The Christian Era of Dionysius Exiguus
The Christian Era of Dionysius Exiguus
Abstract and Keywords
The date of Dionysius Exiguus for the first year of Christ in AD 1 disagrees with almost all other ancient authorities. Many scholars state that Dionysius synchronized his year 1 with the year 754 from the foundation of Rome. That synchronism is a modem inference that nowhere appears in the writings of Dionysius. Some have argued that Dionysius incorrectly dated the years of Augustus. Another hypothesis holds that Dionysius generated his date by adding the 532‐year Paschal period to the presumed age of Jesus at his death of 3l years, then designating as 563 the year in the tables that produced a date for Easter on 25, March. Several scholars have claimed that Dionysius derived his date from Eusebius of Caesarea. One has argued that Dionysius designated as 2000 the year in which he thought the present age would come to and end. None of these hypotheses is persuasive.
The general consensus of authorities earlier than Dionysius is that Jesus was born in the year corresponding to 2 or 3 BC. The date is an inference from Luke's synchronism of the 15th year of Tiberius with the baptism of Jesus at the age of about 30. Most scholars have thought therefore that Dionysius did not follow an established tradition, but generated his own date for the Nativity. Either he misinterpreted the evidence or he deliberately distorted it for reasons of his own. Some more recent scholars have thought that Dionysius did accept an already established date, but they have offered no satisfactory explanation of its origin.
1. THE YEAR 754 FROM THE FOUNDATION OF ROME
Modern reference‐works often state that Dionysius dated the Nativity to December of the year 753 from the foundation of Rome (ab urbe condita) and counted the first year of the Lord from the beginning of the next Roman year in January of 754. Some authorities state that he dated from the Incarnation on 25 March of his year 1, AUC 754.
The Encyclopedia Britannica, in an unsigned article originally published in the 11th edition (1911), says, ‘He [Dionysius] wrongly dated the birth of Christ according to the Roman system (i.e. 754 years after the founding of Rome) as Dec. 25, 753’. The statement remains unchanged in the 2006 Britannica Online.1 The Catholic Encyclopedia (1907–12), in an article signed by John Gerard, says, ‘In chronology Dionysius has left his mark conspicuously, for it was he who introduced the use of the Christian era according to which dates are reckoned from the Incarnation, which he assigned to 25 March, in the year 754 from the foundation of Rome…’.2
(p.340) More recently, Friedrich Wilhelm Bautz (1990) says of Dionysius, ‘He mistakenly placed the birth of Christ in the year 754 ab urbe condita, and specifically at 25.12 of the 1st year of his era’. Jack Finegan said in the original edition of his Handbook of Biblical Chronology (1964: 132), ‘following a reckoning which was evidently current in his time, Dionysius placed these events [the Incarnation and Nativity] in A.U.C. 754 and for the beginning point of the era went back to Jan 1 of that year’. In the new edition (1998: 114), Finegan changed his mind about the year to which Dionysius assigned the Incarnation and Nativity, but retained the synchronism between AD 1 and AUC 754: ‘For the year of the incarnation Dionysius accepted the year AUC 753 and also for the day in that year of the nativity the date of December 25… Dionysius went on to the immediately ensuing first day of January (seven days after December 25), which was the commencement of the regular Roman year AUC 754 (= AD 1) to make this the beginning of the first year of his new era’.
As the variation among these opinions shows, scholars are not agreed on whether Dionysius meant to date the birth of Christ to 25 December of the year corresponding to 753 ab urbe condita, and that the year 1 should therefore start 1 week later on 1 January 754, or dated the Incarnation and Nativity to 25 March and 25 December of the year 754=AD 1. If Dionysius meant to number his years from the Incarnation in the sense of Christ's conception on 25 March, then it is likely that he dated the Incarnation to AD 1. Bede (DTR 47) seems to have understood the year of the Incarnation in this way. Among modern scholars, Ludwig Ideler (1825–6: ii. 384) argued in favour of this view.
If Dionysius meant to number the years from the birth of Christ, rather than the conception, then it is more likely that he considered that event to have occurred on 25 December, seven days before the start of the year that we call AD 1. The consensus of most modern scholars favours that interpretation (Ginzel 1906–14: iii. 179). In fact, we do not even know that Dionysius intended his numbered years of the Lord to be counted from 1 January. They replace years of Diocletian in a table in which those years were synchronized with the indictional year. As Leofranc Holford‐Strevens (2001) has pointed out, it is possible that the year 1 began on 1 September of 1 BC. The issue cannot be decided except on the basis of an answer to the question how Dionysius knew that Diocletian 247 should be followed by the year 532 of the Lord.
Given the frequency with which one encounters the claim that Dionysius Exiguus equated his year 1 with the year 754 from the foundation of Rome, one may be surprised upon actually reading his work to find that Dionysius nowhere offers this synchronism, nor indeed ever in any of his writings expresses a date reckoned from the foundation of the city. It is simply not (p.341) true that numbering the years ab urbe condita was, as the Encyclopedia Britannica puts it, ‘the Roman system’. Dating by consular years was the Roman system, and it is the link that Dionysius gives between AD 525 and the consulship of Probus that connects the ‘Years of the Lord’ to other systems of historical chronology. Only Orosius, among authors earlier than Dionysius, reports a date for the Nativity as an interval from the foundation of Rome. The singularity of that interval shows how unusual such dating was. Dionysius could not have derived his date from Orosius. Orosius (6. 22. 1) dated the birth of Christ to the year 752 from the foundation of the city.
