The Varieties of Prudence

Abstract

We sometimes face personal choices that are so momentous they appear to give rise to an intrapersonal analogue to the non-identity problem. Where the non-identity problem presents as a problem for morality, the intrapersonal analogue presents as a problem for prudence. The analogy has been explored recently by Das and Paul, and although, as this paper argues, their analysis fails—there is no intrapersonal analogue for the non-identity problem—it functions to highlight a persistent and perplexing puzzle for prudential rationality. This paper offers its own explanation: namely, that the phenomena that motivate the purported intrapersonal problem are better accounted for by conceiving prudence as disjunctive. To this end, I sketch a theory of two varieties of prudence.

Keywords:

• transformative choice
• prudence
• well-being
• non-identity

1. Introduction

The non-identity problem prompts us to consider moral obligations to future persons who might be brought into existence by our acts.¹ It is, as such, an interpersonal problem for moral deliberation: an exceptionally fraught one that has no generally accepted solution.²

But what if the trouble didn’t stop in the moral domain? What if there were a prudential analogue for the non-identity problem: an intrapersonal non-identity problem concerned with the commission of self-regarding choices? After all, we sometimes face personal choices that are liable to transform our core values and projects: marry or join a monastery, have a child, or enlist in the space force. These are putatively transformative choices—they threaten to bring something like a new ‘self’ into being. And from the vantage point of one felt ‘self’ weighing the choice of becoming another, we might justifiably feel ourselves confronted by a dilemma for prudential choice—even, as Das and Paul (2020) argue in a recent paper, by an intrapersonal analogue for the non-identity problem.

In this paper, I scrutinise the case for describing the experience of such intrapersonal drama in terms of an intrapersonal non-identity problem—not because the analogy to the interpersonal problem is ultimately successful (I argue that it isn’t); but rather, because considering the positive case for the prudential problem and the source of its failure helps to shed new light on discord within the experience of prudence. In particular, it helps to motivate a challenge to the univocality traditionally ascribed to prudence. Indeed, as I argue, if we conceive prudence as a disjunctive category—as giving rise to at least two distinct forms of prudential care, which I term ‘thick’ and ‘thin’ prudence—then we may both justify our sense that ‘transformative choices’ pose dilemmas for prudence, and better understand the self-regarding difficulties that such choices pose.

To begin our passage to the varieties of prudence, let us first reacquaint ourselves with the basic form of the interpersonal non-identity problem, before moving to consider its intrapersonal analogue. In doing so, it might be helpful to adopt a companion—call her Emma Bovary—who shall confront fateful choices on our behalf.

2. Two Problems

2.1 The Interpersonal Non-Identity Problem

[Emma] wanted a son; he would be strong and dark; she would call him Georges, and this idea of bearing a male child was like an anticipated revenge for all the powerlessness of her past life. A man, at least, is free, free to explore all passions and all countries, to surmount obstacles, to indulge in the most exotic pleasures. But a woman is constantly thwarted. … 

She gave birth one Sunday, about six in the morning, as the sun was rising. ‘It’s a girl!’ said Charles. She turned away her head, and fainted. (Flaubert 1856/2004: 79)

Imagine that our heroine, Emma, is confronted by the opportunity to bring a new person, Berthe, into being—a person whose existence will be unavoidably flawed, but an existence worth having, nonetheless. The only alternatives that Emma has are to bring no one into existence, or to bring a non-identical person, Georges, whose life would not be unavoidably flawed, into existence in place of Berthe.³ Now, if Berthe’s existence, while flawed, is nonetheless worth having, it is hard to see how a moral objection can be raised to the act occasioning it. And yet, faced with the option to bring Georges into existence, it does seem that it would be morally wrong to prefer to bring Berthe into existence. The sting of the non-identity problem, which we feel here, is insured by two key suppositions, which we may read with application to the case at hand. The first of the suppositions is:

The Person-Affecting Principle An act can be morally wrong only if it harms someone.

So, per the Person-Affecting Principle, if we want to say that it would be morally wrong to choose to bring Berthe, rather than Georges, into existence, then we must show that doing so harms someone. Now assume, for the purposes of getting the problem in view, that the only relevant candidates for harm are Berthe and Georges. We may ask: does Emma’s choice wrong Berthe or Georges? But to answer that question, we must know whether Emma harms either of them. And this is where the second supposition, which elucidates the notion of harm, comes into play.

The Comparative Notion of Harm for Persons An act can only harm a person if it makes them worse off than they would otherwise have been.

Where Emma elects the course of action that confers existence on Berthe—an unavoidably flawed existence, but an existence that is, nonetheless, worth having—Emma does not harm Berthe, since she would not otherwise have existed, and hence could not have been better off. Nor does Emma harm Georges, for Georges does not exist, and hence cannot be better (or worse) off. So, we find ourselves confronted by a dilemma: if we wish to hold that it is morally wrong for Emma to elect to bring Berthe rather than Georges into existence, then we must drop at least one of these intuitively appealing suppositions.

