Part III. was at the same time in the hands of the printer, and appeared later in the year (1847). It contains Chapters IV. and V. of the Memoir. The former gives a revised transliteration and translation of the text with an Analysis; and the other a complete revised edition of all the inscriptions previously published by Lassen. Mr. Norris, the Assistant Secretary of the Asiatic Society, saw the whole of this complicated work through the press, and he undertook to alter the transliteration given in Chapter IV. in accordance with the principles laid down in the Supplementary Note. For his services in this matter, he received a vote of thanks from the Society. ‘He unites,’ said the proposer, ‘more varied learning and more rare and extensive research and intelligence than I have ever seen combined in the same individual’;[558] and he subsequently attained an independent position in the first rank of cuneiform scholars.

Part IV. was not published till 1849. It included the sixth chapter of the Memoir and treated of the Vocabulary, but the dissertation was never completed. It breaks off in the middle of a sentence, when the writer had not proceeded further than the words commencing with vowels and with consonants of the first three classes. It purports to give a few brief etymological explanations, but in reality it is admirable as a display of learning in many fields of knowledge, and it is especially interesting for the explanation it affords of the reasons that led to the determination of words of doubtful meaning. While he was engaged in this work, he received Lassen’s Second Memoir. It did not reach him till August 1845, when his own translations were completed and already beyond the reach of alteration. He had little cause, however, to regret the delay that arose from the difficulty in those times of ‘communicating between Bonn and Bagdad,’ for he could have derived small benefit from the very inferior translations of his predecessor. In the philological branch of the subject, however, he found the Memoir ‘of the greatest convenience as a manual of reference,’ and his marginal notes show how carefully he consulted it.[559]

By the publication of this work Major Rawlinson at length took his place among the cuneiform scholars of Europe. We have shown that the study was by that time far advanced, and most of the difficulties of the Persian column were already surmounted. Rawlinson did not, therefore, put forward any pretension to original discovery in that department, but was, he said, ‘content to rest my present claims on the novelty and interest of my translations.’[560] He hoped eventually to earn the higher distinction of an original discoverer, ‘according to the success that may attend my efforts to decipher the Median [Susian] and Babylonian inscriptions.’ It was not, however, without an effort that he presented himself in so modest a garb upon this occasion. He was convinced that he had made each step in the tedious process of decipherment by his own unaided effort; and in whatever light he might appear to the public, he was certainly an original discoverer to himself. He had no doubt that if Grotefend and Lassen had never lived the world would have been indebted wholly to him for the discovery, and, although we think he may have been influenced more than he suspected by other scholars, there is no great improbability in supposing that his own ingenuity would have been quite equal to grapple singly with the task.

