Aspects of Maltese linguistics

« previous post | next post »

[Full disclosure:  the reason I am so consumed by the Arabic vernaculars is because of their own inherent, intrinsic nature, but I must confess that I'm also preoccupied by their comparative parallelism with the Sinitic "topolects".  The workings of both are extremely difficult to comprehend.]

This post is to follow up on VHM's "Arabic and the vernaculars, part 6" (5/12/24) and Mark's "Maltese Arabic: Correction?" (5/13/24), plus J.W. Brewer's excellent first comment to the latter.

Mark ends his post thus:  "…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."  I would not want to do that.

To provide for a more nuanced evaluation of the position of Maltese vis-à-vis the Arabic vernaculars, below I cite several scholarly accounts of the subject and related issues.  Extensive coverage of the history of the languages on Malta is provided.


Maltese language, Semitic language of the Southern Central group spoken on the island of Malta. Maltese developed from a dialect of Arabic and is closely related to the western Arabic dialects of Algeria and Tunisia. Strongly influenced by the Sicilian language (spoken in Sicily), Maltese is the only form of Arabic to be written in the Latin alphabet."

That's the bare bones.  As we shall find in the following paragraphs, the complexities of Maltese are far greater than can be told in such a capsule description.

Maltese, ch. 11 of The Cambridge Handbook of Arabic Linguistics

"A Peripheral Dialect in the Historical Dialectology of Arabic"

from Part II – Arabic Variation and Sociolinguistics

by David Wilmsen

Maltese is one of the so-called 'peripheral' dialects of Arabic, language varieties that are descended from Arabic but that have for various reasons become isolated from contact with its mainland.


David Wilmsen examines Maltese, a peripheral dialect of Arabic. Of those, Maltese stands out as remarkably unusual. Unlike other dialects of Arabic, it is an official language of the state in which its speakers reside, the Republic of Malta, as well as being an official language of the European Union. It boasts a long literary tradition, a language academy, an active press, scholarly journals and societies devoted to it, and an ever-growing digital presence, including a large online, freely accessible corpus encompassing hundreds of millions of words. It is therefore an easily accessible language for linguistic research. The chapter examines Maltese in light of linguistic thinking about so-called enclave dialects, showing that Maltese conforms to the general characteristics of remnant dialect groupings, in that it does borrow from the languages with which it comes into contact, it does undergo independent internal change, and it does retain features of its founder languages. As such, Maltese can be instrumental in demarking the latest date for the emergence of a range of features found variously in mainland dialects of Arabic.

MYL's comment on Arabic "colloquials" vs. Romance languages here:

Contemporary Arabic "colloquials" are approximately as different from one another as the various Romance languages are — having diverged from the original form, sociolinguistically and geographically, at just about the same time as the descendants of Latin did.

Obviously the ideology is radically different — it's as if the official view were that Latin is the only correct language, while French, Spanish, Italian, Romanian, etc. are just the lawless and errorful result of inadequate education.

Sorting out the position of Maltese within the Arabic branch of the Afro-Asiatic language family will profit from a deeper dive into the latter:

Symposia Melitensia (L-Università ta' Malta). 2007, Vol. 4, pp. 1-52

"Between typology and diachrony : some formal parallels in Hebrew and Maltese"

by Alexander Borg


Hebrew and Maltese are obliquely related members of the Semitic language family. Past comparative research inspired by Bible translation highlighted in atomistic fashion a number of common traits in these two languages. The present research probes aspects of selected phonological, morphological, syntactic, and lexical traits in Biblical and Israeli Hebrew from the comparative perspective of contemporary Maltese. Given the fact that the latter may well retain substratal elements inherited from Phoenician and Punic, the parallels tentatively indicated here, particularly in the lexical domain, may provide the basis for a reconstruction of the earliest diachronic stage of the Maltese word stock. If on the mark, it also seriously calls into question claims advanced in recent historical work on Maltese to the effect that the Arab invasion of the Maltese Islands in the 9th century entailed the complete annihilation of the indigenous population thereby breaking the continuity with the linguistic heritage of pre-Arabic ancient Malta.

Returning from this brief, comparative inquiry into one facet of the deep relationships of Maltese, we now take a broader view of the language.

"Maltese, a Language so Unique in Europe:

Arabic origins, Sicilian and Italian vocabulary and a strong influence of English are the characteristics of an original European language with a long history: Maltese."

by Joseph M. Brincat (University of Malta [L-Università ta’ Malta]), Orient XXI (5/7/22)

Because it affords such a comprehensive, informed account of the engrossing linguistic situation on Malta, I will quote extensively from Professor Joseph / Giuseppe Brincat's article.

Situated in the middle of the Mediterranean Sea, Malta and Gozo present an intriguing linguistic picture. Although they are nearer to Sicily (93 km) than to Tunisia (288 km) and Libya (355 km), the Catholic, culturally and genetically European inhabitants still speak a language that is basically a variety of Arabic spoken in the Maghreb and Sicily around the year 1000. Its survival is unique because Sicily, Spain and Pantelleria abandoned Andalusi and Siculo-Arabic, but Malta was spared language shift by the Normans, Swabians, Anjevins, Aragonese and Castilians although it had remained part of Sicily. When Charles V of Spain ceded it to the Knights of St. John in 1530 it became an autonomous state, but the Grand Masters (French, Portuguese or Catalan, with a handful of Italians) did not apply a linguistic policy on the spoken level. They were content to use Latin and Italian as high languages. Besides, education was poor and literacy hovered around 10%, and so the population remained mostly monolingual.

Under Muslim rule (870 – 1091) the population was less than ten thousand inhabitants, but under the Knights of St John it reached 100,000 till 1798. It kept increasing during the British period and now surpasses 500,000. The disproportion between its territory (316 km2) and its population, together with the large number of Sicilian, Italian and English surnames, and a few French and Spanish ones, reveals the importance of immigration in shaping the local language.

