defection across the border of islam and christianity ... · defection across the border of islam...

Defection across the Border ofIslam and Christianity: Apostasy and

Cross-Cultural Interaction inByzantine-Seljuk Relations

By Alexander D. Beihammer

Not a word of goodbye, not even a noteShe gone with the manIn the long black coat.

(Bob Dylan,“Man in the Long Black Coat”)

An Islamic coffin discovered in the church of Maria Spilaiotissa near the oldSeljuk capital of Konya in central Anatolia bears the following Greek inscrip-tion: “Here lies the descendant of men born in the purple, Michael Amiraslan,the grandson of the great-grandson of the blessed emperor born in the purple,Kyr John Komnenos Maurozomes, the son of the humble John Komnenos.”1

This text, dated November 1, 6809 (a.d. 1297), can be characterized as a relicof a family tradition of Greek dignitaries at the Seljuk court who for generationshad maintained their Christian faith and the memory of their imperial ancestry.Paul Wittek, who published and thoroughly examined the inscription in the 1930s,reconstructed a genealogical tree of the family, whose roots go back to the Byz-antine aristocracy of the late twelfth century. But I am not concerned here withthe validity of his arguments. More interesting at present is the fact that the in-scription constitutes a monument in stone to a particular mode of political be-havior that occurred with increasing frequency from the second half of the elev-enth century until the end of the Nicaean Empire and that, therefore, can bedeemed a typical feature of Byzantine-Seljuk relations during this period. I referto the decision members of the Byzantine ruling class made from time to time tosever existing social bonds based on kinship, loyalty, as well as ethnic and reli-

Earlier versions of this paper were presented at a CIHS session organized by the Society for theStudy of the Crusades and the Latin East in Sydney 2005 (University of New South Wales) and at asymposium in honor of Prof. Roderich Reinsch organized by the Department of Byzantine and Mod-ern Greek Studies of the University of Cyprus. I thank all participants for their incisive commentsand suggestions. My warmest thanks go to John S. Langdon (Los Angeles) for his numerous usefulcomments and to Chris Schabel (University of Cyprus) for his encouragement, careful reading, andnumerous linguistic amendments. A constant source of inspiration was the recent book of Jean Schotz,Anadolu Selçuklu devleti zamanında kedi ve köpek arasındaki büyük kavgalar (Mazotos, Gazima-gusa [Famagusta], 2010).

1 Paul Wittek, “L’épitaphe d’un Comnène à Konia,” Byzantion 10 (1935), 505–15, Greek text atp. 507, and idem, “Encore l’épitaphe d’un Comnène à Konia,” Byzantion 12 (1937), 207–11 (pro-poses an emendation of his first reading Μι�α�λ �µηρ&ας to �µηρασλ�νης).

Speculum 86 (2011) doi:10.1017/S0038713411001138 597

This content downloaded from 129.128.216.034 on April 08, 2017 17:44:54 PMAll use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).

gious identity and to join the Seljuk sultan of Rum, the Danismend emir, or otherMuslim lords in Asia Minor in order to obtain their favor and protection. Sim-ilar patterns of behavior can be observed among members of the political elite inthe Turkish states as well. Hence we can talk about a common phenomenon. Un-der certain circumstances, the political and religious archenemy, the morally andculturally inferior barbarian and infidel, could become an appreciated protectorand ally against one’s own relatives, coreligious, and compatriots. Sometimes ref-ugees even found a new home at the court that had given them shelter, con-verted to their hosts’ religion, and stayed there for the rest of their lives, as wasobviously the case with Michael Amiraslan’s forefathers.

Defection in Byzantine-Muslim Political Culture

In the collective memories of twentieth-century postwar generations the no-tion of “defection,” that is, the act of abandoning one’s allegiance to a state or apolitical entity in exchange for another, acquired an emotionally and morallycharged connotation, eliciting manifold ethical judgments and legal consider-ations. A spontaneous train of thought might range from Adolf Hitler’s Field-marshal Friedrich Paulus, who, after his failure at Stalingrad, instead of commit-ting suicide collaborated with the Soviets, to the Cold War period and the BerlinWall as a symbol of East Germany’s coercive measures against Republiksflucht,to contemporary Muslim intellectuals persecuted by radical Islamic organiza-tions.2 Are there any parallels between our modern experiences and medieval or,more specifically, Byzantine and Muslim standards of identity, allegiance, and loy-alty? To judge from the evidence provided by legal texts, obviously there are. Aglance at Justinian’s Digesta and the Byzantine corpus of legal codifications andtreatises leaves no doubt that Roman law had developed a clear attitude in treat-ing defectio (α�τ�µ�λ�α) as a crime approaching high treason, which in all casesentailed the death penalty and, because of its seriousness, especially ignominiousforms of execution. Only in the early tenth century did Emperor Leo the Wiseintroduce some modifications, arguing in one of his novels for milder penaltiesfor defectors who were repentant of their deed.3 The religious-based legal systemof the Islamic world considered defection primarily in terms of spiritual apos-tasy from Islam (ridda or irtidad ), which, because of the theocratic character ofIslamic state theories, implicitly included all forms of political defection from Mus-lim public authorities. Although pertinent statements in the Qur©an are far fromunambiguous, apostasy, from the earliest preserved collections of Islamic law on-wards, came to be deemed as a “sin unto death,” even excluding the possibilityof repentance (tawba), according to some opinions.4 However, the undisguised

2 For modern definitions of “defection” see, for example, J. A. Simpson and E. S. C. Weiner, TheOxford English Dictionary, 16 vols. (Oxford, 1989; repr. 1991), 4:374.

3 For “defection” in Byzantine law see Spyros Troianos, � “π�ιν�λι�ς” τ�υ Εκλ�γαδ��υ: Συµ��λ�εις την ιστ�ρ�αν της ε�ελ��εως τ�υ π�ινικ�� δικα��υ απ τ�υ Corpus iuris civilis µ!�ρι τωνΒασιλικ#ν (Frankfurt am Main, 1980), pp. 21–23.

4 For Muslim attitudes toward apostasy in general, see Samuel M. Zwemer, The Law of Apostasyin Islam (New York, 1924), and the article of W. Heffening on “murtadd” in The Encyclopaedia of

598 Defection


condemnation of defection in both spheres’ moral and legal systems never pre-vented people involved in seditious movements from taking recourse to it, and,therefore, the history of Byzantine-Muslim contacts from the seventh century on-wards provides a long list of prominent apostates, such as the Armenian generalSaborius under Emperor Constans (641–69); the domestikos Manuel and the Per-sian rebel Nas�r-Theophobus during the reign of Theophilus (829–42); the Arabeunuch Samonas, for some years one of the most powerful men at the court ofLeo VI; and Bardas Skleros, the legendary opponent of Emperor Basil II (976–1025). The Byzantine epic of the “twyborn” hero Digenis Akritas, son of an Arabemir and a Greek general’s daughter, perfectly illustrates how familiar a phenom-enon such close interaction across the Anatolian border had become over the cen-turies.5

For the period following the emergence of the Seljuk Turks in Asia Minor, mod-ern scholars have repeatedly noted and discussed the sudden increase of testimo-nies of refugees, apostates, and outlaws living somewhere in the space in be-tween the political and cultural spheres. The two main authorities in the field,Claude Cahen and Speros Vryonis, have considered the phenomenon within thewider context of Christian-Muslim coexistence in the conquered regions of AsiaMinor, namely, with respect to the Seljuk sultanate’s attitudes toward its Chris-tian subjects and, especially, the integration of Christians into the Seljuk courtlife and administration.6 Both authors portray a society that, on the one hand,

Islam, new ed., 13 vols. (Leiden, 1960–2009), 7:635–36. For illustrative examples of early legal opin-ions on apostasy and the death penalty see Malik b. Anas, Kitab al-muwat�t�a©, ed. Najıb Majidı (Bei-rut, 2000), 1444, pp. 411–12, and at�-T�abarı, Jami¨ al-bayan ¨an ta©wıl ay al-qur©an al-ma¨ruf tafsıral-T�abarı, ed. Muh�ammad Shakir, 30 parts in 16 vols. (Beirut, n.d.), 3/2:394–402.

5 For the mentioned incidents of defection see Andreas Kaplony, Konstantinopel und Damaskus:Gesandtschaften und Verträge zwischen Kaisern und Kalifen 639–750, Islamkundliche Untersuchung-en 206 (Berlin, 1996), pp. 51–75; Maria Leontsini, Κωνσταντ�ν�ς ∆& (668–685): � τελευτα��ςπρωτ��υ'αντιν ς αυτ�κρ�τ�ρας [Constantine IV (668–685): The last emperor of the early Byzan-tine period], Institute of Byzantine Research, Monographs 7 (Athens, 2006), pp. 74–79 (Saborius);Jean-Claude Cheynet, “Théophile, Théophobe et les Perses,” in Η Βυ'αντιν� Μικρ� Ασ�α (6�ς–12�ς αι.) [Byzantine Asia Minor (six–twelfth centuries)], ed. Stelios Lampakis, Institute for Byzan-tine Research, International Symposia 6 (Athens, 1998), pp. 39–50 (Nas�r-Theophobus); Juan SignesCodoñer, “Lust am Erzählen: Heiligenviten als Grundlage der Geschichtsschreibung im 10. Jahrhun-dert und der Weg nach Bagdad,” in L’écriture de la mémoire: La littéralité de l’historiographie. Actesdu IIIe colloque international philologique, Nicosie, 6–8 mai 2004, ed. Paolo Odorico, PanagiotisA. Agapitos, and Martin Hinterberger, Dossiers Byzantins 6 (Paris, 2006), pp. 85–106 (Manuel);R. J. H. Jenkins, “The ‘Flight’ of Samonas,” Speculum 23 (1948), 217–35; Shaun Tougher, The Reignof Leo VI (886–912): Politics and People, The Medieval Mediterranean 15 (Leiden, 1997), pp. 200–201, 208–15 (Samonas); and Alexander Beihammer, “Der harte Sturz des Bardas Skleros: Eine Fall-studie zu zwischenstaatlicher Kommunikation und Konfliktführung in der byzantinisch-arabischenDiplomatie des 10. Jahrhunderts,” Römische historische Mitteilungen 45 (2003), 21–57 (Bardas Skle-ros). For Digenis see Digenes Akrites: New Approaches to Byzantine Heroic Poetry, ed. RoderickBeaton and David Ricks, Centre for Hellenic Studies, King’s College London, Publications 2 (Alder-shot, Eng., 1993).

6 Claude Cahen, Pre-Ottoman Turkey: A General Survey of the Material and Spiritual Culture andHistory c. 1071–1330 (New York, 1968), pp. 202–15; idem, The Formation of Turkey: The SeljukidSultanate of Rum, Eleventh to Fourteenth Century, trans. Peter M. Holt (Harlow, Eng., 2001), pp. 123–33; also the same author’s case study on the Gavras family: “Une famille byzantine au service des

599Defection


shows some awareness of traditional Islamic rules concerning the restrictions onnon-Muslim subjects and, on the other, is surprisingly tolerant toward Chris-tians, so that several sultans were born of Greek women and a considerable num-ber of Byzantine aristocrats, both locals and refugees, became high-ranking offi-cials at the court of Konya. Cahen even suggests that Byzantine and Seljuk refugeesshared a special preference for each other, which was stronger than the ties withtheir respective coreligious in the Slavic lands or in Muslim Syria and assumes acommon consciousness of belonging to one cultural entity, the “Roman land”(bilad al-Rum).7 Vryonis, whose view of the Turkish conquest of Asia Minor isoverwhelmingly negative, presenting it as a process of decay of Christian Helle-nism, at least admits that Christian officials and apostates at the Seljuk courtformed a significant force for bridging the gap between the conquerors and thevanquished by offering opportunities to the Turks to become familiar with Chris-tian society’s customs and habits.8

Charles Brand has devoted an article to the opposite phenomenon, that is, theTurkish element in the Byzantine world in the eleventh and twelfth centuries.9

He lists nineteen persons who either lived permanently at or visited the Byzan-tine imperial court, and he briefly discusses related issues such as the integrationprocess through grants of titles and incomes, baptism and marriage, and the roleof Turkish mercenaries in the Comnenian army. More recent studies of Alain Du-cellier and Michel Balivet,10 which are concerned with general aspects of Byzantine-Turkish coexistence, acculturation, and mutual perception, further develop thepositive perspective of Claude Cahen by stressing the manifold phenomena ofcross-cultural contacts and permeation during the twelfth and thirteenth centu-ries: Turkish mercenaries offering their services to the imperial government, Turk-ish merchants living among other Muslim inhabitants of Constantinople and en-joying full religious freedom in the framework of their own mosque, anoverwhelming majority of Greek-speaking subjects in the sultanate of Konya, agreat number of Greek court officials including Greek notaries in the Seljuk chan-cery, and features of Byzantine imperial ideology incorporated into Seljuk cer-

Seldjuqides d’Asie Mineure,” in Polychronion: Festschrift Franz Dölger zum 75. Geburtstag, ed. Pe-ter Wirth (Heidelberg, 1966), pp. 145–49; Speros Vryonis, The Decline of Medieval Hellenism inAsia Minor and the Process of Islamization from the Eleventh through the Fifteenth Century (Berke-ley, Calif., 1971), Greek translation by Katia Galatariotou, 2nd ed. (Athens, 2000), pp. 199–213,esp. pp. 203–8.

