Expression and Self‐Knowledge. By DoritBar‐On and CrispinWright. Wiley &amp; Sons, 2024, xi + 349 pp. <scp>ISBN</scp> : 978111890847

For over twenty years Dorit Bar-On and Crispin Wright have been central figures in debates about self-knowledge. This book gives the reader a fly on the wall view of a high-level back and forth. The back matter claims that the book “focuses pre-eminently on the expressivist idea that a satisfactory explanation [of self-knowledge] will need to assign a leading role to our capacity to manifest our mental state in expressive behaviour and speech”. That is true, but it undersells the breadth of the discussion. Regardless of one's initial sympathies with the expressivist project, this is a book anyone working on self-knowledge should read. From a botanical perspective, the book does a remarkable job surveying both forest and trees, while mostly avoiding weeds. The book divides into four parts. A co-written introduction motivates three theses—the “Trifecta”— that, according to the authors, any acceptable account of privileged access must make sense of. The “Cartesian” approach is discussed and criticized, motivating the search for a non-Cartesian account. Part two, chapters 2–5, mostly co-written, raises objections to a host of accounts of self-knowledge from the recent literature. These chapters motivate 6 additional desiderata an account of the Trifecta must satisfy. In the third part, chapters 6–9, Dorit Bar-On gives an overview of Neo-Expressivism, Crispin Wright raises some objections to this, and Bar-On responds. The concluding part, chapters 10 and 11, involves Wright's attempt at an account of the Trifecta and Bar-On's objections to it. It must be said that Wright's own views here are presented in a much less committal way, humorously taking the form of a “prospectus” for future research of the sort fitting a grant application. Rather than a theory, three hypotheses are put forward. Salience: “we tend, in the round, to know what there is to know: our own mental attributes of the kinds listed do not, in general, elude our awareness, although those of others may often do so” (2). “D3: A satisfactory account should so far as is possible be uniform across the entire range of states with respect to which we enjoy first-person privilege” (101). Ignoring Salience, Bar-On and Wright agree that a satisfactory account of psychological self-knowledge will explain the authority that speakers possess when they avow their conscious mental states and the privilege that their first-personal beliefs about their own mind possess. On the Expressivist view, an avowal of “I am in M” is directly caused by and gives expression to the state of being in M. Avowals are akin to cries and moans. On the classical or radical view, associated with Wittgenstein, avowals express and do not report mental states. Here and in her 2004 book Bar-On makes a compelling case against this view. Wright does the same. A central concern is that the radical expressivist must deny that avowals constitute self-knowledge at all. On the Neo-Expressivist alternative, avowals are directly caused by first-order mental states, but they also express second-order beliefs. The epistemic security of self-beliefs is explained by the fact that they are grounded or based on their truthmakers. So, the Neo-Expressivist can explain authority and privilege. The status of this second-order belief has been a real sticking point in the reception of Neo-Expressivism. Indeed, it is probably the issue for critics. This is the most thorough discussion of it I have seen, much of it centering on what Wright calls “the Routine Testimonial Model.” On the Neo-Expressivist view, when I avow that “I am in M”, you form the belief that I am in M in recognition of the rational causal relation between avowals and underlying mental states. You might think the more natural view is that you believe I am in M because I told you, that is, deferring to an avowing subject is an instance of accepting testimony. But if that is right then the evidential force of an avowal (for an audience) derives from the epistemic standing of the second-order belief it expresses. Expressivists of all stripes deny this. Thus, an argument for the Routine Testimonial Model is an argument against Expressivism. Crucially, as Bar-On repeatedly emphasizes, the question isn't whether a self-belief is involved in the act of avowal—Neo-Expressivism grants it is1—but whether the authority of an avowal, or as they put it, its “evidential force”, must be explained by the epistemic standing of an antecedent or mediating second-order belief. Furthermore, again as Bar-On emphasizes, we have to consider whether the Constitutivist family of views fares any better in explaining this. One of the nice things brought out by the discussion is that settling this issue is going to require more clarity about the nature of belief and the relation between judgment and belief. Wright contends that the denial of an anterior belief gains false support from certain dubious phenomenological reflections about the speed at which avowals are made (271–4).2 Bar-On (300) in turn criticizes Wright's alternative view of belief. Informing much of the discussion is what Bar-On calls “the Episodic Constraint”, that a second-order belief must arise from an