The True Self and Reflective Endorsement

Hello again! Here comes my first substantive post, starting in with the concept of a 'true self.' It's possible that some of you will already be familiar with the ideas discussed in this one, but if you get bored with this first post, I hope I can rekindle on your interest with the next one. Anyway, here goes...

A lot of important work on the notion of a true self has been shaped in various ways by Harry Frankfurt's very intriguing example of the unwilling addict. Here is the basic idea:

Suppose that Mark is a heroin addict and has a very strong desire to get another hit of heroin. Now suppose that Mark also has the belief that doing heroin is deeply wrong. As a result, he wishes that he did not have this desire. He hates this aspect of himself and wants nothing more than to get clean and start a very different kind of life. 

Mark is experiencing an inner conflict, but many people have the intuition that the two sides of this conflict are not simply on a par. Instead, it seems that Mark's true self is the part calling him to get clean and that if he continues using heroin, he will be betraying his own true self. 

 

This strikes me as a very important fact, which reveals something fundamental about our understanding of the self. But what exactly does it reveal? If we start out with the assumption that the self is basically a collection of mental states, we will be immediately drawn to explain this intuition in terms of facts about the type and content of these states (desires, beliefs, second-order desires, emotions, etc.). 

 

Philosophers who take this approach may then be tempted to spell out the intuition that Mark's true self is calling him to get clean in terms of the fact that his second-order desires go in that direction, or that his beliefs do, or in some other way in terms of the type of mental state that is drawing him in this direction as opposed to the other. 

 

This whole approach seems to me to be mistaken. The thing that makes us regard this aspect of his self as the true self is, I think, completely unrelated to facts about which type of mental state is pulling him in which direction. To see this, consider a different case:

Suppose that Mark is gay and has a very strong desire to be with another man. Now suppose that Mark also has the belief that homosexuality is deeply wrong. As a result, he wishes that he did not have this desire. He hates this aspect of himself and wants nothing more than to fall in love with a woman and start a very different kind of life. 

In this latter case, our experimental studies show that most liberals have exactly the opposite sort of intuition. They conclude that Mark's true self lies in precisely the aspect of him that he reflectively rejects. Then they think that if he acts on the aspect of his self that he reflectively endorses, he will be betraying his own true self. 

 

In light of this result and others like it, we are coming to think that the notions of reflective endorsement, second-order desire, etc. just don't have anything to do with people's conception of the true self. Instead, this conception seems to be deeply tied to issues of value. On this hypothesis, the difference between the two cases comes down to something extraordinarily simple. It is that Mark's desire for heroin is bad whereas Mark's desire to be with another man is good.

 

To really spell out this idea in a broader philosophical framework, we are going to need to switch away from this emphasis on thinking of the agent's various mental states and instead adopt an approach that allows us to see why issues of value might be absolutely central here. But for that, we will have to wait till the next post!

Comments

[Jeremy Fantl]

This is great, Josh. I feel like there are cases in which I identify the true self with those non-belief states that are bad. For example, in the Clint Eastwood movie Unforgiven, the main character resists for most of the film a deep internal drive toward killing and drinking (out of a belief that such drives are wicked). The lesson of the movie seems pretty clear that in so doing he is denying his true self. (Whether his drives are wicked can be debated, but imagine the case so that they are, and my intuitions don't really change.) The difference between this and the heroine case seems to be more the artificiality of the internal drive. The heroine addiction (as it is often portrayed) was created by the addict's prior actions, whereas the Eastwood character's drive I think we're supposed to think came naturally. (This is totally consistent with your claim that we shouldn't think of the true self in terms of the kinds of mental states involved. It's more a question about whether value is the right way to think of it, as opposed to the source of the mental states.)

[Joshua Knobe]

Hi Jeremy,

Great to have you as part of this conversation. This is a really good point, and I definitely agree that any adequate theory in this domain will have to accommodate it.

Just to clarify, our study found that people with different values have different intuitions about the true self in these cases. Specifically, liberal participants tend to say that the agent's belief that gay sex is morally wrong is *not* part of the true self, whereas conservative participants tend to say that this belief *is* part of the true self. This result provides support for the claim that value judgments are playing a role in people's true self intuitions.

