Printer Friendly

Kharosti and Brahmi.

THE EMERGENCE OF WRITING (1) IN INDIA and the relation between the two early scripts, Brahmi and Kharosti, have received new attention in the last several years. (2) A consensus has emerged that challenges Georg Buhler's theories that had widely been accepted in Western scholarship for a century: that the Brahmi script was derived for commercial use in the eighth century B.C. from an Aramaic alphabet, and that later, during the Achaemenid domination of Northwestern India, a more modern Aramaic script was introduced into that part of India and subsequently modified under the influence of the Brahmi script. (3) Several Indian scholars (and some early European scholars) considered the Brahmi script an indigenous development, and some tried to derive it from the undeciphered script found on the seals of the Indus Valley Civilization that flourished before 2000 B.C. (4)

One of the problems with Buhler's theory is the oddity that the Brahmi which is better equipped to write an Indian language, would have been replaced by the less apt Kharosti (which would see some secondary modifications under the influence of the Brahmi). Buhler refers to the introduction of the Arabic script after the Muslim conquest, but the parallel is not close: the massive influx of Afghans and Turks and the introduction of Islam and Quran study into India cannot be compared with the few Aramaic scribes who would have served the Persian overlords in the provinces of Gandhara and Sindhu. In fact no Aramaic documents of any kind have surfaced from the period of Achaemenid domination in India. Raj Bali Pandey (5) concluded from this lack of Aramaic documents that Kharosti could not be derived from Aramaic, and that perhaps "the Persians did not rule over India directly." But while no Aramaic inscriptions or other texts are known from the whole eastern half of the Achaemenid empire, the Aramaic inscriptions of Asoka, almost a century later, found in Eastern Afghanistan prove the importance of the Aramaic language and script in that border area.

The distinctive features of both scripts are well known. The Kharosti is more cursive, the Brahmi more monumental. While the Kharosti is written from the right to left, does not differentiate between long and short vowels, and indicates initial vowels with similar signs, the Brahmi is written from left to right, distinguishes between long and short vowels, and uses distinctive letters for the initial vowels. Neither direction of writing offers distinct advantages--it is like driving either on the right side or the left side of the road. The other two features are now seen as improvements of the Brahmi over the Kharosti, but all is not well with the arguments offered.

The Kharosti script used in the inscriptions of Asoka, the Sakas, and Kusanas does not differentiate between short and long vowels. Buhler, who considered the Kharosti essentially a clerk's script, spoke of the "lack of [signs for] the long vowels which are useless in everyday usage," (6) and Pandey argued that "The absence of long vowels in the Kharosthi is due to the fact that it was used for writing Prakrits which avoid long vowels ... not due to any Semitic influence." (7) While long vowels were usually shortened in all Prakrit dialects before a consonant cluster, long vowels in open syllables remained mostly unchanged. The contrast was phonemic and could result in different meanings, e.g., dina "day" and dina "miserable." In the shorthand of accounting and of business notes the ambiguity could be tolerated. But the careful distinction of phonetic and phonemic qualities was essential for maintaining the correct recital of Vedic mantras, and the brahmin phoneticians and grammarians studied the distinctions with great care. The Brahmi script essentially differentiates between short and long vowels, but the distinction of i/i and u/u is not always observed, especially in the Asoka inscriptions at Kalsi and the inscriptions at Sohgaura, Piprawa, and Mahasthan. (8) In the more carefully executed inscriptions the strictly phonemic form of the Brahmi script is maintained: one letter for each phoneme (and only one phoneme for each letter). (9) The lack of differentiation of vowel length in the Kharosti (10) has nothing to do with the phonetic or phonemic reality of the Prakrit languages underlying these inscriptions. It derives ultimately from the technique of Semitic writing that essentially only wrote the consonants--with the occasional option to mark a vowel with the letter yod or waw (for /i/ or /u/), in the so-called plene writing. (11)

It has been suggested--most recently by H. Falk--that these innovations are at least partially due to Greek influence. But R. Salomon has rightly countered that the Greek distinction of vowel length is very haphazard and incomplete, (12) whereas the Indian sound table and alphabet are strictly phonemic and well ordered. At the same time, the Indian scribes did not move on to a letter script (as later the Avestan scribes did, probably under Greek influence, in the fourth century A.D.) (13) but stayed with the semi-syllabic design. (14) The pattern of the phonemic analysis of the Sanskrit language achieved by Vedic scholars is much closer to the Brahmi script than the Greek alphabet.