Now that we have the basic features of the interpersonal non-identity problem in view, we may move to consider the reports of an intrapersonal analogue.

2.2 The Intrapersonal Non-Identity Problem

But, on catching sight of herself in the glass, she was astonished at her appearance. Never had her eyes looked so huge, so black, so unfathomable. Some subtle influence had transfigured her whole being. Again and again she told herself: ‘I’ve a lover! I’ve a lover!’ revelling in the idea as though she were beginning a second puberty. (Flaubert 1856/2004: 144)

Let’s start with an intuitive observation: some choices bring about profound changes in individuals—transformations so profound that, in their wake, some might say that a new self appears. In Das and Paul’s account, an agent’s self is constituted by the

… core values, beliefs, desires, commitments, ideals, character traits, etc., that make her who she is at that time, including the first person phenomenology that is realised by having these core values, etc. (2020: 190)⁴

Transformative choices bring about changes to these ‘self-identity conferring mental states’, and, in doing so, bring a new self into existence (and usher the old self out of existence).⁵ However, this is not to say—and we shall have cause to discuss this further—that Das and Paul believe that an individual who undergoes self-transformation has undergone death; rather, they say that the old self has simply been succeeded by a new. This position is underpinned by a key metaphysical commitment: that selves are constituent parts of persons—and it is the persistence of the person that matters for determining death (2000: 205–6).

For Emma, the possibility of profound change arises in the form of a possible love affair. Having had her baby and feeling increasingly stultified by the institutions of marriage and motherhood, Emma yearns for change: ‘Like a sailor in distress, she cast her eyes in desperation over the solitude of her life, searching the mists of the horizon for some distant white sail’ (Flaubert 1856/2004: 56). The opportunity to take a lover, Rodolphe, appears as such a white sail. But as Emma weighs the decision, a problem arises. She realises that if she takes up with Rodolphe, her life will change radically: it will change to such an extent that, on Das and Paul’s account, she will bring into existence a new self with radically different core values and traits—‘the worldly and fabulous adulteress’—that is wholly distinct from her present self—‘the bored and dutiful wife’.

Emma also recognises that if she doesn’t take up with Rodolphe, she won’t undergo this radical change, and she acknowledges that, given the claustrophobic mores of her society, all things considered, her unchanged self will be better off than the worldly adulteress. Nonetheless, if Emma chooses to take a lover, and undergoes the self-transformation, her new self, although unavoidably flawed, is likely to value its existence, and will not regret having been brought into being (although it may regret events which come to pass). So here she confronts a problem that is structurally similar to the non-identity problem, with prudential concern playing the role of moral concern. The apparent parallels between the two problems are illustrated in Table 1, as follows:

Table 1: Two Non-Identity Problems

Moral Non-Identity Problem	Prudential Non-Identity Problem
(1) Having the child is morally wrong	(1*) Taking a lover is imprudent
(2) Having the child doesn’t harm anyone	(2*) Taking a lover doesn't make your self worse off
(3) If you don’t harm anyone you don’t do anything morally wrong	(3*) If you don't make your self worse off, you don't do anything imprudent

If Emma picks option $B$—continue as before—she will remain self $b$ (the bored housewife), and if she picks option $A$—have the affair—she will become self $a$ (the adventurous adulteress). Since self $b$ would not exist if the option of the affair had been taken, self $b$ is not made worse off by having the affair; since self $a$ would not exist had Emma continued her comfortable existence, self $a$ is not made worse off by refraining. This idea is captured by the following principle:

The Comparative Notion of Harm for Selves ‘Suppose an act or event results in the existence of a future self such that (i) the future self wouldn’t have existed in the absence of the act (or omission), (ii) the future self’s existence is unavoidably flawed, and (iii) the future self’s existence is still worth having for that future self. Then, this act or event does not make things worse for that future self’ (Das and Paul 2020: 193–94).

Finally, if the following principle is also accepted, in confronting the possibility of self-transformation, we face a prudential analogue to Parfit’s Non-Identity problem:

The Self-Affecting Principle ‘An agent can rationally prefer not to perform an available act $A$ only if there exists an available act $B$ and a current or possible future self $x$ of the agent such that the expected well-being of $x$ conditional on the agent’s performing $B$ is greater than the expected well-being of $x$ conditional on the agent’s performing $A$’ (Das and Paul 2020: 193).⁶

Emma’s affair is only imprudent if she has a current or possible future self that would be made better off by the decision to refrain from the affair. But because the affair will bring about a self-transformation—creating a future self that (i) wouldn't otherwise exist, and (ii) whose existence is unavoidably flawed (iii) but still worth having for that self—it brings about a situation in which there is no current or future self which will be better off absent the affair. The future self that would exist were Emma to refrain from having the affair (self $b$) is, arguably, doing better than the future self that would result from the affair (self $a$), but these are two different selves ($b\ne a$). And, according to the Self-Affecting Principle, an act is imprudent only if it makes one of your current or future selves worse off than it would otherwise have been.