The modern claim that Dionysius Exiguus set his year 1 as equivalent to AUC 754 derives from the practice among historians before the end of the nineteenth century of eschewing years BC and using instead Olympiad dates for ancient Greek history and years counted from the foundation of the city for Roman history. The error perhaps derives from misunderstanding of a statement by Ideler, who expressed dates in terms of numbered Olympiads or years from the foundation of Rome. Sometimes he converted the dates to Scaliger's Julian period or to years before Christ. In reference to the Christian era, Ideler concluded a lengthy argument by saying (1825–6: ii. 384), ‘It seems then clearly established: Dionysius set Christ's birth at the end of the first year of his era, the 4714th year of the Julian period, the 754th of the city of Rome’.
The point of Ideler's argument here is not that Dionysius equated his year 1 with the year 754 of Rome, but that he dated the incarnation and birth of Jesus to his year 1—not, as Scaliger and other scholars maintained, to the previous year. Ideler does not mean to say that Dionysius himself dated the Nativity to the year 754 of Rome any more than that he dated that event to the year 4714 of Scaliger's Julian period.
The probable source of the error in the English‐speaking world is the Handy‐Book of Rules and Tables for Verifying Dates with the Christian Era, by John James Bond, published in four editions between 1869 and 1889. Of the Christian era Bond writes as follows:
The first year, or 1 Anno Christi, called by chronologists annus verus, is the fourth year before 1 Anno Domini, hence we find that 33 Anno Christi corresponds to 30 A.D. This difference between the years known as Annus Verus and 1 Anno Domini was caused by Dionysius, when he introduced the present system of reckoning the Christian era, and made 1 A.D. correspond to the 46th year of the Julian era, and 754 A.U.C., instead of 750 A.U.C.=the 42nd year of the Julian era, and thus stopped four years short of the date which, according to the statement of Clement of Alexandria, had been esteemed the true date by the early Christians. (Bond 1875: 213.)
The ‘Julian era’ is another modern convention, referring to Caesar's reform of the calendar in 46 BC. It has an ancient precedent in Censorinus (20. 11, 21. (p.342) 7), who numbers ‘Julian years’ from 1 January of the fourth consulship of Julius Caesar (45 BC), the effective date of the reform. Bond does not mean to say either that Clement dated the Nativity to the 42nd year of the Julian calendar, 750 AUC, or that Dionysius used the expression AUC 754=AD 1.
2. THE 28TH YEAR OF AUGUSTUS
Even if it were true that Dionysius synchronized his year AD 1 with the year 754 from the foundation of Rome, we should know no more than that he synchronized the year 525 with the consulship of Probus and the third year of an Indiction. We should still want to know why Dionysius departed from the well‐established consensus of ancient scholars according to which Christ was born two or three years earlier.
According to Clement of Alexandria (Stromata 1. 21. 145. 5), Jesus was born in the 28th year of the emperor Augustus. Counted from the first year of Augustus at Alexandria in 30/29 BC, the 28th year is 3/2 BC. One theory to explain the alleged ‘error’ of Dionysius Exiguus is that he agreed with Clement in dating the Nativity to the 28th year of Augustus, but that he counted the years of Augustus from 27 BC, when he received the title of Augustus, instead of from his first year at Alexandria.
This hypothesis seems to have originated with Bond (1872: 74; 1875: 203–4). Bond invented an ‘era of Augustus’ dating from 27 BC and argued that Dionysius counted the 28th year of Augustus from that era. Werner Keller also popularized the theory, although without reference to Bond, in his Bible as History: A Confirmation of the Book of Books, first published in 1955 under the title Und die Bibel hat doch recht (p. 353 in the English translation of 1956). John Mosley refuted the idea in an article on ‘Christmas Errors’ published in 1981. According to Mosely, every schoolboy knew the story of the battle of Actium, ‘and a prominent historian working in Rome would not have made such a simple blunder’. Nevertheless, E. G. Richards (1998: 218) has recently reiterated this view:
Dennis had assumed from a reading of Clement of Alexandria that Christ was born in the twenty‐eighth year of the reign of the Emperor Augustus. He assumed that Augustus' reign began in 727 AUC—but there he was mistaken. What he did not realize was that the reign of Augustus was always calculated from his decisive victory over his rivals for power, Anthony and Cleopatra…at the battle of Actium fought on 3 [sic!] September 723 AUC, rather than his acceptance from the Roman people of the title of emperor on 13 January 727 AUC.
(p.343) We do not know that Dionysius had read Clement's Miscellanies. If he did, he would have found in that work, not only the statement that Jesus was born in the 28th year of Augustus, but also that Augustus ruled at Alexandria for a total of 43 years and that he defeated Antony at Actium in the year of his fourth consulship. Clement also reckons intervals from the first Olympiad to the battle of Actium and the fourth consulship of Augustus: 24 years to the foundation of Rome, thence 243 years to the expulsion of the kings, thence 186 years to the death of Alexander, and from there to the victory of Augustus 294 (Stromata 1. 21. 139, 144–5).
