A Zoroastrian Educational Institute



HomeArticlesAuthorsBook ReviewCommunityLibraryProminentsRegisterStoreArticle SubmissionAbout Us




Languages of The Parsi scriptures


Martin Haug

I - The Language of the Avesta erroneously called Zend
II - The Pahlavi language and Pazand

Texts of the Pahlavi Inscriptions at Hâjîâbâd HĀJīĀBĀD

III -The Literature Extant


Related Articles:

Related Links:


ESSAY II. Languages of The Parsi scriptures

The languages of Persia, commonly called Iranian, form a separate family of the great Aryan stock of languages which comprises, besides the Iranian idioms, Sanskrit (with its daughters), Greek, Latin, Teutonic (with English), Slavonian, Letto-Lithuanian, Celtic, and all allied dialects. The Iranian idioms arrange themselves under two heads: -

  1. Iranian languages properly so called.
  2. Affiliated tongues.

The first division comprises the ancient, mediaeval, and modern languages of Iran, which includes Persia, Media, and Bactria, those lands which are styled in the Zend-Avesta airyâo danhâvô, “Aryan countries.” We may class them as follows: -

(a.) The East Iranian or Bactrian branch, extant only in the two dialects in which the scanty fragments of the Parsi scriptures is written. The more ancient of them may be called the " Gatha dialect,” because the most extensive and important writings preserved in this peculiar idiom are the so-called Gathas or hymns; the later idiom, in which most of the books of the Zend-Avesta are written, may be called " ancient Bactrian,” or " the classical Avesta language,” which was for many centuries the spoken and written language of Bactria. The Bactrian languages seem to have been dying out in the third century B.C., and they have left no daughters.

(b.) The West Iranian languages, or those of Media and Persia. These are known to us during the three periods Persia. These are known to us during the three periods of antiquity, middle ages, and modern times, but only in the one dialect, which has, at every period, served as the written language throughout the Iranian provinces of the Persian empire. Several dialects are mentioned by lexicographers, but we know very little about them.[1] Of the ancient Persian a few documents are still extant in the cuneiform inscriptions of the kings of the Achaemenian dynasty, found in the ruins of Persepolis, on the rock of Behistun, near Hamedân, and some other places in Persia. This language stands nearest to the two Bactrian dialects of the Zend-Avesta, but exhibits some peculiarities; for instance, we find d used instead of z, as adam, " I,” in the Avesta azem; dasta, “hand,” in the Ayesta zasta. It is undoubtedly the mother of modern Persian, but the differences between them are nevertheless great, and in reading and interpreting the ancient Persian cuneiform inscriptions, Sanskrit and the Avesta, although they be only sister languages, have proved more useful than its daughter, the modern Persian. The chief cause of this difference between ancient and modern Persian is the loss of nearly all the grammatical inflexions of nouns and verbs, and the total disregard of gender, in modern Persian; while in the ancient Persian, as written and spoken at the time of the Achaemenians (B.C. 500-300), we still find a great many inflexions agreeing with those of the Sanskrit, Avesta, and other ancient Aryan tongues. At what time the Persian language, like the English, became simplified, and adapted for amalgamating with foreign words, by the loss of its terminations, we cannot ascertain. But there is every reason to suppose that this dissolution and absorption of terminations, on account of their having become more or less unintelligible, began before the Christian era, because in the later inscriptions of the Achaemenians (B.C. 400), we find already some of the grammatical forms confounded, which confusion we discover also in many parts of the Zend-Avesta. No inscription in the vernacular Persian of the Arsacidans, the successors of the Achaemenians, being extant, we cannot trace the gradual dissolution of the terminations; and when we next meet with the vernacular, in the inscriptions of the first two Sasanian monarchs, it appears in, the curiously mixed form of Pahlavi, which gradually changes till about A.D. 300, when it differs but little from the Pahlavi of the Parsi books, as we shall shortly see.

The second chief division of the Iranian tongues comprises the affiliated languages, that is to say, such as share in the chief peculiarities of this family, but differ from it in many essential particulars. To this division we must refer Ossetic, spoken by some small tribes in the Caucasus, but differing completely from the other Caucasian languages; also Armenian and Afghanic (Pashtű). After this brief notice of the Iranian languages in general, me shall proceed to the more particular consideration of the languages of the Zend-Avesta and other religious literature of the Parsis.

I - The Language of the Avesta erroneously called Zend

 The original language of the Parsi scriptures has usually been called Zend by European scholars, but this name has never been generally admitted by Parsi scholars, although it may have been accepted by a few on European authority, which is apt to be treated with too much deference by Oriental minds. We shall see, hereafter, that this application of the term Zend is quite inconsistent with its general use in the Parsi books, and ought, therefore, to be discarded by scholars who wish to prevent the propagation of error. At present we need only observe that no name for the language of the Parsi scriptures has yet been found in the Parsi books; but whenever the word Zend (zand) is used alone, it is applied to some Pahlavi translation, commentary, or gloss; and whenever the word Avesta (avistak) is used alone, it is applied to the Parsi scriptures in their original language. The language of the Zend, therefore, is Pahlavi, and this is a sufficient reason for not applying that term to another language, with which its connection is probably slight. For want of a better term, we may follow the example of most Parsi scholars in using the term Avesta for the language of the Avesta; and to avoid confusion, we must discard the word Zend altogether when speaking of languages; although, for reasons given hereafter, we may still use Zend-Avesta as a general term for the Parsi scriptures.

The general character of the Avesta language, in both its dialects, is that of a highly developed idiom. It is rich in inflexions, both of the verbs and nouns. In the latter, where three numbers and eight cases can be distinguished, it agrees almost completely with Vedic Sanskrit, and in the former it exhibits a greater variety of forms than the classical Sanskrit. We find, besides, a multitude of case in Sanskrit, Greek, and Latin; but are now and then confounded, much less, however, in the verbs than in the nouns, where the dissolution first began. The crude form, or original uninflected state of the word, is often used instead of the original inflected forms; thus, we find daęva, “demon, evil spirit,” which is really the crude form of the word, employed as the instrumental singular, which ought to be daęvęna, or at least daęva, and as the nominative plural, which ought to be daęvonho or daęva. The long vowels â and î are out of use in the nominative feminine, so that the gender is not so easily recognized from the termination alone as in Sanskrit; thus we have daęna, “creed, belief,” instead of daęna; moreover, the forms of the dative and instrumental are often confounded, especially in the plural. These deviations from the regular forms, and the confusion of terminations, are far more frequent in the classical Avesta than in the Gatha dialect, where the grammatical forms are, in most cases, quite regular, Notwithstanding these symptoms of decay, the relationship of the Avesta language to the most ancient Sanskrit, the so-called Vedic dialect,[2] is as close as that of the different dialects of the Greek language (AEolic, Ionic, Doric, or Attic) to each other. The languages of the sacred hymns of the Brahmans, and of those of the Parsis, are only the two dialects of two separate tribes of one and the same nation. As the Ionians, Dorians, AEtolians &c. were different tribes of the Greek nation, whose general name was Hellenes, so the ancient Brahmans and Parsis were two tribes of the nation, which is called Aryas both in the classical Sanskrit. We find, besides, a multitude of compound words of various kinds, and the sentences are joined together in an easy way, which contributes largely to a ready understanding of the general sense of passages. It is a genuine sister of Sanskrit, Greek, Latin, and Gothic; but we find her no longer in the prime of life, as she appears rather in her declining age. The forms are not always kept strictly distinct from each other, as is the Veda and Zend-Avesta; the former may be compared with’ he Ionians, and the latter with the Dorians. The most striking feature perceptible when comparing both Avesta dialects with Sanskrit is, that they are related closely to the Vedic form of Sanskrit, but not to the classical. In verbal forms, especially moods and tenses, the classical Sanskrit, though very rich in comparison with modern languages, is much poorer than the more primitive dialect preserved in the Vedas; thus it has lost various forms of the subjunctive mood, most tenses of all moods except the indicative (the imperative and potential moods preserving only the present tense), the manifold forms expressing the infinitive mood,[3] &c. ; whereas all these forms are to be found in the Vedas, Zend-Avesta, and Homeric Greek, in the greatest completeness. The syntactical structure in Vedic Sanskrit and the Avesta is simple enough, and verbal forms are much more frequently used than in classical Sanskrit. There can be no doubt that classical Sanskrit was formed long after the separation of the Iranians from the Hindus.

The differences between Vedic Sanskrit and the Avesta language are very little in grammar, but are chiefly of a phonetical and lexicographical nature, like the differences between German and Dutch. There are certain regular changes of sounds, and other phonetic peculiarities perceptible, knowledge of which enables the philologist to convert any Avesta word easily into a pure Sanskrit one. The most remarkable changes are as follows :-

Initial s in Sanskrit is changed in the Avesta into h; thus soma (the sacred juice used by the Brahmans) = haoma; sama, “together, the same,” = hama; sa, “that, he,” = ha; sach, “to follow,” (Lat. sequi) = hach. In the middle of a word the same change takes place, as in asu, “life,” = anhu; except now and then in the last syllable, as in Av. yazaęsha, " thou shalt worship,” where sh is preserved. At the end of a word sh remains unless preceded by a, in which case the termination ash is changed into o, except when followed by the enclitic conjunction cha, when the sibilant is preserved; thus asura-s “living,” becomes ahurô, instead of ahurash, but we find ahurashcha," and the living.”

The Sanskrit h, when not original, but only a derived sound, never remains in the Avesta. It is generally changed into z, as in , “then, therefore,” = S. hi; zima, " winter," = S. hima ; zbę (root), "to invoke,” = S. hve. The Avesta z is also sometimes equivalent to a Sanskrit j, as in zan, “to produce,” (Pers. zâdan) = S. jan (Lat. gigno) ; hizva, " tongue,” = S. jihva.

In comparing Avesta with Sanskrit words, we often observe a nasal in the former which is wanting in the latter; this nasal is usually followed by h, as in enhu, " life,” = S. asu.

Instead of Sanskrit shv we find sp in the Avesta, as in aspa, " horse,” = S. ashva (Lat. equus, Gr. hippos); vispa, " all,” = S. vishva ; spa, " clog,)) = S. shvâ.

In place of Sanskrit rit, besides the regular change into aret,[4] we find ash as an equivalent in the Avesta, as in mashya, “man,” = S. martya (Lat. mortalis, Gr. brotos); asha, “right, true,” = S. rita.

Instead of Sanskrit sv the Avesta has a peculiar guttural aspirate represented by q, and corresponding in sound probably to qu in Latin and khw in Persian, as in qafna, "sleep,” = S. svapna (Lat. somnus, Gr . hypnos, Pers. khwab).

These are the most remarkable phonetic differences between Sanskrit and Avesta words. By attending to them it is very easy to find the Sanskrit word corresponding to one in the Avesta, and we can thus discover a large number of words and forms similar to those in the Vedas. There are, of course, now and then (as is always the case in the dialects of every language) peculiar words to be found in the Avesta, but these are always traceable to Sanskrit roots.

