Persian: historical sound changes

< Previous | Next >

PersoLatin

Senior Member
UK
Persian - Iran
That's your theory.
Of course, that's my theory, as I have said many times. I think it is helpful in these discussions, to acknowledge it as such until it is proven to be wrong. BTW, it is no different to the mainstream view which itself is a theory, granted, the subscription to mine, is one presently, and it is viewed highly suspiciously by the mainstream believers.

What exactly is Mazdak though?
Mazdak was a Zoroastrian prophet.

Source of your claim? Wiki seems to disagree. It says Greek Manikhaios is from Syriac "Mānī ḥayyā" (The living Mani).
I used the formula which conforms with the mainstream view, i.e. Mānī + ig, the associative suffix.

It is a different breed. The MP form has no final -k/g. So, it doesn't belong here. Maybe discussed in a separate etymology thread, though.
Correct, but the MP drōzan (liar) & druxtan (to lie), seem to have some evidence of g/z/x.

Actually I forgot another consonant that OP words frequently end in: -š.

Consider this text for example (Darius Behistun I.1-15):
TITUS Didactica: Old Persian Text Sample

I am making a couple of small changes to the phonetic interpretation given on the linked page to fit it to the way I prefer it, plus adding punctuation:
Now I understand you, but I find this theory unbelievable, what theory you might say, the theory that the OP script was peppered with vowel 'a', almost randomly. None of those a's are represented in the cuneiform text. Yes there were some real vowels where those a's are, and some could even be a's. The scholars have assumed that all letters of OP alphabet, ended in an a, this is ridiculous.

So if you want to see real OP, remove 100% of a's, then compare them to their equivalent words in NP (there are many), you are very likely to find those missing vowels, you can then apply a 'reasonable' sound change rule and working backwards, find the OP pronunciations. You don't seriously believe OP didn't have any words with consonant endings (bar three).

This is another topic worthy of its own thread, but I can see how that will go if I raise it (& I have been tempted to do so this for while), I will claim "There were no a's in the OP text....", and I will be inundated with people disagreeing with it, I will have no backing, and not because I'm wrong on that point.
 
Last edited:
  • Treaty

    Senior Member
    Persian
    OP script was peppered with vowel 'a', almost randomly
    OP is deciphered using Avestan and Sanskrit because of their proximity (both in time and linguistically). The other two languages (and every recorded ancient IE language) have vowel endings. OP closely matches vocabulary, morphology and grammar of Avestan. So, it would have been random if they hadn't considered that the [a] ending.
    You don't seriously believe OP didn't have any words with consonant endings (bar three).
    In Avestan, apart from a few words ending with certain suffixes (which themselves end only with -sh, -m and -t), almost every single word ends with a vowel. Same goes to ancient Greek (n, s) and Latin (s, m, t, r, apart from prepositions). Same goes for ancient Arabic (only n, except for a few words; though for another reason). Why do you think this can't be serious about OP?
     

    PersoLatin

    Senior Member
    UK
    Persian - Iran
    Why do you think this can't be serious about OP?
    Don't you think this idea (all, or most OP words ended in a vowel) has a direct impact on the topic we are discussing? It creates an unexplained gap between the way OP & MP words were pronounced, of course I'm talking about the g/k ending in MP, and the fact they didn't exist in OP, if we were to believe, all or most OP words ended in a vowel.

    In Avestan, apart from a few words ending with certain suffixes (which themselves end only with -sh, -m and -t), almost every single word ends with a vowel. Same goes to ancient Greek (n, s) and Latin (s, m, t, r, apart from prepositions). Same goes for ancient Arabic (only n, except for a few words; though for another reason).
    And did all words of all these languages end in 'a'? I'd say the answer is no.
     
    Last edited:

    Dib

    Senior Member
    Bengali (India)
    Of course, that's my theory, as I have said many times. I think it is helpful in these discussions, to acknowledge it as such until it is proven to be wrong. BTW, it is no different to the mainstream view which itself is a theory, granted, the subscription to mine, is one presently, and it is viewed highly suspiciously by the mainstream believers.
    Well, neither theory can be proven right or wrong, short of inventing a time-machine. All we can do is evaluate how well the evidence fits the theories, or equivalently how much explanatory power the theories have, i.e. how many observed phenomena they can explain.

