I. INTRODUCTION
Gustaf Arrhenius and Wlodek Rabinowicz (henceforth, AR), with assistance from John Broome in the preparation of a technical appendix, have argued in this journal that one pleasure, or type of pleasure, can be radically superior to another even though no pleasure in a complete and transitive ranking of pleasures descending from the higher pleasure to the lower is radically superior to the pleasure that immediately follows it.Footnote 1 A pleasure e is superior to another e′ in the radical sense ‘if and only if e is better than any number of e′-objects’, that is, better than any whole pleasure, no matter how large, composed of pleasures of the same type as e′. Under AR's definition of radical superiority, the higher pleasure e exceeds, or at least equals, in value the whole pleasure composed of the lower e′-pleasures, even if e′-pleasures are added together endlessly to form the whole.Footnote 2
AR explain that their ‘discussion is to some extent inspired by John Stuart Mill's famous claim that some pleasures are “superior in quality” to pleasures of other kinds’, although their analysis is ‘more abstract and general’ because the ‘objects [of value] that are compared need not be pleasures or even mental states’.Footnote 3 Apparently, a central purpose of their discussion is to show that one pleasure, or type of pleasure, e can be superior in quality to another e′ even though no pleasure e i in a sequence S of gradually decreasing pleasures, S′ (e1, . . . en), is superior in quality to its immediately succeeding neighbour e i+1, i = 1, . . . , n−1, where e1 = e and en = e′. If AR are correct about this, then it must be incorrect to assume that e can be superior in quality to e′ if and only if e is infinitely better than e′. Since that is an assumption that I make in my own interpretation of Mill's doctrine that some pleasures are superior in quality to others, AR suggest that my interpretation, while it might capture what Mill himself had in mind, is not the only possible interpretation of qualitative superiority, and may well be incorrect as a reading of Mill. AR are not particularly concerned about ‘Mill's own views on the matter’, however, and are willing to ‘leave that issue to the experts’.Footnote 4 The more important point, in their view, is that we can properly talk about one pleasure being superior in quality to another in the way Mill indicates, without assuming that the higher one is infinitely superior to the lower.
Remarkably, AR are mistaken about this. Their analysis is infected by an elementary confusion that renders their conclusions inapplicable to classical ethical hedonism, which claims that happiness or utility in the sense of pleasure, including freedom from pain, is the sole intrinsic value. It is undeniable that Mill is a hedonistic utilitarian. He says in Utilitarianism that ‘pleasure, and freedom from pain, are the only things desirable as ends’, and ‘all desirable things . . . are desirable either for the pleasure inherent in themselves, or as means to the promotion of pleasure and the prevention of pain’.Footnote 5 If hedonism is adopted and the objects under consideration are pleasant feelings per se, then we cannot properly talk about one pleasure being superior in quality to another unless we assume that the higher one is infinitely superior to the lower. The argument is fairly straightforward yet there is evidently a need to clarify it and point out some of its implications. Many commentators besides AR are confused about this matter and do not see why there is no other possible way to interpret Millian qualitative superiorities.Footnote 6
II. VALUE ADDITIVITY AND INDEPENDENCE
According to AR, my claim that qualitative superiority in the Millian sense is equivalent to infinite superiority depends on an assumption of ‘value additivity’:
Riley claims that, in the absence of infinite quantitative differences between higher and lower pleasures, a higher pleasure must be equal in value (‘equivalent’) to a finite number of lower pleasures. For if a higher pleasure only has a finite value, then that value sooner or later would be reached, if we started piling up lower pleasures. This reasoning is correct if value additivity is assumed, but without such an assumption it is a non-sequitur.Footnote 7
AR suggest that value additivity is a rather arbitrary assumption, and that G. E. Moore is to be applauded for his ‘important insight . . . that the phenomenon of non-additivity can . . . arise within the realm of intrinsic value’.Footnote 8 The whole value of multiple pleasures of the same type is not necessarily obtained by adding together each pleasure's value on its own, AR insist. Rather, interactions between one pleasure's value and another's may generate a non-additive whole value: ‘That the value of a whole may differ from the sum of the values of its parts is of course an idea that should be familiar to the post-Moorean value theorists.’Footnote 9
AR propose to give up not only value additivity but also the weaker assumption that any one pleasure's positive value contribution to whole value is independent of the positive value contributions made by the other pleasures comprising the whole. More precisely, this ‘independence axiom’, which is implied by value additivity but does not presuppose it, says that if a pleasure e has at least as much value as another pleasure e′, then ‘replacing e′ by e in any whole results in a whole that is at least as good’.Footnote 10 Given independence, the value of a whole must be a monotonically increasing function of the values of the pleasures making up the whole. Since this is a weaker assumption than value additivity, which means that the whole value must be a positive linear function of the values of the component pleasures, giving up independence is a more demanding sacrifice than giving up value additivity.
