For the sake of clarity, I shall examine separately the two essential features of lunar calendar reckoning: This intercalation is necessary in order to remain in line with the solar year. In order for the lunar year to keep up with the solar year, i. There are two ways of making intercalations.
An assessment is made every year—or at most, one or two years in advance—of the discrepancy that has accumulated between the lunar and the solar year. On that basis, the decision is taken whether to intercalate an additional month. In some cases, account is also taken of extraneous considerations, e.
The incidence of intercalated years is thus irregular and largely unpredictable. The length of the cycle can vary cycles commonly found are of 3, 8, 19, 25, or 30 years. The disadvantage of the second procedure is its unavoidable inaccuracy. No cycle, indeed, achieves an accurate synchronization of lunar and solar years. They also form the theoretical basis of cyclical intercalation. But in practice, once a cycle is running, these limits no longer need to be referred to except, perhaps, for the purpose of verifying, from time to time, the accuracy of this cycle.
In our study of intercalation in Jewish calendars, therefore, the two following questions will be considered: The evidence of datings is also too sporadic to be informative: The synchronization of Jewish lunar calendars with the solar year, which is evident in all our sources and documents, leaves at least no room for doubt that intercalation was universally carried out.
But evidence of cyclical intercalation does not go much beyond the books of Enoch and Qumran materials, which have already been surveyed in the previous chapter. More attention will be devoted to this question, therefore, than to the procedure of intercalation. The main point that will emerge from this chapter, indeed, is the diversity of calendrical practice that appears to have existed among the Jewish communities of late Antiquity.
Enoch, q Umran, and Other Sources 2. A number of lunisolar cycles are alluded to in Ethiopic Enoch As I have argued section 1. Whether or not this calendar would have been used in practice remains completely uncertain see section 1. Thus it is unclear whether in practice intercalations were empirical or cyclical in this early period.
One can learn it from what is said by Philo, Josephus, and Musaeus, and not only by these, but also by both of the Agathobuli, who are still more ancient and are surnamed the teachers. One can learn it also from what is said by the excellent Aristobulus. Eusebius, Ecclesiastical History, 7. This may cast doubt on Anatolius' interpretation of the other sources he cites, not least of Aristobulus. It should be noted that the attribution of this rule to ancient Jewish authorities suited the needs of Anatolius, whose whole purpose was to justify the adoption of this rule for his own computation of the date of Easter.
We are entitled, therefore, to question the impartiality of his interpretation. To compound this problem, the only fragment of Aristobulus on the date of Passover which we possess comes from Anatolius himself. This passage is consecutive to the one quoted above: And Aristobulus adds that on the feast of Passover of necessity not only the sun will be passing through an equinoctial sector, but the moon also.
For, since there are two equinoctial sectors, the vernal and the autumnal, and since they are diametrically opposite one another, and since the day of Passover was assigned to the fourteenth of the month in the evening, the moon will stand in the positions opposite and over against the sun, just as one can see it at the seasons of full moon.
So the one, the sun, will be in the sector of the vernal equinox, and the other, the moon, of necessity will be in the sector of the autumnal equinox Anatolius ap. The status of this passage is problematic in many respects. Firstly, it is unclear how much of it is a citation from Aristobulus. Secondly, it is unclear whether the citation is verbatim, or only a paraphrase. Besides, Aristobulus cannot possibly mean that the full moon always occurs on the same day as the equinox.
This rule has no apparent basis, indeed, in biblical sources. Synchronization of the festivals with the solar year is implicit in the Bible, but only in relation to agricultural seasons: Rabbinic sources interpret this phrase as meaning the autumnal equinox, but this is unlikely to have been its original meaning. In the 1st solar month the sun rises from the 4th gate None of these verses, however, suggest the use of the equinox as a criterion for the synchronization of the festivals with the seasonal cycle. Aristobulus' rule of the equinox is more likely to have originated, therefore, from outside the Bible.
This raises the possibility that Aristobulus' concept of when Passover occurs may be no more than theoretical. If so, the rule of the equinox may not have been adopted by the Jews till much later, possibly not before the fourth century CE. Philo, Josephus, and Epigraphic Sources Anatolius is probably right with regard to Aristobulus, but his attribution of the rule of the equinox to Philo and Josephus, in the passage above quoted, is less certain.
This implies, as with Aristobulus, an interest in astronomical concepts which biblical sources do not display. These statements In the rabbinic calendar: Similar statements are found elsewhere in Philo's works. In Special Laws 2. In the Life of Moses 2. Thus Anatolius' attribution to Philo of the rule of the equinox is possible but not really substantiated by the sources.
There is certainly a connection, in Philo's works, between the festivals and the equinoxes; but this connection does not take the shape of a clear or consistent rule. It is also interesting to note that Philo has nothing to say about intercalation, even though intercalation is implicit from the synchronization of his lunar calendar with the equinox and the solar year.
The beginning of Aries was and still is the traditional astronomical designation of the vernal equinox. In Josephus' period, Pharmuthi began on 27 March in the Julian calendar, thus only a few days after the equinox. Josephus may simply be reporting that Passover usually occurs in Aries or in Pharmuthi, after the equinox, but without this constituting a formal calendrical rule.
Eusebius, Ecclesiastical History 7.
- The Book of The Animals - The Collection (Bilingual English-Spanish) (The Book of The Animals (Bilingual)).
See further Loewinger 45—6 n. This festival was apparently Passover, as we shall presently see. The date of Tiberius' death is well known to us from classical Roman sources: Since Passover occurs around the time of the full moon, only two options can be considered, in this context, for the year 37 CE: However, 24 March must be discarded, because news of Tiberius' death could not have reached Jerusalem from Rome in six days.
