a

is, the so-called historical method records a change of sound up to a certain point, and then gives up all further record of change. To take an example: the English word deign is derived from the Latin dignor, or some proviucial form of this word, through the French daigner. Well, the g in deign, like an atrophied member in the animal body, represents the Latin g in diguor, and is in so far source of intense gratification to men of the Trench type of mind. Now, the Latin dignor must have changed its sound in the mouths of the corrupt-Latin-speakers of Gaul, before it could be written daigner instead of dignor. Prima facie, daigner must have been a phonetic representation of a spoken word. Ai in the French of the present day has the sound of é. It is simply impossible that those who first reduced the word daigner to writing contrived a combination of the vowels a and i to represent a sound which could be represented by an already existing distinct symbol e. Ai in daigner must therefore have origi nally had a diphthongal sound, compounded of a and i, and the diphthongal sound must have subsequently been shortened and modified into e, as in what seems to be a parallel case in Sanskrit compounds like Râmesvara (Râma and isvara). The gn, or rather the after the g which wholly lost its sound) further came to bear a peculiar nasal sound which is no way a compound of gand n, but is akin to the Devanagari. After the spelling of the word daigner became fixed, however, although the sound changed, there was a natural tendency to stick to the spelling that had become fixed. To be able to spell the word in the old way became a mark of distinction, a line of demarcation between the well and the ill-educated. Thus, there came to be an arrested historical spelling. Had the spelling gone on adjusting itself to the changing sound, there would have been a system of spelling truly historical. Phonetic writing, therefore, would be only more historical, than that which is now spuriously called such, at the same time that it would be a blessing to mankind by enabling people to acquire spelling almost without effort, and thus sparing their brain power for other acquisitions. An innovation of so radical a character as a change of a nation's alphabet, should, we think, be the proper occasion for sweeping away a vicious system of orthography.

To return now to the subject of a universal alphabet. Looking to existing facts, it seems quite clear, that a universal alphabet must be one based on the Roman. The Roman alphabet has certain inherent merits of its own, but what is of far more importance than this is the fact that all Western-Europe--the chief seat of science, learning, and industry-uses this alphabet (the German alphabet being sul stantially the same as the Roman); and all

America (destined to become hereafter the most populous Quarter of the Globe) and the rising English-speaking communities in Australasia and South Africa use it, too. The Greek or Greckderived alphabets current in the eastern half of Europe (with the exception, and that partial only, of the circumscribed territory now left to the Turks), and in the vast, though now very sparsely peopled Russian dominions in Asia, do not differ very widely again from the Roman alphabet. Every thing, therefore, points to the Roman alphabet, with necessary modifications, ultimately superseding all other forms of writing Signs are not wanting now, that the only highly cultivated nation that does not use the Roman alphabet, viz., the Germans, will in no long time, abandon the caricatures of Roman letters they now use, for the Roman letters themselves.

In making the Roman the basis of a universal alphabet, however, there can be no reason why its patent defects should be cherished and perpetuated. Reason and human happiness demand that its deficiencies should be made good by supplementary letters, in the case of languages whose sounds it cannot adequately represent, and that the superfluous symbols it has be rejected or otherwise utilised. Such obvious defects again as the existing divergence between capital letters and small letters, and between printed and script letters, ought to be got rid of. We fully recognise, with Professor Monier Williams, the utility of a contrivance which enables us to make a distinction between smith and Smith-brown and Brown-bath and Bath. What we maintain, however, is, that the difference between capital and small letters need not be wider than that between s in smith and S in Smith. Such wholly different-looking characters as b and B for one and the same sound are certainly not conducive to mental economy. In choosing between the forms of capital and small letters, there can be little hesitation which to throw overboard. The very difficulty of writing capitals medially or finally led, it appears, to the invention of small letters.+ Small letters are so much simpler in form, and so much more largely employed than capital letters, that to give up the latter would certainly be to work along the line of least resistance. Capital letters have the advantage in respect of symmetry over small letters, in that they are all of the same height. But superior symmetry may here well be sacrificed for larger ends. The present wide difference between printed and script letters

• Preface to Sanskrit English Dictionary, p. 18.

†The initial and medial forms

of Semitic characters appear to have owed their birth to a similar necessity.

may be reconciled. In Italics we have the connecting link between the two sets of symbols. In printing, or in the current hand, the letters need not, however, be slaut, as the Italic characters are. Erect letters, shaped like Italics, would effect a full reconciliation between printing and current-haud writing; while Italic characters in their present slaut forms could be reserved for the purposes they now serve.

The haphazard arrangement of letters in the Roman alphabet, though pre-eminently historical, for its origin can be traced back to even the primitive system of hieroglyphic writing, ought likewise to be abandoned for something like the scientific arrangement of the Devanagari alphabet. The letters of the alphabet again should be named after some uniform system like that which obtains in Devanagari, and not certainly in the unsystematic English way, which in this respect contrasts very unfavorably with that of the rival nation across the Channel.

