Sunday, November 24, 2013

Overriding ignorance and respect for autonomy

by Dylan Popowicz

Last year, I found myself in the minority during a debate on the topic of personal autonomy*. In dispute was the following case:

A woman (let’s call her Mrs. Jones) has been diagnosed with cancer but refuses to believe what her doctor is telling her. A person with cancer, she believes, would not feel as well as she does; a person with cancer, according to her belief, would rapidly lose weight, whilst she has in fact gained a few pounds. According to her own prior experience with cancer (through word of mouth, film, etc.) it isn’t possible that she has cancer—thus, she refuses to take any medicine or follow any course of treatment.

This question followed:

Q1: Assuming that it is certainly true that Mrs. Jones has cancer, but that she doesn’t believe that she does after being proffered the appropriate evidence, would it be right for the doctor to secretly or forcefully administer treatment?

The class’ overwhelming response was that it would be immoral for the doctor to administer any drugs or medication without Mrs. Jones’ permission. The central principle of autonomy, our moral obligation to respect and protect an individual’s autonomy, had won out over any other (benevolence, nonmaleficence, etc.). Even with an important adjustment to the case, the majority opinion would not sway: it made little difference if we could imagine that the cancer could be treated by something as innocuous or unremarkable as a pill. The issue was not pain, or harm of any sort, but autonomy through and through.

It seemed to me that there was a world of difference between the autonomous choice of one holding a skewed relationship with reality, and the autonomous choice of one acting upon reflection of the facts. Of course, any such consideration was answered swiftly: but surely a part of one’s being autonomous is the ability to choose what one believes.

Now, I think that there is a profound philosophical problem at the root of such an untethered conceptualisation of autonomy. It seems peculiar to think that we could choose what we believe, but even if we could, it seems that our range of choice is limited to a certain educational conditioning. A way of teasing this issue out is to change the situation slightly: let us consider that the only reason that Mrs. Jones is denying the doctor’s assertion is due to her distrust of ‘people of colour’, and that the doctor is a black man. A new question could be (and was) asked:

Q2. Assuming that it is certainly true that Mrs. Jones has cancer, but that she doesn’t believe that she does due to the fact that she does not trust the opinion of any non-white person, would it be right for the doctor to secretly or forcefully administer treatment?

Let us assume that there is no other doctor available, no white authority to ask—either Mrs. Jones is treated by the black doctor, or she swiftly and painfully dies. Still, the majority of my class (though a tad fewer) argued that it would be immoral for the doctor to act against Mrs. Jones’ will. Even when I suggested that we certainly knew that Mrs. Jones would accept the treatment if she herself believed that she had cancer, the class still stood opposed to action—the relationship between belief and truth were not nearly as important as the the agent’s ‘choice’.

I’d like to suggest that a more robust conceptualisation of autonomy would take more into consideration than the childish assertion that action x is autonomous because subject y did it because she “wanted to”. I think we need to pay much more attention to how we conceive of agents and their relationship to the world, specifically when it comes to an agent’s epistemic position. An agent chooses based on certain epistemic considerations: we have beliefs or knowledge about the goals of our actions, what we are capable of doing, as well as the present state of affairs, or the point of action itself.

When it comes to these epistemic relationships, it seems difficult to disavow the important role other agents play.

Consider a colour-blind individual, let’s name him Calvin: I’d argue that it is not plausible that Calvin could come to properly understand his own optical deficiencies without the input of a non-colour-blind individual, or what I would like to refer to as an external perspective on his epistemic relationships (a mediating third element). In this situation, the latter individual would have what we could call an epistemic authority. Imagine the (fantastic) situation in which Calvin (who has trouble distinguishing between red and green) has come into contact with a deadly virus, and is sent by a doctor to get the “green” pills from a medical office, only to find that they are located next to a jar of “red” pills. Calvin, using what beliefs he has about the colours “red” and “green”, tragically grabs the wrong pills. Would we consider it a breach of a respect for autonomy if a second individual, let’s call her Claire, who was not colour blind, were to override Calvin’s choice, and make him take the other pills, even against his vehement objections? Remember, it isn’t that Claire is stopping Calvin from reaching a certain end or goal that he desires (the opposite in fact) but rather that she is exercising a certain authority over Calvin’s epistemic position, something that only a perspective from the outside can achieve.

