Nuzhat al-Mushtaq fi'khtiraq al-'Afaq (The Tabula Rogeriana)

Completed in 1154 for Roger II of Sicily, the Nuzhat al-Mushtaq is a seventy-map Arabic geography compiled by the Andalusi scholar Muhammad al-Idrisi from fifteen years of interviews with travelers, verification by agents sent into the field, and the inherited framework of Ptolemaic and Balkhi-school geography. A companion world map said to have been engraved on a silver disc does not survive, and this record has found no primary source that documents how or exactly when it was lost; that part of the story is report, not record, and is presented that way.

The book itself survives only through later copying: about ten manuscripts worldwide, the earliest from around 1325. This record is illustrated from Bodleian Library MS. Greaves 42, a genuine but partial and anonymous 14th- or 15th-century Arabic copy whose own history is undocumented before its arrival at Oxford in 1678. The image most people picture when they hear 'Idrisi map,' Konrad Miller's 1929 composite reconstruction, is a modern scholarly artifact assembled from later copies and is presented here only as a clearly labeled comparandum, never as the object itself.

Read past the popular version and two disciplined findings emerge: the map's celebrated south-up orientation is a documented Balkhi-school convention al-Idrisi adopted rather than an invention of his own, and the frequent claim that his geography fed European Age-of-Exploration cartography does not survive contact with the fact that no Latin translation existed before 1619. What remains, honestly stated, is still substantial: one of the most systematically compiled geographies of the medieval world, made possible by a Sicilian court willing to fund fifteen years of cross-checked, multilingual fieldwork.

A book of seventy maps, and a lost disc of silver

What is usually called the Tabula Rogeriana is not one object but two, and only one of them survives, and not in its original form. The first is a work of geography: an Arabic treatise titled Nuzhat al-Mushtaq fi'khtiraq al-'Afaq, roughly The Entertainment of One Who Longs to Traverse the Horizons, completed in 1154 for Roger II of Sicily. It divides the inhabited world into seven latitudinal climate zones after the Ptolemaic system, each zone cut into ten sections, giving seventy sectional maps, each paired with a two-page written description of the towns, distances, peoples and produce of that section. The second is a report, separate from the book and much less securely documented, that the same commission also produced a world map engraved on a large silver disc.

The book survives, but not as al-Idrisi left it: it survives only in later copies, the earliest of them dated a century and a half after 1154. The disc does not survive at all. This record covers both honestly, which means treating them differently: the book's transmission can be traced through named, catalogued manuscripts; the disc's fate cannot be traced through any comparable evidence and rests on report rather than record.

The commission: fifteen years at the Norman court

Roger II of Sicily ruled a kingdom that had absorbed a substantial Arabic-speaking, Arabic-literate population along with Byzantine Greek and Latin administrative traditions, and his court at Palermo was unusual in medieval Europe for employing Muslim scholars, physicians and officials alongside Christian ones. Al-Idrisi, an Arab-Andalusi geographer of the Hammudid line who had already travelled widely across the Mediterranean and Anatolia, arrived at that court, by his own account and the account of later biographers, after instability in al-Andalus made his position there difficult. He remained at Roger's court for some fifteen to eighteen years working on the geography project; sources differ on the exact figure, and this record has not been able to reconcile the discrepancy from a primary source.

Al-Idrisi describes his method in terms that later scholarship treats as a genuine statement of practice rather than a formula: he says he retained from his informants only that part of their testimony 'on which there was complete agreement and seemed credible, excluding what was contradictory.' Roger II is reported to have taken part personally in some of the interviews with travelers and merchants passing through Sicily, and al-Idrisi is also reported to have sent out agents to check travelers' accounts against each other before anything was recorded as established.

How much of it is fieldwork, and how much is compiled from books

This is where popular accounts of the Tabula Rogeriana tend to overstate the case, and where the honest answer is that both things are true at once. Al-Idrisi's project drew on interviews with sailors, merchants and travelers who had first-hand experience of the coasts and roads they described, and on agents sent to particular regions to corroborate what they said. That is real fieldwork, in the sense that it produced information not available from any earlier written geography, and it shows in the noticeably finer detail given to regions close to Sicily and to the trade routes the Norman kingdom actually used.

