More on "Daesh"

« previous post | next post »

We've had a recent post on the pronunciation of this lightning rod of a word.

"Pronouncing 'Daesh' " (11/15/15)

From a colleague:

Guthrie's article* states:

"And the vowel which begins the word 'islaamiyya' becomes an 'a' sound when differently positioned in a word, hence the acronym being pronounced 'da'ish' when written in Arabic, and  the 'a' coming over into our transliteration of the acronym."

I don't know Arabic, but am guessing this means:

1) the first letter of " islaamiyya' " is < إ >. That's an alef with a hamza under it. Alef on its own would usually be /a/, but alef with a hamza under it is /i/;

2) but in the acronym, the bare alef is used, so it's pronounced /a/;

3) If the under-hamza had been retained, the acronym would be 'di'ish' (that is, / d iː ʕ i ʃ /, rather than  / d aː ʕ i ʃ  / ).

If that is correct, then a question is: did the hamza have to be dropped by convention? (But if acronyms are so rare, do conventions exist?) Or, was dropping it (and altering the pronunciation of the letter) a means of cutting connection to the word 'islaamiyya'. If the latter you'd think she'd point it out.

With the vowel diacritics now implanted (from this Arabic Wikipedia article), here is how it all sounds in Arabic according to Google Translate Audio (click on the speaker icon). First the whole phrase, and then the repeated acronym.

Unrelated to the above, I am under the impression that the Arabic phoneme / i / moves toward an allophonic [ ɛ ] after [ ʕ ], even if the [ ʕ ] is barely (or not at all?) pronounced. Plus which, in localised pronunciations standard  / i / often may be realised as / ɛ / . So, that is why media is full of the spellings of the acronym with <e> rather than <i> .

Someone who knows Arabic will please comment on the above.


* Alice Guthrie "Decoding Daesh: Why is the new name for ISIS so hard to understand?", Free Word 2/19/2015


  1. David Marjanović said,

    December 1, 2015 @ 5:58 pm

    I don't know Arabic either, but…

    1) Alef with hamza is simply /ʔ/. The first vowel of /ʔislaːmijːa/ is not written.
    2) /ʕ/ is commonly an approximant rather than a fricative, so it's more difficult to hear than I used to expect. It's also epiglottal ([ʢ]) rather than pharyngeal in many regions; that means it has, in those places, the opposite of the expected effect on nearby vowels.
    3) French takes its renderings of Arabic from Maghrebi pronunciations, where the three short vowels have merged into a single one which is transcribed as e. Example: Oussama ben Laden.

  2. Ryan W said,

    December 1, 2015 @ 6:41 pm

    In Arabic, most words have a sequence of three consonants that form the semantic root (2 or 4 happen, but are relatively rare), and by using different vowel patterns and adding additional consonants. In the word islaamiyya, these roots are s, l, and m.

    Hamza (the glottal stop) can be written above or below an alef, on its own, or on a waaw or a yaa', depending on what vowels come before and after; in any case, the hamza would normally be considered the important part.

    What I think happens in "Daesh" is that they're conforming to a very common derived form: CaaCiC, which is the active participle of any Form I three-consonant verb. By using just the alef from Islamiyya, the word is easy to pronounce and fits this noun form.

    A lot of Arabic speakers replace a short /a/ with the /e/ phoneme in the colloquial, but here I think it's just to show that there is a new syllable with an unstressed vowel…just a bit easier than writing Daïsh or Da'esh, and less ambiguous than Daish.

  3. Gareth Hughes said,

    December 1, 2015 @ 7:05 pm

    Let's try to explain this. Hamza isn't a letter but a sign of articulation, it represents a glottal stop. The first letter of 'islamiyya' is an alif. So when you take the first letters of the words, you take an alif, and you don't carry the hamza.

    An Arabic word that begins with a vowel sound is written with an alif and a hamza. The tradition is not to write the vowel sign except in the Qur'an, difficult words and children's books. A hamza with an alif is usually written above the letter, but is moved below if the following vowel is an 'i'. That low hamza isn't a vowel sign, just a hint at one. In the middle of a word, alif is usually the sign of a long 'a' vowel (as a mater lectionis rather than a vowel sign itself). It can carry a hamza or madda to represent a short or long, respectively, 'a' vowel following a glottal stop. A lowered hamza does not occur in medial or final positions.

