Kantian Intuitionism as a Framework for the Justification of Moral Judgments

• Citation: Vol. 66 No. 2
• Publication year: 2015

Kantian Intuitionism as a Framework for the Justification of Moral Judgments

Robert Audi

[Page 365]

Kantian Intuitionism as a Framework for the Justification of Moral Judgments

by Robert Audi^*

Intuitionism in ethics has often been thought to lack a way to unify the plural standards it endorses. It has also been taken to have at best meager resources for explaining how we should resolve conflicts between prima facie obligations. On this resolution problem, W. D. Ross appealed to Aristotelian practical wisdom. He argued that neither Kantian nor utilitarian ethics (the two most promising rival views he considered) offers an adequate alternative.¹ There is, however, an interpretation of Kant's humanity formula of the categorical imperative for which this negative assessment is unduly pessimistic. This paper will show why. I am not implying, however, that dependence on Aristotelian practical wisdom in certain cases is a fatal defect in a Rossian intuitionism. There are, moreover, many cases of conflicting obligations for which, even if practical wisdom is required for their resolution, there is no reasonable doubt about what should be done. In any event, no plausible ethical theory makes dealing with conflicts of obligation easy or

[Page 366]

uncontroversial; and dependence on practical wisdom is a central element in any virtue ethics and, in some ways, indispensable for practical ethics even when it is guided by a plausible moral theory of some other kind. My aim, then, is not to eliminate dependence on practical wisdom but to construct a broadly intuitionist ethical theory that helps us to enhance both the unity and the applicability of the intuitively acceptable moral principles it provides. Doing this will bring more resources to guide practical wisdom than Ross and later Rossian intuitionists have provided.

I. Kantian Intuitionism as an Integrated View

In early twentieth-century intuitionism, as in much earlier intuitionist writings, three ideas have been salient. First, basic moral principles have been held to be self-evident. Second, self-evident propositions have been claimed to be unprovable. Third, as suggested by the unprovability claim-which brings to mind the idea of propositions so basic as to be Aristotelian indemonstrables-the self-evident has been represented as a category of propositions whose truth a mature rational person can "just see" and can know only by such immediate insight. All three of these views have led to criticism of intuitionist ethics. But if, as I hold in contrast to Ross, Moore, and others, many self-evident propositions can be evidenced or even proved by other propositions, the way is open both to support Rossian moral principles by appeal to a more comprehensive principle or set of principles and to characterize the self-evident in a way that makes it easy to see why self-evident propositions may be not only far from obvious but also subject to rational disagreement.² Ross's list of apparently basic moral principles consists of eight: they posit prima facie obligations of (1) justice, (2) non-injury, (3) veracity, (4) fidelity to promises, (5) beneficence, (6) self-improvement, (7) reparation (e.g. for injuries to others), and (8) gratitude.³ In earlier work (2004) I have represented a modified Rossian intuitionism, including two additional

[Page 367]

principles, as a good ethical theory, but have argued that a Kantian intuitionism that integrates it with a version of the categorical imperative is still better.

The version I have mainly appealed to is the formula of humanity. This might also be called the personhood formula or, alternatively, the intrinsic end formulation. I understand it along lines that, though they reflect some important elements in Kant's ethical texts, do not presuppose a specific interpretation of Kant. In particular, I have sought to show that its negative injunction-which prohibits treating persons merely as means-is explicable in terms of "descriptive" notions, and its positive injunction, which requires treating persons as ends, is explicable, if not descriptively, then at least without dependence on moral notions.⁴ Showing this is important for defending Kantian ethics as well as for providing objective anchors for these notions. If the humanity formula is to serve as one of our basic guides in making moral judgments, we need a way to understand its requirements that does not depend on prior moral judgments.

