1. Introduction
It often happens in philosophy that an idea is introduced, repeated by others, and takes hold with little regard to its original rationale. One such idea is the so-called principle or criterion of descriptive adequacy for theories of well-being. The criterion was first introduced by L. M. Sumner in his influential book Welfare, Happiness, Morality (1996) and has now become a commonplace in the philosophy of well-being. Very roughly, the idea is that a theory of well-being is adequate only if it best fits with our ordinary experiences and judgments about well-being. If a theory does not best fit with our ordinary experiences and judgments about well-being, then it is inadequate.
The criterion of descriptive adequacy is taken by many to be part of the standard approach to philosophical theorizing about well-being. For instance:
What are the criteria of success for well-being theories in philosophy? […] First, these theories should be adequate to our ordinary understanding and use of the concept; as Sumner puts it, they should be  descriptively adequate. This means, in part, that a theory of well-being should capture (or at least not be at odds with) as many as possible of our ordinary judgments about who has well-being and who doesn’t. (Tiberius Reference Tiberius2018: 19)
It is commonly thought that a philosophy of well-being is a theory that purports to best explain the central intuitions people have about well-being in their ordinary lives (for example, see the criterion of descriptive adequacy in Sumner). (Gowans Reference Gowans and Fletcher2016: 70)
The traditional philosophical approach to the study of well-being encompasses two theses. The Descriptive Adequacy condition holds that the most important demand on a theory of well-being is that it should capture our commonsense judgments. (Bishop Reference Bishop2014: 20)
Philosophers have typically assumed that a theory of well-being consists in an analysis of the ordinary concept of well-being. […] Thus, they have tended to consider “descriptive adequacy” the primary criterion for evaluating such theories. (Prinzing Reference Prinzing2021: 153)
Each author takes a somewhat different attitude towards the criterion: Tiberius endorses the criterion; Gowans simply makes a descriptive claim about what is common in the philosophy of well-being; and Bishop and Prinzing go on to criticize the philosophy of well-being for an over-reliance on the criterion (more on which later). However, despite their differences, each author takes it to be uncontroversial that the criterion of descriptive adequacy plays a central role in contemporary debates about well-being.
Against this, this article argues that theories of well-being are not subject to the criterion of descriptive adequacy. Moreover, it argues that existing debates within the philosophy of well-being are most charitably interpreted as not being committed to the criterion. The reason, I will argue, is that because well-being is an essentially normative notion, well-being theories should be assessed by normative rather than descriptive criteria. A positive lesson of the article is that taking the normativity of well-being seriously at a meta-ethical level has implications for our approach to first-order theorizing about well-being.Footnote 1 The plan for the article is as follows. In the next section, I examine the criterion and its rationale in more detail (section 2). After that, I present my argument against the criterion, drawing parallels between theories of well-being and other kinds of moral theories (section 3). Next, I examine two objections to my argument relating to reflective equilibrium and thick normative concepts (section 4). I then discuss some of the implications of rejecting the criterion (section 5), before finally concluding (section 6).
2. Sumner on Descriptive Adequacy
The criterion of descriptive adequacy is rarely argued for. It is usually assumed. To understand its rationale, we therefore need to examine the reasons for which it was originally introduced. To this end, in this section I examine Sumner’s (Reference Sumner1996) seminal discussion of the criterion. As well as being the origin of the criterion, Sumner’s discussion provides one of the only explicit arguments for the criterion. As such, it will be of particular importance to get a clear understanding of the reasons for which the criterion was introduced.
The criterion originates in a general methodological discussion of the philosophy of well-being. As Sumner sees it, theories of well-being are metaphysical theories that tell us what it is for a life to go well—what a good life consists in, or what is constitutive of a good life (1996: 7). The guiding assumption is that we have a shared network of prudential concepts our use of which is rich and stable enough to support “a high degree of preanalytic agreement on what is to count as faring well or badly” (1996: 8). A theory of well-being then tells us what the world must be like for well-being claims to be true. Sumner then proposes the criterion of descriptive adequacy as the main principle by which we determine which theory of well-being is correct.
According to the criterion of descriptive adequacy, the correct theory of well-being is one that “best captures” (1996: 9), is “most faithful to” (1996: 10), or best “fits” (1996: 10) our ordinary experiences, reflections, deliberative practices, and common-sense psychological explanations relating to well-being. Its degree of fit is “a function of the extent to which the truth conditions it offers can support and systematize our intuitive assessments.” (1996: 11) This will reflect that some beliefs are more central and others more peripheral, that some may or may not survive reflection, and that there may be some degree of indeterminacy or inconsistency in our ordinary opinions (1996: 11). Sumner then draws out some more specific criteria “latent in” the basic proposal (1996: 13). Specifically, a theory must be: (a) general, in the sense that it should explain all kinds of well-being judgments (1996: 13-16); (b) formal, in the sense that it explains the conditions under which someone is benefited rather than listing the particular things that do in fact benefit people (1996: 16-17); and (c) neutral, in the sense that it should be unbiased regarding which ways of life are preferable, aside from what is implied by its formal principles (1996: 17-18).
