The text CTH 532.3 / KUB 34.11 and duplicates is a collection of Hittite lunar-eclipse omens across months 1–13 and can be considered a forerunner to Enūma Anu Enlil tablet 22. The text survives on at least eighteen fragments, all of them in Hittite and mostly in New Script. It is thus the second-best-attested omen text from Ḫattuša by number of manuscripts, surpassed only by the solar-omen text CTH 534.1 with almost thirty witnesses. No Akkadian original has been found in Ḫattuša.

• A. Kloekhorst, ZA 100, 2010: 214 Anm. 47
• E. Laroche, CTH 1971, erwägt als indirekte Joins auch KBo 13.14 (= 114/s) und KUB 8.7 (= Bo 7262)
• G. Torri – F.G. Barsacchi, DBH 51, 2018: 18; 19f; 21f.
• J.N. Postgate, JCS 24, 1972: 175 (Nr. 75); in Gulbenkian Museum; Northumberland Collection; Durham University
• K.K. Riemschneider, DBH 12, 2004: 18f.; 32f.; 34; 74; 74f.; 75; 114-116; 166-169; 169f.; 170
• U. Koch Westenholz, GMS 3, 1993: 241-246

The tablets were copied by H. Otten (KBo 8; KBo 13), H. Otten and C. Rüster (KBo 34), E. Weidner (KUB 8), H. Ehelolf (KUB 34), O. Soysal (CHDS 5), and N. Postgate (JCS 24). The fragment Bo 6688 has yet to be copied.

A comprehensive edition was planned by Riemschneider K.K. 2004a, who first systematically collected the fragments that belong to this text; individual transliterations can be found in Riemschneider K.K. 2004a: 18-19.32-35.65-69.74-75.114-116.118-119.166-170.172.

A partial score of EAE 21 and 22, Emar 6.652, and KBo 13.15 (+) KUB 8.1 is offered by Koch-Westenholz U. 1993a: 241-246. Some exemplars have been transliterated and/or translated in part, e. g. in Riemschneider K.K. 1970a: 33-34 (O rev. III, 7′-10′//B rev. III?, 4′-6′); Otten H. − Rüster C. 1972b: 233 (L rev., 1′-9′); Haas V. 2008a: 138 (J2 obv. II, 5′-6′.16′-19′. rev. III, 8-11); Waal W. 2015a: 382 (colophon A); Torri G. − Barsacchi F.G. 2018b: 19-20.

In addition, many witnesses have been cited and partially translated in works on different issues, e. g. omens that may relate to substitution rituals in Kümmel H.M. 1967a: 25-26 (D 9′; F obv., 9′; J1+2 obv., II 19′, rev. III, 15, rev. IV?, 5′; M 12′-13′), omens about animals, especially vermin, in Collins B.J. 1989a: 252.271.283 (G obv., 16′-17′; I r. col., 8′; J rev. III, 8-11.17-19), or omens about rain in Zinko M. − Zinko C. 2007a: 743 (C rev. IV, 5′-6′; J2 rev., III 8-9).

Several studies focus on rare or unclear words and phrases, e. g. Sommer F. - Falkenstein A. 1938a: 175 pointed out the parallel between Akkadian ulabbar and Hittite para nai- (I obv. II 10′), Beckman G. 1982a: 436 n. 11: noted that ÉRIN pangaui equals Akkadian nakbatu/ina nakbatīšu (I rev. III, 6), Rieken E. 1999a: 431 and Melchert H.C. 2007a: 515 n. 9 have shown that the form šuppalan in the relevant passages is nominative Singular, Collins B.J. 1989a: 238 and Hoffner H.A. 2015a: 64 n. 6 argued that taknaš huidar likely means ‘insects’ or ‘vermin’ (J2 rev. III, 10).

Further collections of phrases and vocabulary can be found with Laroche E. 1948-1949a: 22; Goetze A. 1960b: 72; Kronasser H. 1962-1966a: 515; Riemschneider K.K. 1970a: 79; Monte G.F. del 1988b: 52-54; Schuster H.S. 2002a: 428-429; Taracha P. 2009a: 851-852, and Bachvarova M. 2013b: 141-142.

