Berlin Papyrus P. 13447 was found along with a number of other papyri during the 1906–1908 German excavations at Elephantine. The papyrus was first published by Eduard Sachau in 1911; its first English language edition was A. E. Cowley’s 1923 Aramaic Papyri of the Fifth Century B.C.¹ In the 1970s and early 1980s, more pieces were assigned to the papyrus, culminating in the Corpus Inscriptionum Iranicarum edition of 1982.² The current standard edition is the Textbook of Aramaic Documents from Ancient Egypt (henceforth, TAD).³ There are two texts on the papyrus. The recto and first two columns of the verso contain an Aramaic version of Darius I’s Bisitun inscription (henceforth, DB Aram; the hypothetical urtext is indicated as DB Aram*).⁴ The remainder of the verso contains a set of memoranda (henceforth, Mem.).⁵

Although the texts on the papyrus have been well-studied and published, especially in Cowley’s edition and TAD, these publications obscure the composite nature of the papyrus scroll: the texts on the scroll have been presented in separate sections of the publications. Recently, I had the opportunity to re-examine the papyrus fragments, and now can clarify its previous publications. While TAD remains indispensable for the individual texts, this article aims to highlight the relationship between the texts on the scroll, which is not well presented in either TAD or Cowley’s work.⁶

The papyrus scroll is very fragmentary, but this is not a significant problem for understanding the text of DB Aram, as much of the text can be reconstructed from the other extant versions (primarily the Babylonian one, to which the Aramaic text seems closest).⁷ While Bezalel Porten and Jonas Greenfield argue that the DB Aram* was a translation of the Babylonian version, Philip Huyse argues for the opposite.⁸ Neither argument, however, adequately considers an insertion at the end of DB Aram*, which has been shown to be a translation of the final section of Darius I’s tomb inscription at Naqshi-Rustam (DNb 50–60), and is translated from the Old Persian rather than the Babylonian version of DNb.⁹ Nor do these arguments consider the significant conflation in the final paragraphs. Jan Tavernier’s reconstruction of Paragraph 13 (DB Aram ll. 64–73) does show the relationships between the various texts and versions from a textcritical perspective.¹⁰

Figure 1a presents the recto and verso as mirror images so that their contents may be located with respect to the other side. Figure 1b shows the layout of the recto and verso in schematic form. As Figure 1a–b shows, and contrary to the presentation in TAD,¹¹ the Mem. texts follow immediately upon the end of DB Aram. DB Aram finishes at the bottom of a column, and Mem. begins at the top of the next column, with no more than the usual column space between. The two texts are in different hands.

Figure 2 presents the Old Persian (DB; from Bisitun) and Aramaic (from Elephantine) texts in columns in English translation. Four conclusions may be drawn. First, in DB Aram, DB Paragraphs 55–63 and 64–67 are conflated and DB Paragraphs 56–59 and 62–63 are omitted. DB Paragraph 55 is essentially repeated by DB Paragraph 64, while DB Paragraphs 60–61 are essentially repeated by DB Paragraphs 66–67.

Second, the section from DNb replaces DB Paragraph 65. DB Paragraph 65 enjoins Darius’s successor to protect and not destroy the inscriptions and relief. The paragraph deals specifically with the inscription at the site of Bisitun; it would make sense for it to be replaced in a version that circulated.

Third, DB Paragraphs 66–67 are modified to pronounce blessings and curses upon the successor who makes known or does not make known Darius’s deeds, rather than for protecting or destroying the Bisitun site. And fourth, the much-debated DB Paragraph 70 is omitted entirely (recording the command to translate and circulate the text), as is all of the final column of DB, which records the events of Darius’s second and third years.

