Journal of the Indian Archipelago and Eastern Asia/Series 1/Volume 2/A scheme for representing Malay sounds by Roman Letters
A SCHEME FOR REPRESENTING MALAYAN SOUNDS BY ROMAN LETTERS.
By John Crawfurd, Esq., F.R.S.
Sir William Jones well-known scheme for representing Asiatic sounds by Roman letters seems to me to be chiefly defective from at- tempting more than was practicable. Its basis was the Dewanagri alphabet, and the Arabic, but other Asiatic languages contain sounds not to be found in either, and such an alphabet as would comprehend the whole of them would, I am satisfied, extend to at least 100 cha- racters, and therefore be of intolerable length and prolixity.
Every language, or at least every class of languages requires, in my opinion, a system for itself, and if the native alphabet assumed as a basis be comprehensive and correct, that is, have an invariable character for every sound in the language, and such character is re- presented by an unvarying Roman letter, the task is at once accom- plished. Such a scheme necessarily excludes the use of double let- ters to represent a simple sound, both as cumbrous and superfluous.
The languages of the Archipelago, in so far as the native portion of them is concerned, contain no sounds that do not equally exist in the European languages, and which therefore, with very slight mo- difications may not be easily represented by single Roman letters. The Javanese alphabet, the most perfect of those of the Archipelago, has written characters, not only for every sound in the Javanese lan- guage, but also for all the sounds in the other languages, with a few exceptions easily provided for.
Taking, then, the Javanese alphabet as the foundation of the scheme, the consonants will be 22 in number, and as follows, b. c. d. ḋ. f. g. h. j. k. l. m. n. ṅ. ñ. p. r. s. t. ṫ. w. y. z. The letters b. k. l. m. n. p. 1. and s. having exactly the same sounds as in most of the European languages, require no remarks, except that a final k., and sometimes a. medial, are, in Malay and sometimes in Javanese, pronounced as if they were mere aspirates.
The letter c. being redundant in the English system, and in it or the other languages, having, according to its position, two or more different sounds, I have selected it to represent a consonant for which there is a written character in nearly all the eastern alphabets. The sound is also of frequent occurrence in the European languages, but aukwardly represented by two or three consonants combined, and which do not contain the elements of the sound. The English and Spaniards represent it by ch, the French by tch, and the Dutch by tj. The Italians represent it before the vowels e. and i. by the letter c., and I have adopted this to represent the sound universally, placing au orthographic mark over it, to obviate any ambiguity.
The first d. in the series above given is a dental, and the second marked by a dot is a palatal. They are carefully distinguished by the Javanese, but not by the Malays although both be exhibited in the scheme of their alphabet.
The letter f. as a native sound is unknown to the Malay and Java- nese languages, and is confined to a few of the ruder tongues. I give to the letter g. invariably, its hard sound. The letter h. represents the simple aspirate, aud, fur the most part, occurs, in so far as na- tive words are concerned, after a vowel,—in Malay and Javanese al- ways so. The letter j. represents a sound for which there is a cha- racter in all the native alphabets exactly corresponding to its Eng- lish pronunciation, but not to that of any other European nation. The Dutch represent it by dj, which do not contain the elements of the sound, besides being clumsy.
The ṅ with a dot represents a sound for which there is a charac- ter in all the languages of the Archipelago. It is the ng. of the European systems, and in the native languages is used as an initial, as well as a medial or final. After a vowel and closing a syllable it is represented in most of the alphabets by a dot over the consonant. The sound which I represent thus ñ, is taken from the Spanish or- thography in which it appears as a substantive letter. It occurs in the Spanish word España, it is the gn. of the French and occurs with us in such words as union, onion &c.
In Malay and Javanese, there are two ts., a dental and a palatal, the first being of most frequent occurrence. The Javanese, but not the Malays, carefully distinguish them. I have marked the palatal with a dot as in the case of the palatal d. The letter w. has exact- ly its English sound, and y. is our consonant of this name which the Dutch and Germans represent by j. The letter z., like f., as a na- tive sound, occurs only in a few barbarous and unwritten languages.
The liquids or consonants which coalesce with other consonants are l. r. w. and y.,—the two first very frequently, and the two last rarely. The vowels of the Javanese alphabet, and they correspond exact- ly with the vocalic sounds of the Malay language, are six in number, and with one more will suffice to express all the vowels of all the languages of the Archipelago as far as we yet know them. They will be as follow: a. á. e. é. i. o. u.
The vowel a. has no representative in the Javanese or any other alphabet of the Archipelago except as an initial, but it is understood to follow every consonant, unless supplanted by the mark of another vowel, or suppressed by a sign of elision. Its sound in the Malayan languages is exactly that which it usually has in all the European languages, except our own, as exemplified in the Italian word casa, a house. The sound á. with an acute accent is intended to repre- sent a Malayan vowel which in the Javanese alphabet has a distinct mark. It is the u. of the English words "but" or "cut".