7 Cahen, Pre-Ottoman Turkey, p. 211; idem, Formation, p. 130.8 Vryonis, Decline, p. 205.9 Charles M. Brand, “The Turkish Element in Byzantium, Eleventh–Twelfth Centuries,” Dumbar-

ton Oaks Papers 43 (1989), 1–25.10 Alain Ducellier, “Mentalité historique et realités politiques: L’islam et les musulmans vus par les

Byzantins du XIIIeme siècle,” Byzantinische Forschungen 4 (1972), 31–63; idem, Chrétiens d’Orientet islam au moyen âge, VIIe–XVe siècle (Paris, 1996), pp. 260–75; Michel Balivet, Romanie byzan-tine et pays de Rûm turc: Histoire d’un espace d’imbrication gréco-turque, Les Cahiers du Bosphore10 (Istanbul, 1994), pp. 30–39, 47–53; idem, “Entre Byzance et Konya: L’intercirculation des idéeset des hommes au temps des Seldjoukides,” in idem, Mélanges byzantins, seldjoukides et ottomans,Analecta Isisiana 81 (Istanbul, 2005), pp. 47–79; idem, “Intégration et exclusion des chrétiens dansle sultanat seldjoukide d’Asie-Mineure (XIIe–XIIIe s.),” ibid., pp. 107–24.

600 Defection


emonies of lordship. All these circumstances, their argument goes, point to a highdegree of “Greek-Turkish familiarity” and a relationship “that goes far beyondoccasional alliances.”11 In this respect, the declarations Byzantine authors fre-quently pronounce concerning the barbarian character of the Muslims, describ-ing them as ferocious, cruel, unreliable, greedy, and, in general, much inferior tothe Romans,12 should not be taken at face value since they reflect inherited ste-reotypes and literary conventions rather than everyday experiences and prevail-ing attitudes. Hence, Balivet repeatedly underlines the historical importance ofthe axis of exchange between Constantinople and Konya, even pointing to sym-metrical features in the evolutional process of the states, such as the predomi-nance of centrifugal dynamics in the later part of the twelfth century.13 The com-ing and going of political and military leaders served as an additional means forthe transfer of cultural attitudes and knowledge to each other.14

The present study aims to show that the phenomenon of Byzantine-Turkishdefection encompasses much more ambivalence and complexity than hitherto ob-served. No doubt, there actually existed a strong tendency toward acculturationbetween Byzantine and Turkish elites in the twelfth and thirteenth centuries, butthe interpretative model of cross-cultural interaction by no means explains thesituation exhaustively. In what follows I will argue that defection was not just aresult of acculturative processes but, first and foremost, constituted an essentialelement of Byzantine-Seljuk political culture. It was an important means of ex-erting pressure in several directions, whether as an immediate expression of dis-content or as a threat enabling political opponents to make their claims and toreach compromises. Furthermore, defection was an effective weapon of politicalpropaganda that was frequently accompanied by a sequence of publicly per-formed symbolic gestures and ritual acts and allowed a defector’s hosting courtto celebrate the event as a moral victory over the enemy.

A necessary prerequisite for an appropriate understanding of these forms ofpolitical behavior is the reconstruction of their chronological development. Forexample, the fact that cases of defection are mentioned even in the oldest reportson the first Seljuk invasions into Byzantine territory clearly indicates that the phe-nomenon, at least at this early stage, could not have resulted from a sort of men-tal proximity and common identity. It goes without saying that within a span of

11 Balivet, Romanie byzantine, pp. 52–53; Ducellier, Chrétiens d’Orient, p. 264: “Grecs et Turcs. . . avaient fini par comprendre que leur destin était de vivre ensemble.” For similar views see NevraNecipoglu, “Turks and Byzantines (Eleventh and Twelfth Centuries),” in The Turkish Speaking Peo-ples: 2000 Years of Art and Culture from Inner Asia to the Balkans, ed. Ergun Cagatay (Munich,2006), pp. 255–65, at p. 255: “. . . bonds of solidarity . . . were formed between native Greeks andthe Turkish newcomers through neighbourly relations . . . , a high degree of coexistence and symbi-osis, despite the ongoing Byzantine-Turkish wars.”

12 Ducellier, “Mentalité historique,” pp. 31–35.13 Balivet, Romanie byzantine, p. 46: “Les phénomènes sont rigoureusement symétriques à By-

zance et à Konya.”14 Ibid. p. 47: “les multiples occasions de contacts micrasiatiques . . . finissent par engendrer des

cadres politiques et militaires presque ‘interchangeables.’ . . . [L]a présence et les activités d’une éliteoriginaire du monde adverse . . . joue dans le sens d’une meilleure connaissance de la société rivale etd’une accentuation des échanges turco-byzantins.”

601Defection


almost two centuries the motives of defectors changed according to the evolu-tional framework of particular political and social circumstances. In addition, theintermediary role of the available narrative sources, on which we exclusively de-pend, compels us to shift the focus from the factual level of defection as histor-ical event to the imaginative level of defection as a matter of historiographicaldiscourse. The perceptive patterns of our source material are to a large extentconditioned by the retarding burden of literary traditions and inherited stereo-types. But there is also a close interplay between fact and memory, and in mostcases historical events, especially when they touch upon highly sensitive points,such as political order and moral standards, become vehicles for expressing theviews and interests of contemporary and later observers. Consequently, our in-formants describe the behavior of individual defectors and their hosts in accor-dance with the conceptual and ideological framework determining the narrativemacrostructure and the chief arguments of their text.15 As will be shown below,an essential part of the historiographical discourse in question consists of narra-tive representations of mental conditions and emotional reactions related to thebreaking and the reestablishment of bonds of allegiance.16 Another constantly oc-curing leitmotif concerns the ritualized behavior of all sorts that figures in offi-cial encounters, reception ceremonies, peace negotiations, and reconciliationscenes.17 Therefore, the primary focus of the following analysis revolves around

15 For concepts and patterns in medieval historiography see, for instance, Hans-Werner Goetz, Ge-schichtsschreibung und Geschichtsbewußtsein im hohen Mittelalter, Vorstellungswelten des Mittelal-ters 1 (Berlin, 1999), and idem, “Die Wahrnehmung von ‘Staat’ und ‘Herrschaft’ im frühen Mittel-alter,” in Staat im frühen Mittelalter, ed. Stuart Airlie, Walter Pohl, and Helmut Reimitz, Forschungenzur Geschichte des Mittelalters 11 (Vienna, 2006), pp. 39–58, at pp. 39–44.

16 For emotions in Byzantine historiographical texts see Martin Hinterberger, “Φ � +ω κατασεισθε�ς:τα π�θη τ�υ ανθρ#π�υ και της αυτ�κρατ�ρ�ας στ�ν Μι�α�λ Ατταλει�τη: Τ� αιτι�λ�γικ σ�στηµαεν ς ιστ�ρι�γρ��υ τ�υ 11�υ αι#να,” in Η αυτ�κρατ�ρ�α σε κρ�ση; Τ� Βυ'�ντι� τ�ν 11� αι#να(1025–1081) [The empire in crisis? Byzantium in the eleventh century (1025–1081)], ed. VassilikiVlyssidou, Institute for Byzantine Research, International Symposia 11 (Athens, 2003), pp. 155–67;idem, “Tränen in der byzantinischen Literatur: Ein Beitrag zur Geschichte der Emotionen,” Jahr-buch der österreichischen Byzantinistik 56 (2006), 27–51; and Michael Grünbart, “Der Kaiser weint:Anmerkungen zur imperialen Inszenierung von Emotionen in Byzanz,” Frühmittelalterliche Studien42 (2008), 89–108. For a comparison with the medieval West see the articles collected in Anger’sPast: The Social Uses of an Emotion in the Middle Ages, ed. Barbara H. Rosenwein (New York 1998),and Gerd Althoff, “Empörung, Tränen, Zerknirschung: ‘Emotionen’ in der öffentlichen Kommuni-kation des Mittelalters,” Frühmittelalterliche Studien 30 (1996), 60–79. A very useful general surveyof the current state of research on emotions in medieval studies can be found in Rudolf Schnell, “His-torische Emotionsforschung: Eine mediävistische Standortsbestimmung,” Frühmittelalterliche Stu-dien 38 (2004), 173–276.

17 Basic guides to recent approaches to medieval ceremonies and rituals are Hagen Keller, “Die In-vestitur: Ein Beitrag zum Problem der ‘Staatssymbolik’ im Hochmittelalter,” Frühmittelalterliche Stu-dien 27 (1993), 51–86; Gerd Althoff, Spielregeln der Politik im Mittelalter: Kommunikation in Friedenund Fehde (Darmstadt, 1997); Hagen Keller, “Ritual, Symbolik und Visualisierung in der Kultur desottonischen Reiches,” Frühmittelalterliche Studien 35 (2001), 23–59; Philippe Buc, The Dangers ofRitual: Between Early Medieval Texts and Social Scientific Theory (Princeton, N.J., 2001); and Da-vid A. Warner, “Ritual and Memory in the Ottonian Reich: The Ceremony of Adventus,” Speculum76 (2001), 255–83. The most comprehensive monograph on rituals in the field of Byzantine studiesis Michael McCormick, Eternal Victory: Triumphal Rulership in Late Antiquity, Byzantium and theEarly Medieval West (Cambridge, Eng., 1986).

602 Defection


emotions and rituals as key elements of the preserved narratives on the phenom-enon of defection, interpreting shifts in the discursive approaches as indicationsof changing circumstances and attitudes in the historical process.

As far as the general framework of power politics between Constantinople andthe newly established Turkish lordships of Asia Minor is concerned, the overallsituation after the first great wave of Seljuk invasions from the 1040s until thelate 1070s and throughout the twelfth century, despite tendencies of cross-culturalpermeation and assimilation on some levels, can be characterized as highly un-stable and, in certain periods, even relentlessly hostile. Given the state of affairsfollowing the destruction of the Turkish emirate of Nicaea and the reconquest ofthe western coastal regions in the wake of the passage of the First Crusade (1097),it is obvious that diplomatic contacts and peace treaties never led to a long-termstabilization of the border regions and that temporary alliances mainly aimed attaking advantage of internal conflicts among Turkish rivals without having pos-itive effects on the general conditions of Byzantine-Turkish relations. After Alex-ios I fought his final battles with the sultanate of Konya in Bithynia and Phrygia(1112–13 and 1116), his son John II, either reacting to Turkish attacks or on hisown initiative, undertook a series of seven campaigns within his reign of almosttwenty-five years, which among other directions led him to Laodicea in the Me-ander valley, to Sozopolis in Pisidia (both 1119), to Kastamon and Gangra inPaphlagonia (1130 and 1134), into the valley of the Sangarios River (1138), toNeokaisareia in Cappadocia (1139), and to Lake Pusguse (Beysehir Gölü) in Phry-gia (1141). Manuel I, in the first years of his reign, managed to fortify the bor-derlands of Bithynia and undertook a large-scale campaign against the Seljuk sul-tanate, culminating in a number of military successes in Phrygia and the siege ofKonya in 1146. The difficulties that the Second Crusade encountered during themarch through southwestern Asia Minor, however, clearly indicate how strong apolitical and military factor the Seljuk Turks had become in the decades sinceAlexios I’s death. In the years of his European wars against the Normans and inItaly, Manuel took advantage of the ambitions of Sultan Mas¨ud (1116–55) toexpand his authority over Danismend provinces and Frankish and Armenian re-gions in the southeast of Asia Minor. The peace treaty with Sultan Kılıc ArslanII (1162) had no lasting effect, so that from 1165 onwards the situation wasmarked by an increasing number of Turkish incursions into Byzantine territory,against which Emperor Manuel fortified strongholds in the border region (Neokas-tra, Dorylaion, Subleon) and finally decided to launch a major offensive againstKonya, leading to the catastrophe of Myriokephalon in 1176. The decay of theByzantine imperial government in the last two decades of the twelfth centurybrought about a destructive interplay of local uprisings and Turkish incursions,which did not cause immediate losses of Byzantine provinces to the Seljuk sul-tanate, most likely only because of the succession wars following Kılıc Arslan’sdeath in 1192.18 Hence, the words that Niketas Choniates put into the mouth of

18 For the events see Cahen, Pre-Ottoman Turkey, pp. 91–106; idem, Formation, pp. 7–33; Vryo-nis, Decline, pp. 103–18; and Paul Magdalino, The Empire of Manuel I Komnenos, 1143–1180 (Cam-bridge, Eng., 1993), pp. 35–36, 42, 76–78, 95–98. The most comprehensive account of the eventsfrom the viewpoint of modern Turkish historiography is Osman Turan, Selçuklar zamanında Türkiye:

603Defection


King Louis VII addressing his soldiers at the bank of the Meander River in 1147appear as a programmatic statement on a major goal of Byzantine imperial pol-icy in that period: “I do not know why the Romans, as if they were their sacri-ficial animals, bring up these wolves and shamefully fatten them with their blood.They should protect themselves and think reasonably so that they expel them fromtheir lands and cities like wild animals from the herds.”19

In a prayer inserted into his report on Emperor Manuel’s contacts with KılıcArslan II, Choniates characterizes the Turks as “a silly and unwise people whoare far from being full of pious praise and faith toward you,” as “crafty aliens,”and “our wicked neighbors,” against whom God should protect his heritage andrecuperate the lands and cities of the Romans.20 Accordingly, what “the ruler ofthe Turks,” Mas¨ud, shared with his relatives are in fact “territories of the Ro-mans,”21 and his death as an infidel leads to nothing short of a trip to hell.22

Such statements, to which one could add many similar ones, leave no doubt thatthe official stance of the Byzantine political and ecclesiastical elite toward the Turkswas overwhelmingly hostile. Decades of warfare, in which the empire had suf-fered major setbacks and had lost the greater part of its eastern provinces, madethe traditional discourse on the “barbarian threat” an essential part of the Byz-antine experience and perception of the Turkish foe. In other words, inheritedstereotypes came to express prevailing attitudes and political concepts, which lefthardly any room for justifications of crossing the border to the enemy’s camp,especially in cases where members of the political elite were involved.