The modern analysis of the writing of initial vowels in the Kharosti script has been deeply flawed. "The full or initial vowel signs further differ from those of Brahmi in that they are all constructed from the basic vowel sign for a to which are affixed the postconsonantal vowel diacritics to form initial i, u, and so on: thus [LANGUAGE NOT REPRODUCIBLE IN ASCII] = initial a/a, while [LANGUAGE NOT REPRODUCIBLE IN ASCII] = initial i/i/." (15) This statement of Salomon's echoes similar statements by Buhler, (16) Charu Chandra Das Gupta, (17) and others. (18) As [LANGUAGE NOT REPRODUCIBLE IN ASCII] (ta) with vowel diacritics denotes [LANGUAGE NOT REPRODUCIBLE IN ASCII] (ti) and [LANGUAGE NOT REPRODUCIBLE IN ASCII] (te) and [LANGUAGE NOT REPRODUCIBLE IN ASCII] (tu), we have [LANGUAGE NOT REPRODUCIBLE IN ASCII] (a), [LANGUAGE NOT REPRODUCIBLE IN ASCII] (i), [LANGUAGE NOT REPRODUCIBLE IN ASCII] (u), etc. All these scholars confused the "original" letter [LANGUAGE NOT REPRODUCIBLE IN ASCII] (t) with the syllabic value /ta/ that it has in Kharosti. The vowel diacritics for /i,u,e/ "displace" the basic /a/ in creating syllabic signs for /ti/, /tu/, /te/, etc., and equally these vocalic diacritics are not attached to the "basic vowel sign for a"--they displace the /a/. Then what are these diacritics attached to? The answer has to come from the Semitic writing system, where the vowel onset, the Semitic aleph, is treated as a consonant--the aleph is phonemic in Semitic languages (cf. Arabic ra's "head," qur'an "Koran"). The Kharosti writing of initial vowels continues directly the Semitic way of writing (19) rather than "responding to a desire for simplification." (20)

Why did the creators of the Brahmi go their own way in the denotation of initial vowels, creating discrete letters for each of them? (21) One could suspect Greek influence, but Greek influence cannot explain the precise notation of vowel length in Brahmi, and it would have failed to promote a true alphabetic script. As the notation of vowel length can be fully explained by the advances of Indian phoneticians and grammarians, we should look at these achievements for inspiration when trying to explain the initial vowel signs of the Brahmi. In the "semi-syllabic" Indian scripts (both in Kharosti and Brahmi) the vowels are marked on the preceding consonant: ka (by default), ki, ku, etc. (by diacritics). But how could an initial vowel be marked by a diacritic? The Kharosti simply followed the Semitic model, attaching the diacritic to the sign for the (consonantal) phoneme aleph. But the Brahmi is a phonemic script, and the vowel onset is not a phoneme in Sanskrit (or any Indian language). There could thus be no consonantal sign in the Brahmi for the vowel diacritics to be attached to. To write iyam "this" (the beginning of Asoka's Rock Edict I) it was necessary to create special letters for the vowels in initial position.

Only in the second half of the first millennium A.D. do we come across letters for initial r and au--some with a unique design, and some based on the letter for/a/. Buhler pointed out that in modern Devanagari the letters for /o/ and /au/ (also for /r/!) are modifications of the letter for /a/ and that this trend continued in Gujerati where also the letters for /e/ and /ai/ are formed that way; but the innovation did not spread to the notation of initial /i/ or /u/. The need for letters for initial /r/, /ai/, and /au/ was negligible, since continuous writing made the notation of initial vowels less common than, e.g., in Greek or English--and words beginning with these vowels (i.e., r, ai, au) are not numerous to begin with. Buhler erred when he saw in this trend a parallel to the Kharosti notation of initial vowels--which is not a simplification of Brahmi writing but its forerunner.

(1.) I leave aside here the undeciphered script of the Indus Valley Civilization of a much earlier time.

(2.) Oskar von Hinuber, Der Beginn der Schrift und fruhe Schriftlichkeit in Indien, Abhandlungen der Akademie der Wissenschaften und der Literatur, 1989 nr. 11 (Mainz 1989); Harry Falk, Die Schrift im alten Indien (Tubingen 1993); Richard Salomon, Indian Epigraphy (New York 1998).

(3.) Georg Buhler, Indische Palaeographie (Strassburg 1896), 18-21.

(4.) Raj Bali Pandey, Indian Paleography, 2nd ed. (Varanasi 1957), 51. Pandey (57f.) denies also the derivation of Kharosti from Aramaic for which the evidence, though, is quite strong: Charu Chandra Das Gupta, The Development of the Kharosthi Script (Calcutta 1958), 284-90.