    The "standard" theory makes only one assumption that there was a Proto-Indo-Iranian *-ka- suffix in various meanings, including (almost) no meaning. Then using normal behaviours of language change (common sound changes, morphological generalization), it can explain the following 6 to 8 separately observed phenomena:
    1) Sanskrit -ka- suffix in various meanings, including almost no meaning.
    2) Avestan -ka- suffix.
    3) Old Persian suffix spelt -k(a)- in cuneiform, e.g. v-z-r-k- (great), a-r-š-t-i-k- (spear-thrower < a-r-š-t-i- / Skt ṛṣṭi- = spear), a-r-i-k- (enemy, Skt. ari-), m-r-i-k- (young man, Skt marya(ka)-), b-d-k- (servant, Skt. bandhaka- < *OP band-/Skt. bandh- = to bind) etc. I am not vocalizing them here beyond what appears in spelling, because it is not important for our discussion.
    4.a) Why the ending under discussion was written -k in Pahlavi. (Because that was the late BC pronunciation)
    4.b) and where it might have come from. (Morphological generalization of OP -k(a)-).
    5) Why the ending was written -g in Manichaean. (Because that was the 3rd century AD pronunciation)
    6.a) Why this -k/g does not exist any more in most New Persian words (because it was lost after vowel),
    6.b) but it still exists in others, e.g. bozorg, xošk, etc. (because after consonant).

    The probable counterexamples to this theory awaiting analysis here, as noted before, are:
    NP bārīk, tārīk, nazdīk, tājīk, Mazdak, -ak dimunitives (seemingly violate 6.a).

    What observations does your theory explain or fit to? I guess no. 6. Any other?

    Correct, but the MP drōzan (liar) & druxtan (to lie), seem to have some evidence of g/z/x.
    They do derive from OP d-r-u-g- (a lie) etc. However, the noun differs from the class of words we are discussing in two ways:
    a) It does not contain a -k(a)- in OP. (point 3 above)
    b) It does not contain a -g in Sassanian time Middle Persian. (point 4/5 above)

    Now I understand you, but I find this theory unbelievable, what theory you might say, the theory that the OP script was peppered with vowel 'a', almost randomly. None of those a's are represented in the cuneiform text. Yes there were some real vowels where those a's are, and some could even be a's. The scholars have assumed that all letters of OP alphabet, ended in an a, this is ridiculous.
    EDIT: I Removed my comments about the reconstruction of final -a vowel in OP, because whether there are final -a's was really a passing comment. We can discuss that in a separate thread if needed. The main point I wanted to make was:
    OP would have inflectional endings after those stems.
    in reply to your assertion that:
    ... OP conversations must have sounded a little odd, as every few words ended in g, and peppered with scores of common words with -ig.
    As explained before the -g was a MP feature. The corresponding OP would have been -k(a)-. My point was that stems like b-d-k- would often show up in other grammatical forms like b-d-k-m (acc sing), b-d-k-a (nom/acc plu, inst/abl sing,...), b-d-k-h-y-a (gen sing), b-d-k-a-n-a-m (gen plu), b-d-k-i-y(-a) (loc sing), etc. I am not sure if all of these forms were actually attested for b-d-k- but that's the general idea. The k wouldn't be the final element all that often, no matter whether there is an implied final -a or not.

    Don't you think this idea (all, or most OP words ended in a vowel) has a direct impact on the topic we are discussing? It creates an unexplained gap between the way OP & MP words were pronounced, of course I'm talking about the g/k ending in MP, and the fact they didn't exist in OP, if we were to believe, all or most OP words ended in a vowel.
    There is a gap, yes. But, it is unexplained only if you believe languages do not change
     
    Last edited:

    PersoLatin

    Senior Member
    UK
    Persian - Iran
    You assert:
    6 to 8 separately observed phenomena:
    2) Avestan -ka- suffix.
    3) Old Persian suffix spelt -k(a)- in cuneiform, e.g. v-z-r-k- (great), a-r-š-t-i-k- (spear-thrower < a-r-š-t-i- / Skt ṛṣṭi- = spear), a-r-i-k- (enemy, Skt. ari-), m-r-i-k- (young man, Skt marya(ka)-), b-d-k- (servant, Skt. bandhaka- < *OP band-/Skt. bandh- = to bind) etc. I am not vocalizing them here beyond what appears in spelling, because it is not important for our discussion.
    4.a) Why the ending under discussion was written -k in Pahlavi. (Because that was the late BC pronunciation)
    4.b) and where it might have come from. (Morphological generalization of OP -k(a)-).
    5) Why the ending was written -g in Manichaean. (Because that was the 3rd century AD pronunciation)
    But I think 3 to 5 are very closely linked, i.e. are based on the same single assumption, hence, in my view, don't add more weight to your argument.

