Macnamara's Map of Moral Responsibility

Philosophical Studies is definitely on a roll lately with cool papers relevant to the Garden: Patrick Todd's on manipulation arguments that I brought up in my last post, a newer one on Frankfurt-examples by Chris Franklin, and also one on the nature of holding responsible by Coleen Macnamara. (I can't resist pointing out that all three of these fine individuals are in the department at UC Riverside. The people affiliated with that department sure are smart!)

Although you should definitely read all of the above articles, here I just want to talk a bit about Coleen's paper, and then ask a question for all you readers out there.

In her paper, Macnamara gives "a topology of the terrain of holding others responsible" (p. 3), and it turns out to be quite complex and fascinating. She argues that accepting her map helps both to rectify certain confusions in the literature and to "clarify the scope and nature of our practices of holding responsible" (p. 3). A thumbnail sketch of her map, along with a question, appears below the fold:

As she points out, her map consists of three concentric circles, the innermost of which is itself made up of at least two parts.

Circle 1, the largest: Inside this circle fall all of the attitudes and activities that constitute Strawson's "participant stance", roughly, the stance we take toward other people with whom we are involved in interpersonal relationships. These attitudes include all those that are involved in what Macnamara labels "regarding another as a responsible agent" -- a stance that is contrasted with treating another as an object to be treated, manipulated, and so on.

Circle 2, inside circle 1: Inside this circle are all the attitudes and activities that are centered around deontic and evaluative concepts such as good/bad, right/wrong, and virtuous/vicious. This class is narrower than the class of attitudes that make up the participant stance in general -- for example, accepting promises is a participant-stance activity but seems to have nothing to do, at least initially, with the above deontic and evaluative concepts.

Circle 3, inside both circles 1 and 2: Inside this circle are all the attitudes and activities that are associated either with what Macnamara calls "moral appraisal" or "accountability" -- the two chunks that together make up circle 3.

Circle 3, chunk 1, Moral appraisal: Inside this chunk are such attitudes as unexpressed resentment and gratitude or even outward expressions of praise and blame IF those outward expressions are fundamentally aimed merely at expressing how we feel. (The contrast here is with those outward expressions that are fundamentally aimed, rather, at enforcing an ought or "holding someone to the ought that binds him".)

Circle 3, chunk 2: Accountability: Inside this chunk are those attitudes and activities that aim at enforcing an ought. They aim at "first-personal practical uptake of the ought-violation". Punishment falls into this category, as do certain sorts of overt interactions between agents (those that go beyond mere appraisal).

To help understand the picture, here are examples of attitudes and activities that fall within each of the above realms:

Accepting promises: this is in circle 1 but in neither 2 nor 3.

Trying to explain to someone why what they've done is wrong: this is in both circles 1 and 2, but not circle 3.

Merely feeling resentment, or even saying "You're a jerk!" where this speech-act is aimed merely at expressing my inner resentment: this is in circles 1, 2, and the first chunk of circle 3.

Saying "‘You are a very rude person. A decent person has the courtesy to make a simple phone call!" in response to someone who has stood you up for a date, where this is aimed at getting the offender to feel the force of the ought he has violated: this is in circles 1, 2, and the second chunk of circle 3.

Now, my question: if we were convinced of incompatibilism and the truth of determinism (or some other relevant threat), which parts of the map, if any, would we have to scratch?

(The easy but unhelpful answer: whichever parts of the map implicate desert.)

Comments

[Zac Cogley]

I suppose this may just be evidence that I'm can't imagine being convinced of incompatibilism, but it's not clear to me that any of the responses in your examples would have to be scratched. Maybe I can feel the pull about the 2nd chunk of circle 3 would have to go, but I'm not even sure of that.

I'm assuming that the truth of incompatibilism doesn't show that there are no oughts or wrongs, only that people aren't MR for violating oughts or doing wrong. On the assumption that people still do wrong to each other and violate moral oughts, it's reasonable to explain to people when they've done wrong. (That covers the first cases, setting aside promising.) Further, it also seems reasonable to feel resentment when a wrong is done to me. And so long as "jerk" is accurately predicated of someone who wrongs me, it's hard for me to see any objection to using that term in reference to someone who is, in fact, a jerk.

