By In Discussions, Ethics Discussions at PEA Soup, Metaethics, Normative Ethics, Political Philosophy Comments (31)

Ethics Discussion at PEA Soup: David Enoch’s “Hypothetical Consent and the Value(s) of Autonomy,” with a critical précis by Beth Valentine

Welcome to what we expect will be a very interesting and productive discussion of David Enoch’s “Hypothetical Consent and the Value(s) of Autonomy.” The paper is published in the most recent edition of Ethics and is available through open access here. Beth Valentine has kindly agreed to contribute a critical précis, and it appears immediately below. Please join in the discussion!

Précis by Beth Valentine

“Hypothetical consent is puzzling.” (p.1)  This is how Enoch begins his paper, but by the end I was convinced that this claim is false. “Hypothetical Consent and the Value(s) of Autonomy” motivates this initial puzzlement by pointing to intuitions regarding hypothetical consent that, at first, appear to lack a cohesive explanation. Through examining actual consent and autonomy, he does much to explain away this puzzlement and argues that hypothetical consent can, in some contexts, make a normative difference.

I start my comments at the paper’s conclusion, where Enoch helpfully lists what he takes to be the four lessons of his paper. I’ll then motivate the puzzle of hypothetical consent and discuss lessons 1-3, commenting only briefly and indirectly on 4. Somewhat paraphrased, these lessons are:

  1. Hypothetical consent can still be normatively significant even if it doesn’t do the same normative work actual consent does.
  2. Currently, there is no conclusive argument against hypothetical consent; there are ways to adequately respond to the two main arguments in the literature.
  3. Consent matters for different reasons: sometimes it matters because of non-alienation, sometimes because of sovereignty. Hypothetical consent can matter because of the former but not the latter.
  4. Non-alienation can be understood in terms of Frankfurtean endorsement and higher-order attitudes. (p. 34)

The Puzzle

We sometimes think hypothetical consent makes a normative difference. Imagine there is an unconscious patient who needs a blood transfusion and an unconscious Christian Scientist who needs the same procedure. (Enoch calls these cases The Unconscious Patient and The Unconscious Christian Scientist.) Presumably, it is more permissible to give the transfusion to the unconscious patient than to the unconscious Christian Scientist; the fact that one would not have consented had she been able to but the other would have does some work in explaining this difference. Yet, if I take your property without asking, the fact that you would have consented seems irrelevant. Why this difference?

  1. Hypothetical Consent & Actual Consent: The Same Normative Work?

One of the interesting gems of Enoch’s paper is his admonition that we should not theorize about hypothetical consent with “an overly poor menu of normative upshots.” (p. 11) Our moral practices can have a variety of normatively significant effects, in varying degrees: they can make an impermissible action permissible (or perhaps just less impermissible), create powers to create duties, create duties, defeat reasons, etc. The “upshot” of this observation is that, even if hypothetical consent doesn’t have the same effects as actual consent, it can still have normatively significant upshots.

However, I wonder whether we are justified in giving hypothetical consent the name of consent if its normative work differs from that of actual consent. There is some virtue in having names accurately reflect their concepts. If we are going to call something consent – even if we admit it isn’t actual consent – it should share some resemblance with its namesake. One natural way of establishing this resemblance is by showing that both practices have the same normative upshot for the same, or at least similar, reasons. At this point, though, my worry remains premature. If Enoch is right, hypothetical consent may do at least some of the same work as actual consent, at least some of the time, and this may be enough to justify bestowing the practice with the name “consent.”

  1. Challenges to Hypothetical Consent

Before offering a positive account of why we should let hypothetical consent do (some) normative work, Enoch tackles the two main challenges to the relevance of hypothetical consent.

On the one hand, we have the no real work challenge. This view claims that the reasons for which we impute consent do the real normative work. Any judgment about whether the agent would have consented is merely a by-product of the factors that generate the permissibility. No real work is supported by the Transitivity Argument (p. 8):

(1) The Normative Upshot holds in virtue of Hypothetical Consent.

(2) Hypothetical Consent would be given in virtue of the Underlying Reasons.

(3) The in-virtue-of relation is transitive.

(4) Therefore, the Normative Upshot holds in virtue of the Underlying Reasons. (From (1), (2), and (3)).

(5) Therefore, Hypothetical Consent does no normative work in justifying the Normative Upshot. (From (4))

The main objection Enoch raises against the Transitivity Argument relies on multiple realizability. However, before discussing this topic, I want to suggest that he dismisses too quickly another possible point of push back. Gesturing to the grounding literature, Enoch raises the possibility that transitivity does not hold across different types of in-virtue-of relations. The argument might then equivocate if the in-virtue-of relation in (2) is causal but in (1) is normative. Enoch bypasses this worry by focusing on cases “in which it seems like the Hypothetical Consent holds normatively in virtue of the Underlying Reasons.” (p. 7).

To see if this bypass is plausible, I think it is helpful to (briefly and roughly) introduce a distinction Donald VanDeVeer raises in Paternalistic Intervention (1986, see pp. 71-75). Hypothetical rational consent asks what a fully rational agent (who may be aware of all relevant facts) would consent to. It is the type of consent that commonly pops up in political philosophy and contractarian-like views. Here, Enoch’s bypass seems reasonable. However, there is another type of hypothetical consent which VanDeVeer calls hypothetical individualized consent. This consent inquires if an agent would have consented had she been able to by considering all of her known or likely attributes. Foreshadowing Enoch’s own commitments on non-alienation, this type of consent takes into account the agent’s desires and commitments instead of abstracting away from them. For this type of consent, it seems as if Underlying Reasons are purely motivational and so causal. The reasons themselves have no normative role to play in why we give hypothetical consent force. While I find Enoch’s later comments on multiple realizability persuasive, it is worth noting this other way of resisting the Transitivity Argument for some types of hypothetical consent.

Enoch’s rejection of the Transitivity Argument, and so the no real work objection, resists the move from (4) to (5). He notes that Hypothetical Consent can hold in virtue of different sets of Underlying Reasons. In such cases, it seems plausible that all that matters is simply that there is hypothetical consent; we needn’t further concern ourselves with which set of Underlying Reasons hold. For example, it doesn’t matter why the unconscious patient would have consented; what matters is that she would have. (As Enoch notes, in this way hypothetical consent resembles actual consent.)

The second challenge claims that hypothetical consent is no substitute for the real thing. Since normative powers generally must be exercised if they are to have effect, the gist of this challenge is that hypothetical consent is no more useful than hypothetical water is when one is thirsty – it simply is no substitute for the real thing. In response, Enoch notes that “going hypothetical” is sometimes an acceptable move. Whether such a move is legitimate depends on the philosophical motivations for the original theory. Invoking what one would have seen might be legitimate for a theory of colors, but referencing what one would have valued for a theory of normative concepts may not be. Enoch’s response thus leads us to lesson three, where he explores why we care about actual consent. He’ll conclude that, at least when we care about non-alienation, hypothetical consent is a good substitute for the real thing.

  1. Non-alienation and Sovereignty

Having shown that there is at least hope for normatively significant hypothetical consent, Enoch lays the groundwork for his positive argument by looking at the normative significance of actual consent. Here, he encourages us to embrace the complexity of consent, noting that it would be surprising if it we get a general, context-insensitive account of why consent matters. His account is not surprising, at least not in this regard. He holds that consent matters both because of non-alienation and because of sovereignty. Let’s start with non-alienation. Recall The Unconscious Patient and The Unconscious Christian Scientist. To these cases, add The Conscious Anxious Patient and her unconscious counterpart. The anxious patients fear needles and so refuse or would refuse a blood transfusion on these grounds. Here is Enoch’s assessment of the cases: We can permissibly administer the transfusion to the unconscious patient and the unconscious anxious patient. He’s less sure of administering the transfusion to the others, but he is sure that it is more problematic to administer it to either Christian Scientist than to their anxious counterparts.