AR proceed to show that if the independence requirement is dropped, it becomes possible for a pleasure, or type of pleasure, e to be qualitatively superior to another e′ without assuming that e is infinitely superior to e′. The possibility emerges because as more and more pleasures of the same type are added to the whole, the ‘marginal value contribution’ of each extra pleasure can eventually begin to decrease and go to zero: ‘If this decrease is sufficiently steep, then adding extra [pleasures] will never get us above a finite value limit.’Footnote 11 Suppose that a finite value limit exists as more and more pleasures of the same type are added to a given pleasure e′. This limit, denoted as ve′ by AR, may also be termed a maximum finite value of e′, that is, max V(e′). Let the value of another pleasure, or type of pleasure, e be denoted as V(e). Evidently, e can now be qualitatively superior to e′ in AR's sense provided that V(e) ≥ max V(e′). ‘The manoeuvre of letting the marginal value contribution of extra units of a given kind of good converge to zero is, of course, quite standard, even with respect to intrinsic values.’Footnote 12
AR go on to illustrate that, when independence is dropped, what they call Millian qualitative superiority can emerge without assuming that any pleasure e is infinitely superior to another e′. Since I do not wish to contest the formal result, however, I shall not bother to repeat the simple example they provide. The fatal flaw in their analysis lies elsewhere. More specifically, it does not make sense to drop independence or even value additivity when referring to pleasant feelings of a given type in the context of classical ethical hedonism. A fortiori, qualitative superiority as AR define it cannot be what Mill means by qualitative superiority.
III. A BASIC MISTAKE
Classical ethical hedonism may be defined as the view that pleasure, including freedom from pain, is the sole ultimate good: all other values can be reduced to, and restated in terms of, pleasure, where absence of pain is deemed equal to a zero quantity of pleasure.Footnote 13 Suppose for the sake of argument that pleasure is homogeneous stuff. In other words, there are not plural kinds of pleasant feelings with intrinsically different qualities. Rather, pleasure is assumed to be a single kind of feeling whose intrinsic quality does not vary across its various sources and objects. Call this the standard version of ethical hedonism. Now, since pleasure is a homogeneous feeling and it is the sole basic value, standard ethical hedonism says that one quantity of pleasure is always just as intrinsically valuable as another equal quantity of it, whatever the source or object of the pleasure. Without loss of generality, assume that one unit of pleasure is the relevant quantity so that we can express the point by saying that one unit of pleasure is always just as valuable as any other unit of it.
Before proceeding further, it is worth emphasis that Mill explicitly recognizes this implication of standard hedonism. His texts, including Utilitarianism, cannot properly be interpreted as denying the point. As he says when explaining ‘the very meaning’ of hedonistic utilitarianism, its ‘first principle’ is ‘correctly described as supposing that equal amounts of happiness [pleasure] are equally desirable, whether felt by the same person or by different persons’.Footnote 14 If two pleasures of the same kind are ‘equal in degree’, then the principle of utility requires that the one pleasure must be ‘counted for exactly as much’ as the other.Footnote 15 If the two equal pleasures are felt by different persons, then the principle demands impartiality between the persons: ‘Bentham's dictum “everybody to count for one, nobody for more than one”, might be written under the principle as an explanatory commentary.’Footnote 16 Thus, Mill rejects Spencer's contention that the utility principle presupposes ‘the anterior principle that everyone has an equal right to happiness’. Rather: ‘If there is any anterior principle implied, it can be no other than this, that the truths of arithmetic are applicable to the valuation of happiness, as of all other measurable quantities’.Footnote 17 For precisely the same reason, Mill would reject Sidgwick's suggestion that the utility principle presupposes a prior rational intuition that different persons must be treated impartially. Rather, if their pleasures are equal in quantity and of the same kind, then impartiality follows from the very meaning of hedonistic utilitarianism.
Value additivity and independence follow logically from the standard version of ethical hedonism. There is no question of imposing these axioms from the outside, as if they were foreign elements that standard hedonism could do without. Rather, value additivity is built into the very meaning of standard hedonism. After all, one unit of pleasure contributes just as much positive intrinsic value as every other unit of pleasure does to the value of the whole pleasure comprised of the individual pleasures. There is nothing of intrinsic value besides pleasure and pleasure itself is of homogeneous quality. Thus, n units of pleasure are necessarily equal in value to the whole pleasure formed by adding the n units together. As Mill remarks, this elementary truth of arithmetic applies if standard hedonism is adopted. There is no other basic value with which the single pleasures can interact to yield a whole value that deviates from the sum of the individual pleasures.
AR's discussion fails to recognize this elementary truth, and therefore is fundamentally incoherent in the context of standard hedonism. The elementary truth is obscured by their talk of a decreasing marginal value of an extra unit of pleasure. Their talk is meaningless if the unit of pleasure itself is the sole basic value. We can only speak meaningfully of the decreasing marginal value of an extra unit of pleasure if some value other than pleasure is being referred to in the phrase ‘decreasing marginal value’. But this is merely to reject hedonism, which holds that pleasure is the sole intrinsic value.
Consistently with hedonism, we can speak meaningfully of the decreasing marginal value of an extra unit of some object like money or wine, where the value being referred to in the phrase ‘decreasing marginal value’ is pleasure itself. It seems reasonable to suppose that the marginal pleasure contributed by an extra unit of wine sooner or later begins to decrease, for instance, and that this decrease is sufficiently steep that adding extra units of wine will never get us above a finite limit of pleasure. Indeed, I made this assumption in earlier works, where I took for granted that human beings are only capable of experiencing a finite amount of any given kind of pleasant feeling.Footnote 18
Moreover, if utility is defined in hedonistic terms, as pleasure including freedom from pain, rather than in modern economic terms, as merely the numerical representation of a preference ordering without reference to the motivations or reasons for the ordering, hedonists can speak meaningfully of the decreasing marginal utility of wine or money. In fact, hedonists and modern economists can consistently combine their uses of the term ‘utility’, provided they agree to assume that any person's preferences form a complete and transitive ranking over any suitable domain of objects (wine, money, lotteries, and so forth) and that the sole ultimate motivation or justification for any person's preference ordering is his (expectation of) pleasure from the relevant objects in the domain.Footnote 19
Evidently, the tools of marginalist analysis are helpful even within a monistic theory of value like standard ethical hedonism, in which an object of pleasure or pain makes value contributions to a whole at differential rates depending on the amount of the object contained in the whole, and in which multiple objects must be traded off against one another in order to find optimal bundles of the objects which promote the intrinsic value of the whole better than, or at least as well as, any other feasible bundles. But this use of marginalism has nothing to do with the issue of Millian qualitative superiority of one pleasure, or type of pleasure, over another. One pleasant feeling cannot be qualitatively or intrinsically superior to another under standard hedonism: one unit of pleasure always has the same intrinsic value as any other unit of pleasure.
