The sixth-century version of the Liber Pontificalis contains a reference to a request made by a certain Lucius, a second-century king of “Britanio”, to pope Eleutherius (r. c.174–89). In his Ecclesiastical History of the English People Bede (d. 735) identifies this event as the first evidence for the conversion of the British to Christianity. He writes: “This pious request was quickly granted, and the Britons received the Faith and held it peacefully in all its purity and fullness until the time of the emperor Diocletian”. Bede's narrative was later embellished by Geoffrey of Monmouth (d. c.1155) and became a central plank of English theological debates during the Reformation.Footnote 2 The historicity of Lucius was finally demolished in the mid nineteenth century, but it was not until 1904 that the origins of the legend were established. A. von Harnack discovered that one of the sources for the Liber mentions Britium, the name of the fortress of Edessa ruled in the second century by Lucius Aelius Septimus Megas Abgar IX. He concluded that a later scribe of the Liber had mistakenly rendered “Britio” in the source text as “Britanio”.Footnote 3 This example illustrates the fragile foundations on which remarkably tenacious legends can be built. Most of all, stories and legends persist because they remain meaningful for later audiences.
In this article we discuss another story that has enjoyed a long life in scholarly literature, drama, and the visual arts: the alleged caging of the Ottoman sultan Bayezid I, Yildirim (r. 1389–1402 and d. 1403) by the Central Asian conqueror, Temür (r. 1370–1405, and otherwise known as Timur-i Lang, or in European sources as Tamerlane). A full analysis of the range of primary sources on this event – Arabic, Persian, Turkish, Greek, Latin, Italian, German, and Spanish to name just some of the languages employed by the relevant authors of the fifteenth and sixteenth centuries – is beyond the scope of this article. Rather, our attention is devoted to the evolution of scholarly discourse on the existence (or otherwise) of the cage, with a particular focus on the period from the late seventeenth to the first half of the twentieth century. Significant are the sources available to the scholars who wrote on this subject and the ways in which they exploited them. Like the example of King Lucius, scholarly debate around the captivity of Bayezid must be located within a larger historical framework, particularly the changing relations between the polities of Christian Europe and the Ottoman Empire between the second half of the fifteenth century (following the fall of Constantinople in 1453) and the nineteenth century.
Historical Events and Later Representations
On 28 July 1402 (19 Dhu al-Hijja 804) the armies of sultan Bayezid and Temür met on the field of Çubuk Ovasi near to Ankara. Marking the culmination of a long-standing dispute over the control of the former territories of the Rumi Saljuqs, the battle of Ankara had been preceded by Temür's conquest of the Anatolian city of Sivas in 1400 and Bayezid's retaliatory strike against Erzincan. Relations between the two men had not been improved by a bitter exchange of embassies in the previous years and by Bayezid's support of the Black Sheep Turkomen. True to his epithet, Yildirim (“the thunderbolt”), Bayezid appears to have rushed to battle without proper preparation of his forces. Arriving at Çubuk Ovasi he found Temür's engineers had dammed off the available water and had built substantial ramparts around their own positions. Desertions further weakened Bayezid's position, and the final result of the battle was a decisive defeat for the Ottoman army with many thousands of soldiers left dead on the battlefield.Footnote 4
The battle of Ankara is one of the most important conflicts of the late Medieval period; the impact of the battle can be detected in the political history of the region in the decades after 1402. Some of the more immediate effects of the battle can be briefly summarised: first, sultan Bayezid was captured following the battle and died in captivity in 1403; second, the removal of the sultan sparked off a civil war between his sons that was only resolved with the accession of Mehmed I in 1413; third, the Ottoman empire itself was for several decades greatly reduced in the extent of its territories and its political influence; and fourth, the defeat of the Ottoman army relieved the siege of Constantinople and ensured the survival of the diminutive Byzantine empire for another half a century.Footnote 5
Despite the existence of a wide array of primary textual sources in Persian, Arabic, Ottoman Turkish and European languages, there remains considerable uncertainty concerning what actually transpired when the defeated sultan was brought to the camp of Temür in the evening after the battle. Persian, Turkish and Greek sources concur that the sultan was bound in some manner, and it seems likely that Temür ordered that these bonds be removed while he conducted an interview with his captive. Another recurrent theme is that Bayezid, heedless of his precarious situation, answered the questions posed to him in a haughty manner that displeased Temür. While it had been the latter's practice to seek oaths of submission and re-establish vanquished princes as Timurid vassals in their former territories, it is clear that Bayezid was not released from captivity. He died some months later in 1403 and, at the request of his sons, his body was transported to a mausoleum in Bursa.
