Pathologies of volition, attributability and responsibility

Intuitively, it seems that attributability (the degree to which an action can be attributed to an agent) and responsibility are, if not synonymous, nevertheless inextricably linked. That is, the higher the degree of one of them, the higher the degree of the other. When I am compelled to act by an outside force, if I am not responsible for the action it seems that this is because it is not attributable to me. Here I shall argue that though attributability and responsibility come in degrees, they can come apart. I shall use some pathologies of volition (as I shall call them) to illustrate.

Volition, I claim, can be broken down into three separable elements; which together make up attributability. These elements are (i) the extent to which the action stems from the agent’s action plans, which rationalize the action for the agent and for others; (ii) the extent to which the agent is aware of (i), and (iii) the degree to which the action is experienced as willed. I shall expand on these points as I go.

Now consider some pathologies of volition:

(1) Utilization behavior (UB). In utilization behavior, the agent (usually following frontal lobe damage; its also seen in Alzheimer’s) responds, apparently compulsively, to the affordances (roughly, the possibilities for use suggested by an) an object. So if you put a glass of water in front of them, they drink. If you put a pair of spectacles down, they put them on – even if they’re already wearing a pair.

(2) Anarchic (or alien) hand (AH): this is the kind of pathology seen in Dr Strangelove. One hand of an otherwise quite rational and responsive agent responds to affordances all by itself. It may pick up food from someone else’s plate and put in the patient’s mouth, or make a move in a checkers game (in one case, the hand made the move, the patient took it back with his other hand, only for the hand to repeat the move).

(3) Tourette Syndrome (TS): In TS, the agent experiences a build up of tension, which eventually becomes (apparently literally) unbearable. The tension can only be relieved, usually well before the point it becomes unbearable, by engaging in the (often inappropriate) behavior associated with the syndrome: ticcing, but also swearing or insulting others. Obsessive-Compulsive Disorder seems to exhibit a similar symptomology, though here it is the compulsive act which relieves the tension.

Consider, also, one kind of non-pathological behavior. A great many – probably the great majority – of our actions are routinized, and performed without conscious consideration of our options. Here I have in mind such ordinary activities as switching on a light, driving a car, making a cup of coffee. Though we can be aware of our actions as we engage in these activities, we typically are not. We perform them automatically.

Now this kind of automatic activity is attributable to us. Moreover, we are typically responsible for it (we hold people responsible for their driving). So conscious intending of an action is not necessary for attributability or responsibility. It’s not clear to me whether automatic action is experienced as willed. But the other two ingredients of attributability are in place. The action is accord with the agent’s action plans, and (it seems) the agent is aware that this is the case.

We might say that actions with the highest degree of attributability have all three ingredients. Consciously intended actions are usually like this. Automatic actions have enough of the ingredients to be clearly attributable. But the pathologies of volition have fewer of the ingredients.

In UB, the behavior is not in accordance with the patient’s action plans. She has no propositional attitudes which rationalize her behavior to herself or to others. But she is apparently unconcerned; she does not take her action to conflict with her action plans (perhaps her very short-term action plans are themselves environmentally driven). And she apparently experiences her action as willed.

In TS, the action is not in accord with the patient’s action plans. She does not want to insult this passer-by (for instance). But unlike UB, she is aware that her behavior conflicts with her action plans. Finally, like UB she experiences the behavior as willed. Indeed, she is able (to some extent) to choose when she acts. If the stakes are raised and self-control is required, she can hold out much longer (though she must give in eventually).

Finally, in AH all three ingredients are missing. The patient does not experience the action as willed at all. Instead, she experiences her hand as controlled by an alien force.

Intuitively, then, we can say that AH is the least attributable to the agent. I want to suggest that the extent to which the action accords with the action plans is the most significant ingredient of will, so both UB and TS are clearly unwilled. But TS seems to have a higher degree of volition than UB. Why? Because the agent retains the ability to judge whether or not the action accords with her action plan. This element seems to feed into (iii), the extent to which the action is experienced as willed. Obviously, too, an agent can only attempt to exercise self-control if she is aware that something is at stake.

Somewhat paradoxically, however, responsibility diverges somewhat from attributability. I am inclined to regard the long-term sufferer from AH as most responsible – because she retains overall control over her body, and is aware that toward one part of it she must adopt the objective attitude. UB patients are least responsible, since they do not have intact action plans available to them to which they can compare their moment-by-moment actions. And TS sufferers come somewhere in the middle; responsible for just when (and perhaps how) they give in to their urges.

Comments

[Chris Franklin]

Neil

I think that there is some confusion between your comments that in AH all three ingredients are missing and that, paradoxically, you find the long-term sufferer from AH most responsible. The confusion can be seen in the qualification as to why in AH the agent is most responsible. The reason is because the agent retains overall control of his body. However, all three ingredients of volition are missing only in the case of an action brought about by the archaic hand, not any action that issues from the agent. Is that right?

It seems that your conclusion would only demonstrate that attributability and responsibility can come apart if in AH the agent's actions issuing from the activity of the archaic hand lacked all three ingredients of volitions consituting attributability and yet, he could still be appropriately held responsible for the same action.

Finally, I find it difficult to judge the cases when there is no explicit characterization of what responsibility is. I think such an endeavor to distinguish the two would demonstrate all more the relatedness, possibily even (as you stated although worded a little differently) their identity.

Let me know if I am understanding the cases right.

[Joe]

My inclination is that there is not really a problem here at all (though there are quite a few interesting observations).