What explains these intuitions? For Enoch, the answer relies on the agents’ higher-order desires. Even though both the unconscious anxious patient and the unconscious Christian Scientist would have refused treatment had they been conscious, only the latter would refuse for reasons she endorses with higher-order desires. Administering the transfusion to either Christian Scientist would be a violation of self in a way that administering the transfusion to the anxious patients would not. This explanation, covering both actual consent and hypothetical consent, explains one way in which both types of consent matter: they matter because they prevent treatment which would be “an assault on [the consenter’s] self” and enable us to treat them in a way “they identify with.” (p. 25) Enoch refers to this autonomy-based reason as “non-alienation.”

However, it is important to note that this reason only applies to a subset of hypothetical consents. Enoch points to sexual interactions as one possible limitation. Yet, there is another type as well. Recall the distinction between hypothetical rational consent and hypothetical individualized consent. It seems possible that consent can be so rationalized and abstracted from the individual that she won’t recognize the consent as being reflective of any of her desires. For example, it seems natural to wonder what exactly an agent deciding from behind the veil of ignorance has in common with me. Because of this concern, Enoch’s non-alienation defense of hypothetical consent should be limited to those types of hypothetical consents where the “consenter” actually bears a recognizable resemblance to the agent.

Non-alienation is only part of the story. Actual consent, according to Enoch, also matters for sovereignty-based reasons. There are some contexts, such as whether one has salt at dinner, that are within one’s area of control. Enoch’s daughter’s denial to pass the salt to him doesn’t threaten his deep commitments, but it is nevertheless an assault to his autonomy. It seems plausible that actual consent can promote this form of autonomy – autonomy qua sovereignty. (Enoch doesn’t offer a full treatment of this concept, rightly noting that it is best left to a future work.) Yet, according to Enoch, hypothetical consent fails to be even a “pale form of consent” when measured with this metric.

I grant Enoch this point when hypothetical consent threatens to replace actual consent. Whether he would have asked for the salt shaker under idealized conditions is irrelevant when he is actually asking for it now. However, I suspect that sovereignty concerns can do more work for hypothetical consent than Enoch allows for when actual consent is not possible. Hypothetical consent can allow a person to control what happens to her body and interests when she cannot expressly alter the moral landscape surrounding her. When we give this consent normative force, we thereby recognize that she has the power to create these normative upshots and are responsive to this power. Hypothetical consent may thus enable us to recognize and respond to an agent’s autonomy qua sovereignty instead of treating her as a non-autonomous moral patient (say by adopting a purely welfare view) when she unable to give actual consent.

This conditional consent does not always have to reduce to our deep commitments but instead can be respected purely because of sovereignty-based concerns. Consider Enoch’s Unconscious Weak-Willed Christian Scientist. In this case, the patient would have consented to the transfusion had she been conscious even though doing so conflicts with her higher-order desires. Here, Enoch is not sure what to say, other than such a case is different from the non-weak-willed counterparts. Yet, if we acknowledge that agents have the right to give consent even in contradiction to their higher-order desires, and that such a power is normatively important, we may be able to explain why hypothetical consent makes a difference here. It is because this patient would have exercised her sovereignty by consenting, even though such an act would violate her higher-order desires that makes a difference. True, she can’t actually exercise her normative power now, but by taking into account how she would have exercised it, we are taking into account her ability to act as sovereign over this area of her life. When we turn to the unconscious anxious patient, I suspect any unease we have with sovereignty concerns can be handled by asking if such a patient would consent to the transfusion given that she would be unconscious. If the answer to this query remains no, then I no longer share the intuition that such a procedure would be significantly more permissible than one on an unconscious Christian Scientist. What these cases show, I think, is that the various forms of autonomy may be in tension with themselves, rendering these cases difficult for reasons beyond those relating to hypothetical consent.

However, regardless of whether one is convinced by my attempts to make a place for hypothetical consent in contexts where we are concerned about sovereignty, Enoch’s case for the relevance of hypothetical consent for non-alienation remains sufficient to reassert the importance of this practice.

31 Responses to Ethics Discussion at PEA Soup: David Enoch’s “Hypothetical Consent and the Value(s) of Autonomy,” with a critical précis by Beth Valentine

  1. David Enoch says:

    Thanks, Beth, for starting off the discussion. Your overview of my paper is so well-done that when I read it, it wasn’t entirely clear to me why it took me so many pages to say just that.
    I find the distinction between hypothetical rational consent and hypothetical individualized consent both relevant and important (and I have to say I wasn’t aware of Vandeveer’s relevant discussion, so thanks for the reference). And I agree with most of what you say in applying the distinction – in particular, that some of what I argue for works better for one of these than the other. Still, I want to insist on the following point: I don’t see here a dichotomy, but rather a huge variety of (possible) hypothetical conditions. The term “hypothetical consent” is thus always incomplete – it refers to consent under some hypothetical conditions, but doesn’t (yet) specify which. The relevant conditions can be more or less demanding rationally, they can be more or less individualized, and they can presumably vary along other dimensions as well. I think that the points you’re making utilizing the seemingly dichotomous distinction can be revised with this correction in mind – for instance, the more individualized the conditions, the more adequate the hypothetical consent to deal with alienation, and so on.
    Another point of yours which emphasizes a place where I don’t deal with all the complexity of different hypothetical conditions is the thought that sometimes hypothetical consent can do sovereignty work as well. I think what you have in mind are cases where someone cannot express their actual consent (or lack thereof), where the context is one where sovereignty is very important (say, about body integrity), and we can reasonably find out whether the person would have consented had she been able to express her will. I agree that this may be a case in which hypothetical consent may play a role, despite the value at stake being (perhaps among others) that of sovereignty. But I think that this is a special feature of these specific hypothetical conditions, and so not something that generalizes.
    I think your analysis of the case of the unconscious weak-willed Christian Scientist is especially important. I haven’t quite thought about it in those terms before reading your comment, but I fully agree that we sometimes have the normative power, and indeed the right, to give or withhold consent even in ways that go against our deep commitments. (This actually reinforces, I think, my distinction between autonomy as non-alienation and autonomy as sovereignty.) But I’m not entirely sure about the specific case, because it seems to me the discussion of sovereignty here is somewhat contaminated by our (or anyway, my) independent conviction that the relevant deep commitments are both deeply mistaken, and, if acted on, tragically harmful. Perhaps we should think of cases that don’t have this feature – where our view of the relevant deep commitment is either neutral or positive. What of, say, the Unconscious Weak-Willed Life-Loving Anxious person? He wants to go on living, and his deepest commitments (to projects and people) would lead him to consenting to the blood transfusion in a second. He would even do all he can to overcome his needle-anxiety. But being weak-willed, he would fail. Luckily, though, he’s unconscious. It seems to me permissible (maybe even obligatory) to administer the transfusion. This may be because other values trump here the sovereignty value (after all, had he been conscious and able to express himself, he would have withheld consent). But my intuition here is that there’s no serious loss in sovereignty that we need to compensate for. If so, it’s still not clear to me what exactly the case of the weak-willed Christian Scientist shows.
    Oh, and I fully agree that some of these cases show a tension between different forms of the value of autonomy, in ways that make some of the issues raised here interesting also outside the context of hypothetical consent.

  2. I’m delighted to be invited to chime in over the next few days with comments on David Enoch’s penetrating paper (I will call David “Enoch” and would not mind being called “Estlund” to avoid confusions about which of the Davids is being mentioned as we go forward.) And thanks to Beth for the excellent précis and comments.

    I’d like to look more closely at the “transitivity argument,” which plays an important role in the paper. To be clear, this is not Enoch’s argument. Rather, he emphasizes the limits of this argument as an objection to the normative relevance of hypothetical consent. I’ll come back to his own limited objection to it. I want to start by explaining why I do not understand the argument as stated.

