The Epistemology of Advance Directives

Here’s the familiar hard case: At 50, younger-Meredith (YM), just diagnosed with Alzheimer’s, signs an advance directive (AD) expressing her preference that her doctors do not give her extraordinary treatment to keep her alive if she gets ill after having become demented.  Meredith is an intellectual, someone who values the life of the mind, and she feels that to be kept alive when demented would make of her life a cruel joke.  At 60, older-Meredith (OM) is demented but seemingly quite content with her life when she contracts pneumonia.  When asked if she wants to be kept alive via extraordinary treatment, OM says yes.  Whose preferences are morally authoritative, YM’s or OM’s?

Most people, I believe, share the following Intuition: YM’s preferences should be carried out, i.e., the AD is binding.  How might Intuition be justified?  Many philosophers believe the metaphysics of personal identity is of crucial importance here, that it matters whether or not YM and OM are the same person.  But this doesn’t help to settle the issue.  After all, suppose YM and OM are identical.  While this may seem sufficient to ground Intuition, it’s not, for now we have to figure out why YM’s preferences are authoritative in this case, when in ordinary cases current preferences are authoritative over previous ones.  Suppose, alternatively, that YM and OM are not identical.  While it may seem that this verdict would render Intuition unjustified, it remains perfectly plausible to think that the original directive is instead to be thought of as delivering Substituted Judgment, an articulation from someone else (usually the closest relative) of what the demented patient would have wanted were she not demented.

There are responses offered to each of these problems, but let me just stipulate here that I don’t think any of them are sufficient to ground Intuition in considerations of personal identity (see here for my argument to this effect).  Cases can easily be made for the bindingness or nonbindingness of the AD regardless of whether or not identity is established between YM and OM.  Focusing on the identity of the patient is a dead end.

Perhaps, then, we might learn something from how cases like this are actually treated.  As it turns out, advance directives are regularly ignored by physicians in the real world.  Indeed, they often have to be enforced by court order.  The most-cited reason by physicians is quite understandable (and thoroughly real world): they’re afraid of being sued by relatives of patients who might express wishes contrary to the AD.

I think there’s a much more plausible reason available for ignoring ADs in the hard case, however: how can a physician faced with YM’s directive and OM’s contrary preference really know that the directive expresses what YM would have wanted (either for herself or for OM)?  In other words, the real puzzle for ADs isn’t metaphysical, it’s epistemic.  And barring physicians’ having the requisite degree of knowledge, a “better safe than sorry” mentality rules the day.

Is this a good justification for ignoring ADs?  I think that it is for some types, and this point can be brought out by considering two cases.

Ulysses I: Ulysses tells the sailors under his command to tie him to the mast and stop up their own ears so he can hear the Sirens’ song, but that no matter how much he begs them, they’re not to untie him, unstop their ears (until he tells them), or steer the ship to the shore.  When in the throes of the song, he goes temporarily insane, and he pleads with his sailors to untie him, but they refuse.  Suppose you were a sailor with Ulysses.  It seems clear that you’re right not to untie him when he’s begging you to do so.

Ulysses II: Suppose you’re a sailor lost at sea near the Sirens, and you have a kind of unmusical condition such that their song is more irritating than entrancing.  As Ulysses’ boat sails near you, you climb aboard, and you see Ulysses tied to the mast.  In this scenario, he wrote up instructions to anyone who found him not to untie him, and he sent all his sailors off in a different boat to meet him around Scylla and Charybdis.  He then begs you to untie him regardless, directing you to ignore the instructions (despite admitting they’re in his handwriting), saying that all he wants to do is listen to a little music.  What do you do?

Here I think you’d be perfectly justified in untying him.  Why the difference?  In the second story you simply don’t know him and his history: you weren’t there when he signed the document, you didn’t know his state of mind then, his firmness of purpose, or his awareness of the threat, and so you have no way of knowing now whether or not the previous signer of the document fully appreciated what the current situation would be like and thus really knew what he himself would want when in the presence of the Sirens. 

In Ulysses I, though, you were with him from the get-go: you saw his resolve, you heard his statement of purpose, his concern for your life, and then you saw the change that overtook him.  You’ve experienced the change in him he had predicted, and so you can safely be said to genuinely know what he would have wanted here, namely, to have his earlier preferences respected.  (And this is true, I think, regardless of whether or not we think of the temporarily insane Ulysses as the same person as the original.)

What this suggests is that ADs, in their current form, make too heavy an epistemic demand on physicians, for physicians are often presented with a patient in the ER who is expressing a preference contrary to the preferences written down on some piece of paper by someone they’ve never met several years ago.  In this respect, then, as in Ulysses II, they take themselves to have reason to play better-safe-than-sorry.  How could they know what the patient really would have wanted?

