Function-Based Conceptual Engineering and the Authority Problem

Author: Matthieu Queloz
Published in: *Mind* 131 (524): 1247–1278. 2022. [doi:10.1093/mind/fzac028](http://doi.org/10.1093/mind/fzac028)
DOI: 10.1093/mind/fzac028
Canonical entry: https://www.matthieuqueloz.com/entries/function-based-conceptual-engineering-and-the-authority-problem/
Published PDF: https://philpapers.org/archive/QUEFCE.pdf

Machine-readable text companion generated from the PDF. Page markers follow the journal pagination.

[p. 1248]

The use/mention distinction alone, if it is to be clear-cut, cannot capture the significant difference at issue here because, as Moore (1986/2019, p. 15) argues, that difference turns on what a concept is employed for, and on any account of the use/mention distinction that renders it clear-cut, the mere fact that a concept is mentioned cannot tell us much about what the concept is employed for.

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 This is not merely to question the authority of the concept on a particular occasion (‘Does the fact that a is F really matter here?’). It is to ask why we should ever think in those terms (‘F-ness’) at all. What the authority problem invites us to identify are not, in the first instance, reasons for action or reasons for belief, but reasons for concept use; not reasons operative within a practice of reason-giving, but reasons for a practice of reason-giving, that is, reasons to adopt a concept and become disposed to recognize certain considerations as reasons. Note, however, that the mere ability to mention a concept (‘They propose this concept of “F”, whatever that is’) is not by itself sufficient to address the authority problem. Some grasp is required of how things present themselves to one who actually uses the concept. One has to understand the concept ‘from the inside’ in order to assess its claim to authority. At the same time, one has to find some way of stepping back from the concept sufficiently to acquire critical leverage over it—it would be too uncritically accepting of the concept simply to insist that we should use it because there is so much F-ness around and F-ness is important for the reasons the concept F itself adverts to. To make sense of what is involved in addressing the authority problem, we have to distinguish between two ways of using (as opposed to merely mentioning) a concept. Adrian Moore (2006, p. 137) has usefully marked this distinction in terms of the contrast between disengaged and engaged concept use:2 in Moore’s example, a disengaged user of the concept Sabbath grasps what the concept applies to and what its applicability entails, thereby understanding the concept’s role in other people’s lives, but she does so in the spirit of an ethnographer, without living by the concept herself (she does not observe the Sabbath); an engaged user of the concept, by contrast, actually lives by the concept and is responsive to its concomitant reasons in the conduct of her own affairs. This distinction between engaged and disengaged use allows us to make sense of what is involved in addressing the authority problem. When using the concept F in an engaged way, one looks through it at the world, as if through a lens, and is immediately responsive to the gestalt it gives the world, the aspects it renders salient, and the reasons it adverts to. Yet one can also use the concept in a disengaged way to hold it up to reflective scrutiny, taking a sideways look at the lens instead of peering

[p. 1249]

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 through it. This is what we must do to critically assess a concept’s authority. But to understand what we are looking at, we still have to grasp the concept, and grasp what the world looks like to one who peers through it. We thus have to adopt a stance that lies halfway between full abandonment to and full abandonment of the concept. We use the concept in a disengaged way in order to evaluate whether to use it in an engaged way. A concept will then be authoritative for us in so far as we have reason to use it in an engaged way, that is, in so far as we have reason to be responsive, in the conduct of our own affairs, to the reasons the concept adverts to. As this characterization of conceptual authority in terms of reasons for reasons suggests, I understand conceptual authority as a gradable and comparative notion: I might have more reasons, or better reasons, to use one concept rather than another.

The relevant notion of conceptual authority must be distinguished from a related notion in the vicinity that has received rather more attention, namely, that ‘special sort of authority one becomes subject to in applying concepts’ (Brandom 1994, p. 9, emphasis in original). The ‘normative bindingness’ whereby concept-users become subject to assessment of whether they are applying a concept correctly or not has attracted much scrutiny. But the aim has been to elucidate, in terms meant to apply to all concepts rather than to help us discriminate between them, whether there even is such a normative bindingness, and what its nature and source might be. The authority problem, by contrast, concerns the prior question of what reasons concept-users have to recognize a concept’s normative bindingness. The question, in other words, is not how a given concept binds us, but why we should let it. We can mark this contrast by distinguishing between authority of and authority in use. Questioning a concept’s authority of use problematizes the legitimacy of a concept’s power over us. Questioning a concept’s authority in use problematizes the normativity involved in a concept’s exercise of that power. A concept can then be said to have authority of use in so far as the concept-user has reason to recognize its authority in use. My focus here lies on authority of use. I neither presuppose nor advocate a particular view of authority in use.

A concept’s authority of use can be called into question long after we have adopted it. But the authority problem is particularly salient when one is faced with a proposed concept one does not yet use. Why should I recognize that concept’s claim to authority over my life? Why should I structure my affairs in those terms at all?

One answer suggested by some of the recent literature on conceptual engineering is that the engineered concept is authoritative in so far as,

[p. 1250]

See, for example, Brun (2016, 2020), Cappelen (2018, ch. 2), Simion and Kelp (2020), Eklund (2002, 2019, 2021), Scharp (2013, 2020, 2021), and Dutilh Novaes (2020a, 2020b, 2017).

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 and because, it is more theoretically virtuous than its predecessor.3 Many of our concepts are vague, inconsistent or open-textured, and their relations to other concepts are often indeterminate, unsystematic or incoherent. Fixing those theoretical defects by rendering our conceptual scheme more exact, consistent and systematic might be thought to give engineered concepts a good claim to be recognized as more authoritative than their predecessors. But where, and to what extent, is it appropriate to engineer concepts for theoretical virtues? And why is it that greater precision, consistency, completeness, determinacy, systematicity or coherence should give engineered concepts more authority? Where the realization of theoretical virtues is the dominant concern anyway— in logic and formal semantics, say—the authority problem arguably does not arise, because the aim of engineering new concepts aligns with the aim with which the replaced concepts were being deployed. But the authority problem reappears once the aspiration to engineer for theoretical virtues is generalized beyond the confines of logic and formal semantics.

The trajectory of Kevin Scharp’s work on conceptual engineering offers an illustration of how this aspiration, at home in logic and formal semantics, can come to be generalized to philosophical concepts across the board. In Replacing Truth (2013), Scharp proposes to replace the concept of truth, which he thinks is inconsistent and generates various paradoxes, with two new concepts of truth; but he is clear that the substitution is to be confined to contexts in which consistency and the avoidance of paradoxes are the dominant concerns. In more recent methodological writings, however, this qualification is dropped, conveying the impression that theoretical virtue can be a general answer to the authority problem. Not just logic, but philosophy more broadly ‘is the study of what have turned out to be inconsistent concepts’ (Scharp 2020, p. 398), because ‘truth, knowledge, value, virtue, freedom, justice, etc.’ have turned out to be ‘organized and distinguished by principles that are themselves inconsistent with one another’ (2020, p. 414). These inconsistencies generate various problems that philosophers get tangled up in, and that conceptual engineering promises to resolve. But even if the project of engineering concepts is initially motivated, not by the pursuit of consistency for its own sake, but by the desire to overcome the problems and paradoxes generated by our concepts, the engineering itself is guided by and aims for theoretical virtue. Across the entire range

[p. 1251]

Metrological naturalism has as a methodological principle that philosophers should use measurement theory as a guide or model in philosophical theorizing. … [W]e know pretty well how to do this for things like length and weight. Trying to figure out how to construct a measurement system for something like truth or justice is a lot more complicated, but this isn’t just an analogy. (Scharp 2020, p. 402) Scharp maintains that engineering should always be conducted in a metrological spirit, because ‘engineering without metrological naturalism is blind’ (2020, p. 399): the virtues of a good measurement system are what provide engineers with a guiding sense of what concepts to aim for.

But how plausible is it, outside logic or formal semantics, that an engineered concept will be authoritative because it exhibits the virtues of a good measurement system? Consider the concept of person. In everyday usage, the concept is vague, and its connection to other concepts not very systematic. It indicates a variety of characteristics— self-consciousness, agency, title to respect—that come in degrees, and as debates over abortion show, its ethical implications and relations to other categories such as sentient being or human being are murky and contested. Some, like Michael Tooley, have therefore undertaken to engineer it into a precise sortal notion that sharply delineates a basis for a more systematic way of thinking about issues around personhood.5 But the implications of this precise sortal notion notoriously go drastically beyond anything within the reach of the non-engineered concept. Of course, making a notable difference to the resulting judgements is arguably the point of engineering the concept. But faced with a stark divergence between pre-engineering and post-engineering judgements, the question of authority becomes acute: why should we act on the Examples include Davidson’s (1990) measurement system for belief, desire and meaning. See also Matthews (2010) and Weaver and Scharp (2019).