Finally, moving to the last example, I'm inclined to think that it's not impermissible to try to make people feel the force of oughts that do, in fact, exist. That's the case so long as what you do to try to get the person to feel the force of the ought isn't too harsh/harmful. It would be too harsh to punch the person who stood you up, of course, but if the person has in fact been rude, how would the truth of incompatibilism make pointing out their rudeness impermissible (even when you're actually trying to get the offender to feel the force of the ought)?

All this may just be evidence that I don't feel the force of incompatibilism unless its truth shows, somehow, that there are no oughts or wrongs. But I'm interested to hear what others think.

[Tamler Sommers]

Haven't read the paper yet, but from Neal's description it seems like hard incompatibilists wouldn't have to scratch any of them--and precisely because none of the parts seem to essentially involve desert. If there was a circle for: "should be blamed or punished even if this will have no effect on the future behavior," that would have to go. But at least the second part of circle 3 seems devoted to norm enforcement, which people from all sides of the debate can hold hands and embrace.

I suppose the unexpressed resentment comes closest to something the HI would object to, but the way it's described--thinking or saying "you're a douchebag"--seems perfectly reasonable (as Zac points out) if the person is indeed a douchebag. It may not be his fault that he's a douchebag, but he's still a douchebag.

[Neal Tognazzini]

Thanks for your input, Zac and Tamler. I was thinking along similar lines. To be clear, Macnamara does think that the second part of circle 3 -- Accountability -- will include activities like punishment, which I assume she takes to be in part characterized by an attempt to get the wrongdoer to feel the force of his ought-violation. But if I understand hard determinists properly, even this won't necessarily have to go, right?

Suppose I think that it's sometimes legitimate to throw someone in jail in order to get him to feel the force of his ought-violation even if this may not have an effect on his future behavior. Can I be a hard determinist?

If so, then I guess the question is: what aspects of our practices of holding accountable doesn't Macnamara's map cover? What has she left off?

[Tamler Sommers]

Is it possible to feel the force of an ought-violation without having this affect future behavior? In what sense have they felt the force then? In any case, as long as the person is not being punished because they deserve the punishment, hard determinists can sign off. (Of course, they might have other moral objections to jailing people who don't deserve it just to have them feel the force of their ought-violations.)

"If so, then I guess the question is: what aspects of our practices of holding accountable doesn't Macnamara's map cover? What has she left off?"

Retributive practices, no?

[Mark Young]

Has she missed retibutive practices?

Doesn't arresting/ticketing someone because they broke the law count as enforcing the law? If so, wouldn't punishing someone because they violated a norm count as enforcing the norm?

As I write this I think: the law specifies a penalty for some specified action, so applying the penalty counts as enforcing the law, whereas a norm doesn't specify a penalty for its violation, so applying a penalty doesn't really count as enforcing the norm. On the other hand, maybe a norm does specify a penalty, just not as precisely as the law usually does. After all, we know it's wrong to punch someone in the nose for standing you up on a date....

Anyway, is part of the problem here this bit of the original post: Inside this chunk are those attitudes and activities that aim at enforcing an ought. They aim at "first-personal practical uptake of the ought-violation".

Isn't "enforcing an ought" a wider class than "aiming at first-personal practical uptake of the ought-violation"?

[Mark Smeltzer]

Tamler,

You are right that Macnamara's map is missing anything about retributive practices, but we might disagree on where they would fit.

Macnamara's circle 3 chunk 1 is almost about our retributive practices, except that it says that those expressions of moral appraisal only reflect how we feel. It would seem that this chunk is about subjective content, whereas retributive practices are (supposed to be) based on objective content.

The category of moral appraisal can be thus divide into two spheres: the subjective and the objective.