Contrast this scenario, though, with one in which a patient signs an AD with a physician after serious discussion of the various possible outcomes, and then they continue to see one another over the years.  When the patient undergoes the psychological transformation, the physician is there, and when the patient is in the scenario specified by the AD and expresses preferences contrary to the AD, the physician is also there.  Here, as in Ulysses I, I think the physician has very good reason to respect the preferences of the earlier signer of the directive.  The physician is now in a much better epistemic position, knowing the original patient, her specified wishes, and having seen her undergo the psychological change they had both discussed and predicted.  Here the “better safe than sorry” argument just won’t cut it.

As a result, here’s a modest proposal: advance directives should be restricted only to patients and physicians who are expected to have a continuing relationship through the time of the circumstances specified in the AD.  This allows the physician to serve as the Ulysses sailor, a trusted companion who really does know the patient.  Thus, there could be an implicit (or explicit) “out” clause in the contract: if the patient finds herself in the specified circumstances with another physician at the helm, the contract may be overridden at the judgment of the physician.  This will restrict the range of viable AD’s rather severely, but it seems that it’s the only way to ease the epistemic burden produced by the current system.

As a final note, something interesting is revealed by the argument here.  While metaphysical considerations of personal identity don’t really matter so much when it comes to AD signers, it turns out personal identity may matter when it comes to their physicians: what matters is that the same doctor is by my side as I (or whoever I become) traverse this frightening landscape.  What I want when signing the directive is for my preferences now to be respected in the unknown future, but because in signing the directive I’m really passing a burden on to someone else—the respecter of those wishes—it seems to be quite important that the respecter of the wishes be the same person as the receptor of the wishes.

In other words, when I sign the directive I’m pushing the burden of making a life-or-death call onto my end-of-life physician, but for that person to be able to meet the high epistemic burden demanded by this call, that physician should really have been carrying this burden all along, so that I may be thought to be transferring the burden to that person at the time of signing.  This, I suggest, is the only way to render that AD genuinely binding, for it’s now a document that’s binding on the physician, not the patient.  It’s a promise, in other words, that only one and the same person can carry out.

31 Replies to “The Epistemology of Advance Directives

  1. David,
    I agree with what you say about ADs and personal identity, but I’m not sure I agree with your epistemic argument. You suggest that a physician who has not had a long time relationship with a person currently in dementia does not have enough information to know whether respecting the AD is truly in her interests or not. But isn’t the AD meant to provide just such evidence? Presumably, the person who signed the AD was expressing her wishes about her future life, and we can take the document as evidence of what she most wanted upon signing it. Maybe this kind of evidence is something akin to testimonial evidence?

  2. Hi, Scott. Sure, it counts as evidence. But now weigh the evidence provided by that piece of paper against a real live person laying before you expressing a contrary preference. Suppose further you’ve just met this real live person in the ER. Which evidence will count more heavily for you?

  3. Hi Dave,
    Can you explain why following a “better safe than sorry” policy should lead physicians to not follow ADs. The typical case seems to be one where the physician doesn’t know whether respecting the AD will harm the patient or not. How it’s safe, then, to not respect the AD, when this may harm the patient? And how is untying Ulysses the safe thing to do in Ulysses II?
    I don’t see how you move from ignorance as to whether following the AD will cause more harm than good to the view that the safe thing is to not follow the AD.

  4. Doug: I don’t think the typical case is one where the doc doesn’t know whether respecting the AD will harm the patient or not Instead, I think the initial thought is solely about “respecting the patient’s wishes” (a kind of autonomy consideration), and the epistemic difficulty is that the weight of the evidence of the signed directive is weak compared to the expressed preferences of the patient before him (and here I think physicians often take the expression of preferences to indicate competence, despite what they–physicians–may say otherwise). Given the uncertainty about determining preferential authority/autonomy, then, docs will then typically move to harm-based considerations, but it’s specified in the following way: what’s in the best interests of the patient before me, i.e., what does “Do no harm” mean for this patient here and now? So the “better safe than sorry” argument runs from uncertainty about preferential authority to harm-based considerations.
    The same move, then, is what I see occurring in the Ulysses II case: autonomy/preferential authority considerations are treated separately from harm-based considerations, and uncertainty about the former spurs one to focus on the latter.

  5. Wow, I’m really surprised that you say most people intuitions would favour YM’s directive. Mine say: OM, no-brainer.

  6. As far as possible, current wishes are the ones which should be respected. We can no more legitimately bind our future self than we can legitimately bind someone else. In making this assertion, I have in mind not a philosophical account of personal identity but rather work in experimental psychologists, such as Emily Pronin. (See http://weblamp.princeton.edu/~psych/psychology/research/pronin/pubs/2008TimeDecisionMaking.pdf for a sample article.) When we recall how we were in the past or imagine how we will be in the future, we typically think about those versions of ourselves as though they were other people. We don’t have intimate insider knowledge of the thoughts and feelings of those ‘other’ selves, and so we don’t accord those thoughts and feelings a proper weight. The judgements and decisions that we make would therefore be lacking in an empathy that I feel is essential to good judgement and decision-making.
    (I’m reminded of a scene towards the end of Fight Club, when Ed Norton is desperately trying to get out of the advance directive issued by Brad Pitt.)

