Most of the extant royal letters of Assurbanipal are drafts and archival copies because they were not unearthed at their destination but found in Nineveh, the royal capital during his reign. Only two original royal letters have been found at their destination so far.[[93]] Some royal letters might perhaps be the originals that were not sent out or returned to the capital for some reason. They were first drafted, and then the originals were sent out and copies were archived in the capital. Most of the letters sent to the king were probably original letters, but some of them could be archival copies made in the capital.[[94]] In the whole process of composing the letters, two languages and two scripts were used in the royal correspondence of Assyria: the Neo-Assyrian language and script and the Neo-Babylonian language and script.[[95]] In general, letters in Neo-Assyrian are written in Neo-Assyrian script, and letters in Neo-Babylonian in Neo-Babylonian script.
However, there are exceptions. In the present volume, the following Neo-Babylonian letters are written in the Neo-Assyrian script: 17 royal letters,[[96]] three letters to the king,[[97]] and one non-royal letter.[[98]] Although the native language of the addressees, especially in the south,[[99]] did affect the language choice, the particular language found in the royal letters depended on whether they were drafts or archival copies.
The drafts were made through note-taking or dictation. They would have first been written in Neo-Assyrian because it was the language of the Assyrian ruling elite, and Neo-Babylonian was limited to the south. The former method is found in a Neo-Assyrian letter, SAA 1 1. The king probably gave instructions on the main points to be written to his scribe. The latter method is used in no. 25. It was probably written from dictation in Neo-Assyrian as a first draft because it contains a large number of varied abbreviations; the scribe clearly attempted to reduce the number of signs and sign elements by using unusual logographic writings, by omitting determinatives and some elements of compound logograms, and by not marking the long vowels. The scribe also preferred to use the simpler signs that have a smaller stroke count. However, preliminary drafts are rarely found in the dossier. This suggests that the drafts of the early stages were probably already destroyed in antiquity in order to be recycled for their tablet clay.
After taking a note or dictation, a scribe began editing the letter. In the early editing phase, scribes possibly worked on the drafts written in Neo-Assyrian and using Neo-Assyrian script. Based on no. 25, virtual duplicates exhibited by three other Neo-Assyrian letters, nos. 22, 23, and 24, seem to have been written and edited by different scribes.[[100]] No. 22 is a shorter version of the other two. In the secondary phase of editing, the drafts would have been translated into Neo-Babylonian, but perhaps first written in the Neo-Assyrian script[[101]] and only later using Neo-Babylonian script when the letters were sent to the south, because letters in the Neo-Assyrian language and Neo-Babylonian script are rarely found.[[102]] The fact that Babylonianisms are rarely found in letters written in Neo-Assyrian, while letters written in Neo-Babylonian contain numerous Assyrianisms, supports this interpretation.
When the scribe edited the draft, he made minor changes to the texts. Even the author indicated changes to be made in the text. In many cases, these changes probably caused the erasure of some signs and lines on the tablets. Among the royal letters of Assurbanipal, 15 missives have erasures.[[103]]
Once the scribe and the king were satisfied with a final draft, an expert scribe produced a definitive letter. After the definitive letter had been created, its copy was made and filed away. Or perhaps the final draft was archived instead of making a new copy. In this case, it is almost impossible to distinguish archival copies and very final drafts from originals in the extant corpus of royal letters, because they seem to be almost identical. Then the definitive letter was enclosed in a clay envelope. Only after the letter had gone through the entire process, it was finally ready to be transported to its destination by a messenger.
Two royal letters of Assurbanipal explicitly indicate that they are archival copies: nos. 3 and 33. In no. 3, following the main body and two uninscribed lines, the last three lines read: "Month of Iyyar (II), the 23rd day, eponym year of Aššur-dūru-uṣur (652). Šamaš-balassu-iqbi delivered (the letter)." Since it is likely that these lines are written as an archival note, this letter must be the archival copy of an original sent out to Babylon.[[104]] The text is written in Neo-Babylonian but uses the Neo-Assyrian script and contains numerous Assyrianisms. No. 33 is certainly a copy because the text itself says so. The body of the letter ends at rev. 5 and there is a blank space of one line. After that, the letter has an archival note: "copy (gabrû) of the letters that were brought to the (Chaldean) chieftains and to the land of Akkad. Month of Iyyar (II), the 24th day, eponym year of Nabû-šar-ahhešu (646*)."
