In his Nicomachean Ethics Aristotle argues that there is but a single highest good for humans, namely eudaimonia, that state of flourishing that exists in persons whose practical wisdom is habituated in such a way that they always make choices in accordance with right reason. That eudaimonia is one such "highest good" is virtually trivially true, but must it be the only such good? Can there not be many things that are fully self-sufficient in the sense that we choose them strictly for their own sakes and not for the sake of anything else? Aristotle appears to assume that there cannot be more than one such good, but he offers no argument in defense of this assumption, and that has caused some folks to wonder. Either ethics, as a discipline, is not such as to offer arguments for such things (as Aristotle himself points out in the Nicomachean Ethics, one must not look for greater precision in a given discipline than is possible in that discipline), or the argument for the assumption is, in some sense, obvious.
Since Aristotle is happy to provide arguments for similar assumptions elsewhere in the Nicomachean Ethics, it seems that the argument for this particular assumption must, somehow, be staring us in the face. As it happens, the argument is rather obvious if one is familiar with Aristotle's doctrine of causal explanation, and I was reminded of it recently while perusing Blessed John Duns Scotus' discussion of the unicity of God in his Opus oxoniense (1, dist. 2, q. 3; Vatican ed. vol. II, 222-243). God is, if you will, a kind of "highest good", not just for humans but for everything; yet we assert that there is but one God. Why? Can there not be multiple Gods, all of whom serve as "highest goods" in some theological scheme? In short, why is polytheism impossible? I say "impossible" because, according to Scotus, it is not merely a matter of faith or dogma that there is only one God: it is a matter of logical necessity.
Scotus begins with the assumption that any will that is infinite wills things in the way that they should be willed. This he takes to imply a principle that we may call the principle of natural will: a correct will loves what is lovable to the extent that it is lovable and to the extent that the will is capable of loving, hence an infinite will will love whatever is lovable to the extent that it is lovable without exception. Suppose, then, that we posit two such infinite wills, that is, two Gods, calling one A and the other B. Both A and B, then, will love whatever is lovable to the extant that it is lovable and without exception; since both A and B are infinitely lovable, then each will love the other infinitely. Here Scotus introduces an assumption that must be unpacked. He says that everything loves its own being more than any other, just so long as it is neither a part nor an effect of this other. We may call this the principle of natural love: fundamentally it means that, given a particular nature (e.g., human, dog, divine), the conscious awareness and will of any being with that nature will be most intimately familiar with its own being rather than that of any other particular nature and, hence, most naturally able to will and to do what is best for that particular nature (here "to love" means something along the lines of "having what is best for X at heart and in one's will").
In the case of our two Gods, A and B, we find that each of them is infinitely lovable, hence B is to be infinitely loved by A. And yet A must naturally love itself more than anything else, including B. But if A loves itself more than it loves B, then it does not love B infinitely, even though B is deserving of infinite love from A. If A does not love B infinitely, A is not acting in accordance with its own nature and, hence, cannot be infinite. So either A loves B as much as it loves itself and, hence, violates the principle of nature love; or A loves B less than A loves itself and violates the principle of the natural will. Both are conceptual impossibilities, hence the actual existence of more than one God is conceptually impossible.
The principle of the natural will and of natural love are, I think, unfamiliar to us and yet perfectly acceptable. If they seem strange, though, Scotus offers an ancillary argument based on this one. He remarks that there are two ways in which A may love B. Either A may love B for its own sake, or it may simply use B. If it merely uses B, the love is inordinate. If it loves B for its own sake because of B's nature, then, having the same nature as B, A will love itself for its own sake as well. But this means that A is beatified by two distinct objects, both A and B, neither of which depends upon the other, for A is made happy by itself just as much as it is by B. But it is conceptually impossible to find perfect beatitude in two distinct objects, because either one may be destroyed without any loss of beatitude, hence complete beatitude is not dependent upon either object.
It seems to be something like this latter argument that Aristotle must have in mind in the Nicomachean Ethics: humans have only one final good because the very notion of a "final good" seems to entail that there could only be one such thing. Scotus' arguments, in other words, have that logical flavor that so characterizes Scholastic argument generally.