The Origins of Literary Persian
The Annual Noruz Lecture by a Distinguished Scholar of Iranian Studies
To choose the question of the origin of literary Persian as the subject of a lecture which is supposed to present the most recent results of philological and linguistic research may seem to be something of a paradox, as the origin of the Persian language is indeed well known. It has been known for more than a hundred years, since the great iranist James Darmesteter, in his Etudes iraniennes, published in Paris in 1883, demonstrated that New Persian is the medieval and modern continuation of the Old Persian language of Darius and Xerxes, just as Italian or French are modern continuations of Latin. It is a fortunate circumstance, and a rare one among the languages of the world, that we have monuments of the same Persian language at three stages of its history: Old Persian, with Achaemenian inscriptions, Middle Persian, with a number of inscriptions and the texts of the so called Pahlavi literature, and New Persian, which, for more than a thousand years, has been the main colloquial language of all Iranian countries and a highly valued literary language in a very large area, including half of the Asian continent, from the Balkans and the Persian Gulf to Central Asia and to Bengal.
In fact, the history of languages in the Iranian world is not so simple. It reflects the general history of those countries, with many migrations, invasions, empires built and destroyed, and centers of political power often moved from the North to the South and from the East to the West and back. As the middle of the Iranian plateau is a desert, it is understandable that the dominant power was never in the center, but sometimes on one side and sometimes on the other side. The history of the Persian language was conditioned by this eventful political and social history.
Old Persian was used by Achaemenian kings for solemn official inscriptions engraved on imperial monuments or high on rock cliffs, in a cuneiform script which practically nobody could read. It does not seem to have had any political or administrative function. The administrative language was Arama,ic: it was in Aramaic that royal orders were dispatched to the provinces and messages sent from the provinces to the center. As to the possible oral literary use of Old Persian, we know nothing. However it may be surmised that its use as a colloquial means of communication gradually spread in Fars and adjoining regions, as, after the period of the Greek rule and the subsequent period of Parthian domination, it reappears in the shape of Middle Persian as the dominant language in the South of Iran.
In the third century a. D., the inscriptions of the first Sasanian kings are written in three languages, (Middle) Parthian, Middle Persian and Greek. The archaic orthography of Parthian and Middle Persian is a proof that their tradition of writing had been established several centuries earlier. Parthian was the language of the Parthian rulers, who had progressively extended their empire from northern Khorasan over the whole of Iran and eventually settled their capital in Ctesiphon in Mesopotamia. Middle Persian was the language of the new Sasanian rulers, who were Persians, had risen in Pars, overcome the Parthian king and taken his place in Ctesiphon. At that time, both Parthian and Middle Persian were official languages, and there is no doubt that they were the most wide spread colloquial languages as well, Parthian in the northern part of Iran, Middle Persian in the southern part. Both were also vehicles of important cultural traditions. Although only small remnants of the cultural output of that time have been preserved, there is indirect evidence of intensive religious, philosophical and literary activity in both Persian and Parthian. Particularly Parthian minstrels were famous and their poetry highly appreciated. It was in their country that memories of old legends were transmitted from generation to generation and the epic tradition was constituted which was to receive its final and brightest shape in Ferdowsi's Shâhnâme.
During the four centuries of Sasanian rule, there happened considerable changes in the language situation of Iran. Parthian was soon ousted from official use: from the fourth century onwards, all inscriptions are in Persian only. On the other hand we know for sure, from an old Manichaean text, that in the third century people of East Iran, i.e. Khorasan, spoke Parthian and did not know the Persian language, for, when Mani, the prophet of the new religion, who himself preached in Persian, wanted to send a propagandist of his religion to Khorasan, he had to choose somebody who was conversant with the Parthian language. However, in Islamic times, there is no evidence of any survival of that language in Khorasan, and, at the present time, all the dialects of that region are local forms of Persian. This is all the more remarkable as there are still many Parthian like dialects in the north western provinces of Iran. It is obvious that, in the Sasanian period, Persian as a spoken language spread widely outside southern Iran, towards the North and the East, especially in Khorasan, where it completely superseded the Parthian original language. The Danish iranist Arthur Christensen has hypothesized that this was the result of the establishment of military colonies in Khorasan in order to protect the empire against raids from Central Asia.