occurrent act of judgment at a time (158). The Constraint is a problem for Constitutivist views (on which believing that I am in M or being aware of this is part of what it is to be in M, perhaps in a range of circumstances) and it makes the appeal to anterior second-order beliefs hard to square with the rejection of the Epistemic Approach. (Roughly, if I need to judge that I am in M before I avow that I am in M, and being in M is not sufficient for judging that I am in M, mustn't the judgment be the result of some inner detecting mechanism?) But it is also difficult for the Neo-Expressivist who must explain privileged self-knowledge in the absence of a spoken avowal by appeal to inner expressions.3 The Episodic Constraint merits more discussion. … we argued that so-called epistemic views that treat subjects' first-person authority as a form of self-expertise face a principled challenge. Briefly, we argued that observers ordinarily presumed avowals to be authoritative even in advance of establishing a subject's expertise about their own mind; indeed, it is not even clear how observers could do so without presupposing first-person authority. (226) I am sympathetic with the lesson here, but the argument seems too weak. If you and I were chatting I would treat you as an authority about your birthdate, the age of your children, and a host of other topics, without checking after your credentials. Part of being a competent adult is being able to answer questions about these things. But this tells us nothing of significance about the epistemology of the relevant domains of facts. The fan of the Epistemic Model can grant Bar-On and Wright's sociological observation if they add that self-knowledge is something we expect rational creatures to possess. But that seems uncontroversial. A stronger argument is available, though. It is perfectly coherent to imagine a society that doesn't keep track of birthdates and ages. In such a society someone who claimed to know the precise day they were born would need to show some credentials. It is a contingent matter that people typically possess this knowledge. But the capacity for self-knowledge, it might be thought, is not like that. It isn't just that people don't have to earn first-person authority, it's that it is incoherent to suppose a rational agent could lose it. On the expertise model, the capacity to know about your mental states is entirely distinct from the ability to have those mental states, and so it is intelligible to have mental states but be debilitated in a way that prevents you from knowing them. If that isn't possible, the capacities for having and knowing must be more intimate. An argument like this, most well-known from Shoemaker's discussions of “self-blindness” (1996), has been thought to rule out the Epistemic Approach. But it also seems to require both rejecting the ontological independence of the first and second-order and embracing Salience. But Bar-On rejects Salience and accepts the independence thesis (though see below). Wright thinks the Ground-Clearing Argument clears out Neo-Expressivism. For Neo-Expressivism the authority of avowals is explained by the fact that they are rationally caused by first-order mental states. According to Wright, “the invocation of causation… to underwrite the security of avowals brings with it issues akin to those that afflict” the Expertise Model (222). Avowals have strong evidential force because they are expressive. It wouldn't be rational to treat them this way unless we knew they were, which means that we knew of the relevant causal connection. The problem is that this is the sort of “credential” on which an avowal's authority does not rest.4 Notice, again, the weak version of the argument doesn't convince: there are different explanations for the social practice. What about the stronger version? Here Bar-On may be on firmer ground than the Epistemic Approach. For the Neo-Expressivist can deny the Ontological Independence of the first and second-order, at least in the strong form opted for by the Epistemic Approach. For Bar-On, expression is a mark of the mental in the sense that there is a natural connection between being in a mental state and being disposed to give expression to it in various ways. Those who possess the concepts of the mental are disposed to express by avowing. True, the connection between mental states and their avowals can weaken. This would be a loss in the subject's ability to express their mind. But this is a loss, as it were, in the first as much as the second-order. One important feature of Neo-Expressivism brought out by this maneuver is that, despite being naturalistic, the view agrees with the Cartestian that the mental is, in a sense, an epistemic category. But it denies that this commits one to a Cartesian introspectionist epistemology of mind. I have tried to give a sense of the terrain covered in the book, its overarching dialectic, and some of the main points of contention. But it is hard to do justice to the breadth and depth of the thing. The continued back and forth allows the authors to engage with each other's arguments and cover a lot of territory. There are many fascinating asides and suggestions that are worth pursuing further. The book should be indispensable for people researching self-knowledge and expression.5