The Value of Unity

Abstract

Two debates over value are nearly coeval with philosophy itself. One debate is over what is good for its own sake (intrinsically good), the other is over what contributes to an individual's welfare (“what would make this person's life go, for him, as well as possible”).1 These two debates, over “the good” and “the good-for,” are distinct, yet they have been parallel both in their leading theories and their main objections. Historically, the leading theories in both debates have been hedonist, desire-satisfaction, objective list, and perfectionist theories.2 Hedonist theories claim that what is valuable—either intrinsically or for a person—tracks only experiential quality. One challenge for hedonist theories is that malicious or otherwise anti-social pleasures may seem not to be valuable in either sense.3 Another challenge is to distinguish higher and lower pleasures: to explain how some activities are more valuable even if others are enjoyed just as much.4 Hedonist theories also fail to register non-experiential facts that seem evaluatively relevant, such as that a person has false friends.5 Similarly, hedonist theories suggest that the best life can be had by entering Nozick's experience machine, which many find implausible.6 Desire-satisfaction theories are challenged by what Chris Heathwood calls “defective desires.”7 For example, and parallel to hedonism, some people have anti-social desires, such as to see cruelty, or to be cruel. More, and again parallel to hedonism, some people have what might seem trivial or worthless desires, such as to count blades of grass.8 Other people have masochistic or imprudent or adaptive preferences that are self-destructive or self-negating.9 Moreover, some desires incorporate mistaken beliefs, which appear to vitiate the value of their satisfaction.10 Objective list theories lack explanatory power. As David Brink says, a mere list of purportedly valuable states “begins to look like a disorganized heap of goods.”11 Ben Bradley objects more sharply: a view that “does not tell us why those things are on the list or how to weight them… is not theorizing, but a refusal to theorize.”12 There seems little that proponents can say when their lists diverge; for example, Martha Nussbaum lists sexual satisfaction as a central functioning, while G.E. Moore describes sexual pleasure as “evil in itself.”13 And even where lists have elements in common, such as “success” and “love,” they may seem only to be enumerating judgments shared by those making the lists.14 Perfectionist theories attempt to explain the goods on an objective list as activities that develop or exercise characteristic human capacities.15 Yet the development or exercise of some such capacities (like gratuitously inflicting pain) are anti-social.16 Perfectionism also strains to capture the attractive idea that pleasure is good, as feeling pleasure is not a capacity that can be developed or exercised.17 Even more challenging is the idea that pain is bad, as perfectionism seems stronger as an account of value than of disvalue.18 Perfectionism may also violate a resonance constraint, when it classifies as valuable activities that are not “compelling or attractive” to those engaged in them.19 These debates over “the good” and “the good-for” have continued inconclusively for centuries.20 Each theory has well-known strengths, yet apparently incurable defects. I believe that value theory is at this impasse because we have inherited too few theoretical options at the object level. The familiar hedonist, desire-satisfaction, and perfectionist theories are too simplistic to capture many of our confident evaluative judgments, while no objective list theory can justify the elements it includes and excludes. Because these object-level theories are inadequate, metaethical debates that refer to them (e.g., between objective versus subjective theory, or about the relationships between value and reasons) can inherit their weaknesses. This article outlines a new object-level theory of intrinsic value by setting out a formal model that generates lists of ultimate goods and bads.21 This formal model is extensionally more adequate than familiar hedonist, desire-satisfaction, and perfectionist theories, while also being simpler and more fruitful than an objective list.22 Most of the article will be devoted to defining this formal model precisely and to showing that it appears to capture central features of how we reason about value. I will also interpret this model in ways that suggest that the concept that structures our reasoning about value is unity. Ultimately, I will propose that value acts as if all wills are one, meeting the world together.23 I help them fight child abuse. I help them commit child abuse. I stop them committing child abuse, etc. The model to come will show that the logics that explain such complex evaluations have a formalization that is compact and cognitively ergonomic. More, this new approach