Sumerian

Sumerian was a language spoken in southern Mesopotamia and is most likely first attested in the archaic texts from the end of the fourth millennium bce. By the end of the third millennium, Sumerian had died out as a spoken language. However, it was still used in literary, scholarly, and religious genres, and was preserved in writing until the disappearance of Mesopotamian civilization. Sumerian is not related to any other language. Thus, our knowledge of Sumerian grammar and lexicon is mostly based on bilingual texts in Sumerian and Akkadian, and lexical lists and grammatical texts.

Some scholars believe that Sumerian and its speakers did not enter southern Mesopotamia until around 2900 bce. If so, the archaic texts from the Late Uruk period were probably not Sumerian. However, an important factor is that there are some instances ofphonetic writing in Late Uruk texts, which point to Sumerian as the language of these texts.

Landsberger’s 1974 suggestion of a hypothetical pre-Sumerian substratum has been quite influential. This alleged lexical substratum would constitute the only remains of a hypothetical human group that would have inhabited southern

Mesopotamia before the speakers of Sumerian. The core of this substratum included designations for occupations and trades (a ''s g a b ‘‘leather worker,’’ a z l a g ‘‘launderer,’’ b a h a r ‘‘potter’’). The criterion for the identification of non-Sumerian words was that they are polysyllabic, while Sumerian prefers monosyllables. But one has to conclude that most of these items happen to be Semitic loanwords, Hurrian, words occurring in many languages, words that travel with the objects they name, or Sumerian terms (Rubio 1999).

Nowadays, it is believed that Sumerian died out during the Ur III period. Sumerian was probably still spoken in school, as Latin was spoken in the Middle Ages. Nonetheless, Sumerian remained in use for another two millennia, as a literary, scholarly, and liturgical language. The vast maj ority of Sumerian texts date to the long period between the death of Sumerian as a native tongue and the final disappearance of cuneiform writing and the Mesopotamian languages during the first centuries of the Christian era.

Overview of Sumerian

Cuneiform was the script used for Sumerian. Our knowledge of Sumerian phonology is limited by the nature of its writing. For instance, it is likely to have had a few extra phonemes which are not explicit in the writing. Some final consonants seem to drop (kala ‘‘mighty’’ may be /kalag/). It was suggested that Sumerian may have had lexical tones like Chinese, which allegedly would explain the high number of homophonic terms. However, other factors can explain this.

Sumerian is an agglutinative language, that is, a word consists of a sequence of distinct morphemes, and the lexeme to which the morphemes are attached cannot undergo sound changes or take infixes. Grammatical gender is based on an opposition between animate and inanimate nouns, but this only surfaces in the concord between pronouns and their antecedents. Grammatical number (plural versus singular) does not need to be marked in writing (lugal ‘‘king’’ or ‘‘kings’’), but can be made explicit through suffixes (/lugal-ene/ ‘‘kings’’) or reduplication (lugal-lugal ‘‘kings’’).

The noun has ten cases, which are marked by attaching suffixes to noun phrases. Noun phrases are conventionally called ‘‘nominal chains’’ by Sumerologists, because all the suffixes are heaped at the very end:

/dumu lugal kalam-ak-ak-ene-ra/

‘‘for the son of the kings of the nation’’ son-king-nation-GENmvE-GENmvE-PLURAL-DATrvE

Sumerian is an ergative language, meaning that the subject of an intransitive verb has the same marker as the object of a transitive verb. The subject of a transitive verb has a marker, called the ergative case, that is different from the subject of an intransitive verb. In English, it would be like saying ‘‘him sleeps’’ and ‘‘me sleep,’’ but ‘‘I saw him’’ and ‘‘he saw me.’’ Sumerian has /-e/ as the ergative suffix, and /-0/ as the marker of absolutive case:

Lugal-e c2 mu-un-du3 ‘‘the king built the temple’’ lugal i3-tu's ‘‘the king sat down.’’

In fact, Sumerian exhibits split ergativity. This means that the ergative alignment is followed only in the nominal system. Independent personal pronouns and other forms have an accusative alignment like English. The system of verbal agreement shows a similar split: the hamtu forms (perfective) are ergative, whereas the marii forms (imperfective) show an accusative pattern.

Verbal stems are divided into two major categories: hamtu (‘‘quick, sudden’’ or perfective) and marU (‘‘slow, fat’’ or imperfective). However, these two labels correspond to the understanding of the Sumerian verb by Akkadian-speaking scribes. It is likely that all verbs had two stems. Some verbs marked the imperfective stem with an affix /-e/; reduplication verbs marked it with partial reduplication, but used complete reduplication in perfective forms; other verbs may have used complementary verbs with completely different sounds. But maybe many verbs did not have two different stems, and the only way to distinguish perfective from imperfective in those verbs was through pronominal affixes.

Probably there were only four morphemes used as conjugation prefixes: /ba-/; / imma-/; /i-/; and /mu-/. The prefix /bi-/ would be a combination of the prefix / ba-/ and the locative-terminative infix, and /imma-/ a reduplication of /mu-/. All verbal forms seem to start with an obligatory prefix (/mu-/, /ba-/, or /i-/). The choice of prefix seems governed by focus: /mu-/ is focused for person but not for place, while /ba-/ is focused for place but not for person; and /i-/ is not focused.

The dimensional infixes mark case relations between the verb and noun phrases. The pronominal prefixes agree with the subject of transitive perfective forms and the subject of both transitive and intransitive imperfective forms.

The imperative reverses the order of verbal constituents: it begins with the stem, which is followed by all the prefixes, as in /sum-mu-a-b/ ‘‘give it to me.’’ The suffix /-ed/ can occur in non-finite and finite verbal forms, and can immediately follow the verbal stem and precede the pronominal suffix. Some consider the /-e-/ in /-ed/ the marker of the perfective. Likewise, the nominalizer suffix /-a/ can be attached to both non-finite and finite verbal forms, and can be followed by case endings and pronouns. When the nominalized verbal form agrees with a noun that has an antecedent in another sentence, it is the equivalent of an English relative clause:

/ensi lu e-ninnu in-du-a /

‘‘the ruler that built the Eninnu. . . ’’

The word order tends to be almost always Subject-Object-Verb in all sentences.

The eme-sal dialect

Sumerian is called /eme-gir/ (perhaps ‘‘native tongue’’) in native Sumerian sources. In some Mesopotamian scholarly texts, a few lexical items and grammatical forms are identified as /eme-sal/ (perhaps ‘‘fine language’’). It has been argued that this was a women’s language, especially because the sign involved can also be read as ‘‘woman.’’ Eme-sal is attested in compositions of very specific genres. But no text is entirely written in eme-sal, and there is no true consistency in its use. Eme-sal may have stemmed from an actual regional dialect or from the dialect of a certain group. However, the occurrence of eme-sal forms may be determined by the genre of the text, rather than by the gender of the fictional speaker or the performer.