Blog Coordinator

« My Plan | Main | Symmetry and Responsibility »



Feed You can follow this conversation by subscribing to the comment feed for this post.

A theory of moral responsibility should explain who has moral obligations and (or?) moral restrictions, and deviations from the moral norms of how to treat those people.

That is, why people (or things) are morally responsible in general, and what is it that makes them morally responsible in particular for some action that makes it morally acceptable (obligatory) to treat them in a way that's normally morally forbidden (optional).

Great kick-off question, Matt. Three points.

First, note that your title was on desiderata for a theory of responsibility, but in the text you request desiderata for a theory of *moral* responsibility. One might think that the latter is a proper subset of the former, so that our first aim should be to the former: one desideratum of a general theory of responsibility will be that it tells us a good story about *moral* responsibility.

Second, I'm reminded of Strawson's implied desideratum for a theory of responsibility, namely, that we account for "what we mean, i.e. of *all* we mean, when, speaking the language of morals, we speak of desert, responsibility, guilt, condemnation, and justice." A theory doing less well in accounting for some of those terms might be less plausible than others.

Finally, relatedly, nearly everyone characterizes "moral responsibility" as something for which a range of responses is appropriate. One might then think that one way to figure out the desiderata for a theory of (moral) responsibility would be to articulate all of those responses and then perhaps theorize about their appropriateness conditions.

I am not at all optimistic that we will come to agreement about what needs to be explained by "the" theory or moral responsibility. I think the term 'moral responsibility' is way too indeterminate in extension, admitting of all sorts of reasonable but different precisifications, to make such agreement possible or desirable. Why not just accept massive pluralism about the questions in this area? Our starting-point desideratum is just this: ask clear, well-defined questions about the diverse "moral responsibility" phenomena in the neighborhood that we care about. The good theories are the ones that give good answers to these questions. Some of these questions are about retributivism and lots aren't.

Thanks for the comments, Mark, David, and Nate. I've been without internet all day, struggling with my ISP to get the problem resolved. I hope that it's finally fixed -- but it means I haven't been able to participate in my own post!

I agree a theory of moral responsibility should tell us who is morally responsible for what. But I worry that this invites cross-talk unless we clarify some of the concepts that we take to be interconnected with moral responsibility.

At first I thought you had scored my initial post! You're right to note that I conflated responsibility and moral responsibility. That's in part because I'm an eliminativist about moral responsibility, since I don't think it's a distinctive *kind* of responsibility at all. It's just responsibility for morally relevant stuff. Still, I quite agree that we should want our account of moral responsibility to connect up with a more general account of responsibility.

I like the Strawson quote, and I'd be sympathetic to including desert, guilt, and condemnation in the cluster of concepts related to responsibility. So that's a reasonably good start -- though I doubt it exhausts "all" we have in mind. Surely, though, we have to be open to some revision, no? Our language and practices can't be immune to where our theories take us, can they?

I'm also sympathetic to beginning an investigation into the set of appropriate responses. More on this in a later post -- but then I wonder just what those responses would have to be. Would blame and praise be included? Some theorists seem intent on excluding praise altogether.

Fair enough. So rephrase my question as "what set of diverse 'moral responsibility' phenomena properly belong in the neighborhood?

To all -- just to be provocative, here's a claim: deserved punishment is not something a theory of (moral) responsibility need explain.


I agree with you that BW and PW are closely related. With that said, it seems to me that BW and PW have to do with the morality of an action, which I’m thinking is a fundamentally separate issue from the responsibility of an action. In order to get closer to a “desideratum for a theory of responsibility”, perhaps we need to separate the concepts of “morality” and “responsibility". Here’s why I say that: Responsibility is absolute in nature, whereas morality is relative to the opinion of the observer. So when we combine the ideas of morality and responsibility into a singular term named “moral responsibility”, we’re really talking about an observer’s opinion regarding what some other person is responsible for. That’s why I think the concept of “moral responsibility” can get a little slippery.

To simply communications, perhaps we can separate BW and PW from the concept of responsibility. For example, an observer could say that a person’s physical body is responsible for event A (since their body exerted the forces which caused A), and once we agree on that (i.e., the responsibility aspect), we can talk about whether or not the person’s action is BW or PW (wherein BW or PW is relative to the moral views of the observer).

It’s great to see someone trying to nail down the meanings of terms, thereby helping to improve communications when we’re debating competing theories!

