654 PERSIA [LANGUAGE. importance for the comparative philologist. In age it almost rivals Sanskrit ; in primitiveness it surpasses that language in many points ; it is inferior only in respect of its less extensive literature, and because it has not been made the subject of systematic gram matical treatment. The age of Zend must be examined in connexion with the age of the Avesta. In its present form the Avesta is not the work of a single author or of any one age, but embraces collec tions produced during a long period. The view which became current through Anquetil Duperron, that the Avesta is throughout the work of Zoroaster an Zend, Zarathushtra], the founder of the religion, has long been abandoned as untenable. But the opposite view, which is now frequently accepted, that not a single word in the book can lay claim to the authorship of Zoroaster, also appears on closer study too sweeping. In the Avesta two stages of the language are plainly distinguishable, for which the supposition of local dialectic variation is not sufficient explanation, but which appear rather to be an older and a younger stage in the development of the same language. The older is represented in but a small part of the whole work, the so-called Gdthds or songs. These songs form the true kernel of the book Yasna l ; they must have been in existence long before all the other parts of the Avesta, throughout the whole of which allusions to them occur. These gathas are what they claim to be, and what they are honoured in the whole Avesta as being, the actual productions of the prophet himself or of his time. They bear in themselves irrefutable proofs of their authenticity, bringing us face to face not with the Zoroaster of the legends but with a real person, announcing a new doctrine and way of salvation, no super natural Being assured of victory, as he is represented in later times, but a mere man, often himself despairing of his final success, and struggling not with spirits and demons but with human conflicts of every sort, in the midst of a society of fellow-believers which was yet feeble and in its earliest infancy. It is almost impossible that a much later period could have produced such unpretentious and almost depreciatory representations of the deeds and personality of the prophet ; certainly nothing of the kind is found outside the gathas. If, then, the gathas reach back to the time of Zoroaster, and he himself, according to the most probable estimate, lived as early as the 14th century B.C., the oldest component parts of the Avesta are hardly inferior in age to the oldest Vedic hymns. The gathas are still extremely rough in style and expression ; the lan guage is richer in forms than the more recent Zend ; and the voca bulary shows important differences. The predominance of the long vowels is a marked characteristic, the constant appearance of a long final vowel contrasting with the preference for a final short in the later speech. Sanskrit. Gatha. Later Zend. abhi (near) aibi aiwi iha (work) izha izha. The clearest evidence of the extreme age of the language of the gathas is its striking resemblance to the oldest Sanskrit, the language of the Vedic poems. The gatha language (much more than the later Zend) and the language of the Vcdas have a close resemblance, exceeding that of any two Romanic languages ; they seem hardly more than two dialects of one tongue. Whole strophes of the gathas can be turned into good old Sanskrit by the applica tion of certain phonetic laws ; for example " mat vdo padaish yd frasruta izhaydo pairijasai mazda ustanazasto at vao asha aredrahyaca nemanghd at vao vangheush manaugho hunaretata," becomes in Sanskrit " mana vah paddih ya prac,rutd ihdyah parigachai medha uttanahastah at va rtena radhrasyaca namasa at vo vasor manasah sunrtaya."2 The language of the other parts of the Avesta is more modern, but not all of one date, so that we can follow the gradual decline of Zend in the Avesta itself. The later the date of a text, the simpler is the grammar, the more lax the use of the cases. We have no chronological points by which to fix the date when Zend ceased to be a living language ; no part of the Avesta can well be put later than the 5th or 4th century B.C. Persian tradition at least regards the collection and arrangement of the holy texts as completed before Alexander s time. At that period they are said to have been already written out on dressed cowhides and preserved in the state archives at Persepolis. The followers of Zoroaster soon ceased to understand Zend. For this reason all that time had spared of the Avesta was translated into Middle Persian or PAHLAVI (q. v.) under the Sasanians. This translation, though still regarded as canonical by the Parsis, shows a very imperfect knowledge of the original language. Its value for modern philology has been the subject of much needless contro- 1 The Avesta is divided into three parts : (1) Yasna, with an appendix, VUparad, a collection of prayers and forms for divine service (2) Vendidad containing directions for purification and the penal code of the ancient Persians ; (3) Khordah-Avesta, or the Small Ave.sta, containing the Yasht the contents of which are for the most part mythological, with shorter prayers for private devotion. With verses of rny making, which now are heard, and with praverfu hands, I come before thee, Mazda, and with the sincere humility of the uprieh man and with the believer s song of