You said earlier that is reasonable for someone to think/feel/say that S is a douchebag if S is indeed a douchebag. This kind of schema would be categorized in the category of objective moral appraisal.

My contention is that desert-based ascriptions -- the kind we call retributive -- are always expressions of objective moral appraisal. They aren't about enforcing oughts (which certainly doesn't require desert), and they aren't about sentiments. They are about facts about people.

If we allow ourselves to believe that S is a douchebag, this will undoubtably have repercusions on how we treat S. Even if we are good consequentialists, we will be (perhaps subconsciously) inclinded to treat S disfavorably and we would have to make a conscious effort to overcome that inclination. The fact that we are predisposed to treat S disfavorably because of our beliefs about S, we are talking about retributive practices.

This may not be eye-for-an-eye or revenge killing, but it is retributive nonetheless.

That's why desert is important. If S really is a nice person and we are incorrect in our belief that S is a douchebag, then we are inclinded to treat S unjustly. In other words, if S is not a douchebag, S does not deserve to be treated like one.

Clearly there will be a lot of contention between the subjective and the objective sides of moral appraisal. We may we strongly that S is a douchebag, and we may still be wrong -- maybe we don't have all of the facts or maybe we have the facts wrong.

As regards the subjective side of moral appraisal, the facts don't matter. But for the objective side, which is about how people deserve to be thought of (and thus how they are treated), the facts trump our subjective appraisals.

Even if one is a strict consequentialist, one has strong reason to take retributive desert seriously: when thinking about how acts in the present will unfold in the future, we have to take stock of the present and combine them with our best guesses about cause-and-effect. Included in this mix will be our assements of people. If we believe that S is a douchebag, this will have an affect on our assumptions about cause-and-effect. We may decide to on a course of action that is disfavorable to S based on that belief. If we are wrong about S, then even though we did our best to make an all-things-considered consequential judgment, two things will be true: (1) our judgement about what to do won't be as good as it could have been (because we had the facts wrong) and (2) our judgement won't be fair to S (which is another way of saying that S is not being treated the way S deserves to be treated).

In my view, the whole "consequentialism trumps retributivism" enterprise is bankrupt of justification unless retributivism itself is first debunked. However, it is my suspicion that consequentialism is utterly dependent upon retributivism being true: otherwise, how could we may any reliable assumptions about cause-and-effect regarding people? If we can have reliable assumptions about cause-and-effect regarding people, isn't that simply another way of saying that we can make warranted retributive judgements about people?

[Sofia Jeppsson]

Mark, I don't think I quite follow... I can make educated guesses about the future behaviour of dogs, because I may know facts about their personality. Say that one is outgoing and another one shy, one is energetic and another lazy. All this may be objective facts. Yet I don't think dogs can in any way be responsible for how they are. If I would blame anyone for what a dog does, I'd go to the person who bred the dog, or trained it, or owned it, and sometimes there may be nobody to blame at all.

If one took free will scepticism seriously we should regard everyone as I regard dogs. We could still talk about people's personality traits, and we could say that some traits are agreeable and others unpleasant. But we wouldn't hold people responsible for the way they are, nor for actions they perform because they are the way they are.

I think though, that there is a quite different problem facing consequentialists when discussing free will. If one think an action is wrong if it produces worse consequences than any other option, this means PAP is absolutely necessary for wrong-doing. No PAP, no wrong-doing, but everyone always does the morally right thing.

[Sofia Jeppsson]

I just realized that the last sentences came out wrong. It's supposed to be: "If one thinks an action is wrong if it produces worse consequences than any other option, this means that wrong-doing requires alternative possibilities. No alternative possibilities, no wrong-doing, but everyone always does the morally right thing."

In my post I had alternative possibilities mixed up with the so-called PRINCIPLE of alternative possibilities...

[Mark Smeltzer]

Sofia,

You said, "Yet I don't think dogs can in any way be responsible for how they are. If I would blame anyone for what a dog does, I'd go to the person who bred the dog, or trained it, or owned it, and sometimes there may be nobody to blame at all."

