Editing the Lost Inscriptions from Khorsabad

As a result of the unavailability of so many of the original inscriptions on stone wall slabs, paving slabs, and bull colossi from Khorsabad (in particular text nos. 1–4, 6–14, and 23–41) because of their loss in the Tigris River in 1855 or their reburial at the site of Khorsabad, the unreliability of Botta's copies of these inscriptions, and the uncertainty about the reliability of Winckler's newer copies of the texts, editing these texts presents many problems and challenges, in particular for text nos. 1–4, 6, and 8 (Annals from Rooms II, V, XIII, XIV, and an unknown room of the palace, and the Display Inscription of Room XIV), for which only one exemplar each is known, but also for text no. 7, for which five exemplars are known. As is well known, Botta's copies contain numerous anomalous/unexpected sign forms, which is not surprising since he made no claim to be able to read cuneiform — indeed philologists were still in the process of deciphering Akkadian cuneiform texts at the time — since he was working under non-optimal conditions, and since the originals were clearly found at times in a damaged condition.^[52] In addition, as was mentioned earlier, even though Winckler's copies generally make excellent sense, they are not reliable, even when he claims to have collated the squeezes.

The poor state of preservation of the squeezes that Botta made and that are still extant in the Louvre and the difficulty in working with them means they are not as useful as one might hope. The original squeezes are frequently in a fragile condition and, having been rolled up for decades, do not lie totally flat on a surface; they are unwieldy to work with and require good lighting in order to facilitate the identification of signs and traces versus damage. Fortunately, the squeezes underwent conservation work in the 1990s; all but a few of the squeezes were photographed at that time and the author was able to obtain copies of these photographs with the permission of B. André-Salvini. It is generally both easier and more productive to work from the Louvre photos than the original squeezes since the photos were taken by an excellent photographer under ideal lighting conditions and then printed in reverse; the squeezes themselves of course show a mirror image of the inscription. With the permission of A. Thomas, the author made photographs of those squeezes not photographed in the 1990s.

Regrettably, while examination of the squeezes (either directly or via photos) does allow one to improve or correct Botta's copies at times, the squeezes are often unhelpful, either being damaged or unclear themselves at a critical point. (See Figure 2 below for a photo of section of a squeeze in a moderate stage of preservation.) They often suggest that the original text was damaged at the point where Botta copied an anomalous or unexpected form. They do, however, often reveal traces of signs not copied by Botta, traces that were often, but not always, noted by Winckler. It is likely (and at times proven by squeezes) that signs or parts of signs not indicated on Botta's copies were hidden/not present due to damage to the text even where Botta's copies give no indication that the text was damaged. Thus, numerous signs indicated as being abnormal or incorrect in form or omitted were likely only partially preserved or totally missing due to damage on the original.

Figure 2. Squeeze of Botta, Monument de Ninive 4 pl. 80, middle (text no. 1 section 15). © Musée du Louvre, dist. RMN - Grand Palais / Christian Larrieu.

At times, lengthy discussions would be needed to justify the reading of numerous signs in a single line of a single slab; however, this would make this volume inordinately lengthy and difficult to use. After much consideration and discussion with colleagues, it has been decided that for text nos. 1–4 and 8 a score edition would be prepared and made available online with a simplified notation scheme.^[53] No score is given for text no. 6 since there are no other sources of information on that text apart from Botta's copy in Monument de Ninive (i.e., no original, no earlier copy in Journal asiatique or the Institut de France, and no squeeze). In the scores, one line (indicated by the siglum Bt) indicates the reading based on Botta's copy in Monument de Ninive, with the siglum † used here and elsewhere to indicate an abnormal sign form;^[54] one line (Wi) is based on Winckler's copy, but this line is only present when Winckler claims to have collated the text from a squeeze;^[55] one line (Sq) is based on collation of a squeeze still preserved in the Louvre (mostly examined by means of photographs, but at times from the original squeeze); one line (JA) gives information from Botta's copies published in Journal asiatique when a copy differs from his later one in Monument de Ninive; one line (IdF) gives information from Botta's original copy preserved in the Institut de France when it differs from his copy in Monument de Ninive;^[56] one line (Je) is based on copies by Ch.-F. Jean published in Revue d'assyriologie 24 (1927) pp. 79–80 (only for two slabs in Room II [text no. 1 sections 12–13]); and one line (T-D) gives Thureau-Dangin's collations of individual words and signs on the squeezes published in RA 24 (1927) pp. 76–78 (only for slabs in Room II [text no. 1]). In most cases only one or a few of these lines will appear since the others do not exist for a line (or slab in general). On a few occasions, part of a particular wall slab still exists (see for example text no. 2 section 6 and text no. 3 section 1´) or a squeeze exists outside the Institut de France (see text no. 1 section 5) and lines in the score will be given for them, with the line siglum indicated in the commentary to that text. The dagger siglum (†) is used in the scores to indicate an abnormal sign form in Botta's copy: e.g., i-ta^†; this may be omitted if the sign is only slightly abnormal. However, if the unexpected form is an actual sign, that form is indicated in the score: e.g., a-di*(copy: RU).

