CHAPTER ONE

THE SUMERIAN

Fiir Olympia, die mir seit 1954 aile meine Arbeiten geschrieben hat und die mir auf drei Kontinenten gefolgt ist.

1.1. GENERAL CHARACTERISTICS

Sumerian was called eme-ki--gi-ra "tongue of Kiengir ()" or eme-gi7(-r) in Sumerian proper and lifiin Sumeri(m) "tongue of Sumeru" in Akkadian. The replacing of [S) by [s] in most modern (but not Russian) has its analogue in the change of [S) for [s] in names of the Hebrew bible. Sumerian is characterized by the interaction of word base (nom­ inal, verbal, other) which may be invariable or to variation (e.g., change of , reduction, extension), and intricate system of prefixed and suffixed morphemes. The word base itself is impen­ etrable by other morphemes. Unlike Semitic, no infixes occur. Cf. ha--ab-sum- [ha-m+a-b-sum-e] "he should give it to " (WO 8, 173: 11 b2): precative-to-me-it-base give-ergative (3rd person sing. person class). The number of prefixed morphemes varies between zero and six for the verb, zero and one for the ; the number of suffixed morphemes between zero and three for the verb, zero and three for the noun. Words of considerable length may be built up that way, e.g., hu-mu-na--ib-gi4-gi4 "let him return it to him there" (6 syllables, not comparable, however, with Akkadian ittanablakkatUnikkunufim "they will, over and again, revolt against you", 10 syllables). In both strings of morphemes, prefixed or suffixed, the sequence of the individual elements is unchangeable. The morphemes are mostly monofunctional, as is the rule in agglutinating languages, and very rarely multifunctional as the morphemes of Semitic or Indo-European. Instead of , Sumerian distinguishes a "person" and a "non­ person" class. The case system includes an ergative, and the verbal 2 CHAPTER ONE inflection is characterized by ergativity (whether there was, or evolved, "" is still a matter of debate-see below 12.7.5). There is well developed, but far from perfect, concord between dimen­ sional of the noun and dimensional indicators prefixed to the verb, e.g., DN-ra mu-na(-n)-ru "to god NN, object, he­ built-to-him". As for number, singular and non-singular may be opposed; there are different ways to express plurality. The general word order of Sumerian is S - 0 - V, unless some part of speech is taken to the front for . The nominal or verbal base is the essential carrier of meaning, and only bases are listed as entries in the . Occasionally, meaning is modified by the occurrence of a "frozen" morpheme. Also, the composition of two (rarely more) nominal word bases may lead to a new meaning beyond the sum of the meanings of the individ­ ual parts of the compound, e.g., e-gal "house big" = "palace" (Akk. ekallu) or ma-tur "boat small", a special type of boat (Akk. maturru). In the light of these general aspects, Sumerian may be compared to such languages as Georgian, Basque, or Itelmen and many oth­ ers. Such comparison is, however, purely structural and of no con­ sequence for the question of the linguistic affiliation of Sumerian.

1.2. THE (HOPELESS) QUESTION OF THE LINGUISTIC AFFILIATION OF SUMERIAN

Scholars have wasted much effort looking for living cognates of ancient Sumerian, not realizing that the problem is practically insol­ uble for the following reasons: Sumerian must have separated from a hypothetical language fam­ ily of which it was part in the middle or late fourth millennium B.C. at the latest. We know next to nothing about the sound and struc­ ture of Sumerian before the middle of the third millennium. Thus there is a gap of at least two thousand years between that time and the oldest reconstructible form of any of the languages which have been compared to Sumerian (e.g., Turkish, Hungarian, Sino-Tibetan). Efforts to find cognates have been exclusively based on the sound of individual words. Yet according to W. Deeters (1963, 76) who dis­ cussed the problem of Basque-Caucasian affinities, any words in lan-