    Here is the argument, along with some simple abbreviations that I hope will be efficient:

    1. The Normative Upshot (NU) holds in virtue of hypothetical consent (HC).

    2. Hypothetical consent (HC) would be given in virtue of the Underlying
    Reasons (UR).

    3. The in-virtue-of relation is transitive.

    4. Therefore, the Normative Upshot (NU) holds in virtue of the Underlying
    Reasons (UR) (from (1), (2), and (3)).

    5. Therefore, hypothetical consent (HC) does no normative work in justifying
    the Normative Upshot (NU) (from (4)).

    (1) says “NU holds in virtue of HC.” But then the conclusion (5) is that “HC does no normative work in justifying the NU.” One might have thought that a normative upshot’s holding in virtue of some factor x is just what it means for x to do normative work. “Murder is wrong in virtue of x” would seem to point to some normative work that x does. But that would make (1) and (4) inconsistent. That is, if the hypothetical consent view under discussion is the claim that hypothetical consent does some normative work in justifying certain normative upshots, then unless “in virtue of” and “doing normative work” are somehow distinguished the hypothetical consent position is granted in the first premise, (1), but denied in the conclusion (5).

    Here’s a different but related point about the argument: We have two premises, each giving a different factor that NU is said to hold “in virtue of,” (namely HC in (1) and UR in (3)) and then out of nowhere one of the two factors, HC, is said to “do no normative work” but not the other. Even if we were told how doing normative work is different from being what some normative thing holds in virtue of, it is not explained why the two “in virtue of” cases are treated so differently.

    I’ll stop here, and move to a second post to briefly consider the idea that transitivity may seem to show the underlying reasons to be explanatorily more fundamental than the hypothetical consent, even if both play some explanatory role (still contrary to the transitivity argument’s conclusion).

  3. Following on from my previous post: Why would the fact that HC obtains in virtue of something else, UR, undermine or diminish HC’s explanatory relevance to a given normative upshot (NU)?

    I conjecture that it might seem natural that the distance, as it were, between terms in this transitive relation “in virtue of” marks deeper, or more fundamental, or more explanatory elements. Suppose that claim were itself added as a premise. We would want to consider whether to believe it. But first, even if it were granted it would not support the conclusion (5). It would warrant only a conclusion something like “HC does not do as much normative work as UR in justifying the NU.” That would be important if we were inquiring into the strong claim that HC does all the work, or that UR does no work, or something like that. But I myself think, and here I seem to agree with Enoch, that the fundamental question about hypothetical consent is the one reflected in the original (but insufficiently supported) conclusion (5): Does hypothetical consent do any normative work in justifying certain normative upshots? Is it part of the explanation or not part of it? The additional premise about explanatory depth would not support the conclusion that it is not part of it.

    Second, should we accept that extra premise I have sketched about transitivity and explanatory depth or basic-ness (a premise I have brought up, not Enoch, but as a conjecture about how the argument is meant to cast any aspersions on hypothetical consent). The premise, again, would be that the distance, as it were, between terms in the transitive relation “in virtue of” takes us to deeper, or more fundamental, or more explanatory elements. Maybe that’s so, but it’s not at all clear to me. First, plenty of transitive relations weaken rather than strengthening the binary relation in question. For example: A is descended from B; B is descended from C; So A is descended from C (as well as from A). But A is not more strongly (or whatever) descended from C, quite the contrary. Intuitively, I am more weakly descended from my ancient forebears than from my parents. Now we haven’t considered any account of what kind of “strength” is at issue here, so we don’t know whether the descent case is relevant, but it raises the question pointedly I think. In what sense of “more explanatory” or “more basic” would the transitivity relation support in the case of a fact’s holding in virtue of something? Of course, Enoch hasn’t made the argument from basicness here, and I raise it only as a conjecture about how the explanatory relevance of HC is supposed to be somehow attenuated by the transitivity argument. And to recall, even if hypothetical consent were shown in this way to be less fully explanatory than something else, this would not support the suggestion that hypothetical consent is or is not part of the explanation for certain normative upshots.

    A final note about the transitivity argument: Enoch usefully deploys the possibility of multiple realizability to deny that (4) implies (5), which is an important challenge to the argument. But, to repeat my leading claim, I don’t know how (5) is compatible with (1) in the first place. The way I have proposed of defending HC’s role against the transitivity argument does not need to rely on the presence of multiple realizability and so would apply more generally.

    In wondering how the transitivity argument is supposed to be understood I am not directly disagreeing with anything Enoch says. However, he does claim that in light of the argument a role for hypothetical consent could only be defended in cases where multiple realizability obtains. To support that, he would need to define the different relevant meanings of “in virtue of” and “normative work” so as to render the argument sound except in cases of multiple realizability.

  4. Beth Valentine says:

    Thanks to both Davids for the great comments!

    I’m interested to see how Enoch replies to Estlund’s point. My initial thought is that there is an equivocation of the “in virtue of” relationship, such that one is normative and the other is non-normative. Not all “in virtue of” relationships indicate normative significance, so one way to resolve the tension between (1) and (5) is to claim that the relationship in (1) is not normative. [This would be a reversal of the equivocation worry I raise in the précis.] However, this move would invite complaints of circularity and need further defense. Another potential response would be to accept that all the relationships in the argument are normative but to amend (5) to being a claim about non-derivate normative work. After all, the no real work objection is that HC doesn’t do any real work, not that it doesn’t do any work. Admitting that HC does derivative normative work, in virtue of its relationship to UR, would still – I think – capture the spirit of the Transitivity argument.

    Regarding Enoch’s points, I’m grateful to have an opportunity to think more about the different types of HC, and hopefully one of the upshots of these debates is recognizing that a one-size-fits all approach might not work. Viewing the term HC as always incomplete (or perhaps under-descriptive?) is appealing and raises questions of how best to proceed when trying to say something about HC. Given the variety of options, I’m not sure that we can expect a unifying account of why or even if HC – as a broad category – matters. My hope is that we can impose relevant subcategories on the phenomena (e.g. individualized to commitments, desires, etc. but not to current conscious state or information) to gain more clarity on why it sometimes matters and sometimes doesn’t. These categories need not be a strict dichotomy, but currently I’m suspicious that speaking of HC generally will garner much meaningful progress.

    As for the Unconscious Weak-Willed Christian Scientist and the Unconscious Weak-Willed Life-Loving Anxious person, I would be interested to see what intuitions others have. Perhaps I place too strong a value on sovereignty, but I am inclined to say that if the anxious person would not give consent to X, where X indicates “giving me a blood transfusion while I am unconscious,” then we have a strong autonomy-based reason not to give her the transfusion.

  5. David Enoch says:

    Thank you both. (I will follow David Estlund’s suggestion and call him “Estlund”, and just for unity, I will call Beth Valentine “Valentine”.)

    How to understand the transitivity argument given the tension between (1) and (5)? Well, as a reductio of (1), at least in its most natural reading.

    Estlund asks interesting questions about transitivity in general, and the strength of the relation whose transitivity is invoked. (Recall the plausible claim that, in my words, the ancestor-relation gets diluted over many transitivity moves, not strengthened). I’m not sure I can put the following point in non-metaphorical terms, but for whatever it’s worth: The issue I had in mind here (and the one that’s there, I think, when people in the literature invoke similar worries) is not one of *strength*, but of *subsumption*: The thought is that whatever work you may have thought HC does, it’s entirely subsumed by the work being done by the UR.
    We can see this structure rather clearly in (other) examples. Suppose I put forward a normative theory, according to which we always have reason to do whatever Dave tells us to do. Suppose I then claim that Dave always and necessarily says we should maximize utility. And suppose, further, that I claim that we should always do what Dave says in virtue of the fact that he always directs us to maximize utility. Here a transitivity move seems to work nicely – really, Dave drops out of the picture. Whatever normative work his directions were supposed to do is entirely subsumed by the normative work already done by the underlying principle of utility. I thought the structure here is similar.
    (Of course, with this example too, things can get complicated; If Dave always but contingently says we should maximize utility, a multiple realizability line begins to emerge.)

