In the course of reading The Last Jews in Baghdad (see this LH post), I encountered references to “the Arabic spoken by Jews” and wondered what that was all about: was there really a separate dialect, not just regular Iraqi Arabic with a few Hebrew words tossed in? A little googling turned up a thread at WordReference Forums on this very topic, started by Nun-Translator’s question:

I am reading an autobiographical novel written in Hebrew by Eli Amir that takes place in Baghdad in the 1940s. He frequently refers to “Jewish Arabic” and “Muslim Arabic”. I’m not clear if he is talking about accents or dialects… Are you aware of “Jewish Arabic”? Is it an accent or a dialect? (The way Amir uses the term, it sounds more like a dialect.) Does it still exist? And one more question: Is there a distinctive “Christian Arabic” in Iraq or elsewhere?

The answer turns out to be that there are two major dialects in Iraq, a northern and a southern, conventionally distinguished by their pronunciation of the letter qaf as /q/ (hence “qeltu” for قلت) and /g/ (“gilit”) respectively, and the distinction accidentally became a religious one in Baghdad, as explained by a commenter who goes by the handle smooha:

This happens to be a subject of great interest to me. I had the privilege of having read several books on the subject, most notably Haim Blanc’s Communal Dialects of Baghdad.

As clevermizo noted, the Mesopotamian (Iraq and eastern Syria) varieties of Arabic can be divided into qeltu and gilit types. The gilit type is overwhelmingly of Bedouin origin, from the tribes to the south and west of the rivers. If my memory serves me correctly, virtually all Baghdadis (Muslims, Christians, and Jews) spoke in a qeltu variety, from the Abassid era until the late 19th c (?), which saw an unprecedented process of urbanization, with a vast number of Bedouin tribesmen (previously nomadic) became the majority. The original urban Baghdadi Muslims assimilated with the new majority, while the Christian and Jewish communities maintained their respective dialects (which, though to a significant extent mutually intelligible, contained numerous differences in lexicon).

The dialects were maintained as the language of the home/family, while the Muslim variety was used as a sort of lingua franca in the public sphere. As far as I know, Christian Baghdadis still speak their dialect in their homes, though it seems to me that Christians have tended to assimilate a bit more, because of political and social considerations (anti-Christian sentiment as well as inter-marriage between Christians of different regions)…

An interesting note: the Jewish dialect was 100% mutually intelligible and virtually identical in every non-Kurdish town in Iraq (i.e. a Jew from the northern city of Mosul could communicate fluently with a Jew from Basra, Baghdad, etc.).

It is quite easy to find samples of Muslim Baghdadi speech. If you’re interested in hearing the difference, go to this site.

It’s a comprehensive resource for samples of the (mostly minority) dialects of Arabic… Also, be aware that since most of the Jewish speakers now live in Israel, there are some Hebrew words here and there that are not a part of the original Jewish Baghdadi dialect.

A search for more information on the qaf divide led me to A War of Words: Language and Conflict in the Middle East by Yasir Suleiman, which discusses it on pp. 99-100:

The function of (Q) as a group boundary-setter is exhibited in various speech communities in the modern period. This is attested in a large number of studies… Blanc (1964) reports that the communal dialects of Baghdad are marked by the prestigious variant [g] in the speech of the Muslims and by the stigmatized [q] in that of the Christians and Jews… This pattern of communally based linguistic variation is found in Bahrain. Holes (1983) reports that the prestigious variant [g] is a marker of the speech of the Sunnis in Bahrain, while the stigmatized variant [q] characterizes the speech of the Shi’ites. In Tunis, the prestigious variant [q] is a marker of urban speech, while the stigmatized variant [g] is a marker of the rural or semi-nomadic communities… These, and the other studies listed above, generate four observations:

(1) The variable (Q) serves as a marker of group membership. These groups may be communally defined, as in Baghdad and Bahrain, or ecologically designated, as in Tunisia and in Ibn Khaldun’s study.

(2) The prestige or stigma associated with a variable is contextually defined. Thus while [g] is the prestige variant in Baghdad and Bahrain, it is not in Tunis.

(3) The prestige of a particular variant may change in time. Ibn Khaldun’s observation about the variant [g] exemplifies this phenomenon.

(4) Whenever dialect shift takes place it tends to move in the direction of the prestigious variant in a speech community.

I was going to discuss a recent Jabal al-Lughat thread that involved Iraqi Arabic, but this is quite long enough already, so I’ll save it for its own post!


  1. Charles Perry says

    Blanc found that both Christian and Jewish Baghdadis spoke a qeltu dialect, but the Christian dialect resembled northern Iraqi dialect, showing that the Christians had moved to Baghdad from the north when the city’s population began to recover. The Jewish dialect was not related to any other Arabic dialect, so Blanc concluded that the Jews were the only people who continued to live in Baghdad after its destruction by the Mongols.
    By the way, I believe the title of Blanc’s fascinating book is “Communal Dialects in Baghdad,” not “of Baghdad.”

  2. If I recall correctly, the Christian Baghdadi dialect makes its r’s uvular, whereas the Jewish one doesn’t.

  3. If I recall correctly, the Christian Baghdadi dialect makes its r’s uvular, whereas the Jewish one doesn’t.
    Actually, it’s the other way around. In Baghdadi Judeo-Arabic, r > ġ, i.e. M(uslim)C(hristian) bir, but J biġ “well”, MC arbaʿ, but J aġbaʿ “four” etc.
    I believe I’ve already mentione hered Jacob Mansour’s The Jewish Baghdadi Dialect which is a more detailed description of Baghdadi Judeo-Arabic with a lot of texts recorded by native informants.
    And one should also mention the rich literature written in Baghdadi Judeo-Arabic which includes not only translations of portions of the Bible, targum and various aggadic materials, but also original compositions, mostly legends and alike. For the past few months I’ve been working on a fully-vocalized translation (šarḥ)of targum to Song of Songs, mostly trying to figure out the relation between the language of the šarḥ (which has a lot of strange features, mostly grammaticalizations of earlier hypo- and hypercorrections like the negative marker ליס < Classical Arabic ليس) and the actual spoken language as recorded by Blanc and Mansour. Fascinating stuff.

  4. I remember reading (in Blau?) that Judeo-Arabic literature was legible throughout a larger area in part because the vowelless script disguised dialect features. Of course, English clearly demonstrates that it also works to write the vowels in a way that no one has pronounced them for centuries.

  5. MMcM,
    that certainly applies to what Hary calls “Classical Judeo-Arabic”. But starting with the 15th century, the dialect differences really started to show, especially in the off-center areas like North Africa. There might have been purely linguistic reasons for that, but most agree that it was also motivated by a change in the social structure (the rise of the ghetto). Translations and original compositions written in the Maghribi Judeo-Arabic, whether vocalized or not, would have been largely unintelligible to speakers of Judeo-Arabic in Egypt and Iraq and vice versa. To what extent precisely, that’s difficult to tell, because even though most of those who wrote in Judeo-Arabic no longer tried to emulate Classical Arabic, a large number of them still tried to emulate Classical Judeo-Arabic which in turn emulated Classical Arabic. So in any late (post 15th century) Judeo-Arabic, you are bound to find a mixture of pure colloquial, grammaticalized hypo- and hypercorrections, fusha and what was perceived as standard Judeo-Arabic. And in a high number of cases, some of these categories overlap. No wonder Hary chose to describe it as multiglossia…

  6. Peter Austin says

    Over the past year Eli Timan (a speaker of Iraqi Judeo-Arabic who lives in London) and I have been engaged in a project based at SOAS and funded by the British Academy to document personal history narratives in Iraqi Judeo-Arabic with members of the diaspora community in London, Canada and Israel. Eli has collected hours of wonderful text material and has been transcribing and translating it into English. We plan to work with David Nathan of the Endangered Languages Archive to publish some samples on the Hans Rausing Endangered Languages project website ( We may also develop a small multimedia product presenting the recordings, transcriptions and photographs that Eli has taken.

  7. caffeind says

    You know, “qeltu” sounds kinda like “ghetto”…

  8. David Marjanović says

    A uvular r? That needs to be mentioned in Wikipedia! Trill or fricative?

  9. marie-lucie says

    (David, I wrote back to you about French and German uvulars in The Sound of Home, just before it was closed).

  10. David Marjanović says

    Concerning the uvulars, do you have Skype? We both seem to be underestimating the variation within German, and only acoustic evidence can help at this point.
    In lard rôti, I agree that most speakers use a sound here that is mostly or entirely fricative — and at least partially devoiced. That’s because it’s word-final. The same speakers (those that I’ve heard) use a one-contact trill intervocalically.

Speak Your Mind