Some Of Our Books


« Featured Philosopher: Robert Johnson | Main | Upcoming Ethics/PEA Soup Discussion: Cécile Fabre's "War Exit" »

April 28, 2015


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

So Robert, do you think there is a difference between the moves open to this kind of virtue theory and the moves open to defeasible internalists (a la Michael Smith before he went to the advice model) who think the things we have reason to do are what we would do if we were rational in response to conditional fallacy arguments, and those open to the virtue theorist in response to this line of objection? Since I think that removing the "completely" works for the former (and I know you don't) I'm wondering about whether a parallel move works for the latter. Your being at pains to include the "completely" here makes met think you might think it helps here.

Hi Robert

this is supercool! Conditional fallacy objections to subjunctivist theories generally and your objection to virtue theories of rightness are my favourite philosophical arguments! Just few comments:

1. It seems very implausible to me to understand (V) as a theory of right-makers in the first place. Take a situation in which someone is in extreme pain. The question is, why is helping that person the right thing to do? It beggars belief to answer this question by not saying anything about the patient in pain but rather to insist that helping the person is right because it's what a virtuous agent would do. For me, the right answer has to be something about alleviation of the pain in the first place.

2. I'd like to hear more about what's wrong with the Smith-styled advice-model as a response to this. Here you mention the skill of advising, but that doesn't seem relevant as the view can be formulated in terms of what the virtuous agent would want the less-virtuous agent to do. As I recall, in your original paper, there is another objection that had something to do with full information, but I don't see why the advisor in this case would have to be fully informed. In any case, I would like to hear more about what really is wrong with this strategy. For my part, I would assume that they make the right-making facts even more detached from what really makes actions right.

3. I'd also like to hear something on general philosophical strategies for avoiding the conditional fallacy. All the versions of virtue ethics are formulated here in terms of what Crispin Wright called basic equation (X is F iff if circumstances were C, then Y would be the case). According to Wright, we can avoid these problems by moving to what he calls provisional equations. They have the form If circumstances are C, then (X would be F iff Y would be the case). If we formulate the virtue theory of rightness with this structure, the view becomes immune to the conditional fallacy objection. The cost though will be that the view will not say anything about the rightness or wrongness of the kind of actions that your examples are talking about. But, maybe it is ok for the virtue ethicists to say something different about what makes actions right and wrong in those situations. I should also say that there are also other general philosophical strategies that virtue ethicists could explore.

In any case, this is a great objection to virtue ethics!

Hi Robert,

Thanks for the intriguing post. Could the virtue theorist propose a response of this sort?

Virtue ethics (VE) "demotes" the right, not least by subordinating the theorizing of it to theorizing the good, a good life, a good person, and so forth. Those who are not virtuous might not be able to do the right action in the sense given by criterion (V), and you treat that as a bug rather than a feature.

There might be something that the "learner" ought to do that is different from what a virtuous person would do. Why should we call that action "right" in the same sense? Your brief discussion above reads this position in terms of distinguishing 'right for the virtuous' from 'right for the learner', but why is the latter a "kind" of rightness at all? Could it not be good for learners to act as they ought without thinking that they thereby act rightly in the robust sense given by (V)?

I suppose my worry is that your approach seems to try to re-prioritize the right by assuming that each agent can do what is right, and that VE must theorize this ONE sense of rightness. Do you worry that this assumption might beg the question against VE's fundamental move?

Hi Robert (and Jussi),
Thanks for this post - some very good stuff here!
I was writing to echo Jussi's request for more of your thoughts on the problems facing virtuous advisor/observer approaches.

And Jussi, with respect to your first comment, I think many virtue ethicists would hold that suffering is itself bad (not simply because of the attitudes of virtuous agents), and that virtuous agents would demonstrate an appropriate response to this suffering. That is, they might attempt to characterize the right in terms of the virtuous, but without trying to understand the good in terms of the virtuous. Would this start to address your worry?

Alternatively, I'd actually be inclined to hold that as compassionate, benevolent, etc., the virtuous agent will want to help a suffering person - and that in some sense, it is precisely the suffering that explains why the virtuous agent would want to act, etc. But that the pain itself is bad (and not merely an unpleasant state) and ought to be alleviated in this case is a matter of the attitudes of the virtuous. [After all, there might be other cases where a virtuous agent would not see suffering as warranting alleviation - perhaps the pangs of guilt felt by a criminal as he reflects on his crimes, or the mild pains felt by a training athlete that help her to gauge her performance and not push too hard, etc.]