  7. David: I agree that there is often an epistemic problem, and sometimes there may be a personal identity issue as well (though rarely, I suspect), but suppose that, in the case being discussed, we agree that 1) the patient’s earlier preference is clearly what she would want now if she were competent (no epistemic problem in this case), and 2) that her personal identity has persisted over time (no identity problem in this case). Does that mean there is no problem with the moral authority of advance directives? Not quite, I would argue.
    There are two questions about such cases: 1) Does the advance directive express former preferences or current preferences?; 2) If it expresses former preferences, how can we respect a patient’s autonomy by giving her what she used to want?
    First question: Typically, a former preference is a preference a person once had but has since renounced. However, advance medical decision making does not easily fit this definition, for incompetent patients who made decisions in advance have not changed their mind—they simply lost the mental capacity to comprehend and reaffirm their earlier preferences. If the preference has never been renounced but cannot be reaffirmed, does she still have that preference? One can argue that the patient still has the preference provided the patient would reaffirm if he or she could. However, this move does not work in our cases, which concern permanently incompetent patients. To be disposed to do X, one must be capable of doing X. A patient who is permanently incompetent is incapable of reaffirming her earlier preference, and therefore cannot be disposed to reaffirm it. One might reply that there is a sense in which even permanently incompetent patients can regain competence, given some imaginable medical breakthrough, and therefore they retain their dispositions. However, to say that someone has a disposition to do X is to say that he or she is disposed to do so in his or her actual circumstances, not in other possible world. There is, for example, a possible world where I was raised as a devout Muslim who intends to make a pilgrimage to Mecca, but I do not have that intention–or the accompanying disposition–in this, the actual world.
    Thus, cases of permanent incompetence feature former preferences. Even if there is no epistemic or personal identity concern, the moral authority of the advance directive is still in doubt.
    Second question: Here is an argument that the principle of respect for autonomy sometimes applies to former preferences—that we can respect someone’s autonomy by respecting what she used to want:
    1. When a person has conflicting preferences on an issue, and a third, resolution preference favoring one conflicting preference over the other, respect for autonomy requires respecting the resolution preference.
    2. In cases where the person’s competence declines over time, sometimes the resolution preference is a former preference, existing at the same time as the conflicting former preference.
    3. It is possible to respect a person’s autonomy by respecting his or her former preferences. (Just as we celebrate events long after they happen, we can respect preferences long as they were held.)
    4. Therefore, when there is a former resolution preference favoring an earlier conflicting preference over a later conflicting preference, respecting autonomy requires respecting the former resolution preference and the conflicting former preference it favors.
    Thus, the Intuition is correct.
    (If I may: I’ve discussed these issues in John K. Davis, “Precedent Autonomy, Advance Directives, and End of Life Care” in the Oxford Handbook of Bioethics, and other publications cited therein.)

  8. Suppose there are two boxes: box A and box B. Suppose that I’ve looked in box B and see that there is something in there, but that I haven’t looked in box B. Is it safe to assume that there is nothing in box A? I don’t think so. So why is it safe to assume that the signed directive of Ulysses II/YM doesn’t indeed reflect what he/she really would have wanted?
    Your use of the expression “better safe than sorry” is rather strange. Calling the policy of not respecting the signed directive a better-safe-than-sorry policy would make sense if the only way that the doctor could possibly fail to respect the patient’s wishes is by failing to respect OM’s current wishes. But that’s not the case here. Is it? YM’s AD may indeed reflect the patient’s actual wishes; it’s just that the doctor doesn’t know for sure since the doctor doesn’t know enough about the patient’s past and the history of her AD and her illness.
    So I’m not seeing how the principle “better safe than sorry” is going to get you to your conclusion. It seems that your relying on some other principle, such as, it’s better to do what, given your evidence, is most likely to be the act that would not violate the patient’s actual wishes. But if that’s the principle that you’re relying upon, then I want to hear more about why the presence of a signed directive constitutes worse evidence concerning the patient’s actual wishes than the expressed wishes of a demented patient does.