Avoiding the intrapersonal non-identity problem requires us to either. reject at least one of two putatively appealing principles (either ‘The Comparative Notion of Harm for Selves’ or ‘The Self-Affecting Principle’), or the judgement that Emma acts imprudently in undertaking the affair. But on Das and Paul’s view, a plausible theory of prudential choice is committed to both principles. So, it seems that we must reject the intuitive judgement that Emma’s affair (which will result, among other harms, in an untimely death) is imprudent—indeed, we must reject the thought that any self-transformative act is imprudent. But Das and Paul argue (and I agree) that this is implausible.

Das and Paul have pointed us to a profoundly interesting phenomenon: namely, that when faced with choices that entail a change in self, we are confronted by something akin to the non-identity problem. So, what can we say to Emma as she ponders her fateful choice? Must we accept the analogy to the non-identity problem and succumb to the prudential pessimism it motivates? My answer is no (or at least, not yet): Das and Paul’s puzzle is not an intrapersonal analogue to the non-identity problem after all.

3. How to Reject the Self-Affecting Principle

We left Emma in what appeared to be a serious existential pickle. In this section, I will argue that although she may have ninety-nine problems, an intrapersonal identity problem ain’t one. This is because the Self-Affecting Principle is clearly false, and so is not a principle that any plausible theory of prudential choice is committed to. In the original version of the non-identity problem, the Person-Affecting Principle is quite plausible, and rejecting it is costly—doing so threatens commitments that many hold integral to the moral firmament. But we cannot say either of these things of the Self-Affecting Principle. And we need not confront an intrapersonal analogue for the non-identity problem.

Of course, to say that the intrapersonal analogue fails is not to say that agents like Emma do not face a serious prudential problem when confronted by choices that are liable to transform their core values and traits. They do. The problem is to explain the experience they undergo. In what follows, I show why the intrapersonal analogue for the non-identity problem fails and examine the phenomena that motivated its development. In doing so, I come to advance a radical proposal, one that both explains the emergence of prudential dilemmas in cases of transformative choice, and that stakes out a provocative new possibility within moral thought.

But first, here is why the Self-Affecting Principle should be rejected. According to the principle, an action is imprudent only if there is a current or possible future self that it makes worse off. The principle is plausible if selves are the ultimate objects of prudential concern. But selves are not the ultimate object of prudential concern, persons are. And, therefore, the Self-Affecting Principle is not plausible—and thus there is no intrapersonal analogue of the non-identity problem. Let’s unpack this.

Selves—which, recall, are the things constituted by the core values, beliefs, desires, commitments, ideals, character traits, and so forth, that make one who one is at a given time—can be objects of prudential concern (just as one’s hands might be), but, importantly, they are not the ultimate objects of prudential concern. In saying that they are not the ultimate objects of prudential concern, what I mean is that prudential concern can be extended to something greater than the self—something to which the self stands in a relation of parthood.⁷ This thought can be motivated by means of a simple thought experiment.

Suppose, contrary to what I have said above, that the self is the ultimate or final target of prudence—that nothing beyond the self can motivate prudence. Now, consider someone—call them Lemma—who is confronted by the prospect of imminent self-transformation. By contrast with Emma, no matter what Lemma does, her current self $c$will give way to a new one: either self $d$ or self $e$ (neither of which is unavoidably flawed).

It would be strange to say that, confronted by this situation of necessary self-transformation, Lemma might as well commit suicide. Certainly, if a modern-day Solomon were to appear, metaphysical sword in hand, and offer Lemma a choice: either the continued existence of the person known as Lemma, albeit with a numerically distinct self ($d$ or $e$), or death—the end of the whole Lemma enterprise—the former option should be preferred (and arguably, prudentially preferred) to the latter. But, if selves are, as we are supposing for the sake of argument, the ultimate objects of prudential concern, then it’s hard to see how that could be. Lemma’s current self will cease to be in either case. If c is the ultimate object of her prudential concern, then she should be relatively neutral, at least in prudential terms, about the possibility of becoming self $d$, self $e$, or simply dying. Perhaps she might entertain a preference among the options, but importantly, it would be derivative of her prudential concern for her present self, $c$, which might license conclusions like: ‘dying is best, because at least then $d$ or $e$ won’t sleep with my spouse’. But self-transformation isn’t on a par with death and the prudent actor would choose to become $d$ or $e$.

