Jump to ContentJump to Main Navigation
Medieval Islamic MapsAn Exploration$

Karen C. Pinto

Print publication date: 2016

Print ISBN-13: 9780226126968

Published to Chicago Scholarship Online: January 2017

DOI: 10.7208/chicago/9780226127019.001.0001

Show Summary Details
Page of

PRINTED FROM CHICAGO SCHOLARSHIP ONLINE (www.chicago.universitypressscholarship.com). (c) Copyright University of Chicago Press, 2021. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in CHSO for personal use. Subscriber: null; date: 02 August 2021

Classical and Medieval Encircling Oceans

Classical and Medieval Encircling Oceans

(p.113) 6 Classical and Medieval Encircling Oceans
Medieval Islamic Maps

Karen C. Pinto

University of Chicago Press

Abstract and Keywords

Chapter Six, “Classical and Medieval Encircling Oceans” considers specifically the influence of Greco-Roman and subsequent medieval European cartographic traditions from the perspective of the Encircling Ocean form. The chapter works forward in time to medieval maps of European origin, culminating with a remarkable example of intersection between medieval Islamic and European maps.

Keywords:   Greco-Roman, medieval European, Ptolemy, Okeaynus, Strabo, periodos ges, Isidore, Mappamundi, T-O, Beatus

Ancient and Classical Greek Cartography

The fundamental issue with all ancient Greek cartography is that we have no extant examples of any maps.1 As a result, the exact form that these maps took is based on conjecture and reconstruction from long lost sources.2 As Germain Aujac explains in the History of Cartography series, “We have no original texts of Anaximander, Pythagoras, or Eratosthenes—all pillars of the development of Greek cartographic thought. In particular, there are relatively few surviving artifacts in the form of graphic representations that may be considered maps. Our cartographic knowledge must therefore be gleaned largely from literary descriptions, often couched in poetic language and difficult to interpret. In addition, many other ancient texts alluding to maps are further distorted by being written centuries after the period they record; they too must be viewed with caution because they are similarly interpretative as well as descriptive.”3

Even the earliest manuscript attributed to Claudius Ptolemy, the second-century Hellenistic geographer heralded (p.114) since the Renaissance as the “father of cartography,” is dated to the mid-thirteenth century—more than a millennium after Ptolemy’s death.4 Among scholars of Greek cartography, the question of whether Ptolemy ever drew maps, or whether he just wrote about drawing them, is a matter of considerable debate.5 From the fourteenth century until the rise of Renaissance mapmakers (such as Mercator, Bleu, etc.), there was a wave of map manuscripts of European provenance all claiming Ptolemy as author. The maps in these so-called Ptolemy manuscripts vary from one to another so widely in form and content that the attribution of these manuscripts to a single author is nothing short of astonishing. Visual evidence to the contrary aside, the Renaissance-spawned myth that Ptolemy is the father of cartography still endures.6

In the absence of extant maps, our understanding of Greek cartography is heavily dependent on textual sources. One of our best and earliest textual references comes from the ancient period of Greek history through the Homeric tradition (ca. eighteenth century BCE) and one of our most beloved Iliad characters, Achilles—specifically his shield. According to Homer, Hephaestus fashioned a huge shield for Achilles with concentric rings of metal—gold in the center, surrounded by two rings of tin, and two of bronze—with a three-ringed metal rim. Homer tells us that on the front plate Hephaestus created a schematic depiction of the earth and the universe.7 Arrayed around the edge of the shield was the vast and mighty river of the Encircling Ocean, described thus:

  • First of all he forged a shield that was huge and heavy,
  • elaborating it about, and threw around it a shining
  • triple rim that glittered. …
  • There were five folds composing the shield itself, and upon it
  • he elaborated many things in his skill and craftsmanship.
  • He made the earth upon it, and the sky, and the sea’s water
  • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
  • . . . and the strength of Orion
  • and the Bear, whom men give also the name of the Wagon,
  • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
  • and she alone is never plunged in the wash of the Ocean.
  • . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
  • He made on it the great strength of the Ocean River
  • which ran around the outermost rim of the shield’s strong structure.8

So pervasive is the influence of this shield on the history of cartography that Denis Cosgrove suggests, tongue in cheek, that this shield should be considered the cornerstone of the European cartographical imagination!9 And Yāqūt al-Rūmī, the late twelfth-/early thirteenth-century compiler of the opus magnum of geographical (p.115) dictionaries—Muʿjam al-buldān (Dictionary of Places)—informs us that Achilles’s shield was discussed by medieval Islamic scholars as well.10

The dearth of extant maps means that we have to rely on textual sources for information. For the classical period there are but a few obscure snippets, such as the one from Aristophanes’s fifth century BCE comedy The Clouds. An anonymous student introduces the farmer Strepiades to the location of Athens and Sparta on a map:

Student (pointing to a map):

  • Now then, over here, we have a map of the entire world. You see there? That’s Athens.
  • Strepiades:

  • That, Athens? Don’t be ridiculous. Why, I can’t see even a single lawcourt in session.
  • Student:

  • Nonetheless, it’s quite true. It really is Athens.
  • Strepiades:

  • Then where are my neighbours of Kikynna?
  • Student:

  • Here they are. And you see this island squeezed along the coast? That’s Euboia.
  • Strepiades:

  • I know that place well enough. Perikles squeezed it dry. But where’s Sparta?
  • Student:

  • Sparta? Right over here.
  • Strepiades:

  • That’s much too close! You’d be well advised to move it further away.
  • Student:

  • But that’s utterly impossible.
  • Strepiades:

  • You’ll be sorry you didn’t, by god.11
  • From this reference we get a clear, if comical sense, of the spatial location of places in relation to one another.12 From the will of Theophrastus (third century BCE), a student of Plato and Aristotle, we learn of the commissioning of a painted world map on wooden panels for the Lyceum that is unfortunately no longer extant.13

    With his caustic critique about simpleton mapmakers in Histories, Herodotus (ca. 484–425 BCE) fills out the picture a little more: “For my part, I cannot but laugh when I see numbers of persons drawing maps of the world without having any reason to guide them; making, as they do, the ocean-stream to run all round the earth, and the earth itself to be an exact circle, as if described by a pair of compasses, with Europe and Asia just of the same size.”14 Herodotus held that Europe was twice the length and breadth of Asia and without a confirmed terminus. Therefore he objected to the notion of an Encircling Ocean since he thought that the landmass of Europe extended where the ocean was drawn. Aristotle bolsters Herodotus’s critique by arguing in his Meteorologica that those who drew periodos ges (circuits of the earth) were illogical.15 In this thinking, Aristotle went against the teachings of his mentor, Plato, who also speculated on the shape of the world and invented a fanciful creation myth in Timaeus using information from Solon. In a dialogue with the usual suspects, Socrates and friends, Timaeus explains that


    God placed two other elements of air and water, and arranged them in a continuous proportion—fire:air::air:water, air:water::water:earth, and so put together a visible and palpable heaven, having harmony and friendship in the union of the four elements. … And as he was to contain all things, he was made in the all-containing form of a sphere, round as from a lathe and every way equidistant from the center, as was natural and suitable to him. … All that he did was done rationally in and by himself, and he moved in a circle turning within himself, which is the most intellectual of motions. … And so the thought of God made a God in the image of a perfect body, having intercourse with himself and needing no other, but in every part harmonious and self-contained and truly blessed.16

    This idea of the sphere as the perfect form was picked up by both Christian and Muslim Neo-Platonists and, as Lilley argues, can be seen as the basis for the importance of the circular form in medieval images, including maps.17 Plato does not mention anything specific about the Encircling Ocean except to say that Okeaynus (the Encircling Ocean) was part of the elements mixed into the original cosmogony.

    In his opus Geography, which preserves the works of many ancient and classical Greek authors, Strabo discusses the matter of the ocean in a logical, matter-of-fact manner: “We may learn both from the evidence of our senses and from experience that the inhabited world is an island; for wherever it has been possible for man to reach the limits of the earth, sea is found, and this sea we call ‘Oceanus.’ And wherever we have not been able to learn by the evidence of our senses, there reason points the way.”18 Without extant images, however, we cannot tell for certain if theories of the shape, form, and geographical structure of the earth, such as those of Strabo, were translated into actual maps or not. The most frequent references to maps in the ancient Greek context are to those known as the periodos ges or “Ring-Around-the-World” images possibly made on wooden panels or bronze tablets (pinax; pl. pinaces). At issue is the meaning of the word pinaki, which is the dative singular of pinax and depending on the context can mean “on the pinax,” “for the pinax,” and so on. Pinax has a wide range of meanings, including “board,” “plank,” “writing-tablet,” “votive tablet,” or “board for painting,” and could therefore refer to a picture, plate, or even a public notice board.19

    Periodos ges (pl. periata gaies), as these “Ring-Around-the-World” maps are sometimes called, is not a clear reference to a map either. The same name is used as the title for textual accounts of the world initiated by the sixth century BCE Ionian scholars of Miletus—Anaximander and Hecataeus—and continued by Herodotus’s fifth century BCE contemporary Democritus and the fourth century BCE scholars Eudoxus and Dicaearchus of Sicily. As indicated by Dionysius’s popular poetic version, the tradition of writing periodos ges continues well into the second century CE. Thus, periodos ges can just as easily be translated as a “journey around the world”—an ancient travel account akin to the medieval Islamic riḥla (travelogue) tradition. Discerning the difference (p.117) between a textual reference to a journey and a reference to a physical map is yet another example of how challenging it is to study ancient Greek maps.20

