Counting years by the Christian Era and days of the month by the Christian Calendar have come to be recognised as such useful conventions that questions about their origins and reckoning do not prevent their use in most parts of the world and in fields of scholarship which have little to do with Christianity or with Mediterranean and European history. Rabbis don't like it. Historians have problems with it. Physicists want to start over. Astronomers ignore it at their peril. But even the great mathematician, Carl Friedrich Gauss, could not do better.
The origin of the Chrisian Era is vague and its beginning point is uncertain, but it has become the Common Era for past, present, and future. This was not always so. Georges Declercq explains how all this came about in a short and very useful book.
He begins the story with Scriptural and Patristic evidence buried in the memories of early Christian writers about possible dates for the life and death of Jesus. That evidence is piecemeal and contradictory, for they lived with whatever calendar eras were in local use. It was two or three centuries before they began to create their own Era in both Latin and Greek parts of the Roman Empire and in areas beyond the reach of Rome, Alexandria, Jerusalem, or Constantinople, the four apostolic sees. Any calendar era is a time-series which provides order during a year and in a long series of years. Those who tried to create an Era for the Christian communities wanted to keep track of the Easter celebrations of Jesus' birth, ministry, death, and resurrection and to predict future celebrations accurately.
What could they learn from their own records? Not much. The earliest Gospel and the latest (Mark and John) were interested in the life of Jesus and his death, but not his birth. The Gospel of Mathew added announcements, birth stories, and scenes of recognition in childhood as part of a literary framework, whereas Luke the physician coordinated the same memories more tightly. Thus, only Mathew and Luke place the birth within the reign of Herod, a local king who died in 4 B.C. (N.B. I shall have to express some dates by our reckoning, in order to avoid some complications of the narrative.) But Luke also referred to a census in that year, ordered by Augustus Caesar and directed in Judea by Quirinus, governor of Syria, apparently during A.D. 6 and 7. Thus, his data appear to conflict with themselves and were probably drawn from two different sources. The Synoptic Gospels (Mathew, Mark, and Luke) repeat many of the same stories in the same terms, though not always in the same order. It looks very much like Mathew and Luke shared two documentary sources, one of which was Mark.
All the records say that Jesus and his disciples were Jews, shared in Jewish religious feasts, and went up to the Temple in Jerusalem to pray. They celebrated the Passover at the end of Jesus life. That should provide a secure date for calendar making, for Passover would always fall on the full moon (luna 14) of the month Nisan according to the Hebrew lunar calendar. It is not surprising that the Synoptic Gospels agree that Jesus celebrated Passover and was executed on a Friday, one day later; in that year therefore luna 14 was on Thursday. What is surprising is that John placed that day of Passover, luna 14 of the month Nisan, firmly on a Friday, the same day on which Jesus died. Declercq distinguishes rightly between data from the Synoptics and data from the Gospel of John.
The Christian Gospels do not provide a long year count, but the events recounted by the Synoptics could all have occurred within a single year. Yet John described Jesus' Ministry extending over three years and specified that his death occurred in the 15th year of Tiberius' reign. Thus, the earliest records give dates from four time-series, and it is surely not obvious how they could be synchronised with each other.
It was almost two centuries before Christian scholars tried to settle some of those problems. Sextus Julius Africanus (ca. A.D. 201) and Hippolytus, a presbyter of Ostia also active in Rome (ca. A.D. 222), each thought that Jesus was born annus mundi 5500, a fifth era which was counted from the Creation, according to numbers in the Greek Septuagint translation of the Hebrew Scriptures. Annus mundi 5500 appears to correspond with our year 2 B.C. (There are problems about an arbitrary beginning and the relation of Hebrew lunar years with Roman solar years, but the era of anni mundi is roughly coherent.) They also thought that he died annus mundi 5531, or about A.D. 29. Yet Sextus also gave year 16 Tiberius and Olympiad 202 for the death; by our reckoning the one might be about A.D. 29/30 but the other would p robably be 30/31. Writing in Greek, Hippolytus however placed Jesus' death firmly on Friday, luna 14 of Nisan in solar year 15 of the reign of Emperor Tiberius, following the Gospel of John. It was apparently Hippolytus who first coordinated Passover and the vernal equinox with 25 March of the Roman calendar, adding that not only the death but also the birth of Jesus occurred on the full moon of 25 March.
