In my article “A Scribal Family and its Orthographic Peculiarities” in 2011 I discussed texts from Middle Assyrian Assur, which belong to a corpus written by the sons of the royal scribe Ninurta-uballissu. Thanks to the extant colophons at our disposal we have a rather good grasp of the scribal activity of these scribes. However, due to the state of preservation and the dissemination of parts of the material already in antiquity, future palaeographical studies will most likely reveal more examples belonging to this corpus. (One example, which appears to belong to the corpus copied by the members of this family is a fragment kept in the archaeological collection of the University of Zurich, Switzerland, which, though lacking a preserved colophon, resembles many features of similar tablets in this group.) Although written by “young scribes” (ṭupšarrū ṣeḫrūtu), these texts do not offer many clues about scribal education of some sort, which we can presume most likely happened within the (private) realm of scribes such as Ninurta-uballissu. Putting orthographical peculiarities and other variations aside, there are hardly any significant errors or erasures in the pertinent texts. Unfortunately we lack any archival information on those scribes which would shed some light onto their lives.

The three brothers Marduk-balāssu-ereš, Bēl-aḫa-iddina and Sîn-šuma-iddina copied a great deal of lexical and literary material, all of which has come down to us in bilingual form. Most of these (quite well-preserved) texts represent our earliest sources for Akkadian translations of Sumerian literary compositions such as “Ninurta's Exploits” or “Ninurta's Return to Nippur,” but go back to earlier (bilingual) Middle Babylonian sources, for which evidence is rather scarce (see, for instance, CBS 11153 (+) N 6286). The majority of lexical sources known from Assur and dating to the last centuries of the 2^nd millennium BC already shows signs of “canonization.” A closer glance, however, reveals some differences, as does the text, whose colophon shall be discussed in greater detail below.

The small fragment VAT 9487 originates from the lower left corner of a large tablet with probably either two (see the copy of the second tablet of Diri by Bēl-aḫa-iddina; Ass. 2559; see also Wagensonner 2011: 672, 2.1.2) or three columns per side (see the hand-copy in Wagensonner 2011: 700 and compare to VAT 10172 for a better preserved manuscript of the first tablet of the series Ea with three columns on each side; hand-copy in Wagensonner 2014: 476-477). The copy contains the third tablet of the series Diri, a supplemental sign syllabary, which provides readings and translations of Sumerian compound logograms. By the Middle Assyrian period the sequence of entries and the division into “tablets” already follow the later conventions. It is a “0-1-2-4”-type list (“0” representing the vertical wedge or entry marker; “1” the syllabically written reading of a (compound) logogram; “2” the (compound) logogram; “4” the Akkadian equivalent). Later copies of the 1^st millennium BC add a column “3” containing an analytical writing of the (compound) logogram, traces of which are visible in the copy of the 1^st tablet of the sign syllabary Ea copied by Sîn-šuma-iddina, the third son of Ninurta-uballissu (see VAT 10172, col. ii, 66-67).

Providing readings to Sumerian compound logograms is Diri's most important feature. Single logograms occupying just one slot in column “2” are usually invalid in this list, but are organized within the framework of the sign syllabary Ea. There is also the noteworthy fact that single logograms are “validated” for Diri by adding a classifier or determinative. Thus, frequently both lists attest to the same reading, irrespective of whether a classifier is added in Diri (see Table 1). A similar example comes from entries preserved on our manuscript: Entries 2-5 (Obv. col. i, 3-8) deal with “fire wood.” (see for the composite text Civil 2005: 136). All these entries contain in column “2” the logogram group GIŠ.BAR_tenû.AŠ₂, which is assigned to the readings ki-bir, geš-ki-bir, gi-bil₂, and geš-gi-bil₂. On the 1^st tablet of Ea on the other hand the logogram BAR_tenû.AŠ₂ occurs as well, but appears to be regarded occupying a single slot with the constituent BAR_tenû or ŠU₂ serving as a kind of framing sign (see Wagensonner [forthcoming]).