The master line is based upon the author's best assessment of the sum total of information, although the siglum † is generally placed after the reading of any sign that is less than certain and/or has conflicting information about it that seemed worthy of note.^[57] The reader then needs to look at the score to see the conflicting information on/reading of the sign. This reduces the number of on-page notes arguing about the reading of any one sign. However, where the author felt information about the reading of a sign (or group of signs) was particularly useful, an on-page note on the matter is given. The author is fully aware this is not a perfectly satisfactory manner to deal with uncertain readings, but has decided that it is the most suitable manner to present a manuscript both easily usable by non-specialists and yet able to indicate to the specialist where uncertainty remains about a reading without over-burdening him/her with numerous lengthy discussions that would in many/most cases result in no definite answer.

With regard to the master lines and scores for these texts, it may be useful to explain a number of general procedures:

a) Normally, † will be placed after a sign in the master line if Botta's Monument de Ninive copy has an abnormal or incorrect sign even if a squeeze would allow the correct reading, but cannot confirm it. For example, the score may have a-lid*(copy: TAB) for Bt and a-⸢lid⸣ for Sq, but have a-lid^† in the master line; this would mean that the squeeze indicates the second sign is so damaged (or that the squeeze is so damaged at that point) that although it might well be LID one cannot rule out the possibility of it being TAB. Since Winckler's copies cannot be relied upon (see above), † may still appear in the master line even if a-lid is given in the line for Wi.

b) If Botta's copy has a sign fully preserved, while the squeeze shows it damaged, Botta's copy is followed in the master line.

c) When Botta's copy is only slightly abnormal (e.g., MA for LA, KAL for UN, or IZ for PA), while Winckler's copy has the correct form and the squeeze while indicating that the sign is damaged, supports the correct reading, then the correct reading is given in the master line with an indication that the sign is damaged (e.g., ⸢UN⸣).

d) If Botta's copy in the Journal asiatique and/or Institut de France has the correct/desired sign form of a sign while the later copy in Monument de Ninive has an incorrect/abnormal form, preference is given to the former in the master line and no indication is normally given in the master line that the copy in Monument de Ninive is abnormal, although this will of course be noted in the score.

e) If the master line has † with a sign being fully restored, this means that Winckler's copy shows the sign as fully present, while Botta's copy and the current state of the squeeze do not have any indication of it being there. For example, [a^†]-na in the master line, [x]-na in Bt, a-na in Wi, and [x]-na or [x x] in Sq. However, if Winckler only indicates that part of the sign is preserved, then his reading is sometimes accepted. For example, if Bt has [x x], Wi has [ki]-⸢ma⸣, and Sq has [x x], then [ki]-⸢ma⸣ may appear in the master line.

f) As is the case elsewhere in this series, if the unexpected sign is an actual sign, then that sign is given in parentheses following the expected reading of the sign in the score. However, † will normally be given in the master line. For example, if we expect ia-a-ti while Botta's copy has AD-a-ti and the squeeze is not preserved at that point, then the score will have ia*(copy: AD)-a-ti for Bt and the master line ia^†-a-ti.

g) Frequently, even if there is a squeeze of an inscription, it may not cover the whole area copied by Botta, thus not all of the copy has really been collated. In addition, the edges of the squeezes are generally the most damaged parts of them. Thus, the reader will sometimes find more signs with the siglum † in these areas.^[58]

h) Squeezes and/or originals often show that a damaged area is greater or smaller in size than is indicated on Botta's copies. Thus, where we have only Botta's copy as evidence for a text or exemplar, the amount of restoration given at times in the master line may differ from what Botta's copy would suggest, although for the most part this only amounts to one or two additional or fewer signs.