See, for example, Tooley (1972, 1983) and many of the positions discussed in Merrill (1998), and see Williams (1985/2011, p. 127) for a critique of Tooley on similar grounds.

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 of our concepts, Scharp suggests, philosophy’s guiding ideal should be ‘a consistent conceptual scheme. No paradoxes. No puzzles. Just clarity’ (2020, p. 415).

On Scharp’s view, philosophical theories are to be cast as measurement systems, so that our messy everyday judgements about X can be transposed into a more systematic, precisely defined and consistent language. He calls this view ‘metrological naturalism’ (after metron, Greek for ‘measure’):4

[p. 1252]

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 judgements formed using the engineered concept? Because it is more precise and permits more systematic judgements? But suppose the engineered concept licenses infanticide in situations in which our non-engineered concept suggests that infanticide is abhorrent. Why should we care about the added precision and systematicity when its price is to do something which, from the perspective of ingrained ethical experience, appears deeply revolting? If that is the price of theoretical tidiness, one may well think, then too bad for tidiness.

The display of theoretical virtues such as consistency and precision is not the only thing we want from our concepts, and this makes it an open question whether vague, indeterminate, open-textured or contested concepts might not sometimes serve us best. In Wittgenstein’s analogy: if your concern is to cut bread and you ask for a bread knife, you will hardly thank me if I give you a razor blade because it is sharper (2000, MS 120, 142v).

A particularly instructive example of how engineering for theoretical virtues raises the authority problem is the debate between Ronald Dworkin (2001) and Bernard Williams (2001a) over how best to define the political value of liberty, especially given its well-known tendency to conflict with the value of equality: beyond a certain point, liberty tends to be achievable only at the expense of equality, and vice versa. Dworkin believes that we should define our concepts so as to immunize them against conflict: ‘integrity among our concepts is itself a value’, he writes, ‘so that we have that standing reason for seeking out, for preferring, conceptions of our values that do not conflict’ (Dworkin et al. 2001, p. 127). Accordingly, he proposes to understand the concept of liberty in terms of an equally distributed right, thereby ensuring that, as political values, liberty and equality can no longer conflict (Dworkin 2011, pp. 4, 364–78). Dworkin would not describe this as ‘engineering’, but the difference does not matter for the present point, which is that this way of ‘constructing’ those values, as he puts it, is preferable to others because it yields a coherent and tidy account on which the two values can no longer conflict. But as Williams insists against Dworkin, it is simply no good securing coherence between two concepts if it comes at the cost of severing the ties to the central human concerns that led us to care about these concepts in the first place. I take it that a concept ties in with a concern just in case the effects of employing that concept tend to contribute to the satisfaction of that concern. And what we really want to know is something that Dworkin ignores in his pursuit of conceptual coherence, namely, whether a given elaboration of a concept serves, or fails to serve, the concerns we now have.

[p. 1253]

For further discussion, see Williams (1995, 2001b). I reconstruct Williams’s genealogy of the political concept of liberty in Queloz (2021, pp. 238–42).

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 To assess whether Dworkin’s proposed concept of liberty serves our concerns, we can begin by asking what human beings fundamentally care about in this connection that drives them to think in terms of anything like the concept of liberty to begin with. We can sketch ‘a schema or matrix’, as Williams puts it, a ‘very bare outline of what our central concern is’ (2001a, p. 92). I shall refer to this as the basic matrix of concern that a concept is usefully seen as having developed out of and been moulded by. In the case of liberty, Williams suggests, that basic matrix of concern centrally involves the concern not to be in someone else’s power (2001a, p. 92).6 Such basic matrices of concern are, of course, still too indeterminate to assess whether a particular concept ties in with them. They need, and will have received, further socio-historical elaboration which concretizes the concerns further and gives them a more definite direction. But whatever exactly the concerns that basically go with the values of liberty and equality now go to, Williams’s point is that we cannot redirect those concerns ‘simply nominalistically, by redefining a word’, because ‘an interest in producing a more coherent body of law is not by itself going to stop the concern going to what the concern goes to’ (2001a, p. 94). And if Dworkin’s proposed concepts fail to tie in with the concerns that give us reason to think in terms of liberty and equality in the first place, we have reason not to adopt those concepts, because they would deflect attention away from the satisfaction of our most basic concerns in these connections.

Three insights can be drawn from this example. One is that the merits of a proposed concept ultimately have to be judged on the basis of a prior understanding of the life that this concept is to help us to lead, which is to say, on the basis of the various concerns—the needs, interests, desires, projects, aims and aspirations—that give us reasons to deploy certain concepts rather than others.

The second insight is that those concerns cannot be redirected simply by redefinition and stipulation. To have a claim to authority, an engineered concept must tie in with our concerns as they are before the engineer’s intervention. When these antecedent concerns are themselves primarily directed at the achievement of theoretical virtues such as precision, consistency, coherence and systematicity, the concepts engineered to realize such virtues can be authoritative—that is why, on a concern-based view, Scharp’s twin concepts of truth could well be

[p. 1254]

The argument for this conclusion turns on a complex political account of how the concept of liberty serves our concerns by retaining its inconsistency with the concept of equality (Williams 2001a, 2001b, 2005a). See §4 below.

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 authoritative within the theoretical contexts to which they are confined. But in contexts in which we are not obviously, or not primarily, concerned to realize such theoretical virtues, it is less clear that concepts engineered to be free of generic theoretical defects such as vagueness or inconsistency will carry more authority than the concepts they are meant to replace, and Scharp’s proposal to generalize his metrological approach to all concepts in philosophy then appears questionable. Even if we recast Scharp’s proposal in concern-based terms, as generalizing the concern for consistency to other concepts like liberty or justice, it is doubtful that the concern for consistency is never overridden by other concerns, and conceptual engineering by itself is not going to change that.

Thirdly, the most tidily consistent concepts are no good to us if they do not serve the central concerns that animated our use of anything like these concepts to begin with. Life is not a logic test. A lack of consistency does not necessarily imply a lack of conceptual authority, just as consistency does not by itself guarantee conceptual authority. In fact, if Williams is right, engineering the concepts of liberty and equality to be consistent actually constitutes a form of deteriorative engineering, because it blinds concept-users to a real conflict between their concerns: there comes a point at which the concern for liberty conflicts with the concern for equality, and the satisfaction of one concern must come at the expense of the other.7 In so far as we would be ill-served by a concept that blinded us to this real conflict, we therefore have reason not to adopt the concepts immunized against conflict that Dworkin advocates. They would put us out of touch with our concerns and the conflicts between them. Our concepts of liberty and equality should conflict, because our concerns do. This suggests that consistency is not necessarily something we should strive for across all our concepts. Inconsistent concepts can be the better concepts in virtue of serving our conflicting concerns better.

Thus, engineering that isolates concepts from the practical contexts in which they are put to work and concentrates on the inherent defects of concepts—or what appear as defects when measured against some ideal of theoretical virtue—embodies a strategy that may solve the authority problem in special cases, but cannot hope to do so more widely. By focusing on the theoretical flaws of concepts, we risk neglecting defects

[p. 1255]

arising from the way in which concepts fail to tie in with other concerns of concept-users. To give direction to engineering efforts while also ensuring that their results are authoritative, a more comprehensive approach is needed.

This need is increasingly registering in the engineering literature. Thomasson (2020a, 2020b), Nado (2021), Haslanger (2020), and Simion and Kelp (2020), as well as Riggs (2021) and Jorem (2022), all advocate a functional turn in conceptual engineering. Its catchphrase is that ‘to engineer a concept well, we must attend to its function’ (Thomasson 2020b, pp. 440–1).

What remains somewhat elusive in these discussions, however, is why we should attend to functions: what exactly is the rationale for the functional turn? Its advocates tend to present the functional turn as a way to meet what they call ‘Strawson’s challenge’: sceptical of Rudolf Carnap’s engineering attempts, P. F. Strawson warned that engineering risks changing the subject, thereby failing to address the philosophical questions that first prompted interest in the relevant concepts (Strawson 1963, p. 506). Engineers have countered that we have continuity of subject as long as we have continuity of function across the change from old to new concept (Haslanger 2020; Nado 2021, §3; Thomasson 2020b, p. 443). The functional turn can be used to formulate success conditions for engineering projects that, duly mindful of Strawson’s challenge, take care to preserve continuity of subject.