The vowel e. has peculiar character in the Javanese, and, in- deed nearly all the other native alphabets. It has the sound which oc- curs twice in the Castilian word elevar, to raise. It is the é. in formé of the French, and the unaccented e. of the Italians in the word legge. The é. with an acute accent represents a sound, frequent in the Sunda language, and some other little cultivated tongues, but it has no character in any native alphabet. It is frequent in the Celtic dia- lects, and approaches in sound to the French u.
The sound of the native vowel represented by i. is the Italian one in the word Italia, and not the dipthong sound which we ourselves whimsically give it. The o. has a distinct character in the Javanese alphabet. It is the o. chiuso of the Italians, and the o. which occurs in our word "tone". U has also its peculiar character in the Java- nese, and indeed, in all the other native alphabets. The letter is pro- nounced as in the Italian, Spanish and German, and it is the oe of the Dutch, the ou of the Trench, and the oo of the English.
In writing native words, it is to be observed, that the vowel, ex- cept when long, and sometimes even then, is inherent in each Ara- bic consonant, as in the native systems. The same is the case with i and u at the end of a word.
The dipthongs are but three, ae. ai. and au. In the Javanese al- phabet, they have no peculiar characters to mark them, but are simply expressed by the sequence of the vowels which compose them. It may here be added that neither in the Javanese, nor any other al- phabet of the Archipelago, does there exist a distinction by express characters between long and short vowels as in the Dewanagri. They are long or short, or more correctly, accented or unaccented, according to their position in a word, and in the vast majority of words, the accent of bisyllables is on the first syllable, and of polysyllables on the penultimate. Indeed there are hardly a score of examples in the Malay or Javanese of the accent being on the last syllable.
The system thus sketched would be sufficiently complete, if the paramount language, the Malay, had not been written, as it invaria- bly is, in an Arabian character ill suited to the purpose. The Ara- bian alphabet, imperfect in itself and even for its own particular pur- pose, is preposterous when applied to the languages of the Archipe- lago, the genius of the pronunciation of which is far more at variance with that of the Arabic than of the languages of the south of Europe. It wants eight letters which the Javanese alphabet and the Malay language have viz. c̈. ḋ. g. ṅ. ñ. and ṫ, and it supplies their place by diacritical points over their respective cognates, while it has no less than 12 which no native alphabet possesses, and which, for the most part, are unpronounceable by the inhabitants of the Archipelago. The Malays cannot pronounce most of the peculiar consonants of the Arabian language, and have a repugnance even to several of the vowels. A few only, ambitious of some literary or religious distinc- tion, ape an imitation of the Arabie pronunciation, and the multi- tude reduce the Arabic to the standard of their own enunciation, as we ourselves did in adopting the Norman-French portion of our lan- guage.
The discordance which exists between the orthography and pro- nunciation of Arabian words adopted by the Malayan nations, makes it therefore necessary, in order to preserve the one and exhibit the other, to adopt a double system for every word.
I represent in the following manner the Arabic letters which have no representatives either in the languages or alphabets of the na- tions of the Archipelago ġ. ḣ. ˙k. ḱ. l̇, l̈. ṡ. s̈. ẗ. z̈. ż. I give them in the order of the Roman alphabet, classing them with their cognate letters.
The Arabian f. as I have stated before, occurs, as a native letter, only in the languages of a few of the ruder tribes. In the more cul- tivated it is almost invariably transformed into p. The ġ mark- ed by a diacritical point is the well-known Arabian guttural called ghain, and our Northumbrian r. Its native pronunciation is simply that of g. hard. The ḣ. with a dot is the strong Arabian aspirate; and to which there is no sound equivalent in the Malay languages. The k̈. thus marked is the guttural called kha. It is the sound which is of such frequent occurrence in the Celtic and most of the Teutonic languages, the German ch., the x. of the old, and the j. of the modern Spanish orthography. The k̇. with a single dot is a deep hard guttural which has no equivalent in any European language, and still less is there any approach to its sound in the Malayan tongues. Both these letters have the native pronunciation of an ordinary k.
The two letters l, represented respectively, by a single and by a double diacritical mark are the 15th and the 17th of the Arabian alphabet, called by the Persians zad and za. The Arabs, however, give to both a sound which approaches nearest to the double l in Welsh or Spanish. The Malayan nations pronounce them as a com- mon l.—Of the three s. noted by different orthographic marks, the first is the swad of the Arabs, or 14th of the Arabian series, and the second the 4th Arabian letter, which the Arabs pronounce like our th aspirated, as in the word "thing"; but the Persians like an s. The Malays give to both letters the simple hissing sound of an or- dinary s.