Notwithstanding his avowed disclaimer, he still cherished the opinion that he had really made some important contributions to the determination of the alphabet. On one occasion indeed he went so far as to claim the paternity, directly or indirectly, of at least ten characters, and he referred to his correspondence with Burnouf and Lassen as the medium through which he had made his influence felt. It is clear, however, that in this he was entirely mistaken. According to his own admission, he knew as little of the Continental scholars as they did of him until his first communication to the Asiatic Society, which was received in March 1838. It will be recollected that Burnouf and Lassen had published their Memoirs two years before; so neither of these could have been influenced by Rawlinson. It only remains to inquire whether he could have suggested any of the six values ascribed to Jacquet, whose essay appeared in the course of 1838. Here a comparison of dates is not sufficient in itself to determine the question. Rawlinson’s communication was known in London on March 14, and was submitted to the French Society on April 20. Jacquet began that very month to publish his criticism of Lassen, and his active mind was full of the subject. He was no doubt present at the meeting when Rawlinson’s copy of the inscription was submitted to the Society, and there was ample time for him to profit by any suggestions it contained in his future papers on the subject. We have, however, conclusive proof from Rawlinson’s own admission that the values of these six letters were not then known to him. Nor could they have been communicated to Jacquet through his subsequent correspondence with Burnouf. Jacquet died in July 1838, and Rawlinson’s correspondence with Burnouf and Lassen did not begin till the summer of that year.[561] From that period Rawlinson himself accounts for all the letters in question. Writing after Jacquet’s death, he tells us in a letter to Burnouf that he had just found the value of 16 (𐎨) ch. Two other values, 26 (𐎰) th and 41 (𐏃) h, he fixed still later in the winter 1838-9; another, 27 (𐎹) y, he acknowledges he received from Lassen.[562] The sign 10 (𐎺) he also fixed in 1838-9; but he gave it the same value as Lassen had done in 1836, viz. w. In Germany w was no doubt equivalent to its correct value v, but scarcely so to an English-speaking man, especially as he distinguishes it from his v (𐎻) No. 15. The other letter, 40 (𐎽) r, was known correctly to Grotefend in 1837. With reference, therefore, to the six letters attributed to Jacquet, it is seen that none of them were due to the influence of Rawlinson, either through his Memoir or subsequently by correspondence. One letter (r) was fixed before Rawlinson was known. Three others were first announced after Jacquet was dead (16, 26, 41). One was wrong (10), and the other (27) he acknowledged to have borrowed from Lassen. It is impossible, therefore, to admit the pretension put forward by Rawlinson, that he could ‘fairly claim the paternity, either directly or indirectly, of at least ten characters’ on the ground that ‘it was impossible to say by whom each individual letter became identified.’ On the contrary, the history of the identification is plain enough, and there is no difficulty in assigning the proportion of merit due to each discoverer. It was not till after the essays of Jacquet that Rawlinson bore any share in the general progress of the study; and then not more than four characters remained to be correctly identified.

We have already seen with what conspicuous success Rawlinson had found the true values of two of these, so far back as 1838. One still gave Lassen a great deal of trouble, and he had variously valued it as k (1836), ich (1839), k’h (1844). In his letter to Rawlinson he preferred to leave it undetermined (1839).[563] Rawlinson suggested that it had the sound of t before i, which is so nearly correct and so great an improvement upon all previous attempts that it might almost be conceded to him as an approximate value if he had announced it earlier. He acknowledges that he remained long in doubt concerning it, and there is no evidence, as in the case of the other two letters, that he suggested the emendation to Burnouf.[564] Before his alphabet appeared, in 1846, the true value had been already fixed by Holtzmann in 1845 as d before i.

It thus appears that Rawlinson had a real aptitude for unravelling this kind of puzzle. Only four letters were left to him by his predecessors; and of these he determined two correctly and one nearly correctly. The fourth, 28 (𐎩) the z of Jacquet, he improves to an approximate correct value j’h in his first alphabet; and in his second he gives it correctly as j before a: a correction made simultaneously by Hincks. He may also claim the merit of having restored the sound of k (he writes kh) to 25 (𐎤). The value of this letter had long before been fixed by Grotefend, but since then it had passed through many vicissitudes. St. Martin thought it was h; Burnouf made it q; and Lassen thought, in 1836, it stood for the a in the diphthong au, ô, till at length, in 1844, he reluctantly adopted Burnouf’s q.

Tn addition to these services, Rawlinson contributed two new letters, one of which, No. 43 (𐎵), n before u, has taken a permanent place in the alphabet. The other (𐎾), ñ, is really a Susian letter with a nasal sound, and is found in the Persian column in only two proper names. Oppert suggests that it may be the missing l; and Spiegel is disposed to agree.[565]

If we consult Rawlinson’s alphabet as it stood early in 1846, it will be seen that he was in possession of correct values for the thirty-three signs in Niebuhr’s list, with the exception of two, 10 (𐎺) w for v, and 19 (𐎮) t for d, and both of these may be almost allowed to him as approximately correct. In the case of the first, indeed, we have already conceded it to Lassen, in consequence of the practical exclusion of w from the German language; and we have only denied it to Rawlinson because he distinguishes it from his 15 (𐎻) v.

