Malta speaks the only Semitic language with official status anywhere in Europe. Maltese is a daughter of Siculo-Arabic, written in the Latin alphabet, peppered with Italian and English, and unmistakably a cousin of Arabic and Hebrew rather than of Italian or Greek. A reasonable person might expect the people who speak it to be, in some measure, a Near Eastern or North African transplant on a European island. The genome says otherwise. A modern Maltese sits almost on top of a Sicilian, a short step from a Calabrian, and a long way from any Levantine or Maghrebi population. The deep foundation is the ordinary South Italian one. The single thing that sets the Maltese apart from a Sicilian is a modest North African tint of a few percent. The Semitic tongue is real and it is theirs, but it is a cultural inheritance that the genome never matched.
Key points
- Maltese is the only Semitic language with official status in the European Union, descended from the Siculo-Arabic spoken across medieval Sicily and Ifriqiya. The genome does not follow the language east.
- A modern Maltese is nearest to a Sicilian (about 19 Global25 units), a Calabrian (21) and the wider South Italian average (about 23). They are far from a Lebanese Christian (about 72) and far from a Tunisian Arab (about 133).
- The deep foundation is the standard South Italian one: roughly two thirds early Anatolian farmer, just over a quarter Bronze Age steppe, and a thin Caucasus slice. It is statistically the same as a Sicilian or a Calabrian.
- The one thing that distinguishes the Maltese from a Sicilian is a small North African component, on the order of a few percent of deep Maghrebi (Taforalt) ancestry, or roughly 9 percent read through a modern Maghrebi proxy. This matches the published estimates: under 5 percent Middle Eastern in the Maltese Genome Project, and about 9 to 11 percent Middle Eastern or North African on consumer tests.
- The large eastern Mediterranean component the Maltese share with every Sicilian and Greek (a Canaanite-like slice of about a quarter) is the endemic farmer base of the whole region. It is not a Phoenician fingerprint, and the Maltese carry no more of it than their neighbours do.
- Even Punic Malta was not genetically Levantine. The Punic gene pool of the central Mediterranean was a mixture of Sicilian and Aegean ancestry with a North African slice and almost no Levantine Phoenician ancestry (Ringbauer et al. 2025). The Phoenician network moved culture and language, not people.
- Modern Maltese are not the descendants of the famous temple builders either. The Neolithic population of Gozo sits about 119 units away, farther than a living Sicilian. The islands saw real population turnover.
1. The only Semitic tongue in Europe
Malta is a scatter of small limestone islands in the channel between Sicily and Tunisia, about 80 kilometres south of one and 280 north of the other. That position has made it a crossing point for almost everyone who ever sailed the central Mediterranean. Phoenician traders from the Levant settled it in the early first millennium BCE, and from roughly 480 BCE it fell into the orbit of Carthage, the great Punic power of North Africa. Rome took it in 218 BCE, Byzantium held it after that, and in 870 CE the Aghlabid emirate of Ifriqiya conquered it, opening some two centuries of Muslim rule that continued under the Fatimids. The Norman count Roger I took the islands in 1091, but Arabic speech and a Muslim population persisted for generations afterward, and out of that long Arabic-speaking interval the Maltese language was born. It is a Semitic language, the only one native to Europe, carrying a Siculo-Arabic core under a heavy Sicilian and Italian vocabulary.
The language points clearly south and east. The obvious question is whether the people do too. To answer it we ask the same three things we asked of the Hungarians, whose Uralic language also pointed away from their genome. Who are the Maltese nearest to? What is their deep ancestry made of? And does any part of the genome actually track the Semitic language, or did language and ancestry part ways here as well?
2. Neighbours first: where the Maltese sit
The cleanest way to place a population is to ask what it is nearest to. The chart below gives scaled Global25 Euclidean distances (multiplied by 1000) from the modern Maltese average to a panel of living Mediterranean populations and to the ancient reference points that matter for Malta.
The order tells the whole story before any modelling. The Maltese sit among the people of the central Mediterranean: Sicilians first, then Calabrians and the rest of the South Italian mainland, then the Greeks of southern Italy and Crete. The Levantine populations that share the Maltese language family are between three and five times farther away: a Lebanese Christian at 72, an Iron Age Phoenician from the Levant at 86, a Bronze Age Canaanite from Sidon at 94. The North African Arabs whose medieval ancestors helped carry Arabic to the islands sit farther still, the Tunisians at 133 and the Moroccans at 183. And the Neolithic temple builders of Gozo, the islands' own deep past, lie at 119, farther from a modern Maltese than a living Sicilian is. Wherever the language came from, the autosomes are central Mediterranean.