  9. Dave,
    Let me share my intuitions, and then say a bit from there:
    Ulysses: It seems to me the sailor in both I and II has most reason to respect the AD.
    YM/OM: It seems to me that the doctor (regardless of who s/he is) has most reason not to respect the AD.
    Why the difference? In the Ulysses case, the sailor must be concerned with the following: (a) respecting past Ulysses; (b) respecting/caring for current Ulysses; (c) caring for future Ulysses. Given that the harm to befall future Ulysses is obviously greater than the benefit to current Ulysses of hearing the song, and that doing so would fail to respect past Ulysses desires, it seems the sailor should not untie him.
    But the OM case is different, because there is no reason to believe in an AM (ancient-Meredith) who would differ from OM in any salient way. So the balance is just between respecting YM and respecting/caring for OM. And it seems clear to me that the current desire to live plus apparent happiness with that life is more important than the past desire for dignity (or whatever).
    Now, although considerations of personal identity don’t change my mind in either of these cases, they do affect the strength of my intuitions. And I’m not totally convinced by your argument that PI is irrelevant. You write:
    “While it may seem that this verdict would render Intuition unjustified, it remains perfectly plausible to think that the original directive is instead to be thought of as delivering Substituted Judgment, an articulation from someone else (usually the closest relative) of what the demented patient would have wanted were she not demented.”
    It sounds to me as though Substituted Judgements, as you describe them, only make sense if YM and OM are the same person. Let’s say they’re not: If OM were not demented, she would be a different person, presumably YM. But she is demented, so she is OM. So why should I care what YM’s family thinks she would want (and thus about the AD)? They’re probably biased towards the wishes of some non-existent person!
    If YM and OM are not the same person, then it seems even clearer to me that the AD should not be honored. After all, how could my desire not to become another person less intelligent and rational than myself outweigh that person’s desire to live?
    And this only serves to highlight the disanalogy with Ulysses, for it seems to me that if sane and insane Ulysses are not identical, this only strengthens the case for honoring the AD. In that case, what we have to weigh is the desire of one person (insane Ulysses) to hear music against the torture of another (sane Ulysses).
    One final point about the above: I’m tempted to conclude that, if anything, we should want the doctors who treat YM and OM to be different. If anything, a past relationship with YM would serve to cloud your judgement; you would think too much of your old, intelligent friend who has become the demented yet happy creature before you, and be tempted to relieve her of an imagined burden.
    Oh, and a small, technical quibble with the case: The whole point of Ulysses’ being tied up was so that he could hear the music without being in danger. So why should saying that he wants to hear a little music convince me to release him? I can just inform him that he can hear the music from here.

  10. The physician is now in a much better epistemic position, knowing the original patient, her specified wishes, and having seen her undergo the psychological change they had both discussed and predicted.
    I don’t see why such first-personal evidence is necessary. What’s wrong with reliable testimony? It should suffice if physicians in general can be trusted to only approve patient ADs after extensive discussion (to ensure that the patient has a full understanding of what they are signing). They could even explicitly mention the predicted psychological changes in the AD. If a new doctor receives this detailed document years later, signed by both the patient and a trustworthy physician confirming that the patient is fully informed, why should the identity of the new doctor matter at all?
    Is the thought that physicians cannot generally be trusted to ensure that their patients understand what they’re signing? (This would seem to undermine the institution of ADs.) Or is it that there’s some essential information in these cases that can only be obtained first-hand, and cannot practicably be transmitted via written testimony to future doctors in as much detail as would be necessary for them to judge whether the patient’s recent change of heart casts doubt on the depth/authenticity of their previously expressed preference?

  11. R: I think lots of people think ADs have a point, and at the time of signing believe their wishes will be respected at crunch time. This is to adhere to Intuition. If you think OM’s preferences aren’t authoritative, I doubt you’re on the original AD bandwagon to begin with.
    I’m curious about how the empirical research you cite could imply anything about the legitimacy of self-binding.
    Also, you really believe current preferences are what have credence, even if they’re expressions of the demented?

  12. John: Thanks for the reference. I’m puzzled by two things you say. The first is that, if we stipulate no epistemic problem (i.e., we now know just what the patient would clearly want were she competent), then we could still have a worry about which preferences are authoritative. But then the question you raise about such stipulated cases is “Does the AD express former or current preferences?” Hadn’t that been established? Assuming competent preferences are authoritative over incompetent preferences, why is this question relevant here? In addition, you go on to raise questions about whether or not a permanently incompetent patient still has certain preferences, but this sounds like an epistemic problem to me (indeed, it’s the epistemic problem.
    Second, you say, “It is possible to respect a person’s autonomy by respecting his or her former preferences. (Just as we celebrate events long after they happen, we can respect preferences long as they were held.),” but then you go on to say, “Therefore, when there is a former resolution preference favoring an earlier conflicting preference over a later conflicting preference, respecting autonomy requires respecting the former resolution preference and the conflicting former preference it favors.” But why think that? After all, the earlier claim doesn’t rule out the possibility that one could respect the person’s autonomy in multiple ways, so why think respecting autonomy “requires” respecting only the “resolution” preference?

  13. Doug: I can’t address your “Box A/Box B” case because your second sentence involves a typo that makes it contradictory (have I looked in Box B or not?), and so I can’t follow what the point is, precisely.
    More importantly, here’s the argument on behalf of the ER doc I’m imagining who’s presented with OM: My #1 principle is to respect the patient’s wishes; I don’t have sufficient evidence to determine what the patient’s wishes here and now are (I’ve got a document saying one thing and a perhaps-incompetent patient saying something else); so instead of picking one or the other’s preferences to respect, I’ll err on the side of caution, go with the “Do no harm” considerations (perhaps narrowly construed) and keep her alive (if I go with the earlier preference, I could be wrong and, I suppose, letting someone die who didn’t want to die is worse for her than enabling her to live if she didn’t want to — in part because the latter is rectifiable).
    Incidentally, it’s important that I’m not urging that what the doctor does is what he ought to do, only that it’s reasonable for him to do it, given his epistemic situation.
    Somehow I feel I may not be getting your point, though.
    I’ll try to respond more tomorrow (I’d originally set this up to post last Thursday, when I had more time to deal with responses, but I posted right on top of Dale then and wanted to leave time for his post to breathe, and now I’ve much less time to deal with these helpful comments than I’d like.)

