It is a common idea, inside and outside philosophy, that agents with serious mental disorders impairing agency might not be fully blameworthy for their ethically problematic actions when these are caused by their mental disorders. In a case where, for example, George with schizophrenia attacks his neighbours because of his delusion that the neighbours are trying to kill him, he might not be fully blameworthy for the attacking.Footnote 5 This, of course, does not mean that the action is acceptable. Attacking neighbours is obviously ethically problematic. The idea is rather that George, the agent, is not fully blameworthy for his ethically problematic action.
In this section, we defend a similar claim with respect to whether agents are blameworthy for believing delusional hypotheses.Footnote 6 The claim is that agents with delusions are, at least in many cases, not blameworthy for their delusional beliefs. George, for instance, is not blameworthy for the delusional belief that his neighbours are trying to kill him. This, of course, does not mean that the delusion is not epistemically faulty. In the previous section, we argued that delusions have some epistemic benefits. At the same time, though, we acknowledged that delusions have several epistemic shortcomings; most of them are false and, even when they are not, they are accidentally true at best and, thus, they do not amount to knowledge.Footnote 7 Moreover, we noticed that delusions often do not cohere well with the agents’ other beliefs and are resistant to counterevidence. Thus, we do not wish to argue that delusions are epistemically unproblematic. Rather, we argue that agents are, at least in many cases, not blameworthy for their delusions.
In general terms, when are agents blameworthy for epistemically faulty beliefs, whether these are delusional or not? There are two extreme answers to this question. The first answer is that agents are always blameworthy for their epistemically faulty beliefs. Let us call this first view “universalism”. The second answer is that agents are never blameworthy for their epistemically faulty beliefs. Let us call this second view “nihilism”. And there is a third, modest answer according to which agents are sometimes, but not always, blameworthy for their epistemically faulty beliefs. We call this third view “restrictivism”. In the following discussion, we presuppose that restrictivism is true. The main challenge to restrictivism comes from the famous argument in Alston (1989): (1) doxastic blameworthiness implies voluntary control over beliefs; (2) agents do not have voluntary control over any of their beliefs; (3) therefore, they are never blameworthy for any of their beliefs.Footnote 8 There are many critical responses to the argument in the literature,Footnote 9 but there is no agreement about what exactly is wrong in the argument.
We assume restrictivism because we find it independently plausible. In addition, restrictivism makes our claim interesting. On the one hand, our claim that agents with delusions are, at least in many cases, not blameworthy for their delusional beliefs is true, but only trivially, if nihilism is true. If agents are never blameworthy for their epistemically faulty beliefs, then it is trivially true that agents are not blameworthy for their delusional beliefs. On the other hand, our claim is guaranteed to be false if universalism is true. If agents are always blameworthy for their epistemically faulty beliefs, then it is guaranteed to be false that agents are not blameworthy for their delusional beliefs. Thus, there is no room for a serious discussion about our claim within nihilism or universalism. According to restrictivism, however, there is room for a serious discussion about it. In some cases, agents are blameworthy for their epistemically faulty beliefs. In other cases, they are not. Then, there is room for investigating whether the case of an agent with delusional beliefs falls in the first or the second category.
On the Ability to Believe Otherwise
The crucial notion in our argument is that of the ability to believe otherwise, which we define in the following way:
S, with an epistemically faulty belief that p, was able to believe otherwise iff there is an evidential action A, which was available to S, such that if S had performed (not performed) A, S would have believed otherwise.
An action is an “evidential action” in case it is directly related to searching, gathering, or processing evidence.Footnote 10 Evidential actions include: reading newspapers, paying attention to some objects, careful reasoning, and so on. An evidential action A is “available” to S in case S would have successfully performed A if S had tried to do so. Reading the newspaper in front of me is available to me because I would have successfully read the newspaper if I had tried to do so. S, with the belief that p, “believes otherwise” in case S fails to believe that p. To see how this notion actually works, think about the following two cases.
