Quraysh and confederacy

The name of Muhammad’s tribe, as reported in the Arabic sources, is Quraysh. It’s a word of no certain origin, and the medieval scholars offer a number of speculative and conflicting etymologies, none of which is very plausible. That’s normal in discussions of language: true etymologies quickly recede from the popular memory and are replaced by ‘folk’ etymologies.

Two recent books have proposed a derivation from Syriac, attributing the sense of foederati ‘confederation’. In this conception, the Quraysh were not a tribe so much as a coalition of Arab parties under Roman patronage, situated near the imperial frontier in Syria.

These books are The Syro-Aramaic Reading of the Koran by the pseudonymous Christoph Luxenberg and In the Shadow of the Sword by Tom Holland. (Oddly, Holland doesn’t cite Luxenberg on this point, although their argument is the same.) Neither holds an academic chair, and this particular argument of theirs has not been accepted by specialists; but those aren’t reasons to rashly dismiss it.


The hypothesis is raised…

Luxenberg’s argument (p.237) is as follows. The Arabic root qarasha is (he says) taken from the Syriac qrash. In both languages, it has the basic sense of ‘gather together’. In Syriac, the passive particle is qarīsh(ē), ‘gathered together’, which was taken into Arabic as Qarīsh, with the sense of foederati. In Arabic script, Qarīsh and Quraysh are indistinguishable, and so the mistaken reading of Quraysh was adopted.

Holland, writing for a more general audience, is more succinct: “Qarisha, in the language that had increasingly come to be the Fertile Crescent’s readiest lingua franca, meant ‘collected together’ – ‘confederated’.” (Kindle ed., p.333 ff.) He doesn’t offer an explanation for how Qarīsh(ē) came to be pronounced Quraysh; perhaps he thought it trivial (I would).


…and comes qrashing down

I was surprised, on checking the Syriac lexicons – including those cited in the two works at hand – that the word qarīsh(ē) is nowhere to be found. The meaning ‘confederated’ is not given; not even the purported literal meaning ‘gathered together’. It’s a hypothetical creation of Luxenberg’s and Holland’s, who merely extrapolate from another word, which is attested: qrash, meaning ‘gather’.

(We should beware not to confuse qrash ‘gather’ with qrash ‘be cold, freeze’ and its derivative, ‘congeal, clot’. The two are homonyms, but unrelated. If you conflate the two, you risk looking foolish.)

The word qrash is very restricted in its use. On the one hand, it has no clear derivatives – the only plausible one is qershā ‘trash heap’ – and on the other hand, it’s apparently very rare.

The tenth-century lexicons of Bar ‘Alī and Bar Bahlūl don’t record qrash ‘gather’. The modern lexicons overlooked it at first: Payne-Smith’s Lexicon and Thesaurus and the first edition of the Lexicon by Brockelmann don’t have qrash with this meaning.

It was later added to Brockelmann’s second edition and included in Margoliouth’s Supplement to Payne-Smith. These volumes, along with their successor Sokoloff, are good enough to name their source for this rare word. They all point to the same passage in the Zuqnin Chronicle, written in Iraq in the late eighth century. The passage is about a certain Arab prince who consorts with magicians and sooth-sayers: “they collected stupid and senseless words which they offered him”.

I’m not an experienced Syriacist, so I can’t tell whether the word is a true hapax legomenon, appearing only once in the history of Syriac. I can only deduce that it’s extremely rare. Moreover, the word appears after the Arab conquest; written in the heart of the Arab empire; in a story about the Arab elite. No surprise, then, that Brockelmann and Sokoloff trace qrash to the Arabic qarasha.

Luxenberg, blasé as ever, calls this ‘mistaken’ but doesn’t bother to show why. I think we’ve no choice but to accept that the Syriac word qrash was highly marginal, and perhaps a foreign import.


But was it Arabic?

What’s really strange is that the Arabic root qarasha is also rare. Lane-Poole, the cataloguer of early lexicons, places it in his Supplement, along with many other short entries for peripheral roots. He cites only one source, the Muḥkam by Ibn Sīda (d. 1066), and renders the forms qarasha, iqtarasha and taqarrasha collectively as “He gained, acquired, earned, and collected”.

Likewise, when I run these permutations through the databases of medieval Arabic works, there are depressingly few results. Although I’m happy to invite counterexamples, I think they must have been quite rare in early-medieval Arabic. My feeling is that qarasha is an archaic root, kept alive by the medieval scholars’ relentless antiquarianism; so that its hefty entry in the thirteen-century Lisān al-‘Arab bespeaks a collective effort at conservation rather than a record of living practice.

The trouble is that Arabic (unlike Syriac) is poorly attested before the Conquests. It’s easy to imagine why: the Arabs were a very deprived people, and writing absorbs huge resources of labour and materials. (We sometimes forget that our own, very literate, civilisation is fantastically wealthy in comparison.) If qarasha and its derivatives were used, we’ve no way of proving it: not a trace remains.

