The discovery and decipherment of the trilingual cuneiform inscriptions

by Arthur John Booth

Löwenstern, as we have said, considered that Rawlinson was fundamentally wrong in applying the laws derived from Indo-European languages ‘to a writing and a language that are Semitic.’ He absolutely denied that the vowel is inherent.[791] The signs are simple consonants, and they may be used in connection with any vowel sound. The vowels may or may not be expressed, and the signs for them are to a large extent expressive of any vowel sound. One sign he mentions may convey the sound of hou, a, and ya; another of a or ha, w and ü.[792] The vowels are, he says, by no means limited to the a, i and u of the Sanscrit, but include also the e and o and the diphthong ao. He entirely disagrees with the opinion of Rawlinson that the equivalent signs have any modified value. He compares Assyrian with Egyptian, and regards the signs that are apparently interchangeable as simple ‘homophones.’ He shows the different ways in which the names of the Achaemenian kings are written; and draws the apparently inevitable inference that the different signs have one and the same sound. ‘The variants,’ he says, ‘may be used indifferently without violating the phonetic laws of the language.’[793] He was apparently the first to observe that some signs ‘express different sounds’; and these he calls ‘homotypes.’[794] As was natural, he does not appear to have had any idea of the importance of this discovery. His homotypes seem limited to the signs for vowels, any one of which may express almost any vowel sound, and also aspirates and liquids; and he observed that m and w, and y and i, are each expressed by the same signs.

His present pamphlet indicates how rapidly the study was progressing. He now relies entirely upon the analysis of proper names, in accordance with the suggestion of Longpérier; and he abandons his attempted comparison with the form of the Hebrew letters. He surrenders his reading of ‘Ashdod,’ and suggests ‘No Kaschzar’ in place of it; and he even doubts the identity of Arsak and Sargon. He thinks he has discovered from the Naksh-i-Rustam inscription that the sign he mistook for r is really s, and that his k is certainly n. Accordingly he reads the word s ch(kh) n which somewhat revives his confidence in Sargon.

It is not clear to what extent, if to any, he was indebted to Hincks. His exaggerated Semitism was probably of native growth. He was not yet aware of the age of the Babylonian bricks, as explained by Hincks in 1846: yet he knew the determinative sign for proper names, which apparently was not known to Hincks.[795] Hincks, on the other hand, recognised a sign ‘prefixed non-phonetically to the name Ormuzd, and also used by abbreviation for the word “god.”’ Löwenstern says there is ‘no special sign accompanying the names of the gods.’ He, however, recognised a sign as the monogram for ‘god’; but when he found it in conjunction with the name of ‘Aurmuzd’ he treated it as the initial letter, and gave it the definite phonetic value of a.[796] Such were the difficulties to be overcome before the determinative for ‘god’ was recognised. He observed that a word may be expressed by its first and last signs, an early indication of the phonetic complement.[797] Meanwhile Hincks and Rawlinson announced the discovery already mentioned that the apparent equivalent signs in Persian depended in reality upon the vowel that was associated with the consonantal value. It at once occurred to Longpérier that the great difficulty of the Assyrian homophones might be solved by the application of the same principle. If, he says, there is a separate sign for the consonant m according as it is followed by a, i or u, ‘one can understand how a similar practice, if extended to many consonants, would augment the number of alphabetical signs.’ He warned scholars not to be too ready to accept the existence of homophones, because he observed that ‘according as the work proceeds the number of homophones decreases.’[798]

While these discussions were proceeding, Botta continued his contributions to the ‘Journal Asiatique’ (1847-8), and afterwards published them in a separate ‘Mémoire sur l’Ecriture cunéiforme’ (Paris, 1848). He endeavoured to introduce some degree of order among the profusion of Assyrian signs. He drew up a Table, consisting of a hundred and twenty-five signs that seemed to be most commonly used; and under each of these he arranged the signs which he found were sometimes apparently interchanged for them.[799] In this List of Variants we constantly find six or seven signs—sometimes many more—grouped together as of equivalent value. In view of later discoveries, it will be seen how extremely useful this classification might become, for the signs thus brought together were no doubt usually those that contained the same consonantal values. In the meantime, however, Botta was at a loss to find any reasonable explanation. Like Rawlinson, however, he could not believe that any of them were, as Löwenstern maintained, real homophones, or signs having identical values. They must, he thought, be distinguished from each other by some slight shades of sound that were sufficiently near to be easily confounded.[800] He explained, in anticipation of the discovery so soon afterwards made by Hincks, that ‘it is possible that the language is syllabic—so far, at least, as that each consonant is represented by a different sign, according to the vowel to which it is joined. Thus, for example, there would be one sign for b; others for ba, bi, etc. In Semitic languages the short vowels have little importance, and therefore the syllable ba might be expressed by the sign for b only; by the two signs b and a; and also in certain cases, by the signs that represent b in connection with the other vowels.’ It will be seen that Botta was very far indeed from being the mere painstaking classifier which it was once the fashion to describe him. He and Longpérier were, in fact, the only two Continental scholars, at present occupied with this subject, who were gifted with any real penetration into its difficulties.

