maintains that
[the transparency fact] concerns the relationship, from S’s perspective, between two topics, one pertaining to the world and another pertaining to her own mind (e.g., whether p and whether she believes that p). Explaining this relationship requires explaining how it can be reasonable, from S’s own perspective, to treat these two questions as connected… This is the connection we need to understand…, above all, from the standpoint of the subject who makes the connection: How does she herself understand the relationship between these two topics? (61; emphases added)
Talk of the subject’s own reflective understanding of the intelligibility of the relationship between facts about her mind and facts about the world, or their recognition of what makes their own self-attributions reasonable or justified, seems to me to go against the spirit of Evans’ own discussion of transparency. For, it suggests that the first-person standpoint is a matter of having one’s own—explanatory—take on the mental facts that constitute that take, as opposed to being a distinctive first-order take on the world. Such talk appears to reintroduce the need for the subject herself—as opposed to us, who theorize about the matter—to have special (“transparent”) access to first-order mental states. Footnote 5 And Boyle’s identification of the “root” of transparency as lying in the fact that “the relevant mental states seem to be somehow present to me, just as such” (31; emphasis added) seems to take us back to the Cartesian notion of transparency to oneself which Evans himself rejects.
This return is a defining feature of Boyle’s “reflectivist” approach. On Boyle’s metaphysical view, it turns out that even our first-order mental states themselves (e.g., perceiving x or intending to ɸ appear) essentially involve some form of (implicit yet genuine) self-awareness. As he puts it:
Being in mental states seems itself (normally) to involve some kind of modification of my awareness. They seem to be states of which I cannot normally be oblivious, since I have a special perspective on them “from the inside.” Versions of this point apply, I believe, not only to occurrent states and activities, such as pain and conscious thinking, but also to standing attitudinal states, such as belief and intention.Footnote 6 (32; emphasis added)
I submit that, if our explanatory target is Evans’ transparency to the world, Boyle’s reflectivist account take us in the wrong direction. Instead of explaining why mental self-attributions are, contrary to appearances, essentially outwardly directed, just like the mental states they self-attribute, it portrays first-order states (and their nonself-attributive expressions) as involving Sartrean “nonpositional” awareness, which is inherently a form of self-awareness (albeit implicit). And it adheres to the notion that this form of awareness is constituted by subjects’ “special perspective on their own mental states from the inside.” As I read Evans, by contrast, he seeks to excise all appeal to subjects’ even implicit inward-looking consideration of their own mental states in explaining subjects’ unique ability to make first-person mental self-attributions. He rejects the idea that this ability is underwritten by a special introspective method.
As I see it, the challenge facing Boyle’s Sartrean-cum-Kantian account of transparent self-knowledge is this. To address his own worry about the Kantian account—that it overintellectualizes first-order mentality—and to provide an explanation of Evans’ transparency, he must insist that Sartre’s nonpositional awareness is not a kind of explicit self-knowledge. It is thoroughly world-directed and not conceptually about oneself and one’s mental states. At the same time, if his account is still to vindicate the Kantian claim, he must insist that, nevertheless, our first-order mental states already involve some kind of genuine awareness that can be transformed into explicit self-knowledge through “mere reflection.” In what follows, I argue that Boyle’s has not succeeded in meeting this (admittedly daunting) dual challenge, since his specific blend of Sartre and Kant suffers a certain instability.
