Towards an ethics of conceptual engineering

ABSTRACT

This paper is a prolegomenon to an ethics of conceptual engineering. Ethics is construed as primarily concerned with reasons for action, not belief, and it is argued that most such reasons are to be understood in terms of their connection with well-being. In the case of conceptual engineering, this is the well-being of the engineer and of others. There are alethic reasons for conceptual engineering, but they are derivative, as are many of the philosophical norms applying within conceptual engineering itself, such as parsimony and respect for the history of a concept. There are moral reasons for conceptual engineering, and practical wisdom will be a core virtue for conceptual engineers. Reaching agreement should be a key aim of conceptual engineers, and this is likely to be achieved most effectively by philosophers working together in research groups in close collaboration with other groups. The largely isolationist model dominant in contemporary philosophy itself needs engineering.

KEYWORDS:

• Conceptual engineering, ethics of
• conceptual engineering
• philosophical practice, ethics of
• virtue
• alethic reasons

In an introduction to an excellent recent edited collection, conceptual engineering is helpfully defined by Herman Cappelen as:

(i) The assessment of representational devices, (ii) reflections on and proposal[s] for how to improve representational devices, and (iii) efforts to implement the proposed improvements. (Cappelen and Plunkett 2020, 3)¹

Cappelen goes on to point out that, once one has noted a defect in a representational device, there are four options: live with it, abandon it, modify it, or replace it. In this paper, my aim is to offer a prolegomenon to the ethics of conceptual engineering so understood, both by mapping the various issues that arise, and by proposing various positions on those issues.²

Why the focus on conceptual engineering in particular? Because it is a peculiarly ethically self-reflexive area of philosophy. The aim of most areas of philosophy is to provide us with the truth about some particular domain: the mind, the nature of reality, truth itself. Conceptual engineers seek to do that with concepts, and if necessary to propose reforms to our conceptual scheme which will enable us to understand the world better. But many also intend to assess that conceptual scheme from a broadly ethical point of view, and make ethical recommendations for reform.

One might think that we should see this aspect of conceptual engineering as an import from standard normative philosophical ethics. But in most cases, ethicists do not consider concepts themselves as appropriate objects of ethical reflection. There are a few exceptions, usually consequentialist: Henry Sidgwick’s infamous suggestion, for example, that utilitarians might be required to keep their view to themselves (1907, 489). But here ethics is seen as overriding the alethic aims of philosophy itself. Sidgwick – and other consequentialists – will say whatever it takes to increase net goodness. What makes conceptual engineering especially interesting is its recognition of the possibility of conflict between philosophical truth and broader ethical aims.

But this is not to say that there is no room for an ethics of other areas of philosophy, in particular, perhaps, an ethics of philosophical ethics. Indeed, given that philosophy, which for millennia was seen as a way of life, is now viewed primarily as a profession, it could be said that there is a need for a modern ethics of philosophy in general. If so, this paper might be seen as a prolegomenon to that project also.

1. Ethics, action, and practical reasons

The fundamental ethical question is the first-personal version of what Bernard Williams (1985, 1) calls ‘Socrates’ question’: how should one live (Republic 352)? What primarily matters in our lives, from the ethical point of view, is how we act. As Williams puts it, ‘Should draws attention to the reasons I have for acting in one way rather than another’.³ Williams here uses the singular form of the first-person, though of course Socrates’ question can be asked collectively. The question will then arise for the individuals involved: ‘What part should I play in our collective activity?’; and for the collective: ‘Given that each of us has certain individual reasons to act, how, in the light of those, should we act?’.

Some action-descriptions are very general, and can refer to the spheres that constitute our lives: working, engaging in relationships with others, pursuing a hobby, and so on. Others specify constitutive activities or actions within these spheres (such as drawing up a contract, or having a conversation). These spheres and their constituents constantly overlap and interact, and managing and negotiating them is the central component of practical reasoning. Socrates’ question, then, is not merely a matter of general reflection, prior to action. It is answered, if only implicitly, every time we make any practical decision (Aristotle 1894, 1094a1-3).

