(avg. read time: 18–36 mins.)
References to King Herod when Jesus is born, (Matt 2:1–8, 19–21; Luke 1:5) and to his successor in Judea, Archelaus (Matt 2:22)
No evidence is more crucial for determining the year of Jesus’s birth than the references to Herod the Great. Given the significance of this evidence, and the many details that we need to sort through, we will be spending a disproportionate amount of this analysis here. It was, after all, these references to Herod, and to his death soon after, that led scholars to reassigning Jesus’s birth from 1 BCE to between 7 and 4 BCE.1 As this reassignment is based primarily on the work of Josephus, it was not based on new evidence that came to scholarly attention in the modern era. Josephus was well known, particularly among Christian scholars, as around two dozen Christian writers were already citing him in the second through fifth centuries.2 Rather, scholars gave more focused attention to Josephus’s chronological information and its correlations to establish dates of key events, including Herod’s death. As there is a lot of conflicting data to sort through, this will be the longest and most complicated entry in this series. But each entry will also come with a summary of the basic inference we can make based on the evidence.
Emil Schürer was the founder of the present consensus that dates Herod’s death to 4 BCE.3 The arguments he makes, and the augmentations provided by others such as Timothy Barnes and P. M. Bernegger, are as follows:
1) Herod died 37 years after his appointment as king in Rome (Ant. 17.191; J.W. 1.665). Josephus dates this appointment (incorrectly) to the 184th Olympiad, in the year when Gnaeus Domitius Calvinus was consul the second time and Gaius Asinius Pollio was consul the first time (Ant. 14.389), which Schürer and subsequent scholars determined to be equivalent to 40 BCE.
2) He died 34 years after the execution of Antigonus (i.e., after he had retaken Jerusalem from the Parthians; Ant. 17.191; J.W. 1.665). Josephus dates this event to the 185th Olympiad when Marcus Agrippa and Caninius Gallus were consuls, 27 years (to the day, in the case of the retaking of Jerusalem) after Pompey conquered Jerusalem (63 BCE), and 126 years after the Hasmonean government was first established (Ant. 14.487–490; 162 BCE per 1 Macc 6:20, 48–60), which Schürer and subsequent scholars determined to be equivalent to 37 BCE.4
3) You may notice that the math does not quite add up in the previous two calculations, as the numbers all seem to lead to a date of either 3 BCE (40 – 37; 37 – 34) or 2 BCE (63 – 27 – 34; 162 – 126 – 34). But the Schürer consensus argues that this is because Josephus used inclusive reckoning.5 As other examples, they cite the year and earthquake mentioned in Ant. 15.121 // J.W. 1.370 (the Battle of Actium [31 BCE] occurred in the seventh year of Herod’s reign [ἑβδόμου δ’ ὄντος ... ἔτους]), 18.26 (Quirinius completed his census [presumably in 6 CE] 37 years after Actium [τριακοστῷ καί ἑβδόμῳ ἔτει μετὰ τὴν Ἀντωνίου ἐν Ἀκτίῳ]), and 20.250 (107 years from the reign of Herod to the destruction of Jerusalem by Titus [70 CE]). Alternatively, Hoehner argues that Josephus uses a non-accession year system for dating reigns, wherein any part of the first year prior to the next New Year’s Day counts as one year, and subsequent years are counted beginning with the New Year’s Day.6
4) Josephus tells us that Herod died not long after a lunar eclipse that was visible at night (Ant. 17.167). This is the only eclipse referenced in Josephus, but there were a few that were visible at night from Jericho (where Herod spent his last days, per Josephus, Ant. 17.155–199) prior to the turn of the eras: March 22/23, 5 BCE (total); September 15/16, 5 BCE (total); March 12/13, 4 BCE (partial); January 9/10, 1 BCE (total); and December 28/29, 1 BCE (partial). A few scholars who otherwise agree with the general consensus chronology of Herod’s reign have suggested that the eclipse was the one September 15/16, 5 BCE.7 But the majority have identified this eclipse with the one on March 12/13, 4 BCE.8 Further confirmation for this being the proper eclipse is that a Passover followed not long after Herod’s death (17.213), which scholars typically identify with the one on April 11, 4 BCE.
5) The chronologies of his sons’ reigns further support a date for Herod’s death in early 4 BCE. When Archelaus was deposed from the throne of Judea and banished in 6 CE, Josephus says that he was in the 10th year of his reign (Ant. 17.342; cf. Life 5; though J.W. 2.111 says “9th”). Antipas was tetrarch of Galilee and Peraea until the second year of Gaius Caligula’s reign (38–39 CE; Ant. 18.238, 252, 256), although Ant. 19.351 states that this happened in the 4th year of his reign (40 CE), and coins of his dominion have survived that were minted in “year 43.”9 Philip—tetrarch of Trachonitis, Gaulanitis, and Batanea—likewise had coins made until year 37 of his reign, and Josephus says he died in the 20th year of Tiberius’s reign (33–34 CE; Ant. 18.106).10
6) Philip in particular engaged in projects that would seem to confirm a date of his father’s death in 4 BCE at the latest. He refounded the cities of Julias/Bethsaida and Paneas/Caesarea Philippi, the former of which Josephus claims was named for Augustus’s daughter (Ant. 18.28). As such, this would indicate that the city’s refounding was before her disgrace in 2 BCE. Likewise, the latter city used a dating system that presumably linked to its founding in 3 or 2 BCE as “year 1.”11
One point of ambiguity that has been used both for and against this date is the matter of Herod’s age. Josephus says that he was “about” (περί; Ant. 17.148) or “near” (σχεδόν; J.W. 1.647) 70 years old when he became ill with the condition that eventually killed him. The latter expression could mean that Herod was a little short of 70 before he became ill, while the former expression could mean he was approximately 70, either older or younger by some unclear measure (and in this text Josephus is more clearly reliant on Nicolaus of Damascus, his main source on Herod). Of course, such statements about Herod’s age create a contradiction with Josephus’s statement about Herod being 15 (πεντεκαίδεκα, which could be written in shorthand as ιε’) when his father Antipater appointed him στρατηγός of Galilee (Ant. 14.158) in the ninth year of the high priesthood of Hyrcanus II (47 BCE; Ant. 14.148). The typical solution is to posit that Josephus originally wrote “25” (εἴκοσιπέντε, which could be written in shorthand as κε’). However, as