    Thus, the first letter of 'islamiyya' is alif, and when that initial is used medially, it naturally represents a long 'a' vowel, as in 'Da'esh'. Whereas it could become a hamza followed by an 'a' vowel, this would be more unusual.

    That's a somewhat simplified explanation, but I hope it helps.

  4. Tim Friese said,

    December 1, 2015 @ 8:21 pm

    Arabic speaker here.

    There's some messy business with Arabic orthography here. It is true that alif usually represents /a:/ but initial alif doesn't just represent /i/, it actually represents any initial vowel plus glottal onset, so /ʔu/, /ʔi/, or /ʔa/. (This same prohibition against word-initial vowels can be found in Hebrew and Devanagari-scripted languages, among others.) To add to the confusion, sometimes the glottal is only pronounced utterance-initially and is dropped in liaison – this is called a hamzatu l-waSl 'connecting/liaison glottal stop'.

    So if you want to keep the original diacritics from /ʔisla:miy:ah/, you'd actually end up with the unpronounceable /dʔiʕish/ but if you discard the diacritics you end up with the morphologically-sound looking /da:ʕish/ as Ryan mentioned above.

    I don't see dropping the glottal as related to cutting the connection to the word 'Islamic', but I'd be interested in hearing the perspective of native speakers.

    Regarding the pronunciation of the final vowel, Standard Arabic /i/, this comes out with quite a bit of variety in the various Arabic dialects. Most tend towards [e] or even [ɛ]. My dialect – Damascus – tends to merge /i/ with /u/ and replace them with schwa everywhere except word-finally where they are realized [e] and [o] respectively.

  5. George Gibbard said,

    December 1, 2015 @ 8:33 pm

    >Alef with hamza is simply /ʔ/. The first vowel of /ʔislaːmijːa/ is not written.

    Except it sort of is, if one writes the hamza (glottal stop): hamza is written below rather than above ʔalif just in case there is a following /i/ (who knows why, but it will be relevant that /i/, unlike other vowels or sukuun — the absence of a vowel — is written below the line, in this case below the hamza under the alif).

    > If the under-hamza had been retained, the acronym would be 'di'ish' (that is, / d iː ʕ i ʃ /, rather than / d aː ʕ i ʃ / )

    If the under-hamza had been retained we would I guess have: /d(short vowel)ʔiʕ(short vowel or nothing)ʃ/. But really, word-internally, hamza before /i/ is not written with ʔalif at all but over a dotless yaːʔ (for some people this applies only after a vowel, as would be the case here since a Classical Arabic word can't begin with a consonant cluster; after a consonant, hamza should originally have no "seat"), so this would not really be an option. Other than word-initially, where alif is always used, the rules for writing hamza are basically: write as the word should be spelled in the vernacular, in which hamza has been lost, and then write the hamza back in. So سؤال /suʔaːl/ 'question' is so spelled (with waːw with a hamza over it) because people have usually pronounced it /suwaːl/ since before the invention of the hamza symbol. The orthography of hamza is the clearest evidence for a classical (hamza-ful) : vernacular (hamza-less) diglossia existing since before the time of Muhammad; apparently writing originally assumed the vernacular pronunciation.

  6. hannah said,

    December 1, 2015 @ 9:00 pm

    What? No, this has nothing to do with the alef at the beginning of islamiyya. It's the `ayn at the beginning of `Iraq.

    Arabic (like other Semitic languages) is based on a triliteral root system, and while hamza (the glottal stop on the alef at the beginning of islamiyya), it's not one of the root letters. Those are S-L-M – as in shalom from Hebrew or salam alaykum in Arabic. (Which really should be written as salam `alaykum, because the preposition `ala is used there as 'upon' or 'to,' and it begins with the letter `ayn – we'll get there in a minute.) The triliteral root is manipulated according to set patterns to get other words – nouns, verbs, adjectives, whatever – that are conceptually related to the core meaning of the triliteral root. You see a noun like Islam (S-L-M) turned into an active participle Muslim (S-L-M) via the same principle of manipulating root letters in set patterns.

