The Discovery of the Mecca-Centred World-Maps

David A. King

The two world-maps, hereafter referred to as A and B, became available for study in 1989 and 1995, respectively. My research on these maps progressed in various significant stages, and it seems worthwhile to outline these.

First, in 1991 I conducted a preliminary examination of A without analysis of the geographical data and adopting the false assumption that an accurate cartographic solution was not possible within the context of medieval science. (I had not noticed the fact that the distance-scale on the alidade is non-uniform.) The results are summarized in the Nuremberg GNM 1992-93 Exhibition Catalogue (submitted 1991, published 1992).

Second, in the Spring of 1994 I conducted a detailed analysis of the geographical information on A and of the mathematics underlying the grid. The conclusions of the former were first that the piece was one of a series, and second that the geographical data was several centuries older than the piece itself. Some of the data goes back to the mysterious Kitāb al-Aṭwāl wa-ʾl-ʿurūḍ li-ʾl-Furs, an anonymous work of uncertain date which predates the mid-13th-century tables of Naṣīr al-Dīn al-Ṭūsī since the latter rely heavily on it. But the table underlying the map, which I had identified in a manuscript preserved in London of an Arabic treatise compiled in Baghdad in 1751/52, is at least three centuries older than that treatise and at least two centuries older than the maps themselves.

With the help of computer-generated reconstructions it proved possible to explain how the various markings were constructed, and to demonstrate that the grid, although based on some approximate procedures, functioned accurately for all intents and purposes except on the outer edges (al-Andalus and China).

It also seemed reasonable to suppose that this map was part of a tradition that had endured over several centuries, and it was not hard to find other implicit evidence of that tradition. The geographical data of al-Khāzinī (giving longitudes and latitudes rounded from those of al-Bīrūnī to the nearest 10ʹ as well as qibla-values rounded to the nearest 10ʹ) appeared to satisfy the necessary requirements. The latitudes and longitudes must – so I thought – have been read from a map, and the qibla-values read from the same map, for they are not particularly accurate and are less so at the outer edges. It was – I would still maintain – not unreasonable to suppose the existence of a Mecca-centred map prepared by al-Bīrūnī and used by al-Khāzinī to compile his list, this not least because some of the qibla-values corresponded closely to what one could derive from a map such as A. An account of my findings was presented at the Symposium ‘Science and Technology in the Turkish and Islamic World’, Istanbul, 3-5 June, 1994, and a summary of my conclusions was published in the article “Samtˮ in the Encyclopaedia of Islam(submitted 1994, published July, 1995).

Third, in February. 1995, I had learned from Dr. Tofigh Heidarzadeh of Tehran that the French Capuchin Raphaël du Mans (d. 1696), who was resident in Isfahan for some 40-50 years, had presented a remarkable “compass after the manner of an astrolabeˮ to Shāh ʿAbbās II. Might Abe this “compassˮ? he enquired. No, I replied, for Alies fully within the Islamic tradition, and no European would so have misrepresented the positions of localities in Europe (Cordova is placed slightly north-east of Tripoli, Libya, and Athens is south-east of Constantinople). I obtained the original references in the diary of Jean-Baptiste Tavernier, and also noted that Raphaël du Mans had such a negative attitude towards Islam that he would hardly have taken the trouble to devise a world-map that would be of such potential use to Muslims. By this time I had examined a good number of other Safavid astronomical instruments, and the prime candidate for the maker of world-map AI thought to be ʿAbd al-Alī, who made the superb astrolabe for Shāh Ḥusayn in 1712 that is now in the British Museum. Isfahan seemed an obvious choice for the provenance, being the Safavid capital in the late 17th century and the location of most of the known Iranian instrument-makers during that period.

