[an error occurred while processing this directive]
1- Summary)

Jewish Iraqi, spoken by the Jews of Iraq, is disappearing fast as there are very few native speakers (less than 8) in Iraq. No estimate is available of surviving native speakers, but it is reckoned that there are less than 10,000. Most of the old speakers can be found in Israel and the English-speaking West. They tend to speak either Hebrew or English not only with the younger generations, who are not interested in learning Jewish Iraqi, but also amongst themselves.


The language has a core of Arabic with its own grammar, lexicon and phonetic rules, influenced by Aramaic, which the Jews of Iraq spoke from around 500 B.C. to around 800 A.D. It has many lexical borrowings from Hebrew which was used for religious purposes. It also has imports, albeit fewer, from Turkish, Persian and English.


There is no established orthography for the language. Most Jewish Iraqi writers and poets in the first half of the 20th century wrote in Standard Arabic. There are one or two plays and novels in Jewish Iraqi written in Arabic orthography. Apart from recent recordings made by myself, very few old recordings exist and those are of poor sound quality. Two books about Jewish Iraqi were written during the sixties and early seventies, but one cannot easily recreate accurately the sound of the language from the orthography used in them.


This project was conducted with funding from ELDP. Its purpose was to prepare the ground for a major follow up language documentation project with the aim of preserving the spoken Jewish Iraqi language (JI) by means of audio recordings with time-aligned transcriptions and an audio dictionary. It follows and builds on a previous pilot project funded by the British Academy


Some 43 hours of interviews were conducted in the UK and Israel between March 2008 and February 2009. Handwritten field notes on recording sessions and speakers (metadata) were taken.


As a bonus, the recordings contain rich historical material on the life of Iraqi Jews in the 20th century.


During the trip to Israel, contacts were renewed with Professor Shmuel Moreh, Emeritus Professor of Arabic Language and Literature at the University of Jerusalem and with Mr Ben-Porat the director of the Babylonian Jewry heritage Centre at Or-Yehuda in Israel who opened the library archive to me.





The Iraqi Jewish community have lived in Babylonia (present day Iraq) for over 2500 years. They spoke Aramaic for 1,200 years until Iraq was conquered by the Muslim Arabs. Soon after, Arabic would become the dominant language in the region. One hundred years later, the Iraqi Jewish community (excluding Kurdistan) abandoned the Aramaic in favour of the Arabic, but with borrowings from Hebrew, Aramaic and Farsi.

There are only 7 Jews left in Baghdad at the time of writing this report (December 2009). Most of the community are settled in Israel and the West and have adopted the language of their host countries. It is hard to find anyone under the age of 45 who speaks the language. The major aim behind the project is to preserve the spoken language for the community/communities' future generations. The outcome of this project will also benefit Semitics scholars and Iraqi dialectologists.



3- Interviews, audio recordings and material collection


The largest Jewish Iraqi community today is in Israel followed by the United Kingdom and North America. Interviews were conducted in London (UK) and in Israel mainly in the Tel-Aviv and Jerusalem area.


The age span of the speakers was between 45 and 93 years old. Priority was given to interviewing older speakers for obvious reasons.


Speakers came from a wide variety of social backgrounds. Speakers in the UK tended to be more educated and well off, but Israel offered the widest social and cultural distribution.  As expected there was constant language code switching to and from the host country language as well as to Muslim Baghdadi dialect which was the common dialect for communicating with people outside the JI community in Iraq. Frequent switching to classical- standard Arabic was observed mainly by speakers who were journalists, lawyers, writers or poets. An interview was conducted with a speaker who left Baghdad recently, in 2003 who had lived with her family with little interaction with the rest of the small Jewish community. She uses a lot of standard Arabic in her speech.  It is fascinating to observe the development of the language in 50 years. Although the accent, phonetic and grammar rules of JI were maintained (unconsciously) in her speech, a lot of the morphology was modern Baghdadi with a lot of standard Arabic expressions.


A wide variety of topics were covered, including life in the old oriental houses in Baghdad, schools, marriage, and historical events affecting the life of the Jewish community in Iraq. Also covered were the years just before the mass emigration/deportation from Iraq and the early difficult years in Israel when immigrants were living in tents and tin-roof shacks for up to seven years.

