Sumerian

Overview

Sumerian is the earliest attested written language (late 4th millennium BC), spoken in southern Mesopotamia (present-day Iraq). It is a linguistic isolate: no demonstrated relationship with any other family (proposed connections with Dravidian, Uralic, Caucasian, etc. remain unproven). Typologically, it is an agglutinative language with split ergativity, subject-object-verb order, and a complex verbal chain. It is written in cuneiform, a system it very likely originated and which was subsequently adapted to Akkadian and then to numerous other languages. Sumerian ceased to be a living spoken language around 2000-1700 BC, supplanted by Akkadian, but it survived for nearly two additional millennia as a scholarly, religious, and scholastic language, somewhat like Medieval Latin. Fully deciphered, it is today well documented through vast corpora (tens of thousands of tablets) and reference grammars, even though the reconstruction of its phonology and certain morphosyntactic details remain debated.

History of the people

The Sumerians inhabited lower Mesopotamia (Sumer), between the Tigris and the Euphrates, in present-day southern Iraq. They are credited with some of the world's first cities (Uruk, Ur, Lagash, Nippur, Kish, Eridu), the invention of writing (late 4th millennium), the city-state as a political form, innovations in irrigation, accounting, law, mathematics (the sexagesimal system), and a rich literature (myths, hymns, lamentations, Sumerian narratives concerning Gilgamesh, literary disputations). The political history alternates between rival city-states (Early Dynastic periods), unification under the Empire of Akkad (Sargon, Semitic), a Sumerian renaissance under the Third Dynasty of Ur, and then absorption into the Akkadian-speaking Babylonian and Assyrian worlds. The geographic origin of the Sumerians (indigenous or migrants) and their exact relationship to neighboring Semitic populations remain debated. Their cultural and religious heritage (pantheon, myths, lexical lists) was massively taken up and transmitted by the Akkadians, Babylonians, and Assyrians.

History of the language

Several stages are distinguished: Archaic Sumerian (proto-cuneiform of Uruk, ~3300-2900); Old / Presargonic Sumerian (Early Dynastic period, texts from Lagash, ~2600-2350); Sumerian of the Akkad period (Semitic empire, ~2334-2154); Classical Sumerian of the Third Dynasty of Ur (Ur III, ~2112-2004), a documentary golden age with tens of thousands of administrative tablets; then Post-Classical Sumerian and especially "Post-Sumerian" from the Old Babylonian period onward (~2000-1600), when Sumerian was no longer a native language but a scholastic, literary, and cultic one. Extinction as a vernacular was gradual: under the pressure of Akkadian (Semitic), spoken Sumerian probably disappeared between ~2000 and ~1700 BC, the exact date being uncertain and no doubt variable by region. It then survived as a liturgical and scholarly language, copied and studied in scribal schools until the beginning of our era; the last datable cuneiform tablets (astronomical almanacs from Uruk) date to the end of the 1st century AD (around 79/80). A particular literary register, eme-sal ("fine language"), appears in certain texts (cultic songs, speeches of female deities) alongside the main register eme-gir (eme-gi7(r)).

The script

Sumerian is written in cuneiform script (from Latin cuneus, "wedge"), traced by impressing a stylus with a triangular cross-section into moist clay. The system evolves from the proto-cuneiform pictograms of Uruk (~3300-3100 BC) toward abstract combinations of wedges and chevrons. It is a mixed logo-syllabic system: a single sign may function as a logogram (representing a whole word/morpheme) or as a phonetic sign (syllabic value). Orientation: originally the signs were arranged in columns from top to bottom; toward the middle of the 3rd millennium, reading shifts to run horizontally from left to right, the signs themselves having rotated about 90 degrees (the tablets presumably having been turned by the scribes). A single sign commonly has several values (polyphony), and several different signs may share the same value (homophony), resolved by modern convention through numerical indices (accents and subscript numerals: du, du2/dú, du3/dù, du6, du8...). Determinatives (unpronounced semantic classifiers) precede or follow the word: DINGIR (written d) for deities, ki for toponyms, gesz for wooden objects, uru for cities, ku6 (postposed) for fish, etc. The number of signs actually used in Sumerian is around 600, although the total cuneiform repertoire is larger.

