________________
SEPTEMBER, 1894.)
THE ROOTS OF THE DHATUPATHA.
253
forms of the laukilet bháshd for the use of students of Sanskrit. The road, that leads from the Vedânga to the independent Sabdanuśâsana, is a long one, and has not been traversed in one or a few decades. Centuries were required in order to effect the change. For in India processes of development are particularly slow, except when extraneous impulses come into play. To the conclusion that the prebistoric period of the Vyakarana was a long one, point also Påņini's appeals to the authority of numerous predecessors. He not only mentions ten individual earlier teachers, but also the schools of the North and the East, and his grammar shews indeed very clear traces that it has been compiled from various sources. Now, if Panini's Sûtras are the final redaction of a number of older grammatical works, the same must be the case with his Dhâtupatha. For the arrangement of all Indian Sabdân usasanas presupposes the existence of a Dhâtupåtha, and there is no reason to assume that the older grammars were deficient in this respect. It may be even suggested that the occasional discrepancies between the teaching of the Dhâtupatha and rules of the 'Sabdanuéâsana, the existence of which has been alleged, as well as the inequality in the explanatory notes, appended to the roots, are due to an incomplete unification of the various materials which Panini used. Similar instances of what looks like, or really is, carelessness in redaction are not wanting in other Satras. In the Introduction to my Translation of  pastamba's Dharmasútral I have pointed out that, thongh Âpastamba condemns in that work the raising of Ksbetraja sons and the practice of adoption, he get describes in the Srauta Sutra the manner in which a "son of two fathers " shall offer the funeral cakes, and that Hiranyakesin bas not thought it necessary to make the language of the several parts of his Kalpa agree exactly.
But, if Pånini's Dhatapatha must be considered as a compilation from various works, dating from different centuries and composed in various parts of India, it is only to be expected that it should contain many verbs which had already in his time become obsolete and isolated, many variants or dialectic forms. This supposition becomes particularly credible, if the extent of the territory is taken into consideration, from which the ancient grammarians drew their linguistic facts. It extends from the Khyber Pass and the frontier of Sindh in the West, about 71° E. L., to beyond Patna in the East, in 86° E. L., and from the Himalaya to the Vindhya range, where the Narmada, the melhalá bhurah, divides the Uttarapatha from the DakBhiņa patha, or roughly reckoning from the twenty-second to the thirty-first degree N. L. The Aryan population of this large tract was divided into a very great number of tribes, clans, castes and sects, as well as of schools of Vaidiks, Pandits and poets, and owed allegiance to the rulers of perhaps a dozen or more different kingdoms. In historic India tribal, sectarian, political and other divisions have always strongly influenced the development of the languages, and have caused and perpetuated dialectic differences. It seems difficult to assume that matters stood differently in prehistoric times, when there was not, as later, one single work which was generally considered as the standard authority of speech by all educated Aryans. The diversity of the words and forms in literary works and in the speech of the educated classes probably was very great and the task of the earlier grammarians, who had to make their selection from them very difficult.
This difficulty was, it might be expected, not lessened by their method of working. Even in the present day Indian Pandits rarely use any of the scientific apparatus, of which European scholars avail themselves. Indexes, dictionaries and "Collectanea," such as are at the service of the Europeans, are unknown to them. They'chiefly trust to memory, and work in a happy. go-lucky sort of way. Even when writing commentaries, they frequently leave their quotations anverified or entrust the verification to incompetent pupils. The engrmous quantity of the
. I say advisedly looks like or really is carelessness,' because it is always possible that the Sutrakiras intentionally left contradictory rules unaltered in order to indicate an option. Very clear cases of carelessDess in the working up of different materials, do, however, actually occur, e. 9., in the grammatical and lexicographical works of Hemachandra.
30 Sacred Books of the East, Vol. II. p. xxiii, p. 130, note 7.
