1911 Encyclopædia Britannica/Sumer and Sumerian

SUMER and SUMERIAN. The Babylonian name Shumer was used in the cuneiform inscriptions together with Akkad, viz. mat Shumeri u Akkadī, “land of S. and A.,” to denote Babylonia in general (See Akkad). In the non-Semitic ideographic documents the equivalent for Shumer is Kēngi, which seems to be a combination of kēn, “land” + gi, “reed,” i.e. “land of reeds,” and appropriate designation for Babylonia, which is essentially a district of reedy marshes formed by the Tigris and Euphrates. It was formerly thought that Shumer was employed especially to denote the south of Babylonia, while Akkad was used only of the north, but this view is no longer regarded as tenable. It is more probable that the expression Shumer designated the whole of Babylonia in much the same manner as did Akkad, and that the two words “Shumer and Akkad” were used together as a comprehensive term. That Shumer actually did mean all Babylonia appears evident from the biblical use of Shinar = Shumer to describe the district which contained the four chief Babylonian cities, viz. Babel, Erech, Accad and Calneh (Gen. x. 10), which, according to the Old Testament account, constituted the beginnings of Nimrod's kingdom. The identity of Shinar and Shumer is also demonstrated by the Septuagint rendering of Shinar in Isaiah xi. 11 by “Babylonia.” In short, there can be no doubt that the biblical name Shinar was practically equivalent to the mat Shumeri u Akkadî = non-Semitic, Kêngi-Uri of the Babylonian inscriptions. Furthermore, the fact that the Syriac Sen‛ar = Shinar was later used to denote the region about Bagdad (northern Babylonia) does not necessarily prove that Shinar-Shumer meant only northern Babylonia, because, when the term Sen‛ar was applied to the Bagdad district the great southern Babylonian civilization had long-been forgotten and “Babylonia” really meant only what we now know as northern Babylonia.

The actual meaning of the word Shumer is uncertain. Dr T. G. Pinches has pointed out[1] that Shumer may be a dialectic form of an as yet unestablished non-Semitic form, Shenger, just as the non-Semitic word dimmer, “god,” is equivalent to another form, dingir. Others have seen in the ancient Babylonian place-name Gir-su an inversion of Su-gir = Su-n͡gir, which has also been identified with Shumer. In this connexion Hommel's-theory[2] should be mentioned, that the word Shumer was a later palatalization of Ki-imgir, “land of Imgir” = Shi-imgir, subsequently Shingi with palatalized k = sh and elision of the final r. The form imgir (imgur), however, as a place-name for Babylonia is uncertain. All that can be said at present about this difficult etymology is that in the non-Semitic Babylonian the medial m represented quite evidently an indeterminate nasal which could also be indicated by the combination n͡g. Hence we find Shumer, probably pronounced Shuwer, with a sound similar to that heard to-day in the Scottish Gaelic word lamh, “hand”; viz. a sort of nasalized w. This gave rise to the later inaccurate forms: Greek, Senaar; Syriac, Sen’ar; and biblical Hebrew, Shinar = Shin͡gar.

The so-called “Sumerian problem,” which has perplexed Assyriologists for many years, may be briefly stated as follows. In a great number of Babylonian inscriptions an idiom has long been recognized which is clearly not ordinary Semitic in character. This non-Semitic system, which is found, in many instances, on alternate lines with a regular Semitic translation, in other cases in opposite columns to a Semitic rendering, and again without any Semitic equivalent at all, has been held by one school, founded and still vigorously defended by the distinguished French Assyriologist, Joseph Halévy, to be nothing more than a priestly system of cryptography based, of course, on the then current Semitic speech. This cryptography, according to some of the Halévyans, was read aloud in Semitic, but, according to other expositors, the system was read as an “ideophonic,” secret, and purely artificial language.

The opposing school (the Sumerists) insists that these non-Semitic documents were evidently in an agglutinative language, naturally not uninfluenced by Semitic elements, but none the less essentially non-Semitic in origin and fundamental character. Scholars of this opinion believe that this language, which has been arbitrarily called “Akkadian” in England and “Sumerian” on the European continent and in America, was primitively the speech of the pre-Semitic inhabitants of the Euphratean region who were conquered by the invading Semites. These invaders, according to this latter view, adopted the religion and culture of the conquered Sumerians; and, consequently, the Sumerian idiom at a comparatively early date began to be used exclusively in the Semitic temples as the written vehicles of religious thought in much the same way as was the medieval Latin of the Roman Church. The solution of this problem is of vital importance in connexion with the early history of man's development in the Babylonian region.

