^{page 1 note 1} See about his full name and biography Ethé, H., Gundriss d. Iran. Phil., v. ii, p. 282Google Scholar, and Browne, E. G., A Literary History of Persia, v. ii, p. 270Google Scholar.

^{page 1 note 2} See about him Brockelmann, C., Gesch. d. Arab. Lit., v. i, pp. 200–201Google Scholar. His book is described in Ahlwardt's Catalogue, No. 9972 (v. ix, p. 408 sq.). There is another copy of the same work in the British Museum, Add. 18520.

^{page 1 note 3} I have been unable to find any reference earlier than Jami's. All the posterior information by the authors who lived after him seems always to be derived directly or indirectly from the Nafaḥāt.

^{page 1 note 4} Ed. Nassau-Lees, Calcutta, 1859, pp. 1–2.

^{page 2 note 1} Ibid., pp. 34–312. Some passages, obviously also belonging to the Ṭabaqāt, appear in the later part of the book, especially in the sections, dealing with Anṣārī and his contemporaries' biographies.

^{page 3 note 1} ? Studying the colophons of MSS. prepared in India, one would think that persons of this noble descent were particularly fond of the scribe's profession—so common it seems to have been amongst them.

^{page 3 note 2} Pānīpatī, indeed.

^{page 3 note 3} Apparently the name of a river.

^{page 4 note 1} He mentions the shaykh as his immediate rāwī on ff. 46 v., 115 v., 124 v., 130, 147 v. Allusions to his conversations appear: (33 v.)

, etc.

^{page 4 note 2} Cf. (2) ; (4) ; (99 v.) . See also f. 143 v.

^{page 4 note 3} (114 v.)

^{page 5 note 1} Sometimes he preserves quite obsolete or provincial words in the Nafaḥāt, and a reference from Anṣārī's book invariably proves that they are exactly the same there and in a similar context. They will be mentioned in the vocabulary of rare words later on.

^{page 5 note 2} Jami's words suggest that the Ṭabaqāt were quite popular. It seems strange therefore that copies of them are so rare. It is possible to conjecture that they all were abandoned when Jami's book received its great currency. Most probably therefore the original of our copy was brought to India before the Nafaḥāt came into existence, and was preserved better than other MSS. of the same work because the Indian scribes did not extend their activity to a ‘ correction ’ of the text as well as to its reproduction.

^{page 6 note 1} See about him and his books my article “ A biography of Aḥmad-i-Jām” in JRAS. 1917, pp. 291–365.

^{page 9 note 1} To my knowledge this term is used rather rarely and probably may have been introduced with a desire to say that the changes alluded to were not in wording, but in letters (ḥarf), i.e. orthography of separate words.

^{page 9 note 2} He usually shortens some very long stories and Anṣārī's own à propos, but occasionally the reason of omission is that Jami seems not to trust him entirely.

^{page 9 note 3} Cf. Browne, E., Catalogue of the Persian MSS. in the Library of the University of Cambridge, 1896, p. 19Google Scholar.

^{page 10 note 1} It is difficult to believe that there were any important circumstances which could lead to the opposite state of things, i.e. integration in some dialectic groups. The Persian language is not very flexible and prefers to borrow a word rather than to create a new equivalent for it. As everywhere, the more educated individuals try to use a standard form of speech, and this promotes the gradual disappearance of local peculiarities in the vocabulary, filling it with the words of the ‘ bazar ’ Persian, which, in its turn, brings ready sentences, etc., and destroys in this way the grammatical peculiarities as well. All phases of this process can be watched in various local idioms of Persia. It is remarkable that while all these transition forms of the colloquial Persian meet no resistance on the part of the dialects and even quite different Iranian tongues, as Kurdish, Baluchi, and even Pashtoo, they find a very serious enemy in Turkish. Wherever a Persian- and a Turkish-speaking population lives close by, the former becomes first bilingual and then forgets its original tongue. Besides many provinces where this process is now completed as in Turkistan, Azarbayjan, etc., it can be seen still at work in many places of Persia proper, as along the Turkoman borders, in the Nishapur Valley, in Fars, and especially in the districts of Sultanabad and Hamadan, where often the villagers, who in their appearance, type of face, cranial index, etc., are pure Iranians, speak only Turkish.