After the Arab conquest, the colloquial use of Persian was extended still further East, to Transoxiana and present day Afghanistan, by muslim soldiers and merchants. It may be supposed that Persian, more than Arabic, was the main instrument of *e Islamic propaganda among the peoples of those countries. Eventually it superseded the original local languages, Soghdian and other dialects. And it was there, in the far East, that it became again a literary language. True, literary Middle Persian was not yet out of use. It was practiced by Zoroastrian communities, and even the main body of what is left of the so called Pahlavi literature was produced in the ninth century in the South of Iran. But it was practically restricted to Zoroastrian religious writings. Nearly at the same time a new literary instrument was created in the East on the basis of the colloquial language of those regions. According to the Târikh-e Sistân, it was in Sistan, a little after the middle of the ninth century, that the first lines of New Persian poetry were composed. But it was in Transoxiana and Khorasan, in the Samanid kingdom, that Persian poetry produced its first masterpieces, with such poets as Shahid e Balkhi, Rudaki, Abu Shakur, Daqiqi, and the great Ferdowsi, who subsequently, from the end of the tenth century onwards, inspired many other poets and writers all over the Iranian world, and then also in India, in Central Asia and in the Turkish states of Asia Minor.
Thus, Persian is originally an Iranian dialect of the extreme South West which, as a spoken language, progressively spread from the South to the North and the East, until in the extreme North East it became a literary language, which in turn spread out to the West and the South and ultimately was cultivated in a large part of Asia.
So far the history of the origins of the Persian language ace clear. However there remain a few enigmas, which I now would like to mention.
(1 & 2) The first one, or, to be more exact, the first two ones, are proposed to us by Ebnolmoqaffa`, that celebrated Iranian scholar who was the founder of Arab prose and, among other works, translated the Book of Kalila and Dimna from Middle Persian into Arabic. In an account of the languages of Iran (obviously, as they were towards the end of Sasanian times), he says that there were five languages, to wit Pahlavi, Dari, Parsi, Khuzi, Soryâni:
Pahlavi, he says, is the language of Isfahan, Hamadan, Nehavand and Azerbayjan;
Dari is named after the court, i.e. the royal court (Persian dar "door", meaning also, like dargâh, "court, palace"); it is the language of the court and the cities of Madâyen, i.e. Ctesiphon; and he adds something that seems to mean that, among the languages of Khorasan, that of Balkh is the best Dari;
Parsi is the language of Fars and it is the language of mowbads, i.e Zoroastrian priests, and scholars;
Khuzi is the language of Khuzistan; it is used by kings in domestic occupations;
Soryâni is the language of the plain of Irak, and it is used in letters.
How are we to understand this account? Soryâni and Khuzi do not call for comments: Soryâni is clearly Aramaic; Khuzi, the language of Khuzistan, might be a remnant of Elamite. The other three languages are Iranian. Pahlavi is undoubtedly the Parthian language: Ebnolmoqaffa` ascribes it to regions of ancient Media, where there are still to day dialects that are closely akin to Parthian. By the way, it is worth mentioning that "Parthian" is the original and proper meaning of the word pahla vi, and it is only in the ninth or tenth century that this name, strangely enough, was applied to Middle Persian; it was then taken to mean only "ancient": to people of that time Middle Persian was "the old language".
Now, Parsi, in Ebnolmoqaffa`s account, the language of mowbads and scholars, is assuredly literary Middle Persian. But it is not clear how it can be said to be also the language of the province of Fars. Although it had originated from Fars, by that time Middle Persian was the written, official, administrative and literary language of the whole empire, and it was also widely used in colloquial intercourse: at first sight there seems to be no reason why it should be especially ascribed to Fars. This is our first enigma.
