One of the central propositions in Kant's ethical system is that persons, and persons alone, are proper objects of respect. But since something can correctly be regarded as an object of moral action if, and only if, it is worthy of respect, it follows that persons, and persons only, can be objects of moral action. For Kant this is equivalent to saying that only persons are ends in themselves, or, put otherwise, that persons are the sole limiting conditions on moral action. From this it follows that any being that is not a person can, with moral justification, be used merely as a means.
If Kant is right in his claim that persons alone are the proper objects of respect, then serious consequences follow concerning the moral status of animals. For unless they are rational they cannot be regarded as ends in themselves, and indeed only by the use of a questionable argument can animals be shown to give rise to any moral duties at all. For if Kant, or any philosopher in the Kantian tradition, wishes to say that animals are worthy of moral consideration, he must arrive at this conclusion by showing that our duties towards animals are in some way dependent on our duties towards persons. Thus, in so far as Kant wishes to claim that we have duties towards animals he must take one or other of two lines. He can say that animals are persons. Or he can say that there are moral limitations on our treatment of animals because certain kinds of treatment of animals can involve us in, or lead us to, treating persons merely as means and not at the same time as ends.
(Alexander Broadie and Elizabeth M. Pybus, "Kant's Treatment of Animals," Philosophy 49 [October 1974]: 375-83, at 375-6 [footnote omitted])