^{page 10 note 2} Such are periodical pilgrimages, performed by villagers to the various sacred places, usually combined with large bazars where the result of the harvest, etc., may be realized and the goods required by the peasants purchased. Moreover, there is an old and quite common type of migrations of whole villages, sometimes to very remote localities, in connexion with the exhaustion of the water supply needed for irrigation of the fields. The owners of newly constructed underground canals invite villagers, in similar circumstances, to change their places of habitation, and this is why one meets Isfahanis near Kermanshah or Anarakis near Sabzawar or Meshed.

^{page 11 note 1} Indications of the early differences in the local forms of speech are by no means rare in the Persian literature, although very scattered. Unfortunately the authors of farhanys mention the locality in which a particular word is used only on exceptional occasions, and all their endeavours are usually directed to accumulate these words to reconstruct the ‘ real old’ Persian. Such is the case, e.g., with Asadi (Lughati-Furs, ed. Horn, P., 1897, preface, pp. 13–14Google Scholar), who rarely localizes his idioms, although many of them seem to be provincial. Even nowadays it is possible to trace some of them to a definite locality, as in the case of (f. 24) ‘ good’, (f. 38) ‘ cat’, used in Bukhara; (f. 70 r.) ‘ butter’, (f. 14 r.) ‘squint-eyed’, (chåpulusi or chapuluski) (f. 27 r.) ‘fraud’, (f. 39 r.) ‘frog’, (f. 28 r.), now kalpusü, lizard, etc., are common to Khorasan, especially its southern part. Sometimes even neighbouring districts possessed their peculiar expressions, and an interesting instance from Anṣārī's time may be mentioned here. His younger contemporary, Aḥmad-i-Jām (died in a.h. 536), writes in his book (of which the apparently unique manuscript is preserved in the library of the Asiatic Society of Bengal, E 20), on f. 122 v. :

(a sort of liana) (in the district of Herat ?)

^{page 11 note 2} Geiger, W. (Grundriss d. Iranischen Philologie, v. i, part ii, pp. 412–414Google Scholar) has no doubt as to this traditional language being the form cultivated at the courts of the Persian rulers. He believes that it was originally one of the local dialects of Fars, whence the greatest dynasties of Persia in the pre-Muhammadan period had come. The opinions of such a scholar as W. Geiger are too high authority to be criticized rashly. Therefore here I would like to allow myself to add a remark only because it is based on materials which were not available at the time when he wrote his invaluable book. The usual belief that standard Persian is spoken in its purest form in Fars is probably a product of the native imagination. In fact, not only in the villages the various dialects (quite incomprehensible to anyone who knows only standard Persian, as may be seen from the specimens collected by Mann, O. in his Die Tajik-Mundarten der Provinz Fars, 1909Google Scholar) are very far from this condition, but even the colloquial form used in the cities by more educated people differs from the literary language by far more than in other parts of Persia. From my personal observations I believe that really the ‘ nearest’ form to the standard language must be sought for in the other corner of Iran, especially in the provinces with the old Persian peasant population, i.e. the districts south from Herat, such as Gunābād, Jām, Qāīn, Birjand, Sunnīkhāna, Farāh, and probably Herat itself. It is quite remarkable that the oldest specimens of the Persian literature of the Muhammadan period all come from Khorasan, but Fars, where many books appear in Arabic, begins to contribute to the Persian literary treasure only in quite a late period.

^{page 13 note 1} The attempts to write in a dialect seriously are very rare and only exceptional cases, such as of the Marzubān nāma, the Ḥurūfī literature, the poetry of Amīr Pazewārī, etc., can be recorded in the Caspian provinces (which in many respects are quite different from the remainder of Persia). In other parts of the country writing in a dialectical form of language is calculated exclusively as one of the means to produce a comical effect (just as it is common in all the languages to use for this purpose the rustic or foreign mode of spelling). But as the Oriental ideas about joking are somewhat peculiar, literature of this kind is invariably of extremely obscene character. Such is the poetry of Bū’sḥaq, Yaghmā, Fayyāḍ and Na’īr of Sabzawar, Mulla Ṣabūḥī of Birjand, etc.

^{page 14 note 1} It is often surprising to find a very considerable number of local “men of letters ” practically in every small Persian city. Very few, almost none of them, become known outside a very limited circle of their friends, and their books never survive them very long. Leaking roofs, playing children who tear the book to pieces, use of paper instead of glass in the windows in cold weather, etc., rarely spare these treatises for a very long time.