A minister of defense has some reason to think that large-scale fraud has been perpetrated in the army. A committee is formed which studies the presumed fraud in detail and writes a long report about the situation, which they hand over to him. The minister, however, spends most of his time relaxing on the beach behind his house, enjoying the sun and eating pizzas. Consequently, he does not read a single letter of the report. As a result of that, he is ignorant of the fraud. For instance, he holds certain false beliefs about the army, beliefs for which he might nonetheless have good evidence, as long as he does not read the dossier.Footnote 11
Mark is writing out the shopping list for the weekly grocery shop. He goes to the fridge and sees that there is a carton of orange juice in the fridge. He forms the belief that there is orange juice in the fridge, and hence that he does not need to buy orange juice. As it turns out both of these beliefs are false. One of his housemates finishes off the orange juice, but stupidly put the empty carton back in the fridge. When Mark finds this out, he is irritated at his housemate, but he is also irritated at himself. He did not have to draw the conclusion that there was orange juice in the fridge. He was, after all, living in a student house where people do all sorts of dumb things. That his housemate might have returned an empty container to the fridge was well within the range of live possibilities.Footnote 12
The defence minister was able to believe otherwise. Reading the report is an evidential action that was perfectly available to him. And, if he had read the report, he would have believed otherwise. Again, Mark was able to believe otherwise. Checking the container is an evidential action that was perfectly available to him. And, if he had checked the container, he would have believed otherwise.
It is tempting to think that there is an intimate connection between doxastic blameworthiness and the ability to believe otherwise. In particular, one might think that doxastic blameworthiness implies the ability to believe otherwise.
Principle of Alternative Believability (PAB)
If S is blameworthy for her belief that p, S was able to believe otherwise.Footnote 13
Let us, first, look at the examples above. The defence minister seems to be clearly blameworthy for his false beliefs about the army. And, as we just noted, he was able to believe otherwise. Again, Mark seems to be blameworthy for his false belief about the orange juice, given the fact that, in his household, it was a live possibility (and he knew it) that an empty container is stupidly put back in the fridge. And, as we just noted, he was able to believe otherwise.
Now let us think about a slightly different version of Defence Minister, “Defence Minister 2”, where the defence minister tried to read the report, which is a PDF file, but failed because it was protected with a password that the minister did not know due to some administrative errors by his secretary. The minister consequently retained false beliefs about the army. Assuming that the minister had no alternative source of information about the fraud in the army, he was not able to believe otherwise. Thus, the minister does not seem to be blameworthy for his false beliefs in this case.
Let us consider, again, a slightly different version of Orange Juice, “Orange Juice 2”, where one of Mark’s housemates, who is a magician, finished off the orange juice in the container and put water into the container instead. Next, the water was transformed into something that was totally indistinguishable from orange juice by his magic. Mark checked the container and, being fooled by the magic, retained his belief that there was orange juice in the fridge. Assuming that the Mark had no alternative source of information about the orange juice, Mark was not able to believe otherwise. Mark does not seem to be blameworthy for his false belief in this case.
These cases are therefore consistent with PAB. However, PAB is controversial. A problem comes from notorious Frankfurt-style cases.Footnote 14 Let us consider, for example, another version of Defence Minister, “Defence Minister 3”, where, just like in the original Defence Minister, the minister did not even try to read the report and, consequently, he retained his false beliefs about the army. Unbeknown to the minister, just like in Defence Minister 2, the report was protected with a password that the minister did not know and, accordingly, even if he had tried to read it, he would have still believed the same false propositions about the army. In this case, the minister seems to be blameworthy for his false beliefs. After all, he thinks and behaves in Defence Minister 3 (i.e. spending most of his time relaxing on the beach without trying to read the report at all) in exactly the same way as in Defence Minister where he is clearly blameworthy. There does not seem to be any relevant difference between two cases that makes the minister blameworthy in one case but not in the other. However, assuming that the minister had no alternative source of information, he was not able to believe otherwise because of the password protection. Thus, Defence Minister 3 looks like a counterexample to PAB.
