The Census in the time of Quirinius

Sat, 21/12/2019 - 10:00 -- James Oakley

This post goes through a particular set of arguments regarding the historicity of Luke 2, in some detail, for those who would find those details helpful. I have written another post giving a TL;DR version, distilling these details down to identify the main objections raised and the main replies that can be given. Some readers may prefer to start with that briefer version.

I often hear it said that Luke got his history wrong in Luke 2:1-6. He refers to "the first census that took place while Quirinius was governor of Syria". It is said that there was no such census; it never took place.

This is a serious charge to levy against a gospel writer who is historically impeccable when he records any fact for which we have independent records, and who says (Luke 1:1-4) that he worked hard to check everything out with primary eye witnesses.

Much ink has been spilled on this. However, the commentary on Luke by Darrell Bock is thorough and contains a helpful excursus examining this question at length. For my own future reference, and for the benefit of others, here's a summary of his argument.

His conclusion is that we cannot say for certain precisely how we resolve the difficulties. There are gaps in our knowledge of the history of the period that make it hard to be sure. But there are easy solutions to some of the supposed problems, and enough perfectly plausible solutions to the hardest question of all, that we have no reason to dismiss what he writes as definitely fabricated. Just because we can't pin down exactly which solution is the correct one, it doesn't follow that they are all wrong.

The Alleged Problem

“For many it is the clearest example of historical error in Luke’s Gospel.” (903)

Schürer wrote “History of the Jewish People in the Time of Jesus Christ” in 1890 and the issues haven’t moved on. He identifies 5

“1. Nothing is known of a general, empire-wide census in the time of Augustus.
2. No Roman census would require Joseph to go to Bethlehem.
3. There would be no census in Palestine in the time of Herod the Great.
4. Josephus knows of no census before the Quirinian census of A.D. 6. In fact, this Quirinian census is described as an innovation that caused a revolt ({refs supplied}), meaning that no census could have come before A.D. 6.
5. Quirinius could not have been governor of a census at the time of Jesus’ birth, since the governors’ records of this period are well-known and Quirinius is not mentioned ({refs supplied from Tacitus and Josephus}).” (903)

One thing we know for sure

We can say for a fact that Luke’s census is not that from A.D. 6 because Matthew and Luke both tie Jesus’ birth to the reign of Herod the Great. We know he died at the latest 11 April 4 B.C. because Josephus mentions an eclipse just before his death which can be dated to March, and Herod was still alive the following Passover. The census in A.D. 6 is also precisely dated because it’s 37 years after the defeat of Antony which is known to be in 31 B.C.

Tackling objections 1-4 in order

1. “Augustus is known to have instituted three censuses in this period. In addition, other censuses of a periodic nature also seem to have been in place at this time” such as “in Syria, Gaul and Spain”. So it’s “not unlikely that Augustus could have issued such an edict for Palestine” and the reference to the whole empire could simply refer to the prevailing custom of taking regular censuses at this time. (904)

2a. The Romans sometimes “allowed a census to be taken on the basis of local customs, which in a Jewish culture would require an ancestral registration.” (905)

2b. Why was Mary going too? Maybe she had to. Maybe Joseph didn’t want to miss the birth, so brought her along. Maybe she was actually married to Joseph (since ἐμηστευμενῃ can mean ‘unconsummated’). It’s a relatively modern, Western thing to worry “about a pregnant woman’s ‘tender state’” so the journey was not inherently unlikely. Indeed, if this had happened in the A.D. 6, this would have been far more likely, since Herod the Great’s territory was then divided into different territories, and they would not have been able to cross the boundary to register elsewhere. (905)

3. People say a census was impossible during Herod the Great’s time. But even though he was so kingly as to have his own coins minted, that wouldn’t have meant the Romans left completely alone. Indeed, Samaritan taxes were reduced at the start of Archelaus’ rule, which tells us at least that this part of Herod’s territory was being taxed (and so censuses would be taken). (905-6)

4. Just because Josephus doesn’t mention an earlier census, you cannot conclude there wasn’t one. There were particular historical reasons why the census in A.D. 6 inflamed unrest, and an earlier census done according to Jewish customs may well have not deserved a mention. (906)

Objection 5

The problem of when Quirinius was governor is knottier. The solutions proposed fall into three types: lexical, historical and interpretive, but as the categories overlap it’s simpler instead to outline the solutions proposed.

5a. Was Quirinius governor twice (from 11/10 B.C. to 8/7 B.C., and then in the later period)? The argument for this relies heavily on an inscription called the “Lapis Tiburtinus”, but it’s a partial inscription which makes it hard to rely upon. Also, his role in the Homonadensian Wars places Quirinius in Galatia not Syria during that earlier period (907).

5b. Perhaps Quirinius was a legate between 4 B.C. and 1 B. C. (between Varus and Gaius Caesar), the only gap in what we know of governors. This makes possible an earlier Quirinian census, but moves the date of Jesus to later than the time of Herod. This leaves Luke still with an error, but arguably a lesser one (907-8)

5c. Or, modifying 5b, could Varus have begun the census, but the results were collated by Quirinius who then put his name to it? (908)

5d. Or, modifying 5c, maybe Quirinius was not governor when the census was ordered, but the administrator who organised it. Both 5c and 5d build on the fact that a census was complex and took time. 5c then says that it would have finished some time after it started, whereas 5d says that it may have needed another official to administer. It’s hard to choose between these options without more evidence to help us. (908)

5e. A lexical solution, perhaps the term in Luke 2 for “census” is actually the technical term for the resulting taxation, so that the taxation at the end led to the revolt in A.D. 6, but it started during Herod’s time much earlier. The problem with this is that the distinct terms don’t have such distinct meanings; in particular, Acts 5:34 uses the same term as Luke 2 for A.D. 6, and Josephus uses the terms interchangeably. (908-9)

5f. Maybe there’s an ellipsis in Luke 2:2, so that Quirinius was not connected with the census in the original manuscript. “In this view, Luke says that the census is earlier than the well-known Quirinius census of A.D. 6. The main problem with this view is that the syntax of Luke 2:2 is cumbersome at best. This view is possible, but not likely.” (909)

5g. Alternatively, translate πρωτη as “before” rather than “first”, which it can mean, so that the census took place before Quirinius’ governorship. The question is whether it’s likely Luke used πρωτη with such an unusual meaning. (909)

No one of these options immediately presents itself as the superior option, probably because of “the current historical uncertainty regarding the succession of the governorship in Syria”. The most likely views are 5c or 5d, with 5g also possible.


We note in passing that the one view ruled out by all of this is that Luke means the census in A.D. 6. Whatever else he means, if πρωτη means “first”, then this is Quirinius’ first census, and suggests Luke knew of more than one.

“In the light of this and the various possibilities, it is clear that the relegation of Luke 2:2 to the category of historical error is premature and erroneous.” (909)

Blog Category: 

Add new comment

Additional Terms