I’ve just read T.M. Scanlon’s chapter on blame in his latest book Moral Dimensions. The discussion is subtle, provocative, and quite insightful. It has already caused me to rethink some of my own views on moral responsibility in general and blame in particular. Nevertheless, I have a few questions/worries about the account that may be worth discussing over the course of a few posts. In this post, I’ll just focus on his account of moral responsibility, and in later posts I hope to focus on his accounts of blame and the ethics of blaming. (And I apologize in advance for the length of this post.)
There are three important and related concepts discussed in the account. First, I am morally responsible for X (where X is an action or attitude) just in case X is properly attributable to me. Second, I am blameworthy for X just in case X reveals something about my “attitudes toward others that impairs the relations that others can have with” me (128). Third, for you to blame me for X is for you (a) to judge me to be blameworthy for X, and (b) to modify your relationship with me (or reaffirm a previous such modification) in the way deemed appropriate by your judgment in (a).
While I hope to discuss blame in more detail in a later post (and so will reserve my exposition of Scanlon’s view of blame until then), it’s worth saying something briefly now about the relation between (a) and (b). As Scanlon formulates the view, a judgment of blameworthiness is a necessary condition of blame. But this doesn’t seem right, given that blame primarily consists, allegedly, in a modification of one’s attitudes towards the blamed person, and one could modify one’s attitudes towards someone in this way without, or even contrary to, a judgment about that person’s blameworthiness. Suppose, for example, that I’m in thrall to someone who’s saved my life. As I follow him around, I see him treating certain people very poorly, and while consciously I rationalize his behavior and so make no judgments about his blameworthiness (or in fact I judge that he’s not blameworthy), I nevertheless (subtly, unconsciously) modify my attitudes towards him, eventually perhaps finding myself less enthralled with him. (And it’s still true of me that this modification of attitudes is one that would be made appropriate by a judgment of blameworthiness, were I to make such a judgment.) If blame is merely the modification of certain attitudes, therefore, and such attitudes are modifiable under my conscious radar, so to speak, or independently of my judgment “faculty,” then it’s doubtful that the judgment of blameworthiness is necessary for blame. Of course, one might think that all attitudinal modifications merely reflect shifts in evaluative judgments, but then it would just be unclear what’s added to the conditions of blame by including the reference to judgments of blameworthiness.
My main point here, though, is about moral responsibility, although what I’ll say is connected to the point just made. For Scanlon, to say that I’m morally responsible for X is just to say that X is properly attributable to me for purposes of moral appraisal. This means one can be morally responsible for actions, say, that are morally bad, good, or neutral. To be blameworthy for an action, for example, is for that action to be attributable to one and for it to be morally bad (roughly put). (As an aside, despite Scanlon’s affinity for citing Strawson throughout his discussion, this is, I think, a distinctly non-Strawsonian account of moral responsibility. For Strawson, being morally responsible for X is a matter of being (appropriately) held responsible for X, where Scanlon’s is a view according to which being appropriately held responsible (i.e., blamed) depends on an antecedent judgment of being responsible.)
So what does it mean for X to be properly attributable to me? This is unclear in the current text. One might think that, given his past remarks, Scanlon would maintain that X is attributable to me just in case it is responsive to my considered judgments. But he explicitly denies this view (193), holding instead that something may be attributable to me even if it’s not responsive in this way. For example, he says, I may firmly judge that the fact that hiring some job candidate would please a colleague I dislike isn’t a reason to decide against that candidate and yet this fact may still seem to me to be a reason to do so. My attitude thus resists my judgments and so isn’t under my reflective control; nevertheless, claims Scanlon, this attitude is still mine, an attitude for which I’m morally responsible. Something similar holds for the person who firmly rejects racism but nevertheless occasionally thinks someone’s race is a reason for treating him as inferior: the racist reaction is still his, and so is something for which he’s the appropriate subject of moral appraisal (although we may temper our actual blame of him).
I have a few questions/worries about this story. First, what makes an attitude one’s own in this sense if it’s not in fact connected or responsive to some evaluative judgment? Angela Smith, who has followed earlier Scanlon in her own work (and who Scanlon in the new work cites), suggests that attitudes are one’s own just in case one is answerable for them, which is a matter of their being ultimately grounded in reasons (which, after all, can be dredged up to respond to and defend against the answerability demand). For her, one’s attitudes are reflective of one’s evaluative judgments, and such judgments are simply ways of regarding something as having evaluative significance, and things are regarded as having evaluative significance for reasons. Thus, one could be answerable for an attitude in virtue of its being grounded on such reasons.
It looks, though, as if Scanlon is now disavowing the connection between answerability and responsibility. As he notes, when something seems to me to be a reason, it’s up to me to decide whether or not it’s actually a reason, and its being up to me in this sense is just for me to be answerable for it (something which I can be called on to defend or justify), for such a decision is a judgment that itself is grounded on reasons. But if considerations can continue to strike me as reasons even though I’ve decided that they’re not, then such attitudes look as if they operate independently of any grounding reasons and so aren’t things for which I’m actually answerable any more: in what way could it be subject to a call for defense? So now we have two questions before us: what actually makes such an attitude my own, and can I really be responsible for an attitude without being answerable for it?
Now one might insist that such attitudes are still grounded in reasons, even though they are not grounded in, or responsive to, my reflective evaluative judgments. I must confess that I’m not sure what this could mean, though. I can well understand how there might be explanatory reasons for my holding certain attitudes, but answerability is surely about one’s justifying reasons, and if one has these, then how could they not straightforwardly support evaluative judgments? Is the claim, then, that these attitudes just aren’t responsive to one’s reflective or conscious evaluative judgments? And would this imply, therefore, that they are responsive to one’s unreflective or unconscious judgments? But what could that mean, precisely, especially if one consistently disavowed such “judgments"? Further, if this were the case it seems answerability would still be missing, for how could one answer for or defend one’s unconscious judgments, especially if they were judgments one actively and reflectively disavowed?