At the same time, the intellectual scaffolding of the work is compiled, not observed. Al-Idrisi organizes his seventy sections within the seven-climate system inherited from Ptolemy's Geography, known to him through Arabic translation, and he draws on three centuries of prior Islamic geographical writing, including the tradition associated with the Balkhi school of Baghdad. Earlier Muslim geographers had already corrected some of Ptolemy's errors using exactly the kind of merchant testimony al-Idrisi also relied on, for instance rejecting Ptolemy's claim that the Indian Ocean was a closed sea and revising his gross overestimate of the length of the Mediterranean. Al-Idrisi's achievement was to combine an inherited written framework with a genuinely new round of empirical checking at an unusually large and systematic scale for his period, not to invent either the framework or the method of checking from nothing.

Why south sits at the top

Readers meeting a reproduction of the Idrisi world map for the first time are usually struck by its orientation: south at the top, north at the bottom, with the Arabian peninsula and Mecca near the center. This was not a personal choice by al-Idrisi. The convention was established a good two centuries earlier by the mathematician and geographer al-Khwarizmi in the ninth century, and it was the standard orientation used by the Balkhi school geographers Abu Zayd al-Balkhi and al-Istakhri in the tenth century, whose maps al-Idrisi's tradition inherited.

Why the convention itself exists is argued rather than settled. Some historians read it as an expression of religious orientation toward Mecca, on the model of the qibla; others point out that the earliest of these Muslim world maps carry very little explicit religious iconography, and treat the south-up format as inherited technical convention rather than devotional statement. This record follows the more cautious position: the orientation is well documented as a Balkhi-school convention that al-Idrisi adopted, but the reason for the convention's original adoption is argued, not proven, and this record does not take a side beyond reporting that the dispute exists.

The silver planisphere: what is reported, and what is documented

Modern accounts of the Tabula Rogeriana almost all repeat a striking detail: that alongside the book, al-Idrisi's team produced a second version of the world map engraved on a massive disc of silver, reportedly weighing in the region of 300 lb, and that this disc was lost during unrest that followed Roger II's death. This record has traced that claim as far as it can be traced, and the honest result is that it rests on later report rather than a documented event.

Al-Idrisi's own text, so far as it has been checked for this record, does not itself describe or refer to a silver planisphere. The claim comes from later biographical and bibliographical notices about al-Idrisi, repeated and elaborated in modern secondary literature, not from a surviving eyewitness account of the object or of its destruction. The circumstantial context usually attached to the disc's loss is the period of unrest in Palermo after Roger II's death in February 1154, under his son and successor William I: the assassination of the chief minister Maio of Bari in November 1160, the subsequent revolt led by Matthew Bonnellus in alliance with Simon and Tancred, the seizure and looting of the royal palace, and an anti-Muslim pogrom in Palermo that a contemporary chronicle attributes to the narrowness of the Muslim quarter's streets for having limited its scope rather than to any intervention that stopped it. What none of the primary narrative of that revolt, so far as consulted for this record, actually says is that a silver planisphere was among the palace's contents or was destroyed. The disc's loss is plausible given the palace was looted at exactly the right moment, and it is repeated consistently across modern sources, but it is not, on the evidence gathered here, a documented event in the way the palace looting itself is documented. This record grades it accordingly.

What survives: a manuscript tradition, not a manuscript

Because the original disc and the original presentation copy of the book are both gone, everything known of the Tabula Rogeriana today comes through later copying. Roughly ten manuscript copies of the Nuzhat al-Mushtaq are known to survive worldwide; about half preserve the complete text, and about eight carry at least some of the seventy sectional maps. The earliest known copy, MS Arabe 2221 in the Bibliotheque nationale de France in Paris, is dated to around 1325, some 170 years after composition. A more complete copy, containing the world map and all seventy sectional maps, is held in Istanbul. The Bodleian Library in Oxford holds two separate Arabic copies: MS. Pococke 375, made in Cairo in 1553 and containing descriptions and maps for all seven climates, acquired by the library in 1692; and MS. Greaves 42, undated but assigned on palaeographic grounds to the 14th or 15th century, which is incomplete, carrying only the first three of the work's seven climates, covering territory from the equator to roughly the latitude of North Africa and from the Atlantic coast to eastern China.

No two of these copies necessarily agree in every particular, and the copying process over two and three centuries is exactly the kind of process in which toponyms drift, sections are recombined, and later copyists' own knowledge or ignorance of geography leaves traces. Any specific claim about 'the' Idrisi map has to specify which copy it is drawn from.