3. Nonpositional Awareness Versus Positional Self-Knowledge
Boyle characterizes his proposed Sartrean approach to transparent self-knowledge as follows:
The approach has two main elements: first, the idea that any consciousness of the world (in the broad Sartrean sense of “consciousness”: any form of contentful awareness, whether occurrent or stative) involves nonpositional consciousness of itself; second, the idea that this nonpositional consciousness can be transformed, through an act of reflection, into explicit knowledge of one’s own representational state. It is our ability to bring our nonpositional consciousness to reflective articulacy, I claimed, that explains our capacity for transparent self-knowledge. (74; emphasis added)
By Boyle’s own admission, the Sartrean notion of “nonpositional consciousness (of) our own representational states” is a “dark” one, not the least because Sartre denies that such consciousness is really of that which it concerns (hence the peculiar “(of)”). Sartrean nonpositional consciousness, Boyle explains, is in no way “a further awareness superadded to our consciousness of the worldly objects of these states; it is implied in the very mode of presentation of the relevant objects … and … is integral to our first-order representational states themselves.” Nevertheless, he thinks that “this necessary nonpositional consciousness explains how a subject who possesses concepts of the relevant mental state types can achieve positional knowledge of her own mental states through mere reflection” (74; emphasis added). And that can help vindicate the Kantian idea that, in reflective beings like us, being in a first-order mental state already comes with knowledge of that state.
Since Boyle has already reviewed his construal and use of the Sartrean distinction, I will here only offer a brief recap of the main elements:
-
▪ Positional awareness involves the “positing” of specific referents, or representational targets: it brings certain objects and features under certain concepts.
-
▪ Positional awareness is thus always about some object or feature, which it explicitly represents.
-
▪ The awareness required for explicit self-knowledge is positional in that sense. It involves having propositional, articulate beliefs of the form [I am in M].
-
▪ This, in turn, requires making explicit (which is not to say “spoken”) mental self-attributions and the “positing” of referents for the concepts [I] and [M].
By contrast:
-
▪ Nonpositional awareness requires no “positing” of referents and no application (or even possession) of the concept [I] or [M].Footnote 7
-
▪ It is thoroughly first order and involves focus on the world outside the subject.
-
▪ Nevertheless, nonpositional awareness is a form of genuine self-awareness, insofar as it “presupposes” a subject’s “nonreflective” (nonconceptual?) and “implicit” understanding of their own mental state. This understanding is what enables subjects to achieve positional self-knowledge through “mere reflection.”
For Boyle, self-knowledge properly so-called requires making explicit doxastic self-attributions of mental states. It requires positional self-awareness. Such self-awareness is not part and parcel of being in first-order mental states, since it requires being able to form second-order, warranted beliefs about one’s own mental states and thus involves the application of sophisticated concepts (of self and of mental states). By contrast, the implicit understanding or awareness inherent in simply being in a mental state M is “nonpositional”—it does not involve a doxastic self-attribution of M. Ipso facto, it does not constitute a kind of self-knowledge that one is in M (not even “tacit”).Footnote 8
To use Boyle’s own example, begin with first-order positional awareness. I see a cat. I “posit” the cat as the intentional object of my conscious perceptual state. But I do not thereby posit my own perceiving of the cat. Nonetheless, Boyle maintains, for a subject to perceive a purring cat—as opposed to merely responding differentially to its presence—the subject must present to herself the intentional object of her state, the cat, in a distinctive way: under a visual–perceptual “mode of presentation” (74f.). Thus, seeing a cat, or making a visual judgment that there is a cat in front of one, requires presenting some object in the world in a specific (visual) way. But, Boyle thinks, this, in turn, depends on the subject’s “implicit awareness of their own perceiving.” And likewise for all mental states—occurrent sensations, emotions, feelings, intentions, and thoughts, as well as standing propositional attitudes, such as beliefs, preferences, and hopes, included.Footnote 9
Why does Boyle think that invoking the Sartrean notion of nonpositional awareness is either helpful or necessary in developing a satisfying account of self-knowledge and its philosophical significance? As noted in the Introduction, Boyle ultimately seeks to vindicate the Kantian intuition that the kind of self-knowledge of interest to us is in some way vouchsafed by the very nature of our distinctive mentality. And (as explained in Section 2), he thinks such a vindication ought to provide an explanation of Evans’ transparency. Standard attempts to vindicate the Kantian intuition, Boyle thinks, are guilty of overintellectualizing our own mentality, by portraying being in a mental state as involving believing that one is in the state. Boyle thinks that Sartre’s notion of nonpositional awareness—which is thoroughly nondoxastic, nonreflective, and not (second-order) conceptual—promises to address this worry. For example, adopting Sartre’s notion could allow us to accept that intellectually undeveloped young children (and perhaps even congenitally nonrational beings??Footnote 10) can be in various first-order states despite not being in a position to achieve explicit self-knowledge. At the same time, Boyle thinks he can still vindicate the Kantian intuition by maintaining that, in beings like us, who have mastered the requisite concepts and cognitive–doxastic skills, explicit self-knowledge is readily available through “mere reflection.” All we need to do to achieve self-knowledge properly so-called is to exercise the relevant conceptual–cognitive abilities and “make explicit” a self-consciousness that is only implicit in the first-order mental states.