Nadav Sharon has argued, such an emendation—in addition to having zero manuscript support—misses the point of Josephus in emphasizing Herod’s youth and inexperience, while describing his ability regardless (Ant. 14.159).12 It would be unusual for a man of 25 to take such a position, but not as peculiar as scholars think Josephus’s comment would imply, given that Alexander the Great took the throne at 20 and conquered most of the Persian Empire within a few years thereafter, Scipio Africanus was said to be 24 when he took command in Spain during the Second Punic War (Livy, Hist. 26.18), and Octavian (later, Augustus) was consul before he turned 20.13 But given how Herod being 15 in this time does not comport with the other chronological details of Josephus’s account, it would not work to take him literally either, and he has likely taken over an embellishment from his source, Nicolaus of Damascus, to portray Herod as a Wunderkind. Herod may well have actually been 25 at the time, but such cannot be determined purely on the textual basis. It is reasonable to conclude that he was in fact a young man in 47 BCE, but we have no independent basis for determining his actual age at the time. That determination would need to be made on the basis of other calculations.14
In any case, some scholars, more frequently in recent years, have challenged the prevailing interpretation of the aforementioned evidence with the following arguments:
1) Bieke Mahieu observes that inclusive reckoning does not sufficiently account for discrepancies in the consensus chronology: “The spring of 4 BC is about 35⅓ years (instead of 37) after Herod’s appointment in Rome (on 18 November 40 BC) and about 32½ years (instead of 34) after the conventional dating of the capture of Jerusalem (in July or October 37 BC).”15 But that Josephus used inclusive reckoning for Herod is debatable. In the first example cited, Josephus does seem to use inclusive reckoning for an event, but the noted earthquake was in March/April of 30 BCE, about half a year after Actium, which would imply that Herod retook Jerusalem in 36 BCE (specifically, in March).16 Such a date would also comport with a non-inclusive reckoning of 27 years from the conquest of Jerusalem by Pompey in 63 BCE (Ant. 14.488: ἐκείνου τῇ αὐτῇ ... ἡμέρᾳ μετὰ ἔτη εἰκοσιεπτά) and 126 years from the beginning of the Hasmonean government in 162 BCE (Ant. 14.490: μετὰ ἔτη ἑκατὸν εἰκοσιέξ).17 Indeed, the grammar of these markers, as well as of Ant. 18.26 noted above, with their usage of μετά + accusative (with the sense of “after” in temporal clauses), imply non-inclusive reckoning.18 The last example of Ant. 20.250 does seem to be a case of inclusive reckoning, but this example is complicated by the fact that Josephus is providing a sum for the reigns of 28 high priests from the time of Herod to the destruction of Jerusalem, where he otherwise shows a tendency towards non-inclusive reckoning when he provides the years of their tenures. In fact, his accounts of multiple reigns and tenures on either side of Herod tell against this interpretation. He states that Hyrcanus II was high priest for 24 years following the conquest by Pompey and that Antigonus after him held the office for 3 years and 3 months (Ant. 20.245–246). Likewise, his description of the reign of Herod Agrippa I is non-inclusive, as he says he ruled under Claudius for 3 years, when it was in fact 3 years plus a few months (Ant. 19.351; J.W. 2.219).19
2) The date of Herod’s initial appointment by Antony can be reliably dated to 40 BCE, in the second year from Antony and Cleopatra meeting in 41 BCE (Josephus, Ant. 14.324, 330) and two years after the Battle of Philippi in 42 BCE (J.W. 1.242, 248; if this is not an inclusive reckoning) , despite the attempt by W. E. Filmer to push the appointment to 39 BCE and his first regnal year—following an accession year system in which the first year begins on the New Year’s Day following accession—to Nisan 1 or Tishri 1, 38 BCE.20 While Josephus is mistaken to date this appointment to the 184th Olympiad (which ended earlier in 40 BCE), it makes more sense if he makes the mistake because the last year of this Olympiad was when Herod departed, than if he was off on the appointment by a whole year and the first regnal year by two years. Josephus makes chronological mistakes, as we will note in more detail below, but it is difficult to explain how exactly such an error as Filmer’s thesis needs might have happened. On the other hand, the date of Herod’s capture of Jerusalem could be anywhere from late 37 BCE to late 36 BCE, as the Sabbatical year that Josephus references (Ant. 14.475; 15.7) ran from Tishri 37 BCE to Tishri 36 BCE.21 This data could thus support either view, but given the scope of events surrounding the capture, it is more likely that the siege began in 37 BCE and ended in the third (Roman) month of March the next year.22 Furthermore, the Sabbatical year scenario reflects the siege during a Sabbatical year in 162 BCE 126 years previously (1 Macc 6:49).23
3) The eclipse in March of 4 BCE is unlikely to have been the eclipse Josephus noted. Mahieu thinks that such a date would imply that Herod executed certain rebels (who had resisted his placement of a golden eagle over the great gate of the temple) during Purim of that year.24 Of all the eclipses observed during this time, it covered the smallest portion of the moon. But more importantly, the implication of a gap of less than a month between the eclipse and the next Passover does not leave sufficient room for all the events Josephus describes to have happened. Josephus outlines the following events between the eclipse and the Passover (Ant. 17.168–218):
Herod’s condition worsened and he tried a number of physician recommendations to remedy the situation.
These recommendations included going to the hot baths of Callirrhoe.
After the baths failed to help, Herod returned to Jericho and was bathed in a vessel full of oil.
When Herod acknowledged that there was no cure for his condition, he had letters composed and sent messengers to gather Jewish elders from across his kingdom.
When these elders came to him at Jericho, he had them confined in the hippodrome so that they would be executed when he died, ensuring that the nation would mourn his death (if his officials had actually followed his commands).
Herod received letters from Augustus permitting him to do what he wished with his son Antipater (who had, among many other treacherous acts, attempted to murder Herod), and Herod thereupon had Antipater executed.