I met a Chinese family who happily and successfully spent some years running a business in Malta, using English for their daily needs and transactions, before moving to America, where they are now happily residing and using more Chinese than they did in Malta.

Brincat's article probes the earliest evidence for language usage on Malta and finds that the first inscriptions are in Punic, "the Phoenicians having settled there in the 7th century B.C."  That was succeeded by Greek and Latin.

Then follows an important paragraph on the linguistic history of Malta:

The language spoken today shows no substrate because Arabic was introduced in a sudden manner. Al-Himyari describes a ferocious Muslim raid in 870 (255 A. H.) and a new settlement in 1048-49 (440 A. H.) composed of Muslims and slaves. Depopulation between the conquest and the settlement is suggested by the lack of both Muslim and Christian cemeteries, and also by the local toponyms which recall the Arabic place-names of Medieval Sicily, especially by the component of raħal or Ħal (Rahal gidit and Rachal saphy in Sicily, Raħal Ġdid and Ħal Safi in Malta). The Maltese language, too, has a marked affinity with the Maghrebin variety that was spoken in Sicily in the Norman age. Roger’s conquest introduced Romance elements in Sicily and Malta, where Arabic customs prevailed from 1091 to 1127 until Roger II reasserted his rule and society took on a European lifestyle.

Another vital section of Brincat's article is titled "From Punic to North African Arabic":

Definitions of the Maltese language differed: the locals saw it as unique and called it lingua maltensi (1436), lingua melitea (1549), whereas foreign travellers heard unfamiliar sounds and called it parlata africana (1536), parlar saracino (1558), “un langage arabe corrompu” (1694). Hieronymus Megiser was intrigued: “Although they are Christians, they make use of a language which is Saracen, Moorish, or Carthaginian or ‘lingua franca’, which is a kind of Arabic and which has its origin in Hebrew”. The scientific classification of the Semitic languages was still distant, but Megiser deserves credit for printing a booklet on Maltese, the Propugnaculum Europae in 1606, listing 121 words in German translation. Jean Quintin (1536) associated the Maltese language with the Punic inscriptions and, although the script had not been deciphered yet, the idea pleased many scholars in Malta and the myth was perpetuated for political, religious, and racial reasons. However, the historian Gian Francesco Abela (1647) had understood the real origins of Maltese, was aware of the Arabic substrate in the Sicilian dialect and knew that a similar language was spoken in Pantelleria. In 1810 Wilhelm Gesenius gave scientific proof that the origins of Maltese lied in a North African dialect of Arabic, but the Punic myth was upheld by some local scholars up to the 20th century.

The standardization of Maltese was a slow process. At first only a few isolated words appeared in notarial deeds and in the minutes of the local government, mostly place-names and domestic objects, but around 1470 Petrus Caxaro wrote a poem of sixteen lines, a cantilena, modelled on Romance genres. It is revered as the earliest text in the Maltese language. Occasional phrases in administrative texts show that Sicilian/Italian words were adapted according to Maltese grammar rules. An example from 1473 in a Sicilian sentence shows the word “isfeduene” which is from sfidare, to defy, with initial i for today’s j, a morpheme in the conjugation of verbs in the present tense, third person singular and plural, with the inflectional ending –w, that indicates the third person plural, jisfidaw (they defy), and the pronominal suffix –na, which means “us”, therefore “they defy us.”

Brincat then moves on to outline the impact of war and religion on the development of Maltese: 

A knight from Provence, Thezan, wrote instructions on firing muskets for Maltese-speaking troops, a short grammar and a glossary of 3,000 words in two sections, Maltese-Italian and Italian-Maltese. Thezan faced a dilemma which troubled sixteenth and seventeenth-century writers: the spelling of sounds that are not Latin. Today Maltese has only two such sounds, aspirate h (ħ) and the glottal stop (q), but in those days pronunciation was closer to the Arabic. Thezan added ten Arabic letters to the Latin alphabet, but Gian Pietro Francesco Agius De Soldanis opted for a wholly Latin alphabet in his grammar (1750) and dictionary of 12,000 entries. Later, Michel Antonio Vassalli published his grammar in 1791 and 1827, and a lexicon in 1796 with 18,000 entries, adopting twelve letters from Greek and Punic but not from Arabic.

In the meantime two important steps were taken in the domain of religion. Ignazio Saverio Mifsud wrote sermons in Italian and in Maltese, and enriched the latter with many Italian words and Latin phrases, attempting a literary style that rose above everyday speech. Even more far-reaching was the translation of the Italian catechism into Maltese by Francesco Uzzino in 1752. This had no literary ambition but for the first time all Maltese boys and girls learned it by heart for their First Holy Communion. Religious terms like fidi, liġi¸ sagramenti, Apostoli, Spiritu Santu, etc. were learnt by constant repetition. As for most domains, both high (administration, culture, law and medicine) and low (carpentry, fishing, building), words denoting religion, churches and their furniture, are largely of Sicilian origin.

Brincat then moves on to examine Napoleonic and English attempts to influence the development of Maltese more toward Francophone and Anglophone usages at the expense of Italian ones, but I won't go into the details of that here.

The final sections of Brincat's article trace the growth of English in the language mix on Malta after it became a British colony in 1813 and the rise of the Maltese language among the population.

Knowledge of Italian and English was limited to educated persons: in 1842 the former was spoken by 11% and the latter by only 5% of the population, but when English became compulsory for employment with the British forces, the Police, the Civil Service, and for emigration, the figure rose to 22.6% in 1931. After the Second World War Malta changed completely. English became fashionable, cinemas showed English and American films, pop songs replaced opera and, most important of all, in 1946 primary education became compulsory, teaching only Maltese and English. The battle against Italian was favourable to the Maltese language because the English had realized that its promotion was indispensable to end the long cultural battle.