Arguably, however, this undersells the functional turn. There is another, deeper rationale for attending to the functions of concepts, namely, that it offers us a way to overcome the authority problem. It is a deeper rationale because once we ask why we should care about continuity, it emerges that we care about it not for its own sake but because we want the engineered concept to tie in with the concerns that precede the engineering. Continuity is really a proxy for authority. We only have reason to adopt a proposed concept if it better serves the concerns that gave us reason to use the concept it refines or replaces. If these antecedent concerns can be focused into a concern to have certain questions answered, it will indeed be reasonable to ask of an engineered concept that the answers it allows one to formulate should still count as answers to those antecedent questions—if our concern is to have certain questions answered, and the engineered concept does not answer them, Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 2. The functional turn and the function specification problem

[p. 1256]

The function specification problem: How are the functions that a given concept performs to be specified? A single concept often has unsurveyably many effects that vary widely from context to context. Function specifications therefore risk appearing arbitrary or ad hoc. As Cappelen sees it, specifying conceptual functions must be either uninformative or unfeasible. In so far as it is feasible—that is, in so far as functions are specified ‘disquotationally’ (Cappelen 2018, p. 187) according to the schema ‘the function of the concept expressed by “X” is to allow us to think and talk about X’—it remains uninformative; and in so far as it promises to be informative by venturing beyond disquotationally specified functions, it becomes unfeasible in the face of the overwhelming plethora of functions that concepts might be said to perform in various contexts. Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 why should we adopt such a concept? Where our concerns lend themselves to being expressed in terms of such a question-and-answer logic, therefore, continuity of subject is a necessary condition on authoritative engineering.

But even where Strawson’s challenge gives expression to the authority problem, it remains an incomplete expression of it, because continuity of subject is hardly a sufficient condition on authoritative engineering: a good or helpful answer does more than just qualify as an answer to the question. Moreover, it is not clear that all our concerns can be adequately expressed as questions. Where they cannot, authoritative engineering will have to rely on continuity of function that does not coincide with continuity of subject. Continuity of subject is thus neither a necessary nor a sufficient condition on authoritative engineering.

While the functional turn allows us to meet Strawson’s challenge, then, its chief merit lies in the fact that it allows us to address the authority problem that underlies Strawson’s challenge. To attend to a concept’s function is precisely to attend to how exactly the concept is enmeshed in human affairs, and what, if anything, makes it instrumental, not just in answering questions, but in satisfying a wider variety of concerns. Function-based engineering promises to solve the authority problem by showing that the very thing that made it worthwhile to recognize the authority of the old concept also makes it worthwhile to recognize the authority of the new concept.

If function-based engineering is to deliver on this promise, however, it needs to overcome what Herman Cappelen (2018, ch. 16) has identified as its principal problem:

[p. 1257]

E-Function: A token of type T has the e-function of type B of producing effect E in system S iff (1) tokens of T produced E in the past; (2) producing E resulted in benefit of type B in S/S’s ancestors; and (3) producing E’s having B-benefitted S’s ancestors contributes to the explanation of why T exists in S. (Simion 2019, p. 263) There are many similar definitions of functions in the literature, but a noteworthy feature of this one is that it individuates types of functions in terms of types of benefits and does not reduce benefits to biological See, for example, Simion (2019, MS) and Thomasson (2020b, p. 444). The theory traces back notably to Wright (1973), Millikan (1989), and Neander (2017). See Glasgow (2020). How to define these conditions is a matter of some debate: see Boghossian (2003), Williamson (2003), Fodor (2004), Glock (2006, 2009, 2010), Eklund (2007), and Scharp (2013, ch. 2).

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 A popular solution to this problem has been to invoke the aetiological theory of functions.8 On one influential formulation of this theory, things have functions in virtue of their history of being selected for having certain effects: although the heart has various types of effects— radiating heat, emitting sounds, and pumping blood—it is a function of the heart to pump blood, because that type of effect has been causally responsible for the retention and proliferation of hearts. Likewise, only some effects of concepts have a history of being selected for because they causally contributed to the retention and promulgation of the concepts they were effects of.

We may prefer to say that, strictly speaking, it is not the concept itself which has effects, but the dispositions to conceptualize the world in the terms specified by the concept. The differential promulgation of concepts can then be understood as the differential reproduction of the set of dispositions that constitute a concept’s possession conditions.9 As these dispositions come to be shared across a community, there comes to be a practice of actualizing dispositions of the same type—of using a certain concept. By looking at the causal history of such practices, the aetiological approach hopes to determine which effects it is the function of these practices to bring about. This is an objective matter: the function could in theory be read off the concept’s causal history, though in practice we are epistemically constrained by our limited capacity to access that history.

One of the most detailed elaborations of an aetiological approach to conceptual functions was proposed by Mona Simion (2019). On her account, we can ascribe what she calls ‘e-functions’ to concepts according to the following schema:

[p. 1258]

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 benefits. Accordingly, not all e-functions need to be reducible to biological functions—the account leaves room for genuinely independent epistemic, ethical or aesthetic functions (see Simion MS). In this respect, the account is liberal in its ascription of functions. In other respects, it is restrictive: concepts can have an e-function only in so far as they acquired one through a history of selection. Engineering guided solely by e-functions will therefore be limited to looking to the past effects of concepts, discerning which effects it is their function to bring about, and engineering them to perform those same functions better. While understanding conceptual functions in terms of e-functions yields a robust solution to the function specification problem, it does not at the same time reliably solve the authority problem. It merely kicks the problems towards the concept’s selected effects: granted that certain effects have been selected for, and that this explains how we came by a concept, why should we now care about these effects? Just because a concept’s past effects contributed to its being around now does not necessarily mean that we want to see these effects realized in the future. The concept may have an e-function without promising to be beneficial for us. Of course, there are cases—especially where near-universal human concepts identified in highly generic terms are concerned—in which explanatory and beneficial effects coincide. But their frequent alignment should not distract from the fact that they may also come apart. The same point may be put by saying that engineering based solely on e-functions is structurally conservative in a way that must sometimes hamper its ability to be authoritative. Perhaps the effects which historically account for the spread of a concept are effects that present-day concept-users no longer care about. Or perhaps we, in the contrastive sense that opposes us to other present-day concept-users, do not want to see these effects realized (because they are sexist or racist, for instance). It will then be in our interest to ensure that those e-functions are no longer discharged. The aetiological approach has difficulties accommodating such radical critiques of inherited functions.

Admittedly, this structural conservatism can be mitigated by the fact that the history through which concepts acquired e-functions can be recent and brief, and some of that history is the history of critique and debunking. Even when guided by e-functions, therefore, the engineering of concepts with a history of being selected for their critical effects will aim at the better discharge of critical functions. E-functionbased engineering could seek to amplify the power of the critical tools in our conceptual inheritance and deploy them against other parts of that inheritance.

[p. 1259]

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 Guidance by e-functions alone nonetheless looks like an unpromising way of engineering authoritative concepts when these are ideological concepts promulgated because of their success in stabilizing oppressive regimes. We might, for example, give an aetiological gloss on J. L. Austin’s remark that ‘our common stock of words embodies all the distinctions men have found worth drawing’ (1957/1961, p. 130); but we would then hardly expect a radical feminist critic like Catharine MacKinnon (1989, 1993) to conclude from this that we should tailor our concepts even more closely to men’s ends. To someone who thinks that the history of our concepts is pervaded by dynamics reproducing the patriarchy, engineering for e-functions points us precisely in the wrong direction. To have any claim to authority in the eyes of critics of the patriarchy, concepts should be engineered against the grain of their recent selection history.

It is perhaps because of these limitations of e-functions that, in a paper with Christoph Kelp, Simion introduces the notion of designed function (‘d-function’): the function that something designed is intended to have by its designer (Simion and Kelp 2020, p. 5). If ‘engineers are, in the first instance, designers’ who develop ‘new d-functions for existing concepts or, alternatively, new concepts with new d-functions’ (2020, p. 5), then engineers are not beholden to the past. But neither are they guided by it. So what are they guided by? Simion and Kelp’s answer is: by the potential of d-functions to turn into e-functions (2020, p. 6). Once ‘launched on a competitive market of concepts’ (2020, p. 6), a concept will either go the way of artefacts in the Museum of Failures, which never had more than an intended function, or it will spread because of its effects and thereby acquire an e-function. On this model, engineers are guided by anticipated e-functions.

