Imagine you wrote a letter to your future self ten years ago. In it, you said that if you ever became seriously ill, too ill to recognize the people you love, too ill to care for yourself, you would not want heroic measures taken to keep you alive. You did not want to be a burden. You did not want to linger.

Now imagine that ten years later, you have Alzheimer's disease. You are sitting in your favorite chair. Your caregiver is nearby. You have forgotten much of what you once knew, including, perhaps, that you ever wrote that letter. However, you are not insensate. You reach for your caregiver's hand when she moves to leave. When your grandchildren visit, something in your face lights up. In your more lucid moments, you are scared to die. You say you do not want to go anywhere.

Should the letter govern? Or should you still sit in your favorite chair?

This is not a hypothetical. It is the situation faced by families and caregivers of the more than fifty-five million people living with dementia worldwide (Kerstein, 2023). Both sides have genuine moral force. The letter represents your considered, autonomous wishes. The person in the chair is you, and you are still here, still expressing something that looks very much like a will to live.

Ronald Dworkin has argued influentially that the letter should win. A person in advanced dementia, he contends, has lost the kind of selfhood required for morally authoritative wishes. Her present preferences, however sincerely expressed, do not carry the moral weight needed to override an earlier, carefully formed directive (Dworkin, 1993). I take this argument seriously. However, I want to argue that it rests on a philosophical mistake, one that becomes visible when we examine the very tradition Dworkin appeals to for support.

My central claim is this: the moral standing of persons, on the Kantian account that best supports Dworkin's framework, does not require the capacity to narrate one's whole life. It requires something more minimal, the capacity to genuinely care about things, to hold commitments as worth having, and to recognize their loss as a loss. This is what Kant called our "rational nature," and it is what grounds human dignity. As Agnieszka Jaworska (1999) has shown through careful clinical and philosophical analysis, many Alzheimer's patients retain precisely this capacity. Connecting Jaworska's findings to the Kantian framework developed by Thomas Hill Jr. (1992) and J. David Velleman (1999) is the contribution I aim to make.

The argument unfolds as follows. I will first explain Dworkin's position and target its crucial undefended premise: that generating morally authoritative interests requires a sense of one's life as a whole. Next, I will develop a Kantian account of dignity and show that this premise is incompatible with what that account actually requires, which is something far more minimal than life-narrative coherence. Then I will examine Jaworska's account of what it means to genuinely value something, as distinct from merely desiring it. Later, I make the central argument: that Dworkin is internally inconsistent, appealing to Kantian dignity to authorize advance directives while using a non-Kantian standard to deny that same dignity to the present patient. The second claim I take out here is that Kantian dignity supports giving the past self unconditional priority; it does not, because dignity is unconditional and admits of no ranking between selves. However, I cannot make my claim without acknowledging the strongest objection to my view: that the values a dementia patient holds in her present state were not freely chosen but were dictated by the disease, and therefore lack the authority to override her earlier, deliberately formed commitments. I will then reply that this objection also undermines advance directives, since unchosen factors equally shaped those earlier values, and that the Kantian framework locates moral authority in the present structure of a value rather than in the story of how it came to be held.

Dworkin's framework begins with a distinction that is worth taking seriously, because it captures something true about human psychology. He divides our interests into two kinds. There are what he calls experiential interests, the everyday pleasures and comforts that make life feel good: enjoying a meal, the warmth of company, freedom from pain. Moreover, there are critical interests, the deeper commitments that give a life its shape and meaning: being a devoted parent, maintaining your independence, not becoming a burden to the people you love (Jaworska, 1999). 

Critical interests are not just stronger preferences. They are the values around which a person organizes her life, the things she would sacrifice comfort and convenience to honor.

