As its title suggests, Self-knowledge for Humans attempts to ‘humanize’ recent philosophical discussions of self-knowledge in two particular respects. First, Cassam seeks to put questions regarding the value of self-knowledge back in the philosophical spotlight. Philosophers captivated by the apparent epistemic distinctiveness of self-knowledge have neglected both the kind of self-knowledge that is apparently valuable (‘substantial’ self-knowledge of one's values, say, as opposed to the ‘trivial’ knowledge that one believes one is wearing socks), as well the question of why self-knowledge ought to be pursued. Cassam is certainly correct in this diagnosis regarding the neglect of self-knowledge's value in recent philosophy. Standard reference works on self-knowledge (Brie Gertler's article in the Stanford Encyclopedia of Philosophy) as well as edited collections on the topic (e.g. Wright, Smith, and Macdonald's Knowing our Own Minds (Oxford University Press, 1998), or Brueckner and Ebbs' Debating Self-knowledge (Cambridge University Press, 2012)) exhaustively address the possibility, conditions, etc., of self-knowledge with nary a word about its value. Second, Cassam aims to defend an epistemology of self-knowledge suitable for actual human knowers, not one predicated upon the false assumption that we are ideally rational. Too much of the philosophical literature on self-knowledge, Cassam argues, represents human beings as ‘model epistemic citizens’, ever vigilant in ensuring that their beliefs and attitudes are justified. Cassam thus seeks an account of self-knowledge that ‘doesn't underestimate our cognitive failings and limitations’. (ix)
Cassam's discussion unfolds over fifteen brisk chapters. While there is invariably some repetition in its themes and arguments, Self-knowledge for Humans succeeds both in laying out Cassam's overall position on self-knowledge and in functioning as a stimulating and accessible introduction to the topic. Cassam also deserves credit for his overall diagnosis of what ails the self-knowledge literature, though in the end, I am less confident he has offered a sufficient cure.
In particular, Cassam is more successful on the second front than on the first. Cassam often laments the philosophical obsession with ‘trivial’ self-knowledge of little apparent value, and while the question of self-knowledge's putative value occasionally fights its way onto Cassam's agenda, readers hoping for a sustained and innovative investigation of the value of self-knowledge are likely to be disappointed by Self-knowledge for Humans. He provides a list of the characteristics that distinguishes substantial self-knowledge – the sort of self-knowledge ‘matters in a practical or even a moral sense’ – from trivial self-knowledge (30–32). One oddity of this list is that several of substantial self-knowledge's purported distinguishing characteristics are epistemic – substantial self-knowledge is fallible, corrigible, based on evidence, not acquired merely through reflection on the reasonableness of one's attitudes, etc. – characteristics that Cassam elsewhere argues are shared with trivial self-knowledge. The distinction between these two kinds of self-knowledge is thus narrower than it might appear, and Cassam could have devoted more energy to exploring this distinction. It appears as if the characteristics that actually distinguish substantial from trivial self-knowledge are predominantly ethical, rather than epistemic (that substantial self-knowledge is valuable, ‘entangles’ with a person's self-conception, etc.).
In any event, Cassam immediately pivots away from substantial self-knowledge and its value to ‘self-knowledge for philosophers’, i.e. knowledge of particular occurrent mental states. Self-knowledge's value then disappears until the final chapter. There Cassam argues that ‘high road’ accounts of self-knowledge's value, which argue that self-knowledge is necessary for the realization of lofty ideals such as rationality or authenticity, do not succeed. I concur with Cassam that these high road accounts are too quick, both in their attempts to show that self-knowledge is essential to the realization of such ideals and in their defense of the value of those ideals. Cassam opts for ‘low road’ accounts of self-knowledge, wherein self-knowledge's value consists in how it contributes to well-being. Even here, the value of self-knowledge does not turn out to be very robust, according to Cassam (224–26): More self-knowledge is not necessarily better for us than less. In fact, a bit of self-deception or self-illusion might be better on that score. Furthermore, self-knowledge's contribution to well-being may not rest on its being self-knowledge, i.e., on its being true, but merely on our believing it to be true.
One worry about Cassam's preference for low road accounts of self-knowledge's value is that it proves difficult to see that substantial self-knowledge plays any greater role in contributing to well-being than the supposedly less valuable trivial self-knowledge. No doubt substantial self-knowledge is more central to the agendas of our lives. Through knowledge of our values, emotions, etc., we come to craft what Rawls called a ‘conception of the good’, a ‘more or less determinate scheme of final ends’ that expresses ‘a view of our relation to the world’. (Political Liberalism (Columbia University Press, 1996), 19–20). However, trivial self-knowledge can contribute greatly to well-being in particular urgent contexts. Suppose that a person has had a previous heart attack and so knows what the signs of an attack feel like. In fact, she can differentiate these signs from other maladies that resemble them (indigestion, e.g.). One morning, she experiences these signs and affirms in ‘inner thought’, I know that I am experiencing a heart attack. Here her self-knowledge of her mental states could end up saving her life and so contributes mightily to her survival and well-being. But this looks like an instance of what Cassam would classify as trivial self-knowledge.