Hi Jason

just quickly - I am very sympathetic to the idea that virtue ethicists make sense of right- and wrong-making considerations in terms of the considerations that feature in the deliberation of virtuous agents. This would nicely tie right-making features to the reasons on which virtuous agents act. In this case, however, (V) cannot also be a theory of right-makers or we get awkward kind of overdetermination. So, the point just was that I'd rather have (V) to be a claim about something else like merely which actions are right or what the nature or essence of rightness is. This is nit-picking though.

All of these excellent questions! Let me get a start on them.
Mark, regarding 'completely'. I didn't think weakening in this way would work because of a not-very-imaginative reason: if someone's not completely virtuous, then we'll be hesitant to cast *all* of her actions as right. The 'completely' modifying 'virtuous' is there if we want to capture 'all' modifying 'right actions'. Something such as that. But I'm betting that won't satisfy you.

Jussi and Jason: (Jason we've talked about this before, but I'm not sure I can give you an answer that will satisfy you.) Suppose we marry 'complete virtue' and 'ideal observer' into a single idealization. The question is whether virtue is still doing work in the account. My thought was that it wouldn't be doing any. It wouldn't because in ideal observer views such as Firth's, you have full information, etc. plus benevolence (or something like that). My suspicion is that the set of actions captured by that and the set of actions captured by adding 'complete virtue' are the same. Or at any rate, I'm not convinced there would be a different set. This view just collapses into an ideal observer view. Or so I think.

Also, consider the 'advisor' move. My thought, again, is that complete virtue doesn't automatically give you any advising skills. I just don't see why we would expect that everything the ideally virtuous version of ourselves would want us to do would turn out to be the set of things we ought to do. Sure, our better selves would probably want us to improve ourselves. But whether it would be the right thing for us to do in any case, or what we are to do in such a case, is still up in the air. (In other words, while in general we should improve ourselves, in particular circumstances that might be wrong. Rome might be burning, after all.) To fix that, you need, as Jason himself does, to add stuff like full information. But then we're back to doubts about whether that makes this a view that is genuinely different from ideal observer views.

Jussi I think your third point might well be helpful. This is the sort of move I was thinking about when I mentioned Paul Weirich's suggestion (he works on idealization in game theory.) I don't know what to say about it though at the moment. So say more.

Michael, I don't have much if any argument against the proposal that self-improving actions aren't right except my conviction that, at least sometimes, they are. This might be a signal that people who find virtue ethics attractive and people who do not need to find something more to say than appealing to their intuitions. I don't know what that is yet myself.

Hi Robert

thanks for your response. I worry that there is some miscommunication going on. When we are thinking about the advice-model, we are not thinking about an ideal observers but rather borrowing the structure of Smith's way of dealing with the conditional fallacy in the context of reasons.

The basic structural insight is to separate between evaluating perspective and the evaluated state of affairs. The idea is that we put the virtuous person in the evaluating perspective. There is no need for full information or beneficence but rather all the virtuous perspective and perhaps only the real agent's information if you like or some contextually salient body of information. We then consider what the virtuous agent wants the less virtuous version of herself to do. This allows to avoid the conditional fallacy.

Few observations about this model:
1. There is no threat of collapse to the ideal observer view as you suggest. Virtue is really doing the work.
2. No advising skills are needed. We are not interested on advice but rather what the virtuous self wants the real self to do.
3. I don't quite get the last worry you have about the view. At least it seems that view with this structure can provide an extensionally adequate account. We might not expect this initially but this is up for investigation.

I should also say something more about the third point. I was thinking about the views about colour. The basic equation for red is supposed to be:
X is red iff X were put in broad daylight for normal observers to look then it would seem red to them.
With this view we get the conditional fallacy with the shy chameleon who is green when unobserved but would seem red from blushing were people to look at it. Wright's way to avoid this is to go provisional:
If normal observers look at X in broad daylight, then (X would be red iff X would look red for the normal observers).
This says nothing about the colour of X when it's unobserved. The idea is that the virtue ethicists could avoid your problem in similar ways. They could claim:
If X is an option for a fully virtuous agent, then (X would be right iff X would be the action the fully virtuous agent would do).
Now, this equation still draws an interesting connection between rightness and fully virtuous actions. But, the cost is that it says nothing about actions of less than fully virtuous agents. So, we need some other account of those actions. Here you would need to say something like for less than fully agents actions are right if they are either actions in accordance to virtue or actions that develop virtue. This needs some finetuning for conflict cases. But, even if this is disjunctive we still use virtue to do explanatory work in each disjunct.

