Why Persian?

A Redditor asks:

So I was doing a bit of reading on Islamic Empires and one of the things that I noticed was that a lot of Empires (Mughal, Timurid, Delhi Sultanate….) all chose Persian as their administrative language, but why Persian? Arabic was literally the language during the golden age and the religious language of Islam what made Persian so special?

There are a number of responses; the best is by Draig_werdd, and I’m reproducing it below (silently correcting a few errors in spelling or grammar):

There is a bit of historical context that explains this situation.

At some point soon after the Arab conquests, the area in Central Asia known as Khorasan and Transoxiana but now in Uzbekistan, Tajikistan, North Afghanistan and parts of Turkmenistan and NE Iran became Persian speaking. Other Iranic languages used to be spoken there in the past.

The locals, or at least the elites, also quickly adopted Islam – faster than in the territory of current Iran itself. This created a very unusual situation compared with other areas conquered by the Arabs. This was a very developed region, with an urban culture, a dense network of local nobles speaking a prestigious language but also Muslim. Unlike Aramaic or Greek, this was not a language of a defeated people of a different religion. These people were unhappy with being second class Muslims, so they initiated the Abbasid Revolution (which started in this region and was led by a local Persian speaker). Once in power the Abbasids brought with them a lot of Persian speakers (some of the most powerful people in the early Abbasid state were the Barmakid family, Persian speakers and former Buddhists). Persian culture became accepted as a proper Muslim culture and a lot of traditions were adopted. By the 9th century, the area become more and more autonomous, and eventually four brothers founded the Samanid state. The Samanid rulers viewed themselves as Persian, later as Persian kings descended from the Sasanian Empire. They patronized a vast number of local artists, especially in literature, like the most important figure of Persian literature, Ferdowsi.

The Samanid state is not only important for giving birth to the Medieval Persian culture but it was also the main contact point of the Islamic world with the steppe Turkic people. The Samanid state itself was destroyed by the Turkic people, but they adopted the existing, Persian, administration and language. This was the model for future Turkic states, and they spread it wherever they went, both to India in the East but also to Anatolia in the West. Ottoman Turkish vocabulary was 88% of either Persian origin or Arab origin, but adopted from Persian

The Khorasan and Transoxiana region remained one of the biggest cultural hubs of the Islamic world (indeed a lot of famous medieval scientists and writers were from there) until it was severely destroyed by the Mongol invasions. By the 15th century the area was no longer majority Persian speaking, although many cities remained Persian speaking until the 20th century (under the name of Tajik).

Nothing surprising there, but it’s a useful roundup for those who don’t know the history, and having discovered to my shock that I’ve never mentioned the Samanids at LH, I’m glad to remedy the omission and name-check one of my favorite medieval dynasties. Thanks, Bathrobe!


  1. David Marjanović says

    silently correcting a few errors in spelling or grammar

    You left the very common was lead alone, though. The ambiguity of read has a lot to answer for.

  2. J.W. Brewer says

    “Lead” is ambiguous in its own fashion, representing in some circumstances a word homophonous with “led.” But I don’t think hat was warranting that he’d corrected all such errors, what with him being retired and all. (“No man but a blockhead ever copy-edited, except for money.”)

  3. You left the very common was lead alone, though.

    Fixed, thanks. And JWB is quite right about my warrant. If you want a thorough job, pay me!

  4. David Marjanović says

    “No man but a blockhead ever copy-edited, except for money.”

    Ah, that’s in the real world. I basically copyedit when I review a manuscript, because at most journals it simply wouldn’t get done otherwise.

  5. And “then in the the”. (Meaning “than in the”?) Several frank errors in punctuation also, and sundry infelicities. This:

    “The locals, or at least the elites also quickly adopted Islam, faster then in the the territory of current Iran itself.”

    It might be amended to this instead:

    “The locals, or at least the elites, also quickly adopted Islam – faster than in the the territory of current Iran itself.”

    Assuming that is what’s meant, which is by no means certain.

  6. Oops. I ran out of amending time. Amend to this, rather:

    “The locals, or at least the elites, also quickly adopted Islam – faster than in the territory of current Iran itself.”

