Persian: classical pronunciation

eskandar

Moderator
English (US)
I believe the current and (the classical) pronunciation is:
agar daryâfti bar dâneŝat bus(boos) - vagar qâfel ŝodi afsus afsus
This is of course the current pronunciation, but not the classical. Alfaaz has correctly indicated the classical pronunciation in his post. Notably: the و of بوس and افسوس represents a majhul vowel (long 'o'), not the 'uu' sound as it's pronounced today, and غ is pronounced differently than ق in classical Persian, such that the غ of غافل in classical Persian is like the ق of آقا in today's Tehrani Persian.
 
  • Stranger_

    Senior Member
    Persian
    Notably: the و of بوس and افسوس represents a majhul vowel (long 'o'), not the 'uu' sound as it's pronounced today
    Are you sure of this? I thought this pronunciation is a feature of only Indo-Persian.

    The thing which I am definitely sure about is that words which in modern Persian are pronounced with a "کسره" were pronounced with "فتحه" in the classic times. All words I mean, be they nouns, verbs, adjectives or whatever.

    Examples:
    - residan was proununced rasidan
    - khaneh was khanah
    - sefid was safid
    - yek was yak
    - khandeh was khandah
    etc...

    Most of classic poetry lines support the fatheh-pronunciation, like for example:

    بر آیین خود نیز پیران ندید
    ز پیران سخن سربسر دررسید

    چنین تا شب تیره اندررسید
    از آن بدسگالان یکی را ندید

    Both from Ferdosi. As you can see, you cannot and should not read them "dar resid" and "andar resid" but "dar rasid" and "andar rasid" in order to maintain the meter. I have seen much more examples but cannot remember any right now.

    I am not sure if this way of pronunciation was the norm all over Persia or not, but it seems to have been so in most areas until the Turks came and made the kasreh-pronunciation the common and standard one, because in Turkish all of the Persian words are pronounced with kasreh.

    Of course, this is just my superficial opinion and conclusion and could be totally wrong and I would be much grateful if someone corrected my views and gave a definite and thorough opinion on this matter.
     
    Last edited:

    PersoLatin

    Senior Member
    UK
    Persian - Iran
    Hi Alfaaz,

    I am interested to know why classical Persian pronunciation in Urdu, is so different to the way I know it, I am willing to learn but I need some help.

    Could you provide an English/other example where the Urdu pronunciation for u in shudi (شدی) and o in bos/afsos, can be shown please. I had a quick look in
    Transliteration of Urdu and Hindi as used in this Forum but couldn't see an example.
     

    eskandar

    Moderator
    English (US)
    Are you sure of this? I thought this pronunciation is a feature of only Indo-Persian.
    Yes, quite sure. The majhul vowels were present in classical Persian, so when Persian was introduced to India that pronunciation was dominant. Western (ie. Iranian) dialects started to shift some time later and the majhul vowels shifted with their ma'ruf counterparts, but this shift did not take place in eastern dialects such as those spoken in Afghanistan, Tajikistan, and India.

    The thing which I am definitely sure about is that words which in modern Persian are pronounced with a "کسره" were pronounced with "فتحه" in the classic times. All words I mean, be they nouns, verbs, adjectives or whatever.
    It's not true that all modern Iranian Persian words which have a kasre originally had a fathe, just certain ones (like your examples). Persian words borrowed from the Arabic pattern فاعل , for example, have always had a kasre; in other words, قالب has always been qaaleb and never *qaalab.
     

    Stranger_

    Senior Member
    Persian
    Western (ie. Iranian) dialects started to shift some time later and the majhul vowels shifted with their ma'ruf counterparts, but this shift did not take place in eastern dialects such as those spoken in Afghanistan, Tajikistan, and India.
    Is there a remnant of such words in modern Iranian Persian? I cannot think of any.

    It's not true that all modern Iranian Persian words which have a kasre originally had a fathe, just certain ones (like your examples). Persian words borrowed from the Arabic pattern فاعل , for example, have always had a kasre; in other words, قالب has always been qaaleb and never *qaalab.
    No no, I am not talking about Arabic words at all. I am talking about pure Iranian words. Those words which when you search in a Loghatnameh you see two pronunciations indicated for them.

    A few more example:
    چو سال اندر آمد به هفتاد و یک
    همی زیر بیت اندر آرم فلک
    فردوسی

    در آن دم یکی مرد پوشیده چشم
    بپرسیدش از موجب کین و خشم
    سعدی

    دمان همچو شیر ژیان پر ز خشم
    بلند و سیه خایه و زاغ چشم
    فردوسی


    از این روی تا مرو لشکر کشید
    شد از گرد لشکر زمین ناپدید
    فردوسی

    (you can't read these lines with yek, cheshm, keshid)

    In a lot of Southern Persian dialects the fathe-pronunciation is still the common one. I asked a Kalhori Kurdish speaker from Kermanshah and he said that in Kurdish too they are pronounced with fathe. I do not know about Balochi though.

    I wonder when this shift from fathe to kasre has taken place and whether it is the influence of Turkish or something else.
     

    eskandar

    Moderator
    English (US)
    Hi eskandar, can you provide some rough dates for this shift please.
    According to this source, the shift may have begun by as early as the 13th century CE and continued until the 16th century if not later.

    Is there a remnant of such words in modern Iranian Persian? I cannot think of any.
    I don't think so. Those vowels may have survived in some regional Iranian dialects but I assume by "modern Iranian Persian" you mean primarily the Tehrani dialect and perhaps other major urban dialects.

    No no, I am not talking about Arabic words at all. I am talking about pure Iranian words. Those words which when you search in a Loghatnameh you see two pronunciations indicated for them.
    Arabic-origin words aside, there are still plenty of Persian words of Iranic origin which have always had a kasre. For example, any of the place-names ending in -estaan were originally -estaan, not *-astaan. This applies to words like zemestaan as well. فرشته in classical Persian was fareshta not *farashta. به and که have always been beh and keh respectively and have never had a fathe. چه has also never had a fathe, not even in Pahlavi. I could go on but I think I've made my point: not all instances of [a] have become [e] (not all fathes have become kasres).

    I wonder when this shift from fathe to kasre has taken place and whether it is the influence of Turkish or something else.
    I have seen some sources claim that it was indeed under the influence of Turkish, but they seemed to state the matter as if it were a fait accompli and offered no evidence or argument for why it would be so. I'm really curious about this question myself.
     

    Stranger_

    Senior Member
    Persian
    Those vowels may have survived in some regional Iranian dialects but I assume by "modern Iranian Persian" you mean primarily the Tehrani dialect and perhaps other major urban dialects.
    I cannot think of any word having the majhul vowel, even in my own dialect which is rural.

    Question: were all words pronounced with majhul vowels or only some? is there a poetry line which can support this claim? By the way, even in Urdu not all of them are pronounced with the majhul vowels. For example: دور duur/پیر piir/تیر tiir

    Arabic-origin words aside, there are still plenty of Persian words of Iranic origin which have always had a kasre. For example, any of the place-names ending in -estaan were originally -estaan, not *-astaan. This applies to words like zemestaan as well. فرشته in classical Persian was fareshta not *farashta. به and که have always been beh and keh respectively and have never had a fathe. چه has also never had a fathe, not even in Pahlavi. I could go on but I think I've made my point: not all instances of [a] have become [e] (not all fathes have become kasres).
    No, you got me wrong again :)
    I am not talking about all words but only those words which Loghatnamehs give two pronunciations for them.
    Like:tick:
    خانه
    بهانه
    سایه
    اندیشه
    ریشه
    کشیدن
    رسیدن
    چشیدن
    گرسنه or گشنه
    تشنه
    مرده
    خسته
    زنده
    پرنده
    آشفته
    آهسته
    کِی/کَی when

    Not:cross:
    زدودن
    پریدن
    دست
    بریدن
    مرگ
    درد
    گرفتن
    رفتن
    زرد
    سبز
    مست
    زن
    مرد
    که
    چه
    چرا
    در
    بر

    You see, because these words have only one pronunciation.

    In fact, some Arabic words can also have two pronunciations (of course, the kasre-pronunciation being the standard in Iran)
    Like:
    لقمه
    سکته
    مبارزه
    مطالبه
    جریمه
    حلیمه
    خدیجه
    مدینه
    خجالت

    By the way, the word "به" has two pronunciations :) that's why it is "bah" in Urdu. Urdu has retained the fathe-pronunciation for all of those words I mentioned above and their like.

    In my dialect, it is pronounced "وَ" (which indicates that it was originally بَه) so instead of saying: "به این بزرگی" we say "وَ ای بزرگی/گُتی"

    One more interesting thing is that in Lamerdi, Larestani and Jahromi this "به" has another form too, simply: اَ (a)!
     

    Dib

    Senior Member
    Bengali (India)
    This is an improved version of a post, I made here yesterday and have subsequently deleted, because it contained some basic mistakes. Sorry, if I misled anyone by it.

    ======

    The thing which I am definitely sure about is that words which in modern Persian are pronounced with a "کسره" were pronounced with "فتحه" in the classic times. All words I mean, be they nouns, verbs, adjectives or whatever.

    Examples:
    - residan was proununced rasidan
    - khaneh was khanah
    - sefid was safid
    - yek was yak
    - khandeh was khandah
    etc...

    <...>

    I am not sure if this way of pronunciation was the norm all over Persia or not, but it seems to have been so in most areas until the Turks came and made the kasreh-pronunciation the common and standard one, because in Turkish all of the Persian words are pronounced with kasreh.
    It is of course possible that Turkish influence played a role in these changes, but we should also be aware about the counter-examples, so that the opinions can be properly tempered. While you rightly observe that many Turkish adaptations have /e/ instead of Persian /a/, there are still some exceptions. From the Istanbul variety of Turkish:
    hafta
    hasta (<khaste) = sick
    meyva = fruit
    namaz
    zaman

    While these may be exceptional for Persian borrowings, the short “a” of Arabic is borrowed frequently as an “a” into Turkish. This may be taken to mean that the variety of Persian, the Turks were borrowing from, already had the short /a/ sounding somewhat like Turkish /e/, let's say [æ]. At least, this works for Modern Iranian Persian, where [æ] is a common realization of فتحه, and Modern Istanbul Turkish, where it is an allophone of /e/ (e.g. ben, sen, etc.) The interaction with Arabic might have been more bookish, helping to preserve the a-pronunciation.

    So, what I'm trying to say is: that Persian a is often borrowed as a Turkish e, seems to have a plausible phonetic motivation, without Turkish having to influence Persian pronunciation directly. It is also revealing that so many Persian and Arabic words which have been borrowed with an /e/ in Turkish still contain /a/ in Persian. So, I am not denying that Turkish might have exerted influence on Persian pronunciation, but the details of such influences are not straight-forward.

    ------

    Btw. sefid had a majhuul ē, so classical safēd.

    I am interested to know why classical Persian pronunciation in Urdu, is so different to the way I know it, I am willing to learn but I need some help.
    Urdu mostly retains the classical Persian values of the vowels. If you have access to Thackston's "An Introduction to Persian", you can flip over to the end and read the section on Classical Persian pronunciation. I imagine, there are better descriptions out there, since this book is really for teaching to beginners.

    Question: were all words pronounced with majhul vowels or only some?
    Only some.

    is there a poetry line which can support this claim?
    I'd imagine - no, at east not in a straight-forward way, since all instances of majhul ē have become ma'ruf ī and all majhul ō > ma'ruf ū. So, what scanned and rhymed before, still scan and rhyme. And they always differed from کسره and ضمه respectively in length. The only imaginable way would be to sift through the whole corpus of classical Persian poetry and establish that majhul vowels did NOT rhyme with ma'ruf vowels.

    ======

    Another note about the pronounciation of کسره and ضمه . In Classical Persian they alternated - apparently freely - between short i~short e and short u~short o respectively. Urdu typically uses this first value of the pairs (i.e. short i/u). Turkish also retains this value in some words, most notably "ki", but چونکه is çünkü through vowel harmony.
     
    Last edited:

    Stranger_

    Senior Member
    Persian
    Dib, you have made valid points and have weakened the Turkish influence theory to a great degree but we still cannot entirely exclude it. Why? because it is the only one there. What else could have been the reason for that shift then? A shift which seems to have started from the northern areas of Iran.

    The only imaginable way would be to sift through the whole corpus of classical Persian poetry and establish that majhul vowels did NOT rhyme with ma'ruf vowels.
    Even this, is not possible to do. Because you cannot know which words had majhul vowels and which words did not.
     