Five days after (J.W. 1.665) or on the fifth day after (Ant. 17.191) this execution, Herod himself died, after taking steps to compose a final will and to distribute property to those he considered faithful.25
Herod had wished to be buried at Herodium—200 stadia (a little less than 23 miles) from Jericho—and so arrangements were made for a funeral procession that would set out from Jericho, involving dressing his body in royal regalia, assembling soldiers throughout his kingdom, and 500 servants carrying spices for his body.
Archelaus mourned for his father until the seventh day.
A feast followed this mourning period.
Before the Passover, Archelaus made a number of promotions in the military, liberated some prisoners, promised to ease tax burdens, and did many other things in response to calls from the crowds.
But alongside those seeking to curry his favor and to relieve their burdens was a movement that sought rebellion, which came to fruition at the Passover, and which was itself but the first of the rebellions launched against Archelaus, which followed at Pentecost (Ant. 17.250–298, 324–338, 342–344).
Andrew Steinmann has argued that the shortest span there could have been between the eclipse and the Passover in order to accommodate all these events was 41–62 days.26 Mahieu has further argued that:
at least nine weeks separate the end of the mourning and Pentecost: one week for Ptolemy’s journey, one week for the preparation of the Roman legion, three weeks for the journey to Caesarea, two weeks for the march to Jerusalem and the subsequent military intervention, and two weeks for Sabinus’s occupation. Given that Passover falls seven weeks before Pentecost, the mourning must have ended at least two weeks before Passover and not as late as 8 Nisan.27
Ernest Martin estimated that the events in the list above took place over a period of 10 weeks. However, unlike Steinmann, he bases this calculation largely on the assumption that when Josephus refers to the funeral procession going 8 stades/furlongs towards Herodium (Ant. 17.199), he is implying that this was a daily rate of travel, meaning that the procession would have taken 25 days.28
4) As for the eclipse itself, two candidates are left as the most likely alternatives (as the eclipses of 5 BCE produce the problem of an even more severe restriction on the time window, in the case of the March eclipse, or allowing far too much time, in the case of the September eclipse). Some scholars propose the eclipse of January 9/10, 1 BCE.29 Others favor December 28/29, 1 BCE.30 Against the former is the fact that it was most apparent after midnight while the latter was visible in the early evening. However, contra Mahieu, Josephus’s narrative does not assume that many people observed it; its significance derives primarily from its synchronicity with the execution of Matthias, Judas, and their fellows.31 Either would leave more than sufficient time until the next Passover for all the listed events to unfold.
5) According to Augustus’s Res Gestae 16, he was demobilizing and settling his armies from 7–2 BCE. It would thus seem unlikely that Augustus would continue this policy unabated if the Pentecost War for which Quintilius Varus was dispatched occurred in 4 BCE a couple months after Herod’s death. As such, this would indicate that Herod’s death happened after 2 BCE.32 Furthermore, Martin, arguing from silence, claims that such a victory beyond the official boundaries of his empire would surely necessitate an imperial acclamation. Augustus received one in 8 BCE, but did not receive another until 1 CE.33
6) In order to address the difficulties created for the revisionist chronology by the data of the reigns of Herod’s sons, some revisionists propose that his sons were co-regents with Herod beginning roughly 4 BCE.34 They find precedent for this in Herod’s son Antipater, who Herod had initially named as his heir. At Antipater’s trial for plotting against his father, both Herod and Antipater describe his status in terms that make him sound like a de facto co-regent (J.W. 1.625, 631–632), not merely one promised royal authority, but one who has already acted with royal authority. As Archelaus and Antipas were both named heirs before Herod’s death, even as he went back and forth on his decision, these revisionists suppose that they too reckoned their reigns as beginning in 4 BCE when, presumably, they were given heir and co-regent status.
7) While the chronology of the reigns of Herod’s sons is taken as possibly the strongest evidence in favor of Herod’s death taking place in 4 BCE, given how the dating for each of these three reigns converges at that point, there remain some important problems for the conventional chronology. As noted above, Josephus gives conflicting accounts of when Archelaus’s reign ended, whether in his 9th or 10th year. He does know that his reign ended in the 37th year of Actium (Ant. 18.26; cf. Cassius Dio, Rom. Hist. 55.25.1; 55.27.6), but he seems less clear about when it began and so attempts to calculate it. As a support for this calculation, he cites a dream from Archelaus in which he saw 9 (J.W. 2.112–113) or 10 (Ant. 17.345–347) ears of grain that were devoured by oxen, which Simon the Essene interpreted as the number of years in Archelaus’s reign, the year in question being either the 9th or 10th of that reign (Archelaus would be deposed shortly thereafter). This is obviously linked to Pharaoh’s dream in Gen 41, but Josephus’s source for the dream obviously did not supply either of the numbers he uses, as they were his own attempts to keep his text consistent. In the Genesis text, the number of ears of grain (as well as oxen) are two groups of 7 and the dream source may likewise have revolved around the use of 7 (although Josephus saw a need to “correct” it in two different ways). If this is so, Mahieu proposes that the dream would be referring to the current Sabbatical cycle, which Tishri 1, 6 CE/Tishri 1, 7 CE was the end of.35 This would accommodate her proposal that Herod died in 1 CE and that Archelaus’s reign still ended in 6 CE. Alternatively, if Herod died in 1 BCE, the presumed group of seven could refer to the years of Archelaus’s reign itself and its fitting end in a Sabbatical year.