Thus understood, the criterion places a non-trivial, substantive constraint on theories of well-being. There is a different sense in which a theory might ‘best fit’ our judgments just in case it is the correct theory of what those judgments are about. However, to say that a theory must best fit our judgments in this sense would be trivial and fail to place any real constraint on well-being theorizing. It is also important to distinguish the criterion from the weaker idea that a theory’s degree of fit with our ordinary judgments provides evidence for or against that theory. As well as being relatively trivial, this weaker proposal fails to capture the idea that the criterion places a condition of adequacy on theories of well-being. As Sumner states, “the best theory about the nature of welfare is the one which is most faithful to ordinary experience.” (Sumner Reference Sumner1996: 10, emphasis added) If descriptive adequacy is not a necessary condition for a theory being correct and simply a source of evidence for its correctness, then it is possible that the correct theory of well-being is descriptively inadequate. Thus, the weaker interpretation fails to capture what it is at issue.Footnote 2
The argument for the criterion then arises in a discussion of the normative or evaluative significance of well-being. For unlike other objects of metaphysical inquiry, “it may be doubted whether its nature can be settled independently of assumptions about its value.” (1996: 8) Thus, a question arises concerning what kind of considerations are appropriate for adjudicating among rival theories. As Sumner sees it, we can assess theories of well-being for their descriptive adequacy, for their normative adequacy, or some combination of the two. The criterion of descriptive adequacy is then argued for by a rejection of an approach that assesses theories solely for their normative adequacy.
What would it look like to take a purely normative approach to well-being theorizing? According to Sumner, it would involve assigning well-being a particular role in a normative theory or framework and then asking what best fills this role. For instance, “If it is a theory of rationality it might take the form of prescribing that agents maximize their own welfare, while if it is a theory of morality it might require maximization of general welfare, or the equal distribution of welfare, or maximization of the welfare of the worst off.” (1996: 8) To adapt a phrase from David Lewis: to say what well-being is, we must first ask what well-being does, and then find something that does that.
However, Sumner objects to this approach on the following grounds: “Because the notion of welfare already has a vernacular currency it is not available as a term of art, to be defined in whatever way will best suit some favoured theoretical needs.” (1996: 10) In other words, because the meaning of ‘well-being’ and other prudential terms is already fixed by its use in ordinary language, we cannot stipulate that it is defined by its role in some normative theory or other. To define well-being in this way is simply to change the subject. To illustrate, suppose that we define well-being in terms of its role in some egalitarian theory of justice—namely, as that which requires equal distribution in a just society. The problem is that our theory of well-being will now be determined entirely by our commitments concerning justice rather than anything to do with well-being as ordinarily conceived. This approach is thus unmoored from the proper subject matter of well-being. Hence, theories of well-being need to be anchored in ordinary thought and practice to guarantee that we are talking about well-being and not something else.
In the next section, I will argue that this characterization of a purely normative approach is unduly narrow, since one might instead define well-being in terms of its role in ordinary normative thought and practice rather than some theoretical framework. But Sumner’s own response is to accept the criterion of descriptive adequacy as a way of guaranteeing that well-being theorizing is suitably anchored to our ordinary concept of well-being. For if it is a necessary condition on a theory of well-being’s being correct that it best fits our ordinary convictions, then it seems that there is no question of whether such a theory is really about well-being and not something else. However, while Sumner rejects a purely normative approach to theorizing about well-being, he still thinks that normative considerations have a role to play. Specifically, he thinks that normative considerations play a role in theory construction where the inconsistencies and indeterminacies inherent in our ordinary notion of well-being leave gaps in our theory. Thus, it might be the case that several theories of well-being are equally descriptively adequate (this is why the criterion is only a necessary and not a sufficient condition for adequacy). However, this role is still limited, and descriptive adequacy takes precedence: “Normative considerations properly come into play only around the edges of our ordinary concept, in its disputed or unsettled regions. When the constraints imposed by that concept run out, or when the evidence provided by our ordinary experience is indeterminate or inconsistent, then there is a time for shaping a theory of welfare to fit some favoured normative niche.” (1996: 19) So while normative considerations have some role to play for Sumner, the task of constructing a theory of well-being is primarily a descriptive one. Interestingly, this is quite different from how others conceive of the role of normative adequacy in well-being theorizing. Tiberius, for example, writes “well-being theories should be adequate to explain the value of well-being and why we have good reason to pursue it or why we have good reason to follow the recommendations of the theory; in other words, the theory should be normatively adequate.” (2018: 19; see also Prinzing Reference Prinzing2021: 155)
With the criterion of descriptive adequacy articulated and defended, Sumner then goes on to employ the criterion in arguing for and against various theories of well-being. For example, he argues that experience machine style examples show that hedonism is “too interior and solipsistic to provide a descriptively adequate account of the nature of welfare.” (1996: 98) He also argues that revealed preference theory is “hopelessly inadequate as an account of our ordinary conception of welfare” (1996: 118, fn.14), since there is no conceptual connection between market choice and welfare. And he argues that there is a presumptive case against objective theories, since he thinks that it is central to our ordinary concept that well-being is perspectival or subject-relative, and objectivist theories seem to deny this core feature (1996: 43-44). Further, Sumner defends his own theory of well-being as authentic happiness by arguing that it is superior with respect to descriptive adequacy (1996: 176). However, as Sumner notes (1996: 43), it is not necessarily obvious whether some theory is or is not descriptively adequate, and there may be room for reasonable disagreement. Thus, while Tiberius (Reference Tiberius2018: 21) similarly suggests that we should reject hedonism on grounds of descriptive adequacy, she also suggests (2018: 23) that objective views do better when it comes to descriptive adequacy. However, insofar as both accept the criterion, both are committing to ruling out theories that are not descriptively adequate.