A comparison of the grammatical forms in the Avesta and Sanskrit can be dispensed with. They are so very similar, even when not quite identical, that they are readily recognized by any one who has a slight knowledge of Sanskrit. The strongest proof of the original identity of Sanskrit and Avesta grammatical forms is their harmony even in irregularities. Thus, for instance, the ‘deviations of the pronominal declension from that of the nouns are the same in both languages, as ahmâi, “to him,” = S. asmâi; kahmâi, " to whom,” = S. kasmâi ; yaęshâm, " of whom” (pl.), = S. yeshâm. Also in the declension of irregular nouns we find span, "dog," = S. shvan,[5] sing. nom. spâ = S. shvâ, acc. spânem = S. shvânam, dat. sűnę = S. shune, gen. suno = S. shunas, pl. nom. spânô = S. shvanâs, gen. sunâm = S. shunâm ; likewise pathan, “path,” = S. pathin, sing. nom. paňta = S. paňthas, inst. path = S. pathâ, pl. nom. paňtânô = S. panthânas acc. pathô = S. pathas, gen. pathâm = S. pathâm.

The extremely close affinity of the Avesta language to Vedic Sanskrit can be best seen from some forms of the present tense, in which the classical Sanskrit differs from the Vedic. Compare, for instance, Av. kerenaomi, "I make,” with Ved. krinomi and S. karomi ; Av. jamaiti, " he goes," with Ved. gamati and S. gachchhati ; Av. gerewnami, " I take,” with Ved. gribhnâmi and S. grihnâmi.

With regard to the differences between the two dialects of the Avesta, the language of the Gbthas and the classical or ordinary Avesta, we can here only discuss their relationship to each other in a general way. The chief question is, whether they represent the same language at two different periods of time, or whether they are two contemporary dialects, spoken in two different provinces of the ancient Bactrian empire. Our knowledge of the dialects of the Iranian languages and the periods of their development, previous to the Christian era, is so limited, that it is extremely difficult to decide this question in a satisfactory manner.

The differences between these two dialects are both of a phonetical and grammatical nature. Were the deviations merely of the former kind, we should be fully entitled to ascribe them to two different ways of pronouncing certain vowels and consonants, as generally happens in different districts with nations speaking the same language; but should we discover in one dialect fuller and more ancient forms, and in the other evidently later and more contracted ones, then the difference between the Gatha language and the ordinary Avesta must be ascribed to their being written at different periods.

The phonetical differences of the Gatha language from that of the other books are, at a first glance, so considerable as to induce one to trace them to different localities of the same country, and not to different ages. But on closer inquiry we find that several of these phonetical peculiarities, such as the constant lengthening of’ final vowels, and the severing of one syllable into two (as of the nom. pl. n. of the relative pronoun ya into čeâ), are attributable to the original chanting of the Gathas and other shorter pieces, constituting the older Yasna, and are not to be traced to dialectical differences. These writings are the most important and holiest prayers used in the Zoroastrian divine service, and the way of chanting them was, very likely, analogous to that in which the Brahmans (originally near relations of the Parsis) used to chant the verses of the Sâmaveda at the time of solemn sacrifices, and which is kept up to this day on such occasions. On hearing a Sâmaveda priest chant some verses of this Veda, one notices that he lengthens the final vowels of the words, even when they are short. In Sanskrit, where the grammar was fixed by rules, the texts were not altered according to the mode of chanting them; while in the Avesta, where nothing regarding the grammar and pronunciation was settled, these peculiarities produced by chanting the Gathas and some other pieces crept into the manuscripts, which were generally written from memory only, as is still often the case. Besides these phonetical changes which can be explained as the result of chanting, there are a few other changes of vowels, such as that of a final ô or initial a into ē, as in = , " who ? " and ēmavat = amavat, " strong; " also some changes of consonants, as that of t into s in stavas = stavat, “praising,” and the softening of harsh consonants, as in âdrēng = athras (acc. pl. of âtar, "fire"). These deviations are suggestive of dialectical differences, but they are of no great importance, and no great weight can be attached to them; they are merely such differences as might exist between the idioms of neighboring towns in the same district. That these peculiarities, notwithstanding their insignificance, have been preserved so well, and not been dissolved and changed into the current Bactrian language, which is preserved in the largest portion of the Zend-Avesta, indicates the great reverence in which these hymns mere held by the Zoroastrians. Considering that the Gathas contain the undoubted teaching of Zarathushtra himself (without adverting to other reasons), we do not hesitate to believe that the peculiar language used in the Gathas was the dialect of his own town or district.

As to grammatical forms, the Gatha dialect exhibits not a few deviations from the ordinary Avesta language. Most of these differences evidently represent a more primitive state of the Bactrian language, nearer to its Aryan source; but some might be considered as merely dialectical peculiarities. The genitive singular of masculine nouns in a ends, nearly throughout the Gathas, in ahyâ, which corresponds exactly with the Sanskrit genitive termination asya, while in the ordinary Avesta we always find ahe, apparently a contraction of ahya, thus Gath. dačvahya, "of a demon,” = Av. dačvahe = S. devasya. Again, the first pers. sing. Imperative, expressing intention or volition, requires only the termination a or ai in the Gathas whereas in the ordinary Avesta the derived termination ani prevails, and this is also used in Sanskrit; the usual infinitive formation in the Gathas is that in dyai which is also extremely frequent in the Vedic dialect, while it is nearly unknown in the ordinary Avesta, and wholly so in classical Sanskrit. In the pronouns, especially, the language of the Gathas exhibits more ancient forms than we find in any other part of the Zend-Avesta, as for example maibyâ, "to me,” which ancient form, agreeing so well with Sans. mahyam and Lat. mihi, is nowhere to be found in the ordinary Avesta ; observe also mahyâ, m. maqyâo, f. "of my,” &c. The frequent use of the enclitic pronominal particles î, îm, hîm, &c. (which is a peculiar feature of the Vedic dialect, distinguishing it from classical Sanskrit), and the great freedom with which prepositions are separated from their verbs (a chief characteristic of Vedic Sanskrit and Homeric Greek), indicate a more ancient stage of language in the Gatha dialect than we can discover in the ordinary Avesta, where these traces of a more varied and not quite settled form of expression are much fewer, and only met with, occasionally, in poetical pieces.

Judging from these peculiarities, there seems no doubt that the dialect of the Gathas shows some traces of a higher antiquity than can be claimed for the ordinary Avesta. But the differences are not so great as between the Vedic and classical Sanskrit, or between the Greek of Homer and that of the Attic dialect, the two dialects of the Zend-Avesta being much closer to each other. They represent one and the same language, with such changes as may have been brought about within the space of one or two centuries. The Gatha dialect is, therefore, only one or two centuries older than the ordinary Avesta language, which was the standard language of the ancient Iranian empire.

Much of the difficulty of understanding the Zend-Avesta arises, no doubt, from grammatical defects in the texts extant, owing to the want of grammatical studies among the ancient Persians and Bactrians. Had the study of grammar, as a separate science, flourished among the ancient Mobads and Dasturs, as was the case with Sanskrit grammar among the ancient Brahmans, and had Iran produced men like Pānini, Kātyāyana, and Patanjali, who became lawgivers of the classical Sanskrit language, we should have less ground to complain of the bad condition of the texts, and have found fewer difficulties in explaining them than we have now to encounter. There is every reason to believe that the grammar of the Bactrian language was never fixed in any way by rules; thus the corruptions and abbreviations of forms, which gradually crept from the popular and colloquial into the written language, became unavoidable. In Sanskrit the grammarians built, by means of numerous rules, under which every regular or irregular form in that language was brought, a strong bulwark against the importation of forms from the popular and vulgar language, which was characterized by them as Prākrit.[6] Grammar became a separate branch of study; manuscripts were then either copied or written in strict accordance with the rules of grammar, but always with attention to phonetical peculiarities, especially in Vedic books, if they had any real foundation. To these grammatical studies of the Brahmans, which belong to an age long gone by, we chiefly owe the wonderfully correct and accurately grammatical state of the texts of the Vedas and other revered books of antiquity, In Iran almost all knowledge of the exact meaning of the terminations died out at the time when the ancient Iranian languages underwent the change from inflected to uninflected idioms. Books were extant, and learnt by heart for religious purposes, as is still done by the Parsi priests. But when the language of the Zoroastrian books had become dead, there were no means for the priests, who cared more for the mere mechanical recital of the sacred texts than for a real knowledge of their meaning, to prevent corruptions of the texts. Ignorant of anything like grammar, they copied them mechanically, like the monks of Europe in the middle ages, or wrote them from memory, and, of course, full of blunders and mistakes. On this account we find the copies now used by Mobads and Dasturs, in a most deplorable condition as regards grammar; the terminations are often written as separate words, and vowels inserted where they ought to be omitted, in accordance with the wrong pronunciation of the writer. The best text, comparatively speaking, is to be found in the oldest copies; while in Vedic manuscripts (if written for religious purposes) there is not the slightest difference, whether they are many centuries old or copied at the present day. Westergaard has taken great trouble to give a correct text, according to the oldest manuscripts accessible to him, and his edition is, in most cases, far preferable to the manuscripts used by the priests of modern times, If older manuscripts than those used by Westergaard be known to the Dasturs, they should consider it their bounden duty to procure them for the purpose of collation with Westergaard’s valuable edition, so that they may ascertain all preferable readings for their own information and that of other scholars. Why will they remain behind the Brahmans and the Jews, who have preserved their sacred writings so well, and facilitated modern researches to so great an extent? The era for a sound philological explanation of the time-hallowed fragments of the Zoroastrian writings has come, and the Dasturs, as the spiritual guides of the Parsi community, should take a chief part in it. The darkness in which much of their creed is enshrouded should be dispelled; but the only way of obtaining so desirable a result is by the diffusion of a sound and critical knowledge of the Avesta language.

II - The Pahlavi language and Pazand

It has been already noticed (p. 67) that after the five centuries of obscurity, and probable anarchy,[7] which followed the death of Alexander, when we next meet with the vernacular language of Western Iran, it has assumed the form of Pahlavi, the name generally applied to the language of the inscriptions of the Sasanian dynasty, whether on rocks or coins.

Various interpretations of the word Pahlavî have been proposed. Anquetil derives it from the Persian pahlũ, "side," in which case Pahlavi would mean "the frontier language;" but although this opinion has been held by some scholars, it can hardly be correct, as it is difficult to imagine that a frontier language could have spread over a vast empire. It has also been connected with pahlav, "a hero,” but "the hero language" is a very improbable designation. Native lexicographers have traced Pahlavi to the name Pahlav of a town and province; that it was not the language of a town only, is evident from Firdausi’s statements that the Pahlavi traditions were preserved by the dihgân, " village chief; " it may have been the language of a province, but the province of Pahlav is said to have included Isphâhân, Rai, Hamadân, Nihavand, and Adarbaijan, and must have comprised the ancient Media, but that country is never called Pahlav by Persian and Arab historians. Quatremere was of opinion that Pahlav was identical with the province Parthia, mentioned by the Greeks; he shows, by reference to Armenian authors, that pahlav was a royal title of the Arsacidans. As the Parthians regarded themselves as the most warlike people of the Orient, it is not surprising that pahlav and pahlavan in Persian, and pajhav or pahlav, and pahlavig or palhavig in Armenian, became appellations for a warrior; the name thus lost its national meaning altogether, and became only a title for bold champions of old. It spread beyond the frontiers of Iran eastwards to India, for we find the Pahlavas mentioned as a mighty foreign nation in the Râmâyana, Mahâbhârata, and the Laws of Manu, and we can only understand them to have been the Persians. Regarding the origin of the word, we may compare it with

Pâhlũm, "excellent,” but cannot derive it therefrom.