We have seen already that Josephus provides conflicting information on when Antipas’s reign ended, either in 38/39 CE or late 40 CE (as it was not until then that Caligula returned from expeditions in Gaul and Germany). If the year 43 coins then refer to the years of Antipas’s reign, the conventional chronology would demand that year 1 would cover late March/early April, 4 BCE to late March/early April, 3 BCE and year 43 would then cover late March/early April, 39 CE to late March/early April, 40 CE. However, the conventional chronology cannot explain why there are so many differences between the basic styles of the coins of year 43 and the coins for years prior (for Antipas, we also have coins for years 24, 33, 34, and 37).36 The only other coins from this area that are similarly dedicated to Caligula are the year 5 coins of Herod Agrippa I and they may well have been issued the same year (41/42 CE) under the same regency of Agrippa after Caligula’s death. Despite the damnatio memoriae instituted against Caligula, his successor Claudius did not enforce such (Suetonius, Claud. 11.3; Cassius Dio, Rom. Hist. 60.4.5) and it is understandable why he would allow Agrippa to honor his predecessor so, as the two of them were close friends. The coins from Tiberias may have thus been doubly posthumous, honoring Caligula and Antipas, the founder of the city now referred to with the nominative rather than the genitive (being no longer in possession of the tetrarchy).37
As for Philip, David W. Beyer has presented by far the most extensive text-critical analysis to demonstrate the flaws in the conventional chronology.38 Between the British Library and the Library of Congress, he reviewed 55 editions of Josephus published prior to 1700 to see how each of them rendered Ant. 18.106. Of these, 32 (all but 3 of which were published before 1544) stated that Philip died in the 22nd year (εἰκοστῷδεύτερῳ*, vicesimo secundo) of Tiberius.39 In fact, not a single edition published prior to 1544 printed the 20th year (εἰκοστῷ, vicesimo) in this text. Given the conflicting witness and general scribal tendencies, it is more likely that a number was omitted during transmission than that it was added, especially if the years were written in shorthand (κ vs. κβ). In contrast to the previous two times I rejected this proposal, there is actual manuscript evidence for such a number variation here. The textual picture is further complicated by the variety of combinations the editions have of the year of Tiberius’s reign and how long Philip had reigned. Among the manuscripts that attest his death in Tiberius’s 22nd year (35–36 CE), there are editions that claim that he had a 32-year reign, a 35-year reign, or a 22-year reign (this latter reading is the only “22nd year” reading that is featured in editions beyond 1544; the 37-year reign did not emerge until it appeared alongside the “20th year” reading in 1544 and thereafter).40 The 22-year reign is obviously a duplication error and it has no historical plausibility. The 32-year and 35-year readings could easily be mistaken for each other if written in shorthand (λβ vs. λε; capitals: ΛΒ vs. ΛΕ). Both alternatives have their advocates, with Beyer arguing that the earliest attested reading of 32 years is what Josephus wrote and that it fits with Philip being declared tetrarch in 4 CE, and with Mahieu arguing that 35 years fits with her reconstruction of Philip beginning to reign after Herod’s death in 1 CE.41 Philip’s year 37 coins would thus either need to attest to the year of Herod’s death (Beyer) or to an era Herod established (Mahieu).42
8) Mahieu has argued that various coins and inscriptions suggest that Herod established a new era based not on his regnal years, but on an event sometime in 2 BCE. Coins from Paneas/Caesarea Philippi inscribed with the year 220 refer to two figures—the emperor Macrinus and his son Diadumenian—who would have been celebrated no later than late 217 CE to mid-218 CE. When this is combined with an inscription from Quneitra (from Philip’s dominion) that is marked with the year 463, which dates from January 462 CE, the data entails that Year 1 of this era must have encompassed early 2 BCE and the beginning of 1 BCE.43 Such a date range would correspond to a highly significant event for Herod’s liege, as Augustus was granted the title of pater patriae (“father of the fatherland”) in February of 2 BCE (Res Gestae 35).44
9) If Mahieu’s suggested era is accurate, that would mean that the Paneas coins do not so much signify the year since its founding, but rather the year of the era for its minted coins. Julias/Bethsaida is another story. Josephus provides two conflicting accounts of when the city was founded. In Antiquities, he actually describes it between his references to Quirinius and Coponius (18.26–28), which, by his chronological framing, would be about 8–10 years after when scholars tend to date its founding. In War, he clearly links Philip’s activity of naming the city with the reign of Tiberius, son of Julia, the only Julia named in the text (2.167–168). The Armenian translation of Eusebius (Chron.) links it to the last year of the 200th Olympiad (24/25 CE) while Jerome’s translation (Chron.) in his update of Eusebius links Philip’s founding of the city to the first year of the 201st Olympiad (25/26 CE). Both of these sources are closer to the War account and they receive further support from Philip’s coinage. In year 34 of Philip’s coins, we see him honor Julia Augustus (in Greek: Sebaste) with her image on his coin and with the reverse linking her to the terminology and iconography of the goddess Demeter. If year 34 is linked with 32/33 CE, this would further make sense of coins dedicated to Julia, as January, 33 CE would mark the 90th anniversary of her birth and the 70th anniversary of her marriage to Augustus.45 With the independent evidence on the possible date and Philip’s reverence of Julia/Livia, it is preferable to follow the chronology of the War account.
10) As further confirmation, Mahieu points to the necropolis at Marissa (modern Maresha), which has tombs that plausibly date from Herod’s day. Most interestingly, this necropolis features a sepulchre with an inscription of one Pheroras who died in year 2. Given the location (it was in what was called Idumea), the origins of Herod’s family in Idumea, the characteristics of the tombs, and the rarity of the name Pheroras (there are hardly any records with that name outside of Josephus), it is probable that this Pheroras was Herod’s own brother.46 According to Josephus (Ant. 17.58–60), Pheroras was buried in Jerusalem, which means that this tomb, if it does belong to this Pheroras, could be the site of his secondary burial, as the Jews practiced in this time, and the year could derive from either phase of his burial. The “year 2” inscribed on his tomb is likely part of an era, not a reference to regnal years, since no name is attached to “year 2” or to a few other inscriptions that date between “year 1” and “year 5” using this unknown dating system. Since we know that Pheroras died not long before Antipater’s trial and the events that followed thereafter, it is clear that this “year 2” cannot refer to any regnal year of Herod’s. If it refers to Mahieu’s proposed “pater patriae era” of Herod’s own invention, this would place Pheroras’s death in either 2/1 BCE or 1 BCE/1 CE, depending on whether “year 2” applies to the primary or secondary burial.
