Blog Coordinator

« Featured Author: John Martin Fischer | Main | Moral Responsibility and the Real Self »



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

Indeed, there are general reasons to think that we are likely to be highly unreliable at distinguishing the real self from the ideal self. The notion that we can introspect the real self seems a hold over from some Cartesian notion of transparency. In his recent book *The Opacity of Mind*, Peter Carruthers presents a range of evidence for the view that our access to our own attitudes is interpretative, and it seems likely that it will be shaped by motivational factors concerning what we hope to find there. But that doesn't imply that there can't be cases in which the evidence between real and ideal self isn't overwhelming. Just that cases in which we are mistaken will be common, and there will be no marker for distinguishing the mistaken cases from the genuine ones.

Hey John,

Thanks for the post. One quick question: can you give a case where the two come apart that you think the identification theorist can't handle? Most cases that spring to my mind of agents who want to be different than they are will be cases in which the agents are less than wholehearted about their current motivational structure, and so the identification theorist likely wouldn't count their current motivational structure unequivocally as their "real self".

(Remember, too, that the identification theorist is free to distinguish between the way an agent's will is in fact organized, on the one hand, and the way an agent thinks that his will is organized, on the other. So, an agent can be mistaken about what he's really like.)

Perhaps the “ego ideal” may be modeled as a component of the “real self”.

If the ego ideal is part of what comprises the real self, then maybe it makes sense to believe the “ego ideal” is therefore partially responsible for actions of the agent. We can attempt to subdivide the self into separate entities that cause the self to emerge, but it will probably be difficult to define those sub-entities in a manner that allows intelligent discussion.

Hi John,

This is a juicy issue you raise. Thanks for posting this.

I think we need to distinguish two interpretations of identification. There is a psychological sense that refers to how an agent (consciously) regards her own motives-- Does she endorse them or does she repudiate them? And there is a 'metaphysical' sense in which identification is a relation that holds when a person's underlying 'deep self' is expressed in his actions *irrespective of whether or not he is aware that his actions do in fact concord with his self*. Jaworska and others make this distinction, and she uses the term 'internality' for the latter, metaphysical sense. A man raised with a hyperconservative upbringing may not identify with his sexual desires (in the psychological sense), but these motives may be internal to him nonetheless (in the metaphysical sense).

It is possible to be a Real Self theorist for moral responsibility and adopt either of these two senses of identification, and unfortunately sometimes theorists are less than clear about which sense they intend.

The bottom line is that I think you are right to challenge the Real Self view where identification is construed psychologically, and Neil, Neal and James provide extra reasons for skepticism of this sort of view. But that doesn't need to spell trouble for the Real Self view of MR when identification is construed in the metaphysical internality sense. My own favored deep self view of MR is in fact spelled out along metaphysical, internality-based lines.

Hi John,

Great question. (And I look forward to the next post. I think there are real problems for a theorist who wants to give an account of moral responsibility (as opposed to autonomy) in terms of actions issuing from the "real self.")

Here is one way that one may distinguish between the "real self" and one's "ego ideal": rational norms (e.g., consistency and means-end coherence). Depending on how one specifies identification for the purposes of picking out the real self, it may be the case that one's real self is irrational (violates norms of consistency or m-e coherence) but that one's ego ideal is rational (satisfies these norms). One's real self may be irrational if the proposed account of identification does not have rational norms built into it. And there is a clear sense in which rationality is more ideal than irrationality, so modifications to one's (irrational) real self in order to satisfy rational norms would be an idealization.