Yet contemporary narratives still reveal some serious reasons why high-rankingindividuals under certain circumstances were ready to ignore their scruples andviewed switching sides to the barbarians as a tempting choice. A case in point isone of Choniates’s comments on the political situation in the years of Alexios III(1195–1203), when he refers to the dangers that defectors posed for the safetyof the provinces and the existence of the empire itself: “If there is a main reasonthat the state of the Romans fell to its knees, that lands and cities suffered sub-jugations, and, finally, that the state itself was destroyed, it is these people whobroke away from the Comnenians, being eager to become emperors, because by

Siyâsi Tarih Alp Arslan©dan Osman Gazi©ye (1071–1318), 3rd ed. (Istanbul, 1993), pp. 95–236, esp.pp. 153–61, 167–77, 180–86, 190–211, 213–16. For a brief survey of general characteristics ofByzantine-Seljuk relations in the twelfth century see Ralph-Johannes Lilie, “Twelfth-Century Byzan-tine and Turkish States,” Byzantinische Forschungen 16 (1991), 35–51 (whose views of a peacefulcoexistence in spite of continuous warfare seem to me, however, somewhat questionable); in addi-tion, see idem, “Die Schlacht von Myriokephalon (1176): Auswirkungen auf das byzantinische Reichim ausgehenden 12. Jahrhundert,” Revue des études byzantines 35 (1977), 257–75.

19 Niketas Choniates, Historia, ed. Jan Louis van Dieten, Corpus Fontium Historiae Byzantinae11/1 (Berlin, 1975), p. 70; trans. Harry J. Magoulias, O City of Byzantium: Annals of Niketas Choni-ates (Detroit, 1984), p. 41 (quotations draw on Magoulias’s translation but also include frequentchanges and corrections according to my own understanding of the original text).

20 Choniates, ed. van Dieten, pp. 116–17, trans. Magoulias, p. 66.21 Choniates, ed. van Dieten, p. 117, lines 8–9, trans. Magoulias, p. 66: τ. τ &ων /Ρωµα�ων

σ��ιν�σµατα.22 Choniates, ed. van Dieten, p. 116, lines 70–71, trans. Magoulias, p. 66: 1ν δ2 τ&+ω µ!λλειν 1�

�νθρ#πων γ�νεσθαι κα3 τα&ις 1κε&ιθεν κ�λ�σεσιν 4ς �σε��ς παραπ!µπεσθαι.

604 Defection


becoming guests of nations that did not have friendly intentions toward the Ro-mans they became their fatherland’s ruin, although, as long as they stayed withus, they were completely incompetent and the most useless and foolish of all peo-ple in handling, taking over, and controlling state affairs.”23

Here defection appears as a highly effective instrument for expressing discon-tent and initiating revolts, something that undermines the existing political orderby turning powerful opponents against the empire and, therefore, engenders di-sastrous consequences for the ruling elite’s stability. Equally noteworthy are thecontemptuous and disdainful characterizations Choniates uses for members ofthe Byzantine aristocracy who sought to strengthen their position by securing thesupport of the empire’s enemies. Portraying them as feebleminded, incompetent,and foolish outsiders, Choniates presents them in an extremely negative light,which, instead of revealing to us their actual motives and intentions, reflects thecourt aristocracy’s precarious situation a few years before the final collapse of1204. Contemporary observers like Choniates perceived the tensions and con-flicts among the various factions of the ruling elite as dangerous results of theirrationality and excessive ambitions of certain individuals.

Another passage from the same author informs us that defection was also apowerful instrument of conflict resolution. During the peace negotiations be-tween Emperor Isaac II Angelos and the former supporters of the rebel AlexiosBranas, who in 1187 lost his life on the battlefield at the hands of Conrad ofMontferrat,24 the emissaries threatened the emperor with going over to the en-emy if he was not willing to grant them full amnesty: “. . . they will flee fromthe emperor’s face and seek their last residence with nations hostile to the Ro-mans, and they will do against the Romans whatever is useful to the people theytake refuge with, for it is nothing new if someone goes to his rival and flattershim as a friend, when he finds his relative as an enemy.”25 Obviously, in situa-tions where rebels were caught in a hopeless impasse, the threat of defection couldexert pressure on the other side, so that it prepared the ground for an acceptablecompromise and ensured the opponents’ personal safety.

Another aspect of defection that has to be taken into account is the foreignruler’s moral obligation to grant protection to refugees he has agreed to shelterat his court. This was not just a matter of diplomatic etiquette but a basic prin-ciple of righteous behavior toward a guest, the breach of which could have seri-ous consequences for a ruler’s claim to legitimacy. An illustrative example is Sul-tan Ghiyath al-Dın Kaykhusrau’s decision in 1192/93 to hand over the Byzantinearistocrat Theodore Mangaphas to Emperor Isaac II in return for promises con-cerning the rebel’s security: “. . . his brothers, who had taken over their father’s

23 Choniates, ed. van Dieten, p. 529, lines 25–31, trans. Magoulias, p. 290: Ε5 τι �&�ν α6τι#τατ�ντ�&υ τ�ν /Ρωµα�ων �ρ��ν 1ς γ νυ καταπεσε&ιν κα3 �ωρ&ων κα3 π λεων �ειρ#σεις παθε&ιν, 7στατα δ2κα3 α�τ�ν 1�απ�λωλ!ναι, τ�&υτ� �8 1κ Κ�µνην&ων γεγ νασιν ��ιστ�µεν�ι κα3 �ασιλει&ωντες9 παρ.γ.ρ :θνεσι καταλ��ντες µ� ��λα /Ρωµα��ις ν��&υσι τ&ης πατρ�δ�ς &;σαν παν#λεια, κα�περ 1ν τ&+ω παρ<=µ&ιν µ!νειν πρ�γµασιν 1γ�ειρ�σειν >λε&ιν τε κα3 κρατ�σειν �νεπιτηδει τατ�ι ?ντες κα3 π�ντων��ρει τατ�ι κα3 σκαι τατ�ι.

24 Charles M. Brand, Byzantium Confronts the West, 1180–1204 (Cambridge, Mass., 1968), pp. 80–84.

25 Choniates, ed. van Dieten, p. 390, trans. Magoulias, p. 214.

605Defection


rule along with him, considered the sultan’s deed so repulsive that they wouldhave raised weapons against him, for in return for gifts he seized and gave awayto the emperor of the Romans a man who had come to him voluntarily, if he[the sultan] had not formulated a nice-sounding answer stating that he sent theman on the basis of agreements and by no means gave him away but rather, asthe man was a vagrant, he led him back to his homeland so that he might stoppursuing or being pursued.”26 Thus, treacherous behavior toward a guest not onlydamaged the ruler’s image, which was propagated in the framework of his con-tacts with the outside world, but also detracted from his overall moral standingin the eyes of his own relatives and subjects. But let us turn now to the very on-set of Byzantine-Seljuk contacts in order to explore the grounds on which theaforementioned practices and attitudes were built.

Turkish Warlords and PotentatesVisiting the Imperial Court (ca. 1050–1118)

The first known incidents of defection occurred in the years of the earliest Sel-juk campaigns in the eastern provinces of Byzantium. Along with the increasingintrusion of Turkish tribes into Asia Minor and the creation of Turkish strong-holds and local lordships from the 1070s onwards, the coming and going of high-ranking individuals from both sides became a more or less constant phenom-enon, which engendered the development of new ideological attitudes and patternsof behavior. By the end of Alexios I’s reign one can talk about the existence offirmly established practices in the handling of defectors from either side.

Considering the overall picture provided by the sources, it appears that in thisearly period the Byzantine court was much more attractive for the Turkish in-vaders than Seljuk territories were for the Byzantines. In fact, the empire had anancient tradition of enrolling barbarian forces into its army and of integratingformer prisoners of war into its administrative and military apparatus, therebyenabling them to become members of the Byzantine court elite.27 From an ideo-logical point of view, these practices were underpinned by a strong conscious-ness of cultural, political, and moral superiority.28 This attitude is most clearlyexpressed in idealized portrayals of contact situations, such as Michael Atta-leiates’s description of the encounter between Sultan Togrıl Beg and the Geor-gian prince Liparites, who was taken prisoner by Ibrahım Inal’s forces during thebattle of Kapetrou in 1048. Other sources inform us that the sultan released

26 Choniates, ed. van Dieten, p. 401, trans. Magoulias, p. 220.27 For foreign units serving in the imperial forces see John Haldon, Warfare, State and Society in

the Byzantine World, 565–1204 (London, 1999), pp. 119–20, 125–26, 227–28. For the Byzantinearmy in the eleventh century see idem, “Approaches to an Alternative Military History of the Periodca. 1025–1071,” in Η αυτ�κρατ�ρ�α σε κρ�ση; (above, n. 16), pp. 45–74. For Turks in Byzantineservice see Brand, “Turkish Element” (above, n. 9), pp. 12–25.

28 For concepts of superiority in Byzantium see, for example, Alexander Kazhdan, “The Notion ofByzantine Diplomacy,” in Byzantine Diplomacy: Papers from the Twenty-Fourth Spring Symposiumof Byzantine Studies, Cambridge, March 1990, ed. Jonathan Shepard and Simon Franklin (Alder-shot, Eng., 1992), pp. 3–21, esp. at pp. 14–20; and Jonathan Franklin, “Byzantine Diplomacy, A.D.800–1204: Means and Ends,” ibid., pp. 41–71.

606 Defection


Liparites from captivity without demanding a ransom in the course of the peacenegotiations with Emperor Constantine IX in 1049. According to Attaleiates, how-ever, the sultan’s decision was due to his deep admiration for Liparites’s Romanvirtues, that is, his noble descent, his fame, his bravery, and his steadfastness.29

Attaleiates obviously intended to present the empire’s military defeat as a moralvictory over the Turkish barbarians, who, though superior on the battlefield, werestill overwhelmed by Roman cultural values.

The same concept underlies John Skylitzes’s account of the first recorded at-tempt of defection. Interestingly, according to this source the series of Turkishapostates starts with Togrıl Beg’s famous cousin Kutlumus, who through his sonSulayman became the forefather of the Seljuk sultans of Rum in Asia Minor. Sky-litzes’s highly confused account of the discord between the sultan and his favor-ite commander, which broke out in about 1057, does not deserve much cre-dence. In particular, the detail that Kutlumus sent a message to the Byzantineemperor asking him to grant him refuge so that he would become an “ally andfriend of the Romans”30 cannot be confirmed by other sources. Yet the ideolog-ical framework of the story deserves our attention: the Seljuk family’s seditiousmember is portrayed as a victim of the sultan’s irrational anger, which inducesKutlumus to acknowledge the Byzantine emperor’s superior sense of justice andthe righteousness of his rule.

Kutlumus’s alleged proposal to become the emperor’s subject might reflect laterknowledge about his sons’ temporary alliance with Emperor Nikephoros III. Bethat as it may, in the years between 1056 and 1078 we know about several mil-itary leaders who in fact took refuge with the Byzantine emperor. No matter whattheir original rank within the Seljuk hierarchy had been, they were all receivedwith much honor and respect and subsequently served as commanders in the ranksof the imperial forces. A certain Amertikes, who has been identified with a Tur-coman chief called Ibn Khan al-Turkumanı, the amır al-Ghuzz (leader of theOghuz Turks), was received by Emperor Michael VI in about 1056/57 “with greatfriendliness.”31 Attaleiates calls him a “distinguished man,” but, as his militaryforce after his entering the service of the Mirdasid emirs ¨At�ıya and Mah�mud ofAleppo in 1064 did not comprise more than five hundred warriors, he most likelywas a man of mediocre standing.32

29 Michael Attaleiates, Historia, ed. and trans. Immaculada Pérez Martín, Nueva Roma 15 (Ma-drid, 2002), p. 34. For the event and the other sources see Werner Felix, Byzanz und die islamischeWelt im früheren 11. Jahrhundert: Geschichte der politischen Beziehungen von 1001 bis 1055, Byzan-tina Vindobonensia 14 (Vienna, 1981), p. 170.