Before closing this section, it is also worth remarking that G. E. Moore himself makes no objection to the elementary truth that equal amounts of pleasure must have the same intrinsic values under standard hedonism. He rejects hedonism, however, as an instance of what he calls the ‘naturalistic fallacy’, that is, the fallacy of defining intrinsic value or ‘good in itself’ by reference to something else, including a ‘natural’ object such as a feeling of pleasure or even a ‘metaphysical’ object ‘which is only inferred to exist in a supersensible real world’.Footnote 20 Whatever the bite of the naturalistic fallacy, he argues that ‘good’ is ‘indefinable’ and ‘unanalysable’: it is a unique ‘simple quality’ that we know by intuition attaches to many different things.Footnote 21 With that view of the matter, he claims that the intrinsic value of what he calls an ‘organic whole’ or ‘organic unity’ is not equivalent to the sum of the intrinsic values of the different things that make up the parts of the whole, where he apparently takes for granted that the intrinsic value of a part is identical or proportional to the amount of the thing in question.Footnote 22
On a generous interpretation, perhaps Moore makes such a claim because he sees that the marginal value contribution of any part generally varies as the things that are the parts change in quantity. Thus, the intrinsic value of an organic whole composed of ‘beauty’ and ‘enjoyment’ might differ from the sum of the quantities of ‘beauty’ and ‘enjoyment’, for instance, because the marginal value of ‘enjoyment’ declines as the amount of enjoyment increases relative to the amount of beauty in the whole.
But it is far from obvious that Moore has marginalist analysis in mind. Rather, he seems to assert that the separate intrinsic values of n things in isolation would not add up to the intrinsic value of an organic whole comprised of those n things even if their respective quantities remained fixed.Footnote 23 Such an assertion is simply ‘contemptible nonsense’ if (as Moore informs us) ‘intrinsic value’ is a unique non-natural quality (analogous to a single kind of pleasure) that can vary only in degree. For, in that case, the good of the whole cannot differ from the sum of the goods of the parts: the good of any part is proportional to the quantity of that thing in the whole, and there are not plural kinds of goods whose interactions might produce deviations from value additivity.
IV. INTERPRETING HEDONISM
AR may have made a surprising slip in the context of standard hedonism.Footnote 24 Yet why should we conclude that Millian qualitative superiorities are necessarily infinite superiorities? Can we not speak meaningfully of qualitative superiorities without going so far as to claim that the higher pleasure, or type of pleasure, is more valuable than any finite amount of the lower? Moreover, if we can do this, might it not be possible to rescue AR's argument insofar as the value of a pleasure, or type of pleasure, can then be said to be a function of both the quantity and the quality of the pleasure? Extra units of pleasure of the same type might at some point begin to bring decreasing marginal value overall to the whole, for instance, because their marginal qualitative value contribution turns negative and more than offsets their positive marginal quantitative value contribution. This would not be standard hedonism, where the value of a pleasure, or type of pleasure, depends entirely on the quantity of the pleasure, but is it not a coherent possibility? Even if AR's argument is fatally flawed in the context of standard hedonism, perhaps it still works in the context of such a non-standard version of hedonism, where the quantity and quality of a pleasure, or type of pleasure, are balanced to give it an overall value.
Although AR say they are not particularly concerned with hedonism, it is important for interpreters of Mill, who is a hedonist, to understand why a non-standard version of hedonism along these lines is not a coherent possibility. Such a non-standard approach must either be standard hedonism in disguise or else it is not hedonism at all. A fortiori, AR's analysis cannot be rescued by trying to go beyond standard hedonism in this way. Genuine Millian qualitative superiorities cannot possibly be understood in terms of any such approach.
To avoid needless confusion, a preliminary step is to make clear that even standard hedonism can consistently accommodate some non-Millian ideas of qualitative superiority, including AR's idea, which are redundant if not misleading because they are correctly viewed as ideas of quantitative superiority. They are reducible to relationships of ‘more or less’ between distinct finite masses of a single kind of pleasant feeling. After these superfluous notions are identified, we can turn to different non-Millian ideas of qualitative superiority which are not compatible with any version of hedonism, standard or otherwise. Any attempt to go beyond standard hedonism by relying on some such non-Millian idea simply destroys hedonism. Once these points are appreciated, we can investigate how standard hedonism must be modified to yield a ‘pluralistic’ hedonism that accommodates genuine Millian qualitative superiorities.
Consider, then, a non-Millian notion of qualitative superiority which standard hedonism easily accommodates. Evidently, a single pleasant feeling, or type of pleasant feeling, can have various sources and objects. If the different sources and objects themselves are referred to as different ‘pleasures’, or types of ‘pleasures’, some of which are said to be of higher quality than others, then standard hedonism can accommodate such language even though it opens up myriad possibilities of confusion. Indeed, AR's mistake in the context of standard hedonism is probably traceable to some such confusion of the feeling of pleasure with its sources or objects.
