A number of philosophers have suggested that delusional people do not believe their delusions; they only imagine them and then mistake their imagining for a belief (Greg Currie has a view along these lines: in his recent book, Phil Gerrans defends a related view). What follows are a few inchoate thoughts about views like this.
The main attraction of the imagination view is that it helps to explain the mismatch between patients’ professed beliefs and their behavior: the man who alleges that his wife has been replaced by an alien replica may not show much concern about sharing his home with a replicant or worry about what has actually happened to his real wife. Because imagination plays a different functional role to a belief, we can explain these apparent inconsistencies in behavior. Imaginings can be involved in inferences – hence fictional narratives, which may be faulted for inconsistency – but they don’t underwrite inferences outside their domain (an imagined scenario’s being inconsistent with states of affairs known to obtain doesn’t make it defective). But Currie’s account doesn’t seem able to explain why the imagining would continue to lack the capacity to underwrite inference once it is taken by the person to be a belief. Belief attributions seem self-fulfilling: when an agent takes herself to believe that p, she believes that p in virtue of so taking herself. Of course there are apparent exceptions, like self-deception and perhaps implicit attitudes. The point is that we need an explanation of why so maintaining doesn’t bring about the correlative belief.
In fact, I think things are worse than that suggests: not only is there theoretical reason to think that imagining oneself to believe that p (in a manner which doesn’t recognize that it is an imagining, at least) would tend to bring it about that one believes that p, there is evidence for the claim that imaginings lead to beliefs. In the recent literature on imagination, there has been some debate about the continuum thesis, according to which imaginings are continuous with beliefs (Schellenberg 2013). There are multiple reasons why we should expect the continuum thesis to be true. In fact, I think we should expect something stronger than Schellenberg’s continuum thesis to be true: imaginings are liable to become beliefs. Some reasons to believe this:
(1) cognitive dissonance. Ordinary agents are apt to alter their beliefs to explain their behavior when their behavior is inconsistent with their antecedent beliefs. They do so when they regard that behavior as voluntary. Agents often act in accordance with their imaginings (play acting, for instance). When an agent is immersed in play-acting, the fact that the scenario is imaginary is not salient to them and the mechanisms of cognitive dissonance might kick in.
(2) Spinozist belief formation mechanisms. There is evidence that tokening the thought that p tends to lead to the belief that p. This occurs even when the person is given grounds to undercut any warrant for the belief prior to entertaining it. For instance, Wegner told his subjects that they would receive predetermined feedback about how they performed on a task, and that the feedback would therefore have no relation to their action performance. They formed beliefs that mirrored the feedback nevertheless.
(3) Finally, there is direct evidence of ordinary agents transitioning from knowingly imagining that p to believing that p, even when p has a bizarre content. People who take themselves to be able to ‘channel’ aliens or spirits may begin by pretending to channel these beings. At least that’s an inference supported by the literature which channels have themselves produced on how to channel. They tell novices to stop asking themselves whether they are producing the voices and thoughts themselves: “let that thought go, and for now believe that you have indeed connected with a high-level guide”.
A second example of this same phenomenon is more tragic. Daniel Schacter recounts the story of Paul Ingram, a man accused of abusing his daughters after they ‘recovered’ memories of the events (after a member of their church told them that God had told her of the abuse). Ingram initially denied the accusations, but eventually ‘recalled’ the abuse and confessed. His daughters’ accusations became increasingly bizarre, including the claim that they had been raped and forced to bear children who were then sacrificed in Satanic ceremonies; Ingram confessed even to these bizarre crimes and was sentenced to 20 years’ imprisonment. There is little doubt that Imgram’s imaginings (in which his pastor advised him to engage to ‘recover’ the memories) played a crucial role in his coming to believe he had abused his daughters. Therapists who help patients ‘recover’ memories often ask their clients to imagine what might have happened, saying that many find this helpful to recover the memories.
Of course we might say of these cases what Currie and Gerrans say about delusions: the person doesn’t really believe what they say; they only imagine that they do. But there is little reason to think, though, that someone like Imgram (who confessed to abusing his daughters to police on the basis of recovered memories), didn’t believe what he professed. Its difficult to identify a functional difference between the states he professed and bona fide beliefs.
Imagination – at least insofar as it is appropriate to identify imagination with the mechanisms of mental time travel – seems to have developed by coopting the mechanisms of episodic memory (Suddendorf and Corballis 2007). Further, generating imagery activates the same parts of the brain involved in actual perception in that modality (Van Leeuwen 2013). This may go some way toward explaining our vulnerability to confusing imagination with belief: because we utilize mechanisms that are designed for generating belief, we are disposed to endorse the content of our imaginings under the appropriate conditions. Currie thinks that our capacity to track our own agency plays an important role in allowing us to distinguish imagined thoughts (over which we possess some agential control) from beliefs. But even neurotypical subjects may slip into believing what they initially know themselves to be imagining; for those patients who have problems tracking agency, we ought to expect the mechanism to produce full-blown beliefs.
Perhaps the vulnerability of imagination to slippage is not merely a product of the fact that evolution coopted mechanisms for belief production. Perhaps it comes with the territory. If imagination is to serve its adaptive function, it had better be set up to simulate realistic scenarios. We want to know what would actually happen if we behaved in a particular way, or if some event were to occur. Our imagined scenarios must therefore unfold in way that is as close as possible to how reality would be, given the imagined change. In fact, as Van Leeuwen argues, that’s how minimal imagination – imagination of realistic scenarios – actually works: we import into the world of the pretense background facts about the world, including contingent truths, to fill it out and to constrain our imaginings. To that extent, imagination is itself a mechanism for belief generation, which perhaps renders it less surprising that it may lead us to come to believe in the content imagined, and not just to believe that that’s how the world would be were that content the case.
So we have a problem. Imagined representations tend to become beliefs. But the attraction of the imagination account rests on its generating states that lack the functional role of beliefs. We need some explanation of why delusional subjects continue to imagine that p without ever recognizing that that’s what they’re doing and without coming to believe what they imagine to be the case.
Gerrans’ related account might fill the gap. Roughly, the story is this: delusions are default network thoughts, whereas genuine beliefs are tested for internal and external consistency by DLPC. For one reason or another, the delusional person lacks the capacity to exercise DLPFC supervision over the thought. Why don’t delusional patients come to believe (or to reject) their delusion? Because processing resources are finite. When the delusion is salient, it captures the patient’s attention and monopolizes processing resources. So the thought is not properly tested for consistency with the patient’s other beliefs.
One problem with this account is that seems to attribute more rationality to delusional patients than ordinary people. The rest of us tend come to believe our imaginings; patients don’t (we can’t appeal to the bizarre nature of the delusions to explain why they don’t – we believe bizarre things too, as in the channeling case). It is sometimes objected to two factor accounts of delusion that they postulate reasoning deficits in delusional patients, while the evidence for such deficits is underwhelming. Gerrans needs the evidence to run in the opposite direction, it seems: he needs it to be true that in one respect, delusional patients are more rational than the rest of us. Perhaps this is something he can maintain: he might argue that what is, in some respects, a deficit is nevertheless in other respects protective of the person from a defect of rationality.