These differences between DB Aram and all of the DB versions (Old Persian, Elamite, and Babylonian) at Bisitun demonstrates that Darius’s proclamation went through at least two editions. Either the insertion/replacement was made in the Babylonian version prior to its translation into Aramaic, or it was made during that translation. If, however, the Aramaic version was written prior to the Babylonian, then there were two Aramaic recensions, one attested in the Babylonian at Bisitun (henceforth DB Bab), and one in DB Aram.¹² All of the changes are consistent with a harmonizing and simplifying tendency in scribal practice.¹³ A similar tendency may be seen in 2 Chronicles 32, which takes the complex accounts of Sennacherib’s siege in 2 Kings 18–19 and Isaiah 36–37 and presents a single coherent account.¹⁴

The Mem. texts are much harder to decipher than DB Aram, as there is no other version. Contrary to TAD, there must have been at least thirteen columns of text, not six, as the columns of Mem. are about half the width of each of the DB Aram columns, and fragments are found on the verso of all the recto columns. This text, even in its fragmentary form, shows the greatest number of occurrences of the Aramaic word ‏זכרן‎ (zokrān, “memorandum”) in the entire Egyptian Aramaic corpus. Since there are many other account-type documents in the corpus, what could be the reason? I suggest that as the document was used over a number of years (see below), each use of ‏זכרן‎ allowed a scribe to more easily find previous entries. Figure 3 provides a reconstruction of the text, based on TAD, but with my own modifications to align the reconstruction with the schematic depiction in Figure 1b. Each entry began with zokrān and continued with a note of goods coming in (mostly precious items such as bowls, ladles, incense burners, etc.) or going out (mostly grain, but also “rods”).¹⁵ The majority of the names are Yahwistic and/or West Semitic.¹⁶ The items being deposited look suspiciously like a collection of vessels used in temple worship, as do some of the items received. Notable is the same number of bronze and silver cups deposited by both Hanan and Ah[…], twenty-one bronze and one silver cup (Col. 1). Jedaniah received “rods” (‏קנן‎) (Col. 6), which were among the deposits in other entries. Given the Yahwistic names and the presence of Azariah the servitor (‏לחן‎) (Col. 6),¹⁷ YHW-temple vessels would be the obvious conclusion. As for the grain, contemporary temple practices in Babylon suggest that the grain received by the temple had to be prepared and baked into loaves for grain offerings.¹⁸

Mem. might be the sign-in and sign-out records of the temple vessels and offerings. As with the Collection List (TAD C3.15), where there are women donors, in Mem. there are women depositing goods as well as receiving grain.¹⁹ The presence of these records on the same papyrus as DB Aram suggests that the latter was part of the YHW-temple library. The papyrus itself was found with the other papyri pertaining to the YHW temple, such as the letters dealing with the destruction of the temple; the collection is commonly called the “Jedaniah (communal) archive,” and was found alongside the so-called “Miptahiah archive.”²⁰ However, recently-published Demotic documents from the same period demonstrate that it was possible for institutional documents (such as DB Aram) and private documents such as deeds to co-exist in a temple collection.²¹ We may consider the collection as a temple archive rather than a private one.

The memoranda in Mem. are dated, although none of the date formulae are intact. Nevertheless, the first reconstructable date is the seventh year of Darius II, i.e., 417 bc, in Col. 5; this provides the usual terminus post quem for DB Aram. However, in Col. 1 is the date formula “the month of Epiph, year [x of Dari]us [the king.]” The letters that survive (‏יהו‎) can only have come from the Aramaic spelling of Darius, and no other Persian king. Thus the real terminus post quem of Mem. is 423 bc. Between Col. 1 and Col. 5 is a span of six years at a maximum. The final reconstructed date is 411 or later, in Col. 6; since the text continued for about seven more columns, the terminal date of Mem. cannot be determined. The names do give some hints, especially Ramnadaina, who was frataraka in 420 bc, and is mentioned in Col. 6 in connection with Jedaniah not delivering rods to him at some point after 411 bc. Since Ramnadaina was replaced by Vidranga as frataraka by 410 bc, and as a Persian would likely have returned to Persia, this date—year 13+ (411+ bc)—cannot be higher than year 14 (410 bc). Between the top of Column 5 and the bottom of Column 6, therefore, is a span of, at most, seven years.²² Combined with the span of up to six years for the first four columns, it is possible that Mem. continued to be added to until the end of the documented period of the collection, i.e., 399 bc.