For example, suppose that Jane both identifies with her desire to eat a lot of cake at the party and identifies with her desire to avoid sweets in general. On a view where identification is volitional, such as Frankfurt's, one need not build in rational norms. There is room for inconsistent elements in the set of attitudes with which one is identified. One's will may be in conflict with itself, say, because one endorses desires that, in a given case, such as Jane's party, both apply but cannot both be satisfied. And the identification at issue need not be half-hearted. One may be satisfied both with the higher-order desire to want to eat a lot of cake at the party and with the higher-order desire to want to avoid sweets in general. There is something irrational about being satisfied with both of these second-order desires at once. But given that one is satisfied with them, one is identified with the relevant first-order desires and fails to satisfy the rational norm of consistency. Ideally, one would satisfy this norm, so there is room to claim that one's "ego ideal" would reflect a resolution of the conflict that one's "true self" contains. So, based on the rational norm of consistency, there is a way of distinguishing between Jane's real self and her ego ideal.

Guten tag, Professor Fischer. And speaking of Deutschland, Nietzsche actually asked a similar question: How does one become what one is? Here is a tentative answer, given in my article “Free Will and Indeterminism: Robert Kane’s Libertarianism” (JPR, 1998):

“I suggest that an agent facing a volitional crisis should not be viewed as needing to conduct a “value experiment”, a la Kane. She should be seen, instead, as requiring a deeper understanding of herself, the solution to her crisis being implicit in desires, beliefs and values her eventual appreciation of which will make one course of action appear more attractive than the alternatives. As Charles Taylor has proposed, she must strive to “articulate” preferences her dim awareness of which is responsible for the impasse she must overcome to become who she is. Her efforts here should be taken as an exercise in “depth psychology,” her conflict being resolved just in case she at least provisionally identifies with the insight it yields, having discovered a reason for thinking that she is the sort of person who prefers preferring one of her options to the other. The defeasibility of her identification allows for the measure of uncertainty that Kane believes is required to develop a free will without leaving her decision ultimately arbitrary.”

In short, I see no reason not to believe that one’s true self is simply inborn, a part of one’s nature. I myself have an Anselmian understanding of what it is, viz., the willingness to preserves one’s rectitude for its own sake, but that notion may be dispensed with here. The main point is that certain values are innate and make up one’s true self, free choices a la Frankfurt being deliberate attempts to conform to these ideals, rather than accretions thereto comprising an "ego ideal."

I hope that you and Professor Mele can find time during the conference to have ein gutes Bier.

Thanks to everyone for these helpful comments. And Robert, I certainly hope (and expect) that Al Mele won't be the only one of us drinking good beer at the conference...

Neal: Couldn't a person be wholehearted in the sense of Frankfurt, or even "satisfied" in the sense of Frankfurt (although that might be more difficult), even though he is identifying with his desire (say) to refrain from having sex prior to marriage because he has been raised in a conservative family. Here the agent wants to be a certain way, because that's the way he's been raised, and the mental hierarchy reflects this desired way of being; but in come cases arguably this doesn't reflect the "real" or "true" agent. What the agent really or truly wants is to have the sexual relationship prior to marriage. There can presumably be other, similar kinds of desires--desires for homosexual sex, for instance, where the "real self" may line up with these desires, but the agent does not identify with them, and there is no lack of wholeheartedness (although there is a conflict at the first-order). Does this make sense? Also, perhaps one really doesn't believe in God, but doesn't want to be an atheist, etc. It seems that these kinds of conflicts could be prevalent, although I agree with you that the identification theorist who also accepts some further condition such as wholeheartedness or satisfaction may have ways of addressing these issues.

Chandra: thanks very much. The "metaphysical" notion of identification intrigues me, and thanks for reminding me of my colleague, Agnieszka Jaworska's discussion, which I should look at again. But can you say any more about it? It seems a far cry from "identification" as say Harry Frankfurt conceives it, which is specified in terms of higher-order volitions. I agree that an identification theorist should move away from a purely subjective account, but can you say more about a "metaphysical" approach to identification?

A fascinating question with very interesting responses; but could you say something about how that question relates to your own narrative approach, which seems to me a very insightful way of getting at identity questions (and thus at what Watson calls the attributability aspect of responsibility). Certainly my self narrative is an important element of who I am; but the narrative is not perfectly accurate (for example, it is likely to neglect what T. S. Eliot calls "motives late revealed" in favor of more favorable accounts of my motives which I now mistakenly take to be accurate). In short, how does the question you pose relate to your narrative model? Thanks, best to Al, enjoy Oktoberfest.

