1 The subjectivity argument

Many philosophers think that pleasure and displeasureFootnote ¹ essentially have value (respectively positive and negative), i.e. that value is part of what makes pleasure and displeasure what they are and nothing else (see Fine, Reference Fine1994). Value is an evaluative property, more precisely an axiological property. The value in question may be understood as value for someone (i.e. prudential value or well-being) or impersonal value (i.e. value simpliciter), but for our purposes here the distinction will not matter. The ultimate nature of value remains very controversial, but I take it for granted that value is closely related with, if not reducible to, normative reasons to promote or preserve the thing deemed as valuable.

Pleasure is usually not taken to have value in the same sense that money, for example, has value. Pleasure, it may be thought, necessarily has value. Moreover, its value does not seem to be instrumental. For example, it is not dependent on its being conducive to the subject’s health. Rather, it is the kind of value that something has independently of anything else that it helps achieve. This is why we do not need to explain ourselves when we are asked why we seek pleasure. In fact, it is tempting to go even further and argue that the value of pleasure is somehow inherent to pleasure. In other words, the value of pleasure would be intrinsic, where something has intrinsic value if and only if it has value in virtue of its intrinsic nature alone. This is in contrast with the value that a rare stamp might have, for example, which is derived from the fact that the world does not contain many other stamps of its kind. Suppose that we think that it is necessary, even essential, for a rare stamp to have value. Even so, insofar as it is in virtue of its being rare, it will not be intrinsic to it, because the rarity of a rare stamp is not an intrinsic property of it: describing the intrinsic features of the rare stamp (size, color, texture, etc.) is not enough to infer that it has value. This sense of intrinsic value thus goes beyond essential value. All intrinsic value is essential, but there is essential value that is not intrinsic. This sense of intrinsic value also goes beyond non-instrumental value.Footnote ² Instrumental value depends on the existence of a non-instrumentally good object to which it is thought to lead, so if value is intrinsic in my sense, then it is non-instrumental.

Critics of the thesis that pleasure has intrinsic value might proceed by denying that every pleasure has value, whether intrinsic or extrinsic. For example, one could argue that pleasures taken in bad things, such as the suffering of others, are not good (see e.g. Zimmerman, Reference Zimmerman1980), or that some pleasures are too superficial to be valuable (see e.g. Haybron, Reference Haybron2008). The strategy that I will focus on here, in contrast, grants that pleasure essentially has value, but still maintains that the value in question is always extrinsic. How could it be? The upshot is that pleasure would essentially involve a conative attitude (e.g. a desire) from the subject, and thus would be subjective. This objection, which can be found more or less explicitly in Sobel (Reference Sobel2005), Kagan (Reference Kagan1992) or Dorsey (Reference Dorsey2011: 180), may be formulated as follows:

1. (1) Pleasure is essentially the object of a conative attitude.

2. (2) Pleasure does not have other essential features in virtue of which it has value.

3. (3) If pleasure is essentially the object of a conative attitude and does not have other essential features in virtue of which it has value, then the putative value of pleasure is subjective.

4. (4) The putative value of pleasure is subjective.

Premise (1) paves the way for the value of pleasure deriving from it being the object of a conative attitude, while premise (2) ensures that there is nothing else in pleasure that could explain its value. If this is so, then the value of pleasure – if it has value – can only arise from it being the object of a conative attitude, which entails that its value would be subjective. What is meant by subjective here? There is a sense in which all mental states, including pleasure, are subjective because their very existence depends on their having a subject to experience them (Searle, Reference Searle2000). The sense of interest here, in contrast, has to do with their dependence on subjects’ attitudes. Following Shafer-Landau (Reference Shafer-Landau2003), we can take objectivity and subjectivity to consist (respectively) in stance-independence and stance-dependence. On this view, a property is objective if and only if its instantiation does not depend on any subject’s perspective on it, where a perspective is constituted by all the attitudes (including thoughts and desires) that the subject holds toward the property or the object that instantiates it. In contrast, a property is subjective if and only if its instantiation requires that a subject take a certain perspective on it. Not all experiences we have are subjective in this sense. A perceptual experience – say, an experience as of seeing a tree through the window – may be ours in the sense that we are the subject having it, but it is not necessarily accompanied by any attitude toward it (though it might be if we try to introspect it or if we think about it), so it does not count as subjective in the sense of interest here. Consider the example of friends. Friends are (prudentially) valuable. Now, someone is our friend only if we hold some specific attitudes – trust, familiarity, etc. – toward her. The intrinsic features of our friend (her personality, her interests) are not enough for her to have value for us, so the value of a friend can only be a subjective property, i.e. depending on the subject taking a certain perspective on her. This explains (3). From (1), (2) and (3), we can logically conclude (4), i.e. that the putative value of pleasure is subjective. Note that this trivially entails that the value of pleasure is extrinsic: subjectivity excludes intrinsicality because depending on a subject’s perspective is one way in which a property might depend on extrinsic factors.