That said, I completely agree that there our cases in which people have the intuition that an agent's most wicked desires actually are part of his or her true self. Your example from Unforgiven Is a great case.

Perhaps you would disagree, but my guess is that value judgments play a role in people's true self intuitions but that other factors also play a role. Then, in your example, my guess would be that one of these other factors is leading us to see the agent's desires as part of the true self. But of course, that is just a vague hint at a strategy for addressing your question, and I would definitely be open to other suggestions.

[Kip]

Josh -

Can't wait for the next post.

One question I've had about all of this true self work is: I know Chandra has done work on the "deep self" or true self, yet reaches strongly compatibilist results. I don't see your result here as necessarily compatibilist or incompatibilist, but I think of your earlier work (e.g., on abstract/emotional judgments) as leaning incompatibilist. I'm wondering how this work here fits with Chandra's other work, and where that leaves compatibilism?

(It would also be good to hear your comments on how your work fits with Neil Levy's earlier work on the self and constitutive luck, because he reaches a compatibilist/antirealist conclusion about free will.)

One last point: I've found your work (e.g., with Shaun) to be some of the most important and insightful over the last two decades. One amazing quality is that the results are highly surprising. I never, in a million years, would have hit upon the idea that the structure/type of desires was irrelevant in comparison to questions of value, but it springs out from your groundbreaking research and experiments. Thanks again for doing such great and inspiring work!

[Kip]

Jeremy and Josh -

One more follow up comment, after I read Jeremy's comment:

There is actually a great reply available for Jeremy. I'm a movie buff and I'm familiar with Unforgiven (one of my favorite movies). The reply would be: "Drinking and killing are not entirely wicked. The drinking and killing enable the main character (William Munny) to go on a righteous act of vengeance against the truly wicked character, the villain Bill Daggett, after Daggett brutally murders Munny's friend. Without this darker-but-meritorious Cave Man part of his personality, the friend's murder would never have been avenged, nor would the brutal disfigurement of a town prostitute have been avenged. Although this part of his personality is darker and less savory, we see that it unleashes an awesome display of truly noble power and heroism, consistent with the traditional western ethos of masculine honor culture."

I think this interpretation is consistent with the epilogue that shows before the end credits:

"Some years later, Mrs. Ansonia Feathers made the arduous journey to Hodgeman County to visit the last resting place of her only daughter. William Munny had long since disappeared with the children... some said to San Francisco where it was rumored he prospered in dry goods. And there was nothing on the marker to explain to Mrs. Feathers why her only daughter had married a known thief and murderer, a man of notoriously vicious and intemperate disposition."

The film is implying the following message: the mother of Munny's wife, somewhat ironically, never knew what a true hero he was. She only never about him in a superficial sense, as a thief and murderer, without really knowing what good her daughter saw in the man, or seeing his true power for heroism and righteous vengeance and protection of his family and friends."

Anyway, I had to mention that just because it's a great movie and it invites that rebuttal. Like you, I suspect that there are cases where we identify a true self with true wickedness. I think Hitler might be a counter-example. But, even then, I wonder if Josh could cook up an example where Hitler was torn between the good and bad, and we naturally identify Hitler's true self with the side that wanted to help Jewish people rather than hurt them.

[Jeremy Fantl]

That certainly sounds plausible to me (that value judgments play a role alongside other factors), though it's possible that the value judgments are taken as indicating the presence or absence of those other factors, rather than as themselves essential features of the true self (for example, if people have a presumption that bad drives tend to be artificially created).

[Angra Mainyu]

Hi Joshua

That's a very interesting post; a lot of food for thought.
I'm not entirely sure there is a "true self" (maybe I'm not getting the concept right; I'll have to consider this further), but assuming there is, have you considered a variant of the reflective endorsement theory, on which the reflective endorsement has to be both free and the result of rationally ideal reflection?
That seems to get around the case of Mark, it seems to me.
I don't know about the case of the main character in "Unforgiven". Assuming there is a "true self", is he denying it?
I'm not sure about that.

[Joshua Knobe]

Thanks Kip! I really appreciate the kind words, and I look forward to hearing more about your thoughts on these issues.