3. The foundation is South Italian
Strip the genome down to the three or four deep ingredients every European carries in some proportion and the Maltese resolve into the ordinary South Italian recipe. Modelled against distal ancient sources (early Anatolian farmer, Bronze Age steppe, a Caucasus hunter-gatherer slice, and a trace of Western hunter-gatherer), a Maltese is indistinguishable from a Sicilian or a Calabrian.
Deep foundation: Maltese and their neighbours share one recipe
All four resolve to roughly two thirds early farmer, a quarter to a third steppe, and a thin Caucasus slice. This is the common South Italian foundation, in place since the Bronze Age. On this base the Maltese carry nothing their neighbours lack. Figures are proxy-dependent; read them as directions, not exact percentages.
4. The one real difference: a North African tint
If the foundation is shared, what makes the Maltese sit just slightly off their Sicilian neighbours on a close-up plot? The answer is a small North African component, and it follows a clean geographic logic. When we offer a deep Maghrebi source (Taforalt, the pre-Neolithic North African reference that predates any European farming ancestry) alongside the usual ingredients, the weight it draws rises steadily as you move south through Italy and peaks in Malta.
Deep Maghrebi (Taforalt) signal rises toward the south and peaks in Malta
The deep Maghrebi share is essentially nil in central Italy, around 2 percent in Sicily, and a few percent in Malta. Read through a modern Maghrebi proxy (Tunisian Arab) rather than the deep Taforalt pole, the same excess comes out at roughly 9 percent. Either way it is a minority tint on a South Italian genome, and it is the only feature that separates a Maltese from a Sicilian.
This is exactly the range the published work has found by other means. The Maltese Genome Project reports that most Maltese trace their ancestry to Sicily and southern Italy of about a thousand years ago, with Middle Eastern ancestry, including Levantine, contributing under 5 percent. Consumer genotyping platforms place the typical Maltese at around 88 percent Southern European with some 9 to 11 percent Middle Eastern or North African. Our Global25 estimate of a few percent deep North African, rising to about 9 percent on a modern proxy, sits squarely inside that bracket.
5. The Levantine mirage
Here a careful reader will object. When the Maltese are modelled with a Bronze Age Levantine source such as the Canaanites of Sidon, that source takes a large share, around a quarter of the genome. Does that not make the Maltese substantially Levantine after all, exactly as the language suggests?
It does not, and the reason is important. That Canaanite-like component is not a Maltese peculiarity. It is the shared eastern Mediterranean farmer base of the entire region, and a Sicilian, a Calabrian and a mainland Greek all carry the same amount of it. It is a deep ingredient that arrived with the spread of farming and the Bronze Age, not a recent Phoenician or Arab contribution. Reading it as a sign of Near Eastern descent would force you to call every Sicilian and Greek a Levantine too.
6. Even Punic Malta was not Levantine
Malta spent more than two centuries inside the Punic world of Carthage, and the Phoenician founders of that world came from the Levant. It would be natural to assume that this is where any Near Eastern ancestry entered. The largest ancient DNA study of the Phoenician and Punic Mediterranean to date, published in 2025, shows that the assumption is wrong. Across fourteen Phoenician and Punic sites from the Levant to Iberia, the Levantine Phoenician homeland made almost no detectable genetic contribution to the Punic settlements of the central and western Mediterranean. The Punic people were a genetically diverse population, and their main ingredients were local: Sicilian and Aegean ancestry, with a North African component, and very little of the Levant.
Our own decomposition of the Punic dead reproduces that picture. The Punic populations of Sicily and Tunisia resolve overwhelmingly into a Sicilian or Aegean Greek-like base plus a North African slice, with the Levantine source falling to zero.
What the Punic gene pool is made of
The Levantine source is not needed at all: both Punic groups model as a central Mediterranean base plus a North African slice. The Phoenician network spread an alphabet, a religion and a language across the sea, but it moved remarkably few people. Punic Malta, like Punic Sicily, was central Mediterranean with a North African tint, not a Levantine colony in blood.
7. Where the language actually came from
If neither the deep foundation nor the Punic centuries made the Maltese Levantine, where did the Semitic language come from? From the Arabic-speaking Mediterranean of the ninth to twelfth centuries, carried by people from Ifriqiya and Muslim Sicily. We can see what that population looked like in the medieval Muslim dead of Segesta in western Sicily, contemporaries of the world that gave Malta its tongue. They are a genuinely admixed group: a Sicilian and Aegean base, a substantial North African slice, and, this time, a real Levantine or Near Eastern component.