^{page 14 note 2} It is extremely difficult indeed to draw a proper demarcation line between the ‘standard’ and the ‘popular’ literature because there are many transition forms. Only examination of every individual case can suggest to what class the book belongs.

^{page 14 note 3} The great poem of Firdawsi occupies an isolated position. It seems quite possible that the unsuccessful presentation to Mahmud can be explained chiefly by the unpolished and even ‘ vulgar ’ language in which it was composed. Even nowadays for the greatest majority of the Persians it is not a poem but a versified history, couched in a very coarse and unskilled language which possesses ‘ no beauty ’.

^{page 15 note 1} The question of the lawfulness of using Persian for compositions on religious matters, which was put to the Imams of Khorasan, as told in the preface to the Tafsīr of Ṭabarī translated into Persian in the fourth century a.h. (ff. 3–5 v. of the copy belonging to the Asiatic Society of Bengal, Aa 19, which is about two hundred years older than that in the British Museum, but very defective), suggests that there was still a strong prejudice against the beginnings of the Persian secular literature. The reason for this was probably not its contents (because the same holy divines admired very much the old Arabic poetry which was by no means very religious), but possibly the fact that the memory of its connexion with the pagan tradition was still too fresh.

^{page 16 note 1} The manuscripts of his Munājāt are extremely numerous, but, at the same time, they differ much one from the other and every copy seems to be individual. I never saw any manuscript of this book older than 300 years and the language appears to be invariably quite modern. Perhaps the Ṭabaqāt will help to recover the original version and to reconstruct its language. An interesting question arises in connexion with these pious invocations. Versified ‘ prayers’ almost of the same type are quite common (in the poems of ‘Aṭṭār they occupy hundreds of pages), but in prose, as far as I know, they were not composed by anybody except Anṣārī. A striking likeness exists between them and the Manichean Maḥrnāmag, in Pehlevi, published by Müller, F. W. K. (“ Ein Doppelblatt aus einem Manichäischen Hymnenbuch”, Abhandl. d. k. Preussischen Ak. d. Wiss., 1912)Google Scholar. There is so little originality in the forms of Persian literature that a suspicion arises as to the possible connexion which might exist between both. Is it a simply accidental resemblance or a case of literary atavism, or direct and intentional imitation of a similar form in the old literature ?

^{page 16 note 2} He flourished in the early part of Akbar's reign. About his biography and works see the pamphlet of Dr.Leyden, in the Asiatic Researches, v. x, pp. 363–428Google Scholar, also in C. Rieu's Persian Catalogue, p. 28, and that of H. Ethé (of the library of the India Office), Nos. 2632–8. Almost all his treatises have been lithographed at Peshawar and Lahore.

^{page 18 note 1} The questions of the real position of these prosodic theories and the traditional orthography based on them require a careful revision as soon as they are intended to be applied to colloquial and dialectical Persian. The variations of this kind are some of the most common phonetic rules in all the living Iranian tongues; cf. Geiger, W., “Die Centr. Dialect.” (Grundr. d. I. Ph., v. i, part ii), p. 384Google Scholar, and Lorimer's, D. addenda (“Notes on the Gabri Dialect,” JRAS. 1916), pp. 428–430Google Scholar. The same is very common all over Khorasan and in Fars (see Mann, O., Die Tajik Mundarten, 1909, p. 8)Google Scholar, and even in Kurdish and Pashtoo; cf. W. Geiger (ibid.), Die Sprache d. Afghanen, p. 208, and A. Socin (ibid.), D. Sprache d. Kurden, pp. 265–6.

^{page 18 note 2} Common in Khorasan, si or su.

^{page 19 note 1} In Khorasan both the 1st person sing, and plur. are speltas -um, and this explains why mistakes as above are very common in the MSS. copied there.

^{page 19 note 2} For examples see the section on the negative particles and indeclinabilia.

^{page 19 note 3} The colloquial seems to be fond of changing these forms without any apparent reason, and ishkas(t), istån, ishtew , ushtur, uslåra, etc., are heard as frequently as shikast, etc.

^{page 19 note 4} These changes are also common in the vulgar language.

^{page 20 note 1} It is extremely difficult to form an opinion, based on a study of the old MSS., as to whether there was any difference in pronunciation in the cases of dotted and undotted dāl. Very few MSS. of those which I had a chance to examine followed this system quite thoroughly. Much more common are the cases in which copies of the same age and probably coming from the same province observe this rule with greatest carelessness. In an old Tafsīr (No. Aa 7 of the Asiatic Society of Bengal), the copy of which was apparently made by a professional high-class scribe (as his surname shows) in the beginning of the seventh century a.h. in Khorasan, very often such words as can be found, while on the contrary almost invariably there are , etc.