    Given all the references to the shape, description, and size of the earth, pinaces, and periata gaies, and many others that I have not mentioned in the interest of brevity, it is indeed surprising that no ancient Greek maps have survived. What these textual sources indicate is that from the sixth/fifth century BCE onward ancient Greek society was growing accustomed to quotidian ways of seeing and thinking about the world, if only in theoretical form. Precisely what forms these maps took remain undetermined. How much of the world did they represent, and what were the roots of their inspiration? These questions have formed the basis of some of the most intellectually stimulating work on the history of cartography, but without extant visual examples, no one can know the precise nature of the influence of Greek ideas on Muslim and medieval European cartography.21

    The matter is complicated not only by the lack of extant ancient Greek maps but also by the low survival rate of other key geographical works. While there is no evidence to suggest that the Arabs knew about the work of Strabo or that they consulted it, there is no question that the Arabs knew and preserved Ptolemy’s Almagest, hence its Arabized title. But the Almagest is primarily a work of astronomy while Ptolemy’s Geography was one of latitudinal and longitudinal listings of the location of eight-thousand-plus places. Ptolemy’s work likely influenced the development of the voluminous Islamic zīj tables tradition,22 but the question of influence on Islamic mapping is not an open-and-shut case as has been asserted in the past.23 Although Masʿūdī asserts that the Ṣūrat al-Maʾmūniyya was more beautiful than the maps of Ptolemy and Marinus, none of these maps are extant.24 In the absence of copies of either, we are left in the dark as to the exact nature of this influence. Zuhrī claims that his work is a copy of the Ṣūrat al-Maʾmūniyya but his work does not contain maps either, and Zuhrī’s usage of the Iranian kishwar system (see previous chapter for details) has puzzled scholars who suggest that the Ṣūrat al-Maʾmūniyya may instead have been a mixture of the Iranian and Ptolemaic systems.25 It is just as likely that Masʿūdī and Zuhrī were referring to a KMMS-type world map that may have had its roots in the period of the Abbasid caliph Maʾmūn (r. AH 197–218/813–833 CE). Another possibility is that they were using the word ṣūrat in a textually visual sense to refer to images—the way that the ancient Greeks used periodos ges and pinax. Either way, we cannot look solely to the Greeks for influence on the KMMS world map. Tibbetts goes so far as to suggest that the KMMS maps were “produced independently as a reaction against work dependent on Greek and other foreign agencies.”26

    Roman Cartography

    The case of Roman cartography needs to be considered in the context of Greek cartography. Influential Romans, including the emperors, actively patronized Greek (p.118)

    Classical and Medieval Encircling Oceans

    Fig. 6.1. Itinerarium a Gades Romam. Tabular itinerary showing a list of stops and distances from Cádiz in southwestern Spain to Rome. Circa 330. Line drawing of text on one of the silver beakers from Bagni di Vicarello: Corpus Inscriptionum Latinarum XI 3281, Academia Litterarum Borussica, Leipzig/Berlin 1862–1943.

    Original beaker is in Rome: Museo Nazionale delle Terme, inv. No. 67497.

    scholars. It is generally acknowledged, however, that the Romans approached mapping from a more practical vantage point—that is, for cadastral and architectural surveys.27 The extant record of possible Roman maps is almost as scant as in the case of the Greeks. What little exists speaks of a limited mapping tradition.28

    Map-wise what the Romans were most interested in laying out were their roads, and it is therefore not surprising to learn that they produced itineraries in large volume, at least from the fourth century CE onward when pilgrimage to the Levant became popular (fig. 6.1). Usually a bare list of stops and distances indicating mutatio (change of mount) and mansio (overnight stops), these itineraries took a number of forms. Some were etched on pillars of marble, others jotted down in manuscripts, and still others etched around silver beakers. There are many extant examples of these itineraries, including famed ones such as the octagonal column from Tongeren (Atuatuca Tungrorum) in present-day Belgium, the Antonine Itinerary (Itinerarium Antonini), the Patara monument, and pilgrimage accounts such as that of Theophanes from Hermopolis Magna in Egypt to Antioch in Syria (322–324).29 What they attest to is that

    It was the mind’s eye which ancient chorographers stressed, since written texts—“la carte écrite”—created no less visible, spatial images than maps. … Written itineraries could be more easily memorized than maps, because they were sequential. … Space itself was defined by itineraries, since it was through itineraries that Romans actually experienced space, that is, by lines and not by shapes. … Movement was more important (p.119)

    Classical and Medieval Encircling Oceans

    Fig. 6.2. Close-up of Roma from Konrad Miller’s 1887/1888 line drawing of Tabula Peutingeriana. This is an illustrated itinerarium showing the road network of the Roman Empire. It covers Europe, North Africa, and parts of Asia. Parchment scroll in eleven sections.

    Earliest copy dates to the thirteenth century. 0.34 × 6.75 m. Österreichische Nationalbibliothek, Vienna, Codex Vindobonesis 324, segment 9 (of 11).

    than the form. … The essence of itineraries was space between topographic points—towns, vici, mansions, and so on—which followed in linear succession. It reflected the Roman political view of a world made up of a network of civitates and urban space, between which lay nothing except curiosities for the traveller.30

    The only visual example we have of this Roman zest for routes is the famous Peutinger Table (fig. 6.2). It too suffers from the Ptolemy syndrome (discussed earlier in this chapter) in that the earliest extant copy only dates back to the thirteenth century. Richard Talbert, an ardent proponent of the authenticity of the Peutinger map, while acknowledging the legitimate reservations of skeptics, devotes a book to documenting the history of this map. Talbert argues that although the Peutinger map is clearly (p.120) a medieval production, content analysis suggests that the original copy was done no later than 700 and most likely stems from the period of Diocletian’s Tetrarchy (ca. 300)—in tandem with the pilgrimage boom.31

    The map is composed of multiple sheets and focuses on displaying routes and realms of and for the Roman Empire’s network of roads. It does not contain an Encircling Ocean—unless the green background behind the edges of the landmass is intended to be an indication of the ocean encasing the land. It does, however, contain an image of Roma encased in a double-ringed circle that many scholars believe would have been the center of this map if all its pieces were extant. The crowned female figure seated in the center of this double-ringed circle holds a ball in her right hand that could be interpreted as an orb representing Rome’s dominion over the world. In true medieval European illustration style, as we shall see in the next section, this figure fits in with the portrayals of God and earthly rulers encased by symbolic double-ringed circles. Could this circular form be a resonance of the Achaemenid double-ring motif (discussed in the previous chapter) bequeathed by the Romans to European illuminators? This is a riddle that is unlikely to be resolved unless maps extant from the Roman period are uncovered.

    The same dilemma holds true for the Agrippa world map said to have been commissioned during the reign of the Emperor Augustus. Some suggest that Agrippa never completed this map, while others say it wasn’t really a world map but rather a self-aggrandizing view of only the Roman Empire as the world.32 Agrippa’s world map is sometimes linked with the Hereford world map of the late thirteenth century (fig. 6.10) on the basis of a sketch of Augustus Caesar in the lower left-hand corner issuing an edict for the Hereford map. The connection is tenuous at best. If there is any validity in it, it would suggest that the Romans upheld the concept of an Encircling Ocean.

    This interpretation of the presence of the Encircling Ocean in Roman world maps is reinforced by Eumenius’s panegyric on the rebuilding of a rhetorical school in Gaul in which he mentions a wall map that is on display in the school:

    In [the school’s] porticoes let the young men see and examine daily every land and all the seas and whatever cities, peoples, nations, our most invincible rulers either restore by affection or conquer by valor or restrain by fear. Since for the purpose of instructing the youth, to have them learn more clearly with their eyes what they comprehend less readily by their ears, there are pictured in that spot … the sites of all locations with their names, their extent, and the distance between them, the sources and mouths of rivers everywhere, likewise the curves of the coastline’s indentations, and the Ocean, both where its circuit girds the earth and where its pressure breaks into.33

    There are some extant images such as Roman landscape wall murals that can be interpreted as maplike images with “curves of coastlines” and “mouths of rivers” and mosaics involving routes and places, such as the famous Madaba mosaic map. None (p.121)

    Classical and Medieval Encircling Oceans

    Fig. 6.3. Quartered earth within a circular heaven of Recension A of the Corpus Agrimensorum Romansorum. Occurs in a Roman land-surveying manual. Blue gouache and red and black ink on parchment.

    Courtesy: Herzog August Bibliothek Wolfenbüttel. Cod. Guelf. 36.23 Aug 2°,43b.

    of these, however, show the whole world, so I am excluding them from this discussion on the Encircling Ocean, although they are certainly relevant to the heritage of Roman mapping.