    Also some inconsistencies:
    1
    b-d-k- (servant, Skt. bandhaka- < *OP band-/Skt. bandh
    This is not directly relevant to this thread, but I just noticed there is an 'n' missing in the OP b-d-k, how can this 'n' be present, in Skt bandhaka, (pre OP) and MP & NP (post OP)? Aside from that, the -ak in bandak could be the diminutive marker, maybe to make the 'servant' even lowlier than the Great King.

    2
    6.b) but it still exists in others, e.g. bozorg, xošk, etc. (because after consonant).
    I don't believe every NP (& some OP) word ending in g/k, come under this discussion, at least not my side of it. My theory applies to words ending in strong vowels (i, aa, oo) and the silent h (e or é sound), so bozorg, xošk do not qualify, as well as many others: sabok, ĉâbok, ordak, nik (nikōg qualifies) dōk (spindle), kōdak, pāk, zirak, xunak, xāk, sag, namak etc.

    3 K is diminutive marker in m-r-i-k- meaing small or young man,

    There's not enough evidence to prove g/k words existed in OP, so this shortcoming has been explained by bridging 1) Sanskrit -ka- suffix in various meanings, including almost no meaning. and 4.a) Why the ending under discussion was written -k in Pahlavi. (Because that was the late BC pronunciation)

    The probable counterexamples to this theory awaiting analysis here, as noted before, are:
    NP bārīk, tārīk, nazdīk, tājīk, Mazdak, -ak dimunitives (seemingly violate 6.a).
    I believe we can add yek (one), and Treaty's خوراک and پوشاک

    As explained before the -g was a MP feature.
    I have already said, I meant MP and not OP. However the idea of most OP words ending in 'a', would make OP sound odder than MP, with many g/k words.
     
    Last edited:

    Treaty

    Senior Member
    Persian
    Also some inconsistencies ... an 'n' missing in the OP b-d-k, how can this 'n' be present, in Skt bandhaka, (pre OP) and MP & NP (post OP)?
    OP cuniform comes short of marking nasals (m, n) after [a] (e.g., k-b-u-ji-i-y Cambyses, or z-r-k, Drangiana زرنج), however, these nasals are attested in other contemporary languages (Aramaic, Greek and Elamite).
    so bozorg, xošk do not qualify,
    You can't simply ignore evidence when it doesn't fit your theory. In these words, [k]/[g] is not a part of etymology and not also a diminutive suffix. Therefore, it is the same type of [k] in our topic.
    not sag ... [but] Mazdak, yek (one), and Treaty's خوراک and پوشاک
    In fact "sag" (Av. spa-ka from PIE *kwon) should be included in this list. I'm not sure about yak (one), as k could have come from w>g shift. I'm not sure about خوراک and پوشاک anymore. I was under the impression that they are like بینا but on the second thought, the former group are passive nouns and the latter is active. So, they can be of a different phenomenon. The -ak in Mazdak can also be (or mistreated as) a diminutive suffix (little Mazda) or even if it wasn't, the word is a proper noun. Proper nouns may not necessarily follow rules.
    ... would make OP sound odder than MP, with many g/k words.
    This is a very subjective opinion. OP speakers could have said the same thing about us. They might have not been able to pronounce NP properly because of all the stops at the end of the words. It would have been like how many ESL speakers have difficulty pronouncing the ends of English words. Either they drop it (e.g., East Asians) or add a vowel at the end (e.g., Italians). Even us Persians have problem with pronouncing these endings (e.g., has, raf and gof instead hast, raft and goft. Or omitting 3rd sing. -d from the present verbs).
     

    PersoLatin

    Senior Member
    UK
    Persian - Iran
    You can't simply ignore evidence when it doesn't fit your theory.
    Please don't forget there's no personal agenda behind this, I just want to get to the truth of the matter, and like yourself, when you doubted خوراک and پوشاک, I can dismiss bozorg and xošk, until I can be given a convincing reason not to do so, therefore saying the above (again) doesn't help. I said bozorg and xošk don't fit the rules, I set out in the original thread & also in this one, but if bozorg and xošk do fit, and the k ending is the same category, then great.