Another possible counterexample is the case where the minister was not able to believe otherwise but the minister himself was responsible for the inability to believe otherwise. Let us consider, for example, “Defence Minister 4” in which, just like in Defence Minister 2, the minister tried to read the report, but fails because it is protected with a password that the minister does not know. As a result, he retained his false beliefs about the army. In this case, however, the password had already been given to the minister. He had just lost it because of his laziness. The password was confidential and it could not be retrieved or reissued. In this case, it looks as though the minister is blameworthy for his false beliefs about the army even though he was not able to believe otherwise. Thus, Defence Minister 4 looks like another counterexample to PAB.Footnote 15
One might think, however, that Defence Minister 4 is not a counterexample to PAB because the minister was able to believe otherwise in this case. For instance, if the minister had kept the password, he would have believed otherwise. This objection fails, however, because keeping a password is not an evidential action in our sense. As we already noted, evidential actions are the ones that are directly related to searching, gathering, or processing evidence. Reading an email is an evidential action. Arranging one’s inbox is not. Keeping a password is, just like arranging one’s inbox, related to evidence only indirectly.
To deal with these cases, we retreat to the following claim:
If S is blameworthy for her belief that p, then either (a) it is not the case that S believes that p (partly) because S was not able to believe otherwise, or (b) S is responsible for the fact that S was not able to believe otherwise.
(The right hand side of PAB2 is trivially true in the case where S is able to believe otherwise, since (a) is trivially true in such a case, and it makes the disjunction (a) or (b) trivially true as well.)
Defence Minister 3 is not a counterexample to PAB2 because its right hand side is true in the case. Certainly, the minister was not able to believe otherwise. However, this is due to the fact that the report was locked with a password that the minister did not know, which had nothing to do with minister’s believing the false propositions. Thus, it is not the case that the minister believes the false propositions because he was not able to believe otherwise, which means that the clause (a) is satisfied in this case.
Defence Minister 4 is not a counterexample either because, again, the right hand side is true. The minister was not able to believe otherwise and, in addition, he believed the false propositions partly because he was not able to believe otherwise due to password protection (thus, the clause (a) is not satisfied). However, he is responsible for the fact that he was not able to believe otherwise, which means that the clause (b) is satisfied in this case. He is the person who is responsible for the loss of the password, and losing the password is what makes it the case that he is not able to believe otherwise.
Here is our argument for the claim that agents are, at least in many cases, not blameworthy for their delusional beliefs. Assuming that an agent Sd has a delusional belief that p:
Sd believes that p partly because Sd was not able to believe otherwise.
Sd is not responsible for the fact that Sd was not able to believe otherwise.
If Sd is blameworthy for her belief that p, then either (a) it is not the case that Sd believes that p (partly) because Sd was not able to believe otherwise, or (b) Sd is responsible for the fact that Sd was not able to believe otherwise.
Therefore, Sd is not blameworthy for believing p.
We maintain that this argument is applicable to at least many agents with delusions. The third premise directly comes from PAB2, which we take to be reasonable enough. At the very least, it is not refuted by Frankfurt-style cases. In the following discussion, we aim to support the first and the second premises.
To support these premises, first, we claim that many agents with delusions are not able to believe otherwise or, at the very least, that their ability to believe otherwise is significantly compromised due to some impairments and biases. Second, we argue that the inability to believe otherwise is explanatory of their believing delusional hypotheses (and, thus, the first premise is true), and also that they are not responsible for the fact that they are not able to believe otherwise (and, thus, the second premise is true).Footnote 16
The first factor that compromises the ability to believe otherwise is the inability to regard relevant alternative hypotheses as live possibilities. If one believes a certain hypothesis, but fails to regard alternative hypotheses as live possibilities, then evidential actions are not likely to lead to believing otherwise. Suppose that the defence minister failed to find it a live possibility that large-scale fraud had been perpetrated in the army. In this case, he would not have believed otherwise even if he had read the report. He might have concluded that the report was unreliable. In the study by Freeman et al. (2004), research participants with delusions were asked to come up with alternative, non-delusional explanations of the events on the basis of which their delusions were formed and maintained. Surprisingly, three quarters of the participants failed to report any alternative hypotheses.Footnote 17
Three quarters of the patients reported that there was no alternative explanation for their experiences. The delusion was their only explanation. This matches with clinical experience. Nevertheless, it is a striking finding. By definition a delusional belief is highly improbable. The evidence cited for a delusion is, at best, ambiguous. Yet most individuals could not report any potential alternative explanation for the ambiguous evidence however unlikely that they considered the alternative. (Freeman et al. 2004, p. 677)
The second factor is the inability to examine competing hypotheses carefully. If one believes a certain hypothesis, but is not able to examine competing hypotheses very carefully, then it is not likely that evidential actions will lead to believing otherwise. Suppose that the defence minister is not capable of examining the fraud hypothesis carefully. In this case, he might have failed to believe otherwise even if he had read the report. Huq et al. (1988) revealed that agents with delusions have the bias of coming to conclusions with less evidence in comparison to controls (the jumping-to-conclusion bias). In their experiment, research participants (delusional group, non-delusional clinical group, and non-clinical group) were requested to identify if a given jar was jar A, which contained 85 pink and 15 green beads, or jar B, which contained 15 pink and 85 green beads, on the basis of the observation of the beads drawn from the jar. They found that participants in the delusional group came to a conclusion with less information (2.22 beads drawn from the jar on average) than participants in different groups (3.6 and 4.58 beads for non-clinical group and non-delusional clinical group respectively).