We will return to the accounts of Bayezid's captivity written by Arabic, Persian, Turkish, and Greek authors later in the article, but first it is necessary to examine the ways in which this event was represented in Western Europe during the fifteenth and sixteenth centuries.Footnote 6 The earliest European accounts, which include those of Johannes Schiltberger (d. after 1429) who was captured by Temür's army before being allowed to return to his native Germany and Jean Boucicault (d. 1421), governor of Genoa in the early years of the fifteenth century, provide little detail concerning the treatment of the Ottoman sultan beyond the facts of his captivity and death. Neither mentions the employment of a cage to imprison Bayezid.Footnote 7 The ambassador of Henry of Castille, Ruy González de Clavijo (d. 1412), who visited Temür in Samarqand in September-November 1404, has nothing to add on the existence or otherwise of a cage. Clavijo was in Constantinople soon after the return of Manuel II Palaeologos (r. 1391–1425) from his European trip in July 1403. Clavijo met both the emperor and members of the Byzantine court during this visit.Footnote 8 Had reports of the caging of Bayezid reached Constantinople by the time of Clavijo's visit, one might have expected him to have mentioned it. Notably, none of the Greek primary sources written in the first decade of the fifteenth century contains a reference to the captivity of the Ottoman sultan.Footnote 9
It is clear, however, that more graphic stories were circulating in Europe in the first half of the fifteenth century because they appear in sources such as the Chronicon Tarvisinum and the letters of the Italian humanist, Poggio Bracciolini (d. 1459), in his De Varietate Fortunae. Probably composed around the time of the death of pope Martin V (r. 1417–31), Bracciolini's text makes the following comment about the treatment of Bayezid: “caveaque in modum fere inclusum per omnem Asiam circumtulit”.Footnote 10 More influential than either of these, however, was the account of the fate of the Ottoman sultan in Aeneas Silvius Piccolomini's (Pope Pius II, r. 1458–64), Asiae Europaeque elegantissima descriptio, composed in the late 1450s and early 1460s and first published in Paris in 1509.Footnote 11 Piccolomini brought together the key elements that were to remain fundamental to the European narrative for some two centuries: first, the sultan (often known in European writings as Bajazet) was placed in an iron cage; second, he was forced, like a dog, to eat scraps from under the table of Temür; and, third, Bayezid was employed as the “Scythian” ruler's mounting block when the latter got onto his horse.Footnote 12 Further elaborations can be found in later histories such as the history of the Turks composed by Theodore Spandounes (Spandugino, d. after 1538). The first version of his history was written in Italian and completed in 1509. A French translation by Balarin de Raconis was published in 1519. In this version Spandounes alleges the use of chains of gold and that the sultan was employed as Temür's mounting block.Footnote 13 It is only in later revisions of the text that Spandounes adds the reference to the cage.Footnote 14 He also includes a lurid description of the public humiliation of Bayezid's wives and concubines by Temür (Sachatai):
When Sachatai got back to “Scytia” he staged a magnificent triumph for his victory over Bayezid and a great assembly attended by almost all of the lords and princes of Scytia; and the cage containing Bayezid was brought in. Then [Sachatai] did something very out of keeping with his grandeur and noble character. He had Ildrim's wife, who was also his prisoner brought in and he caused her clothes to be ripped down to her navel so that she showed all her pudenda; and he made her wait upon and serve food to his guests. Ildrim, seeing his wife thus shamed, bewailed his fortune and wanted to kill himself at once. But he had no knife or other means. So he banged his head against the iron bars of his cage so hard that he dispatched himself miserably.Footnote 15
Constantine Mikhailović of Ostrovica (d. after 1563), who had served in the Ottoman army and spoke Turkish, also alleges that Despina was forced to serve drinks to the guests of Temür as a means to humiliate sultan Bayezid. Mikhailović claims that the sultan poisoned himself using his own finger ring. No reference is made to the use of a cage and, in common with the Persian sources, Mikhailović's account indicates that Temür was upset by Bayezid's suicide.Footnote 16
It is worth emphasising that, aside from early primary sources such as Schiltberger, Clavijo, Boucicault, Jean of Sultaniyya (fl. late 14th to early 15th century), and Stefan Lazarević, Despot of Serbia (r. 1402–27), the European authors mentioned above can hardly be viewed as reliable sources on Timurid or Ottoman history.Footnote 17 The last generation of Byzantine historians – Laonicus Chalcocondyles (d. 1490), Michael Ducas (d. c.1470), and George Phrantzes (also Sphrantzes, d. c.1478) – appear to have been better informed about Turkish history than their contemporaries in the Catholic West, probably as the result of direct or indirect access to primary sources in Turkish, Persian or perhaps Arabic.Footnote 18 By contrast, only one Western European scholar of the sixteenth century, Johannes Leunclavius (d. 1593), is known to have consulted Turkish chronicles (see below). A few scholars of Ottoman history – most notably Paolo Giovio (d. 1552) – tried to remedy their lack of knowledge of Middle Eastern written sources by seeking out oral testimony of those who had travelled to the Middle East and visual sources (such as the painted portraits of sultans produced in Turkey).Footnote 19 Where these (often highly inaccurate) European histories are more valuable, however, is in gauging the perceptions of the Islamic world among the literate elite of Europe during the period of the great military successes of the Ottoman empire. Clearly, from its first appearance in the second quarter of the fifteenth century the idea that Bayezid had been confined in a cage proved highly attractive to Europeans as they looked anxiously toward the seemingly inexorable expansion of the Turkish polity.
The sufferings of Bayezid also found expression in Europe in the visual and dramatic arts from the sixteenth to the early nineteenth century. While most printed books on Turkish history were not illustrated, woodcut prints showing the fate of the Ottoman sultan appear in works such as Philip Lonicer's Chronicon Turcicorum (1578). Depictions of the capture and caging of Bayezid were painted for the Neues Palais in Potsdam (by Andrea Celesti) and Schloss Ambras in Graz (by Carl Franz Caspar or Andreas Raemblmayer). Tapestries were also produced on these themes in Antwerp during the seventeenth century.Footnote 20 In Tamburlaine the Great, first performed in 1587, the dramatist, Christopher Marlowe (d. 1593), made his Bayezid function as the footstool to the Scythian ruler's throne. The confrontation between Temür and Bayezid was explored by later playwrights, most successfully by Nicholas Rowe (d. 1718) in his Tamerlane, a Tragedy (first performed in 1701). Robert Irwin notes that Rowe's Tamerlane was performed annually on 5 November (the date of William of Orange's landing in England) until 1815. Thus, for Protestant English audiences the glorious Tamerlane stood for William III (r. 1688–1702) while Bayezid represented Catholic monarchy – the French ruler, Louis XIV (r. 1643–1715), and presumably also the deposed Stuart King James II (r. 1685–88).Footnote 21 Several operas and even a ballet were also composed on the life of Temür and his conflict with Bayezid.Footnote 22
Clearly most of this literary and visual material is pure invention, and careful examination can reveal the initial points of reference. For instance, in about 260 CE the Roman emperor Valerian (r. 253–60) suffered the indignity of becoming the mounting block of the Persian shah Shapur (r. 240–72). The event can be seen on a monumental rock relief at Tang-i Chogan near Bishapur in Iran, while both Valerian and Philip the Arab (r. 244–49) appear in submissive poses before Shapur on a relief at Naqsh-i Rustam. More relevant for the present purposes is the fact that the fate of Valerian became, for Christian moralists from Lactantius (d. c.325) in the fourth century through to Giovanni Boccaccio (d. 1375) in his De Casibus virorum illustrium (1355–74), a symbol of both the transience of earthly power and divine retribution for the Roman emperor's treatment of Christians.Footnote 23 This event is recorded in visual form in illustrated manuscripts and printed versions of Boccaccio's text (and the popular translation by John Lydgate entitled, The Fall of Princes, completed in 1438–39).Footnote 24 Marlowe's vision of Bayezid as the footstool to the throne appears earlier in the woodcut frontispiece to the 1570 edition of John Foxe's (d. 1587) Actes and Monuments of these Latter and Perillous Days (also known as Foxe's Book of Martyrs) in which Henry VIII (r. 1509–47) makes similar use of pope Clement VII (r. 1523–34).Footnote 25
While it is possible to strip away much of the extraneous detail from these European accounts, three accusations – that the sultan was imprisoned in an iron cage, that his harem was subjected to public indignities in his presence, and that his humiliations caused him to commit suicide – are not easily explained away simply as literary references to earlier events. The first of these, Bayezid's imprisonment in a cage, has stimulated the greatest amount of academic controversy from the late seventeenth to the early twentieth century. The next section of this article establishes the scholarly dramatis personae and how the debate over the existence of the cage has evolved over this long period. The final section re-examines some of the oft-quoted primary sources in Greek and Arabic. We state from the outset that, despite the confident assertions made by numerous scholars, we do not believe that the sources allow for a definitive resolution concerning the existence or otherwise of the cage. Of greater interest is the issue of the precise sources and methods employed by the many historians who concerned themselves with the fate of the Ottoman sultan.