Tertullian agreed with Hippolytus that the death fell on Friday, 14 Nisan and 25 March in year 15 Tiberius, adding that this was the consular year of Rubellius and Rufius Gemini and that Jesus had lived for about 30 years. Counting back, he could place the birth in year 41 of Augustus or about 3 B.C. in our terms.
There was also an Eastern tradition among Christians in Syria and Asia Minor which interpreted the conflicting data of the Gospels differently and saw no reason not to celebrated both death and resurrection on the same day: the full moon of Passover, luna 14. Some Westerners cast aspersions on the observation of resurrection on the 14th moon of the Hebrew month Nisan, saying that those "Quartodecimans" were "celebrating with the Jews". In order to find a way out of these complications, Annianos, writing in Greek at Alexandria, proposed that Jesus observed Passover (14 Nisan) on Thursday 22 March and was crucified on Friday 23 March in the Roman calendar, so that one could celebrate both death and resurrection on the same day (15 Nisan), but not with the Jews. He placed those events in a year which would correspond approximately with our A.D. 42.
The two equinoxes and the two solstices were important for all local calendars which were of course seasonal. For the first time, Annianos also applied the Roman vernal equinox (25 March) not to the birth of Jesus but to the day of Annunciation by the angel Gabriel to Mary that she could conceive and bear a child. This was a step toward replacing the Roman "Festival of the Unconquerable Sun" at Winter Solstice (25 December) with a Christian festival of the birth of Jesus, an odd idea which had a future, both sacred and profane.
How did Annianos come to year 42? The full moon of Passover might or might not be counted as luna 14, but it was supposed to be on 25 March in the Julian calendar of the Roman Empire. During the second quarter of the century in which Jesus lived, 25 March could have fallen on a Friday only in years which by our reckoning are numbered 29, 35, 40, and 46; but it could have fallen on a Sunday in years 25, 31, 36, and 42. Amongst Latins, year 29 was in favour, whereas years 31 and 42 were thought to be more fitting by the Greek and Eastern communities of Christians. Annianos organised all the details to fit year 42, and that prevailed as the year of Jesus' death and resurrection in early Byzantine tradition. It is interesting that later reformers of the Latin calendar, Heriger of Laubach (A.D. 990) and Gerland of Besancon (1038-1057) preferred A.D. 42, as being in accord with Gospel truth, and it was offered as a worthy alternative also by Abbo of Fleury (1002-4) and Robert Grosseteste in the mid-thirteenth century.
Perhaps here it could be noticed that, for all of these computists, each day had 24 equal hours; but the count from the first hour began at noon in Alexandria, at dusk when the moon became visible to rabbis in the synagogues, but at midnight for Latin observers in Africa and Italia. This was not mentioned by Declercq, but for application of lunar limits for the Paschal full moon in discussions between Alexandria and Rome, as well as for many other details of Easter tables, it made a big difference, and the neglect of those differences caused many misunderstandings.
The use of luni-solar cycles to adjust lunar months to the solar calendar is well explained by Declerq in terms of decimal calculations. Happily, he also restates the comparison of those cycles during 8, 84, 19 years in terms of days, hours, and minutes which were common for early Christians and other Europeans, as well as for us. (Nothing is gained by using decimal fractions in these computations.) Some Christians tried each kind of cycle before they eventually settled on the 19 year cycle because it would carry on the longest with the smallest loss of synchronisation of sun and moon. An error of only one full day would accumulate after 322 years of 19 year cycles, and such a long-series error was acceptable.