Jussi I think I get the proposal. Maybe it's good enough. But the problem is the bit, "the real agent's information if you like or some contextually salient body of information". Is that *all* of the information? Or just some of it? Or how much? If all of it, then this seems to me to be full information (yes, there's no particle physics knowledge implied, but we probably don't care about it). And then we need to know why full virtue implies full information. I think it doesn't. Fully virtuous people make mistakes of fact. If they do or can, then those mistakes are a problem.
The 'wants/advises' difference isn't a difference. Presumably if the ideal wants the less ideal to do something, she'll advise her to do it (ok, I can manage to construct some counters, but ...). And if she advises her to do it, she wants her to do it as well (again, modulo some counters). What does virtue add, beyond full information and benevolence? This is my only question. I'm not even sure the fully virtuous know what *they* would do in a hypothetical situation, or why they would know this. "You had to be there."
Regarding your elaboration of the 3rd point. How would you respond to my worries about two standards views?

Hi Robert,

I'm not sure I'm unsatisfied with your answer and my question may have been too compressed. As you know I think the way to deal with conditional fallacy issues for internalism is to hold fixed the grounds of the reasons in question and then say that one has reason to do what the most rational/ideal agent would do, consistent with what we held fixed. In those cases where the grounds include features of the agent that make her irrational this means we're talking about an agent who is not completely rational but only as rational as can be given those defects.

I was wondering if the virtue theorist had a parallel move available, by saying the right action is the one that an agent who is as virtuous as possible (holding fixed those features of the situation that ground the reasons to act) would do. If this worked, in cases where the reasons depend on a departure from the ideal, we would be talking about an agent who is otherwise as virtuous as possible, apart from the departures.

But I'm not confident that this works for virtue theory used as a guide to right action, partly because it seems that we need an idea of the relevant grounds for acting to know what we would have to hold fixed. And that (it would seem) requires already knowing the reasons the person has for action. But then we need to answer the question we were using virtue theory to answer.

Since I think internalism is not offered as a guide to how to act, but rather put forth as a constraint on reasons given the explanatory and justificatory roles they must play (if you agree with Williams), it seems to me that the worry doesn't cause trouble there. We can exercise our independent grasp of what it makes sense to do in a situation, along with a grasp of the features that makes this so, to identify the reasons. Internalism just commits us to the claim that the most rational agent in a situation where those grounds obtain would act on those reasons.

So I think that means I'm inclined to agree with your answer here.

If RNJ can persuade MvR of something, RNJ is happy. (Unless someone else out there thinks you shouldn't have been persuaded.)

BTW, Mark, what about Valerie's idea? You're inventive on the reasons stuff. Is there something in that direction?

Hi Robert,

Which of Valerie's ideas are you asking me about? I don't see her posting on this thread which makes me think I should have read something elsewhere that I may not have.

Hi Robert

thanks again. I guess I don't see the force of either one of your objections. Just few responses:

1. I am assuming that there is some kind of connection between what is right and wrong and what agent's can be blamed and praised for. I also assume that we should allow the virtue ethicists to recognise this connection. For this reason, there is no point of stipulating that the virtuous advisor has full information or otherwise what is right and wrong would be totally removed from blameworthiness and praiseworthiness. We would need to blame agents for doing things that are completely accidental from their epistemic perspective.

How much information will the virtuous agent have? This is a good question. Here's one way to answer this: your own objection is based on moral intuitions about what is right and wrong in various situations. It looks like we can generate a version of virtue ethics that's extensionally equivalent with those intuitions: there will be a reflective equilibrium between your moral intuitions and how much knowledge we have to grant for the virtuous agent who wants the real agent to act in some way. I assume that this information will not be full information or exactly the real agent's information but rather what information would be available for the real agent on the basis of the evidence present in her context. This would keep the connection between right and wrong and blameworthiness and praiseworthiness.

2. Beneficence will definitely not be the only relevant virtue and I am not sure why you think it would be. I take it that beneficence is a maximizing virtue - it is beneficent to promote everyone's well-being, and so if the virtuous advisor only had this virtue she would want the real agent to be a utilitarian agent. But, a fully virtuous agent need not want the agent to act in that way in many cases. A fully virtuous agent who is kind, compassionate, just, fair, honest and so on can want the agent to respect moral constraints even when doing so is not maximally beneficent. All the other virtues will add whole range of morally salient considerations that play a part in explaining what the fully virtuous agent would want the real agent to do and thus what is right and wrong in concrete situations.