    I agree with Valentine that if there’s an equivocation on the relevant kind of in-virtue-of, the argument may fail. The advantage of the multiple realizability line is that it may show how the argument fails even in non-equivocation cases, which are sometimes important. (For instance, I think, vis-a-vis Estlund’s theory of normative consent).

    And it *is* interesting that we seem to differ on whether there’s an autonomy-related reason not to administer the transfusion to the weak-willed life-loving anxious! I have to think more about that one.

  6. Nir Eyal says:

    Thanks for the opportunity to join the discussion of this incredibly-well written paper. I am in sympathy with most everything in it, but let me raise two challenges here. One concerns the Weak-Willed Christian Scientist (pp. 29-30), which Valentine and Estlund have already discussed. (BTW Christian Science doesn’t exactly forbid blood transfusion—we are just assuming that for our philosophical purposes.) A clean way of asking this question imagines that the agent deliberating whether to administer blood transfusion knows that no one would ever know or mind that it was administered—so no resentment, hurt feelings, religious wars, loss of trust in physicians, penalties, and so forth.

    Whether the Weak-Willed Christian Scientist is conscious or unconscious, my intuition is that transfusion is permissible in such cases, hands down. Left to decide on his own, the guy wouldn’t die, and it seems crazy to lead to his death out of some idea about his autonomy and when there would be no repercussions. This is a challenge for Enoch, because together, his two autonomy-related concerns seem to suggest otherwise. First, whether this Scientist is conscious or unconscious, nonalienation seems to count against administering life-saving transfusion. After all, transfusion would not gain this person’s consent if, hypothetically, his religious core, stripped of his life-saving weak will, were interviewed (see my “Paternalism, French fries and the weak-willed Witness”, Journal of Medical Ethics 2013). Nonetheless, in the variant where the Weak-Willed Christian Scientist is conscious and begging for transfusion, Enoch rightly concedes, “it seems to me clear that we should administer the transfusion” (p. 29). Enoch seeks to explain this: “that this is so may be the result of factors that are irrelevant for our discussion (perhaps that Christian Science doctrine is wrong, that the reasons for him to go on living are very good ones, or perhaps because of the value of sovereignty…)” (Ibid.). But the explanation fails. Recall that in the standard Conscious Christian Scientist case all agree that transfusion would be highly problematic, and perhaps wrong all things considers. This shows that nonalienation concerns defeat the first two considerations that Enoch’s explanation lists, namely, that Christian Science doctrine is wrong and that saving this person’s presumably good life matters (so perhaps no need for our blog exchange to shift to discussing the weak-willed life-loving anxious person). It is true that sovereignty has great potential force, as illustrated by the Conscious Anxious Patient case. But recall that the intuition on the Conscious Weak-Willed Christian Scientist is not only that the he should get transfusion. It is that there is no question about it; that health as well as any applicable reasonable approach to autonomy all side with blood transfusion in this variant of the case. By contrast, on Enoch’s dual theory of the function of consent, one element (namely, nonalienation) counts strongly against administering transfusion to the Conscious Weak-Willed Christian Scientist; there is allegedly a dilemma or a very hard case. In short, moral intuition seems to endorse administering transfusion more wholeheartedly in this variant than Enoch’s theory permits.

    In the variant in which the Weak-Willed Christian Scientist is unconscious, nonalienation again counts against administering transfusion. Unfortunately, Enoch’s theory seems committed to adding that sovereignty cannot defeat that. in this variant, sovereignty cannot count for or against blood transfusion. It is inapplicable because the person cannot express refusal of transfusion. So Enoch’s theory would seem to suggest that in the unconscious variant of the Weak-Willed Christian Scientist, clearly we should not administer transfusion. This suggestion is counterintuitive as well, in that the intuition is that matters are at least more complex than that. Enoch appears to accept this complexity: “which verdict is more intuitively plausible? I have to admit that this doesn’t seem clear to me” (p. 30). Again, then, moral intuition is more on the side of administering transfusion than Enoch’s theory permits. He may attempt to address this by distinguishing the impact of deep commitment from that of hypothetical consent in this variant of the case (Ibid.). But I am not sure I understand his response because throughout the article, his concrete standard of hypothetical consent is focused specifically on those “hypothetical conditions that filter out [the person’s] superficial desires and emphasize his deep commitments…” (p. 31, and see also p. 27 and elsewhere).

    A second challenge surrounds Enoch’s paternalistic daughter, a simple but quite masterful example to illustrate sovereignty concerns (pp. 31-32). In the case, Enoch’s daughter, seeking to force him to fulfil his core self’s endorsed desire to retain his good health, is refusing to pass him the salt during a meal. While such paternalistic treatment might turn out to be justified all things considered, depending e.g. on how dangerous salt is for Enoch, clearly there is something to be said against this daughter’s conduct. And what is problematic about it cannot be thwarting Enoch’s core desire but thwarting his superficial desire, as well as its superficial owner’s apparent right to some control. By contrast, most examples on related issues in the literature involve a violation of bodily integrity. Notwithstanding Enoch’s admirably light touch, let me point out a complication.

    Suppose, heaven forbid, that immediately after several such interactions with his daughter, Enoch, as the result of a stroke, becomes unable to speak or otherwise communicate his mind. If Enoch’s daughter then does not pass him the salt during meals on the ground that she was not asked to pass it and hence was not violating sovereignty, surely we would still complain of some autonomy-related moral problem in the way she treats him. After all, she has every reason to suspect that if Enoch could still speak, given his well-known savory tooth he would ask for salt. She is exploiting his paralysis to force her will (and his health and nonalienation). The daughter’s actions in this variant do not strike me as much less problematic than in the original variant.

    If the latter point is right, one possible lesson is that the article is too pessimistic about the place of hypothetical consent within the sovereignty concern. The sovereignty concern opposes not only forcing things that people actually refuse. It sometimes opposes forcing things that people would refuse in a hypothetical situation in which they would be free to express their minds, free from paralysis, unconsciousness, and the like. Admittedly, this kind of hypothetical situation is neither the condition of full rationality that Valentine mentions (and which Enoch rejects on p. 24 yet expresses openness to in a post above) nor the condition of nonalienation that the article expounds. It is a third hypothetical condition.

  7. [For simplicity, I will here use “explanation,” “in virtue of,” and “relevant” in interdefinable ways that should, after my saying this, be self-explanatory.]

    Some thoughts, in two related posts (again) on Enoch’s response:

    The issue at the moment is whether the transitivity argument undermines all relevance of hypothetical consent unless there is multiple realizability. I am doubtful. Enoch is not arguing that it does this by itself, but does so along with some supplementary ideas to which he seems to be sympathetic. He says that the contradiction between (1) and (5) can serve as a reductio of (1) if he can avail himself of an additional idea concerning “subsumption.” First, let’s be clear about why there’s no reductio of (1) without the appeal to subsumption.

    Here’s the form of the argument, abbreviated:

    1. NU holds in virtue of HC
    2. HC holds in virtue of UR
    3. Transitivity
    4. Therefore, NU holds in virtue of UR (from 1-3)
    5. Therefore, HC does no normative work in justifying NU (from 4)

    Enoch’s reply to my question about the tension between (1) and (5) grants that they are inconsistent. The idea of normative work is not, then, intended to be a different notion from that “in virtue of” relation as it applies here. So, to make that clear, we can now rephrase (5) as:

    5’. Therefore, NU is not at all (not even partly) in virtue of HC.

    I will come back to this hint that “in virtue” might hold “partly.”

    Let’s suppose that we know something extra such that (5) indeed followed from (4) (which it clearly doesn’t without something extra). This extra thing might be that “in virtue of” is equivalent, here or perhaps always, to “exclusively in virtue of.” But then we don’t need to get to (5) in order to encounter a contradiction: (4), read in that “exclusive” way, is inconsistent with (1). So, we can now ask what this contradiction might be used to show by reductio.

    What we know is that the combination of (1)-(3) implies a contradiction on the exclusive reading of the relation. This does not, then, refute (1), since it would not generate a contradiction but for (2) and (3) AND the exclusive reading of “in virtue of”. What we know is only that one of those four things must be mistaken.