Fourth, in April, 1995, I reinvestigated the geographical data of al-Khāzinī in an attempt to relate his qibla-values to what one might expect from a map with a grid of the same kind as that on A, and found – to my great surprise – that al-Khāzinī’s qibla-values are more readily explained if one assumes that he derived them by using a table displaying the qibla as a function of longitude and latitude and latitude difference from Mecca that was poorly computed. Thus the hypothetical Mecca-centred world-map of al-Bīrūnī disappeared as fast as a computer-screen of calculations appeared before my eyes. (Or to put it another way, if al-Bīrūnī did prepare such a map we have not yet uncovered any trace of it). It is still incontestable that al-Khāzinī read the longitudes and latitudes from a world-map of al-Bīrūnī; we simply do not know what kind of grid that map had. On the other hand, I did ascertain that we can indeed take the theory of projections preserving direction and distance back to al-Bīrūnī, and even byond to a Muslim astronomer of the 9th century, Ḥabash al-Ḥāsib.

Fifth, in May, 1995, as I was still groping for a solution to the problem of the origin of map A, I learned that another world-map of the same kind (B) had surfaced. I immediately went to visit its new owner to compare the piece with photographs of Aand my earlier analysis of the geographical data on it. My conclusion was that Bis some decades later than Aand that Awas probably not made by the same person who signed B, namely, one Muḥammad Ḥusayn, otherwise unknown to the modern literature. The engraving on the two pieces is virtually identical, but that on Bis less precise and less elegant than that on A. And the grid on Ais engraved with more confidence than that on B, which also reveals more errors of construction than that on A. On the other hand, the agreement of the positions of localities on Bwith the ultimate source (the table of which we have a late version in the London manuscript discovered in 1994 that I mentioned above) is slightly better than for A.This notwithstanding, and whilst the bases for both maps are very similar, down to the dimensions, and even the screws, the whole character of the maps is different, Abeing clearly Safavid and Bat least lookingmuch later. It was also obvious that the two maps were independent of each other in detail (choice of localities), different errors occurring in each. This raised the possibility of there having been several of such maps prepared during the 17th and 18th centuries, if not also before. The underlying principle, I still maintained, is older than either of the surviving maps, and maybe older than the underlying geographical data, which in the form in which it was applied to the maps does not predate the 15th century.

Sixth, in June, 1995, a summary of my findings up to that time was presented to the International Colloquium ‘La science dans le monde iranien’ held in Strasbourg. There I learned from Dr. Elaheh Kheirandish that she had read in various Safavid scientific manuscripts that mullāMuḥammad Ḥusayn, still otherwise unknown to me, was the son of the mathematician Muḥammad Bāqir Yazdī, who was known to me, and that he was one of the astronomers from whom Qāsim ʿAlī Qāyinī, also known to me, had learned the science of the astrolabe. Not long thereafter Dr. Kheirandish kindly sent me photocopies and translations of the relevant texts. By this time I had also looked at more Safavid instruments and various other Safavid and Ottoman treatises on the determination of the qibla.

Seventh, in July, 1995, I rewrote my study of the two world-maps, but decided to retain the geographical tables of al-Khāzinī as one of several examples of another trend in medieval Islamic mathematical geography. Instead of being able to trace this tradition of Mecca-centred world-maps back to al-Bīrūnī I can now only raise a series of questions, some of which may be answered by future research and others of which may prove to be irrelevant. On the other hand I can show that al-Bīrūnī, very early in his career, did address the question of how to produce a map preserving direction and distance to an arbitrary central point and that he did come up which a plausible, if messy, procedure for doing just that. Taking advantage of a sabbatical leave in the winter semester 1995/96 I spent six weeks of virtually total isolation in rural France, devoting a great deal of time and effort to finalizing the various appendixes.

Eighth, in December, 1995, I had the pleasure of inspecting the first world-map together with Dr. AssadullahSouren Melikian-Chirvani, one of the world’s leading authorities on Iranian metalwork. He was much taken by the elegance and precision of the engraving and suggested a dating of 1700 ± 20 years; he also inclined to favour a provenance from Khurasan, i.e. North East Iran, rather than Isfahan. In the same month additions and corrections to the article “Samtˮ in the Encyclopaedia of Islamwere submitted to the editors along with the article “Ṭāsaˮ [= magnetic compass].