Interviews were also conducted with speakers who stayed in Iraq and lived during a very difficult period under the Baath regime between 1967 and 1971, including the wives and children of people who were hanged or killed in prison during this difficult period.


Finally, two interviews were conducted with speakers who left Baghdad in 2003. Both came to England with a passport just before the Allied invasion.


There were some 46 interviews conducted with 59 speakers and 43 hours of conversation. The youngest speaker was 46; the oldest was 93 years old. Only two speakers were born in Israel, the rest in Iraq.


Extensive field notes were hand written about the speakers and the recording sessions. These, we have called metadata. They will be digitised in a database for archiving purposes.


Digital images of photographs and paintings have also been recorded.



Sample audio recordings have been assembled and transcribed for this pilot project. They consist of a total of 1.5 hour recordings of 10 sessions with 5 speakers.

The recordings cover life in Baghdad in the 20th century, description of the old houses with illustrations, anti-Jewish laws, revolutions and escape from Iraq.

The software used for transcription is ELAN, developed by the Max Planck Institute for Psycholinguistics which is in the public domain. It allows for several tiers of annotations and for future additions of tiers. The annotations are aligned with the sound segment chosen, normally a sentence. In this pilot project, a tier was chosen for each of the following:

1-     Speaker's audio transcription (transliteration) in roman letters

2-     English (free) translation of the speaker's conversation

3-     Same as 1 and 2 above for other speakers including the interviewer

4-     Notes

5-     Digital images (if any) associated with the interview

For some recordings, transcriptions were done in Arabic to show the potential for future Arabic transcriptions which has the advantage of being the nearest in syntax and lexicography, containing the guttural letters, but suffers from the lack of suitable vowels to represent the JI accent.


More tiers can be added in the future, e.g. Hebrew translations to benefit the younger generation in Israel.


An important aim of this pilot and future projects is the ease of availability of the audio recordings with its transcriptions to the JI speaker community across the world for linguistic and historical research.


With this in mind, published media should be capable of being accessed by means of widely available computer software using English in the main, as this language is widely spoken by the Iraqi Jewish communities in the Diaspora. The internet will be the main method of publishing, though the original material in high quality image and sound will be kept physically at different Jewish Iraqi centres in the US, Canada, the UK and Israel.


A broad phonemic transcription has been used. Since the transcription would always accompany the audio, there is no need for an exact phonetic representation. Future researchers can add a tier for narrow phonetic transcription. Roman letters on a standard QWERTY keyboard have been used with Capital letters to represent long vowels and some specific JI phonemes. Other phonemes are represented by 2 letters using current popular conventions followed in everyday English transcription of Arabic, such as ‘kh’


This will hopefully ensure that the work is accessible to a wide audience including the speaker community and Arabic speakers worldwide. It also has the advantage of searching on phonemic rather than the more complex phonetic transcription.


A detailed description of the transcription convention used with the logic behind follows


Transcription Conventions used



The following general principles apply to the transcription conventions used in this project:


a- JI being a Semitic language, there are a lot of affixes. Very frequently, these will alter the sound of the word by assimilation or elision. In this case, the root and the affixes have been retained, separated by a dash (-). You have to listen to the recording in order to find out how the transcription sounds. However you can easily carry out a search on the proper root of the word using ELAN’s internal search engine.


b- Common with Baghdad Christian Arabic, the JI pronunciation of the letter ‘r’ is not a trill but a sound similar to the Parisian French ‘r’. For clarity and to distinguish it from the letter ‘gh’ which is pronounced the same way, it is transcribed as ‘R’ i.e. in CAPTAL. This will facilitate better searching for the speaker community as well as Arab academics. Note that it is not always the case with JI that ‘r’ is pronounced as the soft ‘R’; examples are Hebrew imports and modern Arabic words. In this case the trill is transcribed as ‘r’.


c- If the word originates from another language, or there is a code switching by the speaker, the following convention applies in a “Note” tier:

(EN) = English

(CA) = Classical Arabic

(H)    = Hebrew

(P)    = Persian

(T)    = Turkish


d- Where a square bracket [..] appears in a translation line, it indicates an addition by the translator for clarification.


Table of Phonemics used

The table below shows the roman characters used on a QWERTY keyboard with the equivalent Arabic script.


