Aristotle reached Latin Europe in Arabic — through Toledo, after 1085

Before: a Latin world with one shelf of Aristotle and no way to get more

In 1120 a master teaching the liberal arts at Chartres, Laon, or Paris could set his entire inheritance of Greek philosophy on a single shelf. It came to two short treatises of Aristotle — the Categories and On Interpretation — in the sixth-century Latin of Boethius; Porphyry's Isagoge, an introduction to the first of them; and three textbooks Boethius had written himself, on categorical syllogisms, on hypothetical syllogisms, and on topical argument ¹. That was the whole of it. The bundle had a name — the logica vetus, the old logic — and it had carried Latin philosophical teaching for five hundred years ².

Of the Physics, the Metaphysics, On the Soul, On the Heavens, On Generation and Corruption, the Meteorology, the biological works, the Nicomachean Ethics, the Prior and Posterior Analytics — of the corpus that would organise European thought for the next four centuries — the Latin West had nothing. Boethius had intended to translate all of Plato and all of Aristotle. He was executed around 524 with the project barely begun, and no one resumed it ¹.

The gap was not a matter of a few missing titles. It was the absence of an entire way of asking questions. A twelfth-century Latin master had a technique for classifying terms and testing inferences, and almost no equipment for investigating the physical world, the structure of the soul, or the principles of being.

What was actually on the shelf, and what was not

The rest of the classical inheritance was thinner still, and mostly second-hand. Calcidius's fourth-century Latin version of Plato's Timaeus ran to about two-thirds of the dialogue and stopped; it was the only Plato most Latin readers ever saw. Pliny's Natural History supplied facts without a system. Isidore of Seville's seventh-century Etymologies supplied definitions. Macrobius and Martianus Capella supplied a Neoplatonic cosmology in literary dress. Together they constituted a library of summaries whose originals no one in the Latin West could read ².

Medicine without a system, astronomy without a model

The medical situation was the same. Salerno had begun to change it: Constantine the African, who arrived at Monte Cassino from Ifrīqiya in the late eleventh century, had already put a first tranche of Arabic medicine into Latin, and that first wave is the reason the second wave had somewhere to land ¹⁰. But Latin medicine still had no organised body of physiological theory, no pharmacological system, and no diagnostic framework of the kind Arabic medicine had possessed for two centuries. A Salernitan practitioner had recipes, regimens, and a reputation. He did not have a book that told him what a disease was.

The astronomical situation was starkest. Latin Christendom had computus — the arithmetic of fixing the date of Easter, taught as one of the seven arts — and it had no working geometrical model of the heavens at all. Ptolemy's Almagest, the book that had defined mathematical astronomy since the second century, existed in Greek, in Syriac, and in three Arabic versions. It did not exist in Latin ⁴. A Latin master could tell you that the planets moved and could not tell you where one would be next spring.

The mathematical situation was worse than absent, because it was actively encumbered. Calculation was done in Roman numerals, on an abacus, by specialists. The Hindu-Arabic positional numerals with their zero had reached the Arabic world from India by the ninth century and had been systematised by al-Khwārizmī; they reached Latin Europe through this same twelfth-century corridor, and the Latin word for calculating with them — algorismus — is his name.

Why the books were in Arabic in the first place

None of this material was sitting in Toledo by accident, and the atlas is careful here because the accident version is the flattering one for Europe. The Arabic corpus existed because a previous society had paid for it. Between roughly 750 and 1000 CE, ʿAbbāsid Baghdad ran a translation movement that put nearly the whole of Greek secular science and philosophy into Arabic — Galen, Ptolemy, Euclid, Aristotle, Archimedes — with Ḥunayn ibn Isḥāq and his workshop producing versions of a technical quality that made them usable rather than merely readable ¹¹.

Dimitri Gutas's account of that movement is the standard one, and its central claim is directly relevant to what happened at Toledo three centuries later: the Baghdad translations were not the work of antiquarians preserving a heritage. They were financed by the ʿAbbāsid state, the court, and a wide social stratum of patrons who wanted the material because it was useful — for medicine, for astrology, for administration, for engineering, for argument ¹¹. Translation on this scale is expensive, and someone always wants something.