All of this is to build up to asking about the case of the kleptomaniac or the obsessive compulsive. For absent a clear story about attributability, it’s hard to see why the kleptomaniac’s desire to steal, or the OCD victim’s attitudes towards handwashing, wouldn’t be properly attributable to these agents, such that they would both be morally responsible for these attitudes (i.e., open to moral appraisal for them). After all, their attitudes do tell us something “interesting about [them] over time” (197), and the fact that these attitudes are unresponsive to their reflective judgments is irrelevant. But this implication would seem quite counterintuitive. Surely the fact that his hands have germs strikes him as a reason to wash them for 12 hours/day is more plausibly thought to be an assailant on the OCD victim’s psyche rather than a constitutive part of it, something for which he’s not responsible. But it’s hard to see just how we could mark a distinction between the OCD victim and the recalcitrant racist, say, given the machinery Scanlon has constructed.
I may have just missed something in the account, though, so I’d be grateful for any help on these points.
Interesting stuff. I haven’t yet had a chance to read Scanlon’s new book, so I’m just going off of your summary, but two things jumped to mind. First, I think that the attempt to distinguish the OCD victim from the recalcitrant racist is harder than it might seem, for reasons that are totally independent of Scanlon’s particular commitments. Just like someone might be an unwilling or alienated kleptomaniac, someone might be an unwilling or alienated racist. Thus people like John Arthur (and others) have argued that truly unwilling racists are in fact not *morally* wrong (because not morally blameworthy) for being racist, and therefore racism is not necessarily morally wrong (though it’s always unjustified). I don’t really agree with this assessment, but the point anyway is that it’s a hard–if sometimes unrecognized–problem that seems to cut across particular theories of responsibility and assessment. That is, the general worry is that for any kind of case where we intuitively want to say that the OCD victim is not blameworthy, we can construct an analogous case of racism.
Second, a thought regarding your opening side-comment. I would think that there could be a third alternative, besides the idea that “all attitudinal modifications merely reflect shifts in evaluative judgments” and the idea that attitude-affecting judgments are all consciously accessible and manifest. Namely, it might be that some (but not all) attitudinal modifications are based on judgmental shifts that are undetected, or, for that matter, simply operative but not occurrently represented either consciously or unconsciously. (This latter idea is roughly the view Horgan and Timmons propose, in another context, in their “Morphological Rationalism and the Psychology of Moral Judgment” (ETMP 2007).) So let’s say that we want every blameworthy attitudinal shift to be produced by a judgmental shift. (We might want to distinguish, for the purposes of understanding blame, attitudinal shifts that are the product of judgmental shifts and those that are the product of, say, situational shifts, e.g., being stuck in bad traffic, smelling freshly baked bread, etc.) We need not therefore say that every attitude-affecting judgment is either accessible or even representational.
Josh: Regarding the first point, a couple of comments. First, Scanlon agrees that, w/r/t the recalcitrant racist, our blame of him may well be appropriately tempered, given that he disowns his racist attitudes (whereas such a response may not be tempered for someone who embraces those attitudes, say). But this isn’t to say he’s not morally responsible for those attitudes: they’re still his, Scanlon wants to say. Now it’s unclear whether or not Scanlon also wants to say that our tempering of blame of this man also means we’ve tempered our judgment of his blameworthiness, but even if it does, we can still say of the man that he certainly has a moral fault, and this undermines the Arthur move, it seems.
The second point on your first comment is that I agree that it may be difficult to distinguish between the OCD case and recalcitrant racist case, but it seems a plausible theory of moral responsibility ought to be able to do so, that these cases fall on different sides of the responsibility map. There are, though, ways to do so. One might have an account according to which those racist beliefs really are reflective of one’s (true?) evaluative judgments — I take this to be Angela Smith’s account — whereas one’s OCD-desires really aren’t (and so represent just a cognitive malfunction). Or one might have a less reason-based view (sort of like my own view of autonomy), according to which one’s real self is located in one’s nexus of cares, and if one’s attitudes are grounded in that nexus (as the racist’s presumably are), they are properly attributable to one, whereas if they aren’t so grounded (as the OCD victim’s desires presumably aren’t), then they aren’t.
As to the second point, this sort of disjunctive account may well be true, but it’s not a third-way of the sort that helps answer the question I asked. If one’s attitudinal shift is accompanied by a conscious evaluative judgment shift, then this isn’t what occurs in the case I’ve given. If, however, the evaluative shift takes place unconciously, then it is, as I said above, unclear what’s added to the account by making judgments of blameworthiness necessary for blame. Or perhaps I’ve misunderstood you?
I think maybe I wasn’t clear enough on the second point, and that I might have buried the key point in the parenthetical comment. Just to set the table properly, go with the second of your (most recent) conditions–the judgmental shift happens either unconsciously or even non-representationally. You’re right that this in itself doesn’t explain how judgment does any needed work; this maneuver is just meant to accommodate any cases where someone might object that judgmental shifts aren’t present. Okay, so with the table set in that manner, why have blame be judgment-sensitive? I was kind of assuming that Scanlon would have provided some rationale for that claim. (Again, I haven’t read his new book yet.) But I can see why it might be at least plausible to say that blame is based on judgment. Our blame-like attitudes are sensitive to all sorts of things, but we might want to segregate some of them as non-blaming attitudes because they don’t fit our judgments of blameworthiness. E.g., you might discover that, despite your harboring blame-like attitudes, you don’t actually blame someone for betraying you, because you don’t judge them to be blameworthy (on reflection you realize that they didn’t really betray you, or intend to betray you, or whatever, but you were really hurt or in a bad mood that day or are sensitive because you were (truly) betrayed just yesterday…). That’s the idea, anyway, but I’m curious: does Scanlon offer a rationale for making blame judgment-sensitive?