To illustrate, assume that x, a hot shower, is a source of pleasant feeling, or type of pleasant feeling e. The feeling of pleasure can be treated as a function of the hot shower, that is: V(e) = e = f (x). More specifically, suppose that the total quantity of e depends on the heat's intensity and the shower's duration. The amount of the whole pleasure will in general rise or fall as these variables vary. As extra units of heat are added to make the water hotter and hotter, for instance, marginal pleasure may initially increase at an increasing rate, then increase at a rapidly decreasing rate until it becomes zero, and beyond that threshold turn negative – transform into marginal pain – at a rapidly increasing rate. A shower for ten minutes at 90 degrees Fahrenheit may give the most total pleasure whereas a scalding shower for ten seconds or a cold shower for ten hours both result in severe pain. These are matters of empirical observation.
Assume also that y, a bottle of Montrachet wine, is a distinct source of pleasure, or type of pleasure e′ such that: V(e′) = e′ = f (y). The total quantity of e′ might be a function of the temperature of the wine and of the number of glasses consumed. As before, marginalist analysis may help us to find the conditions under which maximum total pleasure can be expected from consuming the wine.
We may now speak of two different pleasures, or types of pleasure, in the sense that e is a pleasure of the hot shower whereas e′ is a pleasure of the wine. But the same feeling of pleasure is involved in the two cases, despite the distinct sources: a unit of pleasant feeling from the hot shower is equal in intrinsic value to a unit of pleasant feeling from the wine. Thus, if V(e′) = 20 and max V(e) = 10, for example, then we might refer to the pleasure of the wine as a pleasure of higher quality than the pleasure of the shower. But ‘qualitative superiority’ is a redundant and misleading term in this context because the superiority is purely quantitative. It is strictly correct to say that at least twice as much pleasure of the same kind can be expected from drinking the wine as compared to taking the shower. The wine is a superior source of the single feeling of pleasure to that degree.
If we can drink wine and take a hot shower at the same time, of course, then marginalist analysis can also help us to identify an optimal bundle of wine and shower that maximizes total pleasure viewed as a single kind of feeling.
Consider a second non-Millian notion of qualitative superiority which can be accommodated by standard hedonism. A given quantity of a single pleasant feeling, or type of pleasant feeling, may be produced in various ways even from the same source or object. Suppose, for instance, that the amount of pleasure e is given by the following formula:
where α is the intensity or strength of the pleasant feeling measured in number of units of feeling per unit of time, β is the duration of the feeling, η is what Bentham called the ‘fecundity’ of the pleasure measured as the amount of other pleasures to which it invariably gives rise, and λ is what he called its ‘purity’ measured as the extent to which it is mixed inseparably with pain that must be subtracted from the whole. Clearly, the same amount of pleasant feeling may result from different combinations of these factors. One pleasure may be of mild intensity and long duration, for instance, whereas another pleasure may be of sharp intensity over a brief period.
Even though a single kind of pleasant feeling is involved, we might nevertheless distinguish between mild and enduring pleasures, or types of pleasures e, and sharp but brief pleasures, or types of pleasures e′. Moreover, in cases where we have V(e) ≥ max V(e′), we might speak of the mild and enduring pleasure as being of a higher quality than the sharp but brief pleasure (and vice versa in cases where the inequality is reversed). But, again, ‘qualitative superiority’ is a redundant and misleading term in this context because the superiority is entirely quantitative.
Yet another idea of qualitative superiority that is compatible with standard hedonism is associated with ordinalism, that is, the reliance on an ordering to measure the amounts of pleasure to be expected from some feasible set of alternatives. Suppose that we are estimating the amount of pleasure to be expected from some source or object using the formula given earlier as a guideline. In other words, we are using available information to estimate the intensity, duration, fecundity and purity of the pleasant feeling to be expected from given quantities of, say, a hot shower. True, when dealing with objects such as a hot shower, we must recognize that the quantity of the object may increase in various ways, for example, more heat may be injected into a given volume of water or the shower may be run for a longer period at a given temperature. Moreover, as indicated earlier, marginal pleasure may increase or decrease at various rates as the intensity of the heat or the duration of the shower varies. But, in any case, our estimates of the amounts of pleasure to be expected from different amounts of the object will in principle form an ordering. This ordering ranks the estimated amounts of pleasure such that the highest estimate sits at the top, followed by the second-highest estimate, and so forth down to the lowest estimate (which may well be negative) at the bottom. The ultimate goal of a hedonist is to attain the most possible pleasure, that is, a top-ranked alternative. Thus, the ordering is in effect a hedonistic preference ordering or judgment ordering.
We might refer to such an ordering as a ‘qualitative’ ranking insofar as the information provided by the ordering per se is quite imprecise. More specifically, the ordering is merely informing us that the estimated amount of pleasure at a higher-ranked position is larger than the estimated amount at a lower-ranked position, without being precise about how much the larger estimate exceeds the lower. Even if we say that the estimated pleasure at the higher position is ‘qualitatively superior’ to the estimated pleasure at the lower position in the ordering, however, this notion of qualitative superiority obviously reduces to quantitative superiority: the higher estimate simply reflects a judgment that more pleasure of the same kind is to be expected at the higher position in the ordering.