    Dib

    Senior Member
    Bengali (India)
    Dib, you have made valid points and have weakened the Turkish influence theory to a great degree but we still cannot entirely exclude it. Why? because it is the only one there. What else could have been the reason for that shift then? A shift which seems to have started from the northern areas of Iran.
    Yes, Stranger_, of course. :) I am not trying to deny the Turkish influence as I have mentioned in my original post itself. However, it is a faulty argument that Turkish was the only influence, hence it has to be responsible for the changes in question. Languages change on their own too. True, contact with other languages may accelerate or even give rise to certain changes, but even without external influences, languages evolve. In short, each of the sound changes needs to be evaluated separately and supplied with evidence to conclude or exclude Turkish influence.

    Even this, is not possible to do. Because you cannot know which words had majhul vowels and which words did not.
    Well, fortunately we are not all that helpless. We already have a fairly good idea of the distribution, because Indo-Persian (and hence, loans in Indian languages), Afghan and Tajik Persian still maintain much of the majhul-ma'ruf distinction. Also, even without this knowledge, if you simply analyze the texts, and list all the rhymes involving a written ye, they should naturally divide themselves up into three groups (similarly for written vav): majhul ē, ma'ruf ī and diphthong ay (>ei). This kind of rhyme analysis has been extensively used, e.g. in Chinese historical linguistics. I see no reason why it can't be done for Persian. There are complications, of course, because the distribution may not have been stable over different time periods and regions. But using of the knowledge of the texts' backgrounds, it should also be possible to work out. However, it will miss the instances where the ye is not in a rhyming position. But it will give a good idea about how reliable the modern Tajik, Afghan and Indian evidences are. In any case, I would be surprised if no scholar hadn't attempted this yet. Our resident expert, fdb (and of course, others in the know!), surely can point us to the relevant scholarly literature. :)

    It should also be relatively quick to do for Persian because Persian divans are already arranged alphabetically by their rhymes.
     
    Last edited:

    PersoLatin

    Senior Member
    UK
    Persian - Iran
    Hi Dib, purely playing the devil's advocate here:

    We only have written evidence which unfortunately does not support any of these ideas, fully, because of Perso-Arabic's lack of short vowels.

    We already have a fairly good idea of the distribution, because Indo-Persian (and hence, loans in Indian languages), Afghan and Tajik Persian still maintai.......
    How can we be sure that the Afghan, Tajik and the Urdu accents didn't develop from a regional accent & that the mainstream Persian was different from the aforementioned, from the start? Why, for example, can't we say Shirazi is the classical Persian accent, as it is different, i.e. the same as Afghan or Tajik accents are?

    A few more example:
    چو سال اندر آمد به هفتاد و یک
    همی زیر بیت اندر آرم فلک
    فردوسی

    در آن دم یکی مرد پوشیده چشم
    بپرسیدش از موجب کین و خشم
    سعدی

    دمان همچو شیر ژیان پر ز خشم
    بلند و سیه خایه و زاغ چشم
    فردوسی


    از این روی تا مرو لشکر کشید
    شد از گرد لشکر زمین ناپدید
    فردوسی

    (you can't read these lines with yek, cheshm, keshid)
    Equally, how can we be sure that خشم was not kheshm, فلك was not falek or ناپديد wasn't nâpedid?

    Thanks.
     
    Last edited:

    Dib

    Senior Member
    Bengali (India)
    Thank you fdb, for the link. To give a concrete example of the kind of rhyme analysis I am talking about, refer to Divan-e Shams vers 1057. The rhyme words are: نوش هوش خروش گوش جوش فروش دوش خموش

    They are all pronounced with a majhul ō in Urdu, as corroborated by Platts' dictionary. My personal experience also corroborates that for 5 of the 8 words. I was not familiar with the other three in Urdu (and one خروش even in Persian).

    Now you repeat the same thing for more and more verses, and look whether the groupings arrived at, align with the India, Afghan, Tajik evidence or not. If the majhul-ma'ruf hypothesis is correct, you should find very few violations - I am assuming that a few words might have had variable vowels depending on regions and periods.
     
    Last edited:

    marrish

    Senior Member
    اُردو Urdu
    ^ All of these words are invariably pronounced with a majhūl ō in Urdu. josh-o-xarosh being an example.
     

    Stranger_

    Senior Member
    Persian
    I am not buying the majhul vowels hypothesis because nowhere in Iran can we find a trace of it, even in the most rural areas. Admittedly, most poetic examples are in support of it and there are very rare counter-examples. But isn't one example enough to refute it?

    These are what I could find:

    اگر از خرقه کس درویش بودی
    رئیس خرقه پوشان میش بودی
    (In Urdu, they are darvesh and miish; they clearly do not rhyme)

    که امروز روزی بزرگ است پیش
    پدید آید اندازه ٔ گرگ و میش
    (Same as above)

    گرت زندگانی نوشته است دیر
    نه مارت گزاید نه شمشیر و تیر
    (In Urdu, they are der and tiir)

    کسی کو خرد را ندارد ز پیش
    دلش گردد از کرده ٔ خویش ریش
    (I am not sure though how the word riish, in the sense of "wound", is pronounced in Urdu)

    Equally, how can we be sure that خشم was not kheshm, فلك was not falek or ناپديد wasn't nâpedid?
    Leaving the word خشم aside, the other two words are never pronounced that way, and no Loghatnameh gives more than one pronunciation for them.
     

    eskandar

    Moderator
    English (US)
    I am not buying the majhul vowels hypothesis because nowhere in Iran can we find a trace of it, even in the most rural areas.
    Not so. See here for example, which mentions the retention of the majhuul vowel /ē/ in Bakhtiyari.

    اگر از خرقه کس درویش بودی
    رئیس خرقه پوشان میش بودی
    (In Urdu, they are darvesh and miish; they clearly do not rhyme)
    You are mistaken. میش is pronounced mēsh in classical Persian and Urdu.

    کسی کو خرد را ندارد ز پیش
    دلش گردد از کرده ٔ خویش ریش
    (I am not sure though how the word riish, in the sense of "wound", is pronounced in Urdu)
    It's pronounced rēsh in Urdu and classical Persian which would rhyme with classical pēsh.
     

    Dib

    Senior Member
    Bengali (India)
    I am not buying the majhul vowels hypothesis...
    As I see it now, after reading Fritz Meier's article linked to by fdb (actually, I am still reading, it is a long article dealing with many aspects of Early Modern Persian pronunciation, including majhul vowels), there is no shred of doubt about their existence. Yes, exactly which words contained which vowel at what time and location, may be discussed; but not their existence. The most important source Meier uses is what Hamza al-Isbahani (died before 970 AD), writing in Arabic, notes about Persian pronunciation of his time*. He lists 8 Persian sounds that didn't occur in Arabic, majhul ē and ō among them, and he gives two examples of each of them. Interestingly, one of his examples is bō (smell), which in modern Urdu is usually pronounced with a ma'ruf ū. So, obviously (Western) Classical Persian distribution of majhul vowels does not coincide perfectly with that of Indo-, Afghan or Tajik Persian, and Meier explicitly acknowledges this. However, there is significant overlap. In order to reconstruct the distribution of the majhul vowels, Meiers then discusses rhymes quite a bit, but supplements them with Persian written in pointed Hebrew and Syriac scripts, which apparently can distinguish ē and ī. Arabic transliterations are also potentially revealing, because the majhul ē vacillated between a written yay and an alif, the latter choice coming up because long ā with imala sounded close to majhul ē, while the ma'ruf ī apparently never showed any alif-spelling.

    The occasional "bad rhymes" between majhul, ma'ruf, but also the diphthongs, are also discussed by Meier. Firstly, they are infrequent (especially for ē-ī). But, their frequency is similar in both Eastern and Western poets. Since, East still has a functional majhul-ma'ruf distinction, the status of bad rhymes as proof of absence of the distinction in the West is substantially weakened. It must have had other reasons, like poetic licence.

    Then, Meier talks about Shams-e Qays's 13th century work on Persian prosody. He forbids rhyme between ē and ī, but allows that between ō and ū, which probably indicates that the ō-ū merger was well-advanced at this time, but not the ē-ī merger. However, he himself apparently also criticized Anwari (wrongly!!) of rhyming ō and ū. This likely implied two things:
    1) Qays' living language had already completed the ō-ū merger, leading him to make a wrong call.
    2) The poetic language still preferred to keep ō-ū separate, though (grudgingly?) starting to accept their rhyme.

    Later on, even after the completion of majhul-ma'ruf merger in the living language, the poetic language apparently largely maintained the old tradition of keeping them separate. Shams-e Fakhr-e Isfahani maintains the distinction in his 1344 textbook, "Mi'yar-e Jamali". Even among modern Iranian poets, Malek-osh-Sho'ara Bahar said, he had been able to maintain the old distinction, except in two verses.

    Meier's article really contains a wealth of information. He devotes more than 15 pages to majhul vowels. This is just a very brief summary.

    --------
    * Published in At-tanbih 'ala hudud at-tashif, Baghdad 1968, pp 82-84
     

    Dib

    Senior Member
    Bengali (India)
    گرت زندگانی نوشته است دیر
    نه مارت گزاید نه شمشیر و تیر
    (In Urdu, they are der and tiir)
    eskandar has addressed the other apparent majhul~ma'ruf rhymes. For this one, دیر needs to be read diir. Steingass's classical Persian dictionary lists both diir and der as possible pronunciations. Even Platts' Urdu dictionary mentions diir as a possible pronunciation, though marks it 'rare' (i.e. rare in Urdu).
     

    Stranger_

    Senior Member
    Persian
    What can I say. Your arguments and the evidence you have provided are compelling indeed.

    Anyway, this question is still open: why has the classical pronunciation changed in Iran but remained the same in Afghanistan, Tajikistan, Pakistan and India? Is it because Persian has more speakers in Iran and is used more actively and widely than the rest? i.e. the language is more alive and constantly changing.
     
    Last edited:

    PersoLatin

    Senior Member
    UK
    Persian - Iran
    This may be simplistic but as proof of classical Persian pronunciation, I find majhul, ma'ruf argument very convenient, I think Stranger_ has touched on it too.

    Switching between /e/ and /ii/ to fit the Urdu pronunciation, serves the argument well, but is there any other evidence, other than poetry?

    I think when Platts' dictionary says, دير as diir is 'rare in Urdu', it may mean that it is only acceptable in Persian poetry (to keep the rhyme & meter), where it is absolutely necessary to be read as diir, because the other rhyming word is read with /ii/, even in Urdu. I wonder if it provides examples of these rare situations.

    If you look at the Kurdish پشمرگان, there no ى in it & it can easily be attested that it is pronounced 'pešmergân', I wouldn't read it as pišmargân, nor would anyone else, but you would when it is پيشمرگان.
     
    Last edited:

    eskandar

    Moderator
    English (US)
    why has the classical pronunciation changed in Iran but remained the same in Afghanistan, Tajikistan, Pakistan and India?
    This happens with any language spread out over a large area (or even a small area divided by natural or political factors). Look at English: aspects of its phonology have changed more radically in North America and remain more broadly continuous (though not identical) in places like the UK, Australia, and New Zealand. Such changes also happen within a language over time, geographical features notwithstanding (cf. the English Great Vowel Shift and pin-pen merger, etc).

    Is it because Persian has more speakers in Iran and is used more actively and widely than the rest? i.e. the language is more alive and constantly changing.
    Historically the reverse would have been true: the population of Persian speakers where the majhul and ma'ruf vowels merged (ie. in Iran) would have been less than the population of Persian speakers where they retained the distinction (ie. the combined Persian-using populations of Afghanistan, India, Tajikistan, Uzbekistan, etc). In any case I don't think it has anything to do with the number of speakers. We certainly can't say that Persian was used any more actively and widely in Iran during the time this vowel merger happened than in the other countries.
     

    Dib

    Senior Member
    Bengali (India)
    why has the classical pronunciation changed in Iran but remained the same in Afghanistan, Tajikistan, Pakistan and India? Is it because Persian has more speakers in Iran and is used more actively and widely than the rest? i.e. the language is more alive and constantly changing.
    This happens with any language spread out over a large area (or even a small area divided by natural or political factors). Look at English: aspects of its phonology have changed more radically in North America and remain more broadly continuous (though not identical) in places like the UK, Australia, and New Zealand. Such changes also happen within a language over time, geographical features notwithstanding (cf. the English Great Vowel Shift and pin-pen merger, etc).
    Well, I wouldn't rule out Stranger_'s idea so hastily - at least in the Indo-Pak context. It is quite likely that Persian was largely a learned language, and not a real native language for the most of the Persian-using population in Indo-Pak. Their native language would have been one of the Indian languages, or for some of the royalty and in the early days also beyond, some variety of Turkic. They would have learned Persian principally through formal education. It is similar to what is happening to English in the same region today. The formal learning without native usage can actually provide significant resistance to change. Indian English, e.g. lacks the horse-hoarse merger, which is otherwise a very widespread sound change in English. Something similar might have happened to Indo-Persian. It also might have helped that the ma'ruf-majhul distinction is also present in the Indian and Turkic languages.

    However, this certainly does not explain the Afghan and Tajik situations so neatly. I know Tajikistan (and Eastern Uzbekistan) was a native Sogdian speaking region (Sogdian there still survives in the form of Yaghnobi), and Afghanistan had Indo-Aryans, the ancestors of Pashtuns, and was settled also by speakers of Turkic and Mongolian. But still, I am not aware of any dearth of native Persian speakers in this region in the relevant period. But it may simply be a matter of my lack of knowledge.

    Switching between /e/ and /ii/ to fit the Urdu pronunciation, serves the argument well, but is there any other evidence, other than poetry?
    As I mentioned in post #21, Meier discusses various sources of information:
    1) Direct comments of contemporary authors (10th-14th century)
    2) Persian written in other scripts, e.g. Syriac, Hebrew, etc.

    I think when Platt's dictionary says, دير as diir is 'rare in Urdu', it may mean that it is only acceptable in Persian poetry (to keep the rhyme & meter), where it is absolutely necessary to be read as diir, because the other rhyming word is read with /ii/, even in Urdu. I wonder if it provides examples of these rare situations.
    That sounds like a reasonable assumption to me. But Platts unfortunately gives no such example.
     

    PersoLatin

    Senior Member
    UK
    Persian - Iran
    As I mentioned in post #21, Meier discusses various sources of information:
    1) Direct comments of contemporary authors (10th-14th century)
    2) Persian written in other scripts, e.g. Syriac, Hebrew, etc.
    Thank you Dib, but 'Deutsch ist aller Grieche zu mir'
     

    eskandar

    Moderator
    English (US)
    Well, I wouldn't rule out Stranger_'s idea so hastily - at least in the Indo-Pak context. It is quite likely that Persian was largely a learned language, and not a real native language for the most of the Persian-using population in Indo-Pak. Their native language would have been one of the Indian languages, or for some of the royalty and in the early days also beyond, some variety of Turkic. They would have learned Persian principally through formal education. It is similar to what is happening to English in the same region today. The formal learning without native usage can actually provide significant resistance to change. Indian English, e.g. lacks the horse-hoarse merger, which is otherwise a very widespread sound change in English. Something similar might have happened to Indo-Persian. It also might have helped that the ma'ruf-majhul distinction is also present in the Indian and Turkic languages.

    However, this certainly does not explain the Afghan and Tajik situations so neatly. I know Tajikistan (and Eastern Uzbekistan) was a native Sogdian speaking region (Sogdian there still survives in the form of Yaghnobi), and Afghanistan had Indo-Aryans, the ancestors of Pashtuns, and was settled also by speakers of Turkic and Mongolian. But still, I am not aware of any dearth of native Persian speakers in this region in the relevant period. But it may simply be a matter of my lack of knowledge.
    I find your second language theory a bit too pat for the reasons you pointed to at the end of this post: there was no lack of native Persian speakers, especially in northern Afghanistan, in the relevant period, so this theory doesn't explain their pronunciation. Furthermore, Persian was (and is) "largely a learned language, and not a real native language" for much of the Persian-using population in Iran as well, whose native languages would include non-standard dialects of Persian, closely related languages (Lori, Bakhtiyari, Gilaki, Mazandarani, etc.), and even more distinct languages (Kurdish and Turkic languages, for example). So I fail to see how the second language theory would explain why the pronunciation of native Persian speakers changed in Afghanistan but not Iran, or why it was due to non-native speakers in India but not Iran, for example.
     

    tarkshya

    Senior Member
    Marwari
    According to this source, the shift may have begun by as early as the 13th century CE and continued until the 16th century if not later.
    If the loss of majhul vowels happened between 13th and 16th centuries, then it should necessarily have reflected in Indian Persian, and in Hindi/Urdu consequently. Mughal empire of India was rose and fall *after* 16th century. Majority of the Persian immigrants entered India during the Mughal empire, and continued to trickle in till the fag-end of the empire.

    Also, Mughals themselves had a somewhat low opinion of the Persian spoken in their own courts! Throughout the Mughal empire there were language "reform" movements in the court, where the local innovations were purged from the court Persian, and the language was brought in sync with the Iranian Persian.
     

    eskandar

    Moderator
    English (US)
    If the loss of majhul vowels happened between 13th and 16th centuries, then it should necessarily have reflected in Indian Persian, and in Hindi/Urdu consequently. Mughal empire of India was rose and fall *after* 16th century. Majority of the Persian immigrants entered India during the Mughal empire, and continued to trickle in till the fag-end of the empire.
    The history of Persian dates back to the 12th century or earlier. Since it was already well-established in India by the time the majhul-ma'ruf vowel merger started in the western Persianate lands (modern-day Iran), it's not true that it would have necessarily been reflected in Indo-Persian and Hindi/Urdu, for the same reason that phonological changes that happened in American English were not reflected in British English and vice-versa. In fact, if you read the writings of those Iranian immigrants who visited India during the Mughal empire (for example Hazin Lahiji), you'll see that they already notice a difference and complain about the "bad Persian" spoken by Indians.

    Also, Mughals themselves had a somewhat low opinion of the Persian spoken in their own courts! Throughout the Mughal empire there were language "reform" movements in the court, where the local innovations were purged from the court Persian, and the language was brought in sync with the Iranian Persian.
    As far as I have seen, these reforms had to do primarily with vocabulary, and to a lesser degree with style, syntax, etc., but I have not seen any evidence to suggest that Mughal-era language reformers strove to change the phonology of Indo-Persian. Furthermore, I think the claim that "the language was brought in sync with the Iranian Persian" belies the fact that some of the Persian language reforms in India (cf. texts like Burhan-i Qati' and Dasatir) actually ended up changing Iranian Persian rather than the reverse.
     

    PersoLatin

    Senior Member
    UK
    Persian - Iran
    In fact, if you read the writings of those Iranian immigrants who visited India during the Mughal empire (for example Hazin Lahiji), you'll see that they already notice a difference and complain about the "bad Persian" spoken by Indians.
    This is interesting, I wonder why those immigrants would have found, only the Indian Persian as 'bad', since they must come across other accents in Persia, at the time, which should have seemed 'bad' to them, also.

    Is it possible, that Indian Persian seemed 'bad' to those immigrants, because of the majhul-ma'ruf vowel distinction, prevelant in Indian Persian, plus maybe other really 'bad Persian', some of which were subsequently corrected by those reforms but the majhul-ma'ruf vowel distinction, escaped them?
     
    Last edited:

    colognial

    Senior Member
    Persian
    Could anyone please explain why the vowel 'o' in 'afsos' is called majhul, while the longer 'oo' in 'afsoos' is a maloum vowel?

    And, are the 'و's that appear in the following words also all of them majhul?
    خوب (pronounced 'khob' and not 'khoob' in spoken Iranian Persian)
    خورشید
    خور
    خورش
    خوردن
    خوراک
    چون (meaning 'because')
    مور (transcription of the English word 'moor')
    شوهر
    جور (meaning 'cruelty')
    تو (meaning 'you').

    Thanks.
     

    eskandar

    Moderator
    English (US)
    I wonder why those immigrants would have found, only the Indian Persian as 'bad', since they must come across other accents in Persia, at the time, which should have seemed 'bad' to them, also.
    I wonder about the same thing! But I haven't read many travelogues or other works by travelers within Iran in that period, so I don't know - maybe they (or other Iranians) do, in fact, complain about the Persian spoken in different parts of Iran, too.

    Is it possible, that Indian Persian seemed 'bad' to those immigrants, because of the majhul-ma'ruf vowel distinction, prevelant in Indian Persian, plus maybe other really 'bad Persian', some of which were subsequently corrected by those reforms but the majhul-ma'ruf vowel distinction, escaped them?
    Seems possible to me, though I think there were a lot of other factors, too.

    Could anyone please explain why the vowel 'o' in 'afsos' is called majhul, while the longer 'oo' in 'afsoos' is a maloum vowel?
    The long 'o' in afsōs was a sound that was present in Persian but was unknown (majhūl) to the Arabs, as it was not present in classical Arabic, as opposed to the long 'u' in the modern pronunciation afsūs which was known (ma'rūf) to the Arabs as it was present in their language as well.

    And, are the 'و's that appear in the following words also all of them majhul?
    خوب (pronounced 'khob' and not 'khoob' in spoken Iranian Persian)
    If we are talking about the spoken language, I believe the vowel sound is ma'ruf. We might as well write the word خُب . It's the same 'o' as in the word qor'aan قُرآن , so it's a sound that is represented normally by the Arabic writing system, and is therefore ma'ruf.

    There is another--totally separate--issue here which is that much more recently, in Tehrani Persian (and perhaps other major urban Iranian dialects, I don't know) the diphthong aw as in نو naw "new" has merged with the monophthong o as in نه noh "nine". Which is to say, the vowels in these two words used to be distinct and are now identical, at least in Tehrani Persian. This is totally separate from the majhul-ma'ruf merger we have been discussing. In classical Persian, we had four totally distinct sounds:

    (1) the long 'o' in افسوس afsōs
    (2) the short 'o' in تو to
    (3) the dipththong 'aw' in شوهر shawhar
    (4) the long 'u' in پول pūl

    In contemporary Tehrani Persian, the long 'o' of (1) has become a long 'u' identical to (4), whereas the diphthong 'aw' of (3) has become a short 'o' identical to (2). In classical Persian, only (1) was majhul; the rest were ma'ruf. Then there is the issue of خو which was classically khw as in خور khwor. Here the و is a consonant, and the vowel is a simple ma'ruf short 'o'.
     

    fdb

    Senior Member
    French (France)
    Most of what there is to say on this topic has been said, especially in the learned contributions by Eskandar and Dib, and I would like to mention just a few small details.

    First of all, “classical” Persian is usually used to mean the language of authors of the post-Mongol period like Saʻdī and Ḥāfiẓ, while the language of Firdawsī and his contemporaries is usually called Early New Persian (ENP). Steingass’s Persian dictionary and Platt’s Urdu dictionary give the Indo-Persian pronunciation, as indicated in Indian dictionaries of the Moghol period like Burhān-i-Qāṭiʻ, which is not necessarily identical with Classical, and certainly not with early NP. For these we need to look at other things, especially etymology and the rhymes in early poets.

    For example, the early poets rhyme the reflexive pronoun خوذ xwaδ with baδ, not with šuδ, nor with Arabic asad. Similarly you have xwaršēδ, xwardan etc. In these words the group xwa- regularly becomes xu- in later Persian.

    The pronoun for the 2nd person singular is tō, shifting later to tu. چون is čūn, but چو is ču (always scanned short in poetry). So each of these words has to be looked at separately.
     
    Last edited:

    Stranger_

    Senior Member
    Persian
    For example, the early poets rhyme the reflexive pronoun خوذ xwaδ with baδ, not with šuδ, nor with Arabic asad.
    Yes indeed, I have seen such examples. Here are some:

    بگفتا گر این مرد بد می‌کند
    نه با من که با نفس خود می‌کند
    سعدی

    منادی شد جهان را هر که بد کرد
    نه با جان کسی با جان خود کرد
    نظامی

    But was the last letter really pronounced "δ" and not "d"? The only dialect which has this feature is Lâmerdi and its sub-dialects; of course, the only dialect which I know of I mean, because there might be some more.
     

    Qureshpor

    Senior Member
    Panjabi, Urdu پنجابی، اردو
    Are you sure of this? I thought this pronunciation is a feature of only Indo-Persian......
    aaqaa-ye-PersoLatin, an example of a feature of Classical Persian about which modern day speakers not being aware of.
     

    Qureshpor

    Senior Member
    Panjabi, Urdu پنجابی، اردو
    As I see it now, after reading Fritz Meier's article linked to by fdb (actually, I am still reading, it is a long article dealing with many aspects of Early Modern Persian pronunciation, including majhul vowels), there is no shred of doubt about their existence. Yes, exactly which words contained which vowel at what time and location, may be discussed; but not their existence. The most important source Meier uses is what Hamza al-Isbahani (died before 970 AD), writing in Arabic, notes about Persian pronunciation of his time*. He lists 8 Persian sounds that didn't occur in Arabic, majhul ē and ō among them, and he gives two examples of each of them. Interestingly, one of his examples is bō (smell), which in modern Urdu is usually pronounced with a ma'ruf ū. So, obviously (Western) Classical Persian distribution of majhul vowels does not coincide perfectly with that of Indo-, Afghan or Tajik Persian, and Meier explicitly acknowledges this. However, there is significant overlap. In order to reconstruct the distribution of the majhul vowels, Meiers then discusses rhymes quite a bit, but supplements them with Persian written in pointed Hebrew and Syriac scripts, which apparently can distinguish ē and ī. Arabic transliterations are also potentially revealing, because the majhul ē vacillated between a written yay and an alif, the latter choice coming up because long ā with imala sounded close to majhul ē, while the ma'ruf ī apparently never showed any alif-spelling.

    The occasional "bad rhymes" between majhul, ma'ruf, but also the diphthongs, are also discussed by Meier. Firstly, they are infrequent (especially for ē-ī). But, their frequency is similar in both Eastern and Western poets. Since, East still has a functional majhul-ma'ruf distinction, the status of bad rhymes as proof of absence of the distinction in the West is substantially weakened. It must have had other reasons, like poetic licence.

    Then, Meier talks about Shams-e Qays's 13th century work on Persian prosody. He forbids rhyme between ē and ī, but allows that between ō and ū, which probably indicates that the ō-ū merger was well-advanced at this time, but not the ē-ī merger. However, he himself apparently also criticized Anwari (wrongly!!) of rhyming ō and ū. This likely implied two things:
    1) Qays' living language had already completed the ō-ū merger, leading him to make a wrong call.
    2) The poetic language still preferred to keep ō-ū separate, though (grudgingly?) starting to accept their rhyme.

    Later on, even after the completion of majhul-ma'ruf merger in the living language, the poetic language apparently largely maintained the old tradition of keeping them separate. Shams-e Fakhr-e Isfahani maintains the distinction in his 1344 textbook, "Mi'yar-e Jamali". Even among modern Iranian poets, Malek-osh-Sho'ara Bahar said, he had been able to maintain the old distinction, except in two verses.

    Meier's article really contains a wealth of information. He devotes more than 15 pages to majhul vowels. This is just a very brief summary.

    --------
    * Published in At-tanbih 'ala hudud at-tashif, Baghdad 1968, pp 82-84
    Hi Dib, people from the subcontinent pronounce one of the celebrated works of Sa'di as "Bostaan-i-Sa'di". In my very first school, our head master's name was "Bostaan".

    In Punjabi one of the words for "smell" is "bo".
     

    PersoLatin

    Senior Member
    UK
    Persian - Iran
    aaqaa-ye-PersoLatin, an example of a feature of Classical Persian about which modern day speakers not being aware of.
    On that there's absolutely no doubt & I'll be very surprised if you can find one who is, but why?

    This can not simply be dismissed as a casual forgetting of something that now only exists in Indo-Persian.

    I can not dismiss the existence of the features you have been telling us but why is that they are not in modern Persian. I have asked this before, can you please provide examples of how these features are absolutely necessary and how without them, reading, understanding & enjoyments of Persian works by great poets like Sa'di and Hafez, are spoiled?
     
    Last edited:

    Qureshpor

    Senior Member
    Panjabi, Urdu پنجابی، اردو
    These and other features indeed did exist in the Classical and Early New Persian and I shall provide you two references in your "Classical Persian" thread. They don't exist in the modern language because I suspect there has been a gradual evolution in the language resulting in the current form of the language.

    I don't believe anyone is suggesting that by not reciting Classical poets' works with today's pronunciation "spoils" their enjoyment. However, the reading and understanding could be faulty. For example...

    گل ھمین پنج روز و شش باشد
    وین گلستان ہمیشه خوش باشد
    سعدی

    Would you read the words as shesh and xosh? If yes, they won't rhyme! One needs to read them as shash and xvash.

    گر آمدنم بخود بدی نامدمی
    ور نیز شدم بمن بدی کی شدمی
    به زان نبدی که اندر این دیرِ خراب
    نه آمدمی نه شدمی نه بدمی

    I would suggest a modern Iranian speaker would be confused with words such as بدی، نامدمی، شدن، شدمی، نبدی، آمدمی، نشدی، نبدمی
     

    farzan

    Senior Member
    Standard Iranian Persian
    گل ھمین پنج روز و شش باشد
    وین گلستان ہمیشه خوش باشد
    سعدی

    Would you read the words as shesh and xosh? If yes, they won't rhyme! One needs to read them as shash and xvash.

    گر آمدنم بخود بدی نامدمی
    ور نیز شدم بمن بدی کی شدمی
    به زان نبدی که اندر این دیرِ خراب
    نه آمدمی نه شدمی نه بدمی

    I would suggest a modern Iranian speaker would be confused with words such as بدی، نامدمی، شدن، شدمی، نبدی، آمدمی، نشدی، نبدمی
    Qureshpor, there’s no confusion or a lack of resources with which to (perhaps inwardly) reinstate the rhyme.

    The existence of a standard spelling (presumably imposed at the top by official scribes and translators, and flowing downwards) does not take away the joy of a moderner’s encounter with “slippages”; if anything, the discrepancy makes sure that what’s at variance with today’s norm can continue so unravaged, appearing quaint and correct, never mind that even as Sadi wrote the line you quote he must have known that khvash was already beautifully different — see the line بنده همان به كه ز تقصيرِ خويش in the same text.

    What has happened to Persian is, a careful measure of standardization applied to spelling combined with the omission of diacritics has at the same time ensured that pronunciation of all vowels and of quite a few consonants remain open to variations brought about by accent, constraints of rhythm, and personal choice.

    In the same way the actual sounds attributed to our meagre store of written vowels and of the omitted ones too have been standardized. As a native you have a good idea of how to sound flat when reading from a text or, as an alternative, how to introduce a sort of a twang or some other pecularity into your utterance. I am, for instance, allowed to say either rasseed or resseed as I wish, and the difference would be none too great so as to mark me as modern or old, correct or incorrect, or avoiding or conforming to the dominant Tehrani accent.