Relatedly, agents with delusions have strong desire to reach conclusions. In a study by Colbert and Peters (2002), research participants (non-clinical individuals) were asked to indicate the extent to which they agreed with statements concerning need for closure, which is “the desire for a definite answer on some topic, any answer compared to confusion and ambiguity” (Kruglanski 1989, p. 14). They found that the tendency to delusional ideation is statistically correlated with a high need for closure (although need for closure was not correlated with the tendency to jump to conclusions). Need for closure was also found to be strong in clinical individuals with current or remitted persecutory delusions (Bentall and Swarbrick 2003).
The third factor comes from the motivational states (such as desires and emotions) that drive delusional hypotheses. If an agent’s belief is influenced by strong motivational states, then her evidential actions are not likely to lead to believing otherwise. Let us consider the scenario in Defence Minister again and suppose that the minister’s false beliefs about the army were motivated by his strong desire for the soundness of the army. In that case, he would not have believed otherwise even if he had read the report.Footnote 18 The idea that some delusions are driven by motivational states has received some attention in the recent literature. As we observed in Sect. 1, so-called motivated delusions, such as Reverse Othello syndrome, erotomania and anosognosia, are likely to be driven by motivational forces.
Further, it has been suggested that persecutory delusions are the products of an externalising attribution bias, namely, the bias an agent is subject to if she attributes negative events to other agents rather than to herself. In a study by Kaney and Bentall (1989), research participants (delusional group, depressed group, and non-clinical group) were requested to answer questions about the causes of positive and negative events. They found that participants in the depressed group tended to attribute negative events to themselves, while participants in the delusional group attributed negative events to external causes and positive events to themselves. This externalising bias could be easily construed as a defence mechanism. For instance, one could argue that the bias is a means to protecting self-esteem.
[…] in normal individuals, the tendency to attribute the cause of negative events to external factors maintains self-esteem through the abrogation of responsibility. Since this tendency is significantly more marked in people with persecutory delusions, such delusions can be seen as an extreme method of maintaining self-esteem, a hypothesis which is consistent with Zigler and Glick’s suggestion that paranoia is a form of camouflaged depression. (Bentall et al. 1994, p. 334)
The existence of these three factors strongly suggests that many agents with delusions are not able to believe otherwise or, at the very least, that their ability to believe otherwise is significantly compromised.Footnote 19
Now, it is likely that agents believe delusional hypotheses partly because of their inability to believe otherwise, given the fact that (1) they are unable to think about alternative hypotheses; (2) they are unable to examine alternative hypotheses carefully due to reasoning biases; and (3) there are motivational factors favouring their delusional hypotheses. Although the whole process of delusion formation and maintenance is not fully understood, it would be fair to expect that such impairments and biases play some role in the process. Thus, the first premise of the argument seems to be true. Moreover, agents do not seem to be responsible for the fact that they are not able to believe otherwise. After all, they are not responsible for having those impairments and biases in most cases.