The fourth consulship of Augustus was 30 BC, which is the correct date for the beginning of his first year at Alexandria, but not for the battle of Actium itself. Cassius Dio (50.10, 51.1) dates the battle to 2 September of the consular year corresponding to 31 BC. Clement's interval of 243+186+294=723 from the foundation of Rome to the fourth consulship of Augustus yields AUC 724 for the first year of Augustus. Nothing in Clement's text would lead Dionysius or anyone else to conclude that the first year of Augustus should be synchronized with AUC 727.
In the first book of his History of the Church (1. 5. 2), Eusebius says that Jesus was born in the 42nd year of the rule of Augustus, and the 28th year after his defeat of Antony and Cleopatra. This is consistent with his equation in the Chronicle (p. 163 Helm) between the last year of Cleopatra and the 15th year of Augustus.
Whether from Clement or from Eusebius, Dionysius would have known to count the 28 years either from the battle of Actium or from the first year of Augustus at Alexandria, rather than from his proclamation as Augustus three or four years later. Furthermore, while 27 BC is the modern, scholarly date for the beginning of Augustus' principate at Rome, ancient authorities rarely used that date. Suetonius (Augustus 8. 2) says that Augustus held power alone for 44 years after the defeat of Antony. Cassius Dio (56. 30. 5) says Augustus died on 19 August, having ruled alone for 44 years, less 13 days, from his victory at Actium. Eutropius (7. 8) says that Augustus returned to Rome after the defeat of Antony and Cleopatra, twelve years after his first consulship, and from that time ruled alone for 44 years, his principate being 56 years in total. The only exception is Censorinus, who refers (21. 8) to ‘the years of the Augusti’ as beginning from the consulship of Caesar VII (i.e. Augustus) and Agrippa III (27 BC), when Augustus received that title. He also (22. 16) refers to the consulship of Marcius Censorinus and Asinius Gallus (8 BC), when the month Sextilis was renamed August, as the 20th year of Augustus. These Augustan years refer literally to the years since the conferral of that name and do not correspond to the years that the man so named held power in Rome. (p.344) Dionysius would have been quite alone in counting the regnal years of Augustus from 27 BC.
3. THE 15TH YEAR OF TIBERIUS
Another approach finds the source of Dionysius' error in his reckoning of the fifteen years of Tiberius. This hypothesis originated with James Ussher (1658, at the cosmic year 4015=AD 12) as a way to reconcile Luke's statement that Jesus was about thirty years old in the 15th year of Tiberius, AD 28/9, with the date for the death of Herod in 4 BC.
Clement of Alexandria (Stromata 1. 21. 145) said that Jesus was born in the 28th year of Augustus, that Augustus ruled for 43 years and that the 30 years from the Nativity to the Passion consist of fifteen years of Tiberius and fifteen of Augustus. In the view of Ussher, Clement—and all other early Christian authors after him—erred in counting the 15th year of Tiberius from his accession shortly after the death of Augustus in AD 14, instead of from AD 12, or even earlier, when Augustus made Tiberius his colleague in the exercise of the imperial power. If the first year of Tiberius was AD 12, his 15th year was AD 26, and Jesus was born 30 years earlier in 4 or 5 BC, and therefore within the reign of Herod. Pagi (1689: xiv) and Münter (1827: 3) offered similar arguments.
Philip Mauro (1922: 84–5) applied a variant of this view to Dionysius Exiguus, so as to account for his four‐year error. Unfortunately, Mauro himself committed a combination of arithmetical with typographical error:
Dionysius Exiguus calculated that the year of our Lord's birth was A.U.C. 753. He made his equivalence of dates from Lu 3:1, ‘Now in the fifteenth year of the reign of Tiberius Caesar’ etc., at which time Christ was 30 years of age according to Lu 3:23. But it was ascertained later that a mistake of four years had been made; for it clearly appears by Mt 2:1 that Christ was born before the death of Herod, who died in 749 A.U.C. Tiberius succeeded Augustus, Aug. 19, A.U.C. 767. Hence his 15th year would be A.U.C. 779 [sic!]; and from those facts Dionysius was right in his calculation. But it was discovered in later years that Tiberius began to reign as colleague with Augustus four years before the latter died. Hence the 15th year mentioned by Luke was four years earlier than was supposed by Dionysius and consequently the birth of Christ was that many years earlier than the date selected by Exiguus, which date has been followed ever since.
Suetonius and Velleius Paterculus say that the Roman Senate, at the request of Augustus, conferred upon Tiberius equal power in the management of the provinces and the armies. Velleius Paterculus (2. 121) reports the matter after the disastrous defeat of Quinctilius Varus in Germany and before the (p.345) campaign into Germany of Tiberius and Germanicus. Suetonius (Tiberius 20–1) says the senate passed such a decree two years after that campaign. Cassius Dio (56. 25) dates the campaign to the consulship of Marcus Aemilius and Statilius Taurus, AD 11. Thus Suetonius supports Ussher's date in AD 12, and Velleius Paterculus supports Mauro's date in AD 10.