Al-Andalus inherited that corpus and added to it. The library assembled at Córdoba under al-Ḥakam II in the tenth century is reported in the Arabic sources at four hundred thousand volumes, a figure modern scholars treat as conventional rather than counted, but the working reality behind it is not in doubt: by the eleventh century the Iberian Arabic cities held research libraries with original work in them — al-Zarqālī's astronomical instruments and tables at Toledo, al-Zahrāwī's surgical treatise at Córdoba, and the commentary tradition that would culminate in Ibn Rushd ¹².

So the twelfth-century Latin translators were not recovering Greek learning from a vault. They were arriving, four hundred years late, at a library that another civilisation had built, catalogued, argued with, and improved.

Toledo, 1085: a first-rank Arabic library changes owners

In May 1085 Alfonso VI of León-Castile took Toledo. The city did not fall to assault; it capitulated after a long blockade, on terms. The terms guaranteed the Muslim inhabitants their lives, their property, their religion, and the continued use of their principal mosque ¹⁹. Alfonso styled himself emperor of the two religions. For a moment, the arrangement held.

What Castile acquired was not only a fortress on the Tagus. It was the intact working library of a first-rank Arabic scholarly city — the capital of a taifa kingdom whose eleventh-century court had patronised al-Zarqālī, the astronomer whose Toledan Tables would be recalculated across Europe for two hundred years. The books were not burned and not dispersed ¹². They stayed where they were, in a city that now belonged to a Latin Christian king, three weeks' ride from the schools of southern France.

The city that received them was trilingual in practice. Its Arabic-writing Christians — Mozarabs — had lived under Muslim rule for three and a half centuries and kept their liturgy in Latin and their business in Arabic. Its Jewish community was large, prosperous, and Arabic-literate. Ángel González Palencia's four-volume edition of the surviving Toledan Mozarab archive prints 1,175 legal documents from between 1083 and 1315; all but twenty-five of them are in Arabic, drawn up according to Islamic notarial form, by and for Christians, in a city under Castilian rule ¹⁵. That is the population that made the next hundred and fifty years possible.

The transmission: a conquered city and a relay of two tongues

Gerard of Cremona went looking for one book

Some time before 1144 an Italian scholar in his late twenties left Cremona for Castile. His students recorded why. He had heard of Ptolemy's Almagest, could not find it in Latin anywhere, and went to Toledo to look for it: amore tamen Almagesti, quem apud Latinos minime reperit, Toletum perrexit — "for love of the Almagest, which he could not find at all among the Latins, he made his way to Toledo" ³.

He found it. He also found, in the words of the same account, an abundance of books in Arabic on every subject. So he learned Arabic, and he stayed for the rest of his life. Gerard of Cremona died in Toledo in 1187, roughly forty years later. His companions — the socii who had worked alongside him — attached to his eulogy a list of the books he had translated. It runs to seventy-one titles, and it is known to be incomplete; the modern count of translations attributed to him is around eighty ¹.

The list is not a miscellany. Charles Burnett's reconstruction of it shows a deliberate curriculum: Gerard worked outward from logic through natural philosophy to mathematics, astronomy, and medicine, translating in each field the texts an Arabic student would have read, in roughly the order an Arabic student would have read them ¹. He was not collecting curiosities. He was importing a syllabus.

What crossed under his name includes:

• Ptolemy's Almagest, completed around 1175 — the version Europe used until the fifteenth century, in preference to a Greek-derived Sicilian translation made around 1160 ⁴
• Aristotle's Posterior Analytics, Physics, On the Heavens, On Generation and Corruption, and Meteorology (books I–III)
• Avicenna's Canon of Medicine, which became the standard European medical textbook and stayed on university curricula into the eighteenth century ⁹
• al-Khwārizmī's Algebra, and the arithmetic and geometry that came with it
• Euclid's Elements, in the Arabic recension
• al-Rāzī, al-Kindī, al-Farghānī, Galen, Archimedes, Thābit ibn Qurra, Jābir ibn Aflaḥ

Avicenna, Canon medicinae, trans. Gerard of Cremona. Strasbourg, before 1473, fol. 1r. Bayerische Staatsbibliothek, Munich. Public domain via Wikimedia Commons. · Public Domain

The workflow: Avendauth spoke, Dominicus wrote

The other principal Toledan translator worked differently, on different material, and left a description of his method. Dominicus Gundissalinus, archdeacon at Toledo, translated philosophy — Avicenna, al-Fārābī, al-Ghazālī, Ibn Gabirol — while Gerard concentrated on science and medicine ¹. And Gundissalinus did not work alone.

The dedicatory prologue to the Latin Liber de anima, Avicenna's psychology, addresses the archbishop of Toledo and describes the procedure exactly: one man rendered the Arabic aloud, word by word, into the Romance vernacular; Dominicus archidiaconus wrote it down in Latin. The man doing the rendering signs himself Avendauth Israelita — Avendauth the Jew ⁶.

This is the shape of the Toledan transfer, and it is worth stating plainly: the corpus that would become the operating system of Latin Christian thought crossed into Latin through the mouth of a Jewish scholar speaking a Romance vernacular that neither the source language nor the target language could have recognised as literature. Two tongues, three faiths, one sentence at a time.

Gerard's Almagest went the same way. Daniel of Morley, an English student who heard Gerard lecture in Toledo before 1175, recorded that Gerard was assisted by a Mozarab named Ghālib — Galippus mixtarabe in the Latin — who put the Arabic into Castilian speech for Gerard to write down as Latin ⁵. Recent work on the astronomical tables in Gerard's Almagest shows that some were not merely translated but recomputed, which means Ghālib and Gerard were not copyists but working astronomers ⁴.

Since Marie-Thérèse d'Alverny proposed it in 1954, most scholars have identified Avendauth with Abraham ibn Daud, the philosopher and chronicler who had fled Almohad persecution in the south. The identification has never been proved outright, and it remains contested; Gad Freudenthal's 2016 reassessment marshals the strongest evidence in its favour and still calls it a hypothesis ⁷. The atlas records it as probable, not certain.

Who else was there, and what the archbishops wanted

The translators were not a staff. They were a succession of individuals who came, sometimes overlapped, and left or died: John of Seville, working in the 1130s and 1140s; Hermann of Carinthia and Robert of Ketton in the Ebro valley in the 1140s; Gerard and Gundissalinus and Avendauth in the second half of the century; Alfred of Sareshel and Michael Scot around 1200; Mark of Toledo under Archbishop Rodrigo Jiménez de Rada after 1209 ². Daniel of Morley came from Paris, disgusted by what he found there, because — he wrote — the teaching of the Arabs had made Toledo famous, and he hurried there "to hear the wiser philosophers of the world" ⁵.

The patronage was ecclesiastical. Archbishop Raymond (1125–1152) and Archbishop John (1151–1166) supported the work; the De anima prologue is dedicated to the latter ⁶. Cathedral chapters supplied prebends, which supplied income, which supplied time. This matters for the record's cost framing, because the same institution that paid for Aristotle also paid for something else, and did so in the same decades.

Michael Scot, and the arrival of the Commentator

The second phase changed what the Latin West was reading Aristotle with. Averroes — Ibn Rushd of Córdoba — had died in 1198 after writing the most systematic commentaries on Aristotle produced in any language. Within about thirty years of his death, his long commentaries were in Latin. Michael Scot translated Aristotle's Historia animalium at Toledo around 1220, entered the service of Frederick II in Sicily around 1228, and produced Latin versions of Averroes's long commentaries on On the Heavens, On the Soul, the Physics, and the Metaphysics ¹⁰.

The consequence is easy to state and hard to overstate. Latin Europe did not receive Aristotle and then, later, receive interpretations of him. It received Aristotle already interpreted — wrapped in three centuries of Arabic commentary, with Avicenna's psychology arriving before Aristotle's own text of On the Soul had been digested, and Averroes arriving with such authority that scholastic Latin gave him a title rather than a name. Aristotle was the Philosopher. Averroes was the Commentator ¹⁰.

There was no "School of Toledo"

The phrase is a nineteenth-century invention. In 1819 the French orientalist Amable Jourdain wrote of a collège de traducteurs at Toledo; in 1874 the German philologist Valentin Rose popularised die Schule von Toledo. No medieval source describes any such institution. There was no charter, no faculty, no salaried staff, no building ¹.

What there was, Burnett argues, is better and stranger than a school: a program — a coherent sequence of texts pursued over decades by scholars who knew each other's work, in a city that happened to hold the books and the bilingual population required to read them ¹. The record uses "Toledo" throughout as a place and a corridor, not as an institution.