This illustrates the implausibility of the Self-Affecting Principle. It is clearly imprudent for Lemma to opt for death over self-transformation, but the principle entails otherwise: Lemma has no current or future self that self-transformation will make better off, and so none of her options are, in themselves, any more prudent than the others. But notice that in an analogous interpersonal example, the Person-Affecting Principle gets the right result. Emma can give birth to one of two possible children, Georges or Marcel (neither of whom will have a life that’s unavoidably flawed), or she can elect to not have children at all. Intuitively, Emma does nothing wrong by remaining childless. And that’s exactly what the Person-Affecting Principle predicts: no person is harmed no matter what Emma does, and so all of her options are permissible. The Person-Affecting Principle is independently plausible; the Self-Affecting Principle is not. Selves do not play the vaunted role in our lives that is needed in order for the latter principle to be plausible.

How, then, should we understand Emma’s predicament as she confronts the difficult decision over whether to have the affair? What it pays to remember here, and what the Lemma thought experiment helps to illuminate, is that there is a relationship of personal identity linking self $b$ and self $a$. In Das and Paul’s language, they are both selves of the same person—Emma. If Emma takes option $B$—continue as before—she will be self $b$. If she takes option $A$—have the affair—she will be self $a$. The intrapersonal non-identity problem is posed as a problem for prudence, but the substratum of all those successive selves—what Das and Paul call the ‘person’—can serve as the ultimate target of prudential concern. So, it appears that although Emma’s life might be a bit of a mess, there is no intrapersonal non-identity problem, since the sort of identity that matters—personal identity—holds between the individual in option $B$ and the individual in option $A$.

One might object, however, that whether or not selves are the ultimate objects of prudential concern, nothing precludes one self from caring about another. So Lemma’s self $c$ can prefer self-transformation over death in virtue of caring about her (potential) future selves $d$ and $e$.

I agree that one self can manifest concern for another. The claim, though, is that this concern cannot be prudential concern. After all, by definition, the subject and the object of prudential concern must be identical, or the latter must be a part of the former (otherwise concern would not be prudential, but other-regarding—see footnote 7). If the subject of prudential concern is my present self, then a non-identical future self is non-identical with that object, ex hypothesi. So, as my present self, I cannot feel prudential concern for it. At best, I can feel other-regarding concern, and then I am confronted by the non-identity problem simpliciter, and not some analogue.

To this, it might be objected that the subject of prudential concern is not Lemma’s current self $c$, but rather Lemma herself—the person.⁸ Because future self $d$ and $e$ are (or, depending on Lemma’s choice, would be) proper parts of her person, her concern for them is prudential. One can have prudential concern for one’s proper parts.

In response, we might point out that selves are constituent parts of persons. This makes temporal and spatial parthood analogous. A constituent spatial part of myself can be a fitting object of prudential concern—I can care prudentially about damage to my hand or foot. But it does not follow that a constituent spatial part of myself is the ultimate object of prudential concern—I care about damage to this hand because it is part of my person. There is no non-identity problem when we consider, say, replacing my hand with a prosthetic.

We might say the same of selves. When I care prudentially, I might care about a future self. But it is only because that future self is a part of my person that I am able to care about it prudentially (as opposed to morally). The self, then, is not the ultimate object of prudential concern. But that completely undermines any motivation for the Self-Affecting Principle, which tells us that an agent can prudentially prefer not to perform an act only if the agent has a given current or possible future self that would be made worse off by the decision.⁹ If we accept that persons, and not selves, are the ultimate target of prudential concern, then, as the analogy to spatial parthood illustrates, prudential concern extends beyond a given self to the person, whatever its self-parts may be. But abandoning the Self-Affecting Principle as implausible also means abandoning the conclusion that it supports: that there is an intrapersonal analogue to the non-identity problem in the case of self-change.

4. A Radical Way Forward

I always enjoy an upheaval; I love a change of scene.—Emma Bovary (Flaubert 1856/2004: 72)

As things stand, the argument that there is an intrapersonal analogue to the non-identity problem can be rejected. But perhaps we remain struck by the psychological plausibility of the original statement of Emma’s situation. She appears to face a legitimate dilemma.¹⁰ And that should perplex us—for, if there is no intrapersonal non-identity problem, and as it seems that path $B$ will yield greater overall welfare than path $A$, then why should it seem that Emma faces a difficult prudential choice?

There is a way forward, but it requires a controversial move. So far, we have operated, with our interlocutors, under the standard assumption that prudence is a unified category. But what if prudence were disjunctive? What if there were (at least) two distinct forms of prudence, each apt to a different ultimate target—and together capable, in cases like Emma’s, of suggesting clashing courses of action? What if, moreover, one of these species of prudence was by nature difficult to project across changes of self? As radical as these suggestions might sound, they find significant support in quotidian experiences of choice, as we shall see.

Call the first variety of prudence, ‘thin’ prudence, and the second, ‘thick’ prudence. Each addresses a different aspect of well-being. Thick prudence is concerned with realising the goals, projects, commitments, and values that characterise the self. Thin prudence is concerned with the overall well-being of the individual that persists through self-transformations—with the person. Let’s examine thin prudence first.