Travel by sea at this time of the year or by land would have taken a number of weeks. It is most plausible, therefore, that news of Tiberius' death would have taken just over a month to reach Jerusalem: Passover, in that year, would have been celebrated in Jerusalem on 19 April.
Travel by sea at this time of the year was considered still uncertain. See Seager See Barrett 50—1. See Casson For government couriers, Ramsey estimates an average of 50 Roman miles per day under normal conditions, but this is based on a misinterpretation. News of the accession of Pertinax took over two months for the same journey in the winter from 1 January to 3 March CE, either by sea or by land. However, the relatively late occurrence of Passover in 37 CE does not necessarily prove that such a rule was in existence, as we shall further elaborate below.
Another passage in Josephus, often quoted in this context, must be dismissed as inconclusive. Assuming, furthermore, that the Macedonian calendar used by Josephus was always that of Tyre, Schwartz concludes that in 66 CE 15 Nisan would have occurred on 25 April, more than one month after the vernal equinox. The date of 25 April is, however, untenable, because the full moon—with which Passover should have approximately coincided—occurred in that year on 28 April—as Schwartz himself indicates but yet, remarkably, ignores.
In the Macedonian calendar of Tyre, 8 Xanthikos always corresponded to 25 April. In actual fact, if the calendar was based on sightings of the new moon, 15 Nisan would not have occurred before 30 April for the sources of my astronomical data, see Ch. The correct interpretation of this passage in Josephus is provided by Safrai a: The inscription records a decision by the Jewish community of Berenike to grant regular honours to a Roman benefactor. It begins with an Egyptian date, as was normal in Cyrenaica being a former possession of the Ptolemaic kingdom ; but this date is immediately followed by a Jewish dating: The festival of Tabernacles lasts, in fact, seven days: There is no reason to treat it as part of the narrative which would then refer to some other time.
The text that follows, indeed, is a list of the current archons, standard in the dating of all public inscriptions; this indicates that the narrative of the inscription has not yet begun in this passage. Reynolds, in particular, believes that an era beginning from 95 BCE was also in use in Cyrenaica; which would yield for our inscription the earlier year of 41 Lev.
Note, as she herself points out For conversion tables, see Skeat In this light, the dating of 41 BCE deserves serious consideration. Indeed, it would have been observed much later than the rule of the equinox demanded. Passover So Skeat ; and not the 24th, as erroneously converted in Reynolds —5. The only inscription quoted by Reynolds as evidence of an era from 95 BCE is no.
See Boffo n. I shall argue in section 3. On the method I have employed to determine when new moons would have been visible, see sections 3. The dating of 25 CE see above, n. Within the terms of the rule of the equinox, however, Passover should have been observed in the previous month, on 31 March. Philo is inconsistent regarding the relationship between Passover and the equinox.
Further evidence that will be considered in the next chapter lends support to this conclusion. On the contrary, the date of Tabernacles in the Berenike inscription is so late that it cannot be explained in terms of the observance of this rule. It is possible that intercalation in the Jewish calendar was based entirely, in this period, on the Babylonian system of intercalations.
The dates which we have examined in this section conform, indeed, to the dates of the Babylonian calendar. Thus Nisanu in the Babylonian calendar began in 37 CE on 6 April, and Tashritubegan in 41 BCE on 12 October, exactly as Nisan and Tishre would have done in the Jewish calendar according to the evidence from Josephus and the Berenike inscription that we have considered. Note, however, that the days of the month in Parker and Dubberstein were based on outdated astronomical data, and thus were only intended as approximations as acknowledged ibid.
The names of months, on the other hand, were based on cuneiform and other evidence that is completely secure: In general terms, the Babylonian Nisanu always began after the vernal equinox, thus relatively late. The limits of 1 Nisanu in the Hellenistic and early Parthian periods are approximately 25 March—23 April ibid.
Schwartz has suggested that in the period of the Temple, Passover was celebrated late so as to enable pilgrims to reach Jerusalem on time for the festival. Non-Jewish sources can often be mistaken or unreliable, but this is not a reason to discard them entirely. Like many other scholars, Schwartz follows Anatolius and attributes the rule of the equinox to Philo and Josephus; but he considers this rule to have been only theoretical in the 1st c.
Schwartz's argument is built on the assumption that Passover was celebrated in 66 CE on 25 April i. This is actually incorrect, as I have shown above section 2. Some sources will be discarded entirely for this reason: Julius Africanus and Epiphanius, as we shall see below. Other Christian sources are considerably more reliable; in particular, those pertaining to the fourth-century controversies on the date of Easter, which frequently refer to the dates of the Jewish Passover.
Fourth century sources will be examined in the following section 2. For the second and third centuries, evidence relating to intercalation in the Jewish calendar is at best sporadic. There is a brief but important passage in Galen mid or later second century , already mentioned in Ch. In this context he writes: Those who reckon the months in this manner are forced to make an intercalation as soon as the remnant of the previous years has accumulated into the period of one month.
And it is written in the works of Hipparchus as well of other astronomers when additional months must be intercalated. But alternatively, it may be taken to indicate that the Jews in Palestine did not have a regular cycle of intercalations. They may have intercalated the year by rule of thumb: This would correspond to the empirical procedure that is known from rabbinic sources on Palestinian Jewish practice in the same period. Stern —87 ii, no. Africanus' sole purpose is to justify his use of the octaeteris as the basis for his calculation of the period extending from Nehemiah to Jesus.
