MSA/All dialects: Form I voweling patterns and correspondences

< Previous | Next >

wriight

Senior Member
English (US) / Arabic (Lebanon)
MSA has three possible voweling patterns for each tense of a Form I verb: "fa3ala, fa3ula, fa3ila" and "yaf3ilu, yaf3alu, yaf3ulu". (My understanding is that archaic forms of Arabic also had an alteration yi- for some imperfective verbs.) However, I'll admit I'm not actually positive how these correspond to each other as conjugations of the same verb. Is it possible, for instance, to have fa3ila-yaf3ilu, fa3ala-yaf3ulu, fa3ula-yaf3alu, etc?
And in any case, MSA has a ton of Form I verbs that are identical but for their voweling, usually with different meanings.

And then come dialects. Morphologically, Levantine Arabic has two voweling patterns for each tense: "fi3il, fa3al" and "yif3il, yif3al". The form yif3ul also pops up unpredictably as an equivalent and/or variant of yif3il -- I was looking at the Lughatuna dictionary, though, and it seems to indicate that yif3ul is only particularly common to Lebanon among Levantine dialects, and that Syrian tends toward yif3il? Is this accurate?
And, besides the arbitrary "yif3il/yif3ul" dilemma, some Levantine dialects also have the phonologically-conditioned forms "fu3ul (or fu3il)" and "yuf3ul, yuf3al", which only appear when there are emphatic consonants. (fa3al is unaffected by this.) But they're still morphologically the same as before. Typically, Levantine fi3il-yif3al and fa3al-yif3il correspond to one another, but mix-and-matching is common -- particularly with final-weak verbs, which in Lebanon and I think urban Syria tend toward fi3il-yif3il (i.e. fi3i-yif3i). Does most of the rest of the Levant tends toward preserving the fi3il-yif3al and fa3al-yif3il correspondences?
But anyway, this results in a very-feasibly-countable number of minimal voweling pairs, unlike in MSA. In Lebanese, I think at least the following nine can be identified -- first, the eight I listed in this post, and the following one I noticed today:
  • حكم_يحكم
    • 7ekem-ye7kam: I think only said in the 3rd-person feminine of a heart attack. 7ekmetne sakte 2albiyye aw ra7 te7kamne
    • 7akam-ye7kom: All other usages of the verb. l-wā7ad bte7kmo ẓroufo, l-2āḍe 7akam kaza, etc
(EDIT: There's also the ghala-yeghle/ghele-yeghla pair mentioned by akooha here, which had completely slipped my mind.)

Do more pairs exist in other Levantine dialects? What about in other dialect groups altogether? For instance, from discussing with a Hijazi acquaintance, I'm led to understand that none of those voweling pairs exist in Hijazi Arabic... but it still does seem like Form I verbs have more than one voweling pattern in each tense. Could anyone elaborate?

And I'm curious to know what it's like for other dialects, too.
 
Last edited:
  • wriight

    Senior Member
    English (US) / Arabic (Lebanon)
    Wow, yes! And that looks even cooler than what happens in Levantine — remnants of أفعلَ do show up (a major tell being the presence of a ي in a middle-weak Form I verb where the root has a و... and I see Josh_ noticed this a full 10 years before I did, lol), but its function as a causative has been thoroughly discarded. So, among pairs of verbs in Egyptian Arabic that are only distinguished by their voweling, are there any that aren't linked by a "non-causative vs. causative" relationship? Is it productive, e.g. would you be understood if you were to take a random transitive fa3al verb and use a fi3il variant of it? And are there any EA verbs whose conjugation follows fi3il-yif3il or fa3al-yif3al rather than the two "normal" pairings? (An easy example for the second is قرا-يقرا, but that might be cheating.)

    That post also adds another Levantine example of each verb in a former active/passive pair being lexicalized individually as an active-voice Form I verb, and on second thought I think I do remember reading it in Cowell: 2itil-yi2tal.