Thin Prudence An act $X$ is thinly prudent for an agent if and only if there is no other available act $Y$ such that their (expected) well-being conditional on them performing $Y$ is greater than their (expected) well-being conditional on them performing $X$.¹¹

This notion of prudence relies on there being facts of the matter about how well-off a given person is over the course of their life, even if that person undergoes various self-transformations. And it takes for granted that we can make intrapersonal welfare comparisons. Although such comparisons might not be accepted by everyone, this isn’t a very controversial assumption, given that most of the most promising theories of well-being allow for interpersonal comparisons. Indeed, thin prudence can be described in terms of each of the major theories of well-being.¹² Take, for example, Hedonism. It holds that an individual’s well-being is a function of the balance of pleasure and pain they experience. Accordingly, an act is thinly prudent if Hedonism is true just in case there is nothing else that the agent could do that would result in a better pleasure/pain balance. Or consider the Desire Satisfaction view. According to that view, how well-off an agent is depends upon how well their desires are satisfied. If an act will result in more overall desire-satisfaction, it is thinly prudent to perform it.¹³ Finally, consider an Objective List view, where well-being consists in the individual having those things that it is objectively good for them to have. If an act will help secure things on the list, it is thinly prudent.

Thus, when I consider my future person qua person, and in terms of thin prudence, I insist that I want them to have lots of well-being, but I am relatively agnostic as to what that will look like in practice: my hopes for my person are ultimately abstract and nonspecific. Depending on the theory of well-being, I might aspire to ‘the greatest happiness’, ‘as many satisfied desires as possible’, or ‘friendship and significant knowledge’, but I will leave open the question of the source of that happiness, the specific nature of those desires, or details about the nature of the friendship or knowledge.¹⁴

When the twenty-year-old student is asked what they want for themselves at ninety, and they say, ‘to be healthy, safe, and happy’, they are expressing thin prudential concern. They do not know what it will be like to be the nonagenarian self that they will become, so they stick to generalities that might equally function as responses to the question, ‘what should we want for any ninety-year-old?’ Nonetheless, we may regard this as a form of prudential rather than moral concern because, despite the similarity in the terms that the student will employ in response to both questions, they are demonstrably more invested in the well-being of their ninety-year-old person, than in that of a stranger (they will, for example, be more easily persuaded to save money for their own retirement, than for that of a stranger).

Thick prudential concern, on the other hand, is concerned with one’s ‘self’: ‘the core values, beliefs, desires, commitments, ideals, character traits, etc., that make her who she is at that time’ (Das and Paul 2020: 190). As such, it takes on a rich array of objects: the fulfilment of specific values, projects, and commitments.

Unlike thin prudence, thick prudence is difficult to project across changes of self—if my future person is a different self, with different values, projects, and so forth, then it is objectionably chauvinistic to evaluate their well-being by the yardstick of the values and projects in which I am currently enrapt. Edna Ullmann-Margalit writes of a person who ‘hesitated to have children because he did not want to become the “boring type” that all his friends became after they had children’ (2006: 167). Lo and behold, when this person finally had a child, he found that he was happy, not because parenthood was better than anticipated by his old childless standards, but because his values and projects were transformed. In having a child, he acquired the predicted set of boring preferences, but he also acquired a sense of approval with himself for having that set of preferences. Note that, as each distinct self—as the childless person, and then as the parent—the man was invested as a matter of thick prudence in the realisation of his own values, commitments, and projects, but that attitude could not readily be projected between selves. The fascinating and childless hipster must disdain the boring but dependable Dad.

Thick Prudence An act $X$ is thickly prudent, for an agent at a time, if and only if there are no other acts available that better satisfy the core values, beliefs, desires, commitments, ideals, character traits, etc. that make one who one is at that particular time.¹⁵

It’s not always irrational to perform an action that promises less well-being than an alternative—so long as that act better satisfies your other goals. It’s not irrational to make sacrifices for your children or to jump on the grenade to save your comrades. Such goals might be moral or other-regarding, but they needn’t be—they can be nonmoral and self-regarding too. Doing what best realises one’s core values, projects and commitments is a matter of thick prudence. And just as morality can conflict with self-interest, doing what’s thickly prudent can conflict with thin prudence.

Emma is a case in point. Having an affair is imprudent in the thin sense since it is likely to result in less overall well-being. But having the affair also satisfies her aims for the kind of person she aspires to be—romantic, worldly, interesting—it’s prudent in the thick sense. This explains why Emma’s choice between $A$ and $B$—a choice which is wholly self-regarding—is a difficult one. For both of her options, there is an aspect of prudence that tells in favour of it.