The second one, about Dari, is more important. It directly concerns the origin of New Persian, for we know, from many Arabic and Persian texts, that, during the first centuries of the Islamic period, the name dari was applied to literary New Persian. For instance, Bal'ami, in the Arabic preface of his translation of Tabari's history, written around 963 A.D., says: fa qad tarjamtuhu be lughat al fârisiyya al dariyya "I translated it into Parsi Dari". The translators of Tabari's large commentary of the Koran, who worked at the same time, write: in ketab tafsir e bozorg ast az revayat e Mohammad ebn e Jarir at Tabari tarjome karde be zaban e parsi o dari (or parsi e dari) "this book is the large tafsir transmitted by Mohammad b. Jarir Tabari, translated into Parsi Dari". The author of a poem on medicine called Daneshname, composed in 978 980, says that he wondered whether he should write it in tâzi (Arabic) or dart, and then he adds:
dari guyam sh tâ her kas bedânad
vo her kas bar zabân ash bar berânad
“l will write it in Dari so that everybody may
"Understand it and everybody may recite it"
Ferdowsi in his Shâhnâme several times mentions the Dari language, for example, when relating how the Book of Kalila and Dimna was brought from India to Iran, and the translated into Arabic; then, he says,
be tâzi hami bud tâ gâh e Nasr
bedân gah ke shod bar jahân shâh e asr
gerânmâye Bolfazl dastur e u
ke andar sokhan bud ganjur e u
befarmud tâ fârsi o dari (or fârsi e dari)begoftand o kutâh shod dâvari
"It remained in Arabic until the time of Nasr (the Samanid king), "When he became king over the world:
"Noble Bolfazl (Bal'ami), his minister,
"Who was his treasurer in literary matters,
"Had it translated into Parsi Dari,
"So that the question was solved"
Some eighty years later, towards the end of the eleventh century, the author of the Qâbusnâme, speaking of the art of writing letters in Persian, says that it is not commendable to systematically avoid Arabic words: parsi e motlaq manevis ke nakhosh bovad, khasse parsi e dari "do not write pure Parsi, for it is unpleasant, especially Parsi Dari", which seems to mean that there are other kinds of Parsi.
This name of Dari applied to literary Persian has given rise to many discussions among Iranian scholars and to many hypotheses. Why was Persian not called simply Parsi (or Farsi)?)? The account of Ebnolmoqaffa` led me, twenty years ago, to the idea that, at the end of the Sasanian period, the languages he calls Parsi and Dari respectively were only two varieties of the Persian language: the written one, i.e. literary Middle Persian, which had been standardized many centuries before in an archaic shape, and the colloquial, which was much more advanced and for this reason had been given another name. I still think that this is true on the whole. But it does not explain everything. In particular it does not explain why, in the tenth and eleventh centuries, New Persian was persistently called Dari or Parsi Dari and with which other kind of Parsi it was being contrasted, for instance, by the author of the Qâbusnâme when he says: "Do not write pure Parsi, especially Parsi Dari". On the other hand one may wonder why Ebnolmoqaffa`, speaking of Dari, particularly mentions the capital city and Khorasan and no other part of Iran.
3) Another enigma is the question of the nature of the language used in old Jewish Persian texts, i.e. Persian texts written by Iranian Jews in the Hebrew script. Old translations of the Bible by Jews have been known for more than a hundred years: they were studied by such well known scholars as Noldeke and Horn. Since that time, other texts have been discovered, some of them published: two of them date back from the tenth and the eleventh centuries; other texts bear no date, but they are certainly quite old too.
Now, the language of these texts Is rather disconcerting. It is Persian, but a kind of Persian very different from the language of the literature in Arabic script. It abounds in words and grammatical forms that do not occur at all in literary Persian. For instance the first person plural pronoun is not mâ, but imâ; the verbe "to be" has forms like hoary, hi, him, hand, instead of hastam, hasti, hastim, hastand, and bid, bend, instead of bâshid, bâshand or bovand there are words in eshn or esht, instead of esh, e.g., e.g. âfarineshn "creation", koshesht "fighting", buzesht "receipt", etc. A especially interesting feature is the presence of many words unknown in literary New Persian, but well known in literary Middle Persian, such as ashgahân "lazy", abâri "other", ustâm "confidence", hamimâl "opposite" This fact led some scholars, a hundred years ago, to wonder whether there had been particular connexions between the Jews and the Zoroastrians who wrote Middle Persian books in the ninth century. Other scholars ascribed these archaic features to the allegedly conservative spirit of the Jews. As for me, I wrote some twenty five years ago, and I still think, that the peculiar features of the language of old Jewish Persian texts are not specific to the language of the Iranian Jews, but they are rather local dialectal features of southern Iran: some of these documents were written in Khuzistan, and I was able to demonstrate that some features were skin to the local dialect of Shiraz. Still the question remained why there is no trace of these local peculiarities outside the Jewish Persian texts, and how to explain the similarities with literary Middle Persian.