If an autonomous being is taking a course of action towards a certain goal, but unbeknownst to itself, is mistaken in the course of an action’s relationship to the end, surely we can accept a certain sense of paternalism and epistemic authority whilst still leaving sufficient room for autonomy. It isn’t enough to think about an agent as simply ‘acting’ in a vacuum, or to analyse autonomy without making reference to an epistemic context. Ultimately—and I know a lot of readers will resist this on political grounds—I’m inclined to suggest that a full conception of a respect for autonomy would necessitate the need for certain paternalistic acts, based on the epistemic authority of those who know better than we.

* This discussion and debate took place in Prof. DiSilvestro’s Bioethics course in the summer of 2012. My thoughts on the matter are indebted to Dr. D.’s class.

Dylan Popowicz
Senior Undergraduate
Department of Philosophy
Sacramento State University


  1. I think you should worry a little more than you do about the move from epistemic authority to legitimate force, especially in cases where there’s (reasonable) disagreement about who the epistemic authority is. When you do, the question wouldn’t be whether the conception of autonomy is untethered. It’s whether it should be tethered to *the truth* or something rather less demanding like minimally justified belief. For various reasons, I opt for the latter. We can hold false beliefs with sufficient justification that generate in us reasons for acting that others (even others who know better) should consider as counting against the legitimacy of using force or exercising (practical) authority over us.

  2. Prof. Swan, thank you for the comment.

    I certainly agree with you. I should have clarified in my piece that my only intentions were, first, to analyse what I see to be a fairly ill-defined conception of autonomy as it is seems to be often used in a non-philosophical context and, secondly, to gauge the opinions of a more philosophically minded audience. It was a rather simple minded approach, but I think being clear on the issue will help us better accommodate the practical consequences that you mention, ones that I admittedly didn’t deal with.

    Tethering our conception of autonomy to *the truth*, as you put it, would certainly be too high an order (although I realise I did use the t-word rather irresponsibly) - all I'm asking is the admittance of an external perspective, and the possibility of interference, as being somewhat involved some of the time. I’m attracted to a relatively externalist view on knowledge, especially in regards to our lack of insight into the reliability of our own processes of epistemic acquisition. Recent empirical research at least tends to support such a view of our myopia on such matters. It seems reasonable to me to thus suggest that our rational justifications for action must partially rely on an intersubjective system of checks, agreement and disagreement, where others can correct our mistakes (or help justify our choices). As you point out, the really important question then follows: when and where can such a system legitimately force an individual to not take an action where discourse has failed?

    Obviously, in cases of reasonable disagreement, it would be incredibly difficult to justify any sort of broad, insensitive interference or force. There are also a plethora of other reasons to resist force: for example, we sometimes think it in the best interest for people to learn from their own mistakes, as part of their autonomous development, even if there is no question to our epistemic authority on the matter. Another issue is that of peer/expert disagreement, as discussed in Matt DeStefano’s post a few weeks ago. Etc.

  3. Dylan, I think Kyle's point is interesting, but I don't think you should give up the truth tether too quickly. To me there is something even more poignant about the old lady example if we change it to her having a well-justified belief that she does not have cancer even though we know that she does. In a case like this, the decision to override may be even more compelling. I think when we imagine people engaged in decision making we size them up at the character level in one way or another. Your lady is woefully ignorant, but she also strikes us as someone who is prideful and may value her will being respected more than she does the truth. Hence our reluctance to override her decision is partly due our desire to respect her desire to be respected. But if I sized her up as someone who is fundamentally a truth-seeker, then I would feel a strong inclination to override because I think it is the truth, not respect for her autonomy that really matters to her. (I wouldn't actually override it for practical reasons. I just wouldn't care enough about her to be willing to do something manifestly illegal for her benefit.)

    I think what I'm saying here suggests that respect for autonomy has a strong counterfactual dimension. It is respect for what a person's will would be if she were in possession of the correct information. We could actually figure this out on a person by person basis if we simply gave them a questionnaire to fill out in which they are asked to say how they would want to be treated in the various kinds of scenarios you are considering.

    What you are thinking about here has a lot of relevance for the institutionalized use of placebos.