Medieval Muslim Sicily (Segesta): the kind of population that carried Arabic
This is the demographic signature of the Islamic Mediterranean: a local European majority with a real North African and Near Eastern minority. A population like this brought Siculo-Arabic to Malta. But the Maltese absorbed only a few percent of that incoming ancestry, even as they took the language whole. As with the Magyars in Hungary, the speech survived in full and the genes were diluted to a trace.
The parallel with Hungary is almost exact. A confident, culturally dominant minority arrives, imposes or seeds a language that takes permanent root, and then marries into a much larger resident population until its distinctive ancestry is folded down to a few percent. In Hungary the incoming language was Uralic and the residual eastern signal sits near the noise floor. In Malta the incoming language was Semitic and the residual North African signal is a small but real few percent. In both cases the tongue is the part of the inheritance that survived intact.
8. So, does the genome match the language?
No, and that is the whole interest of the case. The Maltese speak the only Semitic language in Europe, and on the strength of that language alone you might guess them to be a Levantine or North African people perched on a European island. The genome refuses the guess. A modern Maltese is a South Italian, nearest to a Sicilian, built on the same farmer and steppe foundation as every Sicilian and Calabrian, distinguished only by a modest North African tint of a few percent. The Levantine-looking component they carry is the shared eastern Mediterranean farmer base of the whole region, not a Phoenician inheritance, and even the Punic centuries left them central Mediterranean rather than Levantine. The Semitic language is genuinely theirs, a living relic of the Arabic-speaking medieval Mediterranean. It is just not written in their DNA. Malta is the place where a Semitic tongue and a South Italian genome have lived together, intact and unmatched, for a thousand years.
The story in six steps
Claim and reality
Because the Maltese speak a Semitic language, they must be largely a Near Eastern or North African people.
Genome-wide they are South Italians, nearest to Sicilians (about 19 units) and far from Lebanese (72) or Tunisian Arabs (133). The language points east; the genome does not.
The large Canaanite-like component in Maltese ancestry proves a Phoenician or Levantine origin.
That component is the shared eastern Mediterranean farmer base of the whole region. Sicilians and Greeks carry the same share. It is not a Phoenician fingerprint.
The Phoenician and Punic period seeded Malta with Levantine ancestry.
The Punic gene pool of the central Mediterranean was Sicilian and Aegean with a North African slice and almost no Levantine ancestry (Ringbauer et al. 2025). The network moved culture, not people.
Modern Maltese descend directly from the prehistoric temple builders of the islands.
The Neolithic population of Gozo sits about 119 units away, farther than a living Sicilian. The deep islanders were largely replaced; the modern population is a later, South Italian one.
Reproduce it yourself
Paste the coordinates below into Vahaduo (the Global25 tool) to rebuild the comparisons in this article: the modern Maltese average and its Mediterranean neighbours, the deep source poles (Anatolian farmer as Barcin, steppe as Yamnaya, Caucasus as Satsurblia, Western hunter-gatherer as Villabruna, North Africa as Taforalt), the Levantine references (Sidon Canaanite and the Iron Age Phoenicians of Akhziv), the Punic and medieval Muslim Sicilian dead, and the Neolithic islanders of Gozo. Modern averages are scaled Global25 from Davidski; the ancient averages were computed here from the individual coordinates, with the Punic Sicily and Akhziv Phoenician averages drawn from the published individual data.