From d'Herbelot to Köprülü: Changing Academic Viewpoints on Bayezid's Cage
Modern studies of Ottoman and Timurid history exhibit little interest in the alleged caging of Bayezid I by Temür. Passing over the question in silence, they tend to record simply that the sultan died in captivity.Footnote 26 There are good reasons for this lack of elaboration; it is the fact that Bayezid's capture resulted in a debilitating civil war that is of primary significance in the historical record. Today it is left to more journalistic books such as Justin Marozzi's Tamerlane: Sword of Islam, Conqueror of the World (2004) to revive the question of the cage.Footnote 27 The point here is not to denigrate a readable and competent piece of popular history, but to establish that the sources carefully marshalled by Marozzi in support of his interpretation (he concludes that the cage was a fiction) are largely the same as those employed by Joseph von Hammer-Purgstall (d. 1856) in the first volume of Geschichte des osmanischen Reiches (1827). Many of these arguments are advanced earlier by Edward Gibbon (d. 1794) in the final volume of Decline and Fall of the Roman Empire (1778). In turn, both relied upon the collation of written sources presented in Barthélemy d'Herbelot de Molainville's (d. 1695) Bibliothèque orientale, ou dictionaire universel (published in 1697 after having been completed by Antoine Galland [d. 1715], the famous translator of the Thousand and One Nights).
D'Herbelot's Bibliothèque orientale makes a good starting point for an analysis of the scholarly debate over the caging of Bayezid.Footnote 28 The scope of d'Herbelot's reading of Arabic, Persian, and Turkish sources far exceeded that of any previous reference work on Islamic history. In the introductory section to his encyclopaedia he claims to have consulted the writings of four key authors on the life of Temür: Ahmad ibn Muhammad ibn ‘Arabshah (d. 1450), Muhammad ibn Khavand Shah (Mirkhvand, d. 1498), Ghiyath al-Din Khvandamir (d. 1534 or 37), and Sharaf al-Din ‘Ali Yazdi (d. 1454). Of these, only Ibn ‘Arabshah's Kitāb al-‘ajā’ib al-maqdūr fī akhbār Tīmūr was available in printed form (both in Arabic and French translation).Footnote 29 It is also evident that he is aware of some of the Turkish histories of the fifteenth and sixteenth centuries, although he does not name them in the list of authors he had consulted. D'Herbelot mentions the captivity of Bayezid briefly in the entry devoted to Temür (Timour), but the bulk of his comments on this question appear in the entry entitled, “Baiazid, ou Abu Iezid Ben Morad Gazi”.Footnote 30 Without giving specific citations to his sources, d'Herbelot assembles a reconstruction of the encounter between Bayezid and Temür in which the latter treats his captive with respect, offering him a meal in his tent. They have a conversation that encompasses such issues as the government of empires and the vissicitudes of fate. The discourse finishes with Temür posing a question; d'Herbelot writes:
But having finished the conversation with a request as to what would have been the treatment he would have received had he fallen into the same disgrace; this Sultan, who was naturally shy, replied that he would have locked him [Temür] in an iron cage, and carried him about in this state among all the provinces of his empire.
The victor, surprised by the brutality of his prisoner's response, took at the same time the resolution to treat him as he [Temür] would have been treated had he fallen into his [Bayezid's hands] . . .Footnote 31
According to d'Herbelot's account, Temür was unable to take Bayezid all the way to Samarqand because of the latter's death in 804 (this should be 805). D'Herbelot concludes by noting that some Turkish historians claim the sultan died by his own hand. As already noted, Leunclavius was able to exploit these sources with the aid of a translator, one Johannes Gaudier (known as “Spiegel”) in his Annales sultanorum Othmanidarum (1588). Leunclavius also claims considerable familiarity with Medieval Greek histories.Footnote 32 He records the conversation between Temür and Bayezid in similar terms to d'Herbelot, and his version was then repeated in Jean-Jacques Boissard's (d. 1602), Vitae et icones sultanorum (1596) and Richard Knolles’ (d. 1610) The generall Historie of the Turkes (1603).Footnote 33
The first of the major Persian historians of this period to be made widely available to European scholars was Sharaf al-Din ‘Ali Yazdi. His panegyric biography of Temür, the Ẓafarnāma, was translated into French by Alexandre Pétis de la Croix (d. 1751) and published under the title Histoire de Timur-bec in 1722. An English translation of this edition appeared in the following year.Footnote 34 According to Yazdi, Bayezid was brought as a prisoner with his hands tied into the presence of Temür. The latter was touched by pity and compassion, and ordered the sultan's hands to be freed before they started their conversation. Temür's gracious manner leads Bayezid to admit the fault of his actions, and the remainder of Yazdi's account emphasises the good treatment of Bayezid prior to his death from apoplexy (an event that apparently much affected Temür as he had intended to replace the Turkish sultan on his throne).Footnote 35 Pétis de la Croix was well aware of the stories regarding the caging of Bayezid. Adopting a more critical stance to the available Islamic sources than one sees in D'Herbelot's entry on Bayezid and the works of other earlier writers, Pétis de la Croix remarks in the introduction to his translation:
As Timur-Bec had defeated the Turks and Arabs from Syria, and he had taken even the Sultan Bajazet, it is small wonder he was mistreated by the Historians of these Nations, who in defiance of truth, and against the dignity of history, have fallen into [treating] this subject with great excess. We see by the lecture of Condemir, and of many other historians, that everything they have written of the origin and adventures of Timur-Bec, are fables and that their animosity against the Prince made them invent [these]. So to destroy completely the fable, we will attach [to him] the name of Timur-Bec, and lose that of Tamerlane which has been adopted.Footnote 36
The person identified as “Condemir” in this passage is Demetrie Cantemir, Voivode of Moldavia (d. 1723), whose influential The History of the Growth and Decay of the Ottoman Empire was translated into English in 1734–35. Cantemir followed the prevalent view that Bayezid had been confined in a cage.Footnote 37 In the footnote to the paragraph quoted above Pétis de la Croix's offers further thoughts on the cage:
These are passionate historians who invented the fable of the iron cage, in which they say the victor had placed Bajazet. And it was followed by several Europeans, but we see the fallacy in [the writing of] our author (Yazdi), who is contemporary, and reports to the contrary, that Timur-Bec always treated Bajazet as his equal, and that he restored to him all the honours due to the most important kings.Footnote 38