It is worth making two further points in this connection. First, the hedonistic preference ordering can be represented by an ordinal utility function, which is unique only up to a positive monotonic transformation, where ‘utility’ is defined in modern economic terms as a numerical representation of the ordering. Second, it might be wondered why the estimated amounts of pleasure contained within the ordering do not constitute ‘utility information’ that is more precise than purely ordinal utility information. But we must avoid confusing two different meanings of ‘utility’ here. An ordinal utility function merely represents – presents again – in numerical form the ordering of the alternatives, independently of the nature of the alternatives. True, in this case, the alternatives – the estimated amounts of pleasure – which are being ordered seemingly constitute precise quantitative information about pleasure. By considering these estimates, for instance, we may learn that, according to the author of the ordering under consideration, the top-ranked estimate of pleasure is exactly twice as large as the next estimate in the ordering. But the nature of these estimates is irrelevant to the ordinal utility function, which merely represents their ordering in terms of ‘more or less’.
We have now considered three non-Millian ideas which must not be conflated with Mill's idea of qualitative superiorities. The three non-Millian ideas are impostor-ideas in any case. Each of them is correctly viewed as an idea of quantitative superiority and has nothing to do with intrinsically different kinds of pleasant feelings. These impostors can all be accommodated by standard hedonism, which regards pleasure – the sole intrinsic value – as a single kind of feeling whatever its sources or objects.
To move beyond standard hedonism, we may propose to take account of the qualitative superiority of some pleasant feelings, or types of pleasant feelings, in comparison to others. But the idea of qualitative superiority cannot be any of the impostors. Moreover, the new idea must make clear why a qualitatively superior pleasant feeling is intrinsically superior to any inferior pleasures. Now, the proposed non-standard version of hedonism under consideration seeks to introduce a ‘quality factor’ k that attaches to some pleasant feelings and distinguishes them from others such that the qualitatively superior feelings are superior to some finite degree. Without loss of generality, a unit of higher pleasure e might be said, for instance, to be ten times as valuable as a unit of lower pleasure e′, that is, V(e) = e = kV(e′) = ke′, where k = 10. But this non-standard hedonism faces insurmountable problems.
First, what is this ‘quality’ that gives such a finite boost to the intrinsic value of some pleasant feelings but not others? How does it work to make any quantity of e become ten times as valuable as the same amount of e′? We know what it means to say that a pleasure e composed of ten units of e′ is ten times more intense than the pleasure e′, for example, or that a pleasure e composed of one unit of e′ for each of ten minutes is ten times longer in duration than the pleasure e′ for one minute, given that we are talking about a single pleasant feeling, or type of pleasant feeling. But these are evidently quantitative superiorities. How does the new idea of qualitative superiority work given that the ‘quality factor’ is something different from intensity, duration, fecundity and other familiar quantitative factors? Hedonists must provide an explanation from within the phenomenon of pleasure. Otherwise the reference to qualitative superiority is a baseless move that merely distorts the genuine quantities of e and e′.
Second, ignoring the issue of what ‘quality’ means in this context, ‘quality’ is assumed to give a finite boost to the intrinsic value of pleasant feelings. A unit of the qualitatively superior feeling is therefore equivalent in intrinsic value to some quantity such as ten units of the qualitatively inferior feeling. In short, this idea of qualitative superiority, whatever it is, is still reducible to quantitative superiority, albeit a notion of quantitative superiority that involves some as-yet-to-be-explained adjustment to the unvarnished quantities of the pleasant feelings.
Third, given that a unit of higher pleasure is equal in intrinsic value to some finite mass of lower pleasure, it is not the case that a qualitatively superior pleasure is always intrinsically more valuable than qualitatively inferior pleasures. In the example at hand, eleven units of a qualitatively inferior pleasant feeling are intrinsically more valuable than a unit of the qualitatively superior feeling whereas a unit of the superior pleasure is intrinsically more valuable than nine units of the inferior pleasure. But this does not capture what Mill demands of a higher pleasure, or kind of pleasure, as compared to a lower. As he explains, one pleasure, or type of pleasure, is of higher quality than another when ‘those who are competently acquainted with both’ place the one ‘so far above the other that they [or at least the majority among them] prefer it . . . and would not resign it for any quantity of the other pleasure [no matter how large] which their nature is capable of’.Footnote 25 For him, a qualitatively superior pleasant feeling, or kind of pleasant feeling, is intrinsically more valuable than an inferior one independently of the quantities involved: ‘a superiority in quality so far outweigh[s] quantity as to render it, by comparison, of small account’.Footnote 26 Genuine Millian qualitative superiorities are intrinsic superiorities in this strong sense – they are not affected, and cannot be altered, by changes in the relative quantities of the pleasant feelings.
In light of these considerations, the proposed non-standard version of hedonism is deeply problematic and cannot hope to capture genuine Millian qualitative superiorities. On the one hand, if a hedonistic account could be given of what ‘quality’ means in this context, then qualitative superiorities would reduce to a species of quantitative superiorities. This non-standard hedonism would then stand revealed as a disguised version of standard hedonism. Rather than continue to speak of different qualities of pleasant feelings, it would be correct to speak of different finite values of a single pleasant feeling, or kind of pleasant feeling, because the ‘quality factor’ would, like intensity or duration, merely contribute to the finite values of any pleasant feelings. These finite values would be properly viewed as the genuine quantities of the feelings, where ‘genuine quantities’ are the standard quantities adjusted to take account of the quantitative contribution made by the ‘quality factor’.