    My invoice will follow.

  7. Amend to this, rather

    Easily copied and pasted, thanks!

  8. second class Muslims –>

    second-class Muslims

    By the 9th century, the area become more and more autonomous –> EITHER

    By the 9th century, the area had become more autonomous OR

    By the 9th century, the area was becoming more and more autonomous

    a lot of –> many

    the area was no longer majority Persian speaking, although many cities remained Persian speaking –>

    Persian was no longer the language of the majority in the area although in many cities it was still spoken

  9. Eh, people will get the idea. It doesn’t have to be perfect prose, just comprehensible.

  10. a lot of Empires (Mughal, Timurid, Delhi Sultanate….) all chose Persian as their administrative language, but why Persian?

    Not only the imperial elites, but also some subjugated peoples showed a noteworthy preference for Persian. The traveller and adventurer E.B. Soane, in his remarkable To Mesopotamia and Kurdistan in Disguise (1914), describes the attitude of Kurds and Assyrian Christians towards Persian in Ottoman Erbil of a century ago:

    My Kurdish neighbour, I found, knew a little Persian, and had been to Teheran and Kashan. He introduced himself as a Kurd of the Mukri, a native of Sauj Bulaq, the Mukri capital, and lamented the fate that kept him in Turkish territory mending shoes. Here I began to get in contact with the sentiment I found often expressed by Christian and Kurd alike all over southern Kurdistan and eastern Turkish territory, a leaning towards Persian rule and custom, and an emphatically expressed aversion to all things Turkish. Among the Kurds this sentiment takes so strong a form, that many of them set themselves to make a study of the Persian language, and employ it in all transactions requiring writing, never using Turkish unless forced to do so.

    Soane spoke Persian fluently, for he had lived in Shiraz before undertaking his journey in Ottoman Kurdistan and passing himself off as a Persian there. Later in his travels, he stays with Lady Adela, ruler of the Jaff Kurds, who is eager to learn Shirazi idioms from him and remarks upon hearing Soane’s speech:

    “Bravo ! that is the true Persian speech, the sweetest of all God’s languages.”

  11. Very interesting, thanks for that! I have to say Persian is a pleasing language, easily learned (for everyday purposes — I’m not talking about taarof or other subtleties) and the medium of a great deal of splendid poetry.

  12. David Eddyshaw says

    Yes: it is probably not irrelevant that Persian is a (comparatively) easy language, in the sense that relatively little effort is required from the learner before significant benefit accrues (which of course is not at all inconsistent with the fact that the nuances need a lifetime to acquire.) It’s at the English/Spanish/Hausa end of that spectrum, rather than the Russian/Arabic/Fulfulde pole (to say nothing of Navajo and Ingush.)

  13. @DE, “benefit” is often understood as “after a few lessons I will be able to discuss how I spent the summer” and accuracy is understood as “no errors” where an “error” is understood as “what a teacher calls an error”. From this perspective Russian is a nightmare.

  14. Accordingly, Persian is first easy and then a nightmare:)

  15. David Eddyshaw says

    Quite so.

    I have seen it confidently asserted that Russian gets a lot easier after the formidable initial hurdles (from a western European standpoint, anyhow) because Russian syntax is comparatively simple. However, I strongly suspect such statements of being hangovers from the days when the study of syntax in general was much less advanced than it now is. (I never got to the stage in my school study of Russian of being able to have a meaningful opinion on the matter, alas. Never even took O Level in it.)

    There are certainly languages where the reverse is true, though: the initial impression of relative simplicity turns out to be a cruel deception. Hausa is like that …

    Even so, initial appearances of simplicity probably help towards widespread adoption. By the time the Awful Truth becomes clear, it’s too late. You can’t get by in daily life without using the language any more, and have to reconcile yourself to a lifetime of being patronized by L1 speakers for your quaint accent and amusing mistakes.

  16. David Eddyshaw says

    (Probably being patronized by your children, too, though you can at least be revenged on them by discussing them in their presence with your friends in good Old Countryese and knowing that they only understand enough to know that you’re making fun of them in some way.)