Thus, this view definitely does not collapse into full information plus beneficence ideal advisor view. There is a principled reason not to give full information for the advisor and beneficence is definitely not the only virtue doing work. And, extensionally this account seems to match your own moral intuitions. So, the question is: what's wrong with this?

Few other points:
The advising/desiring difference seems to be relevant by your own lights. Above, you said that advising requires a skill that we cannot assume virtuous agent has. Assume that's right. The same cannot be said about desiring.

With the other proposal, I cannot see how these concerns arise there either.

I am trying to work out if Jussi's first complaint above is best understood as the Euthyphro Dilemma.

Also, back in the day I took the sort of conditional fallacy worries that Robert highlighted as a reason to abandon internalism, at least if we are somewhat strict about what we mean by internalism and follow Williams who took internalism to be the view that:

"A has a reason to 0 only if he could reach the conclusion to 0 by a sound deliberative route from the motivations he already has. The externalist view is that this is not a necessary condition, and that it can be true of A that he has
a reason to 0 even though A has no motivation in his motivational set that
could, either directly or by some extension through sound deliberation,
lead him to 0."

With that understanding of internalism, conditional fallacy type worries strike me as dooming internalism but as leaving plenty of room for internalist-like views such as subjectivism that have one's reasons be responsive to one's motivational set in a vaguer way.

Hi David

I hope you are well. I feel bad about highjacking the thread, but here's one thought quickly. I wasn't thinking of the Euthyphro as such. I just have strong intuitions about right-makers: there are many of them and they are fairly basic considerations that explain why some actions are right. This seems to rule for me general theories like virtue ethics as theories about right-makers even if principles like (V) can be understood in many other ways. I also think that there is a connection between right-makers and what virtuous agents think about when they act such that (V) cannot be a view about right-makers or otherwise virtuous agents would be thinking about themselves awful lot.

You are right that this could be developed to a Euthyphro concern. The question is: are some actions right because virtuous agents would do them or do virtuous agents do these actions because they are right? Neither one of these seem appealing to me. The first horn faces the previous problem and second seems to make acting rightly a fetish. But, there are many ways to understand virtue ethics and principles like (V) that don't face this problem.

Hi Jussi,

I am not seeing how virtue can play the role you are suggesting (unless I am missing how this goes). If we take the Smith route in developing things, then we say that rightness is determined by what one's better self would want one to do.

But why think the better self needs to be virtuous rather than just concerned with virtue? One doesn't need to be courageous, for example, to want oneself to act courageously. Virtue seems to be a spare wheel here...

And then there are worries about virtue, rather than concern for virtue, getting the extension wrong. For example some think that humility involves ignorance and that virtue conflicts with epistemic impartiality. Take Julia Driver's view...if she is right the perfectly humble version of myself might be blind to my virtues, such as they are, and this might make us worry about whether my perfectly humble self will want me to do what I should. There might be admirable things that I could and should do but that my humble better angle will not want me to do because he will not recognize that these things are in my power so to speak.

Hi Brad

I agree that wanting oneself to act courageously doesn't require that one is courageous. However, if the advisor was just me, as I am, wanting myself to act virtuously, oftentimes I would fail to want myself to do things that are really at the end of the day courageous. I am assuming that the virtue of courage is a set of certain finely tuned context sensitive beliefs, desires and emotional reactions. Only if the better self has these sensitivities will she want the real self to do the actions which are intuitively right. So, I am thinking that the need for virtue itself is required for due to something like particularism and the context sensitive nature of right and wrong where mere desires to act courageously from the behalf of the non-virtuous person will not be enough to track the salient features of reality. For that, genuine virtue is required.

I need to think about the last point. I'm perhaps not immediately convinced by Julia's view about humility. Also, it seems like the advisor structure avoids the worry. Humility might be ignorance about one's own virtue but I am not humble if I am ignorant of your virtue. However, it's not clear to me that in the theory I've sketched the advisor thinks that the person whose actions she is considering is herself.

I should emphasise that I am not a virtue ethicist myself. I'm just interested in the conditional fallacy and whether virtue ethics can be formulated in a way that doesn't commit it.