It might be protested here that this is too quick—that this calculus actually differs for Emma addressed as self $b$ (the bored and dutiful wife). After all, self $b$ is existent. Self $a$ (the fascinating adulteress) is not. If the affair is not undertaken, then self $b$ can continue its existence. Doesn’t the investment of the existent self in its own continuation count for something in this calculus? Surely self $b$ is worse off where its existence is curtailed—just as I am worse off if I die where I would otherwise continue living. But self-change is not death, and sometimes our dearest projects are projects of self-transformation, whose success consists in the overthrow of the set of values, beliefs and commitments that commenced the process of transformation. This is an admittedly strange feature of our psychology, but a familiar one. Anyone who has ever seen a toddler straining to be seen as a ‘big kid’ (or heard their cries of profound outrage at being called a ‘baby’) appreciates that from an early age our core projects are often projects of self-transformation. For Das and Paul, these sorts of transformative choices precipitate intrapersonal non-identity problems. On my account, they have the potential to wreak havoc in our lives simply because they are big changes—changes that, at least in some cases, will present dilemmas, because prudence is not univocal and is concerned with the individual in at least two registers. When presented with the possibility of a transformative choice, prudence asks both: ‘does doing this help to realise my present hopes, dreams, and values?’, and, in relatively abstract terms, ‘will doing this contribute to the overall well-being of my person?’ And these are two questions whose answers can, as in Emma’s case, come apart.

These dilemmas might be especially prone to emerge in cases of self-transformation, but they are not restricted to them, as the twenty-year-old contemplating his ninety-year-old self illustrated. They also need not involve one’s distant future. We can undergo the relevant sort of conflict when making choices among options with immanent (or even retrospective) significance.

Take for example, an agent, call him ‘Gustav’, who is utterly invested in a present project that he especially values and to which he is deeply committed—say, writing a great novel. On the one hand, Gustav wants a life full of well-being broadly construed: a life of good health, happiness, and stable relationships. On the other hand, he is committed to pursue his art, which he can only do satisfactorily in states of passionate application—even though this makes him unhealthy, solitary, and quite a lot less happy overall than he would be in other potential professions. Gustav faces up to this dilemma by choosing to write:

Since two in the afternoon … I have been writing Bovary. I’ve reached the Big Fuck, I’m right in the middle of it. We are in a sweat, and our heart is nearly in our mouth … Just now, around six o’clock, at the moment when I wrote the phrase ‘a nervous attack,’ I was so carried away, I was making such a racket, and feeling so intensely what my little woman was feeling, that I began to fear I was about to have one myself. I stood up from my writing-table and I opened the window to calm myself down. My head was spinning. I still have sharp pains in my knees, in my back and in my head. (Flaubert 1997: 23 December 1853)

Throwing himself into writing the novel is prudent for Gustav. But he also recognises that there is a sense in which this is imprudent for him, and he frequently complains to his correspondents about the many difficulties that the writer’s life entails.

Now someone might protest that this case, and Gustav’s extreme behaviours, actually function to demonstrate that thick prudence is no prudence at all, and that we should insist on treating a version of thin prudence as the sole form of prudence. They might reason that the mere fact that Gustav (or indeed, Emma) will be happier (or have more satisfied desires, or more items on the objective list) overall, if they were simply to pursue thin prudence, is sufficient to vindicate its claim as the one true prudence.

But to reason thus is to elide the deep significance that our present projects and values and commitments have for us and the rational and prudent demand that they make on us. Imagine, for example, that Emma is offered a third potentially transformative option—a lifetime supply of Quaaludes and a funded residence in a care home. She is assured that, after taking the drugs, she will have desires that are simple to satisfy, and that, indeed, those desires will be entirely satisfied in the care home. Nonetheless, on conducting a thinly prudential evaluation and discovering that the medicalised option promises her the greatest prospect of happiness and desire-satisfaction, Emma might yet disprefer it in terms of prudence—thick prudence. As Max Hayward observes, in a discussion of clashing values, ‘aiming low and abandoning conflicting values’ might set one up for a happier and less challenging life, but it can be rational not to choose that path (Hayward 2017: 12).

On the ‘only thin prudence’ view, if abandoning your current values for different values would ultimately lead to more overall well-being (perhaps, because the resulting values would be easier to satisfy), you should abandon your current values. But, if one is always at the ready to abandon their values for ones that are easier to satisfy, then in what sense can we call them values? Surely, values must be more robust than that!

So perhaps then, we should prefer to elevate thick prudence to the status of the one true theory. At least it assures our dearly held values, commitments and projects of persistence and respect. But that is not a sympathetic prospect either. If we were to stick in a bloody-minded fashion to all of our initial values, commitments and projects, then our lives would be nasty, brutish, and short. And surely choosing a life that is nasty, brutish, and short is also, in some legible sense, imprudent.

Arguably, we need both forms of prudence—even at the occasional cost of insoluble conflict—for the distinct and competing rational perspectives they afford on our lives.