Maltese_(n=8),0.091912,0.146490,-0.002498,-0.043403,0.013271,-0.017814,-0.004406,-0.000923,0.011095,0.020524,0.006150,0.000974,-0.002899,-0.000723,-0.004258,0.000878,-0.000407,-0.003769,-0.000487,-0.002361,-0.006145,-0.002597,0.002819,0.001190,0.002904 Sicilian_(n=6),0.101682,0.146067,-0.003834,-0.038329,0.018619,-0.012597,-0.001293,0.000308,0.010465,0.021808,0.005359,0.003497,-0.004187,0.002500,-0.004094,-0.004840,-0.002369,-0.000591,0.001802,-0.005003,0.000395,0.001752,-0.002547,-0.000603,0.000319 Italian_Calabria_(n=8),0.103437,0.148140,-0.010324,-0.046996,0.012618,-0.015025,-0.002438,-0.001587,0.004448,0.021230,0.002213,0.002997,-0.004887,-0.001737,-0.006548,-0.002784,0.005802,0.000728,0.002938,-0.005362,-0.002605,-0.000325,0.000339,0.000015,-0.002425 Italian_Apulia_(n=15),0.110712,0.148267,-0.003771,-0.040138,0.015839,-0.015116,-0.000956,-0.001615,0.002795,0.020690,-0.000227,0.003876,-0.006472,0.001193,-0.008107,-0.003898,0.001669,0.000600,0.002749,-0.005152,-0.003810,0.001962,0.001758,0.001438,-0.000567 Greek_South_Italy_Apulia_(n=10),0.109384,0.149384,-0.001622,-0.040989,0.011325,-0.014084,-0.001105,0.000739,-0.000184,0.019700,0.001624,0.004121,-0.006838,0.000867,-0.011387,-0.000557,0.006689,-0.000849,0.002262,-0.004002,-0.005939,0.000927,0.000530,0.003567,0.001138 Tunisian_Arab_(n=30),-0.022044,0.135641,-0.013337,-0.068476,0.018085,-0.027815,-0.018535,0.003923,0.047409,0.023108,0.005364,-0.006239,0.019078,-0.010207,0.010116,-0.006241,0.000022,-0.012859,-0.026874,0.006061,-0.008963,-0.021351,0.014810,-0.002386,0.002686 Greek_Crete_(n=12),0.107373,0.146490,-0.017096,-0.047320,0.005950,-0.016292,0.002761,-0.001558,-0.003000,0.012908,0.002016,0.002610,-0.000942,0.004633,-0.011649,-0.003613,0.001945,0.000992,0.003226,-0.005148,-0.004919,0.000886,-0.000493,-0.001466,-0.001876 Lebanese_Christian_(n=25),0.090284,0.144896,-0.051862,-0.080117,-0.014279,-0.027766,-0.000865,-0.004735,0.003600,0.009221,0.005268,-0.004748,0.010608,0.001063,-0.008675,0.006794,-0.001914,0.001464,0.003057,-0.000400,-0.000499,0.001103,-0.001839,0.002612,0.000263 Sardinian_(n=10),0.121335,0.167766,0.028133,-0.051002,0.060380,-0.022228,-0.003925,0.002169,0.041907,0.077560,-0.000373,0.016530,-0.028454,-0.012413,-0.014251,-0.004362,0.010431,-0.001064,0.002250,-0.013406,-0.002146,-0.001175,-0.010230,-0.021039,0.001114 Malta_Gozo_Neolithic_avg_(n=4),0.126344,0.178226,0.015651,-0.070818,0.065166,-0.034722,-0.005111,-0.003923,0.046631,0.084557,0.001177,0.014275,-0.026425,-0.002787,-0.027925,-0.003911,0.023730,0.002629,0.009773,-0.011787,-0.010575,0.004730,-0.007796,-0.008797,0.000359 Anatolia_N_Barcin_avg_(n=22),0.118842,0.181041,0.003548,-0.100835,0.051744,-0.046397,-0.005191,-0.007321,0.036861,0.080971,0.009190,0.012064,-0.023279,0.000982,-0.041820,-0.009106,0.021590,0.000697,0.011753,-0.009470,-0.013102,0.006750,-0.004667,-0.003451,-0.005465 WHG_Villabruna,0.121791,0.114755,0.185920,0.184111,0.156337,0.060798,0.020211,0.035998,0.092445,0.018041,-0.016239,-0.016186,0.016947,-0.010046,0.054017,0.067356,0.000782,0.005448,-0.008422,0.053526,0.100073,0.010758,-0.048313,-0.163517,0.019280 Yamnaya_Hungary_EBA_avg_(n=6),0.122929,0.093598,0.040038,0.106321,-0.022414,0.043368,0.004073,-0.002846,-0.053994,-0.068885,0.001597,0.000799,0.000248,-0.017753,0.034066,0.008618,-0.011995,-0.000485,-0.000084,0.004502,-0.001601,-0.000289,0.010065,0.016368,-0.007265 CHG_Satsurblia,0.092197,0.101553,-0.093526,-0.000969,-0.092633,0.020917,0.030786,-0.001615,-0.139281,-0.085833,-0.002923,0.024278,-0.058424,0.009634,0.036373,-0.022938,0.044591,-0.008488,-0.027025,0.042896,0.045170,-0.009521,0.001849,-0.030245,-0.002515 Taforalt_Maghreb_UP_avg_(n=6),-0.190085,0.082258,-0.024261,-0.084626,0.026620,-0.056057,-0.070464,0.020038,0.156324,0.003493,0.019866,-0.032521,0.075470,-0.051494,0.072248,-0.038849,0.003716,-0.065435,-0.144134,0.037852,-0.038474,-0.124951,0.071894,-0.013817,0.016186 