Blog Coordinator

« Responsibility and Epistemic Confidence | Main | Autonomy and Authority »



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

Hi Suzy,

I’m very interested in the issues you are raising (and plan to raise in future posts), and I’m excited to hear more about what you think about these matters.

For now, I want to ask a question about a nuance in Bratman’s view. Bratman not only claims that a desire endorsed from the perspective of a self-governing policy speaks for the agent but, crucially, that the relevant policy must be one with which the agent is *satisfied*—roughly, the policy’s functioning must not be undermined by any other of the agent’s policies. So, in your description of the case of the conflicted agent, who acts on a desire to sleep with same-sex partners, which s/he has a policy not to treat as reason-giving, we need to add that this policy is not in conflict with any other of the agent’s policies.

[There are some delicate issues relevant to the case. I want to be sensitive to as many of them as possible—for example, I want to be sensitive to the fact that for many people there is social pressure to resist their homosexual desires and that these pressures can be implicated in serious, and sometimes harmful psychological conflicts. So there is good reason to believe that in cases like the one you are describing there are external pressures playing a non-trivial role in the psychological conflict at the heart of the case. These are important issues, and they would seem to play a large role in these cases and our intuitions about them. For present purposes, however, let’s suppose this is not so. That is, let’s suppose that the agent’s externalization of her/his desire to sleep with same-sex partners is internally caused and not due to social pressures (e.g., from her/his upbringing or social context). That will (I hope) allow us to sharpen our focus just on the issues you raise in your post.]

I wonder what you think about the satisfaction condition and its role in this case (and similar ones). Does recognition of the agent’s satisfaction with her/his policy make a difference with respect to your intuitions about autonomy, consent, paternalism and related issues? Do you find the conclusions you mention regarding paternalism and consent as "wildly counter-intuitive" given that this agent's commitment to the policy is wholehearted, so to speak? I’m not sure what to think myself, but it seems reasonable that the nuance might mitigate or undermine the relevant intuitions. I’m not sure.

I look forward to hearing what you (and others) think.

Hi Ben,
Thanks for your comment! You're right that Bratman's view is much more nuanced than I gave it credit for in my post. I don't think that bringing in the idea of satisfaction is going to help with the problem I identified, though. My worry is precisely that Bratman's view will categorize too many actions as non-autonomous, and hence if we used it as the basis for a theory of consent we would be forced to say that the sex from my original example was non-consensual. Once satisfaction is introduced as an additional condition, even more actions will have to be considered non-autonomous. (This is precisely for the kinds of reasons you point to - social pressures are often internalized as self-governing policies, which then conflict with other of our self-governing policies.) But unless we want to say that any time these kinds of conflict arise the agent fails to give valid consent, we're back to square one.

With respect to what we should say about agents who are 'wholehearted' in their commitment to policies that they then violate in their actions: I think these kinds of cases are trickier, but I still think they struggle getting plausible conclusions about paternalism or consent. Imagine I have a self-governing policy never to treat my desire for a second helping of desert as reason-giving, and this coheres with all of my other self-governing policies (would that it were so!). If I one night decide what the heck, I'll have another piece of pie, and you prevent me, it would seem odd to say that your intervention is not paternalistic.

That said, I do think that something like what Bratman calls satisfaction is important in a theory of autonomy, and I'll be laying out my own version of it in future posts.

Hi Suzy,

Thanks for your reply! I'm looking forward to hearing more about your own view. I'm particularly interested in hearing what you have to say about things like social pressures and how they bear on issues related to autonomy and the self.

One more question about the current post. Have you thought at all about how the results Josh Knobe was reporting here last month might relate to the intuitions you are relying on in your arguments here? Josh reported finding that our intuitions about the true self are sensitive to our own values. And he discussed a case similar to the one you appeal to in your post. In his case, the agent is torn between a desire to engage in homosexual acts and a belief that this would be wrong. He and his colleagues found that one who thinks that there is nothing wrong with homosexuality will be more likely to judge that the desire to engage in homosexual acts expresses the agent's true self, whereas one who thinks homosexuality is wrong will be more likely to judge that the belief expresses the agent's true self.

I wonder if you agree that our intuitions about these cases are sensitive to our own views on certain matters. And I wonder if you think this is in any way significant to how we should go about arguing for and against various candidate views. To put a finer point on it, you say that certain intuitions about paternalism and consent in the case of the agent torn about homosexual acts are "wildly counter-intuitive" but similar intuitions about paternalism in the case of a second dessert "seem odd." Perhaps I'm reading into your language here. But if these characterizations of your intuitions reveal something, they might seem indicative of something like the tendencies Knobe reported (given certain assumptions about your own values and perhaps also certain assumptions about the relationship between the notions of paternalism/consent and the true self). Do you have any thoughts about this?

