The following essay surveys all the evidence both to this effect and against all known attempts to reconcile these authors. It was originally written in 1999 and was revised in 2000 to make it more readable and complete, and to take into account new claims and scholarship; two more revisions of the text were made in 2001. Each section of this essay begins with a summary of conclusions in bold type, followed by a sometimes lengthy discussion of the evidence leading to those conclusions. As a result, it is not necessary to read the whole essay if you are looking for quick answers, or only want to read about a particular argument.

Luke 2.1-2 says that "It happened in those days that a decree was issued by Caesar Augustus that a census be taken of all that was inhabited. This census happened to be the first, when Quirinius governed Syria."[1.1] During this census, which we know occurred in 6 A.D. (see below), Jesus is born (2.3-7). Then, eight days later, he is circumcised (2:21), and after the 40th day (2:22; cf. Leviticus 12:2-4) he is publicly presented at the temple in Jerusalem (2.21-38), where two different people publicly proclaim him the messiah (Simon and Anna: 2:25-38), one of whom even continues telling everyone about him in the temple afterward. Then his family returns to Galilee (2.39-40), where Jesus grows up, and his family returns to Jerusalem every year thereafter (2.41) for twelve straight years (2:42).

It is often claimed that Luke has John the Baptist and Jesus born around the same time, but, first, this is not necessarily true and, second, this still would not entail a corroboration of Matthew. The second point is more forceful than the first: namely, that Luke could well be referring to Herod Archelaus, and I think this most likely (see n. 1.1.3 below).

Three months before John is born, Gabriel announces to Mary only that she will conceive (1:31, 36), not that she already has. In fact, Luke never says when Mary conceives. Instead, John appears to have already passed most of his childhood by the time Jesus is born (1:80). Given Jewish law at the time (Mishnah, Abot 5.21), which held that a man becomes subject to religious duties on their thirteenth birthday (which would be John's "day of public appearance to Israel"; we see that day for Jesus in 2:42ff.) and other parallels between Jesus and John (cf. 1:80 and 2:40), it would be reasonable to assume that Luke has in mind that John was nearly twelve when Jesus was born (since "in those days" from vv. 2:1 picks up the "day" of the previous vv. 1:80).[1.1.2] This rescues Luke from charges of chronological error, since Luke reports that John's birth was foretold in a vision "in the days of Herod king of Judaea" (1.5), and if John was born around then, it would be an error to have Jesus born around the same time if Herod the Great were meant, since he was long dead by the time the census occurs--although this may well be Herod Archelaus, not Herod the Great, and in fact if Luke did not mean John to be twelve when Jesus was born, then Luke would most likely have meant Archelaus.[ 1.1.3] Of course, we are not told how much time intervened between the annunciation and John's birth (1.22-24), but if we follow what I just argued could be Luke's meaning, it is notable that he places the birth of John in exactly the same year that Matthew seems to place the birth of Jesus (6 B.C.).

There are two peripheral matters regarding Luke that require brief digression for those who are concerned about them. The first is the date of John the Baptist's ministry, and the second consists of the alleged "errors" of Luke in his description of the census. Each will be addressed here in a separate box, which can be skipped if desired since they aren't essential to the issue of when Jesus was born.

Luke gives us another precise date when he sets the beginning of John's ministry to 28 A.D. (3.1), and this has caused some confusion, though for no good reason. John later says that when Jesus began his own ministry he was "about thirty years old" (3.23), and it is often supposed that this was the same year that John began his ministry. That would make Jesus born around 2 B.C. and maybe as early as 4 B.C. This is so attractive to those who want to reconcile Luke and Matthew that its implausibility is overlooked. Such an interpretation does not solve the many problems created by 2:2 anyway--for it essentially trades a contradiction between Luke and Matthew for a contradiction within Luke. And it is more likely that Luke had in mind the passing of some years between the two inaugurations.

First, it seems unlikely that Jesus should start his ministry in the same year as John: surely John was preaching some years before, preparing the way. It is hard to imagine how he could otherwise have won such widespread and lasting repute (even Josephus sings his praises in Antiquities of the Jews 18.116-19). And we are told Jesus began his own ministry only after John was arrested: Mark 1:14; Matthew 4:12; Luke 3:20; Luke 3:21-2 refers back in time (John contradicts all the others by having Jesus start preaching before John is imprisoned). Josephus allows that this may have happened as late as 32 or 33 A.D. Moreover, all the descriptions of John's ministry imply an extended period, and since we are not told how long John ministered before he baptised Jesus, we cannot assume it happened in the same year. Certainly, Luke 3:21 does not entail a single event of mass baptism (the other narratives of John's ministry also indicate otherwise), but most likely refers to the whole period of the ministry, especially considering the liberal use of the imperfect tense throughout this part of the narrative (3.7, 3.10, 3.14, 3.18), which describes continuous action, and not a single event (Matthew's narrative does this also, cf. 3.1-13). Of course, Luke could always have his chronology wrong, but I believe in the principle of interpretive charity: since he clearly allows an interval of some years, it is unnecessary to assume Luke contradicts himself by dating Jesus' birth first to 6 A.D. and then to several years previous. It is more reasonable to read Luke as dating the beginning of Jesus' ministry between 33 and 36 A.D. (the last years of Pontius Pilate's tenure), which dovetails nicely with Josephus' report that John could have been killed around then.

This also fits the narrative of John the Baptist by Josephus, who reports that a military victory against Herod between 35 and 36 A.D. (Antiquities of the Jews 18.113-19; see also Bowersock's Roman Arabia, pp. 65-6, and Kasher's Jews, Idumaeans, and Ancient Arabs, pp. 177-80) was at the time hailed as God's revenge for executing John, and since people would be calling up the most recent crime of Herod as the cause of his military defeat, John's execution must not have been long before this, certainly some years after 28 A.D.

Some have pounced on Luke's description of the census as being inauthentic and therefore false. There are two problems with such an argument: first, an author who knew Jesus was born during a particular census could still err in describing that census, so such errors would not discredit the entire account (and see [1.1.5]); second, Luke's errors are not that grievous to begin with.

The first "mistake" lies in claiming that the census was of "all that was inhabited," when we know in fact that it was only of Syria and Judaea (and there was no such thing as a universal census at all until many decades after Jesus died, see below). A.N. Sherwin-White has suggested that the fact that there was never in that period a single census of the whole Roman Empire actually confirms rather than refutes the reliability of Luke, since it would betray Luke's or his source's (imperfect) acquaintance with imperial decrees: for it was the standard of the day to preface specific decrees with the general idea behind them, and thus Luke (or his source) could have mistaken a preface for the decree itself, a mistake that one imagines could only be made by someone who had at least glanced at an actual census decree.[1.2] But such an error could simply reflect a common belief among many Jews, due to the usual course of erroneous transmission of popular news, or could have been made, for all we know, by someone a hundred times removed from Luke's immediate source, based on entirely different censuses.

Another possible source of such an error could be the assumption that the first universal census, conducted by Vespasian and Titus in 74 A.D. (which would be fresh in everyone's memory for the remainder of that century), was "typical" when in fact it was not. However, the mistake could also have arisen from Luke himself, for an entirely different reason: by reading Josephus and not realizing that he meant only Judaea. For Luke appears to have drawn on Josephus, and frequently made glaring mistakes when he did so (see my essay Luke and Josephus). On the other hand, there might be no mistake at all: the phrase is pasan tên oikoumenên, "all of the inhabited," where the adjective "inhabited" implies some noun in the feminine, such as "land" or "region," but usually referring to "the whole world." However, this idiom was used not only to refer to the whole Roman Empire, but to regions like "the whole Greek world," and thus may have been meant here as simply the whole Jewish world, or, even more likely, to the whole of Syria (which then included Judaea)--for "Syria" is also a feminine noun.

The second "mistake" lies in supposing that people would be called back to ancestral towns to be counted, rather than be counted in the actual towns they were in. This charge has been formulated a dozen ways, but none of them really carry much force. Though Jesus' family appears to have resided outside Judaea in Nazareth, there could easily be any number of reasons why an ancestral connection with Bethlehem would require them to journey there for a census of Judaea (so much as a tiny plot of ancestral land would be enough, and Judaic law made it unusually difficult to get rid of such properties), though it does seem oddly unnecessary to take a woman on the verge of labor on such a dangerous trip (as all journeys were in such regions). We do know that censuses could have such requirements for travel, not only from papyri [1.3] but also from common sense: it is a well known fact that even Roman citizens had to enroll in one of several tribes to be counted, and getting provincials to organize according to locally-established tribal associations would be practical (see also [8] Mason is wrong in my essay Luke and Josephus; and also [1.3.5]). Finally, even if Luke were making this up, he would sooner make something up that sounded plausible: in other words, such procedures were probably followed in at least one census within the author's memory, and we have no way to disprove the use of such a practice in previous provincial assessments.[1.4] Alternatively, Luke may have deliberately added both of these features to the story for apologetic reasons (see [1.1.5]).

The above two defenses of Luke do not mean that Luke is correct. I think using a census in the story to explain both of Jesus' reputed ancestral homes (the village of Nazareth and the town the messiah had to come from, Bethlehem) is rather convenient and looks more clever than genuine. And Luke's probable use of Josephus suggests a deliberate attempt to paint a veneer of genuine history around an otherwise questionable hagiography. For Luke's accuracy is only provable in details such as these, which were already painstakingly researched by other men and published in Luke's own day. When we get to the question of whether Jesus was actually born during this census or any other, or even to the question of how Luke could possibly know this (and Matthew apparently not know it), we are talking about an entirely different problem. Nevertheless, though Matthew's account looks and smells like a fantastical legend (see below), I do not see Luke's account as historically impossible, as some have tried to argue. To the contrary, I think Luke strained to force his story to seem more plausible than it already was when it got to him. But if one of the two authors must be correct, then Matthew is far more likely the one who has it wrong.

Matthew 2.1 begins by reporting that "after Jesus was born in Bethlehem of Judaea in days of Herod the king" [2.1] the magi reported to Herod that a King of the Jews had been born, stirring Herod and "all Jerusalem" to be troubled (2.3). Herod then seeks to kill the baby by killing all the children in Bethlehem and surrounding areas who would have been born around the right time, i.e. two years earlier (cf. 2.7, 2.16). This would constitute the murder of 10% of the population of that territory [2.2], possibly several hundred children. This is an unimaginable atrocity which probably never happened. A clear and reprehensible crime, it would have almost certainly started a war, and at least been mentioned among the many evil deeds of Herod catalogued by Josephus. Moreover, the myth of the evil tyrant trying to keep his reign by seeking to kill one or more babies (often as a result of an oracle of an usurper's birth), but being thwarted somehow, was a common legendary motif in the period, from Oedipus and Cypselos of Corinth, to Krishna, Moses, Sargon, Cyrus, even Romulus and Remus, just to name the most famous examples. The author of Matthew also gives us some internal evidence that this is mythical: for if so public an event had really occurred (2:3, 2:4, 2:16), it is curious that no one remembers it later (13:54-56), not even Herod's own royal heir (14:1-2). And Matthew is more prone than any other New Testament author to contrive the fantastical--his account of the crucifixion strains credulity, with its rock-splitting earthquakes and hoardes of undead descending on the city (27:51-3), and his account of the empty tomb is as fabulous as legends can get (28:2-5).[2.3] Nevertheless, if we believe Matthew at least has the date right, if not the real events, he is placing the birth of Jesus around 6 B.C., since Jesus was supposedly about two years old when Herod was still alive, and though we do not know how much time passed between this supposed "mass murder" and Herod's death, the narrative implies that it was not long. We might also imagine Herod to have been playing it safe, and thus the date might be 5 or even (though rather unlikely) 4 B.C., or some time could have passed before his death, and thus the date implied could be many years before 6, though the coincidence of Luke's effectively dating the birth of John to 6 B.C. may hint at a possible source of confusion (if there is any true story at all being preserved in these accounts, which is not very likely).

