Constructivism 1 Introduction
4. Objections
4.2. Categoricity
The worry about categoricity runs as follows. It is a conceptual truth that moral requirements are categorical, i.e. if they apply to an agent, they do so independently of that agent’s desires, interests, or ends.72 Since constructivist morality is grounded in the need for social creatures to solve the co-ordination problems with which they are faced, the requirements of
constructivist morality apply to all rational social beings appropriately situated. In this sense, it need not be binding on all rational agents as such, since there is at least conceptual room for rational beings who are not social beings relevantly situated (imagine a race of wholly self- sufficient, independent aliens). But constructivism (at least, as characterised thus far) looks like it will not yield categorical constraints; rather, they will be conditional (if you have a
desire to behave in ways which . . . etc., then you ought to . . .).
One way to make sense of categoricity runs as follows. For a moral requirement to be categorical is for it to be binding on some set of entities independently of their accidental properties. Moral requirements might be understood to be binding on all rational agents as such, for instance. But given the peculiarly social grounding of constructivist morality, this would be implausible. Being a rational being does not, per se, guarantee that that entity will
have any of the required desires, ends, or interests.A second possibility is to understand moral requirements as binding on all social beings as such. This supposes that all social beings will have some shared set of desires, ends, or interests. Understood in this manner,
72 There is another sense in which moral requirements can be said to be categorical, namely the sense in
which they are binding on all rational agents as such. But ‘categorical’ is properly contrasted with ‘hypothetical’, and the primary conception of a hypothetical imperative is one which refers to an agent’s ends. Furthermore, I take it that the central thought behind the attribution of categoricity is that we cannot escape moral requirements simply by ceasing to care about them. I discuss this in more detail below.
categoricity is a matter of inescapability, rather than desire-independence. Social beings, by their nature (and possibly situation), are bound by constructivist moral requirements. In this sense, moral requirements are indeed categorical, i.e. binding on all social beings as such. But the phrase ‘social beings’ is, here, underspecified. In a minimal sense, the phrase includes all agents living within a society. On this construal, it seems hard to claim that there are any inescapable desires (for instance, a desire to behave in ways justifiable to others is widely shared, and prudent – but a particularly talented and capricious knave might very well lack this desire, and have no reason to acquire it). In a more robust sense, ‘social beings’ might be just those with a certain set of motivations and interests. However, this seems like an
objectionably ad hoc stipulation. This being so, the claim that social beings share some set of
properties which grounds constructivist moral requirements turns out to be either trivially true, or false. Bear in mind that the original worry was that moral requirements should not be understood as binding in virtue of one’s desires – they should be non-hypothetical. If the sensible knave claims that she lacks the required motivations, common-sense morality holds that she still ought to (for instance) keep her promises, respect the property of others, and so on. Even if we understand categoricity as a matter of inescapability, the stipulation that social beings share a certain set of motivations will fail to account for this: the sensible knave cannot escape the moral requirement by ceasing to be a social being in the stipulated sense. So the worry about categoricity remains.
We might attempt to tackle this worry head-on. For instance, Korsgaard assumes the existence of unconditional obligations, and then attempts to explain their categoricity:
‘It is our conceptions of ourselves that are most important to us that give rise to unconditional obligations . . . When an action cannot be performed without loss of some fundamental part of one’s identity, and an agent could just as well be dead, then the obligation not to do it is
unconditional and complete.’73
The normative force of these unconditional obligations has to trump any other considerations (or else they would not be unconditional), and hence they must be grounded in something which has this kind of trumping potential; that is, the obligation must arise from something whose value is overriding. For Korsgaard, this is personal identity – because there is nothing worse than death, except for the loss of personal identity, which is itself tantamount to death.
Now this is rather too fast: one might think that there are pro tanto obligations, which bind
unconditionally, in the sense that they provide us with reasons for action no matter what, but
which are not overriding. That an action will harm another person, for instance, may provide us with a (moral) obligation to refrain from that action – but there might, for all that, be overriding reason to pursue that action. Korsgaard’s analysis is an attempt to uncover categorical and overriding reasons for action, and this, I suggest, is too strong.