30 Ioannes Skylitzes, Synopsis historiarum, ed. Johannes Thurn, Corpus Fontium Historiae Byzan-tinae 5 (Berlin, 1973), p. 474, esp. lines 91–93; trans. Bernard Flusin and Jean-Claude Cheynet, JeanSkylitzès, Empereurs de Constantinople, Réalités Byzantines 8 (Paris, 2003), p. 391: πρ�σδε�θ&ηναικα3 σ�µµα��ς κα3 ��λ�ς γεν!σθαι /Ρωµα�ων. For the historical background see Claude Cahen, “Qut-lumush et ses fils avant l’Asie Mineure,” Der Islam 39 (1964), 14–27, repr. in idem, Turcobyzantinaet Oriens Christianus, Collected Studies Series 34 (London, 1974), no. 5.

31 Attaleiates, ed. Pérez Martín, p. 71, line 24: µεγ�λων τυ�@ν δε�ι#σεων.32 On this person see Claude Cahen, “La première pénétration turque en Asie-Mineure (seconde

moitié du XIe s.),” Byzantion 18 (1946–48), 5–67, at pp. 25–26, repr. in Turcobyzantina, no. 1. Themain sources are Ghars al-Ni¨ma, extensively quoted in the chronicle of Sibt� b. al-Jawzı, Mir©at al-

607Defection


In contrast, Arısghı (Erisgen) or Chrysoskoulos, who in 1070 fled to Ro-manos IV Diogenes, is designated by a contemporary Arabic source as “the hus-band of the sultan’s sister,” that is, Alp Arslan’s brother-in-law.33 Hence, he musthave been one of the highest commanders in the Seljuk army. Both Byzantineand Arabic sources agree that his flight into Byzantine territory was due to hisdiscord with the sultan, who had dispatched al-Afshın, another famous militaryleader, to pursue him. After a decisive victory over the Byzantine forces of Man-uel Komnenos, Arısghı came to an agreement with the latter and followed himto Constantinople.34 On this occasion Attaleiates gives us the oldest extant de-scription of an official reception ceremony held for a Turkish refugee during agathering of the senate in the imperial palace’s throne hall, the Chrysotrikli-nos.35 On the one hand, the ceremony emphasized the cultural gap dividing theRoman-Christian and the barbarian worlds through a gesture of undisguised dis-respect, namely, shouts of astonishment pronounced by the attending senatorswhile the “ugly Scythian” was entering the hall. On the other hand, it projectedthe barbarian’s acceptance within the ranks of the Byzantine court hierarchythrough the benevolence of the emperor, who bestowed the title of proedros uponhim.

Eight years later, in early 1078, the sons of Kutlumus presented themselves be-fore Nikephoros III Botaneiates in Nicaea. Attaleiates describes them as Persiannoblemen of royal descent who competed with the Seljuk overlord (Malik Shah)for the sultanate.36 According to the highly encomiastic character of Attaleiates’saccount of Botaneiates’s ascent to the throne, the event appears as one of thegreat successes of the newly acclaimed emperor. The author especially stressesthe significance of the prostration that “these people of royal rank, who wouldnot do this for any Persian king,” performed before the emperor, while the em-peror “made them through his prudent words and gestures more decent and welldisposed, induced them to submit themselves and to be faithful to him, and made

zaman fı tarıkh al-a¨yan, ed. Ali Sevim (Ankara, 1968), pp. 101, 122, 124, 132, 136, esp. p. 122(“Ibn Khan al-Turkumanı and the Oghuz Turks [al-Ghuzz] who were with him, numbering approx-imately 500 ghulam”); and Kamal al-Dın b. al-¨Adım, Zubdat al-H� alab min tarıkh H� alab, ed. Su-hayl Zakkar, 2 vols. (Damascus, 1997), 1:250–54, 256, 259, 269, 289, esp. pp. 253–54 (“Ibn Khanal-Turkı and the Turcomans who have joined him,” “al-Malik Harun b. Khan, the lord of the Turks”).The latter speaks of a “thousand bows,” i.e., “warriors,” but considering the general elusiveness ofmedieval authors regarding figures of military forces, Ghars al-Ni¨ma seems to provide a more reli-able lectio difficilior.

33 On this person see Cahen, “Première pénétration,” pp. 27–28; Brand, “Turkish Element,” p. 2;and Turan, Türkiye, pp. 19–20. For his relationship with the sultan see Sibt� b. al-Jawzı, ed. Sevim,p. 144.

34 Attaleiates, ed. Pérez Martín, pp. 106–7; Nikephoros Bryennios, Histoire, ed. and trans. PaulGautier, Corpus Fontium Historiae Byzantinae 9 (Brussels, 1975), 1.11, pp. 101–3; Sibt� b. al-Jawzı,ed. Sevim, pp. 146–47.

35 Attaleiates, ed. Pérez Martín, pp. 106–7.36 Attaleiates, ed. Pérez Martín, p. 191, lines 9–13: τινες τ&ων ε�πατριδ&ων τ&ης Περσ�δ�ς �δελ��3

κατ. σ�ρκα κα3 ��σιν Bπ�ρ��ντες κα3 τ�ν τ�&υ Κ�υτλ�υµ��ση 1πωνυµ�αν 1κ πατρ+#ας πρ�σηγ�ρ�ας1�!λκ�ντες, τ&ης σ�υλτανικ&ης δ2 µερ�δ�ς κα3 1��υσ�ας �ντιπ�ι��µεν�ι κα3 τ&+ω 1θν�ρ�+η τ&ων �Cννων�ντιταττ µεν�ι κα3 κατ. τ�&υτ� τ&+ η /Ρωµα�ων πρ�σ��ιτ�σαντες γ&+ η.

608 Defection


them fervent supporters of his majesty.”37 Here Attaleiates uses the Turkish de-fectors’ submission as a means of reinforcing the future emperor’s claims to le-gitimacy. Consequently, in contrast to Arısghı’s case, where the narrative dwellson the classical motif of the barbarians’ inferiority, the predominant feature inthe portrayal of the Kutlumus brothers is their outstanding nobility, which causesthem to deny recognition to their own sovereign but, nonetheless, does not pre-vent them from expressing their willingness to become subjects of the Byzantineemperor. Thus we see how incidents of defection, through the pen of one andthe same historian, could be interpreted in totally different ways according tothe conceptual matrix in which the event is embedded. As for the factual frame-work, we can safely assume that the submission of the Kutlumus brothers wasbased on a balanced agreement that in return for their ritual humiliation guar-anteed them important advantages and favors.

Alexios I’s rise to power marks not only a crucial turning point on the politi-cal scene of the empire but also the starting point of Anna Komnene’s indepen-dent narrative of her father’s reign, which for most events of this period is ouronly source.38 During the series of upheavals between Nikephoros III’s march toConstantinople and Alexios’s coronation in early April 1081, the Turkish war-riors of Sulayman created a permanent stronghold in Nicaea of Bithynia, whencethey undertook raids on the coastal areas. In most parts of Asia Minor the po-litical situation was very unstable at that time. Central control had largely col-lapsed, and Byzantine rebels, local governors, and short-lived Turkish lordshipswere alternating quickly.39 Facing the threat of the imminent Norman attack, inabout June 1081 Alexios concluded a peace treaty with Emir Sulayman, whichfor the first time in Byzantine-Seljuk relations included the recognition of a bor-der within the empire’s territory and, thus, the acceptance of a status quo of Sel-juk political authority in the region beyond the Drakon River.40 Turkish auxil-

37 Attaleiates, ed. Pérez Martín, p. 191, lines 14–19: Κα3 Dπερ Eν ��δεν3 �ασιλε&ι τ&ων Περσ&ων,γ!ν�υς ?ντες �ασιλικ�&υ, �Cτε /Ρωµα�ων κατεδ!�αντ� πρ&α�αι, τ�&υτ� πρFς α�τFν παραδ �ως�π!δει�αν9 γ νυ τε κλ�ναντες. . . . �Gς δ� κα3 λ γ�ις κα3 τρ π�ις ε�συν!τ�ις κ�σµιωτ!ρ�υς κα3ε�ν�ϊκωτ!ρ�υς �περγασ�µεν�ς, �7τως ε6ς τ�ν >αυτ�&υ δ��λωσιν κα3 π�στιν διεθ!ρµανε κα3 τ&ης α�τ�&υ�ασιλε�ας 1ραστ.ς �πειργ�σατ�.

38 For recent studies on Anna Komnene see the collective volume Anna Komnene and her Times,ed. Thalia Gouma-Peterson, Garland Reference Library of the Humanities 2201, Garland MedievalCasebooks 29 (London, 2000).

39 For the political situation see Cahen, Pre-Ottoman Turkey, pp. 74–78; idem, Formation, pp. 8–10; Vryonis, Decline, pp. 102–4, 132–33, 140–41, 144–49; Turan, Türkiye, pp. 53–66; Jean-ClaudeCheynet, “La résistance aux Turcs en Asie Mineure entre Mantzikert et la première croisade,” inΕ�ψυ��α: Μélanges offerts à Hélène Ahrweiler, Byzantina Sorbonensia 16 (Paris, 1998), pp. 129–47;George A. Lebeniotis, Η π�λιτικ� κατ�ρρευση τ�υ Βυ'αντ��υ στην Ανατ�λ�: Τ� ανατ�λικ σ�ν�ρ�και η κεντρικ� Μικρ� Ασ�α κατ� τ� �& �µισυ τ�υ 11�υ αι#να [The political collapse of Byzantiumin the East: The Eastern frontier and central Asia Minor in the second half of the eleventh century],2 vols., Byzantine Texts and Studies 43 A–B (Thessalonica, 2007), passim.

40 Anna Komnene, Alexias, ed. Diether Roderich Reinsch and Athanasios Kambylis, Corpus Fon-tium Historiae Byzantinae 40/1–2 (Berlin, 2001), 3.11.4–5, p. 116; trans. Diether Roderich Reinsch,Anna Komnene, Alexias, 2nd ed. (Berlin, 2001), p. 137: τ. περ3 ε6ρ�νης �ν!πεισε τFν σ�υλτ�ν�ν µ�λαθερµ&ως 1�αιτε&ισθαι . . . Jσµεν�ς τFν περ3 ε6ρ�νης δ!�εται λ γ�ν . . . Kµα κα3 δ#ρ�ις δε�ιωσ�µεν�ς,1�ε�ι�σατ� ε6ς ε6ρηνικ.ς �π�νε&υσαι σπ�νδ�ς. κα3 Dρ�ν α�τ�&ις καλ��µεν�ν ∆ρ�κ�ντα π�ταµFνδεδωκ@ς µ� Bπερ�α�νειν Dλως α�τ�&υ µ�τ! π�τε πρFς τ. Dρια Βιθυν&ων 1��ρµ&αν :πεισεν. See further

609Defection


iary forces continued to fight for the Byzantines in the wars against the Normans,but through the treaty of 1081 the former invaders had become much more thanthe emperor’s faithful douloi, who received generous rewards for their services;they had gained a new legal status as officially recognized holders of an auton-omous lordship.

The image of the Anatolian Seljuk Turks as given in Anna Komnene’s Alexiasin several aspects differs significantly from that of previous sources. The perspec-tive of the early 1140s, when this text was written, entails not only a certain num-ber of inaccuracies as far as persons and events are concerned but also differentviews and judgments about forms of political behavior and historical develop-ments. The arrival of armed pilgrims waging a religiously motivated war againstthe infidels from 1096 onwards and the establishment of a Seljuk sultanate inKonya marked by a distinct Muslim character must have had a strong impact onAnna’s Gedankenwelt and in particular on her perception of Anatolian reali-ties.41 Thus, she talks about a soultanikion, that is, “a sultan’s residence,” inNicaea,42 which most probably merely reflects her knowledge about the sultan-ate of Konya. Likewise, Alexios’s efforts to draw Turkish rulers and dignitariesinto his sphere of influence are ascribed to the emperor’s religious zeal, and bap-tism appears as one of the main goals of imperial policy toward the barbarians.43

This new attitude is clearly illustrated by Anna’s depiction of two cases of de-fection during Alexios’s reign, namely, that of the Turkish emissary Siaous, whowas sent to Constantinople by Sultan Malik Shah in the period of his takeoverin Syria in 1086, and that of the local ruler Elchanes, who in about 1092 ruledthe coastal area around Apollonias and Kyzikos.44 Anna’s portrayal of Alexios’s

Turan, Türkiye, p. 61; Franz Dölger and Peter Wirth, Regesten der Kaiserurkunden des oströmischenReiches, 2: Regesten von 1025–1204, rev. 2nd ed. (Munich, 1995), no. 1070a.

