Five years ago I was up-to-date on virtue ethics. I've been ignoring it for a few years, so please forgive me if what I'm about to say is old news.
Rosalind Hursthouse argues: "An action is right iff it is what a virtuous agent would characteristically do in the circumstances, except for in tragic dilemmas, in which a decision is right if it is what the agent would choose, but the action decided upon may be too terrible to be deemed right or good."
One traditional worry with this is a variation of the Euthyphro problem: Does the virtuous agent perform the action because it's right or is it right because the virtuous agent would do it. Hursthouse takes the first horn of the dilemma and tries to argue that this is not a problem for virtue ethics.
However, I think that the analysis of right action is mistaken. There are (I think) counterexamples.
For instance, consider Richard Sylvan's Last Man thought experiment: You know that you are the last sentient being who will ever exist on earth. Beside you is a giant old redwood (or the Mona Lisa or whatnot). You could destroy it for no reason. Would it be wrong to? Some people try to argue that it would be wrong, but another take (defended by Tom Hill, Jr.) is that it would be show bad character, but not strictly speaking be wrong.
Or, suppose I am about to perform a supererogatory action. (E.g., volunteering at a soup kitchen after one's basic duties of charity are discharged.) Someone offers me a small amount of money not to do so. I take the money and fail to perform the supererogatory action. I think a virtuous agent would not accept the money and would instead perform the supererogatory action. It would exhibit vice to take money in this case. However, it's not strictly speaking wrong.
So, I take it that Hursthouse's analysis of right action is wrong because there are actions which are right (permissible) that a virtuous agent would not do.
Has someone already argued this somewhere?
Hi Jason,
This falls outside of the five-years window (so forgive me if this isn’t helpful), but Robert Johnson has some pretty nice counterexamples in his “Virtue and Right,” Ethics 113 (2003). However, his examples are of a different sort than the kind you’re trying to get off the ground. (If memory serves, his are focused on the obligations that are unique to imperfect agents.)
In the first case, I think the virtue ethicist should simply deny that it is permissible to destroy the redwood. In the second case, it is important to distinguish two questions: (1) Does VE imply that it would be wrong to fail to perform the charitable act? (Presumably not.) (2) Does VE imply that it would be wrong to decide whether or not to perform the charitable act based merely on the money? (Yes, but this is not implausible.)
Showing virtue ethics false is, however, harder than one might think. Rather than being obviously false, most people worry that it is empty of content. There’s always many things virtue ethicists can say in these cases.
Here’s couple on the top of my head. In the supererogation case, the basic actions are such that some virtuous agent would do them. Some fully virtuous agents would not work in the kitchen and some fully virtuous agents would take money from others. So, both of the basic actions are permissible by virtue ethicists lights. Maybe they can say that doing combinations of permissible actions is always permissible even if no virtuous agent would do that combination. No wrong actions from permissible ones would be the claim here.
In the first case, it seems like virtue ethicists can say that either it is wrong to destroy the tree (seems right to me) or it can be possible to have all the virtues and yet have the flaw in character that makes one to destroy the tree. The latter option seems fairly plausible: virtues are characteristics that enable us to flourish. Yet, it is not clear whether destroying the last tree has any effect on anyone’s flourishing.
Thanks for the shout-out Josh.
Jason, I think your example is interesting because it is a permissible that a virtuous person wouldn’t do because it is in some way not exemplary. My own examples were of actions that are not only right, but even obligatory (I myself think, more strongly still, exemplary) that a virtuous agent would not do (e.g., a habitual liar should monitor himself in certain ways). But your sort of example seems harder to explain since in effect it admits that the action is a bit unsavory, but falls short of being impermissible.
Jussi, some virtuous agents would take the small cash payment rather than work in the soup kitchen?
I don’t understand how this some/all virtuous agents trick works. Take some paradigmatically supererogatory act, like a Good Samaritan act. Suppose the only reason against aiding is that it’s a lot of trouble. The idea is that some virtuous agents would drive past the stranded motorist in such a situation? That seems wrong to me.