But on what basis does one anticipate that a d-function will turn into an e-function? The answer, presumably, has to be: on the basis of how the d-function ties in with people’s present concerns—or so I will argue in the next section.

The conclusion of the present section, however, is that engineering guided only by selection histories is bound to extend inherited patterns of functionality even when the concepts that would have the best claim to authority are those that radically break with inherited patterns of functionality. For function-based engineering to overcome the authority problem, therefore, it is not enough to secure continuity of function between the engineered concept and the concept it replaces. The function must also be such that concept-users could come to see why they should want to see this

[p. 1260]

3. Concepts and concerns In this section, I sketch an account which, by exploiting the tie between concepts and concerns, reliably allows us (i) to specify the functions concepts perform, and (ii) to say why concepts performing such functions have a claim to authority. I shall refer to it as the concern-satisfaction account.12 The strategy of the concern-satisfaction account is to focus, not on the concerns that acted on concepts’ selection histories, but on the present concerns of concept-users: the needs, interests, desires, projects, aims and aspirations they now have and would still endorse upon critical reflection (a qualification I return to in §4). From these suitably cleaned-up present concerns, one can derive the functions that concepts should perform if they are to serve their users well. Present concerns then serve as a basis both for the concepts’ claim to authority and for the specification of their function. Let us consider the latter first. 3.1 Solving the function specification problem All accounts of functions that identify them with effects require some guiding idea by which to determine which effects count as functions. Aetiological accounts draw on selection histories, and ask which among a trait’s effects would, in past manifestations of that type of trait, have fed back into the differential survival and reproduction of that trait. Causal role accounts draw on the idea of a system—an organism, or the central heating of a building—and ask which among the effects of a As Simion and Kelp (2020, p. 16 n. 5) acknowledge.

See, for example, Haslanger (2020).

For a fuller account, see Queloz (MS).

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 function performed.10 The functions we are led to specify on the aetiological approach do not reliably pass this test. A parallel argument can be run for causal role theories that specify functions in relation to the capacity of some system:11 they similarly invite the question of why we should care about the exercise of the capacity of that system. Extant solutions to the function specification problem do not reliably track solutions to the authority problem. If it is to solve the function specification problem and the authority problem in one fell swoop, therefore, function-based engineering needs to make clear why one should recognize the authority of concepts performing the relevant function.

[p. 1261]

C-Function: A concept X has the c-function of type C of producing effect E if and only if (1) users of X have among their present concerns—their needs, interests, desires, projects, aims and aspirations—a concern of type C; (2) under propitious circumstances, applications of X produce E; (3) E stands in an instrumental relation to the concern of type C, which is to say that under propitious circumstances, producing E contributes to the satisfaction of a concern of type C. For concept X to have a c-function, it need not uniformly produce effect E. Nor need the production of E always stand in an instrumental relation to the relevant concern. As is characteristic of functional accounts, our entry point is the situation in which things come together in the right way and a function is successfully performed, so that we initially tiptoe around unpropitious circumstances, backfiring uses, and deviant causal chains (see Blackburn 2005, p. 28). To claim that concept X has a c-function of type C is to claim only that when circumstances concur, applications of X satisfy a concern of type C by producing E.

To identify this effect and the circumstances that assist its production, it is no use staring at a single application of a concept. We have to stand back and consider the concept’s effects across a wider range of contexts to see whether its application reliably correlates with the satisfaction of certain concerns. Guided by our present concerns, we can then identify patterns of functionality: systematic ways in which, via the judgements of which it is a building block,13 this concept manifests the power to make a useful difference to human affairs. The dispositional notion of a power points us to another feature of the concern-satisfaction account: instrumental relations between As Blackburn (2005, p. 29) notes, making concepts rather than judgements our entry point is compatible with the idea that judgements are the basic units of action.

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 certain component of the system contribute to the realization of some system-level capacity. The concern-satisfaction account, by contrast, takes its guidance more directly from what people care about. Instead of scrutinizing history for mind-independent causal patterns or carving up the world into impersonal systems, it examines which among the effects of using a concept tend to tie in with human concerns in a way that is conducive to their satisfaction.

To solve the function specification problem, therefore, this account relies on what might be called concern-relative functions—c-functions, for short—ascribable to concepts on the basis of their tendency to contribute to the satisfaction of concerns:

[p. 1262]

3.2 Solving the authority problem How does the concern-satisfaction account fare with respect to the authority problem? Focusing on c-functions enables us to overcome that problem as well, because our present concerns yield reasons to use certain concepts rather than others. They do so just in case we are better able to satisfy those concerns if we recognize the authority of the concept than if we do not recognize its authority and try to satisfy the concerns without it. This yields a service conception of conceptual authority:14 a concept’s claim to authority over our lives is vindicated if and to the extent that it demonstrably serves, or, equivalently, ties in with a concern we would endorse upon reflection. Relative to that concern, we are better off living under the authority of the concept than outside it. For such an account to be psychologically realistic and steer clear of reductive instrumentalism, it is crucial that it acknowledge how indirect the connection between our concerns and the reasons that actually figure in our deliberation can be. Except in certain cases—my desire to lose weight can be cited as a reason for me to forgo another piece of cake—the reasons that move me to believe or do certain things make no Compare Joseph Raz’s ‘service conception’ of the authority of persons as issuers of obligatory directives (Raz 1995, p. 214).

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 concepts and concerns obtain, not just when a concern is actually satisfied by a concept’s application, but also when the concern would be satisfied if the concept were to be applied. A concept might possess a c-function even if that c-function had, as yet, never been performed. For engineering purposes, it is a significant advantage of the concern-satisfaction account that it can solve the function specification problem even when we consider concepts we do not yet use, or concerns we do not yet satisfy: if there is a concept that would satisfy a concern of ours were we to come to use it, it has a c-function, and we can try to engineer that concept on the basis of that c-function. This type of forward-looking engineering is closed off to aetiological approaches guided solely by extant e-functions. If we develop a new concern and some concept of ours promises to be instrumental in meeting it, this is a relation that could not possibly show up on the retrospective radar of an aetiological approach, because it is sensitive only to functional patterns that have already been realized. The concern-satisfaction account, however, could pick up on functional patterns that have yet to be realized. This makes it a solution to the function specification problem that is particularly useful for conceptual engineers.

[p. 1263]

• Concerns yield reasons for concept use, reasons to adopt certain concepts rather than others. • Concept use yields reasons for concept application: when something fulfils the criteria guiding a concept’s application, this gives one reason to apply the concept to it. • Concept application in turn yields reasons for belief or action: when one applies a concept to something, this gives one reasons to draw certain consequences of the concept’s applicability in one’s thought and action.15 Therefore, even if concerns hold concepts on a long leash, they are indirectly linked. Take, for example, the concept fire. Our concern to stay out of harm’s way gives us reasons to use the concept fire. To be a user of the concept is to be sensitive to the fact that the presence of certain conditions (such as heat, light, smoke and flames) yields reasons for the application of the concept fire. From the applicability of the concept, it in turn follows that we have reason to steer clear of that to which it applies. On this model, we can say that even when concerns do not themselves show up in our deliberations, they nevertheless explain, vindicate, and lend authority to particular ways of structuring and articulating those deliberations.

Exploiting this tie between concepts and concerns, engineers can offer their addressees arguments of the following form: • You have concern C. • By recognizing the authority of concept X and being responsive to the reasons articulable in those terms, you are better able to satisfy concern C than if you did not recognize the authority of concept X and tried to satisfy concern C without it. • Therefore, you have a pro tanto reason to recognize the authority of concept X.

Here I draw, with modifications, on Queloz and Cueni (2021, p. 766).