While most discussions concern descriptive adequacy as it relates to our ordinary judgments about well-being, it is worth mentioning that Sumner also employs the criterion in relation to the role of well-being in common-sense psychology. For example, he rejects psychological egoism on the grounds that it is “a core element in our ordinary concept” that our own well-being is “just one possible end of action among others” (1996: 13). In the next section, I will suggest that attending to the role of ‘well-being’ in common-sense psychology may actually speak against the criterion.
3. Against Descriptive Adequacy
I will now argue that we should reject the criterion of descriptive adequacy. The argument proceeds as follows. First, I argue that Sumner’s argument for the criterion relies on an overly narrow conception of the purely normative approach and that there is an independently plausible version of the approach that is immune to Sumner’s objection. Second, I argue directly that theories of well-being are not subject to the criterion of descriptive adequacy because they are normative theories, drawing a comparison with the moral case. Third, I offer a diagnosis of why theories of well-being are not subject to the criterion in terms of the meta-semantic differences between descriptive and normative concepts.
3.1. An Alternative ‘Purely Normative’ Approach
First, then, let’s re-examine Sumner’s argument for the criterion of descriptive adequacy. Recall that the argument rested on his rejection of a purely normative approach to theorizing about well-being. We saw that Sumner rejects a purely normative approach because he believes that we cannot define well-being in terms of its role in normative theory. However, this is not the only way in which one might pursue a purely normative approach. Specifically, one might instead define well-being in terms of its role in ordinary normative thought and practice. Moreover, this is not only possible, but it is independently plausible. This is because it is highly plausible that our ordinary concept of well-being is a normative concept (see, e.g., Railton Reference Railton1989, Darwall Reference Darwall2002, Fletcher Reference Fletcher2021, Lin Reference Lin2022; Brown Reference Brown2025).Footnote 3 However, defining well-being in terms of its role in ordinary normative thought and practice is not vulnerable to Sumner’s objection. For to do this just is to define well-being in terms of its “vernacular currency”. Thus, if the ordinary meaning of well-being is given by its role in ordinary normative thought and practice, there is no open question about whether the realizer of the role corresponds to what we ordinarily mean by ‘well-being’.
As an illustrative example, consider fitting attitude accounts of our concept of well-being. This approach defines well-being in terms of its role as the fitting object of certain practical attitudes. The most prominent example is the rational care theory, which maintains that claims about well-being are equivalent to claims about the fitting object of sympathetic concern. As Darwall (Reference Darwall2002) articulates the view, claims about well-being are equivalent to claims about what we should want for someone insofar as we care for them (see also Anderson Reference Anderson1993; Cosker-Rowland Reference Cosker-Rowland2016; Brown Reference Brown2025). Alternatively, we might define our concept of well-being in terms of the role well-being plays in relation to the fitting objects of attitudes like envy or pity (compare Kneer & Haybron Reference Kneer and Haybron2025: 248). Roughly, the idea would be that there is some conceptual connection between someone’s doing well or badly and their meriting envy or pity. If something along these lines is correct, then our ordinary concept of well-being can be defined in terms of its role within ordinary normative thought and practice. And while we might question whether any such account is in fact correct, it cannot be objected that it changes the subject matter by stipulatively defining well-being in terms of some normative theory that has nothing to do with well-being as ordinarily conceived. For it just is an account of our ordinary concept of well-being. And the same will be true of any account that defines well-being in terms of its role in ordinary normative thought and practice, not just fitting attitude accounts.Footnote 4
3.2. Well-being Theory is a Kind of Normative Inquiry
Adopting a more plausible version of the purely normative approach therefore undermines the original motivations for accepting the criterion of descriptive adequacy. However, this is only to criticize the argument for the criterion, not the criterion itself. For all that has been said, it might still be the case that we should accept the criterion. This now leads us to the second part of our argument, which argues that the very same considerations speak not only against the motivations for accepting the criterion, but directly against the criterion itself.