Hi John,

You suggested I say more about internality construed as a metaphysical relation, distinguished from the psychological sense of identification in which a person is said to identify with a particular motive. So here goes.

I follow Jaworska, Shoemaker and others in favoring a ‘care-based’ approach to understanding internality. The basic idea is that a person’s cares enjoy a privileged status with respect to their internality. A person can be alienated from her desires or urges, but she cannot truly be estranged from her cares. Her cares *constitute* her evaluative stance. They are her motivational/evaluative essence, and this is something from which a person, necessarily, cannot be alienated. What is it about cares that imbue them with this special status? People appeal to the ways cares support temporally-extended agency, link to emotions, connect with evaluative judgments, etc... I have my own views as well. This is controversial territory and I won’t try to settle this issue.

The important point is that the notion of internality being put forward by the care-based theorist comes apart from the psychological sense of identification. For example, a person can psychologically identify with a motive. But regardless of how the person subjectively feels about the motive, it is still an open question whether the motive reflects the underlying cares that she does *in fact* have.

"Here the agent wants to be (chaste), because ... he's been raised (in a conservative family), and the mental hierarchy reflects this desired way of being; but ... arguably this doesn't (necessarily) reflect the "real" or "true" agent. What the agent (perhaps) really or truly wants is to have (a) sexual relationship prior to marriage."

If the true self isn't something objective, then who's to say what it is? In what sense could it be discovered? If it is just part of some story I'm fond of reciting to myself (and those I can buttonhole), then I seem to be making it up as I go along. It depends on my audience, mood or emotional needs. A young man's true self could start out being chaste but turn Don Juanish after he begins drinking beer with the older boys at school? Aren't we operating here with a flimsy notion of truth? (The whole thing reminds me of Wittgenstein's private language argument.) Wouldn't this self be better termed 'explanatory'? I use it to get myself out of tight psychological jams or as a somnolent.

Great post John.

Am I right in thinking that identification requires a tighter subjective link to the relevant contents of mental states that ground responsibility than ownership does? What I'm thinking here is that whereas something like a subjective self-reflective endorsement of a second-order desire is required for identification, ownership is a function of a criterion like reason-responsiveness assessable in a more objective way. So whereas by identification criteria one must (in some sense) know or be aware of the fact that one embraces the desire to do what one does, by reasons-responsiveness one must have an intention that is stably counterfactually adjustable to reasons, whether one is self-aware of such adjustability or not. If that's right, then identification criteria require a deeper sense of responsibility (and consequently a deeper probing on the part of others to determine that criteria is satisfied) than does ownership criteria (which in many cases might be more prima facie indicative of reasons-responsive intentions as reflected in the nature of actions).

As well, deep identifications with desires for actions might be perverse themselves (e.g., John Hinckley seems really to have endorsed his desire to shoot a president to impress Jodie Foster). But those same intentions (to shoot a president to impress Foster) might be assessed as not truly reasons-responsive prima facie and much more simply. Reasons-responsive ownership would not require a deeper examination of the question of whether Hinckley properly identified with his desire to shoot the president, which, as I read what I just wrote, sounds pretty damn odd anyway.

If my grasp of any of this is defective, I hope I am reasons-responsive to replies at any rate.

Chandra: Thanks! This is very helpful. I think that a caring-based approach to specifying identification is very attractive and it seems promising in addressing the kind of worry that prompted my post. I like the way it distances itself from an approach that relies solely on an agent's subjective judgments and "higher-order volitions". Perhaps a caring-based approach has the resources to give a more plausible account of identification. I still wonder what the relationship is (or should be) between the real or true self (as given by one's identifications) and moral responsibility--more on that soon.

Robert: although caring is perhaps not "objective" in quite the way you have in mind, it does seem to go in the right direction, in my view.

