In Modern Standard Arabic, phonemic /lˤ/ (a.k.a. "emphatic l") only occurs in one native word: Allah /ʔalˤˈlˤaːh/. (According to the linked article, it also occurs in a few loanwords.)

This seems like a profoundly weird situation. If a phoneme exists at all, you'd expect it to show up more than once. Imagine a time before the loanwords with /lˤ/ were borrowed. Would allah then have been the only instance of /lˤ/ in the entire lexicon of the language? I know that's logically possible, but intuitively it strikes me as completely bizarre.

So what happened? How did this weird situation come about? (Or is it less of a weird situation than I think? I'd be happy to hear about other languages that have similar "one-off" phonemes.)

  • 3
    Disclaimer: I am no linguist and have no scientific evidence of what I am about to say. This might give you a lead to find the answer you seek. I was partially taught reading Quran in some of the 7 readings considered standard. I was told that these readings correspond to old "standard" dialects of Arabic that existed at the time. In one branch of one reading, the phoneme you describe occurs often. More specifically, it occurs when an "l" follows one of the "hard" letters that are pronounced from the back of the palate, like the strong [s] in "Salah" (prayer).
    – aelguindy
    Commented Jun 21, 2012 at 18:30
  • 1
    There are exceptions where it is just pronounced as a normal [l]. The branch allegedly corresponds to the dialect of the area of Medina (carried by an Egyptian reader). Very few places today use this as the standard reading, including AFAIK areas in Northern Sudan and Morocco, but not sure of that either.
    – aelguindy
    Commented Jun 21, 2012 at 18:31
  • 1
    – Houman
    Commented Jul 21, 2018 at 20:09
  • @aelguindy, would you be willing to support an Arabic SE proposal? if so please join this chat room: chat.stackexchange.com/rooms/18664/arabic-stackoverflow-- Commented Feb 11, 2020 at 12:58

3 Answers 3


This is not unheard of in onomatopoeic and other "expressive" parts of the lexicon. English for example has "unh-uh" /ʌ̃ʔʌ/ (meaning approximately "no"), which has a nasalized vowel and a medial glottal stop, otherwise not present in the language (as underlying segments). Allah isn't quite in this category, but (given a cultural background of monotheism etc.) it might be "special" enough to sustain an exceptional phoneme.

(Of course, this is just one possible explanation; I don't know anything about the actual history of the word.)

  • 1
    I added the "wedge-with-a-tilde" you were looking for ;-) Commented Nov 16, 2011 at 22:47
  • perhaps this explains how they are sustained, but do you know how did they occur in the first place?
    – Louis Rhys
    Commented Nov 17, 2011 at 10:11
  • 3
    There is something tempting about the idea that it's phonologically special just because it's theologically special. But does that occur in any other language? Special phonemes that are reserved for culturally important words? Commented Nov 17, 2011 at 18:04
  • 2
    @DanVelleman, the other example that springs to mind is Damin, a ceremonial language/dialect of Australia. It is the only click language outside of Australia (although it is thought to have arisen by conscious invention as opposed to spontaneously).
    – Aaron
    Commented Nov 29, 2011 at 22:08
  • 2
    @Aaron: "Outside of Africa", I'm assuming?
    – jogloran
    Commented Feb 29, 2012 at 3:41

One could hypothesize that the pharyngeality is due to the effect of the two glottal consonants /ʔ/ and /h/ at the beginning and end of the word. This probably isn’t the case however, since this hypothesis would imply that there should be other words with pharyngealization of the middle consonant due to glottal consonants appearing at the beginning and end of the word. I doubt that they exist, given that you say /lˤ/ only occurs in /ʔalˤˈlˤaːh/.

There is an article about the pharyngealized voiced lateral approximant “emphatic l” in Arabic:

  • Ferguson, Charles A. 1956. The emphatic l in Arabic. Language 32.3: 446–452.

Another article discusses the spread of pharyngealization in the context of Grounded Phonology.