Throughout the 19th century Maltese literature grew and found its highest expression in Dun Karm Psaila in the early 20th, while grammars, dictionaries and schoolbooks completed codification. The alphabet was standardized, consisting of Latin letters with a few diacritics: dots distinguished palatal ċ and ġ from velar k and g, voiced s became ż, crossed ħ identified the aspirate, j and w were adopted for the semi-consonants, x was adopted for sh and q for the glottal stop. This followed the principle of one letter for one sound but għ and h, both mute, were retained for etymological reasons. In 1924 cable radio relayed the BBC on one channel and Maltese programmes on the other, and for the first time standard Maltese was heard everywhere by illiterate dialect speakers. A chair of Maltese was set up at the University in 1937 and exams in both official languages were compulsory for jobs in the Civil Service.

How and when did Maltese become an official language on the island of Malta?

Maltese was given official status with English and Italian in 1934. Italian was removed in 1936, but came back in 1957 when Italian television could be viewed from Malta. It assumed a new role now: no longer official, nor limited to culture, it became a passive tool for entertainment and information, kept the highest audience up to the 1990s when local stations got a bigger share, and is still watched by about 20% in prime time. Italian is also the favourite foreign language in the secondary schools. English is the teaching medium of half the subjects in the schools and of all subjects except languages at University. It is also preferred for reading and sending emails and sms, and for ATM banking, but Maltese is spoken by over 90% of the population and boasts two daily newspapers (three in English), three major TV stations, and local radio channels. Production of books and plays is also healthy. In 2003 Maltese became one of the official languages of the European Union.

Where does that leave "peripheral" Arabic, a question that is vital to the debate on the Arabic vernaculars that has been taking place on Language Log for the last few years?  The status of Maltese is crucial for the whole debate, which has significant implications for linguistic taxonomy that go well beyond Arabic and Semitic.

A snapshot of today’s language is revealed by the composition of the lexicon. In Joseph Aquilina’s Maltese-English Dictionary (1987-1990) Arabic words make up 32.4%, Sicilian and Italian 52.5%, English 06.1%. The MED includes archaisms and rare terms among its 41,016 entries, but the Concise version reflects actual usage: its 22,649 entries show more Sicilian and Italian words (61.61%), less Arabic words (22,42%) and a slight increase in English words (8.45%). However, Arabic words comprise grammatical terms and the fundamental vocabulary, and are more frequently used. Together with the rules of grammar (although simplified) they define the language as a variety, albeit “peripheral”, of Arabic. A few examples from the domain of the family will suffice here: members of the inner nucleus have Arabic names: omm, iben, bint, ir-raġel, il-mara, tifel, tifla, (mother, son, daughter, husband, wife, boy, girl), but the word for father is Sicilian, missier. The other members have Sicilian/Italian names: nannu, nanna, ziju, zija, neputi, neputija, kuġin, kuġina, (grandfather, grandmother, uncle, aunt, nephew/grandson, cousin), and the word for family is familja and for relatives qraba. The complementary nature of Arabic and Sicilian/Italian is also seen in the days of the week (Monday to Sunday): it-Tnejn, it-Tlieta, l-Erbgħa, il- Ħamis, il-Ġimgħa, is-Sibt, il-Ħadd; but the months of the year are: Jannar, Frar, Marzu, April, Mejju, Ġunju, Lulju, Awissu, Settembru, Ottubru, Novembru, Diċembru. This stratigraphy marks all the other domains; for example about “bread” we see: ħobż, hobża, qoxra, lbieba, frak, ftira (bread, loaf, crust, crumb, crumbs, low circular shaped bread); bezzun, tal-kexxun, panina, ċabatta, malji (roll, sandwich loaf, flat roll, braided roll); sandwich, toast, baguette.

Informal everyday speech sees a lot of code-switching: firstly, because as most subjects are taught in English their terms are the first words that come to mind; secondly, because young parents prefer to speak English to their babies. At present this is not leading to language shift, because as they grow up children see Maltese as an adult language. Nowadays, the source of innovation is no longer Italian but English: some common words are adapted (kettle > kitla), others are Sicilianized (evaluation > evalwazzjoni), others become false friends (related > relatat, involved > involut), and new formations are created (enforceable > inforzabbli, developers > żviluppaturi, privacy > privatezza, occupancy > okkupanza). This is necessary for the language to keep up with social progress. A tricky problem is how to write unadapted English words, like bicycle and washing-machine (bajsikil? woxing maxin?). It is not easy to decide whether they are irreplaceable or translatable. Otherwise, being recognizably English, they can be attributed to code-switching, in which case they are written in their original spelling.

Brincat's concluding paragraph is profoundly telling:

Understandably, the pressure of English in a bilingual situation like Malta’s is very strong. As the 2011 Census shows, almost all Malta-born citizens know Maltese (99.6%) and English (91.3), many also know Italian (61.3%) and French (21.4%), but few know German (5.1% ) and Arabic (4.3%).

One thing is certain, the Maltese are proud of their language and have done much to study and preserve it, even in the face of overwhelming influence from English.

I am grateful to Mark for prodding me to look deeper into the evolution of Maltese as a language that is indeed "unique" and to examine the island nation of Malta as a fascinating linguistic laboratory.




(this article has extensive coverage of all aspects of Malta's land, climate, flora and fauna, people (including ethnic groups and demographic trends), religion, economy, government, education culture, and history)

Maltese and English are the official languages of Malta as well as official languages of the EU. Maltese resulted from the fusion of North African Arabic and a Sicilian dialect of Italian. It is the only Semitic language officially written in Latin script. English is a medium of instruction in schools. Italian was the language of church and government until 1934 and is still understood by a sizable portion of the population.

