Does the name Córdoba come from the word "Kurd"? What do the pre-Islamic records say?
The subject of this article: the city of Córdoba in Spain. In Islamic history the city is known by its Arabic name, Qurṭuba (قرطبة), and it served as the capital of Umayyad al-Andalus. The article answers two questions:
Is this name related to the word "Kurd"?
Was the name already in use for this city before Islam, that is, before the Muslim conquest of 711?
SUMMARY
The answer to the first question: no. The name "Córdoba" has no linguistic connection whatsoever to the word "Kurd". Not a single specialist in the academic literature has ever proposed such a link. The claim rests on nothing but a resemblance of sound, which makes it a textbook folk etymology.
The answer to the second question: yes, the name was there long before Islam. Bronze coins struck at the mint of Corduba in the mid-second century BC (168-151) already carry the legend CORDVBA. Julius Caesar mentions the city as "Cordubae" and "Cordubam" in his account of the civil war of 49-45 BC. Strabo (c. 7 BC) writes Κόρδυβα; Pliny the Elder (AD 77) writes "Corduba colonia Patricia cognomine". The name is on written record roughly 800 years before the Muslim conquest of 711.
The internal structure of the name confirms this. "Corduba" is not a single block: Cord + uba. The ending "-uba/-oba" recurs throughout the pre-Roman place names of southern Iberia (Tartessian/Turdetanian country): Onuba (today Huelva), Ossonoba (today Faro), Salduba, Maenuba. To see "Kurd" in the name, you have to cut the word in the wrong place.
The strongest etymology available today (Robert Knapp 1983, revived by Jesús Rodríguez Ramos in 2024): Certis, the indigenous name of the Guadalquivir river recorded by Livy, plus -uba = "city on the river". Certuba > Corduba.
The Arabic side settles the matter as well. Qurṭuba (قرطبة) is built on the root consonants q-r-ṭ-b; "Kurd" (كرد) on k-r-d. They share not a single root consonant; even the first letters differ (qāf versus kāf). The resemblance arises only in Turkish spelling, because Turkish renders both Arabic letters with the same "k". It is an accident of the alphabet, not a historical connection.
Finally: in al-Andalus (711-1492) there is no documented Kurdish population, no Kurdish military unit, and no Andalusi individual known in the primary sources by the nisba "al-Kurdī".
1. TWO QUESTIONS, TWO ANSWERS
The two questions are distinct, but their answers reinforce each other:
Was the name in use before Islam? Definitely yes; the evidence is abundant and first-hand.
Is the name related to the word "Kurd"? No; there is no evidence at all.
In fact, the answer to the first question largely resolves the second. The only historical window through which Kurds could conceivably have reached the Iberian Peninsula is the Islamic conquest of 711. If the name is on written record eight centuries before that date, the window closed before it ever opened.
2. THE PRE-ISLAMIC RECORD
In chronological order, from oldest to most recent:
In addition, three celebrated Roman authors were born at Corduba: Seneca the Elder, Seneca the Younger and the poet Lucan, and Roman writers record them as such. After 45 BC the city took the name Colonia Patricia Corduba and became the capital of the province of Baetica.
The coin evidence matters most. According to the catalogues of Hispanic mints, the small bronze denominations struck in the mid-second century BC preserve the indigenous toponym: CORDVBA. In other words, the Romans did not invent the name; they Latinised a name they found already in place. The name is older than Rome itself.
3. THE SETTLEMENT IS OLDER THAN ROME
On the hill known as the Colina de los Quemados ("Hill of the Burnt Ones") in the Parque Cruz Conde area of present-day Córdoba, archaeologists have excavated a Tartessian/Turdetanian settlement reaching back to the Bronze Age. The excavations were carried out in 1964 under Rafael Castejón by Juan Bernier and Javier Fortea, and in 1966 by José María Luzón and Diego Ruiz Mata. In the fifth and fourth centuries BC the site was one of the regional distribution centres for Attic pottery.
Strabo reports that the proconsul Marcus Claudius Marcellus "founded" the city around 169-168 BC (152-151 BC according to some historians). Modern archaeology corrects the picture: Marcellus did not build a city on empty ground; he founded the Roman colony roughly 750 metres from the existing native settlement, and the two coexisted for a time (Ruiz Bueno and Rubio Valverde, Antiquitas 34, 2022). Antonio Monterroso Checa of the University of Córdoba goes further, arguing that the "foundation" was in fact a symbolic refoundation of the existing Tartessian settlement (2026; a recent and still debated proposal).
