V — Chronologica
On the Meaning of the Word Σαρος
Read Before the Philological Society. April 11, 1845
The words σάρος and sarus are the Greek and Latin forms of a certain term used in the oldest Babylonian chronology, the meaning of which is hitherto undetermined. In the opinion of the present writer, the sarus is a period of 4 years and 340 days.
In the way of direct external evidence as to the value of the epoch in question, we have, with the exception of an unsatisfactory passage in Suidas, at the hands of the ancient historians and according to the current interpretations, only the two following statements:—
1. That each sarus consisted of 3600 years (ἔτη).
2. That the first ten kings of Babylon reigned 120 sari, equal to 432,000 years; or on an average 43,200 years apiece.
With data of this sort, we must either abandon the chronology altogether, or else change the power of the word year. The first of these alternatives was adopted by Cicero and Pliny, and doubtless other of the ancients—contemnamus etiam Babylonios et eos qui e Caucaso cœli signa observantes numeris et motubus stellarum cursus persequuntur; condemnemus inquam hos aut stultitiæ aut vanitatis aut impudentiæ qui CCCCLXX millia annorum, ut ipsi dicunt, monumentis comprehensa continent.—Cic. de Divinat., from Cory’s Ancient Fragments. Again—e diverso Epigenes apud Babylonios DCCXX annorum observationes siderum coctilibus laterculis inscriptas docet, gravis auctor in primis: qui minimum Berosus et Critodemus CCCCLXXX annorum.—Pliny, vii. 56. On the other hand, to alter the value of the word ἔτος or annus has been the resource of at least one modern philologist.
Now if we treat the question by what may be called the tentative method, the first step in our inquiry will be to find some division of time which shall, at once, be natural in itself, and also short enough to make 10 sari possible parts of an average human life. For this, even a day will be too long. Twelve hours, however, or half a νυχθήμερον, will give us possible results.
Taking this view therefore, and leaving out of the account the 29th of February, the words ἔτος and annus mean, not a year, but the 730th part of one; 3600 of which make a sarus. In other words, a sarus=1800 day-times and 1800 night-times, or 3600 half νυχθήμερα, or 4 years+340 days.
The texts to which the present hypothesis applies are certain passages in Eusebius and Syncellus. These are founded upon the writings of Alexander Polyhistor, Apollodorus, Berosus, and Abydenus. From hence we learn the length of the ten reigns alluded to above, viz. 120 sari or 591 years and odd days. Reigns of this period are just possible. It is suggested, however, that the reign and life are dealt with as synonymous; or at any rate, that some period beyond that during which each king sat singly on his throne has been recorded.
The method in question led the late Professor Rask to a different power for the word sarus. In his Ældste Hebraiske Tidregnung he writes as follows: “The meaning of the so-called sari has been impossible for me to discover. The ancients explain it differently. Dr. Ludw. Ideler, in his Handbuch der mathematischen und technischen Chronologie, i. 207, considers it to mean some lunar period; without however defining it, and without sufficient closeness to enable us to reduce the 120 sari, attributed to the ten ancient kings, to any probable number of real years. I should almost believe that the sarus was a year of 23 months, so that the 120 sari meant 240 natural years.” p. 32. Now Rask’s hypothesis has the advantage of leaving the meaning of the word reign as we find it. On the other hand, it blinks the question of ἔτη or anni as the parts of a sarus. Each doctrine, however, is equally hypothetical; the value of the sarus, in the present state of our inquiry, resting solely upon the circumstance of its giving a plausible result from plausible assumptions. The data through which the present writer asserts for his explanation the proper amount of probability are contained in two passages hitherto unapplied.
1. From Eusebius—is (Berosus) sarum ex annis 3600 conflat. Addit etiam nescio quem nerum ac sosum: nerum ait 600 annis constare, sosum annis 60. Sic ille de veterum more annos computat.—Translation of the Armenian Eusebius, p. 5, from Fragmenta Historicorum Græcorum, p. 439: Paris, 1841.