The whole argument, then, hangs on premises (1) and (2). In the default version of the argument developed by Kagan (Reference Kagan1992) and Sobel (Reference Sobel2005), they are supported by drawing on a specific conception of pleasure. There is a classic distinction in the philosophy of pleasure between felt-affect theories, according to which pleasure just is a specific phenomenal quality of our experience (see e.g. Broad, Reference Broad2000; Smuts, Reference Smuts2011; Bramble, Reference Bramble2013), and attitudinal theories, which deny that pleasure has any phenomenal character and hold instead that pleasure is any mental state that is the object of a specific attitude (see e.g. Alston, Reference Alston and Edwards1967; Carson, Reference Carson2000; Heathwood, Reference Heathwood2007). The attitude in question is thus extrinsic to pleasure. What kind of attitude do attitudinal theorists appeal to? Because they are keen to explain why pleasure motivates us, they draw on a conative attitude, more particularly a final desire, i.e. a desire whose satisfaction is not only a means of satisfying another desire. With this in mind, we can draw on attitudinalism to complete the default version with the following premise:

(Attitudinalism) straightforwardly justifies (1) and (2) because being the object of a conative attitude is the only essential feature of pleasure according to attitudinalism.

The argument, if successful, can be used to show that the axiological significance of pleasure derives from this conative attitude: pleasure would be good only because it is desired. In practice, it has served more ambitious purposes. First, it can be used to attack ethical objectivism. Some philosophers, like Parfit (Reference Parfit, Egonsson, Josephsson, Petersson and Rønnow-Rasmussen2001, Reference Parfit2011) and Scanlon (Reference Scanlon1998), regard pleasure as a prime example of something that gives us objective normative reasons. Sobel (Reference Sobel2005) criticizes this view by endorsing attitudinalism about pleasure and maintaining that the normative reasons provided by pleasure are of the sort provided by desire, and thus do not constitute a counterexample to ethical subjectivism. Second, the argument can be used to undermine ethical hedonism specifically (Broi, Reference Broi2025). Kagan (Reference Kagan1992) appeals to it to move from hedonism to a desire-based theory of well-being: because it is the existence of a desire-like attitude which makes pleasure good, anything that has this property – i.e. any object of desire – must be good as well, which undermines hedonism.

Ultimately, the argument is only as convincing as attitudinalism about pleasure. Unfortunately, partisans of the intrinsic value of pleasure have little reason to endorse attitudinalism, as this view faces several problems. I will not rehash all of them: suffice it to remember that attitudinal theories, as they are usually conceived of, deny that pleasure has any (proprietary) phenomenal character.Footnote ³ This appears dubious at best. Most people find it obvious that pleasure is genuinely felt: when we have a pleasant experience of a massage, our experience does not have only a sensory phenomenology associated with touch, but also a properly pleasant phenomenology. In fact, it is precisely because of this phenomenal character that we are tempted to assign intrinsic value to pleasure, as many hedonists insist on (see e.g. Mendola, Reference Mendola1990; Tännsjö, Reference Tännsjö2007; de Lazari-Radek and Singer, Reference de Lazari-Radek and Singer2014). The tension between accepting that pleasure has intrinsic value and attitudinalism about pleasure is thus unlikely to trouble any partisans of the former.