On the issue of disassociated racism, I think that there’s a real need here to admit that racism can take exactly the form that (say) OCD or kleptomania can take. It won’t do to say that “we can still say of the man that he certainly has a moral fault, and this undermines the Arthur move, it seems.” To be clear about what I was suggesting, this is the denial of Arthur’s claim, not an argument against it. Regarding whether substantive accounts of responsibility and assessment can handle this kind of racism, I think it’s not as easy as you make it sound. While I like Smith’s work a lot (I’m assuming you’re referring to her 2005 Ethics paper, or at least the ideas contained therein), I don’t think it fully makes sense of the kind of case that concerns Arthur. It can’t just be stipulated that “those racist beliefs really are reflective of [the racist’s] evaluative judgments,” since surely we can imagine a case where this is not so. And the same goes for your view, “according to which one’s real self is located in one’s nexus of cares, and if one’s attitudes are grounded in that nexus (as the racist’s presumably are), they are properly attributable to one, whereas if they aren’t so grounded (as the OCD victim’s desires presumably aren’t), then they aren’t.” To raise Arthur’s worry here, we just need to find a case where a person harbors racist attitudes that aren’t based in the nexus of his cares. This is the kind of case that Arthur suggests, and it really does generate a puzzle. Again, it seems that, for whatever theory of attributability you favor, if it entails that the OCD victim’s desires aren’t attributable to her, we will be able to construct a parallel case for racist attitudes.
Josh: this is good (and thanks for the discussion). To the first point, I’m actually not sure, given what I’ve seen in the text, that one can draw the distinction you want between blame-like attitudes and blame itself (such that what one thought was blame actually wasn’t). Scanlon writes (141), “To blame someone is actually to hold modified attitudes of this kind [indicating a kind of impairment of the target’s relations with others]….” So if the exemptions or excusing conditions you cite occur (e.g., you discover that the target didn’t do what you thought she did), it’s more accurate to say that you withdraw your blame upon such a discovery, rather than saying that you weren’t blaming to begin with. But if that’s the case, then it seems clear that blaming can occur w/o the judgment of blameworthiness. Now at other spots, Scanlon adds an appropriateness condition, such that blaming consists in a modification of attitudes rendered appropriate by one’s judgment of impaired relations (which is one’s judgment of blameworthiness). But I still think there are cases in which one could adopt the modified attitudes without having rendered the actual judgment, where the modifed attitudes are indeed rendered appropriate, but rendered so by the fact of impaired relations, rather than one’s judgment of those impaired relations. Now as far as I can tell, there’s not an explicit argument given in the text for why blame must be judgment-sensitive in the way suggested, but I may simply have missed it (and would be grateful if others know where it might be).
On the second point, perhaps there’s a methodological difference here. I was thinking of cases in which we indeed have fairly settled intuitions that the OCD-person’s attitudes aren’t properly attributable to him, whereas the racist’s (even the recalcitrant racist’s) are, and so then the puzzle was to try to see (a) how Scanlon’s theory might distinguish them, and (b) what other theories might say to distinguish them. On that score, I have a hard time figuring out just how Scanlon’s view distinguishes those sorts of cases, whereas I think Smith’s view (and my view) might do a better job. You, however, are talking about cases in which it may well be difficult to distinguish so easily the OCD-person from the dissociated racist. So I’m saying that, granting a difference between the two, what theory best explains it, whereas you’re saying, “It’s perhaps not so easy to grant a difference between the two.” Fair enough, but then we’re talking past one another.
Let a “blame-like attitude” just be a term of art invented for our discussion: it is an attitude that manifests component (a) but not component (b) of Scanlon’s analysis of blame (as you summarized it initially). We can stipulate this, so we can draw a distinction between blame-as-a-plus-b and blame-like-attitudes. The problem with the quote you just gave from 141 is that it seems to offer a different analysis of blame, one that contains only element (a) and not (b). So I guess the first question is: what is Scanlon’s analysis of blame? Whichever we go with, though, I’m still thinking that neither the a+b nor the merely-a analysis is blatantly implausible, for reasons (we’ve both) given. (That is, I can see some reason to think that judgment-sensitivity might be necessary for truly capturing the ordinary concept of blame.) Also, for what it’s worth, and this is just a technical point that doesn’t impact anything substantive, but by separating (a) and (b), it sounds like Scanlon could agree with you that attitudes can be modified without modification in judgment; it’s just that he wouldn’t call that an instance of blame.
Regarding racism, though, I’m still not inclined to think that the issue can be avoided merely by saying that we’re talking past one another. You say that your or Smith’s view can explain the difference between the OCD person and the dissociated racist better than Scanlon can. I’m saying that this is not true: neither your nor Smith’s view can explain that difference (no matter whether or not Scanlon can). That sounds like a disagreement to me. And this allows that we have “settled intuitions” that the attitudes in question are not attributable to the OCD person but they are to the dissociated racist. I have (or once had, before I was corrupted by philosophy…) those intuitions, too. But the question, as you rightly point out, is how we can explain those intuitions. I claim that we can’t explain them by your or Smith’s views.