41 For Anna Komnene and the First Crusade see Ralph-Johannes Lilie, “Der Erste Kreuzzug in derDarstellung Anna Komnenes,” in Varia, 2: Beiträge von Albrecht Berger, Lucy-Anne Hunt, Ralph-Johannes Lilie, Claudia Ludwig und Paul Speck, Π�ικ�λα Βυ'αντιν� 6 (Bonn, 1987), pp. 49–148;idem, “Anna Komnene und die Lateiner,” Byzantinoslavica 54 (1993), 169–82; and idem, Byzan-tium and the Crusader States, 1096–1204 (Oxford, 1993), pp. 1–95. For a summary of his views seenow idem, Byzanz und die Kreuzzüge, Kohlhammer Urban Taschenbücher 595 (Stuttgart, 2004),pp. 36–63. The most important contributions of English scholars are Jonathan Shepard, “When GreekMeets Greek: Alexios Comnenus and Bohemund in 1097–98,” Byzantine and Modern Greek Studies12 (1988), 185–277; and R. D. Thomas, “Anna Comnena’s Account of the First Crusade: Historyand Politics in the Reigns of Alexios I and Manuel I Comnenus,” Byzantine and Modern Greek Stud-ies 15 (1991), 269–312.

42 Anna Komnene 3.11.1, ed. Reinsch and Kambylis, p. 114, trans. Reinsch, pp. 135–36: τ�&υ Σ�λυµ&α. . . περ3 τ�ν Ν�καιαν α�λι'�µ!ν�υ (�&B κα3 τF σ�υλταν�κι�ν &;ν, Dπερ Eν =µε&ις �ασ�λει�ν Mν�µ�-σαιµεν. For a discussion of the problem see Turan, Türkiye, pp. 62–66.

43 For a detailed analysis see Alexander Beihammer, “Die Ethnogenese der seldschukischen Türkenim Urteil christlicher Geschichtsschreiber des 11. und 12. Jahrhunderts,” Byzantinische Zeitschrift102 (2009), 1–26.

44 On both men see Cahen, Pre-Ottoman Turkey, p. 81; idem, Formation, p. 10; Brand, “TurkishElement,” p. 4; Balivet, “Entre Byzance et Konya” (above, n. 10), pp. 51–52; and Necipoglu, “Turksand Byzantines,” p. 256. The Greek form <Ελ��νης is usually explained as derivative of the Turkishtitle Ilkhan; see Gyula Moravcsik, Byzantinoturcica, 2: Sprachreste der Türkvölker in den byzanti-nischen Quellen, 3rd ed. (Berlin, 1983), p. 124. In Islamic and other Eastern sources, however, the

610 Defection


encounter with Siaous can be viewed as a counterpart of Togrıl Beg’s meetingwith Liparites described by Attaleiates. Here it is the Byzantine emperor who wasimpressed by his guest’s qualities, namely, his prudence and his mixed origin froma Georgian mother and a Turkish father. As a result, Alexios made every effortto persuade him to become a Christian and be baptized. The emperor, of course,was successful. On his behalf Siaous removed Turkish commanders from Sinopeand other Anatolian cities and was rewarded with rich gifts and a post as dux ofAnchialos.45 Likewise, Elchanes’s defection to the emperor was rewarded not onlywith rich gifts but also with “holy enlightenment.” Many of his relatives and com-panions, attracted by the emperor’s generosity, followed his example.46 Alexiostook recourse in exactly the same practices his predecessors had employed so suc-cessfully with Turkish refugees over the previous decades. The only difference liesin the fact that the tightening of diplomatic contacts with Malik Shah’s court andthe increasing number of Turkish local lords in the western parts of Asia Minorenabled him to apply these methods more systematically. What actually changedis the retrospective interpretation of this strategy by Anna Komnene, who, mostprobably under the influence of experiences with the crusades, gave her father’spolicy a religious dimension by presenting him as a preacher of the true faith.47

From the first Byzantine-Seljuk exchange of emissaries in 104948 until AlexiosI’s rise to power, the only representative of the Seljuk Turks whom the Byzantinegovernment recognized as a holder of political authority was the dynasty’s su-preme head and sultan in Persia and, from 1055 onwards, in Baghdad.49 AlexiosI’s treaty of June 1081 with Emir Sulayman marked the appearance of a newindependent ruler representing the Seljuk dynasty’s Anatolian branch. As a re-sult, cases in which these newly recognized sovereigns sought the emperor’s pro-tection demanded a redefinition of their relationship with the imperial govern-ment according to the customs and rules of Byzantine diplomacy. Consequently,

title does not appear before being adopted by the Mongol khan Hülegü in the 1250s (see, for in-stance, David Morgan, The Mongols, 2nd ed. [Malden, Mass., 2007], pp. 130–39); the issue seemsto need further discussion.

45 Anna Komnene 6.9.4–6, ed. Reinsch and Kambylis, pp. 187–88, trans. Reinsch, pp. 215–16:π�λλ�ν πραγµατε�αν 1πεπ��ητ� Nστε τ�&υ θε��υ �απτ�σµατ�ς τ�&υτ�ν τυ�ε&ιν . . . κ�ντε&υθεν τ�&υ θε��υ�απτ�σµατ�ς τετυ�ηκ@ς κα3 π�λλ&ων δωρε&ων 1παπ�λα�σας δ�O� <Αγ�ι�λ�υ πρ��ειρ�'εται.

46 Anna Komnene 6.13.4, ed. Reinsch and Kambylis, pp. 198–99, trans. Reinsch, p. 229: α�τFς δ2µετ. τ&ων καθ< α&8µα πρ�σηκ ντων α�τ�µ�λε&ι πρFς τFν �ασιλ!α κα3 µυρ�ων µ2ν 1παπ�λα�ει δωρε&ων,τυγ��νει δ2 κα3 τ�&υ µεγ�στ�υ, τ�&υ Pγ��υ �ηµ3 �ωτ�σµατ�ς. In particular, Anna mentions a certainSkaliarios (the Persian title of salar, i.e., “commander, chief,” is clearly recognizable in the Greekform) and another person being granted the title of hyperperilampros.

47 Anna Komnene 6.13.4, ed. Reinsch and Kambylis, p. 199, trans. Reinsch, p. 229.48 Dölger and Wirth, Regesten, no. 890d (mission of George Drosos to Sultan Togrıl Beg in about

early 1049). In the same year the sultan replied by dispatching an emissary bearing the title of sharıfto Constantinople; see Claude Cahen, “La diplomatie orientale de Byzance face à la poussée seld-jukide,” Byzantion 35 (1965), 10–15, repr. in idem, Turcobyzantina, no. 3.

49 For Togrıl Beg’s rise to the sultanate see Clifford Edmund Bosworth, “The Political and Dynas-tic History of the Iranian World (A.D. 1000–1217),” in The Cambridge History of Iran, 5: The Sal-juq and Mongol Periods, ed. J. A. Boyle (Cambridge, Eng., 1968), pp. 1–202, at pp. 42–48; and A.C. S. Peacock, Early Seljuq History: A New Interpretation, Routledge Studies in the History of Iranand Turkey 7 (London, 2010).

611Defection


the Seljuk rulers of Asia Minor obtained a higher status at the Byzantine courtand came into closer contact with the imperial sphere. During Alexios I’s reignthe palace of Constantinople witnessed the arrival of two famous potentates, Abul-Qasim of Nicaea (1084/86–1092), lieutenant and successor of Sulayman; andSultan Shahanshah of Konya (1107–16), second son of Sultan Kılıc Arslan I(1093–1107).50 Before their flight to the emperor, both had suffered severe mil-itary setbacks and were threatened by dangerous internal opponents: the first bythe Seljuk commander Bursuq, who on behalf of Sultan Malik Shah besiegedNicaea over several months, the second by his brother Mas¨ud, who was prepar-ing a plot against Shahanshah. In each case the conclusion of a peace treaty withthe emperor preceded the trip to Constantinople.51

The main difficulty for an assessment of these contacts and their consequencesis that Anna highlights different aspects of each event. Abu l-Qasim’s receptionin Constantinople is depicted as a successful ruse of the emperor, who was try-ing to divert the emir’s attention from his efforts to construct a stronghold in thevicinity of Nicomedia. The model she refers to is a maneuver used by the Athe-nian hero Alcibiades against the Spartans.52 The conceptual framework consistsof the polarity between the barbarian’s mental inferiority and the Byzantines’ skill-fulness.53 Anna mainly speaks of the splendid gifts and all sorts of amusementsthe imperial court offered, such as bathing, hunting, and horse racing.54 The lux-uries and comforts of Constantinopolitan court life had a seductive impact onthe barbarian’s mind and made him forget his caution in military affairs. In ad-dition, the author interlaces the motif of the barbarian’s greed for Byzantine lux-uries with allusions to the political situation, which was conditioned by Bursuq’sthreat and the emperor’s support for Abu l-Qasim’s cause.55 These factors, mostlikely, were the true reasons for the emir’s visit to Constantinople. Yet Anna stillkept a crucial detail of Byzantine court ceremonies in her report, namely, the factthat Abu l-Qasim was granted the title of sebastos.56 Given that this title had

50 On these potentates see Cahen, Pre-Ottoman Turkey, pp. 77, 78–80, 88–89, 91–93; idem, For-mation, pp. 9–10, 13, 16–18; Turan, Türkiye, pp. 83–87, 153–58; Jonathan Shepard, “‘Father’ or‘Scorpion’? Style and Substance in Alexios’s Diplomacy,” in Alexios I Komnenos: Papers of the Sec-ond Belfast Byzantine International Colloquium, 14–16 April 1989, ed. Margaret Mullett and DionSmythe (Belfast, 1996), pp. 68–132, at pp. 78–79, 83; and Necipoglu, “Turks and Byzantines” (above,n. 11), p. 256.

51 Turan, Türkiye, pp. 85, 158; Cahen, Pre-Ottoman Turkey, pp. 80, 92; idem, Formation, pp. 10,18; Dölger and Wirth, Regesten, no. 1163 (letter of Alexios I to Abu l-Qasim concerning the treaty),no. 1269 (erroneously mentions Sultan Malik Shah as the emperor’s treaty partner).

52 Anna Komnene 6.10.8–11, ed. Reinsch and Kambylis, pp. 191–92, trans. Reinsch, pp. 219–20.For the comparison with Alcibiades see ibid., p. 192, lines 72–80, trans. Reinsch, p. 221.

53 For this concept and its antique prototype see Yves Albert Dauge, Le barbare: Recherches sur laconception romaine de la barbarie et de la civilisation (Brussels, 1981).

54 Anna Komnene 6.10.9–10, ed. Reinsch and Kambylis, pp. 191–92, trans. Reinsch, p. 220: πε�θεταιQ <Απελ�ασ�µ, κα3 ε6σελθ@ν ε6ς τ�ν �ασιλε��υσαν παντ��ας ��ι�&υται �ιλ��ρ�σ�νης . . . τ&+ω δ! γε<Απελ�ασ�µ καθεκ�στην �ρ�µατα διδ�Oς ��κ 1νεδ�δ�υ ε6ς �αλανε&ια τ2 πρ�τρεπ µεν�ς κα38ππηλασ�ας κα3 κυνηγ!σια.

55 Anna Komnene 6.10.8, 6.11.1–2, ed. Reinsch and Kambylis, pp. 191, 192–93, trans. Reinsch,pp. 219–20, 221–22.

56 Anna Komnene 6.10.10, ed. Reinsch and Kambylis, p. 192, trans. Reinsch, p. 221: πλε��σι δωρεα&ιςτ�&υτ�ν �µειψ�µεν�ς τ&+ω τ&ων σε�αστ&ων τ2 ��ι#µατι τιµ�σας κα3 1π3 πλ!�ν τ.ς συνθ�κας 1µπεδ#σας.

612 Defection


been created by Alexios I for close relatives of the emperor,57 its bestowal uponthe Seljuk ruler of Nicaea signaled his symbolic integration into the imperial fam-ily. Considering that until the middle of the eleventh century even the most im-portant Muslim allies of the empire hardly reached titles higher than magistros,with the exception only of two proedroi, Emir Thimal of Aleppo and the afore-mentioned Arısghı (Chrysoskoulos) in 1051/52 and 1070 respectively,58 the im-pressive upgrading of Seljuk lords within the system of the Byzantine court hier-archy becomes all the more significant.