Nice worry to raise about Hursthouse’s view. I think a close worry was raised by Sobel and Copp in their survey article in Ethics, although, I also think your example adds a nice emphasis. Here is a quote from the Copp & Sobel:
“Even if Hursthouse’s proposal is amended in the ways suggested so far, it is still subject to objections. For one thing, something that an informed virtuous person would do, acting in character, might be morally trivial or even optional, such as carrying an umbrella when it is raining. It might be good to carry an umbrella, but it presumably would not be morally bad or wrong not to do so.”
(2) I suspect that Jason Kawall uses cases like yours to motivate the move to a virtuous observer account of right action…but I can’t find the article on my computer right now.
Jamie,
I didn’t say that. I agreed that no virtuous agent would take a small cash payment rather than working in the soup kitchen.
I was wondering whether that needs to be enough to make the action wrong though even on the virtue ethicist’s view. You could think that that act consists of two basic actions; taking money from another person and working in a soup kitchen.
I was thinking that these would belong to the merely permissible but not required actions. The thought was that the acts belonging to this category are ones which some but not all virtuous agents do whereas required actions are the one’s which all virtuous agents will do. It seems implausible that for every possible situation there is just one act which all virtuous agents can do. You would think that even fully virtuous agents can act differently in many situations.
If all virtuous agents always act in the same way, then there is no room in virtue ethics for merely permissible actions. This would mean that virtue ethics faces the objection from moral options before we even get to the tricky cases.
In any case, if taking money from other people and not helping in kitchen are the kind of things virtuous agents can do, then taking money for not working in the kitchen could be a combination of permissible actions even if no virtuous agent would act in this way.
The Good Samaritan case is admittedly tricky. I think it boils down to what virtue ethicists says about what constitutes virtues. Often people just have a morally perfect person in mind but I’m not sure whether this is the right way to go.
Here’s what one might say. You could take a community that didn’t help the person in need even if this required only little trouble and the stakes for the victim were high. It could be that individuals in such a community could not flourish. People would die, suffer, there would be a lot of anger, people couldn’t be beneficent and so on. So, virtuous people must have the disposition to help in such cases.
Compare this to a community in which all individuals help when this is little trouble but only some when it is a lot of trouble and the stakes aren’t more than being stranded on motorway. Would the individuals in this community be such that they could all flourish as human beings? It doesn’t seem obvious to me that they could not. No one died, people would get helped most of the time, there would be opportunities for beneficence, and so on. So, it seems like flourishing people would not necessarily have to help in your case.
Jamie, I thought the idea was what the virtuous agent *wouldn’t* do was nevertheless permissible. So the virtuous agent wouldn’t drive past. But it’s permissible to do so.
Or are you thinking that this is question-begging?
Ok – here is the one I was thinking of…
“Virtue theory, ideal observers, and the supererogatory” by Jason Kawall.
Abstract: I argue that recent virtue theories (including those of Hursthouse, Slote, and Swanton) face important initial difficulties in accommodating the supererogatory.
In particular, I consider several potential characterizations of the supererogatory modeled upon these familiar virtue theories (and their accounts of rightness) and argue that they fail to provide an adequate account of supererogation. In the second half of the paper I sketch an alternative virtue-based characterization of supererogation, one that is grounded in the attitudes of virtuous ideal observers, and that avoids the concerns raised in the first part of the paper.
Thanks for all the quick responses, everyone. I’ll have to write up some lengthier replies to the first few objections, but I think Jamie has captured what would have been my response to Jussi.
Just to summarize: The idea is that there are actions a virtuous agent would not bring herself to do, but which are nevertheless permissible. If so, then wrong actions are not coextensive with those actions that agents would not perform. Robert already has published some examples of these. My cases are those where the action would betray a defect in motivation and character, but is nevertheless not wrong. The trick, of course, is showing that such actions exist in a non-question-begging way.
Tom Hill’s case goes like this: He had a neighbor who paved over all of his land because he couldn’t be bothered with upkeep and maintenance. The land was beautiful and much of it was natural. We might try to argue that the action was wrong, but it’s more plausible that the action was permissible but not the kind of thing a fully virtuous person would do. If Hill is right, then Hursthouse has a problem–a virtuous agent would refuse to do certain things, but those things are not wrong.
