Luke is the only Gospel that opens by saying what kind of book it means to be. A few lines of formal Greek set out sources, method and purpose — and then, two chapters later, it runs into one of the hardest historical problems in the New Testament.
What the book is
Twenty-four chapters, and by most counts the longest of the four Gospels in Greek — it and Matthew are close enough that the answer depends on the edition. It is also half of a larger work: Acts opens by referring back to a first volume written for the same reader, and the two share vocabulary, style and preoccupations. They are best read as one composition in two parts, and the second is where external checking becomes possible.
The opening verses are a preface of a recognised ancient type, addressed to a man named Theophilus. The writer says that many have already drawn up accounts, that the material came from eyewitnesses from the beginning, that he has investigated everything carefully, and that he writes an orderly account so Theophilus may know the reliability of what he was taught.
Two things follow. First, the writer places himself in a literary tradition — Greek historians and technical writers opened books this way — and does not claim to be an eyewitness, distinguishing himself from them. Second, a preface is a claim, not a verification. Careful research and orderly arrangement were what such prefaces conventionally promised, and those who promised varied greatly in how far they delivered. The preface tells us how the author wanted to be read, not whether the reading is deserved.
The book is nonetheless formally anonymous: Theophilus is addressed, the writer is not named. The attribution to Luke, a physician and companion of Paul, is second-century, and the case for it rests largely on passages in Acts that shift into the first person plural. Whether those mark the author’s presence, a source, or a convention is disputed.
How it is put together
On the majority view Luke used Mark, taking over much of it, keeping the general sequence and improving the Greek. He does not use all of it: a substantial block of Mark has no equivalent in Luke. Luke also shares a large body of sayings with Matthew that is not in Mark. The standard explanation is a lost sayings source, labelled Q — a hypothesis, and always to be named as one; no copy exists and no ancient writer mentions it. The main alternative, the Farrer hypothesis, has Luke use Matthew directly and drops Q; the Griesbach hypothesis reorders all three.
The material found only in Luke is what most readers remember: the shepherds, the good Samaritan, the prodigal son, Zacchaeus, the road to Emmaus. Much of it sits in a long central section in which Jesus travels towards Jerusalem, a journey that is more a frame for teaching than an itinerary.
Two emphases run right through. The first is wealth and poverty, sharper here than anywhere else in the Gospels: the Magnificat sends the rich empty away, the beatitudes bless the poor without qualification and are followed by woes on the rich, and the parables return constantly to money. The second is women — Elizabeth, Mary, Anna, the widow of Nain, the women who funded the travelling group from their own means, the women at the tomb — given weight the parallel accounts do not give them.
The dating question
Conventional ranges put Luke in the 80s CE, the wider spread running from the 70s into the early second century. These are ranges and they are contested. If Luke used Mark, he is later than Mark. Chapter 21 describes Jerusalem surrounded by armies in more concrete terms than Mark’s parallel, which many read as writing after 70 CE, though it can also be read as ordinary siege language.
An earlier date is argued from Acts, which ends with Paul alive under house arrest in Rome and says nothing of his death or of the fall of Jerusalem — silences some take to mean the second volume was finished before either happened. A later date is argued by those who think the author knew Josephus. Neither is decisive. See how we date a text.
The census in chapter 2
Luke dates the birth by a registration ordered by Augustus, calling it the first while Quirinius was governor of Syria, and has Joseph travel to Bethlehem because of his descent from David. This is a real and unresolved difficulty, and should not be smoothed over in either direction.
The problem has three parts. Josephus places the census under Quirinius at the point when Judaea came under direct Roman rule, roughly a decade after Herod the Great’s death — and Luke has already set the birth in Herod’s reign, as Matthew independently does. No empire-wide registration under Augustus is otherwise attested. And Roman assessments taxed property where people lived, which makes a journey to an ancestral town an odd requirement.
Solutions have been proposed for each part: an earlier term for Quirinius, a translation of the Greek making this registration the one before his, a local rather than universal census. Each has been argued seriously; none has secured general agreement. The honest summary is that a writer who opens by claiming careful research has produced a dating that cannot at present be reconciled with the other evidence.
Where the external record touches it
The manuscript position matches the other Gospels and is strong: papyrus fragments from the second and third centuries, complete copies in the great codices of the fourth and fifth. That establishes what the text said early, not whether it was right.
Non-Christian writers confirm the frame, not the story. Josephus mentions Jesus twice, once in connection with the execution of James and once in a longer passage almost universally held to have been reworked by Christian copyists. Tacitus records the founder’s execution under Pontius Pilate; Pliny the Younger describes how Christians met. None of it touches Luke’s narrative. See what external attestation means.
Luke names more datable public figures than the other Gospels — Augustus, Tiberius, Pilate, Herod Antipas, the high priests — most of them independently attested in inscriptions, coins and Roman writers. That shows a writer working with real public history. It does not settle the narrative he sets inside it, and chapter 2 shows the two can come apart.
Archaeology touches the named places. Nazareth and Capernaum have both been excavated, and both were real first-century settlements. The Nazareth synagogue where Jesus reads from Isaiah has not been found, though first-century synagogue buildings have been dug elsewhere in Galilee: the setting is plausible without being confirmed.
Where to start
Read the preface, then chapter 2, then the preface again. That pairing is the whole problem in three minutes: a stated method, and at once a claim the method has not carried.
Then read Acts 1, to see the join between the volumes. Errors here are recorded in the corrections log.