Moving on… You’re wondering if control is required in order for an agent to be responsible. My opinion is yes, but things can get a little complicated depending on how you define the term “control”. I’m thinking there are two different meanings, a “weak” sense and a “strong” sense, just as there are similarly two different meanings to the term free will. For example, in one case we could use the “weak” sense of the term control (i.e., given a choice, an agent can make a decision). In that case, it seems pretty straight forward that the person’s body is responsible for the forces which cause the actions, thereby resulting in the agent being responsible. In another case, however, we could use the “strong” sense of the term control (i.e., given a choice, the agent can make a decision that isn’t controlled solely by the four fundamental forces of physics). In that case, we could say that in order for the agent to be responsible, the strong sense of control must exist, which mankind hasn’t quite proven yet. So in the second case, if you don’t believe the “strong” sense of control exists, then you’ll likely believe that an agent isn’t responsible for his actions.

Finally, you asked “is there anything an account must explain in order to count as a theory of responsibility?”. My opinion is that an account must explain which particular entity (i.e., agent) exerts the forces which cause the action. Otherwise, the account isn’t capable of explaining responsibility for the action.


Wonderful to have you here!

Yours is a great question. Maybe, in some sense, the big question that has to be answered before we even know if all these purportedly competitor theories are even disagreeing.

For what it is worth, I think of my own aspirations as tiered in two directions.

On the first axis of tiers: First, I want to know if certain practices (centrally, but not exclusively, moralized blaming) admit of justification. Second, I want to know if the justification is roughly akin to how we pre-philosophically (or even in the early stages of philosophy) imagine it to be. Third, I want to know how far away a philosophically adequate account is from our pre- or preliminary philosophical assessments, and what we give up in not capturing those views.

There is maybe a second axis of tiered commitments I'm interested in. I care most about an assessment of our doings, and if some of our beliefs have to give way, I'm generally okay with that. On one way of putting it, I think it is a small price to pay for being licensed to praise and blame if it turns out that the basis of these things is at some remove from what we initially thought. So the normative status of the practices matters to me more than the whether or received beliefs are true. Similarly for more coarse grained to finer-grained beliefs. (So, that we are justified in praising and blaming is more important that whether people deserve praise and blame; and that people deserve praise and blame is more important than whether they deserve it on some narrow notion of desert.)

I'm less sure about where the reactive attitudes fit in this picture.

Of course, there are other complexities that your question point us to. Perhaps some accounts of responsibility sufficient to justify moralized praising and blaming may not be beefy enough to support, say, particular retributive state punishment practices. But that strikes me as an indirect cost to a theory of responsibility, not a direct cost.

Concepts/ideas/notions don't operate in isolation from other commitments, and tinkering with one part of the web can have effects in another part. But it seems to me a secondary or tertiary desiderata for a theory of moral responsibility that it should support a given view of punishment. Or, to put the point differently, an adequate theory of moral responsibility may not give us the theory of punishment that we were hoping for.

So yeah, I'm with you deserved punishment not being central to a theory of moral responsibility.

Thanks, James. It looks like you agree with David (and me) that our theory of moral responsibility is beholden to a general theory of responsibility simpliciter. My version of this view is probably pretty idiosyncratic, and I won't speak for David. But on your view you want a theory of responsibility that explains how entities cause events, and in the case of folks like us, how agents cause actions. Do you think this is a special account? Or is it parasitic on a more general explanation of objects causing events?

Thanks, Manuel!

I meant the post as posing a big, central, and important question -- and, for that reason, one that's hard to answer. I appreciate you laying your thoughts.

I think I'm more inclined to take desert claims to be central. Many theorists characterize (or even define) their target as 'responsibility in the desert-entailing sense'. Now, as I've argued in print, this risks trying to illuminate one complex notion (MR) via an equally complex one (desert). So we need to know more about the relevant notion of desert. But I do think a desert-less account of responsibility would perhaps miss the mark.

I like your way of putting it in terms of coarse- to fine-grained. I guess I'm not comfortable going as coarse as you are. Since I think there are lots of ways we could justify our responsibility-related practices, at least in principle, but only some of these justifications will be of the "right" kind. But I'll hold off a bit on this, since I'm probably posting directly about it later.

I'm curious though, do you think the primary aim is justifying our practices because they are so important to human life? Or for some other reason? I guess I'm asking whether there's a point at which certain justifications, even if they obtained, would be of the wrong sort to get the explanations right.