4) The last enigma concerns the bizarre kind of written language called Pazand. It is found in Zoroastrian manuscripts written in India. It is a mere transcription of literary Middle Persian In Avestic script. After the establishment of the Islamic rule, some of the Zoroastrian communities of Iran took refuge in India with their sacred writings. After some time, as literary Middle Persian is to a large extent cryptographic and very difficult to read, they found it convenient to transcribe their texts into the Avestic script, which is fully vocalized and perfectly clear. Thus, the language of those Pazand texts is nothing else but literary Middle Persian. However there are in it some strange words that are found neither in Middle Persian nor in New Persian. For instance, from the verbs udan "to "to be" and shudan "to "to go", we have bahod "he is" and shahod "he goes", instead of classical Middle Persian banged and shawed (corresponding to New Persian bovad and shaved). Where do these forms come from? This question, however minor it may seem to be, is interesting, as we shall see in a few minutes.
A few years ago, a distinguished Iranian scholar, Mr. Ali Ravâqi, discovered a very peculiar manuscript in the library of the Astân e Qods in Mashad, and he undertook to publish it under the title of Qor'ân e Qods. He was kind enough to send me first an extract of his future publication, and then the book itself (two magnificent in folio volumes!). A careful examination of that text led me to conclusions and hypotheses which, I believe, fully solve all the problems I have just mentioned.
This manuscript is a translation of the Koran, or, more exactly, it is a copy of the Koran in Arabic, written in large beautiful script, with interlinear glosses in Persian. Being mutilated at the beginning and at the end, it bears no name or date. I think it may have been written towards the end of the eleventh century. As to its place of origin, on the basis of certain pieces of linguistic evidence I presumed it was in Sistan or some neighbouring region. Ravaqi, independently, reached the same conclusion, so that it may be considered pretty sure.
The interesting point is the very peculiar kind of Persian used in that manuscript. As a specimen I would like to quote a few lines from the beginning of surat 26 (ash shu`arâ):
verse 2 ân âyatha ye ketâb didvar hand "those signs of the book are clear";
verse 3 shâyad ke to koshtâr i nafs e to râ ke naband mo'menân "perhaps you kill your own soul (you are afraid) they might not become believers";
verse 5 nayâmad bedishân hich ayâdkardi az xodây now yâ ney budand az ân ruy gardânestârân "no new reminder came to them from God that they did not turn away from it";
verse 8 dar ân neshâni va na budand gwishtar e ishân mo'menân "there is a sign in it and most of them did not become believers".
In these few lines there are several interesting words. gwishtar instead of bishtar is an example of a characteristic phenomenon in this text: in all words that in standard Persian begin with b from older w , we find here gw or 9 , other examples: gwâd "wind" for bâd, gwâfte "woven" for bâfte, ginad "he sees" for binad, gahisht "paradise" for behesht This is one of the points that directed my attention towards Sistan, as this phenomenon is common is Balochi.
Other interesting forms are ayâd "memory", which is an archaic form of standard Persian yâd, and, from the verb "to be", hi, hand and band for hasti, hastand end bovand You may remember that a moment ago I mentioned similar forms in Jewish Persian. This is indeed the first important result of the investigation of that Koran: its language is very similar to that of the old Jewish Persian texts. It would be tedious to go through a whole list of the peculiarities shared by the Qor’ân-e Qods and Jewish Persian. Suffice it to say that those I mentioned a few minutes ago in connexion with Jewish Persian are also present in the Mashad Koran, among many others. Especially important is the fact that in both kinds of texts a large number of words are found that are usual in literary Middle Persian and do not occur in standard literary New Persian, a fact which points to some kind of special relationship between the language of the Qor’ân-e Qods, Jewish Persian and literary Middle Persian. Since there is no reason to imagine any influence of the Zoroastrians upon the pious Muslim who took pains to copy and translate the Holy Koran, these peculiarities must be explained otherwise.