A: Curved, oblong piece from the left part of what was likely the last column of a two-column tablet, with fifteen lines of spacious script.

B: Trapezoid piece from the middle of the reverse of a two-column tablet; the surface is damaged in places. Remnants of 28 lines in tight, neat script are visible.

C: Two almost directly joining pieces of the top left of a single-column tablet, including the upper and left margin, with a long vertical crack. The surface is rather worn in some places. Contains 36 damaged lines of tightly written script.

D: Small trapezoid fragment from likely the right reverse of a single-column or right of a fourth column of a two-column tablet. Shows remnants of eleven lines of cuneiform in tight script.

E: Pentagonal fragment of a right column, likely from the obverse of a two-column tablet close to the upper edge. Contains skewed paragraph lines and remnants of ten lines of tight, slightly cursive script.

F: Oblong piece of the left of what was likely a single-column tablet. The surface is, in places, abraded and contains remnants of 25 lines of spacious script.

G: Small, roughly pentagonal fragment from perhaps the lower edge of a tablet, with remnants of five lines of spacious, slightly cursive script.

H: Small oval fragment from close to the right edge of the tablet and/or column with remnants of seven lines of tight script.

I: Roughly triangular fragment from the middle of a two-column tablet with remnants of twelve lines of script and an abraded surface.

J1: A roughly pentagonal piece from likely the fourth column close to the left edge. It contains fifteen lines of relatively dense cuneiform.

J2: The lower right quarter of a two-column tablet with a relatively narrow column divider (6-7 mm) and an edge band at the lower margin. It contains 43 lines of spacious script.

K: Small, roughly triangular fragment from close to the left edge of a tablet or column. Contains the beginnings of eight lines of spacious script.

L: Two joining pieces from the lower middle of a tablet with mainly the surface of the reverse remaining. Contains remnants of 22 lines of script.

M: Tall piece from the center of a likely single-column tablet with nineteen lines of spacious, cursive script.

N: Trapezoid fragment with the beginning and end of a two-column or very narrow single-column tablet with the remnants of eleven lines of spacious script.

O: Small oblong fragment from close to the left edge of a tablet or column. Contains the remnants of eight lines of spacious script.

P: Small pentagonal fragment of red-brown clay from close to the left margin of the tablet or column. Contains remnants of eight lines of spacious script.

There is some paleographic evidence for a date earlier than the empire period. N has some Middle Script sign variants and P perhaps even two Old Hittite variants (see the note on the text in Soysal O. – Aykut İ. 2022x), but the fragment is small, and thus it is difficult to say if those are merely graphic archaisms.

The paleographic evidence is supported by a number of signs of pre-New Hittite language: In the (likely) Middle Script manuscript M (KUB 8.4), we find a]n-tu-uḫ-ḫa-aš tu-⸢e⸣-[ek-ku-uš-šu-uš where the (late) New Script manuscript C has UN-aš NÍ.TE‑u[š]‑šu‑⸢uš⸣. The use of the Hittite enclitic possessive falls out of use in New Hittite. Vice versa, later texts have a tendency to replace syllabic spellings with logograms. Similarly, ŠA? ut‑n]e?‑aš a‑aš‑šu‑u in M is replaced by ⸢ŠA?⸣ KUR‑e‑aš a‑〈aš〉‑šu‑u in the New Script manuscript F.

Typical for Middle Hittite is the frequent use of ti-i-e-ez-zi instead of earlier ti-e/iz-zi and later ti-ia-az-zi (Neu E. 1996a: 4; Neu E. 1970a: 52) throughout the text and also in New Script manuscripts like J2. Also in alignment with earlier orthography is ma-a-i-an-[zi] in A (New Script), with the plene spelling and without -ia- to mark the glide, but there are not enough comparable forms to decide whether this is a distinctly Middle or Old Hittite spelling. Once, the local particle =apa is used (I r. col., 2′; J2 rev. III, 10; K 1′), which is frequent in Old Hittite, but in New Hittite only appears in copies of earlier text and is rare in Middle Hittite (HG §28-57-58).