To Boyle’s credit, he is sensitive to various difficulties and challenges facing his Sartrean attempt to vindicate the Kantian claim. At crucial turning points in his discussion, he takes great care to articulate specific questions his account ought to answer. However, the resulting account seems to me to suffer an inherent instability. Somewhat briefly first: Boyle’s appeal to Sartrean nonpositional awareness is designed to avoid the worry of over-intellectualizing mentality and to capture Evans’ transparency. However, insofar as nonpositional awareness is understood to be genuinely more nonintellectual and outward looking, it would seem rather far from explicit self-knowledge, epistemically speaking; and the Kantian claim that there is a necessary link between (our) first-order mentality and self-knowledge would seem much less plausible. Put differently, to the extent that the Sartrean account portrays nonpositional awareness as intellectually undemanding, it becomes less plausible that this awareness could be readily “transformed” into positional self-knowledge without the substantial addition of epistemic and conceptual resources. And to the extent that the account preserves Evans’ observation that the first-person stance is thoroughly outward looking—just a matter of being oriented toward (and representing) the outside nonmental world—it becomes less plausible to regard Sartrean nonpositional awareness as a form of genuine awareness of one’s self. Moreover, the less demanding nonpositional awareness is taken to be, the more arbitrary it would seem to suppose that only reflective beings like us can enjoy this form of awareness, thus failing to vindicate the Kantian intuition of the distinctiveness of our human mentality. Conversely, the more the Sartrean nonpositional awareness that is built into our first-order mental states is taken to manifest some—implicit—kind of self-awareness and self-understanding that could be readily transformed into explicit self-knowledge (through “mere reflection”), the more Boyle’s Sartrean account would be open to the charge of overintellectualizing first-order mentality, as well as of straying from Evans’ transparency to the world. And it would also seem arbitrary to insist that even conceptually innocent or cognitively limited humans enjoy the relevant kind of nonpositional awareness. In the remainder of this section, I elaborate on the instability I see in Boyle’s position.
3.1. Nonpositional awareness as nonintellectual or “practical”
To spell out the instability, consider first Boyle’s ambition to avoid charges he levels against other Kantian proponents and to properly accommodate Evans’ transparency. To paraphrase, Boyle worries that the Kantian proponents simply assume what they need to show and are packing too much rational “oomph” into their account of first-order mentality. By contrast, Boyle’s Sartrean account allows that one can have the nonpositional awareness involved in all first-order mental states without possessing any concept of self or of those states.Footnote 11 As regards awareness of self, Boyle explains that nonpositional awareness is not awareness of a subject, but as a subject. The notion of consciousness as subject is introduced (in Chapter 6) in terms of the familiar contrast between, for example, S wanting to go out and S wanting that they themselves go out. When one wants to do A, Boyle says, “things will present themselves as opportunities or obstacles in relation to” one; one’s orientation is “outward-looking” and “practical” (91), and one is “focused, not on oneself, but on the problems and opportunities presented by one’s environment” (92). Likewise, for all other “subject attitudes.” In the spirit of Evans’ observations, he suggests that the “subjective perspective” is “immanent rather than transcendent: the focus of my concern (…) is not on my being, doing, or undergoing something; it is on aspects of the situation I confront …” (93). So far so good—and transparent. However, Boyle goes on to point out that, although to have consciousness as subject just as such “does not imply knowing that one” is in the relevant state, so it differs from explicit self-knowledge, it also differs from having no consciousness of being in the state (97; emphasis added). For, it necessarily involves awareness of the worldly objects and features the state concerns “under modes of presentation” distinctive of the state. For example, if S consciously wants to do A, A will present itself to S under the mode to be pursued. If S consciously believes that p, p will present itself to S as to believed, and so on.Footnote 12 So far still so good (and transparent).