Botta succeeded in dividing nearly the whole of the Bull inscription correctly into its words, but the difficulty of this task was still so great that even he occasionally fell into error. He also first pointed out that the sign Löwenstern mistook for two signs was one and indivisible. He detected the determinative sign for ‘country’ that is used in the Khorsabad inscription, and he made the important suggestion that the phonetic value of the sign for ‘king’ is ‘sar.’[801] Longpérier at once connected this word with its Hebrew equivalent, and showed that it is used to express the first syllable in the name of the Khorsabad king ‘Sar-gin.’ He made an attempt to decipher an inscription on the leg of the Khorsabad bull, and he was the first to recognise ‘Assur.’ His translation runs thus: ‘Glorious [is] Sargon, King, great King, King of Kings, King of the country of Assur.’[802] He also showed that ‘great’ might be expressed by one sign only,[803] which added another step to the discovery of the phonetic complement begun by Löwenstern. This short contribution to the ‘Revue Archéologique’ shows that Longpérier possessed to a high degree a true aptitude for these studies; and if he had been able to pursue them, he might have vindicated for France a more favourable position than it was her fortune to obtain. The difficulty in these matters of recognising truth from error was nowhere more clearly illustrated than in the case of the identification of ‘Sargon’ by Longpérier. So far was the correctness of this ingenious suggestion from gaining immediate acceptance, that we find Hincks subsequently conjectures that the name of the father of Sennacherib should be read ‘Ni-Shar.’[804] A later attempt, in 1849, which resulted in ‘Kin-nil-li-n’a’ showed little improvement. Even in 1850 Rawlinson is still so far afield that he translates it ‘Arko-tsin’; and it is not till August 1851 that he accepts ‘Sargina,’ the reading given four years previously by Longpérier.[805] Yet the question was of no little interest, for it really settled the controversy between Hincks and Rawlinson as to the date of the Lower Assyrian dynasty in favour of the former. It was not till Rawlinson read ‘Sennacherib’ in a tablet found by Layard at Kouyunjik that he would acknowledge his error, and admit that there was at last found ‘a tangible starting place for chronology.’ Hincks was satisfied, two years earlier, that he had identified the names of Sennacherib and Esarhaddon.

In June 1849 Hincks read a paper on the Khorsabad inscriptions which shows a great advance upon any contribution yet made to the decipherment of the language. There is evidence in his essay on the Van inscriptions that so early as December 1847 he had practically given up the attempt he made in the previous January to assimilate the Assyrian writing as far as possible to the Hebrew—at least as regards the vowel system.[806] He is now satisfied that the Assyrian maintains a clear distinction between the vowels, and also between the surd and sonant consonants at the beginning of a word, though at the end the two sounds were confounded. He points out that no distinction is made between the sounds of w and m, and, he adds, between l and r; but he afterwards correctly admits the independent existence of these sounds in his Syllabarium.[807] He has also definitely arrived at the conviction that the Assyrian characters are wholly syllabic or ideographic—in a large number of cases they are both. He will not now admit that any of them represents a simple consonant. He has still no doubt that there are many homophones. Many characters appear to have precisely the same values, ‘though much fewer than might be inferred from a mechanical comparison of inscriptions and observance of interchanges.’ He recognises the existence of polyphones, already described by Löwenstern in 1847 as ‘homotypes.’

‘Many characters,’ he says, ‘admit of two or more kindred values, the distinction between which would appear not to have been considered so great as to require different modes of representing them.’ This discovery was so perplexing that he doubted how best to present it to the reader: ‘Whether it is more desirable to give different values to the same character, or to give it one value only, with a warning to the reader that he may, under certain restrictions, substitute another for it at his pleasure.’

But the chief importance of his present essay consists in the light it throws for the first time on the nature of the ideograms. The earliest inquirers leant to the opinion that the language was at least partly monogrammatic, though Grotefend was inclined to regard these signs more in the light of abbreviations. He, however, distinctly pointed out the existence of ideographic characters in the shorter inscriptions at Persepolis. In 1846 (December) Hincks recognised a sign that was used ‘by abbreviation for the word “god”’; and he noticed that ‘besides having a phonetic value, it is used as a non-phonetic initial before the name of Ormuzd.’ In the same essay he gave numerous other instances of the existence of ideographic and non-phonetic determinative signs.[808] In 1849 he added for the first time the true phonetic value (an) for the sign for ‘god’;[809] and in the present essay he shows the various uses to which ideographic signs may be applied. He explains that many phonetic characters also express words; these may be considered as abbreviations, though possibly some ‘originally denoted ideas and thence, in process of time, the initial sounds in the words which express them.’ A second class resemble the mixed signs of the Egyptians: they may represent words by themselves, but they sometimes require the addition of complements. Another class never have complements, nor any phonetic value except in compound nouns, of which the word they represent forms an integral part. The ideograms, however, that give rise to the most interesting speculation are those that have phonetic values, but where the words that denote the ideas they express have no phonetic relationship to the phonetic value of the ideogram. For example, we now know that the phonetic value of the ideogram for ‘god’ is an; but this syllable forms no portion of the Assyrian word for ‘god,’ which is il-u. A glance over any table giving the syllabic values of ideograms will show how extensively this peculiarity prevails; and its recognition soon led Hincks to the important deduction that the writing was borrowed from some other people where the phonetic value of the ideograms was in some sort of agreement with the initial sound of the word they represented.[810] Hincks dwelt on the great difficulty of deciphering a language in which the characters are sometimes used as phonetic syllables and sometimes as ideographs. In each case it was necessary first to determine in which sense it occurred, and, if in the latter, the pronunciation could only be ascertained when it was found spelt out phonetically in some known word. For example, the pronunciation of the ideogram for ‘god’ was fixed by finding that it formed the il-u in ‘Bab-ilu.’ When the pronunciation of the ideogram was known, it afforded, as has been said, in the majority of cases no clue whatever to the syllabic value of the sign, and the transliterator was liable to fall into the error of reading a word ideographically instead of phonetically, just as frequently as to mistake ideograms for phonetics. Compound ideograms were also not infrequent, where two or more were used to express an idea, but without reference to the sound. For example, the word for ‘palace’ is composed of two ideograms, bit and rab, meaning respectively ‘house’ and ‘great’; but Hincks warned the reader that he might fall into a serious error if he were to suppose that they are employed in conjunction phonetically, and that ‘bitrab’ is the pronunciation of the Assyrian word for ‘palace.’[811] He shows that several ideographs may be used as simple determinative suffixes to words which are phonetically complete without them. The determinatives probably all originally represented words, and many of them preserved their phonetic values.[812]

This inquiry into the nature and use of ideograms was the first that had been made, and it formed an important contribution to the knowledge of the language. His attempted transliterations are not of equal value. He rejected the reading of ‘Sargon’ given by Löwenstern, and he does not seem to have heard of the solution of the difficulty proposed the year before by Longpérier.[813] He recognised indeed that the sign for ‘king’ with which the word begins forms an integral portion of the name, but he did not perceive, like Longpérier, that its counterpart is the Hebrew ‘sar’; and he was led by other comparisons to assign the value of ‘kin-nil’ to the ideogram. He knew also that the sign Löwenstern had broken up into two and thought signified r, s formed in fact a single sign, which he pronounced ri or li. These efforts resulted in ‘kin-nil-li-n’a,’ which might seem even less manageable than the r, s, k of Löwenstern. But Hincks was quite equal to the occasion, and, with the customary imaginative faculty of the philologist, he found no difficulty in connecting this person with the Chinzirus of Ptolemy, who, it appears, was a contemporary of Porus. He had already detected that the names of the son and grandson of this prince were Sennacherib and Esarhaddon, the builders of Kouyunjik and the South-West Palace of Nimrud. The first we find he transliterated ‘Sanki’ or ‘Sankin,’ with the possible addition of ‘rav’ or ‘ram’—‘Sankin-rav’: the other came out as ‘Adar-ka-dan.’ He also explains how he arrived at ‘Nabiccudurrayuchur’ for Nebuchadnezzar, and how he fancied he had found ‘Jerusalem’; but these instances only serve to illustrate the great obstacles that had still to be overcome. We have already mentioned the suggestions that were made first by Longpérier and afterwards by Botta with the view of reducing the number of homophones. Hincks confesses he had not seen the essay of the former, and we have not observed that he has acknowledged his obligations to Botta, though he was evidently acquainted with his work, as we see from an unpleasant reference to ‘a mechanical comparison of inscriptions.’ It was, however, upon the principle these writers suggested that Hincks was now about to solve one of the greatest difficulties of the language. The solution is contained in an Appendix to the essay just reviewed, and was sent to press on January 19, 1850, the same day that Major Rawlinson read his first paper to the Asiatic Society. Hincks now explains that there are four distinct vowel sounds in Assyrian, ā, a, i and u; but the difference between the first two was not maintained when they preceded a consonant.[814] He laid down that every sign represented a consonant either preceded or followed by one of these vowels. Therefore, each consonant was represented by seven signs, thus: cā, ca, ci, cu; ac, ic, uc. He thought there were at least fifteen consonants, and that the syllabary was of Indo-European origin, and need not therefore, as he had at first supposed, be adjusted to the Hebrew alphabet. The principle thus announced has been accepted with some modifications. The difference between the long and short a has not been maintained in this connection, and consequently the syllabic representation of each consonant is reduced from seven to six. His statement that the difference between surd and sonant is maintained at the beginning but not at the end of the syllable has also been admitted. We have thus separate signs for ba, bi, bu and pa, pi, pu; but the signs for ap, ip and up answer for both. Hincks’s consonants have been accepted without material change. His y has been omitted and h added; z has been substituted for j. Two signs for k to represent Caph and Koph, and two for t to represent Teth and Tau have been added, where Hincks only had one for each, so that the number of consonants is now raised to seventeen. Not only did Hincks arrive at a correct theory of the simple syllables, but he identified correctly a very large number of the signs corresponding to them. Of the seventy-one he gives in his Table at least fifty-seven are accurate, and possibly even more. He closed the essay with a brief specimen of a translation from the Khorsabad inscription.