  14. instead of picking one or the other’s preferences to respect, I’ll err on the side of caution, go with the “Do no harm” considerations (perhaps narrowly construed) and keep her alive (if I go with the earlier preference, I could be wrong and, I suppose, letting someone die who didn’t want to die is worse for her than enabling her to live if she didn’t want to — in part because the latter is rectifiable).
    Are you just assuming that you don’t harm someone by disrespecting their preferences?

  15. Faraci: I guess I disagree with just about everything you’ve said. First, on the point you seem to share with Doug (in the immediately prior comment): it’s not that I’m assuming you don’t harm someone by disrespecting their preferences. I’m instead saying that, for physicians, it seems that autonomy-based-reasons take priority over harm-based reasons for making certain sorts of decisions. This isn’t to say that failure to respect a patient’s autonomy won’t harm her; that may well be true. But these reasons take precedence independently of their connection to harm, I think. So, David, I suppose I’m assuming a similar sort of priority for the sailor in the Ulysses case, and so your appeals to welfare considerations for Ulysses right off the bat jump the gun.
    I’m now seeing a disanalogy between the two cases, though, brought out by Doug’s discussion of the “better safe than sorry” move, for I suppose if the sailor truly went that route, after recognizing his epistemic uncertain regarding the autonomy-based reasons, he’d leave Ulysses tied up, perhaps for the reasons you suggest, David.
    Finally, David, I’m puzzled by your comments about substituted judgment. Surely you’d want as informed an opinion as possible about what the patient would have wanted were she not demented. Evidence for this typically comes from the closest relative, as it’s presumed the relative is in a much better position to offer such a judgment. If YM and OM aren’t identical, then (as was stipulated), some might take YM’s directive to constitute that “insider access” and the best judgment. I think this is not a good conclusion to draw, however, for it would be as if we’d gone back in time 10 years ago to ask YM’s closest relative then what she would want ten years later. But surely that won’t provide us with reliable evidence for a responsible judgment.

  16. Richard: Thanks for your comment. Two things. First, I have a fairly strong intuition that the evidence of “my own eyes” is significantly more powerful than reliable testimony in precisely these sorts of life-and-death cases. That I’ve been the one to have discussed the options and seen the psychological changes will be far more palpable evidence for me to knowledgeably deny the expressed preference of the patient before me. Again, that’s not to say it will be unreasonable for the ER doc unfamiliar with the patient to nevertheless adhere to the directive. I’m just saying that it doesn’t strike me as unreasonable for someone to not adhere to the directive in such an instance, even if he had reliable testimony. But I have no real argument for taking more seriously the first-personal experience as evidence.
    Nevertheless (and here’s the second point), suppose reliable testimony were to be taken just as seriously as first-personal evidence. One might think of the identity of the physician with respect to the role as promise-keeper to the signer of the directive as consisting in a kind of continuity-of-information-regarding-the-patient, and insofar as reliable testimony preserved this (circumscribed psychological) continuity, the eventual physician who receives the relevant testimony and paperwork could well be in the appropriate position to carry out the directive in virtue of his “identity” with the original physician.

  17. “Surely you’d want as informed an opinion as possible about what the patient would have wanted were she not demented.”
    Suppose Frankenstein’s monster—who (let’s stipulate) is not identical to any of the persons that constitute him—contracts pneumonia. We then discover that some of the original body-part “owners” had ADs, all requesting that extraordinary measures not be taken for them. I can’t imagine why these should be relevant. I take it that ADs are about how we should treat a person, not how we should treat their body, no matter whose it later becomes (or, at least, this is the extent to which ADs should be honored, whatever their intent). If becoming demented (becoming OM) is truly becoming another person, then YM’s desires don’t matter, any more than if I were to write an AD for you. In fact, if anything, I would take this AD less seriously than one I wrote for you, since I have some concern and respect for you, but we know that YM finds the idea of OM’s existence distasteful.

