________________
122
THE INDIAN ANTIQUABY.
[MAY, 1904.
first person singular. Moreover, a comparison of the vocabularies of Mundári and Kurukh cannot prove anything whatever, because it is a well-known fact that the former has largely influenced the latter. The comparison would have to be extended to other languages of both families, and even in that case it would not prove much, Dravidas and Mundas must have had early intercourse with each other, as well as with the Aryans; and coincidences between them in vocabulary cannot prove any philological connexion, just as we do not class the Aryan dialects with the Dravidians on account of their having several words in common.
Mr. Hahn himself does not appear to attach much importance to the correspondence in vocabulary, and I therefore pass at once to his principal arguments which are based on an Assumed correspondence in grammar. It will, however, not be sofficient to confine ourselves to those features which have been discussed by Mr. Hahn. It will be necessary to extend the comparison of Munda and Dravidian grammar so as to comprise the most characteristic features of both.
Phonology. - The phonetical system of both families differs in many important eharacteristics. It is much more complicated in the Mundâ languages than in Dravidian.
The vowels are mainly the same in both, though the Mandas possess some shades of pronunciation which do not appear to exist in Dravidian. Thus the short a in Dravidian is pronounced as the u in English 'bat.' The Mundi a is usually the short sound corresponding to the a in father.' It also has, however, another sound, which is much more indistinct. It can be compared with the short indistinct e in French quatre-vingt, bat is pronuanced much farther back.
The Dravidian e bas only one sound, that of e in English ember.' Santali e, on the other hand, has two, or rather four, different sounds. It is sometimes pronounced as the a in English 'hat,' and sometimes as the short sound corresponding to the e in German Segen.' There are, besides, two neatral vowels corresponding to the two full s-sounds.
Similar remarks can be made with regard to o, and so forth.
The vowels of consecutive syllables in Santált are made to agree with each other aecording to a well-defined law. If one syllable contains an open sound, the vowel of the other syllables must also be open, and vice versa. Thus, sän-ak', go; but hoy-ok', beeome. In those instances å denotes the open e-sound of a in " hat," and å the open sound of o in "hot."
E and o are changed to i and w, respectively, when the following syllable contains an . Thus, kora, boy; kuri, girl : bheda, a ram; bhidi, a owe.
It will be seen that these changes are quite different from the interchange between i and in some Telugu and Canarese suffixes.
With regard to consonants, it should be noted that the Mundi languages possess complete sets of soft and hard consonants, with and without aspiration. Thus Santali bas k, kh, 9, gh, and corresponding series of palatals, cerebrals, dentale, and labials. The Dravidian languages on the other hand, are mostly devoid of aspirates, and even the unaspirated sounds are not freely used, but interchange according to fixed rules.
Moreover, the Muņda languages possess another set of consonants, or rather semi-consonants, which are usually written k', ch', t, and p'.
"These sounds are not pronounced like other consonants by successively closing and opening,' and allowing the breath to touch the respective organs at their reopening, but by pertly inhaling the breath and simultaneously closing the throat and the respective organs, and not allowing the breath to touch them at their reopening, but letting it pass unarrested out of the throat: thus an abrupt half consonant is produced." (Skrefsrud.)