This discrepancy is not the only one. Matthew contradicts Luke (above) on other points of detail. Luke describes Jesus being presented in the temple to repeated public pronouncements of his status, which would not have escaped Herod's supposedly murderous eye (or memory). Matthew, in contrast, has Herod only finding out roughly two years later, from foreigners. The family of Jesus, according to Matthew, flees to Egypt and stays there until Herod's death (2.15, 2.21-23). In fact they stayed away from Jerusalem, not only for as many as two intervening years or more, but for the entire ten-year reign of Herod's successor Archelaus (2.22) as well. This flatly contradicts Luke's claim that they stayed in Nazareth the whole time, from the very beginning, and went to Jerusalem every year without fail. The two accounts thus contradict each other in the most fundamental way: if Matthew is right, Luke's account fails to make any plausible sense at all, and vice versa. For the one story entails that Jerusalem was dangerous for the child Jesus, while the other entails that it could not have been. One story has the family start in Nazareth, then journey to Bethlehem and back again, while the other story has the family start in Bethlehem, then flee to Egypt, then find a new home in Nazareth only to avoid the wrath of Archelaus. Other details could be reconciled, but it is strange the authors themselves did not do this. For example, one mentions a visit by magi but not shepherds, the other shepherds but not magi. But these can be reconciled only by supposing that each author by some queer chance failed to mention the other's similar detail, though one wonders how they could know so much and so little at the same time. But the other details are inescapably at odds. They are telling different stories.

Josephus writes that when Archelaus, the successor of Herod the Great, was exiled:

Archelaus's country was assigned to Syria for purposes of paying tribute, and Quirinius, a man of highest rank, was sent by Caesar to take a census of things in Syria and to make an account of Archelaus's estate. [3.1]

And:

Quirinius was a man of the Senate, who had held other offices, and after going through them all achieved the highest rank. He had a great reputation for other reasons, too. He arrived in Syria with some others, for he was sent by Caesar as a governor, and to be an assessor of their worth. Coponius, who held the rank of knight, was sent along with him to take total command over the Jews. And Quirinius also went to Judaea, since it became part of Syria, to take a census of their worth and to make an account of the possessions of Archelaus. [3.2]

The primary function of a census in antiquity was to assess not merely property, but the total manpower for the purpose of direct taxation, and that is why it was routine to conduct one upon assuming control of a province, and obviously why Josephus describes it in just such terms. What is clear here is that Quirinius did not take control of Judaea until after the removal of Archelaus, and Archelaus followed Herod (even Matthew 2.22 confirms this). This entails that some years necessarily had to follow between the death of Herod and the arrival of Quirinius, so on this evidence there is a clear contradiction between the Gospels: Luke places the birth after the death of Herod, Matthew before.

How great is the discrepancy? Josephus writes:

In the tenth year of Archelaus's government the leading men in Judaea and Samaria could not endure his cruelty and tyranny and accused him before Caesar...and when Caesar heard this, he went into a rage...and sent Archelaus into exile...to Vienna, and took away his property.[3.3]

So roughly ten years separate the death of Herod and the arrival of Quirinius. When was the census held in Judaea? Josephus says quite unequivocally that:

Quirinius made an account of Archelaus' property and finished conducting the census, which happened in the thirty-seventh year after Caesar's defeat of Antony at Actium. [3.4]

The victory at Actium is universally agreed to have happened in 31 B.C. The evidence for this is truly insurmountable, confirmed in countless histories, and inscriptions, papyri and other physical evidence. So the census occurred in 6 A.D., which is the 37th year (beginning with 31 B.C., since the ancients reckoned inclusively). This is independently confirmed by Cassius Dio and corroborated by Roman coins [3.5]. It also fits what we know about Quirinius from an inscription. (It is notable that Josephus also attributes the rebellion of Judas the Galileean to this census [3.6], a detail which Acts 5:37 confirms [3.7].) This then puts the death of Herod at 4 B.C. reckoning from the ten years of the reign of Archelaus, and fittingly, in his account of Herod's reign, Josephus places his death in that very year. But even if we could fudge the year of Herod's death,[3.8] we cannot escape the fact that some years must separate his death and the census, since we have to account for the reign of Archelaus in between, so a historical contradiction between Matthew and Luke persists.

Some have tried to reconcile Matthew and Luke by inventing a second governorship of Quirinius, placing it in the reign of Herod the Great. However, we have no evidence at all that Quirinius served as governor of Syria twice, much less that he did so when Herod was king of Judaea. Moreover, no one ever governed the same province twice in the whole of Roman history, making the very proposal implausible. Three inscriptions and a coin have been used to imply otherwise, but none of these items contain any of the information claimed by those who want Quirinius to have been twice governor, and they offer no support to the theory. We also know who was governing Syria between 12 and 3 B.C. and therefore Quirinius could not have been governor then (or before, since he was not qualified before the year 12). Also, in section 3 it will be shown that there was never any such thing as a dual governorship, nor could there have been, given the nature of Roman political and social organization, and even if Quirinius had been governor or co-governor of Syria at an earlier date, no census could have been conducted in Judaea while Herod or his successor Archelaus were alive.

Some Christian apologists, following extremely outdated scholarship, have tried to argue (or have even stated as if it were a fact) that Quirinius was actually governor of Syria on two different occasions--the first time, conveniently, while Herod was alive. Therefore, this argument goes, the census Luke is talking about happened in the days of Herod the Great. Unfortunately, this fails to solve the other contradictions between Luke's and Matthew's accounts. It is also both groundless and implausible. Nevertheless, every single piece of evidence we have about Quirinius has been twisted into "evidence" of a second or earlier governorship of Syria, and evidence has even been invented wholesale--once by an innocent mistake, and once by pseudoscientific insanity. This "evidence" consists of three inscriptions and one coin, which I will examine in detail.

But first I will mention the several preliminary reasons why this "theory" is absurd. First, we know that Quintilius Varus, not Sulpicius Quirinius, was governor of Syria from 6 B.C. to beyond Herod's death in 3 B.C. (and possibly longer), and before him Sentius Saturninus held the post from 9 B.C. to 6 B.C., and he took the post immediately after Marcus Titius, who probably held the post since 12 B.C. (since three years was the average length of a governorship).[4.1] There is no room here in which to fit Quirinius. And since we know he first attained the consulship in 12 B.C.,[4.1.5] and only ex-consuls held the governorship of Syria in the time of Augustus, he could not have governed before that year. This means one would have to propose that Jesus was born between 12 and 10 B.C. even for this theory to be remotely possible, but that still would be ad hoc, involving a truly maverick position regarding the chronology of Jesus, presuming an unusually short tenure for Titius, inventing a spot for Quirinius nowhere attested, and still not solving the problem of the census (below). Second, we do not even have any evidence that anyone ever served as governor of the same consular province twice in the whole of Roman history, so it would have been extremely unusual and quite remarkable--so much so that it would be odd that no one mentions it, not even Josephus, or Tacitus who gives us the obituary of Quirinius in Annals 3.48, a prime place to mention such a peculiar accomplishment. It is certainly unheard of.[4.2] Now for the reputed evidence to the contrary.

Some have tried to appeal to a headless (and thus nameless) inscription as proving that Quirinius held the governorship of Syria twice, but the inscription neither says that, nor can it belong to Quirinius. The inscription in question is a fragment of a funeral stone discovered in Tivoli (near Rome) in 1764, and is now displayed (complete with an inventive reconstruction of the missing parts) in the Vatican Museum [5.1]. We know only that it was set up after the death of Augustus in 14 A.D., since it refers to him as "divine." The actual content of the inscription is as follows:

... ... ... ... ... ... ... ... ...
...KING BROUGHT INTO THE POWER OF...
AUGUSTUS AND THE ROMAN PEOPLE AND SENATE...
FOR THIS HONORED WITH TWO VICTORY CELEBRATIONS...
FOR THE SAME THING THE TRIUMPHAL DECORATION...
OBTAINED THE PROCONSULATE OF THE PROVINCE OF ASIA...
AGAIN OF THE DEIFIED AUGUSTUS SYRIA AND PH[OENICIA]...
... ... ... ... ... ... ... ... ...

The most obvious problem with this piece of "evidence" is that it doesn't even mention Quirinius! No one knows who this is. Numerous possible candidates have been proposed and debated, but the notion that it could be Quirinius was only supported by the wishful thinking of a few 18th and 19th century scholars (esp. Sanclemente, Mommsen, and Ramsay). But it is unlikely to be his. We know of no second defeat of a king in the career of Quirinius, though Tacitus writes his obituary in Annals 3.48, where surely such a double honor would have been mentioned, especially since a "victory celebration" was a big deal--involving several festal days of public thanksgiving at the command of the emperor. We also have no evidence that Quirinius governed Asia. Though that isn't improbable, we do know of another man, Lucius Calpurnius Piso, who did govern Asia and who defeated the kings of Thrace twice, and received at least one "victory celebration" for doing so, as well as the Triumphal Decoration, and who may also have governed Syria.[5.2] Though it cannot be proved that this is Piso's epitaph, it is clear that it would sooner belong to him than Quirinius. Thus, to ignore him and choose Quirinius would go against probability. Yet even if we lacked such a candidate as Piso, to declare this an epitaph of Quirinius is still pure speculation. Even more importantly, this inscription does not really say that the governorship of Syria was held twice, only that a second legateship was held, and that the second post happened to be in Syria.[5.3] From what remains of the stone, it seems fairly obvious that the first post was the proconsulate of Asia. This means that even if this is the career of Quirinius, all it proves is that he was once the governor of Syria.

Another inscription, actually mentioning Quirinius, often called the "Aemilius Secundus" (after the name of the man whose epitaph it is), is quite shamelessly abused by numerous people attempting to reconcile Matthew and Luke. The funeral stone of Aemilius Secundus was acquired in Beirut by merchants from Venice sometime before 1674, and it may have been set up there originally.[6.1] It records, among other things, that Secundus was a decorated officer serving under Publius Sulpicius Quirinius when the latter was governor of Syria, some time during the reign of Augustus, and when in this command Secundus helped conduct a census--taking care of assessing a Syrian city (Apamea)--and took out a bandit fortress in the Lebanese mountains. This confirms only that Quirinius was governor of and conducted a census in Syria. That really does no more than confirm that Quirinius conducted a census in 6 A.D. as governor of Syria. But since no date is given, nor any datable details at all, apologists have tried to invent dates that suit them, and then claimed this epitaph as a bogus source of "confirmation." The inscription, found broken in two but reassembled, reads as follows (small letters in the sketch represent missing or damaged letters):

QUINTUS AEMILIUS (SON OF QUINTUS) SECUNDUS OF THE PALATINE TRIBE, IN THE SERVICE OF THE DIVINE AUGUSTUS, UNDER
PUBLIUS SULPICIUS QUIRINIUS THE LEGATE OF CAESAR IN SYRIA, WAS DECORATED WITH [THESE] HONORS: PREFECT OF A COHORT FROM THE FIRST AUGUST LEGION; PREFECT OF THE SECOND FLEET; ALSO CONDUCTED A CENSUS BY ORDER OF QUIRINIUS IN THE APAMENE COMMUNITY OF 117,000 CITIZENS; ALSO, WHEN HE WAS SENT BY QUIRINIUS AGAINST THE ITURAEANS ON MOUNT LEBANON HE CAPTURED THEIR CITADEL; AND BEFORE HE WAS IN THE ARMY AS OFFICER IN CHARGE OF WORKS, HE WAS DELEGATED BY THE TWO CONSULS TO RUN THE TREASURY; AND WHEN HE WAS LIVING IN A COLONY
HE SERVED AS QUAESTOR, AEDILE TWICE, DUUMVIR TWICE, AND PONTIFEX. QUINTUS AEMILIUS SECUNDUS, SON OF QUINTUS,
OF THE PALATINE TRIBE, HAVING PASSED ON, AND AEMILIA CHIA, [HIS] FREEDWOMAN, HAVE BEEN LAID TO REST HERE. THIS MONUMENT NO LONGER BELONGS TO [HIS] HEIRS.