It is worth noting that several recent theories, such as Lin’s (Reference Lin2018) hybrid theory or Parfit’s (Reference Parfit2011) hedonic liking theory, which can be considered as improvements on attitudinalism, do acknowledge the existence of a phenomenal character of pleasure while retaining the notion that pleasure fundamentally involves an attitude from the subject. These theories, then, could be attractive for partisans of the intrinsic value of pleasure too. Can they help save the subjectivity argument? Unfortunately they will not. Indeed, in contrast with attitudinalism as I introduced it earlier, they do not entail that pleasure is essentially the object of a conative attitude, so they cannot serve to support (1).Footnote ⁴ To better understand this, let us start with Lin’s (Reference Lin2018) hybrid theory, which he presents as follows:

Though Lin introduces his hybrid theory as a promising middle ground between felt-quality theories and attitudinal theories, it actually departs from the latter in a significant way: attitudinalism was traditionally concerned with explaining why pleasure is desired, but this is not the case of Lin’s hybrid theory.Footnote ⁵ In fact, Lin’s hybrid theory is silent on this question. It does postulate the existence of a conative attitude directed toward a state of affairs, but pleasure itself – the tokening of the favorable attitude toward a state of affairs – is not the object of any conative attitude. As a result, in contrast with attitudinalism as I introduced it earlier, Lin’s hybrid theory does not warrant (1). The same conclusion applies to Parfit’s (Reference Parfit2011) theory, which conceives of pleasure as the composite state of a hedonic liking and its object. According to Parfit, we have normative reasons to desire this composite state – when we do so, we have what Parfit calls a meta-hedonic desire. This, however, does not mean that every pleasure is actually the object of such a desire.

Overall, not only do partisans of the intrinsic value of pleasure have little reason to endorse attitudinalism, but the recent theories of pleasure building on attitudinalism, which are arguably more attractive, cannot do the job that attitudinalism does in the subjectivity argument. I therefore conclude that the subjectivity argument, in its current version, has a limited scope.

2 An improved subjectivity argument

I would like to present a new version of the subjectivity argument which applies to a wider range of conceptions of pleasure. It is more successful than the previous one because, by accommodating the existence of a phenomenal character of pleasure, it does not depend on a controversial conception of pleasure. It is not enough, then, to reject attitudinalism to definitively escape the subjectivity argument. The basic idea of this new version is that, by looking into what the phenomenal character of pleasure reduces to at a more fundamental level, we will likely find a physical or functional property that is essentially motivational. This, it turns out, is enough to “trigger” the subjectivity argument.

Let us complete the subjectivity argument with two premises that replace (Attitudinalism):

(Reductionism) is a reductionist thesis that is fairly plausible at first glance. It follows from a more general naturalistic view in philosophy of mind according to which every mental state is identical with or reducible to a physical state or a functional state (i.e. a state characterized by its causal and structural properties). There is little reason to think that pleasure would have a metaphysical status that differs from that of other mental states. In fact, many ethicists do take pleasure to be an ordinary, non-controversial natural property, which strongly suggests that they would endorse the reductionist thesis. This status is considered to favor hedonism as it would make it the most prominent naturalistic view in ethics (see e.g. Moore, Reference Moore and Baldwin1993 [1903]; Rosati, Reference Rosati2003).

(Motivational Character) supplements (Reductionism) with a more specific view: pleasure would reduce to an essentially motivational state, that is, a state that directly contributes to determining the motivational priorities of the individual, either by generating immediate dispositions for a specific class of behaviors or by modifying her short- or long-term desires. Aydede, for example, proposes the following functional role for (sensory) pleasure: a “complex modification (filtering, enhancing, biasing, amplifying, etc.) of the incoming sensory information that will causally influence […] the subject’s motivational, cognitive, and behavioral priorities in such a way that makes some of this available for the subject’s conscious thoughts and conative attitudes” (Reference Aydede2014: 130). Note that (Motivational Character) accounts for the fact that pleasure does not inevitably cause some publicly observable behavior: pleasure only disposes the subject to act in certain ways (see also Pallies, Reference Pallies2021). A monk who took a commitment not to engage in pleasurable activities might thus systematically inhibit his pleasure-based motivation because of a more intense desire to follow an ascetic life. Moreover, I take it that motivational states may have many additional features beside their motivational character: a motivational state may not be just motivational. So (Motivational Character) leaves it open whether pleasure has other features beside its motivational character.