Africanus' statement is also in itself ambiguous: It would be incorrect, therefore, to treat this passage as evidence that the Jews actually used, at any stage in history, the octaeteris. This cycle consists of a regular alternation of day and day months, with one extra day every 3 years hence 28 extra days for the whole year cycle. Every 14 years 5 intercalations are made, hence 30 intercalations in 84 years; in the 85th year of the cycle, a 31st intercalation is made. Epiphanius argues, however, that this calendar was inaccurate, and that Jesus knew the exact reckoning; which would explain why he appears to have celebrated Passover earlier than the other Jews.
Epiphanius' unreliability in historical and chronological matters is, in any case, notorious. The exact reckoning, according to Epiphanius, that Jesus knew consisted not of 28 extra days for the year cycle, but only 24 days—or according to another textual version, 21 days—and 3 double hours.
Thus although the second and third centuries remain obscure, we may accept Galen's intimation that in this period, no regular cycle of intercalation was used. Passover and the Christian Easter The fourth century opens up a corpus of evidence of considerable importance to the history of the Jewish calendar: The date of Easter became the object of intense controversy in the second half of the second century, and then again, following an apparent lull, during the fourth century and later.
Little is known about the second-century controversies; but sources from the fourth century abound. One of the main issues was whether the Christian Easter should be observed at the same time as the Jewish Passover. Frequent reference, in Christian Paschal sources from this period, to the dates of the Jewish Passover provides us with much information about the Jewish calendar that would have been in use.
The nature and history of this important Christian controversy is interesting in its own right. I shall return to it in detail in Ch. This tradition is also cited in later Muslim sources, e. Maqrizi, cited in Bornstein Two hundred years before Islam takes us to the early 5th c. As I have mentioned section 2. This rule he ascribed to the Jews of earlier Antiquity: Passover] by it, commit, as we think, no little and no common blunder. One can learn it also from what is said by the excellent Aristobulus, who was enrolled among the seventy who translated the sacred and divine Scriptures of the Hebrews for Ptolemy Philadelphus and his father.
Whereas Anatolius' main purpose was to legitimize the rule of the equinox by invoking the authority of Jewish sages of old, subsequent sources emphasize the error of contemporary Jews as a polemical argument against those Christians who observed Easter at the same time as the Jews.
I shall return to this letter, and to its historical implications, in further detail below. I would treat this source as reliable precisely because of its polemical character: The next group of sources for us to consider relates to the Council of Nicaea of CE, of which one of the main items on the agenda was the date of Easter. The decisions that were taken at this Council have been preserved in letters sent by Constantine to the Syrian and Palestinian absentees at the Council, that are cited in Eusebius'Life of Constantine and hence by early Byzantine church historians Barnes However, a better edition of this text, based on MS Vat.
According to Grumel It was resolved by the united judgement of all present that this feast ought to be kept by all and in every place on one and the same day. The nature of this rule—and hence, by contrast, the nature of the Jewish custom—is not explicitly stated in this letter, nor in any other contemporary source. Daunoy surmises, in fact, that no rule was ever formulated at Nicaea, for the simple reason that the churches would have been excessively divided on this matter. Rather than risking a heated debate on calendrical rules and cycles, it was safer to agree that Christians would be united in not following the custom of the Jews.
In adopting this course, the Council of Nicaea was aiming in particular at the Christian churches of the East, whose custom, especially in Syria, had long been to observe Easter at the same time as the Jews. For a general introduction, see Simon — Observing Easter with the Jews entailed that it would frequently occur before the equinox.
This is evident from a document from the Council of Sardica that we shall examine in detail below , but becomes even more explicit in later fourth-century sources. Evidence that many Christians did in fact follow the Jews and observe Easter before the equinox can be found in John Chrysostom's third homily against Judaizing Christians, written and delivered at Antioch in CE, in which he upheld the Council of Nicaea and attacked those Christians who were about to follow the Jews and observe Easter one month too early.
For a slightly different interpretation, see Simon — It is worth noting that observance of Passover before the equinox would also have been possible according to contemporary or slightly earlier rabbinic sources. A Tannaitic tradition rules that the year should not be intercalated and thus Passover postponed by one month because of the equinox alone: The year may be intercalated on three grounds: On two of these grounds it should be intercalated, but not on one of them alone.
This may have happened quite frequently. But it is likely that at the beginning of the century, Palestinian rabbis still celebrated Passover in some years before the equinox, just it was practised by other, contemporary Jewish communities elsewhere. For a fuller discussion of this passage, which is only partially quoted here, see section 4.
This suggests, in Jewish calendrical practice, a rather radical change. The theory that the Jewish calendar changed in this manner is of course not new. It lies at the forefront of the argument of Peter of Alexandria, which I have outlined above: A similar idea is also implicit, as we have seen, in the Canons of Anatolius.
Modern scholars such as Grumel have accepted this account at face value; as we have seen, the evidence would appear to vindicate it. This change, however, remains to be explained. I have suggested above section 2. If this is correct, it would follow that after the destruction of the Temple and the cessation of pilgrimage to Jerusalem, the postponement of Passover would no longer have been necessary.
Thus we may endorse the view of Peter of Alexandria that a change in Jewish calendar reckoning occurred after 70 CE, but for entirely different reasons. This change was not the result of error, but simply an adjustment to changing historical circumstances. This would explain why, by the beginning of the fourth century, the celebration of Passover before the equinox had become widespread throughout the Jewish world. Most problematic is his suggestion that at that time a new calendrical computation was formally instituted: Whether the Jews erroneously sometimes celebrate their Passover according to the course of the moon in the month Phamenoth, or according to the intercalary month every third year in the month Pharmuthi makes no difference to us: In order to maintain synchronization with the Egyptian solar calendar, intercalation and occurrence of Passover one month later, in Pharmuthi would often be necessary after an interval of only one year.