How are they to be explained? More generally, what does this state of things mean for the understanding of the emergence of the New Persian language? It means a great deal.
First, it means that the peculiar features of the Jewish texts are not specifically Jewish, but they are indeed local peculiarities.
Secondly, as some Jewish texts originated from Khuzistan and others presumably also from the South West, and the Koran was written in or near Sistan, in the South East, their common linguistic features reflect the kind of Persian spoken both in the South West and the South East, i.e. in the whole southern part of the Iranian plateau. It is true that the language of the Kuran is not wholly identical to Jewish Persian, but, for that matter, there are also small divergences between different Jewish Persian texts. This is because these different Jewish documents and our Koran were written in different parts of southern Iran and, consequently, represent different local dialects of spoken Persian. But these slight divergences are outweighed by the many common features which distinguish all of these texts from standard New Persian.
Thirdly, as literary New Persian was first formed in the North East, and practically all that was written in that language during the first period, i.e. the tenth century, originated in Transoxiana and Khorasan, standard literary Persian primarily reflects the language spoken in the North East of Iran.
Fourthly, as the language of the literature in Arabic script, which is founded upon the spoken colloquial of the North East, and that of the Jewish texts and the Qor~an e Qods, which represents the colloquial of the South, are rather different, one is led to conclude that, during the first centuries of the Islamic era, the Persian language, as it was spoken throughout Iran, was divided into two main dialects, a southern one and another one which was used at least in the North East and possibly also in the North West.
Moreover, it is a striking fact that the southern dialect was much closer to the original form of Persian, i.e. Middle Persian, while the northern dialect had developed further away from it. The South is the original home of the Persian language: it is not surprising that the language in the South remained relatively close to the etat de langue represented by literary Middle Persian, which had arisen and been standardized, many centuries before, presumably in the South West. This is why Jewish Persian and the Qor’ân-e Qods have many words that exist in Middle Persian and are not found in standard New Persian.
On the other hand the northern dialect was a newcomer in its territory. It was a language which in the Sasanian period had spread from its native country in the South into the territory of another language, Parthian, and then, in the Islamic era, further East into Transoxiana. It had progressively eliminated Parthian, at least in Khorasan, and then Soghdian in Transoxiana. No wonder if, in this huge extension of its realm, it had undergone important changes in its grammar and its vocabulary. We know that it absorbed a large amount of loanwords from Parthian. The German iranist Wolfgang Lentz, already in the twenties, published a long list of northern, i.e. Parthian, words in literary New Persian; and other scholars, W.B. Henning and V.A. Livshits also detected a number of Soghdian words: certainly other loanwords may still be discovered.
To what extent literary New Persian was pervaded with words of nonPersian origin is obvious when one considers a piece of Manichaean poetry which was published by Henning thirty years ago. This poem exists in two versions, one in Middle Persian, the other in Parthian. Anybody conversant with New Persian cannot but be puzzled by the following fact. The syntax of the Middle Persian version is undoubtedly more similar to that of New Persian. But it is the reverse with the vocabulary: it is in the Parthian version that words identical or nearly identical with New Persian words are found. It is in that version that "great" is wuzurg, "bright" is rawshan, "hail to thee" is dorud abar tô. The Middle Persian has other words, which sound strange to the ears of New Persian speaking people. This means that many words that belonged to the original Persian stock were replaced in the northern dialect of spoken Persian, and consequently in literary standard New Persian, by other words taken from Parthian or other northern Iranian dialects.
