Jean Astruc, 1753, split Genesis by divine names, Yahweh versus Elohim; Graf and Wellhausen, 1860s-1880s, ordered the strands: J, the Yahwist, southern kingdom, c. 850; E, the Elohist, northern, c. 750, merged after 722; D, Deuteronomy, Josiah's found book, 622; P, the Priestly source, exilic, final form after 539. The evidence: doublets, contradictions, shifting terminology and theology, differing geographic interests. Penn's survey captures the critical consensus: the approach best explains those features, with dates contested but the framework holding.
The rebuttals are equally credentialed: Garrett's Rethinking Genesis, the fragmentary and supplementary alternatives, the argument that repetition is normal Semitic style per Ugaritic archives, the charge that Wellhausen ran Hegel's dialectic on the text, thesis-antithesis-synthesis dressed as philology, and the demolition gibe: if the sources contradict this badly, the redactor was the worst editor in history, yet somehow produced the most read book ever. The hypothesis survives as scaffolding even among scholars who doubt the scaffolding.
The claim: the hypothesis accidentally explains the Bible's durability, a document compiled from rival schools cannot be captured by any single reading, since it contains its own counter-arguments, making it the first proof against proof-texting, an archive in one binding; the seams are not the flaw, they are the anti-fragility. Verdict: the archive endorses the mechanism, competing origins in one cover is this site's entire method, and the Pentateuch, on the critical reading, pioneered it. Whether Moses held the pen or a committee did, the result reads like a design: no single voice gets the last word. Which is, of course, also the archive's rule five.
If the seams are real, why does the stitched book outperform every scroll it was stitched from?