The Elephantine copy of DB Aram is undated. The speculation by Greenfield and Porten, recently taken up by Gard Granerød,²³ that DB Aram* (or the Elephantine copy) was commissioned for the centenary of Darius I’s accession in 421 bc and/or the accession of Darius II in 423 bc, must be discounted; there is no evidence that such commemorative scrolls were produced in Egypt. Additionally, Granerød’s speculation that the satrap Arshama was instrumental in promulgating DB Aram* must also be discounted.²⁴ Greenfield and Porten, as well as Granerød, overlook the fact that DB Aram* must have existed in Egypt in order for the Elephantine scroll to have been produced. ²⁵ (Memphis, the satrapal seat, is a likely candidate for a scribal school.) It would have been part of the imperial scribal curriculum from the origins of the satrapy, not a recent import from Mesopotamia or Persia.²⁶ The erroneous assumption by previous commentators has been that the blank space on the scroll was used immediately after the completion of DB Aram, thus equating the terminus post quem of Mem. with the date of production of DB Aram.²⁷ There is no reason for the blank space to have been used right away; DB Aram could have been copied years or decades earlier. It is possible that the Elephantine copy was more than thirty years old when Mem. was started. In that light, it is more likely that DB Aram formed part of the charter of the Yehudite colony at Elephantine, either as part of Darius I’s efforts to secure the Egyptian border and subdue rebellions within Egypt after 518 bc—as Porten originally posited²⁸—or as part of Xerxes I’s consolidation in Egypt after 486 bc.²⁹

The scroll that was reused for the Ahiqar text found at Elephantine was originally used for shipping records in 476–75 bc.³⁰ Records had an average active life of ten to twelve years at Persepolis, the only well-understood Achaemenid administrative collection of the period,³¹ which suggests that the shipping records scroll was reused before 450 bc for the Ahiqar text. Speculatively, it is possible that the two scrolls—containing DB Aram and Ahiqar—are roughly contemporary, from 450 bc, and that they were brought to Elephantine at that time as part of the standard curriculum for a scribal school located in the YHW-temple. Egyptian practice was for scribal schools to be part of temples. The temple of Khnum, next door to the YHW temple, might have housed the Demotic school, while the YHW temple—associated with the imperial garrison—might have housed the Aramaic school. The school itself—if the find-spot of the papyri is any guide—would have been located in one of the houses next to the temple, as the temple complex itself was not large.³² What Mem. definitely shows is that DB Aram must have been an active text—still in use—up to near the end of the documented period of the colony; otherwise it would have been erased and written over.³³ Instead, DB Aram was left intact and the blank space on the papyrus used. The importance of this fact cannot be overstated: DB Aram was not an obsolete text and neither was Mem.

Jacqueline Du Toit’s work on ancient libraries has made clear that the modern distinction between a library and an archive when dealing with ancient corpora is a false one. Texts were maintained until no longer important, and then discarded; the materials could be re-used.³⁴ If the scroll of DB Aram was at Elephantine as part of the scribal curriculum, was Mem. part of that same curriculum? Were the entries recorded as part of scribal training in maintaining temple records?

Each time the scribe (or student) opened the papyrus scroll to enter another memorandum, he (or she) would have seen the end of DB Aram with the insertion from DNb 50–60, Darius I’s exhortation to his successor. Perhaps the Mem. texts were a way of fulfilling that exhortation, by providing accurate and transparent record-keeping. Further work on the Elephantine library as an operative temple library with its attendant scribal practices and education might shed more light on these matters. The possibility that the “Jedaniah archive” was both a temple collection and an educational collection could also be further explored. Fortunately, the Staatliche Museen zu Berlin has recently received a European Research Council grant that will permit the digitization of all the extant papyri from collections around the world, and the creation of a virtual collection of the entire Elephantine corpus (Verena Lepper, pers. comm.). This will make this sort of work much easier to carry out in the future. In the meantime, this short contribution has begun the work of examining this one scroll both as an individual document and as a part of a larger scribal collection.