    Now, for the actual acronym. Here it is in Arabic and my preferred form of transliteration: al Dawla al Islamiyya fi-al`Iraq w-al Sham. الدولة الإسلامية في العراق والشام

    It's D-A-`i-SH. The letter I've tried to represent here as `i is a weird half-vomit, half-choking noise that doesn't exist in English. I have no idea what the IPA code for it is, but it's sort of what you'd get if you said agggh in your throat without gargling. If you didn't know, the country that's south of Turkey and east of Syria is best pronounced uhh-rock, not eye-rack. There's no way we can say it correctly in English, so the first one's the closest approximation. (Iran uses a different set of letters AND a different language, and you're best saying it ee-ron. Fun!) This letter `i is often romanized as 3 in Arablish, because that is very similar to the letter's shape in Arabic. So da3ish makes more sense in many ways.

    Come to think of it, the same `I letter is the middle letter in the triliteral root for the word al Qaeda. That's not an acronym – it's a word in arabic that means base or foundation, and its triliteral root is Q-`I-D. Interestingly, the word for grammar in Arabic – qawa`id, meaning "foundations" – is from the same root, which makes sense since grammar is so foundational to the language. And since Arabic grammar is a terror to learn. :)

  7. hannah said,

    December 1, 2015 @ 9:10 pm

    Forgot to mention – the justification quoted in the OP is quite wrong. Hamza isn't considered in this acronym. Daesh isn't a triliterally-rooted word because it's an acronym. Arabic acronyms don't follow the same pattern as in English, because so many words start with letters that are part of a pattern – for example, almost every word that appears to us to begin with an M in Arabic would be "alphabatized" under the first letter of its triliteral root, not the M, because that would be impossible. When you learn Arabic, one of your early lessons is how to determine what the triliteral root of a word is, because you can't find it in a dictionary without it, and sometimes weaker root letters (like long vowels) can be ellided in a particular formulation of a new word.

    Another example of how Islam and its linguistic variants are used in Arabic acronyms is Hamas, often rendered HAMAS in English because it's an acronym. The actual name is harakat al muqawama al islamiyya, حركة المقاومة الاسلامية . The S in Hamas comes from the S in islamiyya, or S-L-M. For reasons I cannot explain, they use the actual M in muqawama – even though the root of that word is Q-W-M. I guess Hamas sounds better than Hawas.

  8. hannah said,

    December 1, 2015 @ 9:12 pm

    Haqas. I can't type tonight. :)

  9. Joe said,

    December 1, 2015 @ 9:46 pm

    I will try to address your first question – why does the alif in Daesh represent a vowel instead of a hamza (i.e. glottal stop)?

    To begin with, the letter alif actually serves 3 purposes:
    #1 – to "carry" a hamza – where the hamza is the underlying phoneme.
    #2 – to represent a 'long a' vowel – where the 'long a' is the underlying phoneme.
    #3 – to represent an epenthetic vowel before consonant clusters at the beginning of words (Modern Standard Arabic does not allow initial consonant clusters, so epenthetic short vowels are appended).

    The word islaamiyya begins with an alif. In this case, the alif is serving purpose #3 – to represent an an epenthetic 'short i' before the consonant cluster "sl". The alif does not represent an underlying hamza in any way.

    While the alif serves multiple purposes, modern Arabic speakers view alif as serving purpose #2 by default. That is, it represents a 'long a' vowel sound. Thus, in the acronym Daesh, the alif is interpreted as a 'long a' vowel rather than a glottal stop.

  10. Victor Mair said,

    December 1, 2015 @ 10:00 pm

    From a colleague:

    In Hannic, to the extent that I find it in news media, it's 「Dáyīshā 達伊沙」.

    That said, I have seen at least one instance of 「X āi X / X埃X」, where「āi 埃」, slavishly reflects the seen in most anglophone media (the latter having been slavishly adopted from French convention).

  11. George Gibbard said,

    December 1, 2015 @ 10:16 pm

    ʔalif in ʔislaːm isn't epenthetic (hamzatu l-waṣl): instead, measure IV of the verb has a prefix consonant: ʔaslama. Hence, ʔal-ʔislaːm has a hamza (which you do not get in in measures VII (ʔal-infiʕaːl), VIII (ʔal-iftiʕaːl), IX (ʔal-ifʕilaːl) or X (ʔal-istifʕaːl). It's confusing that the prefix hamza in measure IV is lost when there is a prefix (e.g. the active participle muslim), but it is always there in the maṣdar (verbal noun, the closest thing Arabic has to an infinitive, of which ʔal-ʔislaːm is an example).