The country of Malta became independent from Britain and joined the Commonwealth in 1964 and was declared a republic on December 13, 1974. It was admitted to the European Union (EU) in 2004.


Selected readings



  1. Igor said,

    May 15, 2024 @ 1:28 am

    Just a side note.

    Your rss 2.0 and atom feeds seem to be invalid.

  2. bulbul said,

    May 15, 2024 @ 3:09 am

    The status of Maltese is crucial for the whole debate, which has significant implications for linguistic taxonomy that go well beyond Arabic and Semitic.
    What is even the debate? It is incontrovertible that Maltese is ultimately descended from the same varieties that gave rise to Damascus Arabic, Cairene Arabic, Juba or Uzbekistan Arabic, I don't see anyone arguing against that (well, except for cranks occasionally flooding my FB page, but I don't think they are relevant here). So what is the debate? The details of how it happened and when and what led to it etc. are of course a worthy subject of study, but I am not sure I see a debate. There isn't even a controversy of the type that plagues creolistics (another field of relevance to this).

  3. bulbul said,

    May 15, 2024 @ 3:37 am

    As for the stratification of Maltese vocabulary, Christopher Lucas and I did a metastudy on various works on the subject as a part of our chapter on contact induced change in Maltese. The strata looks different depending on what you look at.
    We (well, in this case more I) also looked through Aquilina and searched for anything that could be identified as Punic/Phoenician. Suffice it to say, we found more Aramaic words than Punic ones, so came to the same conclusion as Grech did back in the 1980s. The Aramaic loanwords somewhat support Alexander Borg's view that Maltese has at least some traces of Eastern (Levantine) Arabic.

  4. Victor Mair said,

    May 15, 2024 @ 5:57 am


    Thank you very much for joining the debate. I will respond to you in detail later today.

  5. bulbul said,

    May 15, 2024 @ 7:00 am

    More from Joe Brincat on Phoenician/Punic ("Maltese and Other Languages", Malta: Midsea Books, 2011 p. 30):

    "Nowadays the origin of Maltese have been firmly traced to a dialectal variety of Arabic and scholars who have seriously examined the Phoenician/Carthaginian question, like Prosper Grech (1961), have not found traces of Punic in Maltese as it is today. Alexander Borg's approach is still open but very cautious: he mentions Jean Cantineau's (1960) attribution of tafxim of /ā/ > /ō/ to Phoenico-Punic substratum (1976a: 195 n. 8), but this is only featured in Maltese dialectal varieties and requires further investigation."

    Brincat refers here to varieties of Maltese where the /ā/ > /ō/ shift is documented even in older literature, e.g. Stumme's collection of fairytales has a text from Balzan (central Malta) with words like /shōp/ < Standard Maltese sħab /sħāp/ "friends" or /kmōmar/ < SM kmamar /kmāmar/ "rooms".

    Brincat also raises the possibility that

  6. bulbul said,

    May 15, 2024 @ 7:03 am

    (ack fat fingers)
    … that if one finds something that looks Hebrew rather than Arabic, it might be the results of contact with the Jewish community of Mdina (op. cit., p. 31). Again, this is just a speculation and Brincat conludes this with

    "For the above reasons we must conclude that the Maltese language does not have a definable substratum because terms of direct Latin , Greek, or Hebrew (Aramaic? Phoenician?) origin are too few."

  7. Jonathan Smith said,

    May 15, 2024 @ 7:38 am

    IDK how parallel, but re: Maltese, compare Taiwanese: while it is similarly obvious (I think?) that Taiwanese descends from the same early language(s) that gave us Mandarin and the rest, scientific terminology is somehow always also ideological. So "Southern Min" (i.e. the southern part of the eastern branch of the Min group of Sinitic) seems neutral to the academic-paper writer, but feels political to many speakers of Taiwanese — the notion of "Southern Min" in Taiwan is associated with the transplanted R.O.C. government / its organs and smells like part of an ethnopolitical sneak attack. This means that associated "scientific"-feeling facts, e.g., that the Taiwanese(s) of Taiwan is 99+% mutually intelligible with Amoy~Xiamen across the strait, aren't exactly embraced by this particular social group… while terms like "Han" and "Chinese" are still harder (impossible?) to wield "objectively" and provoke similarly negative reactions.

    To be emphasized, and to restate what KIRINPUTRA was saying here if I understood correctly, is that, whether in Malta or Taiwan or elsewhere, this is a perfectly legitimate — or at least a perfectly prosaic — sort of sociopoliticolinguistic reaction to have. If one points out that it is not "scientific," one seems not to get it… and one might actually not get it.

  8. Coby said,

    May 15, 2024 @ 8:59 am

    When I visited Malta, I found that I could understand almost all the written Maltese that I saw through my knowledge of Italian and Hebrew (and some knowledge of correspondences between Hebrew and Arabic, though I don't actually know Arabic).

  9. Jarek Weckwerth said,

    May 15, 2024 @ 9:06 am

    @ Jonathan Smith: That's a wise comment. Thank you.

  10. bulbul said,

    May 15, 2024 @ 9:31 am


    precisely. I mean, let's not forget that while we here care about scientific accuracy, this does not apply to, um, normal people. On top of that, there are multiple ways to slice a cake, and they largely depend on the purpose. I have absolutely no problem with people in Malta objecting to their language being called Arabic and I have even less problem with my Maltese colleagues referring to Arabic elements in Maltese as "Semitic".

  11. Mark Liberman said,

    May 15, 2024 @ 11:21 am

    @Victor: "the complexities of Maltese are far greater than can be told in such a capsule description."

    Is there any language (and the associated family of languages) about which this is not the case?

    Linguistic historical relationships are complicated. And attitudes towards linguistic and ethnic identities are complicated. Mix it all up together and it's even more complicated — in ways that seem mostly to obscure rather than to clarity the history…

    But complexities aside, there can be no question whatsoever that Maltese is one of the modern "Arabic colloquials".

