There’s an ethical theory I’ve had interest in for some time, called ‘Desirism’ (or ‘Desire Utilitarianism’). Unfortunately, you won’t find it in any Ethics textbook, nor in any journal, nor even amongst the pages of Wikipedia. Nevertheless the theory has a substantial online presence, and is supported and elaborated by thinkers of quality (e.g. originator Alonzo Fyfe, Luke Muehlhauser) who make a strong case against more mainstream alternatives.
The theory has convinced myself and many others that a robust Moral Realism actually has a shot at being true, whereas prior to encountering it, a factual interpretation of moral statements seemed so absurd to me I assumed Realism couldn’t be true. On the other hand, the theory has never really crystallized in my head – some parts of the account seem to cohere with others, but the logical relations are not made explicit; and in still other parts the generality of the description is unsettlingly vague.
I begin with the briefest of explanations, one that will only get us so far as defining the moral good. But be warned: my explanation will be idiosyncratic, and I doubt the aforementioned thinkers will completely agree with my way of presenting it.
Explanations of Desirism typically give an account of value in general, and then proceed to an analysis of moral value in particular. The rough idea is as follows: Values imply valuers, therefore values are existentially dependent upon these valuers. But it is also true that what is valued is often independent of the valuer – the intentional object valued is part of the external world. So, values are a kind of relation between valuers and what they value, an idea which sits well with our intuitions that values are real, but that they are themselves insubstantial.
What sort of relation could that be? A clue here is the observation that values help to direct our actions in the world, along with beliefs: how we act depends on what we value. But, what is strikingly similar, so too do our actions depend upon what we desire. So it seems a good bet that desire and value are related. The Desirist explanation is that values are relations between desires and the states of affairs desired1: desires confer value, so what is desired is said to have value2. When desires are fulfilled by realizing the state of affairs desired, so too is the value realized. Symmetrically, desires can be thwarted when the desired state of affairs is prevented from being realized – think, for instance, of a party girl wanting to attend a concert, but being grounded instead – and so the value will fail to be realized.
This leads to the obvious distinction between positive and negative value. When we say that what is desired is valued, this typically means that it is valued positively. By contrast, where a desire is thwarted, that positive value fails to be realized, and a corresponding negative value is realized instead: a thwarted desire is a kind of penalty. When a desire is neither thwarted or fulfilled, no value is yet realized by it either way.
This tidy analysis allows us to begin to analyse the language of value. For instance, we know that a generic ascription of ‘good’ is one which evaluates, and so we know that some desire or set of desires are being implicitly referred to in the use of the word. We also know that, because ‘good’ is an adjective, there is some state of affairs which is being related to those desires and described by the word. Further, we know that ‘good’ bestows positive value upon a state of affairs, so that the term has to do with the realization of what is desired. Contraversly, ‘bad’ bestows a negative value, which implies that the desires fail to be realized – it thwarts them.
The term ‘good’ is highly plastic. Though sometimes deployed in an unrestricted sense – evaluating a state of affairs to all desires – it is much more common that a user has a restricted set of desires in mind. Just what desires are to be attended to is something that can be left to the context of use, or signaled by qualifying the term in various ways. One way is describe a state of affairs as “good-for-X”, e.g. “good-for-you”, which restricts the set of desires to those possessed by X. Another way is to adverbially qualify ‘good’, such as when we describe a state of affairs as “good fiscally”, or to affix a category, as when we talk of what is “good healthwise” or “good in an aesthetic sense” – this way of qualifying indicates desires by type, the type being desires to do with, e.g. finance, health, or aesthetics.
A word here is required about just what is meant by ‘desires’, before we get too far. Just as one might say that “there are many beliefs”, and one could be referring to broad kinds of belief everyone shares, or instead individual instances of belief, so too is ‘desires’ ambiguous. At least in the case of the second form of qualification, we should interpret ‘desires’ as picking out (tokened3) types rather than tokens of desire. Whatever is ‘good’ in a particular categorical sense is good because it fulfills the kinds of desires to do with the category, not good insofar as it fulfills multiple instances of a kind of desire.
