The First Year of Herod the Great’s Reign

Share/recommend this article:

Excerpt This article gives some specific rebuttals to the contention that Herod’s reign must be dated from his appointment by the Romans, supposedly in 39 BC... Continue reading

Related Articles
Like this artice?

Our Ministry relies on the generosity of people like you. Every small donation helps us develop and publish great articles.

Please support ABR!

Visa, MasterCard, American Express, Discover & PayPal

A Quick Review

We will begin this article with a brief review of the major points laid out in the previous one (“John 2:12–21 and Herodian Chronology,”, and then go into a detailed analysis of some arguments offered against its conclusion that the beginning of the reign of Herod the Great should be dated to 37 BC, per the reasons laid out by Emil Schürer in A History of the Jewish People in the Time of Jesus Christ.

Josephus Started the Year from Nisan

In the linked article we saw how Josephus’ own words demonstrated that he regarded the month of Nisan as the first month of the Jewish year, hence in Antiquities he followed the ecclesiastical calendar rather than the civil calendar starting in Tishri (emphasis and bracketed comments added; except where noted, the Whiston translation of Antiquities is used throughout this article):

Antiquities 1.3.3 – “But Moses appointed that Nisanshould be the first month…although he preserved the original order of the months [where Nisan, as counted from Rosh Hashanah in Tishri, was taken as the seventh month] as to selling and buying, and other ordinary affairs.”

Antiquities 3.10.5 – “Nisanis the beginning of our year.”

Antiquities 11.4.8 – “the first month; which according to the Macedonians is called Xanthicus; but according to us, Nisan.

It could perhaps be argued that calling Nisan "the first month" is purely a label, and does not entail actually counting years from it. But the fact remains that during the Second Temple period after the Babylonian exile, the Jews themselves, as reflected in both Antiquities 3.10.5 and the Mishnah, regarded Nisan rather than Tishri as "the beginning of our year," and counted their rulers' reigns from it. For our purposes, with a focus on the reign of Herod as a Jewish ruler, that is what matters.

Josephus Used Inclusive Reckoning

In addition it was established, from several examples, that the uncomplicated, plain sense of Antiquities indicates Josephus mainly followed inclusive reckoning. We know that inclusive reckoning was a normal convention in the Bible and in the Roman and Jewish worlds generally; this is not in dispute. The story of Cornelius in Acts 10 was given as a biblical example, while we observed that a straightforward reading of Josephus indicates he followed this convention in the way he handled the time spanned by several pairs of independent, unrelated events: the 27 years from Pompey’s taking of Jerusalem in 63 BC to the start of Herod’s de facto (“in fact”) reign in 37 BC following the siege of Jerusalem, the 107 years from the start of Herod’s reign in 37 BC until the Temple fell to Titus in AD 70, and the seventh year of Herod’s reign matching up with the year of the Battle of Actium on September 2, 31 BC, thus requiring his first year to have been in 37 BC. Since these three examples are obviously discussed in Antiquities —in 14.16.4, 20.10.1, and 15.5.2 respectively, with the last one also addressed in Wars 1.19.3—what Andrew E. Steinmann (“When Did Herod the Great Reign?” Novum Testamentum 51 [2009]: 1–29) states in his note 8 (p. 2) is not borne out by the evidence:

Schürer claimed that Josephus used inclusive reckoning (1896 edition 1.200–201, Vermes and Millar edition, Schürer, History 326–327). If Josephus used inclusive reckoning to arrive at thirty-seven and thirty-four years [the first giving the time from Herod’s Roman appointment to his death, the second from the taking of Jerusalem to his death, Ant. 17.8.1] (i.e., the years 40–4 BCE inclusive totals thirty-seven), such inclusive reckoning is not indicated elsewhere in Josephus (emphasis and bracketed comment added).

The cited instances do indicate Josephus used inclusive reckoning elsewhere. They should predispose us to view the 37 and 34-year time spans the same way. The 34-year figure, commonly dated from the end of the siege, requires one small, but as we shall see, significant clarification: Josephus does not refer simply to the fall of Jerusalem to Herod and Sosius in Antiquities 17.8.1, but to the death of Antigonus: "When he had done these things, he died...having reigned since he had procured Antigonus to be slain thirty four years." It is an important difference.

Josephus Sometimes Gave Time Durations

The few exceptions to inclusive reckoning in Josephus’ writings were seen to reflect elapsed time, or rather, time durations of single events that covered a period of time from start to completion. Such was the case where Herod's 18th year was said to mark the start of the 46 years the Temple was “a-building,” and with the 126 years that had elapsed since the Hasmonean dynasty was first set up until its last king died. Whereas time spans between two different events were reckoned inclusively as expected, the duration of what were essentially single events having a defined start and end point were dealt with differently. They were treated like birthdays; the first example looked at when the Temple became 46 years old, the second at when the Hasmonean dynasty became 126 years old. Since it is self-evident that one would never say a month-old infant was reckoned as being one year of age, exceptions like these to the normal inclusive reckoning approach are to be expected. But in saying this, we must insist that such exceptions to accommodate the duration of an event do not set aside the general principle of inclusive reckoning, nor reckoning specifically from Nisan in the case of the post-exilic (Second Temple) Judean kings.

I initially thought it would be sufficient to press these three simple, positive evidences upon the reader, deeming them adequate in themselves to powerfully call into question the validity of any chronological approaches that began the de facto reign of Herod in 36 rather than 37 BC. But the more I reflected on the matter, the more I realized that it was the very complexity of the arguments put forth by W.E. Filmer (“The Chronology of the Reign of Herod the Great,” Journal of Theological Studies 17.2 [1966], pp. 283–298) and Steinmann that might cause some readers to be swayed by them. Some people are impressed by cleverness, while for others it is far easier to simply accept these scholars’ judgments on authority than to analyze their arguments. Still others may feel a certain loyalty to particular scholars for other reasons, such as Steinmann’s reputation as a conservative scholar with a significant book on biblical chronology to his credit.

But in the end, none of those factors should influence our evaluation of their arguments, only the strength of the logic and data they present to make their cases. Therefore, before turning to consider the death date of Herod, it is necessary to offer some specific rebuttals to the Filmer/Steinmann contention that Herod’s reign must be dated from his appointment by the Romans, supposedly in 39 BC.

The Attacks on the “Schürer Consensus”

The approaches of Filmer and Steinmann involve attacking what Steinmann dubs the “Schürer consensus.” Their arguments primarily involve seeking ways to get around the manifest appearance of non-accession (inclusive) dating of reigns from Nisan seen in Josephus. But when we closely examine the evidence, we see this effort is built on a framework of assumptions, questionable interpretations, and outright accusations of error leveled against Josephus. In what follows we will attempt to demonstrate the existence of this shaky framework, and call for a return to a solid structure based on accepting the fundamental reliability of Josephus as an historian, with his work interpreted in a straightforward manner.

An Insistence on Non-Inclusive Dating

The fundamental way the Filmerians reinterpret Josephus is by insisting that he consistently used non-inclusive dating in Antiquities, with the month of Tishri (September/October) viewed as the first month of the year. In this approach, the year an event occurred in is not included in the counting. This means that in the case of reign lengths of rulers, where this approach is known as accession-year reckoning, counting began with the first of Tishri after the start of a king’s reign, while with other matters actual reckoning (as we count things today) was used, with the first year counted on the first anniversary after the event.

The problem is, the records of the Jews themselves indicate non-inclusive, accession-year, Tishri-based dating was not followed after the return from the Babylonian exile. The Mishnah, the third century AD rabbinic compilation of Jewish oral tradition, clearly supports inclusive, non-accession reckoning from Nisan during the Second Temple era that included Herod’s reign:

On the first of Nissan is the [cut off date for the] New Year regarding [the count of the reigns of the Jewish] kings [which was used to date legal documents. If a king began his reign in Adar even if was only for one day that is considered his first year, and from the first of Nissan is considered his second year…] On the first of Tishrei is the New Year for [the counting of] years [of non-Jewish kings], for the Shemittah and the Yoveil count...” (; brackets with summarized Gemara commentary original, emphasis added).

This tells us there were two main calendars (there were minor ones as well, but they can be ignored for our purposes). One began in Tishri and applied to counting years dealing with civil affairs, specifically in regard to kings of foreign nations, sabbatical (shemittah) years, and jubilee (yoveil) counts; the other began in Nisan and applied to Jewish kings and religious festivals. In emphasizing the primacy of Nisan dating for Josephus and for Herodian chronology in particular, we do not deny that Tishri-based accession-year dating applied to kings’ reigns during the First Temple period, as well as to the reigns of the Babylonian and Persian rulers over the Jews during the Babylonian captivity. But we must insist, on the basis of the Mishnah and the testimony of Josephus, that such reckoning did not apply to Second Temple, post-exilic kings of the Jews like Herod. The evidence clearly indicates that the Jews reckoned Nisan as the first month of their year as far back as the Babylonian exile. We see this unambiguously stated in Esther 3:7 (NASB):

In the first month, which is the month Nisan, in the twelfth year of King Ahasuerus, Pur, that is the lot, was cast before Haman from day to day and from month to month, until the twelfth month, that is the month Adar.

If the first month was Nisan and the twelfth was Adar for Queen Esther, a loyal Jew, the ecclesiastical calendar’s ordering of the months was obviously normative in Jewish minds by that time. There is no clear biblical evidence that the Jews thereafter ever viewed Tishri as the first month of the year for reckoning the reigns of their own rulers.