As one can see, the inscription does not say and gives no clues when the census occurred. The latest scholar to examine the issue, Fergus Millar, has good reason to think that this was actually the very same census of 6 A.D. [6.2], and this is the most logical conclusion: Josephus says that Quirinius was conducting a census in Syria at the time Judaea was annexed, and the Apamea that this inscription refers to (see below) was part of the Roman province of Syria. Nevertheless, I will give one typical example of how apologists abuse this evidence, and how the story gets distorted in transmission by careless authors.

David Allen Rivera reports:

Dr. Jerry Vardaman, an archaeologist at the Cobb Institute of Archaeology, at the Mississippi State University, deciphered a stone tablet known as the "Amelius Secundus," which had been discovered in Beirut, Lebanon, over 300 years ago, and is now at the Venice Museum. Written in 10 BC, Vardaman says it refers to a census ordered by Quirinius, the governor of Syria, which took place before that, and seemed to be the one referred to by Luke.

Rivera's source for this information is an undated article in an obscure newsletter called The Patriot-News (Harrisburg, PA) written by John Goodrich and entitled " 'Comet Sunday' to Draw Attention to Heavens." It does not say where Vardaman argues any of this, but I subsequently traced the claim to its source (see below). Of course, it is called the Aemilius Secundus, not Amelius Secundus, but this error is Rivera's. Everything else Rivera says is entirely correct, except for the date. We do not in fact know the date of its composition, much less that it was 10 B.C., and this is actually unlikely given all the probabilities surrounding Quirinius examined in this essay. It should also be apparent from an examination of the microletters fiasco (see below) that Vardaman's scholarship is not to be trusted, yet Rivera repeats his claims as if they were matters of established fact. The moral of the story is that readers should be very wary of "facts" touted in defense of Biblical inerrancy.

The only other real material evidence mentioning Quirinius to date is a pair of stones found in two different Muslim villages outside Pisidian Antioch in 1912 and 1913. The stones had been removed from wherever they originally lay and then were reused as wallstones. Both are commemorative inscriptions, originally parts of the bases of statues of a certain amply-named Gaius Julius Caristanius Sergius Fronto Caesiano. Both happen to mention as one of his offices the deputy management of a duumvirate held by Quirinius. Once again, the first inscription mentions no date and itself can only be dated by conjecture to sometime between 11 and 1 B.C. The second inscription offers no clues at all, though it was most likely set up after the first, since it mentions additional posts, apparently gained in the interim. Though even earlier dates are remotely possible, later dates are much less likely. Only a sketch exists of one of them, but a photo survives of the other.[7.1] They read as follows:

The only item here that allows any guess at the date the first statue was erected is the fact that it mentions that this was the first man to set up a statue by public decree (in other words, the city legislature voted to pay for it). Since this presumably would not be long after the city was founded (no more than five or ten years), if we can figure out when Pisidian Antioch was established, we will have some idea of when it was set up, though nothing like an exact date. This is not the most famous Antioch (in Syria), founded in 300 B.C. and one of the largest cities in the world at the turn of the era, but "Antioch near Pisidia," possibly as old but refounded sometime after 25 B.C. under the new name "Colonia Caesareia" (Caesarean Colony), for Roman veterans, definitely in the reign of Augustus, but we actually don't know for sure when. Only after this date would "decemvirs" be issuing public decrees, since these were the officials comprising the city council under a Roman colonial charter. When all things are considered, we can speculate Quirinius' duumvirate was held between 6 and 1 B.C. (see below). But even with other dates, the inscription offers no proof of a second governorship of Syria. First, there is no particular connection between being governor and being the Duumvir of a city. The one does not entail or even imply the other. Second, this city is well outside of the Roman province of Syria, on the border between Lycia-Pamphylia and Galatia (near modern day Egridir lake in Turkey). Indeed, it is even Northwest of Cilicia, on the other side of the Taurus Mountains. This makes any connection between this office and a governorship of Syria impossible. No one would range so far from his province or have any major connection with a city so thoroughly separated from his area of control.

Even so, this has not stopped some Christians from telling tall tales about what these inscriptions prove. For example, one Christian periodical reports to its readers, as if it were a simple fact:

Luke had stated that Quirinius was the Governor of Syria at the time of Jesus' birth, however secular records showed that Saturninus was the governor at that time. An inscription was later found in Antioch which showed that Quirinius indeed was governor of Syria at the time.[7.2]

This short statement doesn't even address the possibility that, if Matthew's date is correct, either Saturninus or Varus could have been governor at the time (as mentioned (above). But what it really gets wrong is the claim that the Antioch inscription proves Quirinius "was governor of Syria at the time." Every single thing in italics here is false. As we've seen, the stones (and there are two, not one) only report that Quirinius was a Duumvir, not a governor, and not in Syria, but well outside that province. And they give hardly any reliable clues as to the date. Only pure speculation can set the date between 9 and 4 B.C., and what little argument could be advanced for a date between 6 and 1 B.C. actually goes to prove that Quirinius was fighting a war in Galatia at the time (see below) and that refutes the possibility that he was governing Syria (see below), so there is in the end no evidence in these stones regarding any governorship at any date.

Yet essentially the same claim regarding these stones, with the addition of a false appearance of precision ("Quirinius was indeed governor of Syria in 7 BC as well"), is made by Doug Raymer in his 1999 online essay "The Accuracy of the Bible." I have traced this particular claim to its source, since it also appears in Kirk R. MacGregor's online essay "Is the New Testament Historically Accurate?" MacGregor at least tells us where he heard this: he cites page 160 of John Elder's book Prophets, Idols, and Diggers: Scientific Proof of Bible History (New York: Bobbs Merrill Co., 1960). Elder's credibility is certainly in question. He reports that the Antioch stone (he, too, only seems aware of one) says that Quirinius was Prefect as well as Duumvir. Obviously, Elder never actually read the stones, for they are about the Prefecture of Caristanius, not Quirinius. He also asserts as if it were a fact that the stone "records his election to the post of honorary duumvir...in recognition of his victory over the Homanadenses" yet I have placed in italics precisely what the stones do not say. Scholars only propose these as possible interpretations, yet Elder seems blithely unaware of the difference. Furthermore, Elder thinks this inscription "proves that Quirinius was in the area as a commander" but he does not seem to understand that it places him well outside the province of Syria, where no governor of that province would have been. Thus, when Elder asserts, again without any qualification, that the date of the inscription "can be fixed as somewhere between 10 and 7 B.C." we know he is not to be trusted. He claims that the names on the inscription set that date, yet does not explain how. In fact, the names fix no date at all (see below). It is clear that Elder did not actually read or study the inscription himself--he must be relying on some other scholar, yet he cites no one.

The moral of this story is: always be suspicious of an unsourced "fact" that goes against the common consensus. From an online archive I came across a likely source for Elder's claims: the hopelessly outdated early 20th century work of G. L. Cheesman.[7.3] Cheesman dates the Duumvirate on the conjunction of two speculations: that the Dummvirate coincided in some way with Quirinius' war against the Homanadenses--a likely possibility, but hardly a known fact--and that the war happened between 10 and 7 B.C.--a very unlikely possibility in light of more recent archaeological evidence (See Homanadensian war), but at any rate nothing like a known fact. This teaches us another lesson: in the arguments of Christian apologists, the speculations and inferences of other scholars suddenly and inexplicably become definite facts. For example, on Franke J. Zollman's online "Dustface Chart" of "Biblical Characters Whose Existence has been Confirmed from Archaeological or Secular Historical Sources" he declares matter-of-factly: "Inscription found in Antioch of Pisidia names Quirinius as legatus." The position of Legatus (a title that implied, but did not entail, being the governor of a consular province) is nowhere mentioned or implied in the Antioch stones.

Arguing for the date of these inscriptions as between 6 and 2 B.C. is a long and tedious task and is included next only for those who want to explore the issue in depth. Others can skip it.

G. L. Cheesman, whose own work is already almost a century obsolete, cites the speculations of other scholars that this Antioch was established either when the province of Galatia was created in 25 B.C. or when Augustus visited the region in 19 B.C., but this is pure guesswork, based on no actual evidence.[7.91] The only physical evidence is an inscription that ensures the colony existed as early as 11 B.C. [7.92] This inscription also implies that very date for its founding, and this is in fact the most probable date given the most reasonable circumstances in which Quirinius would hold a Duumvirate there: as commander of the forces we know he led against the Homanadenses, bandits who plagued the large lake regions in south-central Galatia, very near Pisidian Antioch.

We know of no other occasion in which Quirinius would be in such a remote area. Certainly, no governor of Syria would have any cause to range so far from his province--especially with a massive mountain range in the way. Duumvir could be a position of emergency--something likely to be held during a war nearby--or of honor--something likely to be awarded after a great favor, like leading the local veterans in a successful war against a nearby bandit king. The most likely date for this war, however, is 6-3 B.C. (see next). Therefore, if Caristanius' statue was erected during or shortly after that war, then the city of Antioch could not have been founded more than five or ten years previously (as explained in the main text above), and 11 B.C. fits perfectly in that range. This fit argues for a date of 6-1 B.C. for Quirinius' post of Duumvir. In fact, since we know Quirinius was sent to Armenia to be the caretaker of the Imperial hopeful Gaius around the turn of the era, this would be an ideal circumstance in which he would need a deputy ("prefect") conducting his office of Duumvir in Antioch (if that is what the inscription means), in which case the post would date to 2 B.C. or later.

Unfortunately, the other man named does not much help to focus on a date. We know the Marcus Servilius to whom Caristanius was also a deputy was consul in 3 A.D. (Quirinius was consul in 12 B.C.) and was a close friend of Quirinius, having helped him out in a major trial in 20 A.D., as Tacitus reports (Annals 3.22). His late consulship shows that Servilius was probably much younger than Quirinius, and this fits a later more than an earlier date for his own Duumvirate, if we suppose he held one a year or more after Quirinius--the inscription implies this but does not make it clear. Even so, this is inconclusive, for if he achieved the consulship unusually late in life his political career could have begun as early as 20 B.C., though given normal ages of office he would have started around 14 B.C. If any younger than usual it might not have started until as late as 2 B.C., but this is unlikely since the Duumvirate would then have been among his first offices, which is somewhat unlikely given that it was an important post involving high honor. More likely it would be held later in one's career, so that even if he began in 20 B.C. he would not likely hold such a post for another five or ten years at least, but there are always exceptions. So this leaves us with too broad a window to draw any conclusions from, though it does support a range of 6-1 B.C. somewhat more than earlier dates.