I don't think it is possible for anything to be responsible for how it is, but I do think it is possible for some things to be responsible because of how they are. In other words, if responsibility is even possible, it must be because it is based on self-expression and not upon self-definition. Since self-expression is a fairly mundane concept (ie. it's easy enough to be a realist about it), it is likely that lots of things have this capacity, even dogs.

However, the question of moral responsibility probably doesn't pertain to dogs since they lack sufficient mental capacities to grasp universal concepts (dogs are contextual learners after all); especially moral concepts, which limits our ability to make objective moral appraisals of dogs.

If I have misunderstood your criticism, please correct me.

[Sofia Jeppsson]

Mark, I think we both misunderstand each other.

First I think you misunderstood my criticism. I'll try again:
A consequentialist must take other people's tendency to act this way or that into account when deciding what HE ought to do. If the consequentialist's performing action A would cause another person to do action B, and action B has really bad consequences, the consequentialist ought to abstain from A, etc.
If I understood you correctly, you think that this somehow implies that a consequentialist must be able to think about people as "douchebags" for example.
But I don't see how that would follow. It is possible to think about people's personality traits in a purely empirical, not value-laden, way. Let's say that the empirical part of being a douchebag is a tendency to say harmful things to other people. I could use this piece of empirical knowledge about another person when judging which action of mine would have the best consequences; for example, the best consequences might come from excluding this person from parties where he has the opportunity to hurt lots of people's feelings, or it might be best if I said "you're a douchebag!" IF saying so would make him change behaviour. Or perhaps again, the best consequences might come if we all learnt to simply ignore what he says. But I don't need to think about him in a value-laden way, in order to use my empirical knowledge about him when planning my own actions.

The dog analogy was supposed to show that... suppose I have a dog with strong hunting drives. I can use this knowledge to foresee that he's probably gonna chase the moose if I let him off leash in a forest with lots of moose in it. And since I think this would have very bad consequences, I decide not to let him off leash. I could plan like that, using my empirical knowledge about him, without making any value-judgements about him.

Second I must have misunderstood something in your last post, since I simply don't understand what you're getting at. I think it makes sense to hold fairly grown-up, intellectually normal etc people responsible for their actions, and also to some extent for their personalities (we do have some compatibilistic control over the way we are). I don't think it makes sense to hold dogs, tiny babies or severely mentally handicapped people responsible, although all these individuals can have personality traits that affect their behaviour (I suppose that's what you mean by "self-expression" - performing acts because one has certain personality traits). Of course they can be responsible for lots of things in some causal sense of "responsible", but not in a moral sense.

Are you suggesting there is some middle-ground here between causal and moral responsibility, which dogs, babies and the severely mentally handicapped can have? I don't really have any intuition about what that would be... I must be missing something here?

[Mark Smeltzer]

Sofia,

(Sorry for the length of this post, but I wanted to address in detail your concerns and reformulate my proposal in more explicit terms. Hopefully if you read through this you will understand my position more clearly.)

If your point is that consequentialists may object to there being any such thing as objective moral appraisals, I think I see what you're getting at, and I'll address that point here.

However, if you're driving at trying to wedge in some sort of distinction between normal humans, damaged humans, dogs, etc., as regards self-expression, then I am not going to persue a response to that. It is simply obvious that all of these things have the capacity for self-expression, and my prior point is that self-expression itself doesn't get us all of the way to moral responsibility, but it does get us rather close.

For the strict consequentialist (of the act or rule variety), the releavance of the "moral" part of moral responsibility may be somewhat moot. What I am trying to demonstrate here is that the consequentialist is doing the same exact set of mental activities that the moral realist does when making decisions that have retributive consequences.

The only significance difference is that the consequentialist may not "feel" like he's trying to give people what they deserve when he is making his decisions -- after all, he's just trying to make the decision that will achieve the best outcome. However, the very set of mental apparatus that he engages to take into consideration is knowledge of others is the same mental apparatus that makes moral judgements of others.

