Ancient writing systems in the Mediterranean

A critical guide to electronic resources


- Second half of the 24th century B.C. - Second half of the 17th century B.C.

edited by: Amalia Catagnoti (translation revised by Melanie Rockenhaus)

  • Introduction
  • Index
  • Further information

Obverse of the treaty between Ebla and Abarsal (TM.75.G.2420 = ARET XIII 5),Archive L.2769, Royal Palace G, 24th cent. BC

The graphic system used in Ebla (Tell-Mardikh), Syria, during the Early Syrian period (24th century B.C.) is a type of mid-3rd millennium cuneiform, similar to that found in Mesopotamia (from Kish in the north to various Sumerian sites in the south) and in eastern Syria, at Mari (Tell Hariri) and Nabada (Tell Beydar), with some distinguishing differences.
Since this writing system was already known at the time of the archaeological discovery of the Palace G archives (1974-1975), it was easy to decipher. Studies have since focused instead on peculiarities related to the syllabary and to the use of logograms in this complex writing system that was adapted to be used to represent a non-Mesopotamian Semitic language.
The few cuneiforms texts – a dedicatory inscription on the statue of King Yibbiṭ-Lim from 20th century B.C. and various tablets from 17th century B.C. – recovered at Tell Mardikh date to periods later than Palace G. The paleography and syllabary of these demonstrate a clear discontinuity from previous traditions, and can instead be assimilated to those used during the same period in Mari and Alalakh (Tell Atchana).

Go to the online resources.

Online resources


Discovery of Ebla texts and places of finding

The site of Tell Mardikh, Syria (located about 60 km southwest of Aleppo) has been explored since 1964 by the the Sapienza University of Rome under the direction of Paolo Matthiae. The ruins of ancient Ebla were found there.

This urban center first flourished during the zenith of the Early Syrian period (ca. 2600 - 2300 B.C.). At the end of this period a fire destroyed the royal palace (Palace G). Another great fire also ended the successive Ebla (ca. 2300 - 2000 B.C.). The third, and last, Ebla flourished during the early and later Old Syrian period (ca. 2000 - 1600 B.C.), when the city was definitively destroyed by the Hittites. Ebla was frequently a center of international importance during its thousand-year history: a first Ebla princess married the king of Kish, the second Ebla was well-known to the king of Lagash Gudea, and proof of the importance of the final and last Ebla can be seen in the bilingual Hurro-Hittite Song of Release recovered at Hattusa.


The Texts of the Early Bronze Age  
Royal Palace G (A. Archi, SMS 5/2, fig. 1)

Cuneiform texts from the third millennium B.C. have been recovered in nine rooms of the western part of the Royal Palace G. The majority were preserved in the so-called Great Archive (C), to which the tablets recovered in the area of the Court of Audience (D, F) likewise belonged. Smaller archives (B, E) and other tablets (A, G, H) were found up to 40 meters north and 60 meters south of the Great Archive. Only one (older) tablet comes from the northwest wing of the Central Complex (I).
The first 42 tablets were found, with a bulla, in 1974, on the floor of a room in the Administrative Quarter near the bottom of a jar, in L.2586 (A). These administrative texts mostly concern precious metals (MEE 1.1-39; Archi 1986: 75 f., 1996a: 65; Biga 1988: 285-287).

