Blog Coordinator

« Moral Responsibility and the Real Self | Main | The Allure of Initial Design Arguments »



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

Hi John, thanks for another set of great questions for us to consider! Here's the beginning of a response to your worry.

When considering cases like yours it does sound strange to think that we are reflecting on and deliberating about our first-order desires, rather than on our alternatives for action (i.e., our choices). However, I've always thought that the Frankfurt-style models are most apt for the sort of reflective thinking we do when we are planning (Bratman helps here) or even thinking more generally about what sort of person we want to be.

We consider the sorts of desires we feel (or more behavioristically, our memory of the way we've been moved to behave in various situations), and we deliberate about which of them we want to move us in the future (= which represent our deep self?). Do I want to act on my desires to drink more (perhaps have more fun) when I'm in drinking situations or do I want to act on my desire to drink less because of my desire not to be hung over (or be able to work harder or act less stupidly, etc.)?

I'm free and responsible roughly to the extent that I possess (in general), and I am able to exercise (in specific actions), my capacities to carry out such reflective deliberation and to control my behavior in light of it.

I am not sure how different these capacities are from the ones involved in a reasons-responsive model (I'd love to combine the best of all possible worlds/models).

I'll also mention briefly a question I meant to ask about in the previous discussion but applies here too. How come no one is explicitly mentioning consciousness? As one who thinks consciousness (and the causal relevance of conscious reasoning) is essential for free will, it seems like it has to play a role in any deep (or real or hierarchical) model. And I think it is required for reasons-responsiveness too.

Very good point, John. I think you're right that this serious problem is seriously undervalued. Gary puts this problem in terms of the targets of identification: are they desires (putative wills) or courses of action? He plumps (mostly) for the latter, and this seems right for his purposes (to describe an undistorted picture of practical judgment).

But there are other issues at stake in this arena, and here's where the ambiguity between two senses of identification may be relevant. On the one hand, identification may be about some kind of active process, i.e., what is it that I *make mine* by having it be the object of some attitude? On the other hand, identification may be about pure attributability, i.e., what is it that makes some action or attitude mine for purposes of responsibility, where this process may occur entirely passively (i.e., just be about the structural relation between actions and, say, cares)? If we're talking about the former (which is what, frankly, I think Frankfurt initially had in mind), then the more plausible picture of deliberation and practical judgment by far is Watson's. If the latter, though, then the question of what makes some *will* mine is terribly important, and it may make no reference to active processes at all.

I fear this may be a bit obscure (or "vague," as the kids say). This is pretty embryonic. I haven't thought about this criticism of Frankfurt in a while, and I'm glad John has brought it back up. I have no idea why it faded away.

Short term yeah focus on the martini itself. Long term, you had better think about that possibly inordinate desire.

The problem shows up fairly quickly when you try to explain Frankfurt's theory to the uninitiated. To make it compelling, you have to appeal to cases of addiction and repressed sexual desire. These aren't central to normal moral decision making where we are considering courses of action, not our desires. Or you have to give an example of a crisis of desire, lots of clove cigarettes and all that. . . . . The worry is that these kinds of desire-directed desires are rare, and not just for unreflective types. Sure, my cat probably can't have second order desires, but it seems that few people have many at all. Perhaps most of us are wantons. That might be right. But it would have troubling implications for moral responsibility. . .

Sure, it might be better to have gone through a stage of evaluating one's desire, and perhaps progressed to some kind of "higher-order immediacy," or what have you. But that doesn't seem necessary for moral responsibility. I also doubt that it's necessary for authenticity.

Interesting issue John. The problem for Frankfurt seems even worse than this because he proposes that there are desires of still higher-order (e.g., third- or fourth-order desires directed at desires of lower order). These meta-meta-attitudes are even stranger beasts that rarely occur in practical reasoning.

Identification with one's motives seems to be a ubiquitous phenomenon. But the attitudes that are invoked as being constitutive of it (higher-order self-directed attitudes) occur exceedingly rarely in practical reasoning. There does indeed seem to be a disconnect.

'The worry is that these kinds of desire-directed desires are rare, and not just for unreflective types.'

No, Aaron, I think it's not at all uncommon to worry about one's desires, especially recurring ones: one's thematic concerns, as it were. Don't people ask themselves all the time, referring to their peristent desires- Is that really who I am? You will concede that lots of folk spend lots of time trying to break what they consider bad habits. What's going on their if not the culmination of an unfavorable reflection upon a certain desire or cluster of motives?


I think (best guess) that you're write that Frankfurt's view is wrong as a description of practical reasoning. By and large, practical reasoning (deliberation?) involves contemplating courses of action - not first order desires.