According to the colophon, VAT 9487 was copied by a certain Nabû-šuma-iddina (written ^dag-mu-sum-na), son of Badû. The copy was checked by Bēl-aḫa-iddina (written ^d+en-šeš-sum-na), who is certainly identical with the afore-mentioned son of Ninurta-uballissu. It was argued that at the time this list was copied he was not a “young or apprentice scribe” any more, as he is known from copies written on his own (Jakob 2002: 258; Wagensonner 2011: 675). However, the colophon breaks at the right-hand side and therefore it is not clear whether the occupation ^lu₂a.ba is followed by tur or not. But note that there appears to be not much space on the tablet to accommodate this sign or a possible lugal in the preceding line.

Above the reference to the 3^rd tablet of Diri the scribe wrote the catch-line, which refers to the incipit of the subsequent tablet in the series. Although we know from later sources of the 4^th tablet that the first entry containing the compound logogram U₂.KUR.RA.SAR provides a reading šem-bi-ri-da, this reading differs a bit in the catch-line of VAT 9487: šem-bi-ri. Civil 2005: 151 interprets the ensuing damaged sign as possible DA, but the sign remains more likely point to KUR.^⌜RA^⌝, therefore column “2” most likely just contained the compound logogram KUR.RA.SAR. The reading šem-bi-ri-da is based on later sources for both, the 4^th tablet of Diri (see Civil 2005: 150-151) and tablet 4a of Nabnītu, line 260 (77), which reads (after ms. A = CT 12, pl. 36-37: U₂.KUR.RA^{šem-bi-ri-da}.SAR : ^{u2-ku-ur-ra-ni-si-gu-u} ni-nu-u (see Finkel 1982: 86).

Thus far, the beginning of the composition is not preserved on any of the known manuscripts dating to the later “canonical” stage (see Civil 2005: 3). Our text is one of the few preserved manuscripts that provides the series' title in its colophon. Line 3' on the reverse reads:

Miguel Civil in his edition of the text (ms. B) did not read the two signs, which precede the logogram group SI.A (see also the respective entry in the Digitale Keilschriftbibliothek). Although proposing a possible reading UD.DU earlier (Wagensonner 2011: 675, 2.1.6), a re-evaluation of the sign remains suggests a different interpretation now (note that the copy provided by the Digitale Keilschriftbibliothek is misleading, since more remains are visible on the tablet).

Another manuscript of Diri (Ass. 2559) was copied by Bēl-aḫa-iddina himself. Its colophon omits the reading of the compound logogram and only has SI.A at-ru (the tablet needs to be collated; see Wagensonner 2011: 672, 2.1.2).

There are not many possibilities for the two signs on VAT 9487. The second sign is clearly the sign NA, which is testified further down in the text as well. The sign that precedes it resembles EREN₂; it is too short for SUM, which appears as logogram in the sequence sum-na in the names Nabû-šuma-iddina and Bél-aḫa-iddina in the colophon (see above). One might entertain the possibility that the tablet's scribe did write his name directly after the catch line, but decided to include the tablet's reference without erasing his name properly in the still moist clay. There is another insufficient erasure at the beginning of the subsequent line, which (again) looks as if he wanted to start writing the scribe's impressum (qāt PN, “hand of PN”). Instead he inserted the remark al.til representing the Akkadian remark qati, “it (the source) is complete,” followed by the scribal information.

After inspection of the tablet at least the first sign does not appear to be partly erased. Another interpretation for the two signs preceding the compound logogram SI.A is a syllabically written reading, which usually appears in sign syllabaries of this and later periods. But the compound logogram SI.A has no reading ending in the syllable na, and there is no Sumerian equivalent of Akkadian atru with a similar ending either. On the tablet, SI.A is followed by an oddly squeezed sign IR, which should be and has been interpreted as phonetic complement to the reading dir(i).

After re-evaluation I therefore would like to propose that both eren₂-na as well as *ir are traces of the scribe's impressum, which he had not erased properly. Fig. 1,b contains a new hand copy of the respective lines (erasures are drawn red) next to a close-up photo trying to take this possibility into account. Line 3' on the tablet's reverse therefore should be rendered as follows:

The close-up shots of the tablet (Fig. 1,a) show that interpreting the aforementioned signs as traces of older text appears unlikely at first glance. This solution, however, enables us to avoid an unparalleled reading for the compound logogram SI.A and explains the extremely odd-looking insertion of a phonetic complement. And at close examination there are indeed further sign remains visible, which previously appeared more likely to be scratches.