‘To engineer’ is a verb, and Cappelen mentions four actions in his definition of conceptual engineering (assessing, reflecting, and so on), and another four in his list of options. The ethics of conceptual engineering, then, concerns that activity and its constituents: whether to engage in the practice at all, and, given that there is a case for engagement, how to go about it. At this point, it is important to remember that the should in Socrates’s question is not restricted to any particular domain or point of view (Williams 1985, 4–6). It is not asked with reference to, say, self-interest, or moral duty, though it is of course open to answers stated in these and other terms.

Haslanger (2020, 241–2) distinguishes two different forms of conceptual ‘amelioration’: epistemic (in which we improve our understanding of the content of the concept in question) and informational/semantic (in which we change which partition of logical space the concept represents, aiming ‘to change our thought and talk to do better in tracking reality’). She goes on to note that we can do better alethically (improving truth-tracking as an end in itself, e.g. by moving beyond a purely biological account of race), pragmatically (making life easier, e.g. by improving co-ordination), or morally (righting certain wrongs, e.g. by extending the concept of marriage to same-sex relationships).

These distinctions are valuable, but note that the status of alethic amelioration differs from that of pragmatic and moral. Imagine that I persuade you that some form of amelioration will improve our tracking of the truth. It does not follow immediately, even if my claim is true, that there is a reason for promoting such amelioration. There has to be some broadly pragmatic or moral reason for doing so – and then the reason for alethic amelioration will itself be derivative. It might be said, for example, that tracking the truth is valuable in itself; but this will ground a pragmatic or moral reason. When it comes to Socrates’ question, then, there can be no conflict between practical reasons on the one hand, and purely alethic reasons on the other. There is a large set of broadly alethic norms, but any reason we have to act to bring it about that our epistemic states are improved in line with these norms cannot rest merely on their promotion of truth-tracking without further qualification.

2. Absolute value

One obvious way one might view truth-tracking in this context is to see knowledge or understanding as valuable in itself. There are various ways in which this might be done. Consider first a position analogous to that taken by Moore (1903, 83–5) on the ‘absolute’ value of beauty. Moore believed that the mere fact that a world contained beauty made it good, and that this goodness grounded a reason to bring such a world into being, even if that beauty remained unperceived for all time.

This view strikes many as implausible: the only value beauty could have would at least require its being perceived, and is perhaps best understood solely in terms of the value of that perception itself (Glover 1984, 109–10). But it seems anyway that even if there is value in such beauty, this value can be outweighed by relatively trivial amounts of other forms of value. Moore suggests that were you given the opportunity to bring about a beautiful world or an ugly world, you would have a reason to bring about the former. But now imagine you are asked to choose between doing nothing, or bringing about a world containing much unperceived beauty along with one individual whose life involves nothing but suffering. Even if that suffering is fairly minor (e.g. a nagging headache), many who accept Moore’s general position on absolute value are likely to believe that this suffering is sufficient justification for not bringing the world into existence. There is no obvious reason to treat understanding any differently, and so it seems that any absolute value which understanding might have, merely as a constituent of the universe, is, in practice, insignificant, and would offer little justification for conceptual engineering.

3. Well-being

Let us narrow our focus from the universe as a whole to the lives of individuals within any universe. Haslanger’s notion of pragmatic improvement could be specified further in terms of the notion of well-being. Consider first the reasons an agent might have grounded in their own well-being. Some philosophers have argued that, though there are reasons, none of them should be understood in any important sense as self-interested, either because, as utilitarianians claim, all reasons are impersonal or impartial or because, as Kantians claim, all ultimate reasons are moral. But even utilitarians are likely to need an account of self-interest itself, and Kantians might well believe in something like a self-regarding duty of prudence, which again will require an account of the content of that duty.

The debate about well-being – or as the Greeks called it eudaimonia – has largely been between hedonism (the view that well-being consists only in pleasure) and views according to which pleasure is not the only good. Over the last century, a third view developed, according to which nothing is good except the satisfaction or fulfilment of preferences or desires. Since pleasure and commonly alleged non-hedonic goods are standardly desired, we can put this modern view to one side. Nor need we concern ourselves with pleasure, since it is common ground that pleasure is a constituent of the good. If you enjoy some activity (and, let us assume arguendo, it does not, like sadism without consent, involve obvious immorality), this gives you a self-interested reason to engage in it.