11) Herod’s “Year 3” coins likely supply independent confirmation that Herod did in fact create at least one other era and placed dates for an era other than his regnal years on his coins. David M. Jacobson has persuasively argued that the mysterious year 3 coins with their tau-rho (TP) monogram most likely commemorate an era celebrating either Herod’s re-confirmation as king under Augustus (beginning spring 30 BCE) or Herod’s acquisition of Trachonitis (beginning 23/22 BCE).47 The coins are unlikely to refer to Herod’s third regnal year, as, even by the conventional chronology, Herod’s realm would have been in too much turmoil and reestablishment to produce such high-quality coins.48 Furthermore, it is difficult to explain why undated coins from Herod were of lower quality if they came years after Herod consolidated his power (they could not have come before) and his kingdom enjoyed much greater prosperity than in 37 or 36 BCE. On the other hand, this is one piece of data that Mahieu cannot convincingly explain as to why the TP monogram should stand for the Latin pater patriae and why, if the coins were issued by Varus in the third year of such an era, they were issued in the dead Herod’s name and not his own.49 However, she is correct in noting the Augustan (and, in some cases, Herodian) iconography of the coins, and the iconography of these coins is all the more apropos considering that they were issued from Samaria-Sebaste, a city re-founded and renamed with the Greek translation of Caesar’s honorific “Augustus.”50 Herod would hardly have been making Augustan coins while Antony was still alive, not least since Antony is the one who put him in power in the first place. But this iconography is much more explicable after Octavian defeated Antony at Actium in 31 BCE, as well as after he was given the honorific “Augustus” in January of 27 BCE. The coinciding of the latter with a third year from when Augustus reconfirmed Herod’s kingship through ratification by the Senate (J.W. 1.386–393; Ant. 15.183–196) makes this period all the more likely as the referent of the year 3 coins. Conversely, the possession of Trachonitis (J.W. 1.400; Ant. 15.359–360), over which Herod was tetrarch in addition to his already established kingship, was a gift of Augustus.51
12) Most significantly, the conventional chronology is based on Josephus’s chronology, but his chronology seems to feature some significant holes and flaws. We have already noted a few cases concerning the 184th Olympiad as the beginning of Herod’s reign, the end of Antipas’s reign, and Philip’s work with Bethsaida/Julias (as well as, arguably, the consular date for Herod’s capture of Jerusalem), but a few more inconsistencies are pertinent.52 First, if Herod’s death was in 1 CE, as Mahieu argues (though I am inclined to think 1 BCE), his reign would have lasted for thirty more years after Menahem the Essene prophesied to him in 30 BCE that he would reign for “twenty, nay, thirty years” (Josephus, Ant. 15.372–379), which would explain why the prophecy was remembered for so long thereafter. But Josephus’s chronology does not agree with this prophecy and so he had to explain the discrepancy as not actually giving the definite limit for Herod’s reign (15.378). Second, while Josephus provides consular dates for the beginning of Herod’s reign and for his capture of Jerusalem, he does not provide any such information for the year of Herod’s death. In that case, he is relying on either personal calculation work, another source than Nicolaus of Damascus, or both to arrive at the figures he gives for the year Herod died. The 34-year figure probably comes from the Testament/Assumption of Moses or a source with a similar chronology. Chapter 6 of this Testament refers indirectly—in typical apocalyptic fashion—to Herod as the instrument of wrath against the kings who were called the priests of the Most High God (i.e., the Hasmoneans). The reference to 34 years (6:6) is not to Herod’s reign, but follows the statement that he will shatter these leaders and kill both old and young. The timespan rather refers to the complete divestment of the Hasmoneans with Herod’s command for Aristobulus III (still popular with the general populace: Josephus, J. W. 1.437; Ant. 15.30, 32, 52) to have a “drowning accident” in a shallow pool when the latter was 18 in 34 BCE (Ant. 15.56). Josephus himself recognized this as the end of the Hasmonean priesthood (Ant. 20.247–249), although Hyrcanus II, who had already been deposed as high priest, was executed later at 81 in 31/30 BCE (Ant. 15.178).53 Independent confirmation for a similar chronology appears in rabbinic literature in S. ‘Olam Rab. 30 and b. ‘Abod. Zar. 9a, where the sources state that the Herodian era covered 103 years to the destruction of the temple in 70 CE.54
As the reader can see, there is much information to sort through in order to reconstruct a coherent chronology and determine what year Herod died, and thus when Jesus possibly could have been born. One thing that seems clear from the outset is that, if Josephus’s description of Herod’s last days is generally accurate, then the possibility that Jesus was born on December 25, 1 BCE, four days before the eclipse noted above, this would be exceedingly difficult, if not practically impossible, to reconcile with the information in Matthew and Luke. Herod would be unlikely to be in good condition to meet the magi, unless perhaps they showed up in a short enough window of time to rule out the events reported by Luke of Jesus’s circumcision and presentation at the temple (eight and forty days afterwards, respectively).
Otherwise, some of these arguments against the consensus chronology simply do not work. The supposition of a co-regency of Herod with his sons (it would need to apply to at least two of them) is ad hoc and there is significant evidence against it.55 It seems rather odd to assume, if Antipater did take his heir status as a co-regency status, that Archelaus and Antipas so quickly reinforced the way of thinking that led Antipater to his execution, and this while Antipater would still be alive for some years (or are we to suppose they repeated their brothers’ presumptions one more time over). More likely, precisely because of the trial, they remembered that heirs are not co-regents. Augustus had given Herod a directive that he could only appoint for his son(s) to take over the kingdom after him (and not while he was alive; Josephus, Ant. 16.127–129) and the speech of Herod that he was the sole ruler for as long as he lived (Ant. 16.130–135; J.W. 1.457–466) make clear that Herod was to have no co-regents. The rhetoric from Herod and Antipater in J.W. 1.625, 631–632 is highly stylized as speeches made by Josephus (or his source) to reflect the characters of Herod and Antipater, hence why Herod could say that he almost, or in a manner of speaking (depending on the translation), resigned his royal power to Antipater (as a way to deny directly that he violated Augustus’s directive), and Antipater could defend himself from charges of acting treacherously because he “reigned already.” Antipater is not so much confirming that he was co-regent as he is attempting to refute his father’s claim that, though he was heir (1.623), he sought to become king by patricide (1.624). Likewise, Josephus’s roundabout description of Antipater’s status in Ant. 17.3 is not a designation of him as an official co-regent, but a description of the honors Herod gave him, despite his lack of popularity (Ant. 17.1–2). Likewise, Nicolaus of Damascus stated Antipater’s status as successor along with noting that he had been given the opportunity to enjoy a taste of authority in the present that was promised him in the future, not that he had been made co-regent (Ant. 17.115). This is consistent with Josephus quoting Herod as saying that he would give his sons royal honors, but not the kingdom itself (J.W. 1.461). In the summary of the same speech in Antiquities (16.132–134), there is no similar qualification, as Herod simply says that his sons would reign after him and that he alone should be esteemed as king and lord.