We have not included the addition of the aspirate among the number of errors, where it indicated only an unimportant modification of a correct sound. It had its origin in the difficulty that was found in believing that there could be more than one sign in the alphabet to express precisely the same sound. We have seen that Hincks had just shown that these signs do in fact express the same sound, and that their employment depends solely upon the vowel that follows. After Major Rawlinson’s first alphabet was in print, he arrived, independently, as we have already stated, at precisely the same conclusion. He had long been struck with the peculiarity that certain consonants are only to be found followed by a particular vowel, and in his first alphabet he indicated five letters thus distinguished. These were: 19, t with i; 29, m with i; 33, m with u; 40, r with u; 43, n with u; and he observed especially the affinity the vowel i had for certain consonants—a peculiarity he noticed also in some of the Scythic languages.[566] When once his attention was directed to these facts it was not long before he set himself to account for them. One of the most useful contributions to decipherment made by Lassen arose from the suggestion that an a is understood though not always expressed after a consonant, when not followed by another vowel. Indeed until this idea occurred to him the result of decipherment was the apparition of a long series of words consisting of an agglomeration of consonants which no living tongue could pronounce. The next step to be made, resulted from the observation that some letters were always followed by i and others by u. A laborious classification of each letter according as it was followed by each of these vowels was therefore undertaken, and the result was sufficiently remarkable. It showed that in two cases in the grade of sonants (d and m) there was a different sign according as the letter was followed by a, i, or u. Conversely, there were three cases in the grade of aspirates (th, y, sh), where the same sign might be found before any of the three vowels; and finally there were several cases in the grade of surds (k, t, and r), where it was noticed that the same sign was followed by either a or i, and that a different sign was used before u. Taking these facts into consideration, Rawlinson thought he observed sufficient regularity to justify him in formulating the general law that, for some unexplained reason, the grade of surds in each class were expressed by two signs, one used before a and i, the other before u; the grade of aspirates by one sign only, equally available before any of the three vowels; and the grade of sonants by three signs, each applied to one vowel only. He admitted that there were numerous exceptions to the rule; indeed, the class of dentals is the only one where the series is complete, but the exceptions he was inclined to attribute chiefly to the incompleteness of the alphabet.[567]

When the letters of Rawlinson’s original alphabet were distributed into the various classes of gutturals, palatals, and so on, and among the subdivisions of surds, aspirates, and sonants, they were found sufficient to suggest the existence of some such law in the cases that have been named. With Holtzmann’s correction of 19 (𐎮) from t to d, he had the three d’s required to complete his sonants of the dental class. His own list gave him the three m’s required for the sonants of the nasal class; and he already knew that one was used only before i, and the other only before u. He knew also that the aspirates of three of the classes were to be found indifferently before any vowel. In the case of the surds, he had found that five of them (k 𐎣, ch 𐎨, t 𐎫, n 𐎴, r 𐎼) are always to be found before either an a or an i. He knew also that his second sign for n 𐎵 (43) and his second sign for r 𐎽 were only to be found before u; and he observed that the signs he still read (𐎤) kh and (𐎬) th were also only found before u. It required, therefore, no great effort to deprive them of their h, and to range them with the others as the second signs in the surd grades for k and t.

Once the existence of this law was inferred, Rawlinson was led to make other modifications in his original alphabet, in order to bring it into strict conformity, and in every instance the alteration has been confirmed. The following Table shows the distribution of the letters into the various classes and grades, and the modifications they underwent. When they fail to comply with the supposed law, the deficiencies are left blank: when they violate it, the offence is marked by ‘!’.

Rawlinson’s Alphabet, after August 25, 1846[568]

One of the results of the classification of the consonants according to the vowels that follow them was to introduce a considerable change in the method of transliteration. It is only in exceptional cases that the a, following a consonant, is found in the text, but its inherence is inferred. When, therefore, an i or u immediately follows a consonant with an inherent a, instead of transliterating as formerly such a group as 𐎥 𐎡 gi and 𐎥 𐎢 gu, they are now written gai and gau; and this modification has materially assisted the explanation of the words in which such combinations occur. An interesting proof of the accuracy of this system is afforded by the word ‘Kurus,’ which was so long an object of contention. The genitive is denoted by the insertion of an a—‘Kuraus.’ The letter 𐎽 (r), which is followed by u, is used for the nominative ‘Kurus’; but the letter 𐎼 (r), which has an inherent a is substituted in the genitive. Thus:

The interchange of these two signs in ‘Kurus’ finally disposed of a supposition started by Burnouf and supported so lately as Holtzmann,[569] that 𐎽 might be l, and that the Persians pronounced the name of their great king ‘Kulus.’

