The language of the ancient Assyrians and Babylonians of Mesopotamia, Akkadian, subsumes both Assyrian and Babylonian dialects within it. The earliest attested Semitic language, Akkadian comprises the more important member of the eastern branch of the Semitic family (the other member being Eblaite). The name Akkadian (akkadû) derives from Akkad(e), the capital city built by king Sargon in about 2300 BCE. It is not known when speakers of Akkadian or its linguistic predecessor(s) first arrived in Mesopotamia, nor is it known when the last speakers of the language died out; the first written attestations date to the 26th century BCE (in Semitic names; connected texts somewhat later) and the latest to the first century BCE. Several hundred thousand texts have been discovered in excavations, many of which remain unpublished.
Dialects and Genres.
Old Akkadian refers to texts from the earliest attestation of Akkadian down to about the beginning of the second millennium, thus including documents dating to the reigns of Sargon and Naram-Sin of Akkad and to the Ur III period; although the term is used collectively for all such texts, dialectal distinctions, both geographical and chronological, are evident within this corpus. Old Akkadian is attested in letters, legal texts, economic dockets, royal inscriptions, and a few literary texts (such as a love incantation). To the middle of the Old Akkadian period as well date the thousands of texts found recently at Ebla, in Syria, some of which are also partly written in a Semitic language, referred to as Eblaite or Eblaic. Formerly thought to be an early representative of West Semitic, Eblaite is now considered to be a close linguistic relative of Akkadian (or even, though this is less likely, a dialect of Akkadian). [See Ebla Texts.]
In the second and first millennia, two major geographical dialects are attested, Assyrian in northern Mesopotamia and Babylonian in the south. Linguistically these are distinguished by a number of phonological differences, by minor morphological variations, and by certain lexical items. The Assyrian and Babylonian scripts also developed somewhat independently of one another. As is true in all languages, numerous linguistic changes occurred over the many centuries in which Assyrian and Babylonian were spoken, and these appeared (eventually) in the texts. Although these developments arose continuously, so that there are no neat chronological divisions, scholars refer for the sake of convenience to the following subphases, which correspond roughly with political periods:
|Old Assyrian||2000–1500||Old Babylonian|
|Middle Assyrian||1500–1000||Middle Babylonian|
|600 BCE–100 CE||Late Babylonian|
Old Assyrian is known from approximately fifteen thousand letters and legal and economic documents dating from the mid-twentieth to the mid-eighteenth centuries BCE, most of which have been found in Cappodocia (eastern Turkey) at the site of Kaneš (modern Kültepe), although other sites in Anatolia and Assyria have also produced a few similar texts. The majority of these documents concern the business activities of Assyrian merchant houses and their trade with outposts in Anatolia. (The majority of these texts are unpublished.) Also attested are a number of royal inscriptions of rulers of the city of Aššur, commemorative and dedicatory in nature, and a few magical texts. [See Kültepe Texts.]
Middle Assyrian is sparsely attested, although it is known from a variety of genres, including letters, legal texts, economic texts, and inscriptions of the early kings of the nation and nascent empire of Assyria, and a set of harem decrees. Perhaps most famous, however, are the fourteen tablets containing the Middle Assyrian laws, discovered in the city of Aššur.
Neo-Assyrian is the spoken language of the period of the Assyrian Empire in the first millennium until its demise late in the seventh century BCE. There are a great many letters and administrative texts, of both the court and private individuals. Many royal inscriptions and scholarly writings are also attested; as in all periods, literary texts exhibit considerable linguistic influence of the more prestigious Bablyonian dialect.