• Cassius Dio is renowned for mentioning numerous minor wars that are otherwise unattested, yet he fails to note this one. However, due to missing pages from a key manuscript, we lack Dio's account of the years 6-2 B.C. Since we expect Dio to mention wars such as this, it is likely he did, and that would put the war in those missing years.[7.93] 
• Varus is noted as having three legions at his disposal in Syria in 4 B.C., yet the usual complement was four, and at least around 2 B.C. it appears that another legion was absent from its usual post in Egypt. There is no known military operation that would explain these missing legions, making the Homanadensian War a likely candidate. The war was likely to have taken several years and more than one legion, since Quirinius won by surrounding the entire region and starving the bandits out of their mountain fortresses.[7.94]

A Roman road around the region, the Via Sabaste, was completed in 6 B.C., as stated on its milestones recovered in recent years. The course of the road takes it in a hemicircle right around the outside of the very region under attack by the Homanadenses, at the same time linking almost every key military colony that would have been associated with the war (including the Asian Apamea, Antioch, and Caesarea), and connecting to available supply routes in three directions along the open plains. This perfectly matches the circumstances of the war and the manner in which Quirinius fought it (above). Since this particular war was won by investment more than open conflict, and supply lines and access to the region from the key colonies had to be set up before the war began, the construction of this road would be the first thing begun by Quirinius and would likely have occupied his men for the first year or two of any actual conflict. This strongly suggests a date for the war of around 6-3 B.C.[7.95]

Dr. Jerry Vardaman, an archaeologist at the Cobb Institute of Archaeology at Mississippi State University, claims to have discovered microscopic letters covering ancient coins and inscriptions conveying all sorts of strange data that he then uses matter-of-factly to assert the wildest chronology I have ever heard for Jesus. He claims these "microletters" confirm that Jesus was born in 12 B.C., Pilate actually governed Judaea between 15 and 26 A.D., Jesus was crucified in 21 and Paul was converted on the road to Damascus in 25 A.D. This is certainly the strangest claim I have ever personally encountered in the entire field of ancient Roman history. His evidence is so incredibly bizarre that the only conclusion one can draw after examining it is that he has gone insane. Certainly, his "evidence" is unaccepted by any other scholar to my knowledge. It has never been presented in any peer reviewed venue [8.1], and was totally unknown to members of the America Numismatic Society until I brought it to their attention, and several experts there concurred with me that it was patently ridiculous.

Nevertheless, his "conclusions" are cited without a single sign of skepticism by Biblical apologist John McRay, who says "Jerry Vardaman has discovered the name of Quirinius on a coin in micrographic letters, placing him as proconsul of Syria and Cilicia from 11 B.C. until after the death of Herod."[8.2] This actual claim has never been published in any form, but I will address a related published claim by Vardaman, and some background is necessary. I will devote some space to this since, as far as I know, I am the only one who has taken the trouble to debunk this obscure and bizarre claim.

What Vardaman means by "micrographic letters" (he usually calls them "microletters") are tiny letters so small that they cannot be seen or made without a magnifying glass and could only have been written with some sort of special diamond-tipped inscribers. He finds enormous amounts of this writing on various coins supporting numerous theses of his. Vardaman claims that he and Oxford scholar Nikos Kokkinos discovered microletters on coins in 1984 at the British Museum, but Kokkinos has not published anything on the matter. Nevertheless, Vardaman tells us that some coins "are literally covered with microletters...through the Hellenistic and Roman periods"[8.3] and that "whatever their original purpose(s), the use of microletters was spread over so many civilizations for so many centuries that their presence cannot be denied or ignored."[8.4] Such fanatical assertions for an extremely radical and controversial theory that only he advocates, and that has not been proven to the satisfaction of anyone else in the academic community, gives the impression of a serious loss of objectivity. Supporting this conclusion is the fact that he cites one authority in support of his thesis that does not in fact support him, yet he does not qualify this citation for his readers but acts as if this makes his theory mainstream.[8.5]

Apart from the fact that it is totally unattested as a practice in any ancient source and none of the relevant tools have been recovered or ever heard of as existing in ancient times, and it has never been subjected to a professional peer review much less accepted by any expert but Vardaman himself, there are several other reasons to regard this as insanity. First, it is extremely rare to find any specimen of ancient coin that is not heavily worn from use and the passage of literally thousands of years, in which time the loss of surface from oxidation is inevitable and significant. Even if such microscopic lettering were added to these coins as Vardaman says, hardly any of it could have survived or remained legible, yet Vardaman has no trouble finding hundreds of perfectly legible words on every coin he examines. Second, to prove his thesis, Vardaman would at the very least be expected to publish enlarged photographs of the reputed microscopic etchings. Yet he has never done this. Instead, all he offers are his own drawings. Both of these facts are extremely suspicious to say the least. Finally, the sorts of things Vardaman finds are profoundly absurd, and rank right up there with Erich von Däniken's Chariots of the Gods.

Here is a typical example:

Notice that this is merely a drawing, not a photograph, and that he gives no indication of scale.[8.6] He never even properly identifies the coin type, and though he quotes the British Museum catalogue regarding it, he gives no catalogue number or citation, so I am unable to hunt down a photograph of it or to estimate its size. But even if among the largest of coins it would not be more than an inch in diameter. Most coins were much smaller. Yet his drawing (left) has a diameter of 4.75" for a scale of at least 5:1 or more, and his blow-up (right) is a little over three-times that, for at least 15:1. That means that his letters, drawn at around a quarter inch in size, represent marks on the original coin smaller than 1/50th an inch, less than half a millimeter.[8.6.5] It would be nearly impossible to have made these marks, much less hundreds of them, and on numerous coins, at minting or afterward (indeed, even the number of men and hours this would require would be vast beyond reckoning), and it would be entirely impossible for them to have survived the wear of time. Yet Vardaman sees them clear as day.

But this is merely the beginning of the madness. Vardaman's quotation of the coin catalogue establishes this as minted by the city of Damascus in the reign of Tiberius, and the coin itself says "LHKT DAMASKWN" or "328th [year] of the Damascenes," referring to its re-establishment as a Greek city by the first Seleucus, in the last years of the 4th century B.C. However, coins minted in Eastern Greek cities did not use Latin letters or words, they used Greek--one can see even from his drawing that the real letters on this coin are Greek, spelling Greek words--yet almost all of Vardaman's "microletters" for some strange reason appear in Latin. Second, and most humorously, all the Latin letters for "J" appear, as Vardaman reproduces them, as modern J's, yet that letter was not even invented until the Middle Ages! If his J's were genuine, they should be the letter I. This alone makes it clear his claims are bogus.

But in case there is any doubt: Vardaman claims to find in these tiny letters the clear statement that the coin was minted in the first year of king Aretas IV in 16 A.D. But Damascus was not a part of the kingdom of Aretas until after the death of Tiberius in 37 A.D. when it was briefly granted to him by Caligula, so Vardaman uses the microletters as evidence that refutes the accepted history. Yet it is a plain contradiction for the minters to boldly date this coin according to their independent Seleucid heritage, and then microscopically reverse that fact and date it by the reign of a recent king. Yet Vardaman doesn't stop there. The microletters tell him all sorts of new facts about the ancient world, like that the full name of the king was Gaius Julius Aretas, and so on. But even more bizarre still:

The most important references on this coin are to "Jesus of Nazareth." He is mentioned frequently, often in titles and phrases found in the New Testament, for example, "Jesus, King of the Jews," "King," "the Righteous One," and "Messiah." Reference to the first year of his "reign" is repeated often...for example, "Year one of Jesus of Nazareth in Galilee [sic]." [8.7]

The absurdity of all this, officially and microscopically inscribed on every coin by the royal mint of the King of the Nabataeans in 16 A.D., stands without need of comment.

Vardaman also "sees" microscopic letters on inscriptions, even though stone, by its roughness and its exposure to weathering, would be even less likely to preserve such markings, even if they had ever been made. Indeed, stones of the day were not polished, making it literally impossible for microscopic letters to be inscribed on them in any visible way. Yet he finds these tiny letters on the very Lapis Venetus inscription (above) showing "that the text dates to 10 B.C., that the fortress Secundus took on Mount Lebanon was Baitokiki, and that the colony mentioned was Beirut."[8.8] All but the last point would be a valuable addition to our historical knowledge, yet he has never published any papers on these claims. More bizarre still, despite several pages of confused text in his later work on how he arrives at this date, he never even says how he gets the date from the microletters. He merely asserts it over and over again,[8.9] and then appends an unnumbered page with some rough remarks about how his microletters date every office of Secundus by the years from the founding of the Beirut colony in 15 B.C. One could write volumes on the weirdness he finds in his microletters (like the name of the Jewish rebel Theudas on the Lapis Venetus, calling him the king of the Scythians![8.95]). But I think it is clear enough that this is all nonsense on stilts.

Now for the punch line. There is no Quirinius coin. McRay's reference is to an unpublished paper that no doubt comes up with more complete nonsense about Quirinius in the reading of random scratches on some coin or other, twisted into letters by what must be a chronic mental illness. But Vardaman hasn't even published this claim. Instead, almost a decade later, when he did present a lecture on the matter, his paper on the date of Quirinius, though over 20 pages in length, never mentions this coin that apparently McRay was told about. Instead, a date of 12 B.C. is arrived at using nonexistent microletters on an inscription. So we can dismiss this claim of Vardaman's and McRay's without hesitation.

Whether conducted by Quirinius or anyone else, there could not have been a census in Judaea before 6 A.D., since the province had not entered direct Roman control before then. In contrast, we know that such control always entailed a census, because it initiated Roman direct taxation for which the census existed. Since Quirinius is the first Roman magistrate to control the province, we expect a census to occur at that time. So we expect a census in 6 A.D., and not before. This is due to the nature of Roman imperialism. The whole point of a client kingdom, as Judaea was in the time of Herod, was that the kingdom retain its independence while paying a set and agreed annual tribute--many territories received this special status for cooperating with Rome in important wars, or when Rome did not want to trouble itself with running the province directly, and typically these client states surrounded and protected the borders of the Empire, providing a kind of buffer zone against invasions.[9.1]

To conduct a census in contravention of such an alliance would have been a notable event indeed, mentioned in many other places as the peculiar event that it would have been--even if it did not start an outright war, as almost happened when the Romans finally did conduct a census in Judaea, once they were in direct control.[9.2] Why, after all, would Rome want a census of a territory it was not taxing directly? Such a thing was never done at any time in the history of Rome. Horst Braunert's study of the subject "disproves conclusively the notion of a Roman census before the creation of the province" while also demonstrating that a census was "a necessary consequence of the establishment of direct provincial government."[9.3] And as we saw above, Josephus confirms a census at the beginning of Quirinius' reign, just when we would expect it.[9.35]

So not only is a census before the annexation of a Judaean province against all probability and sense, it lacks all evidence of any kind. It is a purely groundless and ad hoc conjecture. Nevertheless, some attempts to "create" evidence for it remain and I will address them here.

Some have tried to argue that the Greek of Luke actually might mean a census "before" the reign of Quirinius rather than the "first" census in his reign. As to this, even Sherwin-White remarks that he has "no space to bother with the more fantastic theories...such as that of W. Heichelheim's (and others') suggestion (Roman Syria, 161) that prôtê in Luke iii.2 means proteron, [which] could only be accepted if supported by a parallel in Luke himself."[10.1] He would no doubt have elaborated if he thought it worthwhile to refute such a "fantastic" conjecture. For in fact this argument is completely disallowed by the rules of Greek grammar. First of all, the basic meaning is clear and unambiguous, so there is no reason even to look for another meaning. The passage says autê apographê prôtê egeneto hêgemoneuontos tês Syrias Kyrêniou, or with interlinear translation, autê(this) apographê(census) prôtê[the] (first) egeneto(happened to be) hêgemoneuontos[while] (governing) tês Syrias(Syria) Kyrêniou[was] (Quirinius). The correct word order, in English, is "this happened to be the first census while Quirinius was governing Syria." This is very straightforward, and all translations render it in such a manner.