(5.) Pandey, 56.

(6.) Buhler, 20: "das Fehlen der, fur den Gebrauch des taglichen Lebens unnutzen, langen Vocale..."

(7.) Pandey, 56.

(8.) Salomon, Indian Epigraphy, 31.

(9.) It has been suggested (M. B. Emeneau, Language 22, PP. 86-93) that n in Sanskrit is not a phoneme, since it is predictably conditioned by its context (rajne, panca). But this is not true for Prakrit (anno, ranno). Panini includes n in his pratyaharasutra-s and uses it as a metalinguistic determinative; is this acceptance of n prompted by a desire for symmetry in the table of consonants or by acceptance of a sound that was phonemic in Prakrit?

(10.) In later times, probably under the influence of Brahmi, Kharosti texts from Niya in Central Asia show notations of long vowels. E. J. Rapson (Kharosthi Inscriptions Discovered by Sir M. A. Stein, part III [Oxford 1927] pp. 298f.) wrote: "It was formerly supposed that the Kharosthi alphabet lacked the means of distinguishing long from short vowels; and the fact that such a means existed, even if it was not commonly used, was first made clear by evidence supplied by Niya documents. The lengthening of any vowel may be indicated by a short stroke written below the line, in form and position like the virama of the Devanagari alphabet; cf. a, 3."

(11.) The Pehlevi script of the inscriptions and books of the Persian middle ages stayed closer to the Semitic pattern where only consonants were written and where virtually no word began with a vowel. Kharosthi innovated with the consistent use of diacritical markers to denote the vowel--but still not its length.

(12.) The distinction of [LANGUAGE NOT REPRODUCIBLE IN ASCII] and [LANGUAGE NOT REPRODUCIBLE IN ASCII] (and o and [LANGUAGE NOT REPRODUCIBLE IN ASCII]) is one of vowel quality as much as length (lengthening of [LANGUAGE NOT REPRODUCIBLE IN ASCII] is often written as [LANGUAGE NOT REPRODUCIBLE IN ASCII], of o as o[LANGUAGE NOT REPRODUCIBLE IN ASCII]), and there is no marking of different vowel length in the ease of [LANGUAGE NOT REPRODUCIBLE IN ASCII] and [LANGUAGE NOT REPRODUCIBLE IN ASCII].

(13.) Though the Avesta script runs from right to left (like the Semitic scripts), it writes all sounds individually, including all vowels. The Avesta alphabet with its phonemic and phonetic distinctions exceeds the precision of the Greek alphabet: it observes the phonemic distinctions like the Brahmi and Devanagari alphabet, adding phonetic (allophonic) distinctions that were noted in India only in phonetic manuals of the Siksa, but were rarely expressed in the script.

(14.) A rare exception in the Mahanistha is recorded by W. Schubring, Abhandlungen der Preussischen Akademie der Wissenschaften (1918), 13, 74ff.

(15.) Salomon, Indian Epigraphy, 48.

(16.) Buhler, Indische Palaeographie, 25.

(17.) Charu Chandra Das Gupta, The Development of the Kharosthi Script, 3.

(18.) Ahmad Hasan Dani (Indian Palaeography [Oxford 1963] 257) similarly writes: "While Brahmi has three basic forms of vowels, a, i and u, Kharoshthi has only one, the forms of the remaining vowels being obtained by the addition of diacritic strokes."

(19.) Seen correctly by M. J. Halevy, Journal asiatique ser. 8, 6 (1885), 264.

(20.) Buhler, Indische Palaeographie, 25: "einem Streben nach Vereinfachung zuzuschreiben." E. J. Rapson (Kharosthi Inscriptions, p. 297) remarks: "Hoernle has shown how the same principle tended to modify Brahmi when it was used for Khotanese in Central Asia, and how it has prevailed in the Tibetan alphabet which was borrowed from Khotan." Is it accidental that these trends were strongest in areas that were constantly exposed to the Semitic way of writing, i.e., the marking of the vowel onset?

(21.) Only a a i u e o are attested in the oldest inscriptions. The letter a is a modification of the letter a, as the rare letters for initial i and u in later inscriptions are modifications of those for i and u.
COPYRIGHT 2002 American Oriental Society
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2002 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:early scripts of India
Author:Scharfe, Hartmut
Publication:The Journal of the American Oriental Society
Geographic Code:9INDI
Date:Apr 1, 2002
Previous Article:Sanskrit for civil servants 1806-1818.
Next Article:Vedisch ksad.

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters