CUNEIFORM (from Lat. cuneus, a wedge), a form of writing, extensively used in the ancient world, especially by the Babylonians and Assyrians. The word “cuneiform” was first applied in 1700 by Thomas Hyde, professor of Hebrew in the university of Oxford, in the expression “dactuli pyramidales seu cuneiformes,” and it has found general acceptance, though efforts have been made to introduce the expression “arrow-headed” writing. The name “cuneiform” is fitting, for each character or sign is composed of a wedge (𒁹 or 𒀸), or a combination of wedges (𒈗), written from left to right. The wedge is always pointed towards the right (𒀸) or downwards (𒁹) or aslant(𒀹), or two may be so combined as to form an angle (𒌋) called by German Assyriologists a Winkelhaken, a word now sometimes adopted by English writers on the subject. The word cuneiform has passed into most modern languages, but the Germans use Keilschrift (i.e. wedge-script) and the Arabs mismārī (مسمارى) or nail-writing.
In Persia, 40 m. N.E. of Shiraz, is a range of hills, Mount
Rachmet, in front of which, in a semicircular form, rises a vast
terrace-like platform. It is partly natural, but was
walled up in front, levelled off and used as the base
of great temples and palaces. The earliest European,
Discovery and decipherment.
at present known to us, who visited the site was a
wandering friar Odoricus (about A.D. 1320), who does not seem
to have noticed the inscriptions cut in the stone. These were
first observed by Josaphat Barbaro, a Venetian traveller, about
1472. In 1621 the ruins were visited by Pietro della Valle, who
was the first to copy a few of the signs, which he sent in a letter
to a friend in Naples. His copy was not well made, but it served
the useful purpose of directing attention to an unknown script
which was certain to attract scholars to the problem of its
decipherment. To this end it was necessary that complete
inscriptions and not merely separate signs should be made
accessible to European scholars. The first man to attempt to
satisfy this need was Sir John Chardin, in whose volumes of
travels published at Amsterdam in 1711 one of the small inscriptions
found at the ruins of Persepolis was carefully and accurately
reproduced. It was now plainly to be seen, as indeed others had
surmised, that these inscriptions at Persepolis had been written
in three languages, distinguished each from other by an increasing
complexity in the signs with which they were written. The three
languages have since been determined as Persian, Susian and
Babylonian. But before the decipherment could begin it was
necessary that all the available material should be copied and
published. The honour of performing this great task fell to
Carsten Niebuhr, who visited Persepolis in March 1765, and in
three weeks and a half copied all the texts, so well that little
improvement has been made in them since. When Niebuhr
returned to Denmark he studied carefully the little inscriptions
and convinced himself that the guesses of some of his predecessors
were correct, and that the inscriptions were to be read from left
to right. He observed that three systems of writing were
discernible, and that these were always kept distinct in the
inscriptions. He did not, however, draw the natural conclusion
that they represented three languages, but supposed that the
proud builders of Persepolis had written their inscriptions in
threefold form. He divided the little inscriptions into three
classes, according to the manner of their writing, calling them
classes I., II. and III. He then arranged all those he had
copied that belonged to class I., and by careful comparison
decided that in them there were employed altogether but
forty-two signs. These he copied out and set in order in one of
his plates. This list of signs was so nearly complete and accurate
that later study has made but slight changes in it. When Niebuhr had made his list of signs he naturally enough decided
that this language, whatever it might be, was written in alphabetic
characters, a conclusion which later investigation has not
overthrown. Beyond this Niebuhr was not able to go, and not
even one sign revealed its secret to his inquiry. When, however,
he had published his copies (in 1777) there were other scholars
ready to take up the difficult task. Two scholars independently,
Olav Tychsen of Rostock and Friedrich Münter of Copenhagen,
began work upon the problem. Tychsen first observed that
there occurred at irregular intervals in the inscriptions of the
first class a wedge that pointed neither directly to the right nor
downward, but inclined diagonally. This he suggested was the
dividing sign used to separate words. This very simple discovery
later became of great importance in the hands of Münter.