  4. Prof. Mayes, thank you for the comment.

    I'm especially glad you mentioned the questionnaire and the counterfactual dimension. Though I didn't develop the idea specifically, my concern here I think stems primarily from this idea that the value of a person’s choice resides in their intended goals, or consequences of actions, rather than the means per se. The means, it seems to me, are generally negotiable, whilst the forced alteration of one’s goals would be tantamount to an egregious undermining of one’s autonomy—to act with force in this case would be disrespectful.

    In the case of Calvin I think it is clear that Claire’s interference in no way alters Calvin’s intentions—she is simply correcting his epistemic blunder. If Calvin were to be properly related to the colour of the pills, he would take the green pill, and this is something that we can easily infer from the situation. As you put it: it “is a respect for a person’s will would be if she were in possession of the correct information”. In the case of Mrs. Jones, we need a more sensitive way to measure her intentions and which counterfactual courses of actions could be applicable to the same ends: possibly a medical questionnaire, as you suggested.

    I think you’re right that the Jones example requires us to more carefully consider her ‘character’. You’re right that my portrayal of her suggests that she most likely values her will (especially her ‘white’ will over the judgement of a black doctor) more than she values the truth. Her intended goal may not be the truth, nor necessarily her good health (if she really believes her own views, then her good health is probably little more than a passing concern in this situation): she is simply reacting in defiance of a figure she does not in any way consider to be an authority, due to the colour of his skin. Your alteration is certainly an attractive one—I hadn’t thought of it, since my initial and intuitive reaction was to reject interference in such a case (in which the agent in question has a well-justified belief). Your Mrs. Jones certainly seems like someone who would be better respected as an autonomous agent if she was corrected in her search for the truth.

    I think I'm being dense: what do you have in mind regarding the use of placebos?

    1. Dylan thank for the thoughtful comments. Regarding placebos I was just thinking that their effectiveness to some extent depends on patients not knowing they are receiving them, so it offers up analogous overriding possibilities. For example, lots of patients are just completely sure they need a certain kind of surgery and are convinced that care is just being denied to to them even though the surgery is known to be inappropriate or ineffective in cases like theirs. It's been shown that placebo surgeries work as well and sometimes better, so why not just give them those without telling them. Same holds true for a variety of medications that don't work and have nasty side effects (SSRI's being a current controversial example). It's fraught with complications, but all the basic considerations you are making seem to apply.

    2. Prof. Mayes, thank you for the clarification. The placebo effect is an interesting one. I think my only reservation in this case would be in regards to the level of deceit. I think Prof. Bellon is right to be concerned with a secretive dimension of interference, so maybe in the case of placebos I would think it in the subject's best interests if they were informed of deceit after the fact (I imagine this is legally how it has to be done anyhow?). In such a case the taking of placebo pills or the procedure of "fake" surgery could be understood as epistemic lessons that would bolster the patient's autonomy in the long run.

      I think application in the field of mental health would be especially interesting (you mentioned SSRIs). To take a personal example: I ended up in the ER during the summer break, convinced that my heart was about to give up on me (it felt like I was having palpitations, my arm had gone numb, I was faint, etc.). Even after an EKG and various blood tests, all of which were clean, I couldn't quite believe that it was "just anxiety" as the doctor(s) were claiming. Obviously this wasn't an option, but I can imagine being in such a state of mind and insisting on being treated for what I believed to be the problem (my heart). A placebo would have done the trick, and, as far as I can tell, I would have realised how foolish I was being if, a week later, the doctor told me that I had taken nothing more than a sugar pill. I would consider such an action in line with a respect for my autonomy. Or let’s assume that SSRIs are effective: another possibility would be for the doctor to force me to take an anti-anxiety pill – if my symptoms were to disappear right afterwards, then I would be afforded a better perspective. (In reality this last example I think would be a breach of my autonomy since I’m rather suspicious of the efficacy of SSRIs myself).

    3. Dylan, that sounds like a scary incident. That's a good example of a base rate fallacy. Heart attacks are so rare in otherwise healthy males of your age group that even if your symptoms were highly indicative of a heart attack, the overwhelming likelihood is that it's a false positive. Easy to say and believe when you aren't experiencing the symptoms.

      On the philosophy, my intuitions run in your direction as well. I'm not sure how easy it would be for me to think they were respecting my autonomy if I was even non culpably misdiagnosed. (I bet this is the sort of thing that would turn up real asymmetries in judgment in a Joshua Knobe type experiment.)