We could be talking past one another if we are thinking of subtly different case-pairs. Since the OCD case is constant for you and me, whether we are talking past each other would come down to whether we are thinking of different cases of racism. Put slightly differently, given an Arthur-style argument that no theory of responsibility can render attitudes non-attributable to the OCD person but attributable to the alienated racist, what we’d need would be some other case of racism (besides the case of dissociation) that can be explained by your or Smith’s theory but not Scanlon’s. Maybe I missed it: is there a case like that, one that’s not about dissociation, alienation, and the like, but about something else that is manifest in the racist but not in the OCD person or vice/versa?
I wonder if the background story for why he takes blaming to require a judgment of blameworthiness has to do with his wider contractualist framework. He could allow that there are many blamelike modifications of interpersonal attitudes that do not require such judgments. But, the kind of change in interpersonal attitudes that is relevant for blame is where one no longer sees oneself to be in a relation with the other person in which justification on non-rejectable grounds is required. And, to be justified to make this kind of change in one’s attitudes requires a judgment that the other person has already acted wrongly in a blameworthy way – in a way in which she has stopped caring about justifying her actions to one in a way that is robustly attributable to her.
About attributability – I’ve always thought that Frankfurt’s higher order account or wholeheartedness account could deal with that problem. After all, they were supposed to be accounts of which desires are one’s own. The proposal is that those desires which one does not desire to have or which one is on occasion moved to stop having are not attributable to one. This condition is weaker that evaluative judgments or reason-judgments. But, if some object of a desire appears as a reason for one, then one seems to endorse desiring this object so Frankfurt’s condition seems to be satisfied there. Yet the OCD victim needn’t have the desires to desire washing hands and she may be trying to get rid of that desire.
I know there are good objections to the Frankfurt account by Watson and others but those seem to be more objections to his account understood as an account of accountability and not attributability.
Jussi: Thanks. I have no doubt the motivation to incorporate a judgment of blameworthiness as necessary for blame rests in the larger contractualist view, as you point out. The question, then, is whether or not one can see oneself to no longer “be in a relation with the other person in which justification on non-rejectable grounds is required” without the judgment in question. After all, if various considerations can strike one as providing reasons even against one’s own contrary judgments (as detailed above), one could well be motivated to modify one’s attitudes in a “blame-like fashion” without the judgment of blameworthiness as well, beneath one’s conscious radar, say.
As to the Frankfurtian account, I’m not convinced that taking up a contrary higher-order desire, or even just failing to be wholehearted w/r/t to first-order desires is sufficient to render the attitudes in question “not one’s own.” The Frankfurt account would render the recalcitrant racist’s attitudes not his own, after all, and that doesn’t strike me as (always) correct.
And speaking of which, Josh, perhaps I’m still not getting it. Someone who avows he’s not a racist but whose attitudes nevertheless seem to say otherwise could well be someone who cares about white people more than minorities (an attitude which, on my story, would manifest itself in his emotional reactions), and this would be an attitude dependent on his cares in a way the OCD person’s attitudes wouldn’t be(i.e., this latter person may have no negative emotional reaction to the news — which he comes to genuinely believe — that he’s germ free).
Nevertheless, I guess I can agree that one could construct cases of “racism” that are analogous to OCD cases, but I’d have a hard time believing that many (any?) of them would be what we would think of as actual racism, rather than a kind of Tourette’s-like syndrome of racist-like behaviors. But perhaps this is a topic for another day.
Dave,
They are unique cases, but they’re not supposed to be like Tourette’s syndrome. What Arthur has in mind is a case like the following. Imagine that someone was brought up on a vicious and extremist racist compound. In his mid-teens, he leaves and is exposed to the world, and he gradually learns that the racism he was brought up to believe in and viscerally feel is both unjustified and morally corrupt. Despite this reformation, he still finds that racist thoughts occasionally ‘intrude’ in his consciousness. But he despises such thoughts now. He is alienated from them in every respect. (He doesn’t identify with them, they violate his cares, etc. You could of course simply stipulate that if he has the intrusive thoughts, then he has the corresponding cares. But then it’s a trivial victory that leaves the problem intact, and anyway you’d have to stipulate the same thing about the OCD person, on pain of being ad hoc.) In order to rid himself of them, he undergoes years of intense psychotherapy. Still, after all these years, they intrude. So after all of this, the questions are: do we still categorize his thoroughly alienated beliefs, hateful episodes, and so on, as both racist and morally suspect, and can we attribute them to him in a way relevant for moral responsibility? Arthur says that we cannot attribute them to him, and so they are not morally suspect (but they are racist and unjustifiable for Arthur). I think that this violates our intuitions (the ones you called upon earlier). I think we still want to say that there is something morally problematic about those racist attitudes. But, no matter what your theory of attribution, it’s hard to say all of this and not say the same thing about the OCD person. (And, specifically, it seems like alienation, dissociation, etc. won’t do the trick.)
Thanks David. The case about recalcitrant racist is interesting. I worry that here our intuitions about the attributability might be distorted by our intuitions about moral responsibility, accountability and blameworthiness. So, we want to attribute the attitude to the agent because we want to hold her accountable and blameworthy for that attitude. I wonder if we could do the latter even when the attitude was an attitude that is not properly the agent’s own. If that was the case, the agent would not get off the hook even if the attitude was not hers but only within her, and so we would not need to try to attribute the attitude for her for the blaming purposes.
Scanlon’s account as you first formulate it seems to allow for this possibility. On that first characterisation, blameworthiness does not require attributability but rather that there is ‘something about [the agent’s] “attitudes toward others that impairs the relations that others can have with” [her] (128).’ A racist attitude must be such an attitude – that attitude deems other people, just like me, inferior and as someone to whom equal justification is not owed. Such an attitude would affect my relation to a person even if she correctly thought that it was an external force working within her over which she had no control.