In sum, unless we can confirm from a squeeze the accuracy of Winckler's copy of a sign against Botta's copy, Winckler's copy is not normally accepted in the master line. In addition, most of the signs in the master line indicated with † or as being totally omitted, were probably damaged on the original (e.g., i^†-(na) should be ⸢i⸣-[na]), but of course this cannot be proven unless the originals are recovered. It must be admitted that some tiny traces (or apparent traces) on the squeezes have not been noted in the scores. This occurs when the author was uncertain if they were or were not actual traces as opposed to damage on the original or to the squeeze.

Since there are five exemplars of text no. 7 (Khorsabad Display Inscription), a score edition of that inscription is presented online; however, only one line is given for each exemplar in that score even though information on each exemplar can come from a variety of sources in addition to Botta's copies in Monument de Ninive: Winckler's copies collated from squeezes, squeezes still in existence in the Louvre, and fragments of the original slabs preserved in various museums (in particular the British Museum, Louvre, Oriental Institute [Chicago], and Museo delle Antichità Egizie [Turin]). The transliteration of each exemplar in the score is thus a conflation of information from all of the sources available for it and has been made to the best of the ability of the author, although he would admit that matters of uncertainty remain over numerous individual signs and words on any one exemplar. Even though the exact reading of the individual exemplars is a matter of uncertainty at numerous places, and the siglum † is thus frequently found in the scores, because there are five exemplars of the inscription, the overall reading of the text (i.e., the master line) is basically secure.

For the inscriptions or exemplars that use more than one wall slab (text nos. 1–4 and 7–8),^[59] the catalogues list each slab separately, calling it a "section" and numbering the sections in consecutive order. For example, text no. 8 is split up into six sections numbered 1–6 to represent each of the six slabs needed for the inscription.

Notes

52 As has been noted by earlier scholars, on occasion Botta's copies misalign fragments of a wall slab (e.g., Room II slab 35 [Monument de Ninive 4 pl. 91]); a double dagger (‡) is employed in the scores to indicate the point in the line of Botta's copy where one must either go up or down a line in his copy in order to continue reading the relevant line.

53 Note the author's article "Lost in the Tigris: The Trials and Tribulations in Editing the Royal Inscriptions of Sargon II of Assyria" in Yamada, SAAS 28 pp. 215–237; some examples of problems with Botta's copies are given ibid. p. 221. For text no. 7, see below.

54 The siglum † is given after the expected/desired reading of the sign. Such abnormal signs are not normally indicated in the lists of variants at the back of this volume, particularly when there is more than one exemplar for an inscription (text nos. 7 and 9–14 and 41).

55 At times Winckler has a line on his copies indicating either where the squeeze he had examined for collation purposes began or where it ended (e.g., his copies of Room II slabs 3 and 9 [Winckler, Sar. 2 pl. 1 no. 2 and pl. 3 no. 6]). In such cases, a superscript right-facing triangular bullet (‣) is employed in the scores to note where Winckler indicates that he began to collate the inscription from a squeeze and a superscript circular bullet (•) is similarly employed to note where he indicates that his collations ended. This latter siglum is also used in the transliteration of text no. 117 (Najafabad Stele) to indicate where the lines of the inscription on the squeezes preserved in the Royal Ontario Museum end.

56 For the most part, the copies in the Institut de France are identical to those published in Journal asiatique, but there are occasional differences between the two. My thanks must be expressed to M. Zink (now secrétaire perpetuel de l'Académie des inscriptions et belles-lettres) through whose support I was able to access the materials in the Institut de France in the summer of 2011 and also to J.-M. Durand for facilitating my visit there.

57 For texts with only one exemplar (e.g., text nos. 1–4, 6, and 8) this will generally mean that P.E. Botta's copy has an abnormal or incorrect sign form and that there is no contradictory information from an original or from a squeeze.

58 These were probably also the most damaged areas on the originals.

59 Text nos. 5 and 6 would have required more than one wall slab, but information only on all or part of one slab is known for either text. As noted with text no. 6, it is possible that that text comes from the same room as text no. 5, the palace throne room (Court VII).

Grant Frame

Grant Frame, 'Editing the Lost Inscriptions from Khorsabad', RINAP 2: Sargon II, Sargon II, The RINAP 2 sub-project of the RINAP Project, 2021 [http://oracc.org/rinap/rinap2/rinap2introduction/lostinscriptions/]