^{page 21 note 1} As, e.g., in ṭampār (82, ₁), ṭambār (81, ₁₀), and ṭamfār (3, ₄), or chīmīt and chīwīt (common), etc.; see Müller, F. W. K., “ Soghdische Texte,” i, Abhandl. d. k. Preussischen Ak. d. Wiss., 1912Google Scholar.

^{page 21 note 2} This process can be traced as continually developing since the earliest period of the history of the Iranian languages. In the Avesta it gains ground from the pre-Iranian (cf. C. Bartholomae, “ Awestasprache,” in G. d. I. Ph., i, pt. i, pp. 163, 165) and further in Pehlevi (cf. C. Salemann, “ Mittelpersisch,” ibid., p. 259). In modern Persian it is prominent even in the standard language (cf. P. Horn, “ Neupersische Schriftsprache,” ibid., pt. ii, p. 48). About this process in the dialects see W. Geiger, ibid., p. 298 (Pamir group), p. 351 (Caspian gr.), p. 386 (Central gr., also D. Lorimer, op. cit., pp. 434–5), O. Mann, op. cit., pp. 12–13 (dialects of Fars), and A. Socin, op. cit., pp. 261, 263 (in Kurdish).

^{page 21 note 3} This is as common in the colloquial of to-day as in old MSS. Cf. ‘Aṭṭār's Tadhkira, ed. R. Nicholson, v. ii, introduction, p. 6.

^{page 22 note 1} This example (as well as further awgandan) resembles Pehlevi forms yāvēṭan, avyandan. These old-fashioned forms are not the only ones given here (on f. 143 appears for , etc.). But it is rather difficult to believe that yāvēṭan became first yābīdan and then came to its previous form. We are compelled to think that either there is some confusion about the early phonetical equivalents of and or that there were (as alluded to later on) two distinct currents of the same process in the development of labials, one of which preferred the labio-dental group while the other transformed all sounds of this class into the interlabial w.

^{page 23 note 1} This depends probably on the physiological influence of the spirant s, which facilitates a tenuis after itself. Although in the Eastern (Khorasani) Kurdish dialect it often appears as siwī, the Persians and especially Herati Parsiwans spell it ispīd, ispand, etc.

^{page 23 note 2} While this way of writing is common in the present copy, I did not notice the same with reference to , although the latter is very common in the old MSS. Cf. Browne, E., “ Description of an old Persian Tafsīr,” JRAS. 1894, p. 433Google Scholar; also R. Nicholson, op. cit., v. ii, introd., p. 7, etc. In the old copy of the translation of Tafsīr-i-Ṭabarī (mentioned above, No. Aa 19 in the library of A.S.B.) is written together with the following word nearly always, while very rarely.

^{page 24 note 1} The frequency of its use may be attributed perhaps to Anṣārī's individual taste as well. I cannot remember to have ever heard in the colloquial similar cases as above (diminutive of is usually and is used only in säräk käshidän in the sense of ‘to spy, to watch stealingly’ in Fars). As far as I could notice, this suffix nowadays is never applied to abstract names in the colloquial, and although it is very commonly used (frequently in its shortened form -ä, cf. P. Horn, op. cit., p. 174), the suffix -chä is preferred with names of inanimate objects.

^{page 24 note 2} Jami (Nafaḥāt, p. 82) reproduces the sentence literally, showing that in his opinion the form was not obsolete.

^{page 24 note 3} I often heard this suffix in its full form in the remote corners of the Southern Khorasan, such as the villages Riqāt, Kubegū, Rich, etc. (district of Khūsp), especially in fairy tales : mardäkäki bud … or ruzäkäki, etc.

^{page 25 note 1} Cf. P. Horn, op. cit., p. 184. At present it seems to be more common in the eastern half of Persia and apparently can be traced to three different forms : (1) in the cases of its use with the vocative probably as a local variation of the vocative -ā used in standard Persian (in Isfahan -ī is in use), especially common with the names of women, e.g., Gawharū, Sultånú, etc.; (2) as an apocopated form of the diminutive suffix -uk, e.g., khurdú, kuchulú, etc.; (3) may be the remnant of the old suffix -ū (the same as in bānū, etc.).