An agent A is [morally] responsible for action (or perhaps outcome) X iff A controls X such that she deserves credit for X [praise if X is good and A had requisite knowledge of that and blame if X is bad and A had requisite knowledge, and (with lots of extra caveats) reward or punishment for X].

There. Now, we just define 'control' and 'desert' (and requisite knowledge and caveats) ... and we're done.

That's a really nice piece of ethnography Eddy. You've perfectly captured a historically recent, culturally localized conception of moral responsibility.. Did you live among these people or just do a lot of sociological and anthropological research?

Hi Matt,

You say:"Just to be provocative, here's a claim: deserved punishment is not something a theory of (moral) responsibility need explain."

This strikes me as a very weird claim. It seems perfectly legitimate to me for somebody to make the following speech: "I'm in interested in a particular form of moral responsibility, namely that form that underlies are punishment practices. We think people some people deserve punishment. Under what conditions is this true?" (Question: do you think that this speech involves a misuse of the term 'moral responsibility'?) I think the speech is totally unobjectionable. On my way of thinking, there are some uses of 'MR' according to which your provocative claim is true--if my "moral responsibility" problem is the problem of saying what is required for certain sorts of character-involving evaluations of actions to be true, e.g. "his act was evil" or "his act was courageous", then punishment has nothing conceptually to do with my question about moral responsibility. But there are some uses of 'MR' (like the one picked out in the above speech) according to which the provocative claim is false.


I think it is helpful to distinguish an inquiry into the concept of moral responsibility from an inquiry into particular conceptions of moral responsibility (to use the "concept"/"conception" distinction introduced by John Rawls). Granted, this presupposes a pre-Quinean willingness to accept something like the analytic/synthetic distinction, in some form or other.

And to employ another bit of Rawlsian terminology, I think we should employ some sort of "reflective equilibrium" in seeking to articulate both the concept and our favored conception of moral responsibility. But here as elsewhere, there will be much room for calling into question our initial judgments and intuitions. (And clearly there is a role here for experimental philosophy in identifying "mechanisms of thought", as Nichols and Knobe put it in their very helpful introductory essay to their anthology on X-Phi).

I'm not sure how much hangs on this, but I tend to disagree with Wallace about how much should be in the scope of moral responsibility. I tend to agree with Peter Strawson's more expansive conceptualization of the range of reactive attitudes pertinent to moral responsibility. I think we can conceive of them as "hanging together" in ways that are not quite as restrictive or constrained as the ways envisaged by Wallace. For example, they are the attitudes distinctive of and appropriate for *persons*, rather than non-persons. This of course includes the "positively valenced" reactive attitudes, as well as the negatively valenced attitudes.

Tamler raises a nice point, as one would expect, about historical and sociological variation in both responsibility-related practices and their underlying judgments. But for these to be variations, there must be some underlying core that's shared. Otherwise, we couldn't say that, to simplify, one group requires control to be responsible for x whereas another group does not. What's that shared element? Desert? That a blamed individual ought to be blamed? That blame is justified?

I like Eddy's simple account. My own favored account is very similar. It might be identical, only filling out the details. But it also seems to me that Eddy's analysis could be compatible with cultural variation, if we allow the details on spelling out control and desert and the like to be sufficiently broad.

Thanks, Nate. I don't think it's that weird a claim. All I mean to say is that a theory of punishment would be largely independent of a theory of (moral) responsibility.

Of course, one answer to "when is punishment deserved"? could be "when the individual is (sufficiently) blameworthy" or "when the defendant is MR for the crime".

But it is also intelligible to think that we can MR for lots and lots of stuff and also think that punishment is never deserved. I take it that means that it wouldn't necessarily count against a theory of responsibility that it didn't also justify punishment. At least, that's my 'provocative' claim.

Good to hear from you, John!

Well, I at least agree with you that I think Wallace's account targets too restrictive a notion. One way my question could be answered is that there is not fixed target notion, so we should just be careful to sufficiently characterize what notion we're targeting with our accounts. This is sometimes how I think about Wallace (and Darwall). Their accounts are reasonable for elucidating a notion of accountability, but I don't take it as a desideratum for a theory of responsibility that it account for accountability.

At the end of the day, I'm somewhat pluralistic about what we're after. I wouldn't want to imply that I think there's only one thing responsibility could refer to and that's the only thing worth explaining. At the same time, I do think it counts against Wallace's view that it ignores praiseworthiness (and seemingly cannot account for it), at least since he's trying to say something about warranted blame.