Considerations of the same kind greatly assisted Rawlinson in rectifying some of his values. For example, in the case of 𐎦, which he had hitherto read gh, he found that it replaced 𐎥 the g before a, in order to form the locative singular ‘Margauw;’ and consequently he had no hesitation ‘in placing the two characters, not merely in the same class, but in the same grade of that class.’ It is therefore now found among the gutturals as g before u.[570]

Precisely the same considerations led to the rectification of 28 (𐎩) j’h to j. He found the locative of Susiana (‘uwajaiya’) written 𐎩 · 𐎡, i.e. ji, which he knew from grammatical considerations must stand for jai; and he therefore concluded that 𐎩 is j before a; which is correct.

A farther result of this classification is to supply the sounds of the missing vowels e and o, for when i or u follow a consonant with an inherent a, the diphthongs ai and au are produced, which correspond phonetically and grammatically to the diphthongs ê and ô in Sanscrit.[571] Such were the results communicated by Rawlinson in his Supplementary Note, which, as we have seen, crossed the detailed account of Hincks’s paper on the same subject. It was received in London on October 8, and its substance was read at a meeting of the Society on December 6, and noticed in the ‘Athenæum’ of December 19.[572] The alterations in the method of transliteration required by the new system were, as we have said, carried out under the supervision of Mr. Norris.[573]

Rawlinson has not drawn up a formal grammar of Old Persian, but he loses no opportunity of comparing its forms with Sanscrit and Zend, and pointing out wherein they agree and wherein they differ. He shows that the initial letter a, so frequently employed, is used to express the temporal augment in the past tenses of verbs, and according to the analogy of the Sanscrit it is short. But the short a of Sanscrit terminations is changed into long a in the cuneiform; and the mute terminal consonants of the former are usually omitted, as in the endings ‘as,’ ‘at,’ ‘an,’ ‘am,’ a rule applicable to both nouns and verbs. He shows also that the suffixes in i, so common in Sanscrit and Zend, are all lengthened into ‘iya’; a rule also applicable to the terminal u.[574] If he had finished his chapter on the Vocabulary, the student might have been able from it to put together a complete grammar. Under their initial letters we find ‘adam,’ the personal pronoun ‘ego’; ‘aniya,’ ‘alius’; the two demonstrative pronouns ‘ava,’ ‘that,’ and ‘iyam,’ ‘this’;[575] correctly traced through all their cases so far as they were known—and the same is done for the cases of the verbs ‘am‘iy,’ ‘I am,’ and ‘thah,’ ‘to say’;[576] and in each he shows the close similarity they exhibit to Sanscrit and Zend. In his notes to the translation he dwells especially on the construction of the sentences and upon the historical questions raised by the subject-matter of the text. He gives an elaborate analysis of each letter of the cuneiform alphabet, comparing its use and pronunciation with those of other languages. He can draw a wealth of illustration at pleasure from the kindred languages of Sanscrit, Zend, Pehlevi, Persian, as well as from Pali, Devanagari, Hebrew, Chaldee, Syriac, and Turkish; and although he displays an amount of knowledge that is truly surprising, he defers with unaffected humility to ‘the more experienced philologist.’[577] He classifies Old Persian as belonging to ‘the Arian type, resembling Sanscrit very closely in its grammatical structure; but in its orthographical development more nearly approximating to the Zend: while in the peculiarity of organisation which requires the juxtaposition of certain consonants with certain vowels it exhibits something of a Scythic character.’[578] He was not disposed to admit the antiquity then beginning to be claimed for Zend. He thought that in comparison with Old Persian it was modern. He imagined that the latter became gradually extinct after the age of Alexander, and that it was succeeded by Zend and Pehlevi, the former as a hieratic and the other as a demotic language but both derived from it.[579] He was clear at least that Old Persian could never have descended from Zend, though he reluctantly admits the possibility of their contemporary existence. His unwillingness to allow the antiquity of Zend was due in great measure to the legendary character of the Zendavesta, a book which he considered could not have been written till after the cuneiform Persian had been entirely forgotten. Otherwise, he said, ‘the priesthood could neither have had the audacity nor the desire to darken authentic history by the distorted and incomplete allusions to Jemshid and the Kayanian monarchs which are found in the Vendidád Sadé and in the ancient hymns.’[580]