The primary alleged non-hedonic constituents of well-being, if we leave virtue and religious faith aside, have been: knowledge or understanding; friendship and other personal relationships; and (again more recently) accomplishment; and autonomy. Here is where alethic amelioration re-enters the picture. Improving truth-tracking just is the advancement of knowledge or understanding. Further, making progress on such matters plausibly counts as an accomplishment, just in so far as it advances knowledge and also, at least in certain cases, in so far as it promotes other goods, such as justice or well-being in general.⁴ And since properly autonomous decision-making itself requires knowledge, conceptual engineering can promote autonomy not only through the decisions involved in the engineering process itself. Finally, since conceptual engineering is a collective enterprise, both within philosophy and in general, as such it provides opportunities for the development of personal relationships.

4. Morality

There seem, then, on any plausible view of well-being, several possible self-interested reasons for engaging in conceptual engineering. But, as we have seen already in connection with Kantianism and with the notion of moral accomplishment, moral theory and the theory of well-being are intertwined. One obvious connection is through the duty or virtue of beneficence. Just as conceptual engineering can in various ways promote the well-being of the engineer, so it can promote that of others. This may be through the goods already identified as internal to conceptual engineering itself, but also through less directly related goods. Consider Haslanger’s example of improved co-ordination. One might think that helping others in some very minor way, such as pointing them in the right direction, is not a moral matter at all; it would be a mistake, perhaps, even to call it supererogatory, in that it is too trivial to deserve praise. But one might prefer a moral theory with wider scope, in which any action that promotes the good of others is to some extent praiseworthy, and perhaps – other things being equal – required.

Haslanger may, then, be marking a distinction between actions which fall below or above some benchmark of moral significance. And there may be strong moral reasons – indeed duties – to engage in conceptual engineering, given the many ways in which injustice and serious harm can themselves involve or depend on conceptual schemes. She mentions the restriction of marriage to same-sex couples, and of course her own work (e.g. 2000) recommends a reform to our concepts of gender and race which would make salient the oppressive moral and political roles of these concepts. In general, it is important to remember that justice is, plausibly understood, not a matter merely of refraining from wrongdoing. It can require the promotion of certain outcomes in the same way as beneficence (see e.g. Rawls 1999, 71). Exactly which concepts require attention will again depend on background moral theory. Someone particularly concerned, for example, at the huge influence of what they see as arbitrary national borders on the well-being of individuals may recommend that we engineer the concept of ‘nation’ to highlight this form of discrimination alongside its role in, say, the grounding of culture and the self-identity of individuals. Or a utilitarian, noting what they see as the terrible effects of the distinction between actions and omissions, might advocate extending the concepts of killing and harming to cover all intentional action which results in, respectively, the death or suffering of another.

The importance of promoting the values underlying certain virtues or duties as well as respecting those virtues in one’s own conduct extends beyond justice. Consider the virtue of veracity. Narrowly construed, it could be said merely to require refraining from telling patent lies. But reflection on the point of veracity is likely to lead us to the value of grasping the truth as at least one of its most significant grounds. Questions may arise in those relatively unusual cases where saying something false actually promotes true belief in one’s hearers, but in general, veracity will promote understanding and knowledge, the potential non-instrumental and instrumental advantages of which we have already noted. But it seems highly plausible that a person committed to truth in their own practice should also be prepared to promote it more generally. And that promotion may be an important result of conceptual engineering. A commitment to truth involves integrity and honesty in one’s own use of language, but also doing what one can to promote conceptual and linguistic care in others.

5. Decisions

As far as values are concerned, then, those that support conceptual engineering are in important part welfarist, involving the engineer’s own well-being and that of others. The potential sources of that well-being include both intellectual and non-intellectual elements, among others pleasure and the diminution of pain. And, on many moral theories (though not of course utilitarianism), there are reasons not only to increase the overall level of well-being, but to improve the justice of its distribution. But there may also be reasons less directly related to well-being, grounded in, say, respect for persons or their dignity, or the value in itself of the possession and exercise of virtues such as veracity as the concern for truth. Further, conceptual engineering, as a practice, has its place in the life of the engineer and the life of their community alongside many other practices, and how any one person should engage in it is a question that cannot be answered without reflection on that situatedness.