Old Babylonian is the language of several tens of thousands of texts, dating to the first half of the second millennium, from the first dynasty of Babylon, from the Isin and Larsa dynasties, and from other southern Mesopotamian sites, as well as from sites outside Mesopotamia, such as Mari and Susa. Not surprisingly, texts from such a range of sites exhibit numerous minor dialectal differences. The best-studied form of Old Babylonian is that of the court of King Hammurabi of Babylon. An extremely diverse variety of genres has been preserved; in addition to thousands of letters, contracts, economic texts, and royal inscriptions, these include the famous “code” of laws of Hammurabi (the longest single Old Babylonian document), and many kinds of scholarly and school texts such as medical, mathematical, and grammatical texts, encyclopedic lists of words (lexical texts), and omens. There are also hymns, prayers, and epic and mythological works of literature, of which the best known are Gilgamesh, Atraḫasis (the flood story), and Anzû (the theft of the “tablets of destiny”); these latter texts, as well as royal inscriptions, are often written in a high literary language full of archaic forms and archaisms.
In later times, Old Babylonian was regarded as the classical period of Akkadian language and literature, and scribes in both Babylonia and Assyria attempted to duplicate it in a purely literary (i.e., unspoken) language called Standard Babylonian (though usually with mixed results, as their own linguistic forms frequently intervened). This is the dialect in which such important works as Enuma elish and the later version of Gilgamesh are written—indeed, all of the literary texts of the first millennium, as well as many royal inscriptions.
Middle Babylonian is the language of texts from the Kassite period. Like Middle Assyrian, it is less well represented than the dialects that precede and follow it. It is known from letters, legal texts, economic texts, a few royal inscriptions, and inscribed boundary stones (kudurrus).
Already in the Middle Bronze period, but especially in the Late Bronze, Akkadian, particularly Babylonian, was used as an international lingua franca; Akkadian texts have been found at a great many sites outside Mesopotamia, including Alalakh, Emar, and Ugarit in Syria, Ḫattuša in Anatolia (modern Boğazköy, the capital of the Hittite land), and el-Amarna in Egypt, to name a few. The language of these texts, which was written by non-native speakers, is usually termed Peripheral Akkadian; the texts vary considerably in their fidelity to normative grammar and frequently betray the influence of the scribes' own languages. [See Emar Texts; Alalakh Texts; Ugarit Inscriptions; Amarna Tablets.]
Neo-Babylonian refers to the spoken, nonliterary language of southern Mesopotamia until the end of the Assyrian Empire, after which the term Late Babylonian is used for the final period of texts written in Akkadian. These dialects comprise large numbers of letters and administrative documents. For literary and monumental texts, Standard Babylonian (see above) was employed.
Akkadian was written in cuneiform, a system of wedge-shaped characters usually pressed into moist clay tablets with a reed stylus. (Other media for inscriptions were wax, metal, and, especially for monumental texts, stone.) The system was borrowed from Sumerian, the other language of Mesopotamia and the language for which it was devised. The writing was at first pictographic, with fairly representational drawings; early on, however, it became more and more stylized until the pictographs ultimately became unrecognizable. As the Old Babylonian (code of Hammurabi) forms in figure 1 show, individual signs may consist of one wedge (such as AŠ and U), of a few wedges (such as BAD and MAŠ), or of many wedges (such as IN and IG). As already noted above, forms of signs continued to change throughout the history of Akkadian, and Assyrian and Babylonian forms differed from one another: the histories of a representative sample of signs appear in figure 2. There are three types of signs:
- 1. Phonetic signs. Syllables or parts of syllables to be pronounced are indicated by phonetic signs that may represent a simple vowel (such as “a”), a consonant plus a vowel (“ba”), a vowel plus a consonant (“ab”), or a consonant-vowel-consonant sequence (“bad”). Individual consonants cannot be written.
- 2. Logograms. These may be used to represent whole words.
- 3. Determinatives. A few signs that serve as classifiers indicating semantic ranges are known as determinatives. One determinative, for example, may precede words for items made of wood, while another precedes names of gods.
The three types of signs are formally indistinguishable and, in fact, some signs are used in all three ways; an example is the sign KI, which, in any given context may, among other possibilities, represent the (part-)syllable “ki,” denote the word for “earth,” or indicate that the word preceding it is a place name. To illustrate the writing system, the beginning of Law 8 of Hammurabi's code is reproduced in figure 3 with a standard Assyriological transliteration, in which phonetic values appear in italics, logograms in roman capital letters (with their Sumerian pronunciation), and a determinative in superscript letters. As the sample in figure 3 indicates, cuneiform is written from left to right.