It does not matter if Luke meant that he knew of a second census under Quirinius, since we have already shown that if there were one it would have occurred some time after 6 A.D. Nevertheless, the passage almost certainly does not mean this. We have no reason to believe Quirinius served as governor again, or long enough to conduct another census, and the Greek does not require such a reading. The use of the genetive absolute (see below) means one can legitimately put a comma between the main clause and the Quirinius clause (since an absolute construction is by definition grammatically independent): thus, this was the first census ever, which just happened to occur when Qirinius was governor. The fact that Luke refers to the census from the start as the outcome of a decree of Augustus clearly supports this reading: this was the first Augustan census in Judaea since the decree. Another observation is made by Klaus Rosen, who compares Luke's passage with an actual census return from Roman Arabia in 127 A.D. and finds that he gets the order of key features of such a document correct: first the name of the Caesar (Augustus), then the year since the province's creation (first), and then the name of the provincial governor (Quirinius). Luke even uses the same word as the census return does for "governed" (hêgemoneuein), and the real census return also states this in the genitive absolute exactly as Luke does.[10.2] This would seem an unlikely coincidence, making it reasonable that Luke is dating the census the way he knows censuses are dated. The only fault with Rosen's thesis (apart from the fact that Luke's passage lacks a lot of other typical features of a census return, e.g. the year of the emperor) is that he assumes the prôtê refers to a year since every province begins with a census. Instead of adopting such an assumption, it is simply more reasonable to take the language at its plain meaning: the first Augustan census, which happened under Quirinius.

But even if one wanted to render it differently, the basic rules of Greek ensure that there is absolutely no way this can mean "before" Quirinius in this construction. What is usually argued is that prôtê can sometimes mean "before," even though it is actually the superlative of "before" (proteros), just as "most" is the superlative of "more." Of course, if "before" were really meant, Luke would have used the correct adjective (in this case, proterê), as Sherwin-White implies, since we have no precedent in Luke for such a diversion of style. But there is a deeper issue involved. The word prôtê can only be rendered as "before" in English when "first" would have the same meaning--in other words, the context must require such a meaning. For in reality the word never really means "before" in Greek. It always means "first," but sometimes in English (just as in Greek) the words "first" and "before" are interchangeable, when "before" means the same thing as "first." For example, "in the first books" can mean the same thing as "in the earlier books" (Aristotle, Physics 263.a.11). Likewise, "the earth came first in relation to the sea" can mean the same thing as "the earth came before the sea" (Heraclitus 31).[10.3]

Nevertheless, what is usually offered in support of a "reinterpretation" of the word is the fact that when prôtos can be rendered "before" it is followed by a noun in the genitive (the genitive of comparison), and in this passage the entire clause hêgemoneuontos tês Syrias Kyrêniou is in the genitive. But this does not work grammatically. The word hêgemoneuontos is not a noun, but a present participle (e.g. "jogging," "saying," "filing," hence "ruling") in the genitive case with a subject (Kyrêniou) also in the genitive. Whenever we see that we know that it is something called a "genitive absolute" construction, and thus it does not make sense to regard it as a genitive connected to the "census" clause. In fact, that is ruled out immediately by the fact that the verb (egeneto) stands between the census clause and the ruling clause--in order for the ruling clause to be in comparison with the census clause, it would have to immediately follow the adjective "first," but since it doesn't, but the entire clause is distinct from the rest of the sentence, it can only be an absolute construction. A genitive absolute does have many possible renderings, e.g. it can mean "while" or "although" or "after" or "because" or "since," but none allow the desired reinterpretation here.[10.4]

John 1:15 (and 1:30) is a case in point: the verb emprosthen is already used (the first "before" found in English translations of the verse) in order to establish the context, and then comes hoti prôton mou ên, "because he was first [in relation] to me." So here we have an example of when prôtos means "before," yet all the grammatical requirements are met for such a meaning, which are not met in Luke 2.2: the genitive here is not a participle with subject, but a lone pronoun (thus in the genitive of comparison); the genitive follows immediately after the adjective; and the previous use of emprosthen establishes the required context. Thus, this is clearly not the same construction as appears in Luke 2.2. Another example is the use of this construction in Acts 16:12, where again the sentence can be rendered "first in relation to" and only then can it be simplified in English to "before." No such license is allowable in Luke 2.2. As a genitive absolute the Quirinius clause cannot have any grammatical connection with prôtê, and "first in relation to the reign of Quirinius" would not produce the meaning "before" anyway.

Ronald Marchant (in "The Census of Quirinius") proposed that the "census of an Apamenian state" mentioned in the Lapis Venetus (above) proves that independent states could be subject to a census.[11.1] But Marchant has his facts wrong. The state in question was not free at the time of the census, but Roman. The city is not precisely named, and we know of four cities named Apamea that were under Roman influence before the 2nd century A.D., but none were free after 12 B.C., and only two were large enough to have the numbers of citizens stated in the inscription. Of these two, the more famous Apamea on the Orontes is the only one in Syria, and it lost its freedom before the 20's B.C. for having sided with Pompey against Caesar (even though it did so under compulsion, having been captured by Pompey in 64 B.C.). The other, Apamea Celaenae (in Asia Minor), lost its freedom sometime in the 2nd century B.C.[11.2] I can only guess that Marchant mistook a reference to "free inhabitants" as a reference to a free state, but such a mistake would betray a profound ignorance of the basic vocabulary of Roman history. Equally so if Marchant thinks that Apamean coins indicate the city was free, for a great many cities subject to Rome were granted the right to mint their own coins.

This is not much of an argument really, but it is a claim that really needs correction since it is so frequently stated, betraying the ignorance of those who state it. Kirk R. MacGregor says in Is the New Testament Historically Accurate? that "archaeological discoveries show that the Romans had a regular enrollment of taxpayers and also held censuses every 14 years. This procedure was supposedly begun under Augustus and the first took place in either 23-22 B.C. or in 9-8 B.C. The latter would be the one to which Luke refers," essentially paraphrasing Elder. John McRay has made a similar claim, stating that "the sequence of known dates for the censuses clearly demonstrates that one was taken in the empire every fourteen years."[12.1] These two men and many other apologists use this claim as part of their argument that a census of Judaea could theoretically have been taken in 8 B.C., fourteen years prior to the census in 6 A.D., perhaps in the very governorship of Saturninus as Tertullian claimed, or in the supposed "earlier" governorship of Quirinius.

Of course, everything covered above already makes this irrelevant with respect to Judaea, and thus of no help in reconciling Luke with Matthew, so there really is no need to debunk it. But it is such a glaring error that it must be corrected. First, all these claims take for granted the reality of an "empire-wide registration" (based on what Luke appears to say, cf. box above), but there never was such a thing until the massive enrollment made by Vespasian and Titus in 74 A.D.[12.15] Thus, since censuses were scattered and never uniform, no "cycle" could ever have been a uniform reality. We know of only two provinces which, owing to their peaceful nature and unusually well-organized infrastructure, were regularly assessed: Sicily and Egypt. Second, the constitutional census cycle for counting Roman citizens was actually five years, and this was actually maintained in Sicily in rare conjunction with a regular census of non-citizens in that province. This was only due to the fact that it had been placed under a special tax system by the kings that ruled the island before the Roman conquest, which the Romans simply continued.[12.2] But regular civil war and the unwieldy size of the empire in the 1st century B.C. resulted in this cycle being disrupted elsewhere. Even after the civil wars were ended, Augustus was only able to complete three of the general censuses in his long reign, which were only of Roman citizens, not provincial inhabitants. These were taken in 28 B.C., 8 B.C., and 14 A.D.[12.3] This flatly refutes any possibility of a fourteen year cycle for these censuses. One comes twenty years after another, then twenty two years after that.

As far as provincial censuses go, we have our second best information from Gaul. Censuses under Augustus were performed there in 27 B.C., 12 B.C., and 14 A.D. (this last was completed only two years later due to local unrest). None of these fits a fourteen-year cycle. Other provinces also fit no pattern. For instance, we also know that a census was taken in Cyrenaica (North Africa) in 6 B.C.[12.4] Our best information comes from Egypt, since from that province alone we actually have countless papyrus census returns. Egyptian administration was unique, for like that of Sicily, it was simply the system employed by its previous ruler (Queen Cleopatra), which the Romans found convenient to continue. In Egypt alone there was a fourteen-year cycle, and this was the direct consequence of a particular capitation tax unique to Egypt in which everyone paid an annual tax after reaching the age of fourteen. This tax may have existed in Syria, but alongside a different capitation tax on women that began at age 12 (Ulpian, via Digest 50.15.3), which would have entailed a 12-year census cycle, or less, if any--but we are unsure when these taxes began, whether they ever applied to Judaea, or whether Syria was as well-organized as Egypt in the first place. No matter how you look at it, a fourteen-year cycle would not apply to any census in which Quirinius was involved. That a census in year 6 matches the Egyptian cycle could well be a coincidence, or the result of a special reorganization of all the Eastern administrations in that year, but it does not entail that the cycle was observed in Syria or Judaea either before or after that year.[12.5] It could have been, or any other cycle, or no consistent cycle at all.

Another proposal is that hêgemoneuontos tês Syrias might mean simply "holding a command in Syria" and since Quirinius is known to have fought a war in Asia Minor between 6 B.C. and 1 B.C. (see above), perhaps Luke means to refer to the time when Quirinius was fighting this war, and not actually "legate of Syria." This doesn't actually solve any of the problems already discussed so far--no census of Judaea could have been held before 6 A.D. But the argument is not even reasonable to begin with. First, it makes no sense to date an event in Judaea by referring to a special command in a war in Asia. Why not simply name the actual legate in Syria? There is no reason to pass over the most obvious man and name another who has absolutely nothing to do with Judaea, much less a census there.

Second, just because Quirinius was probably assigned a Syrian legion to fight bandits on the mountain border between Galatia and Cilicia, it does not follow that he had any kind of command in Syria.[13.1] To the contrary, he was in the province of Galatia, not Syria, and by special command of Augustus. It only makes sense that he was appointed legate of Galatia for this war, for otherwise the actual legate of Galatia would have been fighting it. A Syrian legate would have no business fighting a war in someone else's province, especially in a territory that would leave him cut off from his own province by a large mountain range: for the Homanadenses were active in the mountain-lake valley in Galatia, boxed in by the mountains of Pisidia, Lycaonia, and Isauria--the valley surrounding Egridir lake, Turkey, on a modern map. Every expert familiar with the facts agrees that "only an army coming from the north could subjugate mountain tribes" in that region, in other words an army led from the province of Galatia, not Syria.[13.2] So it would be quite nonsensical of Luke to refer to Quirinius' command and probable governorship in the province of Galatia as "holding a command in Syria," all the more so since "being a ruler of Syria" is what the phrase actually means anyway (since "Syria" appears in the genetive, not dative case).

To get around the problem that all the governors of Syria between 12 and 3 B.C. are already known, some have tried to argue that Quirinius was actually holding a dual governorship with one of those other known governors. This is truly a desperate ad hoc argument, for there was never any such thing as a dual governorship under the Roman Empire, and it would be very strange if there were.[14.1] It also does not solve the problem of the census--for even if Luke was referring to an earlier date, there could not have been a census of Judaea then, as shown above. Nevertheless, it is argued that Roman provincial commands were ambiguous and thus could be held by multiple persons at the same time, though this is argued with no evidence whatsoever, and it is flatly contradicted by the evidence we do have.