It may be too much to insinuate the text is from the ‘epoca antico-ittita’ (Dardano P. 2010b: 53 n. 14, similarly Brosch C. 2014e: 244-245, ‘aH’), unless one makes no distinction between the pre-New Hittite stages of the language. The evidence is enough to assume that the translation of this text dates from the Middle Hittite period.

The phrase to describe the eclipse in all witnesses is D30-aš a-ki: ‘Armaš/The moon dies’, as in CTH 532.2, CTH 532.4, and CTH 532.6. The colophon of ex. A has ma-a-an […] pu-uš-zi where we expect the name of the composition, which implies that ak- and puš- are synonyms when relating to the moon.

A relatively high number of errors is found in the numbering of months and days, mostly negligence errors like leaving out tens in 14, 15, and 16, writing 12 for 21, etc., especially in exemplar J1. A more peculiar case is the ungrammatical writing LUGAL-aš KUR-SÚ-na-aš in exemplar J1, where B has lam-na-aš LUGAL-aš. The latter is a translation of Akkadian šarru ša šuma išû or šar kiššati, as attested by the Mesopotamian tradition. This is best explained by a misreading of a damaged and perhaps rather long sign LAM, that might have looked like K[UR-S]Ú. But this may also hint towards an unfamiliarity with the text genre, at least for this scribe.

No witness with the entire composition survives from Ḫattuša. In sum, every month is attested, although there are substantial gaps in months four, five, and ten. So far, eighteen fragments have been identified as belonging to the text. The only one not published here is Bo 6688, which is to be published by N. Aslantürk.

The exact number of tablets is difficult to determine. I don’t think the join between KBo 13.15 (J1) and KUB 8.1 (J2) proposed by Riemschneider K.K. 2004a: 4.65 is particularly likely. However, I think it is possible that the three fragments KBo 13.14-16 (H, J1, D), all from the Haus am Hang, are part of the same tablet, as they look strikingly similar in writing and layout. The same is true for the likely Middle Script fragments KBo 34.113-114 and KUB 8.4 (K, L, M), and the two pieces CHDS 5.146 and KUB 8.7 (N, O). As indirect joins, especially made based on photographs, are a dangerous undertaking, the fragments will be considered individual witnesses unless another piece can prove they belong together.

To further complicate matters, the text existed likely in a one- and a two-tablet version. Manuscript J2 likely once contained the entire text. A two-tablet version is also attested at Alalakh and probably in Mesopotamia (BM 16775). This is shown by the fact that B rev. IV ends with month six and leaves the rest of the tablet empty. Also, the note QATI, ‘(it is) finished’, is found in exemplar J2 after month seven. Remnants of two colophons survive, one in A and one in N, both after the fifth month.

The colophon of N notes that the tablet was a DUB GÍD.DA, ‘broad tablet’, i. e., a single column tablet, or that it was copied from such a tablet, the latter being more likely due to the short lines. Next to the broad tablets C and possibly F, the text was also written on tablets with at least two columns. With 13 months in total, a distribution of six or five on one tablet and seven or eight on the other seems plausible. Note, however, that QATI is written after months 12 and 13 in L. Although no Akkadian Vorlage survives from Ḫattuša, BM 16775 may be a remnant of an Old Babylonian multi-tablet version of the text that was to become EAE 22: The tablet starts with month 11 on the obverse and ends with month 13 on the reverse. Unless this was an excerpt tablet, it shows that the text of EAE 22 was not yet assigned to a single tablet in a series.