It seems right to suggest that subjective states orient subjects of experience toward the world in distinctive ways; in this sense, different mental states can be said to be associated with different characteristic “manners” or “modes” of presentation.Footnote 13 But as Boyle is well aware, to accept this is not yet to accept that the awareness involved in consciousness as subject is “a genuine form of awareness of [one’s own] mental states” (97). Indeed, in the following chapter (Chapter 7), Boyle details various conceptual pitfalls in supposing too readily that the kind of—essentially practical—orientation toward the world that constitutes nonpositional awareness, which he takes to involve merely representing nonmental objects in distinctive modes, can amount to a genuine form of awareness of our own mental states. Boyle considers in this connection what it takes for a subject to represent something as here. He says:
Surely a small child or a nonhuman animal might represent certain objects as here, and others as over there, without possessing a general understanding of what it is for an object to be in the vicinity of an arbitrary individual. … it is not plausible to suppose that [the] very capacity to represent things as here depends on [a] capacity for … a sophisticated, relational form of egocentric representation. (111)
All that is required are “monadic, egocentric modes of presentation” that are “connected with capacities to engage, perceptually and practically, with objects in one’s environment” (ibid.). All that the conceptually limited or innocent subject needs to harbor are representations of affordances, which are practical representations of worldly objects and features as opportunities or obstacles for the subject but not of the subject. Likewise, though, one might say: a young child or nonhuman animal could represent some things as attractive or scary or credible, others as repulsive, reassuring, or baffling—that is, as presenting distinctive opportunities or obstacles in engaging practically with the world “egocentrically.”Footnote 14 Such representations put subjects in relations to objects around them, yet without necessarily requiring—or amounting to—any representations of these relations to themselves, let alone requiring any representation or understanding (however implicit) of their own mental states as such. Footnote 15
A thoroughgoing “practical” construal of Sartrean nonpositional consciousness along the earlier lines would enable Boyle to escape altogether the charge of overintellectualizing consciousness. However, adopting this construal opens a significant conceptual–epistemic gap between subjects’ first-order conscious mental states and their explicit knowledge of them. Far from being “built into” first-order mental states, the conceptual–epistemic ingredients needed for explicit self-knowledge would in no way be present—not even implicitly—in the various “modes of presentation” distinctive of different types of mental states.Footnote 16 They could not be extracted out of the known states themselves through mere reflection. Coming to have self-knowledge could not then be claimed to be merely a matter of making explicit what is already implicit in first-order mental states. And it would no longer be clear in what sense explicit mental self-attributions qualify as transparent to the world just like the self-attributed first-order mental states. Moreover, the Kantian claim that the first-order mental states of beings like us necessarily come with a form of self-awareness properly so-called would not fall out of a Sartrean understanding of the distinctive nature of human conscious mentality.
A thoroughly non-overintellectualized account of nonpositional awareness would also fail to vindicate the Kantian “distinctiveness intuition,” according to which our own brand of consciousness is sui generis because only we can meet the requirements on genuine first-order consciousness. Boyle’s discussion of consciousness as subject and his own appeal to subjects “egocentric representations of affordances” (in Chapters 5 and 6) directly raises the question of how this Sartrean understanding could possibly vindicate the Kantian intuition of human separateness. But instead of tackling it head-on at that point, Boyle pivots to a different question, namely, the question of “what warrants the step” from the “modally egocentric awareness to explicitly self-ascriptive awareness” (114). This short circuits the more immediately pressing task of explaining why our—and only our—egocentric awareness of the world must qualify as even implicitly awareness of one’s self and one’s mental states and thus serves to separate us from all nonrational beings.