    Notice that even if extra argument were given to put the blame on (1), reading it in the exclusive way, this is so far no problem for the hypothetical consent position as I understand it (along with Enoch, I think), namely the claim that NU holds at least partly in virtue of HC. So, even if it could be shown that the exclusive version of (1) leads to trouble, the interest of this is small since that exclusive claim is not a commitment of the HC position in question.

    So next let’s consider the “non-exclusive” possibility that “in virtue of” is equivalent to “partly or wholly in virtue of.” Now (4) no longer contradicts (1), they simply name two things that NU is at least partly in virtue of. But (5’) (our clarification, following Enoch’s reply, of (5)) now patently does not follow from (4), since nothing has ruled out the possibility that NU is partly in virtue of HC, which (5’) denies.

    It might be best to break here before turning to subsumption in my next post.

  8. Following on my post from a minute ago:

    This brings us to subsumption. To rule out the possibility that NU is partly in virtue of HC, Enoch considers the possibility that the relevance of HC is “subsumed” by the way (that is, exclusive) which HC holds in virtue of UR. What we would need is an argument for this:

    Subsumption Thesis: IF HC holds (wholly or partly?) in virtue of UR then NU does not hold even partly in virtue of HC.

    Roughly, this states, so far without argument, that the relevance to NU of UR “subsumes” and removes any alleged relevance of HC. So, it all hangs on the subsumption thesis. So let’s turn to assessing that claim directly (which is no part of the transitivity argument, but could be combined with it in a certain way).

    Being hypothetical, hypothetical consent is not an actual event. It states a relation between a certain counterfactual case of consent on one hand, and certain reasons on the basis of which that consent would be required. So hypothetical consent theories, according to which hypothetical consent is part of the explanation of certain normative upshots, are all about underlying reasons.

    Now, you might say, aha! If it is “all” about underlying reasons then it is not about consent at all! Not so fast. What I mean is that it is all about underlying reasons to consent. The question is not whether some act of consent to X gets “subsumed” by certain underlying reasons to X. There is no (actual) act of consent to X under discussion. The claim that NU holds at least partly in virtue of the fact that S would, if rational, consent to something, Y, can be understood as the claim that NU holds in virtue of the reasons that S would have to consent to Y. There is no actual consent in the story, but only reasons to consent. But then the claim that the underlying reasons obviate any relevance of hypothetical consent is difficult to interpret. That is roughly like saying that the underlying reasons of the agent’s to consent obviate any relevance of the reasons the agent would have to consent. Hypothetical consent is just the reasons to consent, but that requires essential reference to consent. And it is not any actual consent, but the consent that would, hypothetically be required by the reasons: hypothetical consent.

    If hypothetical consent were to be dropped from the story, as opponents claim it ought to be, then the underlying reasons must not be reasons to consent. What reasons are they in that case, and whose? Such a view may have in mind reasons there are for some other agent, T, such that T is justified (permitted, required, not permitted, etc.) in some normative upshot about doing something to S, or whatever. But this change to reasons for that other thing just changes the subject, and says nothing for or against the relevance of hypothetical consent. No argument has been given that there are such direct justifying reasons for T that need not make any reference to the reasons S would have to consent. Of course there might sometimes be such direct reasons for T, and whether there are might vary by case, but that is not a question about hypothetical consent and whether it is sometimes partly explanatory.

    I take it that the version of subsumption that must be asserted here is something like this:

    Bypass: In any case (or some defined set of cases) where a normative upshot is plausible in light of the reasons S would have to consent to it, the same facts that are the requiring reasons for S will also be—directly and without essential reference to S’s thereby having reasons to consent—reasons establishing the normative upshot.

    We would want some argument for this claim. The point here is that the fact that, as I have put it, hypothetical consent is all about the underlying reasons to consent, is no support for this claim.

    I think the transitivity argument equivocates between reasons for S to consent and reasons for T justifying the normative upshot. Bypass clearly distinguishes them, but stands in need of support. I do not say it cannot be supported, but only that, if I’m right, the transitivity argument drops out of the picture as confused.

  9. David Enoch says:

    Thanks, Nir.
    Let me start with the weak willed Christian Scientist.
    I agree with much of what you say but not quite all.
    You say that the case of the unconscious weak-willed Christian Scientist is an easy one. I agree. But you seem to suggest that in it there’s nothing to be said for not administering the transfusion. And this I don’t accept. I think that the non-alienation concerns still have normative weight here. Perhaps very little weight – but it’s not as if all relevant considerations count in favor of administering the transfusion. Not sure how much this affects your other points.
    But here’s the more important point here, perhaps explaining how – consistently with all I say in the paper – I can be as accommodating in this case (supporting transfusion) as you want me to be:
    What is going on here, I think – both normatively, and psychologically, as exaplaining our responses – is that we *really* want to save him. The (made up for our purposes) doctrine is false and harmful, the life he can lead is a good one, we can’t believe he’s going to refuse the transfusion. We’re just looking, as it were, for a way in. Any way that can make the transfusion permissible – even borderline permissible – and we’re going for it. In the case of the conscious (strong-willed) Christian Scientist, perhaps we don’t have such a way in. (Not entirely sure, but let’s assume so). In the others we do, and then, the full weight of saving the life kicks in, perhaps crowding the normative weight.
    The upshot, I think, is that your reasons-arithmetics fail. The reason interaction here is not just that of balancing. SOmething about autonomy in the case of the conscious strong-willed CS excludes the full weight of the saving-the-life considerations. Once that exclusion is out (in the cases of both the conscious and the unconscious weak-willed CS), that value triumphs,
    Or so it seems to me.

  10. David Enoch says:

    Now, my daughter.
    I like the push that I’m getting from both Nir Eyal and Beth Valentine (the name thing is killing me!) regarding a possible role for hypothetical consent vis-a-vis sovereignty. I did not expect it. So this is nice, and I’m not entirely sure what I want to say about that.
    But I don’t think that the example here works. In the example you mention, your intuition (which I share) that not passing the salt is just as problematic, is motivated, I think, by the fact that we have very strong evidence of *actual* will. The point is not, it seems to me, what I would have said had I been able to speak; the point is rather about what I already, fully actually, want. If you know this – and in the example my daughter does – then the case is indeed very much like the original one.
    Notice, by the way, that there’s some similarity between this case and the point I make in my previous comment: Exploiting the circumstances in order to (permissibly) do good (saving the life of the CS or of the person who really likes salt). I think that sometimes taking advantages of the circumstances in such a way is both admirably benevolent and admirably clever.

  11. David Enoch says:

    Thanks, David Estlund, for the detailed examination of the transitivity argument. Let’s see.
    Note that in my response re the reductio reading, I say something like “at least in the most natural reading of (1)”. That most natural reading is one that attributes real normative work to HC. This is the claim that (5) denies.
    You’re right that the relevant sense of in-virtue-of is the partly-or-wholly one.
    And you’re also right that the initial plausibility of the argument, even before the multiple realizability response, depends on the subsumption point. (As do, I think, similar worries expressed in the literature in the past; don’t you agree?)

    I was surprised, though, to see you assert that when HC matters it *is* just the reasons for consent that matter. I mean, I agree, of course, that reasons for consenting need not be the same as reasons for other things. And so I agree that the concept of consent may be needed here – in order to fix the reference, as it were, to the relevant reasons. But still, then it’s the reasons (for consent), not the consent, that does the work in justifying NU. It now seems to me that you accept this.

    This looks to me like an important concession. Again, you’re right (of course) that HC is not actual, there is no such event, etc. ANd you seem to suggest that for this reason, the role you let HC play is the only role it is at all plausible to expect it to play. But this seems to me false. First, some people seem to want it to play a role closer to that of actual consent, and it doesn’t. Second, if the multiple realizability line works, it shows that *there is* a more serious explanatory role that HC can play, no?