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 reference to my concerns at all. It is one of the remarkable facts about concepts that they have the power to direct our attention away from egocentric concern-satisfaction. And yet, even when the reasons that guide our thoughts and actions make no mention of our concerns, those concerns nevertheless give us reasons to deliberate in terms of certain concepts, and this indirectly determines what reasons to think or do certain things we end up being guided by:

[p. 1264]

4. The authority of concerns While the concern-satisfaction account solves the twin problems we considered, it raises questions of its own, in particular about the authority of the concerns it draws on. The task of this section is to round out that account by addressing three bundles of questions it invites. 4.1 Pluralism First, whose concerns are at issue? It is a basic fact of politics that different people have different concerns, and that the concerns of one group are sometimes satisfiable only at the expense of another group’s. Before focusing on how concerns vary between people, however, it is worth noting how much philosophical work can be done and has been done by drawing primarily on human beings’ more constant concerns. Certain concerns recur or persist across societies and epochs, either because of natural facts about what kinds of creatures humans are, or because of structural facts about what kinds of challenges they nearly invariably face when trying to live together in society. There is a long naturalist tradition of inferring, from the fact that our conceptual scheme does not receive its basic outline from supernatural sources, that we must look to human concerns to understand what gives shape to our concepts.16 The concerns that are most constantly at work in human societies—to avoid violent conflict with others, find out about the dangers and affordances of one’s environment, secure the resources one needs to survive, and establish the conditions of cooperation—will be A tradition running from Hobbes (1651) and Hume (1739–40, 3.2.2) to Hart (1961/2012, pp. 192–3) and Hampshire (1983, p. 128), to name but a few.

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 This pattern of argument helps one overcome the authority problem by showing that one has reason to use the engineered concept in virtue of one’s concerns. Advocates of e-functions might want to exploit the same pattern of argument, since they are using selection histories as a guide to the concerns that concepts serve. But to be authoritative, an engineered concept needs to tie in with the concerns we now have, and the weakness of aetiological accounts is that they are limited to specifying functions relative to historically formative concerns, whether or not these are now ours. Conceptual engineering, being a prospective and evaluative enterprise rather than a retrospective and explanatory one, seems best served by an account that looks forward to how concepts promise to earn their place in our repertoire as we meet the future.

[p. 1265]

See Fricker (2007, 2016, forthcoming) and Pettit (2018, forthcoming). The fact that relationalism—the view that what appear to be n-place relations are in fact n+kplace relations—does not entail relativism is crucial here; see Spencer (2016). See also Fricker (2010, p. 61), Kusch (2011, pp. 9–10), and Hannon (2019, ch. 2).

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 explanatorily basic in this connection. Recent work by Miranda Fricker and Philip Pettit leans on just such explanatorily basic and comparatively constant concerns to vindicate the authority and specify the function of several of our moral and epistemological concepts.17 But not all concepts draw their authority and functionality entirely from such widely shared concerns. Some concepts draw their authority and functionality from concerns that are specific to one group in contrast to other groups. It follows, on the concern-satisfaction account, that these concepts will be authoritative for that group, but not necessarily for other groups: if the authority of some concept X is understood in terms of its tendency to satisfy concern C, X will be authoritative for concept-users if and to the extent that they would endorse concern C upon reflection; consequently, a given concept could be authoritative for some people and not others.

When coupled with a pluralism about human concerns, the concern-satisfaction account thus yields a pluralism about conceptual authority. This does not mean that there is no fact of the matter about which concepts are authoritative for whom. It is merely that the oneplace predicate ‘concept X is authoritative’ turns out to correspond to a three-place relation: concept X is authoritative for some concept-user A if and to the extent that X serves concern C and A would endorse C upon reflection. This is a perfectly objective three-place fact obtaining between a concept, a concept-user, and a concern.18 Of course, the demand to tailor concepts to the distinctive characteristics of the individual concept-user must be balanced against the concern to communicate and cooperate with others by sharing concepts. As Edward Craig has argued in his discussion of the ‘objectivization’ of concepts, even initially fully subjectivized idiolects would be driven by the demands of communication and cooperation to become communally shared conceptual resources (1990, pp. 82–97; 1993, pp. 81–115).19 What the concern-satisfaction account does entail, however, is that when two groups disagree about, say, which concept of liberty to use, that disagreement is not necessarily simply a matter of epistemic error. It might be that their disagreement reflects an underlying difference in concerns. I take it to be a strength of this account that it allows us to capture this political dimension of concept choice. Some people have reason not to use concepts that other people do have reason to use. The

[p. 1266]

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 term ‘ideological’, in one of its uses, captures just this sense in which the use of a concept might be in the interest of one group while going against the interests of another group.

This pluralism about conceptual authority is grounded in a deeper feature of the concern-satisfaction account: it addresses the authority problem in terms of reasons for concept use that are reasons for a set of people individuated in terms of a common concern, and not necessarily reasons for anyone. To think that concepts should be tailored to something as local and variable as human concerns is to adopt a contingent standard that is not only expressive of our humanity, but also the product of more contingent socio-historical forces and distinctive political commitments. Such a standard will not yield many reasons for concept use that would be recognizable to any rational agent. Yet the question, on the concern-satisfaction account, should not be whether anyone has reason to use a given concept, but whether we do, given certain concerns (where the idea is not that the scope of the relevant ‘we’ is independently given, but that it is a function of how far the concerns at issue are shared).

This contrast between reasons for anyone and reasons for us (that is, those of us who share the reason-giving concerns) marks a fundamental divide in what one is prepared to count as a normative resource sufficient to ground the authority of a concept. For some, the only truly authoritative concepts are those that can be validated as timelessly demanded by rational foundations that anyone has reason to recognize as such—foundations in the realm of Forms, human nature, divine commands, natural law, or universal dictates of reason, for example. For others, reasons for concept use can be less than universal. Impressed by the fact that the concerns that those concepts are meant to serve are, for the most part, local concerns, they see no reason to think that the citizens of twenty-first century constitutional democracies, facing problems such as the climate crisis or surveillance capitalism, would have reason to use the same set of concepts as medieval monks or Bronze Age chieftains. If the problems we face today are unprecedented, it would be surprising if the concepts that proved most helpful in tackling them were unresponsive to that fact. It is not just concept-users that are socio-historically situated, but also the normative standards that their ways of thinking must meet. The concern-satisfaction account offers one way of systematically elaborating that way of thinking about reasons for concept use. Does this account force us into a spectatorial relativism when we confront alternative concepts? There is no reason to think so, because there is no reason to think that the conceptual resources of one group,

[p. 1267]

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 just because they are tailored to that group’s concerns, must leave them unequipped to pass judgement on (and perhaps find good reasons to reject) alternative ways of thinking. We need not suddenly demote and recast all the reasons adverted to by our concepts as being merely reasons for us. When using—as opposed to reflecting on— these concepts, we typically do not have the thought, ‘For us, this is a good reason’. We have the thought, ‘This is a good reason’. That is what it means for it to be, for us, a good reason (see Williams 2000/2006, p. 195). And on that basis, we can still full-throatedly insist that others are wrong to think in terms that we have good reason to consider biased, or unjust in their effects, or overly legalistic, or in some other way wrong-headed.

What a concern-based understanding of conceptual authority can do, however, is to change our understanding of what we are at when we disagree with others about which concepts to use. If we are alive not merely to the differences in our concepts but also to the differences in our respective concerns, we shall be able to discriminate between situations in which we can see that others are making a mistake within a collective cognitive enterprise, as when a teacher can see that a student misunderstands the concept that is being taught, and situations in which, while we still want to say that others’ concepts are wrong, we can at the same time recognize that they are not simply wrong: it makes sense to us that it makes sense for them to use the concepts they use, given how different from ours we understand their concerns to be. They are not just confused, or radically deceived, or irrationally clinging to conceptual hold-overs from another age. This yields the kind of understanding of where the other party is coming from that facilitates respectful disagreement. Indeed, some concepts may derive their authority precisely from the fact that different groups within society have conflicting concerns, because some concepts equip us better than others to respectfully accommodate such a pluralism of concerns. Whenever the recognition of the fact of pluralism gives rise to a second-order concern to respect this pluralism, some concepts can derive their authority from their aptness in serving that second-order concern. The very fact that we live in a society marked by intersubjectively conflicting concerns will then give us reason to use concepts tailored to accommodate and negotiate intersubjective conflicts of concerns. This is to derive uniform conceptual authority from a plurality of concerns: precisely because they fundamentally differ in their concerns, everyone in such a pluralistic society will have reason to use certain concepts.