Remnants of rituals are found on B and F. A few minor variants in F that are difficult to explain as errors, but are in line with variations already in the Old Babylonian witnesses, may mean that there was more than one Akkadian origin version. This is also hinted at by Bo 6688, which has an omen on day 20 or 21, which is recorded for month 3, day 16 on other tablets, and with different surrounding omens, as it seems. It is also possible, however, that this fragment belonged to the almost entirely lost version of this text found only on the fragment KUB 8.29.

The copying error in J discussed above shows that copies were made from the Hittite translations.

Within the broader stream of tradition, CTH 532.3 can safely be considered a translation of the most common astrology text from the second millennium. The oldest witnesses are four tablets from Babylonia, in all likelihood from Sippar (BM 16775; BM 22696; BM 86381; BM 109154). Two Late Old Babylonian witnesses are known from northern Babylonia, possibly Dūr-Abiešuḫ (CUSAS 18.13 and 14). Middle Babylonian versions were found in Alalakh (AlT 451-452, with a most puzzling ‘orthography’) and Emar (Emar 6.652) as well as Susa (MDP 18.258) and Qatna (AO 12960). A Middle Assyrian tablet, probably of Babylonian origin (Rochberg-Halton 1988a: 273), is known from Nineveh (BM 12103). One Old Babylonian (lunar) eclipse omen text was found at Mari but is different in structure (ARM 26.248). The first-millennium version from Nineveh has been edited by Rochberg-Halton 1988a.

In addition, there is another entire version in Hittite from Ḫattuša, although a lot less well preserved (CTH 532.8), that shows the popularity of the text among ancient scribes. This, and the fact that already the Old Babylonian witnesses are not exact duplicates, makes it very difficult to pinpoint CTH 532.3 within this stream of tradition. However, there are a few indicators that, in accordance with geography and Hittite geopolitics, the Hittite version CTH 532.3 is closer to the Syrian versions rather than the Babylonian or Northern Mesopotamian ones. First, only the Ḫattuša tablets, AlT 451 and 452 (ištu UD 21KAM adi UD 30KAM), and Emar 6.652 (ina lādānīšu) have entries for the time between the 21st day of a month and the new moon, lacking in all other tablets from the second and first millennium. Second, AlT 452, Emar 6.652, and CTH 532.3 seem to be lacking an entry for month 2, day 20, while having one for the following day – assuming that the first unplaced fragment of Msk 74235a fits in the gap in month 2, which, according to Arnaud’s copy, should work. Strictly speaking, the Hittite text lacks the twenty-first day. But note that the omens of day 20 in CTH 532.3 are those for day 21 in the. g.e fragment from Susa and in the first-millennium version from Nineveh, and thus CTH 532.3 is in fact lacking the omen of the 20th day. Third, only AlT 452 and CTH 532.3 have an entry for day 17 of the first month. In AlT 452, obv. 36 the sign NE survives, which that text uses for zunnu, ‘rain’, likely based on a play with homophones: NE= ŠEG₆‘=’ŠÈG(A.AN). This fits the Ḫattuša apodosis (ḫé-ia-mu-uš DÙ-zi).

There are, however, also agreements of AlT 451-452 with Emar 6.652 against CTH 532.3: Both seem to only have an entry for days fourteen to sixteen and then from day sixteen to 30 in the third month (note that there is a surplus MIN in Arnaud’s copy, which is either an ancient scribal mistake or a copying error by the editor). Furthermore, the Alalakh tablets start with a non-monthly section much like the tablets from Babylonia. The Emar tablet ends with a version of the introduction to Enūma Anu Enlil.

The situation is further complicated when comparing the actual omens. For instance, the very first omen of month one, day fourteen seems to follow the Old Babylonian texts with a tripartite apodosis, but day fifteen is a translation of the shorter apodosis found in the first-millennium version, and day sixteen is close to the Old Babylonian version again. In month six, day fifteen, CTH 532.3 has a rather free Hittite rendering of the later EAE 22, against all Old Babylonian witnesses and the Emar text. But in the same month, day 21, CTH 532.3 agrees with the first-millennium version and the Emar tablet against the Old Babylonian tablets. In month twelve, day fifteen, the Hittite version differs from all other surviving witnesses.