Tychsen also correctly identified the alphabetic signs for “a,”
“d,” “u” and “s,” but he failed to decipher an entire inscription,
chiefly perhaps because, through an error in history, he supposed
that they were written during the Parthian dynasty (246 B.C.–A.D.
227). Münter was more fortunate than Tychsen in his
historical researches, and this made him also more successful
in linguistic attempts. He rightly identified the builders of
Persepolis with the Achaemenian dynasty, and so located in
time the authors of the inscriptions (538-465 B.C.). Independently
of Tychsen he identified the oblique wedge as a
divider between words, and found the meaning of the sign for
“b.” These may appear to be small matters, but it must be
remembered that they were made without the assistance of any
bilingual text, and were indeed taken bodily out of the gloom
which had settled upon these languages centuries before. They
did not, however, bring us much nearer to the desired goal of
a reading of any portion of the inscriptions. The whole case
indeed seemed now perilously near a stalemate. New methods
must be found, and a new worker, with patience, persistence,
power of combination, insight, the historical sense and the
feeling for archaeological indications.
In 1802 Georg Friedrich Grotefend (q.v.) was persuaded by the librarian of Göttingen University to essay the task. He began with the assumption that there were three languages, and that of these the first was ancient Persian, the language of the Achaemenians, who had erected these palaces and caused these inscriptions to be written. For his first attempts at decipherment he chose two of these old Persian inscriptions and laid them side by side. They were of moderate length, and the frequent recurrence of the same signs in them seemed to indicate that their contents were similar. The method which he now pursued was so simple, yet so sure, as he advanced step by step, that there seemed scarcely a chance of error. Münter had observed in all the Persian texts a word which occurred in two forms, a short and a longer form. This word appeared in Grotefend’s two texts in both long and short forms. Münter had suggested that it meant “king” in the short form and “kings” in the longer, and that when the two words occurred together the expression meant “king of kings.” But further, this word occurred in both inscriptions in the first line, and in both cases was followed by the same word. This second word Grotefend supposed to mean “great,” the combined expression being “king great,” that is, “great king.” All this found support in the phraseology of the lately deciphered Sassanian inscriptions, and it was plausible in itself. It must, however, be supported by definite facts, and furthermore each word must be separated into its alphabetic parts, every one of them identified, and the words themselves be shown to be philologically possible by the production of similar words in related languages. In other words, the archaeological method must find support in a philological method. To this Grotefend now devoted himself with equal energy. His method was as simple as before. He had made out to his own satisfaction the titles “great king, king of kings.” Now, in the Sassanian inscriptions, the first word was always the king’s name, followed immediately by “great king, king of kings,” and Grotefend reasoned that this was probably true in his texts. But if true, then these two texts were set up by two different kings, for the names were not the same at the beginning. Furthermore the name with which his text No. I. began appears in the third line of text No. II., but in a somewhat longer form, which Grotefend thought was a genitive and meant “of N.” It followed the word previously supposed to be “king” and another which might mean son (N king son), so that the whole expression would be “son of N king.” From these facts Grotefend surmised that in these two inscriptions he had the names of three rulers, grandfather, father and son. It was now easy to search the list of the Achaemenian dynasty and to find three names which would suit the conditions, and the three which he ventured to select were Hystaspes, Darius, Xerxes. According to his hypothesis the name at the beginning of inscription I. was Darius, and he was ready to translate his texts in part as follows:—
| I. | Darius, great king, king of kings . . . son of Hystaspes. . . . |
| II. | Xerxes, great king, king of kings . . . son of Darius king. |
The form which he provisionally adopted for Darius was Darheush; later investigation has shown that it ought really to be read as Daryavush, but the error was not serious, and he had safely secured at least the letters D, A, R, SH. It was a most wonderful achievement, the importance of which he did not realize, for in it was the key to the decipherment of three ancient languages. To very few men has it been given to make discoveries so important both for history and for philology.