Additional support for Pétis de la Croix's positive assessment of Temür's conduct was provided later in David Price's (d. 1835), Chronological Retrospect, or Memoirs of the principal Events of Mahommedan History (1811–21), a three-volume work largely based on translations of Mirkhvand's vast history, Tārīkh-i rawẓat al-Ṣafā.Footnote 39 Differing only in relatively minor details (such as the cause of sultan's death from asthma and inflammation of the throat) from Yazdi's version of events, Price's collation of Mirkhwand and other Persian authors appears to have been little noted by the nineteenth- and twentieth-century scholars who have dealt with Bayezid's cage. This oversight may have been because Yazdi's biography was so entrenched in European scholarship by this time, and was perhaps also due to the fact that in the Chronological Retrospect it is very difficult to prise apart the original translations from the translator's commentary. Mirkhvand is now regarded as an important source on the life of Temür because of his use of first-hand testimony. In this respect, his assertions – based on the testimony of an eye-witness, Sayyid Ahmad Tarkhan, brother-in-law of Shahrukh (r. 1405–47)Footnote 40 – that Bayezid was brought bound into Temür's presence, but that he was subsequently unmanacled and never confined in a cage, carry considerable weight.Footnote 41 An edition of Khvandamir's chronicle, Habīb al-siyar, was published in 1857, but it had no noticeable impact upon the debate.Footnote 42
The increasingly critical evaluation of source material is a feature of later scholarly contributions in the second half of the eighteenth and the early nineteenth century. Voltaire (François-Marie Arouet, d. 1778) turned his attention to the question of the treatment of Bayezid in Chapter 75 of his Essai sur l'histoire générale (1756 and translated into English in 1782). Without citing specific primary sources, he asserts that the cage is not mentioned by Persian or Arab authors, and appears in the Turkish chronicles “perhaps in order to render Tamerlane odious; or rather because they copied it from the Greek historians”. His Temür is very much the magnanimous ruler drawn by Yazdi, and this characterisation leads Voltaire to conclude that “it is difficult to reconcile the iron cage, and the base affront done to Bajazet's wife, with the generosity which the Turks attribute to Tamerlane”. He extends his scepticism to all Islamic sources noting that “Oriental” historians often put grandiose words into the mouths of their subjects.Footnote 43
Voltaire's assessment of the problems inherent in the interpretation of Islamic histories is noted approvingly by Edward Gibbon in the final volume of the Decline and Fall.Footnote 44 Straying well beyond his remit of writing the history of the Roman empire through to the fall of Constantinople in 1453, Gibbon devotes considerable attention to aspects of Islamic and Asian history. His fascination with the Mongols has been discussed by David Morgan,Footnote 45 and it is little surprise that he should also have been intrigued by the personality and achievements of Temür. Unable to read Arabic or Persian, Gibbon went to considerable effort to consult all of the published translations of Middle Eastern histories, supplementing this with an extensive knowledge of primary sources in Greek and Latin. His treatment of the caging of Bayezid represented the most comprehensive examination yet of the primary sources, and was undertaken with his characteristic acuity. He notes that the story of the iron cage was employed in past times as a moral lesson, and that it “is now rejected as a fable by modern writers, who smile on vulgar credulity”.Footnote 46 In order to come to his own conclusion, however, Gibbon sifts through the information in the primary sources at his disposal.
Gibbon's assesses the potential veracity of those writers who claim the good treatment of Bayezid and of those who assert that Temür confined him in a cage. Emphasis is given to Yazdi's positive portrayal of Temür's behaviour, particularly in view of its early composition (in the 1420s) and the consistencies evident between the Ẓafarnāma and other Persian accounts cited by d'Herbelot. Gibbon is, however, wary of the “flattery” that, in his opinion, pervades such sources, and turns his attention to evidence in favour of the cage. In addition to the quotes of Poggius Bracciolini, Gibbon adduces other references to the cage in European writings of the first three decades of the fifteenth century. He then addresses the important Arabic source, Ibn ‘Arabshah, whose history of Temür is characterised by Gibbon as “florid and malevolent”. That said, he is struck by the correspondence between the broadly contemporaneous testimonies of Europeans like Bracciolini and Ibn ‘Arabshah in the matter of the cage. Ibn ‘Arabshah also adds details concerning the humiliation of the concubines and wives of Bayezid. Of the Greek historians Gibbon singles out Phrantzes, partly on chronological grounds, but mainly because he was sent in 1429 as an ambassador to Murad II (r. 1421–44, 1446–51) and, therefore, may have conversed with elderly janissaries who had also been imprisoned by Temür. According to Gibbon (and many later scholars), Phrantzes is the one fifteenth-century Greek historian to refer to the employment of a cage to confine sultan Bayezid. Lastly, referring to the Turkish historians who write about the cage, Gibbon remarks, “. . .some credit may be allowed to national historians, who cannot stigmatise the Tartar without uncovering the shame of their king and country”.Footnote 47 Faced with two conflicting bodies of evidence Gibbon seeks for a “fair and moderate” conclusion. He writes:
I am satisfied that Sherefeddin Ali [Yazdi] has faithfully described the first ostentatious interview, in which the conqueror, whose spirits were harmonized by success, affected the character of generosity. But his mind was insensibly alienated by the unseasonable arrogance of Bajazet; the complaints of his enemies, the Anatolian princes, were just and vehement; and Timour betrayed a design of leading his royal captive in triumph to Samarcand. An attempt to facilitate his escape, by digging a mine under the tent, provoked the Mogul [sic] emperor to impose a harsher restraint; and, in his perpetual marches, an iron cage or waggon might be invented, not as a wanton insult, but as a rigorous precaution.Footnote 48