Josh: The details help. Thanks. Regardless of my own view of the matter, I think Scanlon will want to say of the racist case that you’ve described that the thoughts are the agent’s, i.e., they’re properly attributable to him, but we may well be justified in seriously minimizing (or altogether eliminating) our blame of him, given how hard he’s trying to get rid of those thoughts. The difficulty, then, is twofold: (a) why exactly should we say this, given that there doesn’t seem to be any direct connection between the agent’s attitudes and his evaluative judgments (at least of the conscious variety), and (b) how, then, could we distinguish this (if at all) from the OCD case, for once you divorce attributability from connection-with-reflective-evaluative-judgments, it looks like there’s no non-arbitrary reason to exclude the OCD case as well. So I take one of your points on this score to reaffirm my original question of Scanlon.
Your larger point is that any theory will have difficulty distinguishing the cases, but I’m still not convinced of that. One thing that matters is what, if any, connection the racist attitudes have to other aspects of his deliberative life: are they susceptible to modification in light of new information, or changes in his cares, etc.? Do the attitudes impair ordinary deliberative functioning? Is he motivated to satisfy the related desires to achieve the object of the desire or just to quell the desire? Insofar as there’s no such connection to ordinary deliberative functioning (as I take to be what’s involved in the OCD case), then sure enough, I don’t see a difference between him and the OCD person (they’re both essentially no different from a case in which an external manipulator implants the thoughts). But once differences in the deliberative functioning are introduced, then the relevant wedge is introduced to distinguish the attributability in either case. The problem I’m suggesting for Scanlon’s view is that once the connection to reflective evaluative judgments is prized apart from attributability, I’m having a hard time seeing how precisely he’d distinguish OCD from recalcitrant racism in cases where there were these deliberative differences.
Jussi: This is an extremely interesting suggestion. I myself have been working on a few papers trying to carve out the possibility of accountability without attributability. However, on the view I’ve been working on, attributability is more substantive than what Scanlon has in mind (more Watsonian than Scanlonian). For Watson, attributability is about tracing some action to a fault, say, in the agent’s character, so an assessment of attributability-responsibility is essentially about moral predication, aseessing someone as “cruel,” or “manipulative,” etc. So the negative assessment is already embodied in the attribution.
For Scanlon, a judgment of attributability is about the assertion that some action/attitude belongs to someone in a way that leaves them open to negative (or positive) appraisal. So for Scanlon, I can properly attribute some action to you without yet attributing it qua vice or virtue to your character.
In light of this, it’s hard to see how one could be accountable for something that isn’t one’s own. After all, whose attitudes are such that they impair relations with others? Now I see what you’re going for: if there are these racist attitudes “within me,” won’t that affect the relations I can have with certain others, regardless of whether or not they’re properly attributable to me? Well, yes and no. Suppose the attitudes are essentially just Tourette’s-like tics. While you may not go with this person to various public events, his tics don’t violate the normative standards of your relationship and so aren’t things for which you’d hold him accountable (in the Scanlonian fashion). And this is in virtue of the fact that the tics aren’t reflective of or on him. If he were to really believe the things he said, however, you would hold him accountable, and this would be in virtue of the fact that what he’s saying really was reflective of or on him. But this is just attributability (of the Scanlonian sort). So perhaps I’m asking just what it would mean for someone to “really believe” — or, more generally, really hold — the attitude in question. (Or perhaps this might be put in terms of the conditions of “sincerity” of attitude.)
Thanks David again. This is interesting. It is an interesting question of how one would react to a racist tic. I think it might change my relation to a person more than just not being with the person in public. I guess this would be to recognise that there is moral luck. In the same way, one seems to be sometimes accountable for accidental actions that are not properly one’s own as one lacks control over them.
Anyway, I think you seem to answer the question of what it to really believe in your answer to Josh. I do like the suggestion of looking at how the beliefs function in deliberation. These questions: “are they susceptible to modification in light of new information, or changes in his cares, etc.? Do the attitudes impair ordinary deliberative functioning? Is he motivated to satisfy the related desires to achieve the object of the desire or just to quell the desire?” seem to be on the right lines. They also seem to be fairly similar to the questions which we use to assess the wholeheartedness of mental states in Frankfurt’s framework.
I guess the question I wanted to ask was whether you think that there is a reason why Scanlon could not add a view like the one you suggest to his account of blame and responsibility. As I remember, attributability did not seem to be one of the main topics of the blame chapter. Maybe this shows that his view is incomplete rather than problematic.
Thanks, Dave, that helps. But how I’m confused a bit, I think. In your initial post, you suggested that the case that poses a problem for Scanlon but not for you or Smith is the “recalcitrant racist” “who firmly rejects racism but nevertheless occasionally thinks someone’s race is a reason for treating him as inferior”. I took this to be a case of an unwilling or alienated racist. And you seemed to agree that this was the issue of contention, as you responded in terms of attribution according to cares, etc. (And you said if it wasn’t a case where attribution holds, it might be like Tourette’s, etc.) When framed that way, I maintain that your care-focused view and Smith’s equally face the problem of not being able to account for the alienated racist. But your latest comment seems to suggest that it’s not about alienation at all. Indeed, focusing on cases of recalcitrance or alienation are sort of beside the point on your latest comment, if I understand it. Instead, the crucial question is whether the attitudes in question are, not attributable to the agent, but part of a functioning deliberative system. (I had a hard time sorting out the last two sentences of your last comment, so while I think that’s what you are saying, please correct me if I’m wrong. Maybe you were saying that for attribution, the attitude in question must both be non-alienated and play a suitable deliberative role.) But then I guess I have something similar to Jussi’s question. Why couldn’t Scanlon incorporate a “deliberative functioning constraint” within his existing view?
