Maltese Arabic: Correction?

« previous post | next post »

In Victor's recent post "Arabic and the vernaculars, part 6", he wrote that "I do not include Maltese because of the Romance superstrata". A more elaborate version of this idea can be found in the Wikipedia article, which tells us that

Maltese […] is a Semitic language derived from late medieval Sicilian Arabic with Romance superstrata spoken by the Maltese people. […] Maltese is a Latinized variety of spoken historical Arabic through its descent from Siculo-Arabic, which developed as a Maghrebi Arabic dialect in the Emirate of Sicily between 831 and 1091. As a result of the Norman invasion of Malta and the subsequent re-Christianization of the islands, Maltese evolved independently of Classical Arabic in a gradual process of latinization. It is therefore exceptional as a variety of historical Arabic that has no diglossic relationship with Classical or Modern Standard Arabic. Maltese is thus classified separately from the 30 varieties constituting the modern Arabic macrolanguage.

Both Victor and Wikipedia are somewhat wrong, or at least misleading — and my main evidence for this is an amusing anecdote. So onwards…

The 2010 LREC conference was held in Valetta, Malta. One of the other attendees was my colleague Mohamed Maamouri, a native speaker of Tunisian Arabic who has extensive research and publications on relevant topics, starting from his 1967 Cornell thesis The Phonology of Tunisian Arabic. His Google Scholar listing also includes "Language education and human development: Arabic diglossia and its impact on the quality of education in the Arab region" (1998),  "Dialectal Arabic telephone speech corpus: Principles, tool design, and transcription conventions"  (2004), "Developing LMF-XML Bilingual Dictionaries for Colloquial Arabic Dialects" (2012), The Georgetown Dictionary of Iraqi Arabic (2013) and The Georgetown Dictionary of Moroccan Arabic (2019).

The area around the conference venue in Valetta was oriented toward tourism, and so the personnel in restaurants, coffee shops and similar sites communicated with their customers in English, French, or other European languages. But among themselves they spoke Maltese — and naturally enough, Mohamed often joined their conversations.

Since he was clearly part of the conference crew, the puzzled response was "How do you know Maltese?" His answer (in Tunisian Arabic) was "Oh, I'm actually speaking Tunisian".

And the reaction was total shock. Eyebrows went up, eyes opened wide, bodies jumped back. This was WTF type shock,, not "how wonderful" type shock — the clerks and baristas were clearly upset at the idea that Maltese and Tunisian are not only mutually intelligible, but are sometimes indistinguishable, at least at the level of brief conversational exchanges.

After a couple of such experiences, Mohamed changed his responses, to something like "Oh, my mother is Maltese…" And of course he also looked into the matter at greater length, including discussions with local Maltese linguists.

The conclusion (at least as far as I recall, 14 years later): Despite the strong socio-political divergence over the many centuries during which Malta resisted first the Arab conquest of the Maghreb and southern Italy and Spain, and then the Ottoman conquest of the eastern Mediterranean and southeastern Europe, some varieties of Maltese Arabic are apparently quite close to some varieties of Tunisian Arabic.

How can that have happened? Apparently, it's partly because of what happened in the 11th century to get the Maltese language started; but it's also because Malta is actually quite close to Tunisia by sea, and over the centuries, there was a lot of under-the-table cultural and commercial contact. This also apparently includes a certain number of families with branches in both countries.

As I said, this is my memory of what I learned in 2010, and it might well be wrong in parts. But it suggests what Victor and Wikipedia said about Maltese may be more a matter of politico-religious prejudice than the conclusion of linguistic analysis.