To summarize: if we adopt a thoroughly nonintellectual construal of Sartrean nonpositional consciousness, Boyle’s account would face two problems. First, it would turn out that explicit self-knowledge is too distant from the kind of nonpositional awareness that is built into our first-order mental states as such. Perhaps reflective beings are guaranteed the ability to achieve explicit self-knowledge, but this would turn out to constitute a substantive cognitive–epistemic achievement—no mere “making explicit of what is already implicit” in being in mental states. Second, it would seem unclear what role the Sartrean notion could play in a vindication of any Kantian human uniqueness claim. Sartrean nonpositional consciousness may well be the mark of the conscious, but it would turn out not to be a mark that could plausibly support a sharp separation between the mental states of rational beings and those of nonrational beings.Footnote 17
3.2. Rational mentality as sui generis
It should be clear by now why Boyle wants more of Sartrean nonpositional awareness than is afforded by the earlier construal of first-order mentality. To vindicate the Kantian claim that our own rational mentality is sui generis in being essentially linked to self-knowledge, he must insist that, in us (though not in nonreflective beings; see next), Sartrean consciousness incorporates the conceptual structure and cognitive ingredients necessary for explicit self-knowledge. This is what motivates Boyle to try to argue that, in beings like us, Sartrean nonpositional consciousness is indeed a form of self-consciousness—that it presupposes some kind of covert cognizance of oneself and of one’s mental states, despite its nonpositional character. Our nonpositional consciousness cannot be fully assimilated to the purely practical, nonself-aware engagement with the extramental world characteristic of nonrational beings.
In earlier work, Boyle defended in this connection the view that our mentality belongs to a different species of mentality from whatever mentality nonrational animals have. Our rationality transforms the very nature of our first-order mental states, so that none of them can be regarded as shared with nonrational beings.Footnote 18 Maintaining that the first-order mental states of rational beings like us are essentially transformed by our reflective–cognitive capacities, so they belong to a different species from the mental states of nonrational being, results in what I have elsewhere dubbed “Mind–mind dualism.”Footnote 19 Such dualism is clearly in the spirit of the Kantian separation of our minds from the minds of nonrational beings. In the present work, Boyle recognizes that his discussion of nonpositional consciousness must engage with this alleged separation. It “invites the question whether nonrational animals too possess” the sort of (“applicative”) understanding implicit in our having “representational states” that involve bring “extra-mental objects and features under ‘modes of presentation’.” In response to this question, Boyle offers a Sartrean spin on his earlier Mind–mind dualism. He suggests that we ought to draw “a distinction between two types of modes of presentation, only one of which implies even a nonpositional awareness of one’s own representational states” (192). Even when it comes to Sartrean nonpositional awareness, then we should appreciate that it comes in two entirely different varieties, only one of which qualifies as a form of mental self-awareness.
If Sartrean nonpositional awareness has the practical, world-directed, and nonreflective character Boyle himself highlights, it is unclear why Sartrean Mind–mind dualism should be more palatable than its Kantian predecessor. In support of the Sartrean version, Boyle proposes that there is a conceptual “link between rationality and the ability to consider” questions relevant to the various first-order attitudes (193). This link, he is confident, supports “a patent distinction of principle between a kind of representer that can merely hold attitudes on (what are in fact) questions and a kind that can pose the relevant questions themselves and consider them in an interrogative spirit” (194). He suggests that the distinction between two species of nonpositional awareness can then be drawn “by appeal to the difference between representing in a way that expresses consciousness of a question and representing in a way that does not express such consciousness” (192). Using this distinction, he thinks we can offer a Sartrean vindication of the Kantian necessary link between rational mentality and self-knowledge. Unlike nonrational animals, we enjoy nonpositional self-awareness that is inherent in our first-order mental states and can be transformed into explicit self-knowledge through mere reflection. Thus, once the essential rationality of first-order Sartrean consciousness in rational beings is recognized—once it is seen that they are inherently “self-knowledge ready”—the Kantian position can be vindicated after all.Footnote 20
Boyle observes in this connection that different first-order attitudes can be differentiated “by appeal to different questions they answer”:
[A]n intention can be thought of as an answer to the question what to do, a belief that p can be thought of as an answer to the question whether p, a desire for something can be thought of as an answer to the question whether that thing is desirable (in some respect), and so on.