The months of January and February, 1996, I spent in France not only to finalize the text but also to upgrade it on a new computer with new software.[1]In the summer of 1996 I was able to incorporate some suggestions and corrections graciously proposed by David Pingree By this time I had worked out to my own satisfaction how the grids might have been conceived and how they might have been constructed It was a good year later (see below) that I found out how they had actually been conceived and constructed. It was only in November, 1996, that I encountered Muḥammad Ḥusayn al-Mashhadī, a mathematician of some renown from 17th-century Meshed, identified by George Saliba in a paper published in 1994. This second Muḥammad Ḥusayn too must be considered as a contender for the honour of having made at least one of the world-maps, if not both. In January, 1997, I actually came across an astronomer named Muḥammad Ḥusayn al-Iṣfahānī, but he was active in Calcutta (and London) and he was involved in the dissemination of the ‘new’ European astronomy in India.

Only in May, 1997, in fact on the day after an article on the world-maps in Imago Mundiwent to press, I finally found how the grids on the world-maps were conceived and how they were constructed. (Fortunately the audience for that article, historians of cartography, has little interest in the mathematics of the grids, already relegated to an appendix in that paper). It was entirely due to the insights of François Charette that I eventually saw the light.

During the summer of 1997, again in France, I prepared the text for my very demanding editor Peri Bearman.François Charette had told me there were too many cross-references. I will never forget the moment when I did a computer check of the number of cross-references indicated by “seeˮ – I had to face the fact that 1,300 was un peu excessif. The reader will still find the same number of cross-references, albeit more subtly formulated.

This study of two world-maps Aand Breplaces all previous studies of map A. Some of these contain early interpretations that were incorrect, and hypotheses since proven invalid.[2]I still cannot claim to know the origin of the tradition of which these two world-maps are the sole surviving examples known to us. But I am convinced that the tradition goes back to al-Bīrūnī and/or Ḥabash and that it was revived in Timurid times. The discovery of a third world-map of the same kind might necessitate reinterpreting some of the commentary to the two maps presented here.

Such is the nature of research on newly-rediscovered primary sources, and historical writing which assumes that it is possible to write conclusively on a small proportion of sources that quite by accident happen to survive, or – even worse – simply rehashes previous summaries from the secondary literature, must be regarded with suspicion. In the case of most modern writings on Islamic geography, a subject on which there has been very little research on the primary sources in the last half century, this is, alas, the case. Thus most of the survey chapters on ‘standard’ or ‘mainstream’ Islamic cartography in the recently-published Historyof Cartographyare not based on any new research; rather, as noted in a review by Karen Pinto, they simply repeat the conclusions of scholars working in the first half of this century and perpetuate the mistakes made by them. (There are, on the other hand, several excellent chapters on more specific topics). In the meantime new primary sources have come to light and await study, but the number of scholars currently doing serious research on Islamic geography is small indeed and bears no relation to the importance of the subject.

The study of these world-maps takes us unto many fields: the history of cartography, the history of mathematical geography, the history of applied mathematics and especially trigonometry, the history of astronomical instruments (European as well as Islamic), the history of technology, the history of clocks and watches, as well as the history of decorated metal-work.[3]And at the same time we must keep in mind the cultural and religious setting in which these maps came into existence, namely, Iran and Central Asia under the Sunnite Timurids, and Iran under the Shīʿite Safavids.[4]Under the Safavids European technology and trade were slowly but surely making inroads in Iran. The accounts of life in Safavid Iran by European travellers prove to be most valuable sources for my study, as they have already for several other disciplines.[5]There is a sense in which this book belongs to the same literary genre.