The calculations of Steinmann and Mahieu do indicate some significant problems that the consensus chronology must overcome, but Martin’s has a fatal flaw. Nowhere in the Greek text of Ant. 17.199 does Josephus indicate that the 8 stades/furlongs was a daily rate of travel. The idea that Herod’s funeral procession spent twenty-five days going from one city to the next is patently absurd, given the number of difficulties it creates. Why would they only walk about a mile per day and then spend the rest just sitting under the sun? How did they supply this massive procession for that long and why did they think this was a good idea? In a politically and religiously sensitive time just after Herod’s reign, why would this procession so flagrantly and extensively violate Jewish law and custom concerning the prompt burial of the dead?56 More likely, Josephus’s statement simply signifies that the entire procession accompanied Herod’s corpse for this long, until he was then transported more expeditiously.57
The argument appealing to Augustus’s demobilizing of his armies is less clearly indicative than the revisionists claim. The fact that Josephus notes the invasion of Varus in response to the revolt alongside the Maccabean War, the invasion of Pompey, and the First Jewish War against Rome (Ag. Ap. 1.34) is not necessarily indicative of scale so much as it is a note of a non-civil war within the bounds of (primarily) Judea. This revolt also took fewer troops and a shorter time to resolve than any of the other events (Josephus, Ant. 17.250–298). Even if Augustus was disbanding most of his army during this time—the Res Gestae makes no clear statement on extent, since it is mostly about disbursement of funds—this revolt on the outer edges of his empire need not have disturbed his overall policy, especially if he could trust a commander like Varus to take care of it.58
The chronological difficulties surrounding when Herod conquered Jerusalem remain problematic for the Schürer consensus. Hoehner has attempted a critique of the argument that Herod captured Jerusalem from the Parthians in 36 BCE by pointing to Cassius Dio, Rom. Hist. 49.24–31. Unlike Josephus, Cassius does not directly assign a year to this conquest, but the indication he gives is that in 36 BCE, Antony could not have spared Sossius and such a large number of his 100,000 troops that he led on a campaign against Parthia in the spring.59 However, if the capture happened in early March, this would not present a clear discrepancy with Cassius. Mahieu has plausibly argued that Josephus misunderstood his sources by assigning the consular year to the end of the siege rather than to its beginning, as he likewise made the mistake of assigning the 184th Olympiad to the beginning of Herod’s reign, rather than to his departure for Rome to be appointed as king.60 Furthermore, if Herod had taken Jerusalem between July and October in 37 BCE—per the consensus chronology—Antony was in Italy, rather than in Antioch, where he executed Antigonus (Josephus, Ant. 15.8–10; Plutarch, Ant. 36.1–2; Cassius Dio, Rom. Hist. 23.1). He was in Antioch during the winter of 37/36 BCE and it was from this region that he launched his campaign against the Parthians the following spring.
More importantly, probably the strongest piece of evidence against the conventional chronology is the timeline established by Steinmann and Mahieu. There is simply not enough time for all the events to unfold between the eclipse of 4 BCE and the following Passover (per Steinmann), whenever in March or April it could have fallen, or even the following Pentecost (per Mahieu).
While at first glance, the chronology of Herod’s sons seems to be the strongest evidence confirming Herod’s death in 4 BCE, the evidence is not nearly so straightforward as the Schürer consensus presumes. Josephus provides conflicting information on Archelaus, probably because he did not have a source that told him how long he reigned, and so his calculations are derived from when he knew Archelaus’s reign ended, as opposed to when it began. He also provides conflicting information on when Antipas’s reign ended (and says nothing of how long his reign lasted or when it began). And, of course, there is much conflicting text-critical evidence on what Josephus wrote about Philip’s reign. The numismatic evidence also may not be straightforward either (especially with Antipas’s year 43 coins), as there are multiple ways to account for the coin information.
In fact, the inscriptional, numismatic, and city founding evidence tips in favor of a date for Herod’s death after 4 BCE. A revisionist chronology that places his death in 1 BCE or 1 CE has greater explanatory scope concerning such matters as Pherora’s tomb inscription, the re-founding of Bethsaida/Julias, the coinage of Paneas/Caesarea Philippi, and the use of at least one other era by Herod than the conventional chronology. The independent character of this evidence presents a multi-faceted challenge to Josephus’s conflicted chronology that serves as the conventional one’s basis.
And therein lies the most foundational set of issues in this debate: the interpretation and appropriation of Josephus. As valuable a source as Josephus is on the events of Jewish history, any chronology that relies too heavily on a straightforward reading of Josephus and copying his chronology is doomed to share his flaws. A full-orbed appreciation of his significant chronological problems—even at the text-critical level, in the case of Philip—should reinforce the need for a critical reading of the dates he provides. The conventional chronology already recognizes some room for error in dealing with Josephus, as their reckoning of years requires some finagling to avoid Josephus’s implication that Herod died in 3 or 2 BCE. But where Josephus is trying to calculate this date and does not have access to a record that tells him this information, we need not necessarily accept his calculations—even with built-in assumptions about his use of accession or non-accession systems and when he considered the regnal year to start—if we have reason to think otherwise. And where he is using a source, it is fair to interrogate if he has understood his source properly.
Summary
Despite the consensus that Herod died in 4 BCE, the revisionists seem to have a good case for presenting an alternative in which Herod died after either the eclipse of January, 1 BCE (and hence his death was in early 1 BCE) or the eclipse of December, 1 BCE (and hence his death was in early 1 CE). I am not as interested in arguing at length which of these death dates is most likely correct, as I am more concerned with when Jesus could have been born. But given the problems cited with placing Jesus’s birth close to the December eclipse, and given that it would be even more difficult to fit Jesus’s birth into the chronology of Herod’s last days following such an eclipse, I think we can rule out 1 CE as the year Jesus was born. As such, I conclude the following for the sake of our analysis of the Gospel evidence: The reference to Herod in the Gospel narratives implies that Jesus could not have been born later than 1 BCE and was likely born before 1 BCE.