Preoccupation with Herod’s Roman Appointment

This brings up an important point about Steinmann’s approach to Herod’s reign. He consistently refuses to treat it as beginning with the deposing of Antigonus (Ant. 17.8.1) following the siege of Jerusalem, but insists it began with the Roman appointment. But why must we suppose that, just because the Romans had something to do with his getting the position, their involvement somehow made Herod a non-Jewish, foreign king to whom Tishri dating applied? Antigonus was likewise made king of Judea by the intervention of a foreign power, in his case the Parthians, yet no one tries to represent him as a foreign king. What makes one a foreign king with his reign reckoned from Tishri is ruling over a country other than Judea. This was not the case with Herod. It has to do with a uniform approach to record keeping in affairs of state.

Moreover, Steinmann’s proposal that Herod dated his reign from when the Romans granted him the throne of Judea is beside the point. For our purposes it ultimately does not matter how Herod may have viewed the start of his reign, but how Josephus did and recorded it in his histories. For argument’s sake, the strongest indicator that Herod may have personally placed priority on his Roman appointment could be seen in his coins. Steinmann cites numismatic evidence in support of Herod using Rome-based dating of his reign on page 27:

Herod’s first coins, issued to replace Hasmonean currency, are also the first dated Jewish coins. They are dated to “year three.”

Pictures of these coins, derived from David Hendin’s authoritative Guide to Biblical Coins, can be found at Then he continues:

Clearly, Herod counted the year he first reigned in Jerusalem as the third year of his reign. This means that he counted his first regnal year as beginning no later than Tishri 38 BCE and issued his first coinage shortly after conquering Jerusalem in 36 BCE (emphasis added).

We may regard the supposition that “Herod counted the year he first reigned in Jerusalem as the third year of his reign” as likely, since there are no known Herodian coins bearing year one or two (or any) dates. But the emphasized statement following it depends on first assuming the Filmerian view that Herod was named king by the Romans in 39 BC with non-inclusive accession-year dating applied (39 BC up to the end of Elul being his accession year), followed by the taking of Jerusalem in the fall of 36 BC. The website disagrees, saying the coin’s date refers to Herod’s capture of Judea in 37 BC. Steinmann has obviously allowed his assumptions to lead his arguments here.

It should also be observed that Steinmann deals with coin matters very carefully, pointing to the “year three” notation as evidence Herod dated his reign from his Roman appointment, yet in the case of Herod’s son Antipas going to great pains to argue against the numismatic evidence:

Antipas lost the tetrarchy of Galilee and Perea in the second year of Gaius (38/39 CE) and the latest coins minted under his authority are dated to his forty-third year. This means that he claimed to have begun his reign in 5/4 BCE. Why would Archelaus and Antipas claim to have reigned from 4 BCE if Herod did not die in that year? Is this not proof that Herod must have died in 4 BCE and not 1 BCE? (p. 20).

It would seem so on the surface, but Steinmann then goes on to argue at length (devoting five pages to this), on a conjectural basis with considerable reading between the lines, that Herod actually did not die in 4 BC, and the evidence of the coins is misleading in this case. This is a discussion that must be tabled for now. My only point in bringing it up is to show that Steinmann is willing to reinterpret the apparent sense of the evidence when doing so will further his argument.

Let us return to our main point, that Herod’s possible view of the start of his reign is less important than how Josephus viewed it. As we saw earlier when discussing the evidence for inclusive reckoning in Josephus, the examples of Pompey’s 27 years, the Temple’s fall after 107 years, and the Battle of Actium assigned to Herod’s seventh year, all rely on dating from 37 BC. So, too, does the rebuilding of the Temple in Herod’s eighteenth year. None of these figures work with Josephus’ own time spans if they are counted from the Roman appointment three years previously. In presenting the synchronisms as he did, it is clear that, so far as Josephus was concerned, the taking of the city and deposing of Antigonus took priority over the Roman appointment for dating purposes. So, if our goal is to understand Josephus rather than read our own preferences onto his records, that is the basis we likewise should prefer for dating.

At this juncture I wish to mention an insight Ed Rickard shared on his The Moorings website, He proposes that the more detailed information in Antiquities was due to Josephus uncovering additional sources of authoritative, reliable information for dating the reign of Herod after Wars was written. Filmer had pointed out (pp. 286–287) that Josephus included nothing in Wars about the consular year and Olympiad synchronizations Antiquities gives for the 40 and 37 BC regnal start dates. Filmer, filtering this observation through his presuppositions, concludes that Josephus devised his own date synchronizations and introduced dating errors into Antiquities in the process.

There is another way of looking at it, however, that does not require demeaning Josephus: to posit that the additional information sources were official Roman records not known to him until after Wars was written, plus what might be called, as Ralph Marcus and Allen Wikgren translate it in the Loeb version of Antiquities 15.6.3, “Herod’s Memoirs.” This idea is attractive. It is highly probable that official records from Herod's time would have reflected consular dating with January as the first month, while information written by Herod himself would also likely have been from a Roman perspective, given how much he owed them for his exalted position. But even if it was, the question still remains: did Josephus himself embrace that perspective in the way he wrote Antiquities? For most Herodian events it makes no substantive difference whether Josephus’ records reflect a January or Nisan first month, but it does matter when we consider exactly when Herod died. That is a matter for future consideration.

It is also worth noting that, notwithstanding the evidence of the coins which, as legal tender of the realm, would naturally have reflected Roman preferences, Herod expended great effort to have the Jews regard him as one of their own rather than a Roman pawn. This is particularly clearly illustrated in the remarks he made in Antiquities 15.11.1 before undertaking the building of the Temple:

I think I need not speak to you, my countreymenOur Fathers indeed, when they were returned from Babylon, built this temple to God Almighty… And it hath been by reason of the subjection of those fathers of ours… I will do my endeavour to correct that imperfection, which hath arisen from the necessity of our affairs, and the slavery we have been under formerly…(emphasis added).

Such self-evident intent to foster an identification of himself with the Jews and gain their loyalty is inconsistent with Steinmann’s insistence that only Roman views mattered to Herod:

Since Herod was appointed by a Gentile power, he probably [assuming a 39 BC appointment as king] began to count his official regnal years as beginning on the following Tishri (September/October) of 38 BCE (since the Jewish civil year began on Tishri). He may have counted his years as beginning in Nisan (March/April) of 38, but this is less likely, since this was the beginning of the religious year, and it would have been unwise to count a Gentile appointment from a sacred Jewish date….This also implies, however, that in Antiquities Josephus numbered Herod’s regnal years from his appointment by the Romans (p. 27, emphasis and bracketed comments added).

All of that is pure conjecture; notice the words “probably,” and “may have.” Consider as well, that apart from being the start of a month—sharing certain prescribed rites with other first days of months (Num. 10:10, 28:11–15), including the first of Tishri—there was nothing especially “sacred” about the first day of Nisan. It was primarily a starting reference point, the first month of the year, from which the Jewish feasts and fasts mandated by the Torah were ordered. It carried with it none of the special “sacredness” that characterized those dates and the weekly Sabbath (Lev. 23).

To wrap up this phase of the discussion, the Mishnah’s stipulations regarding the Tishri dating of kings were for record-keeping purposes relative to other countries, not how kings reigning over the Jews from Jerusalem started their rule. Tishri dating therefore does not apply to Herod as a king of Judea. Yet, despite the united testimony of Scripture, Josephus and the Mishnah in favor of inclusive reckoning from Nisan from the time of Esther onwards for Judean kings, Filmer and Steinmann argue against that evidence, asserting that Josephus used accession-year, non-inclusive reckoning from Tishri. They then use this conjecture as the basis for a frontal attack on the plain sense of Josephus’ records. Filmer was the first modern scholar to take this tack, citing Edwin Thiele to claim that the reigns from Solomon to Zedekiah (pre-exilic rulers of the united monarchy, it should be noted) were reckoned from Tishri (p. 294), and then extrapolating from that to Herod. But surely the knowledgeable rabbis who put the Mishnah together had their own post-exilic history and the dating norms that arose from it down pat, which Josephus’ focus on Nisan dating indicates he likewise followed. Arguing against it appears to be flouting the evidence.

An Insistence on Factual Reckoning: The Actium Issue

Nevertheless, Steinmann chose to align himself closely with Filmer on this and many other matters. One is the Actium issue. He deals with it briefly on pages 5–6 in the context of critiquing the work of P.M. Bernegger, “Affirmation of Herod’s Death in 4 BCE,” Journal of Theological Studies 34 (1983), 526–531. What his objections boil down to is an insistence on looking at Josephus’ data through the lens of factual (date-specific), non-inclusive reckoning from September 2, 31 BC. For example, Bernegger (p. 529) cites Josephus’ discussion about the Roman tax registration in Syria during AD 6:

Josephus stated that the registration was completed in the thirty-seventh year after Actium. The battle of Actium took place in 31 B.C., thirty-six factual years before the completion of the Syrian registration. In this instance, Josephus counted inclusively, and without any ambiguity.

