Self-ignorance

Doris says that human beings are “afflicted with a remarkable degree of self-ignorance” (précis abstract). But while we certainly at times show self-ignorance, there is no absolute metric that allows us to assess the exact degree of our ignorance compared to our self-knowledge. This opens the possibility for researchers, who feed on a steady diet of examples of ignorance, to overestimate its degree. We need to leave open, for example, the possibility that we are dealing not with phenomena that afflict everyone, but with phenomena that only afflict a minority of people, or even a certain personality type. The scope of Doris's skepticism is also broader than it might appear. One sign that we might be overestimating the amounts of ignorance and error is that we have not been moved to enact major changes in folk-psychology to remove dependence on our capacity for self-knowledge. Doris's view seems to commit us not only to being “routinely mistaken” (précis abstract), but also not ever noticing that we are, and attempting to correct it. Doris seems to be neglecting all those times we aren't buffoons.

A comparison with the case of visual perception is illuminating. Even though cognitive scientists have cataloged perhaps hundreds of visual illusions that reveal the seams and flaws of our visual system, the vast majority of our visual perceptions during the day are veridical and serve us quite well. Consider our abilities to visually identify one another. Certainly there are many ways in which the brain systems that achieve this miracle can fail, leading to odd syndromes like prosopagnosia. In the everyday sphere, we have all experienced cases in which we visually misidentified someone. But taken against the overwhelming percentage of correct identifications we make so effortlessly and frequently, these misperceptions are rare. This high rate of effectiveness is due to good equipment.

We think serious cases of ignorance or mistaken self-knowledge are somewhat rare because they typically involve errors at two levels. First, a mistaken impression is created. For instance, it occurs to me that I don't really have to pay back that loan from my friend because he seems to be wealthy, when I would just prefer to keep the money. Then, this error is not corrected (this correction could occur because I note my obligation to repay, or I revise my sense of my friend's situation, or I just realize I am being selfish). The first type of error, where I form a mistaken impression of my own motives, is fairly common; the second, where I fail to correct, or at least where I fail to correct because I cannot correct, less so. And in cases where we have the capacity to correct for our mistaken perceptions, using our brain's prefrontal executive processes, it would seem we are responsible for them (Hirstein et al. Reference Hirstein, Sifferd and Fagan2018). For example, a color-blind person can correct for his problem by memorizing the location of the traffic lights. Doris's view amounts to saying that the entire upper level that has been designed into our brains, including the executive processes and consciousness itself, is of little use or import. This level functions precisely to correct basic errors of perception or memory, as can be seen in the case of confabulation (Hirstein Reference Hirstein2005). This second level tends to only activate when the stakes are appropriately high, so that examples of our failures where they aren't perceived to be high, such as the case of people failing to put money into the office coffee fund, are not showing our cognitive system at its best.

Synchronic assessments of responsibility

Doris argues that moral responsibility for an act depends upon whether the act in question was an exercise of agency (Doris Reference Doris2015b, p. 159). Exercises of agency, according to Doris, are expressions of the actor's values; attributions of responsibility turn on whether an actor's values are expressed in an act (p. 159). However, this sort of view faces clear epistemological difficulties, as Doris notes: It will frequently be difficult to determine whether someone holds a value, and actions often seem related to multiple values, some of which may be unknown even to the actor. Plus, “values are expressed over time, and can, oftentimes, only be identified over time,” and thus “extended observations” may be required to identify patterns to determine if any particular action is of the sort for which an agent can be held responsible. “If one focuses on isolated events, diagnoses may falter” (Doris Reference Doris2015b, p. 162). In the end, attribution of responsibility may require first that “a pattern of cognition, rationalization, and behavior emerges, and that pattern is best explained as involving the expression of some value”; and second, a determination that a particular action expresses that value (p. 164). But why in a revolutionary era of neuroscience assume that we must remain forever locked outside the mind and brain of the subject? Doris's account involves the cognitive sciences, but only those that focus on behavior and outward from there, to society. We suggest that connecting his knowledge of the psychological research with neuroscience, via cognitive neuropsychology, would greatly help resolve the epistemic problems involved in discerning what exactly someone's values are.