This record's manuscript: Bodleian MS. Greaves 42

The photographs used in this record are of MS. Greaves 42, and it is worth being precise about what that manuscript is and is not. It is a genuine Arabic-language copy of al-Idrisi's Nuzhat al-Mushtaq, confirmed as such by the Bodleian Library's own published cataloguing of its Islamic manuscript holdings, not a modern facsimile or a differently titled work that happens to share subject matter. It is partial, containing only the first three of the work's seven climate zones. It is anonymous: no scribe is named, no place of copying is recorded, and its date rests on palaeographic estimate rather than a colophon, placing it one to three centuries after al-Idrisi's own lifetime.

Its documented history begins only in 1678, when the Bodleian purchased the Oriental manuscript collection assembled by the brothers John Greaves, the Gresham College mathematician and astronomer who travelled through the Ottoman Levant and Egypt in the 1630s acquiring manuscripts and instruments, and Thomas Greaves, an Oxford orientalist. That purchase is the origin of the 'Greaves' manuscripts at the Bodleian generally, of which this is number 42. Between the manuscript's copying, sometime in the 14th or 15th century, and its documented arrival in the Greaves collection by 1678, nothing is recorded: where it was, who owned it, and how one of the Greaves brothers obtained it are all open questions this record has not been able to close.

Konrad Miller's 1929 reconstruction, and why it is not the medieval map

The single image most commonly reproduced as 'the Idrisi map,' including on general encyclopaedias and in casual online use, is not a photograph of any medieval object. It is a modern composite drawn in 1928-1929 by the German cartographic historian Konrad Miller, titled Charta Rogeriana: Weltkarte des Idrisi vom Jahr 1154 n. Chr. Miller took the seventy separate sectional maps as they appear in later manuscript copies, principally Bodleian MS. Pococke 375 and/or Bibliotheque nationale de France MS Arabe 2221, and assembled them into a single continuous rectangular world map, transliterating the Arabic place names into the Roman alphabet for a modern readership.

Two things follow from that description and both matter for how this image should be used. First, it is a 20th-century scholarly artifact, not a medieval one: Miller made interpretive choices about how sectional maps drawn as separate two-page spreads, sometimes inconsistently across manuscript copies, should be stitched into one continuous sheet, and those choices are his, not al-Idrisi's. Second, its correspondence with the circular format long associated with al-Idrisi's tradition is itself part of an unresolved scholarly question, addressed below, about whether that circular format originates with al-Idrisi at all. Because Miller's reconstruction is public domain and visually striking, it circulates far more widely than any actual manuscript photograph, and that popularity is precisely why this record separates it out: fame is not the same thing as fidelity to the object.

A rival circular map, a century earlier: the Book of Curiosities problem

A further complication, not usually mentioned outside specialist literature, concerns the well-known circular world map long treated as al-Idrisi's own design. In 2000, the Bodleian Library and the University of Oxford's Oriental Institute began studying an anonymous Fatimid-era Arabic cosmography acquired by the library in 2002, known as the Book of Curiosities (Kitab Gharaib al-Funun), originally compiled in Cairo between about 1020 and 1050, with the sole surviving copy dated to the early 13th century. That work contains its own circular world map, of a type closely resembling the map long labelled 'the Idrisi world map.'

This raises a genuine question of priority that the discovery of the Book of Curiosities put back into play: was the circular world map format invented by al-Idrisi for his 1154 treatise, or did it already exist in the Fatimid geographical tradition a century before him, making the label 'Idrisi world map' a misattribution inherited from modern scholarship rather than medieval practice? The historian Yossef Rapoport, who co-led the Book of Curiosities research project, has argued that the map's different color palette relative to the rest of that manuscript is evidence it was a later interpolation copied from the al-Idrisi tradition rather than an independent precursor to it, which would restore al-Idrisi's priority; but this is Rapoport's argued position, not a settled consensus, and this record has not found a published rebuttal to weigh against it.