Two additional complicating aspects of the writing system are homophony and polyphony. Homophony describes the fact that several discrete signs may have the same phonological value, such as /sa/; these are distinguished in transliteration by diacritical marks as follows: the most common sign found for a particular value receives no special mark, the second an acute accent over the vowel, and the third a grave accent, whereas the fourth and following receive subscript numbers (thus, sa, sá, sà, sa4, etc.; called sa-one, sa-two, sa-three, sa-four, etc.). Many signs are polyphonous—that is, they may represent more than one phonetic value; for example, the sign used to write the value ud may also be used to write the values tam and pir.
The original number of twenty-nine Proto-Semitic consonants was considerably reduced in Akkadian, probably through constant contact with Sumerian, to only twenty, as shown in table 1. Naturally the precise pronunciation of these phonemes is unknown; modern scholars base their pronunciation on parallels with living Semitic languages. The phoneme š is pronounced as “sh” and ḫ as “ch” in the German ach or Scottish loch; the so-called “emphatic” consonants q, ṣ, ṭ were either pharyn-gealized (as in Arabic) or, more likely, glottalized (as in Amharic and modern South Arabian languages). It is likely that Old-Akkadian preserved more of the original Semitic consonants, especially the interdental ⋆θ and some of the so-called guttural phonemes, such as ḥ and c; the writing system was poorly equipped to expressed such sounds, however.
Early Semitic had three short vowels, (a, i, u) and three corresponding long vowels (ā, ī, ū); most dialects of Akkadian preserve these and add a fourth vowel quality, e and long ē. Most Assyriologists distinguish two types of long vowels, marking those inherited as long vowels with a macron (ā, ē, ī, ū) and those that result from the contraction of two adjacent vowels with a circumflex (â, ê, î, û). It has recently been proposed that Akkadian also had a fifth phonemic vowel quality, o (Westenholz, 1991).
Table 1. Akkadian Consonants
Akkadian phonology is marked by a number of sound changes vis-à-vis common Semitic, in addition to the loss of consonants and the addition of the vowel e already described. All dialects exhibit a phenomenon called vowel syncope, in which the second of two short vowels in open syllables is deleted, as in damqum, “good (masculine),” from da/mi/qum (cf. the feminine counterpart, damiqtum) or in napšātum, from na/pi/šā/tum, the plural of napištum, “life.” Also common to Akkadian is Geer's law, according to which two emphatic consonants may not cooccur in one word: thus, for example, Semitic qaṣārum became Akkadian kaṣārum, “to bind”; Semitic ṣabātum became Akkadian ṣabātum, “to seize”; and Semitic qaṭārum became qatārum, “to billow (of smoke).” In Barth's law, the initial m of noun prefixes became n in words in which a labial consonant (b, m, p) followed: narkabtum, “chariot;” narāmum, “beloved”; našpakum, “storage area.” Both Assyrian and Babylonian forms of Akkadian also exhibit distinctive types of vowel harmony: in Assyrian, short a in an unaccented open syllable is replaced by a short-vowel that mimics the vowel of the following syllable, as in iṣbat, “he seized,” but taṣbitī (vs. Babylonian taṣbatī), “you (feminine) seized,” and iṣbutū (vs. Babylonian iṣbatū), “they (masculine) seized”; in Babylonian, the presence of an e vowel tends to cause an a vowel in the same word to become e, as in bēlētum, “ladies” (from bēlātum; cf. šarrātum, “queens”).
The basic morphological categories of pronouns, nouns, and verbs generally exhibit two genders (masculine and feminine) and three numbers (singular, dual, and plural), although the use of the dual is considerably circumscribed after the Old Akkadian period.