Hi Ben,
I should probably be up front about how much work I think intuitions can do, when theorizing about concepts such as autonomy. Though I've used intuitions to motivate an objection to Bratman, I don't think intuitions alone suffice to show that a theory should be rejected. I'm aiming for something like a reflective equilibrium - a theory needs to be balanced against intuitions about particular cases, and I'd also add application to other domains, such as responsibility or paternalism. But it's not always the theory that has to give. This is especially so if intuitions about particular cases cannot be reconciled with one another, in a theoretically coherent way.

To tie this back into the results Josh was discussing last month: I wouldn't be at all surprised if intuitions about the autonomy of particular actions were pushed around by our own values (I think something like this happens quite a bit in discussion of adaptive preferences, for instance). But if we're going to build value judgments directly into a theory of autonomy, we need to first check what effect that will have on a wide array of cases, and how it will gel with other domains (paternalism/responsibility/consent/etc). I think we achieve a more stable equilibrium overall by having a value neutral theory, and rejecting some of our intuitions about particular cases, than we would with a value-laden theory of autonomy.

Hi Suzy,

Thanks for saying more about how you are thinking about the dialectic. I'm very sympathetic to a reflective equilibrium approach like the one you're describing. This all sounds really interesting! I look forward to hearing more.


Thanks for the thought-provoking post. I will take a look at your follow up post next. For now, I just wanted to float a possibility: I have often thought that perhaps free will should be defined negatively. That is, there is no positive account of free will to be had. Rather, free will is what we have when we are not paradigmatically unfree. So, to be free is to be free of constraint, external manipulation, internal compulsion, etc. In short, this way of illuminating free will comports with the way old-school soft determinists like Ayer, Stace, Wittgenstein, and others talked about free will. On this view, to have free will is simply to be free of the sorts of things that undermine free will. Perhaps something similar is the right approach to autonomy--namely, to act autonomously is to act without external constraint, manipulation, compulsion, etc. Some additional conditions will need to be added to differentiate young children from adults (and non-human animals from human animals). But the bulk of the work in understanding autonomy will be spent looking at things that undermine it (rather than trying from the outset to positively define it).

A related question: On your view, what's the difference between acting autonomously and acting freely? Can one exercise free will yet not be autonomous? Conversely, can one act autonomously but lack free will? My worry is that these two words pick out the same concept (or very similar and overlapping concepts). Can you think of cases where the two can be pried apart?

Hi Thomas,
Thanks for your comment. I find myself skeptical of defining autonomy negatively, though I'm going to have to give it a lot more thought. Part of my skepticism may just be temperamental - I'm drawn to looking for a unified explanation of autonomy, and I worry that approaching it negatively will result in an ad hoc collection of factors. That's obviously not an argument, though!

Here's a slightly more considered worry: if we're going to define autonomy negatively, then we'll need some principled way to differentiate manipulation from regular socialization, and internal compulsion from acting on a deep commitment. If we're going to appeal to more than brute intuition, it seems to me that we need to pin down something that the regularly socialized and the deeply committed get right, which is missing in the other cases. At that point, we'll be well on the way to laying out a positive account of free will/autonomy.

This reflects a deeper concern, too: you mention that we'll need some additional conditions to differentiate young children from adults, etc. I think this differentiation needs to be front and center, rather than an addendum to a theory. If to be free/autonomous is just to be free from external and internal constraint, then it looks like a rock or a person in an irreversible coma is free/autonomous. To avoid this conclusion, we need to pick out something the agent has that the rock and the comatose person lack - something like an active will, or a self that's (capable of) governing. Again, it looks like a positive account needs to come first. Do you see a way to avoid these problems?

With respect to the relationship between free will and autonomy: I have to confess that I have a very hard time pinning down what free will is supposed to be tracking, especially if it's meant to be something other than whatever it is that grounds moral responsibility. It may well be that the two concepts are ultimately tracking the same thing, though there do seem to be significant differences in the kinds of questions and problems the respective literatures grapple with.
For instance, the autonomy literature rarely worries about determinism; it's much more focused on social forms of control. More broadly, metaphysical questions about causation seem much more front and center in the free will literature than the autonomy literature (though this is less true of its action theoretic wing).

For what it's worth, I don't know of any accounts of autonomy that aren't (at least implicitly) compatibilist; so I guess an agent could lack free will in a deterministic universe, but still have autonomy. The converse case is trickier: on my own account of autonomy an agent who intends to act and then fails is not autonomous, and my sense is that such an agent could still be exercising free will - though I should stress that this is one of the things that makes my account a little idiosyncratic, since most theories of autonomy focus on desires rather than actions.

The comments to this entry are closed.

Books about Agency

3QD Prize 2014: Marcus Arvan