Around 250 administrative texts recording alimentary goods, which must have been stored on two freestanding shelves, were found in L.2712, the so-called Small Archive (B), in the north corner of the east porch of the Court of Audience (MEE 1.40-620; Archi 1986: 73-75, 1996a: 64 f.; Milano 1988: 288-290).
The most important Palace G Archive, located in room L.2769 (C), in a vestibule in the east porch of the Court of Audience, contained different kinds of texts (administrative, epistolary, diplomatic, lexical, literary). These were for the most part arranged on three wooden shelves placed against three walls of the room (Archi 1986: 77-86, 1996a: 60-62).
Near L.2769, in vestibule L.2875, a smaller number of tablets (D) was found; these included letters, royal ordinances and agricultural administrative texts. The recovery of scribal tools (fragments of bone styluses and a stone smoothing tool) suggest that clay tablets were written in this room. The tablets were probably located on two low banquettes. (MEE 1.6246-6518; Archi 1996a: 62 f., 1986:76 f.; Biga 1988: 291-299; Matthiae 2008: 72).
Near an inner courtyard of the Administrative Quarter, in L.2764, in the so-called Trapezoidal Archive (E), some dozens of administrative texts of various contents were found. Many of these are in extremely fragmentary condition, they were found on the floor but must have been kept on the benches on the west and north walls, in the latter case closed by wooden panels. (MEE 1.621-654; Archi 1986: 75, 1996a: 63 f.; Biga 1988: 300-304; Matthiae 2008: 71-72).
At the center of the Court of Audience, in L.2752 (F), 22 administrative tablets were found on a burnt plank. These concern various subjects, and some are joints of fragments recovered in L.2769 (Archi 1996a: 61 f., 1986: 76; Biga 1988: 305 f.).
Tablets were also present in other parts of Palace G: the Southern Quarter, where, in 1982, five administrative texts (G) were found in L.3143, L.3462 and L.3474 (Archi 1993). These few tablets found in peripheral sectors of the palace dealt with matters, such as the distribution of timber, that were different from those of the records in the other archives of the administrative quarter and may suggest that other small archives could be found in other sectors of the palace, yet to be identified (Matthiae 2008:91). In the two rooms, L.8496+L.8778 and L.8495, built against the east wall of the Throne Room, L.2866, 13 tablets (H) were found in 2004. These texts were considered to be related to operations that were taking place in the Throne Room itself, and it was assumed that after a short time, they would have been transferred to the Great Archives (Matthiae 2008:75).
In L.8606, in the northwest wing of the Central Complex, a tablet (I) was found in 2003, together with a votive plaque; the tablet should pertain to a period earlier than that of the archives (Archi 2016, p. 2 note 8).

Excavations in the Royal Palace G have allowed us to inventory around 17,000 inscribed objects. It can be supposed that there were originally around 4,000/5,000 tablets. The majority of these texts document the spelling of the local Semitic language made by Semitophone scribes according to their own conventions. These scribes were also competent in both Semitic and Sumerian texts originating from distant Mesopotamian areas.
In addition to texts written in cuneiform on clay tablets, cretules with inscriptions and bullas have been found (Archi 1996b; Peyronel 2016).


The texts of the Middle Bronze Age
The dedicatory inscription on the torso of the statue of King Yibbiṭ-Lim of Ebla, found in 1968 (TM.68.G.61), which probably dates to 20th century B.C., is the most ancient of the few documents in our possession that illustrate the Semitic language spoken at Ebla during the Middle Bronze Age. Cultural materials also demonstrate that this was a period of great development for the city.
Very few documents written in the period of Alalakh VII (17th century B.C.) have been recovered at Ebla. One of these, written on a clay tablet, is a brief Old-Babylonian letter concerning agricultural works (Kupper 1980). Other texts have been published later (Kupper 2005; Durand 2018). Brief inscriptions on a seal, on a cup and on a few other tablets recovered during the last archaeological campaign have also been reported.
J.-R. Kupper, Une tablette paléo-babylonienne de Mardikh III, SEb 2 (1980), pp. 49-51.

The tablets of the Early Bronze Age

The comparison between tablets that contain texts of different typologies shows that the scribes of the Royal Palace G of Ebla operated in accordance with strict rules that concerned the standardization of tablet formats, among other things. The analysis of these formats can furnish useful information for the reconstruction of the relative chronology of the Ebla documents and for the identification of the working habits of the various (groups of) scribes.

The main formats can be described as follows (Pettinato, MEE 1, p. XVII; Archi 1996a: 73 f.):

a) roundish tablets, of small and medium-small dimensions;

b) quadrangular tablets with smooth edges and roundish angles, of medium-large dimensions;

c) quadrangular tablets with squared edges and angles, of large and very large dimensions.

 Among the very few exceptional texts, TM.75.G.2414 (= MEE 15.29) is an extremely lengthy monolingual Sumerian lexical list, with only two columns on the obverse, each with 43 places, and with the reverse blank (Archi 1992: 16; Picchioni 1997: 94-96 and Tav. VII). Regarding the dimensions of the tablets, there were:

1) tablets of small dimensions, roundish, loaf-shaped or with a lenticular shape, about 2-7 cm in diameter;

2) tablets of small to medium dimensions, quadrangular, with roundish edges and angles, about 7-10 cm per side;

3) tablets of medium dimensions, quadrangular, with roundish angles, about 10-18 cm per side;

4) tablets of large dimensions, quadrangular, with squared angles, about 18-26 cm per side;

5) tablets of very large dimensions, quadrangular, with squared angles, about 26-37 cm per side.