    The most interesting material from the perspective of the Encircling Ocean form are the images in a set of Roman surveying manuals known as the Corpus Agrimensorum Romansorum. In addition to dealing with land surveying, this manual addresses elements of world cosmography. Particularly intriguing is the image of the quartered earth within a circular heaven (fig. 6.3). The Corpus Agrimensorum Romansorum is also plagued with dating and provenance issues. The only extant copy has been dated variously to between the fifth and seventh centuries and would thus properly fall within the orbit of medieval not Roman painterly technique.34

    From the twilight years of Roman dominion during late antiquity and the early medieval period come three geographic traditions destined for popularity in the later medieval era: Macrobius (ca. 240–320), Orosius (ca. 383 to post-417), and Isidore (ca. 560–636). These traditions spawn a plethora of medieval European maps from the ninth to the fifteenth century (as discussed in the next section). Since no manuscripts are extant from the period of the original authors and the earliest extant maps date only from the ninth century onward, these maps are discussed as part of the medieval European canon not Roman.35 In “Seeing Like a Roman,” Whittaker asserts that while “the conceit of looking down on high, like Icarus or Zeus, was common among geographers,” it “was of limited use in describing the real world of travellers or administrators.”36

    Here, for lack of any other significant extant maps, ends the discussion of Greek (p.122) and Roman aspects of Encircling Ocean iconography. Given the impossibility of clear visual connections between Greco-Roman cartography and medieval Islamic maps, it is perplexing that so much significance has been attributed to the influence of the Greco-Roman tradition on Islamic mapping. A shift in this thinking is an important step forward in the study of the history of Islamic cartography. In contrast, the medieval European mapping tradition is rich in extant images, world maps and others, replete with Encircling Ocean motifs. Yet these are rarely referenced in discussions on Islamic cartography.37 What follows is an attempt to redress the narrative.

    Medieval European Mappamundi38

    Although there are differences between the forms of the Islamic and medieval European mapping traditions, such as the Islamic emphasis on the Indian Ocean, and the sizes and shapes of the continents, the one feature that stands out as a distinct mark of parity between the two mapping traditions is the Encircling Ocean.39

    The best known of all the models of medieval European world maps is the Isidorian Orbis Terrarum tradition that comes from the encyclopedic tradition of Etymologies by Bishop Isidore of Seville (ca. 560–636). Better known as a T-O map, it is so-called because of the shape of the Christian cross implied by the “T” encased within an O-shaped world (fig. 6.4). The image derives from the Mediterranean as the leg of the “T” with the Bosphorus/Tanais/Don40 on the northern end and the Nile on the southern end as the top of the “T.” The “T” is surrounded by a circular band that signifies the Encircling Ocean.41 The circle encases a schematic image of the old world (oecumene) depicting the continents of Asia, Africa, and Europe oriented eastward in the direction of Jerusalem—hence the origin of the phrase “to orient a map.”42 There are in excess of seven hundred examples of T-O maps, which by the fifteenth century were appearing in the first printed versions of Isidore’s Etymologies. The problem is that we have no copies of either Etymologies or De natura rerum contemporaneous with Isidore. The earliest dated Isidorian manuscripts that include T-O maps date from the eighth/ninth century, but even on this dating, there is no definitive consensus.43 Isidore does, however, emphasize the importance of the Encircling Ocean in his textual description of the world. In his discussion of “De oceano” in Book 13 of “De mundi et partibus” (The Cosmos and Its Parts), he says that “Greek and Latin speakers so name the ‘Ocean’ (oceanus) because it goes around the globe (orbis) in the manner of a circle (circulus), [or from its speed, because it runs quickly (ocius)]. … This is what encircles the edges of the land, advancing and receding with alternate tides, for when the winds blow over the deep, the Ocean either disgorges the seas or swallows them back.”44 In keeping with Isidore’s description, most illustrators of Isidorian manuscripts emphasized the Encircling Ocean in their T-O map renditions.45


    Classical and Medieval Encircling Oceans

    Fig. 6.4. Typical Isidorian T-O map. The “T” stands for the Christian cross and places Jerusalem at the center of the world. The “O” enclosing the entire image represents the Encircling Ocean. Late 11th century. Ink on parchment. 6.3 cm. British Library Board, London. Ms. Royal 6 C I, fol. 108b.

    © British Library Board/Robana/Art Resource, NY.

    These Isidorean T-O maps come in many variations including what has come to be known as the Y-O version (fig. 6.5). In this variation the mouth of the Bosphorus, made up of the Marmara Sea, and the Aegean combine to create a Y shape. Another variation is the so-called list T-O in which a long list of place-names is placed on each of the continents (fig. 6.6).46 It is reminiscent of Roman itinerary lists (see fig. 6.1).47 In both of these cases the Encircling Ocean persists as the outlining signifier of the world. With the onset of the Crusades in the late eleventh century, examples of the T-O model multiply and evolve into a visual manifestation of Christian zeal (p.124)

    Classical and Medieval Encircling Oceans

    Fig. 6.5. Example of a modified T-O map called a Y-O map. A more elaborate version of Isidore’s original schema with the Sea of Azov. 11.1 cm. From Isidore of Seville, Etymologies Cologne. 1478.

    This item is reproduced by permission of The Huntington Library, San Marino, California. HEH 89025.

    and a desire for world dominion. See, for example, figure 6.7, which shows the T-O map in a twelfth-century rendition of Isidore’s Etymologies in which the four directions are indicated by crosses.

    The Encircling Ocean in these later medieval European T-O renditions is dotted with islands, resembling late medieval Islamic world maps of the Idrīsī-type (see the next chapter for details). They often include multiple encircling bands that signify both the encircling ocean and the cosmos. The same occurs in eleventh-and twelfth-century examples from Bede’s De natura rerum and Lambert’s Liber floridus. Both show multiple encircling bands of the sun, moon, and planets surrounding the basic T-O model of the earth (fig. 6.8).48 This multiple encircling format resembles the prehistoric, Buddhist, Indic, and Iranian typologies discussed in the previous chapter.

    Even with the transition to complex, multilayered mappamundi in the later Middle Ages, the T-O stamp remains the underlying hallmark.49 Sometimes the T-O stamp is more obvious, at other times less so. The renowned mappamundi the Psalter world map of the mid-thirteenth century (fig. 6.9) and the Hereford mappamundi of the early fourteenth century (fig. 6.10) (p.125)

    Classical and Medieval Encircling Oceans

    Fig. 6.6. Typical list T-O map from Lambert’s Liber floridus. Reminiscent of Roman itineraries, this type of T-O map lists names of key places in the space provided for each continent. It also occurs in Isidorean manuscripts. Circa 1121. Gouache and ink on parchment.

    Courtesy: The University of Ghent. Ms. 92, fol. 19a.

    provide prime examples of this late medieval process of elaboration.50 Although varying greatly in size—by surface area the Hereford map is some 150 times larger than the Psalter map—they both share the same T-O base overlaid by an intricate wealth of detail. One striking feature of the evolution of the Isidorian T-O model into the full-fledged medieval Western European mappamundi is the increased religious signifiers that dot the surface of the map, which do not occur on the Islamic maps. These include features such as church-like place markers and references to key biblical events in the form of Noah’s Ark, the twelve tribes of Israel, biblical monsters, and Earthly Paradise in particular.51

    Both the Psalter and the Hereford mappamundi use a double-ringed band to (p.126)

    Classical and Medieval Encircling Oceans

    Fig. 6.7. Evolution of the T-O map during the Crusades from a copy of Isidore’s Etymologies. Crosses indicating the four directions betray signs of crusading influence. This T-O map is unusual in that it indicates two sets of encircling oceans. Circa 12th century. Ink on parchment.

    Courtesy: British Library Board/Bridgeman. Ms. Royal 12 F. IV, fol. 135v.

    separate the profane world from the sacred garden of Paradise. In the Psalter map, Paradise is a circled space on the main landmass at the top of the map. In the Hereford mappamundi, Paradise is no longer part of the main landmass. It is depicted instead as a double-ringed island within the Encircling Ocean (fig. 6.10). Complete with Adam, Eve, and the sacred rivers, its relocation to the Encircling Ocean emphasizes the extreme otherness and inaccessibility of its sacred space.52 In both cases, God and his angels sit above and beyond the sacred encircling circles. The distance between God and earth is stressed in the Psalter map where we see that in his left hand God is holding an orb with a miniature T-O motif while with his right he is blessing it. A similar double-ringed marker is used in other places on medieval maps to separate sacred spaces from the profane. Most notable of these is the walled city of Jerusalem. In the case of the Hereford depiction, the cross breaks the plane of the earth and rises toward heaven from the center point of the circle (fig. 6.11). Examples abound, especially from the Crusading period, of similar depictions of circular Jerusalem sacrality (fig. 6.12).53

    It is not only cities that find their way into double-ringed circles; as in the Sassanid examples discussed in the previous chapter, people of import do so too. This fits (p.127)

    Classical and Medieval Encircling Oceans

    Fig. 6.8. Planetary circles around Terrum Orbis from Lambert’s Liber floridus. Liber floridus is the earliest medieval encyclopedia containing a universal history that eventually supersedes Isidore’s work. It is lavishly illustrated and contains a variety of T-O maps. The copy at Ghent is an autograph copy in the hand of the author. Circa 1121. Gouache and ink on parchment.