Regarding the question about whether compatibilism is intuitive, it seems like the first thing to notice about that question is that it is deeply *confusing*. A theory of people's ordinary understanding of agency should help us to make sense of this confusion. If a theory just says that people's ordinary understanding is straightforwardly compatibilist or that it is straightforwardly incompatibilist, it seems that this theory is dodging the key phenomenon we need to explain.

I think that an account of people's conception of the self has the potential to illuminate these issues, helping us to see why people often feel pulled in conflicting directions. Maybe I am wrong about that and we ought to be looking for an explanation somewhere else. Still, regardless of whether my explanation is right or wrong, it seems that the phenomenon we need to explain is people's tendency to find these issues puzzling and to be pulled in conflicting directions.

More on the other issues soon...

[Angra Mainyu]

Joshua and Jeremy

I've been thinking about possible examples in which the "true self" (if there is one) isn't a good thing (and probably it's bad), and it seems to me that fictional aliens provide such examples.
For instance, the aliens from "Predator" go around hunting humans, but that is not going against their true self, it seems to me. The aliens from "Independence Day" also just go around trying to exterminate humans, but that's apparently not against their true self, if I make an intuitive assessment on the matter.

[Kip]

Josh - thanks! I think Shaun does a great job of moving us toward understanding that confusion in his book and corresponding articles. Specifically, the free will confusions seems to result from tension between (A) a psychological mechanism that is greedy deterministic in explaining the world and (B) other psychological mechanisms that intentionally hide/distort the underlying causality. Regarding (B), some good candidates are victim/perpetrator bias (i.e., victims think perps act for no reason), the fundamental attribution error, the positive illusions, and reactance.

Angra - I think that is a great example. Fascinating.

[Joshua Knobe]

Jeremy and Kip,

This is really a fantastic example, and I very much hope that it leads to the development of a whole new literature at the intersection of agency theory and film studies. ; -)

I appreciate Kip's suggestion that one might see the act of killing in this case as not truly being morally bad, but let me pursue just a little bit farther the idea that other factors are playing a role.

To give an example of what I have in mind, consider the role of SKILL. It strikes me as important that the Clint Eastwood character is portrayed as being very, very good at killing people. What if it had gone the opposite way? Suppose that he had been a completely incompetent assassin and that he then switched over to taking a job as a farmer, which he did extraordinarily well. Other than that, let's suppose that everything stays the same.

Would we say in this case that, even though he is an amazing farmer, his true self is calling him to go back to being a bumbling assassin? I have not run the study, but my own intuition is that this version makes things very different. More generally, it seems that people's intuitions about an agent's true self are often affected by judgments about what this agent is good at.

Okay, that particular hypothesis might turn out to be wrong, but perhaps it helps to illustrate a larger point. These intuitions are affected by a wide variety of factors, and many of them go far beyond anything about the agent's specific mental states.

[Joshua Knobe]

Angra,

These are both excellent points. Just to clarify, were you thinking that your second point could be explained by the first? Is it that we have different intuitions about the extraterrestrials because we think they would arrive at different conclusions after rationally ideal reflection?

[Angra Mainyu]

Joshua,

Thanks.

With regard to your question, I think the "ideal reflection" theory is compatible with the alien cases. I don't know whether the theory is correct (even assuming there is a true self), so I wouldn't go as far as to say that that is the reason we have different intuitions. I would more tentatively say it might be, and so far I haven't found counterexamples (but I haven't spent a lot of time looking for them, so I'll have to do more digging).

Also, I'd like to elaborate a bit on a couple of other issues involving my examples:

1. I don't know I would say the aliens are behaving in a morally wrong (i.e., immoral) manner. I'm undecided between that, and the assessment that they're not moral agents at all (so nothing they do is immoral, but it can be a bad thing, like a deadly disease spreading among the public, etc.). But either way, I would say that their true self (if there is one), or at least part of it, is a bad thing, either because it makes them strongly inclined to behave immorally, or if they're not moral agents, just a bad thing like the dengue virus can be a bad thing.