The basic thought is a simple one: if well-being is essentially normative, then the task of constructing a theory of well-being is an essentially normative enterprise. As such, we should expect considerations of normative rather than descriptive adequacy to take precedence when assessing theories of well-being. Thought of in this way, constructing a theory of well-being is like constructing a theory of right action or of justice. While such theories might aim to systematize our ordinary moral knowledge, the proper task of a moral theory is not to discover what best fits our ordinary moral judgments. Rather, to construct a moral theory is to engage in a kind of normative inquiry, where the best theory is the one that best fits the relevant normative facts.
To illustrate this point, consider maximizing act-utilitarianism as a theory of right action. While maximizing act-utilitarianism has its roots in certain aspects of ordinary moral thought, it is a radically revisionary moral theory. It asks us to radically reconsider what kinds of things can be justified and what kinds of lives we should lead. We cannot reject such a view simply because it does not fit our ordinary experiences and judgments as well as or better than the alternatives. That is the whole point of radically revisionary theories. Of course, an objector might respond that we know from ordinary experience that, say, there is a distinction between killing and letting die that the theory cannot explain, and that is a reason to reject it. But here the reason for rejecting maximizing act-utilitarianism is that it fails to explain the moral facts, not that it fails to best fit our ordinary judgments. It might also happen to be descriptively inadequate, but this itself is not what makes the theory false.
The same is true of theories of well-being. For instance, consider Socrates’ proposal in the Gorgias that he is worst-off who does wrong and gets away with it. If descriptive adequacy were a central constraint on well-being theorizing, this proposal seems like a non-starter. After all, as Plato would be the first to recognize, very few people would in fact agree with Socrates. But surely Socrates is not trying to offer us a view that best fits our ordinary experiences and judgments better than the alternatives. Rather, he is making a radically revisionary normative claim and asking us to radically re-evaluate our existing priorities. From a normative perspective, we might think that Socrates’ proposal is implausible, given what we think we already know about what kinds of things are good and bad for us. But again, what makes the proposal incorrect, if indeed it is, is that it contradicts the normative facts, not that it is descriptively inadequate. It might also happen to be descriptively inadequate, but this itself is not what makes the proposal false.
Of course, the criterion need not be understood as claiming that descriptive adequacy is what makes a theory of well-being true. But to rule out a theory on the grounds that it is descriptively inadequate is to derive a normative conclusion from non-normative considerations about what we happen to believe. It is important that this point is not obscured by the fact that we commonly appeal to what we know and believe when offering reasons to accept or reject a normative theory. Because knowledge is factive, appeals to what we know can function as appeals to the facts that our purported knowledge is about. And because belief is transparent, appeals to what we believe can also function as appeals to the purported facts that our beliefs are about. But to evaluate a theory of well-being is ultimately an exercise in normative judgment. And as such, it is constrained by whatever criteria are appropriate for assessing normative claims more generally.
Is there any reason to think that normative claims more generally are constrained by descriptive adequacy? It is important to emphasize that rejecting descriptive adequacy does not mean denying that our ordinary judgments about well-being can be used as evidence for or against certain theories or claims. Thus, a purely normative approach is compatible with the idea that considerations of descriptive adequacy might offer some reason for accepting or rejecting a theory. As we observed in the previous section, however, the criterion places a far stronger constraint on well-being theorizing: it offers a condition of a theory’s adequacy. However, from a normative perspective, there is no reason to accept this as a methodological principle. It could only be justified on the assumption that our ordinary judgments about well-being are mostly correct. But it is precisely this assumption that radically revisionary theories reject. Moreover, while the possibility of radically revisionary theories dramatizes the point, all that is needed is the possibility that a theory with sub-optimal fit might be the correct theory of well-being.
Thus, insofar as theories of well-being are normative theories, they must ultimately be assessed by their normative adequacy. And from a normative perspective, there is no reason to suppose that descriptive adequacy necessarily tracks the normative facts. This leads us on to the third and final part of our argument against the criterion, which concerns the deeper question of why normative theories are not subject to the criterion. The explanation I will offer is that normative theories are not subject to the criterion of descriptive adequacy because of the distinctive meta-semantic properties of normative terms and concepts, properties not shared with ordinary descriptive terms and concepts. While I hope others will find this explanation attractive, it should be highlighted that it is independent of the preceding arguments of this section. Thus, one can accept that normative theories are not subject to the criterion of descriptive adequacy without accepting the following diagnosis of why this is so. But if correct, the following diagnosis offers independent support for the above arguments.