As the name of a nation, we can discover it only in the Parthva of the cuneiform inscriptions, which is the Parthia of the Greeks and Romans. The change of parthva into pahlav is not surprising, as l is not discoverable in the ancient Iranian tongues, where r is used instead, and th in the middle of an ancient Iranian word generally becomes h in Persian, as in Av. mithra = Pers. mihir. It may be objected that the Parthians were not Persians but probably a Scythic race, and that Pahlavi could not have been the language of the Parthians. This objection, however, will not hold good when we consider that the Parthians mere the actual rulers of Persia for nearly five hundred years, and made themselves respected and famous everywhere by their fierce and successful contests with the mightiest nation of the ancient world, the Romans. It is not surprising, therefore, that the name which once struck such terror into the hearts of Roman generals and emperors was remembered in Persia, and that everything connected with antiquity, whether in history, religion, letters, writing, or language, was called pahlavi, or belonging to the ancient rulers of the country, the Parthians. Pahlavi thus means, in fact, nothing but " ancient Persian " in general, without restriction to any particular period or dialect. This we may see from the use made of the word by Mohammedan writers; thus, Ibn Hauqal, an Arab geographer of the tenth century, when describing the province of Fârs, the ancient Persis, states that three languages were used there, viz. (a) the Fârsi (Persian), spoken by the natives when conversing with one another, which was spread all over Persia, and understood everywhere; (b) the Pahlavi, which was the language of the ancient Persians, in which the Magi wrote their historical records, but which in the writer’s time could not be understood by the inhabitants of the province without a translation ; (c) the Arabic, which was used for all official documents. Of other languages spoken in Persia he notices the Khuzi, the language of Khuzistan, which he states to be quite different from Hebrew, Syriac, or Farsi. In the Mujmilu-t-tawarikh there is an account of "Pahlavi" inscriptions at Persepolis, but the writer evidently means those in cuneiform characters.

From all this we may clearly see that the name Pahlavi was not limited to any particular period or district. In the time of Firdausi (A.D. 1000), the cuneiform writing as well as the Sasanian inscriptions passed for Pahlavi characters; and the ancient Persian and Avesta were regarded as Pahlavi, equally with the official language of the Sasanian period, to which the term has been now restricted, since the others have become better known. The term Pahlavi was thus, in fact, never used by the Persians themselves in any other sense than that of “ancient Persian,” whether they referred to the Sasanian, Arsacidan, Achaemenian, Kayanian, or Peshdadian times: Any reader of the Shahnamah will arrive at this conclusion. This misapplication of a more recent name to earlier historical facts is analogous to the misuse of the appellation Arumak, “Roman,” which the Parsi writers apply to Alexander, the Macedonian conqueror, because he entered the Persian empire from the quarter where the Roman armies appeared in later times.

However loosely the term Pahlavi may have been formerly applied, it has long been practically restricted to the written language of Persia during the Sasanian dynasty, and to the literature of that period and a short time after, of which some fragments have been preserved by the Parsis, in a character resembling that of the Avesta, but very deficient in distinct letters. These Pahlavi writings are of a very peculiar character: instead of presenting us with a pure Iranian dialect (as might be expected in the language of a period commencing with the purely Iranian ancient Persian, and ending with the nearly equally pure Iranian language of Firdausi), it exhibits a large admixture of Semitic words, which increases as we trace it further back, so that the earliest inscriptions of the Sasanian dynasty may be described as being written in a Semitic language, with some admixture of Iranian words, and a prevailing Iranian construction. Traces of the Semitic portion of the Pahlavi can ’ be found on coins of the third and fourth century B.C., and possibly on some tablets found at Nineveh which must be as old as the seventh century B.C.; so there is some reason to suppose that it may be derived from one of the dialects of the Assyrian language, although it differs considerably from the language of the Assyrian cuneiform inscriptions. Practically, however, our acquaintance with Pahlavi commences with the inscriptions of the first Sasanian kings on rocks and coins.

Since the Mohammedan conquest of Persia, the language has become greatly mixed with Semitic words from the Arabic, but this Semitic admixture is of a totally different character to that we find in Pahlavi. The Arabic element in modern Persian consists chiefly of substantives and adjectives, referring to religion, literature, or science; few particles or verbs have been adopted, except when whole phrases have been borrowed ; in fact, the Arabic words, although very numerous, are evidently borrowed from a foreign language. The Semitic element in Pahlavi writings, on the contrary, comprises nearly all kinds of words which are not Arabic in modern Persian; almost all pronouns, prepositions, conjunctions, and common verbs, many adverbs and substantives in frequent use, the first ten numerals, but very few adjectives, are Semitic; while nearly every Arabic word in modern Persian would be re-presented by an Iranian one in Pahlavi writings. It is optional, however, to use Iranian equivalents for any of these Semitic words when writing Pahlavi, but these equivalents are rarely used for some of the pronouns, prepositions, and conjunctions; so rarely, indeed, that the orthography of a few of them is uncertain. Notwithstanding the Semitic appearance of the written Pahlavi, we find that all traces of Semitic inflexions have disappeared, except in a few of the earliest Sasanian inscriptions, written in a peculiar character and dialect, called Chaldaeo-Pahlavi, in which the Chaldee plural suffix in is still often used, as in mâlkin mâlka, "king of kings,” instead of mâlkan mâlka in the ordinary Sasanian Pahlavi inscriptions of the same age, where the Iranian plural suffix ân is used. Besides this Iranian suffix to nouns, we find the verbs appearing in one unchangeable Semitic form, to which is added certain Iranian suffixes, except in the earliest inscriptions in Sasanian Pahlavi, where these suffixes are wanting. In addition to these indications of Iranian grammar, we also find a prevailing Iranian construction in the sentences, as much in the older inscriptions as in the later writings.

The explanation of this extraordinary compound writing, fundamentally Semitic in its words and Iranian in its construction, is that it never literally represented the spoken language of any nation. The Iranians must have inherited their writing from a Semitic people, and although they were acquainted with the separate sounds of each of the letters, they preferred transferring the Semitic words bodily, so as to represent the same ideas in their own Iranian language, and each Semitic word, so transferred, was merely an ideogram, and would be read with the sound of the corresponding Iranian word, without reference to the sounds of the letters composing it; thus the Persians wrote the old Semitic word malkâ, "king,” but they pronounced it shah. When the Semitic words had more than one grammatical form, they would, for the sake of uniformity be usually borrowed in one particular form, and probably in the form which occurred most frequently in the Semitic writings. As these ideograms were to represent an Iranian language, they would be arranged, of course, according to Iranian syntax. For certain words the writer could find no exact Semitic equivalent, especially for Iranian names and religious terms; to express them he had recourse to the alphabet, and wrote these words as they were pronounced; thus laying the foundation of the Iranian element in the Pahlavi. As the Semitic ideograms remained unchanged, [8]1 it was necessary to add Iranian suffixes to indicate the few grammatical forms which survived in the spoken language ; these additions appear to have been only gradually made, for the sake of greater precision, as some of them are not found in the older inscriptions. In later writings we find a few other Iranian additions to Semitic words, used generally to indicate some modification of the original word; thus abu = pid, “father,” is altered into abidar = pidar; am = mad, “mother,” into amidar = madar; &c. In these later writings, we also find the proportion of the Semitic element considerably reduced, being confined to the representation of some three to four hundred of the commonest words in the language, while all other words are Iranian, written as they are pronounced.

As a proof that the Persians did not use the Semitic words in speaking, we may quote the statement of Ammianus Marcellinus (xix. 2,II). When referring to the war between the Roman Emperor Constantius and Shahpuhar II, about A.D. 350, he says that the Persians used the terms saansaan and pyrosen, meaning "king of kings" and "conqueror.” Both these terms are Iranian, the first being shahan-shah and the latter piruz, " victorious,” and show conclusively that the Persians of those times did not pronounce malkân mlkâ, although they wrote those words, but they both wrote and pronounced piruz, which has no Semitic equivalent in Pahlavi. More than four centuries later, Ibn Muqaffa, a Mohammedan writer of the latter half of the eighth century, states that the Persians ‘possess a kind of spelling which they call zavarish; they write by it the characters connected as well as separated, and it 'consists of about a thousand words (which are put together), in order to distinguish those which have the same meaning. For instance, if somebody intends to write gosht, that is lakhm (meat) in Arabic, he writes bisrâ, but ' reads gosht; and if somebody intends to write nân, that is khubz (bread) in Arabic, he writes lahma, but reads nân. And in this manner they treat all words that they intend to write. Only things which do not require such a change are written just as they are pronounced. It appears from this that the Persians of the eighth century did exactly as a Parsi priest would do at the present time; when they came to a Semitic word while reading Pahlavi, they pronounced its Persian equivalent, so that their reading was entirely Persian, although the writing was an odd mixture of Semitic, Persian, and hybrid words. It was always optional to write the Persian word instead of its Semitic equivalent, and it was only necessary to make this the rule, instead of the exception, to convert the old Pahlavi into pure Persian. This final step became compulsory when the Persians adopted a new alphabet, with which the old Semitic ideograms would not amalgamate, but which facilitated the adoption of Arabic terms introduced by their Mohammedan conquerors. Hence the sudden change from Pahlavi to modern Persian was rather a change in writing than an alteration in speaking. The spoken language changed but slowly, by the gradual adoption of Arabic words and phrases, as may be seen from a comparison of the language of Firdausi with that of recent Persian writers.

Ibn Muqaffa uses the term zavarish for the Semitic element in Pahlavi, and this is the term usually employed in Persian, although written occasionally azvarish or uzvarash; in Pahlavi it is written huzvarish or auzvarishn, but it is doubtful if the word occurs in any very old writings. Several attempts have been made to explain its etymology, but as its correct form is by no means certain, it affords, very little basis for trustworthy etymology. The term Huzvarish is applied not only to the Semitic ideograms, but also to a smaller number of Iranian words written in an obsolete manner, so as to be liable to incorrect pronunciation; these obsolete Iranian written forms are used as ideograms in the same manner as the old Semitic words. The habit of not pronouncing the Huzvarish as it is written must have tended to produce forgetfulness of the original pronunciation of the words ; this was to some extent obviated by the compilation of a glossary of the Huzvarish forms, with their pronunciation in Avesta characters, as well as their Iranian equivalents. When this glossary was compiled is uncertain, but as the pronunciation of some of the Huzvarish words is evidently merely guessed from the appearance of the letters, we may conclude that the true sounds of some of the words were already forgotten.

It has been already noticed (p. 68) that Pahlavi translations of the Avesta are called Zand, and we may here further observe that the Iranian equivalent of Huzvarish is called Pazand, reserving further explanation of these terms for the third Essay. This Pâzand may be written in Pahlavi characters, as happens when single Pâzand words are substituted for their Huzvarish equivalents in a Pahlavi text; or it may be written in Avesta characters, which happens when the whole text is so transliterated, and is then called a Pâzand text; or this Pâzand text may be further transliterated into the modern Persian character, when it is still called Pâzand, and differs from the Iranian element of modern Persian only in its frequent use of obsolete words, forms, and construction. It would be convenient to call this Persian form of Pâzand by the name Parsi but it is not so called by the Parsis themselves, nor in their books; with them, Parsi or Farsi means simply modern Persian, more or less similar to Firdausi’s language. It has been mentioned above that it would be easy to forget the pronunciation of the Huzvarish words, and it is now necessary to explain how this could be. The Pahlavi alphabets, being of Semitic origin, have not only all the usual deficiencies of other Semitic alphabets, but also some defects peculiar to themselves, so that several sounds are sometimes represented by the same letter; this ambiguity is greatly increased, in Pahlavi books, by the union of two or more of these ambiguous letters into one compound character, which is sometimes precisely similar to one of the other single letters; the uncertainty of reading any word, therefore, which is not readily identified is very great. No short vowels are expressed, except initial a, but it is presumed they are to be understand where necessary, as in all Semitic alphabets.