will overcome many of the challenges to the four traditional theories: explaining the badness of anti-sociality, the distinction between higher and lower pleasures, when to enter the experience machine, and so on. Specific logics of unity appear to structure a core of human evaluative reasoning, or so I will venture. Physical pleasure and acts of kindness are good in themselves, we believe, even when they lead to what is bad. Similarly, physical pain and acts of cruelty are bad in themselves, even if they lead to a greater balance of good. Why is this so? As we return to the ancient question of “the good,” we will need some modesty about what we can achieve. First, the main ambition of this article is only to describe a new object-level theory of intrinsic value. Indeed, as is typical with the four traditional theories, the ambition is even less. The ambition is only to describe an object-level theory of basic intrinsic value, to which any (non-basic) intrinsic values will stand in some valid relation.24 To illustrate, if pleasure is a basic intrinsic value then “maximum pleasure” might be an intrinsic value that stands in a mereological relation to it.25 Or again, Thomas Hurka claims that knowledge is a basic intrinsic value and that “loving knowledge” is an intrinsic value that stands in an intentional relation to it.26 Other such relations have also been proposed.27 “X is intrinsically valuable” occurs often in the literature.28 In this article we only say that for any assertion of intrinsic value to be correct, “X” would need to refer to, or stand in some valid relation to, the basic intrinsic values described here. For the second boundary on this article's scope, our study here is value theory, not moral theory. We will draw no conclusions about what is right and wrong, praiseworthy and blameworthy, or what anyone should do. Third, we also set aside political theory. Though the intrinsic disvalue of certain relations (like domination) will be discussed, there is nothing here about what institutions there should be or how scarce resources should be distributed. Fourth, the value theory here is not an ethical theory—a theory of the good person. An ethical theory will explain what character traits a person should and should not have (the virtues and vices). For example, an ethical theory will detail when a person is being attentive or callous: that is, when a person is being appropriately sensitive to the ends of others. Since the value theory here will not speak to what character traits a person should have, it will not aim to explain attentiveness or callousness.29 The theory here will explain the intrinsic value of actions like helping someone, which an ethical theory might use to explain the virtue of helpfulness. The theory here will also explain the intrinsic value of actions like taking care of someone, which an ethical theory might use to help explain the value of caring about someone that, as Stephen Darwall says, “involves a whole complex of emotions, sensitivities, and dispositions.”30 Yet the value theory here can only hope to inform such an ethical theory, not to complete it. The new theory of intrinsic value will be a species of desire-satisfaction theory. As in all such theories, good and bad will be functions of desire-satisfaction and there will be no desire-independent values. Desire-satisfaction theories of value have a distinguished historical pedigree and remain popular in philosophy and economics.31 Part of their attraction has been that, especially in matters of taste and fancy, getting what one wants often seems to be good. Desire-satisfaction theories also obey the resonance constraint, by classifying as valuable only activities that are compelling to those engaged in them.32 Yet traditional desire-satisfaction theories fail extensionally because they attempt to explain value with only a single formal function: good is just “whatever satisfies (informed, etc.) desires.” This simplistic function leaves traditional desire-satisfaction theory vulnerable to extensional objections, especially concerning anti-social desires. A single function cannot capture the phenomena of value, because the value of a satisfied desire depends on its object—particularly when its object is the satisfaction of another desire. What will distinguish the new desire-satisfaction theory from traditional versions is how it characterizes value as emerging from the logical relations that one's desires can have not only to the world, but also to others' desires and to one's own desires. The formal model to come will set out the logics of desire-satisfaction along three distinct dimensions of value: value in our relations to the world, value in our relations to each other, and value in our relations to ourselves. We might call these the “extrapersonal,” the “interpersonal,” and the “intrapersonal” dimensions of value. In addition to this three-dimensional structure, the formal model will be built with further