Is (Motivational Character) plausible? Well, it is widely believed that pleasure motivates subjects, at least sometimes, to act in certain ways: having a warm and pleasant bath may motivate us to linger in the bathtub, taking pleasure in thinking about a coming trip may increase our excitement and eagerness to go for this trip, etc. This was precisely recognized by attitudinalism. However, accepting this as a defining feature of pleasure does not take us back to attitudinalism. Indeed, the latter also held that pleasure has no specific phenomenology: being the object of a conative attitude was supposed to be, at the personal level, the only feature of pleasure. In contrast, (Motivational Character) is compatible with the more plausible claim that pleasure does have a phenomenal character of pleasure that reflects, at least partially, its motivational role, in the same way, say, that the phenomenal character of an itch reflects its motivational role (i.e. the specific action-tendencies it disposes the subject to, like the disposition to scratch or rub one’s skin). In fact, this claim was already hinted at by Brandt (Reference Brandt1979), who is nowadays widely heralded as a leading partisan of attitudinalism. Brandt thought that “for an experience to be pleasant is for it to make the person want its continuation” (Reference Brandt1979: 38), but – and this is generally overlooked – he envisages that this desire might have a counterpart felt quality: “it would be consistent to affirm such a quality as the way it feels when the [motivational] causal process is going on” (38). In psychology and in affective neuroscience, the claim that pleasure is essentially motivational has largely been taken for granted (e.g. Panksepp, Reference Panksepp2005; Frijda, Reference Frijda, Kringelbach and Berridge2009). In fact, from a naturalistic perspective, it is difficult to see how pleasure could have appeared in evolution if it had no motivational character in virtue of which it contributed to guiding the subject’s actions!

Felt-affect theorists usually refrain from explicitly taking a stand on (Reductionism) and (Motivational Character). This is certainly why they have overlooked this version of the subjectivity argument. However, there is no contradiction between these premises and the notion that at the personal level it is as a felt quality that pleasure is presented (and appears good to us), so the premises are very attractive even for partisans of felt-quality theories.

In order to support (1), however, we have to go further and argue that this motivational character motivates us to seek pleasure itself. In other words, we have to endorse the following claim:

This is a fairly intuitive claim. Most philosophers would probably say that when we have a pleasant experience, we are motivated to continue, and repeat in the future, this very experience. Not only is it in line with Brandt’s view, but it also contributes to explaining why pleasure is desired, which was an important goal of attitudinalism. To the extent that this requires, at the functional level, the activation of other conative mental states, it implies that at least part of the motivational character is extrinsic to pleasure.

We might have doubts about the idea that pleasure itself is a motivational object, as this view is not fully uncontroversial. Evaluativists are keen to insist that pleasure is about representing something else as having value, and thus as “to be desired.” According to Carruthers (Reference Carruthers2018), valence (which he equates with pleasure and displeasure) is a “nonconceptual representation of badness or goodness” (Reference Carruthers2018: 6). For example, having a pleasant experience when eating a delicious dish may amount to representing the dish as good. On this view, then, the motivational aspect of pleasure is primarily directed toward a non-mental object, not pleasure itself. Whether evaluativism can make room for motivation directed toward the pleasant experience remains a vexing question, one that exceeds the scope of this article.Footnote ⁶ At any rate, denying that pleasure itself is a motivational object seems largely at odds with ordinary intuitions about pleasure. When we have an intense headache and we take an analgesic, it is explicitly our unpleasant experience itself that we wish to get rid of, not a putative bodily damage that our headache would represent (see Jacobson, Reference Jacobson2013, for an extended discussion of this point).

We might also insist that felt-quality theories are incompatible with (Motivational Character) and (Pleasure as Motivational Object). It is true that historically one major objection against them was precisely that they would fail to account for the fact that pleasure motivates subjects, leaving open the seemingly absurd possibility that pleasure be utterly undesired by the subject (Findlay, Reference Findlay1961; Sobel, Reference Sobel2005; Smithies and Weiss, Reference Smithies and Weiss2019). At best, one might think, felt-quality theories can accommodate a contingent relationship between pleasure and motivation. On this view, it might be true that evolutionary pressures have led the felt quality of pleasure to go hand in hand with motivation in human beings, but it would still be metaphysically possible for human beings to have the felt quality of pleasure without any motivation. This would be the view that complies best with Hume’s dictum, according to which if two entities – in our case the felt quality of pleasure and its motivational character – are distinct in existence, then they do not have metaphysically necessary connections (Wilson, Reference Wilson2010).