One-year intervals are standard, indeed, in cycles of intercalation such as the octaeteris or the year cycle. Passovers Twice in Phamenoth and Once in Pharmuthi. For convenience, I am assuming in this table that the lunar year falls 11 days behind the Egyptian calendar every year, and that intercalated months are 30 days long. An intercalation in year 12 would lead to the occurrence of Passover in Pharmuthi, after only one year's interval.
In any event, therefore, the sequence 2 Phamenoth—1 Pharmuthi would eventually be broken. A system of this kind would lead to a discrepancy from the lunar months of only one day in 18 years. For an illustration, see section 3. Grumel prefers to ignore Schwartz's limits, and proposes instead, somewhat arbitrarily, that in the Jewish Alexandrian calendar the earliest possible 1 Nisan would have coincided with 1 Phamenoth.
Grumel's limits, however, are no less hypothetical than those of Schwartz. This passage cannot serve as evidence of the precise limits of Passover. Nevertheless, his statement can be treated as a reliable indication of the approximate range of dates within which Passover, in early fourth-century Alexandria or Egypt, was likely to occur.
According to Grumel the latest limit is actually 11 Pharmuthi, because for some reason he assumes that the octaeteris would have been used. Alexandrians as 21 March: A facsimile of it is appended to Schwartz's article. The text will be cited in full in the next chapter, where I shall discuss textual and calendrical issues in further detail. In it, they proposed a year cycle for the computation of the date of Easter that would conform with the rule of the equinox and the Council of Nicaea. What is remarkable about this document is that alongside its year cycle, which is presented as the Christian reckoning quo numero facimus nos Christiani , a list of the dates of the Jewish Passover is supplied quibus supputationibus faciunt Iudei pascha.
Calendar and Community
Many of the Jewish Passovers, as by now might be expected, occur before the vernal equinox; indeed, this is what distinguishes them from the Christian cycle. But the most conspicuous feature of the Passover dates is that they all occur, according to this document, in the Julian month of March. These Jewish Passover dates must be considered reliable on a number of accounts. They were listed in this document for the explicit purpose of demonstrating that the Eastern bishops were committed to the rule of the equinox and the decisions of the Council of Nicaea, and would not observe Easter at the same time as the Jews; their year cycle differed, indeed, from the Jewish dates.
It seems that the authors of this document were cautious not to predict when the Jews would celebrate Passover in future years, and to record only those dates of Passover which were known to have actually occurred. Eastern bishops were quite capable of knowing when, in recent years, the Jewish Passover had been observed.
Since the custom of many Eastern Christians was still to observe Easter at the same time as the Jews, the authors Or according to E. Schwartz , in the winter of CE ; see Hess 63—7, —4. If the Jewish Passover occurs towards the end of March, after the vernal equinox, then the Christian date in this document is identical with it e. Otherwise, the Christian date occurs 30 days later, in April.
In these 16 years, the dates of Passover range between 2 March in the 10th year and the 30th of March 2nd year. The decision whether or not to intercalate would have depended entirely on this rule: The Julian calendar, however, would have provided the Jews with a convenient guideline for the synchronization of their lunar calendar with the solar year. An allusion to this rule may be implicit in an obscure passage of Socrates'Ecclesiastical History, in which he describes the difference of opinion among Eastern Christians regarding the date of Easter.
Some, he writes, followed the Jews even when they were in error; whilst others observed Easter after the equinox, saying it should always be kept when the sun was in Aries, in the month of Xanthikos according to the Antiochene calendar or of April according to the Roman calendar. Aries begins at the vernal equinox, thus in the month of March. These attempts fail, however, to convince. This is clearly incorrect: By some inexplicable extrapolation, Grumel infers from this that the Jews must have followed a year cycle similar to the Alexandrian.
This inference, which Grumel reiterates elsewhere The validity of Schwartz's emendations will be discussed in detail in Ch. By a similar extrapolation, Grumel could just as well have inferred that the Jews followed a year cycle, or indeed, any other cycle e. In the absence of a repeated pattern within these 16 years, there cannot be conclusive evidence of any cycle at all. The fact that the sequence of dates in these 16 years is compatible with a section of the Alexandrian cycle could easily be coincidental. The decision whether or not Other comments of Bornstein on this document are also erroneous.
Accordingto him the authors of this document were Quartodecimans on which see below, section 5. Bornstein promises in this article to return elsewhere, in further detail, to the Sardica document; at this stage in his research, he had clearly not studied it in depth. It is interesting to note that two of the most seminal monographs of the 20th c. All that concerns us in this chapter is the fact the Jewish dates occur in March; this is not affected, in any way, by Schwartz's emendations. An empirical procedure of this kind is implicit in the anonymous Paschal homily of CE, which describes the Jewish practice of intercalation as follows: At all events, this passage conveys the impression that the procedure of intercalation was purely empirical, and did not follow a standard cycle.
Moreover, unlike a number of other civil calendars, that of Antioch was identical with the Julian calendar; a Jewish community that used the month of March as the criterion for intercalation is likely to have been in a city where the Julian calendar was in civil use. The choice at Antioch of the month of Dystros March as the Passover month may appear surprising, as it was the following month, Xanthikos, that was normally associated and sometimes equated with Nisan.
It is See above, n. The Antiochene equivalent of March was Dystros. In the Hellenistic period, Nisan was generally equated with Artemisios the month following Xanthikos: In Alexandria, by contrast, the Julian calendar had never been in use; it would not have been practical, therefore, for Alexandrian Jews to use a Julian month as a criterion for intercalation.