Two or three of the earliest rock inscriptions of the Sasanian kings record the names and titles of Ardashir-I Pâpakan and his son Shâhpuhar I. (A.D. 226-270) in three languages, Greek and two dialects of Pahlavi. The Pahlavi versions are engraved in two very different characters, one called Chaldaeo-Pahlavi, from some resemblances to Chaldae in letters and forms, the other called Sasanian Pahlavi, as being more generally used by the Sasanian kings in their inscriptions, both on rocks and coins, This latter character changes by degrees, on the coins of the later Sasanian kings, till it becomes nearly identical with the Pahlavi character in the manuscripts still extant; while the Chaldaeo-Pahlavi appears to have gone out of use before A.D. 300. Two more inscriptions, of greater length, are engraved in both these Pahlavi dialects, but without any Greek translation; of one of these inscriptions only a few fragments are yet known, but the other is complete, and we may take it as a specimen of the Pahlavi writings of the early Sasanian times, as it refers to King Shâhpuhar I. (A.D. 240-270).

This inscription is engraved on two separate tablets (one for each dialect), cut on the rock-wall at the entrance of a cave near the village of Hajiâbâd, not far from the ruins of Persepolis. Copies of the two versions were published by Westergaard at the end (pp. 83, 84) of his lithographed edition of the text of the Bundahish. Plaster casts of the whole of the Chaldaeo-Pahlavi, and of the first six lines of the Sasanian Pahlavi version, are preserved in the British Museum and elsewhere; and a photograph from one set of these casts was published by Thomas in the "Journal of the Royal Asiatic Society,” new series, vol. iii. From a comparison of these copies with the photograph we obtain the following texts, the words of one version being placed immediately below those of the other for the sake of convenient comparison, and short vowels being introduced where they seem necessary.

Texts of the Pahlavi Inscriptions at Hâjîâbâd HĀJīĀBĀD

[SASANIAN PAHLAVI.] - Tagalâhi zenman[9] li mazdayasn bagi Shahpuhari, malkân malkâ Airân va Anirân, mino-chitri min yaztân, barman maz dayasn bagi Artakhshatar, malkan malka Airan, mino-chitri min yaztan, napi bagi Papaki malka; afan amat zenman khitaya shaditun adinan levini shatradaran va barbitan va vacharkan va azatan shaditun, afan ragelman pavan zenman diki hankhetan, afan khitaya lechadu zak chitak bara ramitun, bara valman vayak aik khitaya rami tun, tamman vaydk zak argun l a yehvun, aik hot chitak chiti homan, adin biruni patyak yehvun homan; akhar lanman framat: Mino chitaki aurundari chit, mino yadman ketab homan, zak ragelman pavan zenman diki ayu hankhetun, va khitaya va l zak chitaki ayu shaditun, akhar mino khitaya val zak chitak ramitun; valman yadman ketab.

[CHALDAEO-PAHLAVI.]- Karzâvani zenman li mazdayasn alahâ Shahipuhari, malkain malkâ Aryân va Anâryân, mino-shihar min yâztan, bari maz-dayazn alaha Artakhshatar, malkin malka Aryan, mino-shihar min yaztah,puhari puhar bag Papak malka; va amat Ian zenman khireraya shadit, qadmatman khshatradarin, barbitan, raban va azatan lehad[10] shadit nagarin patan zenman vem haqaimut, va khireraya lehu shiti lebara ramit, bish tamman anu khireraya naflat lehavind atarman le yehut, aik ak shiti banit havindee, kal lebara shadedra akasi yehut havinde; adin lan aupadisht: Mino shiti panman satar banit, avat mino yada kedab havint, nagarin patan zenman vem hip haqaimud, va khireraya kal hu shiti hip shayu mino khireraya kal hu shiti yamzud; lehhp yada kedab havind.

A few words in this inscription are not quite intelligible, but by comparing one version with the other, which corresponds closely in all but two or three phrases, we can arrive at the meaning of most of the obscure passages, and translate as follows: -

“This is an edict of me, the Mazda-worshipping divine being Shahpuar, king of the kings of Iran and non-Iran, of spiritual origin from God; son of the Mazda-worshipping divine being Asrashir, king of the kings of Iran, of spiritual origin from God; grandson of the divine being Papak, the king. And when this arrow [11] was shot by us, then it was shot by us in the presence of the satraps, grandees, magnates, and nobles; and our feet were set in this cave, and the arrow was shot out by us towards that target; but there where the arrow would have dropped was no place (for it), where if a target were constructed, then it (the arrow) would have been manifest outside; then we ordered: A spirit target is constructed in front, thus a spirit hand has written: Set not the feet in this cave, and shoot not an arrow at that target, after the spirit arrow shot at that target; the hand has written that.”

Comparing the two versions of this inscription with the Pahlavi of the manuscripts, it will be noticed that though the Chaldaeo-Pahlavi differs most, it still corresponds with the manuscripts to the extent of about one-third of the words, amongst which the preposition kal, " to, at,” explains the manuscript ghal, which has been often read ghan or ghu, and is used for either val or valman. The construction of the Chaldaeo-Pahlavi resembles generally that of the manuscript Pahlavi, but ‘it does not suffix the pronoun to the initial conjunction or adverb in each phrase, which is a peculiarity of Pahlavi as compared with modern Persian. Furthermore, the Chaldaeo-Pahlavi has begun to use Iranian terminations to Semitic verbs, as t in haqaimut, yehu t, havint ; d in lehavind, haqaimud, yamzud; and the conditional de” in havinde. The Sasanian version has not advanced to that stage in which it adopted Iranian terminations to Semitic verbs, although they are freely used in other inscriptions some twenty or thirty years later; but in all other respects the Sasanian approaches much closer than the Chaldaeo-Pahlavi to the language of the manuscripts, about two-thirds of the words being identical, and the construction of the sentences precisely the same. Thus we find the pronoun suffixed to the initial conjunction or adverb in some phrases, as in afan and adinan, only the pronominal suffix is Semitic; but in later Sasanian inscriptions we find Iranian suffixes, as in afam and afash. This inscription leaves the question of the origin of the idhafat, or relative particle, very uncertain. This particle is nearly always expressed in Pahlavi writings, [12] and not merely understood, as it is generally in modern Persian. In this inscription several words, in both versions, end in i but as this vowel termination cannot be the idhâfat in some cases, it may not be so in any. Thus in the Sasanian version the final i may be an idhâfat in bagi Shahpuhari, napi, Pâpaki, levini, and possibly in chitri, but it cannot be so in diki, biruni, and chitaki, and an idhafat is wanting after malka,barman, Artakhshatar, and lechadu. In the Chaldaeo-Pahlavi version the final i may be an idhâfat in Shahipuhari, bari, and puhari, but it cannot be so in shili and akasi, and an idhafat is wanting after alaha, malka, Artakhshatar, puhar, bag, Pâpak, and lehad, and perhaps after shihar and qadmatman, The omission of an idhâfat after malka is most significant, as it is a position in which it would be expressed even in modern Persian; it is, therefore, very doubtful whether any final i is intended as an idhâfat. In inscriptions a few years later we find the idhâfat in the form of the Semitic relative zi.

To compare with the early Sasanian Pahlavi of the inscriptions, we may take, as a specimen of the manuscript Pahlavi, a passage from the Kârnamak-i Ardashfr Pâkakân, in which the Semitic ideograms are given in italics, and a complete Pâzand version, in Neryoshangh's orthography,[13] is interlined; so that the upper line gives the text as it is written, and the lower as it is pronounced: -

[PAHLAVI]- Papak amatash namak did anduhkun yehevund, afash pavan pasukho val Ardakhshir kard nipisht aigh: Lak ld danakyish kard, amat pavan mindavam mun ziyan la ajash shayast budano, levatman vajurgan stejak yedruntano milaya drusht-advajyish aubash guft. Kevan bujishn yemalelun, pavan pedit-mandak[14] angar; maman danak an guft yekavimuned aigh: Dushman pavan dushman zak la tuban vakhduntano mun[15] asho mard min kunishn-i nafshman aubash rased. Denmanich guft yekavimuned aigh: Min zak aish mustavarmand al yehevunih mun javid min valman la vijared. Va lak benafshman daned[16] aigh Ardavan madam li va lak va kabedan anshuta-i den gehan pavan tanu va khaya va (chabun va khvastak kamkartar padakhshai aito. Va kevanich andarj-i li val lak denman sakhttar, aigh khadukanakih va farman-burdar vaduna nafshman-tanu varz val auben-budih al avaspar.

[PAZAND]. - Papak kash nama did anduhgin bud, vash pa pasukh o Arashir kard navasht ku: Tho ne danaiha kard, ka pa this-e ke zia ne azhash shayast budan, awa guzurgan stezha burdan sakhun durusht-awazhiha havash guft. Nun bozheshn go, pa pashemani angar; chi danaga guft ested  ku Dushman pa dushman a ne tua griftam ke asho mard eazh kuneshn-i qesh havash rased. In-cha guft ested ku: Ezh a kas mustavarmad ma bash ke jad azh oi ne guzared. U tho qad danae ku Ardavan awar men u tho u vasan mardum-i aiidar geha pa tan u jan u khir u qasta kamkartar padishah hast. U nun-cha andarzh-i men o tho in sakhttar, ku euganai u farma-burdar[17] kun qesh-tan varz o avin-budi ma awaspar.

This passage may be translated as follows: - Papak, when he saw the letter, became anxious, and he wrote in reply to Ardashir thus: Thou didst unwisely, when, to carry on a quarrel with the great, in a matter from which there need be no harm. thou spakest words fierce and, loudly about it. Now call for release, and recount with sorrow; for the wise have said that an enemy is not able to take that, as an enemy, to which a righteous man attains by his own actions. This also is said: Be not an antagonist of that person, away from whom you depart not. And thou thyself knowest that Ardavan is a very despotic sovereign over me and thee and many men in the world, as to body and life, property and wealth. And now also my advice to thee is most strongly this, that thou practice conciliation thyself, and act obediently, and yield not to want foresight.