He continues:
In the case of rational animals, however, we can think of their attitudes as answers to questions in an even stronger sense. For it is characteristic of rational animals to understand the questions to which their attitudes respond: they can themselves consider such questions in an interrogative mode, can recognize that a range of answers are possible, can answer these questions in such a way that their responses normally express corresponding attitudes. (192f.; emphases added)
I think this only serves to highlight further the character of the tightrope Boyle is attempting to walk. Very briefly: To support the claim that our first-order attitudes are different from those of nonrational beings, Boyle must identify a feature of those attitudes that qualifies them as in some (implicit) way engaging our reflective capacities. But the feature he settles on—a capacity to understand and consider questions pertaining to their attitudes—seems on its face highly demanding. Even if it does not depend on a capacity explicitly to ask and answer “Why” or “How” questions,Footnote 21 Boyle’s posited capacity still requires, for example, a scared child to be able to consider whether what they are scared of is scary; a young child’s feeling like going outside must be regarded as necessarily a matter of their having settled that going outside would be desirable, and so on. But how plausible is it to credit even prereflective, preconceptual children with manifesting Boyle’s posited capacity (and thereby with having first-order conscious states that have a kind of knowledge-ready self-awareness built into them)? And how reasonable is it to suppose that, in each of us, even a sudden sensation of pain, fatigue, or feeling of nonspecific anxiety must be a matter of our posing some question to ourselves (what question??)?
To avoid the charge of overintellectualizing our own mentality, and to reap the full benefits of invoking Sartre’s nonpositional awareness, Boyle would need to develop a nonintellectual construal of the idea that “the attitudes of rational animals” are “attitudes toward questions” and require “an understanding of the relevant questions, and in a manner responsive to the conscious consideration of them, which rational subjects can in principle undertake at any time” (193). And the construal has to make good on Evans’ transparency-to-the-world observation. In apparent appreciation of this, Boyle assures us that the “relevant questions” concern only “the objects toward which attitudes are held,” not “a subject’s own attitudinal state” (193n21). So “understanding them does not require grasp of concepts of the corresponding attitudes—though it may require grasp of certain concepts (…) whose application to an object normally expresses a corresponding attitude” (193). However, even if the relevant understanding does not require grasp of higher order mental concepts, it still remains to be argued that the concepts it does require do not exceed what all beings like us (very young children included) can be plausibly credited with. Furthermore, suppose the capacity to consider questions can be construed in terms of some form of thoroughly nonintellectual, first-order, practical, and transparent engagement with the world. (That is, it requires no grasp of higher-order concepts, and no ability to engage in any discursive practices, even hypothetically.). Then it becomes unclear how having this capacity is sufficient for bringing in its train explicit (positional) self-knowledge. And it will also become arbitrary to insist that this capacity is the exclusive prerogative of rational beings like us.
To summarize: the more nonpositional awareness is portrayed as informed (even if only implicitly) by reflective capacities of beings like us, the less clear it is how an account that incorporates it can avoid the charges of over-intellectualism and question begging of the sort Boyle levels against other Kantian proponents. And the less clear it becomes what the detour via Sartre is apt to accomplish. Moreover, it becomes unclear how it could be legitimately maintained—as on Boyle’s account—that even conceptually innocent or cognitively limited humans enjoy the relevant kind of nonpositional-yet-reflective knowledge-ready self-awareness.Footnote 22
4. Reflection and Self-Knowledge
By way of conclusion, I return—all too briefly—to the topic of the epistemology of self-knowledge. Recall that Boyle’s Sartrean understanding of the direct link between first-order mental states and self-awareness is intended to support some plausible version of the Kantian claim that there is a necessary link between rationality and self-knowledge. Starkly put, the Kantian claim is this:
Necessary self-knowledge (NSK):
Necessarily, if a rational subject S is in a mental state M, then S has (first-person) knowledge that they are in M.Footnote 23
On the face of it, NSK (necessary self-knowledge) is an exceptionally strong claim. It says, in effect, that one cannot so much as be in a mental state M without ipso facto knowing that one is in M. Since to know that one is in M one must engage in a propositional doxastic attribution of M to oneself, which in turn requires possession of the relevant concepts ([I], [M]), this means that only reflective beings like us, with the relevant cognitive–conceptual wherewithal could have mental states. And, moreover, that such beings always have active, actual self-knowledge of their M’s. This seems implausible on its face.