In some years, they would have observed Passover one month apart. Schwartz has attempted to infer the limits of Passover in Jewish communities of Asia Minor from the Martyrdom of Polycarp 8. The year of this martyrdom has been subject to much debate, which need not detain us here: According to the Martyrdom of Pionus 2: He attributes this to the early occurrence of Xanthikos in the Asian calendar 21 February—23 March. This interpretation must be rejected on a number of counts: See Camelot , Evidence from the Council of Nicaea In his report of the decisions of the Council of Nicaea in CE, Constantine made no explicit mention of the Jews' observance of Passover before the equinox.
He did mention, however, that they observed Passover on an erroneous date, as follows: This expression, however, may be idiosyncratic to Epiphanius. I am grateful to Leofranc Holford-Strevens for his remarks. Further options are also possible, but this lies beyond my scope. Against this argument, however, see Lieu 70—2.
See Beckwith This, however, is hardly more convincing. Even if the writer intended Passover itself, the association of Polycarp's martyrdom with Passover is likely to be a literary topos rather than a factual date. Hence it is that on this point also they do not see the truth, so that, always getting most things wrong, instead of making the appropriate adjustment they celebrate Passover twice in the same year. Why then should we follow those who confess to being sick in grievous error?
Surely we shall never consent to keep this feast a second time in the same year. Modern scholars have generally followed the explanation provided by Epiphanius late fourth century. Accordingly, if the Jews celebrate one Passover after the equinox and the next Passover 12 lunar months later before the next equinox, both Passovers will have occurred within the same year i. To cite Epiphanius' interpretation: For if we always celebrate on the Jewish date, we shall sometimes celebrate after the equinox, as they often do, and we too; and again, we shall sometimes celebrate before the equinox, as they do when they celebrate alone [i.
Therefore if we celebrate then too, we may keep two Easters in one year, one after the equinox and one before it [i. For the year will not be over before the day of the equinox; and the cycle of the course of the sun , which God has given men, is not complete unless the equinox has past Panarion There are, however, objections to this interpretation. Firstly, it is unlikely that Constantine or whoever authored his statement would have assumed that the year began at the vernal equinox. According to Heikel Jones 20; Grumel Epiphanius appears to suggest, at the end of the passage quoted, that Constantine had in mind an astronomical solar year that began from the vernal equinox.
The Julian calendar began the year on 1 January; the Egyptian civil calendar on 29 August; all the Macedonian calendars, in Asia Minor as well as in Syria, began the year in the autumn except for Lycia, where the year began in January, following the Julian calendar, and Attica, where the year began in July. The only calendar in the Roman Empire that began the year on the vernal equinox more precisely, on 22 March was that of the Province of Arabia. Constantine's letter, however, does not refer to this possibility, which is surprising if we consider that not observing Passover at all would surely have been worse than observing it twice in the same year.
Adifferent interpretation of Constantine's statement is suggested by Beckwith According to Beckwith, the Jews in the Diaspora would have kept Passover twice every year, because of their uncertainty as to whether the year had been intercalated by the Palestinian rabbinic court, and hence on which full moon, whether before or after the A suggestion of C. Such a practice would have resembled the rabbinic Diaspora observance of two consecutive days for each festival, which resulted from the uncertainty surrounding the date of the new moon as described in rabbinic sources, and to be discussed in Chs.
Beckwith's interpretation would certainly eliminate the objections I have raised against Epiphanius, but it must be discarded on two grounds. Firstly, this suggested Diaspora observance of two Passovers is completely hypothetical, and has no historical or other evidence to substantiate it. Secondly, Beckwith assumes the dependence of Diaspora communities on the calendrical decisions of the Palestinian rabbinic court.
All the evidence suggests, in fact, that Diaspora communities took charge of their own calendar, without ever referring to Palestinian calendrical authority. The correct interpretation of Constantine's letter, in my view, follows precisely from this point. The context of the council of Nicaea makes this interpretation all the more plausible. One of the stated purposes of the council of Nicaea—indeed, its main purpose—was to achieve the unity of the Church. This is why the date of Easter, together with the Arian heresy, constituted the main items on its agenda. A number of later fourth-century sources, contemporary with Epiphanius, appear not to have connected the notion of celebrating Passover twice in a year with the rule of the equinox.
This indicates, perhaps, that Epiphanius' interpretation of Constantine's statement was not shared by all. Some sources, indeed, may have understood Constantine's statement in the way I am suggesting. It could not occur later, therefore, than the 21st of the moon. It is associated with the celebration of Pascha too late in the wrong week, rather than, as according to Epiphanius, too early and in the wrong month. This passage is only compatible, therefore, with my interpretation of Constantine's letter. Evidence can also be adduced from the anonymous Anatolian homily of CE. Although this homily criticizes the Jews for not obeying the rule of the equinox, it does not mention at any point that this leads to their observance of Passover twice in the same year.
If the See Vita Constantini 3. The equinox is given there as 22 Dystros i. March, in the Antiochene calendar. His omission of this argument suggests, albeit e silentio, that Constantine's letter may have been understood differently, perhaps in the same way as I have proposed. The Persistence of Diversity From the fourth to the sixth centuries, evidence regarding the intercalation becomes rather sporadic.
- Search form.
- Sacha Stern?