As Ludwig Ideler pointed out (1825–6: ii. 418), such a reckoning of the reign of Tiberius is a strictly modern conjecture for which there is no literary or numismatic precedent in antiquity. In fact, if one were to count the years of Tiberius differently than from the death of Augustus, the most legitimate terminus would be his assumption of the tribunician power. Imperial coinage often bears a legend with a numbered year of the emperor's tribunician power (Mattingly 1930). Suetonius (Tiberius 9. 10, 11. 3) says that Tiberius received a grant of the tribunician power about the time of his second consulship (7 BC) and before his retirement to Rhodes. Coinage shows Tiberius in the 12th year of his tribunician power as early as AD 10 (Barnes 1974: 24).
Manipulation of the regnal years of Tiberius cannot produce a credible explanation for the choice of the year we call AD 1 as the first year of Christ.
4. THE GREAT PASCHAL PERIOD
One of the most widely held views among scholars is that Dionysius calculated his date for the year 1 with the assistance of the Paschal period of 532 years.
i. Joseph Scaliger
Scaliger took notice of the remarkable coincidence that Dionysius numbered the first year in his Paschal table as the year 532. The number 532 is the interval of the great Paschal period—the period after which the phases of the moon will return to the same day in the year and the same day of the week in the Julian calendar. It derives, as Bede explained (DTR 65), from multiplying the 19 years of the lunar cycle by the 4 years of the leap‐year cycle and the 7 days of the weekday cycle. On Scaliger's hypothesis (1583, book 2: 157–63), Dionysius thought that the Incarnation and Nativity should have occurred in the first year of a 532‐year period. Accordingly, Dionysius numbered as 1 the first full year of Christ's life on earth in the second year of a cycle.
Scaliger's theory rests partly on the mistaken assumption that Dionysius followed the example of Victorius of Aquitania and composed a Paschal table covering a full cycle of 532 years. St Bede (DTR 65) extended the 95‐year table (p.346) of Dionysius through an entire Paschal period of 532 years, beginning with the year 532 and ending with the year 1063. When Joannes Noviomagus (Johannes von Bronckhorst) first published these tables in 1537 he attributed to Bede the composition of eighty‐four 19‐year cycles (three 532‐year periods), encompassing 1596 years from the birth of Christ to 1595 (PL 90. 826). Misled by Noviomagus and apparently misunderstanding Bede, Scaliger wrongly supposed that Dionysius had drafted tables to the year 1063 and that Bede had extended them to 1595. Indeed, some of the manuscripts of Dionysius' tables carry the heading, ‘Here begins the book of Dionysius Exiguus on the great Paschal cycle of 532 years’ (Krusch 1938: 63). Denis Petau, in book 2 of his work (1703: i. 116), agreed with Scaliger that Dionysius had composed a 532‐year table. By the time he wrote book 12 (1703: ii. 222) Petau had relized the mistake and insisted that Dionysius wrote but one 95‐year cycle. Antoine Pagi (1689: vi) tried to revive the theory, arguing that the 532‐year table began with the year of the Incarnation and was something separate from the 95‐year table that began in 532. Johann Wilhelm Jan in his Historia Cycli Dionysii, published in connection with his 1718 edition of Dionysius Exiguus, finally put to rest the notion of a 532‐year Dionysian period (PL 67. 473–7.)
The idea that Dionysius had composed a 532‐year table actually predates Scaliger by some 700 years in a commentary on Bede's Reckoning of Time. It is preserved as a set of marginal notes in some of the manuscripts. The commentator states that there are ‘now from the Incarnation 873 years’. In the preface to his edition of the text, C. W. Jones (1977: 257–61) suggested the writer might have been Martin the Irishman working at the scriptorium in the cathedral of Laon.
Bede had said that Dionysius, ‘placing the 532nd year of our Lord's Incarnation at the beginning of his first cycle, plainly taught that the second year of his cycle was the same as that when the mystery of the same most holy Incarnation began’.3 Bede apparently thought that Dionysius knew of the 532‐year period, but he did not attribute to him the composition of a 532‐year Paschal list. The commentator misunderstood:
He [Dionysius] composed 28 circuits of 19 years in his great cycle. And in the first of those circuits, immediately after the first year, he placed the year of the incarnation of the Lord—that is the year that corresponds in all conjunctions of the sun and the moon to the year in which the Lord was incarnate. He did not begin from the very year in which he was incarnate, because it was not the first year of a 19‐year cycle. For the (p.347) cycle runs through 28 years and therefore it was necessary that he should begin from the first year of a 19‐year cycle.