See  also Martin Zammit, "The Sfaxi (Tunisian) element in Maltese", 2014:

Much to the frustration of the linguist and the historian alike, the facts surrounding the diffusion into Malta of some form of Arabic dialect, or dialects, are still shrouded in the mists of time. Neither the Arab Muslim geographers and historians nor travelers have been very informative about the Arab experience in Malta. […]

Muslim Sicily was certainly a main source of Arabic diffusion into the Maltese islands, particularly from the year 1053 onwards, in the wake of the fall of Sicily’s central government and the ensuing civil war. This migration coincided with a similar migration which, according to the historian al-Marrākushī, left the coastal towns of NorthAfrica and the interior regions and sought refuge in Sicily, Fez and Spain. These people were fleeing from the Banū Hilāl and the Banū Sulaym incursions which had reached Ifrīqiya by the year 1052. Even though al-Marrākushī does not mention Malta, “… there is a very good probability that many fled to Malta as well as Sicily.” Whatever dialectal varieties had reached the Maltese islands, towards the end of the 11th century these started coming into regular contact with the Romance languages of Malta’s military, civil and religious rulers. By the first half of the 15th century, such linguistic contact ultimately forged the lingua maltensi.

Zammit cites a long list of shared phonological and morpho-syntactic traits between Sfaxi Tunisian and Maltese, and also notes that "Tunisian, particularly sedentary, dialects and Maltese share substantial common lexicon".

A picture of somewhat greater divergence emerges from A. Čéplö et al., "Mutual intelligibility of spoken Maltese, Libyan Arabic, and Tunisian Arabic functionally tested: A pilot study" (2016):

It was found that there exists asymmetric mutual intelligibility between the two mainstream varieties of Maġribī Arabic and Maltese, with speakers of Tunisian and Libyan Arabic able to understand about 40% of what is being said to them in Maltese, against about 30% for speakers of Maltese exposed to either variety of Arabic. Additionally, it was found that Tunisian Arabic has the highest level of mutual intelligibility with either of the other two varieties. 

But Čéplö et al. start their paper this way:

In Neo-Arabic dialectology, the concept of mutual intelligibility is often invoked – whether in positive (Ryding 2005) or negative terms (Abu-Haidar 1992) – to conveniently illustrate various claims about the nature of the complex linguistic landscape that is Arabic and the relationship between the Arabic varieties. As one of those varieties, Maltese is also a topic in the mutual intelligibility discussion, where the claims range from total lack of mutual intelligibility with any variety of Arabic (Owens 2010) to anecdotal evidence asserting that speakers of Arabic (usually Tunisian Neo-Arabic; see Chaouachi 2014) are able to understand Maltese nearly perfectly.

Presumably variations in the various earlier subjective and anecdotal claims, as well as the results of their controlled study, depend partly on the substantial variation in what counts as "Maltese" or "Tunisian" or "Libyan" Arabic. But it seems entirely wrong to exclude Maltese from a taxonomy of Arabic "colloquials" or "vernaculars" (i.e. Arabic languages), purely on the grounds of its borrowings from Italian.


  1. J.W. Brewer said,

    May 13, 2024 @ 2:04 pm

    There are lots of different criteria you can use in classifying language varieties and different ways to be a lumper or a splitter. The wikipedia piece notes the distinctiveness of Maltese in that unlike Tunisian Arabic and many other such Arabics it does not exist in a situation of diglossia with so-called "Modern Standard Arabic," to which you could add that its speakers generally have no knowledge of Classical Arabic texts and do not live in a society where Classical Arabic has a particular prestige or numinosity.

    For purposes of understanding e.g. the synchronic morphosyntax of Maltese, these sociolinguistic-context factors may seem extraneous or irrelevant, but for many other purposes relevant to linguistics scholarship these contextual factors do seem quite relevant, maybe more so than the presence of a given quantum of Italian-origin loanwords, to understanding what makes Maltese distinctive and how it is similar to or dissimilar from Tunisian Arabic. And while many linguistics scholars view writing systems as a trivial epiphenomenon with no real connection to what is interesting about the languages written in them, the fact that Maltese is Latin-scripted strongly reinforces its speakers' lack of easy access to other Semitic tongues, and that lack of access itself has real-world consequences. To take a non-written example – if popular tv shows with dialogue in Egyptian Arabic are watched across the wider Arabic world even where people do not have a perfect understanding of the Egyptian vernacular, that is an obvious vector via which some Egyptian distinctiveness (whether in slang lexemes or particular syntactic formations) may leak into and influence other Arabic topolects in the Maghreb or the Gulf or what have you. But if no one in Malta is watching those shows, Maltese won't be reached by that sort of horizontal diffusion among sister languages.