definitions and axioms, some familiar and some novel. Early on, there will be a distinction between positive and negative desires. The middle of the article will draw out the implications of two axioms: telicism, familiar from Parfit, and telic nesting, which is new. As with any new model, some elements may initially seem surprising; yet, as the bulk of the article shows, the model fits the “data” of our confident evaluations, while remaining relatively simple. Throughout, this new theory of intrinsic value will be purely formal. What is good will turn on formal relations of our desires to the world (Section II), to others' desires (Section III), and to our own desires (Section IV). By the end, these logics of unity will yield an attractive value pluralism. The theory will affirm that many lifestyles and cultural practices are good in themselves, while confirming that bullying and racist domination are intrinsically bad. The formal model will turn on relations that define very specific senses of “unity.” The success of this model naturally raises the question of why these exact relations seem to capture our evaluative reasoning. Toward the end, I respond to this question by sketching an interpretation of the formal model, the “many-one interpretation.” This interpretation explains the structure of the formal model by drawing a parallel between interpersonal value and individual rationality. At the very end, I extend this interpretation by reworking an image from Plato's Symposium. The formal model, and its interpretation, will raise many metaethical questions. For example, like all desire-satisfaction theories, the new theory will need to explain why our phenomenology often seems to favor a “desired because good” explanation within a Euthyphro problem.33 There is also a larger set of questions for all such theories about the relations between value so understood and deontic concepts like “reasons” and “ought.”34 Instead of addressing those questions, I will conclude with the hope that philosophers will welcome having a new approach to intrinsic value to consider. Extrapersonal desires define the first dimension of intrinsic value in the formal model. Extrapersonal desires have the world as their object, for example “that I eat the chocolate,” or “that I go for a row,” or “that the Reef will survive this century.” (Here, “the world” is a technical term in the formal model, comprising any state of affairs that is not the satisfaction or dissatisfaction of a desire.) This extrapersonal dimension of value follows a distinctive logic of unity, which can be described quickly because it resembles traditional (single-function) desire-satisfaction theory. We begin with Derek Parfit's notion of a “telic” desire, which is a desire for some state of affairs as an end, or for its own sake.35 One axiom of unity theory is that the satisfaction of a (positive) telic extrapersonal desire has intrinsic value. Take the example of physical sensations. We define a “pleasing” sensation as one that is the object of a telic desire: a sensation that one wants to have for its own sake. It follows that having a sensation that is the object of such a desire is good. Having pleasing sensations is intrinsically good because one is experiencing bodily sensations that one wants to have for their own sake. Many desires for bodily sensations are ingrained into our common animal nature. Other such desires vary widely across persons. On this account, the value of sensations tracks a person's desires: it is good that you experience the taste of chocolate only if you want to experience the taste of chocolate.36 We will call a satisfied telic extrapersonal desire a “basic unity.” What of the badness of physical pain? Here it is useful to revive a concept of Bernard Williams, of a “negative desire”: a desire that something not occur.37 As a pleasing sensation is one that is the object of a telic positive desire, so an aversive sensation is one that is the object of a telic negative desire. We extend the axioms to say that, just as the satisfaction of a positive extrapersonal desire has intrinsic value, so the dissatisfaction of a negative extrapersonal desire has intrinsic disvalue. In a “basic disunity,” one gets what one wants not to get. Aversive sensations are bodily sensations that, in themselves, one wants not to have. Thus the experience of an aversive sensation (like physical pain) is intrinsically bad.38 The zero point of extrapersonal value is the dissatisfaction of a positive desire (when one does not get what one wants) and the satisfaction of a negative desire (when one does not get what one wants not to get). In these cases, neither a basic unity nor disunity is formed, and (it is axiomatic that) no intrinsic value is present. The satisfaction of a positive telic extrapersonal desire has value, the dissatisfaction of a negative telic extrapersonal desire has disvalue. In contrast to perfectionist theories, the most attractive theses of