Jussi, I see, I misunderstood. But this dividing actions trick doesn’t seem like a good idea to me. Suppose every virtuous agent would push the two blue buttons (with a time separation, so they are clearly two acts) or push the two red buttons, because pushing two buttons of the same color is necessary to save a lot of lives or something. We don’t want it to turn out that pushing a red button and then a blue one is permissible merely because some agents press the red button first and some press the blue one second.
Robert, my understanding was the opposite: what is right (= permissible) is what some virtuous agent would do (characteristically, blah blah blah). No virtuous agent would drive past, but driving past is permissible.
Jason,
So essentially you’re putting cases such as Hill’s into Hursthouse’s view, and what comes out seems wrong. Put in the neighbor case, Hursthouse’s account implies it was wrong. Your objection is that this is much too counterintuitive, I gather. Although it is the work of a flawed character, it wasn’t, for all that, morally impermissible for him to pave his land.
I know. I worry about the action-individuation bit. In your case, you could try to argue that saving lives is a basic action which isn’t just an aggregate of pushing one button and then another in a way that the action of taking money for not working in the kitchen isn’t. But this might not be doable in a non-question begging way.
In contrast, I worry about the consequences of not saying something like this. Say that I only do one bad action during one day. Take then the action of doing all the actions which I’ve done throughout that day. No virtuous person will do that action because I’ve done one bad thing. So, it would be wrong to do all the actions I’ve done that day including the actions which constitute that bigger action. But, this seems odd. All the other actions are right. It’s starting to seem like the virtue ethicist must have some story to tell about act individuation and what thus is the unit of assessment.
Hi Jason,
Perhaps Hursthouse could respond by saying, first, that she intends ‘right’ to pick out the action there is most reason for the relevant agent to perform. Next, she could say that questions about permissibility in the cases you have in mind are best understood, on her view, as issues of some specific virtue, say justice. So she could say that although it would not be right for the person to pave over the land, doing so would not be unjust. I guess the general idea would be that virtue ethics recommends that we leave behind “moral concepts” in something like the way Anscombe was taken to suggest.
I got to thinking about the later part of this (the bit about your case as a case in which the person does the wrong thing but is not unjust) by looking up her comments about Thomson in the footnotes to her (Hursthouse’s) paper on Abortion and VE.
I don’t know if I have any help to provide, but some points that might at least clarify the problem:
– Is ‘right’ in Hursthouse’s formulation equivalent to permissible or to obligatory? Jason, your supererogation counterexample depends on reading it as obligatory, but the language of ‘would’ seems to invite the weaker permissibility reading. If it read ‘would do, believing it obligatory to do so, ….’ would that then answer your counterexamples?
– I’ve not read Jason Kawall’s paper, but it prompts the thought that ‘virtuous agent’ might not be equivalent to perfectly virtuous agent, i.e., that virtue might be a threshold notion allowing for some state of character above it (supervirtue, say). If so, then supererogation could be explicable in terms of what a supervirtuous agent would do in a given set of circumstances, but that a virtuous agent only might do in those circumstances. Perhaps that’s what Jason is trying out in his paper by distinguishing between virtuous agents and virtuous ideal observers.
Perhaps, off the top of my head, the Virtue person would do better to say that an act is wrong if a virtuous person would have a distinctive kind of attitude towards the action (or towards their non-virtuous counterpart doing it), rather than not do it. This would help with the cases where virtuous people would avoid doing stuff that we do not want to say rises to the level of being wrong.
I’ve always thought that once a virtue ethicist reaches for the attitudes, the virtue seems to little if any work in the theory. Why not just identify the attitude that is relavent and construct a qualified attitude view. The virtue part of the story doesn’t help by providing the ideal person with the means to ‘perceive’ the right act, since we’re trying to give a theory of what makes right actions right. Of course, it might be that we have to identify the right-making attitude by way of its being in the fully virtuous person. But that’s a mere heuristic.(Kawall should weigh in on this.)