The study of the Sumerian vocabulary falls logically into three divisions. These are (1) the origin of the cuneiform signs, (2) the etymology of the phonetic values, and (3) the elucidation of the many and varied primitive sign-meanings.

Previous to Professor Friedrich Delitzsch's masterly work on the origin of the most ancient Babylonian system of writing,[3] no one had correctly understood the facts regarding the beginnings of the cuneiform system, which is now generally recognized as having been originally a pure picture writing which later developed into a conventionalized ideographic and syllabic sign-list. In order to comprehend the mysteries of the Sumerian problem a thorough examination of the beginning of every one of these signs is, of course, imperative, but it is equally necessary that every phonetic Sumerian value and word-combination be also studied, both in connexion with the equivalent signs and with other allied phonetic values. This etymological study of Sumerian is attended with incalculable difficulties, because nearly all the Sumerian texts which we possess are written in an idiom which is quite evidently under the influence of Semitic. With the exception of some very ancient texts, the Sumerian literature, consisting largely of religious material such as hymns and incantations, shows a number of Semitic loanwords and grammatical Semitisms, and in many cases, although not always, is quite patently a translation of Semitic ideas by Semitic priests into the formal religious Sumerian language. Professor Paul Haupt may be termed the father of Sumerian etymology, as he was really the first to place this study on a scientific basis in his Sumerian Family Laws and Akkadian and Sumerian Cuneiform Texts.[4] It is significant that all phonetic and grammatical work in Sumerian tends to confirm nearly every one of Haupt's views. Professors Peter Jensen and Zimmern have also done excellent work in the same field and, together with Haupt, have established the correct method of investigating the Sumerian vocables, which should be studied only in relation to the Sumerian literature. Sumerian words should by no means be compared with words in the idioms of more recent peoples, such as Turkish, in spite of many tempting resemblances.[5] Until further light has been thrown on the nature of Sumerian, this language should be regarded as standing quite alone, a prehistoric philological remnant, and its etymology should be studied only with reference to the Sumerian inscriptions themselves. On the other hand, grammatical and constructional examples may be cited from other more modern agglutinative idioms, in order to establish the truly linguistic character of the Sumerian peculiarities and to disprove the Halévyan contentions that Sumerian is really not a language at all.[6]

It is not surprising that Halévy's view as to the cryptographic nature of Sumerian should have arisen. In fact, the first impression given bl; the bewildering labyrinth of the Sumerian word-list is the conclusion that such a vocabulary could never have arisen in a regularly developed language. For example, anyone studying Brünnow's List[7] will find the same sign denoting pages of meanings, many of which have apparently no connexion with any other meaning belonging to the sign in question. A great multiplicity of meanings is also attributed, apparently quite arbitrarily, to the same sign, sound-value or word. In these instances, however, we can explain the difficulty away by applying that great fundamental principle followed by the Semitic priests and scribes who played with and on the Sumerian idiom, and in the course of many centuries turned what was originally an agglutinative language into what has almost justified Halévy and his followers in calling Sumerian a cryptography. This principle is that of popular etymology, i.e. of sound-association and idea-association which has brought together in the word-lists many apparently quite distinct meanings, probably primarily for purposes of mnemonic aid. The present writer in his Materials for a Sumerian Lexicon has mentioned this ruling phenomenon again and again. A very few examples, however, will suffice here. Thus the word ag = the sign RAM =râmu, “love” (proper meaning) is associated with ramâmu, “to roar,” for phonetic reasons only. The word a = the sign A = “water” (original meaning) can indicate anything whatever connected with the idea moisture. Thus, a = “water, moisture, weep, tears, inundate, irrigate,” &c. The word a can also mean “shining, glistening,” an idea evidently developed from the shining rippling of water. Note that in Turkish su means both “water” and “the lustre of a jewel,” while in English we speak of “gems of the first water.” The combination a-mâ-tu, literally “water enter ship, ” means abûbu, “deluge,” ordinarily, but in one passage a-mâ-tu is made the equivalent of shabûbu, “flame,” a pure pun on abûbu, “deluge.” Examples of this, the leading principle which was followed by the framers of the Sumerian system, might be cited almost ad infnitum.

Facts of this character taken by themselves would perhaps be sufficient to convince most philologists that in Sumerian we have an arbitrarily compounded cryptography just as Halévy believes, but these facts cannot be taken by themselves, as the evidences of the purely linguistic basis of Sumerian are stronger than these apparent proofs of its artificial character.