Among commentators, note the examples of Darrell L. Bock, Luke 1:1–9:50, vol. 1, BECNT (Grand Rapids: Baker, 1994), 75; W. D. Davies and Dale C. Allison, Matthew 1–7, vol. 1 of The Gospel According to Saint Matthew, ICC (Edinburgh: T&T Clark, 1988), 227; Joseph A. Fitzmyer, The Gospel According to Luke I–IX, AB 28 (New York: Doubleday, 1981), 321–22.
Louis H. Feldman, “The Testimonium Flavianum,” in Christological Perspectives: Essays in Honor of Harvey K. McArthur, ed. Robert F. Berkey and Sarah A. Edwards (New York: Pilgrim, 1982), 181–85.
Emil Schürer, A History of the Jewish People in the Age of Jesus Christ (175 B.C.–A.D. 135), vol. 1 (New York: Scribners, 1896), repr., rev. and ed. Geza Vermes and Fergus Millar (London: T&T Clark, 2014), 281–82 n. 3, 284–86 n. 11, 294, 326–28 n. 165.
It is worth noting that Josephus contradicts the latter figure in Ant. 17.162, when he describes the Hasmoneans as having 125 years of governance. This would fit with a non-inclusive reckoning from 162 BCE to 37 BCE, but it would also contradict the case for the conventional chronology about the numbers Josephus mentions in direct proximity to his recapture of Jerusalem. It is possible that the different terminology applied to each text implies a different temporal referent. In Ant. 14.490, the 126 years is of the Hasmoneans’ ἀρχή. In Ant. 17.162, the 125 years applies to the action of an imperfect verb (ἐβασίλευον). One also must account for the former being Josephus’s chronological statements, while the latter is indirect discourse from Herod.
Schürer, History, 326–27 n. 165; P. M. Bernegger, “Affirmation of Herod’s Death in 4 B.C.,” JTS 34 (1983): 529–30; Harold W. Hoehner, “The Date of the Death of Herod the Great,” in Chronos, Kairos, Christos: Nativity and Chronological Studies Presented to Jack Finegan, ed. Jerry Vardaman and Edwin M. Yamauchi (Winona Lake, IN: Eisenbrauns, 1989), 104–5; Douglas Johnson, “‘And They Went Eight Stades toward Herodeion,’” in Chronos, Kairos, Christos: Nativity and Chronological Studies Presented to Jack Finegan, ed. Jerry Vardaman and Edwin M. Yamauchi (Winona Lake, IN: Eisenbrauns, 1989), 94.
Hoehner, “Date,” 103–4. Pace W. E. Filmer, “The Chronology of the Reign of Herod the Great,” JTS 17 (1966): 291–93.
Noted by Timothy D. Barnes, “The Date of Herod’s Death,” JTS 19 (1968): 209 and Bieke Mahieu, Between Rome and Jerusalem; Herod the Great and His Sons in Their Struggle for Recognition: A Chronological Investigation of the Period 40 BC-39 AD with a Time Setting of New Testament Events, OLA 208 (Leuven: Peeters, 2012), 236. Mahieu also notes that those scholars favor this date because it would comport with Herod’s death taking place on Kislev 7, as stated in a scholion on the commentary of Megillat Ta’anit for this date. Filmer (“Chronology,” 284–85) makes this reference, along with the reference to 2 Shevat, part of his argument for Herod’s death in 1 BCE.
Bernegger, “Affirmation,” 527; Harold W. Hoehner, Chronological Aspects of the Life of Christ (Grand Rapids: Zondervan, 1977), 13; Alden A. Mosshammer, The Easter Computus and the Origins of the Christian Era, OECS (Oxford: Oxford University Press, 2008), 321; Schürer, History, 327 n. 165.
While Caligula’s fourth year extended to January 24 of 41, there is simply no room in these last few weeks of his reign to fit Antipas’s deposal.
In both cases linked here, note that the dates are written in the shorthand of letters (ΜΓ and ΛΖ respectively).
Barnes, “Death,” 206. Cf. Mahieu, Herod the Great, 256–57.
Nadav Sharon, “Herod’s Age When Appointed Strategos of Galilee: Scribal Error or Literary Motif?” Bib 94 (2014): 49–63.
Ibid., 51–54.
Because it is less directly connected to the Herodian dynasty, I have not explored here the matter that some scholars bring up of the presence of Gaius Caesar—the man who was supposed to be Augustus’s heir—at the hearing of disputes between Archelaus and Antipas (Josephus, Ant. 17.229; J.W. 2.25). He was not, however, present for Philip’s hearing (Ant. 17.300–323; J.W. 2.80–92) As this debate would take us too far afield in an entry that is already in danger of going off track, and as this data is not particularly decisive in any case, I simply refer the reader to those who think this reference indicates the conventional date for Herod’s death (Barnes, “Death,” 208–9; Bernegger, “Affirmation,” 526) and those who argue that it indicates a later date (Mahieu, Herod the Great, 310–24; Ernest L. Martin, “The Nativity and Herod’s Death,” in Chronos, Kairos, Christos: Nativity and Chronological Studies Presented to Jack Finegan, ed. Jerry Vardaman and Edwin M. Yamauchi [Winona Lake, IN: Eisenbrauns, 1989], 87–88).
Mahieu, Herod the Great, 237.
Ibid., 129–32. On this date for Herod’s capture of Jerusalem, see ibid., 60–123. As for the reference to Actium in Ant. 15.121, the syntax (with the use of an aorist participle) seems to imply that the battle had taken place prior to the earthquake that is the focus of the sentence.
Filmer also posits, “27 years is almost exactly 334 lunar months, and 334 lunations require 9,863 days, 5 ½ hours. Since 9,863 days is a multiple of seven, every date in the Jewish calendar in 36 B.C. would fall on the same day of the week as it did in 63 B.C. This coincidence would not apply to 37 B.C., however, so the twenty-seven years interval that Josephus gives looks like being genuine” (“Chronology,” 285–86). For more on these date markers, see Andrew E. Steinmann, “When Did Herod the Great Reign?” NovT 51 (2009): 9–10. However, we must note again the complication of Ant. 17.162. Given the differences between 14.490 and 17.162 noted above, it is unclear what the latter applies to and why the number is different. It could be, as Mahieu (Herod the Great, 83) suggests, that the former is reckoning from the actual date of institution, which she dates to 161 BCE, using a non-accession system, while the latter is an accession year system beginning from the following Tishri New Year.