^{page 25 note 2} This suffix, even if it is of purely Persian origin, is absolutely forgotten now, and, as an interesting example, it may be mentioned that the word (the name of the hero of a popular fairy tale) is invariably spelt Shirwiyä.

^{page 26 note 1} Jami (Naf., 161) reproduces the sentence intact, perhaps because this form did not seem to him obsolete.

^{page 26 note 2} Cf. Nafaḥāt, 243; the difference is only that is used instead of .

^{page 27 note 1} This particular word, as well as generally the suffix -āwand, is quite forgotten (khishú or qo’me khish is used instead). Perhaps it will be not quite useless to recall that the suffix for the names of the various step-relatives, —, mentioned by E. Browne in his description of the old Tafsīr (JRAS. 1894, p. 487), is still common in Southern Khorasan in mårändar, khårändar (), etc.

^{page 27 note 2} Jami reproduces this word as if it were quite all right (Nafaḥāt, 281).

^{page 27 note 3} Sometimes these formations are used in the same way of repetition, just as the monosyllabic onomatopœia: shälåpast-shälåpast, ghurghurrast, etc.

^{page 28 note 1} Perhaps this has something to do with mār, ‘ snake,’ to which the pains of hunger or thirst may be attributed ? In such case the first -ā may be similar to that often used in the copulative compounds, cf. P. Horn, op. cit., p. 196, as in sarāsar, barābar, etc. I remember once having heard the word gushnåmår in Khorasan, but did not pay attention to it because I regarded it as an invention of the man who was talking.

^{page 28 note 2} Ed. by V. Zhukovski, 1899. Unfortunately I cannot cite the page because I have not this book at hand.

^{page 28 note 3} The usual theory, based on the native grammars, that — is to be added to the names of the animate beings and — to those of the inanimate ones, is rarely correct, in any case as far as the colloquial is concerned. One hears invariably mardå, zanå, aspå, dukhtarå, etc. (å = , because the h is imperceptible in the pronunciation). On the contrary, dastú, churåghú, båzårú, etc., are freely used (ú = -ān).

^{page 29 note 1} Needless to say that in the colloquial and the dialects this form is absolutely extinct.

^{page 30 note 1} It is a well-known fact that in the colloquial very often a special suffix α, ä, e, u is used for the accusative : aspa biår, etc. The particle (which becomes rå, rä, ri in various districts) is frequently omitted even if there is stress on the accusative and if it is quite definite, as in ī bedey ‘ give (me) this ’, etc.

^{page 30 note 2} In Khorasan the construction with bär, wär is invariably preferred for the dative and only in one case it is similar to that used by Anṣārī, i.e. in the case of män, as frequently: bedey män, bugu män, etc. The possessive still preserves rå, but the direct construction with dåshtän is more common.

^{page 30 note 3} Cf. Horn, op. cit., 110, and R. Nicholson, Tadhkira, introduction, 9. The verb dåshtän not rarely receives this intransitive sense (as in Samnani, cf. Geiger, op. cit., 367). There are similar expressions in the colloquial, as injå chi dåräd ‘ what is here ’.

^{page 31 note 1} In the present colloquial the preposition is usually omitted when motion towards something is expressed: murum shār , burow khåna , etc.

^{page 31 note 2} The colloquial uses this form frequently, but the final -k is not pronounced, as in bishtärä, durtärä, etc.

^{page 31 note 3} Not used nowadays without the suffix tar.

^{page 32 note 1} Still in common use amongst the Heratis and in the south-eastern part of Khorasan.

^{page 33 note 1} This passage is not reproduced literally in the Nafaḥāt, where another expression is used.

^{page 33 note 2} This use of the reflexive pronoun, as far as I know, has only some analogy in Russian. In the colloquial of to-day in Khorasan, and especially amongst the Parsiwans of Afghanistan, khūd is often simply an equivalent of bā ‘ with ’. Khude qåfilä umad does not mean ‘ the caravan itself arrived ’ but ‘ he arrived with the caravan ’; khude û is frequently not ‘ he himself’ but ‘ with him’, etc. The further development of the same is khot, as in khot mu ‘ with us ’ (Sabzawari).

^{page 34 note 1} Although there are in the present colloquial forms, as ish guf, etc., they probably can be explained as syncopated demonstr. pr. i (in) with pleonastically used -ash, personal suffix of the 3rd person singular.

^{page 34 note 2} Nafaḥāt, 209