2. I'm not sure the "true self" of a human being can be a bad thing (even under the assumption that there is a true self). Maybe, or maybe not, but my doubts on the matter also seem to mirror my doubts about the "Unforgiven" example. I haven't seen the movie, but my worry is a more general one about human psychology. But alien examples seem to get around that difficulty.

[Tamler Sommers]

Jeremy's Unforgiven case is really interesting. I agree with the intuition that his true self is the drinking and killing one. And the "I ain't like that no more" mantra is clearly something his wife drilled into him, not something he came to naturally. Another factor though may be how incompetent he is at living a straight and narrow life. All of his pigs have the flu, he doesn't have any crops, he can't ride a horse properly anymore--the new him isn't working. That's not moral exactly at least in the narrow sense, but it is an example of someone not flourishing. If he was great at farming and his kids were doing well and he seemed happy, I'd be a lot more conflicted.

[Paul Torek]

Tamler,

Bingo: "it is an example of someone not flourishing." I want to expand on that, because I think it provides a way to understand the "true self" that doesn't necessarily involve any metaphysical heavy lifting. It also provides a potential underpinning for Angra's "ideal reflection" theory, viz., ideal reflection would lead me to a certain way of life because (or in part because) that's how I can flourish.

To use Joshua's terms from the previous post, a person isn't just a set of mental states. She's a living organism. Obviously her mental states must to some degree reflect her organic nature, but that leaves a lot of room for an independent reality of "true self" that the mind may not always accurately capture.

Now a human being is a social animal, so that points to an obvious important role for interpersonal morality. But it also, apparently, leaves room for the idea that someone's true self may be called to wronging others.

So, while this discovery of evaluatively laden "true self" thinking may require some philosophers to revise their overly mentalistic theories of appropriately agent-sourced decision making (or else reject large sets of intuitions), I do think it can be reconciled with a relatively modest metaphysics of the person.

[Joshua Knobe]

Hi Angra, Paul and Tamler,

Just a quick note to say that this approach sounds really promising. Of course, the notion of flourishing is in some ways quite different from the notion of being morally good, but on the scale of things, one might regard the difference between them as something of a family squabble. Both of these hypotheses involve a kind of normative evaluation, and the two are therefore much more similar to each other than either is to the hypotheses that emphasize different types of mental states.

[V. Alan White]

Terrific post and great discussion. While I'm waiting for the next post, I'd like to chime in my overall agreement that values are the keys that unlock lots of of these mysteries. But there are lots of keys! One basic question for Joshua: do you think values are more important from an internal, external, or combined perspective? (I suspect the latter, if for no more reason than considering the role of socialization and acculturation in how values actually operate.)

One remark to Angra (glad to see you here again!) on the alien example. Just last night in an email I argued to an anthropologist friend that alien thought-experiment examples might show us how cross-cultural studies might be pursued without positing metaethical cultural relativism. I posited that somewhat less invasive aliens might dispassionately study our differing moral behaviors not because they embrace some sort of Prime Directive or such metaethical stance, but because they completely lack anything like a concept of morality or moral values. They might well qualify as self-reflective true selves in studying us that way, even if they lack an evaluative apparatus to see themselves as embodying a value like curiosity in doing so. (Or, perhaps one could argue that self-reflective true selves simply cannot exist in the absence of an evaluative apparatus.)

[Angra Mainyu]

Joshua and Paul,

The "flourishing" approach sounds very interesting, though I'm not sure how it would handle cases of self-sacrifice that don't conflict with an agent's true self.
For example, one might consider a case in which a person sacrifices her life (or worse, endures something horrible for the rest of it) to save her loved ones from an even worse fate, even in a situation in which she reckons nobody else will know what she did, and yet she's not going against her true self (I can construct a scenario if you like).
I suppose "flourishing" might be said to include self-sacrifice of one's life (or enduring horrible suffering, etc.), but that doesn't seem plausible to me; still, if it does, then the concept of flourishing might end up being no less mysterious (to me, at least) than that of true self.
Regarding the "ideal reflection" theory, I don't have a counterexample yet, but I'm looking for something along the lines of an agent whose true self could be irreflective. If such agents are possible, then the "ideal reflection" theory would be falsified.