As in other Semitic languages, there is a set of independent personal pronouns and there are also sets of pronominal suffixes that indicate possession (when used on nouns and prepositions) or direct or indirect objects (on verbs); the Old Babylonian and Old Assyrian forms are shown in table 2. Indirect object forms are also attested in Eblaite but are not found in other branches of Semitic; they probably represent an innovation within East Semitic (Akkadian and Eblaite).
In early dialects of Akkadian (Old Akkadian, Old Babylonian, Old Assyrian), as in other morphologically conservative Semitic languages, nouns and adjectives exhibit three cases, marked with distinct vowels: nominative, for the subject of a clause (marked with u; in the dual with ā); genitive, indicating possession and also used after prepositions (marked with i); and accusative, for the direct object and various adverbial uses (marked with a). In dual and plural forms the genitive and accusative merge into a single form, termed the oblique (marked with i). In addition to endings that indicate the case, the regular, or free, forms of nouns also bear a final -m in the singular and the feminine plural (but not in the masculine plural) and a final -n in the dual. Feminine nouns and feminine forms of adjectives are usually marked with either -t- or -at- after the base and before the case vowel: bēlum, “lord”; bēltum, “lady”; šarrum, “king”; šarratum, “queen”; ṭābum (masc.) and ṭābtum (fem.), “pleasant”; dannum (masc.) and dannatum (fem.), “strong.” Besides the regular, or free, form of the noun there is also the bound form (or construct form), used when a noun syntactically governs another noun or a pronominal suffix; the bound form generally lacks both the final -m/-n and the case vowel: šarratum, “queen,” and mātum, “land” (both nominative case); but šarrat mātim, “queen of the land” (with “land” appropriately in the genitive case), and šarratni, “our queen.” Adjectives are for the most part declined like substantives, although they exhibit a distinctive ending in masculine-plural forms. Table 3 shows the regular (free form) declensions of the substantives šarrum, “king,” and šarratum, “queen,” and the adjective ṭābum, “pleasant,” in Old Babylonian. Final -m/-n were lost in the late stages of both Old Assyrian and Old Babylonian, so that in later phases of Akkadian the singular šarru/šarri/šarra is found, for example. Still later in the history of Akkadian, the case system itself became defunct as final vowels were lost. (In the literary dialect called Standard Babylonian, the nominative and the accusative are both usually written šarru; one form, the earlier oblique, is usually written for plurals regardless of case.)
Table 2. Old Babylonian (OB) and Old Assyrian (OA) Pronoun Forms (1, 2, 3=first, second, third person; m=masculine; f=feminine; c=common gender; s=singular; p=plural; dual forms are omitted).
|Independent||Possessive||Direct Object||Indirect Object|
In addition to their free and bound forms, substantives and adjectives also enter into a syntactic construction called the predicative form (also called the stative, with adjectives), in which the base of a substantive or adjective is followed by a special set of pronominal suffixes; the two elements constitute a verbless clause, with the substantive or adjective as predicate and the pronominal suffix as subject: šarrāku, “I (-āku) am king (šarr-)”; ṭābat, “it (feminine; -at) is pleasant (ṭāb-)”; rabiānu, “we (-ānu) are great (rabi-).” That this is an ancient construction is suggested by the appearance of a similar construction in Old Egyptian; in other Semitic languages, however, it fell out of use except in restricted circumstances.