Partly for these reasons, contemporary epistemologists and philosophers of mind have shied away from adopting NSK and have instead focused on a more overtly epistemological thesis, viz.:
Privileged self-knowledge (PSK):
Necessarily, if a rational subject S makes a (first-person) doxastic self-attribution to the effect that she is in M, then S is in M (and their self-attribution is knowledgeable).
According to this thesis, subjects’ first-person judgments about their own states of mind (whether shared with others or not) have a special epistemic status. Although not based on any evidence, observation, or inference, they appear to constitute instances of privileged knowledge. And this seems puzzling enough, especially once Cartesian mind–body dualism is rejected. The challenge for contemporary epistemologists has been to explain (or explain away) this puzzling fact, consistently with a rejection of the claim that mental states belong to a separate ontological realm. Importantly, PSK concerns, specifically, beings who are already presumed to meet certain cognitive–epistemic requirements. These are subjects who are assumed to have in fact engaged in a doxastic mental self-attribution, by avowing—to others or just to themselves—“I am in M.” Under normal circumstances, the relevant self-attribution will be “first-person”: it will not be arrived at on the basis of any third-person evidence, observation, or interpretation. Suppose S has produced such a first-person self-attribution that they are, for example, feeling anxious or scared right now. Then, other things being equal, it seems hard to see how S could be simply wrong about how they feel. Hence, PSK.Footnote 24
On the Kantian version of NSK, what is built into the very nature of mental states is self-knowledge; no separate act of reflection is needed to bring about this knowledge. If S is in M, they will already, ipso facto, know that they are.Footnote 25 By contrast, on Boyle’s account, obtaining explicit self-knowledge requires an act of reflection. Boyle’s reflection requires not only possession of certain conceptual–cognitive abilities, but also the actual exercise of those abilities on given occasions, when (and only when) one is in relevant M’s. This means that, absent an act of reflection, even a rational subject who is in M will not actually know that they are. As Boyle himself puts it, such a subject “would be in a position, on reflection, to ascribe perceptual consciousness to herself. She would, that is, be in a position to know her own perceptual state through mere reflection” (150). This, however, means that Boyle cannot uphold an unqualified version of NSK. At most he can claim to have explained how a subject in M will know that they are IF they have the relevant abilities AND they exercise them on the occasion of being in M and upon reflecting on it. Put differently, the Sartrean nonpositional awareness that is built into our being in an M at most makes it possible for us—or puts us in a position—to have genuine knowledge that we are in M. This seems a far cry from the strong thesis of NSK. It is also a thesis that philosophers who are otherwise unsympathetic to Boyle’s metaphysical approach can take on board.Footnote 26 I conclude that, even if Boyle managed to avoid the instability in his Sartrean-cum-Kantian account articulated in the previous section, he could not claim to have fully vindicated the Kantian doctrine.