At the same time, I think that Frankfurt was definitely onto something about moral responsibility. It seems to me that desire-integration is probably a necessary condition for moral responsibility.

I'm not an expert enough about Frankfurt to know whether his account is suppose to describe moral responsibility, practical reasoning, or both. But I think your point about practical reasoning is well taken.

I should add: I think Frankfurt's account has some truth to it as a description of moral responsibility. I think this is especially true if we construe "desires" very broadly. Frankfurt's, and Watson's (etc.) larger point is that our entire evaluative framework should have some degree of consistent integration. This seems to be a necessary condition for responsibility. Whether it is a sufficient condition is another question...

Another great question, John.

I wonder what folks think of Bratman's defense of hierarchical views.

"A hierarchical theory in the spirit of Frankfurt's could could agree that ordinary deliberation is normally first-order deliberation about what to do. It could then say that higher-order reflection on first-order motivating attitudes helps, as it were, to set the stage for such ordinary deliberation: it helps provide the framework within which that deliberation takes place." (Bratman 2007, 229)

Bratman's own view is that self-governing policies are the relevant objects of higher-order deliberation and that they set the stage for first-order deliberation because their content is of the form 'treat desire D as reason-giving in motivationally effective practical reasoning'. Self-governing policies are a type of intention, so they seem to be the sort of thing one actively adopts (as opposed to the sort of passively acquired cares that David Shoemaker has been alluding to).

Bratman's strategy seems to me to be the most plausible way to fix up a hierarchical theory. Does it satisfactorily answer the challenge?

Nice, Ben. I agree that Bratman has been one of the few to wrestle with this problem, and it is to his credit--I admire Michael very much, and not just because he was such a great teacher at Stanford. But I am not sure it really does the trick. What does it mean to say that higher order reflection "sets the stage" for ordinary deliberation? Maybe this is a significantly REVISED version of Harry Frankfurt's view. But on Frankfurt's view, as at least I understand it, *every time* I deliberate I'm really deliberating about which first-order desire of mine to act on. Perhaps Bratman is suggesting that we move away from Frankfurt's actual view and adopt something like a "two-tiered" view in ethics, or the analogue of that for action. But this isn't Frankfurt's actual view, as I see it. Maybe though it is in the spirit of Frankfurt's view, and an improvement upon it.

Bruce: thanks for your great post. It is late here in Germany, and I hope to get to it tomorrow (Thurs). Steubenville (and Weirton, WVA) are special places for me--and almost as super-cool as Youngstown, no?

"Don't people ask themselves all the time, referring to their peristent desires- Is that really who I am?"

No Robert, I don't think so. This sounds like some kind of pronounced mental illness, a strange form of disassociation. In fact, the general model strikes me as slightly insane, as if one were playing a passive role, betting on my first-order desires. But I'm not in the stands with a ticket. I'm on the horse, in the race. Or maybe, I'm one of the horses. I don't know which. I don't have hopes about what desires will motivate me to act. Rather, it seems that I act. Sometimes I act on appetites, other times I seem to be motivated by my evaluative beliefs. . . . .

"You will concede that lots of folk spend lots of time trying to break what they consider bad habits. What's going on their if not the culmination of an unfavorable reflection upon a certain desire or cluster of motives?"

I'm not so sure. I don't think that the common object is one's desires in these cases. Take smoking. I quit 6 years ago. My thought wasn't that I wish I didn't have a desire to smoke. It was that I don't want to smoke. I thought it was bad for a variety of reasons. My desire for cigarettes was secondary. . . . The siren call of lady nicotine didn't make it easy to quit. I wished that it would stop tugging. Then and there I had a second-order desire. I managed it with a lozenge. There was no question about whether the desire for nicotine was really who I am. I'm not a desire for nicotine. In any case, the process of quitting didn't start with any such desire about a desire.


How does one stop smoking without eradicating the very urge to smoke? Isn't that what aversion therapy is all about? Is quitting smoking enough? I would hate to walk around having to fight that urge, even if I was largely successful in not lighting up. I might further think that the urge itself is not becoming a man.

Hi John-

Is the issue whether Frankfurt's particular version of the hierarchical theory can model deliberation, or whether some version (say, the best version!) can model deliberation?

I thought you meant to ask the latter question, but your reply to Ben suggest you meant the former.

For what it is worth, I think the Bratman model is supposed to be one on which ordinarily you decide what to do, and the "stage-setting" point is just that some intentions (policies about deliberation) say treat some things as reasons in deliberation in not others. So, I might deliberate about what to do as I'm drinking coffee in the morning, but my policies about, say, not smoking or drinking before noon "set the stage" by telling me to disregard my then-occuring impulses to smoke and booze. The big raft of policies that operate in my psychology do a lot of work constraining my deliberation, even as I am deliberating about what to do. So, no problem here for the hierarchical theorist.