Sign system

Three functions coexist for a sign. (1) LOGOGRAM (Sumerogram): the sign stands for a word; in scholarly transliteration it is written in capitals when the reading is uncertain (LUGAL "king", DUMU "son/child", E2 "house/temple", DINGIR "god"). (2) SYLLABIC/PHONETIC VALUE: the sign notes a sound of type V, CV, VC, or CVC (a, e, u; ba, du, gi; ab, ur; kur, and lugal as lu+gal). (3) DETERMINATIVE: an unread classifier. Modern scholarly convention uses: roman lowercase for the syllabic, CAPITALS for an unresolved logogram, superscript for determinatives, and numerical indices to distinguish homophones. There are no "Gardiner categories" in Sumerian (those are specific to Egyptian hieroglyphs); cuneiform is catalogued by sign lists (Borger MZL, Labat/Manuel d'épigraphie akkadienne, Mittermayer's aBZL for the Old Babylonian period) assigning to each sign a number and the range of its values.

Phonology

Sumerian phonology is reconstructed indirectly, chiefly via Akkadian syllabic transcriptions and lexical lists; it remains partially uncertain. Vowels generally accepted: a, e, i, u (the existence of a distinct /o/ is proposed by some scholars, including Attinger and Keetman, but is not consensual). The opposition of vocalic length is debated. Reconstructed consonants: stops with an opposition probably of tension/aspiration rather than strict voicing (b/p, d/t, g/k), the sibilants/affricates s, z, sz (/sh/), the nasals m, n, plus a velar nasal ĝ (sometimes written g with a tilde, or ng), the liquids l, r, the fricative h (sometimes written with a dot beneath), and a phoneme sometimes transcribed "dr" (an ambiguous final consonant of certain words). Important morphophonological alternations exist (loss of final consonants before a suffix, assimilation, partial vowel harmony in the verbal chain). A crucial point for a translation engine: the standardized transliteration is a graphic convention and does not faithfully reflect the actual pronunciation, which is largely unrecoverable.

Grammar — essentials

An agglutinative and ergative language. NOUN: no article; grammatical gender of the animate (person) vs. inanimate (non-person) type. Cases marked by suffixes stacked at the end of the phrase: absolutive -0, ergative -e, genitive -a(k), dative (person) -ra, directive/locative-terminative -e, locative -a, comitative -da, ablative/terminative -ta and -sze3, equative -gin7. The genitive follows the construction "possessed + possessor-ak" (e2 lugal-ak = "house king-of" = the king's house), with the -ak often reduced or phonetically masked. Plural of animates by -(e)ne or by reduplication of the logogram (dingir-dingir, lugal-lugal-ene). VERB: the heart of the language, a chain of morphemes around a root. Schematic order: (modal prefix) + (conjugation prefix mu-/i-/ba-/bi-/im-/al-) + (dimensional prefixes echoing the cases: dative, comitative, ablative, locative...) + (agreement marker) + ROOT + (person/aspect suffixes). Two aspects: hamtu (perfective, punctual/completed action) and marû (imperfective, durative/incomplete), often expressed by distinct roots or affixes and by reduplication. Agreement follows an ergative pattern: in the hamtu, the transitive agent is indexed by suffixes, the intransitive patient/subject by prefixes; in the marû, the pattern partially inverts (split ergativity). Negation nu-/la-, modals ha-/he- (precative/optative), ga- (cohortative), u- (prospective). Word order is SOV, with the verb strictly final.

Reading tips

A practical method for reading a Sumerian passage: (1) Break the line down into individual signs using a sign list (MZL/aBZL). (2) Identify the determinatives (d, ki, gesz, uru, etc.): they orient the meaning but are not pronounced and must not be translated as words. (3) For each remaining sign, choose between a logographic and a syllabic reading according to context, period, and attested collocations; in cases of homophony, the numerical index is merely a modern label, not a sound. (4) Reconstruct the noun phrases by locating the case suffixes stacked at the end (mindful of the genitive -ak, often reduced/hidden, and of the possessed+possessor chaining). (5) Analyze the final verb: locate the root, the aspect (hamtu vs. marû), the conjugation and dimensional prefixes, and the ergative agreement markers. (6) Systematically compare with existing editions (ETCSL, ORACC, ePSD2) and with the Sumero-Akkadian bilinguals that serve as an internal "Rosetta Stone". Always flag segments whose reading or analysis remains uncertain rather than smoothing over a hypothetical translation.