    Hm. Re the Najdi Arabic comment showing total preservation of the internal-voweling passive (which confirms both what Wikipedia claims and what I understood from correspondence with a speaker of a Bedouin dialect that also preserves it), do these dialects also have varied active-voice voweling patterns?
     

    analeeh

    Senior Member
    English - UK
    And then come dialects. Morphologically, Levantine Arabic has two voweling patterns for each tense: "fi3il, fa3al" and "yif3il, yif3al". The form yif3ul also pops up unpredictably as an equivalent and/or variant of yif3il -- I was looking at the Lughatuna dictionary, though, and it seems to indicate that yif3ul is only particularly common to Lebanon among Levantine dialects, and that Syrian tends toward yif3il? Is this accurate?
    I think it's more accurate to say that there are three present tense vowellings, yif3el, yif3ol, yif3al (or i u in Lebanon I guess), and that the former two tend to coalesce/co-occur while the latter is much more distinct. In Lebanon and coastal areas of Syria lots of words that in Damascus have yif3el have o instead (yimsok, ya3mol...) and you can make similar comments about distribution across the whole area, but at least within Damascene for example there are yif3ol forms and yif3el forms which are only produced with that vowel. It may be a coastal feature to collapse the paradigms into yif3ol (these two paradigms after all are identical in NL as soon as suffixes are attached and the stem vowel is only distinct in the absence of any sort of suffix).

    And, besides the arbitrary "yif3il/yif3ul" dilemma, some Levantine dialects also have the phonologically-conditioned forms "fu3ul (or fu3il)" and "yuf3ul, yuf3al", which only appear when there are emphatic consonants. (fa3al is unaffected by this.) But they're still morphologically the same as before.
    Which dialects are these?

    Typically, Levantine fi3il-yif3al and fa3al-yif3il correspond to one another, but mix-and-matching is common -- particularly with final-weak verbs, which in Lebanon and I think urban Syria tend toward fi3il-yif3il (i.e. fi3i-yif3i). Does most of the rest of the Levant tends toward preserving the fi3il-yif3al and fa3al-yif3il correspondences?
    Mixing and matching is common even in Classical Arabic, so I'm not sure this is necessarily a matter of dialects diverging from the pattern and creating new irregularities. There are some examples of fi3i-yif3i which are pan Levantine (مشي يمشي, بكي يبكي) but I can't think of very many and I would say the pattern in urban Syrian is still fa3a-yif3i/fi3i-yif3a with the latter typically being intransitive change-of-state verbs (رضي يرضى شفي يشفى etc).

    Do more pairs exist in other Levantine dialects? What about in other dialect groups altogether? For instance, from discussing with a Hijazi acquaintance, I'm led to understand that none of those voweling pairs exist in Hijazi Arabic... but it still does seem like Form I verbs have more than one voweling pattern in each tense. Could anyone elaborate?
    I can't think of any more pairs within the patterns you mention but for hollow verbs there's يقوم يقيم يدوم يديم يدور يدير which are presumably all remnants of 2af3al. There are actually a lot of form I yif3els which correspond to 2af3als in fuS7a but not all of them have an underlying intransitive verb with a different vowelling.
     

    Ghabi

    Senior Member
    Cantonese
    Is it possible, for instance, to have fa3ila-yaf3ilu, fa3ala-yaf3ulu, fa3ula-yaf3alu, etc?
    There are only five possibilities, when one thinks about it:

    1) fa3ala - yaf3al
    2) fa3ala - yaf3il
    3) fa3ala - yaf3ul
    4) fa3ila - yaf3al*
    5) fa3ula - yaf3ul

    1) has to do with the phonetic quality of the second or the third consonant of a verb, not with semantics. 2) and 3) are arbitrary. 4) is often for intransitive verbs. 5) is for intransitives verbs about characteristics of a person or a thing.

    Some verbs have more than one pair of these (usually one intransitive, the other transitive).

    *(There is the interesting exception 7asiba - ya7sib/ya7sab. The variants seem to have been resulted from ancient dialectal differences (cf. noun variants like udhun/udhn, also supposed to be a result of ancient dialects).)
     

    wriight

    Senior Member
    English (US) / Arabic (Lebanon)
    (or i u in Lebanon I guess)
    Oh, no, it's definitely e o in Lebanon, too. (For me, at least. Probably is i u in areas of Lebanon that don't collapse high/mid short vowels like I do and/or tend more toward high vowels.) I just used i and u because it seemed they'd be more general.

    I think it's more accurate to say that there are three present tense vowellings, yif3el, yif3ol, yif3al (or i u in Lebanon I guess), and that the former two tend to coalesce/co-occur while the latter is much more distinct.
    at least within Damascene for example there are yif3ol forms and yif3el forms which are only produced with that vowel.
    Hmm. Are there verbs with different meanings would be pronounced identically if not for that o/e vowel? My assertion that they're variants of the same form is indeed influenced by that (1) they both correspond to the same past-tense voweling, and (2) they're interchangeable in Lebanese; the o/e vowel is never semantic. But your description works too, especially if the vowel choice is semantic elsewhere in Levantine?