The Akkadian Vorlage of CTH 532.3 thus already contained the full system of days fourteen to sixteen, followed by 20 and 21, which had not yet developed fully in Old Babylonian. However, it already sometimes shortens the rather long apodoses like the first-millennium version. It still had an entry for days 21-30, but already eliminated the introductory section about eclipses during the night watches – which is, however, found after the monthly section in the sister version CTH 532.8 (KBo 34.116 and duplicates). It is a good example of how, within the tradition, certain textual elements were modular and could be separated or added to a specific passage while preserving the text at large.

The text is one of three forerunners to Enūma Anu Enlil, tablet 22, from Ḫattuša, the others being KBo 34.116 and duplicates and KUB 8.29. It lists eclipse omens from month one to twelve and in a thirteenth, intercalary month, on days fourteen, fifteen, sixteen, 20, 21, and 21-30 of each month respectively. The phrase to describe the eclipse is D30-aš a-ki: ‘Armaš/The moon dies’, like in CTH 532.2; CTH 532.4, and CTH 532.6. The colophon of ex. A has ma-a-an […] pu-uš-zi where we expect the name of the composition, which implies that ak- and puš- are synonyms when relating to the moon. The apodoses closely follow the Akkadian versions known from Mesopotamia and Syria (see Text transmission) and deal with flooding and rain, harvest, the king, warfare, and the welfare of people and country. When an omen survives both in CTH 532.3 and in CTH 532.8, there are no substantial differences other than the phrase for the eclipse. At times, the Hittite translation seems to have misunderstood its Akkadian original, or that original may have been corrupt. We find tu-li-ia-aš DÙ-ri in month one, day fifteen, where we expect a translation of Akkadian ubbuṭu or ḫušahhu, for example. Month twelve, day fifteen, may have an otherwise unattested apodosis, but the text is fragmentary.

A: New Script: The only surviving diagnostic form is new TAR, with an angular wedge underneath a broken, only slightly sloped horizontal. The first vertical of E is of similar height to the second one, considered diagnostic for NS IIIa by Starke F. 1985a: 24, but cf. The caveats concerning the chronological relevance of the height of inscribed vertical wedges in Leonard T.D. 2025a*: 150.160; G. Wilhelm apud Seeher J. 2005b: 77.

B: New Script: There is KI with two verticals (rev. III? 9′); ḪA with only one Winkelhaken (rev. III?, 11′), KAM is written as ḪI+BAD (rev. III, 7′); there is an old AG in rev. III, 15′.

C: New Script: URU has a protruding middle horizontal (obv., 7′); there is a late LI in obv. 4′.

D: New Script: ID has lost the central broken horizontal and is once written with an additional small vertical between the ḪI-group and the horizontals (3′; HZL 215/25). The heads of the verticals and the tips of the Winkelhaken can be elongated and pointy, very similar to those in manuscripts H and J1.

E: Middle Script or New Script: Good diagnostics are missing. ŠA with a high first vertical has been considered diagnostic for New Script tablets by Starke F. 1985a: 24, but Gordin S. 2015a: 91 pointed out that this form already appears in some older land grant documents. The script is somewhat cursive, with the horizontals sloping slightly downward.

F: New Script: Old AG; The initial verticals of E and URU match the second one, as does the central vertical of RU, but see the caveats mentioned under A. Note that KAM is written ḪI+BAD obv., 7′. If Giorgieri M. – Mora C. 2004a: 38 are correct in their assumption that the variant ḪI+BAD for KAM is an Assyrian import, it would have to be counted among the IIIb or IIIc signs. In the case of tradition-literature such as this text, this can of course stem from an Akkadian Vorlage.

G: Middle Script or New Script: Only a small fragment remains without diagnostic signs; the generous word spacing and appearance of the wedges show that it cannot be Old Script.