  2. Chips Mackinolty said,

    May 13, 2024 @ 5:26 pm

    As a complete amateur, linguistically speaking (but a huge fan of Language Log) my time in Sicily was a fascinating language exercise. As, arguably, the most invaded/occupied country in the world, Sicilia has been influenced by so many different occupier languages over the last few thousand years–least of all forms of Arabic.

    One result is that full-on Sicilian language/dialect is close to incomprehensible to, for example, northern Italians. When visiting friends in Bologna I was often laughed at when speaking pretty lame Italian for my "southern" accent and vocabulary.

    There is still a large Arabic-speaking population in Palermo, where I lived, with major Muslim celebrations in the streets every year.

  3. Philip Taylor said,

    May 14, 2024 @ 4:30 am

    I have never visited Sicily, Chips, but many years ago when I was searching for a particular road in Tunbridge Wells, I asked a passer-by for directions and he replied "Well, you can tell from my accent that I am not local" and asked if I could work out from where he came. I replied that it was almost certainly Italy, probably southern Italy, and my best guess would be Sicily. He was overjoyed and went on to tell me all about the island’s Greek heritage.

  4. Abdulaziz Al-Naqeb said,

    May 14, 2024 @ 8:07 am

    The Maltese Language developed from Kanaani Language, The Kanaani or Vincent People arrived to Maltese before Rome wad founded, the People of Kanaani controlled the Mediterranean around 2 thousands before cr. The. There are many evidence in Maltese say that for Example Al-Qaim Tample, that was founded before crest 1500 years.
    If the language of Maltese came with Muslim, it was gone
    with them.
    But because the Maltese Language was older than Islam, it survived because it is language of the People of Maltese since 4000 years and it is related not to any Religion.

  5. David Marjanović said,

    May 14, 2024 @ 8:14 am

    There is still a large Arabic-speaking population in Palermo, where I lived, with major Muslim celebrations in the streets every year.

    That's not "still", it's "again" – immigration from Libya and beyond in the last few decades, often by way of Lampedusa.

  6. Mark Liberman said,

    May 14, 2024 @ 9:42 am

    @Abdulaziz Al-NaqebL "The Maltese Language developed from Kanaani Language, The Kanaani or Vincent People arrived to Maltese before Rome wad founded, the People of Kanaani controlled the Mediterranean around 2 thousands before cr."

    Can you provide any references? I can find nothing about "Kanaani" or "Vincent" people, or an associated language.

    There are several Maltese megalithic sites, include Ħaġar Qim, but I have not found any theories for the builders' language, or any empirical basis for such theories.

    And whatever survivals there may be from the builders of those structures, there is no doubt that today's Maltese language (in all its varieties) is a form of Arabic.

  7. Heidi Renteria said,

    May 14, 2024 @ 10:30 am

    I'd be very interested to know whether, over the centuries, Arabic has had any influence on Corsican. Does anyone care to comment?