Steinmann protests, “However, Bernegger’s reasoning only works if one forgets about the date of the Battle of Actium, September 2, 31 BCE.” This reveals his Filmerian assumptions. Shelve the idea that factual, to-the-day dating matters, together with the presumption of non-inclusive year counts, and the problem vanishes. Steinmann also projects his own bias onto Bernegger when he writes, “Years after Actium commenced on September 3, not on the following January 1, as Bernegger’s calculations assume.” But as a follower of Schürer’s approach that adhered to Jewish and Roman inclusive dating conventions, Bernegger’s year count would have been inclusive, making the first year “after Actium” begin in 31 BC, not the following January. This makes 30 BC the second year after Actium, and so on until the 37th year in AD 6. Whether Bernegger used a January-to-December Julian calendar or Nisan to Adar, the inclusive approach still makes September 2 part of 31 BC rather than 30 BC. Thus, we see that Steinmann has here criticized a misrepresentation of Bernegger’s position. His so-called “Schürer consensus” increasingly resembles a convenient straw man for him to attack. It bears only passing resemblance to the actual positions of Emile Schürer and those who followed him.

The problems posed by the Battle of Actium against Filmer and Steinmann’s interpretation get worse. For example, Steinmann claimed in note 83 of his article:

The Battle of Actium [September 2, 31 BC] would have taken place at the very end of Herod’s seventh year, since Tishri can begin no earlier than September 20 and no later than October 19. In 31 BCE the Babylonians counted September 21 as the first day of Tishri (Richard A. Parker and Waldo H. Dubberstein, Babylonian Chronology 626 BCE–AD 75 [Brown University Studies 19, Providence: Brown University, 1956] 43). This confirms that Herod started his regnal years in Tishri, not Nisan (emphasis and bracketed comment added).

But this is circular reasoning. All this statement proves is that the Battle of Actium took place in the sixth month, Elul, just before Tishri started. It does not indicate that Actium was at the very end of Herod’s seventh year, or of any year. To claim that means first assuming a Tishri-based year—the very thing Filmer and Steinmann must prove—as well as factual dating. If Nisan-based inclusive dating was used instead, Actium would still have occurred just before Tishri, but would have fallen in the middle of Herod’s seventh year. Not only does this logic fail, the claim again flies in the face of Josephus’ testimony in Antiquities that his historical records for the Jews revolved around the Nisan-based ecclesiastical calendar, not the civil calendar using a Tishri New Year. We should place priority on the source material’s own interpretation of itself, not on a modern scholar’s reinterpretation of it. If it makes good sense as written, there is no real reason to reject it.

Further to this, at, Kenneth Frank Doig observed:

Andrew E. that the Battle of Actium in Herod’s 7th year on September 2, 31 BCE establishes Josephus “confirms that Herod started his regnal years in Tishri, not Nisan.” However, the dating is such that it was Herod’s 7th year reckoned from either Nisan or Tishri. Because of using dating from Tishri Steinmann elsewhere says Josephus “contradicts” himself.

My own independent analysis, put into a spreadsheet long before I read Steinmann’s article, agrees with Doig’s conclusions. Whether Nisan (March/April) or Tishri (September/October) reckoning is used for the start of the year, both of these possible New Year’s dates fall squarely into the January-to-December year of 31 BC.

In the end, what “Alexander” wrote (, quoted in my last article, still applies:

Despite any counting methods that may be employed by various authors, whether Nisan to Nisan, Tishri to Tishri, or even January to January, it holds true nonetheless that if the spring of 31 BCE is his seventh year, then the spring of 32 BCE is his sixth year, the spring of 33 BCE is his fifth year, and so on, making the spring of 37 BCE his first year (emphasis added; the argument would remain true if all instances of “spring” read “fall” instead).

The Filmerian Reinterpretation of Josephus

We will now undertake a point-by-point analysis of Steinmann’s case against the “Schürer consensus.” As observed above, he repeatedly emphasizes the significance of the de jure (in law) date of Herod’s Roman appointment over the de facto (in fact) date of taking Jerusalem. This is intimated as early as page 2, right after summarizing the main points favoring the 37 BC consensus for the start of Herod’s reign:

Despite this widely held opinion that Herod reigned from 40 (37) to 4 BCE, this was neither the consensus before Schürer nor has it gone unchallenged in the last half century. Most disturbingly, the Schürer consensus assigns only thirty-six years to Herod’s reign, thirty-three of them in Jerusalem, whereas Josephus reports the figures as thirty-seven and thirty-four respectively. All early Christian sources place the birth of Jesus after Passover in 4 BCE, with most of them placing it in sometime in late 3 or early 2 BCE (emphasis added).

Concerning these comments, three observations can be made:

First, we need to be clear: Herod did not in fact reign over the Jews from 40 BC on (or 39 BC, if one follows the Filmerian reinterpretation of Josephus), but from 37 BC (36 BC per Filmer and Steinmann). The “widely held opinion”—the “Schürer consensus”—views him in 40 BC merely as king-designate, and in Roman eyes only, until the city was actually taken and placed under his control. Antigonus, as we shall see, was the king of the Jews in every measure of the word—title (including on his coins, which bore the inscription BACIΛEC ANTIΓONOY (of King Antigonus), see, government control, residence in Jerusalem, and acceptance as king by those he ruled—from 40 BC until Herod’s siege removed him and led to his death in 37 BC. This is surely a common sense observation, but it must unfortunately be specifically pointed out, lest the complex arguments put forth by Filmer and Steinmann obscure it.

Second, Steinmann asserts that the “disturbing” Schürer consensus contradicts the figures given by Josephus for the 37 and 34 years of his reign as measured from the Roman appointment and the taking of the city respectively. Actually, what is disturbing is this misrepresentation, for no conflict with Josephus can be found. Schürer himself wrote, in his note 165:

Herod died shortly before a Passover (Antiq. xvii.9.3; Wars of the Jews, ii.1.3), therefore in March or April. Since Josephus says that he reigned thirty-seven years from the date of his appointment, thirty-four years from his conquest of Jerusalem (Antiq. xvii.8.1; Wars of the Jews, i.33.8), it would seem as if, counting thirty-seven years from the year B.C. 40, he must have died in B.C. 3. But we know that Josephus elsewhere counts a year too much, according to our reckoning…The reason of this is that he counts portions of a year as a year [i.e., he counts inclusively]; and, indeed, he probably, according to the example of the Mishna, reckons the years of the king’s reign from Nisan to Nisan. If this be so, the thirty-fourth year of Herod would begin on the 1st Nisan of the year B.C. 4, and Herod must in that case have died between 1st and 14th Nisan, since his death occurred before the Passover. That this is indeed the correct reckoning is confirmed by astronomical date, and by the chronology of the successors of Herod (bracketed comment added).

Since at this time we will not discuss the death of Herod, we will skip over the last two lines (although we agree with them), and just note that there is nothing in what Schürer wrote to conclude that he disagreed with Josephus’ 37 and 34-year figures. His reasoning is actually predicated around accepting them as written. Neither did he accuse Josephus of error, as Steinmann does (“Thus, Josephus is in error,” p. 7) in reference to his matchup of consular and Olympiad dates. (See also page 28, “Josephus made mistakes in Antiquities 14.389, 487 when reporting the consular and Olympian dating of the beginning of Herod’s reign.” The supposition that there was a direct conflict between Josephus’ equating the consular year of Calvinus and Pollio with the 184th Olympiad is addressed under “A Closer Look at the Consular Years,” below.) On the contrary, Schürer fully accepted those numbers and sought to understand them as Josephus and the Jews did, rather than imposing modern non-inclusive dating conventions upon them. The only contradiction is not with Josephus, but with the ultimately unsupported insistence of Filmer and Steinmann on using non-inclusive rather than inclusive reckoning, and that from Tishri rather than Nisan.

Third, he states that early Christian sources place the birth of Christ after 4 BC, generally in late 3 or early 2 BC. It should be pointed out, however, that those sources are not unanimous about a specific year, which indicates they reflect not accurate records but tradition (i.e., early Church hearsay). Jack Finegan’s Handbook of Biblical Chronology (p. 291) gives several dates suggested by early sources. Although it is true that a majority are listed as 3/2 BC, it should not escape our notice that these give a date range, and there are a number of outliers. To take these reports as authoritative is to depend on unproven tradition rather than a single well-attested year. If Luke, arguably the most historically picky of the New Testament writers, did not pinpoint the year for us, nor did any of the other inspired apostles who knew Him (and His mother Mary) best, we have no objective criterion for dating Christ’s birth, only old theologians’ tales. We must conclude, therefore, that early Church tradition gives us no clear year for the Savior’s birth, and therefore no conclusive help in pinpointing the year of Herod’s death.

Reinterpreting Three Incontestable Points

Beginning on page 8 of his article, Steinmann raises several critiques against 37 BC as the start of Herod’s de facto reign. He begins by presenting three incontestable points from Antiquities 14:

- Herod’s siege of Jerusalem ended during the consular year of Agrippa and Gallus, which coincided with the 185th Olympiad, “on the third month, on the solemnity of the fast”

- The city fell 27 years after it had under Pompey, on the same day

- The last Hasmonean, Antigonus, was put to death by Antony 126 years after the Hasmonean dynasty was first set up

Steinmann first admits (p. 9) that “the consular year and Olympiad given by Josephus indicates that Herod took Jerusalem in 37 BCE.” This is objective fact. But then he immediately makes two assertions with no such firm basis: “It was the Day of Atonement (“the fast”) on 10 Tishri in the Jewish calendar, but the third month (September) in the Greek calendar” (parentheses original, emphasis added). In what Josephus wrote there is no discernible indication that Greek months entered into the picture at all, nor reason to pivot from Greek months to Jewish days: “This destruction befel the city of Jerusalem when Marcus Agrippa and Caninius Gallus were consuls of Rome; on the hundred eighty and fifth olympiad; on the third month; on the solemnity of the fast” (Ant. 14.16.4). In the post-exilic era several Jewish months, including the third, were routinely designated by their order in the calendar rather than their Jewish names. Scripture itself demonstrates this in Esther 8:9, “in the third month (that is, the month Sivan),” where the parenthetical clarification is part of the verse, and in Ezekiel 31:1, “in the third month.” We can confidently expect Josephus followed that post-exilic convention. The parenthetical explanation in Esther 8:9 also demonstrates the standardized inclusive counting the Filmer camp denies, for Tammuz, not Sivan, would have been specified as the third month if the Jews had used actual, non-inclusive reckoning.