  12. Coby said,

    May 15, 2024 @ 11:37 am

    I see a certain resemblance between Maltese and Yiddish. Maltese is "Semitic" (or "Arabic") in the same sense that Yiddish is "Germanic" (or "German"). Both outliers in having a different script from the other languages in the family (or subfamily) and in having a large amount of borrowed vocabulary (though of course English shares that feature among Germanic languages).

  13. David Marjanović said,

    May 15, 2024 @ 1:10 pm

    aspirate h (ħ)

    [h], the voiceless more-or-less-fricative, is already aspirate by definition; ħ is IPA [ħ], the voiceless pharyngeal fricative – it is of course where the IPA symbol comes from –, or more likely IPA [ʜ], the voiceless epiglottal fricative (as found in western Arabic and some other kinds of Arabic, but I don't know where exactly), or maybe both and everything in between.

    Historically, għ was the voiced counterpart of ħ; today, says the long Wikipedia article, it

    has the effect of lengthening and pharyngealising associated vowels (għi and għu are [i̞(ˤ)j] (may be transcribed as [ə(ˤ)j]) and [oˤ]). When found at the end of a word or immediately before 'h' it has the sound of a double 'ħ' (see below).

    (…Evidently the first description of għu is missing…)

  14. S Frankel said,

    May 15, 2024 @ 2:34 pm

    Are there any traces of a Punic substratum in Tunisian Arabic?

  15. Scott P. said,

    May 15, 2024 @ 3:01 pm


    See this post:

  16. Victor Mair said,

    May 15, 2024 @ 7:43 pm

    "Trilingual signs in Sicily" (3/17/17) — with four photographs and many learned / informative comments; the scripts are Latin, Hebrew, and Arabic; the texts in Hebrew script may be in Judeo-Italian (Italkian) or Judeo-Arabic (according to different commenters)

  17. Martin Schwartz said,

    May 15, 2024 @ 8:30 pm

    Under family terms, qrab is not Sicilain but of Arabic origin,
    fro, √q-r-b be near, cf. Standard Aarb. qarîb 'relative', etc.
    The words for '4' and '10' show that Arabic 6ayn has merged with Arabic
    GHayn (a spirantic g, in effect; no gamma on this keyboard)), written as gh–I don't know of such a merger in other Arabic dialects. So, Gesenius proved that the origin of Maltese lay in North African.
    Arabic. In my early 'teens I found an early edition of the Gesenius-Rödiger lexicon of Biblical Hebrew, which was replete with cognates in
    various Semitic languages, a great work for its time; my copy fell apart a few years ago. I'm curious–dubious– about the alleged Aramaic words in Maltese. Martin Schwartz

  18. Victor Mair said,

    May 15, 2024 @ 10:04 pm

    This morning, I promised that I would respond in greater detail later today.

    Actually, because of what Jonathan Smith, Coby Lubliner, and others have already explained, there's not too much more that is necessary for me to say now.

    There is enough different / unusual / etc. about Maltese in relation to the "mainland" Arabic vernaculars — characteristics and historical verities that the specialists I cited in the o.p. have described — that I think it counts as a "special case". It is not just a typical Arabic vernacular.

    I agree with David Wilmsen and others that Maltese is a "peripheral dialect' in the historical dialectology of Arabic. To me, it is misleading to say cut and dried that Maltese is a typical Arabic dialect. Indeed, is a quite atypical Arabic dialect, similar to Dungan as an atypical Sinitic dialect (written in Cyrillic, having a very unusual history, many borrowings from Arabic, Persian, Russian, Turkic languages, etc.).

    If only because of its geographic location, Maltese has a more complicated linguistic history than most other languages, so a capsule description of it is naturally unsatisfying, though one may be useful / necessary for introductory / preliminary / summary purposes.

    As for how the speakers of the vernaculars feel about speakers of Arabic, see "Arabic and the vernaculars, part 5" (8/20/22).

    This entire debate has been an edifying exercise, and it isn't over yet. I feel confident that there will be more "Arabic and the vernaculars, part n" to come.

    P.S.: Added to the "Suggested readings" of the o.p.:

    "Punic Survival Part Five: Did Punic Influence Arabization and Maghrebi Darija?" (7/31/13)

    "Did Punic Survive Until the Advent of Arabic? Part 4: The Post-Augustine Evidence" (7/30/13)

    "Did Punic Survive Until the Advent of Arabic? Part 3: The Evidence of Augustine" (7/29/13)

    "Did Spoken Punic Survive Until the Advent of Arabic? Part Two: Punic After Carthage." (7/26/13)

    "Did Spoken Punic Survive Until the Advent of Arabic? Part One" (7/25/13)

    Maltese is mentioned in these posts from the Editor's Blog (Michael Collins Dunn) of the Middle East Institute. Of course, we have to consider where the Arabic that was present in Malta came from. (I see that Scott P. noticed this valuable site too.)

  19. KIRINPUTRA said,

    May 16, 2024 @ 12:48 am

    @ Jonathan Smith

    (I’m behind on other chores, so I haven’t been able to fully engage, but I wanted to stop by and throw a few things into the ring while there’s time.)

    1st — and I put this first b/c it gets lost in the shuffle elsewhere — while Taioanese is clearly descended from Hokkien & apparently Teochew to some extent, why do you think it’s obvious that Hokkien & Teochew are descended from “the same early language(s) that gave us Mandarin and the rest”? What is the basis of this assumption? When & how was this established? Are there exhibits or publications that demonstrate the point with rigor & intellectual honesty?