A proponent of Kantian ethics might accept this constraint but still object that, first, if we adequately understand the categorical imperative, it will suffice by itself to lead us to correct moral judgments in any sufficiently well described case, and, second, that as a principle of final (overall) obligation, it cannot have prima facie principles (such as Rossian ones) as consequences.⁵ Regarding the first, suppose the categorical imperative can lead us to correct (overall) moral judgments. This could be because it enables us to weight and take adequate account of the Rossian principles and thereby the considerations that ground prima facie obligation-something people with practical wisdom would do in any case. This point indicates why the second objection fails: the categorical imperative could be so interpreted as to imply the Rossian principles conceived as each indicating at least one moral element crucial for applying the imperative. In applying it to determine our final obligations, we must, for example, take into account promises, human needs, and potential harms. Indeed, I doubt that the categorical

[Page 368]

imperative can adequately guide moral decision apart from taking account of such factors, and, on the positive side, deeds that enhance non-moral goodness. Even if it could, it would surely do so better as integrated with principles according prima facie obligating force to those and other factors.

Given the detailed development of Kantian intuitionism that I have provided in earlier work, my aim here is to extend and clarify the framework in new ways. Addressing further objections will help in this. One objection is to the effect that the categorical imperative framework is no help in deciding what to do when, as is common, there is a conflict of prima facie obligations and we need to determine what obligation is final,⁶ say an obligation to keep a promise to protect property and an obligation not to harm a person who is fleeing with stolen goods. A related objection is that Kantian intuitionism leaves unclear how close we may come to treating persons merely as means and that, without an account of this, the humanity formula cannot provide grounds for the Rossian obligations.⁷ I take these objections in turn. The response to the first partly deals with the second.

Suppose I am correct in thinking that treating a person merely as a means is roughly treating the person not just solely as a means but (with some qualifications) also with a disposition not to be concerned with any non-instrumental aspects of the treatment. (This negative disposition is needed to account for the force of "merely.") An example of such merely instrumental treatment might be ordering a timid and willing employee to do a risky job, with an intention to let the person struggle alone even if the job becomes highly dangerous. If, as I think plausible, such treatment is prima facie wrong—wrong-making, in another terminology-and prima facie wrong even if the act-type, ordering the job done, is not prima facie wrong in the context, then we have a morally relevant factor that supports fulfilling a Rossian obligation: roughly, a prima facie obligation of sufficient weight to yield overall obligation in the absence of conflicting considerations.

A different example may help: a conflict between the obligation of veracity in making a promise and an opportunity for beneficence. Suppose that, by making a promise I do not intend to keep, I would be

[Page 369]

getting the promisee, who expects to benefit from my keeping it, to do a very good thing for a third party. Making such a promise might also manifest the relevant disposition not to be concerned with non-instrumental aspects of the treatment of the promisee, say the person's suffering from the loss of an important opportunity when I break the promise. Let us assume this. The point that in making the insincere promise one would be using the promisee merely instrumentally weighs against making the promise or, if one does make it, favors reversing one's course and keeping it after all. It also supports the promissory obligation over the obligation of beneficence (though the support may not be overriding). The decision whether to keep the promise may still not be easy, though it might be. In any case, the point is that the added moral ground-that making the insincere, manipulative promise would be, or would at least approach, treating the promisee merely as a means-is helpful and potentially determinative; the point is not that it makes all the conflict cases easy to resolve. That is something no plausible moral view will achieve.

Consider now treating someone as an end, which is mainly a matter of doing-for its own sake-something that is (and is appropriately conceived by the agent as) for the good of the other person. Suppose I have to decide whether to punish a child for bad behavior by keeping the child at home and I know this punishment is reasonable but will make the child suffer. Suppose the retributive considerations, together with the good the punishment will do for the child in the long run, are just strong enough to make the choice between punishing and simply reprimanding difficult, given the desirability of avoiding the suffering. Now suppose that I consider, prospectively, as is appropriate to making a moral decision, not just the two act-types in question-which some other person could realize-but also how I would be treating the child in each case. Treatment is a matter of my conduct, in a sense in...