Brad wrote:
“Perhaps Hursthouse could respond by saying, first, that she intends ‘right’ to pick out the action there is most reason for the relevant agent to perform.”
I worry about this, though, because it’s not obvious that the action that there is most reason to perform is the one a (Hursthousean) virtuous agent would do. I don’t want to assume that morality always trumps non-moral concerns.
More importantly, it’s not obvious that the agent has the most reason not to destroy the last redwood. He has no good reason to do so, but he might also lack a reason not to do so. (Let’s suppose he’s indifferent to the redwood’s beauty, etc.) If so, my take on it is that he’s got some deficiencies of character, but no reason not to destroy the redwood. It’s not wrong to destroy the redwood, but only a person of imperfect character would be willing to do so.
I am sympathetic to your response to the claim about reasons, Jason, but that is because I am a fellow reasons-internalist.
Two further thoughts:
First, I worry this response might leave you with a much weaker argument overall. If you make your overall case by attacking the claim about what the guy has most reason to do, that will not compel externalists. Your initial argument, by way of contrast, was less contentious because it rested on the assumption that the man acted poorly but permissibly – an assumption many externalists would accept.
Second, I fear the problem gets more pressing when you consider that my postulated response might not be the best in the area.
I was looking over Hursthouse’s book last night. She herself seems to have trouble deciding how to elucidate her use of ‘right action’. She seems to reject the use of ‘right’ that implies or entails claims about obligations and permissions. Her positive comments appeal to the concepts of *acting well* and I get the sense she might be ok with talk of the right action being the most choiceworthy.
This suggests she could say this: the man does not act rightly in paving the land, in the sense that he fails to perform the most choice-worthy act, but he does not act unjustly or wrong anyone else.
This seems better than the previous line I suggested as it does not obviously conflict with internalism (about, e.g., reasons and blame).
I’ll second something Brad said. It’s been a while since I read the Hursthouse book, but in general I would be slow to equate ‘right’ in a virtue ethicist’s mouth with either ‘permissible’ or ‘obligatory’ in a Kantian’s mouth. The spirit of VE, it seems to me, involves not thinking of morality as a quasi-legal construct segmented into obligatory/permissible/impermissible actions.
Insofar as a virtue ethics has a use for these concepts, I think Jason is right that one should admit that there are permissible but non-virtuous actions.
Good, so maybe the argumentative tact to take is this:
Either A) Hursthouse means her account of right action to explain the same kind of thing that utilitarians and Kantians want to explain, or B) she wants to talk about what we have most reason to do.
If A) then my objection–there are actions that virtuous person would refuse to do which are nevertheless permissible (and thus right). If so, then her analysis fails.
If B) then I need to argue that her account is implausible if it’s supposed to be an account of what we have most reason to do.
My default take on “morality” is that it refers to a certain system of norms (distinct but possibly overlapping with other systems of norms, such as self-interest, law, etc.), and it’s an open question whether we have most reason all things considered to abide by these norms.
That sounds right to me. The main work will come, I think in defending B and it will help that Foot and Hursthouse have lots to say about the relevant issues.
Another way to go would be to say that there is this important question–what is permissible?–that if (B) is true, Hursthouse does not answer. Then the work would come in defending the importance of the question.
Now that would be an interesting paper!
Hi Jason,
I tend to agree with you that the cases you present pose a problem for a view like Hursthouse’s. But I think shifting to a virtuous observer theory (as Sobel suggests above) can help a virtue theorist to respond. That is, rather than looking at what a virtuous agent would do, rightness (etc.) would be fleshed out in terms of what virtuous observers would approve / disapprove of (and could factor in the strength of these approvals and disapprovals).
Offhand, then, one could hold that in the Sylvan case, virtuous observers would mildly approve of the action (thus the action would be permissible), while disapproving (mildly) of the agent’s motive.
Similarly, virtuous observers would presumably disapprove of the motives of the agent who takes the money not to perform the supererogatory action. They could also mildly disapprove of even taking the money in this case (but perhaps not so strongly as to make the action wrong).