Briefly considered there are six most striking proofs that the Sumerian was based on a primitive agglutinative language. These may be tabulated concisely as follows:—

1. Sumerian presents a significant list of internal phonetic variations which would not have been possible in an arbitrarily invented language. Thus, taking the vowels alone; e = a by the principle of umlaut. Hence, we find the words ga and ge, a and e for the same idea respectively. The vowel i could become e as de = di, &c. Consonantal variation is most common. Thus, b = m, as barun = marun. Compare the modern Arabic pronunciation Maalbek for Baalbek. Perhaps the most interesting of these consonantal interchanges is that occurring between n and the sibilants sh and z; ner = sher; na = za, which by some scholars has been declared to be phonetically impossible, but its existence is well established between the modern Chinese colloquial idioms. For example, Pekingese zhen, Hakka nyin, Fuchow nöng, Ningpo zhing and nying, Wonchow zang and nang all = “man.” This demonstrates beyond a doubt the possibility of a strongly palatalized n becoming a palatal sibilant or vice versa, between which utterances there is but a very slight tongue movement.

The discussion of these phenomena brings us to another point which precludes the possibility of Sumerian having been merely an artificial system, and that is the undoubted existence in this language of at least two dialects, which have been named, following the inscriptions, the Eme-ku, “the noble or male speech,” and the Eme-sal, “the woman's language.” The existence and general phonetic character of the “woman's language” were first pointed out by Professor Paul Haupt, who cited, for example, the following very common interdialectic variations: Eme-ku gir = Eme-sal meri, “foot”; Eme-ku ner = Eme-sal sher, “ruler”; Eme-ku duga = Eme-sal zeba, “knee,” &c. Such phonetic and dialectic changes, so different from any of the Semitic linguistic phenomena, are all the more valuable because they are set before us only by means of Semitic equivalents. Certainly no cryptography based exclusively on Semitic could exhibit this sort of interchange.

It should be added here in passing that the geographical or tribal significance of these two Sumerian dialects has never been established. There can be no doubt that Eme-sal means “woman’s language,” and it was perhaps thus designated because it was a softer idiom phonetically than the other dialect. In it were written most of the penitential hymns, which were possibly thought to require a more euphonious idiom than, for example, hymns of praise. It is doubtful whether the Eme-sal was ever really a woman’s language similar in character to that of the Carib women of the Antilles, or that of the Eskimo women of Greenland. It is much more likely that the two dialects were thus designated because of their respectively harsh and soft phonetics.[8]

2. Sumerian has a system of vowel harmony strikingly like that seen in all modern agglutinative languages, and it has also vocalic dissimulation similar to that found in modern Finnish and Esthonian. Vocalic harmony is the internal bringing together of vowels of the same class for the sake of greater euphony, while vocalic dissimulation is the deliberate insertion of another class of vowels, in order to prevent the disagreeable monotony arising from too prolonged a vowel harmony. Thus, in Sumerian we find such forms as numunnib-bi, “he speaks not to him,” where the negative prefix nu and the verbal prefix mun are in harmony, but in dissimulation to the infix nib, “to him,” and to the root bi, “speak,” which are also in harmony. Compare also an-sud-dam, “like the heavens,” where the ending dam stands for a usual dim, being changed to a hard dam under the influence of the hard vowels in an-sud.

3. Sumerian has only post positions instead of prepositions, which occur exclusively in Semitic. In this point also Sumerian is in accord with all other agglutinative idioms. Note Sumerian e-da, “in the house” (e, “house,” + da, “in,” by dissimulation), and compare Turkish ev, “house,” de, “in,” and evde, “in the house.”

4. The method of word formation in Sumerian is entirely non-Semitic in character. For example, an in determinative vowel, a, e, i or u, may be prefixed to any root to form an abstract; thus, from me, “speak,” we get e-me, “speech ”; from ra, “to go,” we get a-ra, “the act of going, ” &c. In connexion with the very complicated Sumerian verbal system[9] it will be sufficient to note here the practice of infixing the verbal object which is, of course, absolutely alien to Semitic. This phenomenon appears also in Basque and in many North American languages.

5. Sumerian is quite devoid of grammatical gender. Semitic, on the other hand, has grammatical gender as one of its basic principles.