The distinction matters because critical interests and experiential interests can conflict. A soldier may find military service grueling but judge that it expresses what she most deeply values, and we would be right to respect that judgment even if her day-to-day experience is miserable. A parent may sacrifice personal comfort for his children's future, and we recognize this sacrifice as meaningful rather than irrational, precisely because it flows from his deepest commitments. When critical and experiential interests diverge, Dworkin argues, we rightly give priority to the critical ones. To do otherwise would be to treat a person as a bundle of preferences rather than as someone with a life, a coherent set of values that gives the preferences their meaning (Jaworska, 1999).

So far, this is a sensible and largely compelling view. The trouble comes when Dworkin applies it to dementia patients.

He argues that people with advanced Alzheimer's can no longer form morally authoritative critical interests. Generating such interests, he says, requires a sense of one's life as a whole, an ability to situate one's present values within a narrative that connects past to future, that evaluates the overall shape of one's existence. Patients in the late stages of Alzheimer's have lost this. They cannot, in Dworkin's words, have "projects or plans of the kind that leading a critical life requires," and so they have "no contemporary opinion about their own critical interests" (Dworkin, 1993, as cited in Jaworska, 1999, p. 113). The person in the chair, expressing fear and reaching for her caregiver's hand, is only expressing experiential preferences, what feels comfortable right now. The only genuine critical interests she has are those she articulated in the letter: those from before the disease.

This conclusion is highly consequential. It explains why the letter should govern, and it does so without dismissing the present patient as a nonentity; she still has experiential interests that deserve care and attention. However, it does strip her of moral authority over the big decisions. Moreover, I think that is a mistake, one rooted not in bad intentions but in an underexamined assumption about what moral authority actually requires.

To see the mistake, we need to ask a basic question: what makes a person's interests morally authoritative in the first place?

Dworkin's answer ultimately traces back to a Kantian idea, the idea that persons have a special kind of worth, a dignity, that demands respect (Sergeant, 2025). This is the philosophical tradition that tells us people cannot simply be used as tools, that individual human lives are not merely variables in a social calculation, and that every person counts for something, independent of what others think of them. Dworkin invokes this tradition to explain why we should honor advance directives: because respecting a person means respecting her capacity for self-authorship, her ability to shape her own life according to her own values.

However, here is the problem. The Kantian tradition, carefully read, does not ground dignity in the capacity for comprehensive self-authorship. It grounds dignity in something more basic, and that more basic thing turns out to be exactly what many Alzheimer's patients still have.

Thomas Hill Jr. (1992), one of the most careful readers of Kant's ethics, argues that what Kant means by "humanity", the thing that gives us dignity, is a set of rational capacities. Not intelligence, not eloquence, not the ability to write a compelling advance directive. What Hill identifies as central is the power to set ends, to treat things as mattering, to organize one's behavior around what one cares about, to act for reasons rather than from mere instinct (Hill, 1992). Hill is deliberate about how minimal this formulation is. Kant does not say the power to set wise ends, or the power to set ends that cohere into a life plan. He says the power to set any end whatsoever. The threshold is low by design. Dignity needs to be robust, available to all rational beings, not just the most cognitively sophisticated.

There is a further point in Hill's account that I think is philosophically decisive. Kant extends dignity even to people who have behaved terribly, who have acted irrationally, immorally, and self-destructively. The reason is that dignity attaches to the capacity for rational responsiveness as such, not to any particular quality or degree of its exercise (Hill, 1992). The question is not whether you are using your rational capacities brilliantly. The question is whether you have them at all. Dworkin, by contrast, effectively conditions moral authority on the quality of cognitive performance, specifically, on whether the patient can maintain a life-narrative self-conception. This is not a Kantian move. It is the kind of move Kant was arguing against, treating persons as worthy of respect for what they can do rather than for what they are.

Velleman (1999) approaches the same point differently, but arrives at a conclusion that reinforces mine. He claims that persons have what he calls an interest-independent value, a value that inheres in them rather than accruing to them. The point is subtle but important. When we say it matters that something good happens to you, we are implicitly presupposing that you matter, that there is a value in you that makes your welfare worth caring about. That prior value, Velleman argues, is not relative to your preferences or cognitive capacities. It is the kind of value that makes a claim on anyone who encounters you, simply because you are a person. This is Kantian dignity. Moreover, crucially, it is not something that admits of degrees or that can be ranked; the dignity of your past self does not outweigh that of your present self, because dignity is unconditional (Velleman, 1999). You cannot legitimately trade one person's dignity against another's, and the past and present selves are both persons in the morally relevant sense.