Thus, we may represent the language situation of Iran during the first centuries of Islam in the following way. As to the spoken language, if we leave non Persian dialects, Parthian or others, which were in use in different parts of the country, out of consideration, there were two dialects of the common colloquial language, i.e. Persian, a northern one, whose vocabulary and grammar were rather different from Middle Persian vocabulary and grammar, and a southern one, which had remained remarkably close to Middle Persian. As to the written and literary languages, the most widely used was Arabic, but, from the middle of the ninth century onwards, in the East, local poets had begun to compose verse in their mother language, which was the northern dialect of spoken Persian. In the South, Zoroastrians continued for some time to use the traditional literary language, i.e. literary Middle Persian, which for centuries had been their written language; presumably, in the first centuries of Islam, it was still considered a language not really different from their colloquial, but only the high style of their own language.
On this particular point, I mean the question of the sociolinguistic relation of literary Middle Persian to southern spoken Persian, again the Qor’ân-e Qods sheds some interesting light. I said a moment ago that Pazand, which is a transcription of literary Middle Persian has some strange words whose origin is obscure: as an example I quoted bahod instead of bovad. Now, this very form is found repeatedly in our Koran, for instance:
surat VIII verse 7 mi dust dârid ke jod xodâvandân e selâh babud shomâ râ “you want the group other than the armed men to be yours"
surat XIX verse 35 o zamân qazâ konad kâri gohad ân râ bebâsh bebabud "when he decides something, he says to it <, it is".
There are still other peculiar words that are found both in Pazand and in that Koran, and nowhere else.
Again the question is: what does this mean?-
The Indian Zoroastrians who transcribed Middle Persian into Pazand undoubtedly did it according to the traditional pronunciation inherited from their ancestors who had emigrated from South Iran to India. Thus Pazand reflects the reading of Iranian Zoroastrians at the beginning of Islamic times. If in that transcription we find bizarre words, such as bahod, which belonged to the usual spoken language of South Iran, as proved by their occurrence in the Qor’ân-e Qods, that means that the Zoroastrians of South Iran, at that time, pronounced literary Middle Persian (Pahlavi) just like their own colloquial language. Which suggests that Pahlavi for them was not a different language, something like Latin for modem Europeans, but rather the written form of their own usual spoken language, only a more archaic and more dignified style of the language they spoke in every day communication. This is all the more understandable as their colloquial language, the southern dialect of Persian, was not distantly removed from what we modern linguists call Middle Persian.
We are now in a position to come back to Ebnolmoqaffa`s statement. We can understand why he says that Parsi was both the language of mowbads and scholars and the language of Fars. The language of mowbads and scholars was literary Middle Persian. The language of the people of Fars was southern spoken Persian, which obviously, towards the end of the Sasanian period, was not appreciably different from what it was in the ninth or tenth century. If Zoroastrians in the Islamic period considered literary Middle Persian to be the written, learned form of their colloquial language, all the more was it so in the preceding centuries. That language was called Parsi.Its written variety was the official, administrative, religious and literary language of the whole empire.
Now what about Dari? I think the answer is obvious. It emerges from what i have said about the Qor’ân-e Qods, the significance of its linguistic kinship with Jewish Persian and Zoroastrian Pazand and of its divergence from literary New Persian, and the division of spoken Persian into two main dialects. Southern Persian was the direct descendant of the language of the Persian tribes who in ancient times settled in the South West of Iran, among the Elamites, whose language they progressively restricted and eliminated. It also extended towards the East up to Sistan, at a time and under conditions we do not know. The northern dialect was the result of the extension of Persian, during the Sasanian era, into the domain of Parthian, where it was considerably influenced by the dialects it superseded, especially by the Parthian language, lost many original words and took on new words from the former languages of that territory, and thus developed appreciable divergences from Southern Persian. No doubt, at the end of the Sasanian age, people did pay attention to how Persian was spoken in the different regions of Iran and they were aware of the differences. Suffice it to remember how, in the tenth century, the geographer Moqaddasi carefully notices the peculiarities, however minute, of the language of each city he describes.