Bruce: Thanks very much. Well, I still haven't completely thought through the narrativity conception of identity. I would emphasize that my view of the role of narrativity in identity is different from the views of many others who invoke narrativity--my view is perhaps more "minimalistic". I take it that to be morally responsible, one's life has to have a story or narrative, strictly speaking (and not merely a chronicle). Actually, I claim that it is in virtue of acting freely (exhibiting guidance control, in my view), that an agent makes it the case that his life has a story (strictly speaking). The agent thus also thereby makes it the case that his or her life has an irreducibly narrative dimension of meaning or evaluation. Those are the main roles of narrativity. The value of being morally responsible then is the value of being an author of the book of one's life.

I do not however make the further claim that the narrative itself and straightforwardly gives us the "real or true self". Further, my narrative view is silent on the relationship between moral responsibility and expressing the real or true self.

Here's one reason for preferring a narrative conception of the self over a care-based approach: the care-based approach (if I understand it; it's quite possible I don't) seems insensitive to relational facts. Or rather, insofar as it seeks a deep self it is insensitive to these relational facts: by locating the self in the interior, it ignores that the content of cares is deeply constituted by our potentially rapidly shifting external relations (alternatively, it could accept that and give up on the claim that the self is deep). On the narrative account, or at least on *a* narrative account, these swings may be papered over by the stories we tell ourselves about ourselves, which will be partially confabulatiory but which will also have some causal powers to bring it about that the narrated self matches the narrative.

Sorry to come a bit late to this party, John. Very interesting first topic, one that I, uh, care about a great deal. I'm obviously in favor of something like Chandra's approach, and I just wanted to add a few words to what's already been correctly (to my mind) said about it. I think the reasons worries such as yours are relevant is that it has mostly been presumed in the identification literature that the process is active, so that some particular attitude or other must be directed at the motivation or action for one to identify with it. But I have been urging for a few years that identification can be passive, that one may be identified with some motivation or action without having taken up any particular attitude toward it. What I was trying to get at was just the kind of metaphysical, internal structural account of identification Chandra has already discussed, something that obtains or doesn't regardless of one's attitudes toward it. This is the sort of identification that may occur, for example, in cases of volitional necessity, where as one is about to perform an action, one suddenly comes to realize that one is unable to bring oneself to do so. To account for these and other cases, I favor the care-based approach: as long as the action flowed from or expressed (to use Chandra's favored language) one's cares (where these are construed as dispositions to respond emotionally to the up-and-down fortunes of the cared-for objects), one is identified with the action.

And just a quick response to Neil's last remarks: first, I'm not sure what you mean by "deep" or its requirements. I prefer "real" or "true" here when referring to the relevant selves from which I think (one conception of) responsibility flows. So a real self is compatible with lots of change. Nevertheless, I don't think the care-based approach is quite so compatible with rapid change, given that cares aren't desires; indeed, they typically involve lots of stability (and, indeed, recalcitrance in the face of judgments of their objects' low worth). And I worry about narrativity precisely because of Neil's talk of confabulation.

David: thanks very much for these thoughtful and helpful comments. It is I who am late to the party (as usual), not you; and, in any case, better late than a total party-pooper.

David and Chandra have reminded me of an important way of understanding identification that is less subjective and "active" than what is perhaps suggested by the term itself and Frankfurt's account of the notion. On this view, I agree that identification and "true or real self" come closer together (and maybe all the way).

"Identification" is not an element that plays an official or explicit role in my own account of the freedom involved in moral responsibility. I have preferred the paradigm of ownerhip to that of identification, in part because of my worries about the difference between "real self" and "ego ideal". But I wonder: with a more objective, less active account of identification, do identification and ownership (in my account, mechanism-ownership) get one similar results (or play similar roles)? Is there a reason to prefer one theoretical element to the other? (and thanks Neil for the thoughtful suggestion above);

There's a question here about the extent to which the "real self" has to be something that's actually present in the agent. One way of working this out is to suggest that it has two components: there are the underlying motives, which drive our actions but may not be (wholly) transparent to introspection. As a result, the other component--narrative--necessarily confabulates to some extent. One could say that the underlying motives don't make up the self, precisely because they are not accessible enough to form a basis for planning (if I don't know what my motives are, how can I plan for the future on their basis?), so we do the best we can through narrative structures, that only partly reflect but, as Neil notes, also have some power to change, those motives.