  17. I think I meant more or less this:

    There is also inconsistent classification of the language in terms of difficulty for English speakers. It is important to note that Persian (when taught as the modern language of Iran) is relatively easy to begin with, because some of the basic vocabulary is cognate with English and the syntax is similar. Beyond the initial hurdle of a strange alphabet, the student finds relatively simple sentences with familiar structures. Partly for this reason introductory Persian classes are often relatively large. However, few students progress far into the intermediate level because of the increasing need to deal with vocabulary, syntax, and usage that are culturally alien to English speakers as a result of the high degree of convergence with the major non-Indo-European languages in the region, Arabic and Turkish, as well as the importance of imported Arabic vocabulary. Enrollments in Persian (when taken for essentially nonacademic, but now common, purposes) tend to fall off sharply after the first year, further endangering its future in the curriculum.
    (from 1 and also 2)

  18. I thought about differently, though: that at some moment studetns fce the literary register….

  19. Bathrobe says

    an emphatically expressed aversion to all things Turkish

    What’s wrong with Turkish? (From the student of an “Altaic” language.)

  20. @DE, there are easy languages. And there is somethign about Hausa that made me want to play with it.

    But my idea is rather that our expectations and goals, and also what side we approach this forest (I visualize it as a forest…) from – the method if you like – seriosly affect both your measure (“benefit”) and our ability to enjoy the process.

    I know 1 (one) learner who found Russian “easy”. She is 20-something, she speaks ENglish and it is her first foreign language ever. I suspect that unlike some much more experienced learners she did not do what Spanish learners and teachers do, instead she just watched videos like this one and it was easy…

  21. David Eddyshaw says

    From your link (1):

    “Urdu (a creole of Persian superimposed on an Indic grammar)”

    Oh, no it isn’t. I do wish people wouldn’t fling about terms like “creole” like that. I blame that McWhorter fellow (though he at least knows what the word actually means, unlike Spooner.)
    Spooner talks utter nonsense about the Korean script further on, too.

    I don’t think the cognacy of basic Persian vocabulary with English is actually at all helpful in practical terms. There are about a half dozen words where the resemblance might help someone unfamiliar with comparative linguistics, and one of those (bad) is not in fact cognate at all. Nor is the syntax really much like English at all. What is helpful is the absence of grammatical gender and case (and even of complicated plural formations) and the relatively simple verbal system (which isn’t much like English either.) The phonology is relatively simple for English speakers, too (certainly compared with Arabic.) And no nasty mutations or vowel harmony or tones. Or horrid agglutination.

    Is it really the case that practically half the population of Iran have an L1 other than Persian? The CIA seems to agree:


    So, Yes, I suppose.

  22. I don’t know if “…would not be likely to rank very high on an international scale of functionality or simplicity” is utter nonsense (what is “functionality”?) but it’s a poor example.

  23. David Eddyshaw says

    I think it betrays total unfamiliarity with the Korean script. I suspect that he has heard that it’s an alphabet, but that he’s confused by the way it looks, and knows nothing at all about how it actually works.

    The only real complication with it is a tendency to preserve the underlying shape of morphemes in the spelling regardless of internal sandhi changes, which its a pretty venal sin compared with most alphabetic writing systems.

    So yes, utter nonsense. Like calling Urdu a creole.

  24. David Eddyshaw says

    He also rather sloppily implies that almost half of the population of Iran speak Azeri at home (Azeri specifically, as opposed to not-Persian in general), which seems to be pretty far from the case.

  25. I was greatly amuzed by an article that connected the fact that Balochi speakers can’t comfortably read in Balochi to that the Arabic script was intended for Arabic (true) and is not suitable for Indo-Iranian.

    This all may well be true, but an explanation of how the very same speakers manage to read and write in Urdu and Persian is missing:) Compare.