  12. Replying to Enoch 4:45pm:
    (We’re focusing on the case of individual rational hypothetical consent, which is the one where it is claimed that the underlying reasons do all the work.) I argued that an appeal to hypothetical consent in this case is just an appeal to the fact that reasons would require consent. You write, “then it’s the reasons (for consent), not the consent, that does the work…” And you think this would be a significant concession. What does “the consent” refer to? What would it mean to hold the view you mean to oppose—that “the hypothetical consent” does the work, and not only the fact that the agents reasons would require consent?

    Maybe you have this in mind: What “does the work” is not that they would consent if rational, but merely the fact that they would be rationally required to do so. But what is the difference between those? We can’t say that the latter is not about consent but only about the reasons. That it is about consent because it is about reasons to consent. (As I go on to say, if there could be shown to be reasons directly for the other person T (such as the transfusion professional), then no reference to consent may be necessary, but the transitivity argument doesn’t provide any support for that in general as you seem to agree.) And, as you also see, obviously the view never meant to say the NU is supported by consent—that would be an actual consent view.

    As I think you are aware, I am not here taking up the question whether this would be a fact about the agent’s own will, that’s different. I’m arguing that appeal to the concept of consent does not get subsumed by appeal to underlying reasons.

    So I think that there is no clear issue about whether “the consent” that S would give if rational does any work. There is an issue about whether NU can be justified without reference to the consent that would rationally be required. If reference to consent in that way “does work” then it is hypothetical consent. You write just now, “I agree that the concept of consent may be needed here – in order to fix the reference, as it were, to the relevant reasons.” If that much is granted, then I think the relevance of hypothetical consent has been granted, the relevance of the fact that the agent would be rationally required to consent.

  13. David Enoch says:

    Thanks, David (Estlund).
    You conclude by saying that if the *concept* of consent is necessary, then “the relevance of hypothetical consent has been granted”. This seems to me clearly false, at least in the way questions about hypothetical consent have been understood in the literature.
    Of course, nothing (except perhaps “who won”…) hangs on how we use such locutions as “the relevance of hypothetical consent”. The crucial point is that if this is the only way in which we need to refer to the concept of consent, then HC is not relevant in many ways in which many have thought that it was: for instance, as something at all related to actual consent, or in the way my multiple realizability line makes HC normatively relevant. (I’m repeating here something from my previous comment to which you haven’t replied: IT’s not as if the relevance you supply is the only relevance to be had here).

  14. Yes (David Enoch), that’s right, there are indeed various claims that might be made for hypothetical consent. I’m not addressing the whole field. I’m focusing on the argument you make, in a similar line of thought to Thomson and others, that in the case of individual rational hypothetical consent, hypothetical consent drops out (unless there’s multiple realizability)—that it is not something (at least partly) in virtue of which the normative upshot is ever justified. I’ve argued that these arguments fail. I think you now agree, or agree that this “may” be so. Hypothetical consent, being no more than the fact that consent would be rationally required, is something (at least partly) in virtue of which a normative upshot can obtain. If so, the relevance of hypothetical consent is granted in those cases. By “the” relevance, I don’t mean any other relevance than this, but leaving out “the” is weird English. It might have sounded as if I think this would be an answer to “the” question about HC’s moral importance. Not at all, there is no such single question. Other questions about the relevance of hypothetical consent certainly remain, such as whether in some cases HC is morally relevant in order to avoid what you call “alienation,” and whether it is ever relevant in some kind of proximity or similarity to actual exercise of the will, related to “sovereignty.” What I’ve said so far doesn’t bear on those, and I didn’t mean to suggest otherwise. Still, finding a normative upshot to obtain partly in virtue of hypothetical consent in the “rational” case does seem to bear on *one* of the central questions you’re addressing, and, anyway, one that has been disputed: hypothetical consent does not get “subsumed” as claimed. Anyway, it may be that we now agree about this and also agree that it leaves plenty of questions, some of which you also take up.

  15. David Enoch says:

    This is helpful.
    But no, I don’t agree that you’ve shown that those arguments fail. Note the shift in whaty ou say between talk of *some* role for HC, and an *in-virtue-of* role for HC. What I’ve agreed may (!) be true is that there’s some role for the *concept* of consent, in *fixing the reference* of the relevant reasons that do the in-virtue-of work. From this it most certainly doesn’t follow that the normative upshot holds (even partly) in virtue of HC.

    (And for the record, let me place this again in the dialectic of my paper: As you note at the beginning of this exchange, of course I agree that HC may sometimes do actual work – that’s the line of multiple realizability. The disagreement between us is about such work even in the absence of such multiple realizability.)

  16. Nir Eyal says:

    Enoch,

    I. You’re correct to correct my initial, triumphant declaration that there is nothing to be said for nontranfusion in the Unconscious Weak-Willed Christian Scientist variant. Note though that my main point about this variant did not assume this hyperbole. It assumed only that “the intuition is that matters are at least more complex than” judging that this person should definitely be refused transfusion. On the latter point, we seem to see eye to eye. We both see some very limited pull toward nontransfusion. To me, however, this very limited pull is likelier to reflect moral uncertainty on our parts than a first-order pro tanto reason to refuse transfusion.

    II. You made a very fruitful recommendation to handle deliberations on consent to blood transfusion in the following way. If I understand your point, we are and should be looking for an “in” to save the person’s life. Sometimes, we are simply blocked, or at least must acknowledge strong pro tanto reasons not to save his life, e.g. when a decisionally-capacitated Scientist says No to blood transfusion. Nonetheless, we are right in general to use any sufficient “moral loophole” (you might not use this term) to administer life-saving blood transfusion, whenever such a loophole exists.

    Let me explore this intriguing suggestion further.

    a. It seems especially fruitful in contexts where facilitating the patient’s personal autonomy is not our goal but, instead, functions as a constraint on our actions. Am thinking about important medical research on human participants, or important organ and tissue donation. The proper goal of these practices is clearly not fulfilling the participant’s or the donor’s core self, yet they are subject to certain moral constraints. Inter alia they usually cannot violate their consent rights. The most influential approach to consent across bioethics emphasizes realization of personal autonomy. Perhaps, however, Enochism (in Hebrew: Enochiyoot) is to be preferred in e.g. medical research. We should lucidly explain and re-explain details about the trial to willing candidate participants. But if they still don’t get it, and if other safeguards ensure that how the trial treats them is otherwise kosher, we can go ahead and rely on their consent, noncomprehending as it may (compare Gopal Sreenivasan, “Does informed consent to research require comprehension?”, Lancet 2003). Shifting to Enochism in these areas of practice would be big. Medical research is an area where consent practices are especially stringent.
    b. Your suggestion may also help account for the permissibility of certain nudges. Take the nudge of changing cafeteria users’ defeasible defaults from picking pizza to picking salad, e.g. through changes to cafeteria architecture. Even if a young customer simply couldn’t care less about how they would fare when they are old, such that this nudge wouldn’t be nonalientating, and even if there is something manipulative about this nudge such that it transgresses sovereignty a bit, intuitively, this nudge is eminently permissible. Why? A possible account would be that in this setting, autonomy-based reasons kick into action only once they become strong enough that they can no longer be ignored. Maximizing personal autonomy is not a goal of nutrition-informed cafeteria planning. Rather, that practice’s pursuit of its proper goals shouldn’t do too much violence to personal autonomy.
    c. That said, arguably in some contexts, promoting autonomy is among our goals. Enochism should not extend to these contexts.
    d. It also remains the case that this is about how to decide what to do, not about what to do.

    III. Finally, your daughter. What element is responsible for our intuitions in the Paralyzed Enoch variant? Enoch’s hypothetical request to have salt (as I suggested), or his known actual desire to have salt (as you suggested)? Perhaps a way to test whether any responsibility is borne by the hypothetical request is to imagine a third variant. In Laconic Enoch, your daughter knows that you like and desire salt, but she also knows that it is a core value of yours not to mince words. Brevity at all costs is a value that you both exercise and endorse. So she knows that, even stripped of your weak will, you wouldn’t ask her for salt—even though you would still crave it. She exploits this in order not to pass you the salt, hoping to remain off the moral hook for any transgression of your sovereignty rights.