The severity of this treatment leads Gibbon to conclude that Temür could be held responsible for the Ottoman sultan's untimely death.Footnote 49 Both Voltaire and Gibbon portray the actions and character of Temür in a relatively positive light. Their comments should be seen in the wider context of the debate concerning Oriental despotism in eighteenth-century Europe. In their favourable assessment of some Muslim rulers Voltaire and Gibbon can be grouped with such Orientalists as Abraham-Hyacinthe Anquetil-Duperron (d. 1805) and Antoine Isaac Silvestre de Sacy (d. 1838) on the opposite side of the European view of despotism from Charles de Secondat, baron de Montesquieu (d. 1755).Footnote 50 Further confirmation of the statesmanlike character of Temür was provided by the publication of A Specimen of the civil and military Institutes of Timour, or Tamerlane (1780), translated from a Persian manuscript by Joseph White. Although it is now recognised as a work of the Mughal court in the seventeenth century, the Institutes of Timour exerted some influence over scholarly views of Temür in the late eighteenth and nineteenth centuries.Footnote 51 Notably, Gibbon had also written admiringly of the law code (yasa) attributed to Chinggiz Khan (Genghis Khan, d. 1227), even comparing it to the Constitutions of Carolina (1669) by John Locke (d. 1704) and Anthony Ashley-Cooper, first earl of Shaftesbury (d. 1683).Footnote 52
An influential contribution to the study of the primary sources dealing with Bayezid's cage was made by Joseph von Hammer-Purgstall in the first volume of his Geschichte des osmanischen Reiches (1827).Footnote 53 Aside from his examination of the Turkish historians, Ashikpashazade (d. 1481), Mehmed Neshri (d. c.1520), and Hoca Sa'd al-Din Efendi (d. 1599), the key element of his argument is philological. He focuses his attention on the meaning of the Turkish word, qafes, that was commonly understood to mean simply “cage”. Relevant too is the Arabic term qafas used by Ibn ‘Arabshah in his life of Temür. Crucially, Von Hammer points out that qafes (he transcribes the word as kafes) does not simply have to be read as “cage” (Käfich), but can also be understood as “a barred room” (ein vergittertes Zimmer) or “a barred litter” (Sänfte). He argues that this misunderstanding has led to the perpetuation of the “fairytale” (Mährchen) of the iron cage in the writings of Ibn ‘Arabshah, Phrantzes and numerous Western European historians.Footnote 54 Von Hammer points to a passage in the chronicle of Neshri that is translated in Geschichte des osmanischen Reiches as, “Timur liess eine Sänfte machen, in der man ihn (Bajesid) wie in einem Kafes zwischen zwey Pferden trug”.Footnote 55 Thus, Neshri's account appears to be specifying a litter held between two horses. In addition, this term was often applied to the litters used to carry women of the harem, and could even be applied to the residence (Wohnung) of the princes in the palace of Constantinople. In this last context, qafes implies also the ritual seclusion that was established around the sultan and the princes of the Ottoman dynasty. According to Von Hammer's interpretation qafes loses most of its negative connotations. Thus, Temür's actions were, as is implied in the Persian accounts, consistent with the necessary respect for the dignity of this royal captive.
While Von Hammer-Purgstall's analysis of the Turkish word qafes was widely accepted throughout the nineteenth and early twentieth century, a note of dissent was offered by Gustav Weil in Volume 5 of his Geschichte der Chalifen (1862).Footnote 56 Weil pointed to the fact that the Arabic term qafas lacks the secondary meanings found in Ottoman Turkish. He concluded that Ibn ‘Arabshah, an author conversant with Turkish source material, would have been able to substitute a suitable Arabic term for “litter” (or, at least, some further clarification in the text) if that were what he actually wanted to indicate. Weil's scepticism concerning Von Hammer's approach is echoed by Herbert Gibbons (d. 1934) in The Foundation of the Ottoman Empire (1916). Gibbons also suggests that for Bayezid the humiliation of being placed in a “harem litter like a woman” would have been no less than being placed in a cage meant for a wild beast.Footnote 57
Two points may be offered in support of Weil's viewpoint. The first relates to the semantic range of qafas in Medieval Arabic sources. For scholars of the late nineteenth and early twentieth century there existed two readily accessible reference works – Edward Lane's (d. 1876) Arabic-English Lexicon (1863–93) and Reinhart Dozy's (d. 1883), Supplément aux dictionnaires arabes (1881) – that offered definitions for qafas as well as examples of primary texts in which the word appears.Footnote 58 Curiously, we have found no evidence that these works were consulted by any scholar writing about the caging of Bayezid. Lane's Lexicon does record a meaning of “a thing composed of two curved pieces of wood between which is a net”, but this does not appear to be something sufficiently substantial to carry or enclose a man. Furthermore, the first definitions offered in this entry in the Lexicon indicate relatively small objects such as a cage or coop (made of wood or reeds) for confining birds or animals. Lane also notes that the word can refer to the “the cage-formed structure of the bones of the thorax”. The second point is that had Ibn ‘Arabshah intended his use of qafas to bear the meanings it carries in Turkish, one might have expected the word to have been repeated in subsequent retellings of the captivity of Bayezid by other Arab historians. No mention of a qafas is found in Ibn Taghribirdi's (d. 1470) account of these events in al-Nujūm al-zāhira fī mulūk misr wa'l-qāhira, even though he was personally acquainted with Ibn ‘Arabshah.Footnote 59 Ibn Taghribirdi records only that the sultan appeared before Temür “hobbling in shackles (qayd pl. quy d)” and that he was held in this manner at the time of his death in prison in the month of Dhu al-Hijja 805 (June 1403).Footnote 60
Between the 1860s and the 1940s most of the significant contributions to the debate over the caging of Bayezid focused upon the information in the earliest Turkish histories.Footnote 61 The relevant sections of Neshri's account of the Ottoman dynasty were edited with a German translation by Theodor Nöldeke (d. 1930) in 1861.Footnote 62 Franz Babinger (d. 1967) produced an edition of the chronicle by ‘Uruj b. ‘Adil (fl. late fifteenth century) in 1925 which was exploited by Nicholas Martinovitch in a short article published in the Journal Asiatique in 1927.Footnote 63 ‘Uruj b. ‘Adil records the alleged conversation in which Bayezid, in answer to Temür's question of what would have happened if their situations had been reversed, carelessly responded that he would have imprisoned his captive in an iron cage (qafes). This remark seals his fate as Temür immediately orders a cage to be made in order to confine the Ottoman sultan. Martinovitch confidently asserts that this relatively early account (in fact, written several decades after the earliest European references to the cage) definitively solves the question of the cage. He appears to have been unaware that European scholars including Leunclavius in the late sixteenth and d'Herbelot in the late seventeenth century were already aware of Turkish sources recording this supposed conversation. In his article Martinovitch also surveys the Russian publications that discuss Bayezid's cage.