hedonism, about the intrinsic value of physical pleasure and pain, follow directly from the definitions and axioms so far.39 Extrapersonal desires need not have sensations as their objects: the objects of a person's world-oriented desires can be states of affairs beyond her current or possible experience. One can want to see the cherries blossom next spring, for example, or want that one's book stay in print after one's death. Any such desire may integrate some factually incorrect beliefs, which raises a classic problem: is what satisfies a desire based on false beliefs good or not? Parfit advances a thesis that handles desires based on false beliefs about means. Parfit says that only what satisfies telic desires matters—only what is wanted for its own sake.40 Call this thesis “telicism” for intrinsic value. Say a woman falsely believes that this train will take her to her lover. Then, by telicism, her taking this train will not help to satisfy her telic desire, and will have no value in itself. And yet, as Parfit says, one may also have telic desires with false beliefs incorporated in their end.41 Consider an ancient Canaanite, say, who wants to worship the god Baal. Or consider Arneson's wife, who (Arneson imagines) wants to construct a huge monument to his virtue, which she wildly overestimates.42 Some theorists move from such examples to “informed desire” (or “ideal advisor”) theories, which attempt to filter out desires based on false beliefs.43 Yet we might instead stay within actual desire theory, and look to the teleological nature of desire.44 Desiring entails having a disposition to attend to an end, and to realize (or prevent) it when one believes that one can do so.45 The end of a desire sets its telos: its aim or object. The actions taken by the Canaanite or Arneson's wife cannot achieve the aims of their telic desires, because the definite descriptions within their ends (“the god Baal,” “Arneson's great virtue”) have no referents. So their actions cannot be valuable with respect to these aims, which is the correct result. What of wishing? A wish is a desire whose end the agent believes she cannot act to achieve. When an agent has a wish, her underlying disposition may keep her attention directed at the end, even as her practical reasoning keeps churning out the result that she can take no action to achieve it. When a wished-for state of affairs comes about (“I wish the sun would shine”), a basic unity is created just as when an end is achieved through action. For extrapersonal desires, it is good when one gets what one wants, whether through action or by happenstance.46 How good? Degree of value varies with strength of desire. A stronger desire disposes an agent to attend more to things associated with its end; and a stronger desire motivates the agent more to realize or prevent the end, modulo the agent's subjective probability of attaining the end.47 The pursuit of a weak desire is easily abandoned for other desires, while a life project may be pursued across a wide range of circumstances at cost to other ends. The stronger the desire, the more valuable is the achievement of the end.48 This account of extrapersonal value captures what has attracted many to traditional desire-satisfaction theories. Here value tracks only what one finds compelling, so the account obeys the resonance constraint. Moreover, this account affirms that, in many cases, getting what one wants can be very good.49 Of course, telicism says that what is truly valuable turns on what we actually want, which can be a deep fact unknown even to us. Yet if we give ourselves poetic license for a moment and call everything that is not desire-satisfaction “the world,” we can say that, on this account, value flows from unities between will and world. Kindness is good and cruelty is bad—and why? These look to be objective, desire-independent judgments of value. I will suggest that while these judgments are indeed objective, analysis shows that they track purely formal relations among desires. There is a logic—or, if you like, a geometry—that defines the relations that desires can have to each other, and once one sees it, one sees it everywhere in everyday interpersonal evaluations. The “formal” structures our “ordinary” evaluative concepts, and many of our confident judgments of value have formal relations between desires at their core. Interpersonal desires have the desire-satisfaction of others as their objects. The logics of interpersonal desire are more complex than the logic of extrapersonal desire, because interpersonal desires are all higher-order desires, and they have two possible structures. In the simple form of the first structure, the Tau- or T-structure, the object of one person's desire is another person's desire-satisfaction. In the simple form of the second structure, the Lambda- or Λ-structure, two people share the end of acting together.50 The T-structure models what we want to do to and for each other; the Λ-structure models