  18. David: There are a number of issues here on which I’m unclear. First, I don’t get the Frankenstein case. He’s surely not constituted by persons; rather, he’s constituted by the body parts of former persons. ADs aren’t written about the dispensation of body parts, however; they’re written about the dispensation of patients. So I don’t see the relevance here. Second, if the monster is truly a distinct person who’s not demented (i.e., not incompetent), then of course his preferences should be respected.
    As to another point, it’s unclear whether or not ADs are about how we should treat persons (despite being about how we should treat patients). Indeed, they’re often called “living wills,” precisely because, for some, they’re about how I want my living body treated in cases where I — the psychological person currently inhabiting it — have ceased to exist. Furthermore, “persons” is ambiguous between competent and incompetent persons. I’m talking specifically about cases in which the person before the ER doc is incompetent.
    Finally, I’m still unsure why you think that, if an identity-change occurs, YM’s desires are irrelevant. Suppose I were to fission. The products wouldn’t be identical to me (on pain of being identical with each other — set aside the four-dimensionalist “solution” to this case as well). Suppose they fell into comas. Wouldn’t an AD I wrote still provide evidence for what they would have wanted had they been competent? So why wouldn’t what YM wrote at 50 provide some evidence for what her psychological descendant would have wanted were she competent, even on the assumption that YM and OM aren’t identical?

  19. Dave,
    I take it that in the case where YM and OM are not identical, this is because YM’s growing dementia caused her to cease to exist at some point, a point at which OM came into existence. If this is the case, then I don’t think it makes sense to ask what OM would want if she were not demented. Being demented is, in a sense, essential to her existence; for were she not demented she would cease to exist. Now, one could argue that a competent person like YM who has desires about how her body should be treated once she no longer inhabits it should outweigh the current desires of an incompetent person like OM. But barring that, it still seems to me that YM is irrelevant, because to ask “What would OM want if she weren’t demented?” is really just to ask “What would YM want for OM?” But since YM and OM are not the same person and, as I’ve suggested, YM is in a sense hostile to OM, I can’t see why her desires should carry any weight.
    Your fission case is different because there is no direct tie between the loss of competence and the change in personhood. Suppose that you fission as you suggest and that the new persons fall into comas. At that point, I agree that the AD should be carried out. But this is not because your wishes should be in any way respected (any more than YM’s), but because they do (as you say) offer good evidence as to what the counterparts would want for themselves. After all, they bear (nearly if not exactly) the same level of psychological similarity to you at the point when you wrote the AD as you yourself do (how much difference depending on how long after the fission they fall into comas). And thus the AD is as good an indication of their desires as it would be of yours had you fallen into a coma. (This is all ignoring the further wrinkle of whether entering the coma itself constitutes a further shift or loss of identity.)
    So, the upshot: I still think that personal identity matters in the case of ADs generally. Let’s take a generic case. P1 writes an AD and P2 is incompetent. If they are identical, then honoring the AD is a matter of (a) respecting P1‘s desires for his future self and (b) indicating what P2 would want were he competent. If they are not identical, then honoring the AD is a matter of (a) respecting P1‘s desires regarding the treatment of his body once he no longer inhabits it and (b) indicating what P2 would want were he competent. While (b) is the same in both cases, the difference in (a) seems to me very important. What’s more, cases like YM/OM point out that a link between shift in personhood and loss of competence render (b) incoherent.
    I also don’t think this linking problem is that special to YM/OM. In fact, I suspect most real cases are going to have it. The fission case is special because there is a change in personhood that has nothing to do with the loss of competence. Not only does this rarely (if ever) happen in real AD cases, but the it is odd because it’s hard (at least for me) to accept that the persons are non-identical in the first place. The fact that the new persons could bear precisely the same relevant psychological relations to your past self who wrote the AD as you do make it tempting to assign identity (this is why the four-dimensional solution is the only one that has ever made any sense to me, anyway).

  20. Being demented is, in a sense, essential to her existence; for were she not demented she would cease to exist.
    In what sense is it essential to her existence? In the sense that, if she didn’t have that property, she’d cease to exist?
    I’m still not seeing how what you say challenges the point about identity in my post. My point about identity there was that, when it comes to defending Intuition considerations of personal identity are just irrelevant. Take your generic case. In the scenario where P1 and P2 are identical, we still have to establish why P1’s preferences are authoritative here. Indeed, identity just makes this harder, given that we typically view the preferences of a later self as authoritative. If they aren’t identical, then one possibility is that P2 is someone else (the other possibility is that P2 is no longer a person, but I won’t deal with that here). But even if P2 is someone else, we can still try to defend Intuition by appealing to substituted judgment (where P1’s AD counts as evidence for what P2 would want were she competent). Now I’ve said that this is problematic too, because why should we think that what a ten-years-ago different person wanted should count as good evidence for what P2 would want now? (This doesn’t undermine the possibility, however, that sometimes such a directive could count as some evidence in cases of non-identity.) So my point is that, even if we could establish identity/non-identity in these cases, it’s not relevant to what will have to do the real work in defending Intuition (which will be instead an argument about preferential authority or substituted judgment). My suggestion, then, was that perhaps we should shift from focusing on the metaphysics of these cases to focusing on the epistemology of them (which, it turns out, reveals the more basic problem).
    What you do in your original comment is simply deny Intuition, but that’s not a direction most people want to go in. So I’m trying to see if we can rescue Intuition for some small number of ADs. (I should have noted this point earlier.)