Jussi, I think your view is coming more into focus for me. I see where you want to go, and maybe that's the way to go. So I'll just say a few more things.
Regarding 1. I'm not sure what connection the virtue ethicist would like between right and wrong and praise and blame. It wouldn't be Millian, since that would give away the game, wouldn't it? In any case, either its full information or not that your idealized helper has. If the information is compromised, I don't see how that could give us a set of right actions. If it is all of the relevant information, then we're back to ideal observers. (I'm not sure how reflective equilibrium helps that issue.) Plus whatever sentiments we need to add in there so that they're wanting what is good for ordinary agent to do. I don't care what that is. My point is that it ain't full virtue that will give us the set of actions we want to call 'right' for the less than virtuous.
Regarding 2. I didn't suggest the virtue of beneficence was required. It's the sentiment of benevolence (again, that's Firth, arguably Hume, et. al.) Some sentiment that insures the idealized version of me or whatever *cares* about me and my situation. You're probably right that this isn't the only sentiment that would be required. After all, many of such views were meant as accounts of a person's good, not as accounts of what is right. The issue for me is, Why is complete virtue required for that idealization? Unless it is and not just in a trivial way, it collapses into the idealized view.
Regarding advising and wanting. The difference: information as far as I can tell.

HI David,
I think the issues around the conditional fallacy and internalism are different from the worries I have about V. There's an important normative point in the latter: Don't be trying to do what Jesus would do. Probably we should do what Jesus (or whatever version of deity you like) would advise us to do...after all, he's omniscient and omnibenevolent. I think Julia Markowitz has nicely pointed out the connection of the former to the explanatory desiderata, and the challenges of hanging on to that. But that's not what's at stake for V.
I think you're right. V has to be read right to left to be a genuine version of virtue ethics. If praiseworthiness and blameworthiness get in there, well, why not drop out the 'completely virtuous' altogether?

I like the perfectly humble self point. But maybe that shows that humility aint a virtue.

Thanks for the discussion, especially to Robert for discussing my piece from JESP. It's nice to know that someone other than my p&t committee read that piece (and I'm not too sure about them!).

I'm still awash in grading, but I wanted to make a quick reply, especially to the objection that (P) is ad hoc. I was inspired by Christine Swanton's "Profiles of the Virtues" piece there. What unifies the unholy trinity of promotion, respect, and expression is that they're all ways of responding to value, so (P) can be motivated independently of responding to Robert's incisive criticism.

I was thinking of expression along these lines: Kant bids us not only to respect or honor duty by not acting contrary to it (where to respect value V is not to act un-V-ly (I think I first encountered this idea in Slote, Pettit, and Baron's Three Methods of Ethics) but also to express the value of duty (or our commitment to it) by acting from it. Aristotle makes a similar distinction in NE VI.13 between natural virtue, where actions are in accordance with right reason (kata ton orthon logon), and virtue proper, which implies the presence of right reason (meta tou orthou logou).

That's probably not satisfactory, but it's the gist of the response I'd want to make if there weren't a stack of essays calling my name.

I wonder if one thing to be said about this discussion is that by abstracting virtues and situations into little algebraic figures, it has forgotten what virtues and situations are actually like.

"Suppose A lacks benevolence. It would be right for him to cultivate it. But that is not what a benevolent person would do—they’re already all full up with benevolence."

This is not obviously true. It seems to me that one very reasonable view of virtues is that we work on them continuously - there is no such thing as being so benevolent that you don't have to work at being benevolent any more, and even if there was, it is in the nature of the benevolent person that they would always want to be more benevolent (i.e. you'd still do the learning even if you actually couldn't get any more virtuous). Wise men who say "I'm wise enough, no need to learn any more" are generally regarded as unwise; courageous warriors who decline to test their own courage are not really brave; etc.

"The completely virtuous wouldn't waste their time on self-improvement, even though the rest of us might be doing the right thing by doing so."

There are two c-related problems here. The first is that the original formulation refers to a specific c. Your friend needs your help, but you have work to do. A fully generous person would help the friend; but it's not the case that an imperfectly generous person would spend their time learning generosity instead. That's not an option on the table. So in a specific context, the learning problem might never arise. (Moreover, for the most part we tend to think that you learn to be virtuous by practising the virtues, so in fact the learner's choice in any specific situation might well be the same as the virtuous person's.)

(The other problem with c is that c may include the subject, so a fully virtuous person and a learner would never find themselves in the same c, because they are always different. This problem threatens the original V formulation, so c may have to be defined to exclude the subject.)

The comments to this entry are closed.

Ethics at PEA Soup

PPE at PEA Soup

Like PEA Soup

Search PEA Soup


  • Unless otherwise indicated, the views expressed in any given post reflect the opinion of only that individual who posted the particular entry or comment.