So if neither narrative nor the underlying motives give us a complete picture, why not say that the "real self" is just the conjunction of those, even if in practice the two dimensions never fully merge? If in looking for the real self we are trying to find something that's fully formed and devoid of ambivalence, we're going to be disappointed.

John: I've had your same wonder with respect to your account. I had noticed your preferred talk of "ownership" over the years (of the relevant mechanisms) and had idly ruminated on whether a more objective understanding of identification would be okay with you. I think it might, although I haven't thought sufficiently about it. For my part, I've found myself talking much more about "ownership" in recent years too, although I've basically been treating it as synonymous with the more objective understanding of "identification." Perhaps because of all the "activity" associated with talk of "identification" over the years, I've unconsciously thought of "ownership" as rhetorically less misleading. At any rate, I've always meant to talk with you about what you had in mind by "ownership," primarily because I'd probably like to steal your answer.

"To account for these and other cases, I favor the care-based approach: as long as the action flowed from or expressed (to use Chandra's favored language) one's cares (where these are construed as dispositions to respond emotionally to the up-and-down fortunes of the cared-for objects), one is identified with the action."

I would hate for my real self to be based ultimately upon my concerns. I have cared (even deeply) about lots of silly things over the years, still do. (I was class sports fan in high school; I still spend way too much time sweating the Tigers.) Now someone might say, 'You don't really care about those things; by your own admission they are trivial matters'. But aren't we in Cartesian territory here unless there are others concerns that I ought to have, being based upon the objective order of things? If there is no such thing, then I have myself a nice little private language going and my self is what I say it is, subject to further subjective approval. Am I getting any nearer to the truth about me? Or do we not think that there is here a fact to the matter? Thus, I don't see this approach to self-hood as an improvement upon Narrativism.

"If in looking for the real self we are trying to find something that's fully formed and devoid of ambivalence, we're going to be disappointed."

Where is the argument for this claim? Is it impossible that my real self should be something I have no hand in forming, simple and waiting to be discovered? Were those following the Delphic oracle completely misguided? What of St. Paul's writings or the inner journey described in The Confessions?


First, here's a great quote from "The Faintest Passion" that (I think) illustrates that Frankfurt himself wants to make a distinction similar to the one between "active" and "passive" identification discussed above:

"Wholeheartedness (the opposite of ambivalence) “is not a measure of the firmness of a person’s volitional state, or of his enthusiasm. What is at issue is the organization of the will, not its temperature” (NVL, p. 100).

But second, I'm really glad you raised the issue of how identification accounts contrast with your ownership-based account, because this is something I've always been puzzled about. It's not clear that they are competitors in the way that, say, event-causal and agent-causal libertarian views are competitors.

I've always thought of the two elements of your view -- moderate reasons-responsiveness and mechanism ownership -- as ways of saying that the MR agent has to be unified in various ways: across modal space (in the case of MRR) and across time (in the case of ownership). Does that seem accurate? Get rid of either one and you've got someone who is just too disjointed to be the proper source of his actions.

If that's right, then identification (at least as early Frankfurt uses the term) wouldn't be analogous to ownership, since (as you've pointed out in several places) identification in that sense is independent of history. But in Frankfurt's later work, he introduces the notion of caring as the thing that unifies an agent over time. This is perhaps simply a superficial analogy, but it's something. And I think it's not for nothing that the process of coming to own one's mechanism might be described in terms of one's coming to be confident in one's own agency. For Frankfurt, it's the things about which we cannot help caring that allow us to be confident in our own agency, and it's the quest for confidence which most of all seems to drive Frankfurt's work.

"Where is the argument for this claim? Is it impossible that my real self should be something I have no hand in forming, simple and waiting to be discovered?"

