A number of ethical theorists in southern California (myself included) have recently convened a reading group to consider Derek Parfit’s unpublished manuscript Climbing the Mountain. The book appears to defend a normative ethical theory that borrows elements of contractualism (in the Scanlonian reasonable rejectability vein), consequentialism, and Kantianism. I don’t want to spill many beans, since the book is unpublished, but in his second chapter, Parfit presents some intruiguing ideas about Kant’s Formula of Humanity.
That Formula famously states that we are always to treat humanity, whether in ourselves or in others, always as an end in itself and never as a mere means. The orthodox readings of the Formula tend to see Kant as proposing two distinct ways in which an action (or perhaps speaking in a more strictly Kantian sense, a maxim) can be wrong: It can treat another as a mere means, or it can fail to treat another as end in him/herself. One advantage of interpreting the Formula thus is that it establishes a parallel with Kant’s distinction between perfect and imperfect duties, where violations of perfect duties are ways of treating a person as mere means (in the way that Kant thought that lying treated others as mere means) and violations of imperfect duties (e.g., when we fail in every case to give aid to persons in need) are ways of not treating persons as ends in themselves.
Now Parfit doesn’t quite come out and say this directly, but the best reconstruction I can manage is that he believes he most plausible interpretation of Kant’s Formula identifies only one overarching way in which our actions may violate it. (Parfit’s emphasis is on actions, rather than on maxims or the attitudes that might be associated with actions.) That is, an action that does not treat a person as an end in itself is (or can be) wrong. Furthermore, an action that treats a person (a) as a means, and (b) fails to treat her as an end in itself, is (or can be) wrong. So on the view I attribute to Parfit, failing to treat a person as an end in him/herself is the fundamental concept behind Kant’s Formula, and treating as a mere means is a species of not treating a person as end in itself. So to Parfit’s eye, what distinguishes treating a person as a mere means from permissibly treating them as a means is that the former fails to treat the person as an end.
This intrigues me, simply because it strikes me as very standard to think that Kant meant to capture two species of wrong with the Formula, not one genus comprising two species. And it rejects the notion that there’s something distinctive about treating a person as a mere means. But I’d be curious to know how others understand the Formula of Humanity, and in particular, how we ought understand the ‘mere’ in Kant’s idea of treating as a mere means.
I’m no Kant scholar, but I thought that the idea was that the conjunction of using someone while refusing to take their status as an end in themselves was what was wrong. So there is one two part test as I read Kant. I suspect that this is related to his consequentialism. If not worrying about the ends of others was intrinsically wrong, then one would be required always to worry about how other people were faring. Admittedly two principles might allow this to be trumped by the not-using principle, but since I think it is clear that the no-using principle is not general principle against using people but rather one which restricts using unless one *also* treats those one uses as ends in themselves, it isn’t obvious how or why that version of the other principle would override the requirement always to treat another as an end.
I actually think this is the part of Kant that comes the closest to working, but I’m still no Kantian. So I’m sure others (Robert?) will have more reliable info on this.
I hesitate to claim any sort of expertise, as I don’t have an extensive background in Kant, let alone from the German, but I had intuitively understood the two as halves of one thought. Means vs ends seems an automatic dialectical pairing, where to every end belongs the several means available to achieve that end, be they necessary or merely sufficient. Subjectively, X is either an end, or a means to that end.
People are rational beings; Kant sets great store by rational beings. Inanimate objects may be mere means, non-rational animals may be mere means, but to treat a rational being as a mere means is, by way of the categorical imperative, to deny the right to be treated as anything other than a means. If you are not morally obliged to respect the ends of another rational being, no other rational being is morally obliged to respct your ends. All rational beings become mere means. The only CI-consistent way of achieving one’s ends is to respect the ends of every other rational being. As we cannot know another’s subjective ends — but we must treat them as though their ends are equal to our own — despite the fact that another rational being may be necessary to the process of achieving an end, we are morally obligated to respect the sovereignty of every rational being to receive their assistance toward our end. So rational beings synthesize means and ends, and the achievement of any end becomes contingent upon that sovereignty, the respect of that end-in-itself nature.
I guess I have always read Kant as Parfit does, more or less. ‘Using rational nature as a mere means’ is nothing more nor less than ‘not treating it as an end it itself’. ‘Treating something as an end in itself’ is the primary notion here; when you use something as a means and don’t also do that, you’re treating it as a mere means. There’s nothing more to it. But I could be wrong.
As I understand it, the distinction between perfect versus imperfect duties on the second formulation concerns two differrent ways you can fail to treat something as an end in itself. There is a difference between ‘not treating as an end in itself’ when it means ‘acting in a way that *conflicts* with that status’ versus when it means ‘acting in a way that does not *harmonize* or *positively agree with* that status. (OK, that’s foggy. So think here of not using a large, beautiful diamond as a plug for your sink (conflicts with its value) versus failing to polish it, put it on display, etc. (does not ‘positively agree with’ that value) Something like that.)
I think the second, ‘positively agree with its status as end in itself’ idea *can*, as Mark suggests, be read as offering an openning for a consequentialist (see, for instance, David Cummisky’s reading of it). But I think that requires the idea that you must think that it requires maximizing.
Let me add that, although on your reconstruction Michael, Parfit thinks there’s only one overarching duty, Parfit himself talks as if there’s a lot more ground covered under this formula. In any case, apparently he thinks the only true bit of Kant here is that we must not *regard* people as mere means or things, and must also *regard* them as ends in themselves, neither of which helps to tell us what we must *do*. Or that’s my reading of him.