    Courtesy: The University of Ghent. Ms. 92, fol. 94v.


    Classical and Medieval Encircling Oceans

    Fig. 6.9. Psalter world map. Among the most famous of the medieval European mappamundi, this tiny but extremely detailed world map depicts Christ towering over the world flanked on either side by two angels. In spite of the significant expansion of information from the simple Isidorean T-O map, it is still encircled by an double-ringed Encircling Ocean now dotted with a plethora of islands. Circa 1265. Gouache and ink on parchment. Diameter 9 cm. British Library Board, London. Add. Ms. 28681, fol. 9a.

    © British Library Board/Robana/Art Resource, NY.


    Classical and Medieval Encircling Oceans

    Fig. 6.10. Hereford mappamundi. One crucial difference between this and the Psalter map, which is some 150 times smaller but shares the same T-O base, is that in the Hereford mappamundi, Paradise is no longer shown on the easternmost edge of Asia. Instead it is detached and indicated as one of the many islands in the Encircling Ocean. Circa 1300. Gouache and ink on parchment. 158 × 133 cm.

    Courtesy: The Mappa Mundi Trust and Dean and Chapter of Hereford Cathedral.


    Classical and Medieval Encircling Oceans

    Fig. 6.11. Island of Paradise with encircling ring on the Hereford mappamundi. A double-ringed motif encircles Jerusalem suggesting that it was used to separate sacred spaces from the profane. Circa 1300. Gouache and ink on parchment.

    Courtesy: The Mappa Mundi Trust and Dean and Chapter of Hereford Cathedral.

    with the well-established Christian iconographic practice of surrounding the heads of God, Christ, prophets, and angels with halos. So ubiquitous are examples of these that there is no need to prove this, but it does raise the possibility that the halo—like the Ahura Mazdā symbol—originates in the basic double-ringed Encircling Ocean marker for the world. Depictions of rulers and religious figures within double-ringed circles date back at least to the silver plates of the Sassanid period (see fig. 5.13 of the previous chapter). The images from Lambert’s medieval encyclopedia Liber floridus—the earliest autograph edition—depicts the ruler (specifically Caesar Octavian Augustus but in fact an allusion to the ruler of the region in which the manuscript was made) with a T-O orb in one hand and a sword symbolizing earthly dominion in the other (fig. 6.13).54 Note the parallels with God in the Psalter map holding a T-O orb in his left hand (see fig. 6.9).

    As Gurevich reminds us, “The highest position is occupied by the monarch appointed by God. Just as the world (macrocosm) is directed by God, and the human body (microcosm) is directed by the soul, so is the body politic directed by the monarch, whose relation with his subjects might be compared to the relation between the head and the limbs. The power of the monarch does not depend on the will of the subjects. The monarch is subject to one God alone, whom alone he serves (rexminister Dei).”55 It is no exaggeration to say that the Liber floridus is replete with images employing Encircling Ocean forms, suggesting that the double-edged circle as a symbol of separation of the sacred from the profane was a well-established device by the twelfth century.

    Where medieval European maps craft their own unique dictum is with the “T” in the “O” standing as it does for the connection between a crucified Christ-man and the (p.131)

    Classical and Medieval Encircling Oceans

    Fig. 6.12. Jerusalem encircled with a double ring. Indicates an imagined symmetry between the shape of the earth and an idealized earthly Jerusalem, a form that gained popularity in the wake of the crusaders’ withdrawal. Circa late 12th century. 31.3 × 21.0 cm.

    Courtesy: Bibliothèque royale de Belgique, Brussels. b. Regia No. 9823-24, fol. 157a.


    Classical and Medieval Encircling Oceans

    Fig. 6.13. Ruler encased by circle of power holding T-O map from Lambert’s Liber floridus. Symbolic use of the double-ringed motif as a metaphor for power reinforced by the encircling ocean encasing the T-O map held by the ruler. Circa 1121. Gouache and ink on parchment.

    Courtesy: The University of Ghent. Ms. 92, fol. 138b.

    cosmos.56 This connection is articulated in an illustration from a twelfth-century copy of a Hildegard of Bingen manuscript. In this image of the world, encircled by multiple ripples of an encircling ocean, a distinct T form of Christ as man connects with God and Christ the Son of God. In this way the sacred firmament is connected to the earth through the corporeal body of Christ (fig. 6.14). His veins represent the course of the sacred waters that are shown cycling through the body, beginning and ending in the Encircling Ocean.57 The connection between this rapturous Bingen vision of the world and T-O maps is apparent when we compare it to an image of a T-shaped figure with outstretched arms gripping the “O” of the Encircling Ocean in (p.133)

    Classical and Medieval Encircling Oceans

    Fig. 6.14. Macrocosm, microcosm, and winds from Hildegard of Bingen’s Liber divinorum operum. World encircled by multiple ripples of the encircling ocean is connected to God through a “T” formed by the body of Christ. 13th century.

    Courtesy: Biblioteca Statale di Lucca. Cod. lat. 1942, fol. 9r. Photo credit: Scala/Art Resource, NY.


    Classical and Medieval Encircling Oceans

    Fig. 6.15. Floriated “T” with human figure gripping the “O” of the Encircling Ocean from Isidore’s Etymologies. The Encircling Ocean is stressed with six points highlighting its importance for framing the world and humans within it. 13th century. Blue, green, and yellow gouache and red and blank ink on parchment.

    Courtesy: Biblioteca Medicea Laurenziana, Firenze. Ms. Conv. Soppr. 319, fol. 90b.

    a floriated “T” for “Terrat/es” from a thirteenth-century illuminated copy of Isidore’s Etymologies (fig. 6.15).58

    Sources of the European Mappamundi Encircling Ocean Motif

    What are the sources of the medieval European T-O model with its single and multiple encircling bands? There are two likely candidates, the Macrobian and Beatus traditions, heralding from the late antiquity and the early medieval period respectively. The former is a legacy of the twilight of Roman dominion and the latter has (p.135)

    Classical and Medieval Encircling Oceans

    Fig. 6.16. Zonal image of the world from Macrobius’s Commentary on the Dream of Scipio. Circa 9th or 10th century. France. Red and black ink on parchment.

    Courtesy: Bibliothèque royale de Belgique, Brussels. Ms. 10146, fol. 109b.

    Visigothic roots. One developed during the heyday of the Patristic period (ca. 400–700 CE); the other grew out of adversity and a time of war and conflict between invading Germanic tribes and invading Muslims, giving rise to apocalyptic views of the world. Both models employed an Encircling Ocean motif and highlighted its significance through space and prominent embellishments.

    The Macrobian tradition, based on the early fifth-century commentary by Macrobius Ambrosius Theodosius on Cicero’s Dream of Scipio (51 BCE), takes a visual (p.136)

    Classical and Medieval Encircling Oceans

    Fig. 6.17. Mozarab world map from Beatus Liebana—Facundus. Mozarabic map depicts the world in an oblong shape with dramatic markers for Jerusalem and Paradise. It is surrounded by an ultramarine Encircling Ocean filled with large fish outlined in white. 1047. Gouache on parchment. 31.2 × 28 cm.

    Courtesy: Biblioteca Nacional de España, Madrid. Ms. Vitr. 14-2, fols. 63b–64a.


    (p.138) approach to Scipio’s “God Trick” of looking back at the earth from outer space while listening to the music of the spheres. Alfred Hiatt emphasizes the importance of the ocean in the maps of the Macrobian tradition. He argues that “the primary purpose of the image was to illustrate the direction of ocean flows, the formation of seas, and the relationship of the known world to unknown but hypothesized regions.”59 This emphasis is clear from the earliest extant Macrobian map dated approximately to the ninth/tenth century (fig. 6.16).60 The copyists of this French rendition reinforce the importance of the Encircling Ocean in the Macrobian vision by highlighting the ocean with squiggly lines indicating waves and currents.

    The Visigothic Beatus tradition begins toward the end of the eighth century after the lightning Muslim invasion of the Iberian peninsula, when it was rumored that the events of the Apocalypse as written in the biblical book of Revelation were about to unfold. To this end, in 776 CE, the monk Beatus prepared his Commentary on the Apocalypse. To date we know of twenty-six elaborately illustrated copies of this manuscript. It is important to note that none of the copies date to the time of Beatus.61 The Beatus maps come in two key models: square and oval. Both models display a world surrounded by a prominent Encircling Ocean.62 The example shown in figure 6.17 is from the Beatus of Liebanna and is of Mozarab provenance. Among other things, this example of a Beatus world map shows the influence of local Andalusi Mozarab artists,63 and parallel elements from this rendition can be found in other Islamic maps. (See, for instance, fig. 7.1 for parallels in the concept of gigantic fish that populate the waters of the Encircling Ocean.)

    Like the illustrated Liber floridus tradition, the Beatus manuscripts too are filled with images incorporating sacred double-ringed circles. In the “Vision of God Enthroned” in the 1220 CE Las Huelgas manuscript God is seated within a circle surrounded by the prophets, apostles, and saints.64 A similar image is that of the Lamb of God (fig 6.18) symbolizing Christ in a Heavenly version of Jerusalem surrounded by the angels.