Concerning Sultan Shahanshah, Anna gives a detailed description of his firstencounter with Emperor Alexios on a plain near Akroinon in 1116,59 but shecompletely omits the sultan’s subsequent sojourn in Constantinople, which wecan derive only from other sources.60 The focus lies on the peace negotiationsfollowing the campaign of 1116, and, therefore, rituals and gestures play a cen-tral role in the description. What matters, above all, is the public spectacle pro-jecting the relationship between the two rulers in the framework of a carefullychoreographed sequence of ritual acts performed in the presence of a large crowdof court officials, military commanders, and common soldiers. While the mem-bers of Shahanshah’s entourage prostrated themselves before the emperor in thecustomary way, several times Alexios prevented the sultan himself from gettingoff his horse in order to follow his companions’ example. Eventually, the sultandemonstrated his deference by kissing the emperor’s foot. In return, Alexios tookhim by the hand and ordered him to mount one of his dignitaries’ horses. Whilethe sultan and the emperor were riding side by side, Alexios took his cape andthrew it on the sultan’s shoulders.61

What at first sight might appear as a spontaneous performance conveys mes-sages that are too complicated to be the result of mere coincidence. The wholescene obviously aimed at propagating a new hierarchical relationship betweenthe Byzantine emperor and the Seljuk sultan, which was to be confirmed for-mally by the subsequent conclusion of a peace treaty. The emperor rewarded thesultan’s gesture of humiliation with symbolic acts of graciousness, displaying signs

57 For the significance of the title, which in Alexios’s early years was limited to a rather small groupof dignitaries related to the emperor by blood or marriage, see Lucien Stiernon, “Notes de titulaireet de prosopographie byzantine: Sébaste et gambros,” Revue des études byzantines 23 (1965), 222–43; Magdalino, Manuel Komnenos, pp. 181–83; Stauros G. Georgiou, “�ι τιµητικ�� τ�τλ�ι επ�Κ�µνην#ν (1081–1185)” (Ph.D. thesis, Thessalonica, 2005), pp. 110–38 (until the end of the twelfthcentury only about 10 percent of the known sebastoi were foreigners).

58 For court titles bestowed upon Muslim rulers see Franz Dölger and Andreas E. Müller, with Al-exander Beihammer, Regesten der Kaiserurkunden des oströmischen Reiches, 1/2: Regesten von 867–1025 (Munich, 2003), nos. 758b, 790c, 793a, 801b, 801c; and Felix, Byzanz und die islamische Welt,pp. 113, 117, 121.

59 Anna Komnene 15.6.5–6, ed. Reinsch and Kambylis, pp. 477–78, trans. Reinsch, pp. 530–31.60 Michael the Syrian, Chronique, ed. and trans. Jean-Baptiste Chabot, 4 vols. (Paris, 1899–1910),

15.10, 3:194–95 (trans.), 4:593, col. B, lines 10–18 (Syriac text).61 Anna Komnene 15.6.5, ed. Reinsch and Kambylis, p. 478, trans. Reinsch, p. 531: τFν δ2

σ�υλτ�ν�ν π�λλ�κις τ�&υ Rππ�υ �π��&ηναι 1πι�ειρ�σαντα Q α�τ�κρ�τωρ �� υνε�#ρει. �λλ< 1κε&ιν�ςτα�O πε'ε�σας τFν π δα τ��τ�υ ;σπ�σατ�. κα3 Sς �ε&ιρα τ2 δ�Oς α�τ&+ω κα3 Rππ�ν τ&ων 1κκρ�των1πι�&ηναι 1κ!λευσεν. 1πι��ντα δ2 τ�&υτ�ν κα3 περ3 θατ!ραν πλευρ.ν τ�&υ α�τ�κρ�τ�ρ�ςπρ�σεγγ�σαντα, παρα�ρ&ηµα τF Jµ�ι�ν S περιε�!�λητ� λ�σας, τ�&ις Tµ�ις 1κε�ν�υ 1π!θετ�.

613Defection


of proximity and intimacy to the sultan. The Seljuk lord, who acknowledged theemperor’s superiority by kissing Alexios’s foot, has become a fully recognized sov-ereign and friend of the emperor.

In summary, a clear line of development can be observed from the first Turk-ish apostates, who from the 1050s onwards were accepted in Constantinople asthe emperor’s allies for military purposes, to Turkish local rulers and dignitaries,who in return for their voluntary surrender were granted incomes, posts, and sen-atorial titles, to Seljuk chiefs, who from the treaty of 1081 onwards were de factorecognized as independent rulers and on the occasion of meetings with the em-peror were granted titles reserved for the dynasty’s inner circle and symbols offriendship. Moreover, Anna Komnene ex eventu interpreted her father’s atti-tudes toward Turkish apostates in terms of religious zeal aiming at the barbari-ans’ conversion and baptism.

Christian Frontier Lords Turning to Islam (ca. 1050–1118)

As regards the first known incidents of Christian defectors who took refugewith the Turks in the early period of Byzantine-Seljuk contacts, we do not knowof any member of the empire’s court elite who took this action. As will be shownbelow, this phenomenon appears for the first time during the reign of John II(1118–43).

The well-known episode of Emperor Romanos IV’s one-week captivity in Sul-tan Alp Arslan’s camp following the battle of Mantzikert (August 1071) resultedin a face-to-face meeting of the supreme heads of both empires.62 Later works ofMuslim historiography celebrate this event as a major triumph of Islam in gen-eral and the Seljuk sultan in particular over the most powerful representative ofthe Christian world, portraying the latter as a humiliated subject full of grati-tude for the sultan’s generosity and leniency. Byzantine sources agree with theMuslim authors on the sultan’s magnanimous and respectful behavior toward theemperor. Both traditions, along with the peace treaty, refer to a series of ritualacts projecting the new relationship between the rulers, such as gestures of com-passion, common meals, and the bestowal of honorary clothes.63 On his way back

62 On the battle of Mantzikert and Romanos IV’s captivity see Claude Cahen, “La campagne deMantzikert d’après les sources musulmanes,” Byzantion 9 (1934), 628–42, repr. in idem, Turcoby-zantina, no. 2; Jean-Claude Cheynet, “Mantzikert: Un désastre militaire?” Byzantion 50 (1980), 410–38; Speros Vryonis, “The Greek and Arabic Sources for the Battle of Mantzikert, 1071 A.D.,” inByzantine Studies: Essays on the Slavic World and the Eleventh Century, ed. Speros Vryonis (NewRochelle, N.Y., 1992), pp. 125–40; idem, “A Personal History of the History of the Battle of Mantzi-kert,” in Η �υ'αντιν� Μικρ� Ασ�α (above, n. 5), pp. 225–44; idem, “The Greek and Arabic Sourceson the Eight Day Captivity of the Emperor Romanos IV in the Camp of the Sultan Alp Arslan afterthe Battle of Mantzikert,” in Novum millennium: Studies on Byzantine History and Culture Dedi-cated to Paul Speck, ed. Claudia Sode and Sarolta Takács (Aldershot, Eng., 2001), pp. 439–50.

63 The main source, on which almost all later Byzantine accounts are based, is Attaleiates, ed. PérezMartín, pp. 121–23, for example, p. 122, lines 4–5: �νθρωπ�νως κα3 ν�υνε�&ως τF πρ�τ!ρηµα τ&ηςν�κης �8 Τ�&υρκ�ι 1δ!�αντ�, µ�τε µεγαλαυ��σαντες, and p. 122, lines 18–22: Qµ δ��ν κατ. τ�ντιµ�ν π�ιησ�µεν�ς . . . µηδ2 µ!�ρι κα3 �ρα�υτ�τ�υ λ γ�υ πρFς τ�&υτ�ν 1µπεπαρ +ωνηκ@ς. The mostimportant Syrian and Armenian sources are Michael the Syrian 15.3, ed. Chabot, 3:169–70 (trans.),

614 Defection


home Romanos IV wore “a Turkish dress,” and Muslim sources maintain thathe was accompanied by Turkish guards holding Islamic flags.64 Henceforth thisevent served as a precedent for Muslim attitudes toward defeated Christian op-ponents and as a model case for idealized presentations of contact situations be-tween victorious Seljuk lords and Byzantine men of rank.

Leading personalities of the imperial entourage in the 1070s, such as Alexios’sbrothers Manuel and Isaac Komnenos and Caesar John Doukas, were taken cap-tive during their expeditions against the Turks, but none of them seems to havebeen tempted to switch sides in that period.65 Things were different among thelocal aristocracy in the eastern border region, mainly Armenian noblemen, whoin the general disorder of the 1060s and 1070s were acting as independent com-manders or local lords. Unfortunately, there is no extant Muslim source of thatperiod revealing early Turkish attitudes toward Christian defectors and their mech-anisms of integrating them into their ranks. Our knowledge, therefore, is basedexclusively on Byzantine or Oriental Christian (Syrian or Armenian) authors, whoprimarily treat these people as traitors against their compatriots and faith. WhileTurkish refugees could be presented as symbols of the empire’s moral strength,Christian defectors would prove the opposite and, therefore, had to be broughtinto disrepute. Moreover, historians relating the fate of these men were forced tofind sound explanations for their deviation from commonly accepted norms ofbehavior. Generally speaking, the immediate threat to life, rank, and propertywas considered a reasonable justification for attempts to come to terms with theenemy, but it did not excuse defection or conversion to Islam.

A case in point is a diplomatic maneuver of the Bagratid king of Kars, Gagik-Abas II, who shortly before his abdication in 1064 received an envoy of Sultan

4:578–79 (Syriac text), for example, p. 578, col. B, lines 7–9: “malka d� en d-Turkaye estawdı d� -enzk� a l-Romaye wa-lb�ak� l-Dıyugın rah�ma ¨b�ad ¨law(hy) w-b�a-slama mhappek� leh l-at�ra” (The king ofthe Turks promised that, if he defeated the Romans and took Diogenes captive, he would show himclemency and send him back home peacefully); Aristakes of Lastivert, Récit des malheurs de la na-tion arménienne, trans. M. Canard and H. Berbérian, Bibliothèque de Byzantion 5 (Brussels, 1973),25, pp. 127–28: “Il inspira au féroce roi de Perse l’amour et la sollicitude que l’on éprouve pour unfrère bien-aimé”; and Matthew of Edessa, Chronicle, trans. Ara Edmond Dostourian, in Armeniaand the Crusades, Tenth to Twelfth Centuries: The Chronicle of Matthew of Edessa (Lanham, Md.,1993), 2.57, p. 135. For the Islamic sources see the text collection with Turkish translations by FarukSümer and Ali Sevim, Islâm kaynaklarına göre Malazgirt savası (metinler ve çevirileri), Türk TarihKurumu Yayınları, 19. Dizi–sa. 3a, 2nd ed. (Ankara, 1988). For a new English translation of thesesources with extensive commentary see Carole Hillenbrand, Turkish Myth and Muslim Symbol: TheBattle of Manzikert (Edinburgh, 2007). For the diverging opinions on the interpretation of this sourcematerial see Cahen, “Campagne de Mantzikert”; Vryonis, “Greek and Arabic Sources”; and Hillen-brand, Turkish Myth, pp. 3–25. Speros Vryonis is currently finishing his new monograph on Manzi-kert, containing the results of his almost lifelong research on the topic.

64 Attaleiates, ed. Pérez Martín, pp. 106 and 123, line 21: µετ. τ�υρκικ&ης τ&ης στ�λ&ης. Sibt� b. al-Jawzı, ed. Sevim, p. 151, mentions a qaba© (an Islamic dress) and a qalansuwa (a hat worn by Mus-lim dignitaries).

65 Attaleiates, ed. Pérez Martín, pp. 104–5, 106 (Manuel Komnenos is taken captive in the battleof Sebasteia, 1070), 136; Bryennios 2.5, ed. Gautier, pp. 149–51 (Isaac Komnenos is taken captive ina battle near Kaisareia); Attaleiates, ed. Pérez Martín, pp. 140–41; Bryennios 2.18, ed. Gautier,pp. 179–81 (John Doukas and Rouselios are taken captive).

615Defection


Alp Arslan, to whom he demonstrated his grief for Sultan Togrıl Beg’s death bybeing “dressed up in a black garment of mourning.”66 Deeply impressed by thisgesture, the sultan “offered Gagik his friendship,” which was confirmed by anexchange of gifts and a rich banquet. Matthew of Edessa presents this ritual act,by which the king expressed emotional commitment and deference to the Seljukdynasty, as a clever ruse creating the basis for a peaceful relationship with thesultan. In this way the author praises King Gagik’s sagacious handling of the Turk-ish threat but by no means criticizes him for his appeasing attitude.

Offended honor was another motive that in the eyes of contemporaries mightoffer a reasonable explanation for collaboration with the enemy. The Armeniannobleman George Shirakats©i of Ani, for example, was severely maltreated andinsulted by the people of Antioch. As a result he allied himself with a group offive hundred Turkish warriors, ravaged the territory of Antioch, and slaughteredhis captives in front of the city’s gate. Matthew of Edessa, instead of condemningthis behavior, rebuked the townspeople for doing injustice to the nobleman.67

Military defeat, under certain circumstances, might result in forced conversionto Islam, as we learn from an incident recorded by Bar Hebraeus. In a.h. 460(November 11, 1067–October 30, 1068), while on patrol, the Armenian patri-cian Aristakis and his two hundred soldiers were attacked by the Seljuk emir Shir-wan Shah.68 In an attempt to save their lives, the Armenians declared that theywere on the way to the sultan in order to become Muslims. Thus, against theirwill they were received with honor at the sultan’s court and circumcised, and thepatrician was awarded an annual income of twenty thousand dinars. Eventuallythey managed to escape and return to their Christian faith. Defection and con-version are presented here as acts of despair, which subsequently were corrected.Even high honors and rich grants could not keep them from returning to theirpaternal faith, so that their initial failure is excused.

