Luke 2's census: history or harmonization?

The fixed points

Four pieces of evidence have not been in serious dispute since the late nineteenth century. Herod the Great died shortly before Passover in 4 BCE. The dating is anchored by Josephus's report of a lunar eclipse that fell shortly before Herod's death (Antiquities 17.6.4), which most chronologers identify with the partial eclipse of 13 March 4 BCE. Quirinius (Publius Sulpicius Quirinius) became legatus of Syria in 6 CE, after the deposition of Archelaus, and immediately conducted a provincial census of newly Roman Judea. The census appears in Josephus, Antiquities 18.1.1-2.6, and triggered the revolt of Judas the Galilean. Acts 5:37 refers to this same revolt and to 'the days of the taxing' as background known to the reader.

What is also stable is the line of Syrian governors between the two dates. Sentius Saturninus governed Syria from about 9 to 6 BCE. Quinctilius Varus succeeded him from 6 to 4 BCE; Varus is the same governor whose later defeat at the Battle of Teutoburg Forest in 9 CE became proverbial. After Varus, the post is held by an unnamed governor or possibly by Quirinius in an unrecorded earlier term, and then by Saturninus again or by another candidate, and then by Quirinius from 6 CE forward.

The Lapis Tiburtinus, an inscription found at Tivoli near Rome in 1764, names a Roman official who was 'legate of Augustus, of Syria, for the second time' (legatus Augusti pro praetore Syriae iterum). The name is missing from the surviving stone. The candidates for the missing name include Quirinius, Lucius Calpurnius Piso, and several others. If the inscription refers to Quirinius, it would prove he held a Syrian legateship before his documented 6-9 CE term. The identification is contested.

Five positions on Luke 2

Timeline of the documented evidence

Three documents, side by side

The grammar of prōtē

Almost all standard English translations render Luke 2:2 with 'first.' The KJV has 'first,' the ESV has 'first,' the NIV has 'first,' the NRSV has 'first.' The Sherwin-White reading does not change the Greek; it argues that prōtē in this construction can carry the comparative sense 'before' rather than the superlative sense 'first.'

The case rests on parallel constructions in John's Gospel. John 1:15 has John the Baptist say of Jesus 'prōtos mou ēn' (he was before me); John 1:30 repeats the construction. Both are clearly comparative; John is not saying Jesus was 'first of me' but 'before me' in time or rank. Sherwin-White argued that the same comparative sense can be read into Luke 2:2, especially given the following genitive participle hēgemoneuontos. On this reading, Luke is saying 'this census happened before Quirinius was governing Syria,' clearly distinguishing it from the 6 CE census he himself references in Acts 5:37.

Critics of the reading note that the grammatical possibility is not the default lexical sense. BDAG, the standard Greek lexicon of the New Testament, lists 'before' as a possible but secondary sense for prōtos and treats Luke 2:2 as ambiguous. Some Greek scholars accept the reading; others reject it. The dispute is not over whether the reading is grammatically possible (it is) but over whether the reading fits the Greek and whether Luke would have chosen an ambiguous construction if he meant 'before.' This is the linguistic core of the harmonizing argument.

What the patristic record says

The patristic tradition does not settle the question. Justin Martyr (First Apology 34, c. 150 CE) refers to records of the Quirinius census as available in Rome but does not address the Herod chronology. Tertullian (Against Marcion 4.19, c. 207 CE) names Saturninus rather than Quirinius as the governor at the census; this is the strongest patristic testimony for the 'before plus Saturninus' reading. Eusebius (Ecclesiastical History 1.5.2, early fourth century) cites Josephus on Quirinius's census but treats it as the same event Luke describes, attempting a harmonization that suggests Eusebius read Luke 2:2 in the 'first' sense and was reconciling it with Josephus by other means.

Origen (Against Celsus 1.45, mid-third century) responds to a critic who challenges the census detail but does not develop a grammatical defense. The patristic record, taken as a whole, shows that the apparent conflict between Luke 2:2 and the 6 CE date was visible early but was addressed in different ways by different writers, with no single solution achieving consensus.

What each reading has to live with

The 'before Quirinius' reading has to live with the lexical default. Most readers, ancient and modern, take prōtē as 'first.' The reading depends on a grammatical possibility that requires defending against the standard sense, and on the argument that Luke would have used the ambiguous construction deliberately to distinguish his census from the famous 6 CE one. This is the simplest harmonization if the grammar is accepted, and the heaviest lift if it is not.

The 'earlier Quirinius governorship' reading has to live with the absence of direct evidence. The Lapis Tiburtinus is genuinely interesting and may name Quirinius, but the stone's missing name is the entire case. Tacitus mentions Quirinius's Cilician campaign without dating it or naming it a Syrian governorship. The reading turns on inferring an administrative arrangement from indirect evidence, which proponents (Ramsay, Hoehner, Maier) argue is reasonable for the period and which critics argue stretches the evidence further than it goes.

The 'Luke conflated' reading has to live with Luke's general accuracy and with Acts 5:37. If Luke knew the 6 CE census well enough to refer to it in Acts as a known event, attributing a chronological confusion to him at Luke 2 requires explaining why his Acts reference is precise while his Luke 2 reference is not. Brown's answer is that Luke's narrative purposes at Luke 2 led him to use the census as a framing device for the Bethlehem journey, and that the precise dating was secondary to that purpose. Critics of this reading hold it concedes too much to Luke's literary intentions over his historical claim.

The 'literary device' reading has to live with Luke's preface and with the independent Bethlehem tradition in Matthew. If Luke wrote with the historical intent his preface claims, treating the census as pure literary construction reads against the grain of his own self-presentation. The Bethlehem location is also older than Luke; Matthew has Jesus born there by a different mechanism. The reading is internally coherent for readers who treat the Gospels as primarily literary, but it is hard to integrate with Luke's stated purpose.

The 'before plus Saturninus' combined reading has to live with the compounding of disputed elements. It combines a contested grammatical reading, a contested epigraphic identification, and a reconstructed administrative arrangement. Each piece is plausible on its own; combining them produces a position that requires several uncertain elements to align. Proponents argue this is the cost of the data, since the data themselves are uncertain at several points. Critics argue the combined reading multiplies hypotheses past what the evidence supports.

Reading the verse with the question open

Five positions, all of which preserve the central claim of the narrative (that Jesus was born in Bethlehem during the reign of Herod the Great), differ on what to do with the census detail. The grammatical reading turns the verse into a clarification. The earlier-governorship reading turns it into a record of a previously unattested administrative event. The conflation reading accepts a chronological error in the framing. The literary reading treats the framing as constructed. The combined reading patches the pieces back together with multiple harmonizing moves.

Reading Luke 2 with the question open means knowing that the Greek of verse 2 admits more than one rendering, that the Roman administrative record of the period is incomplete in ways that matter, and that the five positions above have been live in the conversation for over a century. Each position has serious defenders and serious problems. Each is a way of holding together the documentary evidence we have. The narrative itself, of a poor family forced by an external administrative event to travel to an ancestral town for the birth of a child, holds whichever historical frame the reader assigns to the census detail.