    In his book City and Cosmos, Keith Lilley expounds on the multiplicity of double-ringed circles in everything from T-O maps to images of the cosmos and renditions of Jerusalem on earth and in the heavens. Lilley’s answer to this widespread phenomenon is the influence of Neo-Platonic thought—a key influence in medieval Islamic culture as well.65

    What we can deduce from comparing images of the Christian world with contemporary depictions and descriptions of earthly and heavenly cities is that they share a sacred geometry. Particular shapes—circles and squares, the cross and cardinal axes—connected city and cosmos, urban form and cosmological form were analagous, and their shared geometrical forms conveyed a common symbolic meaning. Both formed part of a hierarchy of concentrically ordered spaces from center to edge, and both were marked with (p.139)

    Classical and Medieval Encircling Oceans

    Fig. 6.18. Apocalyptic vision of Heavenly Jerusalem with the Lamb of God from Maius’s Morgan Beatus. Sacred hierophantic circles abound in the medieval European illustrated manuscript tradition. Circa 940–945. TaÌbara, Spain. 38.1 × 27.9 cm.

    Courtesy: Morgan Library and Museum, NY. Ms. M.644, vol. 1, fol. 87a.

    (p.140) the sign of the cross, symbolizing Christ himself. These Christian urban and cosmological imaginings, with their common structuring “geometric schema,” were drawing upon Neoplatonic and Aristotelian sources to elucidate and conceptualize the form and the order of the cosmos. … They forged a symbolic link between city and cosmos in the medieval imagination.66

    Lilley reinforces his argument with images such as the well-known one from the Bible Moralisée (fig. 6.19), which depicts God as architect, with compass in hand, measuring the earth with its distinctive blue Encircling Ocean band.67 The four-petal enclosure encasing God in the likeness of Christ of the Bodleian version resembles Indic cosmographic forms discussed in the previous chapter.68

    In an attempt to reconcile certain passages in the Bible with the known world, medieval European scholars proposed a variety of innovative interpretations that resemble other premodern religio-cultural traditions. Among the key passages of the Hebrew Bible that refer to the Encircling Ocean are the following:

    Genesis 1:6–7: “And God said, ‘Let there be a dome in the midst of the waters, and let it separate the waters from the waters.’ So God made the dome and separated the waters that were under the dome from the waters that were above the dome. And it was so.”

    Genesis 1:9–10: “And God said, ‘Let the waters under the sky be gathered together into one place, and let the dry land appear.’ And it was so. God called the dry land Earth, and the waters that were gathered together he called Seas. And God saw that it was good.”

    Psalm 148:2: “Praise him, you highest heavens, and you waters above the heavens!”69

    One gets from the Bible a sense of a primordial ocean from which the earth was created—an idea that was prevalent in the ancient period (see chapter 5 for a detailed discussion of the Judaic vision of the world in the cosmos). These verses mention not one but multiple Encircling Oceans. Evidence that medieval scholars wrestled with this idea is indicated by their rationalization of the phenomenon. Isidore, for instance, says: “Our dwelling-place is divided into zones according to the circles of the sky.”70 Explanations such as these help us link the medieval European picture of the primordial Encircling Ocean with what came before and what was transmitted after (see chapters 5 and 7).

    One of the most common presumptions was that all the waters of the earth must be interconnected, resulting in a “congregation of waters” emanating from a single source. This theory is bolstered by Genesis 2:6: “A spring rose out of the earth, watering all the surfaces.” On the basis of this, St. Augustine posited that it was via this spring that water cycled back and forth from the rivers to the ocean, and that there (p.141)

    Classical and Medieval Encircling Oceans

    Fig. 6.19. Architect of the world from Bible Moralisée. Mid-13th century. Reims, France. Gouache and gold on parchment. 24.4 × 26 cm.

    Courtesy: Oesterreichische Nationalbibliothek, Vienna, Austria. Cod. 2554, fol. 1b. Photo credit: Erich Lessing/Art Resource, NY.

    must have been a subterranean passage through which the water passed.71 (See reflections of these concepts in Judaic conceptions of the world and the cosmos as discussed in the previous chapter, in particular figure 5.16.) Bernard of Clairvaux portrayed Christ as an unending source of virtue and knowledge among men, in the way that “the sea is the source of fountains and rivers.” Likewise, Hildegard of Bingen assumed that the water of wells, springs, and rivers is derived from the ocean, which surrounds the earth.72 It is an idea that is reminiscent of the Mazdean concept of a single water source recycled between Mt. Hara and the Vurukasha Sea (refer to the discussion in the previous chapter and fig. 5.15).73

    Similarly, the waters of the universe, within which the earth was said by the Bible to be contained, were a matter of speculation. Ambrose suggested that they were (p.142) “intended to cool the axis of the universe, overheated by its perpetual rotation.” Others argued that it was intended “to screen the earth from the fiery heat generated by the stars and the sun.” Ambrose went so far as to suggest by analogy that “if the earth can hang in the center of the universe without support so also can the waters hang unsupported above the firmament.”74 Isidore extolled the virtue of water in a like manner:

    The element of water rules over all the rest, for water tempers the sky, makes the earth fertile, gives body to the air with its exaltation, ascends to the heights, and claims the sky for itself. Indeed, what is more amazing than water standing in the sky? And it is not enough that it reaches such a height, but it snatches a school of fish with it and when poured out becomes the cause of all growing things on earth. It brings forth fruits and trees, produces shrubs and grasses, cleans away filth, washes away sins, and provides drink for all living creatures.75

    All this water, above, below, and beyond, necessitated a concept of a ball-like earth bobbing on the surface of an all-encompassing body of water like the Encircling Ocean. So ubiquitous, in fact, is the Encircling Ocean in medieval European motifs—mappamundi and otherwise—that in her analysis of the Hereford mappamundi, Naomi Kline devotes an entire chapter in her book Maps of Medieval Thought to the subject of “the circle as a conceptual device.”76 She asserts that “that mappae mundi … functioned in a way similar to medieval rotae, flat, circular wheels that made understandable, in a simplified graphic manner, concepts that explained the way the world worked. … In effect medieval mappae mundi share with medieval rotae the circular format as containers of diverse information. Our consideration of the medieval mappae mundi is based upon the idea of the circle as a shorthand device for assisting and learning and memory and situates medieval mappae mundi within the medieval realm of wheels of memory.”77

    In an extensive archaeology of Encircling Ocean–like images, Kline shows that the concept of a world within a multiplicity of heavenly spheres was connected to the medieval European imagining of the cosmos. She elucidates her thesis by marshaling the evidence of multiple rotae-type cosmographic images that were found in monastic schoolbooks such as William of Conches’s Summa magistri Wilhelmi de Conches and Gauthier of Metz’s Le Romounce del ymage du monde.78 Indeed, Kline’s entire Maps of Medieval Thought is a contemplation of the primacy of double-ringed circles in medieval Christian illuminated manuscript art.79 She does not, however, make the connection between the Encircling Ocean and the symbolic stamp of the imago mundi80 as the origin of this medieval Christian obsession. Nor does she take the route of Lilley and others who see it as a Timean reflection of Neo-Platonic influence.81


    Classical and Medieval Encircling Oceans

    Fig. 6.20. Macrobian-like world map from al-Sijzī’s Tarkīb al-aflāk. The Northern Hemisphere has a faint latitudinal grid inked in red. Circa 12th century or earlier. A reader’s note on the opening folio indicates a date of 846/1442. Diameter 13.25 cm.

    Courtesy: Leiden University Libraries. Cod. Or. 2541, fol. 21b.

    Islamic and Medieval European Mappamundi Connections

    Just like the medieval Islamic world maps, all medieval European mappamundi contain the same dominant feature: an Encircling Ocean that frames the image of the world—ergo imago mundi. Did medieval European maps influence the Islamic ones or vice versa? Or were they mutually exclusive?82 The Mozarabic maps of the Beatus tradition, discussed above, hint at a cross-pollination of decorative motifs at least within the artistic milieus of Islamo-Christian Spain. There are also some Islamic maps reminiscent of medieval European Macrobian models that show a fourth southern continent.83 One such example is found at the end of an incomplete Sijzī astronomical (p.144) manuscript, Tarkīb al-aflāk (Composition of the Spheres), housed in the Oriental manuscript collection in the library at Leiden (fig. 6.20), recently redated to between 1041 and 1165.84 The map is labeled in Arabic even though the manuscript is in Persian. It shows a fourth continent separated from an overextended Africa by a large body of water. The main landmass is broken up into distinct segments for North Africa, Egypt, the Arabian Peninsula, Sind, Sarandīb (Sri Lanka), which is unusually shown as attached to the South Asian mainland, and Central Asia with its customary Turkic tribes and requisite rivers (Jayhūn and Sayhūn), ending with the landmass of al-Andalus at the northwestern edge.85

    In addition to the unusual world map from the Persian astronomical manuscript discussed above there is another surprising T-O map from an eighth-or ninth-century Isidorian manuscript that points to Islamo-Christian transcultural cartographic connections:86 a basic Isidorian T-O world map labeled in Arabic (fig. 6.21).87 For the first time we have proof of a definitive, early medieval tie between the two cartographic traditions.88 The Arabic notations placed on the map in a rough list T-O fashion (compare with fig. 6.6) record the names of key Muslim places on each of the three landmasses, such as Mecca (Makka) and Medina (Yathrib) on the Levantine side, the Berbers and the Ḥabasha (Ethiopians) on the African flank, and Gog and Magog along with al-Rūm (Byzantium) and al-Andalus (Islamic Spain) on the European side. Equally interesting are the notations in the Encircling Ocean, which give the dimensions of this ocean on three flanks. At the top of the circle, we read in Arabic that the earth is collectively 24,000 farsakhs, and in Latin that it is 180 furlongs.89

    The Arabic commentary ranges from the names of the landmasses to a listing of place-names to ecumenically related symbolic narratives. Aside from the heavily annotated map, there are marginalia comments in Arabic scattered throughout this manuscript beginning early on folio 4v and occurring throughout the manuscript to the final colophon folio of 163v. This map confirms something that historians of medieval Islamic history have long sensed but could not prove: namely, that Arabic speakers in western parts of the Islamic world not only knew about medieval European maps but interacted actively with them. It suggests that the Arab reader and owner of this manuscript took an active interest in the geographical material of this manuscript. He marked folios 70–150 with names of places in Arabic that match the Latin text, and he took enough of an interest in the T-O map to add his own personal notations.