These two archival copies are dated, and the dates are visually separated from the text itself by a blank space. As stated above, letters are usually undated because a messenger would have informed the recipient when and where the letter was written. However, archival copies cannot transmit such information. Thus, the date of the letter was sometimes added to an archival copy or archived letter. Of Assurbanipal's royal letters, 15 missives are dated and all of them have some visual indications before the date such as a blank space of one or more lines,[[105]] a blank space of erased lines,[[106]] a horizontal dividing line,[[107]] or an indented date line.[[108]] Among them, nos. 3, 15, and 43 are written in Neo-Babylonian using Neo-Assyrian script. Furthermore, though nos. 27 and 75 are Neo-Babylonian letters in the Neo-Babylonian script, their dates are inscribed in the Neo-Assyrian script. This means that the scribe switched the script from Neo-Babylonian to Neo-Assyrian for the entire text or a portion thereof when he made the copy. These visual indications and switching of scripts were probably useful to easily distinguish the text of the letter from archival notes in the royal archives.
93 SAA 10 295 = PBS 7 132, a royal letter to the exorcist Urda-Gula, found in the University Museum excavations at Sippar, and SAA 13 1 = KAV 114, a letter of Assurbanipal addressed to Aššur-mudammiq, possibly a priest of the Aššur Temple, Aššur-šarru-uṣur, a temple steward of the Aššur Temple, and Aššur-hussanni, a goldsmith of the Aššur Temple. The letter was unearthed in the archive N1 containing the library of the Aššur Temple in Assur. See Pedersén 1986, 28 (149).
94 For instance, nos. 106, 109, and 117. One of Nabû-ušabši's letters, ABL 269, is also written in Neo-Babylonian but with Neo-Assyrian script. I assume that these Neo-Babylonian letters in the Neo-Assyrian script were originally written in Neo-Babylonian using Neo-Babylonian script, and then copied in the Neo-Assyrian script, the original Neo-Babylonian letters being now lost. See also Worthington 2006, 63. There are two examples of which both the original and the copy are extant. A Neo-Babylonian letter exists in two different copies: ABL 960 (no. 106) in the Neo-Assyrian script and CT 54 189 (no. 107) in the Neo-Babylonian script; ABL 751 + CT 54 429 (letter from Nabû-ušabši to Assurbanipal) in the Neo-Babylonian script and ABL 268 (archival copy) in the Neo-Assyrian script.
95 Opinion is divided on whether Neo-Assyrian is a language or a dialect. See Radner 2014, 66; Parpola 2007, xi-xii; Worthington 2006, 59; Luukko 2004, 2-3.
96 Nos. 1 (SB in NA script), 2-5, 8-10, 16, 36, 43, 52, 54, 69, 85, 87-88. As pointed out by Reynolds (2003, XVI), most of the letters written in Neo-Babylonian and in Assyrian script are copies of royal letters sent to the south.
97 Nos. 106, 109, 117.
98 No. 153.
99 Assyrian kings accepted Neo-Babylonian letters from the south and sent Neo-Babylonian letters to recipients in the south.
100 Parpola 2004a, 228-229. Cf. Frame 1986, 267-269.
101 For instance, no. 8 could have been written as a draft and no. 9 was drawn up as a revised and shorter version of no. 8.
102 Except for SAA 10 183, a Neo-Assyrian letter but written in Neo-Babylonian script. It includes two messages but both of them concern the same issue, the extispicy on the roof of the temple of Marduk. It seems to me that a Babylonian diviner in royal service at Nineveh wrote the first message as an initial draft and the second message as a revised version based on the first message, see also Parpola 1983b, 357 on LAS 340 = SAA 10 183. Cf. Worthington 2006, 63.
103 Nos. 3, 6, 23, 24, 26, 27, 42, 48, 51, 55, 57, 58, 62, 66, and 83. However, nos. 3 and 27 appear to be copies because they are dated. See the discussion below.
104 A similar statement is also found in SAA 19 1. As Luukko pointed out, since that statement does not seem to be part of the original letter, it was added as an archival note when SAA 19 1 was made as a copy.
105 Nos. 3, 15, 32, 33, 43, 60, 61, 80, and 82.
106 No. 27.
107 Nos. 63 and 65.
108 Nos. 43, 57, and 75.
Sanae Ito
Sanae Ito, 'The Process of Composing the Royal Letters; Languages and Scripts', The Correspondence of Assurbanipal, Part I: Letters from Assyria, Babylonia, and Vassal States, SAA 21. Original publication:Winona Laka, IN, Eisenbrauns, 2018; online contents: SAAo/SAA21 Project, a sub-project of MOCCI, 2020 [http://oracc.org/saao/saa21/composingtheroyalletters/]