But the impression conveyed is that the situation remained largely unchanged. The rule of the equinox appears to have remained ignored in most Jewish communities—in spite of its possible adoption in fourth-century rabbinic circles. This obscure passage in Procopius'Anecdota or Secret History has often been quoted by historians as an example of Justinian's maltreatment of the Jews, but to my knowledge, has never been properly interpreted or understood: If it ever happened, for instance, that the year in its recurring rounds brought on the Feast of Passover before the festival of the Christians, he would not allow the Jews to celebrate this at the proper time not to make any offering to God at that feast nor to perform any of the rites customary among them.
Juster's suggestion that on another occasion Justinian postponed the date of Easter by one week which he did not so that it should not coincide with the Jewish Passover is a complete misinterpretation of his sources. Justinian's persecution is often viewed as an expression of Christian prejudice against Passover, which was to become commonplace in medieval Europe. However, it is unclear why the Jews were prohibited, in this instance, from celebrating Passover only if it occurred before Easter. The reasons behind such a decree—which Procopius himself may not have fully understood—are to be sought in the context of the Christian Paschal controversies of the later Roman period.
As we have seen, a number of dissenters from the council of Nicaea, in the East but also in parts of Asia Minor, perpetuated the custom of observing Easter at the same time as the Jewish Passover i. The Novatians, in particular, adopted this custom in Bithynia and Phrygia in the second half of the fourth century; in the sixth century, under Justinian, this heresy was still in existence in some areas of the Byzantine Empire. Justinian may have reasoned that if the Jews were forced to observe Passover in the same month as the Orthodox, Nicene Easter, i.
If the Jewish equinox was later than according to Christian reckoning, then Passover could have been postponed, on occasion, to the following month. Even then, a postponement of this kind would only have been rare, as the discrepancy between the Jewish and Christian equinoxes could not have been great; whereas Procopius appears to imply, quite on the contrary, that it was the occurrence of Passover before Easter that was less frequent. If this interpretation is correct, it would mean that Justinian's persecution was not anti-Jewish per se, but rather an indirect, political measure designed at eliminating anti-Nicene heresies.
More important to us, this decree would suggest that as late as the sixth century, Jews were still widely observing Passover before the equinox, as we have seen was the case in the fourth century. Whether Justinian was successful in imposing the rule of the equinox upon the Jews is a tantalizing question which must remain unanswered.
The rule of the equinox was also promoted by the rabbis, perhaps as early as the fourth century CE, and yet their success appears to have been also limited. The precise date of this document involves detailed analysis that pertains, above all, to the problem of the beginning of the lunar month. This analysis will be left, therefore, to the next chapter section 3. My conclusions, however, are also relevant to the question of the intercalation; they are presented here in summary.
This means that the preceding Nisan in CE began on 4 March. Passover, 14 Nisan, thus occurred in that year on 17 March, which means that the rule of the equinox was not observed. Contemporary reckonings of the equinox, however, were all later than 19 or 20 March: These inscriptions are unusual in that their dates, exclusively Jewish, are recorded in detail and with much precision.
Since then, many more tombstones from Zoar have come to light.
Four of these have been published by myself S. Yannis Meimaris, in association with Sebastian Brock, is currently working on their publication. One of the most important questions this will raise is whether or not this southern Palestinian community conformed to the calendar that can be inferred from contemporary rabbinic sources.
My conclusions in this chapter are restricted to calendrical matters pertaining to intercalation; further aspects of the Zoar calendar will be discussed in the next chapter. The site of Zoar has not been systematically excavated; most of the tombstones collected by Politis were discovered on the site by local peasants see Politis All the tombstones are made of local sandstone, similar in size 30—40 cm long, 25—35 cm wide , and bear similar inscriptions; they date from the 4th—6th cc.
The correlation Aramaic—Jewish and Greek—Christian is almost completely consistent. It is attested for instance in the ketubah of Antinoopolis see section 3. See also Bornstein a —38 and Friedman i. The synchronization of Sabbatical cycle and era of destruction of the Temple is generally, but not always, consistent. Numerous attempts were made by archaeologists and calendar experts to reconcile their datings, often with most ingenious emendations.
All previous theories, therefore, can be swept away as irrelevant and obsolete. The minority that deviate from this scheme do so only by one year, either positively or negatively Naveh These deviations, I have argued elsewhere, are most likely the result of error see Table 2. Following earlier scholars, Naveh has suggested that the year from the destruction began on 9 Av, the anniversary of the destruction of the Temple, thus slightly earlier than the Sabbatical year that began on 1 Tishre; this would explain the one-year discrepancies in some of the tombstones that are dated to the intervening month of Elul i.
It is just as plausible, therefore, that years from the destruction were in fact conterminous with Sabbatical cycle years and ran from 1 Tishre, and that discrepancies were generally the result of error S. Cassuto a , Akavia , Wacholder —4, and Assis For more extensive bibliography, see Naveh N1—3 are the three inscriptions originally discovered and S inscriptions are numbered S13—16, in sequence with the numeration of Naveh Inscriptions are listed in chronological order.
I have omitted fragmentary inscriptions where essential elements of the date are missing e. As I suspect further inscriptions may be discovered and published in the near future, this list must be treated as provisional. Corrections required to obtain a consistent synchronism between era of destruction and Sabbatical cycle years. However, we cannot exclude the possibility that in inscriptions dated to the middle of the Sabbatical cycle, such as Z, N10, S14, it was the Sabbatical cycle year that was wrongly entered.
This will not affect, however, our conclusions in this present chapter. For the purposes of our analysis, we now need to convert the years of these inscriptions into years of our era CE. Most scholars have assumed that the Sabbatical cycle was reckoned in the ancient world the same as it is today. This entails that the Sabbatical-year laws would have been applicable at Zoar I am grateful to Philip Alexander for this observation.