There was an administrative and scribal reform when Yiṯġar-Damu (11-ar-da-mu) became king and Yibrium (Ib-rí-um) prime minister (Archi 1991:212).

Traditions and textual typology

With the exception of royal inscriptions, cuneiform texts belonging to all the main textual typologies were found in Ebla. It is useful, on one hand, to contrast the texts of local traditions with those of Mesopotamian tradition, and on the other, the textes de la pratique with the literary texts.
The main group is comprised of local texts, with the majority being documents that refer to events in Ebla itself, or in neighboring areas of interest for the Eblaite elite during the lifespan of the archives. There are a few thousand bookkeeping accounts, but also letters, accounts of rituals, and documents of varying content (treaties, ordinances etc.) with common juridical bases (ARET XI, XIII, XVI, XVIII). These materials furnish important quantitative and qualitative information on the economic, political and social situation of the Ebla kingdom and of the Syrian kingdoms in general. However, it should be noted that there are also some local literary texts, for example Semitic incantations in reference to Syrian cults (ARET V).

Not surprisingly, the texts of non-local tradition form a smaller group. In fact, they are texts of Mesopotamian tradition, where the content is mainly either religious (ARET V: Sumerian incantations, but also Semitic compositions of literary form, like hymns to the gods) or lexical (MEE 3, MEE 15: Sumerian lists belonging both to the Uruk and Fara traditions). However, the Sumerian-Eblaite bilingual lexical lists (MEE 4) occupy a special place, between the texts of local and those of Mesopotamian tradition. These texts belonged to a non-local tradition, but they were largely reworked, most probably in Ebla. These texts list Sumerian terms that are often translated (or glossed) with local Semitic terms, and they are organized according to the acrographic principle, where items are grouped according to the first sign used to write them. In general, all of the texts of Mesopotamian tradition elaborated in Ebla allow a glimpse into the practices of the scribes of the Palace G; the number of these scribes appears to be particularly low in comparison to that of the great Mesopotamian centers of the third millennium.
As for textual typologies, administrative documents form the main group in all the archives and in the lots of tablets from the Palace G. In particular, the texts listing alimentary product deliveries (ARET IX, X) were preserved in L.2712, while lists of cereal quantities and their places of production were part of the group of tablets found in L.2764, and lists of deliveries of wine jars were found in the rooms of the Southern Wing of the Palace G (SMS 5/2).

In contrast, the Great Archive L.2769 + L.2752 was not specialized. It contained literary, lexical, and chancery texts (letters, real ordinances, treaties, donations) as well as texts of rituals in which the last two kings of Ebla participated. It likewise included various kinds of administrative documents (ARET I, II, III, IV, VII, VIII, XII, XV, XIX, XX; MEE 2, 7, 10 12), mainly concerning the bookkeeping of textiles and metals, but also agricultural matters (such as cultivation and breeding).

Considering the texts of this archive more in detail, the most ancient monthly accounts of textiles (MAT) were written on quadrangular tablets (16-18x19-21 cm) with right angles, flat on the obverse and centrally convex on the reverse. More recent ones, however, were characterized by an increase in the dimensions of the tablets (up to around 21 cm per side) and by smaller writing. As for the annual accounts of metals (AAM), the older ones were written on large rectangular tablets with smoothed corners (around 20 cm per side). While the dimensions of later AAMs increased (up to 32 cm high and 37 cm wide) and the angles became squared, the reverse convexity remained and was now found on two-thirds of the obverse as well. The annual recordings of metals and textiles were written on tablets of medium-sized dimensions (around 13x16 cm), with smoothed angles and convex on the reverse, as were the accounts of agricultural products, which were, however, written on tablets that were not baked. The monthly accounts of the distribution of sheep were written on similar tablets, although a little bit larger (around 19x22 cm). The lexical lists were written on tablets (in general around 15x17 cm) with right angles, convex on the reverse, using pure clay and very accurate writing of signs. Well-baked, they faithfully reproduced the Sumerian lists in a number of columns on the obverse, and on the reverse only the colophon, which in most cases carried the name of the Eblaite scribe.

The graphic system

The terms used in the Eblaite cuneiform texts of the third millennium were written using both syllabograms and logograms.

The syllabograms were normally used to denote terms in the local Semitic language, including proper names (the personal name Ìr-kab-da-mu, the geographical name Ib-laki, the divine name dRa-sa-ap and the month name Ig-za) and common names (ʾà-da-umtúg, "mantle"; da-mu, "blood; stock"; du-ba-lum, "pasture"; wi-rí-gúm, "orchard, garden"), pronouns (an-da, "you"), adjectives (ra-gu, "fine, thin"), prepositions (si-in, "to, toward; for"), conjunctions (ù, "and"), adverbs (a, "where?") and conjugated or declined verbal forms (ne-sa-bar, "we send"; wa-ba-lu, "to bring", mu-da-li-gú, "he who continually moves").

In particular, note:

- Ìr-kab-da-mu, means Yirkab-damu, "The-lineage-has-triumphed": the name of the penultimate king of Ebla (in which are distinguishable the terms rakābum, "to mount", in this case meaning "to triumph", and damum, "blood", in this instance meaning "lineage") is written using a syllabogram of the type V(ocal) C(onsonant), ìr, a syllabogram CVC, kab, and two syllabograms CV, da and mu;

- du-ba-lum should mean dubrum, "pasture" (meaning suggested by the equivalence ú-síg = du-ba-lu in the bilingual Sumerian-Eblaite lexical list, and from the comparison with the Semitic *dubr-): in this case note the use of lum to express /rum/ and the use of the sign Ca (here -ba-) to express only the consonant /b/;

- mu-da-li-gú means muhtallikum, from *muhtanlikum (participle tn/1 of the verb halākum, "to go"): the missing notation of the double consonant has similarities in Akkadian texts. Instead, the use of to express /k/, opposed to the use of gu to express /q/ (for example in the Sumerian-Eblaite lexical equivalence níg-sal = ra-gu, "fine, thin"), is a typical feature of the syllabary used in Ebla.

In addition to others (for example the use of the value ru12 of the sign EN), these cases illustrate some peculiarities of the Eblaite syllabary, distinguishing it from the contemporary Semitic syllabary of Mesopotamia. Common features include the plurality of readings for single signs (the sign LUM, for example, could be read as lum, gúm, núm, ḫum) and the plurality of phonetic values for individual readings (for example, ʾà-la-gúm, /halāk-um/, "to go", where the reading gúm of the sign LUM appears, with the value /kum/). The writing does not distinguish among voiceless, voiced and emphatic consonants, except some cases that show the ability of the Eblaite scribes attempting to adapt the writing to their own language (the sign gu for the phoneme /q/, distinguished by for the phonemes /g, k/, see above).

Some Sumerograms appear in some of the Semitic terms listed above as determinatives: ki, "place", in Ib-laki, dingir, "god", in dRa-sa-ap and túg, "cloth", in ʾà-da-umtúg.

As always in documents written by Semitophone scribes, also at Ebla a large number of Semitic terms were noted down with what was considered to be their Sumerian equivalents. The practical purpose of the Sumerian-Eblaite bilingual lexical lists was to provide the local equivalent of the Sumerograms used in the Eblaite textes de la pratique, for example in the case of šeš-mu = a-ḫu-um, "brother". The use of Sumerograms was also extended to frequently used words such as en "king", ninda "bread", níg-ba "gift", šu-mu-taka4 "to deliver." Pluralization of Sumerograms was done by duplicating the word (lugal-lugal "the lords", dumu-nita-dumu-nita "the children", giš-gígir-sum-giš-gígir-sum "the transport wagons"). Rarer, but equally attested, are the Semitograms. Some can be defined as Akkadograms, or examples with a clearly Semitic aspect, but which were treated as logograms and not as terms written syllabically. Some examples are Semitograms that end with a genitive inflection (bu-di "pin"; na-se11 "person, people", whose vocalism contrasts with that of the Eblaite corresponding term, *nišum), while others ended without inflections (li-im "clan") and still others with the nominative ending (ma-lik-tum "queen", da-mu "lineage"). All of these Semitograms show a doubled form in the plural, thus they are treated as common Sumerograms.