    We can further deduce that at some point this manuscript returned to Latin hands. The notation above the top of the “T” in Latin was written after the Arabic notations. Furthermore, the Arabic notations themselves are in two different hands: one for the internal regions of the map and the other for the external Encircling Ocean band listing the measurements. We cannot, however, be sure of the date either of this manuscript, its T-O map, or the Arabic notations. Latin experts have pegged this manuscript vaguely as sometime between the eight and the ninth centuries. Oddly, (p.145)

    Classical and Medieval Encircling Oceans

    Fig. 6.21. T-O map with Arabic notations from a copy of Isidore’s Etymologies. Circa 8th–9th century. Ink on parchment.

    Courtesy: Biblioteca Nacional de España, Madrid. MS. Vitr. 014/003, fol. 116v.

    until recently historians of medieval European cartography mentioned neither this map nor this manuscript in their accounts of medieval European mappamundi.90

    This map is nothing short of stunning both for specialists in Islamic cartography and medieval European cartography and for the field of history of cartography in general. It implies an early intersection between the medieval European and Islamic mapping traditions (at least in the Maghrib).91

    (p.146) Transcultural Cartographic Drops

    At the XVIIIth International History of Cartography conference in Athens, in July 1999, Christian Jacob gave a visionary paper on the subject of transcultural influences on Hellenistic Greek cartography in which he asked: “If I pour some drops of Chinese cartography on Greek Hellenistic cartography, how does the latter react? What does the Chinese case reveal when applied to the understanding of Hellenistic cartography?”92 Sitting in the audience, I found myself thinking: “If you add additional drops from the Indian, Iranian, Mesopotamian, Egyptian, and Central Asian cartographic traditions, then you have Islamic maps!”

    This broad yet necessarily cursory transcultural survey of the Encircling Ocean in this and the previous chapter reveals that it was a widespread motif best classified as a metaform—nurtured by the observation that no matter what direction one travels in, one eventually reaches the sea. It emerges as an important component of water-based hylogonies—namely, cosmogonies that are based on water as the source of life and death of the world.93 The Encircling Ocean becomes the basic outline stamp of all premodern imagines mundi and turns into a useful signifier of power and domination for ruling groups seeking an easy stamp of world dominion.

    What this iconography seeks to prove is that the “Encircling Ocean” feature of Islamic maps cannot be exclusively pegged to the Greeks, Ptolemy, or Neo-Platonists. Rather, it is the most common of transcultural metamotifs of premodern world maps. My purpose is not to deny the connections between Greek and Islamic cartography but only to set them in a metacultural perspective. Greek influence is present, but so are Indian, Iranian, Central Asian, Chinese, Egyptian, Roman, and medieval European influences.

    The Muslims retained the Encircling Ocean in their maps as a metacartographic feature, but to understand how they thought about it, we need to examine their textual descriptions of this timeless ocean. Did the Muslims share the ancient Iranian, Indian, and early Semitic view that all the water in the world stemmed from a single source linked to the Encircling Ocean? Did they believe in one or multiple Encircling Oceans? Can the primary texts help us to determine the specific cultural sources for their ideas on the ocean, or are they as metaculturally diverse as the images? Only a close reading of some of the key Arabic and Persian sources can help us fully address the significance of the Encircling Ocean in the Islamic cartographic imagination.


    (1.) There are only two extant examples of ancient Greek drawings from the classical period: the so-called Romance Papyrus (Bibliothèque nationale de France, Paris, Co. Suppl. Gr. 1294) and the Heracles Papyrus Oxford, Sackler Library, Oxyrhynchus Pap. 2331). In the 1990s, researchers discovered sketches on a papyrus fragment known as the Artemidorus Papyrus, now housed at the Museo Egizio in Turin. Views on this papyrus range from genuine to nineteenth-/twentieth-century fraud. The maplike sketches are so vague and fragmentary that they are hard to identify. Guesses range from all Spain to part of it, to possible details of the river Ebro or a view of Huelva in southwest Spain. See Brodersen and Elsner, “Artemidorus Papyrus,” in particular Talbert, “The Map,” 63–64.

    (2.) Reconstructions of ancient Greek maps on the basis of textual sources are problematic because we cannot avoid the influence of present-day cartographic knowledge. Textual readings of the ancient Greek sources suggest that there was significant dispute about the size of the continents. Some scholars presumed Europe, for instance, to be much larger than Africa and Asia, but this is not reflected in the reconstructions of Bunbury and others. Bunbury, Ancient Geography, 1:148. Richard Talbert concurs with my view on the matter of reconstructions. See Talbert, “The Map,” 57.

    (4.) Urbinas Graecus 82, Bibliotheca Apostolica, Rome. Dated to the late thirteenth century, it is thus far the earliest known copy of Ptolemy’s Geography.

    (5.) There is a voluminous literature on both sides of this debate. See, for instance, Polaschek, “New Light”; and Bagrow, “Origin.” The suggestion that Ptolemy had to draw a map first in order to compile his eight-thousand-strong latitudinal and longitudinal listing of places is one of the many strange arguments used to justify the thesis that Ptolemy accompanied his (p.319) treatise with long since lost maps. A. Jones, “Ptolemy’s Geography,” 122–23. Aside from the fact that it would have been well nigh impossible for any scholar prior to satellite mapping and Google Earth to construct a map—or even a series of maps—with eight thousand places, Muslim scholars in the medieval period filled many volumes (called zīj tables) with the latitudes and longitudes of thousands of places. They did not use maps to compile these tables, so why Ptolemy would have needed to do so is puzzling. For more on zīj tables, see the work of David King, in particular EI2, s.v. “Zîdj,” by D. A. King and J. Samsó. Sezgin includes a comprehensive overview on the debate surrounding the thorny question of whether Ptolemy’s Geography was accompanied by maps in which he suggests that it would not have been possible to draw the rivers on Ptolemaic maps because of the lack of sufficient data points and that the source of the extant models only date back to the fourteenth-century cartographic interpretations of the Byzantine monk Maximos Planudes. Sezgin, Mathematical Geography, 1:38–56.

    (6.) See Aujac, Claude Ptolémée. Albeit a terrific article in other respects, Alexander Jones falls prey to the pitfall of using Renaissance maps to explain what Ptolemy’s second century CE maps would have looked like. Jones, “Ptolemy’s Geography,” 123–27.

    (7.) For illustration, see Willcock, Companion, 210. Georgia Irby discusses this shield too although she avoids reconstructing it. Irby, “Mapping the World,” 87–88.

    (9.) Cosgrove, Apollo’s Eye, 29–36. This shield captured the Western imagination from the Enlightenment onward, and to date, elaborate reconstructions of it are still being produced. Gentleman’s Magazine, 19 (1749), 392, engraving used by Cosgrove, Apollo’s Eye, 33. See also the thread “Achilles’ Shield” on the Pirates and Revolutionaries blog, which provides a rich array of the many interpretive renditions to date (http://piratesandrevolutionaries.blogspot.com/2009/03/achilles-shield-shield-of-achilles-as.html, accessed March 16, 2015).

    (10.) Yāqūt also tells us that his fellow geographers were aware of other ancient notions of the shape of the world, such as a table, a drum, a dome, and a stone cylinder or pillar, although he does not tell us the exact ancient sources of these ideas. Yāqūt, Muʿjam, 1:31; Yāqūt, Introductory Chapters, 19–20.

    (12.) On the basis of this 423 BCE comic portrayal of the political impact of maps, Irby argues that “large scale maps were known in Athens from the fifth century onward, and they were symbolically powerful.” Irby, “Mapping the World,” 81–82.

    (13.) Ibid., 82.

    (15.) Aristotle, Meteorologica, 2.5.362b.13; and the Stoic philosopher Geminus (first century BCE) followed suit. Geminus Introduction to the Phenomena, 16.4.5.

    (19.) I thank my colleague Dr. Brett Rogers, former member of the Classics Department at Gettysburg College, now at the University of Puget Sound, for this crucial grammatical clarification.

