The first authors who wrote in chapters were not storytellers. They were compilers of knowledge, either utilitarian or speculative, who used chapters as a way of organizing large miscellanies. Cato the Elder’s “De Agri Cultura” (“On Farming”), from the second century B.C.E., was organized in numbered units with titles; Pliny the Elder’s great compilation of Roman science, “Naturalis Historia” (“Natural History”), from the first century C.E., came with a summarium of topics similar to a modern table of contents; Aulus Gellius, a collector of legal and linguistic arcana in the second century C.E., divided his “Noctes Atticae” (“Attic Nights”) into “capita” with long descriptive titles.* These chapters, unlike the “books” of epic poetry, were what we would now call finding aids: devices for quickly locating specific material in long texts that were not meant to be read straight through.I am not sure whether this is true, but it is interesting.
For more, see this article in the New Yorker.
It's amazing to me how recent are such things as spaces between words, chapters, alphabetization of materials, and other organizational principles that characterize the modern books really are. Perhaps there are some fundamental ordering principles that we overlook in ancient texts. In any case, the transition from oral organization to mere bookishness was more gradual than we realize.