Attributability is one of at least four necessary conditions for moral responsibility.

Other necessary conditions include: (a) epistemological conditions, (b) freedom relevant conditions, and (c) additional cognitive conditions. I don't want to suggest that these conditions are well defined, or that they do not over-lap. But it seems to me that in each of the (very interesting) cases that you mention above, one of the other necessary conditions—(a), (b), or (c)—for moral responsibility that I have mentioned is violated.

Perhaps I've missed something, though!?!

[Neil]

I think the really hard one is TS. In TS, agents experience their actions as voluntary - "I do the ticcing". Often, they are morally problematic; eg, shouting a racist epithet (sometimes, violence to self or others can also be traced to TS). Moreover, TS sufferers do have a degree of control over their actions: they can (with effort) suppress the tic, for a greater or shorter period. People close to TS sufferers are generally disposed to excuse them of responsibility for their actions. Should they be, and why?

Tim Schroeder has a paper coming out in Phil & Phen Res, in which he argues that we ought to excuse in these cases because their actions do not reflect the quality of their will. Schroeder notes that TS is a product of disorder of the basal ganglia, which is involved in the production of actions through operant conditioning. The hypothesis is that in normal individuals the basal ganglia inhibits the motor cortex commands, whereas in TS the inhibition fails, and the motor cortex issues commands to perform conditioned actions. These commands do not lead directly to action, but are difficult to resist. In resisting, Schroeder claims, the agent demonstrates a good will, but the urge itself, since it is a conditioned or automatic response, should not be seen as a desire and therefore does not reflect a bad will.

Its an interesting account. But I'm not convinced that we should excuse in TS *because the urge is not an expression of the person's will.* Suppose the TS sufferer is racist, but attempts to suppress the racist epithet because he knows it is socially inappropriate. Surely there is some tendency to diminish repsonsibility? What about cases in which he is not racist, but has had a racist upbringing - so that his racism is able to lead to first-order, but not higher-order, desires? Finally, what about normals in similar kinds of cases? Mightn't they, too, feel strong urges to express morally problematic views, which they know to be morally problematic (think of the perverse urge we sometimes feel to laugh at funerals).

I think the best explanation of excuse in TS will instead turn on the difficulty of suppressing the urge, not its origin. But if it is the difficulty of expressing the urge, then I see no reason to limit the excuse to cases of TS. In fact, peremptory urges (indeed, sometimes with basal ganglia involvement) are a feature of many impulse-control disorders (eg. kleptomania and pedophilia) and probably of the immoral desires of perfectly normal people as well. Of course, the stakes involved in, say, pedophilia, are much higher than in TS, so we rightly demand much more strenuous resistance from sufferers.

[Lindsay Beyerstein]

Very interesting post, Neil. I have a question about accountability over time. What would your theory say about accountability for behaviors which stem from dispositions that the agent was responsible for before he became incapacitated?

For example, When I volunteered in a senior's home, there were a couple of really nasty older guys who would have been condemned as bigoted sexual harassers if they hadn't also been suffering from brain damage. At the time, the guys weren't really agents, so we couldn't blame their current selves for their current infractions.

I often wondered whether it could ever be reasonable to blame a person's former self for their current automatic or anarchic behavior. Obviously, you wouldn't blame someone whose beliefs and desires changed as the result of their brain injury. But what about people who cultivated bad character traits all their lives? Before their brain injuries, we would have held them responsible for the types of sexist and racist attitudes that they acted out after they became disinhibited by brain malfunctions. Or, in the case of Dr. Strangelove, DS may have been accountable (at some point) for his hand going Nazi on him, even though the saluting at time T wasn't attributable to him.

[Eddy Nahmias]

Lindsay raises an interesting question. Consider that it seems like we have two basic ways of supressing our bad behavior. First (and perhaps more common), we don't act on the bad thoughts or desires we have. A sexist comment comes to mind but we don't say it, etc. One common way to think about this is that we have second-order volitions not to act on certain first-order desires. Second (and seemingly more difficult), we try to change the way we think and the desires we have. We try to suppress certain thoughts or habituate ourselves to react in certain ways. We have second-order desires to eradicate certain first-order desires. This sketchy account certainly doesn't do justice to the complexity of this form of self-control.

But it seems plausible that certain brain disorders, perhaps some brought on by senility, or even 'cognitive overload' in normals, will undermine one's ability to inhibit one's behavior when bad desires or thoughts arise. In such cases, as Lindsay suggests, we may want to hold the agent accountable because they did not do enough (when they were not cognitively impaired or overloaded) of the second kind of suppression. They did not suppress their bad thoughts and desires, or habituate themselves to 'think right', etc.

Of course, we know enough (as did Aristotle) to recognize that most of this habituation occurs before we are fully autonomous, so the historical aspect of tracing back accountability will often extend to the parents and teachers of the agent. As a relatively new parent I certainly feel like I have greater freedom and responsibility in instilling the right values in my children than I do in changing my well-ingrained values and habits. Frankfurt's hierarchical model has always seemed to me to be more effective cross-generationally than intra-individually--that is, second-order volitions to act on or to develop first-order desires (or to eradicate first-order desires) seem much more effective when the first-order desires are in (or not yet in) a pre-habituated youngster. Paternalism at its best! (Similarly, I think that given our tendency to act on immediate desires at the expense of our future desired goods, it also seems appropriate for us to ask our leaders to be paternalistic by enforcing our long-term interests--and our children's--even when this conflicts with our immediate desires. Suppressing our addition to oil comes to mind). On that political note, I'll sign off.