What is more, even setting aside the Kantian doctrine, without further development, it is not clear how a properly epistemological account of self-knowledge can be said to fall out of Boyle’s Sartrean metaphysical account and how it can provide a viable alternative the various accounts Boyle rejects. To provide such an epistemological account, Boyle has to tell us, at a minimum, two things. First: What are the conditions that ensure that one will engage in a cognitively effortless act of reflection upon being in M, so they produce a doxastic self-attribution of M? And second: When one in fact engages in an act of reflection that transforms nonpositional M awareness into an explicit belief that one is in M, what renders that belief warranted? Absent answers to these questions, all we have is, at best, a story about the potentially immediate or purely reflective character of our explicit mental self-attributions. This falls short of explaining what epistemic warrant accrues to such attributions—and thus what renders them instances of genuine knowledge. At the very least, we need to be assured that reflective acts do not regularly occur when one is not in M. Are we to suppose that one simply cannot, as a matter of necessity (of what kind?) arrive at an explicit mental self-attribution unless one is M? Why? It is not even clear why it should be presumed that subjects’ relevant explicit self-attributions must be true (let alone warranted).Footnote 27 And, even supposing that the relevant self-belief could of necessity only “come from” being in the relevant state, so it has to be true, we still need to learn what renders this belief justified—and (as Boyle requires) justified from the subject’s point of view. On pain of circularity, the answer cannot be that the subject is justified in their belief that they are in M because they know that it is “grounded” in, or comes from M.
Boyle’s discussion does not provide direct answers to these epistemological questions. Nor do possible answers seem to fall directly out of the Sartrean account he develops. I thus conclude that, even setting aside the internal difficulties I have raised for Boyle’s Sartrean attempt to vindicate Kant and explain Evans’ transparency fact, Boyle’s positive account does not provide us with a viable alternative to contemporary epistemological accounts of self-knowledge.
***
It is a mark of a fine philosopher to be able to keep firmly in view not only the ideas they wish to develop and defend, but also the challenges that face them in so doing. By this measure (as well as others) Boyle’s Transparency and Self-Knowledge is the work of a fine philosopher. Without a doubt, and my criticisms notwithstanding, anyone interested in Kantian and neo-Kantian claims about the nature of our own rational mentality (whether they wish defend or reject them) would do well to take the measure of Boyle’s book and engage with the distinctive and original Sartrean take on these claims it articulates.Footnote 28
Dorit Bar-On is Professor of Philosophy at the University of Connecticut and the Director of the Expression, Communication, and the Origins of Meaning Research Group.
Appendix:
A brief overview of central claims in Transparency and Self-Knowledge
-
(1) Necessarily, to be in M is to represent some aspect of the world with a particular (M-relevant) mode of presentation.
[Is this true of, e.g., all sensations, fleeting entertained thoughts, and other occurrent states that do not appear bring a present worldly object or feature under a “mode of presentation”?]
So,
-
(2) Necessarily, to be in M is to have nonpositional awareness of extramental objects and features.
[Why is not this just a matter of having first-order practical representations of affordances that is entirely world directed?]
-
(3) Nonpositional awareness comes in two distinct varieties or is bifurcated into two different species: the rational and the nonrational. Only the former qualifies as a form of self-awareness.
[Does this distinction correspond to a natural dichotomy? Does it amount to more than an arbitrary conceptual stipulation?]
-
(4) Our first-order mental states all fall under the rational species. They necessarily involve some form of mental self-awareness.
[Why accept that our nonpositional awareness of the world as such necessarily involves more than first-order practical, nonconceptual, egocentric representations of the world (with no self-awareness), which can in principle be shared by nonrational beings].
[Can this claim avoid all charge of overintellectualizing or question begging? Do prereflective, preconceptual children fall on the right side of the divide?]
-
(5) Our nonpositional self-awareness can be “transformed” into explicit doxastic self-attributions through an act of mere reflection.
[What if it is not so transformed? Must even cognitively equipped subjects who are in M always engage in transformative reflective acts? If they do not, do they still have actual knowledge, as opposed to just being in a position to know?]
-
(6) The resulting doxastic self-attributions will be not only true but also warranted self-beliefs, hence, qualify as instances of genuine (self-)knowledge.
[Wherein the presumption that subjects first-person self-attributions must be true? And wherein the warrant of even true mental self-attributions? How exactly does nonpositional self-awareness warrant the explicit self-attributions?]