    EDIT: Oh! Right! I forgot the other big reason for my referring to them as variants of one another: they both collapse into yif3il as soon as anything is suffixed, whereas the a of yif3al remains distinct.

    Which dialects are these?
    Well, mine, for one. :p I don't have concrete stats, but it absolutely exists elsewhere in Lebanon according to that survey thing I keep mentioning and I'm 99% sure I've heard it from a Syrian, too.

    but for hollow verbs there's يقوم يقيم يدوم يديم يدور يدير
    Oh yeah, neat! The only three of these pairs for me are عاد and دار and هان -- with يديم existing marginally/in a religious context only, like the "to guide religiously" meaning of يهدي.

    There are actually a lot of form I yif3els which correspond to 2af3als in fuS7a but not all of them have an underlying intransitive verb with a different vowelling.
    Absolutely. There are 5 possible outcomes for أفعلَ by my count, and it's a bit tricky to determine if you don't have an obvious tell like that middle hollow-verb vowel.
     
    Last edited:

    analeeh

    Senior Member
    English - UK
    Hmm. Are there verbs with different meanings would be pronounced identically if not for that o/e vowel? My assertion that they're variants of the same form is indeed influenced by that (1) they both correspond to the same past-tense voweling, and (2) they're interchangeable in Lebanese; the o/e vowel is never semantic. But your description works too, especially if the vowel choice is semantic elsewhere in Levantine?
    There might not be many verbs distinguished by the pattern, but there are verbs that even in Lebanese only occur with one or the other - does anyone really say ما بعرُف for example? So it's lexically determined even if it isn't often semantically contrastive.
     

    lukebeadgcf

    Senior Member
    English – US
    As far as I understand, in formal Arabic the perfect and imperfect stem vowels of a verb don't change depending on whether there are attached object pronouns. However, I've heard Syrians from Latakia use different vowels for the verb قري depending on whether an object pronoun is attached. I asked about this, and (I think) they said that the مضارع could take the following forms:

    For "I read" without an attached object pronoun: بِقْرا only

    For "I read" with an attached object pronoun: بقراها OR بقريها

    So you could say: عم بقرا لكتاب لي عم بقريا

    Relatedly, Cowell says the following (p. 63):

    1577571944492.png


    This is all to say that there are perhaps additional considerations (such as whether an object pronoun is attached) to take into account when discussing form I vowelling patterns in non-formal varieties of Arabic.
     

    wriight

    Senior Member
    English (US) / Arabic (Lebanon)
    Yeah, if you Google "ba3rof" site:ask.fm you'll get at least two hits that read as distinctly Lebanese, which of course indicates a much-larger offline presence for بعرُف.

    Cowell is on point here, although he doesn't connect the dots entirely: what this "obliteration" actually shows us is that yif3ul is morphologically yif3il, and that the arisal of the u is only a surface-level effect on pronunciation. My comments here basically explain it; there's no precedent or reason for a non-epenthetic u as in *yiktubu to be elided, but there definitely is precedent for elision of a non-emphatic non-epenthetic i as in *yiktibu, showing us that when an extra syllable is appended to yif3ul verb it automatically becomes *yif3il- before other processes affect the new i.

    So it's lexically determined even if it isn't often semantically contrastive.
    Lexically determined per speaker, for sure, but my point's that there's no use attempting to distinguish them in any bigger-picture context.

    (Also, clarifying my "I just used i and u because it seemed they'd be more general" above: I'm using i u because they're completely allophonic with e o except word-finally, and it can be determined completely automatically which one will surface in a particular context. Well, I think. I'm not sure I really understand Cowell's notation fi3el, because at least in Lebanon, it's impossible for /i/ to be stressed unless it precedes /j/ -- stress is one of the contexts that conditions the allophone /e/ instead. Is this moot in Syria, and is the Syrian pronunciation of pronunciation fi3el actually audibly [fi]-initial?)
     
