Autonomy Consequentialism

Maximizing respect for others' self-regarding preferences

In How Intention Matters, I lamented the common myth that concern for people’s intentions and quality of will was inherently “Kantian” or otherwise non-consequentialist. Today we do the same for autonomy. TL;DR: you can optimize differently, so valuing autonomy is not by itself any reason to reject consequentialism.

Valuing Autonomy

Valuing autonomy over utility can mean a couple of different things. As a first step, it can mean rejecting paternalism: refusing to impose your judgment of a person’s best interests in a way that overrules their choices or personal priorities. This in turn can be broken down into: (i) deferring to their ultimate preferences rather than your preferred theory of welfare; and (ii) deferring to their right to make outright imprudent (or diachronically instrumentally irrational) choices, such as prioritizing a weak present preference over a strong future preference.

As a second step, it can mean rejecting “busybody” (other-directed) preferences, so that each person is treated as the final authority over what would be non-instrumentally good to happen within the sphere of their own life. I find this pretty appealing, and it certainly seems an important corrective over unrestricted preference utilitarianism: it shouldn’t matter how many other people maliciously desire that I suffer, those malicious outsider preferences should not override my own preferences about my own life.

Similar principles would seem to carry over to less extreme or overtly vicious cases of sticking one’s nose where it doesn’t belong. Consider a young fellow who wishes to leave his traditional (perhaps Amish or ultra-orthodox) community. Suppose everyone in his community dearly wishes him to stay. Those wishes may be well-meaning, and even grounded in love of a sort. But in valuing autonomy, we hold these third-party wishes to lack weight. (We may even hold them to be viciously controlling if held unconditionally, i.e. if the other community-members want the young fellow to remain with them no matter what he wants for his own life. A better love, we of liberal bent are wont to say, would hold our preferences for his life to be conditional upon deference to his, such that our will serves only to amplify his—so far as his own life is concerned.)¹ It doesn’t matter how many others wish him to stay. If he wishes to leave, that suffices to settle that his leaving is non-instrumentally better, as far as the value of autonomy is concerned.

The agent’s personal preference is decisive

That’s not to say that self-interested decisions are always all things considered better. Instrumental value matters too. For example, if a magical curse will condemn the remaining community members to a painful death if ever one of their clan leaves them, then that changes things! They may reasonably wish not to die. And that is a wish about their own life. So the autonomy consequentialist may advise our young man to give up his dreams and remain behind in order to save his community, if it comes to that. But they won’t advise him to remain merely because others want that for him. Only suitably self-directed preferences count.

What Autonomy Isn’t: deontological

Crucially, there is nothing about the value of autonomy that should lead to distinctively deontological verdicts. Consider, e.g. Trolley Footbridge. We maximize autonomy by pushing the fat man off the bridge and saving the five on the tracks. Why? Because this way five people get the future they autonomously desire, rather than just one of the six. (Duh.)

It’s weirdly common for people to not notice this. They’ll say things like, “Respect for autonomy requires you to defer to the fat man about what should be done with his body.” They never think, “Autonomy requires us to defer to the five on the tracks about what we let happen to their bodies.”

Of course, there is a metaphysical difference here: we are not personally acting upon the five when we leave them to die by trolley collision, whereas we would be acting upon the one if we pushed him off the bridge. But that’s just a fact about the causal situation. It’s not, on its face, anything to do with valuing autonomy as opposed to other stuff.

Now, as it happens, I think there’s a strong instrumental case for instituting negative rights to bodily autonomy, so that other people can’t kill you or use your body without permission.² But that’s a matter of the instrumental value of negative rights (as part of a principled pragmatic proceduralism), not a matter of valuing autonomy as opposed to other stuff.

Crucially, you can’t explain the positive/negative distinction by reference to autonomy.³ Rather, the “commonsense” conception of bodily autonomy presupposes that we value negative over positive rights, while the only good justification for this—it seems to me—is instrumental, i.e. fundamentally consequentialist.

Conclusion

There are two separate questions that are too often conflated. One is whether you value autonomy over “utility” (whatever might be meant by that). The second is whether you ascribe negative (rather than positive) rights to bodily autonomy, and if so whether your basis for doing so is fundamentally instrumental or non-instrumental. Believing in fundamentally non-instrumental negative rights is distinctive of deontology, but this difference is not best characterized as “valuing autonomy”. An autonomy consequentialist may endorse negative rights on instrumental grounds, but most fundamentally should want people to get to live the lives they want whether this is threatened by action or by inaction. Killing one to save five serves just as well to maximize others’ autonomy as it does to maximize their well-being.⁴

1. An exercise for the reader: Does valuing autonomy in this way imply that open/poly relationship norms are more ideal than exclusive/monogamous ones?

2. I take it there’s not the same instrumental value to giving people such strict positive rights to bodily autonomy that the rest of us must ensure that things go for their body as they wish. Such positive rights are hard to satisfy, and there’s more instrumental reason to limit than to expand agents’ discretionary powers so far as their impact on others is concerned. So, that’s why (I take it) sensible people endorse commonsense negative rights.

3. Here’s a quick objection to Kantianism (channeling Kagan’s classic objection to rights): Kantians claim to derive all of ethics from the categorical imperative—either universalizability or treating people as ends in themselves, depending on which formulation they favor. But it seems like they really presuppose (rather than derive) the positive/negative rights distinction, when they suppose that treating people as ends in themselves requires consent in order to cause them harm or use their body to save others rather than requiring consent in order to allow them to be harmed or affected bodily in ways that are contrary to their will. It’s not clear that there’s any basis in the basic notion of consent or in the status of rational agents as ends that explains the deontologist’s asymmetric treatment of causing versus allowing harms. It’s arbitrary intuitionism masquerading as pure reason.

Or so it seems to me, at least. Let me know if you can think of a compelling counter on behalf of the Kantian.

4. Which is to say: it serves either well enough in theory but may undermine both in practice, depending on the details of the case.