Deuteronomy is shaped like an ancient treaty, and that resemblance is used to argue both that it is very early and that it is very late. Evidence pulling two ways deserves more attention than evidence pulling one.
What the book is
Thirty-four chapters, almost all of it speech: oratory rather than narrative, in a register unlike anything before it in the Pentateuch. Israel is camped on the plains of Moab, Moses is about to die and will not cross the Jordan, and the book is what he says before he does: a retelling of the journey, a restatement of the covenant, a body of law, blessings and curses, the commissioning of a successor, two poems and a death.
The Hebrew title is devarim, “words”, from the opening phrase. The English one derives, through Greek and Latin, from a phrase in chapter 17 rendered as something like “second law” — a mistranslation that misleads productively, since the book is a second version of law already given, and does not always agree with the first.
How it is put together
The frame is a series of addresses, marked off by editorial headings in the third person announcing a new speech. Chapters 1 to 4 are historical retrospect with a sharp interpretative edge. Chapters 5 to 28, the longest address, restate the Ten Commandments, exhort on loyalty and memory, present the law code in chapters 12 to 26, and close with blessings and curses. A third address sets the choice between life and death. The final chapters commission Joshua and record his death.
The law code covers much of the same ground as the Covenant Code in Exodus — altars, slaves, festivals, homicide, damages, the poor — and differs from it repeatedly. Its slave law releases women on the same terms as men and requires the freed slave to be provisioned. It is hortatory where the Exodus rulings are terse, giving reasons, usually that you were slaves in Egypt.
The largest difference is about place. The Covenant Code permits an altar of earth or unhewn stone and says that in every place where the divine name is remembered God will come and bless — many legitimate sites of sacrifice. Deuteronomy 12 forbids that: worship goes to the one place the Lord will choose to put his name there, and every other site is destroyed. The book never names Jerusalem, which is worth noticing.
Centralisation is a programme rather than a single rule, and the code adjusts to it throughout. With one altar, meat can no longer be slaughtered sacrificially, so secular slaughter is permitted at home if the blood is poured out. Passover ceases to be the household rite of Exodus and becomes a sacrifice at the chosen place. Levites, deprived of local altars, become a landless class commended to charity beside the widow and the orphan. The code hangs together, which is itself an argument that it was designed.
The dating question
Ancient Near Eastern states bound their subordinates by written treaty, in a recognisable order: preamble naming the great king, historical prologue rehearsing what he has done for the vassal, stipulations, provision for depositing and reading the text publicly, witnesses, and blessings and curses. Deuteronomy has almost all of that, in almost that order, with a god in the position of the great king. The parallel is neither vague nor disputed. What it means is, because two bodies of treaty material are involved and they point opposite ways.
- The early argument. Hittite vassal treaties of the second millennium BCE carry the historical prologue as a standard element and keep blessings and curses in rough balance. Deuteronomy does both. First-millennium Assyrian loyalty oaths largely drop the prologue and run to curses at length. If form follows the conventions of its age, Deuteronomy looks second-millennium — close to the period it claims for itself.
- The late argument. The curses of chapter 28 have been compared closely with the loyalty oaths Assyrian kings imposed on vassals in the seventh century BCE, including a stretch where the misfortunes fall in a similar sequence. Judah was an Assyrian vassal in exactly that period. On this reading Deuteronomy takes the form of an imperial loyalty oath and turns it against the empire: the loyalty owed is to another sovereign.
Both are positions, argued by serious people. The treaty form fixes the book’s genre and dates it only as tightly as one’s prior view of which comparison matters more. See how we date a text.
Where the external record touches it
No external document mentions Deuteronomy, Moses or the plains of Moab. The record supplies comparison only — the treaty material above, and the ancient law collections against which the code can be read. That is context, not attestation; the distinction is set out at what external attestation means.
The nearest thing to an anchor is inside the Bible. In 2 Kings 22, during repairs to the temple in the eighteenth year of Josiah — conventionally 622 BCE — the high priest reports finding a book of the law. The reform that follows destroys the high places, clears the temple, centralises sacrifice in Jerusalem, and keeps a Passover such as the text says had not been kept since the days of the judges.
That is Deuteronomy’s programme, item by item, and identifying the scroll with some form of this book is the mainstream critical position — an inference first pressed in the early nineteenth century and still the foundation of most Pentateuchal dating. On this view the legal core was written in the seventh century, and the reform is its implementation.
The traditional alternative deserves its strongest form. Kings describes a book being found, not written. A law scroll neglected through the long reigns of Manasseh and Amon and recovered during building work is not implausible, and the reform then matched the book because the book was the ancient law it claimed to be. It is fair to note what the critical position must assume: that a document presenting itself as the words of Moses was produced by people who knew it was not.
Either way, Kings does not identify the scroll. The identification rests on the fit between reform and book; the fit is close; and that is an argument from correspondence, to be graded as one.
The death of Moses, and the books after
The book ends with Moses climbing a mountain, seeing the land and dying there. The notice is in the third person: thirty days of mourning, no one knows his burial place to this day, and no prophet has arisen in Israel like him since. “To this day” and “since” look back from a distance. Ancient Jewish tradition already debated who wrote the closing verses, so the question is old rather than an invention of modern criticism.
Its influence on what follows is the largest single fact about it. Its vocabulary is distinctive — love the Lord, with all your heart and all your soul, the place he will choose — and the sequence from Joshua to Kings is written in that language and judges every king by its standard, above all by whether the high places were removed — a criterion that exists because chapter 12 created it. The mainstream account of that fact is that Joshua to Kings was shaped as one long work under this book’s influence; that account is a scholarly reconstruction rather than something any of the books states, and it is held with varying confidence about how many hands were involved and when. The fall of the kingdoms is narrated as the working out of chapter 28, exile being the final curse. The prose of Jeremiah shares the diction closely enough that the relationship is a standing question; the oracles of Isaiah come from another stream.
Where to start
Read Exodus 20:22-26, then Deuteronomy 12 immediately after. Two pages, and the central legal problem of the Pentateuch is in front of you in the text’s own words. Then read Deuteronomy 28 slowly, and 2 Kings 22 to 23 after it. Anything we have had to revise is recorded at corrections.