The appearance of Rawlinson’s work was received with feelings of enthusiasm in Germany. The reproach that England had hitherto neglected the cuneiform records was at length effaced, and in such a manner as to entitle her to claim the first place in the roll of discovery. Benfey declared that few, if any, of the contributions made in recent times in the field of Oriental research could compare with it in importance.[581] Major Rawlinson, he says, displays an extraordinary aptitude for decipherment, and an accuracy and depth of philological learning that render it peculiarly fortunate that such an important document should have fallen into his hands. In mere length, the inscription exceeds by more than a hundred lines all those published by Lassen put together. It consists of five columns of about four hundred and ten lines, and, although there is considerable repetition, it nevertheless offers a great variety of words and phrases which added immensely to the knowledge previously acquired.[582] One great difficulty with which he had to grapple was the very imperfect state of the text. A glance over the plates will show the numerous blanks left in the writing in consequence of injury to the rock. Rawlinson’s copy was so carefully executed that he committed only one serious error, the omission of a line in the fourth column. The other imperfections are due entirely to the ravages of time. In the first column there is a large fissure on the right hand, extending from the top to the twenty-fifth line, and again from the sixty-third line to the end, besides numerous occasional gaps elsewhere. But the second column is in a much worse condition. ‘A fissure, varying in breadth, caused by the percolation of water, bisects it and destroys the continuity of the writing throughout its whole extent.’[583] The third is nearly perfect, except at the bottom, where several lines are wholly lost. The fourth column is worse than the second; ‘a fissure transects the tablet longitudinally,’ and in the lower half ‘the rock is more or less broken by the trickling of the water.’ But when we come to the fifth, we find ‘a state of such deplorable mutilation that it would be waste of time and ingenuity to undertake an analysis of the text, or to attempt anything like a connected and intelligible translation.’ In the face of these difficulties he was obliged to have recourse to very elaborate and ingenious restorations. At the end of the second column, for instance, he found the Susian copy perfect, and this enabled him to ‘restore’ the Persian text. It is one of the first instances of a long translation from the Susian, and his version of it turned out afterwards to be correct.[584] In the numerous repetitions that occur so frequently, he found a safe guide in other passages of the inscriptions. Sometimes, however, he had to work on much less solid foundation, as when he sought help from other sentences that were only ‘of nearly similar construction’; or when his restoration was ‘generally borne out by the context’; or merely by considerations of ‘grammatical propriety.’ In such cases he could never arrive at more than a high probability. He had frequently to measure the length of a blank and then tax his memory to supply one or more words with the required number of letters that would fit into the vacant place and at the same time make sense. Sometimes, as in the fourth column, the sense was so obscure that he feared his ‘restorations will be considered rather bold than felicitous.’[585] Occasionally his courage failed him altogether, and he was obliged to confess that ‘I cannot restore the [passage] even conjecturally.’ It is remarkable how uniformly successful his ‘conjectural restorations’ were found to be. He imposed the most admirable restraint upon the intuitive faculty with which he was so eminently gifted; and his emendations exhibit a patience and sobriety that many scholars engaged in similar work might advantageously study. When he had surmounted the imperfections of the text so far as possible, he set himself to the task of translation, and achieved the most notable success in this department of literature. When we consider that he had to unravel the intricacies of long sentences, determine the grammatical relations of multitudes of new words and fix their meaning by a patient comparison with Zend or Sanscrit analogies, the unfailing divination he displays is absolutely marvellous. A careful comparison of this first translation with that now accepted as correct will show comparatively few alterations, although the labours of many scholars have since been devoted to a rigorous study of the same text. The main body of the translation remains the same, word for word, down to the minutest particulars. Some doubtful passages, concerning which Rawlinson himself entertained doubts, have been cleared up; but it rarely happens even in these that the original translation was in fault as to the general meaning.