3.3. Descriptive Adequacy and Normative Meta-Semantics
If we think about descriptive adequacy as a condition for theorizing more generally, there are many cases where the condition seems more plausible. These are cases in which our thought and talk about the object of inquiry is grounded in a certain meta-semantic picture. Specifically, for many ordinary descriptive terms and concepts, it is plausible to think that what a term or concept refers to is determined by an individual or community’s dispositions to apply that term or concept. For instance, and very roughly, it is facts about our dispositions to apply ‘table’ to certain things that determines the reference of ‘table’. If someone were disposed to apply ‘table’ only to Faberge Eggs, we would probably do well to interpret them to be not simply mistaken about tables but to mean something different by the term (if they aren’t simply exceptionally confused or incompetent). If something like this were the correct view of our thought and talk of our object of inquiry, it would make sense why theorizing about that thing would be subject to the criterion of descriptive adequacy. For in order to discover the nature of that thing, we would need to know what best fits our ordinary judgments (i.e., applications of the relevant term or concept), as it is precisely this that determines what those judgments are judgments of.Footnote 5
By contrast, it is widely held that normative terms and concepts are not like this. This is because normative terms and concepts allow for radical disagreement without change of subject or conceptual incompetence (see, e.g., Hare Reference Hare1952; Horgan & Timmons Reference Horgan and Timmons1992; Björnsson Reference Björnsson, McPherson and Plunkett2017; Williams Reference Williams2018; though see Dowell Reference Dowell2016 for dissent). For example, imagine that Carl the capitalist applies the term ‘good’ only to things that promote free markets, and that Mel the Marxist applies the term ‘good’ only to things that will bring about a communist state. If ‘good’ were an ordinary descriptive term, then we would interpret Carl and Mel to be talking past one another. However, assuming that ‘good’ plays the same commendatory role for both, we naturally interpret Carl and Mel as disagreeing about something—namely, about which things are good. This suggests that the extension of the term ‘good’ is not determined by facts about what we are disposed to apply the term to. However, if the extension of ‘good’ is not fixed in this way, it is hard to see what could justify the assumption that a theory of the good is adequate only if it best captures our ordinary judgments about what is good. For on this picture, the reference of ‘good’ might diverge radically from what we are ordinarily disposed to apply ‘good’ to.
The same seems to be true of prudential terms and concepts. For instance, when Polus says that he is worst-off who is unjustly killed and Socrates says that he is worst-off who unjustly kills, we naturally take them to be disagreeing with one another. Indeed, Darwall (Reference Darwall2002: 10-12) and Fletcher (Reference Fletcher2021: 40-42) argue that the fact that well-being discourse allows for this kind of disagreement is itself a reason to think that our concept of well-being is essentially normative. However, if ‘worst-off’ were an ordinary descriptive term like ‘table’, we would more naturally take Polus and Socrates to be talking past one another. Thus, the possibility of radical disagreement and radical moral error strongly suggests that the meta-semantic story for well-being vocabulary must be different to that of ordinary descriptive terms. If, however, the criterion of descriptive adequacy presupposes that meta-semantic picture, the justification for applying the criterion to theories of well-being is undermined. For again, the reference of ‘worst-off’ might diverge radically from what we are ordinarily disposed to apply ‘worst-off’ to.Footnote 6
4. Reflective Equilibrium and Thick Concepts
In this section, I consider two objections to my argument so far. The first objection is that my argument implausibly rules out the method of reflective equilibrium as a legitimate approach to moral and well-being theorizing. The second objection is that my argument can be resisted if well-being is a thick normative concept. I argue that neither objection is sound.
4.1 Reflective Equilibrium
It might seem to some that Sumner’s descriptions of the criterion of descriptive adequacy are most charitably interpreted as descriptions of the method of reflective equilibrium. However, given that the method of reflective equilibrium is widely considered to be the most plausible approach to normative theorizing, the arguments against the criterion of descriptive adequacy would show too much. For they would show that the most plausible approach to normative theorizing in general is mistaken. As Scanlon writes, “it seems to me that this method, properly understood, is in fact the best way of making up one’s mind about moral matters and about many other subjects. Indeed, it is the only defensible method: apparent alternatives to it are illusory.” (2003: 149)
Now, there are different ways of understanding reflective equilibrium. Most notably, proponents typically distinguish between narrow and wide versions of reflective equilibrium (Rawls Reference Rawls1974). We therefore need to know which version of reflective equilibrium is closest to Sumner’s own views. What we see is that his own approach more plausibly aligns with narrow reflective equilibrium. Narrow reflective equilibrium starts with our considered judgments about a subject matter, for example justice, which we would make in favourable circumstances. We then aim to construct a scheme of principles that deliver the same verdicts of those judgments. Because even our considered judgments made in favorable circumstances will display various irregularities and distortions, we will have to make adjustments to some of our considered judgments to allow for “the smoothing out of certain irregularities” (Rawls Reference Rawls1999: 43). By contrast, wide reflective equilibrium involves considering a much wider range of considerations in constructing one’s scheme of principles and adjusting one’s considered judgments, including an examination of “all possible descriptions to which one might plausibly conform one’s judgments together with all relevant philosophical arguments for them.” (1999: 43) Crucially, Rawls points out that unlike narrow reflective equilibrium, wide reflective equilibrium allows that a person’s considered judgments might undergo a “radical shift” (1999: 43).
