Two dialects may be recognized in the Avesta; one, the `Glitha dialect,' or the language of the oldest parts of the Gathiis, or metrical sermons of Zoroaster; the other, 'younger Avestan,' or the 'classical dialect.' This latter is the lan guage of the great body of the Avesta. The Gatha dialect is more archaic, standing in the re lation of the Vedic to the classical Sanskrit, or the Homeric Greek to the Attie. Possibly the Gatha language may also owe some of its pecul iarities noticed below to an original difference of locality. The Gatha dialect was the speech of Zoroaster and his followers. Its grammatical structure is archaic and remarkably pure. The younger Avesta. in its very latest compositions, owing to decay, shows many corruptions and confusions in its inflections. All that is written in metre, however, is old and is generally correct and accurate. Inaccuracies that have crept in we must generally attribute to the carelessness of the scribes. In its forms, as a rule, the Avesta is extremely antique; in general it stands on the same plane as the Vedic Sanskrit, and oc casionally, though not often, it shows even more ancient forms.
The language of the Avesta is most closely allied to the Sanskrit, though individually quite distinct from the latter. Together they may be classed as making up an old Indo-Iranian group. Almost any Sanskrit word may be changed at once into its Avestan equivalent, or vice versa, merely by applying certain phonetic laws. As an example may he taken the metrical stanza Ys. 10, S. in the Avesta: yo yaOd puBrant taurunam haomam ma§yo fra abya tanabyo haw() visaite baCs'azai.
`Whosoever kindly welcomes lIaoma, even as a tender stripling. to such person Hoama ap proaches bringing health'—becomes when ren dered word for word in Sanskrit, yo yathil putram tarunam sonium vuntRta nuirtyah prd Obhyas tanubhyalt slim() vi..:ate In its phonology the Avesta agrees with the Sanskrit in its vowels, but the Avesta shows a greater variety in its c- and o- sounds instead of a. Final vowels, except a, are shortened as a rule. The Skt. diphthong a appears in Av. as us., Of, c (final ; Skt. a as Av. no 5//, a (final). A striking peculiarity in Av. is the introduction of epenthetic vowels and help sounds, giving rise to improper diphthongs, bavaiti, he becomes'= Skt. bhdrati; Av. haurra-, 'whole' = Skt. servo-; Av. rakhddhra-, 'word'=Skt. vaktra-; Av.hrara; `sun'=Skt. svar. The Skt. voiceless stops, /c, t,p, generally become spirants, kit, th, ph (i.e. x, B, f) in Av. before consonants. The original voiced aspirates pit, dh, bh, became in Av. simply voiced stops g, b, d. They are so pre served in the old Gatha dialect; the younger dia lect, commonly resolves the latter before conso nants and between vowels into spirants y, s, The sibilant s in Skt. initial becomes Av. h, as in Greek. Thus, Ay. hapta, `seven'=Skt., saptd. %Vhen internal, Skt. s may also appear as oh. Thus Av. va hana-, `vesture'=Skt. vasana-. Final -as of Skt. appears regularly as -a. Thus Ay, asps., 'horse' = Skt. (Hras.
The Gatha dialect regularly lengthens all final vowels. It frequently inserts the anaptyctic vow els: G Av., lard, Y Ay., fra, 'forth'=Skt.. pra. Original Its appears as ng; (1 Ay., daf'v.ing (ace. pl.), Y Av. dat'vOn, `demons'=Skt. divan.
In inflection the Avesta shows nearly the rich ness of the Vedic Sanskrit. There are three gen ders—masculine, neuter, feminine; likewise three numbers, singular, dual, plural. The dual is not extensively used. There are eight well-developed
cases of the noun and the adjective; the normal endings arc: Singular, Nom., s; Ace., -am ; it ; Dat. -at; Gen. -a (-as) ; Loe. -i;Voe.
--. Dual, Nom., Ace., Voe. Instr., Dat., Abl. -bya; Gen. as; Loc. Plan Nom. Voc. -O (-as), -O; Ace. -5 (-as), -a; Instr. Dat. -byo 1 byas) ; Gen. am; Loc. -so, -hu, -sea. The classes of declension agree exactly with the Sanskrit; the method of forming comparison of adjectives likewise corresponds. The numerals answer to Skt. forms, except Ay. aara-, 'one,' but Skt. aka-, and Av. '10,000; but Skt. ayfita. The Av. pronouns closely resemble the Skt., but show also individual peculiarities.
Noteworthy is the remote demonstrative Av. ava, Mu, 'that, yonder,' contrasted with Skt., amd, asau. The verbal system in Av. and in Skt. is in general identical. The roots are chiefly monosyl labic and are subject to similar modifications in both. In voice, mode, and tense, and in their con jugation-system the two languages quite agree. The endings show equal antiquity with the San skrit. The primary active endings in Av. are: Sing. 1, -mi, 2, -hi, 3, -ti; Dual, 1, -vahi, 3, -to, -Bs.; Plur. 1, -mahi, 2, -tha, 3, -nti. The Av. possesses like facility with the Skt. in forming words by means of prefixes, and by adding suf fixes of primary and secondary derivation. The same classes or compounds may be recognized in both tongues. The rules of Sandhi, or joining to gether of words in a sentence, so universal in Skt., are almost wanting in Avestan. The Avesta separates every word; the vowels are fully ex pressed as in Greek, etc., by individual letters. No diacritical points or accents are written in the texts. The metres in which the Gathns are composed have analogues in the Veda. Almost all the metrical parts of the younger Avesta are in 8-syllable lines. The syntax, however, differs in a number of points from the Sanskrit, and shows certain marked individualities, especially in the later portions.
The best information iu general will be found in the contributions by various scholars in Geiger and Kuhn, Grundriss der imnischen Philologic Strassburg, 1895-1902). A new and complete edition of the Avesta texts has been published by Geldner (Stuttgart, 1885-96). The first full editions were by Westergaa•d (Copenhagen, 1852-54), and by Spiegel (Vienna, 1853-58). The latter is complete only for the Avesta in its nar rower sense (Vs. Vsp. Vd.), the Khordah Avesta being omitted; but it gives the Pahlavi version. —Translations: Best for reference is the transla tion by Dormesteter and Mills in the Sacred Books of the East (3 vols., Oxford, ISSO, seq.), and especially the French version by Darme steter,Lc Zend-Avesta (Paris. 1893-94).—Lexic6n and Grammars, F. Justi, Handbuch der Zend sprarhe (Leipzig, 1864) ; De Ilarlez, Manuel do l'Avcsta (Paris, 1882) ; Kanga, Avesta Diction ary (Bombay, 1899) ; 1(artholorme„41tiranischcs Worterbuch (Strassburg, 1902) ; Schuyler, Index of .1restan Fragments (New York, 1902) ; Bar tholmmr, (irundriss der iranischea Philologie 1895) ; Jackson, Avesta Grammar (Stuttgart, 1891) ; Gleldner, Jletrik (Tubingen, 1877).—Literature and Antiquities, Anquetil Duperron, Zend-Avesta. Ouvruge de Zoroastre, etc. (Paris, 1771) ; Hang, Essays on the Parsis, new edition. by E. W. West (London, ISS4). See also the introduction of De Darlez, Avesta (Paris, 1881) ; E. Spiegel, Eranischc kundc (Leipzig, 1871-78) ; Sanjana-Geiger, Civ ifivition of Eastern Iran (London, 1886).