  26. Oh, Spooner wrote about Balochi!

    Meanwhile, although literacy rates in the region have risen significantly over the past generation, the rise in Balochi literacy has been slow, and generally negligible. Literacy for most Balochi-speakers is not in Balochi, but in Urdu in Pakistan and in Persian in Afghanistan and Iran. Even now very few Baloch read Balochi, in any of the countries, even though the alphabet in which it is printed is essentially identical with Persian and Urdu. Despite efforts to make Balochi a medium of written and print communication, Baloch who read Urdu or Persian comfortably and could easily make out Balochi on a printed page, claim to find it illegible, though there is no difficulty for a non-native speaker. This is a situation which needs to be explained both in terms of the way people read and of the historical status of the three languages. In general people read by unconscious recognition of the shapes of common combinations of letters, rather than by phonetic construction of words from individual letters. This is especially true of languages written in the Arabic script, because writing works in terms of established penstrokes including particular serial combinations of letters, rather than by the simple connection of whatever letters are required to form any particular word. Most of the established pen-strokes in Urdu are different from those in Persian. Those needed for writing in Balochi are different again, and therefore unfamiliar to readers already accustomed to either Persian or Urdu and difficult to read. In general also there is still an historical expectation that writing should be in Persian or Urdu, rather than Balochi. But without a growth in literacy rates or increase in use in electronic (non-oral) media Balochi is unlikely to achieve standardization, or to increase in national significance. Currently, however, its use is central to local political as well as cultural identity within Baloch-majority areas. This is due to its historical use as the medium of public life among the Baluch, and as the medium of interaction between subcommunities who have come to consider themselves Baloch even if they retain the use of another language

  27. Very interesting!

  28. @DE, how do you create a writing system?

    Step 1: invite specialists with degrees in linguistics (sometimes enthusiasts)
    Step 2: they determine phonemes of the language (ignoring autosegmental stuff)
    Step 3: they determine a set of diactitics for Latin characters for those phonemes.

    The problems:
    – creativity is excluded. New alphabets are not created.
    – community is excluded.
    A writing system is an artistic endeavour.

    – science is misrepresented. When people act like this, it implies a belief: “this new writing system is sceintific because phonology is a science”. NO. And enother belif is “it should be done by linguists”.
    Phonology is a sceince. Making an object of art reflect a pattern studied by science is not.

    What would be “science” is studying writing systems and their advantages and disadvantages (and I expect the same writing system to have BOTH rather than be just “better”) and sharing your findings with people who use writing systems.

    I agree with Spooner, but he gave a shitty exmaple:) Hangul is just not based on Latin script.

  29. David Marjanović says

    Quoth Wikipedia: “The Balochi vowel system has at least eight vowels: five long and three short. These are /aː/, /eː/, /iː/, /oː/, /uː/, /a/, /i/ and /u/. The short vowels have more centralized phonetic quality than the long vowels. The variety spoken in Karachi also has nasalized vowels, most importantly /ẽː/ and /ãː/.” All dialect writing in German is hard to read, first out of unfamiliarity, second because the orthographic conventions to represent Standard German vowel systems fail at those of almost any dialect.

  30. The context of my comments (because I think not everyone read the first article by Spooner):

    Spooner said that switching to Latin/Cyrillic was expected to help with literacy rates and that there are not studies supporting this idea and that Thai and Korean scripts are not simple, while literacy rates are high

    I strongly agree with that we should study the question first – and I think the question is complex enough.

    But I agree with DE, that Hangul is a terrible example.

  31. I also think that intuitively what he said about recognition of shapes makes sense (or why else the Chinese can read at all?) and I think that having a thriving community of users of a script (any script), moreover, growing up in this community is especially important. Not the writing itself:)

    You would never think it is even possible to read and write like Chinese if there were not such a community. And this belief: “reading anything but alphabetic writing is difficult” would excacly be unscientific.

  32. David Eddyshaw says

    Yes, I agree that while intrinsic simplicity can hardly hurt when it comes to acquiring a script, other factors, not least motivation, may be as important in reality.

    After all, children usually learn to read at an age when they are still able to acquire extremely complex spoken languages, and even adults display much greater ability to pick up a new language, if they really have to, than the citizens of our smug effectively-officially-monoglot Western states usually imagine. (I’ve seen this often in Africa.)