There is a counter-view as well
Alicia M. Canto (Latomus 50/4, 1991) argues that the word ktisma in Strabo means "urban embellishment" rather than a foundation from scratch, and that the Marcellus in question may have been Augustus's nephew (d. 23 BC). This debate concerns the date of the foundation, not the age of the name: whichever view is correct, the coins already read CORDVBA in the second century BC.
4. INSIDE THE NAME: WHY IT IS NOT "KURD"
This is the heart of the matter.
"Corduba" is not an indivisible word. Its ending "-uba" (elsewhere "-oba") is a recurring element in the pre-Roman place names of southern Iberia:
Onuba : today Huelva
Ossonoba : today Faro, in Portugal
Salduba : in Baetica; its exact site is uncertain, usually placed near Marbella
Maenuba : near Málaga, together with a river of the same name
Corduba : Córdoba
All of these places sit on the sea, on rivers or on estuaries. That is why philologists assign the element "-uba" the meaning "water" or "settlement, town".
A simple analogy
The words "Kurdistan", "Pakistan" and "Hindustan" share the ending "-stan", meaning "land, country". Nobody argues that "the stan in Pakistan is the same as the stan in Kurdistan, therefore Pakistan is Kurdish in origin". A suffix is a suffix; the meaning lives in the root.
In Corduba the suffix is "-uba", which leaves "Cord-". And for that root, the strongest explanation is this:
The Knapp-Rodríguez Ramos thesis (the best explanation available today)
The indigenous, pre-Roman name of the Guadalquivir was Certis. Livy records it (Ab Urbe Condita 28.22: "Baetim fluvium, quem incolae Certim appellant", that is, "the river Baetis, which the locals call Certis"). Certis + uba = "city of the Certis", in other words "city on the river". The form Certuba became Corduba as it was Latinised.
Robert Knapp first proposed this in 1983 (Roman Córdoba); for forty years hardly anyone noticed. In 2024 Jesús Rodríguez Ramos of the University of Barcelona revived and defended the thesis in the journal Antiquitas ("Una etimología para Córdoba, Tartessos, Taršiš y algunas cuestiones de lengua turdetana"), calling it "muy plausible y la mejor existente, con gran diferencia" (very plausible, and by far the best available).
The serious rival proposals likewise explain "Cord-" from an indigenous root. Francisco Villar suggests the pre-Indo-European or Indo-European element kar-/kort- "rock, stone"; on that reading the name means something like "rocky place by the river".
No proposal contains "Kurd", and that is no accident. To get "Kurd" into the name, one would have to divide the word as Kurd + uba, then explain where the vowel of the first syllable (o/u) came from and why every neighbouring town carries the same ending. No linguist has done this, because it cannot be done.
5. THE PHOENICIAN THEORY: POPULAR BUT WEAK
Another frequently heard theory derives the name from Phoenician/Punic Qart-Ṭūba, "good town". It was first proposed in 1799 by José Antonio Conde and still circulates in tourist guides and blog posts.
Modern specialists do not accept it. The formula of C. F. Seybold and Manuel Ocaña Jiménez is quoted again and again in the literature: the name is "certainly not Semitic but Old Iberian".
There is also a "Kart-Juba" variant: the claim that the Carthaginian commander Hamilcar Barca named the city around 230 BC in honour of a Numidian general called Juba. This, too, has no primary source behind it.
Why does this matter? Because folk etymology always works the same way: a similar-sounding word is found, a meaning is attached to it, and a historical narrative is invented to prop it up. "Qart-Tuba" and "Kurd-uba" are two versions of the same mistake.
The Spanish Wikipedia sums it up honestly: "El significado etimológico del nombre de la ciudad ha sido largamente discutido en la historiografía y no existe en la actualidad consenso al respecto" (the etymological meaning of the city's name has long been debated in the historiography, and there is at present no consensus). No consensus; but on no side of the disagreement do Kurds appear.
6. THE ARABIC SIDE: THE ROOTS DO NOT MATCH
Arabic builds its words from root consonants. Root consonants are like the foundation stones of a house: the paint may look similar, but if the foundations differ, the houses are different.
Not a single root consonant is shared. Even the first letters differ: one is qāf (ق), the other kāf (ك). To an Arabic speaker, the two words do not evoke each other at all.
So where does the resemblance come from? From Turkish spelling
The Turkish alphabet has a single letter "k", which renders both Arabic qāf and kāf. As a result, Arabic Qurṭuba is written "Kurtuba" in Turkish and looks, to the eye, like a relative of "Kürt" (Kurd). In Spanish ("Córdoba") and in Arabic (قرطبة) the illusion does not exist. It is born of the Turkish alphabet.