[Angra Mainyu]

Thanks, glad to see you here too!
Regarding your aliens, it's a difficult and interesting question. I don't want to stray too far from the topic, so I'll try to be brief:

I'd say actual advanced aliens probably are social beings if they evolved (even after millions of years of genetic engineering and integration with machines), so they probably have something akin to moral rules (though the rules might be quite different from ours; that might lead to some sort of species relativism); alien AI might not, though they still would need to have some way of making evaluations from their own preference structure, in order to make decisions.
All that said, there is the question of whether it's metaphysically possible that they have no evaluative apparatus at all. I'm not sure. Even if it's metaphysically possible, I'd bet against it when it comes to actual aliens, even AI, cyborgs, etc. But they might still study us in a dispassionate fashion; their psychology might be very different from ours.

[Ben M-Y]

Hi Josh,

Thanks for launching this fascinating discussion here.

I wonder if you've done any studies or thinking about wantons. These are creatures with no reflective attitudes. Given Frankfurt's account, they have no higher-order volitions. But we might conceive of them as having no evaluative attitudes about their own desires. Their mental economy is devoid of reflective endorsements about their own motivations.

There is a significant difference, it seems to me, between the case of a wanton and your Mark cases. A wanton cannot experience the kind of conflict involved in your Mark cases. A wanton would not be aptly described as wishing that she were rid of certain desires, because she simply does not care about her own desires. So it seems as though thinking about wantons might be useful because it would provide a new set of cases on which to test our intuitions.

Now, I wonder what people's intuitions are about whether wantons have true selves. And, if people do think that wantons have true selves, I wonder if their conception of these true selves are sensitive to normative considerations in the same way. To report my own initial intuitions: I do not think that wantons have true selves of the sort in question here. But I really do wonder what others think.

Suppose for a moment that folks agree with me, and we conceive of wantons as lacking true selves. I think this intuition has some bearing on the conclusions you draw from your work on Mark-type cases. In particular, I think that this might show that the agent's attitudes do play some role in determining her true self. In describing your findings, you say that "notions of reflective endorsement, second-order desire, etc. just don't have anything to do with people's conception of the true self." But if wantons are thought not to have true selves, this would seem to suggest that we should temper your conclusion a bit. Perhaps our conceptions of the true self do have something to do with reflective endorsement, but also something to do with values (or norms). As an initial attempt at a description of what's going on, we might think that our conceptions of the true self depend on reflective endorsement as a ticket into the game--only those creatures who are reflective in the way that Mark is (and the wanton is not) have true selves.

This is all fairly sketchy. I would love to hear what others think.

[Joshua Knobe]

Hi Ben,

Wonderful to see you here. Very much looking forward to hearing more about your thoughts on this!

Just as you suggest, if it turns out that people have the intuition that wantons do not have true selves, that fact would provide strong evidence against my view. So all I can say is that I do not think people would have this intuition.

Let us imagine a person who has no second order desires of any kind. She has first order desires and first order moral beliefs, but she has no attitudes at all about her own mental states.

Suppose that this person grows up in a racist society. She is taught to treat people of other races as inferior and sincerely believes this to be the right thing to do. Then, one day, she experiences a surge of compassion and radically changes her desires and beliefs. She now works for racial justice, believes that her past actions were morally wrong, and feels guilty for the things she has done.

However, because she has no second order states, she has no attitude at all about either her previous mental states or her present ones. So she believes that her past actions were morally wrong and that her present actions are morally better, but she has no attitudes about the mental status that generated those actions - only about the actions themselves.

Although it is clearly an empirical question, my guess is that people would say in such a case that her desire for racial justice is part of her true self. Then, more generally, my sense is that second order states play very little role at all in people's ordinary understanding of agency. In all of the famous philosophical examples that involves such states, my guess is that people's intuitions are actually driven by some other factor entirely.

[Paul Torek]

Angra,

I think we could use "flourishing" to include self-sacrifice where the alternative is, per your example, death of one's loved ones. And sure enough, that makes it just as mysterious as true self. But when we acknowledge both biological and cultural, both physical and mental, aspects of the self, I think we generally know how to recognize advances in true-self-understanding - even though it often can be quite difficult to decide prospectively, and even though philosophical metacognition on this topic is much harder.