H: New Script?: Only a small fragment remains without diagnostic signs. The fragment looks very similar to the two other fragments from the ‘house on the slope’, manuscripts D and J1 (note the sometimes drawn-out and pointy Winkelhaken and heads of the verticals), and may in fact stem from the same tablet. The other two show clear New Script sign forms.

I: Middle Script or New Script: There are not many signs left, no helpful diagnostic signs, and the surface is very abraded. The spacing and the heads of the verticals show that it is unlikely to be an Old Script tablet.

J1: New Script: The fragment uses old LI, but DI has a second vertical after the initial Winkelhaken, and URU has a protruding middle horizontal. The only visible instance of AZ has no subscript. It may belong to the same tablet as KBo 13.14 and KBo 13.16.

J2: New Script: New AG; DU is written with a fourth horizontal instead of a Winkelhaken; Á and DA have lost their broken central horizontal. Like D, H, and J1, the Winkelhaken can be pointy and elongated, but in J2 they frequently extend over the paragraph lines, which they tend not to do in KBo 13.15 (J1). Different from J1 in their executions are the signs Á, DA, E, NA, IŠ, ŠA, and UŠ, and the spacing of their lines differs substantially. I therefore reject the indirect join proposed by Riemschneider K.K. 2004a: 4.65 (some of the differences were kindly brought to my attention by Jared Miller via personal communication).

K: Middle Script or New Script: ID has a fourth Winkelhaken in the upper row of the ḪI-Group. TA has very small verticals, but according to Weeden M. 2011a: 45-46, this is only relevant for separating Old from Middle Script. The signs look somewhat low and compressed, cf. especially the elongated MA with an indented central horizontal. This form of MA and overall compressed sign appearance can be found in Middle Script (cf. the solar omen tablet KUB 8.17), but there is yet no systematic analysis of sign height and this type of MA to tell whether this is a feature of Middle Script tablets in particular. The script looks similar to manuscripts L and M; perhaps the fragments once belonged to the same tablet.

L: Middle Script or New Script: In the text body, we find the late Middle Script/IIc forms of ḪAR and URU (Starke F. 1985a: 24) and the old forms of IG and KÙ. The small remaining section of the colophon, however, seems to write QA-TI using the QA-version with two crossing horizontals, considered to be a very late sign by Theo van den Hout (Weeden M. 2011a: 49, citing the unpublished part of van den Hout’s dissertation). Either this is an earlier instance of this sign variant or a very diligent copy of an older text. The script looks similar to K and M; perhaps they were part of the same tablet.

M: Middle Script or New Script: We find the Late Middle Script forms of AḪ and URU. GAR still has a very small central upper vertical, but the first vertical of E is of similar height to the second one, which could speak for a later date (Starke F. 1985a: 24), but see the comment on A. The signs tend to look ‘compressed’, see especially the A in 16’, with vertical consisting of almost nothing but the head. This could perhaps belong to the same tablet as the fragments K and L.

N: Middle Script?: Not many signs are preserved, but there is an old AḪ in obv. 6, and, although damaged, URU in obv. 4 has a visibly smaller first vertical.

O: Middle Script?: Less than 20 signs are preserved; URU in 3′ has its old form with a small first vertical as is often found in Middle Script, AZ is written with a subscript. The fragment could belong to the same tablet as manuscript N.

P: New Script?: Only a small fragment remains without signs that are conventionally considered diagnostic. However, KAM is once misspelled as ḪI, which points to the use of KAM as ḪI+BAD. If Giorgieri M. – Mora C. 2004a: 38 are correct, this is an Assyrian loan, which would make an Empire-period date likely.

While some fragments are too small to evaluate their paleography precisely, in sum, we can assume that there existed a Middle Script and a New Script manuscript tradition of this text. Of the New Script tablets, B, C, D, J1, and J2 contain signs that are considered to be diagnostic for the latest period of Hittite writing (IIIb in C, D, J2; IIIc in B, J1) by Starke F. 1985a: 25; Klinger J. 1996a: 38, see also Gordin S. 2015a: 90-91; van de Peut L.E. 2022a*: 19-29.