  12. George Gibbard said,

    December 1, 2015 @ 10:55 pm

    Make that read "the prefix hamza in measure IV is lost when there is *another* prefix".

    Historically, what's going on is that the causative prefix (which was a sibilant in Proto-Semitic, i.e. *s or something similar) has turned into glottal stop at the beginning of an Arabic word (counting the definite article before the maṣdar as a separate word) but was lost with the following vowel when a vowel preceded it within a word. In Akkadian (Akkadian forms are usually cited with the root P-R-S), the 3rd person masculine singular of the preterite of the causative (cf. Arabic "lam yufʕil") is u-ša-pris (with the same vowel pattern as Arabic "lam yufaʕʕil" or "lam yufaːʕil").

  13. Leo said,

    December 1, 2015 @ 11:23 pm

    Good comments have been made about the hamza and the i vowel. I think that simply seeing the letters د ا ع ش in succession naturally produces the CaaCiC-type word (like kaatib, maalik, saakin, etc.) as Ryan W said, especially since the hamza is usually omitted from the alif. The pronunciation of written i as e, as has been noted, is very common.

    I was interested in why ISIS has been said to dislike this acronym, so I looked up some folk etymologies. Professor Mair, if you are frustrated by this process in Chinese, you would really have fun with what people come up with in Arabic. I'm not going to provide links to any of these sources, for obvious reasons. Here are some of the ways this word has been back-projected: None of the following reflects any sympathy for those involved in this discourse.

    – It is a conspiracy of Mossad, and actually stands for dawlah (د) al-yahūd (ا) ˁibādah (ع) šayṭān (ش), or "The State of the Jew Cult of Satan." Here one can see that the alif is not the first letter of the word, but taken from the definite article al-.

    – Erroneously attributed as originating with the name of a pre-Islamic family, Dāḥis, riven by infighting and greed in their war with Ġabrāˀ; those are actually names of horses who were raced by Hādhifah and Qays that resulted in a forty-year war between ˁAbs and Dhubyān.

    – Dāˁiš is a name chosen by Zionists, and God desires that sorcery should be turned back upon the sorcerer, so this name has now instead become a terror to the opponents.

    – Dāˁiš originates (inexplicably) from the English ISIS; Isis and Adonis (Ῑzīs wa Adūnīs) were the names of two gods worshipped by people in the Jāhiliyyah (period of "ignorance" prior to the rise of Islam), as well as in the Stone Age and even prior to the Stone Age.

    – It was originally used by the Syrian activist Khālid al-Ḥajj Ṣāliḥ on May 27, 2013, and carries the connotation of "denial and disdain." At that time, the organization was called the State of Islamic Iraq and Syria (dawlah al-ˁirāq wa al-šām al-islāmiyyah) and so the original acronym was Daˁšā (the alif of islāmiyyah again manifesting as the sound ā).

    – The meaning of the word dāˁiš in Arabic is "the pleasant wind that precedes nourishing rains." However, some places I see this add "we need to check the source of this definition." Go ahead, we're waiting…

  14. Victor Mair said,

    December 2, 2015 @ 12:16 am

    From an Arabist colleague:

    Some very unscientific notes here [VHM: in the OP].

    The "e" represents the 'ayn in Iraq. The "e" key is where 'ayn is on my transliteration keyboard for typing Arabic on a Mac (as opposed to the 'real' Arabic keyboard), for what that's worth.

    I presume that the "a" is simply an alif, which in the word Islamiyya is the first letter. In that word (Islamiyya) the alif has a glottal stop (hamza) and the i-vowel, but if you just used the unpointed alif in the acronym it would look like a long a.

    I think people are taking their name too seriously.