[p. 1268]

4.2 Non-ideal circumstances A second bundle of questions arises from the worry that one’s concerns will sometimes have been formed under non-ideal circumstances in a way that makes them questionable bases for conceptual authority. What of concerns that have been instilled in the oppressed by the oppressors so as better to oppress them? Or what of ‘adaptive’ concerns that one only has by way of adapting to non-ideal circumstances, such as the gay man’s prudential concern to be perceived as a ‘real man’ in a violently homophobic society? Does the fact that this prudential concern is best served by adopting the concept of real man suffice to render that concept authoritative for him?20 In considering such cases, it is important to see that the formula ‘a concept is authoritative if and to the extent that it serves a concern one would endorse upon reflection’ is meant to carry two qualifications. By only allowing concerns one would endorse upon reflection, it rules out concerns that merely reflect easily corrected misapprehensions, or I am indebted to a reviewer for this example.

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 An example of this structure can be found, again, in Williams’s debate with Dworkin. A ‘thoroughly political concept of liberty’, Williams insists, ‘acknowledges in its construction the on-going existence of political conflict’ (2005a, p. 126). But in the process of smoothing out the conflict between the concepts of liberty and justice by equating liberty with rightful liberty, Dworkin shaves off the political character of the concept of liberty. A concept of liberty that is mindful of pluralism needs to equip citizens to acknowledge the real costs in liberty incurred by those who end up on the losing side of a political decision, however just. Telling them that they do not understand the concept of liberty will not do. Losers’ consent is crucial to the functioning of democratic politics, and whether that consent can be secured depends notably on what we can say to those who feel they have incurred a cost in liberty as a result of a decision going against them. ‘Telling these people that they had better wise up and revise their definition of the values involved’, Williams cautions, is hardly ‘prudent, or citizenly, or respectful of their experience’ (2001a, p. 102). A genuinely political concept of liberty needs to accommodate the fact that people with different concerns will sometimes reasonably resent losses in liberty that are perfectly rightful implementations of the majority will. And it can only do that if it remains wider in scope than Dworkin’s more legalistic concept of rightful liberty.

[p. 1269]

I ignore here the well-known difficulties involved in specifying what is allowed to go into such a process of reflection, and in excluding problematically self-validating forms of that process. See Williams (2002, pp. 191–98), Fricker (2007, pp. 52–53), and Pettit (forthcoming).

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 concerns that one has been deceived into forming against one’s own interests, but that would immediately go dead on one the moment one realized how one came by them (see Fricker 2020). This also helpfully underscores that one’s identification with and endorsement of a concern is not static, but dynamic and responsive to reflection and to changes in one’s sense of what is possible. The maelstrom of our inchoate wishes arguably only settles into fully formed concerns in response to social pressures, such as the demands that others make on us to present ourselves one way or another, declare what we want, and take a stand.21 Even once settled, however, a concern can be unsettled again, called into question by the critical prodding of others, or by the experience of an uncomfortable tension between that concern and the rest of one’s concerns. One’s identification with a concern is thus responsive to critical reflection on it in light of one’s other concerns.

Again, some would insist that critical reflection should look beyond human concerns, to some more independent standard of rightness for concepts, such as the realm of Forms, human nature, the Mind of God, or the dictates of universal reason. In resisting this, the concern-satisfaction account expresses a humanistic commitment to the sovereignty and autonomy of human beings.22 It offers a picture on which the demands our concepts should make on us depends entirely on what demands the world makes on us given the demands that we make on the world. In the end, we are the ones who authorize our concepts, in both senses of that useful term: we are their authors, and we lend them authority. The other crucial qualification is that a concept’s serving a concern yields only a pro tanto reason to consider it authoritative. This reason may be outweighed by countervailing reasons deriving from the concept’s obstruction of other, more important concerns. Even an ideological concept foisted upon me by my oppressors may be authoritative if and to the extent that it serves my concern to stay safe by not stepping out of line. But this falls far short of making it authoritative for me all things considered, since the same concept, if it merits being described as an ideological concept deployed under circumstances of oppression, will at the same time radically frustrate many of my other concerns. There are many concepts we have reason to use that we have even stronger reason not to use.

[p. 1270]

4.3 Wrong kinds of reasons One might worry, finally, that by taking an instrumentalizing view of concepts as tools of concern-satisfaction, the present account ends up encouraging people to employ all their concepts in such a self-consciously prudential spirit. This would seem to leave the account vulnerable to the charge of yielding the ‘wrong kinds of reasons’: it may yield various prudential pro tanto reasons to recognize the authority of the concept moral rightness, for example, but for many philosophers, moral rightness is precisely not authoritative merely prudentially—to do the morally right thing merely out of prudence is to do it for the wrong kind of reason. The first point to emphasize in response to this worry is that we must resist the slide from concept to object: to think instrumentally about concepts is not necessarily to think instrumentally about their objects. On the contrary—what value concepts possess as instruments of concern-satisfaction often depends on those concepts’ ability to make their users think non-instrumentally about their objects and conceive of them as things that possess intrinsic value, or that simply must be done. In such cases, we have instrumentality without instrumental mentality: the concepts may still be instrumental to concern-satisfaction, but the view one takes of things when thinking in those terms is not Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 Because concerns can conflict not just between groups, but also within the breast of one individual, one can be pulled in different directions when assessing the authority of a concept. This forces one to reflect on which concerns one most identifies with and wants to prioritize. Does one identify sufficiently with a concern to be willing to sacrifice other things one cares about to its pursuit? How high a price is one willing to pay? In view of the possibility of such questions, the Manichaean expectation that a concept such as real man must either be or not be authoritative for someone proves too simplistic. The gay man in the violently homophobic society is caught in a bind between conflicting concerns: his concern for self-preservation means that he cannot afford simply to become oblivious to the concept real man and the behaviour it prescribes; at the same time, he cannot fully embrace it without betraying some of his other concerns, such as his concern for authenticity. This calls for a correspondingly complex adjustment in his cognitive economy. He may, for instance, come to acknowledge the authority of the concept real man only in a self-consciously prudential spirit, rendering the acknowledgement of its authority conditional on its serving his concern for self-preservation in a given context, and thereby drastically reining in the concept’s influence on his life and self-conception.

[p. 1271]

See Williams (1985/2011, p. 7). The phrase ‘conceptual ethics’ hails from Burgess and Plunkett (2013a, 2013b).

Korsgaard (1996) might be thought to exemplify conceptual morality. But even if one accepts that the reasons for action bearing on my practical deliberation on what to do should be limited to moral reasons, it is at any rate not obvious that this restriction to moral reasons must extend to the reasons for concept use bearing on the question of what concepts to recognize as authoritative. Reasons for concept use are only indirectly connected to reasons for action (see §3 above). Philosophical work is required to show that the question of conceptual authority is even a practical question in the relevant sense, or that it always makes sense to speak of ‘moral obligations’ to use certain concepts.

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 an instrumental view. Just because the value of people’s concepts is grounded in their concerns does not mean that the only thing that has value, from the point of view of those who think with rather than about those concepts, is the satisfaction of their own concerns, or even of any human concerns—some concepts enable one to see value in states of the world that do not have any human beings in it. Environmental values, for example, need not, at the level of their conceptual content, be centred on human beings; yet, just by not being centred on human beings, they might serve human concerns all the same.

Another important point is that the concerns in relation to which concepts prove instrumental can be the most high-minded moral concerns, and when they are, the resulting reasons for concept use will be moral reasons—we might have reason to abandon one concept in favour of another out of a concern for fairness, or justice, or impartiality. Nothing restricts the concern-satisfaction account to reasons of a prudential kind. If we have reason to use a certain concept of moral rightness to realize our moral concerns, that makes neither the reasons for concept use nor the concerns prudential.