    To be clear, there is a difference here in that the zeer is ever so slightly shorter, more curt and less drawled, than the zebar, but that’s about the only consideration with which to determine and receive the choice of pronunciation. What I mean is, it is insignificant socially and - dare I say it! - linguistically, being merely useful, that resseed sounds slightly shorter than rasseed.

    I am still hugely and indescribably overwhelmed that you think Khayaam confuses. It would be untrue, I totally agree, to claim that nothing has been lost or that there have not been shifts. History ensures this just could not be. However, as I fancifully say to myself, just temporarily rub out the borders between all our countries in the Persian language domain, and you will be amazed at how much has so well endured, thanks to, well, history.
     

    Qureshpor

    Senior Member
    Panjabi, Urdu پنجابی، اردو
    Hi Farzan. With due respect and if you don't mind my saying so, your response is nothing more than a "diplomatic coverup" and it does not accurately address the point I am making.

    As a matter of curiosity, what is the word for "verbosity" in Persian?
     
    Last edited:

    PersoLatin

    Senior Member
    UK
    Persian - Iran
    These and other features indeed did exist in the Classical and Early New Persian and I shall provide you two references in your "Classical Persian" thread. They don't exist in the modern language because I suspect there has been a gradual evolution in the language resulting in the current form of the language.

    I don't believe anyone is suggesting that by not reciting Classical poets' works with today's pronunciation "spoils" their enjoyment. However, the reading and understanding could be faulty. For example...

    گل ھمین پنج روز و شش باشد
    وین گلستان ہمیشه خوش باشد
    سعدی

    Would you read the words as shesh and xosh? If yes, they won't rhyme! One needs to read them as shash and xvash.

    گر آمدنم بخود بدی نامدمی
    ور نیز شدم بمن بدی کی شدمی
    به زان نبدی که اندر این دیرِ خراب
    نه آمدمی نه شدمی نه بدمی

    I would suggest a modern Iranian speaker would be confused with words such as بدی، نامدمی، شدن، شدمی، نبدی، آمدمی، نشدی، نبدمی
    My answer is here: Classical Persian: What does it mean?
     
    Last edited:

    farzan

    Senior Member
    Standard Iranian Persian
    Hi Farzan. With due respect and if you don't mind my saying so, your response is nothing more than a "diplomatic coverup" and it does not accurately address the point I am making.

    As a matter of curiosity, what is the word for "verbosity" in Persian?
    A diplomatic coverup? Perhaps you would care to explain yourself!
     

    Derakhshan

    Senior Member
    Arabic, Persian
    I am not buying the majhul vowels hypothesis because nowhere in Iran can we find a trace of it, even in the most rural areas.
    We have a majhul ē in a very small number of words.

    For example: abēnom = می‌بینم

    This is consistent with the MP stem wēn-.

    But it's not as pervasive as in Eastern Persian. For example, شیر (lion) is šir.

    We don't have any majhul ō... as far as I know.
     

    Derakhshan

    Senior Member
    Arabic, Persian
    Also, it's worth noting that not all the majhul vowels are historical CP relics in Eastern Persian. The Cyrillic е in Tajiki is always a majhul /ē/, so you get сешанбе /sēšambē/ and беҳтар /bēhtar/, these don't have majhul vowels historically.
     

    Qureshpor

    Senior Member
    Panjabi, Urdu پنجابی، اردو
    I cannot think of any word having the majhul vowel, even in my own dialect which is rural.

    Question: were all words pronounced with majhul vowels or only some? is there a poetry line which can support this claim? By the way, even in Urdu not all of them are pronounced with the majhul vowels. For example: دور duur/پیر piir/تیر tiir
    The "izaafat" sound vowel is pronounced majhuul, e.g kitaab-e-surx (The red book)

    No, not all words were pronounced with majhuul vowels. Some were ma3ruuf and others were majhuul. Examples of majhuul

    nek, eraan, hameshah, hame/me, be-baak (fearless), aavextan, aavez, aamext, aamez pesh, resh, mez (table), guft-em (we said) raft-ed (you went), mard-e (a man), raft-e (he/she/it used to go),شیر lion

    o (he/she/it), bo (smell), go (say), koshish, posh, aamoxt, josh, hosh, dosh, gosh, noshiidan

    کارِ پاکان را قیاس از خود مگیر
    گرچہ باشد در نوشتن شیر، شیر

    آن یکے شیریست کآدم مے خورد
    و این یکے شیریست کآدم مے خورد

    That one is a lion which eats man
    This one is milk which man eats (drinks)

    مولانا جلال الدّین رومی
     

    Qureshpor

    Senior Member
    Panjabi, Urdu پنجابی، اردو
    Hi Dib, purely playing the devil's advocate here:

    We only have written evidence which unfortunately does not support any of these ideas, fully, because of Perso-Arabic's lack of short vowels.

    How can we be sure that the Afghan, Tajik and the Urdu accents didn't develop from a regional accent & that the mainstream Persian was different from the aforementioned, from the start? Why, for example, can't we say Shirazi is the classical Persian accent, as it is different, i.e. the same as Afghan or Tajik accents are?
    See, تاریخِ زبانِ فارسی، پرویز ناتل خانلری page 52 volume 2 where the learned author quotes
    اسیبویہ، حمزہ اصفہانی and خواجہ نصیر who describe the existence of these vowels in Persian.
     

    Derakhshan

    Senior Member
    Arabic, Persian
    But was the last letter really pronounced "δ" and not "d"? The only dialect which has this feature is Lâmerdi and its sub-dialects; of course, the only dialect which I know of I mean, because there might be some more.
    In fact, we can find very clear evidence of this by examining Larestani. A minority of Larestani dialects have /z/ instead of /d/ in many words. For example, boda بودن, nada نهادن, xarda خوردن, in these z-dialects become boza, naza, xarza. This is clear evidence to me that these words were once pronounced boδa, naδa, xarδa.

    This happened in Persian on a smaller scale. Postvocalic /d/ in MP became /δ/ in ENP/CP. Then, it went back to /d/, except in a few words like گذشت where it became /z/.

    What happened in these z-dialects is that all of these /δ/ became /z/ instead of going back to /d/.

    Many dialects in southern Iran also retain postvocalic /δ/.
     
    < Previous | Next >
    Top