While the criterion of descriptive adequacy may plausibly be interpreted as narrow reflective equilibrium, it is not plausibly interpreted as wide reflective equilibrium. Like narrow reflective equilibrium, the criterion can allow that some ordinary judgments will not stand up to scrutiny, for example judgments grounded in bias, self-interest, or made under unfavorable conditions. But, like narrow reflective equilibrium, the criterion does not allow that our considered judgments might undergo a radical shift. Both narrow reflective equilibrium and the criterion provide us with a descriptive method which generate principles to represent, say, our sense of justice or well-being. While Sumner allows for normative considerations to play a role in theory choice, this only happens when the criterion of descriptive adequacy leads to theory underdetermination due to the irregularities inherent in our ordinary judgments.
Importantly, proponents of reflective equilibrium typically reject narrow reflective equilibrium in favor of wide reflective equilibrium. For example, Rawls (Reference Rawls1999: 43) asserts that “Clearly it is [wide reflective equilibrium] that one is concerned with in moral philosophy,” and Scanlon (Reference Scanlon and Freeman2002: 147) describes wide reflective equilibrium as “primary”. The reason is that moral philosophy is ultimately concerned with principles that are normatively justified, and descriptive considerations about what best fits our judgments just aren’t the kind of thing that provide such justification. As Daniels (Reference Daniels1980: 21) comments, narrow reflective equilibrium confuses moral anthropology with moral philosophy. By contrast, wide reflective equilibrium selects principles not solely on the basis of our considered judgments, but also on the basis “all relevant philosophical arguments for them,” as well as any relevant moral or non-moral background theories, and so on. While wide reflection equilibrium might involve articulating principles that best fit our considered judgments, the process of going back and forth between principles and judgments and adjusting them is through and through a normative exercise, one that is informed by a much wider range of considerations than simply what principles best fit our considered judgments. It is this feature that allows for radical shifts in our considered judgments.
Thus, while our initial considered judgments still play an important role in wide reflective equilibrium, wide reflective equilibrium is not constrained by the criterion of descriptive adequacy. Insofar as the criterion of descriptive adequacy can be interpreted in terms of the method of reflective equilibrium, it is only plausible to interpret it in terms of narrow reflective equilibrium. But narrow reflective equilibrium is and should be rejected for the same reasons we should reject descriptive adequacy: it merely offers a descriptive representation of our moral sensibility rather than any normative justification. Thus, rejecting descriptive adequacy is compatible with accepting reflective equilibrium for normative theorizing.
4.2. Well-being as a Thick Normative Concept
My argument against the criterion of descriptive adequacy relied on the idea that well-being is a normative notion. As a normative notion, I argued that theories of well-being are constrained by normative rather than descriptive considerations. However, one might accept the claim that well-being is a thick normative concept that encodes both normative and descriptive content. This is proposed, for example, by Tiberius (Reference Tiberius and Kirchin2013). While Tiberius offers no account of the descriptive content of ‘well-being’, in the context of the present discussion, one might think that if our concept of well-being is indeed a thick concept, then descriptive adequacy will have a prominent role to play after all. Specifically, the thought might be that the descriptive content of ‘well-being’ constrains theories of well-being in such a way that a theory of well-being is adequate only if it best fits our ordinary judgments.
We see, however, that we have no reason to accept the criterion of descriptive adequacy even if we assume that well-being is a thick concept. To see why, consider an uncontroversially thick concept like selfishness. Suppose that to say that someone is selfish is to say that they inappropriately prioritize their own interests over those of others. If this is the right way of thinking about selfishness, then it seems that a theory of selfishness must be descriptively constrained, because it must apply only to instances of agents prioritizing what they take to be in their own interests, which seems to be a descriptive property. However, this way of being descriptively constrained is not the same as being subject to the criterion of descriptive adequacy. This is because the question of what counts as inappropriately prioritizing one’s own interests cannot be settled simply by working out what best fits our ordinary experiences and judgments about selfishness. It is, rather, a robustly normative question.