A second worry one might have about low road accounts is whether the putative value is actually contributed by one's knowing a given mental state or attitude as opposed merely to having the mental state or attitude. In many pedestrian cases, it seems as if all that is necessary for a choice or act to advance the agent's well-being is that she have particular states or attitudes that motivate such a choice or act, not that she know that she has those states or attitudes. If I desire a slice of toast with jam, and I believe that, inter alia, there is jam in my refrigerator, these states appear sufficient to motivate me to prepare myself toast with jam. I may well also know that I have these beliefs and desires. But assuming that Cassam has in mind that self-knowledge contributes to well-being in something like a causal way, that parcel of self-knowledge seems superfluous to the value in question. Low road accounts thus seem vulnerable to the worry that they are unable to show that self-knowledge has any distinctive value, value that is not ultimately reducible to whatever contribution to our well-being is made by the mental states of attitudes that we know ourselves to have.
That said, what is striking about Cassam's discussion of value is (once again) just how under-developed the ethics of self-knowledge is in contrast with its epistemology. (Only a few sources are cited in Cassam's final chapter.) At the very least, the question of self-knowledge's distinctive value seems ripe for the picking.
Cassam is much more successful on the second front. Much of the early chapters consist of a sustained attack on the ‘rationalist’ theory of self-knowledge, associated with Richard Moran, wherein self-knowledge is acquired via the ‘Transparency Method’, i.e. a person comes to answer the ‘inward-directed question’ whether she has a particular attitude by answering the ‘outward-directed question’ whether that attitude is warranted by her evidence. Cassam's rejection of rationalism is multipronged, but its core is that rationalism cannot be squared with the abundant evidence that we do not form our beliefs on entirely rational grounds (chapters 2, 5, and 6). There is, according to Cassam, a ‘Disparity’ between the homo philosophicus represented in rationalism and the biased, often uncritical, and self-ignorant creatures we actually are. I was less moved by these broad empirical arguments against rationalism than by the simple observations that (a) we often find ourselves with evidence regarding our attitudes (phobias, for instance), at odds with what we take to be justified and (b) the Transparency Method seems patently ill-equipped to account for self-knowledge of attitudes besides belief (do we really ascertain who we love by looking to the evidence we believe would justify loving them?). Cassam also rejects ‘inner perception’ theories. Cassam analyzes such theories as resting on the inference that because self-knowledge is immediate, and perception is immediate rather than inferential, self-knowledge must have a perceptual character. (123–24) Cassam rejects both the immediacy claim and the thesis that perception provides immediate knowledge of ‘inner’ states.
Against rationalism and inner perception theories, Cassam prefers ‘inferentialism’, that ‘inference is a key source of intentional self-knowledge for humans’. (137) How are we to gloss his talk of inference as a ‘source’ of self-knowledge? On its face, inferentialism is not a view regarding the source of self-knowledge in the sense that either rationalism or inner perception posit sources of self-knowledge. Those views claim that there is a particular kind of premise or evidence that forms the basis of our self-knowledge. Inferentialism, in contrast, is a claim about the justificatory structure of self-knowledge, one that rejects the ‘myth … that intentional self-knowledge is normally “immediate”.’ (141) But as Cassam observes, ‘inferentialism is an inclusive doctrine that keeps many doors open’. (140) Indeed, rationalism and inner perception could themselves be construed as inferentialist. A rationalist like Moran might propose that we infer our self-knowledge from the evidence we take to justify our attitudes, i.e., that from the fact that evidence indicates to me that P, I infer that I believe that P. (5) Similarly, defenders of perceptual theories could argue that from our awareness of inner states, we infer that we know those states. (Cassam argues as much in chapter 10.) All three theories could thus represent self-knowledge as an inference from evidence. What issue therefore divides these theories? My sense is that once Cassam embraces inferentialism, the issue that divides his view from its rivals is whether the evidential bases of self-knowledge are singular or plural.
On this point, Cassam's ‘source pluralism’ has great appeal. Indeed, a great deal of the philosophical literature on self-knowledge has operated on the unstated (and mostly undefended) assumption that there is a single source of self-knowledge. But it is mysterious (at least to me) why we ought to investigate self-knowledge beginning from that assumption. The contents of our mental attitudes do not seem to constitute a domain of truths known through a particular method or on the basis of a particular body of evidence (in the way that there might be a particular method or body of evidence relevant to knowing arithmetic or the geography of Chile, say). Moreover, the assumption that self-knowledge has a single source may explain why so many epistemologists of self-knowledge seem content to develop theories that even they concede only account for our self-knowledge of some kinds of states (beliefs) but not others (emotions, desires, or values). A pluralistic account of the sources of self-knowledge, while less tidy, would likely do better in accounting for how we know the wide array of states or attitudes that we are capable of knowing. (47)
The relative inattention Cassam pays to the ‘nature, scope, and value’ of substantial self-knowledge (viii), along with the exhaustive attempts to refute Rationalism, result in a work that is not likely to assuage the disappointment that (Cassam notes) laypeople are likely to undergo when they confront philosophical work on self-knowledge. I would have preferred Cassam devote less effort to grinding old axes and more effort to developing new tools. These criticisms notwithstanding, Cassam admirably captures the present state of philosophical thinking about self-knowledge and does so in an accessible and engaging way. Self-knowledge for Humans does not fully embody the ‘humanistic’ revolution in the philosophy of self-knowledge that Cassam seeks, but it should inspire reform in that direction.
