Sanskrit is written in various Indian characters, but the character which is peculiarly its own is the Nágarí or Deva-nágarí, i. e. that of the divine, royal, or capital city.' The earliest form of this character can scarcely be traced back to a period much anterior to the 3d century B. c.*; and the more modern, which is one of the most perfect, comprehensive, and philosophical of all known alphabets, is not traceable for several centuries after Christ. The first is the corrupt character of the various inscriptions which have been discovered on pillars and rocks throughout India, written in Mágadhí Prákṛit, spoken at the time of Alexander's invasion over a great part of Hindústán. These inscriptions are ascertained to be addresses from the Buddhist sovereigns of Magadha to the people, enjoining the practice of social virtues and reverence for the priests. They are mostly in the name of Piya-dasi † (for Sanskrit Priyadarśí), supposed to be an epithet of Asoka, who is known to have reigned at some period between the 2d and the 3d century B. c. by his being the grandson of Candra-gupta, probably identical with Sandrakottus, described by Strabo as the most powerful Rájá, immediately succeeding Alexander's death. He was one of the kings of Magadha (Bihár), whose court was at Páli-bothra or Páțaliputra (Patna), and who claimed the title of Samráts or universal monarchs; not without reason, as their addresses are found in these inscriptions at Delhi, and at Kuttack in the south, and again as far west as Gujarát, and again as far north as the Panjáb. The imperfect form of Nágarí which the corrupt character exhibits is incompatible with Sanskrit orthography. It may therefore be conjectured that a more perfect alphabet existed, which bore the same relation to the corrupt form that Sanskrit bore to Prákṛit. later period. It is to be observed, that in the South-Indian dialects the Scythian element constitutes the bulk of the language. It may be compared to the warp, and the Sanskrit admixture to the woof. In the Northern dialects the grammatical structure and many of the idioms and expressions are still Scythian, but the whole material and substance of the language is Sanskrit. See, on this subject, the able Introduction of the Rev. R. Caldwell to his 'Comparative Grammar of the Dravidian or South-Indian Languages,' lately published. * Mr. James Prinsep placed the earliest form as far back as the 5th century B.C. ↑ The regular Prákṛit form would be Pia-dassi. Probably the spoken Prákṛit of that period approached nearer to Sanskrit than the Prákrit of the plays. Nor does it militate against this theory that the perfect character is not found in any ancient inscription, as it is well known that the Bráhmans, who alone spoke and understood the pure Sanskrit, and who alone would therefore need that character, never addressed the people, never proselytized, and never cared to emerge from the indolent apathy of a dignified retirement. An interesting table of the various modifications of the Devanágarí alphabet, both ancient and modern, from the date of the earliest inscriptions to the present time, may be seen in Mr. Edward Thomas' edition of Prinsep's Indian Antiquities,' vol. II. p. 52*. The perfection of the modern character, and the admirable manner in which it adapts itself to the elaborate and symmetrical structure of the Sanskrit language, will be apparent from the first chapter of the present Grammar. * This table, by the kind permission of Mr. Thomas, was lent to me by Mr. Stephen Austin of Hertford, the printer of the above work, and inserted in my second edition; but as the table is more interesting to scholars generally than useful to the student of Sanskrit grammar, and as the increase of matter in the present volume makes space an object, I have preferred referring to the table as exhibited in Prinsep's Indian Antiquities.' NOTICE TO THE STUDENT. THE publication at the Oxford University Press of the Story of Nala' (confessedly the best reading-book for beginners), as a companion to the present volume, with full vocabulary and copious grammatical references, has almost superseded the need for the exercises in translation and parsing appended to the previous editions of the Grammar. They have, therefore, been much abridged in the following edition. When the Sanskrit-English Dictionary, now being printed under the patronage of the Delegates of the Oxford Press, is completed, the student will be supplied with such facilities for translating the literature that a delectus at the end of the Grammar will be rendered unnecessary. Observe-The Sanskrit Manual,' by the author of the present work, contains a complete series of progressive exercises intended to be used in connexion with the rules in the following Grammar, and adapted to facilitate its study. This Manual may be obtained from W. H. Allen & Co., London, or any bookseller. SANSKRIT GRAMMAR. CHAPTER I. LETTERS. I. THE Deva-nágarí character, in which the Sanskrit language is frums-combination. written, is adapted to the expression of almost every known gradation cannot like ato of sound; and every letter has a fixed and invariable pronunciation.