The last major salvo in this scholarly controversy was penned by the Turkish scholar, Mehmet Fuad Köprülü (d. 1966) for the journal, Belleten. Entitled, “Yildirim Beyazıd'ın esareti ve ıntıhari hakkinda” (‘On the story of Bayezid Yildirim's captivity and suicide”), Köprülü's article represents the most sustained Quellenkritik of the Turkish, Persian, and Arabic sources that discuss the treatment of Bayezid from his initial capture through to his death. A second article by the author revisiting the question of Bayezid's alleged suicide appeared in the 1942 issue of Belleten.Footnote 64 Split into two halves (the first dealing with the cage and the second with the allegation that the Ottoman sultan killed himself), the 1937 article starts with a brief summary of the major primary sources and of the interpretations offered by earlier scholars. Köprülü disagrees with Von Hammer's conclusion that the cage was no more than a legend. He also questions Martinovitch's assertions and the reliability of Babinger's edition of ‘Uruj b. ‘Adil as a tool for the study of the period. In common with Herbert Gibbons he believes that the question of the historicity of the cage “is yet far from being resolved”.Footnote 65 Notable too is the praise he has for Edward Gibbon's undogmatic methodology and conclusions regarding the cage.
Köprülü's analysis of the Turkish sources is of great importance.Footnote 66 He divides the accounts into those that make no mention of the cage or seek to discount its existence from those that write of it as an historical event. The first group includes the very earliest Ottoman chronicles, Sukrullah's Behcat-ul Tevarikh and the history of Muhammad b. Haci Halil Konevi'im, and other later works by Idris Bitlisi, Ibn Kemal (Kemalpashazade), and Sa'd al-Din Hoca. What holds this first group together is that they were either court historians (who had a vested interest in expunging evidence of the humiliation of a sultan of the dynasty) or simply writers aiming to praise the Ottoman rulers. Sa'd al-Din, considered by Köprülü as an official historian of the Ottoman state, went to particular lengths to try to show the cage to be a rumour (including consulting Yazdi's Zafarnāma), and his scepticism probably influenced the approach taken by Von Hammer. Among this first group Köprülü identifies one exception: Sehnameci Lokman who in the late sixteenth century wrote that “Timur had a cage made out of iron that looked like a throne (taht) and put Yildirim in it”. As Köprülü notes, although the account claims the object looked like a throne, it is clearly a type of cage (qafes).
The second group is composed of independent historians from among the Anatolian Turks who felt no need to praise the Ottoman sultans. These include the anonymous, Tevarikh-i Al-i Osman, the histories of ‘Uruj, Ashikpashazade, Muhyiddin, Lutfi Pasha, Neshri, and the verse history of Hadidi. Ashikpashazade's account is interesting in that he claims it to have been based on the testimony of a left-handed individual who was in the service of Bayezid from the battle of Ankara through to his death in captivity. While Von Hammer inferred it was a palanquin – not a cage – that was made for the Ottoman sultan, Köprülü draws attention to Ashikpashazade's actual statement: “the cage was not a palanquin (taht-i revan) that was merely aimed for the transportation of the captive sultan, but a vessel that would prevent him from escaping”.Footnote 67 From these sources Köprülü concludes that the rumour about the cage was circulating in Anatolia in the early fifteenth century. According to Köprülü further confirmation of this comes from the chronicle of Phrantzes, who had visited the Ottoman court in 1429 and Ibn ‘Arabshah, who had also spent time in Anatolia and was conversant in Turkish. Another Arabic historian, Ibn Iyas (d. after 1521), also records the story of the cage in his chronicle, perhaps drawing on Ibn ‘Arabshah or traditions circulating among the Ottoman conquerors of Egypt. Köprülü mentions the Greek historian, Spandounes, as another who had been in Turkey and had close relations with Ottoman court circles (though Köprülü was unaware that the earliest published editions of Spandounes’ history lack any reference to the cage – see above).
On the basis of his analysis of the sources Köprülü concludes that the cage was most probably an historical fact. He favours the reconstruction offered by Edward Gibbon. That the Persian sources – he cites Yazdi, Nizam al-Din Shami (fl. 1404) and Hafiz ‘Abru (d. 1430) – make no mention of can be explained by their desire to show Temür in the best possible light. The omission of the cage from many of the Byzantine and Western European accounts is not seen by Köprülü as sufficient reason to deny its existence. Finally, he identifies examples from Arabic and Persian chronicles of the caging or placing in chains of captured princes and sultans. In addition, Köprülü cites an example of similar treatment of Christian merchants from the writings of Jean de Joinville (d. 1317), companion of Louis IX (r. 1226–70) on the Seventh Crusade, and the popular Sufi verse that “the physical body is an iron cage that imprisons the spirit”.Footnote 68
While no major advances have been made in the analysis of Islamic primary sources surrounding the captivity of Bayezid since Köprülü's 1937 article, mention should be made of the collation of European and Middle Eastern primary sources offered in Marie-Mathilde Alexandrescu-Dersca's 1942 publication, La Campagne de Timur en Anatolie (1402).Footnote 69 Michele Bernardini has conducted the most detailed analysis of the primary and secondary sources on the cage in recent years, also making numerous important observations about the portrayal of the conflict between Temür in European dramatic traditions.Footnote 70 Most of the significant Arabic, Persian, and Turkish sources are now available in critical editions. An English translation of Ibn ‘Arabshah's life of Temür by John Sanders was published in 1936, and modern popular biographies of Temür were written by Harold Lamb (1928) and Hilda Hookham (1962).Footnote 71 Adam Knobler has brought to light additional European primary sources dating from the last years of Temür's rule.Footnote 72 Lastly, the question of Bayezid's cage has also been explored by scholars of Christopher Marlowe. The most thorough examination of the sources consulted by Marlowe is provided by Una Ellis-Fermor (d. 1958) in the introduction to her edition of Tamburlaine the Great (1930).Footnote 73 Although she incorrectly identifies the chronicle of the Armenian, Het'um (d. c.1311), as a primary source for the life of Temür (this final section of the early printed versions of Les flors des histoires is, in fact, a later additionFootnote 74), her discussion of the sources is, nevertheless, intriguing as a guide to what an educated reader of the late sixteenth century would have been able to learn about the lives of Temür and Bayezid through consultation of European printed books.