Steinmann’s assertions about the day and month of Jerusalem’s fall appear to arise not from what Josephus wrote, but from the Filmerian preoccupation with Tishri dating. There is no reason to suppose that Josephus, a Romanized Jew whom we already know—from his own words, no less—viewed Nisan as the first month of the year, would refer to a Greek month out of the blue, particularly without also naming it for his readers unfamiliar with Greek conventions (recall how he explained Xanthicus earlier). It also makes little sense that Josephus would flip-flop with his calendars, giving the month in Greek terms, but the day in Jewish terms. The odds are overwhelmingly against it. A straightforward understanding of the passage indicates the Jewish month of Sivan, the third month of the ecclesiastical calendar, was meant. This means “the fast” is impossible to assign to the Day of Atonement in Tishri, the seventh month. “The fast” has another more likely meaning, to be discussed later.

Steinmann then (p. 9) presents two other considerations which, he claims, “contradict” the 37 BC date indicated by both consular year and Olympiad reckoning. First, he says, the 27 years that passed after Pompey takes one to 36 BC, not 37 BC. Two problems exist here: first, he again assumes without supporting evidence that it was the Day of Atonement in Tishri, and second, he makes the further undemonstrated assumption that non-inclusive, actual dating was used. In short, he is using his (and Filmer’s) assumptions as the basis for claiming Josephus was in error, instead of trying to understand the data as Josephus understood it.

As for the second “contradiction,” Steinmann says that there is no evidence of any government by the Hasmoneans until 162 BC, therefore the 126 years had to be reckoned from 36 BC, not 37 BC. But as discussed in the previous article of this series, this overlooks the detail that Antiochus IV Epiphanes died in 163 BC, vacating the Syrian kingship over the Jews and defaulting to leaving the Hasmoneans in power. By recognizing this we can say that the Hasmonean dynasty endured for 126 years.

Three Considerations Favoring Actual, Non-Inclusive Years?

At this point in his article, it is apparent that Steinmann feels the pressure of the “Schürer consensus” against the Filmerian position he has staked out: “Nonetheless, the Schürer consensus could hold that the data given by Josephus here were reckoned by inclusive reckoning, making no conflict” (p. 10). Indeed, it not only could, but it does. But then he adds, “However, that Josephus was not using inclusive reckoning and that these data should be seen as reporting actual years is demonstrated by three more considerations” (p. 10). What are these considerations?

Supposed Conflict of the High Priest Chronology with the Consular Years

The first he owes directly to Filmer (p. 287): “Josephus also contradicts his own consular year for Herod’s conquest of Jerusalem by his chronology of the high priests.” The main assumption behind this is that Josephus used factual, to-the-day dating for the reigns of high priests, similar to the way the Romans reckoned the reigns of their emperors. But another assumption is less obvious: that Josephus’ account of Hyrcanus II and Antigonus views their “reigns” only as those of high priests, not kings. We will discuss this matter in detail below.

The Alleged Passivity of Sosius

The second consideration arises from a conflict Steinmann, again following Filmer (p. 286), sees in Dio’s Roman History, which he claims “casts doubt on the Schürer consensus that the conquest of Jerusalem occurred in 37 BCE.” Here I quote his argument in full (p. 11):

Concerning 37 BCE Dio states (49.23.1–2):

…during the following year [37 BCE] the Romans accomplished nothing worthy of note in Syria. For Antony spent the entire year reaching Italy and returning again to the province, and Sossius [sic], because anything he did would be advancing Antony’s interests rather than his own, and he therefore dreaded his jealousy and anger, spent the time in devising means, not for achieving some success and incurring his enmity, but for pleasing him without engaging in any activity (emphasis and brackets added).

Thus, Sossius would not have helped Herod—a man favored by Antony—capture Jerusalem in 37.

Rather than teaching that Sosius was entirely passive during 37 BC, the Dio passage merely tells us, and quite specifically, that the Romans accomplished nothing of note in Syria. The sense is that, lest personal successes in Syria might inadvertently offend the uninvolved Antony (which had happened earlier with Ventidius, costing him the Syrian governorship, Dio 49.21.1), Sosius likewise did nothing there. This text does not address activities Sosius might have pursued in Judea at Antony’s specific behest, however. Making the assumption that Sosius’ fear of affronting Antony paralyzed him into inactivity everywhere is entirely unwarranted. Since Dio emphasizes Sosius’ desire to please Antony, if Antony wanted him to help Herod with the siege of Jerusalem, of course he would! That is exactly what the text says in Antiquities 14.16.1: “Sosius [was] sent by Antony, to assist Herod.” The supposed problem Filmer and Steinmann see is nonexistent. Consistent with this, in Wars 1.17.2 Josephus elaborates a little further:

For after the taking of Samosata [in 38 BC], and when Antony had set Sosius over the affairs of Syria, and had given him orders to assist Herod against Antigonus, he [Herod] departed into Egypt; but Sosius sent two legions before him into Judea, to assist Herod, and followed himself soon after with the rest of his army (emphasis and brackets added).

This information is in direct conflict with the idea that Sosius “would not have helped Herod” to capture Jerusalem in 37 BC. Therefore, we can dismiss the Dio “problem” as being nothing of the sort for the “Schürer consensus.”

The Sabbatical Years

The third consideration adduced by Steinmann against a 37 BC fall of Jerusalem is tied to the Jewish sabbatical years. The sabbatical years are brought up in two places in Antiquities in conjunction with the start of Herod’s reign. Describing the siege undertaken by Sosius and Herod, Josephus records:

Now the Jews that were inclosed within the walls of the city, fought against Herod with great alacrity and zeal...and making use of brutish courage, rather than of prudent valour, they persisted in this war to the very last. And this they did while a mighty army lay round about them; and while they were distressed by famine, and the want of necessaries: for this happened to be a sabbatick year (14.16.2, emphasis added)

He continues the story in the next chapter, observing:

Nor was there any end of the miseries he [Herod] brought upon them [the defeated Jews]: and this distress was in part occasioned by the covetousness of the prince regnant [Herod was confiscating silver and gold wherever he could find them]; who was still in want of more; and in part by the sabbatick year, which was still going on, and forced the countrey to lie still uncultivated: since we are forbidden to sow our land in that year (15.1.2, emphasis and bracketed comments added).

These details give us a way to determine the year that Herod took Jerusalem, but only if we can identify with confidence at least one other post-exilic sabbatical year to synchronize with it. Once again following Filmer’s lead (pp. 289–291), Steinmann presents this argument (p. 11):

Finally, it should be noted that Herod besieged Jerusalem at the end of a Sabbatical year when food supplies were running low. This was the same situation in mid-162 BCE near the end of a sabbatical year. Thus, Tishri 163–Elul 162 was a Sabbatical year. Since the summer of 162 BCE fell during a Sabbatical year, the summer of 37 BCE could not have been a Sabbatical year. Instead, Tishri 37 BCE–Elul 36 BCE was also a Sabbatical year. Since food supplies would have been adequate at the beginning of the Sabbatical year, Jerusalem could not have fallen to Herod in Tishri 37 BCE as the Schürer consensus holds. Instead, Jerusalem fell at the beginning of the following year (Tishri 36), with the siege taking place during the summer of the Sabbatical year (summer of 36 BCE).

To begin with, I have no idea how Steinmann concluded that his “Schürer consensus” holds that Jerusalem fell to Herod in Tishri. None of the references I looked at that accept a 37 BC de facto start for the reign of Herod place the taking of Jerusalem in Tishri, but in early summer, generally the month of Sivan (June of 37 BC, cf. This conclusion follows the logic that the “siege of five months” (Wars 1.18.2) began “after the rigour of winter was over” (Ant. 14.15.14) around February, and concluded in the “summer time” (Ant. 14.16.2), “in the third month” (Ant. 14.16.4) of the Nisan-based ecclesiastical calendar Josephus favors. No, those who follow Schürer do not think Jerusalem fell in Tishri.

Now, the validity of all of Steinmann’s sabbatical year reasoning, including the assertion that 163–162 BC was a sabbatical year, depends on first accepting the foundational premise of a Tishri-based, actual/accession-year/non-inclusive dating scheme, and then presuming on the accuracy of the sabbatical year determinations made by Ben Zion Wacholder (“The Calendar of Sabbatical Cycles During the Second Temple and the Early Rabbinic Period,” Hebrew Union College Annual 44 [1973], 153–193). Quite possibly under the influence of Filmer’s 1966 article, Wacholder decided to revisit the careful earlier study done by Benedict Zuckermann (Treatise on the Sabbatical Cycle and the Jubilee, translated by A. Löwy from the German original of 1856), and concluded Zuckermann’s dates for the sabbatical years were a year too early. Of particular note is that he concluded that 37-36 BC, Tishri through Elul, was a sabbatical year, aligning it with Filmer’s date for Herod taking Jerusalem.

