Matthew stood first in the Christian canon for most of its history, and for most of that history was read as the earliest of the Gospels. Modern scholarship reversed that judgement, and much of what is interesting about the book follows from the reversal.
What the book is
Twenty-eight chapters covering the birth, ministry, teaching, execution and resurrection of Jesus of Nazareth, written in Greek for readers who already knew a great deal about Judaism. That governs the whole book. Matthew explains almost nothing about Jewish law, festivals or scriptural argument, assuming the explanation unnecessary, and it argues in terms a Jewish audience would recognise as arguments.
The book is formally anonymous. It never names its author, and the title the manuscripts carry — “according to Matthew” — is an attribution, not a signature. The identification with Matthew the tax collector, a minor character inside the narrative, is a second-century tradition, as the attributions of all four Gospels are. Papias, whose words survive only because a later historian quoted them, says Matthew compiled the sayings in the Hebrew language and others translated them as best they could. What work he meant is unclear, and the Greek of our Matthew does not read like translation.
The most visible feature of the book is its use of the Hebrew scriptures. At intervals the narrative stops to observe that something happened “to fulfil what was spoken by the prophet”, and quotes the passage. These formula quotations run from the infancy chapters to the passion, and they are why Matthew is called the Gospel of fulfilment. Several quote texts that in their original setting are not predictions at all, and Matthew’s use of Isaiah has been argued over for as long as there have been commentaries.
How it is put together
The book opens with a genealogy running from Abraham through David to Jesus, then tells the reader how to read it: the list is arranged in three groups of fourteen generations. The arrangement is deliberate and the text says so, so the genealogy is doing theological work as well as recording descent. Four women appear in it, all with irregular or foreign associations, which is a choice rather than an accident.
After the infancy chapters, the teaching is gathered into five long blocks, each closed by a variant of the same sentence — “when Jesus had finished saying these things”. They are the Sermon on the Mount, the instructions to disciples sent out on mission, the parables of the kingdom, the discourse on life inside the community, and the final discourse on judgement and the end. The five-fold arrangement is often read as an echo of the five books of Moses, which is a suggestion rather than a demonstration; the blocks themselves are unmistakable, and they are Matthew’s editorial signature.
The narrative between the discourses is largely Mark’s. On the majority view Matthew had Mark in front of him and reproduced the great bulk of it, tightening the Greek and redistributing material to fit the discourse structure. Matthew also shares a body of sayings with Luke that is absent from Mark, and the usual explanation is a lost sayings source, conventionally labelled Q. Q is a hypothesis: no copy has been found and no ancient writer mentions it. The alternatives are named alternatives, not fringe positions — the Farrer hypothesis dispenses with Q by having Luke use Matthew directly, and the two-Gospel or Griesbach hypothesis reverses the order, making Matthew first and Mark an abridgement.
The dating question
Conventional ranges put Matthew in the 80s CE, with the wider spread running from the 70s to the early 90s. Those are ranges, and contested. They rest on two supports, both arguments rather than measurements.
The first is Markan priority. If Matthew used Mark, any date given to Mark pushes Matthew behind it, with time added for a copy to circulate. That chain is only as strong as its first link.
The second turns on a single verse. In the parable of the wedding banquet at 22:7, the king whose invitation has been refused and whose servants have been killed sends his troops, destroys the murderers and burns their city. The argument is that the detail sits oddly in the parable — burning a city is not required by the story — and reads like a writer glancing back at the destruction of Jerusalem in 70 CE. It is widely accepted, but it remains an argument. The counter-case is that punitive burning is stock imagery in ancient accounts of royal reprisal, and that a stock image cannot carry a date on its own.
The latest possible date is set from the other end: Christian writers of the early second century appear to know the book and quote from it. The general reasoning is set out in how we date a text.
Where the external record touches it
Three kinds of external evidence bear on a Gospel. Manuscripts come first, and the position is strong: fragments survive on papyrus from the second and third centuries, and complete copies in the great parchment codices of the fourth and fifth. By the standards of ancient literature that is exceptional attestation. It should be said plainly what it establishes: what the text said early, not whether what it said was true.
Non-Christian writers come second. Josephus mentions Jesus twice; the shorter notice, on the execution of James, is generally accepted, while the longer passage is almost universally held to have been reworked by Christian copyists. Tacitus records that the founder of the movement was executed under Pontius Pilate in Tiberius’s reign. Pliny the Younger, as a provincial governor, describes what Christians did when they met. Together these establish that the movement existed, that its founder was executed by a Roman official, and that outsiders knew roughly what Christians believed. They do not corroborate the narrative content of Matthew. See what external attestation means.
Archaeology comes third, and works on places and people rather than events. The Galilean villages Matthew names have been dug: Capernaum, where a later synagogue stands on earlier foundations, and Nazareth, where first-century domestic remains confirm a small settlement. A stone found at Caesarea names Pontius Pilate as the Roman official in Judaea, fixing a figure otherwise known only from texts. This is confirmation of setting, which is worth something and not everything: a story can be set accurately in a real place and still be told for reasons other than reporting.
Where to start
Read chapters 1 and 2 together, then the Sermon on the Mount in chapters 5 to 7 at a sitting. The infancy chapters show the fulfilment machinery at its densest; the Sermon shows what Matthew thinks the whole thing is for.
Then read Mark straight through and come back. The overlap is what the synoptic problem is made of. Errors on this page are recorded in the corrections log.