What we are concerned for in the mode of thick prudence, where the target of concern is the self and the realisation of specific goals, values, and projects, is quite different to what we are concerned for in the mode of thin prudence, when we are concerned for the achievement of very broadly construed goods. Indeed, another way of arriving at the conclusion that prudence is disjunctive might be by means of exploring dissent in the landscape of well-being. As Shelly Kagan observes about this mode of approach, ‘If we have difficulty settling upon a single conception of well-being, perhaps this is because “welfare” judgments do not actually have a single subject’ (Kagan 1994: 309).¹⁶ Now if we, like Kagan (and others) think that welfare judgements do not have a single subject, and if doing what is prudent involves doing what enhances one’s well-being, then perhaps we should expect prudence to have more than one sense.

Where prudential concern is disjunctive, it also poses conceptual issues for forming and acting upon due moral concern, perhaps especially when it comes to the matter of how we should act towards those to whom we are partial. As Joseph Raz observes,

Caring about people consists in respecting them and engaging with them in various ways. What people care about in their own life is an important guide for their friends, for those who care for them by engaging with them. (2004: 292)

Emma’s friends will care about her, and want what’s best for her, and in so doing they will consult what they know about what she wants in her own life—they will consider her core values, projects and commitments. But if what’s good for the furtherance of Emma’s projects and values is to have the affair, and if what’s good for Emma’s overall well-being as a person is not to have the affair, then the friend who is partial to Emma, and sensitive to her circumstances, may find themselves gripped by a dilemma for action. There might be two competing ways to care for Emma, corresponding to the two varieties of prudence. Doing what’s best for Emma may mean both helping her to set up a secret meeting with Rodolphe and thwarting her attempts to do so.

5. Conclusion

The distinction between thick and thin prudence acknowledges that there is a certain kind of self-regarding concern that can’t be projected into abstract evaluation of well-being, and there is another kind of concern for one’s person that can. It supplies a fruitful means of illustrating and explaining conflicts in our experience and conduct. It explains why, for example, it is intelligible to think these two things at once: ‘I’d rather die than become a fascist’ and ‘whoever I become in the future, I hope that I’m happy’. And it sheds light on why, as the transformative choice literature illustrates, we often struggle to perform effective cross-self comparisons. One variety of prudence is invested in the projects and values of the current self, jealously guards those projects and values, and thus cannot be readily projected to take up the values of past or future selves. Another variety of prudence takes overall well-being as its end, but in doing so, abstracts away from the particular investments of disparate selves.

Acknowledgements

I am very grateful to Marcus Arvan, Olivia Bailey, Nilanjan Das, Kimberly Dill, Ryan Doody, Max Khan Hayward, Megan Hyska, L.A. Paul, Geoffrey Sayre-McCord, Briana Toole, and Nikhil Venkatesh, for stimulating conversations and helpful advice. I am also indebted to two anonymous reviewers and the editors of the Australasian Journal of Philosophy for their very generous and detailed feedback. Thanks are also due to audiences at the 2022 Central Meeting of the American Philosophical Association, the 94th Joint Session of the Mind Association and Aristotelian Society, and the 2020 Eastern Meeting of the American Philosophical Association.

Disclosure Statement

No potential conflict of interest was reported by the author.

Notes

1 Parfit (1984) names the ‘non-identity problem’ in Reasons and Persons, but those who contribute to its initial characterization include Adams (1972), Parfit (1976), Schwartz (1978), Kavka (1982), and Woodward (1986).

2 This is not to say that philosophers have failed to suggest ingenious solutions, but rather to observe that, despite this valiant work, the problem persists as a widely appreciated problem. See, for surveys of extant responses: Boonin 2014 and Roberts 2019.

3 I use this example because it forces us to attend to what it is that makes an existence ‘flawed’ in most cases advanced by philosophers—and that is not any essential feature of the individual per se, but rather the anticipated reaction of those with whom they must live. This holds for the disability cases that Paul and Das develop, as well as for their touchstone—Parfit’s famous case of the juvenile parent confronting circumstances of economic deprivation.

4 They also give a technical definition: selves are temporal parts of a person, $S$, (here, they endorse a perdurantist metaphysics), that have (i) prudential status, (ii) constancy, and (iii) maximality (see Das and Paul 2020: 191–92). While, in what follows, I develop an argument that responds to Das and Paul’s account on its own terms, I ultimately want to remain agnostic about the metaphysical commitments that they adopt. When I refer to the ‘self’ and ‘person’ in developing my alternative account of the varieties of prudence, I aim only to speak to their normative significance, and to describe our moral psychology and phenomenology. But I aim to be agnostic about which metaphysical account of persistence is correct. Understand, then, my use of terms like ‘temporal part’ to be a façon de parler.

5 In other work, Paul suggests that such transformations occur where a person undergoes experiences that bring about both epistemic and personal change, where, for an experience to be epistemically transformative, it must be such that the only way to know what it is like is to have had it yourself (e.g. seeing a particular shade of blue for the first time); and, for an experience to be personally transformative, it must bring about a change in your core preferences and attitudes (e.g. visiting a factory farm and being so profoundly moved by the suffering that you encounter that you become a vegan) (2014: 15–16). Experiences that bring about self-transformation unite epistemic and personal transformations—Paul’s examples include becoming a parent, and, more extravagantly, a vampire (2014: 1, 71).