    2nd — We should be level-headed enough to look past the claim that Amoy Hokkien is unintelligible to Taioanese speakers. Amoy Hokkien & Taioanese are very much mutually intelligible, but large swathes of Hokkien are functionally unintelligible to speakers of either. And, numbers-wise, Amoy Hokkien represents a very small (& ever-shrinking) slice of the Hokkien pie; linguistically, it’s kind of like a “continental dialect of Taioanese”. So the issue as commonly presented … is cherry-picked. (To what end, if we may ask?)

    3rd — Is “the southern part of the eastern branch of the Min group of Sinitic” valid science in itself? If so, where is the evidentiary basis? If not, we have more to worry about than what to call it.

    4th — As for what to call it, what is so scientific about “Southern Min”? “Hoklo” is arguably more scientific, and less “politically motivated” — although it could be a poor fit for other reasons, as “Southern Min” is as well.

    5th — (As applicable here….) My point was firstly that what contemporary human society calls “languages” are not purely a matter of mutual intelligibility, . In fact, it is “unscientific” to pretend otherwise; and even the academic establishment only does so selectively (such as in the context of “Sinitic”, but not with Philippine languages), which seems hypocritical. (Nor do “scientists” ever object to the social intuition that Penang Hokkien & Amoy Hokkien are varieties of the same language, even though the intelligibility is not mutual. Why the bias?) … So “political animus”, or Taioanese-speaking hotheads claiming in forums online that they can’t understand a word of Amoy Hokkien … has to be seen as distinct (even if related) from the question of whether Taioanese & Hokkien are one language or two on observable, social terms. And that observation is also Science.

  20. bulbul said,

    May 16, 2024 @ 7:15 am


    it is misleading to say cut and dried that Maltese is a typical Arabic dialect
    The idea of "typical" Arabic dialect is something you mention here for the first time. At no point did I say that, nor would I ever, because the notion is, plainly, laughable.

    [VHM: Good to hear you say that.]

    Maltese has a more complicated linguistic history than most other languages
    No one is denying that, least of all me.

    [VHM: Glad that you agree.]

    That does not change the fact that, as I said, Maltese is ultimately descended from the same proto-language that gave rise to Damascene, Cairene, Juba etc. Arabic.

  21. bulbul said,

    May 16, 2024 @ 7:29 am

    Punic Survival Part Five says:

    "There are those who argue that a Punic substratum can still be identified in the colloquial dialects of North Africa (plus Maltese)."

    Excellent, so what are they? The blog post cites this article noting "Elimam presents a table showing vocabulary in common between Punic and Maghribi" which consists solely of common Semitic words like ab and um. This is, naturally, par for the course.

    I do not think it is completely impossible for there to be traces of Punic in Maltese. But in all those years, no one has ever come up with any. If anyone does in a convincing and scholarly manner, there is a bottle of my father's moonshine in it for them.

  22. Victor Mair said,

    May 16, 2024 @ 10:50 am

    Lumpers and Splitters

    I would invite all who are interested in having a productive debate on the distinctiveness of Maltese to go back and carefully reread J.W. Brewer's elegant, informative, reasoned first comment to "Maltese Arabic: Correction?" (5/13/24)

    Also relevant are KIRINPUTRA's thoughtful, sensible analysis of the relationship of Taioanese and Amoy Hokkien and other "Southern Min" languages, plus Mandarin (cf. Arabic) and Sinitic (cf. Semitic), with cogent remarks on intelligibility, Science, and “political animus” in his enlightening comment to "Aspects of Maltese linguistics" (5/14/24)

  23. Victor Mair said,

    May 16, 2024 @ 3:09 pm

    From GeorgeW:

    Years ago, in Saudi Arabia, I took a Qur'anic Arabic course at a local university. I was the only non-Muslim in the class made up primarily of South Asians. I was shocked to realize that they could read and recite the Qur'an very well, but understand hardly a word of what they were reading. I mostly understood the texts, but was slow on pronunciation.

    (a comment to this post: "Nonword literacy" [5/16/24])

  24. Victor Mair said,

    May 16, 2024 @ 3:11 pm

    Talk about intelligibility!

  25. Levantine said,

    May 16, 2024 @ 7:13 pm

    What’s the relevance of GeorgeW’s comment to the present discussion? That a good number of non-Arabophone Muslims learn to read Qur’anic Arabic without understanding it doesn’t tell us anything about the so-called dialects.

  26. Victor Mair said,

    May 16, 2024 @ 8:44 pm

    It has something profound to say about the gulf between "Arabic" (whether MSA, Qur'anic, or Classical) and the vernaculars. Again, a parallel with the topolects and Literary / Classical in Sinitic. How many times have we seen in our discussions on Language Log that people do not understand the nonintelligibility of these two types of language.

  27. Levantine said,

    May 17, 2024 @ 2:39 am

    But the students he was referring to were South Asian, not speakers of Arabic vernaculars. It’s not at all uncommon for non-Arabophone Muslims to learn to read and sound out Qur’anic Arabic (which, unlike most written Arabic, is scripta plena) without actually learning the language.

    Most speakers of Arabic vernaculars grow up with a high degree of exposure to Qur’anic Arabic and (the classically derived) MSA, which means the language of scripture is far, far more accessible to them than to the South Asian students mentioned by GeorgeW.

  28. KIRINPUTRA said,

    May 17, 2024 @ 5:43 am

    @ Victor, thank you for the helpful summary on this fascinating theme (Maltese; Arab speech). I don’t speak anything that could be called Arabic, and that seems to have always made it tricky to understand the subject matter. I guess this mirrors what goes on with “Chinese” speech.

    It strikes me that when the Arabic vernaculars — the more populous ones, at least — are being discussed, the question of mutual speakability tends to be on the table, even when the consensus is ultimately no. Whereas when it’s, say, Teochew vs Hokchew (Foochow) — supposed semi-close relatives according to modern Sinology — mutual speakability is out of the question. Is there a basic difference in the imaginative caliber of the ideas of “Arabic” & “Chinese”?