    My intuition is that in Laconic Enoch, your daughter’s conduct is slightly less condemnable than in Paralyzed Enoch. (What do others’ think?) If so, it would seem as though Paralyzed Enoch involves transgression of your sovereignty and not only transgression of your alleged entitlement to the satisfaction of your actual desires. For what otherwise would make the daughter in Paralyzed Enoch worse?

  17. OK, maybe you don’t grant my whole argument, so I wonder where we part ways. My talk of the concept was responding to the idea that in that case it’s not about “the consent” but about the reasons. It’s about consent, hypothetical consent, which just is (I argue) the fact that consent would be rationally required. If that’s all right, then transitivity arguments fail if they purport to show that (except for multiple realizability) normative upshots do not obtain partly or wholly in virtue of hypothetical consent.

    It looks like my argument goes like this, and I wonder where you get off the boat:

    1. Hypothetical (rational) consent (HC) is nothing but the fact (F) (in certain cases) that consent would be rationally required

    2. Transitivity arguments do not refute that normative upshots (NU) can obtain partly in virtue of the fact (F) that consent would be rationally required.)

    3. Therefore, transitivity arguments do not refute that normative upshots (NU) can obtain partly in virtue of hypothetical consent (HC).

  18. David Enoch says:

    Thanks, Nir.
    I like the term moral loophole here – so long as it’s kept in mind that it’s not a case of looking for a loophole in order to, say, maximize self-profit, but in order to promote the good, or some such.
    I think that the nudging case is an excellent example.
    I don’t know enough about the medical research. It seems to me many factors are relevant here, and they interact in complicated ways. But yeah, I’m sure there are giong to be cases where this applies.
    Let me note, though, that I wasn’t making a general point about the value of life. I was specifically speaking of the Christian Scientist, whose life, we seem to assume, can go quite well. For that guy, any loophole that will make saving him permissible is a blessing, yes.

    To the extent that I understand talk of what our goal is, and its relation to autonomy, I don’t think that this loophole consideration only applies when autonomy is not our goal. For one thing, we may have to balance the autonomy value of many people, and then we may look for permissible ways around the autonomy of one to promote that of many others. For another, we may need to balance one person’s present autonomy with her future autonomy (many paternalism cases are like this.

    And I’m not sure I see why you think this is about how to decide what to do, not about what to do. I meant it as the latter – it’s a part of the story in virtue of which some actions are permissible, no?

    And I, for one, don’t share your intuition re “Laconic Enoch”. But it’s not just about desires – I do sense some sovereignty concern. The point is not just that I want the salt, or some such, but that I will the salt, and indeed will my daughter to pass me the salt. My act of will is fully present – only the words are missing, I think.

  19. David Enoch says:

    Estlund – I reject (2).
    The subsumption point, as far as I can see, stands threatened.

  20. David Enoch says:

    Contd, to Estlund:
    I may be lost in the dialectic now, but:
    We can talk about the reasons for the HC, and we can talk about the fact that there is HC. Perhaps these are extensionally equivalent, but there’s a Euthyphro contrast that’s important here: Presumably, Consent would be given (in the relevant, full rationality conditions) because of the reasons, not the other way around.
    So there’s no symmetry. (I’m not sure whether you were implying that there was).
    This is what open the door to the subsumption point.

  21. Nir Eyal says:

    Dear Enoch,

    Let me move on to other questions. Please feel free to address only some.

    1. When nonalienation and sovereignty generate duties that bind others than the potential consenter, are some of these duties positive, e.g. duties to promote nonalienation and promote fuller control? If so, can the latter duties be strong enough to override the negative duties, such that, in the name of nonalientation and/or sovereignty, we would be obligated to force things on people without their consent?

    2. Do you see your paternalistic daughter’s transgression as one of a negative duty, or of a positive duty?

    3. If the state banned manufacture or sale or purchase of very salty food products, based on evidence that such a ban would do marvels to public health, that would intuitively seem less problematic than your daughter’s refusal to pass you the salt. Some of the reasons for this are obvious and boring, e.g. the state can have better information on what would promote health effectively and sustainably. But some of the differences are ethically more interesting, and may speak to when sovereignty duties are stringent. Care to consider some?

    4. How much are duties generated by sovereignty conventional or otherwise non-universal? To illustrate what I have in mind, focus on the duty to pass you the salt when your superficial self requests it. One account says that it is a duty only insofar as acceding to your requests about such matters is a conventional mark of respect for your core/autonomous/metaphysical self (or any other self different from the superficial self), who is the true object of respect. A very different account says that merely analyzing some key concepts like self/respect/agent/… would reveal that passing you the salt upon your request is a duty.

    5. How much does consent that is unfree, uninformed, or incapacitated count as consent for your purposes?

    6. Are you saying that nonalienation and sovereignty are the only grounds for consent norms? If not, why did you focus on them?

  22. Enoch
    (I might as well stay in the third person since I started that way.)

    The Euthyphro framing suggests some concerns that I hadn’t been focusing on, namely whether there are some deeper metaethical stakes about voluntarism vs. realism. I don’t know for sure if these are part of David Enoch’s concern, but in any case there are a few puzzles about that.

    First, hypothetical rational consent does not let anything normative be settled at will, but only by what would be rationally required. What is rationally required is not up to anyone’s will. More on that in a moment.

    Second, even if that were not so, whatever else might be said against hypothetical consent, it would be weird to be troubled by the “at will” implications of hypothetical consent even as it is being granted that many important normative matters arise from actual consent—actual will. (Granting the moral importance of consent hardly refutes metaethical realism—it’s not at stake in that way.)

    Third, in the original Euthyphro context the issue was whether certain normative matters were settled (so to speak) by actual will (or at least attitudes, we can fudge that difference here)—namely of gods—not hypothetical will.

    Expanding a little on the “first”: I have never thought, and indeed have sometimes troubled to deny, that hypothetical consent or deliberation theories put things in the hands of the willers or deliberators. In democratic theory, some authors such as Habermas had long hoped to *avoid* realist (in the metaethical sense) metaphysics suggesting, he thought, that we were sadly under the authority of external standards—and thus unfree. But in his mature view what is politically justified is settled by what would hypothetically be decided in a certain ideally rational and deliberative procedure. My point about that, as here, is that what that result would be is not up to us, or up to any actual procedure. So that is, in a relevant respect, an external procedure-independent standard after all, despite its crucially mentioning procedures. I say the same thing here about what would be consented to by a fully rational agent: that is not up to any agent, but is a will-independent standard in a relevant way, despite its making essential reference to willing.

    Expanding a bit on the Euthyphro analogy: I suggested that hypothetical rational consent is nothing but the fact that consent would be rationally required. In response, Enoch writes, plausibly, “Presumably, consent would be given (in the relevant, full rationality conditions) because of the reasons, not the other way around.” I guess the other way around would be if certain things count as reasons “because” consent would be given by the fully rational.

    I find this to be a difficult question. But I also don’t think much hangs on it for our purposes, as I’ll try to explain below. But here’s what I find difficult. In one sense, it must be right that a rational agent acts “because of the reasons,” but this doesn’t yet show that they are reasons independently of what the rational agent would do (but keep in mind this doesn’t threaten voluntarism either way). She acts because of those facts, which are (in fact; put aside what description is operative in her thinking for now) the ones that fully rational agents will act because of: facts like “this action would ruin my health,” “that action would devastate my family,” “this action would not do either of those,” and so on. Does she act on those because they are, independently of what even a fully rational agent would choose, reasons to choose? Or is their being reasons to choose (after all, they are not reasons simpliciter, but reasons for agents to choose) simply their being the considerations that would move fully rational agents (which is, importantly, no empirical matter)? I’m not sure.