To Grotefend it was, however, not given to translate a whole text, or even to work out all the words whose meaning he had surmised. Rasmus Christian Rask (1787–1832), who followed him, found the plural ending in Persian, which had baffled him; and Eugène Burnouf (1801–1852), by the study of a list of Persian geographical names found at Naksh-i-Rustam, discovered at a single stroke almost all the characters of the Persian alphabet, and incidentally confirmed the values already determined by his predecessors.
At the same time as Burnouf, the eminent Sanskrit scholar Professor Christian Lassen (1800–1876), of Bonn, was studying the same list of names; and his results were published at the same time. The controversy which resulted as to priority of discovery may be here passed over while we sum up the results in general conclusions. Lassen may certainly claim in the final court of history that he discovered independently of Burnouf the values of at least six and possibly of eight signs. But in another respect he made very definite progress over Burnouf. He discovered that, if the system of Grotefend were rigidly followed, and to every sign were given the value Grotefend had assigned, some words would be left wholly or almost wholly without vowels; and therefore unpronounceable. As instances of such words he mentioned ÇPRD, THTGUS, KTPTUK, FRAISJM. This situation led Lassen to a very important discovery, towards which his knowledge of the Sanskrit alphabet did much to bring him. He came, in short, to the conclusion that the ancient Persian signs were not entirely alphabetic, but were at least partially syllabic, that is, that certain signs were used to represent not merely an alphabetic character like “b,” but also a syllable such as “ba,” “bi” or “bu.” He claimed that he had successfully demonstrated that the sign for “a” was only used at the beginning of a word, or before a consonant, or before another vowel, and that in every other case it was included in the consonant sign. Thus in the inscription No I. in the second line the signs should be read VA-ZA-RA-KA. This was a most important discovery, and may be said to have revolutionized the study of these long puzzling texts.
During the entire time of this slow process of decipherment, from the first essays of Grotefend in 1802 until the publication of Lassen’s book in 1836, there were more sceptics than believers in the results of the deciphering process. Indeed the history of all forms of decipherment of unknown languages shows that scepticism concerning them is far more prevalent than credulity or even a too ready acceptance. There was need for a man of another people, of different training and a fresh and unbiased mind, to put the capstone upon the decipherment, and he was already at work when Lassen’s important researches appeared.
Major (afterward Sir) Henry Rawlinson had gone out to India, in the service of the East India Company, while still a boy. There he had learned Persian and several of the Indian vernaculars. That was not the sort of training that had prepared Grotefend, Burnouf or Lassen, but it was the kind that the early travellers and copyists had enjoyed. In 1833 young Rawlinson went to Persia, to work with other British officers in the reorganization of the Persian army. While engaged in this service his attention was drawn to the ancient Persian cuneiform inscriptions. In 1835 he copied with great care the texts at Hamadan, and began their decipherment. Of all the eager work which had been going on in Europe he knew little. It is no longer possible to ascertain when he gained his first information of Grotefend’s work, for Norris, the secretary of the Royal Asiatic Society, has left us no record of when he began to send notices of the German’s work. Whenever it was, there seems to be no doubt that Rawlinson worked independently for a time. His method was strikingly like Grotefend’s. He had copied two trilingual inscriptions, and recognized at once that he had three languages before him. In 1839 (Journal of the Royal Asiatic Society, x. pp. 5, 6) he thus wrote of his method: “When I proceeded . . . to compare and interline the two inscriptions (or rather the Persian columns of the two inscriptions, for, as the compartments exhibiting the inscription in the Persian language occupied the principal place in the tablets, and were engraved in the least complicated of the three classes of cuneiform writing, they were naturally first submitted to examination) I found that the characters coincided throughout, except in certain particular groups, and it was only reasonable to suppose that the grounds which were thus brought out and individualized must represent proper names. I further remarked that there were but three of these distinct groups in the two inscriptions; for the group which occupied the second place in one inscription, and which, from its position, suggested the idea of its representing the name of the father of the king who was there commemorated, corresponded with the group which occupied the first place in the other inscription, and thus not only served determinately to connect the two inscriptions together, but, assuming the groups to represent proper names, appeared also to indicate a genealogical succession. The natural inference was that in these three groups of characters I had obtained the proper names belonging to three consecutive generations of the Persian monarchy; and it so happened that the first three names of Hystaspes, Darius and Xerxes, which I applied at hazard to the three groups, according to the succession, proved to answer in all respects satisfactorily and were, in fact, the true identification.”