    The boundary of religion is interesting to ponder. Attention must be paid, since there’s a “religious” (in Western terms?) tinge to nationalism in East Asia.

    Unlike with Maltese & Std Arabic, Taioanese has largely been in diglossia with the Chinese Book Koine & later Mandarin throughout its existence. Another comparand for Maltese — spoken by Theravada Buddhists, no less — would be Kelantan Peranakan “Hokkien”. But KPH is linguistically much more distant from its “continental” cousins.

    I was wondering what everyone here thinks about the idea that “Lebanese” speech is not actually descended from Proto-Arabic as currently formulated. I find it hard to evaluate such claims; I guess this is the opaqueness that plagues both “Arabic” & “Chinese”? Is there anything to Nassim Taleb’s (“No, Levantine is not a ‘dialect of’ Arabic” — really nasty article in places) intriguing description of the Arabic linguistics establishment, with its tendencies toward “overfitting”, and flatly “ignor[ing] developments that are uniquely local”?

  29. KIRINPUTRA said,

    May 18, 2024 @ 10:27 am

    There is this. Interesting comments too.

  30. Lameen said,

    May 19, 2024 @ 1:59 am

    To give Taleb his due, I don't think he's ever knowingly claimed that "“Lebanese” speech is not actually descended from Proto-Arabic as currently formulated" – because neither he nor anyone else outside academic linguistics has a notion of "Proto-Arabic" in his head to begin with. If he stopped fulminating about Phoenician long enough to think seriously about what a Proto-Arabic reconstructed from the contemporary colloquials would look like, he would have to write a very different piece. He shares with most of his enemies the assumption that any "Proto-Arabic" would simply be Standard Arabic as he learnt it at school, which is, of course, false.

  31. Jonathan Smith said,

    May 19, 2024 @ 7:36 pm


    I think a chunk of the answer to your comments 1-4 (below addressed kinda backwardly) is comparative phonology. However else the term "Southern Min" might be used, it’s used by historical phonologists to refer to a group of languages of ≈Southern Fujian which feature lots of basic vocabulary displaying highly regular phonological correspondences such that the languages in question are taken to constitute a phylogenetic clade. So in Bit-Chee Kwok’s book (Southern Min, 2018), a hypothetical ancestral system called "Proto Southern Min" is reconstructed via the comparative method by reference to data from points named "Quánzhōu", "Zhāngzhōu", "Dàtián", "Jiēyáng", "Cháoyáng" and "Léizhōu".

    One could include more or different points (like your Teochew instead of / in addition to Jieyang) in approaching "Proto Southern Min", but terms like "Hoklo", "Hokkien", "Taiwanese", etc., aren’t seen in this literature because they don’t name single, (relatively!) coherent sampling sites.

    Also in light of comparative phonology, Min as a group is IMO highly coherent. A well-known and highly readable paper showing that a language traditionally regarded as Hakka-adjacent was in fact "Min" in this technical, "scientific" :D sense is Norman (1982), "The classification of the Shaowu dialect." One major takeaway from Norman’s early work is that certain languages of northwestern Fujian constitute a coherent subgroup within Min — of late usually called "Inland Min" — which would have split from the rest of the group ("Coastal Min", which above I unscientifically referred to "eastern") at an early stage.

    And finally it’s also largely because comparative phonology that people think that (for example) "Mandarin" and "Taiwanese" are "obviously" related. However TBH I remain a bit hesitant to say "obvious" here because the correspondences at this higher level, while numerous and largely regular, are nonetheless imperfect. For example, it's not at present understood why Mandarin onset /ʂ-/ ("sh-") can correspond to Taiwanese /s-/ (e.g., 'mountain'), /ts-/ (e.g., 'water'), and also /tsʰ-/ (e.g., 'hand'). An example of a study in which the author tries seriously to use phonological correspondences involving everyday words (as opposed to old rime books) to compare Min to the rest of Chinese is Norman (2014), "A Model for Chinese Dialect Evolution." (The results are mixed because Norman tries tricky cases like the above — as do Baxter & Sagart [2014], IMO less successfully.)

    Re: your point 5, I totally agree — actually "mutual intelligibility" is not even a consideration per se on the above comparative approach. Instead the idea is to compare points for which at any given point such intelligibility should be a given but across which it is irrelevant. E.g., while I know / care for personal reasons that Teochew remains damn near impossible to understand given decent knowledge of Taiwanese :(, the "scientist" above need not know / care. S/he only observes that language facts point to recent divergence from a common ancestor and seeks to understand its nature.

  32. KIRINPUTRA said,

    May 21, 2024 @ 6:02 am

    @ Lameen

    Thank you. Taleb does seem to have become vaguely aware of Proto-Arabic (the concept), but threw himself into "proving" that "Classical Arabic is … not an ancestor of Levantine"…. (It would be cool to be able to understand his statistical args completely, though.)

  33. KIRINPUTRA said,

    May 21, 2024 @ 6:41 am

    @ Jonathan Smith

    Thank you for the reply. Some issues bled together and I’d like to tease them back apart afresh.

    1 — The point about mutual intelligibility was (of course) unrelated to our discussion about ling. kinship, the comparative approach, etc.

    2 — I don’t doubt at this p’t that metropolitan Hokkien & Teochew (& thus Taioanese too) spring from one language in the standard sense. The running-standing tone patterns (what Sinologists call “sandhi”) line up too well to have been borrowed. (Why don’t experts talk about this more?)

    3 — Yes, Teochew & Hokkien are related. But what is the evidentiary basis for Mandarin being related to the two? Borrowed cognates follow the same laws as inherited ones. Just b/c sound correspondences work out between languages X, Y, & Z … doesn’t mean they’re kin.