On the other hand, if no hedonistic account can be given of what ‘quality’ means in this context, as I suspect is the case, then it becomes obvious that the reference to qualitative superiority is merely a way to smuggle in some intrinsic value other than pleasure so as to modify hedonistic assessments based on quantities of pleasure. If two objects or activities are sources of equal quantities of pleasure, for instance, a hedonist cannot coherently maintain that one of the pleasures is qualitatively better than the other because of some quality inherent in the one source independently of the feelings of pleasure. That is simply to reject hedonism, which holds that pleasure, including freedom from pain, is the sole intrinsic value. The hedonist must explain what it means in terms of pleasure to say that one pleasant feeling, or type of pleasant feeling, is qualitatively superior to another.
Three general claims relating to the proper interpretation of hedonism can be drawn from my argument so far, assuming it is correct. First, standard hedonism, which assumes that pleasure is a single kind of feeling whose intrinsic quality is independent of its sources and objects, is nevertheless sufficiently flexible to accommodate various non-Millian ideas of qualitative superiority. Yet these non-Millian ideas are really impostor-ideas because they are properly ideas of quantitative superiority. Second, standard hedonism cannot accommodate genuine Millian qualitative superiorities because they involve intrinsically different kinds of pleasant feelings whose different intrinsic values cannot be reduced to quantitative differences. AR's failure to recognize this has led them to mistake standard hedonistic quantitative superiorities for Millian qualitative superiorities. But, third, a non-standard version of hedonism that seeks to accommodate Millian superiorities cannot import ‘qualities’ or values that have no explanation or justification in terms of pleasant feelings. Such a move would only destroy hedonism.Footnote 27
V. PLURALISTIC HEDONISM
Mill indicates that there is only one possible meaning that can be attached by a hedonist to the proposition that one pleasure, or type of pleasure, is ‘qualitatively superior’ to another apart from the issue of quantity: ‘If I am asked, what I mean by difference of quality in pleasures, or what makes one pleasure more valuable than another, merely as a pleasure, except its being greater in amount, there is but one possible answer.’Footnote 28 The one possible answer, I maintain, is that the higher pleasure is infinitely more valuable than the lower, that is, more valuable irrespective of the finite quantities of the two pleasures: any finite amount of the higher pleasure, however small, is more valuable than any finite amount of the lower, however large. This idea of qualitative superiority gives rise to a non-standard ‘pluralistic’ version of hedonism.
Unlike either standard hedonism or an ethical pluralism in which one of the plural irreducible values is pleasure viewed as a single feeling that feels the same independently of its sources or objects, Mill's pluralistic hedonism holds that pleasure is the sole basic value but views pleasure as a family of distinct kinds of feelings some of which are felt to be infinitely more pleasant and thus always more valuable than others. In other words, the term ‘pleasure’ refers to a vector of plural types of pleasant feelings e = [e A, e B, . . . , e K] such that the intrinsic value of the highest type V(e A) = e A is qualitatively or infinitely superior to that of the next highest type V(e B) = e B, and V(e B) is infinitely superior to V(e C), and so forth, down to the lowest type of pleasant feeling V(e K). It is an empirical question how many intrinsically different qualities of pleasant feeling exist.
Perhaps the easiest way to appreciate what this plural hedonism involves is to briefly discuss some non-hedonistic ethical theories which seem to have exerted an important influence on Mill's thinking, to wit, the theories of Aristotle and Hutcheson. These non-hedonistic theories also involve higher and lower kinds of pleasant feelings such that higher kinds are always more valuable than lower kinds.
In Book X of Nicomachean Ethics, Aristotle outlines a fixed hierarchy of different kinds of activities from highest to lowest such that the highest kind, namely, contemplative activity, is said to be the best for human beings in light of the essential nature or form of the human species, followed next by virtuous activity, and so forth, down to sensual and beastly activities in which non-human animals without intellectual capacities can also engage. Each of these kinds of activities is completed, he tells us, by its own distinctive kind of pleasure: ‘the activities of thought differ from those of the senses, and among themselves, in kind; so, therefore, do the pleasures that complete them’.Footnote 29 Indeed, ‘pleasures . . . derived from noble sources are different from those derived from base sources, and one cannot get the pleasure of the just man without being just, nor that of the musical without being musical, and so on’.Footnote 30 Moreover, it is clear that, for Aristotle, a higher kind of activity with its distinctive pleasure is always better than a lower kind. If it were possible, for instance, a life spent in continuous contemplation would be the best kind of life for a human being. By suitably developing his intellectual capacities so that he could correctly perceive and contemplate the forms of the various things, including the human species, found in the natural world, a rational individual might even experience a quasi-divine kind of happiness or flourishing (eudaimonia):
If happiness is activity in accordance with excellence, it is reasonable that it should be in accordance with the highest excellence; and this will be that of the best thing in us . . . intellect or something else . . . which is thought to be our natural ruler and guide . . . [The] activity of this [best thing] in accordance with its proper excellence will be complete happiness . . . [T]his activity is contemplative.Footnote 31
If he could do so and survive, a rational person with ‘practical wisdom’ (phronesis) would invariably prefer the higher pleasure of contemplative activity over any competing lower pleasures.Footnote 32