The most renowned defector of this period, no doubt, was the Armenian lordPhilaretos Brachamios, who had been appointed domestikos by Romanos IV.69

Following the overthrow of his former lord, he rose up against the regime of Mi-chael VII and created a short-lived lordship in Cilicia and northern Syria com-prising major urban centers such as Mar¨ash, Edessa, Melitene, and, in 1078, evenAntioch. All sources are very ill disposed toward him, and they agree on the factthat, because of the increasing pressure exerted by the Seljuk leaders in the Syr-ian border region, in about 1085 Philaretos decided to submit to Sultan MalikShah and convert to Islam, something that eventually caused the collapse of hislordship.

66 Matthew of Edessa 2.23, trans. Dostourian, p. 104.67 Matthew of Edessa 2.79, trans. Dostourian, p. 148.68 Barhebraeus, Chronicon Syriacum, ed. Paul Bedjan (Paris, 1890), p. 244; trans. Ernest Alfred

Wallis Budge, The Chronography of Gregory Abû©l-Faraj, 1225–1286, the Son of Aaron, the He-brew Physician Commonly Known as Bar Hebraeus (London, 1932), p. 218.

69 For Philaretos see C. J. Yarnley, “Philaretos: Armenian Bandit or Byzantine General?” Revuedes études arméniennes 9 (1972), 331–53; and Gérard Dédéyan, Les arméniens entre grecs, musul-mans et croisés: Étude sur les pouvoirs arméniens dans le Proche-Orient méditerranéen (1068–1150), 2 vols. (Lisbon, 2003), 1:287–357; for his title see Anna Komnene 6.9.2, ed. Reinsch andKambylis, p. 187, trans. Reinsch, p. 214.

616 Defection


Matthew of Edessa connects his fate with the deplorable state of affairs in hishometown during the period of Philaretos’s rule, thus calling him an “impiousand most wicked chief,” “the very offspring of Satan,” and a “perfidious man,who indeed was a precursor of the abominable Antichrist.”70 According to Mat-thew’s version, Philaretos first sought Sultan Malik Shah’s benevolence; but whenthe sultan became angry because of Edessa’s decay, Philaretos fell into completedespair: “At that moment [he] abjured his Christian religion, renouncing the faithof Christ . . . for by so doing, he thought he would be honored by the Persians,but this was not the case . . . and he came to be cursed and despised by bothGod and men.”71 Hence, conversion to Islam appears as a self-destructive act ofsurrender to treacherous forces that eventually leads to material and moral down-fall. Michael the Syrian arrives at a similar conclusion, stating that Philaretos,having returned from his trip to Baghdad and Khurasan, found all his territoriesin Turkish hands: “When he gave up his faith for his lordship, he was deprivedof his lordship as well.”72 Anna Komnene establishes a direct causal link be-tween Philaretos’s defection and the ruin of Antioch. Philaretos’s son, she tellsus, failed to dissuade his father from his decision to become a Muslim. Hence hefled to Emir Sulayman of Nicaea to offer him rule over Antioch.73 The sequenceof events as presented by Anna is highly illogical, but the conceptual frameworkof the story is much the same as in the versions of Matthew and Michael.74 De-fection and conversion are patterns of behavior that characterize people who havelost self-control and are led into a state of despair. Their decision, unavoidably,has destructive results for both themselves and the people depending on them.

By the beginning of the twelfth century defection had become a common ex-perience for both sides. Contemporary and later chroniclers had developed inter-pretative patterns and conceptual frameworks, on the basis of which these expe-riences were transformed into discursive events and narratives. The ideologicalbackground was provided by the ancient model of barbarism, by questions oflegitimacy and lawfulness, and by moral judgments.

Comnenian Aristocrats and Court OfficialsWandering about Anatolia (1118–1260)

In the following sections of this essay I will examine the phenomenon ofByzantine-Seljuk defection from Alexios I’s death in 1118 until ca. 1260, whichcan be considered a logical turning point inasmuch as the Mongol protectorateover the Seljuk sultanate of Rum and Michael VIII Palaiologos’s rise to power inthe Empire of Nicaea brought about a totally new political equilibrium in AsiaMinor. The available source material resembles that of the previous period in thatwe depend mainly on Byzantine historians, most essentially Niketas Choniates

70 Matthew of Edessa 2.60, trans. Dostourian, pp. 137–38.71 Matthew of Edessa 2.85, trans. Dostourian, pp. 152–53.72 Michael the Syrian 15.4, ed. Chabot, 3:173 (trans.), 4:580–81 (Syriac text): “w-k�ad� l-haymanut�a

d� -ıt�-(h)wa leh awbed� met�t�ul resanut�a w-ap men resanut�a pad� .”73 Anna Komnene 6.9.2, ed. Reinsch and Kambylis, pp. 186–87, trans. Reinsch, p. 214.74 For Philaretos’s struggle with the Seljuk Turks see Dédéyan, Les arméniens, 1:338–43, 346–49.

617Defection


and George Akropolites,75 while Oriental authors, whether Christian or Mus-lim, provide us only a few glimpses of the phenomenon of defection. Our per-spective, therefore, remains to a certain extent biased and unbalanced. Only thechronicle Al-Awamir al-¨ala©ıya fı l-umur al-¨ala©ıya, written in about 1281 byIbn Bıbı, the first historical work exclusively devoted to the Seljuk sultanate ofRum,76 gives us more substantial information in this respect.

Like the sources discussed above, whenever twelfth-century narratives refer toincidents of defection, they are mainly concerned with aspects of political ideol-ogy and offer moralizing interpretations. Niketas Choniates, as will be shownbelow, also introduces innovative features into the description of these events,placing special emphasis on the role of emotions as motives for action and formsof behavior. While leaving aside or intentionally ignoring matters of internal strife,power politics, and military conflicts, he highlights spontaneous irrational reac-tions and emotional conditions dominating the actions of men, like fear and envy,as the main factors causing people to disregard their natural bonds and take ref-uge with the enemy. Thus, Niketas’s approach, though not very helpful for pro-viding a look behind the scenes, reveals a new form of perception that embedsapostasy and defection into a context of mental conditions. Interestingly, this ap-proach is by no means applied uniformly; the quality of emotions and the degreeof an individual’s irrationality vary according to his social status and his prox-imity to the empire’s inner circle of power.

In addition one has to take into account the intrinsic logic and thought struc-ture on which a given work’s historiographical discourse is based. Choniates’sadmiration for Emperor John II, for instance, necessarily makes his brother Isaacappear as a disturbing dissonance within a perfect political harmony, whereasGeorge Akropolites’s hatred for Theodore II Lascaris shows Michael Palaiologosin the light of a righteous opponent of a tyrannical regime.77

On account of these parameters we might distinguish several types of Byzan-tine defectors during the period in question. (See the Appendix below for a syn-opsis of the instances of defection considered in this essay.) Undoubtedly, the mostprominent defectors were members of the ruling dynasty and close relatives ofthe emperor. In fact, this series of defections starts with members of the Comne-nian family, namely, John II’s brother, the sebastokrator Isaac, and his two sonsJohn and Andronikos. Choniates repeatedly underlines Isaac’s love for and de-votion to his brother and stresses the decisive role he played in the establishmentof John’s rule during the succession crisis of 1118.78 As a sign of honor John con-

75 George Akropolites, Historia, ed. August Heisenberg, in Georgii Acropolitae opera, rev. ed., 2vols. (Stuttgart, 1978), 1:1–189; trans. Ruth Macrides, George Akropolites: The History (Oxford,2007).

76 Ibn Bıbı, Mukhtas�ar-i Saljuqnamah, ed. Martijn Theodoor Houtsma, in Histoire des Seldjouci-des d’Asie Mineure d’après l’abrégé du Seldjouknameh d’Ibn Bıbı, Recueil de Textes Relatifs à l’Histoiredes Seldjoucides 4 (Leiden, 1902); trans. Herbert W. Duda, Die Seltschukengeschichte des Ibn Bıbı(Copenhagen, 1959).

77 For these two cases see the discussion below, pp. 619–21 and 627–29.78 For a biographical sketch see Konstantinos Varzos, Η γενεαλ�γ�α των Κ�µνην#ν [The geneal-

ogy of the Comnenians], 2 vols. (Thessalonica, 1984), 1:238–54, no. 36.

618 Defection


ceded to his brother a place of equal rank on the imperial throne and at the im-perial table.79 Nevertheless, some years later—from Michael the Syrian’s accountthe incident can be dated to the summer of 1130, during Emperor John’s cam-paign against the Danismends in Kastamon80—“Isaac broke off relations withhis brother and departed from the land of the Romans, taking with him John,his eldest son, as companion and fellow wanderer.”81

Choniates and other Byzantine authors, reflecting the imperial court’s view-point, solved the problem of giving a reasonable explanation for Isaac’s behav-ior by keeping silent about the true causes for the brothers’ discord and by min-imizing its real dimensions. These can only be discerned through reports fromoutside the empire, namely, through Michael the Syrian, who explicitly refers toa “conspiracy” of Isaac and some magnates.82 Obviously, the emperor’s brotherwas the leader of an unsuccessful coup d’état supported by elements of the Con-stantinopolitan ruling elite. Whether they were motivated by personal ambitionor by political aims, there must have been a deep-rooted rivalry between AlexiosI’s two sons. This conflict most probably took shape long before its outbreak in1130, so that Choniates’s insistence on the brothers’ equality in rank seems toreveal the deeper causes of a serious dynastic problem. When Isaac eventuallywas forced to leave the empire, he consequently chose to take refuge with Em-peror John’s most dangerous eastern enemy at that time, Gümüstekin Ghazı(1104–34), son and successor of Danismend. Thus, for the first time a memberof the Constantinopolitan elite adopted a practice that until then had been em-ployed only by Seljuk apostates fleeing in the opposite direction from Asia Mi-nor to the imperial court.83

Byzantine accounts, in their attempt to downplay the extent of the discord,restrict themselves to elusive hints. Choniates talks about a “futile grief,”84 andin one of his poems Theodore Prodromos refers to the fateful power of envy(�θ ν�ς),85 without going into further detail. Instead they focus on the reconcil-

79 Choniates, ed. van Dieten, p. 6, trans. Magoulias, pp. 6–7.80 Michael the Syrian 16.4, ed. Chabot, 3:230 (trans.), 4:612, col. B, lines 29–32 (Syriac text).81 Choniates, ed. van Dieten, p. 32, lines 33–36, trans. Magoulias, p. 19: τ�&υ Qµ�γν��υ δια'ευ�θε3ς

κα3 �υγ.ς �π�ρας 1κ τ&ης /Ρωµα�ων, συν!κδηµ�ν κα3 συµπλαν�την :�ων τFν <Ιω�ννην τFν τ&ωνπα�δων πρωτ τ�κ�ν.

82 Michael the Syrian 16.4, ed. Chabot, 3:230 (trans.), 4:612, col. B, lines 31–32 (Syriac text): “et�nk�el¨law(hy) ah� u(h)y w-nasın men rawrb� anaw(hy)” (his brother and some people of his great men con-spired against him).

83 For Isaac’s conspiracy and flight see Ferdinand Chalandon, Les Comnène: Jean II Comnène (1118–1143) et Manuel I Comnène (1143–1180) (Paris, 1912), pp. 83–85; Cahen, Pre-Ottoman Turkey,p. 94; Ralph-Johannes Lilie, “Des Kaisers Macht und Ohnmacht: Zum Zerfall der Zentralgewalt inByzanz vor dem Vierten Kreuzzug,” in Varia, 1: Beiträge von Ralph-Johannes Lilie und Paul Speck,Π�ικ�λα Βυ'αντιν� 4 (Bonn, 1984), pp. 9–120, at pp. 60–61; Varzos, Γενεαλ�γ�α, 1:239–43; Chey-net, Pouvoir, p. 105; and Necipoglu, “Turks and Byzantines” (above, n. 11), p. 258.

84 Choniates, ed. van Dieten, p. 32, line 33: κατ. γ.ρ µικρ�λυπ�αν.85 Theodoros Prodromos, Historische Gedichte, ed. Wolfram Hörandner, Wiener Byzantinistische

Studien 11 (Vienna, 1974), 41, lines 10–12: �λλ< W τ��ης κ�λινδρ�ς, W τρ��Fς ��υ, / ε5λη�εν 6σ�Oνκα3 καθ< =µ&ων Q �θ ν�ς / κα3 �ε&υ διε&ιλε τ�ν ��υ συµψυ��αν. For other allusions to the event seeibid. 2, line 26, and 40, lines 53–54 (Isaac’s return).