Consonants (For the IPA equivalent, ctrl + click)

‘ =   ع  voiced pharyngeal fricative

b = ب  voiced bilabial stop, similar to English “b”  in bake

B =     emphatic (velarized) “b”

ch =  چ  alveolar affricate, similar to English “ch” in chime. Used mostly for imported Persian or Turkish words or Muslim Iraqi   pronunciation of the “k”.

d = د  voiced dental stop, similar to English “d” in dad.

dh = ذ  voiced interdental fricative similar to the English “th” in than.

DH =  ظ velarized “dh”, voiced inter-dental fricative , no equivalent in English.

f  = ف  unvoiced labio-dental similar to English “f” as in fan.

g = گ  voiced velar stop similar to English “g” in get. Used for imported words from Persian, Hebrew or Muslim Iraqi where the “q” is pronounced as “g”.

gh = غ voiced velar fricative, no English equivalent.

h  =  ه  voiceless glottal fricative similar to English “h” in home. 

H = ح  voiceless pharyngeal fricative; no English equivalent.

j =  ج  voiced alveolar affricate similar to English “j” in jail

k = ك  voiceless velar stop, similar to English “k” in kit

kh =  خ voiceless velar fricative similar to Scottish “ch” in Loch

l = ل  voiced alveo-dental lateral similar to English “l” in lame.

L =    velarized “l”  similar to English “l” in full.

m = م  voiced bilabial nasal similar to English “m” in man.

M =    emphatic “m” .

n = ن  voiced dental nasal similar to the English “n” in neat.

N =    emphatic “n” .

p = پ voiceless bilabial stop similar to English “p” in patrician

P =    emphatic “p” similar to English “p” in Path

q = ق  voiceless uvular stop; no equivalent  in English

r  = ر  alveolar trill similar to Spanish “rr” in burro. Note that where the JI pronunciation             of the trill “r” has been rendered as “gh”, it is transcribed as Capital “R”

R =     the JI pronunciation of “r” .It sounds like “gh” almost like the Parisian French “r”             in ratter

s = س   voiceless alveolar fricative, similar to English “s” in sad.

S =  ص emphatic or velarized “s”.

sh = ش  voiceless post-alveolar fricative, similar to English “sh” in sheep.

t =  ت voiceless alveolar stop, similar to English “t” in take.

T  = ط  emphatic or velarized “t”, no equivalent in English.

th = ث voiceless interdental fricative similar to English “th” in thank.

v =     voiced labio-dental fricative, similar to English “v” in volition. It is used for            import words from French and English.

z = ز voiced alveolar fricative similar to English “z” in zebra.

zh = voiced alveolar fricative similar to French “j” in Je. Rare; used for imported            French names such as Giselle or French sentences. It occurs in JI where the        affricate j is followed by a consonant, e.g. Jdidi > zhdidi.


w = و  voiced labial-velar approximate, similar to English “w”  in way. Note this is          equivalent to the Arabic consonant “و” and not the long vowel “و” transcribed         as “u”; see below.

y= ي  voiced palatal approximate, similar to English “y” in Yes


The Vowels

Short vowels:

a  low front/back unrounded, similar to English “a” in snap.

o mid back rounded, similar to English “o” in tomato.

i  high front unrounded similar English “i” in tip.

e  the central vowel, similar to French ‘e’ in je. Also used for anaptyctics

u  high back rounded, similar to English “u” in bull.

Long Vowels: The vowels are capitalized to render them long.

A  similar to English “a” in father.

I   similar to English “ee” in cheese.

O  similar to English “o” in  horn.

U   similar to English “oo” in choose.


ai   mid front, unrounded, similar to English “a” in late.


Note on the Arabic Hamza (the glottal stop):

Spoken JI has no glottal stops except at the start of a word . In this case the short vowels will be used. For example, ana (I am), enta (you, masculine, sing.), u (and). Where code switching to standard Arabic is found with a glottal stop in the middle of the word, the symbol /~/ is used.

Epenthetic (helping) vowel

Generally no epenthetic vowels have been used for transcription. This is in order to keep to the phonemic integrity of the word. Some exceptions occur, for example the Arabic prefix “b” (meaning “in” ) when followed by a consonant, would require an epenthetic vowel to sound “eb” . This will be transcribed as either “eb-“ or  more frequently as “b“ followed by the noun.


For a table of consonants used laid out in IPA form see Consonants table.

[an error occurred while processing this directive]