For the purpose of this discuss, let's think of morality in purely pragmatic terms (as the consequentialist will). Suppose when we say that S is a douchebag, we simply mean that S will act in ways W more than C percent of the time. Define this way, there's nothing value-laden about calling S a douchebag, and if we suppose that all objective moral appraisals can be cached out in a similar manner, then all objective moral appraisals can be defined empirically.

At this point, I want to drive in the point I made in my original post: this just is retributive thinking. Giving people what they deserve and treating people according to our best understanding of them equate to the same thing. If this is correct, then even consequentialists that accept self-expression are retributivists.

So now let's talk about the kind of consequentilist that could avoid the charge of being a retributivist. This would require acceptance of an extreme form of nominalism that would rule out self-expression, along with all other forms of normative logic. On this view, we have a strictly Humeian type view of causality: all we can say is the way things have been, not the way things will be. On this view, we cannot make even pragmatic objective moral appraisals because all we can say is what has happened. On this view, we could define a douchebag as someone who has acted in ways W more than C percent of the time, but this definition gives us no reason to think that S will act in ways W more than C percent of the time. On this view, the consequentialist's decisions are ultimately arbitrary and guided only by sentiment, since the decision maker will never be mentally equiped with the necessary tools to make the leap from the descriptive to the normative.

Now, in this last example we were thinking of an extreme form of nominalism combined with consequentialism. The net result is that in such a world it would be logically impossible to make a pragmatic decision that is any better, in principle, than any other. In this way, all decisions would be objectively arbitrary. It would still be possible to distinguish between good and bad decisions purely on terms that good decisions are those that actually achieved the desired outcome, but there would be nothing underneath to explain why things turned out well (remember that we're still assuming an extreme nominalism).

Now, Hume himself made the point that even if we become aware of extreme nominalism, we will still live our lives as if there is a normative basis for cause and effect. So even if extreme nominalism is true, we can still expect there to be retributivist consequentialists running around -- even if they accept this extreme nominialism in principle, it is difficult to imagine anyone actually living it out.

As soon as the consequentialist starts thinking about people in ways that involve normative cause and effect, then they will attempt to treat people in ways that accord with their best understanding about what others are like, and this (again) is the same thing as treating people according to what they deserve.

I made this point to Saul Smilansky sometime in the last year or two on this blog. I presented him with a case where he finds out that his baby sitter had made sexual advanced on his daughter. Would he treat the baby sitter differently going forward? I suggested to Saul that if he never hired the baby sitter again or told his friends never to hire that baby sitter, then he would be engaging in retributive practices. Moreover, those are both very normal reactions to such a situation, and it is hard to imagine any consequentialist not having sufficient reason to choose at least one of them.

Of course there are ways out of that situation without taking a retributive response, but they all center around the need to educate rather than to punish. Suppose that Saul decided that the baby sitter just needed to have the ought (never make sexual advanced on the kids you are baby sitting) enforced in order to internalize the principle. As such, Saul could choose to scold the baby sitter or to not hire the baby sitter for three months. So long as Saul chooses something that is meant to be temporary and educational in nature, Saul needn't worry himself over what the baby sitter deserves, because he is merely taking an appropriate corrective response.

So suppose that Saul refrains from taking the retributive response, hires the same baby sitter again three months later, and the baby sitter makes another sexual advance against Saul's daughter. What now? Saul again has to decide between taking a corrective response or a retributive response, and the question comes down to what Saul believes about the baby sitter: whether Saul believes that the baby sitter is always likely to make sexual advances, or whether Saul believes that the baby sitter just needs more education to internalize the principle.

So long as Saul remains committed to the idea that the baby sitter needs more education, Saul could conceivably keep rehiring the same baby sitter every three months (or maybe he would increase the penalty each time: three months, then six months, then a year, etc.) until his daughter becomes old enough not to need a baby sitter.

If Saul follows that regimen, he would be have avoided any retributivist actions against the baby sitter, but the question then is: does that seem like responsible parenting? In other words, the question is now about how we *can* act (I admit that it may be possible to live a life free of retributivist actions); no, it is about how we *should* act.