Rawlinson’s next work was the copying of the great inscription of Darius on the rocks at Behistun (q.v.). He had first seen it in 1835, and as it was high up on the rocky face, and apparently inaccessible, he had studied it by means of a field-glass. He was not able to copy the whole of the Persian text, but in 1837, when he was more skilled in the script, he secured more of it. In the next year he forwarded to the Royal Asiatic Society of London his translation of the first two paragraphs of the Persian text, containing the name, titles and genealogy of Darius. This was little less than a tour de force, for it must be remembered that this had been accomplished without the knowledge of other ancient languages which his European competitors had enjoyed. The translation, received in London on the 14th of March, made a sensation, and a transcript sent in April to the Asiatic Society of Paris secured him an honorary membership in that distinguished body. He was now known, and many made haste to send him copies of everything important which had been published in Europe. The works of Burnouf, Niebuhr, le Brun and Porter came to his hands, and with such assistance he made rapid progress, and in the winter of 1838–1839 his alphabet of ancient Persian was almost complete. In 1839 he was in Bagdad, his work written out and almost ready for publication. But he delayed, hoping for more light, and revising sign by sign with exhaustless patience. He expected to publish his preliminary memoir in the spring of 1840, when he was suddenly sent to Afghanistan as political agent at Kandahar. Here he was too busily engaged in war administration to attend to his favourite studies, which were not renewed until 1843 when he returned to Bagdad. There he received fresh copies and corrections of the Persepolis inscriptions which had been made by Westergaard, and later made a journey to Behistun to perfect his own copies of the texts which had formed the basis of his own first study. At last, after many delays and discouragements, he published, in 1846, in the Journal of the Royal Asiatic Society, his memoir, or series of memoirs, on the ancient Persian inscriptions, in which for the first time he gave a nearly complete translation of the Persian text of Behistun. In this one publication Rawlinson attained imperishable fame in Oriental research. His work had been carried on under greater difficulties than those in the path of his European colleagues, but he had surpassed them all in the making of an intelligible and connected translation of a long inscription. He had indeed not done it without assistance from the work of Burnouf, Grotefend and Lassen, but when all allowance is made for these influences his fame is not diminished nor the extent of his services curtailed. His method was adopted before he knew of Lassen’s work. That two men of such different training and of such opposite types of mind should have lighted upon the same method, and by it have attained the same results, confirmed in the eyes of many the truth of the decipherment.
The work of the decipherment of the old Persian texts was now complete for all practical purposes. But in 1846 there appeared a paper read before the Royal Irish Academy by the Rev. Edward Hincks of Killyleagh, County Down, Ireland, whose keen criticisms of Lassen’s work, and original contributions to the definite settlement of syllabic values, may be regarded as closing the period of decipherment of Persian cuneiform writing.
The next problem in the study of cuneiform was the decipherment of the second language in each of the trilingual groups. The first essay in this difficult task was made in 1844 by Niels Louis Westergaard. His method was very similar to that used by Grotefend in the decipherment of Persian. He selected the names of Darius, Hystaspes, Persians and others, and compared them with their equivalents in the Persian texts. By this means he learned a number of signs, and sought by their use in other words to spell out syllables or words whose meanings were then ascertained by conjecture or by comparison. He estimated the number of characters at eighty-two or eighty-seven, and judged the writing to be partly alphabetic and partly syllabic. The language he called Median, and classified it in “the Scythian, rather than in the Japhetic family.” The results of Westergaard were subjected to incisive criticism by Hincks, who made a distinct gain in the problem. It next passed to the hands of de Saulcy, who was able to see further than either. But the matter moved with difficulty because the copied texts were not accurate. By the generosity of Sir Henry Rawlinson his superb copies of the Behistun text, second column, were placed in the hands of Mr Edwin Norris, who was able in 1852 to present a paper to the Royal Asiatic Society deciphering nearly all of it. Mordtmann followed him, naming the language Susian, which was met with general acceptance and was not displaced by the name Amardian, suggested by A. H. Sayce in two papers which otherwise made important contributions to the subject. With his contributions the problem of decipherment of Susian may be considered as closed. The latter workers could only be builders on foundations already laid.