It will be noticed that many of the words in this Pahlavi text, such as did, kard, nipisht, &c., are Pazand, although they have Semitic or Huzvarish equivalents, such as khaditund, vadund, yektibund, &c., which might have been used. This is generally the case in Pahlavi manuscripts, as it is quite optional for the writer to use either the

Huzvarish word or its Pazand equivalent, except perhaps in the case of some of the particles and detached pronouns, which are hardly ever used in their Pazand form in Pahlavi writings. It is necessary to observe that the proportion of Huzvbish words in a manuscript is no criterion of its age, but merely an indication of the style of its writer, for it is not unusual for a manuscript of yesterday to contain more Huzvarish than one of the same text written five hundred years ago; though sometimes the case is reversed. The reason for this uncertain use of Huzvarish is obvious; the copyist either knows the text by heart, or reads it from a manuscript, but in either case he repeats it to himself in Pazand, so that he has nothing but frequent reference to the original to guide him in the choice between Huzvarish and Pazand modes of writing, and for want of frequent reference he will often substitute one for the other, or even use a wrong equivalent (if he does not quite understand his text) when there are two Huzvarish forms with nearly the same Pazand, or when he has misread a Huzvarish form which has two meanings. Thus we often find the Huzvarish amat, " when,” confounded with mun, " which,” because the Pazand of both is ka or ke; and sometimes the Huz. aigh, “that,” is similarly confounded, owing to its having been read ki instead of ku; on the other hand, as the Huz. vakhdund, "taken,” cannot be distinguished from vadund, "done,” they are both liable to be read and written either kard or grift, according to the knowledge or ignorance of the copyist.

III -The Literature Extant

Pahlavi writings may be divided into two classes: first, translations from the Avesta; and, secondly, writings of which no Avesta original is known. The translations are always written in sentences of moderate length, alternating with those of the Avesta text; they are extremely literal, but are interspersed with short explanatory sentences, and sometimes with long digressions, serving as a commentary on the text. The Pahlavi writings without an Avesta original are nearly entirely of a religious character, though a few are devoted to historical legends. Pazand versions of some of these writings, as well as of the translations, exist both in the Avesta and modern Persian characters. Sometimes the Pazand, when written in the Avesta character, alternates with a Sanskrit or Gujrati translation; and when written in the modern’ Persian character, in which case we may call it a Parsi version, it is usually accompanied by a Persian translation, either alternating with the Parsi sentences or interlined ; in the latter case, it is a literal translation, and in the former it is more of a paraphrase. Some writings are found only in Persian, and this is more especially the case with the Rivayats or collections of memoranda and decisions regarding ceremonial observances and miscellaneous religious matters; these are generally very free from Arabic words, but some of them contain nearly as much Arabic as is used in Mohammedan Persian writings. These Rivayats also contain metrical Persian versions of some of the more popular Pahlavi and Pazand books ; these distant imitations of the Shahnameh are generally from two hundred to three hundred and fifty years old. Having thus taken a brief survey of the Pahlavi writings and their connection with Parsi literature generally, me may now proceed to give further details of such works as are known to be still extant, beginning with the translations from the Avesta.

The Pahlavi Vendidad is probably the most important of these translations, and extends to about 48000 words.[18]  Each sentence of the Avesta text is continuously followed by a literal translation, or attempted translation, in Pahlavi, interspersed with short explanations of unusual words, and often concluding with an alternative translation, introduced by the phrase, “There is (some one) who says.” In many places the translation of a sentence winds up with a longer commentary, containing Avesta quotations, and citing the opinions of various old commentators who are named, but regarding whom very little is known. As the next sentence in the Avesta text follows without break of line, it is often difficult to distinguish it from one of the Avesta quotations before mentioned. In the translation there are probably fragments of various ages, as some of the commentaries bear traces of translation from Avesta originals, while many of the shorter explanations appear more modern, but they must have been brought together in their present form before the Mohammedan conquest. All the known extant copies of the Vendidad with Pahlavi appear to have descended from a manuscript of herbad Homast, from which a copy was made in Sistan in A.Y. 554 (A.D. I 185) by Ardashir Bahman, and taken to India by herbad Mahyar Mah-mihir, who had been passing six years with the herbads of Sistan, whither he had come from the town of Khujak on the Indus. After the arrival of this MS. In India it was re-copied by Rustam Mihirapan, who has forgotten to mention the year, [19]  and from his copy the oldest manuscript now extant was copied by herbad Mihirapan Kai-Khusro (who was probably his great-grand-nephew) in A.Y. 693 (A.D. 1324) in the town of Kambay. This manuscript is now in the University Library at Copenhagen, but is very defective; the first portion of the manuscript (Vend. i. I-V. 78, Sp.) having fallen into other hands, probably on some division of property among brothers; and nearly half the remainder is SO much damaged, by the ink corroding the paper, that it is almost useless. Another manuscript, which appears to be in the same handwriting, but the colophon of which is missing, is in the India Office Library in London; this is also defective, as the folios containing Vend. I. I-iii. 48 and IV. 82-viii. 310 have fallen into other hands, and have been replaced by modern writing; the folios containing Vend. iii. 49-iv. 81, and a few others, are also damaged by the corrosive action of the ink used by Mihirapan Kai-Khusro. From a comparison of these two manuscripts, we can ascertain the state of the text 5 53 years ago, except with regard to Vend. i. I-iii 48 and a few other short defective passages, for which we must refer to other old manuscripts. One of these was formerly in the library of Dastur Jamasp Asa at Nawsari, and is said to have been transferred from Bombay to Teheran in Persia some twenty years ago. It was copied, probably from the Copenhagen MS., in A.Y. 963 (A.D. I 594), by herbad Ardashir Ziva, in the town of Bhroch; it is rather carelessly written, and many of the later copies are descended from it.[20] Another old manuscript, now in the University Library at Bombay, was obtained at Bhroch; it corresponds very closely to the one last mentioned, and is probably about the same age, but its colophon is lost. The Pahlavi Vendidad was printed at Vienna separate from the Avesta text, and was published by Spiegel in 1853, but his text can be much improved by careful collation with the old manuscripts above mentioned. None of these MSS. contain the twelfth fargard of the Vendidad, so that the Pahlavi translation of this fargard, which occurs in a few modern MSS., is Probably the work of some Dastur in India. It is difficult to account for the omission of the twelfth fargard in the old MSS., as the fargards are all numbered, so that any accidental leap from the eleventh to the thirteenth ought to have been soon discovered; and it is unlikely that the twelfth fargard would have occupied exactly the whole of any number of folios which may have been lost from some original manuscript before it was copied.

The Pahlavi Yasna contains about 39000 words, exclusive of the kiriya or introductory prayers. It is written alternating with itsAvesta,in the same manner as the Vendidad, but the long interpolated  commentaries are much less common, and fewer commentators are quoted; so it may be suspected of containing less old matter than the Pahlavi Vendidad. For the oldest manuscripts of this text we are again indebted to herbad Mihirapan Kai-Khusro who copied at Kambay a manuscript of the Yasna with Pahlavi (now in the University Library at Copenhagen) in A.Y. 692 (A.D. 1323) from a manuscript written by Rustam Mihirapan; in the same year he also wrote a second manuscript of the same, which is now in the library of Dastur Jamaspji Minoehiharji in Bombay, and is dated only twenty-two days later than the first, but it does not mention whence it was copied. Both these manuscripts begin with a series of introductory prayers in Avesta and Pahlavi, of which the commencement is lost; some of the folios are also damaged in both by the corrosive action of the ink used by the writer; and one folio in the middle of the Bombay copy is lost, and many others are worm-eaten. Several more modern manuscripts of the Yasna with Pahlavi exist but they are less common than those of the Vendidad. The Avesta and Pahlavi texts were printed separately at Vienna, and published by Spiegel in 1855, but his text would be improved by collation with the old manuscript in Bombay.

The Pahlavi Visperad contains about 3300 words, and resembles in character the Pahlavi translation of the Yasna. Probably the oldest copy of this text extant is contained in a manuscript of miscellaneous texts brought from India by the author of these Essays; this copy was written by Peshyotan Ram Kamdin at Bhroch in A.Y. 766 (A.D. 1397). The Avesta and Pahlavi texts were printed separately at Vienna, and published by Spiegel, along with the Yasna texts, in 1858.

The Hadokht nask in Pahlavi is a mere fragment, containing about I 5 30 words, and consisting of three fargards which were probably not consecutive in the original Nask. The first fargard details the value of reciting the Ashem vohu formula under different circumstances, and is probably an extract from the first division of the Nask. The second and third fargards describe the fate of the souls of the righteous and wicked respectively during the first three days after death; but their contents do not agree very well with the description of the Nask in the Dinkard, where it is stated to have consisted of three divisions containing 13, 102, and 19 sections respectively.[21] The oldest copies of the text known to be extant are contained in the manuscript of miscellaneous texts written in A.D. 1397, which includes the Visparad, as mentioned above; also in a very similar manuscript in the University Library at Copenhagen, which must be about the same age. Avesta and Pahlavi texts, alternating as in the manuscripts, were printed at Stuttgart, and published with the Arda-Viraf Namak in 1872, and a translation of the Avesta text will be found in the third Essay.

The Vishtasp yasht is found with a Pahlavi translation of about 5200 words, but only one manuscript has been examined; this is in the library of Dastur Jamaspji in Bombay, and is said to have been written some thirty-five years ago. The Avesta text is probably descended from the Kirman manuscript used by Westergaard, and now at Copenhagen, and the Pahlavi text has the appearance of a modern translation.

Pahlavi translations of other Yashts also exist such as those of the Auharmazd yasht, about 2000 words: the Khurshed yasht and MMah yasht, each about 400 words; the Srosh yasht hadokht, about 700 words ; the Haptan yasht Behram yasht, and probably others which have not been examined. In these, as in all the other translations the Pahlavi alternates with the Avesta; and there seems little doubt that most of these Yasht translations are old.

Among the remaining translations are the Pahlavi texts of the Atash hyayish, about 1000 words; the Khurshed nyayish, about 500 words; the Aban nyayish about 450 words ; the Afringan gatha, the Afringn gahanbar, the Afringan dahman (Yasna, lix. 2-15 Sp.), the last containing about 450 words ; the Afrin myazd, also called Afrin Zaratusht ; the Sirozah in both its forms, containing about 530 and 650 words respectively; and many short extracts from the Yasna which are much used in the Xhurdah Avesta, such as the Ashem-vohu Hatha-ahu-vairyo, and Yenke-hatam formulas; Yasna, v. I, 2; XXXV. 4-6, 13-15; i. 65-67, sp.; &c.

The Chidak avistak-i gasan, or selection from the Gathas, is an old miscellaneous collection of short passages, sometimes merely single lines, from various parts of the Gathas, alternating with the usual Pahlavi translation. Altogether 76 lines are quoted from the Avesta, and the Pahlavi translation of about, 1100 words does not differ materially from that given in manuscripts of the Yasna. Several copies of this selection exist, but the oldest seems to be that in the manuscript of miscellaneous texts written in A.D. 1397, as mentioned above.

Intermediate between the translations and the purely Pahlavi works, there are those which contain many Avesta quotations, which are often translated, but do not in themselves form any connected text, as the bulk of the work is Pahlavi. The following three are of this class: -

The Nirangistan contains about 30,000 words, including the Avesta quotations, many of which are no longer extant in the Zend-Avesta It consists of three fargards, and treats of a great number of minute details regarding rites and ceremonies, and precautions to be adopted while performing them. Its contents correspond very closely with the description of the second section of the Husparam Nask, as given in the Dinkard; and the name of that section was Nirangistan. The opinions, of many of the old commentators mentioned in the Pahlavi Vendidad are also often quoted in this work. A manuscript of the Nirangistan was brought from Persia to India by Dastur Jamasp Wilayati, A.D.I. 720 ; this was copied from a manuscript dated A.Y. 840 (A.D.I. 471), but whether it still exists is uncertain ; it was re-copied by Dastur Jamasp Asa of Nawasari in A.Y. 1097 (A.D. 1727), and this copy is now in the library of the Khan Bahadar Dastur Noshirvanji Jamaspji at Poona. Several later copies exist, but owing to the text being difficult and little known to copyists, their variations from the original are unusually numerous.