    (20.) Separating what was just a textual account or a reference to an actual journey from (p.320) what was actually a map is yet another example of how hard it is to analyze Greek maps, if one cannot know if these “circuits around the world” actually existed in visual form. For an excellent discussion on the linguistic problems involving periodos ges, see Romm, Edges of the Earth, 26–31. Georgia Irby suggests that these geographic treatises may have been accompanied by maps. But she concurs that in the absence of extant examples, “we cannot know how many Greek maps were produced, or what exactly their content and purpose may have been.” Irby, “Mapping the World,” 83–84.

    (21.) See, in particular, the voluminous and imaginative work of Christian Jacob. Two examples among many are “Inscrire la terre habitée”; and “La carte du monde.” For an overview of the Greek mapping tradition, see the two stimulating articles by Irby and A. Jones in AP (“Mapping the World” and “Ptolemy’s Geography,” respectively).

    (22.) See note 5 above.

    (23.) EI2, s.v. “Ḵẖarīṭa (or Ḵẖāriṭa),” by Maqbul Ahmad.; Tibbetts, “Beginnings,” 94–95.

    (24.) For a detailed discussion on this matter, see chapter 3.

    (25.) EI2, s.v. “Ḵẖarīṭa (or Ḵẖāriṭa),” by Maqbul Ahmad. There is no clear agreement on this matter. Kramers maintains that the primary influence was Iranian; Barthold, preface, 9, argues for Greek. Nazmi proposes that Iranian influence held sway up to the twelfth century and that Greek took over thereafter. Nazmi, Muslim Geographical Image, 75–77. Nazmi is, however, basing his post-twelfth-century thesis on the basis of the work of Idrīsī, Ibn Saʿīd, and Qazwīnī. When, in fact, KMMS maps after the twelfth century display more Iranian influence rather than less.

    (27.) See Dilke, “Service of the State,” 201–11. For a specific discussion on the Severan Marble Plan of the imperial city of Rome, carved in 203–211 CE, see Trimble, “Severan Marble Plan.”

    (28.) See, for instance, the stone Dilke identified as a Roman map of Gaul. Dilke, “Service of the State,” 207. For further examples of questionable cadastral surveys, see Dilke, “Roman Large-Scale Mapping,” 213, 219, 221, 223–31. Talbert objects strongly to Dilke’s identification and labels it spurious. For a flavor for the debate between scholars on this matter, see Talbert, “Twenty-First Century Perspectives.” See also Bekker-Nielsen, “Terra Incognita,” 155–57, for further support of Dilke’s argument.

    (31.) Talbert, Rome’s World, 133–36. Disagreeing with Talbert, Emily Albu asserts that the Peutinger map is actually a Carolingian product. See Albu, “Imperial Geography”; and Albu, “Rethinking the Peutinger Map.” In “Urbs Roma,” 182–87, Talbert imagines a possible location for the Roman prototype of the Peutinger map.

    (32.) Dilke, “Service of the State,” 207–8; Berthon and Robinson, Shape of the World, 27; Talbert, “Twenty-First Century Perspectives,” 13–14, where he discusses the debate between Dilke and Janni and Kai Brodersen’s book Terra Cognita, in which Brodersen argues that the Agrippa map was never a world map but only an extensive text, res gestae. With the publication of Rome’s World, two years later in 2010, Talbert goes further and states definitively—as if the Janni and Brodersen dispute never existed—that “large ‘world’ maps were unquestionably to be found on public display in Rome and elsewhere, and as such were of potential value for the making of the (p.321) Peutinger map’s original. Most famous of its type was the map of the orbus terrarum commissioned by M. Vipsanius Agrippa.” Talbert does not explain what caused him to switch to such a definitive assertion for the existence of the Agrippa map. Talbert, Rome’s World, 136–37. For a different approach, see Nicolet, Space, Geography, and Politics, 95–114 and 167–221.

    (33.) Eumenius translated by Richard Talbert, Rome’s World, 138; Talbert’s brackets; emphasis my own.

    (34.) For an interesting discussion and analysis of the dating and provenance issues, see Carder, Codex Arcerianus A, Wolfenbüttel, 1–35. See also Campbell, Roman Land Surveyors.

    (35.) See Woodward, “Medieval Mappaemundi,” 299–301, for an overview of these three traditions.

    (37.) Edson and Savage-Smith’s book, Medieval Views of the Cosmos, is to date the only attempt to juxtapose Christian and Islamic mapping traditions. Even in this case there is less of an attempt to study one tradition in light of the other and more of a straight juxtaposition of chapters on each tradition separately. Still the authors are to be credited for making an effort to work together and for breaking ground on this important subject. I have conducted extensive research on this subject and intend to publish my findings in a book tentatively titled Islamo-Christian Cartographic Connections.

    (38.) The literature on the medieval European mapping tradition is voluminous. As such, this section cannot be a comprehensive recapping of the medieval European tradition. I have aimed at bringing to the attention of the reader relevant highlights.

    (39.) There is only one known medieval European world map example that does not feature an Encircling Ocean: the diagrammatic Y-O Isidorian mappamundi sketch on the inner fly of the St. Gall manuscript. Even with this map, it is possible to argue that the thick encircling black line was intended to be representative of the Encircling Ocean. The image resembles the work of a student scribbling what his teacher was explaining in class. Discussions on this St. Gall map and manuscript are too numerous to list completely. For starters, readers may consult the reference to it in Woodward, “Medieval Mappaemundi; and for a recent discussion on the manuscript, see Lozovsky, “Medieval St. Gall,” 65–82. Viewers can examine the image online courtesy of the Stiftsbibliothek: St. Gall, Codex 237, fol. 1r, http://www.e-codices.unifr.ch/en/csg/0237/1/small (accessed March 16, 2015).

    (40.) Some scholars refer to this as the Tanais or Don River, but given the enormous importance of the Bosphorus in comparison to the rivers, along with the fact that the Byzantine empire straddled it, it seems logical to presume that this was customarily meant to represent the Bosphorus.

    (41.) The best discussion on the meaning of the “T” symbol in T-O maps is by J. T. Lanman, “T in T-O Maps.”

    (42.) Evelyn Edson traces the T-O format back to the Roman geographers Sallust and Lucan. But since the manuscripts containing maps that she uses for her argument date to the eleventh and fourteenth centuries, the argument is hard to prove. Edson, Mapping Time, 18–35.

    (43.) I speak here of the T-O map in the Escorial manuscript, R. II. 18, and of the palimpsest map in the St. Gall manuscript. Regarding the Escorial manuscript, Williams, “Isidore, Orosius,” 13, argues that the ink matches the text and that it should be dated to a few decades (p.322) after the death of Isidore in 632. Dalché, on the other hand, argues in “De la glose à la contemplation,” 749–52, that all three T-O diagrams in this manuscript belong to a later addition. The T-O map in the Gall manuscript is subject to a similar dating controversy. The form of this map is a Y-O map, which is considered a later innovation. Evelyn Edson returns to the problem in “Maps in Context.” As with the dating of medieval Islamic manuscripts, the dating of the early medieval European maps is complicated by the lack of clearly dated colophons, causing the dates assigned to vary significantly from one publication to another.

    (44.) Isidore, Etymologies, 277; translator’s parentheses and brackets.

    (45.) I say most because the palimpsest Y-O map in the Stiftsbibliothek St. Gall manuscript (Codex 237, fol. 1r) does not include a double-lined circle for the Encircling Ocean. See earlier discussion in note 39 above.

    (46.) For more examples and a further discussion on List T-O maps, see Edson and Savage-Smith, Medieval Views of the Cosmos, 41.

    (47.) Derolez suggests influence from the Late Antique cosmography of Ps.-Aesthicus. As noted by Derolez himself, the copy that Lambert probably consulted is an eleventh-century copy that is still in the library. Derolez asserts the influence of the Late Antique work of Orosius’s Historiae adversus paganos but does not comment on the existence of a Late Antique copy that Lambert could have consulted. Derolez, “Liber floridus,” 48–49.

    (49.) See, for instance, Dalché, “Maps in Words,” 227.

    (50.) Both these maps have been the subject of studies too numerous to list here. For a recent summary on the central place that this Psalter map occupies in the canon of medieval European visual studies and the many studies on it, see Whittington, “Psalter Map.” For the Hereford mappamundi, see HWM.

    (51.) Contrary to naming practices, medieval European cartography is heavily laced with ecumenical symbols whereas Islamic cartography, in spite of the name, has no overt religious symbols. I have discussed this in “Traces of the Diabolic and the Divine in Islamic Maps,” and “What Is Islamic about Islamic Maps?”

    (52.) For an in-depth discussion of the Hereford map, consult Scott Westrem’s place-by-place analysis of the layout in Hereford Map, 34–37.