But we’re interested in the ‘morally good’. That it is a species of good indicates positive value, and thus a relation of desire fulfillment. The adverbial qualification is a clue that a type of desires is being appealed to. But what type of desires? According to the Desirist, all types of desires4. I do not know of any persuasive argument for this point5, except perhaps to point out that “moral good” and “good” are terms often used synonymously, and the unrestricted scope of ‘moral good’ can perhaps be inferred from this. If there is a difference between the good and the morally good, the difference may be in that the latter picks out types, whereas the former is ambiguous as to between types and tokens.
Given that what is morally good is to be evaluated against all (tokened) desire-types, and given that it fulfills these being ‘good’, is there anything else to add? Yes, there are at least two more elements required. The first element is the Desirist thesis that moral concepts like ‘morally good’ are primarily applicable to desire-types6. So, for example, the ‘desire to give charitably’ is a desire-type, and will count as morally good if instantiation of this type fulfills the other desire-types. This primary evaluation allows derivative evaluations, most notably, the evaluations of actions according to the desire-types that we take those actions to exhibit. So, what we have here is a kind of virtue theory, based on desire-types, rather than character traits.
The second element we need is, well, subtlety. A desire-type is going to fulfill some desire-types, and thwart others, but it is not therefore both morally good and bad. Our evaluation needs to account for the preponderance of desire-fulfilling over desire-thwarting such that, all else being equal, a desire-type which fulfils more desire-types than it thwarts is a morally good desire. Further, some desire-types are stronger than others, and therefore their fulfillment is of greater value than weaker peers – we need to weight the various types. And finally, though one desire-type may fulfill another desire-type, greater value accrues to a desire-type which more often fulfills that other type, which means that the propensity to fulfill or thwart other types must factor in the weighting of types, too. These adjustments greatly complicate the process required to assess moral value, but I think it can be seen that each one is necessary to an accurate analysis, given our understanding of value.
Thus we arrive at a mature account of moral good, though it becomes more naturally expressible in terms of an equation than a definition. A desire-type is morally good iff, as a function of .. (i) the numerical preponderance of desire-types it fulfills over those it thwarts, (ii) the comparative strength of desire-types it fulfills to those it thwarts, and (iii), the tendency of the desire-type to fulfill or thwart other desire-types, .. the desire-type has a net positive value. If as a function of those three conditions a desire-type has a net negative value, then that desire-type counts as morally bad, or evil.
1 It is natural to talk of objects as desired and valued, rather than states of affairs. However, any expression of this sort can be easily translated into the philosophical vernacular: “I want ice-cream” can be interpreted as a desire to realize the state of affairs in which the person ‘I’ refers to has icecream.
2 An analogy might help. Justification is a relation between (e.g.) a belief and adequate grounds for that belief. It is the grounds which confer justification, and it is in virtue of the grounds being so related to the belief (sometimes by way of other beliefs) that the grounds confer justification upon the belief. Having justification conferred on it, we now call the belief ‘justified’. Notice that ‘justified’ here masquerades as a non-relational property, when in fact justified beliefs are always justified in relation to their grounds. So too with ‘value’: that something has value depends upon its relation to some grounds, and the grounds are desires.
3 Why tokened types, rather than all types? Because if a value statement is to be true, the desire-types which form one of the two relatum had better exist.
4 I’m uncertain whether this analysis is defective – one naturally thinks that if ‘epistemically good’ picks out desires to do with knowledge, then ‘morally good’ should pick out desires to do with morality. The analysis so far is compatible with many other ethical theories, e.g., that ‘morally good’ picks out other-regarding desires which exhibit sympathy.
5 Some suggest that only the consideration of all desires could escape the charge of being arbitrary, since any restriction would need to be justified (and what could justify it?), yet at best this is a weak prima facie justification for wide scope. Another reason given is that morality is universal. I think this is true, but it is not a finding of the Desirist analysis of moral concepts, and so I do not think relativism should be ruled out.
6 I offer no argument here, as I have none. All I can offer is that the thesis may be justified in hindsight, by the fruit it bears.