I have seen only two examples offered as evidence to the contrary, but they do not make the case. The first comes from Josephus, where the casual remark "there was a hearing before Saturninus and Volumnius, who were in charge of Syria"[14.2] is taken to imply that there were two governors in Syria at the same time. But when we read Josephus' account of another hearing within a year or two of that one, we are given more specific information: it was held before "Saturninus and the senior colleagues of Pedanius, among whom was Volumnius the procurator."[14.3] The procuratorship was a post held by men of the equestrian class (sometimes even ex-slaves), who were ineligible for the position of governor, and subordinate to him (the relationship is almost identical to that between Quirinius and Coponius, as described by Josephus above). So Volumnius was not and could not have been a governor of Syria, much less co-governor. The second "example" makes exactly the same error.[14.4] Could Luke mean that Quirinius was prefect of Syria under a superior official? Besides the fact that it is illogical to name the second in command rather than the actual governor, Quirinius had been a Senator of consular rank since 12 B.C. and thus could never have been a prefect, who had to be of equestrian rank. It simply makes the most sense to read Luke as saying just what he says, rather than trying to lap on layers of undemonstrable and implausible hypotheses. And the likely fact that Luke is borrowing from Josephus further undermines all such attempts at a solution (cf. Luke and Josephus).

Some are tempted to propose the notion that Luke made a mistake: that he really meant Publius Quinctilius Varus rather than Publius Sulpicius Quirinius. Of course, this would entail that Luke is wrong, and thus would admit that the text as we have it (which reads "Quirinius") contradicts Matthew. And it fails to solve the census problem anyway (as discussed above). But it is also not a plausible conjecture, although it is a very old one: around the turn of the 3rd century, it is believed that Tertullian claimed the Lukan census occurred during the tenure of Gaius Sentius Saturninus (who was governor of Syria from 9 to 6 B.C.).[15.1] Of course, Tertullian is not very reliable,[15.2] and makes this claim in the context of antiheretical rhetoric, so he is not to be trusted as an authority on this point. Moreover, this would have been an easy mistake to make: the governor relieved by Quirinius was Volusius Saturninus, who governed Syria from 4 to 5 A.D. But in fact Tertullian does not link Sentius Saturninus with the census in Luke 2:1, as is commonly supposed by those who ignore the context of this passage. Rather, he says censuses (plural, not singular) prove that Jesus had brothers, in defence of Luke 8:19-21. Since Tertullian believed Jesus was the first born, just as Luke says he was, there could not be any record of his brothers in the census of the nativity. Therefore, Tertullian could not possibly have been thinking of the census during which Jesus was born. So he may well mean another Sentius Saturninus (an ancestor of the other), who was governor of Syria in A.D. 19-21, a plausible time before which Jesus' siblings would have been born. For the sentence sed et census constat actos sub Augusto nunc in Iudaea per Sentium Saturninum, apud quos genus eius inquirere potuissent, can be translated "But it is also well known that censuses were conducted under the Augustus in that time in Judaea by Sentius Saturninus, consulting which they can investigate his family."[15.3]

But it is not very likely that Luke made either mistake, for a number of reasons, including the possibility that Luke borrowed his information from Josephus (cf. Luke and Josephus). It is not likely that Luke was mistaking the men, rather than making a mistake in writing the name: Quirinius is a cognomen, but Quinctilius is a nomen. In Roman nomenclature there were three names: praenomen (a "first name" or "casual name"), like Publius (the fact that the two men have the same first name cannot explain the mistake--that would be like confusing Kevin Klein with Kevin Sorbo); nomen ("family name," our "last name"), like Quinctilius or Sulpicius; and cognomen (originally a "nick name," but in this period it is a broader clan name), like Varus or Quirinius. It would be odd for an author to confuse one man's cognomen for another man's nomen. This is all the more so since Varus was a notoriously famous general, the only one to have lost legions in the early Empire (his three legions were destroyed by Germans in 9 A.D.), a disaster echoing ominously across the Roman world and credited with scaring Augustus into a non-expansionist policy, and leading to the universal glory of Germanicus, beloved all over the empire, who recovered the lost legionary standards in the reign of Tiberius. Why someone like Luke would mistake a famous man for a relatively unknown one is hard to fathom. The reverse would have made more sense. Finally, a mere transcriptional "error" would not likely produce "by chance" the correct name of an actual governor under whom a census was known to be made. But for those die hards who want to argue that anyway, see next. Others can skip it.

Even if we suppose a remarkably coincidental transcription error, such a mistake is palaeographically unlikely. In Greek books of the time, which were written in uncial (all-capitals), Quirinius is written KURHNIOS and Quinctilius is written KUNTHLIOS. In order for a mistake to be made by Luke, or by a scribe very early in the tradition (since all manuscripts agree completely on the spelling), RHN would have to be seen in NTHL, which is all but impossible. Although it is common for LI to be mistaken as N, if this mistake were made here then the name would have ended up KURHNOS rather than KURHNIOS, since the I would be lost. But this letter is there, requiring a second mistake. And this would not be enough. Four more mistakes would have to be made as well in seeing the single letter R in the letters NT, which even a modern English reader can see is quite implausible: it requires missing two entire vertical strokes and then an entire grave, and then reading a T as a completed R. And the notion, actually said to me by someone desperate to defend this argument, that the N and T might overlap, and thus "look like" a R is silly even in principle, and has no precedent in any known writing--the letters never overlapped in written Greek. Perhaps these theorists want to suppose Luke was half blind? Since he spells every other proper name correctly, even that disperate thesis fails to make sense.

Due to popular and foreign usage the -c- tended to be dropped from Quinctilius, and if Luke had read the Greek or Latin of the proper spelling, then his mistake would be even more implausible, since KUNKTHLIOS is even farther from KURHNIOS. In Latin inscriptions these names would have appeared as QUINTILIUS (or QUINCTILIUS) and QUIRINIUS respectively. A mistake from the Latin would require the unbelievable reading of NTIL as RIN. The way these letters are always composed in stone or print would make a mistake of NT for R very unlikely: the grave stroke of the R is smaller and lower than it is on the N; likewise, the horizontal stroke on the T never crosses to the left of an N even when the letters are merged; and how the remaining right side of this stroke could be ignored is difficult to imagine. Moreover, the second vertical of the N is straight, so mistaking that as curved, as is the case in the loop of an R, is most unlikely. Finally, a mistake of L for N is well nigh impossible, as is self-evident.

One might try to argue that the census was actually not Roman at all, but conducted by Herod. Of course, this is prima facie implausible, for it is most strange that Luke would not simply say this, but instead associate it with a Roman governor and Imperial decree: the plain language and obvious context leaves no reasonable interpretation but that Luke meant a Roman census (and again, the possibility that Luke is drawing on Josephus would support this). Even if some other nations held their own censuses, the Jews held a negative attitude toward taking a census in peacetime.[16.1] Had Herod conducted such a census on his own initiative, it would have been a truly remarkable event, and could not have escaped mention by historians such as Josephus. Indeed, the only case we even know of a client king trying to conduct a Roman-style census actually ended in just such a disaster: the people of Cappadocia were so hostile to the idea when their king Archelaus tried it in A.D. 36 that they fled to the mountains in open rebellion. Roman legions had to be sent to put down the revolt (Tacitus, Annals 6.4). And whereas Cappadocia was then already under more direct, punitive Roman control (Annals 2.42), Herod the Great enjoyed the greatest favor and freedom of any client king ever under Roman influence and thus any Roman attempt to "force" Herod to run a census would have been inexplicable and unprecedented.

Nonetheless, Brook Pearson argues for this interpretation and his arguments will be addressed here.[16.2] First, Pearson traps himself in a false dichotomy: arguing that Luke can only be seen as "historically accurate" if Quirinius governed Syria in the reign of Herod or the census took place before Quirinius, he then argues for the latter. But he ignores the easiest solution of all: that Luke is right and Matthew is wrong. Of course, it is just as likely both are wrong, but if one's goal is to defend Luke, one need only reject the historical accuracy of Matthew. But having set up the goal of defending the only thesis he hasn't seen refuted, Pearson is left to try and build a case on poorly-researched conjectures. His relevant points are:

• Pearson claims that "Josephus records a great deal of indirect evidence that a careful and detailed system of census and taxation existed under Herod" (p. 265). But he then presents evidence only of the latter: a careful and detailed system of taxation. He assumes that taxation entailed a census (p. 266). But it did not. Only capitation and corvee taxes (taxes paid per person) required a census. Most taxes had no need of census returns, which were very expensive to administer: land taxes were based on records of property ownership, which were maintained regularly throughout the year; tariffs and other taxes on transportation or sale were levied on the spot; rents of royal land or animals were collected by contractual agreement peculiar to every case; fixed tribute assessed on townships and metropoleis was due in full no matter who had to pay it, and the general point of such taxation was to leave the central power without the expense of having to worry how they came up with it; taxes on produce were based on annual outcomes or predictions, and even when the productivity of land was based on scheduled assessments, this had nothing to do with counting people (and so would not require Joseph to travel).[16.3] What Pearson needs to show is evidence that any sort of capitation or corvee tax was ever levied by Herod, but he doesn't give a single piece of even indirect evidence of this. Pearson fails to recognize that this is why Egypt had a 14-year census cycle: the only reason a census was taken there at all was that a tax levied on each person began when boys reached the age of 14 (cf. How Often Was the Census Held?, above). There is no evidence of any such cycle or capitation tax in Judaea in the time of Herod, and as noted above it is particularly improbable for a Jewish king, and even more improbable that Josephus should never once mention or allude to such a thing despite covering in detail every year of Herod's thirty-seven-year reign.
• Pearson cites Heichelheim on the account books of Herod, claiming they "could not have been possible" without a census (p. 266).[16.4] Of course, Heichelheim's work is rather outdated, and plainly false from what we know of ancient taxation: as just noted, most taxes required no census. And in fact, the reference in question is only to records of Herod's own estates and the annual income of the whole kingdom, which affords no evidence at all of any capitation or corvee taxes (or even land productivity assessments for that matter), and thus provides no evidence of census-taking.[16.5] In fact, one passage Pearson cites implies the opposite: Herod's kingdom was taxed by assigning fixed annual tribute to each territory, and thus not based on a census. This is clear when Augustus reduces by a quarter the annual tax owed by the whole "Samaritan territory," rather than, for instance, reducing by a quarter the tax each Samaritan had to pay. In fact, Josephus uses here the word for tribute, which was not capitation or corvee tax but a fixed annual amount levied on cities or groups of cities, who then had to figure out on their own how to come up with it.[16.6]
• Pearson cites evidence that there were 'town scribes' (kômogrammateis) under Herod and therefore census records and therefore a census (p. 271). It never occurs to him to investigate what else such scribes did: for in fact, government-employed scribes kept property records and records of contracts, records pertaining to market taxes and tariffs, and so on, and thus had plenty to do and plenty reason to be in every town. Their existence does not in the least entail a census (indeed, one wonders just what Pearson thinks they did in the years of down-time between censuses). Moreover, not all town scribes were government employed: many were private businessmen who read and wrote documents for a fee, hired by the great majority of people, who were illiterate. Thus, once again, non-evidence is blown way out of proportion and turned into a phantom 'proof'.
• Pearson tried to dismiss the obvious objection, that Luke refers the census to an Augustan decree and the governorship of Quirinius, by insisting that ancient people dated events by reference to the most memorable event near it, and thus Luke meant to refer to an earlier census by reference to a later 'more memorable' one (pp. 277-8). Here all evidence contradicts Pearson, and it is not surprising that he does not give a single analogous example to demonstrate his generalization. Not only would people not have any good idea eighty or so years later when 'the census of Quirinius' occurred without reference to some standard chronology, but it was routine to date all events by regnal years of important persons (or, in some places, from the foundation of important cities): coins, handled by everyone, routinely did this, and it would be second nature to understand such a dating scheme. And in fact this is just what Luke does in 3:1. If Luke is comfortable dating events by a precise and normal means there, why would he use an imprecise and abnormal means in 2:1-2? Thus Pearson's reasoning is entirely shot through. To the contrary, Luke associates the census not with a private action of Herod, but with a decree of Augustus: a Roman action, not a Herodian one. And Luke says it happened when Quirinius was governor of Syria, not before (as has been amply shown above). Indeed, if we follow Rosen, the passage may be taken to mean the first year of Quirinius' reign, making 2:2 just as precise as 3:1.