ii. Gustav Oppert
Gustav Oppert (1900: 116–17) recognized that Dionysius did not compose a table of 532 years, but nevertheless argued that he knew of the great Paschal period and used it in fixing the era of the Incarnation. On Oppert's theory, Dionysius knew that Jesus had been born a little more than 500 years before his own time. He also accepted the Alexandrian tradition that Jesus rose from the dead on Sunday, 25 March. So he looked at his 95‐year extension of the Alexandrian tables to find the first year when appropriate data for the Resurrection would recur—a year when Easter would fall on a Sunday, 25 March, shortly after the 14th day of the moon. He found those data at what would correspond to the year 279 of Diocletian. Dionysius accepted the shorter chronology for the career of Jesus and assumed that he would have been 31 years old at the time of his death. Applying the 532‐year period, Dionysius decided that the year corresponding to Diocletian 279 should be numbered as 31+532=563. He accordingly numbered the first year of his table, corresponding to Diocletian 248, as the year of the Lord 532. Oppert's argument has been widely accepted.4
Georges Declercq (2002: 213–24) has discussed Oppert's hypothesis in detail and argued that Dionysius neither used the great Paschal period for any computistical purpose nor recognized it as a true cycle. He acknowledges that Dionysius knew the work of Victorius of Aquitania and his claim that the data for Easter would repeat after 532 years. But since Dionysius regarded Victorius' calculations as fundamentally flawed, he probably also rejected the claim of a 532‐year period. In the prefatory letter to Petronius, Dionysius points out (Krusch 1938: 64) that a cycle of 95 years does not constitute a true period. Had Dionysius accepted the 532‐year period of Victorius, Declercq says, he would certainly have mentioned it in this context.
Declercq also argued, quite rightly, that Dionysius, working in the sixth century, would not have followed the short chronology for Jesus' public ministry. Prosper of Aquitania (Mommsen 1892: 409) affirmed the validity of the long chronology some 70 years before Dionysius published his tables. In his chronicle written shortly before Dionysius Exiguus published the Paschal tables, Cassiodorus accepted the year corresponding to AD 31 as the date of the Passion, but he dated the Nativity 33 years earlier to 3 BC (Mommsen 1894: 135–7).
The Alexandrian cycle had a base‐date corresponding to the first year of Diocletian. The numbered year of Diocletian when divided by 19 therefore yields as a remainder the number of the year in the cycle—the so‐called ‘Golden Number’ (see Ch. 6). Marius Chaîne (1925: 64–6) argued that Dionysius wanted the difference between his new Christian era and the era of Diocletian to be a multiple of 19, so as to preserve this computational tool. The difference between 532 and 248 is 284, not 285. Chaîne accounted for the discrepancy by appealing to the difference between the Alexandrian year beginning 29 August and the Roman year beginning 1 January. ‘The years of the era of Diocletian give us the same Golden Number as the years of the Dionysian Christian era, taking account of the difference due to the change in the starting‐point of the epacts.’
The arithmetic will not support the hypothesis. Dionysius should have made his table begin with 533 if he wanted to maintain the same correspondence with respect to the Golden Number.
6. EUSEBIUS AS POSSIBLE SOURCE
For anyone working in Rome in the early sixth century, the obvious place to look for a date for the birth of Jesus or for an interval between that date and the first year of Diocletian was Jerome's Latin translation of the Chronicle of Eusebius. In that work, the interval between the Nativity in the 42nd year of Augustus, at the third year of the 194th Olympiad, and the first year of Diocletian, at the second year of the 266th Olympiad, is 287 years—not, as in Dionysius' system, 284 years. Three hypotheses have nevertheless been offered explaining how Dionysius might have derived his date for the Nativity from Eusebius.
i. Gustav Teres
Gustav Teres (1984), a Norwegian astronomer, sought to rescue Dionysius' reputation for mathematical precision by arguing that he was a careful scientist who made no mistake at all. Teres asserts that ‘Dionysius was neither a chronologist nor the establisher of the Christian era’ and that ‘it was neither his intention nor his task to determine the birth year of Christ’. In this claim (p.349) eres is fundamentally correct. His method for showing how Dionysius obtained a date from the Chronicle of Eusebius is, however, seriously flawed. According to Teres, Dionysius knew that the consulship of Probus was equivalent to the year AUC 1278. He then consulted Eusebius and found that Augustus had died on 18 August 767, and that the 15th year of Tiberius therefore began in August of 782 and ended in August of 783. Dionysius simply subtracted 30 from 783 to find that Jesus was born in 753 and that the year 1 should therefore begin a week later on 1 January 754. The consulship of Probus in 1278 would correspond to the year 525 from Christ.
As Declercq (2002: 228–9) has pointed out, Teres himself makes a mathematical mistake in computing the 15th year of Tiberius. To get the 15th year one should add 14 to 767, not 15.
Teres does not state what edition of the Chronicle he is using. Neither Eusebius nor Jerome included a column of years numbered from the foundation of Rome. The only modern editor to supply such numbers was Angelo Mai in his 1833 edition of the Chronicle of Jerome, reprinted in volume 19 of the Patrologia Graeca. Mai noted years of Rome every 10th year beginning at Abraham 1264. The death of Augustus appears at Rome 765. In short, as Declercq notes, the data that Teres claims for the Chronicle of Eusebius cannot in fact be found in any version of that text.