( 1. There are fourteen vowels (or without rí* thirteen) and thirty- £79. three simple consonants. To these may be added the nasal symbol, called Anusvára, and the symbol for a final aspirate, called Visarga (see rule 6). They are here exhibited in the dictionary order †. All the vowels, excepting a, have two forms; the first is the initial, the second the medial or non-initial. VOWELS. Hă, M 1 á, z fi, §t í, F ̧«, F ̧ú, H ̧ ri, He ऌ re kri, ल्ह औ au. rí, Nasal symbol, called Anusvára, * m. Symbol for the final aspi The character is not given, as being peculiar to the Vedas. See 16. a. In the previous editions this letter was represented by ch, out of deference to The compound or conjunct consonants (see rule 5) may be multiplied to the extent of four or five hundred. The most common are given here; and a more complete list will be found at the end of the volume. dy, nt, nth, nd, or a tr, ♬ tv, a ts, ¤ ǹć, sa ǹćh, ¤ ǹj, e ṭṭ, ☎ɔ ty, ♬ dg, nn, шny, att, tth, tn, ♬ tm, ♬ ty, thy, dg,ddh, ☎ dbh, ☎ dm, ■ dy, è dr, ☎ dv, w dhy, sa dhv, ☎nt, ♬ nd, ☎ nn, ♬ ny, ■ pt, ■ py, ¤ pr, ♬ pl, ♬ bj, z bd, u̟ by, br, bhy, bhr, ↔ mbh, ➡ mm, ♬ my, ♬ ml, ■ yy, â rk, ✯ rm, & lp, all, a vy, a vr, a ść, qu by, a śr, ■ śl, Tshth, ष्ण sheার, घ्प shy, स्क sk, रख skch, स्त st, स्थ sth, sy, ♬ sr, ☎ sv, ♬ ss, a hm, hy, ☎ kshn, TM kshm, TM kshy, TM♫ gny, г hl, kty, gbhy, TM gry, ♬ śv, v shț, न sn, स्म sm, ktr, ■ ktv, n∙kt, n.ky, try, attr, thy, chr, ndy, tsn, tmy, try, attv,addy, a ddhy, dbhy, dry, πnty, TM mby, ✯rdr, ✯ryy, वे rvy, ष्ट्र shitr, स्थू sthn, सत्य sty, स्त्र str, त्स्य tsny, न्य ntry, त्यै rtsy, rtsnya The letters (except r) have no names like the names, in Hebrew or Greek, but the consonants are enunciated with a; and it is usual in designating any letter to add the word kára; thus, a-kára 'the letter a,' ka-kára 'the letter ka." The letter r, however, is called repha. Observe-In reading the following pages for the first time, it is recommended that the attention be confined to the large type. OF THE METHOD OF WRITING THE VOWELS. 2. The short vowel a is never written unless it begin a word, because it is supposed to be inherent in every consonant. Thus, ak is written, but ka is written; so that in such words as kanaka, nagara, &c., no vowel has to be written. The the usage of English Orientalists. In an essay 'on the use of the Roman character' prefixed to my edition of the Bagh o Bahár, published in 1859 (p. xxviii), I have explained my reasons for preferring ć. mark under the k of, called Viráma (see rule 9), indicates a consonantal stop, that is, the absence of any vowel, inherent or otherwise, after the consonant. a. The other vowels, if written after a consonant, take the place of the inherent a. They assume two forms, according as they are initial or not initial. Thus, ik is written, but ki is written fa. b. Observe here, that the short vowel fi, when initial, is written in its right place, but when not initial, is always written before the letter after which it is pronounced. Hence, in order to write such a word as iti, the letters would have to be arranged in Sanskrit thus, iit इति. c. It is difficult to assign a reason for this peculiarity. The top of the noninitial fi, if written in its right place, might occasionally interfere with a subsequent compound letter, but this tells both ways; as in the word af tarhi, where the i would come more conveniently in its right position. Possibly the peculiarity may be intended to denote a slight drawing back of the breath, in the pronunciation of short i; or it may be merely a method of marking more decidedly the difference between the short and the long vowel. In the Bengálí character this artifice for distinguishing more forcibly between the length of vowel sounds is not confined to i. 3. The long vowels † á and ↑ í, not initial, take their proper place after a consonant. The vowels u, ú, ri, rí, lri, not initial, are written under the consonants after which they are pronounced; as,ku, kú, ♣ kṛi, ☎ kṛí, ☎ klri; except when u or ú follows r, in which case the method of writing is peculiar; thus, ru, ☎ rú. a. The vowels ri, rí, lri and Irí are peculiar to Sanskrit. See rule 11. c.ri only occurs in the root, to make,' and its derivatives. b. The long ri is not found except in technical grammatical phraseology; strictly it has no existence, and is useless except as contributing to the completeness of the alphabetical system. c. The vowels e and ai, not initial, are written above the consonants after which they are pronounced; thus, ke, kai. The vowels o and au (which are formed by placing and over 1 á), like 1 á, take their proper place after their consonants; thus, a ko, kau. OF THE METHOD OF WRITING THE CONSONANTS. 4. The consonants have only one form, whether initial or not initial. And here note this peculiarity in the form of the Devanágarí letters. In every consonant, except those of the cerebral |