A NOTE ON TURKISH SOURCES. Turkish literature on Islamic history almost always writes the city as "Kurtuba" (for example the TDV Encyclopaedia of Islam, in its "Kurtuba" entry). The spelling is not wrong, but it has a side effect: it nudges Turkish readers toward the Kürt/Kurtuba association. One look at the original forms (Latin Corduba, Arabic قرطبة, Spanish Córdoba) dissolves the association instantly.
7. WHY DOES "CORD-" SOUND FAMILIAR?
The intuition is not empty; it is simply knocking on the wrong door. Kurdish history does contain ancient names with the sound "Kord-/Kard-":
So the sound "Cord-" genuinely exists in the geography of Kurdistan. But it does not travel to the Corduba of Spain. Similar sounds in two different places are no proof of linguistic kinship. Turkish "kar" (snow) and English "car" are pronounced almost identically; they have nothing to do with each other.
8. EVEN THE EQUATION "KARDUCHOI = KURDS" IS CONTESTED
Honesty is required here. This is one of the most contested files in the linguistics of Kurdish history. All three positions belong on the same table.
THESIS: the continuist reading
G. R. Driver's 1923 article "The Name Kurd and its Philological Connexions" (Journal of the Royal Asiatic Society 55/3, pp. 393-403) builds a chain linking Sumerian Karda/Qarda, Assyrian Qurṭie, Greek Karduchoi and the word "Kurd". Driver himself notes that the link is "uncertain" because of doubts about the exact value of Sumerian palatals and dentals.
Mehrdad Izady, in The Kurds: A Concise Handbook (1992), continues this line: he treats the Kurti/Qurṭie record of c. 1100 BC as "the oldest record of the Kurdish name", positions the Karduchoi as one of the populations later absorbed into the Kurdish ethnic pool, and proposes a layered continuity model reaching back to the Halaf and Hurrian strata.
ANTITHESIS: the philological minimalist reading
Theodor Nöldeke rejected the Karduchoi link as early as the end of the nineteenth century. In his view, "Kurd" derives from an older form kurt, which appears in Greek as Κύρτιοι (Kyrtioi): "Kurd ist also aus einer älteren Form kurt entstanden, die uns in der griechischen Form κύρτιοι begegnet."
D. N. MacKenzie, "The Origins of Kurdish" (Transactions of the Philological Society 60, 1961, pp. 68-86), calls the Karduchoi identification the perennial "favourite choice" but shows that it rests on nothing more than a superficial resemblance of sound, and argues that the only clear references to Kurds in the classical sources before our era are the Kyrtioi of Polybius, Livy and Strabo.
Garnik Asatrian, "Prolegomena to the Study of the Kurds" (Iran and the Caucasus 13, 2009, pp. 1-58), declares the link "obsolete" (p. 25) and argues that in pre-Islamic Pahlavi texts the word kurt was not yet an ethnonym but a social label meaning "nomad, tent-dweller, shepherd" (pp. 23-24).
The Encyclopaedia Iranica entry "Gordyene" (Michał Marciak) is blunt: the ancient root Qardū "is not akin to the root Kurd-. Therefore, ancient Gordyene and its people cannot be held as ancestors of the modern Kurds."
SYNTHESIS
Three things can be true at once, and most likely are.
First, Asatrian is not a neutral referee. He is the editor of the journal Iran and the Caucasus, and his polemical stance toward Kurdish studies is well documented; he goes so far as to call Izady's Handbook "phantasmagoric". That does not refute his arguments, but it does mean they should not be taken as the only voice.
Second, Izady is weak on this particular point. His continuity model is highly ambitious, and its linguistic support rarely goes beyond resemblance of sound. Izady's real strength lies in the geography, demography and historical periodisation of Kurdistan, not in pure etymology.
Third, and decisive for this article: whichever side of the debate one takes, the conclusion does not change. Neither Driver, nor Izady, nor Nöldeke, nor MacKenzie, nor Asatrian, no one, proposes any link reaching the Iberian Peninsula. Even the most generous version of the continuist reading never leaves the Zagros-Taurus line, Upper Mesopotamia and the Caucasus.
9. WERE THERE KURDS IN AL-ANDALUS?
This is a separate and legitimate question. The answer: there is no documented Kurdish presence.