    Last edited:

    analeeh

    Senior Member
    English - UK
    As far as I understand, in formal Arabic the perfect and imperfect stem vowels of a verb don't change depending on whether there are attached object pronouns. However, I've heard Syrians from Latakia use different vowels for the verb قري depending on whether an object pronoun is attached. I asked about this, and (I think) they said that the مضارع could take the following forms:

    For "I read" without an attached object pronoun: بِقْرا only

    For "I read" with an attached object pronoun: بقراها OR بقريها

    So you could say: عم بقرا لكتاب لي عم بقريا
    I'm quite familiar with Latakian forms. The system overall is that final -i becomes (or can become at least in more traditional dialects) -aa when suffixes are added. This extends as far as words like فيه, rendered in Latakian as فاه faa (the existential I mean). I think this has contributed alongside the general pull of regularisation to the anomalous -a -a pattern of قرى being regularised as 2ara/yi2ri when suffixes aren't present. This is combined with a broader coastal tendency to have all suffixes trigger a stress shift to the final stem syllable (as opposed to e.g. Damascene where the suffix only triggers a stress shift if its structure closes the final syllable and means it is stressed by default according to the normal rules of prosody). I seem to remember you saying something about Lebanese suffixes that reminded me of this?

    Cowell is on point here, although he doesn't connect the dots entirely: what this "obliteration" actually shows us is that yif3ul is morphologically yif3il, and that the arisal of the u is only a surface-level effect on pronunciation. My comments here basically explain it; there's no precedent or reason for a non-epenthetic u as in *yiktubu to be elided, but there definitely is precedent for elision of a non-emphatic i as in *yiktibu, showing us that when an extra syllable is appended to yif3ul verb it automatically becomes *yif3il- before other processes affect the new i.
    I'm not sure I agree with you here. What is the u form a surface effect of in your analysis? I'd be more inclined to agree if there was any evidence that the vowel is determined non-lexically. (To be fair I think our disagreement might just be rooted in different approaches to describing morphology rather than anything more meaningful). All this 'obliteration' shows us to my mind is that u/i merge when stressed. Cowell also ignores the existence of forms without elision: those suffixes that don't cause a stress shift don't necessarily lead to 'obliteration' of the difference: btiktubi exists alongside btikitbi (apologies for imprecise transliteration).
     

    wriight

    Senior Member
    English (US) / Arabic (Lebanon)
    Pardon my typo of "epenthetic" -> "emphatic" above; edited & fixed. (I've been having lapses like that more and more lately, never used to happen. :()

    btiktubi exists alongside btikitbi (apologies for imprecise transliteration).
    Well, this may indeed crush my argument! My iffily-expressed point was just that there's no precedent for elision of u as in *tiktubi -> tiktbi, but that there's tons of precedent for elision of i as in *tiktibi -> tiktbi, so therefore tiktub must be becoming tiktib- underylingly (in response to the addition of a syllable) in order for that elision to happen. This naturally lines up with the surfacing of the i in cases where the new syllable causes it to be stressed. I'm not sure what to make of tiktubi, though, because it of course indicates otherwise.
    Are you sure this also doesn't appear when the problem vowel is stressed, i.e. that there's no variety in Syria that uses tiktubi and also tiktubha rather than tiktibha? If there is, then yeah, we can say that there are varieties where yif3ul is its own morphologically-distinct nonpast voweling, and I guess that it's on a spectrum where there are also varieties somewhere in between. If not, then would you happen to know whether pronunciations like tiktubi are a newer innovation rather than an older form?

    What is the u form a surface effect of in your analysis?
    The general phenomenon of free variation between i and u in North Levantine. Consider the variation in nouns on the patterns of lugha/ligha, kutub/kitib, 3umr/3imr, as well as the leveling of the contracted hollow verb's u to i. The only true i vowels we find (where there's no variation whatsoever and they can only surface as i no matter who you ask) end up being the same i vowels that are elided when they're unstressed + not word-final. I think "Lexically determined per speaker, for sure, but [...] there's no use attempting to distinguish them in any bigger-picture context," sums up my standpoint here, with the addition of reiterating that there's also no underlying, morphological difference for speakers without the tiktubi realization that you mention.
     

    analeeh

    Senior Member
    English - UK
    Are you sure this also doesn't appear when the problem vowel is stressed, i.e. that there's no variety in Syria that uses tiktubi and also tiktubha rather than tiktibha? If there is, then yeah, we can say that there are varieties where yif3ul is its own morphologically-distinct nonpast voweling, and I guess that it's on a spectrum where there are also varieties somewhere in between. If not, then would you happen to know whether pronunciations like tiktubi are a newer innovation rather than an older form?
    I don't know how old they are,, although Cowell doesn't mention them. But given that separate u/i present paradigms exist in fus7a (and perhaps more importantly in SL), I would read them as distinct paradigms that are brought together by two main features: 1) the systematic merger of the two vowels in stressed syllables (although this feature is far from universal thanks to fuS7a loans it seems to be pretty consistent morphologically at least) and 2) the general impulse towards u/i being somewhat but not entirely interchangeable, which is a phenomenon found across Arabic (2umm/2imm, jibne/jubne, etc) and not just in NL.

    The general phenomenon of free variation between i and u in North Levantine. Consider the variation in nouns on the patterns of lugha/ligha, kutub/kitib, 3umr/3imr, as well as the leveling of the contracted hollow verb's u to i. The only true i vowels we find (where there's no variation whatsoever and they can only surface as i no matter who you ask) end up being the same i vowels that are elided when they're unstressed + not word-final. I think "Lexically determined per speaker, for sure, but [...] there's no use attempting to distinguish them in any bigger-picture context," sums up my standpoint here, with the addition of reiterating that there's also no underlying, morphological difference for speakers without the tiktubi realization that you mention.
    That makes sense, but the usual explanation in North Levantine (for which see e.g. Cowell) is not that i and u are in free variation but that they merge in stressed syllables to a vowel that is typically realised as either a schwa or a high front /i/, with this realisation determined by phonological context. It's true that this explanation is in itself incomplete because this vowel appears in all sorts of different unstressed contexts as well, but I'm not sure that's directly relevant to the argument - unless I'm misunderstanding you.
     

    wriight

    Senior Member
    English (US) / Arabic (Lebanon)
    I just read this passage from the Levantine Arabic chapter of the new 2nd edition of The Semitic Languages, an otherwise-fantastic book. It was the final straw in convincing me that the schwa is a harmful, pointless abstraction that has no place in discussions of Levantine-in-general:
    1583015471259.png


    <3omor> very much does reflect the Lebanese non-urban "normal pronunciation" /ʕumər/ [ʕomoɾ], which contrasts with the urban "normal pronunciation" /ʕimər/ [ʕemeɾ]. (I'm using /ə/ for the CVCC epenthetic and nothing more.) Whatever their supposed "normal pronunciation" <ʕəmər> is meant to be (considering it's apparently distinct from both of the actual pronunciations), it's certain that it doesn't actually exist anywhere in Lebanon. It seems to me that Cowell's schwa, in order to bring about disappointing misanalyses like that one, improperly merges a few distinct phenomena:
    1. The regional and lexical variation /i/~/u/ we've talked about.
    2. This unconditioned, Damascene-specific centralization(?) of /i/, which among other things apparently creates two distinct pronunciations ʔə́za/ʔíza of إذا. I don't believe this makes any sense when discussing Lebanese. In Damascene, it should overwhelmingly be considered a newer reflex of /i/ and of /i/ alone (except for minor confounding cases like "mə́mken" < múmken? I see that "memken" occurs in Jordanian, so Damascene mə́mken < mémken < múmken is plausible IMO, but I'm not sure) — that means it only operates after the shift /i/</u/, so it shouldn't have any bearing on diachronic stuff that concerns the distinct treatment /i/ and /u/.
      And yeah, it has nothing to do with stress. Cowell gives the examples "ləzzēʔa" (not sure why ē rather than ā, huh) and "təlmīz", both of which show this apparent ə<i in an unstressed syllable.
    3. The centralization/backing of /i/ that's conditioned by proximity to a pharyngeal and occurs in all regions, as seen in /nƏħna/ "we", /rƏħət/ "I went", etc. (Capitalizing it just to distinguish it from the epenthetic, because unlike the epenthetic, this one actually does phonetically approach [ə].)
    4. The lowering+rounding of /i/ that's conditioned by proximity to an emphatic and occurs sporadically in all regions. This is in fact just an emphasis-conditioned /u/</i/, and that's supported by 3arabizi transcriptions, where writers seem to always represent this using <o>.
    5. And the CVCC epenthetic, because even though Cowell is careful to only write it as a superscript ᵊ, it seems that that memo didn't reach the authors of this Levantine Arabic chapter.
    None of these have any business being lumped under the same symbol or the same description as any of the other three. But even if we extracted and separated #2, it would still be misunderstood to be a merger of /i/ and /u/ — when, again, it looks like it's overwhelmingly only a reflex of /i/.

    but the usual explanation in North Levantine (for which see e.g. Cowell) is not that i and u are in free variation but that they merge in stressed syllables to a vowel that is typically realised as either a schwa or a high front /i/, with this realisation determined by phonological context.
    So this is moot (in the American definition ;)). The 'vowel that is typically realised as either a schwa or a high front /i/' is just an /i/ (which is perhaps centralized somehow in Damascus, but is still morphologically an i), and entertaining the possibility that it's some special schwa vowel just obscures what's actually going on.