ii. Georges Declercq
While rejecting Teres's solution, Declercq (2002: 230–46) has himself argued that it was from the Chronicle of Eusebius that Dionysius derived his interval from the birth of Jesus to the first year of Diocletian. Declercq says that ‘the key’ to this whole problem can be found in a text written by Felix, the abbot of the monastery of Gillitanus in North Africa. Victor of Tonnena says that Felix was driven into exile in 553 and died in Sinope on the Black Sea in 557 (Mommsen 1894: 203–4). The text in question is a prologue prefixed to an Easter table covering the years 627 to 721, under the heading ‘preface of Felix, Abbot of Cyrillitanus’. Jones (1943: 73–4) argued that Felix wrote the preface when he introduced the 95‐year table of Dionysius Exiguus to North Africa. The text (PL 129. 1331) can be translated as follows. The heading appears only in one manuscript (Oxford, Bodleian Library, Digby 63):
Whence Dionysius took the beginning of his computation
In the middle of Olympiad 194, the 42nd year of the emperor Octavian Augustus, our Lord Jesus Christ was born in the flesh. From the Nativity of the Lord to Olympiad 265, that is to the first year of Diocletian, there are 284 years. And from the first year of Diocletian to the first cycle of Dionysius you will find 248 years, for a total of 532 (p.350) years. From here Dionysius in the year 532 from the Incarnation of the Lord began his first cycle. For the cycle of Saint Cyril ended in 247. Add 1, where Dionysius begins, and you will find 248 years.
Declercq argues that this reckoning by Olympiads must have been based on the Chronicle of Jerome. In that work one will find the 42nd year of Augustus noted ‘in the middle’ of the 194th Olympiad (Olympiad 194. 3). The first year of Diocletian, however, will be found not in Olympiad 265, but at the second year of Olympiad 266. Thus Jerome has 72 Olympiads and 288 years between the birth of Jesus and the first year of Diocletian, instead of the 71 Olympiads and 284 years that Felix claims. Declercq believes that Felix must here be repeating uncritically an explanation that he learned from someone else—namely, from Dionysius Exiguus himself. According to Declercq, Dionysius correctly calculated an interval of 287 years from the Chronicle of Jerome, but then ‘deliberately tampered’ with the data ‘in order to arrive at a date that was ideal from a computistical point of view’. Having altered the interval to 284 years, Dionysius then took advantage of the lack of mathematical skill in the West at that time ‘to convince those, who, like Felix, were eager to know how he arrived at the date AD 532 that it was simply based on a conversion of the Olympiads in the well known Chronicle of Eusebius/Jerome’.
Declercq explains the reason for the tampering. The interval of 287 years that Dionysius found in the Chronicle of Jerome between the 42nd year of Augustus and the first year of Diocletian suggested that the first year of Diocletian should correspond to the year 288 from Christ and that Diocletian 248 should be numbered as 535 from Christ. But Dionysius wanted that number to be divisible by 4, in order to preserve the rule, already embedded in the numbering of the years from Diocletian, that leap‐years would be those years in the table whose number was divisible by 4 without a remainder. The year 536 would have been the candidate closest to the Eusebian interval between the Nativity and Diocletian, Declercq argues, but ‘Dionysius must have been drawn as a magnet’ to the year 532 instead. For although Dionysius did not understand the construction of the great Paschal cycle, he must certainly have noticed that 532 was the product of the 19‐year lunar cycle times the seven days of a week and the four years of the intercalation. Accordingly, with the first year of his table numbered as 532, that year and every first year of a cycle thereafter would be divisible by 19 without a remainder. ‘Thus’, Declercq concludes, ‘Dionysius was first and foremost guided by computistical and practical rather than historical considerations when he decided to synchronize the 248th year of Diocletian with the year AD 532.’
Declercq argues that Dionysius would have liked the resulting year 1 for another reason. In that year he could have computed that the Incarnation (p.351) took place under the same calendrical data as the traditional Roman date for the Passion—Friday, 25 March, on the same day of the week as the creation of the first man, and on the same day as the traditional Roman date of the vernal equinox. Dionysius' remark in the letter to Petronius that the years numbered from the Lord will help remind people both of the ‘commencement of our hope’ and of ‘the Passion of our Redeemer’ shows, according to Declercq, that ‘Dionysius was well aware that the date he proposed for the incarnation coincided with the traditional date for the crucifixion in the West’. Declercq thus agrees with Ideler that Dionysius numbered as 1 the year of the Incarnation and birth of Christ—not, as Scaliger, Oppert, Ginzel, and most other scholars maintain, the first full year of his life on earth beginning a week after the Nativity.
Declercq is certainly right that the data in the preface of Felix must have been drawn from the Chronicle of Jerome. The use of reckoning by Olympiads suggests that conclusion. The notion that Felix received this information from Dionysius Exiguus, rather than directly from Jerome, and that Dionysius deliberately deceived Felix about the origin of his 284‐year interval strains credulity. We should rather regard Felix as the first of a long line of scholars to try to figure out how Dionysius chose a date in which Dionysius himself shows so little interest. Felix must have looked in the Chronicle of Jerome himself to try to verify the interval of 284 years. When he found the Nativity in Olympiad 194 and the first year of Diocletian in Olympiad 266, he decided to count the intervening Olympiads beginning with Olympiad 195 and ending with Olympiad 265, for a total of 71 Olympiads and 284 years.
Declercq's argument rests less on the text of Felix and data drawn from the Chronicle of Jerome than on computistical considerations connected with the interval of 532 years. Declercq rejects Oppert's thesis that Dionysius used the 532‐period to calculate the year of the Passion, but nevertheless maintains that Dionysius must have been attracted to the number 532 like a magnet. The only substantive difference between the two hypotheses is that Oppert believes that Dionysius understood the 532‐year cycle as the product of 19 and 28, while Declercq argues that Dionysius was unfamiliar with the 28‐year solar cycle, but recognized 532 as the product of 19×7×4.