Think of it this way. If a healthy, cognitively intact person says "I am scared to die and I do not want to go anywhere," we treat that as a morally weighty expression of her will. We do not override it with a document she wrote a decade ago, especially if her circumstances and values have shifted. Why should the same expression of fear carry no moral weight when it comes from a person with dementia? Dworkin answers that the dementia patient lacks the cognitive architecture to generate authoritative wishes. However, if the Kantian account of dignity, which is what grounds the moral authority of persons' wishes in the first place, does not require that architecture, then Dworkin's answer fails on its own terms.

This is where Jaworska (1999) becomes indispensable. She has done the philosophical and clinical work of showing what it actually looks like for a person to have values, rather than mere preferences, and has shown that this capacity survives dementia in many patients.

Here is the core distinction, stated as plainly as possible. Think about your craving for a particular food. It might be intense. However, if you found out tomorrow that the craving had simply disappeared, that you no longer felt it, you would probably feel relief, or at most mild indifference. You would not feel that you had lost something of yourself. Now think about a deep friendship, or a commitment to your family, or your sense of yourself as someone who does not give up on people. Imagine those were simply gone. 

That feeling would be completely different. You would feel that something had been taken from you, that you were somehow less yourself. That is what Jaworska means by a value rather than a desire (1999). Values are the commitments you would view as an impoverishment. They are the attitudes you hold as correct, not just "I want this" but "this is worth wanting."

Values, unlike cravings, are subject to rational assessment. You can be right or wrong about whether something is genuinely worth caring about. Values generate an internal pressure for consistency; if you truly value loyalty, you cannot easily value betrayal. Moreover, values are typically tied to one's sense of self-worth: you evaluate yourself by how well you live up to what you care about (Jaworska, 1999). None of this requires having a coherent life plan. A person can hold something as something worth caring about, be moved by it, and feel the prospect of losing it as a genuine loss, all without being able to situate that commitment within a narrative of her whole life.

Jaworska goes on to argue, with clinical evidence, that many Alzheimer's patients retain this capacity. She describes a patient who cannot name the day of the week, form new memories, or find her way to the bathroom, and yet who volunteers to participate in a research study because she believes she should help her fellow man (Jaworska, 1999). That is not a craving. That is a normative commitment, a sense of what is right, being expressed under conditions of significant cognitive impairment. She describes another patient who pauses mid-sentence, his voice trembling, and says that one can see Alzheimer's at work in his own failing speech (Jaworska, 1999, p. 114). He is not just uncomfortable. He values his intellectual capacity; he holds it as something worth having; and the prospect of losing it, including his awareness of already losing it, registers as a genuine loss rather than a matter of indifference.

What these patients are doing is not, on Dworkin's own terms, merely expressing a preference for comfort. The attitudes Jaworska documents have the structure of genuine values: held as correct, resistant to simple elimination, tied to a person's sense of her own worth. Here is what I think Jaworska's account implies, though she does not draw the connection herself: this structure is exactly what Hill argues Kant has in mind by the power to set ends; holding something as worth having, and acting on it for that reason. If that reading is right, then a patient who still genuinely values is, in the relevant Kantian sense, still an agent. Moreover, agents, on the Kantian view, have dignity.

Dworkin grounds his case for precedent autonomy in the Kantian idea that persons deserve respect in virtue of their capacity for self-determination. That capacity, he argues, is what makes a person's critical interests morally authoritative. However, when Dworkin then argues that this capacity is destroyed by dementia, he sets the bar for what counts as morally authoritative agency at life-narrative coherence. Moreover, here he parts ways with the very Kantian tradition he relies on. Kantian dignity, the thing that makes persons' wishes morally authoritative in the first place, does not require life-narrative coherence. It requires the capacity to hold ends as normatively correct and to act on reasons. That capacity, as Jaworska shows, is retained by many patients with dementia. They are still Kantian agents. Furthermore, Kantian agents have unconditional, incomparable dignity that cannot be ranked below that of their earlier selves.