We might be tempted to apply an argument from silence, and insist that, since the Arabic root is unattested, it probably didn’t exist; whereas the Syriac is attested, however weakly, and therefore the Arabic was borrowed from the Syriac! Rather, we would expect the root not to appear in early Arabic, because the pool of evidence is tiny; whereas we would expect the root to appear in early Syriac, because its canon is relatively large.

On balance, its near-absence from both Syriac and Arabic condemns the Syriac hypothesis alone.


Occam’s razor and Morris’ chainsaw

In Arabic, the root qarasha is most readily found in discussions about the etymology of Quraysh. For instance, the tenth-century historian al-Ṭabarī explains that the Quraysh were so called because they were gathered together by the patriarch Quṣayy, and “to come together (al-tajammu‘) is to gather together (al-taqarrush)”. Luxenberg agrees with the story, but insists on a Syriac origin; Holland doesn’t address the issue.

The obvious conclusion is that, if the word Quraysh had the sense ‘gathered together’, then it’s much easier to believe that this took place within native Arabic than in collaboration with Syriac. Whatever job the Syriac passive can do, an Arabic passive (albeit much rarer) can do just as well:


It’s not inconceivable. But I can’t believe it.

Certainly, Ṭabarī’s narrative requires fewer assumptions than Luxenberg’s and Holland’s. There needn’t have been contact between Arabic and Syriac in order to form this word, the Syriac origin of which is dubious anyway. But it’s still so much speculation. I don’t mean to say that it definitely isn’t true; merely that the evidence isn’t strong enough to support it, no matter what route the etymology took.



‘Quraysh’ is not, as far as we can tell, derived from Syriac; nor does it mean ‘confederated’.

The medieval linguists, trying to understand the ‘age of ignorance’, used to fashion (somewhat) plausible etymologies based on the flimsiest evidence. Our homage to them is that we’re still at it today.

Comments and friendly criticism are welcome.

Postscript: Etymon! (gotta catch ’em all)

Ian D. Morris

Ian D. Morris

Historian and occasional human, powered by tea and cat videos.


  1. Maybe Quraish is the diminutive form of Qarīsh like Hassan and Hussein, Hirra Hureyra, Nasser and Nusayr.

    • Yes, it does look like a diminutive form. Unfortunately, without an attested usage of qarīsh, there’s not much we can do with that.

  2. Hi! I don’t know if this is the right place to ask this question, so I apologize if I did wrong.

    It’s about the Arabic language and its dialects and their usage of the relative pronouns. As we know, in Classical Arabic and Modern Standard Arabic alladhi gets declined (alladhina etc.), as for the dialects (actually all of them) ,that’s not the case.

    That’s the same for the other Semitic languages, except old Akkadian. If Classical Arabic is really the common origin of all Arabic dialects, why they do not decline alladhi?

    Also it seems that their relative pronoun sounds very differently, thus we can’t really reconstruct back alladhi at all!

    As far as I know, there’s only one pre-islamic Thamudic inscription mentioning the female allati,, and also both alladhi/alladhina are found e.g. in the Qurra papyri of the early 8th century.

    This German linguist, however, thinks that this relative pronoun rather was “invented”:

    He thinks the allati in the Th. inscription developed independently from the actual R.P., but he does not say anything about the Qurra papyri. He also says the Arabic script orginated somewhere in Mesopotamia, though as we all know its derived from Nabatean script!

    This guy indeed says a lot of false things, but the case with the relative pronoun is interesting. This one is I guess become quite long, so I apologize.

    Do you, however, know some academics who could deal with this and recommend it to them?

      • My friend, the historical linguist Alexander Magidow of the University of Rhode Island, responded on Facebook – and I quote –

        “I think it’s actually al-la-dhii where the first element is a definite article, basically “agreeing” with the definite antecedent a la Naamas theory [Naama Pat-El], the la is totally obscure and opaque and the dhii is a normal demonstrative, so tii is expected also.

        “The plural demonstrative forms, though, I do think weren’t generally conjugated for case, and I think the dialect diversity may have been re-interpreted as case conjugation. There’s a footnote in my dissertation, search for al-farra I think.

        “Dhuun and dhiin are probably legit dems in some dialects (Maltese I think).”

  3. Ancient language geeks are always fascinating, as is this discussion (and various others in Islamic literature). I wonder though if there’s a good reason the author and all commentators ignore the various uses of the stem “qarasha” as opposed to the forumation “taqarasha” (which incidentally doesn’t only mean “to gather”)… since Arabic is (mostly) a living language: clear advantage in “seeing” or listening to how native speakers use… and numerous dictionaries… (though clearly it complicates “laboratory-like” isolation of ancient languages).
    The Lisan al-Arab, gaves numerous meanings, uses, and the same two possibilities of the origin of the name as Ibn-Hisham (from the verb to gather and shark: “marine animal which eats all others and of which other animals area afraid”; curously there’s some connecton between the two: the shark also gathers –to eat! similarly the root q-r-sh can also mean to grind and to scatter; it’s a long stretch if one lets imagination run wild, the shark does all these–gathers, grinds and scatters! ).
    Is there a good reason why we’re ignoring all the other diverse possibilities, many perhaps more remote but rooted in continuous usage?

Leave a Reply

Your email address will not be published. Required fields are marked *