  8. Victor Mair said,

    May 14, 2024 @ 12:45 pm

    I think that, by "Kanaani", Abdulaziz Al-Naqeb is referring to Canaan:


    Canaan (/ˈkeɪnən/; Phoenician: – KNʿN; Hebrew: כְּנַעַן – Kənáʿan, in pausa כְּנָעַן‎ – Kənāʿan; Biblical Greek: Χανααν – Khanaan; Arabic: كَنْعَانُ – Kan‘ān) (Wikipedia)


    And that, by "Vincent People", he is referring to the Phoenicians (deformation of the labial plus some other transformations):


    The Punic religion, Carthaginian religion, or Western Phoenician religion in the western Mediterranean was a direct continuation of the Phoenician variety of the polytheistic ancient Canaanite religion. However, significant local differences developed over the centuries following the foundation of Carthage and other Punic communities elsewhere in North Africa, southern Spain, Sardinia, western Sicily, and Malta from the ninth century BC onward. (Wikipedia)


    As I will show in a post that I will make later today, it is held by specialists on Maltese that it "may well retain substratal elements inherited from Phoenician and Punic".

  9. Chips Mackinolty said,

    May 14, 2024 @ 8:11 pm

    @ Philip Taylor

    Your email made me laugh! I just love the way languages/accents can be shared and celebrated rather than derided!

    @ David Marjanović

    It's both "still" and "again"! There are many street/area names in Palermo, for example, that derive from Arabic language origins–often with bilingual street/place-name signs. And yes, Lampedusa has been a major source of Arabic-speaking migrants to Italy in recent decades.

    A recent mayor of Palermo, Leoluca Orlando, has been a strong promoter of links to Northern Africa and the Arabic-speaking world.

  10. Julian said,

    May 15, 2024 @ 12:44 am

    "speakers of Tunisian and Libyan Arabic able to understand about 40% of what is being said to them in Maltese, against about 30% for speakers of Maltese exposed to either variety of Arabic."
    Just curious: how is this measured? Is there an industry standard for measuring it? Presumably it might vary according to the setting and the nature of the conversation.

  11. bulbul said,

    May 15, 2024 @ 2:28 am

    Full disclosure: I am Čéplö of Čéplö et al. 2016.

    To me, the whole discussion of whether Maltese is Arabic or not is somewhat pointless and when the issue is raised, it usually is political. Hell, there is a reason Maltese linguistic tradition prefers to speak of "Semitic" words, roots, structures etc. instead of Arabic ones. My original training is in Arabic and diachronic linguistics, so to me, the answer to "Is Maltese Arabic" is obviously yes. Same as Kormakiti Arabic, Afghanistan Arabic, Uzbekistan Arabic, Kinubi or Juba.

    FYI, I am currently working on computational phylogeny of Arabic varieties (coming soon to AIDA which this years also takes place in Malta). Here is a sneak peak. The data favors innovations, so it is not surprising clusters with all the other peripheral varieties, but stands separately even from them.

    This is the paper which describes the methodology. tl;dr is that there is no industry standard, but there have been many mutual intelligiblity studies, especially on Nordic languages (like this classic) and Slavic and Romance languages; you may want to check out Charlotte Gooskens work, she is the leading expert. I modelled my study on one by Vincent van Heuven (Charlotte's long-time collaborator) on, funnily enough, topolects of Chinese. I adapted the method a little bit, though.

  12. bulbul said,

    May 15, 2024 @ 2:36 am


    As I will show in a post that I will make later today, it is held by specialists on Maltese that it "may well retain substratal elements inherited from Phoenician and Punic".
    Now that I am curious about, because to my knowledge, no specialist on Maltese holds that. In fact, it is somewhat of an orthodoxy in Maltese historical linguistics, that Malta was depopulated – perhaps not completely, but certainly enough to ensure a complete break from Punic – some time in the 9th century. See Joseph Brincat's "Malta 870-1054. Al-Himyarī's Account and its Linguistic Implications".

  13. bulbul said,

    May 15, 2024 @ 6:02 am

    Re: "may well retain substratal elements inherited from Punician"

    Now that Victor's other post is up, I see what he meant, namely Alexander Borg's 2007 statement that "Given the fact that [contemporary Maltese] may well retain substratal elements inherited from Phoenician and Punic…"
    With all due respect to Victor, that is one specialist. And with all due respect to Alexander Borg, this is nothing but baseless speculation. It is tantalizing, but as with his 1996 speculation "On some Levantine linguistic traits in Maltese", it remains just that, speculation. Except with the purported Levantine traits, there is at least some evidence. The Phoenician/Punic claims are baseless.