    I’m not sure that alphabetical writing necessarily is the simplest method for everyone; that may be an illusion generated by our own standpoint. There seems to be some reason to think that syllabaries are more natural for many people, especially if they speak languages with a relatively simple syllabic structure.

  33. David Eddyshaw says

    how do you create a writing system?

    The history of Kusaal orthography is not atypical for its period and place.

    The Agolle Kusaal orthography up until 2016 was basically the creation of SIL missionary linguists in the late 1960’s. They did a pretty good job of identifying the phonemes, but deliberately chose to ignore quite a number of distinctions in their practical orthography (tone, naturally, but also some fundamental vowel quality distinctions), I think largely to make it typeable on English typewriters, though they mysteriously kept ŋ, which would actually have been quite easy to represent with ng with just minimal tweaking of their orthographic rules. They certainly consulted with L1 speakers about all this, though I don’t have any information about the details of the decision-making.

    The resulting system is … quirky. On the plus side, they made some positively ingenious choices which mean that in practice there is much less actual ambiguity in the system than you would have expected, given that they use five vowel symbols for a basic seven-vowel system and do not systematically distinguish length in diphthongs, despite the fact that the language has several dozen of these, with three contrasting degrees of length.

    The 2016 revision was the work of an L1-speaking committee, I think largely following the work of Anthony Agoswin Musah, a Kusaasi linguist and the author of a full-dress Agolle Kusaal grammar. The most striking change is that it introduces the symbols ɛ ɔ, which over the years have become familiar to many Ghanaians from seeing them used in Twi. However, they are used in a way which is in fact redundant, because there are no actual e/ɛ o/ɔ contrasts in Agolle Kusaal. A step forward is that /ʊ/, which contrasts with /u/, is now written ʋ in accordance with local common practice; mysteriously, the equally important contrast /ɪ/ versus /i/ remains unmarked in any systematic way, with [ɪ] being written as any of i e ɛ in various contexts. I’ve no idea why this happened.

    The writing of vowel nasalisation, which admittedly was pretty counterintuitive, has been “simplified” in a way which unfortunately introduces serious ambiguities which did not arise in the old system. Tony Naden has unfortunately adopted this system in his dictionary, thereby losing significant information about word pronunciation.

    Toende Kusaal has largely been described by French speakers, and the orthography used in Burkina Faso has nothing at all to do with the standard orthography of Agolle Kusaal (to be honest, it’s a lot more sensible, using separate symbols for all nine basic vowels, and a tllde for nasalisation.)

  34. January First-of-May says

    @DE, how do you create a writing system?

    Step 1: invite specialists with degrees in linguistics (sometimes enthusiasts)
    Step 2: they determine phonemes of the language (ignoring autosegmental stuff)
    Step 3: they determine a set of diactitics for Latin characters for those phonemes.

    There’s some creativity in the specific diactritics (more so in Cyrillic than in Latin), but otherwise, yeah.
    Of course if you approach it too scientifically you get something like Marr’s Abkhazian (which is very scientific and logical but a crappy writing system).

  35. David Marjanović says

    I’ve seen this often in Africa.

    Now that I teach German to adults, I’m seeing it here too.

  36. David Marjanović says

    if you approach it too scientifically

    That’s actually not scientific enough: given that all the palatalization and labialization is spelled out, most of the vowel symbols could be dropped.

    On the practical side, I’m pretty sure ° (for labialization) could be replaced by the already used letter w without causing ambiguity…

  37. Reproducing a pattern (one pattern) observed in nature because there’s a science that studies this pattern is not “scientific”:(

  38. But: scientists or not, but let us observe that many writing systems do approach the structure of the langauge closely.

    Many modern writing systems have diverged: they still reflect the langauge in that they contain etymological information.

    Does not it mean that we can postulate two periods for a writing systems: convergence and divergence? It is trivial when your system was invented by someone (then the process of inventing it is “convergence”) and then immediately codified and kept intact since then (then it drifts towards “etymology”). But when a system is a collective work, and “convergence” is a natural process (many people write a letter 1 instead of 2 ‘because it sounds like 1’, and with time 1 cames to mean what it means now) you can speak about such periods.

Speak Your Mind