But a life of continuous contemplative activity is possible only for immortal gods. Such a life is too high for mortal man, who needs not only food and shelter but also members of the opposite sex to reproduce as well as comrades to help protect his family from enemies. Unless such needs are met, the human species cannot survive. Thus, rational human beings must settle for another kind of life as best, namely, a life as a member of a political community in which the individual pursues contemplative activity as far as he can consistently with the display of ‘moral excellence’ in the interactions he has with the fellow members of his community in order to meet his various bodily and emotional needs. The kind of eudaimonia associated with this best possible life for humans involves not only contemplative activity in accordance with intellectual excellence, the ‘highest excellence’, but also virtuous activity in accordance with ‘moral excellence’, a ‘secondary degree’ of excellence. Such happiness entails some blend or balance of contemplative and virtuous activities and pleasures. Even so, Aristotle stresses that ‘we must, so far as we can, make ourselves immortal, and strain every nerve to live in accordance with the best thing us’.Footnote 33 Contemplative activity is proper to man, in other words, ‘by nature best and most pleasant’ for the human species as opposed to other species of animals: ‘for man, therefore, the life according to intellect is best and pleasantest, since intellect more than anything else is man’.Footnote 34 A human life will be the more truly happy the more the individual is able to engage in contemplative activities and pleasures: ‘Happiness extends, then, just so far as contemplation does, and those to whom contemplation more fully belongs are more truly happy.’Footnote 35
Aristotle suggests that some higher kinds of pleasures are desirable in themselves because they complete excellent activities, that is, contemplative as well as virtuous activities. But he is not an ethical hedonist. For him, ‘neither is pleasure the good, nor is all pleasure desirable’.Footnote 36 Lower kinds of pleasures can be worthless or even bad. The worth of any kind of pleasure is determined by something other than pleasure itself, roughly, an objective ideal of human nature. The essential nature or form of the human species determines a fixed ranking of different types of activities and associated pleasures as more or less intrinsically valuable for human beings. As he puts it:
Now, since activities differ in respect of goodness or badness, and some are worthy to be chosen, others to be avoided, and others neutral, so, too, are the pleasures; for to each activity there is a proper pleasure . . . The pleasure proper to a worthy activity is good and that proper to an unworthy activity bad.Footnote 37
Good pleasures, those kinds which are desirable in themselves, are those ‘which appear so to . . . the good man’ who displays intellectual and moral excellence:
Whether, then, the complete and blessed man has one or more activities, the pleasures that complete these will be said in the strict sense to be pleasures proper to man, and the rest will be so in a secondary and fractional way, as are the activities.Footnote 38
Strictly speaking, the pleasures that complete contemplative activities are the most worthy for human beings, and the pleasures that complete virtuous activities are worthy to a ‘secondary’ degree. Lower kinds of pleasures may be merely neutral such as the enjoyments taken in, say, the sensations of taste, or they may be unworthy such as the pleasures taken in vicious activities.
Although he implies that higher pleasures are always more worthy than lower pleasures, Aristotle does not say that higher pleasures are incomparably or infinitely more valuable than lower ones. Rather, as we ascend the fixed hierarchy of different kinds of activities, the corresponding kinds of pleasures apparently become increasingly worthy to some finite degree. In a sense, these higher pleasures are ‘larger’ than lower pleasures, if by ‘larger’ we mean ‘larger in worth’ as opposed to ‘larger in quantity of pleasant feeling’. In this sense, the largest or greatest pleasure completes the most excellent kind of activity, namely, contemplative activity. Aristotle does speak of ‘greater’ pleasures but it is important to keep in mind that he is not talking about greater quantities of a single kind of pleasant feeling. Rather, he is employing an independent measuring rod of the worthiness of pleasure. For him, higher and lower pleasures seem to be qualitatively different feelings whose relative worth depends on their underlying sources or activities. Qualitative superiority is finite superiority in terms of an intrinsic value other than pleasant feeling per se.
Aristotle combines this non-hedonic idea of qualitative superiority with an unusual phenomenological view of pleasure according to which pleasant feelings do not vary in quantity. In his view, the pleasant feeling that completes any activity is felt as an indivisible whole whenever it is felt:
Pleasure . . . is a whole, and at no time can one find a pleasure whose form will be completed if the pleasure lasts longer . . . [O]f pleasure the form is complete at any and every time.Footnote 39
Thus, a pleasant feeling does not increase or decrease with its duration. Nor does the feeling vary in intensity – it cannot be divided into parts. Rather, the feeling is experienced as a complete indivisible whole. Its quantity is fixed. In this regard, Aristotle distinguishes between wholes and ‘movements’. Whereas a whole is complete and indivisible, a movement or motion is a ‘coming into being’ that is divisible into parts and incomplete at any time before it ends – achieves its purpose or value – if it ever does. A whole is necessarily finite but a movement may be infinite in the sense of an endless expansion or contraction of parts. Pleasure, he insists, is a whole rather than a movement: ‘there is no coming into being of seeing nor of a point nor of a unit . . . therefore there is none of pleasure either; for it is a whole’.Footnote 40 By implication, any kind of pleasant feeling is always experienced as an indivisible finite mass or unit – there is no variation between different kinds of pleasures in that respect. The higher kinds of pleasant feelings that complete the higher kinds of activities are ‘larger’ than the lower kinds insofar as the higher pleasures are of greater worth in terms of the essential nature or form of the human species. But the higher pleasures are not ‘larger’ in the sense that they embody a greater quantity of pleasant feeling than the lower pleasures do. Quantitative superiority of one pleasure, or kind of pleasure, over another has no meaning for Aristotle in terms of pleasant feeling per se.