  15. Leo said,

    December 2, 2015 @ 1:20 am

    Here's a question: does dāˁiš / da'esh just sound bad? Why? I have a sense that what's offensive to ISIS about it may just be that it's an ugly sounding word in Arabic, though I can't find anything to support this very qualitative claim. I've often vaguely felt that some sounds just don't go well together, though it may just be selection bias or the power of suggestion. Could parameters for "naturally ugly" words in Arabic, or any language, be identified? One other example that comes to mind is مبغض mabġaḍ, 'repugnant,' which kind of just sounds like what it means. I know this sounds like a naive idea of a non-native speaker, but I'd be interested to hear other ideas. None of the sites I looked at for the above post, all in Arabic, mentioned this possibility, but it might just be taken for granted. A native speaker's perspective would be great.

  16. maidhc said,

    December 2, 2015 @ 3:52 am

    So whatever the derivation, we have settled on the name "Hamas" in English. That's fine, I will call them by that name.

    I really don't like the name "Isis" because it is used as a personal name and a business name in English, deriving from the Egyptian goddess. Let's call them something that doesn't mean anything in English. "Da'esh", "Da'ish" … I'm not going to be able to pronounce "a weird half-vomit, half-choking noise that doesn't exist in English". If the experts would decide on an approximation that is suitable for us barbaric plebes, I'd be grateful to adopt it.

  17. Yuval said,

    December 2, 2015 @ 6:33 am

    Does the U in POTUS lose its/y/ status from "United" because it's pronounced as a schwa in the acronym?

  18. hannah said,

    December 2, 2015 @ 8:45 am

    maidhc – Try ISIL. That's my preferred acronym, which is Islamic State of Iraq and the Levant, which is a better translation of Sham than just Syria.

  19. Victor Mair said,

    December 2, 2015 @ 8:46 am

    From a colleague:

    I hear this / i / = [ e ] movement also at in the pronunciations of ' عِيسَى ' (Jesus), which, as non-arabophone, I'd think 'should' be [ ʕiːsaː ]. I might guess from this that / ʕ / necessarily always pushes /i/ down toward [ e ] or even [ ɛ ].

    However, in Qur'anic recitation, I do hear it as [i], or close to it. For example —

    — which are recitations of three different persons of Qur'an 3:49, the third word of which is ' عِيسَى ' (Jesus).

    What is behind this phenomenon?

  20. Eli Nelson said,

    December 2, 2015 @ 8:59 pm

    @Victor Mair: There are special rules about pronunciation for Qur'anic recitation, called tajwīd. That probably explains why your colleague heard a difference between the pronunciations on Forvo and in the recitations.

  21. Neil Dolinger said,

    December 2, 2015 @ 10:39 pm

    "Here's a question: does dāˁiš / da'esh just sound bad? Why? I have a sense that what's offensive to ISIS about it may just be that it's an ugly sounding word in Arabic…."

    Yeah, maybe whenever they hear that name, it sounds "moist" to them.

  22. Tim Friese said,

    December 3, 2015 @ 4:00 pm

    @Victor: you're absolutely right that Qur'anic recitation (tajwiid) tends to pronounce the six vowels of Arabic as [a] [a:] [i] [i:] [u] [u:], and realizations of /i/>[e] and /u/>[o] are generally viewed as more 'dialect'.

    For what it's worth, your Arabist colleague's claim that the 'e' in 'daesh' comes from Iraq is just not accurate. That's a vowel added in to make it conform to an established Arabic form. The apostrophe in some English spellings corresponds to the first sound of Iraq, a pharyngeal approximant.

    I agree with others that too much time and energy is spent debating this organization's name. As I wrote in my blogpost on this issue, it's pretty standard to refer to most groups by the names they use for themselves, even when we strongly disagree with them, such as with the Shining Path, the FARC, or the Islamic Republic of Iran.

    Read more here:

  23. Boyd said,

    December 5, 2015 @ 10:46 pm

    I agree with Hannah, the apostrophe (and the following e) aren't a glottal stop, there an ayn (as has been mentioned, this is the first letter of Iraq in Arabic), both of which are unfortunately often written the same way when someone tries to transliterate into English. But as has also been said, the glottal stop isn't a letter, whereas ayn is one.

    Your article is the first time I had seen anyone consider the middle of Da'esh to be a hamza instead of an ayn. It never entered my mind that it could be the initial for Islamiyyah, because to me it's clearly there for Iraq.

RSS feed for comments on this post