I will put aside any question of possible imprecision of the formulations, I will assume that the Kant's intentions can be understood intuitively reasonably precisely.

Kant gives two forms of the categorical imperative:

  1. Behave in such a way that a reasonable generalization of your action to a universal rule will lead to a benefit to a generic person under this universal rule.
  2. Always treat others as ends and not means.

Kant then claims that 1 is equivalent to 2. Is this correct?

Suppose there were a class of people who liked to be ends. Suppose these people, under certain special circumstances, would like to be treated as objects, for example, as tables. They enjoy being tables, and have tablecloths and wine glasses put on their backs, it does them no harm, and they enjoy the experience, and talk about the experience with joy and regard.

Under these circumstances, knowing that you too might want to be a table at some point, would it be ok to treat these people as means and not ends, at least temporarily?

Is there a precise sense in which 1 and 2 are equivalent, as they seem completely different to me.

Perhaps the answer is that by respecting the wish to be means, not ends, you are treating the people as ends, not means. But then it becomes very difficult to actually determine when you are acting correctly according to imperative 2.

To give a more realistic precise examples, here are some things that are ok under 1 and not ok under 2:

  • purchasing blood plasma from a poor, willing donor.
  • lying to someone about something painful (like whether this person has cancer, or whether she is attractive in that dress, etc).
  • prostitution, dwarf tossing, and other superficially exploitative professions.

In general, I find that formulation 2 is the ethical intuition people are comfortable with (seeing as all three points above are objectionable to a lot of people), and the claim that formulation 2 is equivalent to formulation 1 I think is essentially unfounded.

Can someone who knows Kant better explain?

3 Answers 3


There's quite a bit of scholarly debate concerning what Kant means when he insists on the unity of the various formulations of the categorical imperative; the general consensus seems to be that they all would generate the same duties.

It's worth remembering that Kant does not say that we should only treat people as ends, and not as means—rather, we should not treat them as mere means.

The secondary literature on Kant's moral system is enormous, and ever-growing; Derek Parfit's On What Matters is perhaps the most significant and recent development. A good starting point, however, would probably be the relevant Stanford Encyclopedia of Philosophy article.

  • I was wondering if someone could analyze Kant's argument and see whether it is cogent, in particular concerning the apparent counterexamples I gave. Thanks for the SEP link, there is something there about this, but it does not analyze specific cases like those in the question.
    – Ron Maimon
    Commented Apr 15, 2012 at 0:53
  • 2
    The second paragraph is a really important thing to emphasize. Lots of people misunderstand this when reading Kant. Commented Apr 15, 2012 at 10:08
  • 2
    @RonMaimon: As the SEP article points out, Kant doesn't really make an argument that the two forms are equivalent-- he merely claims that they are. There are, as I pointed out, various ways to interpret the claim (and its validity), but none (that I know of) are syllogistic. Commented Apr 15, 2012 at 10:12
  • I'll accept this, if you put the comment in the answer, and if it is actually true--- I had to read Kant regarding this twenty years ago (as part of the required moral philosophy course they made me take as an undergrad, my only philosophy course), and I remember he had an A implies B implies C implies D implies A argument, which is syllogistic, but informal, and I didn't buy it.
    – Ron Maimon
    Commented Apr 15, 2012 at 17:29

I think there is a point or two to pull out of the original question and supporting expansion of the question. First, you state, 'suppose there are people who like...' If we take part 1 of Kant and apply a universal rule evaluation, we could say it is good to give people what they like. So satisfying part 1 eliminates the negative of part 2. In other words you are not treating those people like 'means', your treating them like ends. Second, 'treating someone' like a mean versus and end is a rather complicated by the more imprtant fact of allowing the other party to participate in the decision.

Sadly, I am not more well schooled in Kant, but I do enjoy pondering his universal rule.


As to the example about people who like to be tables - well, as this is their life choice, then treating them as tables is exactly who and what they are by their own admission, therefore we are doing what they want and like..... there would be no real gain for us, except perhaps the convenience of having an extra table to hand, but really the gain is for them.

So we are not treating them as 'means'. We are in fact behaving in a way that promotes their desired goals, i.e. to be 'used' as tables, thus we are meeting their ends.

Moreover, it would fit with the first CI, that the action must be one to which universality can be applied, in that we all have ways in which we like to be treated, even if it is not to be used as a table! Thus we can claim that we are behaving in a way that satisfied the way in which one group of people prefers to be treated, which in itself is subsumed under the universal rule that we take note of everyone's preference/s.

  • As already pointed out by @MichaelDorfman in his answer, it is not about not treating persons as means, but not treating them merely as means, not ends in themselves
    – Philip Klöcking
    Commented Nov 30, 2015 at 8:14

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .