A linguistically isolated, extinct language, Sumerian is preserved only on clay tablets, in a considerable corpus of texts, written in cuneiform. The tablets, dating from about 2,800 to 100 BCE, have been found in excavations in Iraq and, to a lesser degree, in other areas of the Near East, most notably northern Syria, but also as far as Susa in Elam, Boğazköy in Anatolia, Palestine, and el-Amarna in Egypt.
Linguistic and Historical Background.
The native name of Sumerian is eme-gir (Akk., lišan šumeri, “language of the Sumerians”). A dialect used by women is called emesal (Akk., lurǔ, “woman's language”). All attempts to connect Sumerian to known linguistic families have so far failed. Typologically, it belongs to the languages with a subject-object-verb order, postpositions instead of prepositions, and adjectives following the noun. The roots tend to be monosyllabic, but the language can be considered to have an agglutinative morphology—one in which a word may consist of more than one morpheme, but with clear-cut morpheme boundaries.
From the beginning of historical times, Sumerian was in contact with languages of the Semitic family. It appears, in fact, that the early texts, largely logographic, could be read either as Sumerian or as Semitic. After 2,000 BCE, Akkadian, a Semitic language, became dominant and Sumerian was relegated to the status of a literary language. [See Akkadian.]
The phonology of Sumerian can be reconstructed up to a point with the help of native syllabary tablets. The tablets provide a phonological definition of a word in terms of a set of basic syllabograms, as shown in the examples provided in figure 1. The basic syllabograms in the left subcolumn are the same ones used to write Akkadian, a Semitic language whose phonetic structure is better known. This allows an approximate reconstruction of the Sumerian words, although some phonological details are undoubtedly lost in the process. The resulting phonological inventory of Sumerian consonants is provided in table 1. There are some uncertain points, such as the exact nature of /h/ (glottal?), and /š/ (interdental?). This Akkadian interpretation of the Sumerian phonological system quite possibly involves some degree of underdifferentiation. In its earlier stages, the writing system did not distinguish between voiceless and voiced stops. It is not known whether this type of simplification extended to other phonological features. There is also a tendency to simplify consonantic clusters in writing. The existence of glides or semivowels (*w, *y) is suggested by indirect orthographic evidence.
Structural conditions limit the coexistence of some consonants in the same root. For instance, in a root of the form C(onsonant)1-V(owel)-C(onsonant)2, the two consonants cannot be labial; if C1 is /h/, C2 cannot be a velar; if C2 is /h/, C1 cannot be nasal or velar. The consonant /r/ is almost never found as word initial.
The writing shows four vowels—a, e, i, and u—but there are indications that the system was more complex. Bound morphemes and compound words exhibit vowel harmony. There is no indication in the writing of dynamic stress or tone; given the structure (mostly monosyllabic) of the roots, the presence of a tonal system is a reasonable, but unprovable, hypothesis.
Table 1. Sumerian Consonants
The discussion of Sumerian syntax may be divided into nominal phrases, verbal phrases, and clause syntax.
A simple nominal phrase consists of a stem optionally followed by a possessive suffix, and/or a plural suffix, and obligatorily by a postposition. The distinction between verbal and nominal stems is functional, with no phonological or morphological marks. There is no gender distinction either in the form of the stem or by affixation. If the gender needs to be stressed, this is done by adding the appropriate adjective (nita, “male”; or munus, “female”). The nominal stems are animate (humans, deities, personified beings) or inanimate; this feature governs the choice of certain affixes. There is no general system to express number. In nouns of the animate class, the plural can be expressed optionally by the suffix -e-ne. Nouns, regardless of class, and adjectives can be reduplicated, often with a nuance of totality (with nouns), or intensity (with adjectives). Nouns, with no plural marker, can function as collectives, taking plural markers in the verb. The personal pronouns and possessive suffixes are shown in table 2.
The postpositions are
|-zero||subject of intransitive verbs, direct object of transitive ones|
|-e||subject (agentive) of transitive verbs, second object with verbs that take two objects|
|-ta||ablative (“from”), instrumental|
The genitive is marked by the suffix -ak (written simply -a, unless followed by a vowel). A genitival relationship can also be expressed by moving the head of the genitive construction to second position and adding a possessive: giš-hur é-a(k), “the plans of the house,” can be expressed also by é-a(k) giš-hur-bi, “of the house its plans.”
A stem in a verbal function is preceded and followed by several classes of affixes, as shown in table 3. Not shown in the table are a postradical infix, -d-, indicating obligation, and pronominal infixes (-e-, -n-, -b-) that can be inserted before the comitative, the directive, and the stem. There are two tenses: a past/complete action and a present-future/incomplete. They are marked by affixation of -e, reduplication, or, in the case of some common verbs, different stems (suppletive paradigm). Some verbs have different stems for the singular and plural. The imperative is formed by moving the verbal stem to initial position: muna-ab-sum (“he gave to him”), sum-mu-na-ab (“give to him!”).
The verbal agreement is of the ergative type. Subordinate clauses (completive or adverbial) have, as a rule, a nominalizing suffix, -a, followed by a postposition, and are thus treated as any nominal phrase. Subordinate clauses precede the main clause. Relative clauses may have an explicit head (lú, “the one, the man”; or nig, “the thing”) or an unmarked head noun and also have the suffix -a.
A very large number of the preserved Sumerian texts, probably about 90 percent, are practical texts: administrative accounts, letters, and contracts and other legal documents, public or private. The rest are “scientific” texts (mostly in the form of lexical word lists and other inventories of knowledge) and literary texts.
Table 2. Sumerian Personal Pronouns and Possessive Suffixes
|3d person animate||e-ne||e-ne-ne||-a-ni||-e-ne-ne|
|3d person inanimate||-bi||-bi|
Table 3. Sumerian Verb Forms
|zero||inga||V (i-, e-, a-)||-a-||-da-||-s̆i-||-ni-||| |||zero|
Among the oldest texts, Uruk III–IV (c. 2900 BCE), a large number of word lists is already found, and these are still found among the latest tablets, almost at the beginning of the Christian era. Besides their didactic value in training scribes, and their linguistic contents, these texts are remarkable for their preoccupation with classifying and making an inventory of human knowledge. Lexical texts assume standard, traditional forms that are kept unchanged for centuries over large geographic areas. They thus played an important role in the shaping and transmission of Mesopotamian culture.
The oldest lists are unilingual, but there indications that they could be read either in Sumerian or in Semitic. Later lists tend to be bilingual—Sumerian and Akkadian. Some of them include a subcolumn with the pronunciation, as in the example above. There are even a few lists in several languages: Sumerian-Akkadian-Hittite, or Sumerian-Akkadian-Ugaritic-Hurrian. [See Hittite; Ugaritic; Hurrian.] The lexical lists can be classified into two major types: lists ordered by signs and thematic lists. As an example of the first, the collection á = Á = nâqu gives, in forty-two tablets, the pronunciation and Akkadian translation of about fourteen thousand entries. The most remarkable of the thematic lists is the collection HAR-ra = hubullu. Twenty-four tablets list the terms designating, for instance, trees and wooden implements, stones, metal objects, fish, and birds. The complete collection has about 9,700 entries.
Tablets with texts of an undeniably literary character already appear in the Fara period (c. 2500 BCE). They are written in an extremely abbreviated form and can be only partly understood—except in a few cases in which later recensions are preserved. Some better-understood literary compositions date to the Sargonic period, and the lengthy commemorative inscriptions of Gudea, ruler of Lagash, shortly after the Sargonic period, show a full-fledged literary language and style. The Ur III Empire seems to have been a period of intense literary creativity, but few tablets from this period are preserved. Its literary works are known principally from tablets from two or three centuries later, possibly in an updated and revised form. The majority of literary works are known from tablets dating to the eighteenth century BCE, mostly from Ur, Nippur, and Babylon. No major works were created after this time. Copies of older texts, sometimes with a corrupt text, recompilations of liturgical chants, and a few royal inscriptions are all that were produced in later periods.
The average Sumerian literary composition is between one hundred and three hundred lines long. There are exceptions: short texts and exceptionally long stories like The Deeds of God Ninurta, with 728 lines. Some compositions are preserved in a single copy and may be only partially recovered; others are attested in eighty or more duplicates and their text may be complete. The main reason for the high number of duplicates is the use of literary texts for school practice; a well-educated scribe was apparently expected to memorize the main works of literature. Some catalogs give the titles (i.e., the opening lines) of the compositions in the order in which they were studied. The written, preserved works seem to be only a part of a more extensive oral literature, now lost. [See Scribes and Scribal Techniques; Writing and Writing Systems.]
The literary style often reflects this oral origin. Some texts are known in different recensions, which write down divergent oral forms of what is essentially the same composition. The main features of style are parallelism of various sorts and simple, direct similes of the type “like a cloud drifting in the sky.” Many are taken from a traditional stock and appear again and again in different works. Verbatim repetitions of speeches and event descriptions are common in narrative texts. Some texts show an evident strophic structure and even indications of some sort of meter. The imperfect knowledge of Sumerian phonology makes almost impossible any prosodic reconstruction.
Major literary genres
About two dozen narrative texts describe events with divine protagonists that take place in a supernatural world. They revolve around the gods Enki, Enlil, Inanna (and Dumuzi), Ninurta, and other deities.
Two cycles of narrative texts with superhuman protagonists have been preserved: one deals with Gilgamesh (five compositions) and the other with Lugalbanda and Enmerkar (four compositions). Both cycles originated in Uruk.
There is a very large number of commemorative royal inscriptions, from all places and dynasties. Some of them can be quite extensive, such as Gudea's cylinders commemorating the building of Ningirsu's temple. In addition, there is a dynastic list and more literary texts centered around historical events, such as an explanation of why the city of Agade was destroyed, and laments about the destruction of other cities. Under this rubric is a huge number of more or less poetic texts in praise (sometimes self-praise) of the rulers and prayers of, or for, the rulers to various deities.
Besides business letters—which are almost always very terse, factual messages—about seventy-five more-extensive letters, used as school exercises, are preserved. They include royal letters and letters of prayers to the gods.
About 125 texts praise various deities or make petitions to them. A collection of forty-two brief songs celebrates the most famous shrines at Sumer, and an archaic composition commemorates the holy city of Kesh.
The Instructions of Shuruppak, perhaps dating as far back as 2400 BCE, is a collection of aphorisms giving ethical advice. Man and His God is a series of reflections on how God allows evil and misfortune to befall a faithful man. On a mundane level, the Instructions of a Farmer explains the proper way to cultivate barley. Also included in the category of didactic texts are collections of medical prescriptions and hemerologies (lists of unfavorable events on given days).
Half a dozen texts are literary contests between personifications of natural entities (Summer and Winter, Bird and Fish, Tree and Reed, Grain and Sheep, Silver and Copper) or crafted implements (Hoe and Plow). They conclude with a verdict, given by a deity, explaining the reasons for the superiority of one above the other.
Although practically all literary texts are preserved in copies used for didactic purposes, one group deals specifically with school activities. Half a dozen texts describe life in school directly or indirectly, give advice to students, and describe the conflict between a father and his son, an unsatisfactory pupil.
The preceding school texts make frequent use of dialogue, but five compositions consist of nothing but dialogue: two between students, one between apprentice musicians, and two between women.
Tales, fables, and proverbs.
Popular tales, such as The Three Men from Adab, are very rare, and there are a few fables (The Crane and the Raven, The Heron and the Turtle), but many hundreds of proverbs were used as relatively elementary exercises in school. About thirty riddles are preserved.
Many texts are not easily adscribed to the preceding genres: two elegies of a scribe lamenting the death of his parents, a message of an exiled scribe to his mother, a song in praise of the hoe, and the songs used by the king in the ritual opening of the agricultural season.
Compositions sung on ritual occasions are the most common form of literary text between the sixteenth and second centuries BCE. They are mainly laments over the destruction of ancient shrines and prayers for their reconstruction.[
- Civil, Miguel. “Lexicography.” In Sumerological Studies in Honor of Thorkild Jacobsen on His Seventieth Birthday, June 7, 1974, edited by Steven J. Lieberman, pp. 123–157. Chicago, 1976. .
- Edzard, Dietz O. “Literatur.” In Reallexikon der Assyriologie und Vorderasiatischen Archäologie, vol. 7, pp. 35–48. Berlin and New York, 1987. .
- Hallo, William W. “Toward a History of Sumerian Literature.” In Sumerological Studies in Honor of Thorkild Jacobsen on His Seventieth Birthday, June 7, 1974, edited by Steven J. Lieberman, pp. 181–204. Chicago, 1976. .
- Römer, W. H. P. Die Sumerologie: Versuch einer Einführung in den Forschungsstand nebst einer Bibliographie in Auswahl. Alter Orient und Altes Testament, vol. 238. Neukirchen-Vluyn, 1994. .
- Thomsen, Marie-Louise. The Sumerian Language: An Introduction to Its History and Grammatical Structure. Copenhagen, 1984. .