The long list of commentators begins with Benfey, whose tract on the subject was sent to the press in January 1847, when he could only have seen the first part of Rawlinson’s work. Some of Benfey’s suggestions have been accepted, others definitely rejected. A few instances will illustrate the nature and extent of the earliest attempt at revision. Almost the first error occurs in Column I. par. 10, where we find: ‘The troubles of the state ceased which Bardius excited.’ Rawlinson warns us that ‘this sentence cannot be read with any certainty,’ on account of a blank in the inscription, and the doubt attaching to the word ‘azada.’ Benfey derives this word from the Sanscrit ‘ajatá,’ which signifies ‘deprived of children,’ and he translates the passage: ‘The kingdom was deprived of heirs because Bardius was killed.’[586] But this alteration turned out to be quite unauthorised. ‘Azda,’ as the word is now written, is identified with the Armenian ‘azd,’ ‘information,’ corresponding to the Sanscrit ‘addhâ,’ ‘certain’; and Rawlinson himself corrected the passage in 1873 to ‘It was not known to the state that Bardius was killed.’ The accepted version given by Spiegel is ‘The army had no information that Bardius was killed.’[587] Soon after, we are told that ‘Cambyses, unable to endure his [misfortunes], died’ (par. 11). Benfey attributes his death to ‘overwhelming anger’ (übergrossem Zorn); but Oppert showed that the word really means ‘suicide.’ In 1873, Rawlinson had made the rectification himself.[588] Not more happy was the substitution of ‘Liebet mich’ for the often repeated exclamation Rawlinson renders ‘Hail to thee!’ and which properly signifies ‘Go forth!’ as Rawlinson said in 1873.

There are, however, some instances where the commentator makes a useful correction. The erroneous reading, ‘he would frequently address the state’ (par. 13), he altered to ‘er möchte das Reich mit Macht vernichten,’ which approaches Spiegel, ‘er möchte viele Leute tödten,’ or, as Rawlinson said in 1873, ‘he slew many people.’ So also the passage ‘afterwards Dadarses remained away from me in the field’ (Col. II. par. 9), is improved to ‘dann erwartete mich dem Befehl gemäss Dadarses.’ This Rawlinson changed in 1873 to ‘afterwards Dadarses waited for me there,’ the accepted version being ‘dort erwartete mich Dadarses so lange’ etc. It will be observed that the critic was himself liable to fall into superfluous additions not far removed from error. Sometimes, however, he avoided this trap, as when he substituted ‘he lied’ (Col. IV. par. 1) for Rawlinson’s ‘impostor,’ and ‘if you so think’ (par. 5) for ‘if it should be thus kept up.’ But in a large proportion of cases Benfey follows his leader into error with perfect complacency. In one or two places Rawlinson is obliged to confess that the difficulties are so great that translation is almost impossible. One of these occurs in the description of the religious reform after the death of Gomates, the Magian (Col. I. par. 14). Referring to this passage, he says: ‘Of several of the most important words the orthography is doubtful; of others the etymology is almost impenetrable, and the construction, moreover, in some parts renders the division into sentences a matter of serious embarrassment.’[589] But the difficulties have been found almost as insuperable by his successors. They have not been able to derive assistance from the Susian and Babylonian columns, where the difficulties are even greater. The meaning of some words still remains a mystery, and even the general drift of the passage is open to discussion. Spiegel[590] warns his readers that the explanation of the whole paragraph is still so uncertain that no opinion as to the religious history of the times can be prudently based upon it. But such passages as these are fortunately of rare occurrence; and from the moment of Rawlinson’s publication the contents of the inscription were known as thoroughly as they are at the present day. The careful study of two generations of scholars has changed a word here and a word there, and cleared up the meaning of a few doubtful passages, but in all substantial respects the translation remains unaltered.