  • Davis, Stuart. 1995. Emphasis spread in Arabic and Grounded Phonology. Linguistic Inquiry 26.3: 465–498.

These might at least give you some references to dig further.

  • 1
    The Ferguson paper has a really nice statement of the problem. It doesn't seem to offer a solution, but it's still good to know I'm not completely off-base in being confused. Commented Nov 17, 2011 at 17:54
  • 3
    Also, since you mentioned spreading: There are (in many modern dialects) words where [lˤ] shows up as a predictable allophone of /l/, conditioned by an emphatic consonant elsewhere in the word. The puzzling thing is that "Allah" doesn't have any other emphatic consonants in it. /ʔ/ and /h/ don't trigger this sort of predictable emphasis spreading. So in this one specific case, you have to treat the /lˤ/ as phonemic. Commented Nov 17, 2011 at 17:57
  • Note that the glottal stop is normally only pronounced when it's at the beginning of the sentence, otherwise it would be absorbed by the final vowel of the preceding word, like "Abdullah", and the l will still be emphatic in those cases.
    – Louis Rhys
    Commented Nov 18, 2011 at 18:13
  • @LouisRhys: You should know that "Abdullah" is in fact derived from "Allah" - it means "servant of Allah". At least in the local pronunciation where I live (Israel), the last part is pronounced as is "Allah".
    – dotancohen
    Commented May 2, 2012 at 21:12

I'm not an expert in the classical Arabic language nor in the "Tajweed" science but I can contribute with the little bit I know ... my clarification can be summarized in some points:

First of all, we have to know that the word "Allah" is a "classical" Arabic word (very ancient) consisting of the word "ilah" (god) linked to the definite article "al" to say "The God" (The One [not any one]) : that's a proper noun (so, without plural nor feminine) and it's not the equivalent of the English word "god" or "dieu" in French ...

That's an exceptional word that has an eccentric spelling - did you even notice? - and, an eccentric pronunciation (and that's the subject of your question). If we know that the "L" (ل) [lām] in almost all Arabic recitations of the "Qur'an" and, before that, in almost all dialects is not an emphatic consonant (it's rather pronounced with some thinness), how this happened with the word "Allah" ?

There are two reasons for the "tafkhim" (thickening) of the "lām" in the word "Allah" :

  • the first is to avoid the confusion with the name of an historical
    pagan divinity called "allat" - which is linguistically a word from the same roots, ends with "t" instead of an "h" and having a thin "L" .
  • the second is the Glorification of the name of The God (yes, almost all Arabian tribes were not monotheistic but they knew that "Allah" is "The authentic God")


I have to mention that the "lām" in the word "Allah" is not always thickened, that's to say there are cases making the "lām" thin if the word "Allah" comes after a "Kasra" (almost after an [i]), as in [ لِلَّهِ مَا فِي السَّمَاوَاتِ وَمَا فِي الأرْضِ] or in [بسم الله], and I have also to mention that in the method of recitation of "Warch" (a famous reader of the Qur'an), the emphatic "L" , thus the famous /lˤ/ , happens in other cases (known as the "L"s of warch : لامات وَرْشٌ), like in: يصلى,فصل ,ويصّالحا,فلما فصل طالوت بالجنود ,صلى,طال ...


Finally, it's worth saying that language it's not a Darwinian realm, it's rather a complex confluence of historical and geographical events added to some rare "mutations": some phonemes or letters can persist without being even used (the German alphabet contains the letter x but it is used only for borrowed words ...)!

(Sorry for the links, they are in Arabic, I will try to find English links if there are any ones.)

  • Is there a source for the claim of the hard laam being used to differentiate from Al-Lat? This seems significant, as it alludes to a deliberate conception of the word Allah rather than it evolving over time as most language does. Commented Sep 6, 2019 at 22:34
  • @jihed gasmi, would you be willing to support an Arabic SE proposal? if so please join this chat room: chat.stackexchange.com/rooms/18664/arabic-stackoverflow-- Commented Feb 11, 2020 at 12:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.