So, which is more accurate for the post-exilic period, the sabbatical year determination of Wacholder, or the one by Zuckermann? A detailed discussion of the issues involved is given in a Wikipedia entry ( that references the work of many acknowledged authorities, such as Parker and Dubberstein, Edwin Thiele, and Jack Finegan. It also draws on the work of ABR's Dr. Bryant Wood and several ABR Associates, including biblical chronologist Rodger Young and Dr. Douglas Petrovich. After acknowledging that the geonim (medieval Jewish scholars) and the modern state of Israel follow Zuckermann’s approach, this significant admission seems to be depreciated in favor of a sympathetic focus on Steinmann’s views favoring Wacholder. Arguments are also presented there based on the Seder Olam in support of Wacholder’s dates, but they are ultimately rendered indecisive by translation uncertainties. Their uncertainty is compounded by the fact that the Talmud demonstrates that even the leading rabbis could not agree on when the sabbatical years after the Second Temple should be observed (for example, see the convoluted discussion given in Mas Arachin 12a–12b, Without a solid, objective basis for translating its problem passages bearing on the sabbatical years, the Seder Olam provides no conclusive help in choosing between the approaches of Zuckermann and Wacholder. We must look elsewhere for a basis to make the choice.

I believe we find this basis in the detailed analysis of sabbatical year evidences of Wacholder and Zuckermann presented by Bob Pickle. The above-cited Wikipedia article is incomplete without considering the balanced treatment Pickle sets forth (as well as Blosser’s study, see below), which would help readers better appreciate why, as the Wikipedia article admits, “there are many prominent scholars who still maintain a cycle consistent with Zuckermann’s conclusion of a 38/37 BCE Sabbatical year.” Pickle’s two online articles, “When Were the Sabbatical Years?” ( and “Which Years Were the Sabbatical Years?” (, are of such high quality, they should have been published in a scholarly journal. Here we merely summarize some of their key points.

Wacholder had presented ten lines of evidence for his sabbatical year determinations, and Pickle bases his study on them. They are:

The Pledge of Nehemiah 10:31
Alexander’s Grant of Tax Exemption
Judah Maccabee’s Defeat at Beth-Zur
Simon’s Murder
Herod’s Conquest of Jerusalem
Herod Agrippa’s Reading of the Law
Note of Indebtedness from Nero’s Reign
Destruction of Second Temple
Land Contracts of Bar Kochba
Tombstones from Zoar

For each of these topics Pickle examines the data, contrasts how the Zuckermann and Wacholder approaches deal with it, and draws conclusions about which does a better job at explaining the data. In his Introduction he asks, “So which proposal is correct? First of all, why does it matter? It matters because this question is pertinent to a study of the 70 weeks of Daniel 9.” (This is exactly right, and the reason why we will have further reason to address the sabbatical years as part of “The Daniel 9:24–27 Project.”) He then goes on:

Since Daniel 9 begins with a reference to the 70 years of Babylonian captivity, this conclusion [of Wacholder, that the Hebrew word for “week” in Daniel 9 is used in Jewish writings to refer to sabbatical cycles] seems certain. The reason the Jews were sentenced to a captivity of 70 years is because they had not kept that many sabbatical years and had to catch up (Lev. 26:34–35, 43; 2 Chr. 36:21–23). So Daniel 9 begins with a reference to the missed sabbatical years of 70 sabbatical cycles, and ends with a discussion of another 70 sabbatical cycles (bracketed comment added).

It therefore follows that if a particular interpretation of the 70 weeks coincides with known sabbatical years, then that interpretation has additional merit. This approach requires the positive identification of at least one sabbatical year sometime in history.

This paper assumes that Wacholder’s ten lines of evidence for his position are the best possible case against Zuckermann’s dates. What follows is a discussion dealing with each of these lines of evidence as they appear in his 1973 paper. A careful re-analysis of this data seems to indicate that Zuckermann’s dates are the correct ones after all.

It is outside the scope of this study to detail all the various points Pickle covers. We will only note here that, when the actual data showing the confluence of the consular year of Agrippa and Gallus with the 185th Olympiad is allowed to stand as Josephus wrote it, it indisputably indicates that 38/37 BC was the sabbatical year indicated by Antiquities 14.16.2–4. As Pickle puts it:

The 185th Olympiad began in July 40 BC and ended in June 36 BC. Agrippa and Gallus were consuls in 37 BC. Thus Jerusalem was besieged by Herod in the spring and summer of 37 BC. Since 38/37 BC was a sabbatical year according to Zuckermann, Zuckermann’s sabbatical dates must therefore be correct.

I should mention here my own approach to the challenge of determining a solid basis for post-exile sabbatical year calculation. I first tried to identify the most likely first sabbatical year after the return of the Jews from the Babylonian captivity. (This was done before looking at either Wacholder’s or Zuckermann’s dates, so they did not “prime my pump.”) The details of how I made my determination must wait for a later time, but briefly, I assumed that the 70 years the land lay “desolate” and uncultivated during the Babylonian captivity served to eliminate the Jews’ sabbatical year debt. When the Jews returned to the land and commenced agriculture once again, this initiated a new sabbatical cycle from Tishri. (I later discovered that Zuckermann had made the same deduction.) On that basis, after considerable study of Ezra and Nehemiah, I determined a likely first post-exilic sabbatical year. It was only after this that I discovered Pickle’s website, and was gratified to find that Benedict Zuckermann’s data reproduced there gave the same conclusions I had independently arrived at. This gave me a considerable measure of confidence than my determination was correct, despite the fact that it disagreed with Wacholder.

Building upon that determination—and independent of my studies of Herodian chronology that came later, by the way, so the latter in no way influenced my sabbatical year conclusions—I determined from Zuckermann’s dates that there was a sabbatical year running from the start of Tishri in 38 BC (October 7) to the end of Elul in 37 BC (September 25). This meant, since Josephus tells us that Herod took Jerusalem during a sabbatical year, the siege must have ended during the summer of 37 BC. The 36 BC date preferred by Filmer and Steinmann would place the siege after the sabbatical year had ended, so it would not work with Zuckermann’s formulation. If one instead opts for Wacholder’s scheme, the sabbatical year would have run from Tishri 37 BC to Elul 36 BC (Steinmann, p. 11), theoretically reconciling with it.

However… Steinmann holds that the “fast” mentioned in Antiquities 14.16.4 was the Day of Atonement, 10 Tishri, 36 BC. This would have been ten days after Wacholder’s sabbatical year had ended. Josephus said in Antiquities 15.1.2 that immediately after the end of the siege, while Herod was plundering Antigonus’ loyalists, “the sabbatick year…was still going on.” This means problems exist for Steinmann’s approach even if we accept Wacholder’s 36 BC sabbatical year. To nullify it, Steinmann would have to drop his insistence on the “fast” being the Day of Atonement, but doing so would seriously undermine his thesis keyed on factual, non-inclusive dating.

In the end, even if Steinmann should discard the Day of Atonement idea, all of the other problems discussed earlier with the Filmer/Steinmann approach still remain, as well as the many conflicts Bob Pickle points out that exist with Wacholder’s ideas. It is beyond the scope of this article to detail them; the interested reader is referred to Pickle’s website. Suffice it to say that there are far fewer problems involved in utilizing Zuckermann’s sabbatical year pattern, and it offers an independent confirmation that 37 BC marked the start of Herod’s reign.

We will close this sabbatical year discussion by noting that the highly-respected Jack Finegan pointed out on page 116 of his Handbook of Biblical Chronology (revised edition):

In 1979 Donald Wilford Blosser published a new study of Jubilee and Sabbatical years, with a calendar of Sabbatical years extending from 171/170 B.C. to A.D. 75/76, a tabulation which is contrary to Wacholder and in exact agreement with Zuckermann (“Jesus and the Jubilee: Luke 4:16–30, The Year of Jubilee and Its Significance in the Gospel of Luke” [Ph.D. diss., St. Mary’s College, The University of St. Andrews, Scotland, 1979], 113, emphasis added).

Finegan's conclusion on sabbatical year matters is to side with Zuckermann over Wacholder:

Since we have taken Yose ben Halafta as an early and dependable authority, we accept the date of 68/69 [as the sabbatical year before Titus took Jerusalem] and also use it as basic for the determination of several other Sabbatical years in what follows, all thus, in fact, in accordance with Zuckermann (and Blosser) (p. 122, § 226, bracketed comment added).

Antigonus and the “Times of Herod”

Steinmann then goes on to attempt to justify dating Herod’s reign from his de jure appointment as king by the Romans rather than his de facto rule after Antigonus' death, by appealing to the 28 high priests and a strange definition of “the times of Herod”:

However, a closer examination of Antiquities 20.250 [20.10.1 in Whiston’s edition] demonstrates that Josephus was reckoning Herod’s years from his appointment by the Romans. This is shown by Josephus’ noting that there were twenty-eight high priests from “the times of Herod” until the destruction of the temple in 70 CE. When Herod conquered Jerusalem he appointed Ananel to be high priest. Counting high priests beginning with Ananel and ending with Pannias, the last high priest before Titus conquered Jerusalem, there were twenty-seven high priests. This means that Josephus was including Antigonus in his reckoning of twenty-eight high priests during the reign of Herod. To confirm this, note that above it was demonstrated that Antigonus reigned in Jerusalem as high priest from Tishri 39 to Tishri 36. Therefore, Josephus began the “times of Herod” with Herod’s appointment by Rome three years earlier than his conquest of Jerusalem, and the beginning of his reign according to official regnal years overlapped the high priesthood of Antigonus by about two years (1 Tishri 38 BCE–10 Tishri 36 BCE). However, if one were to date the “times of Herod” to his appointment by the Romans according to the Schürer consensus (40 BCE or perhaps 39 BCE in official regnal years) this would mean that there were 110 or 109, not 107 years from “the times of Herod” to Titus’ conquest of Jerusalem (p. 28, emphasis original, bracketed comment added).