The decipherment of the third of the three languages found at Persepolis and Behistun followed quickly on the success with Susian. The first worker was Isadore Löwenstern, who made out the words for “king” and “great” and the sign for the plural, but little more. The first really great advance was made by Hincks in 1846 and 1847. In these he determined successfully the values of several signs, settled the numerals, and was apparently on the high-road toward the translation of an entire Assyrian text. He was, however, too cautious to proceed so far, and the credit of first translating a short Assyrian text belongs to Longpérier, who in 1847 published the following as the translation of an entire text: “Glorious is Sargon, the great king, the (. . .) king, king of kings, king of the land of Assyria.” It was nearly all correct, but it advanced our knowledge but slightly because it did not give the forms of the words—because (to put it in another way) he was not able to transliterate the Assyrian words. This was the great problem. In the Persian texts there were but forty-four signs, but in the third column of the Persepolis texts Grotefend had counted one hundred and thirty different characters, and estimated that in all the Babylonian texts known to him there were about three hundred different signs, while Botta discovered six hundred and forty-two in the texts found by him at Khorsabad. That was enough to make the stoutest heart quail, for a meaning must be found for every one of these signs. There could not be so many syllables, and it was, therefore, quite plain that the Babylonian language must have been written in part at least in ideograms. But in 1851 Rawlinson published one hundred and twelve lines of the Babylonian column from Behistun, accompanied by an interlinear transcription into Roman characters, and a translation into Latin. That paper, added to Hinck’s still more acute detail studies, brought to an end the preliminary decipherment of Babylonian. There were still enormous difficulties to be surmounted in the full appreciation of the complicated script, but these would be solved by the combined labours of many workers.
The cuneiform script had its origin in Babylonia and its inventors were a people whom we call the Sumerians. Before the Semitic Babylonians conquered the land it was inhabited by a people of unknown origin variously classified, by different scholars, with the Ural-altaic or even Origin. with the Indo-European family, or as having blood relationship with both. This people is known to us from thousands of cuneiform inscriptions written entirely in their language, though our chief knowledge of them was for a long time derived from Sumerian inscriptions with interlinear translations in Assyrian. Their language is called Sumerian (li-ša-an Su-me-ri) by the Assyrians (Br. Mus. 81–7-27, 130), and its characteristics are being slowly developed by the elaborate study of the immense literature which has come down to us. In 1884 Halévy denied the existence of the Sumerian language, and claimed that it was merely a cabalistic script invented by the priests of the Semites. His early success has not been sustained, and the vast majority of scholars have ceased to doubt the existence of the language.
The Sumerians developed their script from a rude picture-writing,
some early forms of which have come down to us. In
course of time they used the pictures to represent sounds, apart
from ideas. They wrote first on stone, and when clay was
adopted soon found that straight lines in soft clay when made
by a single pressure of the stylus tend to become wedges, and the
pictures therefore lost their character and came to be mere
conventional groups of wedges. Some of these wedge-shaped
signs are of such character that we are still able to recognize
or re-construct the original picture from which they came. The
Assyrian sign 
, which means heaven, appears in early texts
in the form 
in which its star-like form is quite evident
(star = heaven) and from which the linear form 🞻 may be not
improbably pre-supposed. A number of other cases were
enumerated by the Assyrians themselves (see Cuneiform Texts from Bab. Tab. in Brit. Museum, vol. v., 1898), and there can be
no reasonable doubt that this is the origin of the script.