The upshot is that Dworkin cannot consistently invoke Kantian dignity to justify honoring advance directives, only to use a non-Kantian standard to strip the present patient of the very dignity that would make her wishes count. He is, in effect, borrowing the moral authority of the Kantian framework for one purpose, authorizing the letter, and then abandoning it for another, silencing the person in the chair. 

Imagine someone who does not have dementia expressing the same things Mrs. Rogoff expresses: fear of death, attachment to her caregiver, and a wish to remain where she is. Nobody would say that person's wishes carry no moral weight. Nobody would say that a document she wrote a decade ago automatically governs. We would treat her expressed wishes as evidence of what she actually values now. The question we need to answer is: what exactly is different about the patient with dementia? Dworkin's answer is cognitive; she can no longer maintain a life-narrative self-conception. However, if that is not the Kantian threshold for moral standing, then cognitive decline in that specific dimension does not strip her of the relevant capacity. Moreover, if she retains the relevant capacity, if she still genuinely values, then her present voice has a moral claim that cannot simply be canceled by the letter. None of this means advance directives are worthless. In cases where a patient's capacity to value has genuinely collapsed, where there is no evaluative standpoint remaining, only basic sensation, the earlier directive may be the best guide we have. My point is narrower: precedent autonomy cannot be given unconditional priority on Kantian grounds, because the Kantian framework gives both the past self and the present self equal standing wherever the present self retains the capacity to value.

The most compelling objection to my argument goes like this. Even if Alzheimer's patients can retain something like genuine values, there is still a morally important asymmetry between those values and the ones expressed in the advance directive. The directive was the product of free, reflective deliberation. The patient sat down, thought hard about what mattered to her, and made a considered judgment. Her present values, by contrast, are largely the product of neurological deterioration. As the disease destroys some of her earlier commitments, the remaining ones take on greater salience, not because she reasoned her way to new priorities but because her brain's architecture for maintaining earlier commitments has been dismantled. These are not chosen values. They are disease-caused values.

This is the "losing and choosing" distinction developed in the philosophical literature (Sergeant, 2025). The idea is that there is a difference between the athlete who, after a career-ending injury, deliberately revises his values and builds a new life around different commitments, choosing new values,  and the dementia patient whose values shift because the disease has taken some of them away. The first change is authentically the person's own. The second is something that happened to her. Moreover, if we grant authority to disease-driven value changes, we may end up granting the disease itself authority over her care, which is not what anyone wants.

There is something intuitively important about the difference between choosing a new direction and being redirected by illness. However, despite its initial appeal, I think this objection fails for reasons that are both philosophically important and practically consequential. The objection rests on the premise that genuine moral authority requires that a value have been freely and reflectively chosen. However, this is a premise the Kantian framework explicitly resists. Hill (1992) and Velleman (1999) ground dignity in the present capacity for rational responsiveness, not in the causal history of how one came to be the rational agent one is. Kant extends dignity even to agents whose values were shaped by forces entirely outside their control, such as upbringing, culture, temperament, and circumstances. The question in Kantian ethics is not how you got your values. The question is whether, right now, you hold them with the normative structure of genuine values: as correct, as worth having, as something whose loss would impoverish you. If you do, your dignity demands respect, regardless of what produced the values you hold.

Moreover, here is where I think the objection becomes self-defeating. The values expressed in advance directives were also not freely chosen in any deep sense. A person's commitment to independence, or her distaste for being a burden, was shaped by her upbringing, cultural context, relationships, and temperament; none of which she selected from a neutral standpoint. She did not choose to be someone who values independence any more than the dementia patient chose to lose her earlier commitments. If the causal history of a value determines its authority, then the advance directive is in the same boat as the present patient's expressed wishes. The objection, pressed consistently, would undermine the very directives it is trying to defend.