A couple of people have kindly mentioned my Virtue Ethics, Ideal Observers, and the Supererogatory – some of what I say there might be relevant to the issue of the supererogatory for virtue ethics (though I don’t say much about the kinds of approvals that would need to be involved in distinguishing right and supererogatory actions).
Uhm – this is perhaps poor form, but I have an earlier paper, Virtue Theory and Ideal Observers (Phil Studies 2002, 109(3), 197-222) that uses a few counterexamples to motivate the shift from a Hursthouse-style view to a virtuous observer view. I think some of these counterexamples might pose a problem even for claims that what a virtuous person would characteristically do in a situation would be right (even in this is not treated in terms of the obligatory/permissible, as Heath suggests) – consider what a virtuous agent would characteristically do if very drunk, or under the influence of a drug that increases aggression, etc.
You might look at Gregory Mellema’s book, Beyond the Call of Duty – he tries to address cases like the ones you discusss (from a broadly v.e. perspective) , and makes use of such notions of the suberogatory in so doing.
A final plug – I try to address some of the Euthyphro-style worries in a paper now out in JESP “In Defense of the Primacy of the Virtues”.
Heath,
That sounds like a great suggestion.
Thanks to everyone else, too. Lots of stuff to think about. I think I’ll just try writing a paper on this and see what happens.
Jason: One quick note regarding the dilemma you’re thinking of foisting on the virtue ethicist — I suspect that at least some virtue ethicists will be happy to say that moral considerations are not always overriding, and so grant that in any particular situation it’s an open question whether we have most reasons to abide by moral norms. Foot, in *Natural Goodness*, explicitly says as much. The question, then, is whether there are any examples of the kind you need to illustrate B). For in those cases where it seems most plausible that the thing to do is not act on the relevant moral reason, the virtue ethicist might well agree with you (i.e. they will agree that the virtuous agent will not so act).
The larger issue in the background, I suspect, is that many virtue ethicists will simply disagree with your way of construing morality as a system of norms that might overlap with, but is ultimately distinct from, the norms of self-interest. But perhaps that’s neither here nor there for the issue at hand.
Long-time reader of PEA Soup, first-time commenter. Hope no one minds the intrusion!
Hi everyone,
This is a very interesting discussion.
Michael is correct to ask whether ‘right’ means
permissible or obligatory for Hursthouse. And I would agree with Brad that she seems to have trouble dedicing how to elucidate the concept. At some points she says simply ‘right action’, at others she speaks of a ‘morally right action’. I think it’s important to keep in mind that she wants VE to be a ‘genuine rival’ to utilitarianism and deontology. To do that, something like morally obligatory action must be covered, even if other things are also covered.
Many cases discussed involve the right action which the exemplar wouldn’t do. I think some of the cases show problems with H’s principle. I think a somewhat cleaner counterexample would be one where x is what the exemplary agent would characteristically do, but x is not related to morality at all. It is amoral. Jane is a businesswoman who exhibites fortitude in getting her new business up and running. It might be replied that x is permissible. But the example shows us how far the virtues will take us from the moral concerns of the desired rivals, utilitarianism and deontology, for Hursthouse must say that Jane exhibits a good moral character. And when merely permissible action flows from good moral character it seems to be a problem for the theory.
Just to add to what I said earlier: Heath’s right that some V-ethicists want to junk the standard deontic statuses, but by providing an analysis of ‘right’ action, Hursthouse seems to be placing herself outside that circle of V-ethicists. Given this, if ‘right’= obligatory here, my main worry about her formulation is that speaking of what the virtuous agent ‘would characteristically do’ doesn’t provide a sufficiently fine-grained explanation of the facts that move the virtuous agent to do what she ‘would characteristically do’. It doesn’t differentiate among moral and non-moral reasons, for instance, so that ‘right’/obligatory could sometimes designates what is prudentially best.(I imagine some V-ethicists would welcome collapsing such distinctions. I don’t.) At that point, either ‘right’ doesn’t mean morally obligatory or ‘virtuous agent’ doesn’t mean morally virtuous agent. But as Jason’s post points out, it also doesn’t differentiate among the deontic statuses and the reasons that underlie them. At the very least, the righthand side needs some reference to the virtuous agent’s states or reasons.