Calendar and Community - Sacha Stern - Oxford University Press
Stern , and in documentary evidence from the end of the 1st millennium CE Friedman i. Wacholder has argued, consequently, that the Sabbatical year was reckoned in Antiquity one year later than it is today. Firstly, the meaningof its reference to the Sabbatical year is unclear, as it is not part of the dating at the beginning of the document. Secondly, the second regnal year of Nero may have begun, according to the Jewish computation, in Nisan of 55 CE, which would still have been, according to the traditional computation, a Sabbatical year.
Wacholder's theory is also problematic in that it implies that at some stage in late Antiquity, the computation of Sabbatical years was changed and brought back by one year, thus becoming the computation that is in use today. No explanation is provided as to how or why such a radical change should have been made. For argument's sake, I shall also take account of Wacholder's cycle: However, it will become evident from my analysis of Zoar inscriptions, both in this chapter and in the next, that the Jews of fourth- to sixth-century Zoar are far more likely to have followed the generally accepted Sabbatical cycle.
Most relevant to us, at present, are the inscriptions that indicate intercalated years. See also Wacholder — See Bickerman 66, Meimaris — For rabbinic evidence of regnal years beginning in Nisan, see M. Note, for instance, that N2 is dated to the same year as N12, but without reference to intercalation; the same applies to Z99 and Z The latter is perhaps less likely because it would entail a one-year discrepancy with the Sabbatical cycle year.
These dates are unlikely to be erroneous, because the synchronization of years from the destruction with Sabbatical years is consistent in all cases except for Z, where the Sabbatical cycle year is not extant. All that is provided, indeed, is the date of death of the deceased; the inscription itself may have been written many months later, after the intercalation of the year had occurred. From Z99 to N11 the interval is 95 years, whereas after 12 cycles of 8 years, the interval between two intercalated years should have been 96 years.
Although is less likely, allowance will be made for this uncertainty. The overall conclusions, however, will not be affected. The datings of Z99 and Z, however, are somewhat uncertain. The year of this inscription—assuming it is not erroneous—is from the destruction, two years after the intercalated year of Z The following year , therefore, must presumably have been intercalated, since in any lunisolar calendar intercalation must been made at least every three years.
If we add to our chart, we obtain a sequence of intercalated years that is incompatible both with the year cycle and with the year cycle. Although other cycles e. Firmer evidence can be drawn regarding the rule of the equinox. The inscriptions suggest, indeed, that this rule was not observed.
In that year, the day of Passover 14 Nisan could only have occurred c. Now since the year The sequence ———— can be reduced, by removing all multiples of 19, to the sequence of 18—2—4—18—9; each of these numbers representing the position of intercalated years in a hypothetical year cycle. From year 2 to year 9 inclusive of such a cycle, there would be a period of 8 years comprising4 intercalations 2, 4, 9, and an additional intercalation, necessarily, between 4 and 9 , which is excessive in any lunisolar cycle. The existence of such a cycle is therefore implausible.
This argument depends entirely, however, on the reliability of the dating of Z99 i. The sequence —— see chart in main text can be reduced, by removing all multiples of 25, to the sequence of 3—21—23; each of these numbers representing the position of intercalated years in a hypothetical year cycle.
From year 21 to the following year 3 inclusive of such a cycle, there would be a period of 8 years comprising4 intercalations 21, 23, 3, and an additional intercalation, necessarily, between 23 and 3 , which again is excessive in any lunisolar cycle. This argument still depends, however, on the reliability of the dating of Z99 i. Note, however, that this date of Passover is only approximate, because in the majority of cases the month began at Zoar slightly earlier, as will be shown in the next chapter, section 3.
Calendar and Community: A History of the Jewish Calendar, 2nd Century BCE to 10th Century CE
Account will be taken of this below. But as we shall see in the next chapter, in the majority of cases from Zoar the new month would have begun earlier, between conjunction and visibility of the new moon. Passover is likely, therefore, to have occurred in CE on the 16th of the month, even further apart from the equinox. Secondly, and more importantly, it is highly unlikely that the Jews of Zoar referred to the astronomical true equinox, which occurred, as stated, on 19 March. Instead, they would have referred to the equinox that was assumed by various calendars, Jewish or non-Jewish, in the same period.
The Easter cycle of the Alexandrian Church assumed, from the beginning of the fourth century, an equinox on 21 March Grumel The Jews of Zoar could not have failed, moreover, to realize that they were in breach of this rule. The proportion of tombstones discovered at Zoar indicates that the Jews lived in the midst of a predominantly The new moon would have been visible at Zoaron 3 March in the evening; but see previous note. This equinox date went back to Anatolius in the late 3rd c.: Grumel 31—2; Strobel The same is implicit in rabbinic sources, B. Stern —6, and below, section 5.
The Christians at Zoar are likely to have followed the rule of the equinox, as did most Christians in the fourth century and beyond. The evidence is not restricted to N Inscription Z99 yields a similar result: The Christians, besides, would have observed Easter one month later. N12, on the other hand, does not provide further information in this respect.
Passover would have occurred, in its year, on 10 or 11 April. In the year preceding this inscription, CE, Passover would have occurred on 21 March. It is unlikely, indeed, to have occurred in the next lunar month, for then Passover of the previous year in CE , 12 lunar months earlier, would have occurred very late, around 25 April; this is well over 5 weeks later than the date of Passover inferred from N11 17 March , and is therefore unlikely.
