38 SPELLING

{b) The idshdid is inserted to indicate that the letter over which it is placed, whether vowel or consonant, is sounded twice, e.g. u as uw, i as iy, s as ss, ng as ngng, and so on.

cdIj buwat, jj duwa, Jj diyam, (J'l iya, iu-usiyang, sj^ suddah,^I!j bessar, ^k:i tengngah.

But the use of the tashdid to double the consonant, i.e. when it follows a short vowel (as in our words better^ fellow, galant or gallant') is not generally observed in Malay I\ISS. and is not Arabic, nor does it represent any real phonetic doubling of the consonant. (One is tempted to compare with it the double kk in such words as ^^^^is^ilij pertunjokkan, which are often said to be in imitation of Javanese double-letter forms and to exhibit a survival probably due to the existence of two forms of k ; but Javanese influence hardly existed at all in Acheen, and it is more probable that the first k j never represented anything but the mere glottal check (§9).)

[c) The omission of final \ j and ^ :

(j^ kayu, e)l aku, J^^j' terlalu, c:>.;i pintii, uj^J;-^ meribu, -yj bertemu, i-Jl api, oU jadi, ^L mati, d^ laki, ^.9 pergi, ^y bumi, e^U hati, JiC-. sa-kali, dj-suka, y dua, s^i tuba.

Exceptions :

(i) A final vowel is inserted when required for the explanation of a foreign word. In the Bodleian Sn' Rama the Sanskrit puri is written sometimes .ys but oftener i^.f and pei'kasa is written \j6j.

(2) It is inserted when 1 represents a uniform long a, and J , {J the diphthong sounds au, ai.

\S kera, \\ bela, lb depa, ^S kilau, y^, risau, ^y tupai, f\jt helai.

{d) The omission of an\' S}'mbol to represent the indeterminate short vowel e unless the iashdid as used above {U) can be taken as an illegitimate symbol.

SPELLING 39

(e) The omission of 1, ^ and ij as medial vowels in closed syllables.

cijyj berat, ^.b belum, c:*.*.*. sebut, c^^l orang, A-..»5 kambing, oXc^ ringgit, A.^_ bimbang, !u*) tumboh.

Exceptions :

(i) When the spelling of Arabic words is retained (though Arabic pronunciation may not be followed), e.g. ^)kJ\ islam, ^^i.9 fakir, t_>Uj kitab — in Arabic of course one would get kitabu, fakiru, &c., and the syllables would not be closed.

(2) The monosyllables ^jb dan and ^ys pun.

(/') 1, J and 1^ are inserted in an open syllable upon which the accent falls — in Malay usually the penultimate.

^y ttian, ^_^j5. riias, A^.-ft.^ kepiting, UU.j telinga, j:>^jS^ sekedudok, ^c-\J larangan, ^^cLj tambangan, _i^».. raja-raja, J^.J1 apa-bila, klLU hulubalang.

Exceptions :

A few words like dL« maka, si pada,,j dari, 5j« siidah, JX^ segala, in which the vowel may have been omitted because it was short and had little stress on it.

[In certain derivative words formed by the addition of suffixes, 1, . and ij are shifted to the penultimate of the derived word: oi;l ingat, ^jAiA ingatan, ,^*»a. chemburu, ^jj^-j*** ehemburuan, ili tahu, (j^^-4--^ ketahui, .>U jadi, ^^--Jajadikan, :,Ji kuda, j\jS kuda-nya, ui;0 kata, ^^U5 katanya. Many have thought that the accent does not now, in the Peninsula at any rale, fall on that syllable, but still remains on the penultimate of the root word. Consequently it has been suggested that this shifting arose from vicious analogy with Javanese forms and phonetics, which seems improbable considering the small influence Java had in the north of Sumatra : or that it might have arisen from vicious analogy with the spelling of jMalay roots, but that again seems very improbable.

42 SPELLING

must be identical, viz., AJji' and J^y or else the well-established and convenient use of the huruf saksi to represent the accented vowel, i. e. the penultimate, must be abandoned. The theory is artificial, arbitrary and impracticable.

(2) Another theory would revert for guidance in the employment of the huruf saksi to the same system that determined the use of the voivd points in the seventeenth century ; namely, for the determination of vowel sound in rare words. It has historical basis, but what are rare words ? For the pundit, simple terms of husbandry ; for the peasant, the Sanskrit and Arabic loan-words of literature.

(3) A third theory would employ huruf saksi to distinguish words like tolong and tulang which otherwise would be spelt alike. This might be a serviceable empirical device, but strict scientific uniformity \vould entail encyclopaedic knowledge of every word in the language.

(t) Modern spelling has adoptetl hamzah *, which is rare in early MSS., to indicate :

( 1 ) As in Arabic a break between two vowels, of which the first is a and the second begins with the spiriius lenis, or in other words when an a sound passes on to another vowel only by means of a gentle aspirate — it never marks a break between vowel and semivowel. For instance JjIS^ = kaiyil, ^ip ■= kail a diphthong ; JjIS^ with hamzah =. ka'il ; and so too \\ith lain ^^{3 and lavit ^y), though such words are conmionly written without hamzah. j^i-^ sa-ekur, c^^i— sa-orang, oX^^ sa-ikat, ^^ l^ kealatan, J^~SS kekayaan, jlS keenam, ojb ia itu, {j\^^ namai — it must be noted that when a word beginning with alif is preceded by sa the alif of that word is omitted.

(2) As an alternative to (j, indicating the glottal check, j^-.) tirok, Jjb datokj J-svil inchek.

SPELLING 43

§ 19 . The following combinations of words are found : (a) sa is joined to the following word j-^h*» sa-ribu,

((5) The prepositions di and ka are similarly joined i.^jji ka-rumah, sJcJj di»tengah.

(c) And so too, the unemphatic short forms of the personal pronouns eJ ku and j5^ kau, c:*^^^ ku-lihat, IjLu.^ kau-bawa.

((/) ^ mu, Q nya (and ku and kau \\hen they belong to a preceding word) are joined to the word preceding them ; also daku, dikau and dia after akan and dengan. ^\^\ oleh-mu, (jA^^>. bawah-nya, jX),?.i diri-ku, lss^^ akan dia, ^^a:^^ dengan dikau.

{e) The particles lah, kah, tab, pun are joined to the jireceding word, sXiJl alang-kah, ^-^.sl apa-tah.

( /) M yang ii'iay be joined with the preceding or following word, or with both. Ar^^.j' orang yang, c^-sl^.::.*-) yang terdapat, ^i\^skJ-j^\ orang yang di-dapati.

(;') I'he component parts of compound words are joined J^sl apabila, iJLU hulubalang.

The impossibility of formulating a logical system of spelling shows the inadaptability of Arabic characters for the expression of Malay phonetics. But the above general principles, such as they are, are better guides than merely empirical rules which are broken through and through by exceptions.

§ 20. Romanized Malay.

Two main systems of rendering Malay in Roman characters have been practised :

(i) The old Dutch method of reproducing the Arabic letters in Roman equivalents; the indeterminate vowel represented generally by e (and according to Dutch spelling, ancient and modern, u represented by oe, y by j, j by dj, ch by tj, kh by ch).