Even if we are consistent consequenialists and define that impetus in terms of what patters of action are likely to produce the best outcomes, does the scenario where Saul's daughter has to deal with that creep over and over sound pragmatically favorable to just banning the baby sitter permanently after the first offense?

For any consequentialist who thinks there should be a cut off point, that person is also a retributivist.

[Lyndon Page]

Hi Mark and Sofia, enjoyed the posts, thought I would chime in.

Mark: I think your hypothetical example falls apart a little bit if you apply more generous non-retributivist practices to the situation. I think Saul (or some other "non-retributivists") would take a more calculated measurement for the correction of someone who "molests a child." We would certainly add in a high degree of psychological counseling and a little more stringent "big brother" watching of a possible offender. Also, in the controversial style of Pereboom, I think one could isolate an individual, or even deny certain privileges of interaction with others, without being retributive. Such a practice would need to be performed with full clarity to the molester that such isolation is for a total societal benefit (of not having molesters watching children) without going so far as to be retributive in its scope, although that is a blurry line, as you display above. In any case, I think there must be a way to protect children and run a stable society without retribution.

(Kip Werking has a paper on the possibility of a "medicalized state" that I think takes a fairer stance on the possibility of non-retributive practices.)

[Mark Smeltzer]

Lyndon,

Thanks for the reply. I understand your point, and the basis for it, on a societal level. We can distinguish between personal and societal retributive practices. Obvious Saul would not have the right to put the babysitter in jail, bonk him on the head, or what not - those are punishments/tellishments reserved for the (civilized) society to administer.

Within the scope of personal rights, Saul has the right to exercise several types of retributive actions against the babysitter. Keep in mind that I did not go so far as to suggest that anything strictly illegal happened, just that (minimally) something happened that Saul would not approve of. Given these constraints, Saul could ban the babysitter for life, hate the babysitter, tell others never to hire the sitter, etc. Those are all retributive responses that are within the scope of personal rights.

Moreover, I think it is possible to accept retributivism on the personal level and reject it on the societal level. So Kip may be talking past this distinction... it is not clear.

[Sofia Jeppsson]

Okay Mark, I think I understand what you mean by retributivist praxises better now.

As I see it, there are two problems with the claim that consequentialism can happily co-exist with determinism in a way that other ethical theories cannot.

Problem 1: In most utilitarian theories an action is considered wrong if the agent could have done something else that would have had better consequences. I e, wrongness strictly implies the ability to do otherwise on these theories, and there's this big debate on whether that is compatible with determinism or not. This falls outside what you discuss in your posts, but I do think it's a problem that ought to be taken seriously.

Problem 2: Closer to what you discuss. You call all kinds of punishment (whether social or legal) "retributive actions". A consequentialist wouldn't agree there, at least not the consequentialists I know. They would only call punishments "retributive" if they were done from a certain motive, namely to give people what they deserve. If they were done from another motive, namely to deterr people from future bad acts, to rehabilitate people or something else which is forward-looking, they're not "retributive" any longer. So you seem to be using the term "retributive" to encompass much more than the consequentialist means by that term.
One could of course ask why it would matter what kind of justification one gives for a certain punishment. If everyone agrees on what action Saul should take against the baby-sitter, then why on earth does it matter if he performs the action because he wants to give the babysitter what he deserves (retributive) or ONLY because he wants to protect his and other children (consequentialist)? I guess the motives would be mixed in most people. And besides, it seems pretty non-consequentialist in itself to care about people's motives.

But when it comes to self-expression I realize that I have no clue what you're talking about. Typically compatibilists have requirements for moral responsibility that only a sub-section of human beings can fulfill; one needs some level of rational thought and capacity for reflection for example. These are not mysterious or metaphysical concepts. And it's obvious that most animals and some people lack these qualities.
So the self-expression you talk about must be something else - but what? If it just means acting from one's personality traits, then even fish have self-expression, but I have no idea how that in itself could get us even close to moral responsibility...