For the reader unfamiliar with Islamic science in general it may be well to point out that Western historians have been preoccupied with the scientific heritage that was transmitted to the West. We are dealing here with materials that were not transmitted to the West, indeed – at least for the time when the two world-maps were made – we are dealing with the time when European science and technology had surpassed Islamic science and technology. This actually happened no earlier than the 16th century, and we shall be dealing mainly with what are ostensibly two late-17th-century objects. These could not have been devised by Muslims in the 17th century, and it still surprises me that the Safavid instrument-makers were in a position to make them, and to make them as well as they did. There were a few Europeans in the 17th century – in such places as Paris, Oxford and Leiden – who might have been able to have devised a cartographical solution to the qibla-problem – if they had set there minds to it, but there is not a shred of evidence that that any of them ever did. It is also highly unlikely that any European resident in or visitor to Iran was in a position to have devised a cartographic solution to the qibla-problem. Therefore, as we shall see, for the inspiration behind these objects we have to look back in time, to an age when Islamic science was still flourishing and innovative and particularly concerned with the kind of mathematical procedures that underlie the Safavid world-maps.

[1]Since such stories seldom appear in print and for the benefit and amusement of others I record my own experiences. Before the Autumn of 1995 Iused a Mac LC with MSWord 4.0. and a font for representing Arabic transcription called METimes. Since the disc-drive went defunct a few days before I was leaving for France I bought a new Power Mac 7200/75; this, however, would not accept MSWord 4.0. I had to pass through Switzerland to purchase an English version of MSWord 5.1a (on special offer because it was already out-of-date). I used this during my first sojourn in France, but the computer was not at all happy. An update to MSWord 6.0 brought by my son to Frankfurt from London at Christmas failed to initialize. The update to MSWord 6.0.1 subsequently provided by Microsoft in England failed to appreciate the special font METimes; somehow the font actually damaged the computer’s main-frame. Also the update interfered with the formating of all of the tables in the appendixes, which had to be reset. In early February 1996, after weeks of concentrated anguish and frustration, I received a new updated transliteration font ‘sabzevar’ from Professor Jamal Elyas of Amherst Collage in the US. Thinking this to be the answer to my dreams I restyled the entire text, only to find that I could not print more than about 10 pages at a time. A week later I received an METimes laser-printer font from Pim Rietbroek of Leiden. All I had to do was to restyle the entire text once again. It took three weeks for Gravis of Frankfurt to repair the main-frame. Needless to say, my new computer accepts none of my old software for writing Arabic (at-Kaatib 3.0) or for calculating (MSBasic 3.0) or for graphics (Superpaint), but I was naïve to ever think that it would. I yearned for my old computer so much I gave it to Gravis for repair but they went bankrupt and disappeared from Frankfurt, with or without my computer, I shall never know.

[2]See n. 4:1 below.

[3]Not a single Islamic astrolabe has ever been investigated satisfactorily by a historian of Islamic art. Yet the instruments with which we are dealing are indeed scientific works of art. One eminent historian of Iranian art whom I addressed on this issue stated that he felt that it was difficult for a “laymanˮ to decide whether a given instrument was genuine or fake (and there are indeed plenty of fakes in circulation on whose authen-ticity even the specialists on instrumentation cannot agree). Yet the ground has been laid – see, for example, Gingerich et al, “ʿAbd al-Aʾimma Forgeriesˮ and n. 7:7 below – and the time is ripe for Islamic instruments to be taken seriously by art historians, if only by those who can read the Arabic or Persian inscriptions. (On fake instruments it is not only the astronomical markings that may be inept but also the inscriptions.) In many cases, and this is particularly true of Safavidinstruments, they constitute the only available signed and datedpieces of Islamic metalwork.

[4]An informative introduction to Shīʿism is Donaldson, ShiʿiteReligion.

[5]The potential of this corpus of literature for contributing to our knowledge of such aspects of contemporaneous life in Iran as diverse as linguistics and architectural decoration is already well illustrated by such publications as Perry,“Persian during the Safavid Periodˮ, and Necipoğlu,Topkapi Scroll. Sonja Brentjes has embarked upon a systematic study of what these travelers recorded about scientific activity, and some of her discoveries relating to geographical knowledge in Safavid times have been incorporated in this study: see further n.8:1.

Source note:
This was published in:
DAVID A. KING, 1999, World Maps for Finding the Direction and Distance to Mecca: Innovation and Tradition in Islamic Science