    It's true that this explanation is in itself incomplete because this vowel appears in all sorts of different unstressed contexts as well, but I'm not sure that's directly relevant to the argument - unless I'm misunderstanding you.
    And then yup, my argument both depends on and explains the merger's appearance in these yif3ul verbs' unstressed syllables. Like I was saying above, the elision from yiktub + -u = *yiktVbu to yiktbu is literally all we need to prove that that unknown underlying vowel is an /i/, because there's no precedent for anything but an /i/ getting deleted in an unstressed open syllable such as that one. The schwa is a red herring, and discussing stressed u/i separately from unstressed u/i is also a red herring.

    All this 'obliteration' shows us to my mind is that u/i merge when stressed.
    (Mild tangent, but the Cowell screenshot this is in response to doesn't even discuss u/i when stressed. He only shows the "obliteration" in terms of unstressed syllables.)
     

    LiliaGaripovaRadikovna

    Senior Member
    Russian, Tatar
    *(There is the interesting exception 7asiba - ya7sib/ya7sab. The variants seem to have been resulted from ancient dialectal differences (cf. noun variants like udhun/udhn, also supposed to be a result of ancient dialects).)
    May i specify, yahsib- it's past? And yahsab - present?
     

    wriight

    Senior Member
    English (US) / Arabic (Lebanon)
    Nope, they're both present! 7asiba ("hasiba") is what the past form is. ya7sib/ya7sab ("yahsib/yahsab") are just ancient dialectal variants of المضارع for this verb.

    BTW, some belated corrections to my last post:
    1. The Damascene schwa is definitely nothing other than a reflex of */i/, so my hedging with "overwhelmingly" wasn't necessary. The example of ممكن that I thought was confounding is actually normal, because you can actually hear mimken for this word in a whole lot of other places.
    2. The reason ləzzēʔa has the long vowel ē rather than ā is because it actually turns out to be an *ay diphthong, not an alif. See this thread.
    3. Typos: "the distinct treatment" -> "the distinct treatment of", "other three" -> "other four".
    A neat conclusion that that also helps us derive (especially when looking at interesting, half-anomalous Lebanese forms like yista3mul "he use" or yiz3uj "he annoy" or kitob "books") is that the outcome of the North Levantine /i u/ merger is really determined by positioning: it tends, or at least has tended, towards /u/ in final syllables and /i/ elsewhere. Probably explains why the suffixes -(h)un, -kun have remained unshifted to -(h)in, -kin, for one.
     
    Last edited:

    Wadi Hanifa

    Senior Member
    Arabic
    I think these so-called pairs arise for two reasons:

    1) a 'native' form of intransitive verbs such as ghilii co-existing with an innovative form ghala (the latter arising by analogy with transitive forms of the pattern fa3al)

    2) verbs that come with both transitive and intransitive forms (e.g. ghili el-dahab vs [fulan] ghala el-mayy)

    I'm no longer able to concentrate on lengthy linguistic discussions like this one so I can't contribute much except to say that in Arabian dialects (and probably traditional forms of non-Arabian dialects as well), the distinction between transitive and intransitive forms has survived, with two vernacular forms mapping onto three fuSHa forms:

    Intransitive:

    Classical: fa3ila and fa3ula > Vernacular: fa3il

    Transitive:

    Classical: fa3ala > Vernacular: fa3al (or f3al) (vowel situation gets complicated once you begin to attach the pronouns)

    One difference between conservative Bedouin-type dialects and 'innovative' urban ones like Urban Hijazi is that the latter merge the fa3al and af3al patterns, so:

    Najdi: arsalt al-risalah <> Urban Hijazi: rasalt al-risalah

    This has a knock-on effect on all the derived patterns after that (so a sender is mirsil in Najdi but raasil in Urban Hijazi and so on).
     
    Last edited:
    < Previous | Next >
    Top