    Just to be clear, this is what I’m asking, in other words: How can we be sure that (1) the basic Chinese etyma in Hoklo were not borrowed, at various points; AND that (2) (modern) Hoklo doesn’t have a Chinese-based creole in its lineage?

    (I’ll have to get a hold of “A Model for Chinese Dialect Evolution”. Thanks.)

    4 — As far as naming, my (previous) question was: How is “Southern Min” any more scientific (or less sloppy) than “Hoklo” as a name for this family of languages?

    5 — What do you think of the claim that Coastal & Inland “Min” don’t reconstruct to any common ancestor shared exclusively by the two?

    6 — If linguistic kinship can be determined solely on the basis of phonological correspondences & lexicon, why isn’t this standard practice for all the languages of the world? (Is it really a coincidence when an expert who claims “Chinese languages are grammatically very similar” speaks Hokdarin or Mantonese, blissfully unaware?)

    (I’m looking at page 143 in Kwok where he says “綠” is a “basic wor[d] and can hardly be borrowed from one language to another”. Even lexicon is a plaything in Neo-Sinology.)

  34. Jonathan Smith said,

    May 21, 2024 @ 9:40 am


    To large degree I share your mindset re: these questions ("show me the money"), but the details suggest that the "standard model" is at least a good starting point —

    You say (under 2) that "sandhi" patterns "line up too well to have been borrowed" across Teochew-Hokkien. This general point is the crux of the more general matter: PATTERNS quickly become too intricate to have been borrowed… like say the geochemistry of stones in the foundation of my house could reveal a common origin with the stones in yours.

    In the case of "Min," Norman noticed most importantly that certain etyma — whose Taiwanese reflections include words like (here using ph- examples and Taiwanese-style tone numbering) phe5 'skin', phang7 'crack (n.)', and phak8 'to dry in the sun' — are involved in intricate correspondence patterns implicating languages of (e.g.) northwestern Fujian province: thus, in Shaowu (citing from Norman 1982 pp. 555ff.), apparent cognates of the above three words have ph- + Tone 7 (phei7 'skin') for Tw. ph- 5 (NOT ph- 2, which = Tw. p- 5 in 'flat', etc.), ph- + Tone 5 (phiung5-ə 'crack [n.]') for Tw. ph- 7 (NOT ph- 6, which = Tw. p- 7 in 'cooked rice', etc.), and ph- + Tone 7 (phu7 'to dry in the sun') for Tw. ph- 8 (NOT ph- 6, which = Tw. p- 8 in 'white', etc.)

    Such a pattern — and there are of course more etyma + other similar patterns — is not something which borrowing of individual items will ever end up imitating, period. Actually, even the three pairs above taken alone would give a faint funky smell of some deep relationship between Shaowu and Taiwanese. Indeed, borrowing/contact on the contrary obscures such patterns over time: languages like Shaowu tend to move items like 'skin' into their Tone 2's, for instance, "where they belong" according to their loud non-Min neighbors.

    Such correspondence patterns are not affected by the fact that, as you state correctly (under 3), "[b]orrowed cognates follow the same [sound change] laws as inherited ones." The existence of regular sound change "laws" is indeed why correspondence patterns are retained to the extent that they are.

    Re: your 5 — yeah Norman ran into problems that led him to wish he had treated "Inland" and "Coastal" Min separately. But, here eliding many details, the above comparanada alone already convincingly unite the two. Part of the field's (such as it is) task at the moment IMO is to try to find better solutions to some of these problems.

    Re: your 3 and your central query — yeah, correspondence patterns are less clear and convincing when we compare Min as a whole, or Taiwanese individually, to languages like Mandarin. We shall see if these can be resolved in future. It is totally fine with me if one prefers, given the state of the field, to regard Min as separate from (the rest of?) Chinese/Sinitic. Just as it is fine if one prefers given the state of the field(s) to regard "Sinitic" as separate from "Tibeto-Burman." I (unlike many?) am super interested in the (numerous!!!) "non-Sinitic" features of Min languages — but on present evidence, the foundation stones of Taiwanese, etc., look to be "Sinitic," not something else… quite unlike say Vietnamese, where work of exactly the same kind as the above led to the realization that the foundation stones were geochemically "Mon-Khmer," despite the lovely Chinese exterior :D

    —–below for the sake of thoroughness

    your 2 – yes so-called "sandhi" (I also hate this term) is an interesting unifying feature of say Teochew and Taiwanese… but also of lots of other languages of the area including most obviously Hokchiu (also some Mien languages!). Probably it is under-talked-about in comparative work because the focus there is on phonological comparison of independent words (see above). For what it's worth, I discussed "the possibility of [a Northern Min]-inclusive account of general Min stress/sandhi" in a paper reconsidering some of the weird features of (northern Fujian) Jianyang, etc., that so troubled Norman, but this needs work…

    your 6 — phonlogical comparison of the above kind is standard practice wherever you look… of course, (esp. inflectional) morphology would arguably be even better for demonstrating historical relatedness, but such is largely absent in the China-MSEA area. You are right though IMO that borrowing can absolutely occur throughout the lexicon — and for langauges of this general kind, borrowing syntax isn't categorically different from borrowing words… syntactic features will thus be of very limited utility for determining deep historical affiliations.

    your 4 — "Hoklo"/"Hokkien" tend to be used to refer to Zhangzhou+Quanzhou and related diasporic/"mixed" varieties (right?). "Southern Min" (1) when used technically doesn't mean some particular language, like "Hoklo"/"Hokkien" often can, but rather a (sub)group of languages, and (2) is bigger than the term "Hoklo"/"Hokkien" suggests, extending west to incorporate not only Chao-Shan but also languages of the Leizhou peninsula / Hainan island. Incidentally, this means that the tendency in some quarters to refer to Taiwanese as "Taiwanese Southern Min" is irregular… like referring to "American West Germanic" or something.

RSS feed for comments on this post

Leave a Comment