It must first be pointed out that Steinmann’s claim about his self-defined “Schürer consensus” again misrepresents what Schürer and those following him believe. See again what Schürer wrote in his footnote 165:

Again, from the conquest of Herod down to that by Titus he [Josephus] counts 107 years (Antiq. xx. 10)…(bracketed comment added).

“From the conquest of Herod.” This starting point is unambiguously referring not to the Roman appointment the Filmerians focus on, but to Herod’s de facto rule over Judea following the siege and putting to death of Antigonus. Given the convention of using inclusive dating from Nisan for the reigns of Jewish kings, 107 years is exactly right for the time spanning 37 BC to AD 70.

Apart from that, what Josephus wrote “demonstrates” none of the things Steinmann claims in his quote above. His statement rests on the extremely shaky foundation of two unsupported assumptions. First, it depends on using accession-year, non-inclusive dating from Tishri. We already showed this runs counter to normal dating conventions for post-exilic Jewish kings, and as discussed above, this is apparently one reason why Steinmann wants to reckon Herod’s kingship from his Roman appointment, not the death of Antigonus: he does not want to view Herod as a Jewish king subject to the Nisan ecclesiastical calendar, but as essentially a Roman vassal assigned to Judea, with the unspoken implication that the Tishri civil calendar applying to foreign rulers applies to him as well. Thus, moving the beginning of Herod’s reign forward by six months, from Nisan in the spring to Tishri in the fall, and then not counting inclusively, gives Steinmann a theoretical mechanism for down-dating Herod’s reign from 40 to 39 BC.

Second, this reinterpretation of Josephus specifically depends on the atypical list of high priests put together by James C. VanderKam (From Joshua to Caiaphas: High Priests after the Exile [Minneapolis: Fortress, 2004] 394–490, 492). That list diverges from the widely accepted ones given by Whiston ( and Wikipedia (—which happen to be identical in the persons included, though presented differently—by leaving off Jonathan (AD 58), and supplying in his place Antigonus as the requisite 28th high priest. Doing this, however, requires supposing that the “times of Herod” (as king) included the period 39–36 BC (by Steinmann’s reckoning) covering Antigonus’ reign (an important little word!), a period when Herod was neither in control of Jerusalem nor viewed as king by its inhabitants (see the Strabo quote below). It further requires viewing Antigonus merely as a high priest, though Josephus plainly describes him as the factual king of the Jews immediately preceding Herod! How could Antigonus, a king of Judea whose coins, we saw above, declare him to be such, simultaneously be regarded as a high priest during Herod’s “times”? When we read Antiquities 20.10.1, we find Antigonus described in this manner:

Barzapharnes and Pacorus, the generals of the Parthians, passed over Euphrates, and fought with Hyrcanus, and took him alive, and made Antigonus, the son of Aristobulus, King [not merely high priest]. And when he had reigned [as king] three years and three months, Sosius and Herod besieged him, and took him (emphasis and bracketed comments added).

We see that after deposing John Hyrcanus II—who had also basically functioned as a king, save that the Romans kept a close eye on him and gave him the lower title of ethnarch, governor of the nation, and denied him the privilege of wearing a kingly crown—the Parthians made Antigonus king. Although he also functioned as a high priest, Josephus actually ignores that aspect. That his kingship is emphasized is crucial to recognize, it depreciates his being high priest to irrelevancy. Steinmann, however, overlooks this detail and instead, following VanderKam, emphasizes only the high priest side of the man, in order to justify using Herod’s Roman appointment as the true start of his kingship, with the dating implications inherent in that choice. In so doing he is utterly silent regarding Whiston’s earlier, longstanding, well-accepted listing of the 28 high priests that left off Antigonus and included Jonathan.

Filmer and the “List of High Priests

In his 1966 paper, Filmer likewise downplayed the kingship roles of Hyrcanus and Antigonus, solely emphasizing their function as high priests. From that standpoint he used the mention of the 3-1/2 year reign of Antigonus, considered only as a high priest, to argue for factual, non-inclusive reckoning instead of inclusive reckoning from Nisan:

In fact the dates 40 and 37 B.C. for the accession of Herod are at variance with the chronology of this period as given by Josephus himself. In a list of high priests and the periods for which they held office, he gives Hyrcanus twenty-four years and Antigonus three years and three months. Now Hyrcanus was appointed by Pompey in 63 B.C., whence we deduce that Antigonus began his reign in 39 and was removed by Sosius in 36 (emphasis original). 

Notice the emphasis on the words, “list of high priests.” Although it is true that the section of Antiquities where these matters are discussed (20.10.1) has a certain focus on high priests, Josephus discusses Hyrcanus and Antigonus as kings, not high priests! This is obvious when we look at the broader context of what Josephus wrote in 20.10.1, noting particularly how he uses the words “reign” or “rule” to focus our attention on the function of kingship rather than of a high priest:

Judas…kept the priesthood, together with the royal authority... Alexander had been both King and High Priest…Aristobulus…did himself both reign [i.e., act as king], and [in contrast to reigning] perform the office of High Priest to God. But when he had reigned [as king] three years, and as many months, Pompey came upon him: and…restored the High Priesthood to Hyrcanus; and [in addition to being high priest] made him governor [de facto political ruler] of the nation: but forbad him to wear a diadem. This Hyrcanus ruled [as effective king, albeit crownless], besides his first nine years, twenty-four years more, when Barzapharnes and Pacorus, the generals of the Parthians…made Antigonus, the son of Aristobulus, King. And when he had reigned [as king, not as performing the office of high priest] three years and three months, Sosius and Herod besieged him, and took him. When Antony had him brought to Antioch, and slain there, Herod was then made King by the Romans…(emphasis and bracketed comments added).

It is very important to observe above that Josephus uses “reign” specifically to describe the kingship function, while he uses “perform the office”—not “reign” or “rule”—to describe the separate high priestly function. Plus, when Sosius and Herod besieged him, it was not because Antigonus was functioning as a high priest, but as king of the Jews. Not only did Filmer, like Steinmann after him, downplay the obvious kingship role of these men that Josephus highlights in Antiquities, he used that minimizing of the kingship aspects as an excuse for suggesting factual reckoning was used here; because he wanted to start Herod’s reign from his Roman appointment in 39 rather than 40 BC, he needed to count those three extra months after Nisan (which marked Antigonus’ third year as king, more evidence the Nisan calendar for dating Jewish kings was used, not Tishri). This was a pretext for ignoring the Roman and Jewish convention of reckoning the rule of kings inclusively.

What all of this means is, Antigonus reigned as king until his removal by Sosius and Herod in 37 BC. It was at this point that the “times of Herod” began, not before. This is why I wrote earlier that Steinmann used a strange definition of the “times of Herod”:  it includes a different king as part of Herod’s “times.” We cannot emphasize this too strongly: Antigonus cannot be regarded as a high priest under Herod.

Filmer continues:

These two terms of office [Hyrcanus plus Antigonus] together total twenty-seven years, and so conform to the twenty-seven years’ interval between the two captures of Jerusalem by Pompey and Sosius which historians reject. Neither can it be argued that the twenty-four years of Hyrcanus were inclusive of an accession year, for, as we shall see, Josephus did not use that system of reckoning, and furthermore, Antigonus’ term of office is given as three years and three months, which is clearly factual (emphasis added).

“Term of office” was emphasized above to point out that Filmer is talking about Antigonus only as a high priest, completely ignoring the fact that Josephus does not use “term of office” in regard to him, but “king” and “reigned.” The three years measure the Nisan-based reckoning of Judean kings’ reigns, while the additional three months accounts for the fall of Jerusalem in Sivan.

Filmer does not tell us who the historians are who supposedly “reject” the 27 years from Pompey to Sosius/Herod; it must be people unwilling to use inclusive reckoning, which easily explains the 27 years (the three months then being part of Antigonus’ third year as reckoned inclusively from Nisan, not part of a fourth year by accession reckoning from Tishri). We can grant that the above is a single clear instance where Josephus gives a factual reign length for Antigonus, but what of it? We must insist that it is quite insufficient, on the basis of Esther 3:7, the Mishnah and Josephus’ instances of obvious inclusive dating discussed earlier, to overrule the abundant evidence for a regular pattern of inclusive dating. Those give us a firm foundation for asserting that inclusive reckoning does apply, notwithstanding the denials of Filmer and Steinmann. Why, then, does Josephus use factual dating that includes the months here? Simply because in this particular context, he is concerned with pinpointing the time “Sosius and Herod besieged him [Antigonus], and took him,” leading directly to Herod becoming king. The focus here is on determining the time of the siege relative to Antigonus’ kingship, not the counting of his regnal years.