Pitfalls

Frequent pitfalls: (a) confusing transliteration with pronunciation — the values (with accents/indices) are modern conventions, not the actual phonetics. (b) Translating determinatives as words. (c) Missing the genitive -ak, often phonetically reduced or "hidden" (e2-lugal may hide e2 lugal-ak). (d) Mis-segmenting the verbal chain and assigning a syllabic value where a grammatical prefix is required (mu-, ba-, bi2-, im- are not full words). (e) Ignoring split ergativity: the agent and the patient are not indexed in the same way depending on aspect. (f) Taking homophony indices (du3, du6...) for phonetic information. (g) Forgetting the eme-sal register in certain cultic texts. (h) Over-interpreting unproven genetic connections (Dravidian, Uralic...): Sumerian is an isolate.

Decipherment history

Sumerian was deciphered indirectly, through Akkadian. The decipherment of Akkadian cuneiform took place in the mid-19th century thanks to Edward Hincks, Henry Creswicke Rawlinson, and Jules Oppert, drawing on the trilingual inscriptions (Old Persian / Elamite / Akkadian), notably that of Behistun copied by Rawlinson. As early as the late 1840s and during the 1850s, Hincks established that the signs have multiple values (logographic and syllabic) and sensed that Akkadian was not the original language of cuneiform. Scholars discovered bilingual lexical lists and school texts contrasting a non-Semitic language with Akkadian. In 1869, Jules Oppert proposed the name "Sumerian", drawn from the royal title "king of Sumer and Akkad". A controversy (led notably by Joseph Halévy) for a time contested the very existence of the language, viewing it as merely a sacerdotal cryptography of Akkadian; this thesis was refuted by the accumulation of purely Sumerian texts. The modern grammars were then built up from Poebel (1923) to Thomsen (1984), Edzard (2003), and Jagersma (2010).

Common formulae & expressions

lugal

king (lu2 'man' + gal 'great')

dumu ... -ra

'son of ...', frequent in filiations and dedications

e2

house / temple (E2)

dingir / d-

'god'; unpronounced divine determinative before theonyms

nam-lugal

kingship (nam- = abstract prefix, forms abstract nouns)

lugal kalag-ga

'mighty king', a recurring royal epithet

... mu-na-du3

'he built (it) (for the deity)', building-dedication formula

a mu-na-ru

'he dedicated/consecrated (it) (for her/him)', votive formula

nin

'lady, mistress', a female divine or royal title

KI.EN.GI(-r)

Sumer (the land's autonym); the ki-en-gi region of the south

Corpora & reference editions

• ETCSL - Electronic Text Corpus of Sumerian Literature (Oxford, edited and translated Sumerian literature)
• CDLI - Cuneiform Digital Library Initiative (catalogue and images of cuneiform tablets)
• ePSD2 - electronic Pennsylvania Sumerian Dictionary (University of Pennsylvania)
• ORACC - Open Richly Annotated Cuneiform Corpus (annotated editorial projects)
• ETCSRI - Electronic Text Corpus of Sumerian Royal Inscriptions (an ORACC component)
• RIME - Royal Inscriptions of Mesopotamia, Early Periods (printed editions of royal inscriptions)

Bibliographic references

• Marie-Louise Thomsen, The Sumerian Language: An Introduction to its History and Grammatical Structure, 1984 — Mesopotamia: Copenhagen Studies in Assyriology 10, Akademisk Forlag. Long the standard reference grammar; reprints (2001) with updated bibliography. Verified.
• Dietz Otto Edzard, Sumerian Grammar, 2003 — Handbook of Oriental Studies, Section 1, vol. 71, Brill (Leiden). A major synthetic grammar. Verified.
• Bram Jagersma, A Descriptive Grammar of Sumerian, 2010 — Doctoral dissertation, University of Leiden (~747 pp.); the most complete and rigorous description to date. Verified (available in open access, Leiden repository).
• Daniel A. Foxvog, Introduction to Sumerian Grammar (Cuneiform Digital Library Preprints 2), rev. 2016 — Open-access pedagogical manual (CDLI/CDLP), widely used for teaching. Verified.
• John L. Hayes, A Manual of Sumerian Grammar and Texts, 2nd ed. 2000 — Undena Publications (Aids and Research Tools in Ancient Near Eastern Studies 5). A learning manual with annotated texts (26 lessons). Verified.
• Rykle Borger, Mesopotamisches Zeichenlexikon (MZL), 2004 — Alter Orient und Altes Testament 305, Ugarit-Verlag (Münster). The reference cuneiform sign list (numbering, values). Date corrected: 1st ed. 2004 (2nd revised ed. 2010), not 2003.
• Arno Poebel, Grundzüge der sumerischen Grammatik, 1923 — Selbstverlag / Winterberg, Rostock. The first systematic modern grammar of Sumerian (of historical value). Verified.