The conquering army of 711 consisted of Berbers under Tariq ibn Ziyad and the Arab reinforcements of Musa ibn Nusayr. The Syrian jund of Balj ibn Bishr, which arrived in 741 and became the "Syrian" stratum of al-Andalus, was drawn from the five military districts of Syria: Damascus, Jordan, Qinnasrin, Homs and Palestine, plus Egypt. The Jazira (Upper Mesopotamia), the region with the dense Kurdish population, is absent from this list.
The standard ethnic taxonomy of Andalusi society runs: Arabs (Baladis and Syrians), Berbers, Muwallads (converted natives), Saqaliba (European slave soldiers), Mozarabs (native Christians) and Jews. Kurds appear in none of these lists.
The striking detail is this: even the "Kurtuba" entry of the Turkish TDV Encyclopaedia of Islam, when it enumerates the city's population (Arabs, Franks, Slavs, Syrians, Berbers, Greeks, Goths, Hispano-Romans and Blacks of West African origin), does not mention Kurds. The absence of Kurds is therefore not the preference of any partisan source; it is a consistent gap across all source families.
Nor can any Andalusi individual bearing the nisba "al-Kurdī" be found in the primary sources.
10. THE ZIRYAB FILE: THE CLAIM THAT ACTUALLY CIRCULATES
This section is an appendix, because there is a Turkish internet narrative far more widespread than the claim about the city's name: "Ziryab was a Kurd."
Ziryab (Abu l-Hasan Ali ibn Nafi, d. 857) was the musician and cultural figure who moved from Baghdad to Córdoba, credited with bringing the oud, new fashions and refined table manners to al-Andalus.
Turkish popular publications routinely call him "Ziryab the Kurd". Their evidence is thin:
An unattributed assertion that "most English and Spanish sources say he was Kurdish".
A rival poet's satirical jibe (Abu Dulama's reference to "your Kurdish ancestors") treated as genealogical proof. An insult in a poem is not an identity document.
The trail of the "Kurdish" claim on English Wikipedia leads back to a single source: John Gill, Andalucía: A Cultural History (Oxford, 2008), p. 81. That is a popular travel-culture book, not an academic study.
The serious scholarship says something else. Dwight Reynolds, "Al-Maqqarī's Ziryāb: The Making of a Myth" (2008), shows that the medieval primary sources (Ibn Abd Rabbih, Ibn Hayyan) describe him as black or dark-skinned (aswad), and argues that the "Persian" pedigree is a later back-formation. The TDV Encyclopaedia's "Ziryâb" entry likewise offers only two options: Persian (weak, name-based) or African (stronger, based on physical description). A Kurdish option does not appear. No medieval Arabic biographical dictionary gives him the nisba "al-Kurdī".
In short: the question about al-Andalus genuinely worth debating is not the name of the city but the identity of Ziryab. And there, too, the Kurdish thesis is at present unsupported.
11. A NOTE ON METHOD: HOW TO TEST CLAIMS OF THIS KIND
Whenever a claim of the type "place name X comes from people Y" appears, three steps suffice:
Find the earliest written form. Not today's Turkish or English spelling, but the original. For Córdoba, that means the coins of the second century BC.
Cut the word in the right place. In place names the suffix usually sits at the end. Seeing "Kurd" in Corduba without first removing "-uba" is like seeing "Kurdistan" inside "Pakistan".
Check the time window. When was the alleged contact possible? If the name predates the window, the claim collapses at the outset.
12. WHAT WE DO NOT KNOW
For honesty's sake, these are the limits of this article.
The exact meaning of the element "Cord-" is unknown. The Certis thesis is the strongest available, but it is not certain, because the Turdetanian language remains insufficiently deciphered. Rodríguez Ramos himself claims no certainty.
The exact meaning of "-uba" is debated: "water", "town", or a more general place-name suffix. The existence of the element is not in dispute; its meaning is.
No pre-Roman coin or inscription in Iberian or Punic script carrying the name has been found at Córdoba. This is a gap; but numismatic scholarship accepts that the Latin coins preserve the indigenous name.
The date of Marcellus's foundation (169-168 BC, 152-151 BC, or an Augustan-era embellishment) remains debated.
None of these uncertainties opens a door to the "Kurd" thesis. The entire debate plays out within the indigenous languages of the Iberian Peninsula.
CONCLUSION
The name Córdoba is unrelated to the word "Kurd", and it is on written record roughly eight centuries before Islam. Its root belongs to the pre-Roman languages of Iberia itself, most likely to the indigenous name of the Guadalquivir. The resemblance produced by the Turkish spelling "Kurtuba" is an accident of the alphabet, not a historical connection.