Rawlinson’s revised translation of the inscriptions published two years before by Lassen forms the concluding chapter of his Memoir. We have already had occasion to contrast the merits and demerits of the two translators. Rawlinson’s transliteration presents an entirely modern appearance in consequence of the correctness of the values of so many signs. There are still a few errors, owing chiefly to the prolongation of the final syllable in such words as ‘thatiya,’ ‘tyaia,’ and others, an error already signalled by Hincks. But these are of small importance. As regards the translations, he brought from the study of the Behistun record a knowledge of the language that no one else then possessed, and he was able at once to resolve difficulties that had baffled all previous attempts. Such expressions as ‘generosus [sum]’ finally disappeared for the correct rendering ‘he says.’ Sentences that were hitherto entirely misapprehended now appear in their correct form. ‘The son of Arcis’ gives place to ‘the son of a Persian, an Arian of Arian descent.’ The last paragraph of the same inscription is satisfactorily explained. Even where he fell short of success, as in the end of the I inscription, he made important contributions to the elucidation of intricate passages.

Rawlinson added the inscription on the Venice vase, not known to Lassen. It had been recently published by Longpérier in the ‘Revue Archéologique’ (1844), who thought it should be referred to Artaxerxes I. Rawlinson translated it ‘Artaxerxes the King,’ and assigned it to Artaxerxes Ochus. Opinion has since been divided upon the subject; Spiegel and Menant follow Longpérier; Oppert and Weissbach follow Rawlinson.

Very few additional inscriptions in Old Persian have been brought to light since the date of Rawlinson’s Memoir. But a good deal of labour has been spent in clearing up the doubtful passages in those already known. In some cases the text was, as we have said, so much mutilated as to defy intelligible translation. This was the case with the fifth column of the Behistun; and Rawlinson thought it best to omit it altogether from his revised edition of 1873.[591] This diffidence, however, may have stimulated M. Oppert to attempt a restoration of the text. We have already described the process followed by Rawlinson. It consisted in selecting a word or words containing the number of letters required to fit into the space left vacant by the erasure of the text, and which would at the same time make some kind of sense. It is obvious how much of the success of this operation will depend upon the ingenuity of the restorer; and still more upon the restraint he exercises over his imagination. No one is more distinguished than M. Oppert for the ingenuity of his conjectural restorations; and the column that Rawlinson abandoned as hopeless appears in Oppert’s edition as the ‘Complementary Behistun Text,’ and in a comparatively perfect condition. In this case he could receive no assistance from the translations, because both the other columns are destroyed; and not much by comparison, for there are few parallel passages. His work is, therefore, the more admirable as a display of the imaginative faculty. But his reading has not been accepted by Menant or Spiegel, or by Weissbach.[592] No doubt we owe the ‘Testament of Darius at Naksh-i-Rustam’ to the same method which appears to have been exercised upon the few lines copied by Westergaard from the long inscription below the one known and which Spiegel declares it is clearly impossible to translate.[593] One of the three short inscriptions engraved above the figures of the Tomb has also given rise to some discussion. Mr. Tasker, an English traveller, sent them to Rawlinson, by whom they were published in the ‘Journal of the Asiatic Society,’ and an improved version appeared some time afterwards.[594] Even at that period considerable difference of opinion existed as to the meaning of at least one of these short legends. According to Rawlinson, Norris, and Oppert, Aspathines is the ‘keeper of the arrows’ of King Darius; but Norris adds ‘chamberlain’ and Oppert ‘quiver-bearer.’ Elsewhere Oppert translates it quite differently: ‘Aspathines, minister of King Darius, who makes the law observed.’ Spiegel’s version is: ‘Aspacanâ des Königs Darius genosse, Zügelhalter (?).’[595]