Is it not also interesting that in this passage, Josephus says Herod was “then”after the siege ended and Antigonus was slain at Antioch—”made King by the Romans”? Although Filmer and Steinmann are silent on this comment by Josephus, this remark tells us that Josephus recognized the pragmatic reality that, although the kingship was technically granted to Herod three years beforehand by the Romans, it had no practical force until Antony sent Sosius to help Herod depose and replace Antigonus. We should recognize this reality as well. It was the successful siege, removal, and putting to death of King Antigonus that made Herod king, not merely a ceremony in Rome three years earlier. This is also indicated by an easily overlooked detail in Antiquities 15.1.2:

Nor was there any end of the miseries he [Herod] brought upon them [the defeated Antigonus loyalists]: and this distress was in part occasioned by the covetousness of the prince regnant [Herod]; who was still in want of more…

Notice those easily overlooked words, “prince regnant”? They are fraught with meaning. A prince regnant/regent is defined as “a prince who rules during the minority, absence, or incapacity of a sovereign” ( The term was used here for Herod by Josephus for a specific reason—it recognizes the reality that Antigonus was in the hands of Antony at the time, but had not yet been put to death and was still regarded as the true sovereign of Judea. Because he was still alive, Josephus applies to Herod the precisely correct term “prince regnant” here (contra the Filmerian understanding that Herod had already been king for three years), until after Antigonus was dead: “When Antony had him brought to Antioch, and slain there, Herod was then made King by the Romans…(Ant. 20.10.1). Then, not before. Consistent with this, again recall Antiquities 17.8.1, “he [Herod] died…having reigned since he had procured Antigonus to be slain,” not “since the siege ended.” It is little details like this which show that Josephus did not intend to portray Herod as king going all the way back to his Roman appointment. Developing an entire narrative around that view is building a house on sand. In this connection, a detail Josephus adds in Antiquities 15.1.2 is worth pointing out:

And Strabo of Cappadocia attests to what I have said; when he thus speaks; “Antony ordered Antigonus the Jew to be brought to Antioch, and there to be beheaded. And this Antony seems to me to have been the very first man who beheaded a King; as supposing he could no other way bend the minds of the Jews, so as to receive Herod, whom he had made King in his stead. For by no torments could they be forced to call him King: so great a fondness they had for their former King. So he thought that this dishonourable death would diminish the value they had for Antigonus’s memory; and at the same time would diminish their hatred they bear to Herod.” Thus far Strabo (emphasis added).

Do you notice that Strabo views Antigonus not as a high priest but as king of the Jews, as we took pains to point out above? This offers independent confirmation from another ancient historian that we are on the right track in rejecting all attempts to minimize Antigonus’ status as king and instead elevate his supposed status as a high priest, all in an effort to justify expanding the “times of Herod” to 39 BC.

No, Herod’s “times” did not include the period when Antigonus was in office. They must therefore be reckoned from the start of his practical rule when the city was taken and King Antigonus was put to death. We must insist that the evidence, free from conjectures, begins the “times of Herod” in 37 BC, and all attempts to reconcile events tied to various years of Herod’s reign must be tied into that date via inclusive dating to avoid intractable problems. We must conclude that the Filmerian position is erroneous.

The High Priests and Aggregate Dating

Returning again to the high priests…we must not overlook the fact that in Antiquities 20.10.1, from the outset of Herod’s de facto kingship over Jerusalem and Judea until the fall of Jerusalem to Titus, only the aggregate totals of the high priests’ times in office are given:

Accordingly the number of the High Priests, from the days [or “times,” Loeb edition] of Herod, until the day when Titus took the temple, and the city, and burnt them, were in all twenty-eight. The time also that belonged to them was an hundred and seven years.

Aggregate dates were also the rule of thumb from the time of Aaron until Antiochus IV Epiphanes removed Onias, putting him to death, and installed Jacimus (also called Alcimus) into the office, of whom it is said that he “retained the High Priesthood three years” until his death began a seven-year interregnum. Thus, Josephus covers many centuries without listing a single high priest’s individual term in office, establishing this as his standard approach. That did not change until the Hasmoneans, who as Josephus puts it “had the government of the nation conferred upon them” and were already the de facto rulers of Judea, also took on the high priestly duties beginning with Jonathan Maccabeus.

The aggregate treatment of those who were nothing more than high priests thus stands in marked contrast to the specific reign lengths assigned to the Maccabean hybrid kings/high priests, most notably Hyrcanus II and Antigonus. This indicates those time periods applied to them in their roles as kings, not as high priests (besides, high priests do not “reign,” they serve in office!). Generalizing from the reign lengths of the Maccabean rulers to factual reckoning of high priest’s terms in office goes beyond the evidence.

In all events, with Antigonus’ death making Jerusalem both kingless and high priestless, Herod becomes king in Antigonus’ place. Up to this time Herod has had nothing whatsoever to do with appointing high priests, but now it is his job. Having just overseen the demise of a Hasmonean king/high priest, he was not much of a mind to put another Hasmonean in that position, so he began his custom of making “certain men to be [high priests] that were of no eminent families,” the first of which was Ananelus. (The lists at, and differ only in that Wikipedia also mentions that Ananelus, Joazar ben Boethus, and Jonathan ben Ananus had short periods of restoration, a detail ignored in Whiston’s margin note in Antiquities). VanderKam errs in including Antigonus in his list while inexplicably leaving off Jonathan (a different man than Jonathan ben Ananus), who is included in both the Whiston and Wikipedia lists (the latter assigning him to AD 58). By including Jonathan, the matching lists of Whiston and Wikipedia have the expected 28 high priests from Ananelus to Phannias, so there is no need to include Antigonus as VanderKam and Steinmann do. This removes any need to synchronize the start of Herod’s reign with the 39 BC (actually 40 BC) beginning of Antigonus’ term in office, and in turn eliminates the requirement Steinmann sees to base Herod’s reign on his Roman appointment. The interrelationships of various kinds of data these scholars have proposed DOES fit together—after all, they are clever!—but it is all based on a foundation of conjecture that requires jettisoning known dating conventions and tinkering with Josephus’ data.

Before moving on, here is an observation made by Kenneth Frank Doig in chapter 4 of his book, New Testament Chronology, posted online at, which effectively wraps up this topic:

In 63 BCE Pompey “restored the high priesthood to Hyrcanus,” (Ant. XIV 4:4) and “this Hyrcanus ruled, besides his first nine years, twenty-four years more...and the Parthians, passed over the Euphrates, and fought with Hyrcanus, and took him alive, and made Antigonus, the son of Aristobulus, king; and when he had reigned three years and three months, Sosius and Herod besieged him.” (Ant. XX 10:1) Twenty-four inclusive years from 63 BCE end in 40 BCE, the year the Parthians attacked Judea and established the rule of Antigonus. The three years and three months of Antigonus’ rule were, inclusively, 40, 39, 38 and the first three months from Nisan, or in Sivan of 37 BCE.

This straightforward solution is so simple, it is a wonder why any prefer the complexities of the Filmer/Steinmann/VanderKam synthesis in its place. Surely Occam’s Razor applies here: the simplest solution, having the fewest assumptions, is most likely the correct one.

A Closer Look at the Consular Years

Besides the support for the traditional date of 37 BC for the start of Herod’s reign given by Josephus’ list of 28 high priests (plus the kings Hyrcanus and Antigonus prior to the “times of Herod”), there are other factors to consider. Steinmann defends his 36 BC date by claiming that the consular years of Calvinus and Pollio at Herod’s Roman appointment, and of Agrippa and Gallus when he took the city after the siege, were in error by being one year too early. But to imagine that Josephus, with access to detailed official records and surrounded by many Romans who could have corrected any calendar errors obvious enough for Filmer and Steinmann to find, would mess up such a fundamental criterion as consular year dating, strains credibility. We observed previously ( that consular years were named after the “ordinary consuls” of a given year, and had nothing to do with when in the year they were named:

A consul elected to start the year—called a consul ordinarius (ordinary consul)—held more prestige than a suffect consul, partly because the year would be named for ordinary consuls (emphasis added).

Yet Steinmann claimed (p. 7),

The one hundred eighty-fourth Olympiad ended on June 30, 40 BCE. However, Calvinus and Pollio were not appointed consuls until after the Treaty of Brundisium on October 2, 40 BCE. Thus, Josephus is in error.

This reveals once more a preoccupation with factual, to-the-day date reckoning according to modern standards. Due to the civil war between Octavian and Antony, there were no consuls named for 40 BC until after the treaty was signed. Hence, Calvinus and Pollio were the ordinary consuls of record for that whole year, notwithstanding the fact that they did not take office until after October began. It is not Josephus who is in error here. The consular year named for Calvinus and Pollio indeed overlapped the latter half of the 184th Olympiad, exactly as Josephus wrote (Ant. 14.14.5, “And thus did this man receive the Kingdom; having obtained it on the hundred eighty fourth olympiad; when Caius Domitius Calvinus was consul the second time; and Caius Asinius Pollio [the first time]”). This was true regardless of the circumstance that their factual terms in office were delayed by war. Since he has other motives for wanting Herod to take office in 39 BC, Steinmann presumes to accuse Josephus of error, and therefore bumps Calvinus and Pollio from 40 BC to 39 BC only to defend his thesis. In the process he unabashedly disregards Josephus’ confirmation of the date by double-dating it to the 184th Olympiad, asserting Josephus really dropped the ball by committing a double error! How likely is that?