Facing the same question as did all bishops as to when to advise the churches to celebrate Easter, Eusebius of Caesaria also tried to synchronise these divergent data in a 19-year Easter Table and in his Greek Chronological Canons, according to the report of Jerome. It is from Eusebius that we learn of the work by a former Alexandrian scholar, Anatolios, who became bishop of Laodicea (A.D. 269-280). From the venerable Bede we learn that Anatolios accepted the vernal equinox on 22 March rather than 25 March of the Roman calendar, that is, the date used by his compatriot, Ptolemaios of Alexandria. In his time, the scholars of Alexandria were using Easter Tables based upon an 8-year cycle, notably Demetrios (d. 231) and Dionysios (d. 264) who became patriarchs. In fact the marble tables of Hippolytos (d. 234) which you may see beside the steps leading up to the elevator, upon entering the Vatican Library, used a double or 16-year cycle in Rome. Those data cannot be interpreted in terms of the later introduction of the more efficient 19- year cycle. Eusebius' reports about Anatolios have attracted modern Irish scholarship because his work may have been the basis for a later Irish Easter table about A.D. 590, but always on the assumption that he used the 19-year cycle which never quite works out. In my opinion, the fragmentary data surviving from the work of those earliest Christian scholars should all be reassessed with this in mind, that the 8-year and 16-year cycles were applied for a long time before earliest evidence of the 19-year cycle appears in Alexandria about A.D.303/4 (or was it 322/3?).
Contrary to numerous reports by modern writers, Decercq makes it clear that surviving records from the Council of Nicaea (A.D. 325) show that its "318" bishops and abbots had little to do with any of this, other than expressing the need for Christians of the whole world to worship together. Other bishops and abbots are still trying to do that today, I believe.
Declercq explains these and even more dreadful complications in good style, passing over Victorius of Aquitaine in order to deal with his main theme: Dionysius Exiguus. The contribution of that Scythian canonist in the Roman curia has been explained well by C.W. Jones (1939, 1943) and W.M. Stevens (1993, 1995), but Declercq's work discusses more fully two "eras" which come into play: Alexandrian and Byzantine. He presents the Alexandrian system as "remarkably coherent", adding that "The same cannot however be said of the Roman system". (57)
It is too simple however to say that there was only a single "Roman system", for the skimpy data of tables and practices which survive from Carthage, Rome, Ravenna, and Milan do not cohere enough to be one system or even to have ever been variants of a single system. Too many people in too many places were working at these questions, and their tables and proposals differed from each other. Declercq points out that Easter Sunday could not fall before 15 Nisan in Rome, and he also says that "the church of Rome refused, until the middle of the fifth century, to go beyond 21 April". There is evidence however of both earlier and later celebrations of Easter in Rome. Behind his statement about a late date for Easter Sunday are the letters of Leo, bishop of Rome, asking about the Alexandrian practice which would allow Easter Sunday to be celebrated as late as 25 April. Actually, they expected 24 April in A.D. 455, and that would have been rather inconvenient where pagan festivities for the founding of the City (21-25 April) would overlap and interfere with a peaceful Christian Holy Week during 17-24 April. But there was no refusal. Rather, Bishop Leo accepted instruction from Alexandria in an ecumenical spirit, perhaps recognising also that outside the city of Rome, neither 25 nor other Easter Sundays on 21 to24 April would seriously bother any of the churches. Nevertheless, mathematicians were found to study these matters in the West.
The fourth chapter of this book focuses upon the Easter Table of Dionysius Exiguus which places the Incarnation, specifically the Birth of Jesus, at the head of a longer series of years to be counted of course from Year One. It was this Dionysian data table which guided the Bedan time-series. With adjustments by Bede, it gradually displaced Roman systems, both Julian and papal, made other royal or imperial systems passe, and gradually overrode most other local practices to become in our time the Common Era of most of the world. It was also this Dionysian data table to which later historically-minded and mathematically-careful Christian schoolmasters reacted, leading to the changes we call the Gregorian Calendar of 1583. Did Dionysius err, as the scholars argued?
Declercq carries this discussion forward along interesting paths that I shall not recount further. He emphasises that Dionysius' concern was not for historical accuracy but for quite different considerations. It is a well written tale and provides access to information which has not been easy to find but by anyone using historical documents. Libraries need this book.
Jews and Arabs may agree that a time-series so unavoidable may fairly be used in terms of the abbreviations CE and BCE, rather than AD and BC, thus avoiding any religious implications of Christian terminology. Historians of any faith nevertheless are stuck with one fact of the matter: however weird it seems and however uncertain the earliest evidence, the Common Era begins with the Incarnation. So be it.