Photograph by Richard Mortel. Mosque of Bāb al-Mardūm (Cristo de la Luz), AH 390 (1000 CE), Toledo. CC BY 2.0 via Wikimedia Commons. · CC BY 2.0

The second phase: Alfonso X, and the corridor changes language

The Latin phase ran roughly to the middle of the thirteenth century. Then the corridor did something unusual: it switched target languages. Alfonso X of Castile, who reigned from 1252 to 1284 and is known as el Sabio, sponsored a further wave of translation out of Arabic — but into Castilian, not Latin. The astronomical compilations made for him, the Libro del saber de astronomía and the tables known ever after as the Alfonsine Tables, were assembled by teams that included Jewish scholars working in the vernacular ¹².

The consequences ran in two directions. Scientific Castilian was, in effect, invented for the purpose — a European vernacular given technical vocabulary and expository prose two centuries before most of its neighbours had either. And the Alfonsine Tables, translated onward into Latin at Paris in the 1320s, became the standard European astronomical tables, used for planetary positions across the continent until the sixteenth century. Copernicus computed with them.

It is the same corridor, the same city, the same intermediary population, and a different destination language. It is also the last phase: Arabic-to-Latin translation activity ceased almost entirely after about 1300 and did not resume until after 1480 ¹⁰.

What changed and what was replaced

From prohibition to prescription: Paris, 1210 to 1255

The Latin institution that received the corpus tried, at first, to refuse it. The sequence is documented and short enough to set out plainly:

Forty-five years separate the prohibition from the requirement. The commission of 1231 never produced an expurgated Aristotle. What happened instead is that the arts faculty read the books anyway, the books proved indispensable, and the institution wrote the fact into its statutes.

What the university became

The corpus did not merely enter the curriculum; it supplied the curriculum's shape. An arts degree at Paris after 1255 meant working through a fixed sequence of Aristotelian books with fixed lecture allocations. That structure — a defined corpus, a defined order, a defined method of disputation over contested propositions — is the ancestor of the syllabus, the examination, and the academic degree ¹⁸.

It also made the arts faculty something it had not been. Under the older cathedral-school arrangement the liberal arts were preparation for theology. After the reception of a complete natural philosophy, the arts faculty had a subject matter of its own — the physical world, the soul, the heavens — that did not reduce to theological propositions. The condemnations of 1270 and 1277 are best read as the theology faculty noticing this ¹⁷.

The vocabulary that came with it

The loanwords are the fossil record of the transfer, and they are still in use. From the mathematical and astronomical texts: algebra (al-jabr), algorithm (from al-Khwārizmī's name), zenith, nadir, azimuth, almanac, and the star names — Aldebaran, Altair, Betelgeuse, Rigel, Vega, Deneb. From the medical and chemical texts: alkali, alcohol, elixir, syrup, soda. The philosophical borrowings are subtler and more consequential: technical Latin terms such as intentio, coined to carry Avicenna's maʿnā, entered scholastic psychology and stayed there for four hundred years ⁸.

What it did to medicine and astronomy, concretely

The philosophical reception is the famous half of the story. The practical half is larger. Avicenna's Canon of Medicine in Gerard's Latin gave European medicine what it had never had: a single systematic work covering general principles, materia medica, diseases by body part, generalised conditions, and compound drugs, organised so that it could be taught in sequence and examined. It became the backbone of medical teaching at Montpellier, Bologna, Padua, and Paris, and it was still being lectured on in European universities into the eighteenth century ⁹. No Latin work displaced it for five hundred years, because no Latin work of that scope existed.

Astronomy was reorganised just as completely. Before the Almagest, European astronomy was calendrical. After it, European astronomy was a mathematical discipline with a geometrical model, a star catalogue, and a computational apparatus — epicycles, eccentrics, and equants — that practitioners could criticise, adjust, and eventually reject. Al-Zarqālī's Toledan Tables were adapted for Marseille, London, Paris, and Toulouse; the Alfonsine Tables superseded them; and the whole enterprise Copernicus was trained in, and then broke with, is the one that arrived through this corridor ⁴.

The point is worth stating without romance. The transfer did not give Europe a set of conclusions. It gave Europe a working technical apparatus — models to compute with, a corpus to teach from, and a commentary tradition demonstrating that authorities could be argued with. The seventeenth century's rejection of Ptolemy and Galen was only possible because the twelfth century had imported them in a form precise enough to be proved wrong.

The Greek route, and an honest accounting of the debate

The Arabic route was not the only one, and the atlas does not claim it was. Greek-to-Latin translation ran in parallel throughout the twelfth century. James of Venice translated the Posterior Analytics, Physics, On the Soul, and part of the Metaphysics directly from Greek before about 1150; Burgundio of Pisa worked on Galen and John of Damascus; and in the 1260s William of Moerbeke produced the direct-from-Greek Aristotle that Aquinas preferred and that displaced several Arabic-derived versions ¹⁰.

Two things are true at once. First, the direct Greek versions of several key texts existed early and were widely used — James of Venice's translations were read by Grosseteste, Bacon, and Aquinas. Second, for a long list of works the Arabic-derived version was the one Europe actually used, and for the sciences the Arabic route carried material that had no Greek route at all: al-Khwārizmī's algebra, al-Zarqālī's tables, Avicenna's Canon, al-Zahrāwī's surgery, and the whole commentary tradition. The Almagest is the clean test case: a Greek-derived Sicilian translation was made around 1160, and it lost. Gerard's Arabic-derived version of about 1175 is the one that circulated, the one that was copied, and the one that was printed in 1515 ⁴.

This has been a live controversy. In 2008 Sylvain Gouguenheim's Aristote au Mont Saint-Michel argued that the Greek route, and specifically the work done at Mont Saint-Michel, was the decisive channel and that the Arabic contribution had been inflated ¹⁴. The response was immediate and overwhelmingly negative: a collective volume, Les Grecs, les Arabes et nous, edited by Philippe Büttgen, Alain de Libera, Marwan Rashed, and Irène Rosier-Catach, dismantled the book's handling of the manuscript evidence, and a petition of medievalists and historians of philosophy repudiated its use ¹³. The scholarly consensus after the episode is not that the Greek route was unimportant. It is that the two routes ran together, that the Arabic route dominated in the sciences and in the commentary tradition, and that the arithmetic of which corpus arrived when is not in serious dispute — 131 Arabic philosophical texts are catalogued as having crossed into Latin ¹⁰.

What was displaced

The reception replaced things, and the atlas names them:

1. The Platonising cosmology of the twelfth century. The Chartrain reading of the Timaeus — the world-soul, the cosmos as ordered artifact — was the most ambitious natural philosophy Latin Europe had produced on its own. After the Physics and On the Heavens, it had no future as a research programme.
2. Monastic lectio as the dominant mode of study. Slow devotional reading of authoritative texts gave way to the disputation: a contested question, arguments on both sides, a determination.
3. Salernitan practical medicine as the top of the field. Avicenna's Canon supplied a complete theoretical system, and the system won ⁹.
4. The arts as mere preparation. The trivium and quadrivium became a philosophy faculty with its own object.
5. Latin astronomy as computus. After the Almagest, European astronomy meant geometrical modelling, and it meant working in Ptolemaic terms until Copernicus ⁴.

What did not cross

A transmission is defined as much by its filter as by its contents, and Toledo's filter was narrow in a specific way: it passed science and philosophy, and almost nothing else.

• Arabic literature did not cross. No Latin translation of the muʿallaqāt, of al-Mutanabbī, of Ibn Ḥazm's Ring of the Dove, of the maqāmāt. The Arabic that Europe learned to read was technical prose.
• Islamic theology did not cross as theology. Kalām — the discipline in which Muslim thinkers argued about God's attributes, causation, and the created world — arrived only in fragments, filtered through philosophical texts.
• The great internal critique of philosophy did not cross. Al-Ghazālī's Maqāṣid al-falāsifa, a summary of Avicennian doctrine written as the prelude to a refutation, was translated at Toledo at the end of the twelfth century — without its introduction. Latin readers therefore took "Algazel" for a faithful disciple of Avicenna and used him for four hundred years as a convenient compendium of Arabic philosophy. His actual argument against the philosophers, the Tahāfut al-falāsifa, was not translated in this period at all. Albert the Great and Aquinas both cite Algazel; neither knew what he thought.
• Arabic historiography did not cross. Latin Europe took the astronomy of a civilisation without taking its account of itself.

The filter is not neutral. A society that imports another society's instruments while declining its poetry, its theology, and its history has already decided what the other society is for.

Aquinas, and the settlement that held for four centuries

Albert the Great, teaching at Paris and Cologne from the 1240s, set out to expound the whole Aristotelian corpus in Latin. His student Thomas Aquinas, writing the Summa theologiae between 1265 and 1274, built a theology whose framework, vocabulary, and method were Aristotelian and whose interlocutors were Arabic. Aquinas cites Avicenna and Averroes constantly, agrees with them frequently, and refutes them precisely; his treatise against the Averroist reading of the intellect is an argument conducted inside the Commentator's own terms ¹⁰.

The synthesis held. Despite 1277, despite the fourteenth-century critics, the Arabic-transmitted Aristotelian corpus remained the operating system of European natural philosophy, medicine, and metaphysics until the seventeenth century displaced it — which is to say, until Galileo, Descartes, and Newton were arguing against something specific, and what they were arguing against had come through Toledo.

What the cost was

The transmission itself was peaceful. Its precondition was not.

Nobody died for the Almagest. There is no massacre attributable to the translations, no expropriation of books that can be documented as theft, no coerced labour in the scriptoria. Gerard came voluntarily and stayed voluntarily. Avendauth initiated the Avicenna project rather than being conscripted into it ⁷. On the evidence, the working relationships in Toledo were collegial.

The cost is elsewhere, and it is real. The corridor existed because Toledo had been taken by conquest; the communities that made the corridor passable were the ones the conquering society progressively destroyed; and the credit for what crossed was systematically withdrawn once the crossing was complete.

The terms of 1085, and what happened to them

Alfonso VI's capitulation terms guaranteed the Muslims of Toledo their principal mosque. In October 1087, while Alfonso was away, Archbishop Bernard de Sédirac and Queen Constance sent armed men to seize it, set up a Christian altar, and hang a bell in the minaret. The chronicle tradition reports that Alfonso returned furious and close to punishing those responsible, and that the city's Muslim jurist Abū al-Walīd talked him down and persuaded his own community to accept the loss ¹⁹. The mosque became the cathedral. The treaty had lasted two years.

The Muslim population that had been guaranteed its lives, property and religion did not stay to test the guarantee for long. Across the twelfth century the Muslims of the conquered centre either emigrated south to Granada and across to North Africa or remained as mudéjares — legally tolerated subject communities with their own quarters, mosques, and jurisdiction, and no political future ¹⁹. Toledo, a first-rank Islamic city in 1084, was a Christian city with a Muslim minority within two generations. The libraries stayed. The society that had produced them left, or was reduced.

The Mozarabs — the Arabic-speaking Christians without whom the translations could not have been made — were dismantled in a different register. The Council of Burgos in 1080, five years before Toledo fell, had already abolished their Hispanic rite in favour of the Roman one throughout Castile and León, at the urging of Alfonso VI and Gregory VII and under Cluniac pressure. Toledo's Mozarabs resisted and won a partial exemption — six parishes — which is to say they were permitted to survive as a curiosity ¹⁹. The linguistic record in González Palencia's archive shows the community writing Arabic into the fourteenth century and then ceasing to ¹⁵.

The same corridor produced the first Latin Qur'an

In 1142 Peter the Venerable, abbot of Cluny, travelled to Spain and commissioned a team — Robert of Ketton, Hermann of Carinthia, Peter of Toledo, and a Muslim identified only as Muhammad — to produce a Latin dossier on Islam. Robert of Ketton's translation of the Qur'an, finished in 1143 and titled Lex Mahumet pseudoprophete, was the first into any western European language. It survives in twenty-four manuscripts, was printed by Bibliander in 1543, and was the version Europe read for four hundred years ¹⁶.

Its purpose was stated: to make Christian refutation of Islam more effective. In 1210 Mark of Toledo produced a second, more literal Latin Qur'an for Archbishop Rodrigo Jiménez de Rada — the same archbishop who two years later rode with the Christian coalition at Las Navas de Tolosa ¹⁶. The translation movement and the crusade shared patrons, a city, and a decade.

This is not a hypocrisy to be scored. It is the actual structure of the transfer: the same skills, the same intermediaries, and often the same men served scientific appropriation and religious polemic at once. Thomas Burman's study of how these Qur'an translations were actually read shows the two motives intertwined — the philological seriousness was genuine, and so was the polemical purpose ¹⁶.

The intermediaries: from indispensable to suspect

The communities that made the relay possible — Toledo's Jews and Mozarabs — were tolerated while they were useful and destroyed after they were not. The dates are specific.

• 1391. The pogroms that began in Seville in June reached Toledo within weeks. The city's judería, the most important Jewish quarter in Castile, was sacked; its great synagogues were burned or converted to churches; the dead ran to hundreds in Toledo and to thousands across Castile and Aragon, and the survivors converted in the tens of thousands ²⁰.
• 1449. On 5 June, during an anti-converso rising led by Pero Sarmiento, Toledo promulgated the Sentencia-Estatuto — the first blood-purity statute in European history. It barred conversos of Jewish descent from municipal office, from judgeships, and specifically from the notariate: the profession that had drawn up the Arabic legal instruments in González Palencia's archive ²¹.
• 1492. The Catholic Monarchs expelled the Jews of Castile and Aragon. Toledo's Jewish community, which had been continuously present since before the Visigoths and had supplied the Latin West with the men who read Avicenna aloud, ended.

Three hundred years separate the Liber de anima from the expulsion. That is a long causal chain, and the atlas does not claim the translations caused the pogroms. It claims something narrower and harder: that the transfer depended on a plural society, that the receiving society did not preserve that plurality, and that the same city produced both the corridor and, in 1449, the legal instrument that made ancestry a permanent disability.

Suppression after absorption

The final cost is a matter of the record itself. Arabic-Latin translation ceased almost entirely after about 1300 and resumed only after 1480 ¹⁰. The Renaissance revival was substantial: new translations, multi-volume editions of Avicenna and Averroes, Arabic authorities cited across medicine, philosophy, and astrology.

And alongside it ran a campaign to write the debt out. Dag Nikolaus Hasse's Success and Suppression is the definitive study, and its finding is deliberately two-sided: the Renaissance was simultaneously the peak of Arabic influence in Europe and the period in which the West began to deny it. The denial was not a broad consensus but the work of a small group of hard-liners who attacked the Arabic-Latin corpus for barbarous style, for plagiarism from the Greeks, and for irreligion — and who were arguing so hard precisely because the influence was so strong ⁹. Humanist editors purged Arabic names from title pages, replaced Arabic-derived versions with Greek-derived ones where they could, and taught that the Middle Ages had recovered nothing and the Renaissance had recovered everything.

That campaign succeeded well enough that the standard European story of the classical revival still begins in fifteenth-century Florence rather than twelfth-century Toledo — three hundred and fifty years late, and in the wrong language. The nineteenth century then invented a "School of Toledo" that had never existed, and in 2008 a French medievalist could still publish a book arguing that the Arabic route had been overstated, and be answered by a book-length collective rebuttal ¹³¹⁴. The argument about who is owed what has run for five hundred years and is not finished.

What this record does not claim

The atlas is not describing a tolerant paradise despoiled by Christian conquest. Al-Andalus in the twelfth century was also the Almohad state: it forced conversion or exile on Jews and Christians from 1148, drove the young Maimonides's family into a decade of wandering, and in 1195 tried Averroes himself, burned his philosophical books, banned philosophy, and exiled him to Lucena. He was rehabilitated two years later and died in 1198 ¹². The Latin West received the Commentator's work partly because the society that produced it had just repudiated him.

Nor does the record claim the translators were agents of empire. They were, for the most part, scholars pursuing books, and several of them — Gerard above all — gave their working lives to a city not their own and died there. The cost framing is structural, not personal.

The bill

Cost severity 2. No deaths attributable to the transmission. A treaty broken within two years of the conquest that opened the corridor, and a principal mosque taken by force in October 1087. A liturgical tradition abolished by council in 1080 and reduced to six exempted parishes. Two intermediary communities — Mozarab and Jewish — that were indispensable to the transfer and were, within three centuries, one assimilated out of existence and one massacred in 1391, legally disabled in 1449, and expelled in 1492. And a five-hundred-year campaign of credit erasure that is still shaping how Europeans narrate their own intellectual origins.

Against that: the European university, the recovery of Greek science, four centuries of natural philosophy, and the intellectual equipment that the seventeenth century needed in order to argue with it.

Both columns are real. The atlas prints both.