A similar accusation is leveled against Josephus concerning the consuls in the year Herod and Sosius took Jerusalem. Josephus unambiguously stated that the consular year of Agrippa and Gallus coincided with the 185th Olympiad (Ant. 14.16.4: “This destruction befel the city of Jerusalem when Marcus Agrippa and Caninius Gallus were consuls of Rome; on the hundred eighty and fifth olympiad; on the third month [i.e., Sivan, May/June, counting inclusively from Nisan]; on the solemnity of the fast”). Since Steinmann had bumped Calvinus and Pollio from 40 BC to 39 BC, but has to maintain the three-year time difference between Herod’s Roman appointment and the actual taking of Jerusalem, he is forced to also bump Agrippa and Gallus to 36 BC. This supposed error is also pure conjecture on Steinmann’s part, but a necessary ingredient in the complex interrelationships of events he and Filmer constructed. The weight of the internal evidence is against these machinations. Ronald Syme (The Augustan Aristocracy [Oxford: Clarendon Press, 1986], pp. 455-458), whose list of Roman consuls from 80 to 1 BC is reproduced at, had no hesitation in assigning the year of Calvinus and Pollio to 40 BC, and that of Agrippa and Gallus to 37 BC.

The odds are overwhelming: Filmer and Steinmann got this wrong.

Incidentally, Filmer (p. 285) also states the following:

The principal historical evidence for the date of Herod’s death is provided by Josephus in two statements in which he says that he reigned thirty-four years from the death of his predecessor Antigonus, but thirty-seven years from the time when he was declared king by the Romans. It is therefore important to establish precisely these two dates for his accession. Regarding his appointment in Rome Josephus says: “Thus did this man receive the kingdom, having obtained it on the hundred and eighty-fourth Olympiad, when Gaius Domitius Calvinus was consul the second time, and Caius Asinius Pollio the first time.” This makes it 40 B.C., but in this he is contradicted by Appian, who mentions Herod’s appointment in a context that can be dated from Dio’s Roman History to 39 B.C. (emphasis added).

Leaving aside the fact that Filmer sets himself up as an arbiter to pick winners and losers between ancient historians, do the records actually show that Appian “contradicts” Josephus? The citation of Appian, The Civil Wars 5.8.75, is the following:

He [Antony] set up kings here and there as he pleased, on condition of their paying a prescribed tribute: in Pontus, Darius, the son of Pharnaces and grandson of Mithridates; in Idumea and Samaria, Herod; in Pisidia, Amyntas; in a part of Cilicia, Polemon, and others in other countries (emphasis added).

Rather than Josephus making an error here as Filmer claims, the situation Appian describes has nothing to do with Judea at all. It simply reflects the situation before 40 BC, antedating Herod’s Roman appointment, when he was increasingly being recognized by Antony as an ambitious, capable leader utterly loyal to Rome, and as a result was given control over Idumea and Samaria. That Octavian, who collaborated with Antony in proclaiming Herod as king of Judea, is not mentioned here, is yet another reason to see what Appian reports as a different, earlier situation. We must likewise not overlook the fact that Judea is specifically left off Appian’s list. In short, we have no justification for drawing out, as Filmer has, an implication that his Roman appointment took place at that time. This is an example of something we should not do: reading our own conjectures onto the ancient records, with an agenda to nuance—or even misrepresent—them to support an outcome we favor. We should respect the ancient records as written by those much closer to the events, with only overwhelming evidence sufficient to make us question them.

“On the Solemnity of the Fast”

We have one more loose end to tie up. Earlier it was stated that “the fast” in Antiquities 14.16.4 should not be understood as the Day of Atonement on 10 Tishri, an assumption both Filmer and Steinmann make and use as support for dating Herod’s de facto reign from Tishri. In the interest of transparency, I must confess I also previously saw “the fast” as the Day of Atonement, though based not on Tishri dating but on what Josephus wrote in Wars 1.7.3–4 (

for it was in the third month of the siege before the Romans could even with great difficulty, overthrow one of the towers, and get into the temple (emphasis added).

In this place Josephus was apparently not talking about the fall of Jerusalem in the third month, Sivan, but in the third month of the siege itself. However, he states in Wars 1.18.2 that the total time covered by the siege was five months:

Indeed though they had so great an army lying round about them, they bore a siege of five months, till some of Herod’s chosen men ventured to get upon the wall, and fell into the city, as did Sosius’s centurions after them…

and that the siege began right after “the rigour of winter” was over (Ant. 14.15.14), “as the winter was going off” (Wars 1.17.8). These statements imply that Herod’s efforts to take the city began in the late winter month of Shebet (February of 37 BC). I suggest that if we want to be generous to Josephus and not presumptuously accuse him of error, in Wars 1.7.3–4 he was ignoring the two months (Shebet–Adar) of preliminary preparation prior to the arrival of Sosius and the Roman army. During that time Herod apparently occupied himself with setting up for the main effort later—cutting down trees, shipping in and setting up siegeworks from Tyre, etc., even taking time to get married. Then after Sosius arrived the assault on the walls really got underway, lasting the three months from Nisan to Sivan.

So, if “the fast” was not the Day of Atonement (Yom Kippur), what might it have been? As pointed out, it would be most natural for us to expect, since Josephus regards Nisan as the first month of the year and demonstrates that he consistently follows the normal Jewish/Roman convention of using inclusive dating, that the “third month” in which the “solemnity of the fast” took place—essentially, reprising the situation that unfolded during the 63 BC siege by Pompey, 27 inclusively-counted years previously—was the Jewish summer month of Sivan (June, 37 BC).

There is historical evidence that “the fast” in Antiquities 14.16.4 should not be equated with a defined Jewish fast date, but as the regular Sabbath. Granted that this is not the first thing most people (myself included) would think of, some solid evidence does point in that direction. Cassius Dio, who as a non-Jew may have misconstrued the no-work Sabbath as a “fast day,” called the day Pompey took the city 27 years earlier the “day of Saturn” (i.e., Saturday) in his Roman History (37.16.4):

Most of the city, to be sure, he [Pompey] took without any trouble, as he was received by the party of Hyrcanus; but the temple itself, which the other party had occupied, he captured only with difficulty. For it was on high ground and was fortified by a wall of its own, and if they had continued defending it on all days alike, he could not have got possession of it. As it was, they made an excavation on what are called [by the Romans] the days of Saturn, and by doing no work at all on those days [note the plural—it was not just a single special fast day] afforded the Romans an opportunity in this interval to batter down the wall. The latter, on learning of this superstitious awe of theirs, made no serious attempts the rest of the time, but on those days, when they came round in succession [i.e., once a week], assaulted most vigorously. Thus the defenders were captured on the day of Saturn, without making any defence, and all the wealth was plundered. The kingdom was given to Hyrcanus, and Aristobulus was carried away (emphasis and bracketed comments added).

The Jewish Encyclopedia article on “Pompey the Great” ( accepts this understanding, observing:

Pompey declared Aristobulus a prisoner and began to besiege the city. Although the party of Hyrcanus opened the gates to the Romans, the Temple mount, which was garrisoned by the people’s party, had to be taken by means of rams brought from Tyre; and it was stormed only after a siege of three months, and then on a Sabbath, when the Jews were not defending the walls. Josephus calls the day of the fall of Jerusalem “the day of the fast” (νηστείας ἡμέρα; “Ant.” xiv. 4, § 3); but in this he merely followed the phraseology of his Gentile sources, which regarded the Sabbath as a fast-day, according to the current Greco-Roman view. Dio Cassius says (xxxvii. 16) correctly that it was on a “Cronos day,” this term likewise denoting the Sabbath (emphasis added).

Josephus was a very knowledgeable Jew who would almost certainly have clarified that the fast in question was the very important Day of Atonement, if such it was. Since he did not, it implies this day was not the high holy day of Yom Kippur. Further, it supports the Jewish Encyclopedia’s suggestion that he was using pagan Roman sources to put together this part of his history, reflecting their inaccurate understanding of the Sabbath. Some have suggested that a little-known fast commemorating Jeroboam’s forbidding sacrifices in Jerusalem on Sivan 22 might be a possibility (see, note 14), but I think this is a “hail Mary” suggestion of little likelihood. Dio 37.16.4 gives us an entirely satisfactory answer. Thus, we can understand Antiquities 14.16.4 as telling us that Pompey and Herod/Sosius both took Jerusalem 27 years apart, in the same summer month of Sivan, on a Sabbath day when the Jews’ religious scruples would not let them oppose construction of the siegeworks.

Looking Ahead to Herod’s Death

I think we have now said all that is necessary about the start of the reign of Herod the Great. Many different details point to its beginning in the summer of 37 BC, in contrast to the Filmer/Steinmann approach, which requires ignoring some details, interpreting others in non-straightforward ways, and proposing errors by Josephus even for events he double-dated. We have not comprehensively looked at every single point that Josephus’ modern-day critics attack him for, but enough of them—and I hope in convincing fashion—to set aside their 39 BC proposed start date for Herod’s reign, together with the idea of non-inclusive, accession-year dating from Tishri that underlies it. The conclusion for today is this: Herod became king of the Jews in the summer of 37 BC following the death of Antigonus.

In the next phase of this study, I plan to examine the other end of Herod’s life—his death and the events leading up to it. That discussion will regard the start of Herod’s reign in 37 BC as established, and build upon it.

Research RSS Feed

AddThis Feed Button

Recent Articles

A message from ABR director, Scott Lanser.
This article gives some specific rebuttals to the contention that Herod’s reign must be dated from his...
Join co-host Scott Lanser as he interviews special guest Pastor Bryan Windle, as shares the top 10 archaeological...
Many biblical scholars who interpret the genealogies of Genesis 5 and 11 as yielding a continuous chronology...
Associates for Biblical Research
  • PO Box 144, Akron, PA 17501
  • Phone: +1 717-859-3443
  • Toll Free: 1-800-430-0008
Friend ABR on Join us on Twitter Join us on Twitter