what we want to do with each other. We will examine the T-structure throughout most of Section III. This sub-section begins to set out its logic, the next two sub-sections test the power of this formalism to explain ordinary evaluative judgments. We then turn to multi-leveled T-structures with “telic nesting.” After this, we explore telicism in more depth and make some basic contrasts between the right and the good, before ending this Section with a “unity” interpretation of the T-structure formalism and how it introduces the Λ-structure formalism. Let us start on the T-structure with a simple two-person case. P (Patient) yearns for a certain pleasing sensation—a sensation like the warmth of the sun on his body. As we have seen, this means that P has a telic extrapersonal desire for this sensation: his desire is “that I feel S.” Now consider the second person, A (Agent), who has a telic interpersonal desire: that P have the experience that he wants to have. The object of A's telic desire is P's desire-satisfaction: A's final end is “that P feels S (because he wants to feel it).” When P feels the sun's warmth, we say that P is in unity with the world and that A is (in a different sense) in unity with P. The satisfaction of A's desire generates its own positive (interpersonal) value. To calculate the total value in a two-person case like this one, two separate analyses are needed: one “on the P (Patient) side” and one “on the A (Agent) side.” It is good “on the P side” when P experiences the sensation that he wants to have (extrapersonal value). Leaving strengths of desires aside, let us stipulate that “P feels S” has value +1, and “P does not feel S” has neutral value 0. Then it is also separately good “on the A side” when A's desire that P experience this sensation is satisfied (interpersonal value). On the A side, “P feels S” has value +1, and “P does not feel S” has value 0. The satisfaction of A's kindly desire generates a separate interpersonal goodness, to be added to the extrapersonal goodness of P's pleasure. When P experiences the sensation, the total of value (extrapersonal plus interpersonal) is +2.51 Say now that A wants P to feel an aversive sensation like pain. In this case, P wants “that I not feel S*” and A wants this negative desire of P's to be dissatisfied. As above, P's feeling the aversive sensation is bad: there is extrapersonal disvalue in P experiencing sensations that he wants not to have. “On the P side,” the dissatisfaction of P's negative desire (P feels S*) has value −1 and the satisfaction of this desire (P does not feel S*) has value 0. Yet A takes as her end the dissatisfaction of P's desire: A's will is disunified with P's will.52 Here, the satisfaction of A's desire has value −1, and its dissatisfaction has value 0. The satisfaction of A's malign desire generates its own, interpersonal, badness. When P feels the aversive sensation, the total of value is −2. Filling in more of the logical space fills out the formal model. Say that A desires that one of P's negative extrapersonal desires be satisfied: A (like P) wants P not to feel a particular pain. Then when A's compassionate desire is satisfied, the P value is 0 and the A value is +1.53 Say now that A desires that one of P's positive extrapersonal desire be dissatisfied: A (unlike P) wants P not to experience a pleasure. Then when A's spiteful desire is satisfied, the P value is 0 and the A value is −1.54 This formal model has a neat compactness. For notice that when A wants a desire of P to be satisfied, the satisfaction of A's desire generates interpersonal value whether P's desire is positive or negative. That is, whether A wants P to feel pleasure, or A wants P not to feel pain, the satisfaction of A's desire is an interpersonal unity (+1). Similarly, when A wants a desire of P to be dissatisfied, the satisfaction of A's desire has interpersonal disvalue whether P's desire is positive or negative. Whether A wants P not to feel pleasure, or A wants P to feel pain, the satisfaction of A's desire is an interpersonal disunity (−1). This is a central feature of the model, which will be important for the interpretation of “telic nesting” below. The analysis so far may bring to mind an axiological principle familiar from Hurka and others: the recursive principle (“loving the good is good, loving the bad is bad,” etc.).55 While the recursive principle may parallel some of the logic of value described here, it will not be useful in unity theory. Unity analysis turns not on attitudes (like “loving”) but on desire-satisfaction itself (“getting,” we might say). Moreover, the recursive principle's reference to “the good” can blur the mathematical values at stake; it may not distinguish “feeling pleasure” (+1) as a good from “not feeling pain” (0) as a good. As we continue to set out the formal model, there will be ever more divergences between unity theory and recursive theories, which will be tracked in the footnotes.56 Even in the