The Farhang-i oim khaduk or vocabulary of Avesta and Pahlavi, so called from its first words being oim khaduk, consists of about 3300 words, including the Avesta, and contains several words and phrases which are no longer. extant in t he Avesta texts. Very old copies of this vocabulary exist in two manuscripts of miscellaneous Pahlavi tests, one brought from India by the author of these Essays, and written in A.D. 1397 and the other at Copenhagen, written about the same time. Dastur Hoshangji’s edition of this vocabulary, printed at Stuttgart, and published in 1867 with the title of " An Old Zand-Pahlavi Glossary,” could probably be improved by collation with these old copies of the text.

The Afrin-i dahman, including the aogemadaecha Avesta quotations, contains about 2,000 words. The first of the quotations is Yasna, vii. 60 Sp., but most of the others are no longer extant in the Avesta. They are also found with alternating Pazand and Sanskrit translations, and without the introductory sentences of the Afrin We may now proceed to notice the purely Pahlavi works, which contain but few quotations from the Avesta, and those are generally references to the proper texts to be recited on particular occasions. There is much diversity in the style of these compositions, some being merely descriptive, in which the language is easy, and the construction simple; while others are more philosophical, and their language difficult and obscure.

The Vajarkard-i d in i, containing about 19,000 words, might almost be classed with the preceding, as the latter part of it contains several quotations from the Avesta. I t is a very miscellaneous collection of injunctions and details regarding religious matters, resembling a Rivayat, and divided into three chapters, professing to have been written by Medyomah, one of the old commentators quoted in the Pahlavi translations and other works, An old manuscript of the work, written in Kirman, A.Y. 609 (A.D. 1240), is said to have been brought to India and deposited in the library of the Mody family in Surat, where it was copied A.Y.I. 123 (A.D. 1754) by an uncle of the late high-priest of the Parsis in Bombay; from this copy the text was edited by Dastur Peshotanji, and printed in Bombay in 1848, as already mentioned (p. 59). This work includes three or four of the minor texts hereafter mentioned, as will be noticed when we come to them. The Dinkard is the longest Pahlavi work extant, although the first portion of it, containing the first and second books, is missing; the latter part of the work, consisting of books iii.-ix., contains about 170,000 words. The third book consists of a series of explanations of religious matters and duties, for general information and removal of doubt, concluding with a description of the solar and lunar years, and a legendary history of the Dinkard which is evidently identified with that of the Nasks generally ; this book contains 73,000 words. The fourth book contains various statements selected from the religious books by Adarfrobag-i Farukhzadan, the original editor of the Dinkard (see p. 55), extending to about 4000 words; these statements commence with the characteristics of the Ameshaspends, and in discussing those of Shatrovair the third Ameshaspend, an account is g given of the endeavours of various sovereigns, from Vishtasp to Khusro-i Kavadan (Noshir-van), to collect and preserve the national literature. The fifth book contains the sayings of the same Adarfrobag from a book, called Simara,[22] and his replies to many questions on obscure and difficult matters in history, astrology and religious customs, extending to about 6000 words. The sixth book contains the opinions of the poryodkeshan (professors of the primeval religion of Zarathushtra) on all matters of tradition, customs, and duties, with many sayings of Adarpad Maraspendan ; the whole extending to about 23,000 words. The seventh book contains an account of the wonders, or miracles, of the Mazdayasnian religion from the time of Gayomard, the first man, to ‘that of Soshans, the last of the future prophets ; including many details of the life of Zarathushtt and extending to about 16000 words. The eighth book contains an account of the twenty-one Nasks, giving a short description of each, but going into more details of the four Nasks xv.-xviii. Which constitute the majority of the seven "legal Nasks; this book consists of. about 20,000 words. The ninth book contains a much more detailed account of the contents of each fargard of the first three Nasks, concluding with some remarks upon selections from the whole. Yasna, and extending to about 27,000 words. The work concludes with colophons to the extent of nearly 1000 words, which relate that this latter part of the Dinkard was copied at the place where it was found, Khushkand in Asuristan, from an original which had been written by elders of the family of Adarpbd-i Maraspendan, by Mahvandad Narimahan Behram Mihirapan, and finished on the 24th day of the 4th month A.Y. 369 (7th July A.D. 1000). From this copy others dated A.Y. 865, 1009,[23] and 1038 I have descended, and the last appears to have been brought from ‘Persia to Surat in A.Y. 1152 (A.D.I. 783) by Mulla Bahman, and about four years afterwards some copies of the manuscript of A.Y. 1038 (A.D. 1669) were spread among the Parsis; but before any of these copies were made, the manuscript from Persia had been lent to various parties, and more than one-sixth of the whole had been abstracted, so that all the manuscripts are now deficient to’ that extent; but out of 69 folios missing, 64 have been discovered, though they still remain in various hands. The, manuscript itself is in the library of Dastur Sohrabji Rustamji, the high-priest of the Kadmi sect of Parsis in Bombay. Dastur Peshotanji is publishing an edition of the text, with Gujrati and English translations, as has been already mentioned (p. 59), but it will be many years before he can complete his task.

The name Dadistan-i-dini is usually confined to a work of about 30,000 words, written by Dastur Minochihar Yuda-daman, who was high-priest of the Mazdayasnians in Fars and Kirman about A.Y. 350[24] (A.D. 981). It consists of 92 questions and answers about religious duties, customs, and legends ; the last of these answers seems to be incomplete, so that a portion of the original work may have been lost. The oldest manuscript of this text that has been examined was written in Kirman by Marjpbn Fredun in A.Y. 941 (A.D. 1572); his writing was to supply the deficiencies in a still older manuscript, of which only 28 folios now remain; and his manuscript has, in its turn, had its deficiencies supplied ‘from later copies. In this manuscript the text of the Dadistan-i-dini is preceded and followed by other somewhat similar writings by the same Dastur, and by Zadsparam-i Yudan-daman, who appears to have been his brother. The first part of these extra writings contains about 23,000 words, and the last part about 30,000 words, of which 5000 are lost; if these writings be taken as part of the Dadistan-i-dini, the whole work contains about 78,000 words extant. The author of these Essays recommended the Parsis, twelve years ago, to have this work translated, and it is said that a translation was prepared, but has not been published. If the nonappearance of this translation be due to any of the opinions of the old Dastur of Kirman differing from those of Parsis of the present day, it is to be regretted, as the proper course in such a case would be to publish a correct translation, and point out the probable cause of the original writer’s errors in notes; this is all the more necessary as none of the Pahlavi books are free from statements which would be considered heterodox nowadays. Thus, whenever they give details regarding khvetuk-das, or next-of-kin marriage, they describe it as applying to closer relationships than present customs tolerate; but whatever may have been the reasons for the establishment of this custom when the Zoroastrian faith was in power,[25] it is evident that when the faith was held merely by a persecuted remnant of ‘the Persian people, their priests advocated the custom as a specially meritorious act, with the view of discouraging intermarriages with their Mohammedan neighbors, which would have led to the final extinction of Zoroastrianism. That the present customs of the Parsis are not quite the same as those of eight or ten centuries ago is not surprising, when we consider that it was the usual practice of all Christian sects who had sufficient power, two or three centuries ago, to put heretics and witches to death by burning or otherwise; such practices were t he n not only legal, but were considered highly meritorious; now they would be called judicial murders.

The Shikand-gumani vijar is a controversial work of about 18,000 words, written by Mardan-farukh-i Auharmazd-dad, who acknowledges the instruction he has received from the Dinkard of Adarfrobag-i Farukhzadan, which contained a thousand chapters (dar), as well as from the Dinkard [26] of Adarpadyavand, a work no longer known, unless it be the book of the Mainyo-i-khard, mentioned hereafter. The writer begins by answering, some questions of Mihiryar-i Mahmadan of Ispahan regarding the existence and work of the evil spirit being permitted by Auharmazd; he then proceeds to prove the existence of God, and to disprove the arguments of atheists, and of those who disbelieve in the evil spirit, and attribute both, good and evil to God; and he concludes by criticizing the doctrines of the Jews, Christians, and Manichaeans. Most of the manuscripts of this work are incomplete, and only the first 3600 words are found in the Pahlavi character; the more complete manuscripts are in Pazand with Neryosangh's Sanskrit translation, but there are evident indications of the Pazand text having been originally transliterated from Pahlavi An edition of the Pahlavi and Pazand texts has been prepared by Dastur Hoshangji, but is not yet printed.

The Bundahish calls itself " the Zand-akas[27] (zand-knowing, or tradition-informed), which is first about Auhar-mazd's original creation and the antagonism of the evil spirit, and afterwards about the nature of the creatures from the original creation till the end, which is the future (existence, just as it is revealed by the religion of the

Mazdayasnians. The contents of this book are too well known to require further description; it contains about 13,000 words, but the manuscripts do not agree either in extent or arrangement. The most complete and best-arranged text, but not the most accurately copied, is that in the manuscript of miscellaneous Pahlavi texts at Copenhagen, which is about five hundred years old, and has lost one or more folios in the middle of the text of the Bundahish but contains more sections (chaps. xxviii., xxix., xxx., and xxxii. of Anquetil) than are found in other independent copies. The text is found differently arranged, without those sections, but more accurately copied, in the similar manuscript of miscellaneous texts brought from India by the author of these Essays, and written in 1397. Most of the manuscripts in India seem to have been copied from the latter of these two old manuscripts, but they sometimes vary further in their arrangement. The Copenhagen text was lithographed in facsimile and published by Westergaard in 1851; a French translation was published by Anquetil in 1771, and German translations by Windischmann in 1863, and by Justi in 1868.

The Minok-i [28] khard, called in Pazand Mainyo-i khard, or Spirit of Wisdom, consists of sixty-two answers given by the said Spirit to the inquiries of a wise man regarding the tenets, legends, and morals of the Mazdayasnian religion. It contains about 12,000 words, but the text ends abruptly, as if incomplete; and its introduction bears some resemblance to that of the Shikand-gumani, so as to lead to the suspicion that it may be the first portion of the Dinkhard consulted by the author of that work. An old manuscript of the Pahlavi text was brought by Westergaard from Persia, but the Pahlavi versions in India are probably merely translations from the better-known Pazand text which generally alternates with Neryosangh's Sanskrit translation; a manuscript of this Pazand-Sanskrit text, written in A.D. 1520, is preserved in the India Office Library in London. A few fragments of the Pazand text were published, with a German translation, by Spiegel in his "Grammar of the Parsi Language" (1851) and his "Traditional Literature of the Parsis" (1860); and the whole text, both Pazand and Sanskrit, was published by West, with an English translation, in 1871.

The Shayast la-shayast, or Pahlavi Rivayat, contains about 10,000 words, and treats, of sins and good works, the proper treatment of corpses and other kinds of impurity, with the proper modes of purification, the proper use of the sacred thread and shirt, other customs and rites, with the reasons for reciting each of the Gathas, and details of the extent of those hymns; all subjects which are generally explained in the Persian Rivayats; but here the statements are enforced by quotations of the opinions of several of the old commentators, and by references to some of the Nasks no longer extant. The oldest extant copies of this work are contained in the two manuscripts of miscellaneous Pahlavi texts, written about five hundred years ago, which have been already mentioned. In these manuscripts the text appears in two detached portions of about 7500 and 2500 words respectively.