Of course, there are many different ways of analyzing thick concepts, so one might reject the analysis of selfishness given above (see Roberts Reference Roberts2013, Reference Roberts, McPherson and Plunkett2017). However, regardless of our particular understanding of thick concepts, it seems independently plausible that they allow for radical disagreement and radically revisionary applications in much the same way as thin normative concepts. For instance, an effective altruist’s and a rational egoist’s application of the term ‘selfish’ might radically diverge from ordinary usage of the term. Yet we can intelligibly disagree with them about what is genuinely selfish, as they can with each other. Thus, even if well-being is a thick normative concept, it does not follow that we should accept the criterion of descriptive adequacy.
5. Implications of Rejecting the Criterion
I have argued that because well-being claims are robustly normative, theories of well-being are not constrained by the criterion of descriptive adequacy. In this regard, well-being is quite different to the nearby notion of happiness. Although ‘happy’ admits of multiple, distinct senses, many think that ordinary claims about happiness are typically interpreted as psychological claims (see Haybron Reference Haybron2008; Kneer & Haybron Reference Kneer and Haybron2024, Reference Kneer and Haybron2025). To say that someone is happy is to ascribe a certain kind of attitude, mood, or emotion to them. Insofar as psychological claims are descriptive claims, theories of happiness are plausibly subject to the criterion of descriptive adequacy (see Haybron Reference Haybron2008 for explicit employment of the criterion in this context).Footnote 7 Given the normative importance of happiness, we might think it appropriate to appeal to normative considerations where descriptive considerations give out. However, while happiness claims are normatively relevant, they differ from well-being claims in that they are not normatively contentful (compare Schroeder Reference Schroeder2007: 80). In this section, I will examine three implications of rejecting descriptive adequacy relating to (i) evaluating theories of well-being, (ii) the possibility of other kinds of descriptive constraints, and (iii) integrating the philosophy and science of well-being.
First, the most general implication of rejecting the criterion of descriptive adequacy is straightforward: the criterion should not be used to determine whether to accept or reject a theory of well-being. Because theorizing about well-being is an essentially normative activity, theories or theoretical claims about well-being should be primarily assessed on normative grounds from an engaged normative perspective. Thus, if we want to reject a theory like hedonism or Socrates’ proposal in the Gorgias, it is not enough to simply point out that such views are descriptively inadequate. This could only be justified on the assumption that our ordinary judgments are mostly correct. However, it is question begging to make this assumption in response to such views. To reiterate, this does not mean that considerations of descriptive adequacy cannot be used as evidence for or against a theory. But the mere fact that a theory best fits with our ordinary experiences and judgments is not a criterion of adequacy; it is at most defeasible evidence.
How, then, do we assess theories of well-being from a purely normative perspective? The argument against the criterion of descriptive adequacy does not presuppose any positive answer to this question. It is only committed to the negative claim that, whichever methods are appropriate for assessing normative theories, they do not include the criterion of descriptive adequacy. It seems to me, however, that there are many ways in which we assess normative theories. We see what tallies with our ordinary convictions and we make considered and reflective judgments. We employ philosophical argumentation and appeal to general theoretical considerations. We engage our normative sentiments and tell stories and offer compelling pictures. We engage in hermeneutical interpretation of our values, as well as ideological critique and other kinds of debunking explanations. We appeal in various ways to empirical considerations in disciplines from psychology to anthropology. And so on and so forth. While I am inclined to accept a broad, pluralistic approach, the core point is simply that whichever of these methods are appropriate for assessing normative theories, none of these methods involve accepting or rejecting theories on the grounds that they do or do not best fit our ordinary judgments and experiences.
Second, nothing I have said here rules out the possibility of other kinds of descriptive constraints playing a role in theory construction. For example, we saw above that if well-being is a thick normative concept, theories of well-being might be descriptively constrained by whatever descriptive content is implicit in our concept of well-being. Further, some might be attracted to the idea that there are ‘prudential fixed points’ (compare Cuneo & Shafer-Landau Reference Cuneo and Shafer-Landau2014 on ‘moral fixed points’) or that the intelligibility of well-being claims presupposes some minimal degree of fit with our ordinary judgments (compare Foot Reference Foot1959 on moral beliefs). While I am skeptical on both counts, the argument against the criterion of descriptive adequacy is compatible with accepting that there may be other kinds of descriptive constraints on theories of well-being besides the criterion of descriptive adequacy. Thus, even if there are prudential fixed points and even if theories of well-being require a minimal degree of fit with our ordinary judgments to be intelligible, it remains the case that the best theory of well-being need not be the one that best fits our ordinary judgments.
One might wonder whether a weaker version of the criterion of descriptive adequacy might be defended if the principle takes a satisficing rather than maximizing form. In other words, perhaps a theory of well-being is correct only if it captures enough of our ordinary judgments about well-being. An immediate problem with this proposal is that we do not seem to have any means for specifying the relevant sufficiency threshold in a non-arbitrary way. But this problem aside, the proposal faces the same problem as the maximizing version of the principle. For assuming that the relevant sufficiency threshold is stringent enough, it will ineligibly rule out the possibility of radically revisionary theories. And if it is relaxed enough to rule in such theories, then the principle is too weak to play the theory-guiding role ascribed to the criterion of descriptive adequacy.