On most views, then, the reasons for conceptual engineering will emerge from several different and competing sources, and any decisions on how to engage in it will require a capacity to weigh these values against one another, that is, something like what Aristotle called phronēsis, or practical wisdom (1894, book VI). In broad terms, the more the proposed piece of engineering is directly pragmatic or moral, rather than merely alethic, the greater the role for practical, as opposed to intellectual, wisdom within the engineering itself. The best pragmatic and moral engineers, that is to say, will be virtuous. But even virtue and philosophical ability will be insufficient on their own. In most cases, the engineer will have to master a good deal of empirical information on the methods of conceptual improvement, and the probabilities of success and failure of any proposed programme (how likely is it, for example, that a widespread liberal attempt to change our concepts of race and gender would result in an illiberal project to reform these concepts as slurs?).

6. Principles of conceptual engineering

As Haslanger notes, the justifications for conceptual engineering can be seen as broadly alethic, on the one hand, or pragmatic or moral, on the other. As we have seen, alethic justifications cannot themselves provide ultimate reasons for engaging in conceptual engineering, but given the links between knowledge and other important goods they can provide norms for internal governance of conceptual engineering as a general project supported by pragmatic or moral reasons. As far as the focus of any particular conceptual engineering project goes, there is a rough contrast to be drawn between those which are ‘merely’ philosophical (concerning, say, the concept of ‘supervenience’ (Chalmers 2018, 3)), and those which are both philosophical and directly supported by moral reasons (such as Haslanger’s project on race and gender).⁵ This raises the question whether moral projects should be given priority. The answer here is that individual cases must be judged on their merits. Moral projects may be or less significant (or feasible), and the same is true of the philosophical. Also relevant of course is the expertise of the researcher. There is a stronger case for an individual with a deep knowledge of metaphysics to work on supervenience than someone without that knowledge, and the same point applies to moral cases such as Haslanger’s for a person with, say, a deep grasp of the mechanisms and sociology and linguistic reform.

There may even be cases where the reasons in favour of truth-tracking, veracity, and so on conflict with non-alethic moral concerns. Consider, for example, the notion of human rights as understood by utilitarians. Even if there are no such rights, it seems highly plausible that the practices based on such ‘fictions’ have greatly promoted human good and will continue to do so, and hence that utilitarians might be required to encourage others to use substantively what they see as empty or ‘fictional’ concepts. But these cases are rare. In general, conceptual engineering of any kind, whether ‘moral’ or not, should aim at truth-tracking and the advancement of knowledge.

Which concepts might be strong candidates for engineering? We have to admit that philosophical projects standardly emerge from the particular interests of those engaged in them, and these interests themselves, through the motivation they provide, are important factors in the success of such projects. Nevertheless, these interests can often be directed, and this should be towards significant concepts. There would be little to be said for spending large amounts of time on, for example, attempting to ameliorate the concept of a shovel, or some arcane concept specific to a game played at some particular school. The significance in question should again be assessed in light of the reasons and values outlined above, and may be entirely philosophical, or extend more directly into the moral sphere. And judging that significance will again require not only practical wisdom, construed in terms of virtues of moral character, but also its intellectual analogue.

As far as internal, alethic principles are concerned any project within conceptual engineering should seek to meet the standards required of any inquiry: consistency, coherence, explanatory adequacy, accuracy, and so on. One of its most significant guiding principles should be parsimony. Concepts are complicated and hard to understand, and are often related to other concepts through inference relations which can lie hidden, or will at least require their own elucidation (see e.g. McPherson and Plunkett 2020a). In this connection, it is also crucial to recognize and to understand the history of any concept, prior to any attempt at engineering it (Plunkett 2016). In any area of philosophy, then, the key or basic question should be identified, and properly understood, before new concepts are introduced only if necessary to rephrase or to answer that question. Analysis should then be carried out on these new concepts, which may raise further questions to be answered in the same way. Because philosophers are often unreflectively unparsimonious, their debates almost certainly rest on misunderstandings of the question at issue or of the position of their opponents in that debate. According to Socrates in, for example, Plato’s Meno, interlocutors in some philosophical argument should ensure, at each stage of their argument, that they are in agreement before moving on. This agreement should include not only the content of the philosophical claims under discussion, but the concepts used in those claims. But it is important to provide a balance to the cautiousness of parsimony: the virtue of creativity, involving the recognition of problems and the finding of plausible solutions, is also highly significant in the ethics of conceptual engineering.