  5. Randy, I think coercion/interference has to be justified for a person in order for it to be legitimate. It must be the case that I have sufficient reason, in some sense, to do it. I also think that whether I affirm it or not can't be the right test for whether it's justified for me/I have sufficient reason to do it. We agree about that. Actual Calvin doesn't affirm taking the green pills, but it seems right to say he has reason to do that. Counterfactual/idealized Calvin does affirm taking the green pills. Maybe the thing to say is that our reasons for action are constituted by what idealized versions of ourselves would affirm. It's what is justified for *that guy* that matters rather than what's justified for our actual selves.

    But we should worry more about the scope of the idealization. Think of it as a continuum. There are big differences among actual me, me after some careful thought and reflection, me absent egregious errors, me after consulting a questionnaire I've filled up designed to identify things I'm rationally committed to, me with only correct information, and me with *all* the facts and free from rational mistake. Focus on this last guy, hyper-idealized me. I have no idea what he would affirm, but it's plausible that it's stuff that I just don't get or see the point of. He's just way too different from me. I doubt I'll very often have sufficient reason to do what he affirms. In such cases, forcing me to do it is disrespectful.

    Idealized conceptions of agency are abused in moral and political philosophy. Too often it conveniently turns out that "reasonable" people are the people who agree with the theorist advancing the coercion/interference.

    1. Prof. Swan,

      Please correct me if I'm misunderstanding you. I could be wrong, but I don't think that what I have in mind can be construed quite as strongly as your idea of the idealised agent. I think we should always ground our analyses of an agent's choices by what she is aiming to accomplish - this means that we can't override Mrs. Jones' or Calvin's choices by going so far as to stop them from reaching their goals. The way I see it, it isn't that Calvin 'affirms' taking the red pill or green pill: Calvin affirms taking the pill that will cure him, and he is simply mistaken in the facts. We aren't overriding this choice when we force him to take the other pill - we're actually affirming it, and his autonomy.

      Now, let's say we have a different situation in which Mrs. Jones is told she has cancer, and believes that she does, but doesn't want to go through any treatment procedure (there are no easy shortcuts, just brutal chemo). I don't think that there is anything in my account that suggests that we can legitimately step in and force her to be treated: to do so would be to choose her goals for her, to change the ends that she affirms. No matter how "rational" we may think the desire the survive, it's not our place to say different. What I think I want to say is that an agent's values and moral perspectives aren't within reach of force or coercion here - just their ability to deal with real facts, or information about the world.

  6. Dylan, very nice. The problem you construct is similar to the problem of False-Consciousness, made (in)famous by Marx in his analysis of Alienation. But it's also a problem that even the most libertarian of thinkers have grappled with (here, for example, Ayn Rand's egoism is similarly famously constrained by the self-interest in question being rational self-interest). Just about every conception of autonomy I know similarly attaches to it that the individual is rational, and, as a knowing being -- even if knowing is limited to practical knowledge and the rationality to practical rationality -- is interested in more knowledge rather than less. So what to do with someone who either lacks sufficient information (you reference her evidence to support her believe that she does not have cancer, but not the evidence the doctor provides in support of the claim that she has cancer) or denies the information available to her?
    So, what is the proper epistemic and moral response? Epistemically, we owe to her the evidence that will persuade her to reconsider her assessment of the situation. This might require epistemic and psychological counseling depending on the source of the denial. The moral situation changes though when you shift from example 1 to 2. In example 1 the question is whether to administer her treatment absent her consent. In example 2, the question is whether to correct Calvin's reach for the wrong medication. Consent is absent in both. Indeed, Calvin might actually react with affront at having his autonomy impeded by Claire as she corrects his reach. But that is different from the first case in that he is reaching for pills, wants to take them, knows that he ought, but is mistaken about which pills are the ones he needs. A more compelling comparison would be if Calvin insisted on the other bottle, yet Claire gave him those in the other bottle. His mistaken belief that they were the wrong pills would render Claire's attempt to give him the correct ones seem hostile interference and coercion, if not active poisoning. If she were to switch the bottles without his knowing, and he then correctly took the pills he ought, then no harm was done morally or epistemically, though he remains epistemically challenged.
    Here's the common thread then. In my modification of the Calvin example, he takes the correct pills though was ignorant of the switch, so he remains ignorant of the fact that the bottle he reached for was the wrong bottle at the time of the reach. That remains uncorrected. Mrs. Jones, however, refuses treatment, believes she is not getting treatment, and remains convinced that she does not have cancer, all mistakenly. Her secret treatment does nothing to remedy her epistemic condition. It also, therefore, does nothing to remedy her failed autonomy. She believes herself to be autonomous only because she is ignorant of the Dr deceiving her with the secret treatment. To those who value autonomy, this constitutes a harm -- she is being deceived about the veracity and efficacy of her own choices while being left in ignorance about her judgement of her own interests. You cannot claim the moral high ground in treating her cancer while doing so is premised on the continued deception and ignorance of the one whose interests you claim to be advancing. The point of autonomy is to reserve to the individual the determination and priority of values as they affect her interests. Assuming there is no evidence of psychological impairment, the first duty would be epistemic -- remedy the ignorance and false beliefs -- and with it, re-establish the conditions for autonomy in her case (also for Calvin). If she still refuses, we should ask the question why she continues not to believe, not to hang her out to dry by her own misplaced demand for autonomy nor to deceive her into doing as we think best. At least, that's what I'm thinking.