In addition to the moral concerns that may give us reason to use that concept, however, there may indeed be prudential reasons to use it, and it is in whether one recognizes these reasons as pertinent, or dismisses them as the wrong kind of reason, that genuine differences in philosophical outlook will emerge. Perhaps, echoing Williams’s distinction between ethics and morality, we can conceive of conceptual ethics as the broad-minded enterprise of evaluating concepts on the basis of all kinds of considerations that can inform how one should think and live, and contrast this with the more narrowly focused enterprise of conceptual morality, understood as an evaluation of concepts that insists on drawing a sharp boundary between moral and non-moral considerations and only counts moral considerations as pertinent.23 The concern-satisfaction account, in its willingness to recognize prudential concerns as legitimate normative input, can then be identified more specifically as exemplifying conceptual ethics as opposed to conceptual morality.24

[p. 1272]

5. CONCLUSION

I have argued that the problem of showing concept-users why they should recognize the authority of concepts cannot generally be solved by appealing to theoretical virtues. Outside contexts in which we are anyway already concerned to realize such virtues, solving the authority problem requires conceptual engineering to attend to the functions of concepts. What is required, however, is an account of the functions of concepts which shows why concepts discharging such functions should be recognized as having authority. I have argued that to meet this combination of demands, we should specify the functions of concepts directly in terms of the present concerns that are ultimately the source of their authority. The resulting concern-satisfaction account can guide function-based engineering while ensuring that its results can be recognized as authoritative, and it can do so in a way that offers fruitful ways to address the challenges posed by pluralism, non-ideal circumstances, and worries about drawing on the wrong kinds of reasons. This solves the twin problems we considered in principle. But it also indicates that the bulk of the task for engineering still lies ahead, in gaining a nuanced understanding of the variety of concerns our concepts answer to, and of the complex interactions between those concerns and the terms in which we cast our thoughts. For if we base conceptual engineering on the functions that concepts can serve in relation to human concerns, it will only be as good as the understanding of human affairs it draws on.25 I would like to thank two anonymous reviewers as well as the Editors and Associate Editors of MIND for their exceptionally helpful comments. I am also indebted to Damian Cueni, Francesco Testini, David Plunkett, Herman Cappelen, Max Deutsch, Mona Simion, Joseph Chan, David Enoch, Amie Thomasson, Manuel Gustavo Isaac, Alexander Prescott-Couch, and Cheryl Misak for stimulating discussions of these topics. The paper has also benefited from discussion at the workshop on Conceptual Engineering and Pragmatism organized by Yvonne Hütter-Almerigi and Céline Henne at the University of Bologna. My work on this paper was supported by grant N° P400PG_199210 of the Swiss National Science Foundation.

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 But as long as we remain mindful of the difference between instrumentalizing a concept and instrumentalizing its object, and as long as we remember that nothing in the concern-satisfaction account entails that the concerns at issue must be limited to self-interested concerns and the prudential reasons they engender, we have little reason to resist taking an instrumental view of what are, after all, our instruments of thought.

[p. 1273]

REFERENCES

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 Austin, J. L. 1957/1961: ‘A Plea for Excuses’. Proceedings of the Aristotelian Society, 57(1), pp.1–30. Reprinted in his Philosophical Papers, pp. 123–52. Edited by J. O. Urmson and G. J. Warnock. Oxford: Clarendon Press. Page references are to the reprint. Blackburn, Simon 2005: ‘Success Semantics’. In Hallvard Lillehammer and D. H. Mellor (eds.), pp. 22–36. Ramsey’s Legacy. Oxford: Oxford University Press. Boghossian, Paul 2003: ‘Blind Reasoning’. Proceedings of the Aristotelian Society Supplementary Volume 77, pp. 225–48. Brandom, Robert 1994: Making It Explicit: Reasoning, Representing, and Discursive Commitment. Cambridge, MA: Harvard University Press. Brun, Georg 2016: ‘Explication as a Method of Conceptual Re-engineering’. Erkenntnis, 81(6), pp. 1211–41. ――2020: ‘Conceptual Re-engineering: From Explication to Reflective Equilibrium’. Synthese, 197(3), pp. 925–54. Burgess, Alexis, and David Plunkett 2013a: ‘Conceptual Ethics I’. Philosophy Compass, 8(12), pp. 1091–1101. ――2013b: ‘Conceptual Ethics II’. Philosophy Compass, 8(12), pp. 1102–10. Burgess, Alexis, Herman Cappelen, and David Plunkett (eds.) 2020: Conceptual Engineering and Conceptual Ethics. Oxford: Oxford University Press. Cappelen, Herman 2018: Fixing Language: An Essay on Conceptual Engineering. Oxford: Oxford University Press. Chang, Ruth 2009: ‘Voluntarist Reasons and the Sources of Normativity’. In David Sobel and Steven Wall (eds.), Reasons for Action, pp. 243–71. Cambridge: Cambridge University Press. Craig, Edward 1990: Knowledge and the State of Nature: An Essay in Conceptual Synthesis. Oxford: Clarendon Press. ――1993: Was wir wissen können: Pragmatische Untersuchungen zum Wissensbegriff. Wittgenstein-Vorlesungen der Universität Bayreuth. Edited by Wilhelm Vossenkuhl. Frankfurt am Main: Suhrkamp. Davidson, Donald 1990: ‘The Structure and Content of Truth’. Journal of Philosophy, 87(6), pp. 279–328. Dorsey, Dale 2016: The Limits of Moral Authority. Oxford: Oxford University Press. Dutilh Novaes, Catarina 2020a: ‘Carnap Meets Foucault: Conceptual Engineering and Genealogical Investigations’. Forthcoming in Inquiry.

[p. 1274]

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 ――2020b: ‘Carnapian Explication and Ameliorative Analysis: A Systematic Comparison’. Synthese, 197(3), pp. 1001–34. ――and Erich Reck 2017: ‘Carnapian Explication, Formalisms as Cognitive Tools, and the Paradox of Adequate Formalization’. Synthese, 194(1), pp. 195–215. Dworkin, Ronald 2001: ‘Do Liberal Values Conflict?’ In Lilla, Dworkin and Silvers 2001, pp. 73–90. ――2011: Justice for Hedgehogs. Cambridge, MA: Harvard University Press. Dworkin, Ronald, Bernard Williams, Mark Lilla, Thomas Nagel, Richard Wollheim, Frances Kamm, and Steven Lukes 2001: ‘Pluralism’. In Lilla, Dworkin and Silvers 2001, pp. 121–39. Eklund, Matti 2002: ‘Inconsistent Languages’. Philosophy and Phenomenological Research, 64(2), pp. 251–75. ――2007: ‘Meaning-Constitutivity’. Inquiry, 50(6), pp. 559–74. ――2019: ‘Inconsistency and Replacement’. Inquiry, 62(4), pp. 387–402. ――2021: ‘Conceptual Engineering in Philosophy’. In Justin Khoo and Rachel Katharine Sterken (eds.), The Routledge Handbook of Social and Political Philosophy of Language, pp. 15–30. London: Routledge. Fisher, Justin C. 2006: ‘Pragmatic Conceptual Analysis’. Ph.D. thesis, University of Arizona. ――2014: ‘Meanings and Methodologies’. In Mark Sprevak and Jesper Kallestrup (eds.), New Waves in Philosophy of Mind, pp. 54–76. London: Palgrave. ――MS. ‘The Authority of Pragmatic Conceptual Analysis’. Unpublished manuscript. Fodor, Jerry A. 2004: ‘Having Concepts: A Brief Refutation of the Twentieth Century’. Mind and Language, 19(1), pp. 29–47. Fricker, Miranda 2007: Epistemic Injustice: Power and the Ethics of Knowing. Oxford: Oxford University Press. ――2010: ‘Scepticism and the Genealogy of Knowledge: Situating Epistemology in Time’. In Adrian Haddock, Alan Millar, and Duncan Pritchard (eds.), Social Epistemology, pp. 51–68. Oxford: Oxford University Press. ――2016: ‘What’s the Point of Blame? A Paradigm Based Explanation’. Noûs, 50(1), pp. 165–83. ――2020: ‘Bernard Williams as a Philosopher of Ethical Freedom’. Canadian Journal of Philosophy, 50(8), pp. 919–33. ――forthcoming. Blaming and Forgiving: The Work of Morality. Oxford: Oxford University Press.