Table 3. Old Babylonian Noun Declension
Akkadian verbal morphology is complex, as in other Semitic languages. The verbal “root” usually consists of three consonants, as in r-k-b, “ride.” There are four main finite forms (or tenses, or aspects): (1) an imperfective form denoting the present, the future, the habitual or circumstantial past, and a variety of modal functions, that is characterized by doubling in the second of the three root consonants, as in arakkab, “I ride/am riding/will ride/used to ride (etc.),” in which the prefix a- denotes “I” (cf. nirakkab, “we ride,” with ni- for “we”); (2) a punctive form called the preterite, for the past, as in arkab, “I rode, did ride”; (3) the perfect, a form in which -t- is infixed after the first root consonant, as in artakab, “I have ridden”; and (4) the imperative rikab, “ride!”. The imperfective form arakkab and the preterite arkab were undoubtedly inherited essentially unchanged from common Semitic. Akkadian shares the arakkab form with Ethiopian and modern South Arabian Semitic languages (compare classical Ethiopic 'ǝrakkǝb, “I attain”), but the form was lost in the Central Semitic languages (such as Arabic, Hebrew, and Aramaic). The preterite form arkab is found in most other Semitic languages in restricted uses, but has been replaced as the main past-tense form by a conjugation in which the subject pronouns are indicated by suffixes rather than prefixes (derived ultimately from the predicative form of Akkadian, discussed above), as in Arabic rakibtu or Hebrew rͻḵaḇti, “I rode.” The form called the perfect is not found in other Semitic languages and is generally considered an Akkadian innovation. Finite verbs may be marked with a morpheme called the ventive, which indicates motion or activity in the direction or proximity of the speaker: illik, “he went,” but illikam, “he came”; tuṣi, “you went forth,” but tuṣiam, “you came forth.” The ventive is unknown in other Semitic languages and its presence in Akkadian may reflect the influence of Sumerian.
In addition to the finite forms there is a verbal noun or infinitive, rakābum, “to ride”; an active participle rākibum, “riding, rider”; and a verbal adjective rakbum, “ridden.” These forms have analogues or parallels in the other Semitic languages.
In common with other Semitic tongues, Akkadian exhibits the modification of verbal roots, by means of prefixes or other changes, to produce a series of augmented stems with meanings derived more-or-less predictably from that of the basic stem; for example, from the basic form parāsum, “to separate,” are derived a passive verb with a prefixed n, naprusum, “to be separated”; a causative verb with a prefixed š, šuprusum, “to cause to separate”; and an iterative verb with an infixed syllable, pitarrusum, “to separate repeatedly.” The last of these is an Akkadian innovation, unknown elsewhere in Semitic.
In many features of syntax Akkadian resembles other Semitic languages closely. For example, attributive adjectives follow their head nouns and must agree with them in gender, number, and case; and verbs must agree with their subjects in gender and number. Akkadian makes frequent use of a genitive chain involving juxtaposition of two nouns to indicate governance of one by the other, the lead or governing noun appearing in the bound (construct) form, as in šarrat mātim, “queen of the land” (see above); but a pronominal form, ša, also evolved for the expression of the same relationship, as in šarratum ša mātim, “queen of the land” (a similar development occurred in many other Semitic languages).
In other syntactic features Akkadian deviates markedly from its Semitic relatives. In all prose dialects and texts, the verb occupies the final position in the sentence, a feature undoubtedly borrowed from Sumerian: šarrum ekallam ina ālim ibni, “king-palace-in-city-built” (i.e., “the king built a palace in the city”). Already noted above was the use of the predicative construction in verbless sentences in which the predicate is an adjective or a noun: Ḫammurapi rabi, “Hammurabi is great”; šarrāku, “I am king.” There are two main clause coordinators: u (Semitic wa-) connects nouns in phrases (šarrum u šarratum, “king and queen”) and clauses that are not semantically connected (abni u arkab, “I built and I [also] rode”), whereas the enclitic particle -ma connects clauses in which the first is logically subordinate to the second (allikam-ma abni, “I came and then built” or “having come, I built”). The use of -ma is so common that the wide range of subordinating conjunctions is used only infrequently (subordinate clauses, like other parts of speech, usually precede the main-clause verb). The syntax of the infinitive is complex; for example, the phrase “on reaching the town” may appear in the following three forms: ina kašād ālim (lit., “in the reaching of the town”), ina ālim kašādim (“in town-reaching”), ālam ina kašādim (“the town in reaching”).
The majority of the lexicon in all Akkadian dialects is inherited from common Semitic. In the earliest period, however, there are already many loanwords from Sumerian, nearly all of which are nouns, covering a wide semantic range. Beginning in the Neo-Assyrian and Neo-Babylonian periods, many Aramaic words begin to enter the language as well.
