|
Article Excerpt
Two apparent paradoxes lie at the heart of discussion of self-deception, one focusing on belief, the other on intention. The belief paradox concerns how the self-deceived can combine the belief that p and the belief that not-p. The intention paradox concerns how the self-deceived can intend to believe that p, and manage it, without knowing what they are up to and vitiating it. Both are said to be paradoxes because, on the one hand, self-deception seems possible and, on the other, it can seem to require combinations of states that render it impossible.
The first choice point for debate is whether to divide or dilute. Dividing presses on the analogy with the deception of others. There is no problem with Jo believing that p and Josephine believing that not-p; nor is there a problem with Jo intending to bring it about that Josephine believe that not-p and managing to make it so. The division strategy seeks to appeal to this fact and relocate the division, in some attenuated sense, within subjects so that they can genuinely count as self-deceived. Dilution explains how the allegedly paradoxical combination of states is not required. Instead, self-deception involves something less that is not paradoxical.
Both approaches suffer from a problem--in fact, the same one. Each gets rid of the paradoxical character of self-deception at the price of losing the instability that is essential to it. The problem with the self-deceived is that they seem to avoid accepting a certain proposition and have anxiety over, or lack confidence in, what they are up to. It seems as though the project may fail or requires work. I put all this in terms that are as neutral as possible. It is pretty clear how the two paradoxes with which I began involve a more precise articulation of it. Their anxiety or lack of confidence stems from the fact that, deep down, they believe the proposition and have intentionally produced a belief in the opposite whose work will be unpicked if they appreciate what they have done.
Thus, I say, characterize the instability and the essential work of characterizing self-deception is done. The approach has a number of advantages that I seek to bring out in the course of the paper. The first is that it provides unity where other accounts of unity fail. As we shall see, there are a number of different ways people can be self-deceived (in terms of combinations of states) and it is a mistake to try to single out one form only. The second is that diluted accounts are subject to counterexamples, or have unfortunate commitments, which can be avoided if an appeal to instability is added.
The second point needs careful handling. Self-deception is plausibly seen as one kind of theoretical irrationality: irrationality about what to believe, sincerely avow, or in some other way cognitively endorse as true. To keep the options open, I shall talk of cognitive endorsement and understand this to cover the other two types of states just identified, and take endorsement to be to endorse as true. Self-deception is often contrasted with wishful thinking (in which subjects cognitively endorse a proposition because they want it to be true) and full-blown delusion (which I shall discuss more fully in section 4), but in some way involves a subject losing grip on reality with regard to a certain subject matter so that they have little chance of being able to make appropriate cognitive adjustments to the way the world is. As the American Psychiatric Association's Diagnostic and Statistical Manual of Mental Disorders puts it, delusion is:
A false belief based on incorrect inference about external reality that is firmly sustained despite what almost everybody else believes and despite what constitutes incontrovertible and obvious proof to the contrary. (1)
The careful handling to which I refer is partly due to the fact that since we are in the empirical business of identifying, presumably, sometimes instantiated mental kinds, there is no reason to suppose that our present judgments about what is involved in self-deception, in contrast to these other phenomena, will survive scrutiny. An appropriate taxonomy may point in another direction. One of the aims of this paper is to argue that an appropriate demarcation should appeal to a certain kind of instability.
Careful handling is also required because it seems that our talk of self-deception may involve two substantially different senses of deception. According to the first, external, way, the crucial difference is that the resultant cognitive endorsement is false. Subject are deceived by themselves because they are, in some way, responsible for the fact that they have arrived at a false cognitive endorsement. According to the second, internal, way, subjects are deceived by themselves because they are, in some way, responsible for the fact that they have arrived at a cognitive endorsement that, by their own lights, they take to be or suspect is false.
This is compatible with the cognitive endorsement being true. Maybe Alfred Mele is right that we should add that the cognitive endorsement is false simply because that's what deception means. (2) Yet, arguably, this would lose an important dimension of commonality between true and false cases of internal self-deception.
The most significant contribution of diluting accounts is that they have rightly questioned whether all self-deception must involve the allegedly paradoxical combinations of states even if they have failed to establish that no self-deception can include the combinations in question. Indeed, the foremost proponent of a diluting account--Mele--seems to accept that at best the paradoxical combinations are empirically unjustified rather than logically or metaphysically impossible. (3) This raises the question of why attempt dilution at all. The right answer is that the diluters are onto something--as I already mentioned--but have fumbled the justification for their position. They can argue that weaker combinations are all that is required when they give rise to the essential instability of self-deception. They don't have to grit their teeth and say: "limit self-deception to that if you want to but it is not obvious that the case you have provided should be described in that way."