Re-evaluating the Greek and Arabic Sources
It is beyond the scope of this article to present a critical analysis of all the primary sources dealing with the treatment of Bayezid during this captivity. A detailed examination of the pertinent material in the earliest Turkish histories is certainly required, and the Arabic, Persian, and Eastern European primary accounts also warrant further attention. What will be offered in this section is a re-evaluation of the Greek and Arabic sources that have most frequently been employed by Orientalists in their discussions of the alleged caging of the Ottoman sultan.
The major Greek historians of the last phases of the Byzantine empire have often been cited in the debate over the existence of Bayezid's cage. While the Greek sources of the early fifteenth century and the biography of Mehmed II (r. 1444–46, 1451–81) by Critovoulos of Imbros (d. c.1470) have nothing to add to what is known elsewhere,Footnote 75 the histories of Ducas, Chalcocondyles, and Phrantzes deserve more attention than they have previously been given. In his chronicle Ducas provides a relatively detailed account of the battle of Ankara and its aftermath. He makes the following remarks about the capture of Bayezid by the troops of Temür (he calls them Scythians):
And Bayezid's misfortunes became so great that the Scythians approached him and told him “Descend from the horse, lord Bayezid, and come. Temir-chan is calling you”. Then, even though he did not want to for the horse was Arabian and very valuable, he descended from the horse. They laid [a saddle] on a small pony (ἱππάριον σμικρὸν), sat him on it and took him away to Temir-chan.
When he (i.e. Temür) was informed that Bayezid had been captured, he ordered the setting up of a tent, and sat in the tent playing chess with his son, declaring that “I do not care at all about Bayezid's capture since through my immeasurable force I [already] had him like a small sparrow in a trap (ὡς στρουθίον ἐν παγίδι)”.Footnote 76
Ducas continues with an account of the first meeting between Temür and Bayezid in which the former treats his captive graciously. Temür places him in tents and orders the digging of a ditch around them, with soldiers stationed both inside and outside this boundary. Following an attempt to mine under the tent and release the Ottoman sultan, Temür takes further measures. Most pertinently, Ducas claims, “Since then a large prison was made for him and iron collars and handcuffs (σίδηροι κλοιοὶ καὶ χɛιροπέδαι) for the night”.Footnote 77 It was in this condition that Bayezid finally died, possibly poisoning himself. Ducas concludes his account noting that Temür had intended to take Bayezid back to “Persia” in order to exhibit to his people the “sort of beast he had gained power over”.Footnote 78 Chalcocondyles offers less of interest, though he does mention that, following his capture and first interview with Temür, Bayezid was paraded on a mule (ἐπὶ ἡμιόνου) around the Timurid army camp. He also confirms the detail that Despina (called by him “Lazar's daughter”) “was paraded with him in the camp, and they had her pour wine opposite her husband”. Like Ducas he records the digging of a tunnel under the tent of the Ottoman sultan.Footnote 79
Phrantzes has the least to say. In the Chronicon minus he merely records that “the amir Bayezid was killed by Temir on 28 July”.Footnote 80 The content and brevity of Phrantzes’ treatment of this theme accords well with one of the earliest Turkish sources to deal with events in the first decade of the fifteenth century, the anonymous work entitled Aḥvāl-ı Sultān Mehemmed bin Bāyezīd Ḫān (translated recently under the title, “The Tales of Sultan Mehmed, Son of Bayezid Khan”). Completed soon after the termination of the civil war in 1413, this semi-mythic biographical text differs in character from conventional chronicles. It perhaps reflects, however, the sort of retelling of events that Phrantzes may have encountered when he sent as ambassador to sultan Murad II in 1429. In common with the chronicle of Phrantzes, “the Tales of Sultan Mehmed” deals with the fate of sultan Bayezid in the first few lines. The first paragraph deals with the victory of Temür at the battle of Ankara, and the scattering of the Turkish forces. Mehmed had found his way back to Rum, and from there he received news that his father had been taken captive and that the whereabouts of his brothers were unknown. The author continues by quoting the words apparently uttered by Mehmed in which he laments the collapse of the empire of Osman, and concludes with the following statement: “My father the sultan has been captured by the enemy. Past pleasures have turned to pain, and joy has been replaced by grief”.Footnote 81
The more extensive account of events can be found in another Greek work, based on Phrantzes’ original history, known as the Chronicon maius. As has been recognised by scholars of Byzantine history since the 1930s (but was unknown to the authors surveyed in the previous section), the Chronicon maius is an expansion of Phrantzes’ text by Makarios Melissenos (or Melissourgos, d. 1585), metropolitan of Monemvasia. The Chronicon maius was written in Naples where Makarios and his brother Theodoros had fled after leading an unsuccessful insurrection against Ottoman rule in the Peloponnese in the aftermath of the battle of Lepanto in 1571. Makarios Melissenos may also have travelled to Spain.Footnote 82 Like the history of Spandounes, another emigré Greek, Melissenos’ Chronicon maius betrays an awareness of sixteenth-century Italian writing on Turkish history.Footnote 83 The first printed edition of Melissenos’ history (credited to Phrantzes) appeared in Venice in 1604. A critical edition containing both the Chronicon minus and Chronicon maius was prepared by Vasile Grecu (d. 1972) and published in 1966. Melissenos places the alleged caging of Bayezid after his first interview with Temür (Demiris). The Ottoman sultan's haughty responses seal his fate: “When Demiris heard Bayezid's arrogant words he was angered and made an iron cell (κουβοúκλιον ἐκ σιδήρου), put him inside and after a while killed him”.Footnote 84