[See also Akkadians.]
- Aro, Jussi. Studien zur mittelbabylonischen Grammatik. Studia Orientalia, 20. Helsinki, 1955. The most recent monograph-length treatment of Middle Babylonian grammar.
- Borger, Rykle. Handbuch der Keilschriftliteratur. 3 vols. Berlin, 1967–1975. Indispensable guide to the publication of Akkadian and Sumerian texts and all matters pertaining to the reading of cuneiform texts, updated in the periodical Archiv für Orientforschung.
- Borger, Rykle. Assyrisch-babylonische Zeichenliste. 4th ed. Kevelaer, 1988. One of two sign lists that give both phonological and logographic values and histories of individual signs (the other is Labat, 1988; see below).
- Ebeling, Erich, et al., eds. Reallexikon der Assyriologie und Vorderasiatischen Archäologie. Berlin, 1932– . Encyclopedic reference work of all things Assyriological.
- Gelb, Ignace J., et al. The Assyrian Dictionary of the Oriental Institute of the University of Chicago. Chicago, 1956– . Encyclopedic reference work, devoting one volume to each “letter” of the Akkadian phonological system; sixteen of twenty-one volumes have appeared to date.
- Gelb, Ignace J. Old Akkadian Writing and Grammar. Materials for the Assyrian Dictionary, 2. 2d ed. Chicago, 1961. Description of the earliest phase of Akkadian.
- Groneberg, Brigitte. Syntax, Morphologie und Stil der jungbabylonischen “hymnischen” Literatur. 2 vols. Wiesbaden, 1987. Up-to-date description of the important literary dialect known in English as Standard Babylonian.
- Hecker, Karl. Grammatik der Kültepe-Texte. Analecta Orientalia, 44. Rome, 1968. Comprehensive description of the Old Assyrian dialect.
- Labat, René. Manuel d'épigraphie akkadienne. 6th ed. by Florence Malbran-Labat. Paris, 1988. One of two sign lists that give both phonological and logographic values and histories of individual signs (the other is Borger, 1988; see above).
- Mayer, Walter. Untersuchungen zur Grammatik des Mittelassyrischen. Kevelaer, 1971. The most thorough treatment of the Middle Assyrian dialect.
- Reiner, Erica. A Linguistic Analysis of Akkadian. The Hague, 1966. The most competent linguistic presentation of Akkadian yet published.
- Reiner, Erica. “Akkadian.” In Current Trends in Linguistics, vol. 6, Linguistics in South West Asia and North Africa, edited by Thomas A. Sebeok, pp. 274–303. The Hague, 1970. History of Akkadian grammatical and lexical research, and a review of topics requiring new or additional research.
- Reiner, Erica. “Die akkadische Literatur.” In Altorientalische Literaturen: Neues Handbuch der Literaturwissenschaft, edited by Wolfgang Röllig, pp. 151–210. Wiesbaden, 1978. Comprehensive review of the types of Akkadian literature.
- Röllig, Wolfgang. “Überblick über die akkadische Literatur.” In Reallexikon der Assyriologie und Vorderasiatischen Archäologie, vol. 7, pp. 48–66. Berlin, 1987. Survey of Akkadian literary works.
- Soden, Wolfram von. Akkadisches Handwörterbuch. 3 vols. Wiesbaden, 1965–1981. The only complete modern dictionary of Akkadian.
- Soden, Wolfram von. Grundriss der akkadischen Grammatik samt Ergänzungsheft. Rome, 1969. Standard reference grammar of Akkadian.
- Soden, Wolfram von, and Wolfgang Röllig. Das akkadische Syllabar. 4th ed. Rome, 1991. Authoritative reference for phonetic sign values in Akkadian.
- Westenholz, Aage. “The Phoneme /o/ in Akkadian.” Zeitschrift für Assyriologie 81 (1991): 10–19.