Declercq's comment about the calendrical data for 25 March in Dionysius' year 1 is interesting, but no more persuasive than his claim that Dionysius was fascinated with the number 532. It is true that some Paschal calculators liked to find parallels among the calendrical data for the Creation, the Incarnation, and the Passion. The problem is that the data for the year 1 do not in fact anticipate the data for the Passion. In the year 1, which would have data (p.352) equivalent to the year 533 in the Dionysian table, Friday, 25 March, was the 14th day of the moon. In AD 29, the traditional Roman date for the Passion, corresponding to the year 561 in Dionysius' table, the Paschal full moon would fall on 15 April. Instead of inspiring the reader to think about the Passion, a date for the Incarnation on 25 March of the year 1 would only lead the reader to discover, as Bede did to his dismay (DTR 47), that the data in the tables do not produce a 14th or 15th day of the moon on 25 March in the year 29, 31, 34—or any other year within an acceptable interval.
iii. Daniel McCarthy
McCarthy (2003) has claimed that Dionysius was one of several authors who received a system of numbering the years from the birth of Christ that originated with Eusebius, but not in the Chronicle.
From the silence of Dionysius about how he knew to synchronize Dionysius 248 with the year 532 of the Lord, McCarthy concluded (2003: 38) that he must have ‘had access to a source which had already identified both the Julian year which we call AD 1, and the principle of dating events by counting the years from that epoch’. As evidence for the existence of such a system, McCarthy adduces the entry for the birth of Christ at the consulship of Caesar and Paullus in the Chronograph of 354 and a set of entries in the Irish Annals of Tigernach.
In the fragments of the Annals of Tigernach for the period between the birth of Christ and the reign of Constantine, there is a set of notes numbering the current year from the creation of the world and from the Incarnation. These notes appear at years numbered from the Incarnation as 10, 20, 30, 63, 76, 96, 115, 134, 267, and 324. Some of the numerals are corrupt, and sometimes the position of the notice does not correspond with the numbering in the text.
McCarthy (1998; 2003: 44–6) has shown that the notes for the years 10 and 30 are interpolations and that the original series appeared at 19‐year intervals beginning with the year 1 from the Incarnation and continuing through the series 1, 20, 39, 58, 77, 96, 115, 134, and so on. The numbers 267 and 324 come from a different manuscript, but belong to the same series of 19‐year intervals. That the year 1 in these annals corresponded to the same year as the Dionysian AD 1 follows from the data indicating the weekday of 1 January for the various years.
The fact that these entries appeared at 19‐year intervals suggests, according to McCarthy, that they were derived from a Paschal table. They cannot have been derived from the Paschal table of Dionysius Exiguus, because the series numbers as 1 what was the second year of a Dionysian cycle. McCarthy (p.353) believes (2003: 46–51) that the likeliest source was an otherwise lost Paschal table of Eusebius of Caesarea. The Annals of Tigernach have a note at the year corresponding to AD 309 that Eusebius of Caesarea composed a 19‐year cycle. Jerome (de viris illustribus 61) attributes such a table to Eusebius. In a letter preserved by Bede (HE 5. 21), Ceolfrid also attributed a Paschal list to Eusebius, although without reference to a 19‐year cycle.
McCarthy argues that both the entry at the year 309 and the series of notes at 19‐year intervals beginning in AD 1 came to the Annals of Tigernach through a now lost chronicle composed by Rufinus of Aquileia. He suggests that Rufinus used the Chronicle of Eusebius in its original Greek text, his own translation of the Ecclesiastical History, and several other sources. McCarthy believes that this work was the common source for entries in the Irish annals and in Bede's chronicle that are not included in Jerome's chronicle. McCarthy (2008) has explored that thesis further in a paper on Bede's chronological sources.
As evidence that Eusebius was interested in establishing a Christian era, McCarthy adduces the singularity of the fact that in the Chronicle Eusebius includes an explicit count of the years since Abraham only in the notice for the Nativity. He also takes Eusebius' interval in the Ecclesiastical History of 305 years counted from the birth of Christ to the outbreak of the Great Persecution as representing the use of a Christian era. McCarthy points out that the year 305 since Christ fits the series of numbers appearing at 19‐year intervals in the Annals of Tigernach beginning with the Christian year 1.
Following Burgess (1997), McCarthy believes that Eusebius composed the Chronicle in its first edition between 304 and 311, then the Paschal table between 309 and 312, followed by the Ecclesiastical History in its first edition about 313. In the Chronicle, McCarthy says, Eusebius dated the Nativity to the year 2015 of Abraham and the destruction of the churches in the 19th year of Diocletian to 2320, an interval inclusively counted of 306 years, while in the Ecclesiastical History he dated the 19th year of Diocletian to the 305th year from the Nativity. McCarthy suggests that it was the construction of the Paschal cycle that led Eusebius to make this revision. This latter equation also appears in the Annals of Tigernach, where the destruction of the churches is noted at the year corresponding to AD 305.