What does Ebnolmoqaffa` say about Dari? He says that Dari is the language of the cities of Madâyen, that it is spoken by those who are at the royal court, and that it refers to presence at the court. The last remark obviously is meant to explain the name Dari as derived from dar "door" and "royal court". It appears that Dari was only a spoken language: wa bihâ kâna yatakallamu man bi bâb al-matik. It was spoken at the court and it was the language of the capital. What other language can it be than the northern dialect of Persian? Madâyen, or Ctesiphon, before the rise of the Sasanian dynasty, had been the capital of the Parthian kings, and presumably at their court, and also in the city, people had spoken Parthian. When it became the capital of Persian kings, the Persian language must have become the usual language of the court, and to a certain extent also that of the city. Thus, in Ctesiphon, as in the North of Iran, the common colloquial was Persian established on the former domain of Parthian. Dari is nothing else but the northern dialect of Persian. As, towards the end of the Sasanian period it was felt to be rather different from the southern dialect, it was given an new name: it was called Dari "the court language", or perhaps Parsi Dari "court in contradistinction to Parsi proper.
As for the mention of Khorasan and the East and Balkh, we know that these were regions where Persian (Dari) had entirely superseded the former languages, be it Parthian or Bactrian. What Ebnolmoqaffa` says implies that he takes for granted that Dari is the only language there, for he only says that the people of Balkh are the best spoken people of the East. This remark is corroborated by Moqaddasi, who, in the tenth century, states that the language of Balkh is the best one, especially for writing letters!
There remains of course the question of the North West, i.e. ancient Media, the region of present day Tehran, Isfahan, Hamadan and Tabriz. We do not have information about how far Persian, as a colloquial language, was in use in that part of Iran. Ebnolmoqaffa` says that Pahlavi, i.e. Parthian, was the language of the region, which is true, as many Parthian like dialects are still alive there in villages. But it is likely that Persian, in its Dari variety, had penetrated there also. Ebnolmoqaffa` says nothing about it: perhaps he was better satisfied by neatly assigning a specific geographical domain to each one of the languages he mentions, Fars to Parsi, Media, which he calls Fahla, i.e. Pahlow, to Pahlavi, and the capital (and the East) to Dari.
Since Dari was the name of northern Persian, it was only natural that literary New Persian, which was established on the basis of north eastern dialects, should also be called Dari, especially at a time when literary Middle Persian was not yet completely out of use: Zoroastrians had written books in it in the ninth century, and Arabic writers very often make mentions of Parsi books, meaning of course books in literary Middle Persian, and we know that in the tenth and eleventh centuries, many people in Iran were still fond of Pahlavi literature, as Gorgâni, the poet of Vis o Râmin says:
dar in eqlim ân dafter bekhânand
bedân tâ pahlavi az vey bedânand
kojâ mardom afar in eqlim hamvâr
bovand ân lafz e shirin râ kharidâr
In this country they read that book
"In order to learn Pahlavi in it,
"For people in this country always
"Are fond of that sweet language."
Other writers also bear witness to the relative popularity of Middle Persian literature down to the eleventh century at least. This is why those who wrote in the new literary language, in New Persian, found it useful to specify that they were not using the old literary Parsi, but another kind of Persian, written in Arabic script and called Dari. It was Parsi too, but of a special kind, Dari Parsi. Later, when literary Middle Persian was practically forgotten, except among Zoroastrians, the name Dari became useless: it was dropped, and New Persian was called simply Parsi or Farsi, as it is to day.
It is time to conclude. The establishment of a national language is always and everywhere a complicated matter. It implies the choice of a basic dialect, then contributions by other dialects, often the imitation of another more prestigious language, sometimes the struggle with a concurrent language, etc. The story of the origins of literary Persian is especially complicated, due to the vicissitudes of the history of Iran. But it is also typical. It is rooted in the long and glorious past of Iranian civilization. It involves a multifarious heritage: a descendant of south western Achaemenian Old Persian, a continuation of the official language of Sasanian Iran, the Persian language was impregnated with contributions from other Iranian languages, especially Parthian. Thus it is heir to both main branches of the Iranian languages, just as the great Persian literature is heir to both Middle Persian wisdom and Parthian poetry.
I have finished. I thank you very much for your kind attention. And, since this is a Nowruz lecture and Nowruz is indeed in a few days, I wish you a happy New Year. Nowruz e shomâ farkhonde bâd!