As it stands, Doris's theory indicates that synchronic assessments of responsibility are often impossible. However, the most common and important responsibility attributions are synchronic. Take, for example, criminal verdicts. Judges and juries do not, and ought not in most cases, focus on past behavior as a means to indicate responsibility for a particular crime.Footnote ¹ A criminal court is asked to determine whether a defendant held a particular mental state and whether this mental state is causally related to the criminal harm. Such canonical cases of responsibility attribution are considered so secure we use them to deny defendants' liberty and even life. If Doris's theory is correct, and responsibility assessments rest on extended investigations into a person's values, then it would seem our current system of generating verdicts and punishing offenders is likely to attribute responsibility to persons when it has not been proven they deserve blame.

Doris indicates that he is a pluralist about responsibility, and thus “sympathetic” to the possibility that there may sometimes be warranted attributions of other types of responsible agency, including reflectivist agency (Doris Reference Doris2015b, p. 174). However, he also feels that a pluralistic account must place dialogic agency in an “appropriately prominent” position (p. 175). To vindicate the thrust of his theory with regard to criminal verdicts, Doris should provide an account of how a synchronic act must be related to dialogic agency. Further, this account must explain how a synchronic act can be seen as an expression of such agency without an exhaustive review of the agent's history. But if this were possible, then it would seem that Doris's requirement of “extended observations” would, in most cases, be unnecessary because a less burdensome, synchronic assessment would suffice.

In a similar vein, the reactive attitudes, which Doris acknowledges are important first indicators of responsible agency (Doris Reference Doris2015b, pp. 23–24), are typically generated in synchronic cases without information about character. They depend on the brain's mindreading (or theory of mind) capacities, through which we attribute motives behind a person's actions, sometimes using fairly few behavioral cues. If these motives are selfish, for example, a strong negative reactive attitude will follow. In the criminal law, we feel stronger condemnation where an agent directly desired criminal harm (committed the act “purposely” under the U.S. Model Penal Code) than in cases where an agent merely ought to have known there was a risk of substantial harm (committed the act “recklessly”).

It isn't clear that Doris's weakly proposed pluralism, which encompasses his dialogic view and reflectivism (Doris Reference Doris2015b, p. 174), can generate many of the synchronic responsibility assessments made in the criminal law. As Doris argues, many culpable actions do not seem connected in the right way to reflective judgments, which are often confabulated. Thus, if extensive investigation of dialogic agency is not done, on what grounds are criminal verdicts generated? For example, in a case where the fire was due to the building owner's forgetting to check the functioning of the water sprinklers, a synchronic assessment of the defendant's conscious mental states with regard to the criminal harm will not secure a responsibility assessment. In our view, only an account that provides a synchronic assessment of capacity for responsible agency, where that capacity is more expansive than just the capacity for conscious reflection, can ground criminal verdicts of negligence.

Personal versus subpersonal

As we noted, Doris chooses to keep his analysis at the personal, rather than the subpersonal level, by using information largely from social psychology. But sometimes, simple knowledge of the person's brain can clear things up. For example, Doris notes that “the valuational account says if your action properly expresses your values, it's an exercise of agency, regardless of whence your values came” (Doris Reference Doris2015b, p. 30). But what about someone with Tourette's whose outbursts do express his values, but not in a way he wanted? Or a person whose sleepwalking actions do express his values, but are horrible, and which he would never do when awake? In both of these cases, responsibility does not seem to rest with the actor, due to volitional incapacity, despite the alignment of the action with the actor's values. If we could “see” the actors lack of control via evidence of brain function (or dysfunction), we might correct mistaken assessments of responsibility.

Doris searches everywhere for help in attributing psychological states such as motives, including other people (the dialogic part of this theory), except in neuroscience. There is useful information at the subpersonal level, from neuroscience, cognitive neuropsychology, and from historical neurology, that is vital to gaining a full understanding of the relevant phenomena. Neuroscience can provide valuable data regarding synchronic assessments of responsibility. For instance, it might be able to tell whether an action is “done habitually” (i.e., what the neuroscientists call an action done “in routine mode”) or done as a result of conscious reflection, which involves quite different and more extensive brain processes. While Doris avows materialism, it is difficult to see how his theory, as stated, can be put into stark, materialistic terms. What concrete things, states, processes, and events do claims about “values,” “desires,” “plans,” “self-awareness,” and “the exercise of agency” refer to? We are not done with the project of building a theory of responsibility until we can do that.