McCarthy argues that one of Dionysius' predecessors in the transmission of the Alexandrian cycle accepted from the Paschal table of Eusebius the date for the Incarnation corresponding to AD 1. That person corrected the Eusebian equation between Diocletian 19 and AD 305 by recalibrating the tables such that the 19th year of Diocletian now corresponded to the Julian year that we call AD 303. Dionysius found a note to this effect in the prologue to the Cyrillan table.
(p.354) The weaknesses in this ingenious set of arguments are manifold. We do not know that Eusebius composed a Paschal table of any kind. Jerome thought that he did, but he was mistaken (see Ch. 10). The hypothesis of a lost chronicle of Rufinus as the source for the entry in the Annals of Tigernach at the year 309 is plausible. But Rufinus or whoever was the source for this entry is more likely to have received the idea that Eusebius composed a Paschal table from Jerome, rather than from a copy of the Chronicle of Eusebius that included a notice absent from both Jerome's translation and the Armenian version.
The manuscripts that preserve this portion of the Annals of Tigernach date from the twelfth and fifteenth centuries. The correspondence between the year 1 of the Incarnation in this text and the Dionysian AD 1 is therefore likely to derive from the Dionysian Christian era. The fact that there is a series of entries at 19‐year intervals beginning at 1, instead of 2, does not nullify this likelihood.
Furthermore, there is in fact no change in the Eusebian interval from the Nativity to the 19th year of Diocletian between the Chronicle and the Ecclesiastical History. In Jerome's version of the Chronicle, the two entries appear (pp. 169, 227 Helm) at Abraham 2015 and 2320, as McCarthy states. The Armenian manuscript breaks off at the 16th year of Diocletian, aligned with the year 2319 from Abraham (p. 227 Karst). The 19th year of Diocletian would therefore be Abraham 2322. Jerome's version is presumably the more reliable witness to the original Eusebian alignment of these entries. If 2015 is the year 1 of Christ, then 2320 is the year 306. In the Ecclesiastical History, Eusebius does not in fact equate the 19th year of Diocletian with the year 305 from Christ. What he says is that his narrative in the first seven books covers 305 years from the birth of Christ to the destruction of the churches and that he will proceed in the next book to the terrible ordeals of his own time. In the eighth book, he describes those ordeals which began, he says, with a decree issued in the month of March during the 19th year of Diocletian. The span of 305 years is therefore the period covered in the first seven books, and the eighth book begins in the next year, the 19th year of Diocletian and the year 306 from the birth of Christ.
7. V. GRUMEL
Like Oppert, Grumel (1958: 224) suggested that Dionysius followed the shorter chronology for Jesus' ministry and accepted the traditional Alexandrian date for the Resurrection on 25 March of the year corresponding to AD (p.355) 31. Dionysius then subtracted the 30 or 31 years of Jesus' lifetime, choosing to designate the year that he did as number 1 because it was the first year of a Julian leap‐year cycle.
Grumel's theory suffers from some of the same weaknesses as Oppert's. We do not know what date Dionysius accepted for the Resurrection, and it is unlikely that Dionysius would have followed the shorter chronology. Nevertheless, as we shall see in Ch. 17, Grumel's suggestion has merit if applied as part of the solution to the question of the Christian era of Julius Africanus.
8. SEPP ROTHWANGL
Sepp Rothwangl offered a theory of an entirely different kind in a paper presented to a conference on ‘Cosmology through Time’ at the Astronomical Observatory of the University of Rome in June of 2001. Rothwangl believes that Dionysius wanted to alleviate anxiety arising from the fact that the year 6000 in the chronology of Julius Africanus had already passed. He decided to postpone the second coming of Christ from the year 500 since the Incarnation to the year 2000. He observed a planetary alignment on 31 May in the year corresponding to 531. Dionysius calculated that a conjunction of all planets would occur some 1469 years later on 5 May. He identified that event with the end of an age and assigned to the year the number 2000. The year of the observed phenomenon was therefore 1469 years earlier in 531.
Rothwangl is apparently unaware that Dionysius completed his work and published his tables in the consulship of Probus, the year AD 525. Rothwangl attributes to Dionysius both an astronomical knowledge and an interest in millennialism for which there is no direct evidence.
In the letter to Petronius, Dionysius remarks on his decision to renumber the years with words so unobtrusive as to suggest that his date for the Incarnation was not an innovation:
Because St Cyril began his first cycle from 153rd year of Diocletian and ended the last in the year 247, I began from the year 248 of that same man—tyrant, rather than prince. I did not wish, however, to perpetuate the memory of that impious persecutor in my cycles, so I have chosen rather to number the years from the Incarnation of our Lord Jesus Christ. (Krusch 1938: 64.)
(p.356) As Leofranc Holford‐Strevens has stated (1999: 778), Dionysius does not ‘explain or justify the underlying date, or even claim it for his own discovery, but treats it as an unproblematic fact, corresponding to current knowledge or belief’. Although Daniel McCarthy's identification of the source of that common knowledge as Eusebius of Caesarea cannot be sustained, his intuition that Dionysius found his equation between years of Diocletian and years since Christ in his Alexandrian sources points in the right direction.