One of the most powerful pieces of evidence in favour of the caging of Bayezid is the testimony of Ibn ‘Arabshah. Although the evident hostility to Temür in the Kitāb al-‘ajā’ib al-maqdūr fī akhbār Tīmūr means that it cannot be viewed as an unbiased text on the subject, the relatively early date of composition (in the 1430s) and Ibn ‘Arabshah's knowledge of Persian and Turkish sources give his account considerable credibility. What is curious, however, is how little attention has been paid to what Ibn ‘Arabshah actually writes in his description of the capture and subsequent humiliations of Bayezid. The crucial passage in this respect relates to the moment in the aftermath of the battle of Ankara when the sultan was apprehended. Before reviewing the words Ibn ‘Arabshah uses, it is also worth noting that, unlike the Persian and Turkish histories, he does not discuss the first time that the Ottoman sultan is brought into the presence of Temür. Instead, the brief mention of Bayezid's capture is followed by two short chapters discussing the ensuing chaos within the Ottoman territories and the feuding of the sultan's surviving sons. Following this, Ibn ‘Arabshah records Temür's practice of daily bringing the shackled Bayezid into his presence. According to his account Temür “received him with kind and cheerful speech and marks of pity, then derided and mocked him”.Footnote 85 It is on the occasion of a public banquet that Bayezid endured the added indignity of witnessing his wives and concubines serving as cupbearers to Temür's guests. Much of this detail is recorded in abbreviated form by Ibn Taghribirdi.Footnote 86
Therefore, in Ibn ‘Arabshah's version of events the reference to the cage (qafaṣ) appears before the first recorded meeting between Bayezid and Temür and not after they have already conversed (as is suggested in the Turkish accounts). Equally significant is the precise context in which the term qafaṣ is employed. The relevant passage reads (according to the Sanders translation):
Then their arms being exhausted and the front line and reserves alike decimated, even the most distant of the enemy advanced upon them at will and strangers crushed them with swords and spears and filled pools with their blood and marshes with their limbs and Ibn Othman (Bayezid) was taken and bound with fetters (muqayyad) like a bird in a cage (ka'l-ṭayr fi'l-qafaṣ).Footnote 87
In order to render the last clause into idiomatic English, Sanders changed the expression from the definite article (as it appears in the Arabic) to the indefinite article. If one examines this clause in both the modern Arabic edition and the English translation, the most natural interpretation would be that “like” (ka) refers not simply to “a bird” but to “a bird in a cage”. In order to transform it to mean that Bayezid was indeed placed within a cage, one might expect “fi'l-qafaṣ” to be placed before “ka'l-ṭayr”, and not after. In any case, what this clause is being compared to (using ka) is the present situation of Bayezid following his capture: that he is bound or shackled (muqayyad). There is no verb in the sentence that would indicate the sense of being placed or confined within something. In other words, Ibn ‘Arabshah is evoking the image of a caged bird as a simile for the reduced state of the captured sultan. In this context it is also worth repeating that a common meaning attributed to the word, qafaṣ, in Medieval Arabic is that of a cage or coop for birds (but not a large structure meant to imprison a man).Footnote 88 If one looks elsewhere in Ibn ‘Arabshah's text, it is apparent that he frequently employed similes and metaphors as a means to heighten dramatic pitch. By contrast, the more sober historian, Ibn Taghribirdi, retained the content of Ibn ‘Arabshah's account, including the humiliating encounters between Bayezid and his captor, but omitted the more literary flourishes such as the likening of the Ottoman sultan to a bird in a cage. It is intriguing that Ducas also employs the motif of an ensnared bird in his report of Temür's reaction to the capture of Bayezid. Ducas’ metaphor also appears in the narrative before the Ottoman sultan is brought into the presence of Temür (though the context in which it is used is rather different from Ibn ‘Arabshah's).
Why then have scholars persisted with this misinterpretation of Ibn ‘Arabshah's text? The answer appears to lie in the overwhelming reliance (even by Orientalists conversant in ArabicFootnote 89) upon the earliest European translation, Pierre Vattier's (d. 1667), L'histoire du grand Tamerlan, published in Paris in 1658 rather than the Arabic edition produced by Jacob Golius (or Gool, d. 1667) and published under the Latin title, Vitae et rerum gestarum Timuri, in 1636.Footnote 90 In his translation Vattier introduces punctuation – a feature not found in Arabic texts prior to the twentieth century – which in this case significantly revises the original meaning. He translates the relevant passage rather freely as: “Le fils d'Othoman tomba dans le piège, et se trouva enfermé, comme l'oyseau, dans la cage”.Footnote 91 One may reasonably conclude that the simple addition of two commas has affected more than three hundred years of scholarship; this punctuation led many readers to assume Ibn ‘Arabshah believed that Bayezid was, like a bird, locked in a cage!Footnote 92
In conclusion, it should be emphasised that the precise details of the captivity of Bayezid are of little importance to an understanding of the early evolution of the Ottoman empire. Furthermore, it will only be with a complete analysis of the available primary sources – particularly those in Turkish – that we may be able to form more accurate conclusions concerning the ultimate origins of the story of the cage as well as its historical veracity. One might ask, therefore, whether there is any value in pursuing what might seem like such a recherché topic. Perhaps the main reason for focusing upon this issue is the light that it sheds on European Orientalist historiography in the early–modern and modern periods. Although scholars such as d'Herbelot, Voltaire, Gibbon, and Von Hammer had unburdened themselves of the baggage of Christian moral philosophy that conditioned the interpretation of Asian history in the fifteenth and sixteenth centuries, they were not immune to other prejudices. In particular this can be seen in the changing evaluation of the personality of Temür. During the Renaissance he had functioned both as the Machiavellian prince and the “scourge of God”,Footnote 93 while in the eighteenth century it was not uncommon for scholars to view him in some respects as an embodiment of Enlightenment values. The coarse barbarity of parading a captured ruler in a cage fitted uncomfortably into this vision, and was either suppressed on linguistic grounds or justified as a somewhat reluctant response to the haughty behaviour of Bayezid. Finally, another recurrent phenomenon uncovered in this research is the willingness of scholars to cite one another rather than the original sources, forming chains of transmission not unlike the isnāds found in the works of Islamic historians. Of course, it would be dangerous to assume that any of us are immune to the same flaws.