The Arda[29] Viraf namak; or book of Arda Viraf, contains about 8800 words, and describes what was seen by a chosen high-priest in a vision of the other world, where he was shown the rewards of the righteous, the punishments of the wicked, and the neutral state of stationary expectation of those who belong to neither extreme. It is stated in this work that Arda Viraf was called Nikhshapur by some; this is not only the name of a town, but is also that of one of the old commentators, sometimes quoted in the Pahlavi Vendidad, and very often in the Nirangistan; it is possible, therefore, that this commentator may have written the book of Arda Viraf. Copies of this text are found in the two old manuscripts of miscellaneous texts written about five hundred years ago, which have been already mentioned. A manuscript of a Pazand and Sanskrit version, written A.D. 1410, was also brought from India by the author of these Essays; and Persian versions, both in prose and verse, are likewise extant. The Pahlavi text was printed at Stuttgart, and published, with an English translation, in 1872.

The Madigan-i Gosht-i Fryano, of about 3,000 words, is a tale of the evil Akhtya of the Aban Yasht (81-83), propounding thirty-three enigmas to Yoishto-yo-Fryananam, to be solved on pain of death; after this is done he has to solve three enigmas in his turn, but fails and is destroyed. The enigmas are generally of a very trivial character, and nine of them seem to be omitted. This text accompanies that of the book of Arda Viraf in the two old manuscripts before mentioned, and was published with it in 1872.

The Bahman yasht, of about 4,200 words, professes to be a revelation from Auharmazd to Zaratusht of the sufferings and triumphs of the Mazdayasnian religion, from his time to the end of the world, apparently in imitation of part of the Sudkar Nask. As it mentions the Musalman, and gives many details of the sufferings occasioned by them, it must have been written a considerable time after the Mohammedan conquest. It details how the power of the Mazdayasnian religion is to be restored by the victories of Vahiram-i Varjavand, a prince (kai) of the Kayan race, who at the age of thirty is to put himself at the head of Indian and Chinese armies, whose power will be felt as far as the banks of the Indus, which is called the country of Bambo. Foreigners should be careful not to confound this name with Bombay, which is merely a European corruption, through the Portuguese, of Mumbai; a corruption which native writers still avoid when writing in the vernacular languages. The Pahlavi text of this work is found in the old manuscript of miscellaneous texts at Copenhagen, and its two copies, one of which is at Paris, but no other copies have been met with; a Pazand version is, however, common in India. Spiegel has given a German translation of extracts from the Bahman Yasht in his "Traditional Literature of the Parsis."

In the same old manuscript at Copenhagen is the Andarj-i Hudavar-i [30]1 danak, containing about 1800 words, of which one-third have been lost, as two folios are missing. This admonition (andarj) is given in reply to questions asked by his disciple (ashakard). No other copy of this work has been met with, but it will be found, of course, in the two copies of the Copenhagen manuscript.

In the same inanuscript is also a copy of the Madigan-I gujastak Abalish, containing about 1200 words. The accursed Abalish appears to have been a zandik or heretic, who relied upon later corrupt traditions in preference to the true faith. In the presence of Mamun, the commander of the faithful (amir-i muminin) at Baghdad, he proposes seven questions to a Mobad, who replies to the satisfaction of Mamun and the confusion of Abalish himself. The writer concludes by blessing Adarfrobag-i Farukhzadan (the author of an old edition of the Dinkard) for having destroyed Abalish; and he could not have written this work before A.D. 830, as Mamun was living at that time. Many copies of it exist in Pahlavi, Pazand, and Persian.

The Jamasp namak consists of Jamasp's replies to King Gushtasp's questions regarding creation, history, customs of various nations, and the future fate of the religion. The most complete manuscript examined contains about 5,000 but seems unfinished. The Pahlavi text is rare. A Dastur Peshotanji's library in words. A very old manuscript in Dastur Bombay contains about one-fourth of the text, but no other copy has been met with. The Pazand and Persian versions are found in many manuscripts.

A very old manuscript in the library of Dastur Jamaspji in Bombay has been called the Pahlavi Shahnamak, as it contains several short tales connected with the kings of Persia. Its colophon states that it was finished in India, in the town of Tanak,[31] on the 19th day of some month A.Y. 691 (A.D. 1322), by Mihirapan Kai-Khusro, the copyist who wrote the oldest manuscripts of the Yasna and Vendidad that are still extant. The handwriting, however, more nearly resembles that of the old manuscript of miscellaneous texts at Copenhagen, which contains several copies of Mihirapan's writing , with his colophons attached; so that the Pahlavi Shahnamak may also be a copy of his manuscript, but, like that at Copenhagen, it is certainly about five hundred years old. This manuscript is much worm-eaten, but a copy of it exists at Teheran, made one hundred and ten years ago before the original was much damaged, which will probably supply most of the deficiencies in those texts of which no other copies are known to exist.

Of the texts contained in this old manuscript and its single complete copy, the following are not known to exist elsewhere in Pahlavi: - (1.) Yadkar-i Zar iran, of about 3000 words, containing an account of the war between King Vishtasp and Arjasp. (2.) Cities of the Land of Iran, about 880 words, giving their names and a very brief account of each. (3.) Wonders and Prodigies of the Land of Sistan, in about 290 words. (4.) Khusro-i Kavadan (Noshiravan) and the Slave-boy, who replies to the king's thirteen inquiries as to what things are the most pleasant, about 1770 words. (5.) Admonitions to Mazdayasnians in six separate paragraphs, about 940 words. (6.) Andarj-I  Khusro-i Kavadan (Noshiravan), about 380 words, said to contain the dying injunctions of that monarch. (7.) Sayings of Adarfrobag-i Farukhzadan and Bakht-afrid, about 320 words. The following texts, contained in this old manuscript, are also found in Dastur Peshotanji’s old manuscript, which has been already mentioned as containing part of the Jamasp-namak ; but they are not known to exist elsewhere in Pahlavi: -  (1.) Madigan-i si roz, about 460 words, is a statement of what ought to be done on each of the thirty days of the month; at the end it is called an admonition (andarj) of Adarpad-i Maraspendan to his son, which leads one to suspect that it may be a detached portion of his Pandndmak. (2.) Dirakht-i Asurik, about 800 words, is a debate between a tree and a goat as to which of them is the more worthy. (3.) Chatrang namak, about 820 words, relates how a chessboard and chessmen were sent by Devasarm, a great king of India, to Khusro-I Anoshak-ruban (Noshiravan), with a request for an explanation of the game, which was given by Vajurg-mihir-I Bukhtakan, who afterwards takes the game of Niv-Ardashir to India, as an effectual puzzle for the Indian sages. (4.) Injunctions given to men of the good religion, about 800 words. (5.) The Five Dispositions of priests, and Ten Admonitions, about 250 words, which also occur in the Vajarkard-i-dini (pp. 13-16 of Dastur Peshotanji’s edition). (6.) Daruk-i khursandi, about 120 words. (7.) Anecdote of King Vahiram-i Varjavand, about I 90 words. (8.) Advice of a certain man (fulan gabra), about 740 words. Of the following texts contained in the two old manuscripts of Dastur Peshotanji and. Dastur Jamaspji, a third copy exists in the library of the latter Dastur: - (1.) Forms of Letters to kings and great men, about 990 words found also in the vajarkard-i-dini (pp. 102-113 of Dastur Peshotanji’s edition). (2.) Form of Marriage Contract, dated A.Y. 627 (A.D. 1258), about 400 words. (3.) Vachak aechand (some sayings) of Adarpad-i Maraspendan, about 1270 words. (4.) Stayishn-i dron va sipasdari-I myazdpan, about 560 words.

Of the following texts contained in Dastur Jamaspji's old manuscript many copies exist :- (1.) Pandnamak-I Zaratusht, about 1430 words, contains admonitions as to man’s duties. A copy of three-fourths of this text exists in the University Library at Copenhagen. (2.) Andarj-I Adarpad-i Maraspendan, about 1700 words, is sometimes called his Pandnamak, and contains his advice to his son Zartusht; but the last quarter of the text is missing in the old manuscript, and the end is very abrupt in other manuscripts, which makes it probable that the next text in the old manuscript, the Madigan-i si roz may have been originally the conclusion of this, as has been already noticed. This Pahlavi text was printed in Bombay, and published, with a Gujrati translation, by Shahryarji Dadabhai in 1869 ; and an English version of this Gujrati translation, by the Rev. Shapurji Edalji, was published in 1870, but being a translation of a translation, it differs considerably from the meaning of the original. (3.) Kar-namak-i Artakhshir-i Papakan, about 5600 words, records many of the actions of King Ardashir and his son Shahpuhar, beginning with the discovery of Sasan, the father of the former, among the shepherds of Papak, and ending with Auharmazd, the son of the latter, ascending the throne; but this is not the original work, as it begins with the phrase, In the Karnamak of Artakhshfr-i Papakan it was 'thus written.’ A Gujrati translation of this text was published by Dastur Peshotanji in 1853. (4.) Pand-namak-i Vajurg-mihir-i Bukhtakan, the prime minister of King Khusro Noshirvan, contains about 1690 words, but seems to be merely a fragment of the work, as it ends very abruptly. This text is also called the Ganj-i shigan, because it states that it was placed in the royal treasury (ganj-i shahakan in the old manuscript).

The other old manuscript in Dastur Peshotanji’s library, which includes some of the above mentioned texts, likewise contains the following :- (1.) Madigan-i si yazadan about 80 Words, stating the one special quality of each of the thirty Pazads who give their names to the days of the month. Another similar statement, in the old manuscript of miscellaneous texts brought from India by the author of these Essays, specifies different qualities in most cases. (2.) Madigan-i mah Farvardin roji-i Horvadad, about 760 words, which details all the remarkable occurrences said to have taken place, at different periods, on the sixth day of the first month of the Parsi year. A Persian version of this text is found in the Rivayats. (3.) Another Madigan-I si roz, about 1150 words, detailing the proper business and duties for each of the thirty days in the Parsi month and the five Gatha days at the end of the Parsi year. This text is also contained in the Vajarkard-i Dini (pp.113-125 of Dastur Peshotanji’s edition).

Copies of the remaining texts are numerous both in Pahlavi and Pazand. The Madigan-i haft Ameshaspend about 990 words contains a detail of the various duties of the seven Ameshbspends, as revealed by Auharmazd to Zaratusht. The Andarji danak mard, about 520 words, details the advice of a wise man to his son.

The, about 1300 words is the glossary of Huzvarish and P&and edited by ‘Dastur Hoshangji and published in 1870. It is called the Mariknamak-i Asurik or Assyrian vocabulary, by Dastur Peshotanji in the list of Pahlavi works given in the introduction to his Pahlavi Grammar; but Che origin of this name requires explanation, as it appears to be unknown to the Dasturs generally.