Matt, is another way to put your claim about deserving punishment this: We might want, and develop, and account of MR, while also believing that, because of various moral or pragmatic reasons, no one ever deserves the harm of punishment?

Thanks for answering Tamler's "response" ;-)
[he loves emoticons so make sure to include them in any response to him.] As you say, my definition (and yours) is amenable to cultural/historical variation, mainly regarding how 'desert' is spelled out. But we also should be allowed to offer a definition of a culturally limited concept that need not also apply to what might be a different concept in other cultures.

Hi Eddy,

Yes, that's another way to put it. Though I will get to this in later posts, the idea is that there is lots of space between, say, judging that someone is blameworthy and how we ought to treat them in light of that fact. In my view, very little of that space is occupied by the blame that is deserved.

But, yes, I deny that 'A is morally responsible for something wrong' tells us even prima facie that they deserve punishment. (Though it may be true that in order to deserve punishment they must be responsible for something wrong.)

Thanks for your reply, Matt.

I think it’s reasonable to believe that “agents cause actions” is a special account of the broader scope of “entities cause actions”, since agents are a subset of entities. I also think it’s reasonable to believe that some entities aren’t alive, and therefore it’s appropriate to make the distinction between the broad class of all entities and the narrower class of living entities (i.e., agents).

Some people may believe the term “responsibility” applies only to human beings (and perhaps to other living things). But I’m thinking it’s appropriate to use the term in a broader context – it applies to all entities that exert forces. For example, if lightning strikes someone walking down the street, it’s appropriate to believe the lightning was responsible (at least to some degree) for the person’s injuries. To me, understanding responsibility for a given action is about identifying the entities which exert the controlling forces (which can be a daunting task).

Matt, I have a couple more ideas regarding responsibility.

It’s possible to trace back causal chains, and therefore think you’re tracing back responsibility. When doing that, there are generally two perspectives you may have in mind. 1. The four fundamental forces of physics (4FFOP) were solely responsible for controlling the path (i.e., history is predeterministic in nature), or 2. Life exerts *new* emergent forces (forces which don’t result solely from a direct sum of preexisting forces), and those forces add together with the 4FFOP thereby helping to determine the path. If you believe in 1, then you believe responsibility really only exists in the “weak” sense, and the 4FFOP are truly responsible for all actions that occur in reality. In a sense, it becomes pointless to trace back responsibility if you believe in 1.

If you believe in 2, then you believe responsibility exists in the “strong” sense, and when you identify certain levels of agency (i.e., living systems), those agents are truly responsible for the forces they exert. Yes, there are causal chains, but those chains aren’t predeterministic in nature. They are affected continuously by input from new life. So when we’re tracing responsibility for a given action, there is likely life involved in the sequence of events, and said life is partially responsible for determining the path forward.

Thanks, James. We certainly talk about the responsibility of inanimate objects (as in your lightning example). But to my mind, this is just code for "The event of the lightning strike caused the person's being injured." That is, we use it to replace event-causal talk with object-causal talk. So I actually don't think the notion of responsibility does any interesting work here. (I would wager it's a leftover from our historical practices of imputing agency to all manner of natural phenomena, but this is pure speculation on my part.)

I'll have to quibble as well with equating the class of living entities with the set of agents. Plants are alive but not agents; there could be agents who aren't alive (e.g., androids).

I just saw your followup comment. I don't agree that if we can give reductive scientific pictures of the causal order that it becomes pointless to ascribe responsibility to agents. Reductive explanations don't thereby eliminate the phenomena they explain.

But in any case, I don't mean to be equating a theory of responsibility with a theory of causation. While there may be a necessary causal component to responsibility (as captured in control conditions, e.g.), I'm not here inquiring primarily about fixing the targets of our theories of causation.

Matt, good point, there does have to some shared core element but it has to be extremely minimalistic. Maybe a modification of yours: blame is justified, whether or not there are any consequentialist/forward-looking reasons to blame. I would leave a control condition out of it entirely, since it's an open question whether justified blame requires control. The same certainly applies to the knowledge condition (think Oedipus). Anything more that just that basic condition involving justified blame will beg questions.

The other option is to go pluralist as Nate suggests. I have a lot of sympathy with his views here.

Thanks, Tamler. That's helpful. But I wonder whether what you've proposed could really be the core of what a theory of responsibility should explain.