  21. Dave,
    Okay, sorry if I went off too much on attacking Intuition in general, rather than focusing on the possible relevance of personal identity.
    If I understand you correctly, you want to argue that personal identity is irrelevant because answering the question of personal identity is not sufficient to render Intuition justified or unjustified. Rather, Intuition must be justified or defeated by an epistemic argument. My position is that at least in the case of YM/OM (and I suspect in many other cases as well), while it is true that if we were to learn that YM and OM are identical this would not be sufficient to justify Intuition, were we to learn that they are not identical this would be sufficient to defeat Intuition. I believe this because I do not agree with you that if YM and OM are not identical, then YM’s AD can be treated as Substituted Judgement for OM.
    If I’m right, then YM/OM is one case where one could only defend Intuition if identity holds. You can maintain, of course, that there is still a class of cases where Intuition can be defended either by arguing for Preferential Authority or Substituted Judgement. This would provide the opportunity to “rescue Intuition for some small number of ADs” without having to deal with personal identity.
    The challenge I mean to offer, then, is to identify concrete cases that are non-controversially in this class, and to offer reasons for thinking that they are sufficiently prevalent. After all, it might yet turn out that most cases are more like YM/OM than anything else.

  22. I recognize, of course, that your basic point holds whether or not this challenge is met. Certainly, if you just want to show that in some cases, a defense of Preferential Authority or Substituted Judgement will allow a bypass of the personal identity issue, then I have no objection. But I do think it is interesting that, at least so far as I can tell, the case you offer is not such a case. I am further suspicious that such cases, at least in the real world, are going to be hard to come by. But, again, assuaging my worries about the applicability of your theoretical point is not your job. So, anyway, this is all basically just a means of saying that “challenge” was perhaps the wrong word for me to use.

  23. Why not simply administer a Roth/Meisel type competency test and if OM is deemed competent then allow OM’s present wishes to decide what happens next. I presume that part of the test would revolve around a discussion of OM’s present understanding of the AD she signed earlier. It seems that the only issue is if OM still agrees with YM and if not, why not? It does not follow that if OM does not want the AD to be implemented at the present time that the AD is still not applicable when OM is deemed incompetent.

  24. David: Fair enough. Thanks for the discussion.
    John: My stipulation was that OM is incompetent. But it’s not at all clear to me (nor is it all clear to the ER docs I’ve spoken with about this) that incompetence renders the expression of a preference irrelevant to what a competent version of the patient would want for herself.

  25. David
    Sorry that I misread your argument. I am at a loss to see how a moral problem arises in this situation. One question I now have is why would an ER doc ask OM what her preferences are in the 1st place? What is their motivation? If they know that OM is not competent it certainly cannot be to gain reliable information on what M wants in so far as for the information to be reliable OM would have to be competent at some relevant level. I presume the ER docs know that OM is incompetent so it seems that asking OM the question as to her preferences is misguided and out of place. We would not allow OM to sign an AD so it does seem that any response OM makes would not be relevant to what we should do with her. If there was no AD signed by YM would we allow OM to have any say in determining how she should be treated or would we, as ER docs, utilize our best medical judgment in determining how to treat her? I am reasonably confident that we would do the later. If OM did not have an AD and OM requested that she not be treated, the ER docs would override this request based on OM’s condition and do what they thought was in her medical best interest and probably keep her alive especially if she were enjoying some quality of life. The fact the OM has an AD signed when she was competent to give an advance directive places a binding constraint on how ER docs can treat her so that they cannot substitute their judgment for hers. I take it that is why YM signed the AD in the 1st place – to make sure that her competent wishes (preferences) are followed even if the ER doc does not want to do so. So, I guess I do not see this as a ‘hard case.’ What am I missing?

  26. John: What makes the AD that YM signed a “binding constraint” on how ER docs can treat OM? This is precisely the problem here: what makes YM’s preferences authoritative over OM’s? It’s irrelevant why OM is asked what her preferences are. Perhaps she’s not asked, but simply offers up the information that she’s content and peaceful and doesn’t want to die. To say that this information is irrelevant to the decision of the ER doc is to assume that such physicians are inhuman: how could you allow someone to die before you who’s saying again and again that she wants to live? And why think the preference she’s expressing isn’t some evidence for what she’d prefer were she in fact competent? After all, YM may not have fully appreciated what life would be like for her were she in this condition. And so forth. These are the genuine epistemic puzzles facing physicians in this position, and these strike me as genuinely hard ones.