praise." versy amongst European scholars. It is only a secondary means towards the comprehension of the ancient text, and must be used with discrimination. A logical system of comparative exegesis, aided by constant reference to Sanskrit, its nearest ally, and to the other Iranian dialects, is the best means of recovering the lost sense of the Zend texts. The phonetic system of Zend consists of simple signs which express the different shades of sound in the language with great precision. In the vowel-system a notable feature is the presence of the short vowels c and o, which are not found in Sanskrit and Old Persian ; thus the Sanskrit santi, Old Persian Jiantiy, becomes henti in Zend. The use of the vowels is complicated by a tendency to combinations of vowels and to epenthesis, i.e., the transposition of weak vowels into the next syllable ; e.g., Sanskrit bharati, Zend baraiti (he carries) ; Old Persian margit, Zend mtiurva (Merv) ; Sanskrit rinakti, Zend irinakhti. Triphthongs are not uncommon ; e.g., Sanskrit a^vcbhycts (dative plural of rtfra, a horse) is in Zend aspaeibyo ; Sanskrit krnoti (he does), Zend kerenaoiti. Zend has also a great tendency to insert irrational vowels, especially near liquids ; owing to this the words seem rather inflated ; e.g. , savya (on the left) becomes in Zend Jidvaya ; Ihrdjati (it glitters), Zend bardzaiti ; gnd (yvvvi), Zend gend. In the consonantal system we are struck by the abundance of sibilants (s and sh, in three forms of modification, z and sh) and nasals (five in number), and by the complete absence of 1. A characteristic phonetic change is that of rt into sh ; e.g., Zend asha for Sanskrit rta, Old Persian arta (in Artaxerxcs) ; fravashi for Pahlavi fravardin, New Persian fcrvcr (the spirits of the dead). The verb displays a like abundance of primary forms with Sanskrit, but the conjugation by periphrasis is only slightly developed. The noun has the same eight cases as in Sanskrit. In the gathas there is a special ablative, limited, as in Sanskrit, to the " a " stems, whilst in later Zend the ablative is extended to all the stems indifferently. We do not know in what character Zend was written before the time of Alexander. From the Sasanian period we find an alpha betic and very legible character in use, derived from Sasanian Pahlavi, and closely resembling the younger Pahlavi found in books. The oldest known manuscripts are of the 14th century A.n. 3 Although the existence of the Zend language was known to the Oxford scholar Hyde, the Frenchman Anquetil Duperron, who went to the East Indies in 1755 to visit the Parsi priests, was the first to draw the attention of the learned world to the subject. Scientific study of Zend texts began with E. Burnouf, and has since then made rapid strides, especially since the Vcdas have opened to us a knowledge of the oldest Sanskrit. 2. Old Persian. This is the language of the ancient Persians Old properly so called, 4 in all probability the mother-tongue of Middle Pers: Persian of the Pahlavi texts, and of New Persian. We know Old Persian from the rock-inscriptions of the Achremenians, now fully deciphered. Most of them, and these the longest, date from the time of Darius (Old Persian, Darayavaush) ; but we have specimens as late as Artaxerxes Ochus. In the latest inscriptions the language is already much degraded ; but on the whole it is almost as antique as Zend, with which it has many points in common. For instance, if we take a sentence from an inscription of Darius, as " Auramazda hya imam bmnim add hya avam asmanam add liya martiyam add hya siyatim add martiyahyd hya Darayavaum khshayathiyam akuuaush aivam paruvnam khshayathiyam," it would be in Zend " Ahuro mazddo yo imam bumim adat 5*6 aom asmanem adat yo mashim adat yo shaitim adat mashyahe yo darayatvo hum khshaetem akerenaot oyum pouru- udm khshaetem." 5 The phonetic system in Old Persian is much simpler than in Zend ; we reckon twenty-four letters in all. The short vowels e, o are wanting ; in their place the old "a" sound still appears as in Sanskrit, e.g., Zend bagem, Old Persian brrgam, Sanskrit b/mgam; Old Persian hamarana, Zend hamcrcna, Sanskrit samarana. As regards consonants, it is noticeable that the older z (soft s) still preserved in Zend passes into d, a rule that still holds in New Persian ; compare Sanskrit. Zend. Old Persian. New Persian. jiasta (hand) zasta dasta dast jrayas (sea) zrayo daraya daryd aham (I) azem adam .... Also Old Persian has no special /. Final consonants are almost entirely wanting. In this respect Old Persian goes much farther than the kindred idioms, e.g., Old Persian abara, Sanskrit abharat, Zend abarat, Zfape ; nominative baga, root -form baga-s, Sanskrit 3 Grammar by Spiegel (Leipsic, 1S67); Dictionary by Justi (Leipsic, 18G4); edition of the Avesta by Westergaard (Copenhagen, 1852), translation into German by Spiegel (Leipsic, 1852), and into English by Darmesteter (Oxford, 1880) in the Sacred Books of the. Efist. 4 And perhaps of the Medes. Although we have no record of the Median language, we cannot regard it as differing to any great extent from the Persian. The Medes and Persians were two closely-connected races. There is nothing to justify us in looking for the true Median language either in the cuneiform writings of the second class or in Zend. 6 "Ormuzd, who created this earth and that heaven, who created man and man s dwelling-place, who made Darius king, the one and only king of many."