The bottom line is that there is no evidence of a Herodian census, and no reason to believe there ever was one. And even if there were, there is no way Luke's reference could be to such a census, and thus Pearson's argument is baseless.

For the sake of completeness I will address an argument some Christian apologists advocate out of desperation to preserve Biblical infallibility, drawing on a particular work by Jack Finegan (Handbook of Biblical Chronology, rev. ed., 1998). This is the assertion that in fact Herod the Great was still living in 2 B.C., and since we do not know who was governing Syria then, it was "obviously" Quirinius. Besides having no evidence whatsoever for either fact (and thus it is an entirely ad hoc theory), the evidence we do have stands against such a thesis. For example, Josephus says, point blank, that Varus, not Quintilian, was governor when Herod died (BJ 1.9-10). But the argument becomes the most convoluted and obscure we've yet seen, demanding a rather lengthy treatment.

First, Herod simply could not have been alive in 2 B.C. (see 3.8). Josephus' principal source for the reign of Herod in AJ books 14-17 (and presumably for the parallel material in the earlier BJ) is the Histories of Nicolaus of Damascus, a close friend of Herod, who in turn relied on first-hand knowledge and Herod's own Memoires.[17.1] In fact, we know Josephus consulted Herod's Memoires directly, and "others" (tois allois) who wrote about Herod's reign (AJ 15.174). Thus, to propose that he erred in dating the king's death by a full two years (actually three, as Finegan places his death in 1 B.C.) is incredible. Josephus says in AJ 17.191 and BJ 1.665 that Herod died thirty-seven years after he was proclaimed king by Rome (40 B.C., a date confirmed by Appian, BC 5.75; and Josephus agrees, with a very precise date: AJ 14.389, so there is no room to move here), and thirty four years after he assumed the crown (37 B.C., as Josephus himself says: AJ 14.487), and since Josephus accurately proceeds through the years of his reign, including several that have independent corroboration (such as "the seventeenth year" of Herod's reign, securely placed by Josephus in 20 B.C., see n. 17.4), it is absurd to suggest he made any mistake greater than a single year.

Finally, we cannot trust the reported coincidence of a lunar eclipse near to Herod's death (AJ 17.167). Only a partial eclipse is astronomically confirmed for March 13, 4 B.C., which makes this an unlikely candidate, and it is unclear how much time followed the event and his actual death anyway. But that kind of claim was commonly made for great events (in this case a notorious murder) and thus is often not genuine, as I explain in my essay on Thallus. Even if accepted, the only total eclipse in this period fell on 23 March 5 B.C., which would allow his death to fall in 4 B.C., and, in fact, all the events supposed to happen in the interim more easily fit this than the partial eclipse of 4 B.C. Of course, Finegan latches onto a total eclipse in 1 B.C. for his theory, but even to use this he is forced to go against evidence in Jewish literature for the actual day of Herod's death (§ 506) which preceded that eclipse.

Besides this hand waving, Finegan's case is built largely in § 228, where he attempts to re-date Herod's coronation, against all evidence, to 35 B.C. He notes that Josephus reports Jerusalem was taken in the 185th Olympiad (AJ 14.487), which ended in 36 B.C., but this Olympiad also includes 37 B.C., and as Josephus also gives the precise consular year he can only mean that year, disallowing Finegan's hypothesis from the start. But Finegan then notes that Josephus says this was "after twenty seven years" from when Jerusalem was taken by Pompey in 63 B.C., which would refer to 36 B.C. Of course, he fails to admit that this could just as well mean Jerusalem was taken in 64 B.C., or that Josephus erred in his math, or that a scribe mistook eikosiex for eikosiepta, and since Josephus otherwise names the year exactly, the latter two errors would be more likely. But in actual fact, Josephus always counts the short portions of a Roman calendar year as complete years, as many scholars of Josephus have noted, and thus he reckons inclusively, so that "after twenty seven years" actually points to 37 B.C., not 36 as Finegan thinks.

This is obvious even on Finegan's own calculations: consular years began January 1 (§142, §172), but Jewish years began in March (§165, §513), and the Olympiads began in July (§185), and all three used calendars assigning different days to the months, so they often fell out of alignment even with our Julian reckoning (all the more so before the reforms of Julius Caesar in the 40's B.C.). Josephus uses all three schemes simultaneously, thus errors (or ambiguities) of some months are to be expected in any date he gives. Josephus says Pompey took Jerusalem in the third month (AJ 14.66) by Jewish reckoning, in the year 63 B.C. by Roman reckoning, and Herod took the city on the same calendar day in the third month by Jewish reckoning, in the year 37 B.C. by Roman reckoning, which on both occasions was a day of fasting (AJ 14.487). The only known fast in the third month (the month of Sivan, which crosses May and June) is that of the apostasy of Jeroboam "who made Israel to sin" (Sivan 22 or 23).[17.1.5] It cannot be the Day of Atonement as Finegan presumes (§227), which fell in the seventh month. This means that Herod was crowned after twenty seven Roman consular pairs held office, hence after twenty seven calendar years on the Roman system. This is also twenty seven years on the Olympiadic system, since the year 63 includes two Olympiadic calendar years for any event that happens before July. Finegan also argues that Herod might have dated his regnal years from the following new year (Nisan), but even if he reckoned this way he would sooner date his reign from the previous new year, so that there would be no year zero for any of his official acts as king, and so his coins could begin right away showing year 1 of his reign, without having to wait nine months to start counting his years of rule. Indeed, we have no evidence that any ruler in antiquity employed any other practice.

So, the fact of the matter is, Josephus reckoned Herod's reign as beginning in 40 B.C. with a coronation in 37 B.C. There is no way around this, and thus when he dates Herod's death, he can only mean 4 B.C., since he relates it to both events precisely (and one is confirmed by another extant historian). That Josephus is wrong about something so central to his histories and for which he had such good, eyewitness sources is simply not credible. Finegan knows all scholars agree with this.[17.2] In fact, Finegan knows that all external and circumstantial evidence is against him. For example, it is a fact that all three regnal dates of Herod's successors match a coronation date of 4 B.C. (§ 516). This includes Archelaus, whose dates are also corroborated by Cassius Dio (55.27.6), and Josephus does not have Archelaus declared king until Herod dies (BJ 1.670), but has Archelaus deposed in 6 A.D. after 10 years rule (see above), which also puts Herod's death in 4 B.C. And then there is Antipas, whose dates are confirmed in extant coinage, according to Finegan himself. Finegan tries to suggest against this evidence that all three of these kings were made co-regents with Herod in 4 B.C. until his death in 1 B.C., a claim that is groundless and prima facie absurd. With Antipater, that would make five kings ruling simultaneously! It is inconceivable that Josephus would not mention such a remarkable action. Indeed, the political atmosphere of heated tensions and indecision about who would inherit makes such a massive coregency profoundly unthinkable for Herod--his coregency with Antipater (the only one Josephus mentions) was already such a disaster that Herod had him executed a week before he himself died, and the other three were only assigned their territories by Herod's will and confirmed by Augustus after Herod's death. Josephus is absolutely clear on this. And it is the only logical way things could have happened.

Apart from all this ad hoc assertion, Finegan's only 'case' for his hypothesized mass-coregency is an attempt to redate the reign of just one successor, Philip, according to an obscure textual variant (§ 218). He proposes that in AJ 18.106 "in the twentieth year of Tiberius" should be read as "in the twenty-second year of Tiberius," so that Philip's "thirty-seven year" reign would have begun in 2 B.C. (and thus, so the argument goes, Herod must have died then). The original basis for all this tinkering is the fact that Philip's obituary is indeed placed in Josephus' narrative seemingly around the year 35 or 36. But it is clear that Josephus wrote "twentieth" and not "twenty-second," and analysis shows that Josephus is either wrong about the dates of all the events he places in this year, or else he is compressing many years together, or both. It is therefore most likely that Josephus is correct about when Philip began his reign, just as he is with all the other tetrarchs, and simply misplaced (or loosely placed) his obituary among external Roman events he knew less well.

As evidence of Josephus' confusion about events, Cassius Dio dates the Vitellian parley, which Josephus places before Philip's death, to the reign of Caligula, several years after Philip's death (59.17.5, 59.27.2-3). And it appears that Tacitus may have, too: Vitellius, as a future emperor, is an important person, yet the event is not recorded by Tacitus for the reign of Tiberius, while Tacitus' account of Caligula's reign is lost. Likewise, Tacitus (Annals 6.31) and Cassius Dio (58.26) both date the other Parthian events to 34/35, which Josephus places after 36/37. Thus, while Josephus dates the death of Philip as having happened "about the same time" as all these Parthian affairs (AJ 18.96-105), they did not happen in the same year. Indeed, it appears that the Parthian king Artabanus established his son Arsaces as ruler of Armenia in 33 or 34 A.D., not 36 as Josephus' narrative implies (s.v. "Artabanus" and "Armenia," Oxford Classical Dictionary). Since Josephus clearly did not have a good idea of when the surrounding events actually happened, or else is not discussing a single year at all, he is certainly being too vague to pinpoint an exact year when he says Philip's death happened "around" then. Likewise, right after Philip's obituary, Josephus says "around the same time" Herod and Aretas began to have a falling out, but the narrative of this event begins with 28 A.D., and spans from then all the way to 37 A.D. in a matter of a few paragraphs. Thus, very little can be concluded about the date of Philip's death from where Josephus has placed it in his narrative.

What about that obscure textual variant? Finegan's only source for this claim is a mysterious, unpublished speech given by David Beyer.[17.3] In Finegan's summary, he never identifies any actual manuscripts, and though Beyer names them he does not identify their relationship to other mss. or their known quality or origins. All Finegan (and Beyer) does is "count manuscripts" and argue that older mss. are the most reliable. But neither is true, as any palaeographer knows. We have no way of knowing which of the mss. Beyer counted were copies of other extant mss. (and thus completely irrelevant to the question), and we have no idea whether the mss. he looked at are known to be reliable or unreliable or to what degree or in what ways. Older mss. can sometimes be poorer than new mss., since newer ones can be based on even older but more reliable archetypes (see "On Calvinist Scorn of Textual Criticism" for more about textual analysis), and older ones may stem from especially faulty textual traditions. Moreover, Beyer examined only mss. in the British Museum and the Library of Congress--yet the best mss. are in France and Italy--one of which is the oldest, Codex Ambrosianae F 128, inscribed in the 11th century (the oldest mss. Beyer examined was 12th century); and another is the most reliable: Codex Vaticanus Graecus 984, transcribed in 1354; both confirming a reading of "twentieth," and thus invalidating all his conclusions from the start. Finegan and Beyer seem ignorant of all of these issues. Consequently, we cannot put any trust whatever in what he claims here.