Let's adopt the minimal core of 'explaining when blame is justified'. This is what I took you to be suggesting. This is truly minimal, as you note, for it eschews mention of control, knowledge, desert, guilt, etc. But maybe that's worth it in order to be consistent with the highly variable social practices we find throughout history and, perhaps, their underlying judgments.

Still, this core looks to me more plausibly a description of what we might minimally want from a theory of blameworthiness. Should we give a parallel minimal core for praise? Or should we forgo such unification and treat them separately? Even if we give the same kind of description for praise (i.e., 'explain when praise is justified'), this doesn't tell us anything about what ties blame and praise together. I grant that one needn't accept my suggestion that BW and PW should be treated together and jointly illuminated by a theory of moral responsibility, but if one rejects the connection, where has responsibility gone?

Moreover, given that the practices vary so much, there'd be the very real worry, even on this quite minimal proposal, that 'blame' and 'praise' aren't referring to the same responses, attitudes, or actions between the various groups. It doesn't seem to me that anyone was indignant over Oedipus' actions, for example. Ancient Greek praising was more like honoring than showing gratitude or lauding. So, we'd only have a shared core if blame and praise were sufficiently shared. But it seems to me that much of the evidence for the significant cultural variation concerns variation over blame and praise themselves. That would mean we can't even get the minimal shared core going, and so the accounts wouldn't really be competitors. (As I'm not an expert on Ancient Greek moral practices, however, perhaps I've got it all wrong.)

(I should note that I'm highly suspicious that the cultural variability both (1) is as robust as some claim and (2) supports relativism about responsibility. Since I don't think our own cultural practices are all coherent, justified, or even possibly identifiable at any helpful precisification, I don't think there's much threat from the practices of "others". But this doesn't amount to an argument for any claim -- just reports my position.)

I feel like I perhaps should make clear that I'm not a skeptic about whether there is a set of explanatory desiderata for a theory of (moral) responsibility. And that's even if we can't agree on what it is.

So, for example, when someone like Wallace says praiseworthiness just isn't part of a theory of responsibility, I think he's just wrong about this. And this isn't just intuition; I have arguments!

But whether I'm right or wrong, I wanted to make clear that I do think there must be something any account of responsibility should explain, and that we're not nearly careful enough in getting this out into the open in advance. And this is important -- for every time I read Wallace I wonder whether he isn't trying to account for something very different from what I'm interested in.

At the end of the day, all these details may wash out. If what we want is a comprehensive moral picture which gives a coherence explanation of all our moral notions while justifying a complete set of moral practices (not necessarily *our* set), then we will plausibly check off all the boxes anyhow. But how we arrange which boxes belong under which heading still seems a worthwhile task to me.


If a person believes that the laws of physics solely control the actions of a lightning bolt, and they also believe the laws of physics solely control the actions of a physical human body, then wouldn’t it make sense for the person to also believe that the concept of responsibility applies to both the lightning bolt and the human body?

Granted, we’re only talking about responsibility in the “weak” sense.

When developing a theory of “responsibility simpliciter”, I’m wondering what qualifications you envision in order for an entity to be considered responsible? (I’ve left morality out of this comment.)

It seems to me that our intuition generally tells us there’s a difference between that which controls the actions of a lightning bolt, and that which controls the actions of a human body. But in order to justify our intuitions (and thereby believe humans qualify for responsibility in the “strong” sense), it seems like we need to believe that human actions aren’t controlled solely by the laws of physics.

Responsibility is a great (and difficult) topic!

Perhaps tangential, but here's a bit on blame in the law I found interesting. Judge Barbier just ruled against BP in the spill and fined them $18 billion, but BP was not the only one to blame: "He ruled that BP bears 67% of the blame, Transocean bears 30%, and Halliburton is responsible for 3%.... BP made profit-driven decisions that 'evince an extreme deviation from the standard of care and a conscious disregard of known risks."

(Why do we like thought experiments when the real world gives us so much to work with?) Not only is there the divvying up of blame in precise percentages (which I hadn't seen before--anyone know if it's just civil cases that do this?) but there's the reference to *conscious* disregard of risks (but by whom? not the corporation--they aren't conscious persons, just persons, right?--so presumably individuals within BP, but then does the blame get transferred to any of them and divvied up some more?)

Anyway, no expectation for anyone to respond to any of this--just thought it was interesting.

And while I'm at it, thanks to Alan White for his nice comments on Leiter (where he, John and I have been touting this wonderful blog--thanks to Thomas!).

The comments to this entry are closed.

Books about Agency

3QD Prize 2014: Marcus Arvan