    1. Prof. Bellon, thank you for your response.

      I think you're right in regards to the secret modes of interferences I alluded to. I had initially failed to consider the temporal significance of a secret manipulation, but I can see now that changing an agent's course of action without them knowing it would allow their ignorance to thrive in the long run. We would be putting them in a worse epistemic position than we found them in. Mrs. Jones would believe that she was right to trust her own epistemic view over the doctor's, making it all the more difficult for her to tether her future choices to the correct information, or to reality. At best, I think that deception could only be justified in the short run (possibly, we could swap the pills on Calvin, and then tell him afterwards). But as I'm seeing it right now, it would seem, rather counter-intuitively, that a forceful interference would be more respectful that any other underhand tactic.

      I'm not sure exactly where this leaves us in the Jones example. I think in the racist version the issue is complicated by our understanding of her motives. I think it's possible that the goals and values at play aren't related to her health, but rather to her ideological perception of the doctor's, and her own, race – either way, it isn’t clear that we can have adequate access to her “determination and priority of values”, as you put it. Since we could probably never prove to her that our forced application of medication did anything for her, it might be best just to let her die. In the non-racist version, or perhaps Prof. Mayes' version in which Mrs. Jones has a well-justified belief on the matter, things, may be different. You mentioned that I didn't list the evidence the doctor provided to her. Let's assume that he shows her an X-ray of a tumour but Mrs. Jones believes that tumours "don't look like that". In this case, if she isn't convinced by our epistemic inquiries, we could forcefully administer some kind of treatment and then show her the appropriate evidence afterwards: another X-ray without the tumorous growth. We would still be taking a bit of a risk, and perhaps Prof. Swan’s worry about over-idealisation comes in to play: perhaps we shouldn’t assume that she will come to her senses regarding her epistemic perspective after the fact. If we were to cure her, and she did not accept the evidence for this either, then we would be harming her sense of autonomy (and she would probably lose trust in medical opinion as a result).

      The issue is complicated. But I think recognition of these elaborate concerns is useful: I don’t think we can simply conceive of an autonomous agent as one who picks x because y – there are always many other factors at play.

  7. Kyle, that's a great point. I think failure to both embrace the reality of and explore the significance of the continuum is the number one failing of a great deal of traditional philosophical inquiry.

    I don't think I am too interested in promoting solutions to problems like Dylan's by reference to an idealized version of the agent, so much as the decisions the agent would embrace under conditions more conducive to rational inquiry (though again not idealized). For example, Dylan's old lady might very well give different advice to a friend of hers in exactly that situation. We're all familiar with how easy it is to give sage advice (and I speak without irony) when we're not involved in the situation.

    The questionnaire approach would give us the ability to dictate the preferences we have when we are capable of thinking clearly to that compromised self of ours that is not capable of it. This, in some ways, is the whole point of the institution of promise-making. And it also obviously meshes with the self-binding techniques that are popular among behavioral economists, Ulysses binding himself to the mast with clear instructions to the crew not to release him under any circumstances.

    I think part of belonging to a culture that values rationality is, at the very least, to be bound by decisions you make under conditions conducive to good decision-making. We need to get clearer about what those are (again an active area of interest in economics, too) and become more comfortable with holding people to those decisions even if they repudiate them in times of crisis.