Even if that's right (and that seems dubious, because I suspect our conscious, reasoning activity has some effect on who we are), we engage in all sorts of diachronic activity that seems to require us to have some conscious sense of what our "real self" is: so something has to mediate between the underlying motives or dispositions and their manifestation in agency. But I'm not sure I see your point. Are you suggesting that there is a self lying in wait inside of us that we might never discover? Is it causally inert? Or does it just find subtle ways of manifesting itself when we find it? What does our discovering it change about our agency?

'Are you suggesting that there is a self lying in wait inside of us that we might never discover?'

That's exactly what I'm saying. Thanks for your thoughtful response. For chronicles of its effect upon agency, I would again suggest St. Paul's letters or Confessions by St. Augustine.

Thanks for getting the conversation rolling John! In regard to the issue about whether the grounds of identification are subjective or objective or both, I admit to loosing my grip on the very notion of identification as one waxes increasingly objective. I find this particularly true in Jaworska's PPR paper. Usually one tries to get others thinking about identification by giving cases in which the agent *experiences* a sense of passivity in acting (at least this is generally true of Frankfurt's work). There are times when our motivation leads us to act and we experience the motivation as moving us into action without our participation. Such cases seem to strongly suggest a subjective reading of the grounds of identification. I wonder what kind of sense we can give or what kind of cases can be offered to fix a notion of identification that is not (at least partly) subjective.

So while one may try to avoid John's worry by waxing objective, I worry (let me stress this is just a worry) that this undercuts either our ability to grasp the notion of identification or to evaluate claims about when an agent is or is not identified with an attitude.

An interesting point that Jaworska makes in this vein is that the subjective state of the agent may be evidence of the metaphysical fact that an agent is identified with the attitude. Indeed, it seems to me that she must rely on something like this to give the idea of identification some specific content.

I'd be curious to her if David or Chandra are worried at all by this or if they already have a sense about how'd they answer.

Some thoughts on this great discussion:

@David: One quick point about Deep vs. Real. People use the terms Real vs. True. vs. Deep Self more or less interchangeably. I think that is fine. I prefer Deep because its opposite is ‘surface’, while the opposite of Real and True is something like ‘false’. I like to be able to formulate my preferred criterion for MR like this: You are morally responsible for those *surface attitudes or actions* that express your Deep Self. Using the terms Real Self or True Self does not allow me the needed distinction. It makes no sense to say you are morally responsible for those *false attitudes or actions* that express your Real Self. It’s a small point, but I thought I’d throw that out there.

@John, Neal, and David: I don’t think this passive sense of identification (what I am calling internality) amounts to the same thing as ownership, not at least when ownership is construed along the lines that John does in Responsibility and Control. As Neal points out, internality is supposed to be ahistorical. Ownership obviously is not. Neal says that the later Frankfurt emphasizes caring and starts to incorporate diachronic notions into his picture of agency. I think that is right. But the diachronic elements that Frankfurt emphasizes are *forward-directed*, not historical (they are cross-temporal connections rather than continuities, to use Bratman’s distinction). For example, later Frankfurt talks a lot about the role of commitment in caring—to care for something is not just to desire it but to desire to keep on desiring it. The person is “disposed to take steps to refresh the desire if it should tend to fade”, etc.

Another difference is that internality operates at the grain of individual actions and attitudes. Ownership, as John characterizes it, operates at the level of entire mechanisms (which in turn generate different individual attitudes or actions in different possible worlds). So it is possible to construct cases where an action or attitude is internal but not owned, or owned but not internal, because they operate at very different grains.

Finally John gives ownership an active gloss when he talks about 'taking ownership', though admittedly there might be subtleties in how this gets unpacked. This passive notion of identification we have been discussing is supposed to be far more, er, passive.

Ooops. I said 'taking ownership' when I meant 'taking *responsibility*'.

The comments to this entry are closed.

Books about Agency

3QD Prize 2014: Marcus Arvan