________________
JUNE, 1920)
THE WIDE SOUND OF EANDO
111
tion in Prakrita and the vernaculars is a very important one which, if not properly attended to, will leave unexplained several points of word formation in our vernaculars.
With this preliminary remark I proceed to examine a few words to find out the influence of accents in determining the process of corruption :
(i) Skt. khadira> (i) Pr. khaïra > (iii) kha ‘ra > (iv) H. P. M. khaira, a (usually
BO written, but strictly, khair, the final vowel a being dropped and thus the
word becoming practically monosyllabic). Here in the second stage, (i) khaïra evidently contains three syllables (kha-s-rah). Now in the third stage, the word kha'ra, which is dissyllabic (khai-ra), could by no means be formed unless the second of the three syllables in the second stage was reduced in quantity and . thus easily blended with its preceding syllable or vowel (i.e. a in kha-), making the word a dissyllabic one, and this reduction in the quantity of the second syllable in the second stage, (kha-i-na) becomes logically impossible unless the quantity of the first syllable is somewhat strengthened and lengthened by some stress or acoent on it.
Similarly (i) Skt. chatushka > (ii) Pr. chaikka > (iii) chakka > (iv) H. P. M. chawka >chauka,
4 (-chauk, ). Here the existence of the third stage which is a dissyllabic one depends entirely on the reduction in the quantity of the second syllable in the second stage which is effected by the accent on the first syllable.
In a similar way we may explain the cases where there are aya or ava in a Sanskrit or Prakrita word. Thus:
1. (i) Skt. nayana> (ii) Pr. nayaņa or nayana (with ya-sruti)>(iii) nay na> (iv) nay
na>(v) H. P. na ina>(vi) nain or nain ( 47, 4 ). 2. (i) Skt. kavala > (ii) kavala > (iii) kavla > (iv) ka“la>(v) karl or kaul (sim,
In the above instances the ya, ya, or y, and ra, ea, or v have gradually become i and w respectively, because they are not accented. It is evident from Vedic texts that ya and va undergo samprasarana only when they are not accented; on the other hand when they are accented they do not do so. For example, from V. yaj, to sacrifice', we have islá on the one side and iyáj-a on the other; from V vad, 'to speak', udi-tá, udya-te, when the root is unaccented, and váda-ti when it is accented. This fact has been noticed by Dr. Tessitori, too, when he says (ante, Sept. 1918, p. 231) "80 long as the v in kasa-vagt retains the stress or emphasis ... it can never undergo samprasdrana." The principle working here is not difficult to understand. Accent or stress strengthens a syllable, which then cannot be subject to a change.
Let us now illustrate what has been said before regarding the wide and narrow sounds of è, ò and é, ó in Marwari and Gujarati.
(i) Skt. vaira (dissyllable)>(ii) Pr. vaira (trisyllable) <(iii) G. vaira (dissyllable)> (iv) vair =vairs (with a pronunciation different from that of the Skt. diphthong ai)
ver, PC Here in the third stage, vasra, which is derived from the second (vaira, trisyllable), as shown above, there are two syllables, vai and ra; the acoent falls here on the first part of the first syllable, vai, i.e. on the va or a of va and not on the whole syllable vai. The accent having
