I had some thoughts on responsibility and mental disorders today. I’m going to approach the subject from the three frameworks discussed earlier: character theory (deep-self theory), control theory, and the conversational approach (viz., McKenna).
On the earliest versions of character theory proposed by Frankfurt (1969, 1971) and Watson (1987), internal consistency is the hallmark of responsible agency. A responsible agent has internally consistent desires and/or values. Internal fragmentation is the sign of lack of responsible agency. A person also needs higher-order reflective agency, of course, but this agency functions to consolidate one’s mental states into a coherent framework. From a certain natural perspective, this makes good sense. The unwilling alcoholic is not responsible because his alcoholic impulses are recalcitrant, external to his set of internally-consistent desires (see Frankfurt 1971). By the same token, someone with schizophrenia is not responsible at all, because the person is too deeply psychologically fragmented for any ‘deep self’ to exist (on the standard logic). This may explain why Strawson held that “psychologically abnormal” individuals are not responsible agents; they warrant the ‘objective attitude,’ as opposed to the ‘participatory perspective’ that we take to all members of the moral community. To psychologically disordered individuals, we withdraw the reactive attitudes (praise, blame, approbation, resentment, etc.)
Prima facie, the control view seems to exclude psychologically disordered people as well, insofar as these individuals lack control over their choices and actions. If we assume something like a Strawsonian picture, it is natural to think that either we see someone as a member of the moral community, deserving of the reactive attitudes, or we see someone as external to this community. This is the natural perspective on a ‘threshold’ account of responsibility, on which an agent is either a responsible agent or not. And this is the stance that our medical and juridical institutions take. Either a person is mentally ill, just in case she satisfied the diagnostic criteria of the DSM-5, or she is mentally well. Either a person has decisional capacity and decision-making authority, or she does not. On a control view, we might similarly think that people with mental illness lack the sufficient degree of control to count as responsible agents, and thus cannot be held responsible, tout court.
There are obvious problems with the all-or-nothing view, some of which I addressed in my last post. But here, I want to show that the character and control view can accommodate the idea that people with psychological disorders have responsible agency, but it is a domain-specific kind of agency. However, we need to assume a notion of responsibility that eschews the criterion of psychological congruence, or at least allows for local (as opposed to global) congruence.
The first thing to note on this score is that it is reasonable to think that all kinds of agency are domain-specific. As Doris noted in an early paper, character often does not transfer from one context to another, and intelligence appears to be domain-specific: “mathematical acuity in the classroom, for example, may not strongly correlate with ability to apply similar computations in work-related settings” (1998: 511). Doris develops this view further in his 2008 and 2014 books. The general idea is that character traits and control are domain-specific, and do not translate smoothly across different contexts. If this is right, then even psychologically normal people may be quite internally fragmented. Susan at a party may be different than Susan at work, who may be different than Susan at the playground with her children. This insight helps to explain how some people can be functional alcoholics: they are alcoholics in private and high-performing career-people at the office. Caroline Knapp, herself once a high-functioning alcoholic journalist, speculates that most alcoholics might be the high-functioning type (‘Drinking: A Love Story’ ). I can’t say one way or the other, but certainly many alcoholics are very functional in certain capacities: Hemingway, Humter S. Thompson, and Tennessee Williams are just a few examples of alcoholics who were some of the world’s greatest writers. This would not make sense if human capacities were not highly context-sensitive and compartmentalised.
Now, if situationist psychology and recent research in cognitive science is to be believed, ordinary people’s competencies are quite domain-specific, and this may be highly adaptive – it allows us to perform well in our various roles. But this domain-specificity may be even more adaptive in people with mental illness, inasmuch as it may permit them to excel in certain capacities in spite of having a psychological disorder that impairs some of their basic functions – functions needed to live independently in some cases. I once dated someone who had previously been hospitalised for a year with schizophrenia, and during that time wrote a 4oo-page autobiography. He had an impressive, indomitable writing capacity, even though he could not, during his psychotic episode, live a normal life. Van Gogh was mentally ill and a great painter. Some philosophers, I am sure, have a clinically-significant mental illness. (See Daily Nous’ post on philosophy and depression). The point of the argument about domain-specificity is that mental illness does not necessarily impair all of a person’s capacities – it does not cause global impairment. If this is right, then we cannot endorse a global psychological congruence principle, which might be suggested by certain formulations of character theory and control theory. And we cannot endorse an all-or-nothing interpretation of the reactive attitudes, since a person is not either psychologically congruous or psychologically fragmented: everyone is psychologically fragmented. We need to assess, not the person, but the person’s capacities. A person might be responsible for capacity A but not for capacity B (or for lacking capacity B). My friend may have been praiseworthy (and thus responsible) for writing an insightful autobiography that captures life in mental health systems, and yet not responsible for certain actions he took or failed to take only as a result of schizophrenic symptoms.
I think that it is easy to lose sight of the lessons of situationist psychology, because we all have a natural tendency to construe a person’s character as situation-invariant, and even if we are aware of the relevant literature, this tendency persists. Like many Canadians, I was shocked when Gian Gomeshi, a staunch feminist in his public life as a CBC radio host, was accused of sexual assault by multiple women and men, with allegations stemming all the way back to his academic tenure at my Alma Mater, York University. I was shocked because I had implicitly assumed that his public persona was also his private persona – his only persona – and this was a false assumption. And I am in the best position to know that character is domain-specific; yet somehow this fails to register in my daily dealings with people.
Similarly, if we know that someone is ‘mentally ill,’ it may be easy to assume that this person is dysfunctional in a domain-general way. And perhaps this explains a lot of the stigma and prejudice against people with psychological disorders. But we need to resist this way of thinking. We need to be vigilant about our implicit biases, and we need to talk and write about agency in a properly nuanced manner. I think that most philosophers do this, but it is worth having still more conversations about agency and the reactive attitudes.
I haven’t said anything about the conversational model advanced by McKenna, so I will say a few words here in closing. The conversational model is, I think, more permissive than the character theory or the control theory, because it only requires that the agent be capable of engaging in conversation to be responsible. This capacity is more basic than the capacity for character or control (on the strict conception of these capacities that I have been discussing). Arguably, even children have this capacity to some extent. So this model might have less troubling implications for cases of mental illness than the other two. One thing that I think this view captures particularly well – or at least, is capable of accommodating – is the role that conversation can play in enhancing agency. Research shows that talk therapy can enhance treatment outcomes. Likewise, peer-support is shown to enhance outcomes (CAMH). As I have argued elsewhere (2016), the efficacy of conversational, interpersonal treatments may be due to their ability to foster positive narratives about agency and recovery, which are agency-enhancing. Vargas comments briefly in his book on the role of positive ‘scrips’ in agency cultivation (2013). If this is right, the the conversational model might best capture responsibility-attribution as an agency-enhancing practice, provided that the conversation follows certain principles (e.g., is respectful).
In sum, I think that all of the models discussed here can accomodate the idea that mental illness does not entail lack of global responsibility, on the condition that we eschew any commitment to global capacities; yet this commitment is, I believe, deeply psychologically entrenched. Nonetheless, we can overcome it with a little cognitive effort.