When, instead, we examine all existing critical editions of Josephus, composed by scholars (Niese, Naber, and Thackeray) who themselves looked at the manuscripts, and properly, identifying relationships among them and assessing their reliability, we find a very different story. First of all, little more than a handful of mss. are worth even examining for this passage--yet Beyer is counting dozens (none of which are even among the best), proving that his investigation is completely disregarding the proper criteria of textual analysis. Second, all scholarly editions agree: the word for "twentieth" (eikostô) exists in all extant Greek mss. worth considering. Where does the reading "twenty-second" come from? A single mss. tradition of a Latin translation (which reads vicesimo secundo). Beyer's case completely falls apart here. The Latin translations of Josephus are notoriously inferior, and are never held to be more accurate than extant Greek mss., much less all of them. Indeed, this is well proven here: whereas the Latin has 22 for the year of Tiberius, it also has 32, or even in some editions 35, as the year of Philip, not the 37 that Finegan's argument requires. Thus, clearly the Latin translator has botched all the numbers in this passage. Any mss. that Beyer examined were no doubt either from these inferior Latin mss., or Greek translations from these Latin mss. Therefore, there is no basis whatever for adopting "twenty second" as the correct reading. Philp was crowned in 4 B.C. exactly as Josephus says, and just as all the other tetrarchs were who inherited portions of Herod's kingdom. This means Herod died in 4 B.C., exactly as Josephus claims.

Even allowing such an inconceivable error on the part of Josephus, this whole theory runs afoul of the problem that Quirinius could not have been governor of Syria twice (see Section II above) nor could there have been a census in Judaea when Herod was still alive (see Section III above). But Finegan employs two ad hoc maneuvers in an attempt to bypass these facts:

(1) Finegan's response to the first conundrum is that Quirinius was actually prefect or procurator of Syria in 2 B.C. (§ 522), not an actual governor. But that is definitely impossible: those were offices held only by the knights (men of the equestrian class), never by senators, much less senators of the most prestigious consular rank, and Quirinius had been a consular since 12 B.C. This mistake is similar to that made by those who want Quirinius to have been a co-governor. It just isn't possible or logical, and of course has no evidence of any kind in support of it.

(2) Finegan's response to the second conundrum is that Luke was referring to some sort of other 'counting' by Herod the Great himself. This could not be a census (see Was it a Census Conducted by Herod the Great? above). Finegan argues that it was when "the people of Rome" proclaimed Augustus Pater Patriae (§ 525), but he has badly erred here: this is a reference to a vote by Roman citizens, which would have nothing whatever to do with Judaeans. By confusing a vote with an oath-taking, he conjures the false claim that Luke is referring to the registration of oaths of loyalty. Of course, this is already shot down by the fact that Herod was not alive in 2 B.C., as we've seen. And we have no record of such an oath in Judaea in that year or any year near it, despite the fact that Josephus usually records them: the last such oaths commanded by Herod were in 20 B.C.[17.4] and in 8 or 7 B.C.[17.5] Worse, this thesis is inherently implausible: Luke does not use the vocabulary of oath-swearing, nor does he describe such a process. For example, Joseph would not travel to Bethlehem if all he had to do was swear an oath of allegiance--that had to be done where he lived.[17.6] The travel only makes sense in the context of a census, where family land could require his presence (see second part of Luke's Description of the Census above). Likewise, "that all that was inhabited be recorded," using apographô as the verb, repeated again as the noun apographê, can only refer to a census: a register made for taxing. Indeed, the word does not even mean "count," but a detailed record-making, and this term is never used in reference to registering oaths. Rather, some form of eunoeô is the correct word (AJ 17.42; e.g. epi eunoiai, AJ 18.124; enômoton tên eunoian, AJ 15.368). Indeed, typically, oaths were not registered at all: one swore before a magistrate and received a diploma attesting to the fact that you swore, which you could present if anyone challenged the fact, as is shown in detail in the martyrologies of those who refused to swear for Decius in 249 A.D. (and in accounts of Christians avoiding martyrdom by buying forged diplomas). Certainly the burden is on those who claim otherwise to present evidence, and I have never seen any.

[1.1] 2 ¹ egeneto de en tais êmerais ekeinais exêlthen dogma para Kaisaros Augoustou apographesthai pasan tên oikoumenên ² autê apographê prôtê egeneto êgemoneuontos tês Syrias Kyrêniou. I have attempted to render all Biblical translations as literally as possible without spoiling the genuine meaning in Greek.
Jacques Winandy recently pointed out that a connective de after autê is required by the rules of Greek grammar (it fulfills the role of a period in English), and thus its absence here is peculiar, cf. "Le recensement dit de Quirinius (LC 2,2): une interpolation?" Revue Biblique July 1997 (104:3), pp. 373-7. He takes this as evidence that this entire line began as a marginal line note made by a scribe that was later accidentally assimilated into the text (for an example of this very thing, see my discussion of Phlegon in my article on Thallus). But given the evidence of borrowing in Luke from Josephus, I find this unlikely (see my article on Luke and Josephus). It would be more probable that Luke simply made an error or the de was lost in transmission. But in fact, this is simply a Lukan affectation, for he often omits a connecting particle when beginning a sentence with a pronoun (cf. Luke 1:32, 7:27, and esp. 23:51 for a close parallel; Acts 1:14, 4:11, 6:6, and esp. 16:17 for a close parallel), and it is common in Koinê Greek for the nominative pronoun to mark transition. Winandy also regards the absence of an article in this phrase as indicative of an interpolation, but his argument (and suggested re-translation) are misguided: the absence of an article is perfectly acceptable for a predicate adjective construction with a form of "to be" in Koinê Greek. Moreover, the use of the reflexive autos as outos is not unusual for Luke but would be unusual for a scribal note. I must conclude that the phrase is genuine.

[1.1.2] One might suppose that Luke 1:42, Elizabeth's greeting to Mary, implies that Mary is already pregnant, but it does not entail that. She may merely be anticipating the future, as is the case in Deuteronomy 28:4, where the same present participle construction is used in the Septuagint clearly in reference to future generations and not to present conceptions. And this is likely, since at 1:45 she actually uses the future tense in reference to Mary's conception.
Some have noted that this entails a twelve year-long betrothal between Joseph and Mary, but there is nothing incredible about a long betrothal--anyone familiar with societies where marriages are arranged knows that well enough, especially when the would-be husband is not yet financially sound or the woman's parents can't yet come up with an adequate dowry, and as we have reason to presume Jesus' family very poor, we can expect such complications. We are never told when or even if Mary and Joseph marry--they are still unmarried when Jesus is born (2:5)--so the fact that she was celibate would be a given, unless she was raped or fornicated contrary to Jewish law.
Others have questioned this theory because no one else has suggested it. But the reason so few have proposed this is that so few are willing to admit one author is wrong. Instead, everyone almost to a man tries to reconcile Matthew and Luke, which means forcing Luke into a 4 BC date (since Matthew cannot be forced into a 6 AD date). That is a universally misguided approach, but not a surprising one. The only thing we can rightfully presume is that an author will strive to be consistent with himself, not that he will strive to be consistent with another author, especially one of whom he shows no awareness and who otherwise tells an entirely different story. Thus, if there is any way to read Luke as consistent with himself that is correct on the Greek and plausible in the context he describes, logic would lead us to read Luke in that way. To deny this and instead apply the same principle to something Luke did not even write is illegitimate procedure.

[1.1.3] Mark Smith has composed a good article explaining in his own terms why attempts to reconcile Luke and Matthew fail, while concluding with strong support for the accuracy of Luke as against Matthew ("Of Jesus and Quirinius," The Catholic Biblical Quarterly, 62:2; April, 2000; pp. 278-93). He argues that Luke may have meant by "Herod the King" not "Herod the Great" but "Herod the Ethnarch," in other words Archelaus, Herod's successor (ibid. pp. 285-6). There is something to be said for this possibility, and Smith makes a decent case. I originally decided against it because I thought Luke was otherwise very precise with the titles of men in power throughout Luke and Acts (a fact that Smith himself documents), but Luke fails to be precise in naming the office of Quirinius, too, and Archelaus only called himself Herod on his coins (Burnett Roman Provincial Coinage 1992, nos. 4912-17) and the historian Cassius Dio also knows him only as such (55.27), while even Josephus, who otherwise refers to Archelaus as ethnarch, could still call him a king (AJ 18.93), facts that slipped my notice before. I believe Smith may well be right, and thus the date of John's birth comes some time after 4 B.C.

[1.1.5] Robert Smith, in "Caesar's Decree (Luke 2:1-2): Puzzle or Key?" (Currents in Theology and Mission 7:6; Dec. 1980, pp. 343-51), recapitulates a theory proposed by a few other scholars that Luke created (or emphasized, if they are true) these features in order to paint Jesus, via his family, as peaceful and obedient to Caesar, in contrast to rebels and radicals with whom Christianity was often in danger of being compared. Smith shows very clearly Luke's preoccupation with this theme, for Luke alone mentions numerous times census-related rebellions and the theme of disobeying Caesar, while never showing Romans committing an injustice (in all the trials related in Luke-Acts, the Romans come out as superlative judges), and repeatedly emphasizing the unrebellious and obedient nature of Christians, etc. Although I think it is clear that Luke has these apologetic concerns, I don't think he needed to emphasize, much less invent, a universal census or a journey of Joseph and Mary to convey them. Mark Smith and other scholars agree with me (op. cit. 1.1.3, p. 284).

[1.2] Roman Society and Roman Law in the New Testament 1963: p. 168.

[1.3] The famous P. Lond. 904, discussed in F. Kenyon and H. Bell Greek Papyri in the British Museum 3 (1907), p. 125 (with plate 30), and in George Milligan Greek Papyri (1910) pp. 72-3. Rostovtzeff cites other evidence proving that there was an idea of a return to the idia, "place of origin," employed in some censuses conducted in the east, cf. Studien zur Geschichte des römischen Kolonates, pp. 305 ff. Also, cf. Rosen, op. cit., 10.2.

[1.3.5] Mark Smith (op. cit. 1.1.3, pp. 287-91) even proposes a possible "tax dodge" was at work, and otherwise explains how this migration is not as incredible as critics propose. Robin Lane Fox, on p. 31 of The Unauthorized Version: Truth and Fiction in the Bible argues that "Roman censuses cared nothing for remote genealogies, let alone false ones," but this isn't true. Membership in Roman tribes was certainly a fiction, yet essential for census taking. It is possible that tribal fictions were maintained in Judaea for similar purposes, and Romans would see an obvious utility in exploiting them to ease the counting of a highly itinerant population (a great many Jews were herdsmen or pilgrims). Likewise, it is very doubtful that Luke would use a system of census taking that his every reader would know didn't exist--certainly, his apologetic aims would fail if his "explanation" held no water. Therefore, there must have been something to it--for even if fiction, it had to play on some fact or else the lie would be obvious to everyone. At the very least, we can suppose many Jews believed they could trace a lineage to some ancestor in the town of their family land, so as to justify their belonging there and to secure their claim in perpetuity, not to mention the mere glory of tracing one's line to some tribal hero (many Greeks and Romans did just the same). We can suppose that Luke believed (or wanted his readers to believe) that Joseph had family land in Bethlehem, and that this was because it was a portion of David's land, and since Jewish Law required the return of sold land every seven years, it was impossible to ever be dispossessed of it--thus, it might have seemed obvious to every Jew that any family plot cou