partial model that we have so far, the explanatory power of the formalism is emerging. For notice how cleanly the analysis captures the conviction that it can be bad when people get what they want. Within this model, for instance, to act cruelly is to thwart for its own sake certain strong desires of another person—say, their desires not to feel pain. The objective badness of cruelty follows from the formal relations between the desires, and indeed the analysis says that the more a person wants to act cruelly and does, the worse things get. We reach this result directly, without needing to “launder out” cruel desires with an informed desire theory, and without simply stipulating that cruelty is bad with an objective list theory. In this model, the objective badness of cruelty is captured by the same formal analysis that captures the objective value of kindness, of compassion, of spitefulness, and so on. The model shows directly when it is bad that people get what they want, using the same analysis that shows when it is good that people get what they want. This formal model is, like perfectionist theories, attempting a systematic account of the elements on an objective list. Yet the model is achieving a better extensional fit with our evaluative judgments.57 The model so far appears to track some of the thin structure of our reasoning about interpersonal value. Moreover, notice how the theory is finding this logic embedded in our thick evaluative concepts as well. Acts of kindness and compassion are good in themselves, while cruel and spiteful and malicious acts are bad in themselves. These firm judgments of value are being explained in terms of merely formal relations between desires. The more ordinary value-laden concepts we can account for in this way—in terms of formal relations between desires—the more evidence we will have that this model tracks the underlying structure of our evaluative thought. Sadistic torture, for instance, is bad. On a formal analysis, sadistic torture is bad because of its combination of cruelty and pain. The torturer takes as her target the victim's aversions: the torturer wants to force her victim to experience sensations that he cannot help wanting not to have.58 The cruelty “on the A side” combines with the pain and other aversions “on the P side” to produce an action that is highly negative overall. Turning to the positive end of the value spectrum, to take care of someone involves preventing and alleviating his pain, both of which increase interpersonal and extrapersonal value. Let us for the rest of this sub-section focus on cases where the satisfaction of the patient's desires will produce positive value (e.g., the patient will achieve some telic extrapersonal end). In these cases, many ordinary interpersonal value concepts can be analyzed as modalities by which an agent furthers the satisfaction of a patient's desires. For example, successful acts of aid and support help the patient to achieve his ends. Facilitation and encouragement are interpersonally good for the same reason. Generosity and beneficence provide another with what he desires, which is in these cases good.59 All of these ordinary value concepts yield to analysis in terms of purely formal relations between desires. We can continue to find “the formal in the ordinary” with analyses of familiar cases where an agent succeeds in frustrating a patient's desires. Start with gratuitous deception. Here A instills a false belief in P in order to thwart P's end-achievement. Say P is in SoHo, trying to get home to Brooklyn and walking toward the Brooklyn Bridge. P stops to ask A whether he is on the right route. Then A's willfully deceiving P will be straightforwardly bad, as A intentionally thwarts P's desire-satisfaction. Again, the badness of A's deception is “on the A side”—to be added to the separate loss of value “on the P side” of P's getting lost in the city. It is bad in itself for A gratuitously to mislead P. The analysis of this last case says that the badness of A's lie “on the A side” turns on P not achieving P's (good) end. And this is like the badness of other things that A might do to P in order to frustrate P's ends. For example, A might obstruct or impede or injure or kill P. Here A may be aiming to cause P pain (which we know is bad); or, as with deception, A may be aiming to frustrate P's achievement of his ends. With deception, it is the false belief; with obstruction or impediment or injury or killing, it is the physical incapacity. In all these cases, A's telic desire is to cause P not to achieve P's ends, and A's success will be bad (“on the A side”) over and above the badness of P's lost end-achievement (“on the P side”). A similar analysis holds of telic coercion. In any coercion, the agent's successful threat will result in the patient satisfying at least one fewer of his desires (e.g., his desire for life only, instead of his desires for life and money). So when A coerces P for its own sake, this