  14. Victor Mair said,

    May 15, 2024 @ 6:10 am

    @bulbul / Čéplö

    Thank you very much for your input. I will respond to you in detail later today.

    I notice from your comment just posted that you did look at my long piece on Maltese linguistics. No cranks who "flood" your FB page there. All good people that I think you would approve of.

  15. bulbul said,

    May 15, 2024 @ 6:37 am


    I am looking forward to your responses. Your longer piece is indeed nothing but respectable scholars, though some of their arguments are unconvincing.

  16. Yuval said,

    May 15, 2024 @ 7:03 am

    I don't think "Slavomir" begins with an A.

  17. Jarek Weckwerth said,

    May 15, 2024 @ 9:29 am

    @bulbul Forgive the intrusion, but now that you have made your disclosure, may I ask what the meaning is of the ö in Čéplö? Innocent bona fide question!

  18. bulbul said,

    May 15, 2024 @ 10:16 am


    not at all: my father's family is Hungarian, but has lived in the territory of today's Slovakia since mid 18th century. The original spelling of the last name is Cséplö or Cséplő (cf. item 16 here), but when I was born, there was a terrible typewriter accident. We don't speak of it.

  19. David Marjanović said,

    May 15, 2024 @ 1:36 pm

    A recent mayor of Palermo, Leoluca Orlando, has been a strong promoter of links to Northern Africa and the Arabic-speaking world.

    Ah, so he had the bilingual signs put up, and there's no evidence for continued presence of Arabic in Sicily in the last thousand years…

    especially on Nordic languages (like this classic)

    I thought you were going to link to the sketch that says nobody understands Danish…

  20. Victor Mair said,

    May 15, 2024 @ 9:59 pm

    See "Aspects of Maltese linguistics" (5/14/24)

  21. Jarek Weckwerth said,

    May 16, 2024 @ 10:57 am

    @bulbul: Thank you, I suspected something along these lines but couldn't quite put my finger on it. The idea of using a diacritic from the "recipient" script while leaving in one from the "donor" script just didn't occur to me…

  22. bulbul said,

    May 16, 2024 @ 12:04 pm


    I thought you were going to link to the sketch that says nobody understands Danish…
    Why would I? All I need to do to drive this point home is quote from the Wiki entry on the TV show Borgen:

    Borgen (Danish pronunciation: [ˈpɒˀwn̩]) is a Danish political drama television series.

  23. Mike Maxwell said,

    May 21, 2024 @ 9:34 pm

    @Mark, when I saw Victor's later post on this topic (LanguageLog lends itself to reading from new to old, sort of like top-posting in email), I was going to respond with something like your anecdote–until I saw you had already done so. I have a vivid memory of Mohamed Maamouri at a restaurant near the LREC meeting pantomiming a Tunisian saying that literally meant s.t. like "full and overflowing" but was used as a metaphor for thanking someone. It turned out Maltese had the same idiom. (For all I know, lots of Arabic varieties do.)

  24. Mike Maxwell said,

    May 21, 2024 @ 9:41 pm

    @Julian, you ask whether there's a measure of mutual intelligibility. SIL (the Christian missionary organization that does Bible translation into very low density, often previously unwritten, languages) has a career specialty in this intelligibility testing, which helps them determine whether a translation is needed for a given language variety. The results make their way into the Ethnologue, as one way of determining the answer to the vexed language vs. dialect question.

    Afaik, this is not an "industry" standard (your other question), since in most cases it's irrelevant to any language _industry_ (with the exception maybe of Arabic varieties). It does of course make for contention among linguists, native speakers, and politicians, since no matter how you measure it, you're likely to offend someone.

RSS feed for comments on this post · TrackBack URI

Leave a Comment