    (53.) These examples are too numerous to enumerate. For more information, see von den Brincken, “Jerusalem on Medieval Mappaemundi.” For an intriguing study on the representation of Jerusalem from the Old Testament to the Beatus manuscripts and Crusader representations, see Kühnel, Earthly to the Heavenly Jerusalem. Especially useful also is Kühnel’s Real and Ideal Jerusalem. Keith Lilley discusses the connections between the heavenly depiction of Jerusalem and Jerusalem on earth at length. He too sees the connection between the double-ringed circular marker for Jerusalem, the world, and the cosmos. See, in particular, Lilley’s chapter “Urban Mappings: From the Heavenly to the Earthly City,” in City and Cosmos, 15–40.

    (54.) Another example from a Lambert manuscript: “Ruler Encased by Encircling Ocean,” Lambert, Liber floridus, ca. thirteenth century, Paris, Bibliotheque Nacionale, Ms. Latin 8865, (p.323) fol. 45b. Albert Derloez, the reigning expert on this Ghent autograph copy of the Liber floridus, tells us that the inscriptions allude to the world historical significance of Augustus and his role in the creation of world peace through his edict ordering a census and the putting down of the civil wars. He states that “the map in the ruler’s hand, replacing the customary sphere of the world, constitutes another link between history and geography.” Derolez, Liber floridus, 126.

    (55.) Gurevich, Categories, 163. To show the influence of the double-ringed circle, Lilley brings to our attention medieval English municipal seals situated within imago mundi–type stamps. Lilley, City and Cosmos, 22. Kline raises similar examples with the seals of Edward II. Kline, Maps of Medieval Thought, 59.

    (56.) Charles Burnett explores the “strangeness of Hildegard of Bingen’s cosmology” in “Science of the Stars.”

    (57.) For a discussion of this image and other images in Hildegard manuscripts, see Caviness, “Hildegard as Designer.” See also Denis Hüe’s interesting discussion on imago mundi images that incorporate anthropomorphic forms in “Tracé, écart,” 134–35.

    (60.) Ibid., 166–68. Hiatt discusses the problems of dating and developing a stemma for the tradition.

    (61.) For details on this debate, see the most up-to-date article on it by the reigning expert on the subject of the Beatus tradition: Williams, “Isidore, Orosius.”

    (62.) Many books on the history of cartography include reprints of Beatus maps. By far the best and most comprehensive reprint of all the Beatus manuscripts worldwide is the five-volume set by John Williams, The Illustrated Beatus.

    (63.) For an excellent discussion on this map and other geographical Mozarabic manuscripts, see Aillet, Les Mozarabes, 167–69.

    (64.) To see this image, please consult http://www.themorgan.org/collection/Las-Huelgas-Apocalypse/13 (accessed May 2, 2015).

    (65.) For starters, see Netton, Muslim Neoplatonists; Morewedge, Neoplatonism; and Adamson and Taylor, Arabic Philosophy. There is no question that medieval Islamic maps are just as affected by Neo-Platonist ideas as medieval European ones. The Encircling Ocean is the most obvious manifestation, but there are others. This book is not the place to delve further into aspects of Neoplatonic influences on Islamic cartography. I do, however, have plans to explore the impact of Neoplatonic thought on medieval Islamic mapping at a future date. I touched on some of these Neoplatonic overtones in an honorary talk that I gave in the Virginia Garrett lecture series at the University of Texas, Arlington, in October 2006, entitled “Traces of the Diabolic and the Divine in Islamic Maps.”

    (67.) Ibid., 37–40 and 126. On the one hand, Lilley argues that this image of “God as Architect” reinforces the Neo-Platonic Timean one. On the other hand, Lilley argues that “the compass as an instrument used by God in Creation to trace out cosmological form would seem to owe more to medieval Hebrew sources” (126). This dichotomy mirrors some of my findings in this and the previous chapter.

    (69.) Holy Bible: New Revised Standard Edition with Apocrypha [NRSV] (Oxford: Oxford University Press, 1977). All subsequent biblical scripture references are to this verson.

    (72.) Ibid., 200 (Bernard of Clairvaux), 201 (Hildegard of Bingen).

    (73.) Wright does allude, albeit briefly, to possible connections between ancient Egypt and Persian cosmology, but no one has followed up on this. Ibid., 58.

    (74.) Ibid., 59.

    (77.) Ibid., 10–13.

    (78.) Ibid., 26–29.

    (79.) Readers interested in further exploring the double-ringed circle in the medieval Christian imagination should consult Kline, Maps of Medieval Thought.

    (80.) I use “imago mundi” as the way to describe the symbol even though, as I acknowledge in chapter 5, note 2, there are some occasions in which “imago mundi” was used to title a book on geography of the world. I do not believe that the exceptions should preclude the use of the term to label a symbol that most readers familiar with the history of cartography would recognize.

    (82.) In “Géographie Arabe et Géographie Latine,” Dalché argues for the latter. He seems not to be aware, however, of the Isidorian world map with Arabic notations discussed later in this section. Nor does Dalché take into consideration cartographic connections outside of the twelfth century.

    (83.) For an example of a Macrobian world map showing multiple southern continents, see Hiatt, “Map of Macrobius,” 152, fig. 3, and 157, fig. 5.

    (84.) Radiocarbon dating by the DFG-ANR Project Coranica, 2015, at the request of Karen Pinto and Arnoud Vrolijk, curator of the Oriental Manuscripts and Rare Books Special Collections, at Leiden University Libraries. Study of this map and the manuscript within which it is housed is part of an ongoing project.

    (85.) This manuscript is incomplete and therefore doesn’t contain a colophon. The notation on the opening folio, however, indicates two dates: one of AH 846/1442 CE and the other, more significant one that lists the name and author of this manuscript (Kitāb tarkīb al-aflāk by Sijzī) and another book, Kitāb al-talwīhat (Book on the Intimations), by the famous philosopher and founder of the philosophical school of Illumination, Shihāb al-Dīn al-Hakīm, known as al-Maqtūl, with the date 578/1191—the year that he was put to death by the Ayyubid sultan Ṣalāḥ al-Dīn’s son al-Malik al-Ẓāhir, the viceroy of Aleppo. This suggests that the manuscript was completed sometime before the late twelfth century. It is highly unusual for a map to be included in an astronomical manuscript. This is in addition to the fact that the text of the main manuscript of Sijzī’s Tarkīb al-aflāk is in Persian but the map is annotated in Arabic. Little is known about Sijzī other than that he was an astronomer especially interested in the problems associated with the measurement of spheres. Pinna reprints this map and devotes a page of discussion to it but does not consider the possible Macrobian connections. Pinna, Il Mediterraneo, 1:41–42. Pinna incorrectly lists the date of Leiden’s Sijzī manuscript as 1248 CE based on a misreading of one of the dates on the opening folio.

    (86.) The term “Islamo-Christian” comes from another one of Richard Bulliet’s seminal contributions to the field: Islamo-Christian Civilization.

    (87.) This map was first studied in the context of medieval European cartography in 1965 by Pidal, in Mozarabes y asturianos, and introduced two years later to the Arabic-speaking public by Muʾnis in Tarīkh al-jughrāfiya, fig. 10. Then it was forgotten for half a century. I first stumbled upon this map through the 2010 Guggenheim fellowship announcement for Simone Pinet’s forthcoming research and book The Task of the Cleric: Three Studies on the “Libro de Alexandre”(in progress) (http://www.gf.org/fellows/all-fellows/simone-pinet/, accessed May 22, 2015). This map is a truly exciting renewed find that is destined to open up the discourse on medieval Christian and Muslim cartographic connections as is evidenced by the recent outpouring of interest and writing on the piece: Chekin, Northern Eurasia, 59–61; Aillet, Les Mozarabes, 167–70; and Schröder, “Kartographische Entwürfe,” 268–76.

    (88.) This finding runs counter to Dalché’s skepticism in “Géographie Arabe et Géographie Latine” regarding connections between the two traditions. Aillet places the manuscript within the Mozarabic culture of Andalus, and Schröder follows suit. Both argue that the annotations were done by Arabic-speaking Christians who were trying to introduce Latin texts into the Ibero-Islamic environment of Andalus. Aillet, Les Mozarabes, 167–73; Schröder, “Kartographische Entwürfe,” 274–76. Having examined the marginalia notations of place-names and other comments along with the colophon of this manuscript, I think there is more to the author of the Arabic notations than the average Mozarab trying to fit in with the new Arabo-Islamic elite of southern Spain.

    (89.) Detailed templates translating and transcribing the Arabic are in preparation and forth-coming in the author’s separate essay on this map, titled “A 9th Century Isidorian T-O Map Labeled in Arabic.” Farsakh is a Persian word of Parthian origin referring to the measure of distance based on time. Originally akin to the “marching mile,” the distance has been deteremined to be approximately 6 km. We should keep in mind that this was a guesstimate of the marching mile that must have varied over time, especially with changing material conditions. See EI2, s.v. “Farsakh,” by Hinz.

    (90.) Chekin rectifies this half-century hiatus and brings this map back to the attention of historians of medieval European cartography in his Northern Eurasia, 59–61.

    (91.) Further collaborative research is underway with Visigothic Latin specialists. This is part of a larger book project on connections between medieval European and Middle Eastern maps tentatively titled Islamo-Christian Cartographic Connections.

    (93.) For more on this, see, in particular, Eliade, “Waters and Water Symbolism,” in Patterns, 188–213; and Eliade, “Symbols and History,” in Images and Symbols, 151–78.