This objection, however, does not succeed because the motivational character in question is not distinct in existence from the phenomenal character of pleasure. It is an automatic, involuntary motivational character which is essential, though extrinsic, to the physical or functional state that pleasure reduces to. Accordingly, felt-affect theorists should insist that it is metaphysically impossible for human beings to have the phenomenal character of pleasure without motivation. (Motivational Character) and (Pleasure as Motivational Object), then, do not lead to violating Hume’s dictum.

In a last attempt to support the objection, we could draw on a view of phenomenal character as constituted by qualia, taken to be causally – and thus motivationally – inert. Committing to qualia in this sense would go against (Reductionism), not (Motivational Character), as they would not admit of any physical or functional reduction. But felt-quality theorists are under no pressure whatsoever to commit to qualia in this sense. All in all, there is no decisive reason for partisans of felt-quality theories to reject (Motivational Character) and (Pleasure as Motivational Object).

As for the move to (1), we need to show that the motivational aspect of pleasure does constitute a bona fide conative attitude directed toward pleasure. This is not trivial though, as we might have misgivings about whether the motivational role of pleasure really amounts to an attitude. To start with, we might argue that its location at the subpersonal level makes it impossible to count as an attitude. Attitudes, we might think, are personal-level mental entities, the kind of entities that are introspectively accessible. This objection, however, does not succeed because the motivational character of pleasure does manifest itself in phenomenal experience – thus at the personal level – under the guise of the felt quality of pleasure. Next, we might argue that it departs too much from typical conative attitudes to count as a proper one. Certainly, the motivational aspect of pleasure does not amount to a propositional attitude: having pleasure does not require the use of concepts. It is less sophisticated than ordinary propositional conative attitudes and lacks some of their typical features such as responsiveness to reasons, so the conative attitude associated with pleasure would probably be sui generis. But this does not raise any problem: what matters is that this attitude is indeed conative in the sense that it belongs to the conative genus of mental states. How much it differs from other typical conative attitudes is thus mostly irrelevant. I conclude that there is little reason to deny that the motivational character underlying the phenomenal character of pleasure qualifies as a conative attitude.

As a result, the move to premise (1) is warranted, and the subsequent steps of the argument follow unproblematically. We can infer that its value is subjective too, which finally undermines the claim that its value is intrinsic.

If this argument is convincing, it seems that the value of pleasure can no longer be thought as intrinsic or objective. At any rate, I take this to be a live option. This, however, comes with significant drawbacks. Objectivity is generally considered as an important requirement that value must satisfy. For example, ethical hedonists are particularly keen to insist on the objectivity of the value of pleasure (see e.g. Katz, Reference Katz and Sinnott-Armstrong2008; Mendola, Reference Mendola1990; Tännsjö, Reference Tännsjö1998, Reference Tännsjö2007; de Lazari-Radek and Singer, Reference de Lazari-Radek and Singer2014). Another option, then, would be to reject the starting point of the argument, namely (Reductionism). Doing this would undercut the argument and definitively secure the objectivity and intrinsicality of the value of pleasure by preventing us from equating pleasure with a subjective state at a more fundamental level. In this case, we might still hold that there are correlates of pleasure, i.e. neurobiological or functional states that tend to correlate, at least in typical human beings, with the phenomenal character of pleasure, but we would insist that the phenomenal character of pleasure is objective even though its correlates are subjective. The value of pleasure, being attached to the phenomenal character of pleasure, would be objective as well.

Of course, it would be odd to reject the reductionist thesis only for pleasure. Non-reductionism about pleasure should likely be derived from non-reductionism about phenomenal character in general. Though non-reductionism departs from the naturalistic spirit that animates most partisans of the intrinsic value of pleasure, if the improved subjectivity argument here outlined is correct it might be their only viable option.