2. Berosus—σάρος δέ ἐστιν ἕξακόσια καὶ τρισχίλια ἔτη, νῖ ρος δὲ ἕξακόσια, σώσσος ἑξήκοντα.—From Cory’s Ancient Fragments.
Now the assumed value of the word translated year (viz. 12 hours), in its application to the passages just quoted, gives for the powers of the three terms three divisions of time as natural as could be expected under the circumstances.
1. Σώσσος.—The sosus=30 days and 30 nights, or 12 hours × 60, or a month of 30 days, μὴν τριακονθήμερος. Aristotle writes—ἡ μὴν Λακωνικὴ ἕκτον μέρος τοῦ ἐνιαυτοῦ, τοῦτο δέ ἐστιν ἡμέραι ἑξήκοντα.—From Scaliger, De Emendatione Temporum, p. 23. Other evidence occurs in the same page.
2. Νῆρος.—The nerus=10 sosi or months=the old Roman year of that duration.
3. Σάρος.—The sarus=6 neri or 60 months of 30 days each; that is, five proper years within 25 days. This would be a cycle or annus magnus.
All these divisions are probable. Against that of 12 hours no objection lies except its inconvenient shortness. The month of 30 days is pre-eminently natural. The year of 10 months was common in early times. In favour of the sarus of five years (or nearly so) there are two facts:—
1. It is the multiple of the sosus by 10, and of the nerus by 6.
2. It represents the period when the natural year of 12 months coincides for the first time with the artificial one of 10; since 60 months=6 years of 10 months and 5 of 12.
The historical application of these numbers is considered to lie beyond the pale of the present inquiry.
In Suidas we meet an application of the principle recognised by Rask, viz. the assumption of some period of which the sarus is a fraction. Such at least is the probable view of the following interpretation: ΣΆΡΟΙ—μέτρον καὶ ἁριθμὸς παρὰ Χαλδαίοις, οἳ γὰρ ρκ´ σάροι ποιοῦσιν ἐνιαυτοὺς βσκβ´, οἳ γίγνονται ιε´ ἔνιαυτοὶ καὶ μῆνες ἕξ.—From Cory’s Ancient Fragments3.
In Josephus we find the recognition of an annus magnus containing as many ἔτη as the nerus did: ἔπειτα καὶ δι’ ἀρετὴν καὶ τὴν εὐχρηστίαν ὧν ἐπενόουν ἀστρολόγιας καὶ γεωμὲτριας πλέον ζῇν τὸν Θεὸν αὐτοῖς παρασχεῖν ἅπερ οὐκ ἧν ἀσφαλῶς αὐτοῖς προειπεῖν ζήσασιν ἑξακοσίους ἐνιαυτούς· διὰ τοσοῦτον γὰρ ὁμέγας ἐνιαυτὸς πληροῦται.—Antiq. i. 3.
The following doctrine is a suggestion, viz. that in the word sosus we have the Hebrew שֵש = six. If this be true, it is probable that the sosus itself was only a secondary division, or some other period multiplied by six. Such would be a period of five days, or ten ἔׁτηׁ (so-called). With this view we get two probabilities, viz. a subdivision of the month, and the alternation of the numbers 6 and 10 throughout; i. e. from the ἔτος4 (or 12 hours) to the sarus (or five years).
After the reading of this paper, a long discussion followed on the question, how far the sarus could be considered as belonging to historical chronology. The Chairman (Professor Wilson) thought there could be no doubt that the same principles which regulated the mythological periods of the Hindoos prevailed also in the Babylonian computations, although there might be some variety in their application.
1. A mahayuga or great age of the Hindoos, comprising the four successive yugas or ages, consists of 4,320,000 years.
2. These years being divided by 360, the number of days in the Indian lunar year, give 12,000 periods.
3. By casting off two additional cyphers, these numbers are reduced respectively to 432,000 and 120, the numbers of the years of the saroi of the ten Babylonian kings, whilst in the numbers 12,360 and 3600 we have the coincidence of other elements of the computation.