6 Although Das and Paul put the principle in terms of what an agent can ‘rationally prefer’, I think it’s better to characterise the principle in terms of prudence. After all, as many philosophers contend, what one can rationally prefer is fairly capacious. And even if it is irrational to prefer the destruction of the whole world to scraping my finger, it’s surely not always irrational to prefer an outcome in which I have less well-being to one in which I have more well-being—so long as the former contains other things that I value more highly than my welfare. It’s not irrational, for example, to make sacrifices for those that we love—but it might be imprudent.

7 One can have prudential concern for one’s limbs, but because one’s limbs are proper parts of something else that one can have prudential concern for—namely, the person—one’s limbs are not the ultimate object of prudential concern. Of course, an ultimate object of prudential concern can be a proper part of something larger than one has concern for (e.g., a family, a nation), but that concern is not prudential. To illustrate: Marta and Mia might both be members of the same soccer team. But the mere fact that they are both parts of the same thing does not establish that Marta’s concern for Mia is prudential.

8 Thanks are owed to an anonymous referee for pressing this point.

9 To be fair, Das and Paul offer another source of (indirect) motivation for the Self-Affecting Principle. They say that if it is rejected, some other principle of prudence must be adopted instead. But the three alternative views that they address—a ‘Presentist’ theory, a ‘Totalist’ theory, and an ‘Averagist’ theory—each have problems of their own. I agree that, if the Self-Affecting Principle is false (which I’ve argued it is), it would be good to offer a plausible alternative—and in what follows, I attempt to do so, forging a new path not directly countenanced in their paper.

10 There are various ways to explain why there appears to be a dilemma, and I do not claim to argue for the only possible explanation. For example, as an anonymous referee helpfully points out, we might attribute Emma’s sense of turmoil to the fact that although avoiding the affair is ultimately ‘best’ for her, there is a ‘prudential remainder’—a dimension along which this path is worse than the alternative (life remains boring!). Indeed, Emma’s sense of dissatisfaction with the prudent path may be further aggravated by the fact that the alternative—having the affair—is only imprudent due to prevailing unjust social conditions. But I would resist such an explanation: some features of the case, such as the context of social injustice, are inessential for the problem as I take it. There being a dimension along which the ‘best’ choice is worse than the alternative is arguably necessary, but pointing to it is not sufficient to explain the conflict that Emma encounters. There are many cases involving prudential remainders that lack the distinctive sort of existential drama present here.

11 I’ve put ‘expected’ in parentheses because I wish to remain neutral about whether to understand this notion of prudence objectively (and hence, without the ‘expected’) or subjectively (including ‘expected’). I think that this distinction, while interesting, is orthogonal to issues in this paper. This definition mirrors Das and Paul’s Self-Affecting Principle but drops the restriction to selves. Note that the statement explains our judgments about their Surgery I and Surgery II cases (2020: 193–94) just as well as their Self-Affecting Principle.

12 Today, theories of well-being are standardly divided into three competing camps: hedonism, desire, and objective list (Crisp 2018: 803).

13 One might worry that intrapersonal welfare comparisons are a particular problem for desire-satisfaction views, given that such views famously have trouble making sense of interpersonal comparisons (e.g., Harsanyi 1955; Hausman 1995: 2011; Weintraub 1998). See, however, Briggs (2015) who argues that ‘preference satisfaction theorists have a range of promising solutions to the puzzle of how to assign degrees of wellbeing to people who undergo transformative experiences … ’ (27).

14 This is not to say that the realization of specific projects and values will not matter for thin prudence, but because thin prudence takes whether something is best for a person’s overall well-being as its ultimate object, it can commend courses of action that frustrate specific projects or values whose realization is nonetheless prudent for the agent in a different sense.

15 Here, in referring to what makes one ‘who one is’ at a given time, I borrow language from Das and Paul’s description of the self, which I simply intend to mark, in metaphorical terms, the sense we have of being a bearer of a rich set of present attributes (like core values, etc.), without directly implicating any claims about personal identity. Note too, that thick prudence is concerned not with the self per se, but with satisfying one’s core values, beliefs, desires, commitments, ideals, traits, etc.

16 Although it is a standard assumption that well-being is unified, several eminent moral theorists, among them Shelly Kagan (1992, 1994), James Griffin (2007), Joseph Raz (1994), Thomas Scanlon (1998), and Anna Alexandrova (2017), have expressed scepticism about the unity of well-being. This scepticism has met with only limited engagement, ‘perhaps because’, as Dan Haybron observes (echoing Kagan) ‘distinguishing multiple concepts of prudential value makes the already difficult job of giving a theory of well-being harder’ (2020). Velleman (1991) and Schectman (2022) also offer glimpses of the disunity in well-being in their discussions of temporal self-experience (thanks are owed to an anonymous referee for suggesting these two resources).