The Patit-i Adarpad-i Maraspend, about 1490 words is a form of renunciation of every possible heinous sin, to be recited by the sinner. The Patit-i khud about 1000 words, is a similar form of renunciation, but somewhat abbreviated. Avar chim-i dron, about 380 words, regarding the symbolism of the, ceremonial wafer-cakes and the use of them in the myazd, or sacred feast. The Pahlavi ashirvad, or marriage blessing, about 460 words. The Nam-stayishni, or praise of Ahuramazd, about 260 words. The Afrin-I “tu peshgah-I khada,” so called from its first prayer, which has not been examined.

A Pahlavi version of the Saddar Bundahish is also said to exist, but must be a modern translation, for the Saddar itself although often written in Avesta characters, seems to bi rather Persian than Pazand, as it contains many Arabic words. Dastur Peshotanji mentions a few more Pahlavi texts, some of which may be included among those described above, but under different names. There are also several Persian texts, such as the book of Dadar bin Dad-dukht, &c., which may have originated in Pahlavi.

From the above details we may form some idea of the probable extent of the scanty remnants of Pahlavi literature. Without making any allowance for works, which remain unexamined or have escaped observation, it appears that the extant Pahlavi translations from the Avesta exceed 104,000 words, and the other Pahlavi works exceed 413,000 words, making a total of upwards of 517,000 words in all the extant Pahlavi writings, which have been examined. This total is nearly eleven times the extent of the Pahlavi Vendidad, or forty times that of the Bundahish.

The Parsi community has been doing a good deal, of late years, for the preservation of the last remnants of their national literature, but it would be better if their efforts were of a more systematic character. Before much more is done for encouraging the publication of isolated texts, a systematic inquiry for manuscripts should be set on foot, for the purpose of ascertaining which are the oldest and best manuscripts, so as to avoid the error of editing texts without reference to the best materials. Influential members of the Parsi community, assisted by the Dasturs, ought to have but little difficulty in inducing all possessors of manuscripts to supply a properly organised committee. Such catalogues need only be lists of the names of the works, with the names and dates of the copyists when these are recorded; but all undated manuscripts supposed to be more than a century old should be specially noted. From such lists the committee could easily prepare a statement of all extant texts and of the owners of several of the more valuable manuscripts of each text. Possessed of this information, the next step would be to obtain a copy of the oldest manuscript of each text, beginning with the rarest works, and have it collated with one or two of the next oldest manuscripts (not being copies of the first). These collated copies, if correctly made without any attempt at emendation, would form standard editions of the texts, and should be carefully preserved in some public institution accessible to all members of the Parsi community, such as the Mulla Firuz Library.

It can hardly be expected that Westergaard’s edition of the Avesta texts can be much improved from any manuscripts to be found in India; although copies from Yazd or Kirman, in Persia, might afford valuable emendations coming from an independent source, but it is generally understood in India that there are very few such manuscripts still existing in Persia. Justi's Old-Bactrian Dictionary is a tolerably complete collection of the Avesta words, but requires to be supplemented by the addition of many words contained in the Nirangistan, Farhang-i oim khaduk, and Aogemadaecha; and the meanings attached to the words want careful revision.

With regard to Pahlavi texts, it would be important to discover any Pahlavi Vendidad or Yasna descended from any other source than the manuscripts of Mihirapan Kai-Khusro, also to find the first three fargards, missing from his manuscripts in Europe, in his own handwriting. The first two books of the Dinkard, the Pahlavi text of the latter part of the shikand-gumani, chaps. xxviii.-xxx. Of the Bundahish, and a complete Pahlavi version of the Jamasp-namak, are all desiderata regarding which some information might be obtained by a systematic inquiry for manuscripts. Hitherto the Parsis have had to rely upon Europeans for all explanations of their literature, beyond the merely traditional learning of their priesthood; they may always rely upon some European being ready to carry on such investigations, provided the materials be forthcoming; and Europeans, in their turn, ought to be able to rely on the Parsis for the discovery of all existing materials, and for rendering them accessible.

[1] In Sayyid Hussain Shah Hakikat's Persian grammar, entitled Tuhfatu'Ajam, there are seven Iranian languages enumerated, which are classed under two heads, vie. (a) The obsolete or dead, and (b) such dialects as are still used. Of the obsolete he knows four: sughdi, the language of ancient Sogdiana (Sughdha in the Zend-Avesta); Zauli (for Zâbulî), dialect of Zâbulistân; Sakzî, spoken in Sajastan (called Sakastene by the Greeks); and Hiriwi, spoken in Herat (Haroyu in the Zend-Avesta). As languages in use he mentions Parsi, which, he says, was spoken in Istakhar (Persepolis), ancient capital of Persia; then Darî, or language of the court, according to this writer, spoken at Balkh, Bokhara, Marv and in Badakhshân; and Pahlavi, or Pahlavâni, the language of the so-called Pahlav, comprising the districts of Rai (Ragha in the Zend-Avesta), Isphahan, and Dînũr. Darî he calls the language of Firdausi, but the trifling deviations he mentions to prove the difference between Darî and Pârsî (for instance, ashkam, “belly,” used in Darî for shikam, and abâ, "with," for bâ), refer only to slight changes in spelling, and are utterly insufficient to induce a philologist to consider Darî the an idiom different from Pârsî.

[2] This is distinct from the usual Sanskrit, which alone is studied nowadays by the Brahmans. The most learned Pandits of the present Brahmanic community, who are perfectly acquainted with the classical Sanskrit language, are utterly unable to explain the more ancient portions of the Vedas, which consist chiefly of hymns, and speculations on the meaning of ceremonies, their effects, &c. They learn them parrot-like by heart, but care nothing about understanding their prayers. If they are asked to explain the meaning, they refer to a commentary made several hundred years ago by a highly celebrated Brahman (Sayana), which often fads to give a complete insight into Vedic antiquity.

[3] The Sanskrit vowel ri is always represented by are or ere; rit itself in a corruption of art.

[4] In the Vedic dialect eleven such forms can be found, which are reduced to one in classical Sanskrit.

[5] Spelt as pronounced, sh representing the palatal sibilant, and sh the cerebral sibilant.

[6] One must not, however, lose sight of the fact that a language is not made by grammarians, but by the common people whom they despise, The work of grammarians is merely to take the language as they find it, and try to ascertain what rules they can manufactures to account for the various forms and idioms used by the people around them. So long as such rules are laid down merely as explanations of existing facts, they will be useful to the scholar, and will not impede progress; but once let them be enunciated as inflexible laws, unalterable as those of the Medes and Persians, and then they hinder progress, ossify thought, and stop discovery. Grammar is no exception to the general rule that lams are hurtful unless subject to constant revision; for a law that cannot be altered becomes a dogma, an impediment to discussion, progress, and improvement, whether it be grammatical, medical, legal, scientific, social, or religious. Whether the stoppage of Hindu progress in knowledge beyond a certain point is not due to the excessive systematizing adopted by their writers when they approached that point is a matter worth consideration. Arrived at a certain amount of progress, they ceased to look forward, but contented themselves with surveying and arranging what they already knew.

[7] “In the Kârnâmah of Artakhshîr-i Pâpakân it was written that after the death of Alexander of Rum, there were 240 small rulers of the country of Airan. The warriors of Fars and the borders adjacent to it were in the hands of a chieftain of Ardavan. Papak was governor and sovereign of Fars, and was appointed by Ardavan.” Karnamak-i A.P.

[8] The only exceptions extant seem to be a few Semitic plurals in -in found in the Chaldaeo-Pahlavi inscriptions before mentioned (p. 82); but even these are used in phrases of Iranian construction.

[9] The syllable man, is represented by a single letter in both characters, which evidently corresponds with the common Pahlavi termination man, as we find it here in the common Pahlavi words zenman (= denman), barman, ragelman, valman, tamman, homan, lanman, and yadman, as well as in the uncommon forms qadmatman, qadmatman, atarman, and panman. In tamman the syllable man corresponds to man in Chaldae, but in other words we must suppose it to represent an original van, vain, or an. Thomas reads the letter i, because it resembles i in some old alphabets. For a similar reason Adreas reads it a. Thomas points to the correspondence of barman, in one dialect of our teat, with bari in the other. Adreas points to a similar correspondence of yadman with yada; he also shows that the reading a overcomes many etymological difficulties. We adhere to the traditional man on the authority of the Chaldee tamman, and because we do not see why there should be a second a in the alphabet.

[10] Adreas reads this word lechad, as the his peculiarly formed, and may perhaps represent the letter tsade, or ch in Pahlavi

[11] The form of the word is plural, we adhere to the but used probably for the singular.

[12] A few exceptions to this general rule, besides unintentional omissions, may be discovered, especially in manuscripts from Persia.

[13] Derived from other works, as no version of the Kârnamak by Neryosangh is known.

[14] A doubtful word, and pashemani no difference between these words in is merely a guess.

[15] All MSS have kardano min, and no doubt some old copyist has read vaduntano (= kardan) instead of vakhduntano (= griftan), there being.

[16] Plural used for the singular.

[17] So in all MSS., but the text is either corrupt, or the construction peculiar.

[18] In estimating (more or less accurately) the number of words in each of the works he has examined, as the best standard of their length, the editor has not included the conjunction va and idhafat i; and he has counted compounds as either one or two words according to the usual mode of writing them.

[19] He copied the Arda-Viraf namak in A.Y. 618 (A.D. 1249), and had visited Persia.

[20] The descent of manuscripts can generally be traced by their copying errors, which have been insufficiently erased; or by their misreading ill shaped letters; but it is hazardous to such blunder.

[21] The total number of sections is given as 133; so there must be an error of one in some one of these four numbers

[22] There are, of course, many ‘other ways of reading this name.

[23] These dates no longer exist in the manuscript brought from Persia, but are taken from the copies and from the account given by Mulla Firuz in his Avijeh-Din.

[24] Altered to 250 in the old manuscript written by Marjpan Fredun, but whether the alteration was made by the original writer or not is uncertain.

[25] They had probably something to do with the dislike of Eastern nations to any absolute alienation of family property; a feeling, which led even the Jews to adopt stringent exceptional marriage laws, in case of a failure in direct heirs.

[26] The Mulla Firuz library in Bombay contains two modern Persian manuscripts, named respectively Dinkard and Dinkhird; these were written by Mulla Firuz to describe his voyage to Persia and the answers he obtained to seventy-eight questions Proposed by the Indian Dasturs. These Persian works must not be confounded with their namesakes in Pahlavi.

[27] The word min, “from,” with which many of the manuscripts commence, appears to be a later addition, as it is not found in the Copenhagen manuscript, and has evidently been added by a later hand in the only other manuscript of equal age mentioned in the text.

[28] This word, which is traditionally rend madonad, has been pronounced minavad, or mainivad, and traced to a supposed ancient Persian form, mainivat. Whether such a form actually existed is not known and if it did, we should expect to find its final letter represented by d = t in Pahlavi, and not by d. On the other hand, the Persian min o must have been minok in Pahlavi; this would be liable to be written minog, and the addition of circumflexes (all the uses of which, in Pahlavi, are not thoroughly understood) changes this word into the traditional madonad.

[29] Sometimes written Ardai, which should perhaps be read Ardak having been altered into ardag, which is not distinguishable from ardai. It is no doubt merely a title meaning "righteous;" the Parsis say, however, that it is also a name.

[30] This name may also be read Khushvar-i, or otherwise.

[31] In another colophon, in the middle of the manuscript, this place is called Tamuk in Jazirak (or Gujirak) zilah, the date being the 6th day of the sixth month A.Y. 691.