Aristotle's unusual view of the phenomenon of pleasure was not adopted by Hutcheson in his otherwise Aristotelian System of Moral Philosophy, where it is maintained explicitly (apparently for the first time) that higher pleasures are ‘incomparably’ superior to lower ones in the sense that no quantity of the lower, however large, can be as valuable as the higher. Unlike Aristotle, Hutcheson takes for granted that any kind of pleasant feeling is divisible into parts and can vary in intensity and duration: ‘As to pleasures of the same kind, 'tis manifest their values are in a joint proportion of their intenseness and duration.’Footnote 41 But pleasure is not the good for Hutcheson any more than it is for Aristotle, and quantity of pleasant feeling plays a subsidiary role in his ethical system. Like Aristotle, he claims that different kinds of pleasures can be ordered from best to worst in terms of human ‘dignity’ or worth: ‘In comparing pleasures of different kinds, the value is as the duration and dignity of the kind jointly.’Footnote 42 The dignity or ‘beatifick quality’ of a higher kind of pleasure makes it ‘incomparably’ superior to lower kinds:
We have an immediate sense of a dignity, a perfection, or beatifick quality in some kinds, which no intenseness of the lower kinds can equal, were they also as lasting as we could wish . . . By this intimate feeling of dignity, enjoyments and exercises of some kinds . . . are incomparably more excellent and beatifick than the most intense and lasting enjoyments of the lower kinds.Footnote 43
Thus, for Hutcheson, qualitative superiority so far outweighs quantitative superiority as to render it, in comparison, of small account. But qualitative superiority is based on something other than pleasant feeling per se, to wit, an objective ideal of human nature analogous to Aristotle's ideal.Footnote 44
In Hutcheson's ethical system, the ‘moral pleasures’ of the ‘virtuous affections and actions’ are ranked as the highest kind, followed by the pleasures of the ‘sympathetick and kind feelings’ (typically confined to kin and a few close friends), and then by the pleasures of the intellect and of the imagination, with the ‘sensual pleasures’ (largely of the palate and of sex) as ‘the meanest’ kind. Within the moral kind, moreover, the pleasures of contemplating, worshiping and serving God are classified as the most valuable on the grounds that virtue is inseparable from faith in a benevolent, omniscient and omnipotent deity. (There is also a corresponding ranking of different kinds of pains down to the most unworthy kind.) This fixed ranking of different kinds of pleasures is determined by an infallible ‘moral sense’ or ‘instinct’, implanted in us by God, that immediately perceives the dignity or worth of any kind of pleasant feeling with the important caveat that its operation is dependent on our fallible powers of intellect to ascertain the kinds and amounts of pleasures reasonably to be expected from any given activities or characters. By developing his intellectual capacities and acquiring knowledge of the pleasures associated with different types of acts and dispositions, a person can learn to become a good human being in the sense of achieving the highest excellence – which, for Hutcheson, is ‘moral excellence’ – of which humans are capable: the person's innate moral sense encourages him to desire moral excellence, including worship of the deity; indicates to him the natural rights and correlative obligations which individuals have in a state of nature and which civil societies must try to better secure in establishing positive laws; and, consistently with that, directs him to choose those virtuous activities and dispositions which his developed powers of reasoning tell him will produce the greatest amount of the moral kind of pleasure and thereby promote the most extensive public happiness.Footnote 45
Again, like Aristotle, Hutcheson affirms that the ‘good’ men who display moral, aesthetic and intellectual excellence are ‘the best judges’ of the worth of ‘all enjoyments’.Footnote 46 He even subscribes to the Aristotelian view that humans can to some degree achieve a quasi-divine kind of happiness, not so much by contemplating the forms of the various natural substances which the deity has conceived but rather by sharing in God's perfect benevolence by striving to do good for others so far as is consistent with our survival and the satisfaction of our bodily and emotional needs. Indeed, it is so that humans may taste this true happiness that the deity – the god immanent in nature, Hutcheson emphasizes, not some silly projection of man in the heavens above – has given us the moral sense:
The external advantages we procure to each other by our active virtues, God could have immediately conferred by his power without any action of ours; but, such was his goodness, he chose that we should enjoy some share of that divine and honourable pleasure of doing good to others; and, by the exercise of our kind affections and by our moral faculty, we do partake of it.Footnote 47
Despite his rejection of hedonism, Hutcheson's view that some kinds of pleasures are qualitatively superior in the sense of being ‘incomparably’ more valuable than any quantity of lower kinds of pleasure may well have influenced Mill.Footnote 48 Like Hutcheson and Aristotle, Mill claims that there are plural types of pleasant feelings which can be arranged in a fixed hierarchy such that higher types are always intrinsically more valuable than lower types. Unlike his two great predecessors, though, Mill is a hedonist. Thus, he must provide a hedonistic account of qualitative superiority in this sense, showing how it can be justified and why it is compatible with the standard hedonistic idea of quantitative superiority. This requirement leads Mill to his great hedonistic innovation, to wit, the higher kinds of pleasant feelings are held to be infinitely more pleasant than the lower kinds. With this hedonistic idea of qualitative superiority as infinite superiority, he is able to consistently maintain that higher pleasures, including the pleasures of the moral sentiments and of the aesthetic emotions, are always intrinsically more valuable in terms of pleasure than lower pleasures such as mere bodily sensations of pleasure, no matter how large the finite mass of lower pleasures becomes.
In Part II of this two-part article, I shall attempt to further clarify the nature of Mill's hierarchy of different kinds of pleasant feelings, give some indication of why the infinite superiority of a higher pleasure over a lower may be reasonable, and relate his pluralistic hedonism to the structure of his utilitarian doctrine.