In section 1, I will focus on the question of whether self-deceptively supported cognitive endorsements that p are the result of a desire that p, a desire for the cognitive endorsement that p or an intention that one cognitively endorse that p. Broadly, there are two arguments for holding that it is a desire for the cognitive endorsement that p or an intention that one cognitively endorse that p. First, it is suggested that an action, or more generally purposive, explanation of the cognitive endorsement that p is not available if a subject desires that p. (4) In this respect, appeal to a desire for the cognitive endorsement that p is better. Second, it is suggested that appeal to a desire for the cognitive endorsement that p, or an intention that one cognitively endorse p, is preferable, because it presents the best chance of self-deception being a unified phenomenon. (5) Twisted cases of self-deception (in which subjects self-deceptively believe what they don't want to be true) vitiate this prospect if we take self-deception generally to involve a desire that p, since then we are forced to treat twisted cases as special. (6)
I shall argue for the following three claims. First, it is compatible with self-deception involving a desire that p that an agent-style explanation can be provided of the resultant cognitive endorsement that p. Second, Mele's appeal to desires' influence on confidence levels for, specifically, believing that p or believing that not-p provides limited insight into the nature of self-deception, and hence it is not clear that an alternative to agent-style explanation has been identified. Third, it is a mistake to attempt to unify cases of self-deception by appeal to a desire for belief, or cognitive endorsement more generally.
The upshot of section 1 is that things look good for those who try to unify self-deception by appeal to agency. In section 2, I argue that this masks a diversity of kinds of agent-style explanation that may be provided.
So we need to look for unity elsewhere. In section 3, I discuss how appeal to the essential instability of self-deception in attentive consciousness can provide this unity and, importantly, a come-back for Mele's approach appealing to desires' influence on confidence levels. I develop this point to illustrate further how such an appeal enables diluting accounts of various kinds to avoid counterexample, in particular, those that reject the idea that the self-deceived must both believe that p and believe that not-p. So I make good on my claim that there is unity in the face of considerable diversity. In section 4, I discuss the differences between self-deception and delusion. I compare, favorably, my instability-based account with Mele's appeal to motivational factors.
1. Agency versus Non-Agency Views of Self-Deception
Agency views of self-deception hold that self-deceptively favored cognitive endorsement is produced by some, perhaps attenuated, form of agency. The most straightforward way in which this may be understood is that such cognitive endorsements are intentionally produced. Anti-agency views deny that these cognitive endorsements are produced as a result of agency. Instead, anti-agency views take desires and, perhaps, emotions more broadly to have a direct influence upon self-deceptively favored cognitive endorsements. (7) Both agency and anti-agency views have an explanatory burden: what is the mechanism by which desires and other emotions influence subjects' cognitive endorsements? The attempt to satisfy this explanatory burden is one way to get traction on the issue of what is the most plausible attribution of states in virtue of which the product of self-deception is achieved: via an intention to cognitively endorse that p, a desire that p, or a desire for a cognitive endorsement that p.
Suppose that I want to believe that I am a good driver or want to believe that I am a good judge of character. Why do these (when they do) result in the belief that I am a good driver or the belief that I am a good judge of character? The problem is that the desires seem to concern states of the world and not beliefs we may have. Yet, the desires are supposed to result in beliefs, or cognitive endorsements more generally. We don't get what we desire and yet the desires are supposed to explain what we get.
In the case of agency accounts, one way in which there would be a connection between subjects' desire that p and belief that p is if they had the means-end belief that if they believe that p, then p is the case. Nevertheless, this would be a most peculiar means-end belief for subjects to have. It attributes a power to thought that, sadly, is rarely evident. (8) Fortunately, there is a better alternative. If we desire that p, then, in general, we also desire to believe that p. If you want the world to be a certain way, then you don't want it to be that way without also believing it to be that way. There will be exceptions--indeed we could fix up a fantasy case in which coming to believe it would destroy the very thing we want--but agency explanations will be available for all the others. S desires that she believe that p, believes that by doing such and such, she will have the belief and hence,...
|