    In order to move on, I'd appreciate it if you could identify (follwing list) which word fits which category(1): alak(sieve), barg, ĉâbok, ĉâk, ĉak, dōk (spindle), farhang, kōdak, mang (dizyy), maŝk, meŝk, namak, nik (nikōg qualifies), ordak, pāk, rag (vein), rang, sabok, sang, tāk, xāk, xunak, zirak.

    EDIT: tagarg, tang, tong, nang
    EDIT: tumbag (drum, NP dombak )

    (1) - It is reasonable to assume, that not all words ending in g/k belong to the group under discussion, or the diminutive group, so we should have Normal, g/k and Diminutive.

    I took 'sag' out of this list and will add it to the other, also I now think Mazdak should belong to the diminutive group.

    I think the suffix nâk, added to these Arabic words asafnâk, xatarnâk, is important, although I can't find early recorded use of these.


    This is a very subjective opinion. OP speakers could have said the same thing about us. They might have not been able to pronounce NP properly because of all the stops at the end of the words. It would have been like how many ESL speakers have difficulty pronouncing the ends of English words. Either they drop it (e.g., East Asians) or add a vowel at the end (e.g., Italians). Even us Persians have problem with pronouncing these endings (e.g., has, raf and gof instead hast, raft and goft. Or omitting 3rd sing. -d from the present verbs).
    The reason I mentioned this was because there was a suggestion (withdrawn now) that all OP words ended in 'a' with exception of a few words. I am familiar with the phenomenon you are describing.


    OP cuniform comes short of marking nasals (m, n) after [a] (e.g., k-b-u-ji-i-y Cambyses, or z-r-k, Drangiana زرنج), however, these nasals are attested in other contemporary languages (Aramaic, Greek and Elamite).
    I just want to make sure the established sound rules changes, are not being ignored, just to get a word in a category (something I get accused of).

    I'm asking as I don't know. Is it normal to see an unwritten nasal m/n, as in say, OP and be fully pronounced and written in MP, without breaking any sound change rule? Bearing in mind that Skt. bandhaka (with 'n') must have been closer to OP version, so, in effect, we went from full 'n' (pre OP) to nasal 'n' (OP) then back to full 'n' (MP & NP).
     
    Last edited:

    Treaty

    Senior Member
    Persian
    I can dismiss bozorg and xošk, until I can be given a convincing reason not to do so,
    Have you read the link in post #25? It provided the etymology for both. bozorg comes from vazra + ka and xošk from haos- + ka. Basically, the topic of this discussion is when the apparent -ka suffix which is retained until MP but lost in NP, which you argue wasn't there in MP at all. So, your list includes the words which have this -ka in both MP and NP, minus the diminutive -ka. These two -ka are not definitely diminutive and so should fall into your category. I don't understand why you don't find this convincing.
    kōdak, mang (dizyy), maŝk, meŝk, namak, nik (nikōg qualifies), ordak, pāk, rag (vein), rang, sabok, sang, tāk, xāk, xunak, zirak.
    kodak and zirak are probably diminutives. sang, rang and rag have g as a part of their root. namak and xonak can be either of -k suffixes. I don't know about others.
    I think the suffix nâk, added to these Arabic words asafnâk, xatarnâk, is important, although I can't find early recorded use of these.
    No idea about it.
    I'm asking as I don't know. Is it normal to see an unwritten nasal m/n, as in say, OP and be fully pronounced and written in MP, without breaking any sound change rule? Bearing in mind that Skt. bandhaka (with 'n') must have been closer to OP version, so, in effect, we went from full 'n' (pre OP) to nasal 'n' (OP) then back to full 'n' (MP & NP).
    Please note that both [m] and [n] are nasal. We don't have a non-nasal [n]. Anyway, Persian (even now) has at least two qualities of ن pronunciations. Before vowels, it is pronounced with full alveolar contact (that is [n]) while before consonants, the contact is reduced (like اخفاء in Arabic). Writing it seems to have been a matter of how strong they perceived this difference and how mindful they were of their language morphology (well, of course I can't read their mind).

    By the way, please note that Manichaean and Sassanian MP are not necessarily descendant of Royal Achaemenid OP, but a general OP whose only attested dialect/accent is the Royal Achaemenid OP (inscriptions and loanwords) plus sporadic instances of personal names, some of which, possibly including pre-Darius kings, might have been of another OP dialect/accent.
     

    PersoLatin

    Senior Member
    UK
    Persian - Iran
    Have you read the link in post #25? It provided the etymology for both. bozorg comes from vazra + ka and xošk from haos- + ka. Basically, the topic of this discussion is when the apparent -ka suffix which is retained until MP but lost in NP, which you argue wasn't there in MP at all. So, your list includes the words which have this -ka in both MP and NP, minus the diminutive -ka. These two -ka are not definitely diminutive and so should fall into your category. I don't understand why you don't find this convincing.
    I am afraid you misunderstood me, please read what I wrote again, fully:
    Please don't forget there's no personal agenda behind this, I just want to get to the truth of the matter, and like yourself, when you doubted خوراک and پوشاک, I can dismiss bozorg and xošk, until I can be given a convincing reason not to do so, therefore saying the above (again) doesn't help. I said bozorg and xošk don't fit the rules, I set out in the original thread & also in this one, but if bozorg and xošk do fit, and the k ending is the same category, then great.
    kodak and zirak are probably diminutives. sang, rang and rag have g as a part of their root. namak and xonak can be either of -k suffixes. I don't know about others.
    Thank you.
     

    PersoLatin

    Senior Member
    UK
    Persian - Iran
    Would you believe Turk is in the e/k category, as Tur & Turân are the geographical areas and Turag is the associative vesion.
     

    Treaty

    Senior Member
    Persian
    I can not access this, is there another way to get this please:
    Outcomes of the Indo-Iranian suffix *-ka- in Old Persian and Avestan
    You simply need to sign up in the website and download it (both are free). You can even connect to it with Facebook without signing up for the website. Anyway, you should also be able to read it online without signing in. Otherwise, you may need to update or change your browser (unless the problem comes from country-based copyright). Alternatively, you can download another more-related article by the same author (even without signing in). It has information about some of those words you asked me about (I was incorrect about sang).
    Would you believe Turk is in the e/k category, as Tur & Turân are the geographical areas and Turag is the associative vesion.
    Turk (a Turkish word) and Tur (an Iranian word) are very likely to be unrelated, both etymologically and ethnologically. There is no reason to believe in their relation. However, تورج (is Turag the MP versian?) is the associative/diminutive version of Tur, comparable to Iraj (Av. airika) that is of Air/Arya.
     

    PersoLatin

    Senior Member
    UK
    Persian - Iran
    Turk (a Turkish word) and Tur (an Iranian word) are very likely to be unrelated, both etymologically and ethnologically.
    This image is the basis of my #post 61, which shows three consecutive lines from MacKenzie Pahalavi dictionary. Unless this dictionary is no longer viewed as reliable/accurate.

    Also from here:
    tūrestān name of a land < *tugra-stāna- (cf. Parth. tugrestān). This is just an observation, it seem the modern word ترکستان hasn't changed much.

    Either way, تورج then can be added to the list.

    However, تورج (is Turag the MP versian?) is the associative/diminutive version of Tur, comparable to Iraj (Av. airika)
    Would be correct to say Iraj refered to an Iranian, airika and Turaj to someone from the land/people of Tur which, by all accounts, spoke Turkish?
     

    Attachments

    Last edited:

    Treaty

    Senior Member
    Persian
    During the middle ages, Persian did confuse between the identity of ancient Turs in Avesta and contemporary Turks, which were considered as hostile to Iranians as the Turanians. So, if تور/تورج is used to connote Turk, it is not an etymological connection but an anachronic confusion by Iranians. It's not helpful in our discussion.

    Tugrestan (Tugran, Turgestan or Turestan) was a land in modern Balochistan/Pakistan. It is different from Turkestan of Central Asia. I don't know about its etymology (my dictionary suggests possible connection to Tochar).
     

    PersoLatin

    Senior Member
    UK
    Persian - Iran
    Found this here (p15):
    xiŝt ‘brick’ < *iŝta- (cf. OPers. iŝti-, Av. iŝtya-, Parth. hiŝtig).

    The OP and Av. versions, are missing the g ending, is this consistent with the mainstream theory?
     
    Last edited:

    Treaty

    Senior Member
    Persian
    Found this here (p15):
    xiŝt ‘brick’ < *iŝta- (cf. OPers. iŝti-, Av. iŝtya-, Parth. hiŝtig).
    The OP and Av. versions, are missing the g ending, is this consistent with the mainstream theory?
    So is the MP word. You seem to have forgotten what you were reading.
     

    PersoLatin

    Senior Member
    UK
    Persian - Iran
    So is the MP word. You seem to have forgotten what you were reading
    So are you saying Parth. haštig is not MP? I was trying to say that the mainstream theory claims that the g/k ending was in prevalent in OP & Avesta too, but that's not the case with xišt.
     
    Last edited:

    Treaty

    Senior Member
    Persian
    So are you saying Parth. haštig is not MP? I was trying to say that the mainstream theory claims that the g/k ending was in prevalent in OP & Avesta too, but that's not the case with xišt.
    Parthian is technically MIr not MP, but it's irrelevant. Suffix -ka was common in OP and Avestan but this doesn't mean we have all their -ka words available to us (though in this case we have Skt. iśtika "brick"). Neither all the g/k words in MP had an existing -ka ancestor in OP. For example, گاه was common in early and classical NP; however, this doesn't mean they had the word زایشگاه which is found in Modern Persian. To put it another way, what you are doing in regard to gotcha-comparing MP and OP seems the wrong approach.
     
    Last edited:

    PersoLatin

    Senior Member
    UK
    Persian - Iran
    (though in this case we have Skt. iśtika "brick").
    Thank you, I can see the mistake.

    To put it another way, what you are doing in regard to gotcha-comparing MP and OP seems the wrong approach.
    May be the reason for this is, the seemingly contradictory information that this thread may be suffering from . There was an assertion by Dib (post #54), that g/k existed in OP and that (fact) is then used to add weight to the mainstream argument, which says (or does it?), g/k ending didn't just appear in MP and was the continuation of a process, started in previous language(s). Anyway isn't the case that the Av. script was developed sometime in the MP period?

    For example, گاه was common in early and classical NP; however, this doesn't mean they had the word زایشگاه which is found in Modern Persian.
    I was hoping, by now, I have earned enough credit such that you don't have to use examples like this to make a point, (I need to try harder).
     
    Last edited:

    Treaty

    Senior Member
    Persian
    May be the reason for this is, the seemingly contradictory information that this thread may be suffering from . There was an assertion by Dib (post #54), that g/k existed in OP and that (fact) is then used to add weight to the mainstream argument, which says (or does it?), g/k ending didn't just appear in MP and was the continuation of a process, started in previous language(s).
    I don't see any contradictory information, apart from the etymology of some of the examples (which comes from my lack of knowledge), and a dozen of exceptions. Dib's claim was about the continuity of the suffix, not the continuity of every word with that suffix.
    Anyway isn't the case that the Av. script was developed sometime in the MP period?
    Yes it was. However, you should note that the accurate and correct pronunciation of Avestan words (as well other recited religious languages like Quranic Arabic and Sanskrit) was of utmost importance and religious duty (this is the only thing that Magi were really conservative about). The magi, like their Hindu counterparts, were trained from childhood to recite every single verse in Avesta by heart and accurately. It only makes sense for them to show the same pedanticness to its transliteration. The systematic and accurate corresponding between Gathic and Vedic, which were separately gone through transliteration, attests to this pedanticness even 1500-2000 years after those two languages.
     

    PersoLatin

    Senior Member
    UK
    Persian - Iran
    apart from the etymology of some of the examples (which comes from my lack of knowledge)
    I, I hope, like others, allow for this. Anyway, I really can't think of many of these, if you are referring to your examples of خوراک and پوشاک, then I certainly believe they belong to the group under discussion, to me they are good, but if some are not 'good' they'll stay under discussion. Anyway the good few words we've collectively found, haven't affected the wholesale mainstream theory, yet.

    The magi, like their Hindu counterparts, were trained from childhood to recite every single verse in Avesta by heart and accurately.
    I don't know this, do you know if in modern recitations of Av., these g/k endings are still used?
     

    Dib

    Senior Member
    Bengali (India)
    I took the previous couple of days away from here to educate myself a bit more on Middle Persian. One of the things, I realized, is that the MP situation is quite complicated, and with my almost zero knowledge (beyond some grammar) of MP and Avestan, and rather basic knowledge of OP and NP, I can't possibly contribute to this discussion too much that is meaningful as well as accurate. Check this paper out to get a sense of the complications, we are talking about:
    C.Ciancaglini. "Variability of Middle Persian continuants of the Old Iranian suffix *-ka-"

    So, I'll wind down my participation on this thread, and move to being more of an observer than contributor.

    There's not enough evidence to prove g/k words existed in OP
    The -k(a)- suffix is attested in OP, as has been given examples of in this thread. I hope you could also access Ciancaglini's other paper, the first one that Treaty linked to. It contains a good survey of the attested forms. If you haven't still been able to get it, let me know. I can send it to you. Also note, it seems to me (of course, I may be wrong) that you are distinguishing between diminutive -k and other (meaningless) -k. Within the standard theory, they are not distinguished. That's why I listed NP diminutive -ak as a difficulty for the theory, as it should have been lost. They are taken to belong to the same range of meanings of the -ka- suffix. The basic marker of its usage in Old Indic and Iranian (with a small set of exceptions) seems to have been low sociolinguistic status. However, it is true that OP -k(a)- suffix is relatively rare, which is not unexpected given the formal/administrative characters of the bulk of the attestations of the language. Ciancaglini addresses this point in this paper as well.

    With this reservation in mind, I think your folowing idea is basically correct:

    ... so this shortcoming has been explained by bridging 1) Sanskrit -ka- suffix in various meanings, including almost no meaning. and 4.a) Why the ending under discussion was written -k in Pahlavi. (Because that was the late BC pronunciation)
     

    PersoLatin

    Senior Member
    UK
    Persian - Iran
    Hi Dib,

    Without what you brought to this (and Treaty) we wouldn't have got this far, some might argue we haven't moved at all, but I believe we have. You certainly know a huge amount more than I do about all of these languages, but that won't stop me, unless of course I get thrown out of the forum. Thank you for your contributions so far, and I hope in the future. BTW, any information is good, it opens up the discussions, and in most cases, when not good/accurate, leads to accidental finds.

    This is very apt:
    Check this paper out to get a sense of the complications, we are talking about
    I tested & can access your link.

    especially considering that this semantic change happened long after very different abāg and pad changed to more similar and ba in Dari
    I believe the 'unexplained' د -d, in the following three words بدان, بدین, بدو standing for به آن, به این, به او, is the legacy of the /d/ in the MP word for 'to' i.e. 'pad'. I said 'unexplained' as I have not seen any explanation of this د -d before.
    So far, I can only think of 3 words (mentioned before), that fall in this category. The survival of د -d in those words, is very interesting and my theory is that, the /d/ in was added as a liaison when it was used before a word starting with a vowel, like آن, این, او. It is also interesting that it hasn't survived in similar words like به من or به تو as بد من and بد تو, since these don't start with a vowel.

    Edit: In fact I just thought of another possible word بداسمان standing for به آسمان and after a quick search, I found it here, with a similar explanation,
     
    Last edited:

    Dib

    Senior Member
    Bengali (India)
    I believe the 'unexplained' د -d, in the following three words بدان, بدین, بدو standing for به آن, به این, به او, is the legacy of the /d/ in the MP word for 'to' i.e. 'pad'.
    Yes. I believe so too. This MP 'pad' in its turn, I guess, would have derived from Old Persian pati(p-t-i-y)=various meanings including to, at, etc.
     

    Derakhshan

    Senior Member
    Arabic, Persian
    MP xāyag "egg" has survived in Iranian dialects as xāg, and this form is listed in dictionaries too. Would this count as evidence of final -g being pronounced?
     

    PersoLatin

    Senior Member
    UK
    Persian - Iran
    MP xāyag "egg" has survived in Iranian dialects as xāg, and this form is listed in dictionaries too. Would this count as evidence of final -g being pronounced?
    Yes that appears to be the case, although why did /ya/ disappear? Is it possible that xāg was re-borrowed say, from Armenian, where some words with g/k ending have survived? In this case shortened too.

    It is interesting that xāg survives in modern Persian in the form of خا گینه/xâginé/xâgina, an omelette made from beaten eggs only.
     
    Last edited:

    Derakhshan

    Senior Member
    Arabic, Persian
    Also, there is MP paristōg "swallow", which in NP is usually parastu, but the form parastuk exists as well (and in some dialects farastuk/faraštuk).
     

    Derakhshan

    Senior Member
    Arabic, Persian
    پرستوک is listed in dictionaries as an alternative form of پرستو.

    Southern Lori has faraštuk.

    Edit: there's even a form farastog فرستگ is listed in Dehkhoda.
     
    Last edited:

    PersoLatin

    Senior Member
    UK
    Persian - Iran
    پرستوک is listed in dictionaries as an alternative form of پرستو.

    Southern Lori has faraštuk.

    Edit: there's even a form farastog فرستگ is listed in Dehkhoda.
    Thanks, I saw them.

    More and more of these word types, where g/k ending are pronounced, are appearing, I don't know which side of the argument this supports.
     

    Derakhshan

    Senior Member
    Arabic, Persian
    More evidence of final -g:

    بیوگ bayūg, alternative form of بیو bayū "bride" (< MP wayōg).

    Occurs in Bushehri as beyg.

    Lari sāg "shadow" < sāyag

    Khargi hamseyg "neighbor" < hamsāyag

    All of these dialectal forms were preserved through deletion of the final vowel.
     

    Derakhshan

    Senior Member
    Arabic, Persian
    Is this not the Turkish بیگم and بیگ?
    It is from بیوگ bayūg "bride" (look this up in the dictionary), from MP wayōg "bride". It occurs in Lori as bayū and Lari as baü or bay with same meaning. Seems to be one of those rare words that didn't completely drop final -g from all it's forms, like فرستگ above.

    Both words have a -ōg ending (wayōg, faristōg) so perhaps that has something to do with it?

    More on MP wayōg, you can see here that it's from proto-Indo-Iranian *wedʰ-úHs “bride” and in fact cognate with wedding: Reconstruction:Proto-Indo-European/wedʰ- - Wiktionary

    If you mean that Bushehri beyg in particular is from Turkish, I doubt it since it carries the same meanings (both "bride" and "doll") as عروس and عروسک.

    As you know some words of this kind were re-imported into Persian from Arabic, Armenian, I am not saying sāyag is definitely one of them but the geography of Kharg points that way.
    Doubt it since I haven't heard of همسایه ever being borrowed into Arabic, and if it were then we should expect something like همسايج or همسايق, while these words retain a [g].

    sāyag > sāg also mirrors the development of xāyag > xāg discussed in a post above.
     

    Derakhshan

    Senior Member
    Arabic, Persian
    Oh, another good one:

    MP dīg and parīg (دیروز and پریروز) still retain the final -g in some dialects such as Kazeruni and Lori. In Kazeruni they are dīgru and parīgru.
     

    Treaty

    Senior Member
    Persian
    Turkic (or Mongolian?) بیگ is used for men (the ـم suffix makes it for women, cf. خان vs خانم). It is definitely not the same Iranian words for "bride".
     

    PersoLatin

    Senior Member
    UK
    Persian - Iran
    I am adding two more words to the tally of words which kept their g/k ending:
    روا borrowed into Arabic as ‏رواج where the g/k ending changed ج, reborrowed into Persian
    آواز the same as آوا, with a non-etymological /z/ ending which must have come from g/k ending
     

    fdb

    Senior Member
    French (France)
    آواز the same as آوا, with a non-etymological /z/ ending which must have come from g/k ending
    This is a bit more complicated. There is a sound-law in proto-Indo-Iranian that says that /k/ becomes /č/ before IE front vowels. Thus the term for “voice, word”, with the root *wak in ablaut with *wāk, has the nominative singular *wākš (Avestan vāxš, like Latin vox) and the accusative singular *wāčam (Avestan vāčəm; compare Latin vocem). Middle and New Persian āwāz is from the accusative, with pre-verb ā and the regular MP change of post-vocalic /č/ to /z/. This is all perfectly regular. What is irregular and non-etymological is the loss of the final consonant in the (poetic?) NP form āwā.
     

    PersoLatin

    Senior Member
    UK
    Persian - Iran
    Many thanks for the clarification fdb.

    In fact before posting I looked for آواختن/āwāxtan with آواز/āwāz being its present stem, but couldn't find anything. I have since found واژیدن/wāžidan "to speak/talk" in Dehkhoda.
     
    Last edited:
    < Previous | Next >
    Top