Third, another implication relates to proposals to integrate the philosophy and science of well-being. For example, Bishop (Reference Bishop2014) and Prinzing (Reference Prinzing2021) argue that traditional debates in the philosophy of well-being have resulted in stalemate precisely because of the tendency in such debates to rely on the criterion of descriptive adequacy as a means for choosing between theories. They argue that while each of the major theories of well-being does well to explain a certain amount of the intuitive data, none manages to explain it all, and none provides a better fit overall. Bishop then argues for his preferred theory—the network theory of well-being—on the grounds that it ties explaining the intuitive data but better accounts for the scientific data. Similarly, Prinzing proposes that we break the stalemate by conceptually engineering new concepts of well-being by taking input from both normative and empirical theorizing. Thus, both respond to the stalemate problem by introducing further conditions of adequacy relating to empirical investigation into well-being.
We can grant that if descriptive adequacy were the main criterion for evaluating theories of well-being, then this would lead to philosophical stalemate. However, descriptive adequacy is not the main criterion for evaluating theories of well-being. True, there remains deep and pervasive disagreement about the nature of well-being. But this is simply our epistemic predicament in the normative realm. This is not to deny that there are a number of ways in which the philosophy of well-being could benefit from a closer engagement with empirical research on well-being. For example, empirical research might confirm or deny a theory’s empirical presuppositions, offer more robust evidence about lay intuitions, and provide new theoretical perspectives (Prinzing Reference Prinzing2021; Brown & Potter Reference Brown and Potter2024). However, certain other ways of appealing to empirical research in well-being theorizing do not seem to be appropriate if we adopt a purely normative approach to well-being theorizing.
For instance, Bishop argues for his preferred theory of well-being on the grounds that (a) it is no less descriptively adequate than other theories and (b) it explains the scientific data better than other theories. The sense in which his theory explains the scientific data is that it fits what well-being science actually studies and unifies what otherwise seems like a disparate subject matter (2015: 110-111). However, it is unclear from a normative perspective why facts about what explains scientific practice in this sense would provide evidence for thinking that some theory of well-being is correct. Indeed, this would not even be justified on the assumption that empirical researchers have mostly correct views about well-being, since the criteria of fit and unity do not concern the beliefs of well-being scientists, but how their object of study is best characterized.
Further, if our concept of well-being allows for radical disagreement in the way that is characteristic of normative concepts more generally, then we cannot conclude that entrenched disagreement about well-being is a result of our ordinary concept being inconsistent or defective. Perhaps we need some way of making progress, and perhaps some ways of engaging with empirical research will facilitate this. But if there is nothing inconsistent or defective with our concept of well-being, then it is unclear why engineering new well-being concepts will allow us to make progress on the question of which theory of well-being is correct. Prinzing’s proposal is that we create new well-being concepts whose adequacy should be assessed in terms of its (i) normativity, (ii) measurability, (iii) precision, and (iv) mutability (2021: 158). If our aim is to study well-being from an empirical perspective, then this seems like a fruitful suggestion. However, if our aim is to answer the question of what well-being is, then it is unclear from a normative perspective why we should assess theories of well-being according to (ii)-(iv). For there is no reason to suppose that (ii)-(iv) will track the normative truth.
6. Conclusion
In conclusion, a theory of well-being is adequate if and only if it adequately captures the relevant normative facts. Such a theory might turn out to be descriptively adequate. But it seems unlikely that the best way to arrive at the normative truth is simply to look to our beliefs about what makes our lives go well. Rather, we should look to the facts themselves. And there is no reason from a normative perspective to accept the criterion of descriptive adequacy.
Perhaps there is a way of interpreting descriptive adequacy such that it requires us to be responsive to our ordinary judgments and practices in some weaker sense. But it is hard to see what this would be if it is to be distinguishable from normative adequacy. Moreover, perhaps some proponents of the criterion will insist that they only take descriptive adequacy to provide defeasible evidence of a theory’s correctness. But this would make the criterion trivial and uninteresting. For all it would say is that we have reason to think that a theory of well-being is true insofar as it accords with what we generally take to be true about well-being.
All in all, we have no reason to accept the criterion of descriptive adequacy for theories of well-being. There is a question about the extent to which the arguments of this article concern what philosophers think they are doing when theorizing about well-being as opposed to what they are in fact doing. But our conceptions of our practices can sometimes affect the reality of those practices. Although couched in the negative, we should see the conclusion of this article as a positive for well-being theorizing. For it can free us from the conservative strictures of much contemporary theorizing and allow us to think more boldly and radically about what makes our lives go well and badly.Footnote 8
Open access