  27. David
    Thank you for responding. I am enjoying our exchange.
    You wrote: “What makes the AD that YM signed a “binding constraint” on how ER docs can treat OM?”
    I would think that it is the fact that when YM develops the AD and signs it, it is with the expectation that it will be followed when the conditions stipulated by her in the AD are met that make the AD a binding constraint on how M wants to be treated when she changes from YM to OM. If that is the function of an AD, and I cannot see any other function that it could serve, then the epistemic issues of the ER docs are irrelevant.
    Only YM is in a position to competently (by your stipulation) evaluate the worth of her life. The use of ‘may not have’ does not change this because she might very well ‘may have.’ I would think that an epistemologist would want to eliminate the ‘may or may not’s at some point of the discussion (or we are stuck in a ‘what-if’ game) It seems that is what you are trying to do by suggesting that the preference she states as OM, if it is a preference, is some sort of evidence that should be considered. I think that if something is to be considered as evidence we should consider the source and if that source is not competent then we should rightfully ignore it.
    Also, I can be genuinely puzzled by something without that thing being a genuine puzzle. There has to be some epistemic basis for a puzzle and I do not see that this problem demonstrates that such a basis exists. If OM (under your stipulations) came in without an AD and said repeatedly she wanted to die and that is all she said, would the ER doc be justified in letting her die using her stated ‘preference’ as the only evidence for such justifying his/her action? I think we would question that ER doc’s decision (Do you remember the “It’s all over now Debbie” controversy a few years back? Similar issue I think to what you are trying to develop.)
    Throwing in terms like ‘inhuman’ does not further your argument. Why would it be inhuman to follow the directives of the competent YM based on what we know YM thought about the conditions under which she wanted to not have extraordinary treatment and those conditions have been met? Why would it not be inhuman to force her to live in a condition she clearly stated, while competent, that she did not want to live in? If you are correct that this is a serious epistemic issue then why have an AD at all if epistemic issues can arise so easily simply as a result of a statement made by someone who is, by stipulation, incompetent. What role does competence (expertise) play in epistemology and decision-making? Who is the expert regarding OM?
    Now, maybe some work needs to be done on how AD’s are developed, but that is a different issue.

  28. David
    You rely on UI and UII to make a distinction between bystanders and what they should do, or what is morally permissible for them to do, regarding untying Ulysses. Are you stipulating that Ulysses in UII is not competent? It seems to me that he might well pass a Roth/Meisel type test and be found competent by the sailor. If so then his AD certainly could be set aside. I do not see this example (UII) having a clear relationship to YM/OM.
    Also, why should we not simply view the doctors not fulfilling the AD of M as an example of moral cowardice in so far as it seems they are disregarding the AD not because YM expresses a preference that needs to be taken into account, but that they fear being sued and want to pass the responsibility for executing (no pun intended) the AD to someone else, namely the courts?
    How are the doctors refusing to abide by the AD any different then nurses refusing to provide service to a competent client they disagree with as to the chosen course of treatment, the so called ‘right of conscience?’ I have argued elsewhere that professionals do not have the right to not provide service to a competent patient even if performing that service would violate a deeply held moral belief. If the service is one that is offered by the practice that the nurses knowingly and freely joined knowing what to expect within the normal course of functioning as a professional in that practice then the nurses must abide by their promise to provide competent care to their patients/clients.

  29. I am an attorney who works with terminally ill clients on a regular basis to draft advance directives and durable powers of attorney for health care and so come at this discussion from a different perspective. Under the law a person remains capable of renouncing their AD as long as they retain the mental competence to provide meaningful, informed consent to the proposed course of treatment. In this case OM may or may not retain such competence. It would depend on whether she retained the mental capacity to understand at even a basic level the risks of either treating (pain, recovery period, reduction in capacity) or not treating (death.) If she retains that mental capacity then both the law and medicine would say that the decisions of OM rule. And I think philosophy would agree. If she does not have that level of mental competence than her stated preferences are in my opinion not meaningful and “respecting her wishes” is illusory. In all such cases the doctor must rely upon substituted judgement. That might come from an AD or it might come from the person that YM designated as her health care agent. In my opinion the best decisions come from an agent guided by an AD which is why all my clients execute both. The Agent, who ideally does know the circumstances and motivations of the AD and who has talked over those choices with YM, knows the system of decision making M used while she was competent but also has new information not available to YM, such as OM is happy as a demented person and does get enjoyment out of life in that state despite her expectations. The agent can combine those two source of knowledge along with the judgement and wisdom that led them to be selected as an agent to make the best guess as what OM would want if she had the ability to truly understand the choice in front of her. That is why almost all of my client put a clause into their AD that says that their agent is free to vary from it if they think that would be a better choice at the time. Without an agent, doctors do often treat, even without a request to do so, despite an AD. In my experience that is often out of cowardiness and a desire not to be sued and often does lead to harm being done such as painful procedures being done an a frail and terminally ill patient who never is able to regain an semblance of the life that made them a happy demented OM. I think the better analogy is to a sailor who knows quite clearly that UII is out of his mind and will meet a bad end but unties him because he is afraid he will otherwise be sued. I don’t think there is much of an argument that he is doing the right thing. The agent introduces a new player — the last sailor left on the ship who knew UI and can tell the person boarding the ship “He left this note, he meant it, and he put me in charge of deciding when he get untied.” Better for everyone on board the ship.

Comments are closed.