From W. Robertson Smith’s Lectures on the Religion of the Semites (1889):

In connection with every religion, ancient or modern, we find on the one hand certain beliefs and on the other certain institutions, ritual practices, and rules of conduct. Our modern habit is to look at religion from the side of belief rather than of practice; for, down to comparatively recent times, almost the only forms of religion seriously studied in Europe have been those of the various Christian Churches, and all parts of Christendom are agreed that ritual is important only in connection with its interpretation. Thus the study of religion has meant mainly the study of Christian beliefs, and instruction in religion has habitually begun with the creed — religious duties being presented to the learner as flowing from the dogmatic truths he is taught to accept. All this seems to us so much a matter of course that, when we approach some strange or antique religion, we naturally assume that here also our first business is to search for a creed, and find in it the key to ritual and practice. But the antique religions had for the most part no creed; they consisted entirely of institutions and practices.

No doubt men will not habitually follow certain practices without attaching a meaning to them; but as a rule we find that while the practice was rigorously fixed, the meaning attached to it was extremely vague, and the same rite was explained by different people in different ways, without any question of orthodoxy or heterodoxy arising in consequence. In ancient Greece, for example, certain things were done at a temple, and people were agreed that it would be impious not to do them. But if you had asked why they were done, you would probably have had several mutually contradictory explanations from different persons, and no one would have thought it a matter of the least religious importance which of these you chose to adopt. […] The myths connected with individual sanctuaries and ceremonies were merely part of the apparatus of the worship; they served to excite the fancy and sustain the interest of the worshipper; but he was often offered a choice of several accounts of the same thing, and, provided that he fulfilled the ritual with accuracy, no one cared what he believed about its origin.

[…] It follows that mythology ought not to take the prominent place that is too often assigned to it in the scientific study of ancient faiths. So far as myths consist of explanations of ritual their value is altogether secondary, and it may be affirmed with confidence that in almost every case the myth was derived from the ritual, and not the ritual from the myth; for the ritual was fixed and the myth was variable, the ritual was obligatory and faith in the myth was at the discretion of the worshipper. […] The conclusion is that in the study of ancient religions we must begin, not with myth, but with ritual and traditional usage.

Two years earlier, Friedrich Nietzsche writes in the second essay of his Genealogy of Morals (1887, trans. Walter Kaufmann, 1966) §12–13:

A word on the origin and the purpose of punishment — two problems that are separate, or ought to be separate: unfortunately, they are usually confounded. How have previous genealogists of morals set about solving these problems? Naïvely they seek out some “purpose” in punishment — for example, revenge or deterrence — then guilelessly place this purpose at the beginning as causa fiendi of punishment, and have done. The “purpose of law,” however, is absolutely the last thing to employ in the history of the origin of law: on the contrary, […] the cause of the origin of a thing and its eventual utility, its actual employment and place in a system of purposes, lie worlds apart; whatever exists, having somehow come into being, is again and again reinterpreted to new ends, taken over, transformed, and redirected […]

However well one has understood the utility of any physiological organ (or legal institution, social custom, political usage, form in art, or religious cult), this means nothing regarding its origin; however disagreeable this may sound to older ears [who] had always believed that to understand the demonstrable purpose, the utility, of a thing was also to understand the reason why it originated — the eye being made for seeing, the hand made for grasping. […]

To return to punishment, one must distinguish two aspects: on the one hand, that which is relatively enduring, the custom, the act, the “drama,” a certain strict sequence of procedures; on the other, that which is fluid, the meaning, the purpose, the expectation associated with the performance of such procedures. [T]he procedure itself will be something older, earlier than its employment in punishment; the latter is projected and interpreted into the procedure (which has long existed but been employed in another sense); in short, the case is not as has hitherto been assumed by our naïve genealogists of law and morals, who have thought of the procedure as invented for the purpose of punishing, just as one formerly thought of the hand as invented for the purpose of grasping.

To find out Nietzsche’s alternative hypothesis, his idea of the causa essendi, you’ll have to read the essay (particularly §5–6).

In §13, Nietzsche writes:

All concepts in which an entire process is semiotically concentrated elude definition; only that which has no history is definable.

Which reminds me of a bit that stuck in my head from Plato’s Laws (Book IV, 712d–713a):

ATHENIAN: Well then, which of you would be prepared to answer first and tell us which of these terms [aristocracy, oligarchy, democracy, tyranny] fits the political system of your homeland?

MEGILLUS: When I consider the political system in force at Sparta, sir, I find it impossible to give you a straight answer: I just can’t say what one ought to call it. You see, it really does look to me like a dictatorship (it has the ephors, a remarkably dictatorial institution), yet on occasions it gets very close to being run democratically. But then again, it would be plain silly to deny that it is an aristocracy; and there is also a kingship (held for life), which both we and the rest of the world speak of as the oldest kingship of all. So when I’m asked all of a sudden like this, the fact is that I can’t distinguish exactly which of these political systems it belongs to.

CLINIAS: I’m sure I’m in the same predicament as you, Megillus. I find it acutely difficult to say for sure that the constitution we have in Knossos comes into any of these categories.

ATHENIAN: And the reason, gentlemen, is this: you really do operate constitutions worthy of the name.

The Athenian, who in this dialogue is the discussion-leader or dialectician who acts as Plato’s inerrant mouthpiece, goes on to point out that each of these “theoretical” constitutions merely puts one group of citizens in charge over the rest, and is named after that group; any “real” constitution would put the law of the gods in charge over all men, and would rightly be named for that.

Notice that the Platonic notion of law and its origins diametrically opposes the Nietzschean notion — and, per our first two quotations, we shouldn’t make the mistake of confusing their notions with their theories. The notion is prior to the theory, and determinative of.