S13 suggests, therefore, that Passover of CE may have been observed as early as 15 March. If we assume, for argument's sake, Wacholder's Sabbatical cycle, then all our dates must move forwards by one year. The calculations above, relating to the rule of the equinox, must then be altered. However, this would not affect our main conclusion, that the rule of the equinox was breached. The year of N12, for instance, would become —16 CE.
In CE, Passover would have occurred on the 30th of March one lunar month later, on 29 April, is less plausible as it is well over 5 weeks after the equinox—see below. In the absence of evidence to the contrary, we may thus assume—though not without some reservations—that Palestinian Christians followed the normative Easter cycles Alexandrian or Anatolian in conformity to the Council of Nicaea.
Similarly in CE , Passover, according to those cycles, would have occurred on 19 March whilst Easter was postponed to Sunday 25 April see references above, section 2. In CE, Passover would have occurred on 25 March one lunar month later, on 23 April, is less plausible as it is 6 weeks after the Passover date inferred from N12—see below. If so, Passover of the previous year, 13 lunar months earlier, would have coincided with 7 March, thus in further and more pronounced breach of the rule of the equinox. The alternative but less plausible option is that Passover of N12 occurred on 29 April, and Passover of N11 on 23 April, with Passover of the previous year occurring on 5 April.
This would mean that most Passovers occurred in April, often more than a month after the equinox, and hence often one month after the Christian Easter—a rather unlikely occurrence.
- Jilted City!
- Nail Your Novel - Why Writers Abandon Books and How You Can Draft, Fix and Finish With Confidence.
- Maximum Performance Basketball.
- Product details?
I would prefer, therefore, to conclude that the limits of Passover according to Wacholder's cycle would have been around 7—30 March. Although the early date of 7 March cannot be dismissed as impossible—as we have seen, the Jews of fourth-century Antioch would have celebrated Passover as early as the beginning of March—it is perhaps less likely than the later dates which I have inferred as limits of Passover according to the generally accepted Sabbatical cycle.
Wacholder's cycle, in the context of Zoar, is therefore less likely. In practical terms, this would have meant that in CE, CE, CE, and presumably also in other years, the Jews of Zoar celebrated Passover one month earlier than in the Palestinian rabbinic communities. The Zoar calendar differed in further ways, indeed, from the rabbinic calendar, as will be seen in the next chapter. This challenges the common assumption that by the later Roman period, the rabbis and rabbinic Judaism had become a dominant force in Jewish Palestinian society. What is particularly unique about Zoar, indeed, is its proximity to the main centres of rabbinic activity in Judaea and Galilee.
Their ability to ignore the authority of their rabbinic See Bornstein —3, citing B. RH 21a, and below, section 4. In Galilee itself, by contrast, there may be evidence that the rule of the equinox was observed. In this period, the rule of the equinox was far from universally observed. See illustrations in Nagy et al. These factors may be political, as in ancient Greek calendars, or religious, as in the rabbinic calendar. In almost all ancient lunar calendars, the month begins around the time of the new moon. Precisely when this is, however, varies considerably from one calendar to the next.
This occurs invariably in the evening, not long after sunset, and close to the point on the horizon where the sun has set. Saadya and Ben Meir. The organizing argument of the book is that the emergence in the ninth and tenth centuries of a normative Jewish calendar "epitomizes the gradual development of solidarity and communitas among the Jewish communities of Late Antiquity and the early Middle Ages, and hence, the development of an increasingly united culture and religion" p.
To avoid skewing the discussion in favor of the calendar in use today, Stern defers a discussion of the rabbinic calendar until the final two chapters. The first three chapters of his study deal mainly with non-rabbinic evidence. His first task is to explain how solar calendars, the existence of which is well documented in the period of the Second Temple, fell into disuse by the end of the first century C.
The deviation of the day solar year from a real solar year makes it unlikely that the solar calendars evidenced in the Book of Enoch and Qumran were ever anything more than theoretical constructs.
Also Available In:
But whether theoretical or real, Jewish solar calendars fell out of favor by the end of the first century—this, at a time when, under the ascendancy of the Julian calendar, the solar calendar was becoming increasingly fashionable elsewhere. In Stern's view, the Jewish tilt to a lunar calendar may thus suggest "a deliberate attempt, on the part of first century Jews, to distinguish themselves from Roman culture and the increasingly expanding Graeco-Roman world" p.
The lunar calendar, whose adoption had earlier signified Jewish integration into the dominant Near Eastern culture, was now a sign of cultural difference. Without intercalation, the discrepancy between the lunar and solar year inevitably produces undesirable results, the most conspicuous of which would require the observance of Passover in a season other than spring. The present-day rabbinic calendar follows a system designed to [End Page ] ensure that Passover never occurs before the calendrical vernal equinox. But as late as the sixth century, there was no uniform system of intercalation.
In the lunisolar cycles mentioned in Qumran and in Jewish pseudepigrapha, it is uncertain whether the intercalations were empirical or cyclical. Stern also doubts that the equinoctial rule was as widely observed in the first century as has sometimes been postulated. Following a suggestion by E. Schwartz, he proposes that the late celebration of Passover in the period of the Second Temple was not connected with the equinox at all; it was only to allow time for pilgrims to reach Jerusalem. Since the destruction of the Temple removed the need to accommodate the travel plans of pilgrims, Jews after 70 allowed the observance of Passover to recede to earlier dates in the solar year.
Far from a fixed system, calendrical intercalation was simply an ad hoc expedient designed to compensate for discrepancies that had built up over time. Nor was there anything approaching a standard calendar, especially one reflecting the rabbinic model. In fact, Christian writers of the fourth century regularly accuse the Jews of ignoring the equinox. If there was any rule at all, it was simply to make sure that Passover was celebrated in the month of March.