7. Self-defeat and disagreement

Conceptual engineering consists of actions that make up a life, and as we have seen answering ethical questions about it is part of an attempt to answer the Socratic question: ‘how should one live?’. That ‘should’, I suggested, following Bernard Williams, is to be understood in terms of reasons for action. So our task, as rational ethical agents, is to assess which reasons we have, and their relative strength, and then to act in the light of them.

One immediate worry here is the possibility of self-defeat. First, there is a danger of a conceptual engineering project’s undermining itself. Imagine that an engineer decides that the concept of a reason should be entirely abandoned. It may then appear as if they have no normative position at all, and must, say, accept the implications of nihilism. But this is to forget that abandonment might be followed by replacement, in this case by what the engineer hopes will be a normatively more adequate concept. (Consider here the analogy with the objections made in the philosophy of mind to the earlier versions of eliminativism by the Churchlands that they could not believe their own position (see Baker 2004)).

There is also a risk of practical self-defeat. It may turn out that reflecting on one’s reasons actually leads one to act in ways that are less in line with those reasons as one conceives of them. Consider for example a person who is faced with a life-choice between work at a high level in the ‘effective altruism’ movement, or a philosophy research fellowship. They are attracted to working for EA, but, wanting to understand more about ethics, they choose the latter, and begin a life of work developing the foundations of utilitarianism. They know that, by their own lights, it would have been better for them never to have entered philosophy, and would be better now to switch tracks. But they cannot now muster the motivation to do so. If we assume the correctness of utilitarianism, this is a case in which it would have been better to have abandoned the very concept of the unsubscripted ‘should’, or, while living with it, to have spent a good deal less time with it. But such cases are, we might plausibly assume, unusual. In general, rational reflection on one’s life, including one’s role in perpetuating or altering our conceptual scheme, is likely itself to be supported by reasons.

I have suggested that philosophers should begin with the most significant concepts and questions in their area of philosophy, and then develop their accounts as parsimoniously as possible. In ethics, that is the Socratic question. But what about the concepts involved in asking that very question – in particular, those of ‘reason’ and ‘action’? What seems clear is that even these basic concepts should not be given a free ride. Some attempts should be made to elucidate them, but it seems fairly obvious, given our situation, that it would be a mistake never to make any practical decisions before being certain about them.⁶ We must be open to the possibility that these concepts themselves require engineering, perhaps of some radical kind (for example, if we were to discover that hard determinism were true, then perhaps the basic question in ethics would axiological: ‘which events would be best?’). Even if, as Derek Parfit has suggested (2011, 31), the concept of a reason is fundamental and irreducible, analogous to that of time, this does not make it a necessary element in our thought.⁷ (The same is true, of course, of the concept of time; see e.g. McTaggart 1908.)

It is hard to conceive of any strong objection to beginning ethics with the Socratic question, though of course there may be objections to particular conceptions of reasons. Normative nihilism is a possibility, but its practical implications – that I have no reason to avoid agony, for example – explain why it has seemed unacceptable to nearly all philosophers over the centuries. But once philosophers do start answering the Socratic question, of course, there will be a great deal of disagreement, both about which concepts we should be using, with what content, and in what way, and about the claims made with these concepts. Conceptual engineering should proceed, I suggest, somewhat along the lines of contemporary science. Research groups (philosophical ‘schools’) should be permitted to develop, but these groups should ensure that they are in dialogue with other groups (see McPherson and Plunkett 2020a, 275n3). If Kantians find themselves attending conferences only with other Kantians, Humeans with other Humeans, utilitarians only with other utilitarians, this is a sign of unjustifiable intellectual isolationism. Only one view can be right, and members of any group should reflect upon the fact that they are likely to be on the wrong track and developing a view only so that – if things go well – it can later be discarded. Openness and humility, then, are key virtues for conceptual engineers as they are for all philosophers, and the largely isolationist model dominant in contemporary philosophy itself needs engineering.