Glad to see you guys are gravitating in the same direction now so I can just address you both at once.
Here’s what I was thinking. These sorts of normal deliberative connections and functions are necessary conditions of genuine attributability, so if they’re absent w/r/t to some particular attitude, this alone can undermine attributability. Nevertheless, I’d like to be able to maintain that it’s not solely in virtue of its having a place in normal deliberative functioning that an attitude is one’s own; rather, it’s because of its connection to some specific aspect of that deliberative mechanism (e.g., cares, evaluative judgments) that’s taken to be representative of one’s “real self.”
So yes, it’s open to Scanlon to appeal to the deliberative condition as I’ve articulated it, but it’s still an open question how, if at all, the attributability of attitudes is connected up specifically to evaluative judgments, even with that constraint. (To that extent, Jussi, my charge is, as you say, just one of incompleteness.)
The reason this matters is that, in What We Owe to Each Other, Scanlon seemed to connect up attributability to answerability (and this is something Smith does quite explicitly as well): that an attitude is yours means you’re in principle an appropriate target of a demand for explanation/justification of it. Presumably, then, attributability still needs to be connected up to evaluative judgments to provide this answerability aspect of moral responsibility (how could one be answerable for an attitude that wasn’t connected up to evaluative reasons, say?). Merely being part of a normally functioning deliberative mechanism is insufficient for that, then, and I’m merely asking how the connection to evaluative judgments — and thus answerability — is preserved in the new account.
I hope that makes sense. I’m just thinking through many of these ideas.
You probably know this, but….Wallace raises something like your worry in his author meets critic paper in Ethics (2003) & Scanlon provides some enlightening comments about in his response. Specifically, he discusses why wayward desires are attributable to us even if we are alienated from them.
D’oh! No, I didn’t know that, Brad, but thanks. I’m just now getting around to investigating this particular sort of account of responsibility (which is very different from my Strawsonian roots), so I’m still catching up on all the relevant literature.
I do not have the Ethics issue here, but I seem to remember that he says the desires and actions at issue are attributable insofar as they are (or are appropriately connected with) judgment sensitive attitudes – i.e. ones that are “in principle” responsive to judgment.
I have not had a chance to read Scanlon’s new book until this summer, so thanks for the post – it has me excited to get started reading!
David,
that makes sense. I was looking back at WWO and there is a discussion of an example by Nagel that is similar to the racist case we have been talking about (pages 273-274). Here Scanlon tries to explain why the tendency to be greedy can be attributable to the person even if the person affirms the opposite judgment (at this point Scanlon explicitly identifies the person (the real self) with the considered evaluative judgments). It seems that the reason why the tendency can be attributed to the person is that the agent (the considered judgments) is at least struggling to modify and retract the tendency. This is why we can praise the agent for trying to deal with the attitude even if he would have a better person without that attitude (so he is accountable for that too).
I wonder if this is enough to connect the attitude to the person’s real self. First, if the agent was not taking the greedy attitude to be part of himself, it’s not clear why he would try to struggle to modify that attitude. In this sense the agent is taking a responsibility of that attitude. And, you might think that taking responsibility is enough for responsibility (on the lines of Fischer and Ravizza).
Second, this seems also imply that the agent does regard the tendency to be greedy to be a judgment-sensitive attitude even if the struggle shows that irrationality is blocking that sensitivity at the moment. So, this seems to connect to his view of accountability being based on reason-responsiveness.
Hi David,
I was wondering about the “primarily” in this sentence:
“…given that blame primarily consists, allegedly, in a modification of one’s attitudes towards the blamed person…”
Does Scanlon say or imply that the judgmental component of blame is only secondary? I’ve read the book and don’t recall him insinuating this. But perhaps I just missed it.
Sorry, I was away at a conference and then had to catch up on other things. Perhaps the conversation is now stale, but nevertheless…
Jussi: if you’re still interested, perhaps you could clarify a few things in your last remark. First, I’m not sure why you think that Scanlon thinks the reason an attitude is attributable to someone is that he’s struggling to modify/retract it. I just thought it was a matter of it’s being connected to an evaluative judgment. Second, I’m not exactly sure what you’re saying in your second point: perhaps you could expand a bit?
Steve: I’m not entirely sure what you’re asking either (my uncertainty knows no bounds). If you mean by “secondary” “lower in importance,” I surely didn’t mean to imply that that’s what Scanlon has in mind w/r/t judgments of blameworthiness. All I meant in the quoted sentence was that Scanlon spends most of his discussion of blame talking about the modification of attitudes aspect.
David,
I just thought that this seems to be what he says on page 273 in WWO but maybe I’m reading that passage wrong.
Well, I was thinking that the agent would not even try to modify/retract this attitude if she didn’t think that the attitude was not even in principle (or in ideal conditions) responsive to her judgments about reasons. And, if the agent regards the attitude as being modifiable at least in principle, then this seems to be the right sort of connection to evaluative judgments and the attitude being the agent’s own. This would explain why the agent can be held accountable for the attitude.
Hi guys,
Here is a relevant snippet from Scanlon’s 2002 response to Wallace (sorry – got the year wrong above) that supports Jussi’s reading and adds a related point about alienation:
“Wallace agrees that even wayward desires that are not responsive to an agent’s judgment “must in some sense be attributable to the agent, otherwise there would be no special need for the agent to struggle against them” (p. 441). The question is, in what sense are they attributable? I maintain that it is the sense in which all of a person’s judgment‐sensitive attitudes are attributable to him or her. There is a need to “struggle against” wayward desires just because they involve judgment‐sensitive attitudes, that is to say, ones that are in principle responsive to one’s judgment. Even the idea that we are “alienated from” such desires presupposes this: we are alienated from them because they conflict with, but resist, our judgment. Without this element of conflict they would be “alien” only in the sense that our height is—that is to say, merely in being beyond our rational control.”