The assumption underlying the "losing and choosing" objection is that a person's values are most authentically expressed under conditions of full cognitive health and formal deliberation, that the letter, written in careful reflection, is more genuinely her than the person in the chair. However, this assumption deserves scrutiny. There is a reasonable case that what a person values most deeply is revealed not only in her moments of greatest cognitive sophistication but also in her behavior under conditions of loss, fear, and vulnerability. When Mrs. Rogoff reaches for her caregiver's hand, cries when she leaves, and says she does not want to die, she is telling us something about what matters to her. That expression is not merely a reflex. It has the structure of a value, something held as worth having, whose loss would be an impoverishment. The idea that the formal document is more authentically her than the frightened woman in the chair is not a Kantian idea. It is an intellectualist bias, a preference for explicit, articulate expression over embodied, affective expression. Moreover, I see no good reason to accept it.

To conclude, this essay has argued that while Dworkin correctly identifies the importance of self-determination, his exclusion of moderately to severely impaired Alzheimer's patients from the sphere of moral authority is misplaced. By revisiting the Kantian roots of dignity through the lenses of Hill and Velleman, and grounding these theories in Jaworska's clinical observations, we find that the capacity to value persists far longer than Dworkin allows. 

Ultimately, if the Kantian dignity that justifies the past self's letter still resides in the person currently in the chair, then that present voice cannot be silenced by a previous directive. Dworkin is right that persons deserve respect as self-determining agents, and right that advance directives are a serious attempt to extend that self-determination into conditions where it might otherwise be lost. What he is wrong about is the claim that Alzheimer's patients in moderate to advanced stages have lost the capacity that grounds moral authority. The Kantian framework, which is the philosophical tradition best equipped to explain why persons' wishes deserve respect in the first place, does not require life-narrative coherence as the threshold for moral standing. It requires the capacity to hold things as worth caring about, to act on reasons, to recognize the loss of one's commitments as a genuine impoverishment. We see this through Thomas Hill and J. David Velleman, when we read Kant, and what Agnieszka Jaworska shows us when we look carefully at dementia patients: many of them still meet this threshold. Many of them still genuinely value.

When we put these two bodies of work together, the result is an internal critique of Dworkin's position. He borrows the authority of the Kantian framework to justify respecting advance directives, then denies that same framework's implications to the present patient. He cannot have it both ways. If Kantian dignity is what makes persons' wishes count, then Kantian dignity is also what gives the present patient's expressed wishes a moral claim that cannot be automatically overridden by a letter from her past.

The practical implication is not that advance directives should be thrown out. It is that they should be understood as one morally significant voice among others, to be weighed thoughtfully alongside the present patient's ongoing values rather than treated as automatically decisive. In the hardest cases, this means caregivers and physicians cannot simply defer to a document. They must also attend, carefully and with genuine moral seriousness, to the person in front of them.

The person in the chair is still there. The question is whether we are listening.

References

Dworkin, R. (1993). Life's dominion: An argument about abortion, euthanasia, and 

individual freedom. Knopf.

Hill, T. E., Jr. (1992). Dignity and practical reason in Kant's moral theory. Cornell University 

Press.

Jaworska, A. (1999). Respecting the margins of agency: Alzheimer's patients and the 

capacity to value. Philosophy & Public Affairs, 28(2), 105–138.

Kerstein, S. J. (2019). Hastening death and respect for dignity: Kantianism at the end of 

life. Bioethics, 33(5), 591–600.

Kerstein, S. J. (2023). Dignity, dementia and death. Kantian Review, 28, 221–237.

Sergeant, A. (2025). The problem of value change: Should advance directives hold moral 

authority for persons living with dementia? Bioethics, 39, 381–388.

Velleman, J. D. (1999). A right of self-termination? Ethics, 109(3), 606–628.