The third s. is the 12th letter of the Arabian series, the English sh, the French ch, the Dutch sj, and the German sch. With the Malayan nations it is, like the two letters which precede it, only a simple s.
The ẗ. thus marked to distinguish it from the native palatal ṫ, is the toe of the Arabs, but, in utterance the natives of the Archipela- go make no distinction between it, and an ordinary dental t. The z. without a diacritical mark corresponds with a sound which exists in some of the ruder languages of the Archipelago, but is un- known to the written ones now in use. It is the 11th letter of the Arabian series immediately following r, and distinguished from it only by a dot above it. The ż with a diacritical mark is the 9th of the Arabian series immediately following the letter d. and distin- guished from it only by the point. Both these letters are pronounc- ed in the same manner, and as an ordinary z. They may indeed be said to be the only peculiar letters of the Arabs that can be pronoun- ced without much effort by the Malayan nations, but even these they not unfrequently convert into a j.
The difficulty of rendering the Arabic vowels by Roman letters is still greater than of the consonants. Of the three vowels which form substantive, alif, wa and ya, the first alone invariably expresses one sound, which is that of the inherent a in the Hindu and insular al- phabets, but, as in these, it is for the most part expressed in the con sonant, at the end of words, always.
With the letter wa, the power is not only that of two vowels u. and o, but also of the consonant w. As a vowel, it may be invariably expressed by u, for the natives themselves do not always clearly dis- tinguish between the u and o.
The letter ya is not only the two vowels e and i, but also the con- sonant y. As a vowel I represent it invariably by its most frequent sound of i. In the Arabic Malayan alphabet, in order to distinguish the vowels from the consonants in the case of both letters now un- der consideration, a double letter is used at the beginning of words, which consists without distinction either of the letter, a. or the soft aspirate humza. I have no doubt that this was borrowed from the lost alphabet of the Malays, on which the Arabian scheme was form- ed, for it is analogous to the practise of the Javanese and other na- tive alphabets in which the vowel a is considered as a substantive, and its power changed by the application of the other vowel points. This native a is evidently represented by the Arabian alif aspirate which is never pronounced.
In the system adopted, therefore, the Arabic letters wa and ya will be represented, according to their position in a word, either by w. and y. as consonants, or u. and i as vowels, just as v. and y. are sometimes made to serve with ourselves.
The short vowels of the Arabic alphabet are noted by orthogra- phic marks, three in number, which represent, respectively, the sound which in the native alphabet I have given as á,—that of e. or i., and that of o. or u. But these orthographic signs are hardly ever written, so that it is impossible to tell which of the vowels is to be expressed. I shall represent all of them by the vowel a short, hav- ing the true sound as pronounced by the natives of the Archipelago as is done in dictionaries in the oriental character.
There still remains one letter of the Arabic alphabet to be re- presented, the peculiar guttural vowel called ain. It is a substantive letter of the alphabet, and, according as the supplementary vowel marks are applied, it is a. e. i. or u. I distinguish it by a grave accent, leaving the sound for the explanatory word. The natives treat it as an ordinary vowel without any attempt at a guttural pro- nunciation.
The whole of the Roman letters necessary to represent the Ara- bico-Malayan alphabet will be 40 in number, and as follow a. á. ä. à. b. c̈. d. ḋ. e. é. f. g. ġ. h. i. j. k. k̇. k̈. l. l̇. l̈. m. n. ṅ. ñ. o. p. r. s. ṡ. s̈. t. ṫ. ẗ. u. w. y. z. ż. It is fortunate, however, that this cumbrous alphabet will not be very often called into use, for the actual number of Arabie words in the ordinary written language is not ve- ry considerable, and to the oral language, in which they are trifling, it has no reference.
A very few examples will show how much better adapted the Roman letters are than the Arabic to the expression of the native sounds of the languages of the Archipelago. The Malay words bawa, to bring, and bau, odour, are in the latter represented by the very same let- ters. So are bunga a flower, and buang, to throw away, so are arti meaning, and arta, goods, both Sanskrit words, the word bágánda, prince or highness, is written bgnd, that is without any vowel at all, although it contains three.
I conclude with a few examples of the changes which Arabic words undergo in Malayan pronunciation. Kábár, news, becomes ka- bar, fákár, to think, pikir, sábáb, cause, sabab in Malay and sawab in Javanese, rálá leave or permission, lila, and wakt, time, waktu, fá- luli, to meddle, paduli. The more the language abounds in vowels the greater the havoc committed. Májád, a mosk, become in Bugis, masigi, sálám, peace, solon, bárkát, a blessing, baraka, and kártás paper, karotusa. The Arabic word wákt time, had been pretty well mutilated by the Malays and Javanese when they reduced it to waktu, but the Bugis thoroughly disfigure it when they make it watüe.