In the autumn of 1846, when Rawlinson’s ‘Memoir’ was passing through the press, Dr. Frederick Hitzig published a tract on the Persian Text of the Tomb Inscription of Darius; but his translation shows little advance beyond the point reached by Lassen.[596] The edition of the Persian inscriptions by Theodore Benfey made its appearance early in 1847. His transliteration suffers from having been made before the method of writing explained by Hincks and Rawlinson in the previous year had become generally known. We have already sufficiently noticed Benfey’s translations, which show considerable improvement on those of Lassen, but fall far short of the comparative excellence obtained by Rawlinson. In the course of the summer, another writer appeared whose name has been already mentioned and who was destined to occupy a very important place in the future history of Cuneiform Research. M. Oppert was born in Hamburg in 1825, and studied successively at Heidelberg, Bonn and Berlin. At Bonn he was a pupil of Lassen, and to this circumstance probably we owe his early interest in cuneiform and the publication of a tract on the subject at the age of twenty-two. The promising youth was precluded by his Jewish faith from holding a professorship in Germany, and consequently he went to Paris in 1847. His later writings have all appeared in the French language, and for this reason he is generally included among the number of French scholars, of whom, till quite recent years, he was by far the ablest representative.[597] So far as we are aware, the pamphlet on the ‘Lautsystem des Altpersischen’ was his first effort in this department of knowledge.[598] In it he explains the principle that regulates the employment of the consonants in substantially the same manner as Hincks and Rawlinson had done in the previous year. He had evidently arrived at his conclusion independently, and it is remarkable that he was not better informed of the progress that had recently been made. It might be supposed that the attention of everyone interested in the subject would be directed to the appearance of Rawlinson’s ‘Memoir,’ which was then eagerly expected. It was in fact published before May 1847, and Oppert’s tract was not sent to press till July.

With the publication of Rawlinson’s Memoir, in 1846-7, the decipherment of the Persian inscriptions may be considered accomplished. In 1850, he could write that there ‘are probably not more than twenty words in the whole range of the Persian cuneiform records upon the meaning, grammatical condition or etymology of which any doubt or difference of opinion can at present be said to exist.’[599] The value of his own contribution to the general result received the fullest recognition. Professor Max Müller declared to Canon Rawlinson that, ‘thanks mainly to your brother, we have now as complete a knowledge of the grammar, construction and general character of the ancient Persian language as we have of Latin.’[600] He was greeted as the Champollion of the new decipherment, a position he has retained to a large extent in Germany and France. So late as 1895, M. Oppert found occasion to remark that, ‘after Rawlinson it was only possible for other scholars to obtain gleanings in the field of Persian cuneiform interpretation.’[601] In his own country, however, he seems to have suffered for a time from the singular affectation that was so long in fashion, of looking to Germany alone for all the springs of knowledge. As a matter of fact, in this department at least, few Germans, with the exception of Grotefend, made any important contribution. Rask, Lassen, Westergaard, were all Scandinavians, and it is certain that for many years Rawlinson continued to be the source whence Continental writers drew most largely; and the neglect into which he fell at home occasioned the surprise of at least one eminent Frenchman. ‘Young English and Germans,’ says M. Oppert, ‘pretend not to know him. An Englishman once told me he had never read a line of Rawlinson. I replied: “I supposed just so; if you had read him, your papers would be less imperfect than they are.”’[602] This testimony to the great services of Sir Henry Rawlinson is given by the scholar who for many years occupied by far the most prominent position among Continental writers upon cuneiform subjects, and who has himself contributed largely to the progress of the study.

In 1851, M. Oppert undertook a complete revision of the whole series of Achaemenian inscriptions. The work appeared in the ‘Journal Asiatique,’ between February 1851 and February 1852, and was afterwards published in a separate form.[603] He introduced a considerable number of alterations, in both the transliteration and translation, and, so far as we have noticed, a large proportion of them have been accepted. He criticised some of Rawlinson’s opinions, but rarely with asperity, and he generously acknowledges that the English scholar ‘a grandement mérité de l’histoire du genre humain.’[604] He subsequently revised his own translations for his book on the ‘Expédition en Mésopotamie,’ published in 1858. In the same year Rawlinson made an amended version of the Behistun Inscription for the edition of Herodotus published by his brother; and again in 1873 for the ‘Records of the Past.’[605] The other inscriptions were once more revised by Oppert in 1877 for the same useful publication.[606] But meanwhile a complete edition of the whole Persian inscriptions had been published by Spiegel in Germany in 1862,[607] and in France in 1872, by M. Menant.[608] The latest publications on the subject that have come under our notice are the valuable treatises on the languages of the first and second columns by Weissbach,[609] and of the third column by Dr. Bezold,[610] accompanied in each case by a revised text, transliteration and translation.