Sir William Jones initiated the method of transliteration which, with various modifications, is now employed in representing the sounds of the Sanskrit and other oriental languages. Professor Lepsius of Berlin has put forth a more ambitious system of his own which aims at being a standard universal alphabet. Missionary alphabets have been formed for representing the sounds of divers languages on the method of more than one scholar.

The adoption by common agreement of one uniform method of modifying some of the letters of the Roman alphabet for representing sounds wanting in the Latin language must be a work of time, and must require the co-operation of savans of different nationalities, assisted by competent natives of countries that are too backward yet to have savans. The Germans and the French represent the same sound by ö and eu, respectively. It is only by common concert between Germans and Frenchmen that some common modification (say—ö, or ë, of a Roman vowel letter could be made to represent the same sound in both German and French, Concert with other nations using the Roman alphabet would also be necessary in order that the same modified Roman character might not be employed to represent a different sound in some language spoken by any of those other nations.

The present writer is not ambitious of propounding such a universal scheme for modifying some of the Roman characters of adding, where necessary, to their number-a task for which he kucws he is unqualified. He wants only to throw out a few suggestions for securing uniformity and accuracy in the system

* Vide Trübner's Grammatography, p. 8.

of transliteration which has, since Sir William Jones's time, been so successfully applied to Sanskrit and other Indian languages. Although he does not believe that the time has yet come for replacing the Indian alphabets by the Roman, yet the Romanisation of Indian proper names is a present necessity; and orientalists likewise do Romanise. Some tentative method of Romanisation, therefore, appears to be necessary even now.

Sanskrit is properly written in the Devanagari character, and the Devanagari system of writing is apparently phonetic. The system of transliterating Devanagari letters now generally adopted, appears to be, on the whole, very well devised. A few defects, however, call for remark. should not by any means be represented by ch, for stands for a simple sound and not for any combination of any sound with h. Professor Monier Williams represents by c'. C standing for either k or s, however, would have no distinct function to perform, and would therefore be wholly superfluous. It would be best then to turn it to account by making it the equivalent of, as indeed it has already been partially made in modern Italian. The transliteration of and, either by sh both or by sh aud sh' respectively, is to be reprobated for the reasons urged above in regard to . A simple sound like that of should have a single letter to represent it. S with a श dot below or s or ç for T, and s with a dot above for, would answer very fairly. Ch and sh, it may further be stated, are purely English conventions, and cannot be acceptable to continental scholars. Modification by dots above or below a letter is a thing quite familiar to Indian populations, and is therefore to be preferred to accent marks, which have long had another function ssigued them. It can only cause confusion to press them into other service. On abstract grounds, independent characters would in all cases be preferable, indeed, to any dotted variations of letters. But agreement about characters to be newly coined would be harder to arrive at than about the employment of dots or other marks; and dots have been in familiar use in India and all Muhammadan countries, and been found to answer in a way. If by a concert among civilised Governments some congress of scholars and scientists were to lay down one uniform system of writing for the civilised world, the adoption of some of the existing characters of uou-Roman alphabets and their adaptation to the Roman system of writing might in some cases be preferable * The absence of proper types ics, &c., necessary. renders these shifts and that of Ital

to dotting, or otherwise marking, Roman letters. But till such time comes, dotting would, the present writer thinks, be the provisional arrangement that could be most easily applied. Whatever be the arrangement adopted, it is certainly desirable that that arrangement should universally prevail.

The semi-vowels ri ́and rí, lri and li, and the Vedic L should be represented after some uniform method. and should likewise be represented, each by an appropriate symbol. One remark on the system of transliteration advocated by Professor Williams may not be inappropriate here. That every second and fourth letter of the five groups of Devanagari consonants is unnecessary, as representing a sound which is compounded of that of the next preceding letter and of h, was a very sound theory of Sir William Jones, who practically acted upon it in devising his system of Romanisation. Long before Sir William Jones set foot on the soil of India, however, the Muhammadan conquerors of the country had, in writing Hindustani in the Persian character, followed precisely the same method. In all likelihood, it was the Urdu system of writing that suggested the idea to Sir William Jones's mind. Professor Williams, from a misapprehension, as we conceive, of the character of the aspirate sounds, advocates the transliteration of by k', &c., and not by kh, &c. Mr. Beames takes the same view of the sounds of , &c. He says: "The aspirates, it must however be remembered, are never considered as mere combinations of an ordinary letter with h. It is quite a European idea so to treat of them; kh is not a k-sound followed by an h, it is a k uttered with a greater effort of breath than ordinary. The native name for these aspirates is mahaprana, 'great breath,' as opposed to alpaprana, little breath' letters. The European method of speaking is used in this section as being likely to be more familiar to the reader; but it must ever be borne in mind that the aspirate is uttered by one action of the mouth; there is not the slightest pause or stop between the k and the h; in fact, no native ever imagines that there is a k or an h either in the sound. The difference between 'eat' and 'cause to say' is extremely well marked, even in the most rapid speaking." That the idea of aspirates, being compound sounds is not quite a European idea, is conclusively proved by the shifts employed in Urdu for representing the sounds of,, &c. That

Comparative Grammar of the pp. 264-265.

AYS anguages of ludia, Vol I,