I have worries about this doctrine, though.
I bite my finger nails. I presumably have desires to do so, which explain my doing so. But I am reluctant to say that my desire to bite my finger nails is judgment-sensitive.
And yet I act as if my desires and the resulting behavior are attributable to me in some sense; I am sometimes ashamed of my bitten finger nails and of biting them, I struggle against my desire to bite my nails, and I feel alienated from the desire (in some sense).
Not sure what Scanlon would say about this case – perhaps he discusses one like it in his reply to Copp and Sobel when they press his rationalist account of desires? Will have to look that up…
Ah. Thanks for the clarification, David. Well, as you mention, Scanlon explicitly treats blame as involving both a judgmental and an attitude-modificational component. That seems pretty solid textual evidence that, on his account, the judgmental component is necessary. The mere fact that he later talks much less about this component and sometimes fails to mention it doesn’t strike me as a good reason to question the necessity of it. So, I’m yet not sold on the points you make immediately following the in-the-thrall example. But, I guess these aren’t the main claims of the post either.
David (and others):
Thanks for this very interesting discussion. I am very pleased that you are taking up these issues. Here are a few thoughts in response.
First two small points of clarification. I say that a modification of one’s understanding of one’s relationship with a person counts as blame only if one makes this modification because one sees it as appropriate in the light of a judgment of blameworthiness that one accepts. I say this to distinguish blame from cases in which a person modifies his or her understanding of his or her relationship with a person for other reasons, such as because he or she has come to be envious of the person, or has come to feel inferior to him or her, or has learned that that person’s ancestor belonged to a racial or ethnic group that the agent dislikes. The modifications involved in these cases can be similar in immediate content to some of those involved in blame: ceasing to wish the person well, ceasing to trust the person, ceasing to see the person as someone to be associated with, join projects with, and so on. This does not mean that in order for a case of modification of one’s relationship to count as blame the judgment of blameworthiness on which it is seen to be based must be correct, or justified.
A second point about Strawson. David is correct that I do not share Strawson’s methodological position about responsibility. I find the aspects of his article dealing with the threat of determinism the least persuasive parts. But I referred often to what he says about the content of particular reactive attitudes because my views on these matters are in some respects close to his but different in other important ways. I wanted to be clear about these similarities and differences in order to make my own view clear, and also to acknowledge my indebtedness to him on this question.
About attributability (Here I am trying to think through the question, not to look back and try to interpret what I have written. I think Jussi gets my published views pretty much right.) An attitude is judgment sensitive in my sense if (in virtue of its content) it is the kind of attitude that would be responsive to the agent’s judgments about reasons if the agent were fully rational. This does not mean that the attitude must in fact be responsive to the agent’s judgments about reasons (we are not fully rational) nor does it mean that the attitude must be “grounded on reasons” if that means that it must arise from some judgment about reasons that the agent holds.
It seems to me a mistake to say that an agent is not morally responsible for (open to moral reactions on the basis of) ny attitude that is not in fact responsive tohis or her judgment. The significnce of the unreflective attitudes of friends and lovers, for example, seems to me to show that this is a mistake.
If the OCD person thinks that he has reason to wash his hands, for example, then this is a judgment sensitive attitude, even if it is the result of his abnormal condition. Is this attitude “attributable to the agent?” Well, it would be odd to say that it is not an attitude he has, since he acts on it, and may not even disown it in the way that the unwilling racist does. One possibility, which I suggested in my essay in the Frankfurt festschrift, is that any attitude that occurs in the agent’s mental life is “attributable to him.” (That it is a mistake to look for a criterion of ownership of the kind that Frankfurt sought for many years.) There is no “on/off” distinction between those that do and those that do not belong to the agent in some stronger sense. The differences between attitudes for purposes of moral response are, rather, qualitative and matters of degree, having to do with the significance of those attitudes for our relations with the person. So for example, a tendency to have hostile thoughts about certain kinds of people has greater significance for our relations with a person than a tendency to think that one’s hands are dirty. This significance is also affected by the degree of integration with other attitudes. So, for example, a momentary nasty thought implanted by electrical stimulation belongs to the agent, but because of its transitory and isolated nature it does not matter for our relations with him. Anyone would have such a thought when so stimulated, and the fact that he had this thought at that moment is not a basis for reinterpreting his motives and attitudes toward us or others over time (past or future.) A lasting tendency to nasty thoughts, whatever its origin, matters more, although it origins may also matters for our overall reaction to the person (whether we should see him as a victim, e.g.)
The class of judgment sensitive attitudes is significant on this view because such attitudes are particularly significant for our relations with each other (not because they are “attributable” or “belong to” the agent in a sense in which other attitudes to do not. Or, to put the same point differently, we tend to think of these attitudes as “attributable” in a way that others are not only because their content has this greater significance.
I am not certain about this, although I find it very plausible. So I will put it as a question or hypothesis (this is pretty much what I said in my paper on Frankfurt): Do we need a tighter criterion of “ownership” of attitudes, which singles out those for which the agent is open to moral responses? The suggestion is that we do not.
Tim: Thanks so much for the contribution. You’ve clarified a number of key points. As I suggested at the beginning of the post, I find your view on blame to be extremely plausible, and I’ve been seriously rethinking my Strawsonian commitments in light of it. So at this point, I’m merely pressing at the margins of the view to see how it handles some key cases in the literature.