(For the record, my policy is to never smoke—unless it is a Cohiba, or I've been manipulated or am regressing.)

Great post, John!

Aaron (if I may), if Robert asked (1)"is that who I really want to be?" instead of (2) "is that who I really am?" when thinking about desires and past actions, would you still think his proposal is "insane"?

Frankfurt acknowledged that many adults were wantons, to a degree. In most of our daily actions we might be wantons, but, when we make important life decisions (i.e go to grad school, change careers, have children, become philosophers) we often consider (1), even if only indirectly, we do seem to consider it though. If that is true, then it seems plausible to think that the actions that aim to sustain those important decisions are in some way identifiable to us, at least in weakly. (Consider studying to stay in grad school or going on interviews to establish that career change you decided on as identifiable to us if they related to the example important life decisions I mentioned).

This might happen more often than only those important cases as well. For instance, tomorrow, if I wake up overly-tired, I might decide not to stay up reading blogs until 1:30am the next time the first-order desire to chime in arises. I may not want to be the kind of guy who is tired at work the next day since I experienced what that person was like and identified better with the guy who isn't tired at work (I’ve had both experiences to compare it to as an adult who has many life experiences). Again, I don’t think we think like this, actively. But, it seems plausible that our decisions are colored by such considerations, indirectly at least. This might be all you need for some variation of the identification to get off the ground.

Your proposal that we “act” seems much closer to metal illness than Robert’s contention that we often reflect. Most who have mental illness will act without considering their evaluative judgments, not vice-versa. Sure, there are extreme forms of reflective deliberation that constitute mental illness, but, it’s a stretch to paint Frankfurt’s view with that brush.


Thanks. And also, sorry: your thoughtful comment points to a general issue about which I have not been sufficiently careful in this blogging context: sometimes I seem to be raising a question about Frankfurt's views in general, or even hierarchical or identification theories in general, but then it might well turn out that my point only applies to early Frankfurt, or perhaps Frankfurt (as opposed to other identification or hierarchical theorists). Good point, and it is important to be more careful or vigilant than I have been here.

Yes, one could combine some of the resources of the hierarchical approach with the idea that in ordinary contexts of practical reasoning one considers courses of action, rather than one's own mental states. Perhaps on this sort of approach, there are some times in one's life when one "steps back" (to use that metaphor) and reflects at a second- or even higher-level on one's first-order desires, but this need not happen in every deliberative context. (Again: think two-tiered normative ethical theories.) Perhaps this more nuanced (or complex) kind of strategy can work--I'll think about it. I just didn't see that Frankfurt himself could adopt it, but maybe so, and thanks again for the suggestion.

If I may add: presumably the reason this is a particular problem for Frankfurt is that his account of acting freely (and moral responsibility) has it that an agent acts freely only if she identifies with the first-order desire on which she acts, i.e., on the early, unvarnished account, only if she forms a second-order volition to act in accordance with her "will" (the first-order desire that does in fact move her all the way to act, or would under certain [unspecified] conditions). Of course, Frankfurt allows that these second-order volitions can, like first-order desires, be unconscious or not part of explicit conscious processing. But still: he must think of these second-order desires as present, and presumably they must be similarly present in deliberation.

Of course, it is always possible to make a simple, elegant, but problematic theory immune to counterexamples by making it more nuanced, as, again, with two-tired or even fancier normative ethical theories. But one is in danger of throwing the baby out with the bathwater--losing the intuitive point with which one started.


Well, I'll take advantage of this natural segue (is that how you spell it?). Right: there's gotta be a connection between moral responsibility and consciousness. I think, although I'm not sure how to argue for this, that a being cannot be morally responsible without being conscious; no being that lacks any consciousness throughout its existence can be morally responsible. But of course this leaves lots of room for being morally responsible in a particular context for behavior which is the result of spontaneous, unplanned, non-reflective processes, and even for behavior that is the result of unconscious processing--as in the work of Nomy Arpaly.

If reasons-responsive mechanisms can be unconscious, then it is perhaps an interesting question why we need consciousness at all to be morally responsible. I am not sure how to answer that. But I admit that it is hard to imagine a creature or entity that is never conscious and yet morally responsible. I guess I agree with you about the necessity of consciousness for moral responsibility, but the relationship can be quite loose, I think, and I'm not sure what the explanation for the necessity is.

To continue. Perhaps the answer to the question about why consciousness is necessary for moral responsibility is that it is hard to imagine the whole "conversation" of moral responsibility--as envisaged by Strawson, Watson, and recently McKenna--with the reactive attitudes playing a crucial role without consciousness. So if moral responsibility is properly modelled along the lines of a conversation, then the participants, it would seem, would have to be conscious; otherwise, the distinctive communicative aspects of moral responsibility would be impossible.