In the case of 18.26, Steinmann notes the following “Since the first year after Actium ran from September 2, 31 BCE to September 1, 30 BCE, the thirty-seventh year ran from September 2, 6 CE to September 1, 7 CE. Thus, in this case Josephus was not counting inclusively, since one-third of the thirty-seventh year after Actium took place in 6 CE” (“Reign,” 6).
Mahieu, Herod the Great, 241–42.
Filmer, “Chronology,” 293. Cf. Steinmann, “Reign,” 7, 25–28. For a corrective, see Mahieu, Herod the Great, 49–54.
Mahieu, Herod the Great, 62–64; Filmer, “Chronology,” 289–91.
Mahieu, Herod the Great, 64–74.
Steinmann, “Reign,” 11.
Mahieu, Herod the Great, 238.
Mahieu (ibid., 293) notes this distinction whereby Josephus sometimes uses inclusive reckoning in Antiquities (14.330, 416; 17.200, 348) where he uses non-inclusive reckoning in War (1.248, 305; 2.1, 113).
Steinmann, “Reign,” 13–16.
Mahieu, Herod the Great, 240.
Martin, “Nativity,” 88.
Filmer, “Chronology,” 283–84, 293; Jack Finegan, Handbook of Biblical Chronology: Principles of Time Reckoning in the Ancient World and Problems of Chronology in the Bible, rev. ed. (Princeton: Princeton University Press, 1998), 299–301; Steinmann, “Reign,” 17.
Mahieu, Herod the Great, 290–91; Marshall, Eternal City, 49.
Mahieu, Herod the Great, 290.
Ibid., 304; Martin, “Nativity,” 86–87.
Martin, “Nativity,” 87.
Filmer, “Chronology,”296–98; Steinmann, “Reign,” 20–23.
Mahieu, Herod the Great, 397–98.
Ibid., 274–75.
Ibid., 272–87.
David W. Beyer, “Josephus Reexamined: Unraveling the Twenty-Second Year of Tiberius,” in Chronos, Kairos, Christos II: Chronological, Nativity, and Religious Studies in Memory of Ray Summers, ed. E. Jerry Vardaman (Macon, GA: Mercer University Press, 1998), 85–96.
The earliest editions reviewed here are Latin. Only 11 Greek manuscripts of Josephus contain this portion of Antiquities (one from the 11th century, the rest from the 14th–16th centuries), and to my knowledge, none of them attest to this reading. However, the dating of these manuscripts may further vitiate this point, and I am also not aware of if any systematic review of these manuscripts for this portion of the text has yet been done. Furthermore, as Mahieu observes, “Analogously, the Greek version of B.J. 1.68 assigns 33 years to Hyrcanus I and is inferior to the Latin version by Hegesippus (the second half of the 4th century AD), which records 31 years in accordance with the (Greek and Latin) texts of A.J. 13.299; 20.240: Schürer, History, 1:200 n. 1” (Herod the Great, 399 n. 25).
Beyer, “Josephus Reexamined,” 91–93.
Ibid., 87; Mahieu, Herod the Great, 402–3. Finegan (Handbook, 301) suggests what would effectively be a hybrid reading, in which Philip had a 37-year reign and died in the 22nd year of Tiberius, which would fit with a date of Herod’s death in 1 BCE.
Beyer, “Josephus Reexamined,” 96; Mahieu, Herod the Great, 403. For more on the chronology of Philip’s death, see Mahieu, Herod the Great, 398–410.
Mahieu, Herod the Great, 256–60.
On the significance of such an event for Herod, see ibid., 260–62.
Ibid., 271.
Ibid., 266–69.
David M. Jacobson, “Herod the Great, Augustus Caesar and Herod’s ‘Year 3’ Coins,” Strata 33 (2015): 89–118. The former date is also advocated by Adam Kolman Marshak, “The Dated Coins of Herod the Great: Towards a New Chronology,” JSJ 37 (2006): 212–40. Marshak (ibid., 235) also hypothesizes that the ΤΡ monogram represents the initials of the mintmaster who made these coins.
Ibid., 222–23.
Mahieu, Herod the Great, 359–78. For a number of other proposals for identifying year 3, see Jacobson, “Year 3,” 90.
Marshak, “Dated Coins,” 225–35. On the iconography, see Jacobson, “Year 3,” 91–102.
As an analogy, Jacobson (“Year 3,” 107) cites Cleopatra VII receiving Ituraea from Antony and issuing coins in year 6 of a new era.
Beyond what is discussed below, Mahieu (Herod the Great, 354–55) lists the following chronological inconsistencies in Josephus’s chronology of reigns: Jonathan (Ant. 13.212, 228; 20.238); John Hyrcanus I (J.W. 1.68; Ant. 13.236, 299; 20.240); Salome Alexandra (Ant. 14.4, supplying the incorrect year of her death); Hyrcanus II (Ant. 15.178, 180; 20.245).
Mahieu, Herod the Great, 350–56. On the latter, as Mahieu (ibid., 62 n. 58) notes, Hyrcanus II may have been old, but not that old, as he would then be born several years before his parents met or married.
Mahieu (ibid., 355) suggests that Josephus’s source connected the 34 years to an execution by Herod and he mistakenly applied it to the execution of Antigonus (Ant. 17.191; J.W. 1.665). However, in the other cases where he mentions the event without the 34 years, he correctly notes that Antony executed Antigonus, not Herod (Ant. 14.490; 15.8–9; 20.246).
Barnes, “Death,” 206; Hoehner, “Date,” 109; “Johnson, “Eight Stades,” 94–96; Mahieu, Herod the Great, 254–55.
For these and more problems, see Johnson, “Eight Stades,” 98–99.
Ibid., 99.
Hoehner, “Date,” 110.
Ibid., 103.
Mahieu, Herod the Great, 73–74.