Just a couple of points in response, then, to what you say here. First, as to judgments of blameworthiness being necessary for blame, I appreciate your point about wanting to distinguish blame from other cases in which the modification of attitudes is identical in immediate content. What I was trying to get at was a case in which a modification occurred in the face of an explicit judgment that the target was not blameworthy (because of the appraiser’s being enthralled with the target). Here I want to say that he’s nevertheless blaming the target, i.e., that his actually modified dispositions and attitudes tell the true story. Now perhaps you’d want to say that, in such a case, this person must, then, *really* judge the target to be blameworthy, despite his avowals otherwise. But then I’m not sure that the judgment of blameworthiness is doing the work I thought it was supposed to do, namely, as a more robust precondition of blame.
With regard to your final question — “Do we need a tighter criterion of ‘ownership’ of attitudes, which singles out those for which the agent is open to moral responses?” — I’m not yet convinced that we don’t. I take it the point of talking about attributability is its intimate connection with moral responsibility, that is, if X is properly attributable to one, then that means one is thereby answerable for it — the appropriate subject of demands for explanation or justification — and one is also thereby open to moral appraisal for it. But it seems to me that there could be many sorts of attitudes, say, for which one just isn’t open to moral appraisal (or, more generally, normative appraisal), and not just that the modification of someone else’s attitudes towards one might be significantly lessened. The attitudes of the OCD victim strike me as just this sort of thing, as do the one-time hateful thoughts of the externally manipulated agent. It’s not that, once I find out what’s happened to you, I merely scale back the modification of my attitudes; it’s rather that I see you as exempt from such modifications altogether. Indeed, part of the reason I still find Frankfurt’s bullet-biting strategy plausible w/r/t external manipulation in “Three Concepts of Free Action” — in which the agent who’s given a stable character can “own” the implanted attitudes — is that it marks an important distinction between attitudes one has and attitudes for which one takes responsibility. And if the point of talk of attributability is to tell a story about answerability, this distinction is critical, I think, for it marks a subset of attitudes for which the agent can plausibly be called on for defense or justification. So I have no problem agreeing with you that, in one sense, these are indeed attitudes the agent has, but in another sense — the sense relevant for moral appraisal and answerability — I still feel like something more is needed.
David–
As to your first point, my response is essentially the one you have predicted. I would say that if the person, because in thrall, refuses to admit the other person’s blameworthiness, but really recognizes that he has been treated badly and is responding to this, then the attitude in question is blame. But if it reflects a shift in attitude for some other unconscious reason (e.g. that he is beginning to be in thrall to someone else), then it is not blame. I don’t see why this dependence undermines the role of blameworthiness, since it is still whether the person, unconsciously perhaps, judges the other to be blameworthy that determines whether it is a case of blame or not.
On your second point, I should have said more about “answerability.” In the most basic sense an agent is answerable for those attitudes about which it makes sense to ask him for a justification. Answerability in this sense is a corollary of judgment sensitivity. So this covers more than attitudes that are the basis for moral appraisal (such as beliefs about the big bang, etc.) Among the attitudes for which one is answerable in this minimal sense (i.e. judgment sensitive attitudes) those that are the proper objects of moral response are singled out by the fact that their content makes them significant for our relations with one another. If we could provide a general characterization of this class of attitudes that is more specific than the one I just gave it would be morally significant and interesting, but it would not, I would claim, be a criterion of “ownership.” It would not tell us when an attitude “belongs to the agent” in the sense required for it to be an object of moral response, but would identify (on the basis of their content) which attitudes, among those that are so attributable to the agent, are morally significant.
A final point that may seem niggling: I say “an appropriate object of moral response” rather than “an appropriate object of moral appraisal” because it has come to seem to me important that blame is not (merely) a form of appraisal. It is not a negative correlate of praise.
Thanks again, Tim. I’m sure you have many other pressing demands, so I don’t expect you to want to continue the discussion much (or any) further. I’ll just say two things briefly. On the first point, I was, I suppose, thinking that the judgment of blameworthiness was thought to be a conscious judgment, such that it would help meet certain epistemic demands relevant for the ethics of blame you go on to discuss. This was what I meant by a more “robust” precondition of blame. Whether or not we can morally assess someone’s modification of attitudes under the rubric of the ethics of blame depends on whether or not we have some sort of access, then, to its necessary conditions. But in the “enthralled” case, whether or not the agent’s attitudinal modifications are blaming responses depends on facts about him to which none of us (including the agent himself) have epistemic access (or even evidence). So there could be two such agents, both of whose attitudes shift in identical ways, and both of whom have and express conscious judgments that they don’t blame the target, but only one of whom is unconsciously responding to the poor treatment and the other of whom has unconsciously become enthralled with someone else. This would render only the former the appropriate subject of judgments of blameworthiness for his blaming behavior, but any such judgment made would have to be evidentially ungrounded. But as I said, this seems just an epistemic worry that probably doesn’t attach solely to your view and anyway is irrelevant to the metaphysical or conceptual structure you’re interested in.
As to the second point, your distinction between thin and more normatively robust answerability demands is helpful. I see more clearly, in particular, its connection to the much larger normative theory at stake, and I don’t want to question that here, certainly. I will try to address this general point in more detail in a future post, however. In particular, I want to explore in more detail what a view like yours says about psychopathy.
http://www.3quarksdaily.com/3quarksdaily/2009/09/3qd-philosophy-prize-2009-finalists.html
Hello, you have won second place in our philosophy blogging prize contest, as judged by Dan Dennett. See here: http://www.3quarksdaily.com/3quarksdaily/2009/09/the-winners-of-the-3-quarks-daily-2009-prize-in-philosophy.html (we’ll appreciate if you leave a comment there)
Please contact me to claim your $300 prize money.
Thanks, and congratulations!