[p. 1275]

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 Glasgow, Joshua 2020: ‘Conceptual Revolution’. In Marques and Wikforss 2020, pp. 149–66. Glock, Hans-Johann 2006: ‘Concepts: Representations or Abilities?’ In Ezio Di Nucci and Conor McHugh (eds.), Content, Consciousness, and Perception: Essays in Contemporary Philosophy of Mind, pp. 36– 61. Cambridge: Cambridge Scholars Press. ――2009: ‘Concepts: Where Subjectivism Goes Wrong’. Philosophy, 84(1), pp. 5–29. ――2010: ‘Concepts: Between the Subjective and the Objective’. In John Cottingham and Peter Hacker (eds.), Mind, Method, and Morality: Essays in Honour of Anthony Kenny, pp. 306–29. Oxford: Oxford University Press. Hampshire, Stuart 1983: Morality and Conflict. Oxford: Blackwell. Hannon, Michael 2019: What’s the Point of Knowledge? A Function-First Epistemology. Oxford: Oxford University Press. Hart, H. L. A. 1961/2012: The Concept of Law. 3rd edn. Oxford: Oxford University Press. Haslanger, Sally 2020: ‘How Not to Change the Subject’. In Marques and Wikforss 2020, pp. 235–59. Hayward, Max Khan 2019: ‘Immoral Realism’. Philosophical Studies, 176(4), pp. 897–914. Hobbes, Thomas 1651: Leviathan or The Matter, Forme and Power of a Common Wealth Ecclesiasticall and Civil. Edited by G. A. J. Rogers and Karl Schuhmann. London: Continuum, 2006. Hume, David 1739–40: A Treatise of Human Nature. Edited by David Fate Norton and Mary J. Norton. Oxford: Oxford University Press, 2000. Johnston, Mark 2001: ‘The Authority of Affect’. Philosophy and Phenomenological Research, 63(1), pp. 181–214. Jorem, Sigurd 2022: ‘The Good, the Bad and the Insignificant—Assessing Concept Functions for Conceptual Engineering’. Synthese, 200(106), pp. 1–20. Korsgaard, Christine M. 1996: The Sources of Normativity. Cambridge: Cambridge University Press. Kusch, Martin 2011: ‘Knowledge and Certainties in the Epistemic State of Nature’. Episteme, 8(1), pp. 6–23. Lilla, Mark, Ronald Dworkin, and Robert Silvers (eds.) 2001: The Legacy of Isaiah Berlin. New York: New York Review of Books. MacKinnon, Catharine A. 1989: Toward a Feminist Theory of the State. Cambridge, MA: Harvard University Press. ――1993: Only Words. Cambridge, MA: Harvard University Press.

[p. 1276]

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 McPherson, Tristram 2018: ‘Authoritatively Normative Concepts’. In Russ Shafer-Landau (ed.), Oxford Studies in Metaethics, 13, pp. 253– 77. Oxford: Oxford University Press. Marques, Teresa, and Asa Wikforss (eds.) 2020: Shifting Concepts: The Philosophy and Psychology of Conceptual Variation. Oxford: Oxford University Press. Matthews, Robert J. 2010: The Measure of Mind: Propositional Attitudes and their Attribution. Oxford: Oxford University Press. Merrill, Sarah Bishop 1998: Defining Personhood: Toward the Ethics of Quality in Clinical Care. Amsterdam: Rodopi. Millikan, Ruth Garrett 1989: ‘In Defense of Proper Functions’. Philosophy of Science, 56(2), pp. 288–302. Moore, A. W. 1986/2019: ‘How Significant is the Use/Mention Distinction?’ Analysis, 46(4), pp. 173–9. Reprinted in his Language, World, and Limits: Essays in the Philosophy of Language and Metaphysics, pp. 11–38. Oxford: Oxford University Press. Page references are to the reprint. ――2006: ‘Maxims and Thick Ethical Concepts’. Ratio, 19, pp. 129–47. Nado, Jennifer. 2021: ‘Conceptual Engineering, Truth, and Efficacy’. Synthese, 198(7), pp. 1507–27. Neander, Karen 2017: A Mark of the Mental: In Defense of Informational Teleosemantics. Cambridge, MA: MIT Press. Pettit, Philip 2018: The Birth of Ethics: Reconstructing the Role and Nature of Morality. Edited by Kinch Hoekstra. Oxford: Oxford University Press. ――forthcoming: When Minds Speak: The Social Practice that Enables Humanity. Oxford: Oxford University Press. Plunkett, David 2020: ‘Normative Roles, Conceptual Variance, and Ardent Realism about Normativity’. Inquiry, 63(5), pp. 509–34. Queloz, Matthieu 2021: The Practical Origins of Ideas: Genealogy as Conceptual Reverse-Engineering. Oxford: Oxford University Press. Available at https://fdslive.oup.com/www.oup.com/academic/pdf/ openaccess/9780198868705.pdf. ――MS: Reasons for Reasons: An Essay on the Authority of Concepts. Unpublished manuscript. ――and Damian Cueni 2021: ‘Left Wittgensteinianism’. European Journal of Philosophy, 29(4), pp. 758–77. Raz, Joseph 1995: Ethics in the Public Domain: Essays in the Morality of Law and Politics. Revised edn. Oxford: Clarendon Press. Ridley, Aaron 2005: ‘Nietzsche and the Re-Evaluation of Values’. Proceedings of the Aristotelian Society, 105(2), pp. 155–75.

[p. 1277]

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 Riggs, Jared 2021: ‘Deflating the Functional Turn in Conceptual Engineering’. Synthese, 199(3), pp. 11555–86. Scharp, Kevin 2013: Replacing Truth. Oxford: Oxford University Press. ――2020: ‘Philosophy as the Study of Defective Concepts’. In Burgess, Cappelen and Plunkett 2020, pp. 396–416. ――2021: ‘Conceptual Engineering for Truth: Aletheic Properties and New Aletheic Concepts’. Synthese, 198(S2), pp. 647–88. Simion, Mona 2019: ‘Knowledge-First Functionalism’. Philosophical Issues, 29(1), pp. 254–67. ――MS: ‘Epistemic Norms and Epistemic Functions’. Unpublished manuscript. ――and Christoph Kelp 2020: ‘Conceptual Innovation, Function First’. Noûs, 54(4), pp. 985–1002. Spencer, Jack 2016: ‘Relativity and Degrees of Relationality’. Philosophy and Phenomenological Research, 92(2), pp. 432–59. Stampe, Dennis W. 1987: ‘The Authority of Desire’. Philosophical Review, 96(3), pp. 335–81. Strawson, P. F. 1963: ‘Carnap’s Views on Constructed Systems versus Natural Languages in Analytic Philosophy’. In Paul Arthur Schilpp (ed.), The Philosophy of Rudolf Carnap, pp. 503–18. La Salle, IL: Open Court. Thomasson, Amie L. 2020a: Norms and Necessity. New York: Oxford University Press. ――2020b: ‘A Pragmatic Method for Normative Conceptual Work’. In Burgess, Cappelen and Plunkett 2020, pp. 435–58. Tooley, Michael 1972: ‘Abortion and Infanticide’. Philosophy and Public Affairs, 2(1), pp. 37–65. ――1983: Abortion and Infanticide. Oxford: Clarendon Press. Weaver, Bryan R., and Kevin Scharp 2019: Semantics for Reasons. Oxford: Oxford University Press. Williams, Bernard 1985/2011: Ethics and the Limits of Philosophy. London: Routledge. ――1995: ‘Saint-Just’s Illusion’. In his Making Sense of Humanity and Other Philosophical Papers, 1982–1993, pp. 135–50. Cambridge: Cambridge University Press. ――2000/2006: ‘Philosophy as a Humanistic Discipline’. Philosophy, 75(294), pp. 477–96. Reprinted in his Philosophy as a Humanistic Discipline, pp. 180–99. Selected, edited, and with an introduction by A. W. Moore. Princeton and Oxford: Princeton University Press. Page references are to the reprint. ――2001a: ‘Liberalism and Loss’. In Lilla, Dworkin and Silvers 2001, pp. 91–103.

[p. 1278]

Mind, Vol. 131 . 524 . October 2022 © Queloz 2022 ――2001b: ‘From Freedom to Liberty: The Construction of a Political Value’. Philosophy and Public Affairs, 30(1), pp. 3–26. Reprinted in Williams 2005b, pp. 75–96. Page references are to the reprint. ――2002: Truth and Truthfulness: An Essay in Genealogy. Princeton, NJ: Princeton University Press. ――2005a: ‘Conflicts of Liberty and Equality’. In Williams 2005b, pp. 115–27. ――2005b: In the Beginning Was the Deed: Realism and Moralism in Political Argument. Selected, edited, and with an introduction by Geoffrey Hawthorne. Princeton and Oxford: Princeton University Press. Williamson, Timothy 2003: ‘Understanding and Inference’. Proceedings of the Aristotelian Society Supplementary Volume 77, pp. 249–93. Wittgenstein, Ludwig 2000: Wittgenstein’s Nachlass: The Bergen Electronic Edition. Edited by The Wittgenstein Archives at the University of Bergen. Oxford: Oxford University Press. Wodak, Daniel 2019: ‘Mere Formalities: Fictional Normativity and Normative Authority’. Canadian Journal of Philosophy, 49(6), pp. 828–50. Wright, Larry 1973: ‘Functions’. Philosophical Review, 82(2), pp. 139–68.