    But suppose we leave that hanging. So we aren’t sure, or ready to agree, about whether hypothetical consent is something over and above the facts of what would be rationally required. But if it is something more, it is not troubling (to Enoch, I mean) as some kind of voluntarism, since he has no problem with actual consent, and this would be less, or surely no more, voluntaristic than that. And we are agreeing, I think (or at least the transitivity argument does not refute it) that some normative upshots hold in virtue of the fact that it would be rationally required to consent, where mention of consent is not optional. We must also agree, I think, that since the consent non-optionally mentioned is not actual, it is hypothetical. This may not quite…quite…get us to the normative upshot holding in virtue of precisely the hypothetical consent. But the space remaining, if any, seems to me very small. A role, let’s call it, for hypothetical rational consent in the account of how certain normative upshots are justified seems to be granted. If this is right so far, we should register the extent of our agreement, even if the residual space that remains (or might remain) is also important in some way, which I’m not sure about.

  23. David Enoch says:

    Thanks, Nir. Let’s see.

    1+2. I’m not sure I fully understand the distinction between positive and negative duties. But I certainly think of autonomy in general as generating both deontological constraints, and at least reasons to promote, perhaps also duties to promote. This is why there is a reason to protect people from their poor choices, but still, not any amount of protection can justify action – because of the more deontological side of autonomy.
    If we’re now using the distinction between non-alienation and sovereignty, do both generate both kinds of noramtive considerations? Good question! I think I want to offer a tentative yes. It’s tentative, because with both, I think that what we’re directly promoting is not non-alienation or sovereignty directly, but rather the conditions conducive to both. (I think this is a central point in Joseph Raz’s discussion of Autonomy). So you promote my autonomy, say, by forbidding boxing, because you increase the chances that I won’t suffer from the kind of brain damage that precludes autonomy. Makes sense?

    3. Great example. I agree that the state regulation of salt seems less offensive than my daughter’s. This is not because it’s the state exactly – if the state did not regulate salt consumption in general, but sent an official to my dinner table, this would have been more, not less, offensive to my autonomy.
    I think that the (non-boring) explanation is closely related to your question 4: I think that the significance of autonomy is non-conventional, but that its manifestations often have a conventional side to them. If in a specific culture or environment practices developed that are understood to be central manifestations of autonomy, then intruding on them will be more offensive, due to their symbolic effect, than others. (This again is a point from Raz).
    So this is a part of the answer. Another part may be somewhat less interesting: If the state regulates salt, it doesn’t express a specific, personal distrust as does my daughter, right?

    So – 4 – I think I would go for a middle course between your two suggestions. The underlying value is entirely convnetion-independent; the specific manifestations often are.

    And 6: I don’t know whether there are other values that are relevant. I mean, I don’t have an argument up my sleeve showing that it’s just non-alienation and sovereignty. On the other hand, I can’t think of others. Can you?

  24. David Enoch says:

    Thanks, David Estlund.

    In a second I’ll return to our main exchange, but I think it’s worth while to note that when you say that according to the picture of hypothetical rational consent nothing depends on anyone’s will you are either exaggerating, or (more plausibly) relying on very strong claims about rationality: You may think, for instance, that rationality is silent on some issues, like for instance whether we should go for vanilla ice cream or chocolate. Then, if you prefer vanilla and I chocolate, now we are both unconscious or something, and Beth has to buy the ice cream for tomorrow’s party, she may think that you would rationally choose vanilla and I chocolate (and similarly for rational consent). You only get to no-dependence-on-will if you assume that rationality is never silent on anything or indifferent as between two options – certainly a possible view, but a highly unobvious one, I would think.

  25. David Enoch says:

    (Contd., to Estlund)

    I understand that your theory of normative consent is not voluntarist at all – you’re clear and explicit about this in Democratic Authority, as I recall. I think voluntarism is irrelevant in our context – none of us committed to it, nor does any of us understand the other as committed to it. (And I am really not competent at all to comment on Habermas). My invocation of the Euthyphro was meant to be more general than this – just about the direction-of-because claim, not necessarily when one of the things flanking the “because” is will-related.

    I’m not sure I fully followed the Euthyphro-expansion paragraph. I think that you are close here to something like the de-re – de-dicto distinction, asking whether the agent would act for the reason de re (namely, that it would devastate my family), or also because of the reason de dicto, roughly, because it is a reason that so-and-so, or maybe because she has a reason, or some such. This raises issues that I don’t think I need to commit on here.

    The crucial thing for me is, I think, the point about the interaction between the point you accept: namely, that consent would be given (in the relevant, full rationality conditions) because of the reasons, not the other way around; and your claim “but this doesn’t yet show that they are reasons independently of what the rational agent would do”. I want to ask about the notion of independence you’re working with – if it’s a modal notion, then I agree. But once we’re in the context of Euthyphro contrasts, a modal dependence is not enough – the whole point is to distinguish a priority relation within a modally symmetric relation. Or are you perhaps working with another understanding of independence?

    So, while I’m as happy as the next guy to register agreements, I don’t accept something you say I do: “that some normative upshots hold in virtue of the fact that it would be rationally required to consent”. Even if mentioning consent is non-optional, this in no way entails – I want to continue insisting – any such “in virtue of” claim.

  26. Nir Eyal says:

    Enoch,

    Thanks for yours. Totally agree with what you say on the state vs. your daughter. (Australia has a fraught recent history of attempting to implement certain anti-alcohol abuse policies in aboriginal communities only.)

    On conventional vs. nonconventional, thanks for your answer. The question that most intrigues me is whether there is something merely conventional about translating the value of self-control into a duty that is (primarily) negative. Would a culture that understood self-control only as “to be promoted” and not at all as “not to be interfered with” be wrong, on analytical grounds?

  27. David Enoch says:

    Nir:
    The Australian example is an excellent one.
    I don’t know what you mean by “on analytical grounds”, but I do think (on intuitive grounds, without a fancy argument I’m afraid) that a society that doesn’t recognize (and somehow incorporates in its practices) something like an autonomy-related deontological constraint is missing something important. Don’t you think?

  28. Nir Eyal says:

    I share your intuition that such a society is missing something. Am trying to look at that shared intuitive feeling or judgment critically. Can we derive the conclusion that something is missing in such a society? If not, then our shared intuition may reflect merely our cultural/evolutionary-psychological/other biases, as opposed to philosophical truth. Some philosophers would readily admit that this conclusion cannot be derived, but insist that it is only fitting to engage with the value of self-control in that deontic form. They would argue that your paternalistic daughter example is philosophical, as opposed to e.g. culturally-interpretative. But such philosophers may lack any ground for holding that their intuition reflects moral truth not these biases.

  29. David Enoch says:

    I think it’s always possible that something seems to us true only because …
    This justifies modesty, but not skepticism.
    I think of such seemings as deafisibly good starting points, no?
    And derivations – though more respectable-looking – are not at the end of the day on more secure grounds. After all, they have premises, which may be either based on intuitions or derived from other premises, …

  30. Nir Eyal says:

    This one is a bit different than other seemings in that there is a live debate between philosophers, and live discrepancy between cultures, on what is the right way to engage with the relevant value. Crudely generalizing, Americans venerate negative freedom far more than lots of other cultures.

    Suppose that an argument existed that flawlessly derived the conclusion that (special emphasis on) noninterference is the correct way to engage the value of autonomy, from e.g. what it means to be an agent. Suppose also that even cultures that hadn’t put faith in that conclusion accepted that premise. That would be very respectable.

  31. David Enoch says:

    Yeah, sure, arguments are great, especially if they proceed by indubitable steps from true, known, consensual, clear and distinct premises. How many of those have you come up across with informative conclusions, in ethics or frankly anywhere?
    Everything is controversial – among and within cultures. Acknowledging this, it seems to me, is a part of growing up.
    (And don’t get me started on theories in political philosophy that haven’t, it seems, internalized this lesson.)
    (Oh, and nothing starting from the nature of agency can do serious normative work. But this is the beginning of a different story: https://www.jstor.org/stable/20446897?seq=1#page_scan_tab_contents)