BIBLIOGRAPHY
Ancient sources
Titus Livius, Ab Urbe Condita 28.22 ("Baetim fluvium, quem incolae Certim appellant").
Xenophon, Anabasis 3.5.15-17; 4.1.3-4.3.2 (the Karduchoi).
Polybius, Histories 5.52 (the Kyrtioi).
Julius Caesar, De Bello Civili 2.19; 2.21 (Cordubae, Cordubam).
Strabo, Geography 3.2.1 (Κόρδυβα); 11.13.3 and 15.3.1 (Kyrtioi); 16.1.24-25 (Gordyene).
Pliny the Elder, Naturalis Historia 3.10 ("Corduba colonia Patricia cognomine").
Claudius Ptolemy, Geography 2.4 (Baetica) and 5.12.2 (Gordyene).
Etymology and archaeology of Córdoba
Knapp, Robert C. (1983), Roman Córdoba, University of California Press. (First version of the Certis + -uba thesis.)
Rodríguez Ramos, Jesús (2024), "Una etimología para Córdoba, Tartessos, Taršiš y algunas cuestiones de lengua turdetana", Antiquitas 36. Dialnet record: dialnet.unirioja.es/servlet/articulo?codigo=9262972
Ruiz Bueno, Manuel D. and Rubio Valverde, Manuel (2022), Antiquitas 34. (On the relationship between the Roman colony and the native settlement.)
Canto, Alicia M. (1991), "Colonia Patricia Corduba: nuevas hipótesis sobre su fundación y nombre", Latomus 50/4, pp. 846-857.
Villar, Francisco: studies on prehistoric hydronymy in the Iberian Peninsula (the element kar-/kort-).
Conde, José Antonio (1799): the Punic proposal qart ṭūbah "good town". (Historically important; rejected today.)
Seybold, C. F. and Ocaña Jiménez, Manuel: the counter-formula; the name is "certainly not Semitic but Old Iberian".
Cordópolis / eldiario.es, "El origen del nombre de Córdoba": cordopolis.eldiario.es
Tesorillo, catalogue of Hispanic mints, CORDVBA entry: tesorillo.com/hispania/2latinas1.htm
Spanish Wikipedia, "Córdoba (España)", Toponimia section.
The debate on the ethnonym "Kurd" and Corduene
Driver, G. R. (1923), "The Name Kurd and its Philological Connexions", Journal of the Royal Asiatic Society 55/3, pp. 393-403.
Nöldeke, Theodor (1898): the derivation Kurd < kurt < Kyrtioi (quoted in Asatrian 2009, p. 26, n. 31).
MacKenzie, D. N. (1961), "The Origins of Kurdish", Transactions of the Philological Society 60, pp. 68-86.
Minorsky, Vladimir (1927 ff.), "Kurds, Kurdistan", Encyclopaedia of Islam.
Izady, Mehrdad R. (1992), The Kurds: A Concise Handbook, Taylor & Francis.
Asatrian, Garnik (2009), "Prolegomena to the Study of the Kurds", Iran and the Caucasus 13, pp. 1-58.
Marciak, Michał, "GORDYENE"; Schmitt, Rüdiger, "CYRTIANS" (2013); Dandamayev, Muhammad, "CARDUCHI": all in the Encyclopaedia Iranica.
James, Boris (2014), "Arab Ethnonyms (ʿAjam, ʿArab, Badū and Turk): The Kurdish Case as a Paradigm...", Iranian Studies 47/5.
Al-Andalus and Ziryab
Reynolds, Dwight (2008), "Al-Maqqarī's Ziryāb: The Making of a Myth", Middle Eastern Literatures 11/2.
TDV Encyclopaedia of Islam, entries "Kurtuba" and "Ziryâb": islamansiklopedisi.org.tr
Gill, John (2008), Andalucía: A Cultural History, Oxford University Press, p. 81. (The sole popular basis of the claim that Ziryab was Kurdish; not an academic study.)
A note on source criticism
This article gives primary weight to the English- and Spanish-language academic literature. Turkish sources were used only in the section on the history of al-Andalus, and even there the problem created by the spelling "Kurtuba" was explicitly flagged. In the debate on the ethnonym "Kurd", both the continuist wing (Driver, Izady) and the philological minimalist wing (Nöldeke, MacKenzie, Asatrian) were presented from their own sources, and Asatrian's documented polemical stance toward Kurdish studies was noted separately.