What the map gets right, and where it becomes speculative

Judged against later, more precisely surveyed maps, the Idrisi tradition's accuracy is uneven in a pattern that tracks the availability of first-hand testimony. The Mediterranean basin, the regions the Norman kingdom traded with and travelled through, and the general shape of the connected Indian Ocean and Atlantic are rendered with a level of relative correctness that impressed 19th-century Orientalist scholars when the work was properly studied again. Al-Idrisi's team also arrived at a calculated figure for the Earth's circumference; modern secondary sources disagree on which measurement should be quoted and in which units, which is itself worth recording rather than silently resolving, and this record has not been able to establish a single authoritative figure from a primary calculation.

Further from Sicily and from the trade routes reachable by informants at Roger's court, the map becomes progressively more schematic and, in places, simply wrong by inheritance rather than by new error: it repeats Ptolemy's fictitious 'Mountains of the Moon' as the source of the Nile, and it omits central and southern Africa almost entirely rather than guessing at their shape. This unevenness is not a flaw specific to al-Idrisi; it is the expected signature of a compilation method that mixes verified local testimony with an inherited classical framework, and reading the map's reliability as a gradient from the Mediterranean outward, rather than as a single verdict of 'accurate' or 'inaccurate,' is the more defensible way to use it.

An influence claim that needs qualifying

It is often suggested, in popular accounts, that al-Idrisi's geography fed into the European voyages of exploration of the following centuries. The chronology does not support a strong version of that claim. No medieval Latin translation of the Nuzhat al-Mushtaq is known to have circulated in Europe during the relevant period; the first Latin edition, by the Maronite scholar Gabriel Sionita, was printed in Paris only in 1619, more than a century after Columbus and Vasco da Gama, and well after the main phase of Iberian oceanic exploration that a strong influence claim would require. Wider European recognition of the scale and quality of al-Idrisi's achievement had to wait longer still, for the rise of academic Oriental studies in the 19th century, which is also when serious critical study, translation into further European languages, and eventually Miller's reconstruction all followed.

None of this diminishes the work on its own terms. It means the accurate claim is narrower than the popular one: the Nuzhat al-Mushtaq was, by the assessment of the scholars who examined it once it became properly accessible, one of the most detailed and systematically compiled works of geography produced anywhere in the medieval world, and it stands as material evidence of exactly how much cross-cultural scientific exchange the multi-confessional Norman court of Sicily was capable of sustaining. It is not, on the evidence available, a demonstrated causal input into the specific navigational project of the Iberian voyages of discovery three and a half centuries later.

After Roger: a second geography, now lost, and the geographer's own exile

Roger II died in February 1154, close to the completion of the project bearing his name. Al-Idrisi remained at the Sicilian court under Roger's son and successor, William I, known to later historiography as William the Bad, and is reported to have produced for him an enlarged and updated version of the geography, titled Rawd al-Uns wa-Nuzhat al-Nafs, roughly The Garden of Intimacy and the Delight of the Soul. That work does not survive in any known copy; it is known only by title and description in later bio-bibliographical sources, which places it in a strictly weaker evidentiary category than the Nuzhat al-Mushtaq itself, which at least survives in ten manuscript witnesses.

Al-Idrisi's own departure from Sicily is usually connected to the same period of anti-Muslim unrest under William I discussed above in connection with the disputed fate of the silver disc, though, again, no source consulted for this record specifically names al-Idrisi as a witness to or victim of that unrest; the connection is inferential, based on timing, not documented by a source describing his departure directly. He is reported to have died in 1165 or 1166, and even his place of death is not agreed in the sources consulted: some give Sicily, others his birthplace of Ceuta in North Africa. Beyond the two geographical works, he is also credited with a pharmacological treatise cataloguing simple medicinal substances with their names given in multiple languages, reflecting the same comparative, cross-checking habit of mind visible in the geography.

What is being done with it now

The most active recent work on this material is not textual editing but physical reconstruction. The Factum Foundation, a Madrid-based organization specializing in digital recording and physical facsimile production of cultural heritage, has an ongoing project titled Re-creating the Lost Silver Map of al-Idrisi, which uses the surviving sectional maps from the Nuzhat al-Mushtaq manuscript tradition to attempt a reconstruction, focused initially on the Mediterranean and Arabian Gulf sections, of what the lost silver disc's map content might have looked like. This is explicitly presented as a modern reconstruction project working from the surviving book, not as a discovery of the disc itself or of new documentary evidence for its fate, and this record treats it accordingly: it is evidence of continuing scholarly and conservation interest in the object, not evidence that resolves the underlying disputes about the disc's original appearance or its loss.