6. Furthermore, in a real cryptography or secret language, of which English has several, we find only phenomena based on the language from which the artificial idiom is derived. Thus, in the English “Backslang,” which is nothing more than ordinary English deliberately inverted, in the similar Arabic jargon used among school children in Syria and in the Spanish thieves' dialect, the principles of inversion and substitution play the chief part. Also in the curious tinker’s “Thary” spoken still on the English roads and lanes, we find merely an often inaccurately inverted Irish Gaelic. But in none of these nor in any other artificial jargons can any grammatical development be found other than that of the language on which they are based.

7. All this is to the point with regard to Sumerian, because these very principles of inversion and substitution have been cited as being the basis of many of the Sumerian combinations. Deliberate inversion certainly occurs in the Sumerian documents, and it is highly probable that this was a priestly mode of writing, but never of speaking; at any rate, not when the language was in common use. It is not necessary to imagine, however, that these devices originated with the Semitic priesthood. It is quite conceivable that the still earlier Sumerian priesthood invented the method of orthographic inversion, which after all is the very first device which suggests itself to the primitive mind when endeavouring to express itself in a manner out of the ordinary. For example, evident Sumerian inversions are Gibil, “the fire god,” for Bil-gi; ushar for Sem. sharru, “king,” &c.

It is, moreover, highly probable that Sumerian had primitively a system of voice-tones similar to that now extant in Chinese. Thus, we find Sumerian ab, “dwelling,” “sea”; ab, “road,” and -ab, a grammatical suffix, which words, with many others of a similar character, were perhaps originally uttered with different voice-tones. In Sumerian, the number of conjectural voice tones never exceeds the possible number eight.

It is also clear that Sumerian was actually read aloud, probably as a ritual language, until a very late period, because we have a number of pure Sumerian words reproduced in Greek transliteration; for example, Delephat = Dilbat, “the Venus-star”; Illinos = the god Illil = Bêl; aidô = itu, “month,” &c.

In view of the many evidences of the linguistic character of Sumerian as opposed to the one fact that the language had en grafted upon it a great number of evident Semitisms, the opinion of the present writer is that the Sumerian, as we have it, is fundamentally an agglutinative, almost poly synthetic, language, upon which a more or less deliberately constructed pot-pourri of Semitic inventions was superimposed in the course of many centuries of accretion under Semitic influences. This view stands as a connecting link between the extreme idea of the Halévyan school and the extreme idea of the opposing Sumerist school.

Literature.—Radau, Early Babylonian History; Lenormant, Études accadiennes, ii. 3, p. 70; Eberhardt Schrader. Keilinschriften u. das Alte Testament, ii. 118 sqq., Keilinschriften u. Geschichtsforschung, pp. 290, 533; Weissbach, Zur Lösung der sumerischen Frage; T. G. Pinches, “Language of the Early Inhabitants of Mesopotamia,” in Journ. Roy. Asiatic Soc. (1884), pp. 301 sqq.; “Sumerian or Cryptography,” ibid. (1900), pp. 75 sqq., 343, 344, 551, 552; article “Shinar” in Hastings’s Dict. Bible, iv. 503–505; Halévy, Journal asiatigue (1874), 3rd series, vol. iv. pp . 461 sqq.; Comptes rendus, 3rd series, vol. iv. p. 477; 3rd series, vol. iv. pp. 128, 130; Journal asiatique, 7th series, vol. viii. pp. 201 sqq.; Recherches critiques sur l’origine de la civilisation babylonienne (Paris, 1876); J. D. Prince, Journal of the American Oriental Society, xxv. 49-67; American Journal of Semitic Languages, xix. 203 sqq.; Materials for a Sumerian Lexicon, with grammatic introduction (Leipzig, 1905-1907). Compare also the material cited in the footnotes above, and note the correspondence between Brünnow and Halévy in the Revue sémitique (1906).  (J. D. Pr.) 

  1. Hastings's Dict. Bible, iv. 503.
  2. Ibid. i. 224b.
  3. Die Entstehung des ältesten Schriftsystems oder der Ursprung der Keilscfzfiftzeichen (Leipzig, 1897).
  4. Die sumefischen Familiengesetze (1879). Die akkadische Sprache (Berlin, 1883). Akkadische und sumerische Keilschriftlexte (Leipzig, 1881). See especially his Sumerian grammar in this latter work, pp. 133–147.
  5. Cf. A. H. Sayce's interesting article in Philological Society (1877–1878), pp. 1–20.
  6. Prince, Materials for a Sumerian Lexicon, pp. 18, 21.
  7. R. E. Brünnow, A Classified List of all Simple and Compound Ideographs (1889).
  8. Prince, Materials for a Sumerian Lexicon, p. 14.
  9. Ibid. pp. 20–34.