John asks: "Why is so much attention paid to the regress problem and manipulation problems [for hierarchical models of agency, particularly Frankfurt's], and so little attention to weakness of the will and the apparently distorted picture of practical reasoning?" I think part of the answer lies in the piecemeal way in which we tend to do philosophy (owing to space constraints, for example). Here's an example. In my first book, *Irrationality* (most of which is on weakness of will and self-control), I argued that appeals to higher-order desires won't generate a perfectly general solution to various paradoxes about self-control. There's a section of my *Autonomous Agents* entitled "Higher-Order Desires: Are They Essential to Continent and Incontinent Behavior" in which I argue that the answer is no. About half of that book is on weakness of will and self-control (or "continent" and "incontinent" action, as we sometimes said in the old days). In *Motivation and Agency*, I defend a view of practical reasoning or deliberation. One of its planks is that practical reasoning typically is about what to do; so I'm in agreement there with what John says in his post. But when it comes to Frankfurt on free action or moral responsibility, rather than attacking various aspects of of his hierarchical model of agency, I go directly after what he says about these topics in particular. And in that connection, certain manipulation cases seem to me (but definitely not everyone) to be extremely effective as counterexamples. Of course, if you put all these pieces together, you have a more general critique of Frankfurt's model of agency. By the way, my latest book, *Backsliding* is about weakness of will. (I just can't seem to shake that topic.) There I don't say anything about hierarchical models of agency, as I recall. I guess I had already persuaded myself that I had a better model and didn't need to rehash old critiques.

I said "part of the answer" above. I think that another part of the answer is that free will and moral responsibility get significantly more attention than models of agency in general and weakness of will in particular. So putting the two parts together, one expects to see significantly more about manipulation cases than about weakness of will and agency in general in critiques of Frankfurt's work.


Thanks for your nice and thoughtful comment. Although I think Frankfurt was primarily interested in giving an account of moral responsibility, it would seem also to imply a picture of practical reasoning. I also agree with you that, without refinements and qualifications, the picture of practical reasoning is somewhat implausible.

How much desire integration is necessary for moral responsibliity? I agree that if someone is radically inconsistent in this regard, he or she may not even be an agent at all. But I also note that even on Frankfurt's and Watson's (different but related) pictures, agents can have very significant conflicts in desires--especially but not exclusively at the same level.

It is distinctive of Frankfurt's view that there is an impetus toward "wholeheartedness". This has always struck me as intriguing. Perhaps it is a "regulative ideal" toward which we do or should strive--wholeheartedness. On the other hand, if values are complex and messy, and it is desirable that our minds reflect the facts--including value facts--in the world, then perhaps our normative orientations should also be messy. Maybe it is too fussy--and implies a kind of Procrusteanism--to strive for wholeheartedness.


Thanks for your extremely helpful comment! I'll think more about the "piecemeal" idea, which seems to me to be right. Also, allow me to emphasize that I was contending that there hasn't been much focus on Frankfurt on weakness of will, not weakness of the will per se. And, in my view, no one has done as much good work on weakness of the will than you! (And I don't mean the instantiation, but, of course, the analysis!!)

As I walked by my bookshelf just now, the title of the 2006 volume containing Frankfurt's Tanner Lectures, "Taking Ourselves Seriously & Getting It Right," caught my eye. Then I opened the book and read the first line:

"I suppose some of you must have noticed that human beings have a tendency to be heavily preoccupied with thinking about themselves."

On the matter of consciousness and reponsibility: this summer I was privileged to offer comments on a book-length manuscript by Neil Levy that argues that consciousness is a necessary condition for moral reponsibility. I did some quick research and couldn't establish that it's out. He does have a podcast on Philosophy Bites on the same material. I don't wish to steal his thunder but I thought his book was terrific, and his account of access consciousness especially illuminating for this issue.

I have recently been reminded of an excellent and important 1990 article by Philip Pettitt and Michael Smith, "Backgrounding Desire", in which they thoughtfully discuss the problem first noted by Irving Thalberg and also Gary Watson--the problem of the apparent "solipsism" of Frankfurt's conception of practical reasoning: Philosophical Review, Vol 99, 1990, pp. 565-92.

I have benefitted from Pettit's very elegant Muester Lecture and follow-up discussions. And I highly recommend the Philosophical Review paper--as well as the Pettit/Smith Journal of Philosophy paper also in the 90's, "Freedom in Desire and Belief".

The comments to this entry are closed.

Books about Agency

3QD Prize 2014: Marcus Arvan