8. Conclusion

This essay has argued for the following as a basis for a philosophical or practical ethics of conceptual engineering:

1. Ethics is primarily concerned with our reasons for actions. Since conceptual engineering consists of actions, it is open to ethical assessment.

2. All reasons for action are, by definition, practical. There are no non-derivative ‘alethic’ or ‘epistemic’ reasons for actions.

3. Any notion of ‘absolute’ value (e.g. the absolute value of knowledge or understanding) provides little, if any, justification for conceptual engineering.

4. Well-being can ground reasons for conceptual engineering, in so far as it promotes pleasure, and, perhaps, understanding, accomplishment, autonomy, and personal relationships.

5. The well-being in question may be that of the engineer (grounding self-interested reasons) or of others (grounding other-regarding ‘pragmatic’ or moral reasons, of beneficence).

6. Given the instrumental, and possibly non-instrumental, value of knowledge or understanding, there are potentially strong derivative reasons for conceptual engineers to follow the alethic norms intrinsic to philosophical inquiry. These reasons may be of sufficient strength to justify purely philosophical conceptual engineering projects over less philosophically significant projects with pragmatic or moral significance.

7. There may be moral reasons for conceptual engineering, grounded in, for example, the promotion of justice, respect for others, and veracity or intellectual integrity. The nature of these reasons depends on which moral theory is correct, as that of the reasons of well-being mentioned above depends on which theory of well-being is correct.

8. To engage well in conceptual engineering is likely to require practical wisdom in judging the pragmatic and moral reasons at stake. Further, any recommendations for conceptual reform beyond philosophy will need empirical as well as evaluative and philosophical support.

9. Candidate concepts for conceptual engineering should be judged in light of both their philosophical and their directly pragmatic or moral significance.

10. Conceptual parsimony and respect for the history of a concept, balanced with intellectual creativity, are crucial to effective conceptual engineering.

11. Philosophical argument should proceed on the basis of agreement between interlocutors on the concepts in play, and reaching that agreement is itself a form of conceptual engineering.

12. Conceptual engineering may be self-defeating, in cases where the reasons supporting it provide stronger reasons against. But, unless utilitarianism is correct, such cases are likely to be unusual.

13. Conceptual engineers, like all philosophers, need to strike a balance between working within a group of like-minded researchers, and seeking criticism and inspiration from groups with whom they are in fundamental disagreement.⁸

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 The introduction is by Cappelen and Plunkett, but some sections are written by one or other author.

2 Plunkett provides his own account of ‘conceptual ethics’ in the section following Cappelen’s; see also Burgess and Plunkett (2013). In Burgess and Plunkett (2020), it is argued that conceptual engineering consists in mereologically complex activities whose parts fall into the categories associated with conceptual ethics, conceptual innovation, and conceptual implementation.

3 It may be that various considerations count in favour of my being a certain sort of person (say, generous or kind, or one with consistent beliefs). But being is not the same as doing; the only way I can be a certain sort of person is to act in certain ways that result, instrumentally or constitutively, in my being of that sort. On the implications of this for conceptual engineering, and reasons for engaging in it, see the following section.

4 It may that any individual can make little significant difference (see e.g. Cappelen 2018, 199). But it is commonly held that merely playing a part in some significant activity (such as defeating the Nazis, or advancing gender or racial justice) can count as an accomplishment or give meaning to an individual’s life.

5 Harder to place is, for example, the discussion on the ‘consequentializing’ of moral theories (see e.g. Dreier 1993).

6 As McPherson and Plunkett put it (2020b: sect. 5), we are justified here in a form of ‘epistemic self-trust’, allowing for the possibility that this trust may in fact be ungrounded.

7 In general, I am inclined to accept Cappelen’s somewhat Quinean axiom (‘It is conceptual engineering all the way down’) rather than Chalmers’ hypothesis that some concepts might constitute ‘bedrock’ (Cappelen 2018, 95–96; Chalmers 2011, 543–57). This axiom is of course consistent with the claim that some concepts are more basic (that is, other things being equal, significant) than others (at least in certain contexts).

8 For very helpful comments, discussion, and encouragement I am most grateful to Herman Cappelen, Matthew McKeever, and David Plunkett.