Today I was reading this text from Book I of Manilius' Astronomica. I'll translate the relevant part below.
The boldface text could be (loosely) translated like this: "Nature follows paths that she herself has made, and she does not stray as the inexperienced do."Nam neque fortuitos ortus surgentibus astrisnec totiens possum nascentem credere mundumsolisve assiduous partus et fata diurna,cum facies eadem signis per saecula constet,idem Phoebus eat caeli de partibus isdemlunaque per totidem luces mutetur et orbeset natura vias servet, quas fecerat ipsa,nec tirocinio peccet, circumque feraturaeterna cum luce dies, qui tempora monstratnunc his nunc illis eadem regionibus orbis,semper et ulterior vadentibus ortus ad ortumoccasumve obitus, caelum et cum sole perennet.
On the one hand, this sounds like an early articulation of the idea of laws of nature. If nature follows paths laid down by nature itself and from which it does not deviate, that would be compatible with our idea of a natural law.
Yet there's an important difference between law and path. As you walk along a path, your feet may fall more to one side of the path or the other; and over time, paths may shift, broadening with use or narrowing with desuetude. Charles Peirce, responding to advocates of Hume's argument against miracles, argues that nature is like this as well, and that what we now call laws were once called the course of nature, and Peirce thinks of them as habits that nature has taken on.
The name makes a difference, if only a slight one. Peirce was not trying to argue for particular miracles, but he was urging students of science not to insist that nature behave according to their preconceptions of law. Manilius' idea is not so far-fetched: the laws might not have been there at all, but nature took them on and then, once they became habits, nature has stuck to them. Thinking about science this way alerts us to two possibilities: first, that things might not always have been as they are, and second, that nature might still be taking on new habits. We shouldn't expect nature to stray far from its habitual paths, but on the other hand, what would prevent it from doing so?