The number of the original picture-signs cannot have been
great, but the development of new signs never ceased till the
cuneiform script passed wholly from use. The simplest
form of development was doubling, to express plurality
or intensity. After this came the working of two
Development and characteristics.
signs into one; thus 𒀀 “water,” when placed in 𒅗
“mouth” gave the new sign 𒅘 “to drink,” and many others.
Other signs were formed by the addition of four lines, either
vertically or horizontally, to intensify the original meaning.
Thus, for instance, the old linear sign 
means dwelling, but
with four additional lines, thus 
, it means “great house.”
This sign gradually changed in form until it came to be 
.
This method of development was called by the Sumerians gunu,
and signs thus formed are now commonly called by us, gunu
signs. They number hundreds and must be reckoned with in
our study of the script development, though perhaps recent
scholars have somewhat exaggerated their importance. The
process of development is obscure and must always remain so.
The script as finally developed and used by the Assyrians is cumbrous and complicated, and very ill adapted to the sounds of the Semitic alphabet. It has (1) simple syllables, consisting of one vowel and a consonant, or a vowel by itself, thus 𒀀 “a,” 𒀊 ab, 𒅁 ib, 𒌒 ub, 𒁀 ba, 𒑙 bi, 𒁍 bu. In addition to these the Assyrian had also (2) compound syllables, such as 𒂍 bit, 𒁄 bal, and (3) ideograms, or signs which express an entire word, such as 𒎏 beltu, lady, 𒀜 abu, father. The difficulty of reading this script is enormously increased by the fact that many signs are polyphonous, i.e. they may have more than one syllabic value and also be used as an ideogram. Thus the sign 𒆳 has the ideographic values of matu, land, shadu, mountain, kashadu, to conquer, napachu, to arise (of the sun), and also the syllabic values kur, mad, mat, shad, shat, lat, nad, nat, kin and gin. This method of writing must lead to ambiguity, and this difficulty is helped somewhat by (4) determinatives, which are signs intended to indicate the class to which the word belongs. Thus, the 𒁹 is placed before names of persons, and 𒆳 (the ideogram for matu, country, and shadu, mountain) is placed before names of countries and mountains, and 𒀭 (ilu, god) before the names of gods.
The cuneiform writing, begun by the Sumerians in a period so remote that it is idle to speculate concerning it, had a long and very extensive history. It was first adopted by the Semitic Babylonians, and as we have seen was modified, developed, nay almost made over. Their inscriptions History. are written in it from circa 4500 B.C. to the 1st century B.C. From their hands it passed to the Assyrians, who simplified some characters and conventionalized many more, and used the script during the entire period of their national existence from 1500 B.C. to 607 B.C. From the Babylonian by a slow process of evolution the much simplified Persian script was developed, and with the Babylonian is also to be connected the Susian, less complicated than the Babylonian, but less simple than the Persian. The Chaldians (not Chaldaeans), who lived about Lake Van, also adopted the cuneiform script with values of their own, and expressed a considerable literature in it. The discovery in 1887 of the Tell-el-Amarna tablets in upper Egypt showed that the same script was in use in the 15th century B.C., from Elam to the Mediterranean and from Armenia to the Persian Gulf for purposes of correspondence. There is good reason to expect the discovery of its use by yet other peoples. It was one of the most widely used of all the forms of ancient writing.
Bibliography.—The history of the decipherment may be further studied in R. W. Rogers, History of Babylonia and Assyria, vol. i. (N.Y. and London, 1900); and in A. J. Booth, The Discovery and Decipherment of the Trilingual Cuneiform Inscriptions (London, 1902), which is very exhaustive and accurate. The Sumerian question may best be studied in F. H. Weissbach, Die Sumerische Frage (Leipzig, 1898), and Charles Fossey, Manuel d’Assyriologie, tome i. (Paris, 1904). For development and characteristics, see Friedrich Delitzsch, Die Entstehung des ältesten Schriftsystems (Leipzig, 1897); Paul Toscanne, Les Signes sumériens dérivés (Paris, 1905). (R. W. R.)