On the other hand, if the construction procedure depends on widespread but contingent desires – the desire to behave in ways which are justifiable to others, for instance – then it seems as if the resultant obligations will fail to provide reasons for those who lack the relevant desire. This leaves us with two problems: how to account for the appearance of categoricity within moral discourse, on the one hand, and the issue of whether moral
constraints are necessarily categorical, on the other. Now it does seem that categoricity marks an important distinction between norms which are ‘merely’ conventional, and norms which are specifically moral. We ought to keep our promises even if we don’t want to; we cannot escape our moral obligations by ceasing to care about them. There is a shared, strong,
intuition that we cannot escape the demands of morality simply by failing to care about them: the thief who decides not to abide by prohibitions on theft still ought to keep to them. But note that this entails only that moral demands are independent of one’s immediate desires – not that they are categorical. We may also think that we cannot escape the demands of
morality simply by being appropriately positioned, but this intuition is much less strong: there seems to be ample scope within common-sense morality for a degree of relativism, such that the demands of morality may vary across societies.
So there is a sense in which categoricity is a necessary component of morality. Making moral obligations dependent on individual whims would defeat the point of the practice. But the existence of entirely amoral persons might nevertheless be thought to pose a problem for the constructivist; this is the familiar worry about external reasons discussed in the previous chapter. If it is true that we ought not to steal, then everybody has a reason to refrain from stealing. But the hardened amoralist might entirely lack the relevant motivations. Hence, for him, our moral discourse amounts to ‘mere browbeating’.
Accepting this conclusion is, I think, innocuous. It is worth bearing in mind that hardened amoralists of the required sort are few and far between, hence moral discourse does not, in general, amount to ‘mere browbeating’. Even those who do not care about morality per se
still, by and large, have reason to behave morally; even if a ‘sensible knave’ sees nothing wrong with theft, it will still (in general) be in her interest to refrain from stealing, because thieves tend to be punished, if not formally (by the law) then informally (by the disapproval of others). She may have no reason not to steal where she can get away with it, but since
people tend to be poor judges of when they can get away with it, she is best advised to adopt a policy of refraining from stealing. Given the existence of a widespread moral practice,
knavish behaviour is generally (although not always) imprudent.
Nonetheless, let us allow that there are cases where agents have no reason to act in
accordance with the constructivist part of morality. Since the truth of these moral principles is fixed by their acceptability as solutions to co-ordination problems, it is true to say of the knave that she ought to refrain from stealing. That this amounts to ‘mere browbeating’ is unproblematic, because the practice as a whole does not amount to ‘mere browbeating’.
Constructivism need not be understood as committed to thoroughgoing categoricity, in the sense of the provision of reasons for action for all rational agents independently of their particular ends. But it is clear to see why the discourse should retain the appearance of thoroughgoing categoricity: after all, the principles under discussion are proposals for mutual co-ordination of action. As such, they are general, of the form ‘one should not break promises unnecessarily’, and are binding on agents qua social beings, appropriately situated. This, in
turn, results in the appearance of categoricity. Strictly speaking, the reasons generated by these requirements will not be categorical: the highly sophisticated amoralist may have no reason to comply with their demands. But this is innocuous; we still have overwhelming reason to talk of these requirements as categorical.74
4.3.Moral Twin Earth
Constructivism, at least in the formulation set out by Scanlon, is not to be understood as an attempt to provide an analysis of the concept of moral wrongness – of what is meant by ‘morally wrong’. Rather, it is supposed to be an account of the property of moral wrongness. Given this, Moore’s ‘Open Question Argument’ does not apply. However, there is a successor argument to the OQA, namely Horgan and Timmons’ ‘Moral Twin Earth’ argument(s).75 Initially designed to work against synthetic ethical naturalism (according to which moral properties can be identified a posteriori, yielding property identities akin to ‘water is H2O’),
the argument runs as follows. Suppose that Earthlings discover that moral wrongness, for instance, is to be identified with causing harm to innocents. Suppose also that on Moral Twin Earth it turns out that the term ‘moral wrongness’ picks out a different property: that of violating the Ten Commandments. Aside from this, Earthlings and Moral Twin Earthlings are alike in all relevant respects. Now if the synthetic account were correct, then Earthlings and
74 This position is therefore close to the proposal outlined in Joyce 2001, which presents an error theory
regarding the commitment to categoricity coupled with a non-revisionary proposal for moral discourse.