Levant_MBA_Sidon_Canaanite_avg_(n=5),0.081270,0.146845,-0.061320,-0.097934,-0.010525,-0.038933,-0.004794,-0.006600,0.011453,0.010169,0.010490,-0.009771,0.021585,0.005505,-0.007573,0.006523,-0.000156,-0.001191,0.001509,0.004252,0.004267,0.007543,-0.001257,0.001591,-0.000503 Levant_IA_Phoenician_Akhziv_avg_(n=11),0.081746,0.150391,-0.052625,-0.097282,-0.009261,-0.036636,-0.002350,-0.009293,0.013778,0.012939,0.010201,-0.010981,0.018812,0.001964,-0.004725,0.002700,0.001529,0.002557,0.003348,0.004377,0.002019,0.002518,-0.001120,-0.000548,-0.001796 Sicily_Greek_Himera_480BCE_avg_(n=7),0.118701,0.162049,-0.005980,-0.061462,0.025279,-0.025738,-0.002384,-0.002802,0.006633,0.043815,0.004083,0.009977,-0.018540,-0.003578,-0.018574,-0.009262,0.011995,-0.000398,0.007775,-0.008647,-0.010250,0.005317,-0.000916,0.006025,-0.004704 Sicily_Punic_Marsala_avg_(n=14),0.101953,0.154941,0.005657,-0.052580,0.032314,-0.022570,-0.005304,-0.003264,0.025332,0.037684,0.004280,0.002837,-0.008888,-0.006429,-0.007125,-0.002102,0.006137,-0.000090,-0.001077,-0.002269,-0.004786,-0.001616,-0.000088,-0.005474,-0.000855 Tunisia_Punic_Kerkouane_avg_(n=7),0.073985,0.155376,0.010829,-0.052326,0.041062,-0.022630,-0.009669,-0.002769,0.042219,0.043164,0.004338,0.001542,-0.006265,-0.010420,-0.003335,-0.007103,0.007059,-0.005357,-0.013001,-0.002162,-0.005330,-0.008232,0.003011,-0.002685,0.000274 Sicily_Muslim_Medieval_Segesta_avg_(n=5),0.056001,0.137096,-0.011314,-0.045285,0.013972,-0.019076,-0.009823,-0.001800,0.015421,0.017786,0.002696,0.002428,0.001159,-0.002862,-0.002443,0.001936,0.006285,-0.004054,-0.005883,0.001376,-0.007736,-0.004575,0.004634,-0.003013,0.001820
References and sources
- 1 Ringbauer, H., Salman-Minkov, A., Regev, D., Olalde, I., et al. (2025). Punic people were genetically diverse with almost no Levantine ancestors. Nature 643, 139 to 147. link
- 2 Ariano, B., Mattiangeli, V., Breslin, E. M., et al. (2022). Ancient Maltese genomes and the genetic geography of Neolithic Europe. Current Biology 32, 2668 to 2680. The source of the Neolithic Gozo islanders and the evidence for their isolation. link
- 3 Fernandes, D. M., Mittnik, A., Olalde, I., et al. (2020). The spread of steppe and Iranian-related ancestry in the islands of the western Mediterranean. Nature Ecology and Evolution 4, 334 to 345. Context for Punic-era ancestry in the central and western Mediterranean. link
- 4 Sarno, S., Boattini, A., Carta, M., et al. (2014). An ancient Mediterranean melting pot: investigating the uniparental genetic structure and population history of Sicily and Southern Italy. PLoS ONE 9, e96074. Background on the shared Sicilian and South Italian genetic landscape into which the Maltese fall. link
- 5 Maltese Genome Project (A. E. Felice and colleagues), University of Malta. Uniparental and genome-wide work reporting Sicilian and South Italian ancestry for most Maltese, with Middle Eastern contribution under 5 percent. project
- 6 Global25 coordinates: Davidski (Eurogenes), with modern and ancient averages drawn from the Moriopoulos 2025 collection. Global25 spreadsheet tooling: Vahaduo. G25
Modern and ancient Global25 coordinates: Davidski (Global25) and the Moriopoulos 2025 collection. Maltese, Sicilian, Calabrian, Apulian and South Italian Greek averages computed here from the individual coordinates; Punic Sicily (Marsala) and Iron Age Phoenician (Akhziv) averages computed from the published individual data; Tunisia Punic, Sidon Canaanite, Taforalt, Anatolia Neolithic, Yamnaya, Satsurblia, Villabruna, Muslim Sicily and Neolithic Gozo from the collection averages. Global25 spreadsheet tooling: Vahaduo. Analysis: scaled Global25 Euclidean distances and non-negative least squares modelling in Python. Ancestry fractions are proxy-dependent and best read as directions rather than exact percentages.