619Defection


iation that took place in 1138/39, perhaps shortly after John II’s stay in Anti-och.86 The rejoicing over the reestablished peace turns the previous hostility intoa temporary misfortune, the reasons for which have no further significance. Thisattitude seemingly reflects propagandistic efforts that celebrated Isaac’s return andthe reestablished dynastic order as part of the emperor’s triumphal entrance inConstantinople. Thus, the inhabitants of the Byzantine capital not only hailedthe emperor’s military exploits but also expressed their joy at Isaac’s arrival.87

Because of the sources’ elusiveness it remains unclear to what extent Isaac’sexile in Asia Minor had an impact on the political situation in the region. Yethousing a guest who was second in rank to the emperor no doubt was an eventof the highest significance for the sovereigns in the East. Michael the Syrian statesthat Gümüstekin Ghazı “was very glad of his presence and honored him mag-nificently,”88 and accordingly Choniates observes that Isaac “was treated withrespect, for he was most imperial in bearing and very noble because of his fam-ily.”89 His high standing also allowed him to come into contact with other Chris-tian and Muslim lords of Asia Minor who had a common hostility against Con-stantinople.90 Isaac visited Constantine Gabras of Trebizond, an ally of theDanismend emir, and spent the winter together with Gümüstekin and his son-in-law, Sultan Mas¨ud of Konya, in Melitene. His next stop was the court of Leo ofCilicia, who at that time was a tributary of the Danismends. An alliance basedon a marriage between Leo’s daughter and Isaac’s son failed on account of a dis-agreement between the two potentates, so Isaac went back to Sultan Mas¨ud.Choniates’s statement that “he was planning to make attacks against Roman ter-ritories and to become Satan for John”91 suggests that Isaac constantly endeav-ored to pursue his conspiratorial plans and to create a network of adversariesagainst his brother, but we do not know how active a role he was in fact able toplay within the existing political equilibrium.

From the perspective of Byzantine imperial propaganda the power and pres-tige that Emperor John II had gained on account of his military exploits posedtoo dangerous a threat for the eastern lords to offer any effective support to theirguest. On the other hand, the fact that Isaac stayed nearly a decade in Asia Mi-nor seems to reveal a certain inability on the part of the emperor to prevail uponhis opponents. Consequently, the reconciliation between the brothers coincidedchronologically with the culmination of John’s expansionist policy toward Anti-och and the Crusader States. The event was preceded by a pilgrimage by Isaac tothe Holy Sepulchre, on the occasion of which he financed the construction of an

86 See Michael Italikos’s speech to Emperor John II on the occasion of his campaign to Syria in1138: Michel Italikos, Lettres et discourse, ed. P. Gautier, Archives de l’Orient Chrétien 14 (Paris,1972), 43, p. 265, lines 11–17: Συνηδ µεθ� σ�ι, �ασιλε&υ, κα3 συνε�α�ρ�µεν κα3 1π3 τ&+ η τ&ης δραγµ&ηςεBρ!σει9 δραγµ�ν δ2 λ!γων, τ�ν �ασιλικ�ν ε6κ να σηµα�νω, τFν σFν τριπ θητ�ν �δελ� ν.

87 Choniates, ed. van Dieten, p. 33, trans. Magoulias, p. 19.88 Michael the Syrian 16.4, ed. Chabot, 3:230 (trans.), 4:612, col. B, lines 32–33 (Syriac text): “wa-

h�d� ı b� eh saggı w-ıqreh rawrb� a©ıt�.”89 Choniates, ed. van Dieten, p. 32, lines 45–47, trans. Magoulias, p. 19: κα3 παρ< �&8ς κατ!λυε δι<

α6δ�&υς �γ µεν�ς 4ς τF ε&6δ�ς τυραννικ#τατ�ς κα3 τF γ!ν�ς 1πισηµ τατ�ς.90 Varzos, Γενεαλ�γ�α, 1:240–42.91 Choniates, ed. van Dieten, p. 32, trans. Magoulias, p. 19.

620 Defection


aqueduct for the monastery of John the Baptist on the Jordan River.92 These weremost likely preparatory measures that can be interpreted as both gestures of pen-itence and signs of agreement with the aims of Byzantine imperial policy, namely,the emperor’s ascendancy over the holy places. In this way Isaac purified himselffrom the stigma of having mingled with Muslims for almost a decade and re-stored his image as an Orthodox Christian and member of the imperial family.Especially noteworthy is Choniates’s focus on emotional aspects, which in his nar-rative constitute causal links in a sequence of decisions and actions. Grief leadsto hatred, which is finally overcome by the inborn feeling of love between thetwo brothers, which, in turn, is celebrated with their subjects’ overwhelming joy.

The reconciliation of 1138/39, however, did not result in a lasting peace be-tween the two branches of the Comnenian family. Already the following year, dur-ing the siege of Neokaisareia, another incident occurred. This time it was Isaac’sson John93 who, shortly before the battle with the Danismend forces began,switched sides in the framework of an exciting spectacle: “He mounted anotherhorse and full of annoyance he took the lance under his arm and rushed in thedirection of the enemies’ ranks. After advancing a short distance, he turned theshort lance to the rear, placing it on his shoulder and, removing the helmet fromhis head, he defected to the Persians.”94 Once more, Choniates’s argument goes,it was an insignificant occasion that triggered an excessive emotional reaction:the emperor had asked for John’s horse, to hand over to a horseless Frankish mer-cenary. What followed was due to John’s “unreasonable and self-willed vanity andhis wholly uncontrollable passion,” “his obstinacy and excessive arrogance.”95

Without doubt the incident shows that the reconciliation of the preceding yearwas built on fragile foundations and that, during their stay in Konya and otherMuslim centers, Isaac and his son had developed close ties with their hosts. John’sdefection seems to have been provoked by an offense against his personal feelingof honor, while it remains unclear whether the emperor insulted his nephew onpurpose or not. Obviously, Choniates tried to conceal the persisting dynastic ten-sions by ascribing the incident to John’s personal deficiencies. John’s estrange-ment from his family and his advanced assimilation into the Muslim-Turkish mi-lieu are evidenced by the fact that, in contrast to his father Isaac, John brokewith his past forever and eventually converted to Islam, marrying the Seljuk sul-tan’s daughter.96 For the first time in the history of Byzantine-Seljuk relations amember of the ruling dynasty became a permanent resident and fully integrated

92 Varzos, Γενεαλ�γ�α, 1:242–43 (with a different interpretation of the event).93 For a biographical sketch see Varzos, Γενεαλ�γ�α, 1:480–85.94 Choniates, ed. van Dieten, p. 36, lines 51–55, trans. Magoulias, p. 21: µετακελλητ�σας δ< ε6ς

Xτερ�ν �θυµ�ας πλ�ρης τF δ ρυ �γκ�ινησ�µεν�ς κατ. τ&ης τ&ων π�λεµ�ων παρατ��εως �!ρεται.µικρFν δ! τι πρ��.ς τFν λ�γ��ρη κ�ντFν ε6ς τ��π�σω �ντ!στρεψε κα3 τ&+ω ν#τ +ω 1παναθ!µεν�ς πρFςΠ!ρσας α�τ µ�λ�ς γ�νεται, τF κρ�ν�ς τ&ης κε�αλ&ης ��ελ µεν�ς.

95 Choniates, ed. van Dieten, pp. 35, lines 40–41, and 36, line 46, trans. Magoulias, p. 21: τ&υ��ςJλ�γ�ς α�τ�θελ�ς κα3 θυµFς παντ�πασιν �κυ�!ρνητ�ς . . . Q δ2 �ρ�νηµατ�ας Wν κα3 γα&υρ�ς πλ!�ντ�&υ δ!�ντ�ς.

96 Choniates, ed. van Dieten, p. 36, trans. Magoulias, p. 21. See also Chalandon, Les Comnène,pp. 179–80; Turan, Türkiye, p. 176; Varzos, Γενεαλ�γ�α, 1:482–83 and 484–85 (concerning the myththat regards John as ancestor of the Ottomans); and Necipoglu, “Turks and Byzantines,” p. 258.

621Defection


dignitary of the sultan’s court in Konya. Accordingly, Choniates’s emphasis onemotional aspects in comparison to Isaac’s case is here even stronger.

Irrationality and uncontrollable passion appear as the predominant motives foraction for other Comnenian defectors as well. No less spectacular an event thanJohn’s flight is that of his younger brother, the enfant terrible Andronikos Kom-nenos,97 who, after an adventurous life full of conspiracy, imprisonment, and ex-ile, became reconciled with his cousin Manuel I in 1166 and was appointed gov-ernor in Cilicia. It comes as no surprise that this crucial position, which gavehim control over the revenues of Cyprus and enabled him to pursue his own po-litical ambitions with respect to Thoros of Cilicia and the principality of Anti-och, very soon led him to a new clash with the emperor, which compelled himonce more to flee from the emperor’s wrath, first to the Kingdom of Jerusalemand subsequently to several eastern lords, eventually ending up at the court ofthe Turkish emir Saltuk of Erzurum, where he stayed until shortly before Man-uel’s death in 1180.98

The extant narratives on these events, written by William of Tyre,99 John Kin-namos,100 and Niketas Choniates,101 pass over in silence all political aspects andfocus exclusively on Andronikos’s love affairs with Philippa, sister of BohemondIII of Antioch and Manuel’s wife Maria, and with Theodora, daughter of An-dronikos’s cousin Isaac and widow of King Baldwin III of Jerusalem. While Wil-liam contents himself with underlining the fraudulent and shameless nature ofAndronikos’s deeds,102 the Byzantine reports blame the impulse of erotic passionand sexual licentiousness, which made Andronikos forget decency, lawfulness, andloyalty, so that he rose up against his brother’s will. According to Choniates, An-dronikos, facing defeat at the hands of the Armenian lord Thoros, forgot aboutwarfare and devoted himself to the orgies of Aphrodite, being “love-smitten byhearsay” and behaving himself “as a horse in heat covering mare after mare be-yond reason.”103

97 For his biography see Oktawiusz Jurewicz, Andronikos I. Komnenos (Amsterdam, 1970); Varzos,Γενεαλ�γ�α, 1:493–638; and Magdalino, Manuel Komnenos, pp. 197–201. For aspects of Androni-kos’s image in Choniates’s work see Niels Gaul, “Andronikos Komnenos, Prinz Belthrandos und derZyklop: Zwei Glossen zu Niketas Choniates’ Yρ�νικ� δι�γησις,” Byzantinische Zeitschrift 96 (2003),623–60.

98 For the events see Varzos, Γενεαλ�γ�α, 1:518–28. For a thorough discussion of the political back-ground and possible Byzantine reactions to Andronikos’s reception by King Amalric, who enfeoffedhis guest with Beirut, see Lilie, Byzantium and the Crusader States, pp. 193–95.

99 William of Tyre, Chronicon, ed. R. B. C. Huygens, CCCM 63–63A (Turnhout, 1986), 20.2.1–14, 21.12.40–46, pp. 913–14, 978–79.

100 Ioannes Cinnamus, Epitome rerum ab Ioanne et Alexio Comnenis gestarum, ed. August Meineke(Bonn, 1836), 6.1, pp. 250–51; trans. Charles M. Brand, Deeds of John and Manuel Comnenus byJohn Kinnamos (New York, 1976), pp. 188–89.

101 Choniates, ed. van Dieten, pp. 137–42.102 William of Tyre 20.2.9–13, ed. Huygens, p. 914: “dominam Theodoram . . . fraudulenter, ut

dicitur, abduxit et in terram hostium, Damascum prius, deinde in Persidem favente Noradino trans-vexit”; ibid. 21.12.44–46, ed. Huygens, p. 979: “dominam Theodoram . . . inpudenter nimis sed etinpudice non minus clam abduxit.”

103 Choniates, ed. van Dieten, p. 139, line 36, trans. Magoulias, p. 79: &;ν γ.ρ 1� �κ�&ης1ρωτ ληπτ�ς; and Choniates, ed. van Dieten, p. 141, lines 88–90, trans. Magoulias, p. 80: 4ς Rππ�ς

622 Defection


Meanwhile, Manuel is presented as growing more and more enraged at his cous-in’s indecent and unlawful liaisons, so that he made every possible effort to cap-ture and punish him. The emperor’s wrath, in turn, made Andronikos fear forhis safety, so he left Antioch for Jerusalem, where he became involved in a newaffair with Theodora. An imperial letter signed with red ink, ordering the lordsof the kingdom to imprison and blind his cousin,104 horrified Andronikos to suchan extent that he made the decision to desert to the empire’s enemies. But evenunder these precarious circumstances he still demonstrated his deceitfulness, se-ducing Theodora to follow him as companion and fellow wanderer.105 The restof the story is quite reminiscent of sebastokrator Isaac’s exile in the 1130s. An-dronikos was received and greatly honored at many courts; if we trust Kinna-mos, from Saltuk’s realm he undertook a series of incursions into Byzantine ter-ritory, for which he was excommunicated by the church of Constantinople.106

Seemingly, just like his father, for some years Andronikos attempted to exertpressure on the imperial government in order to force the emperor to make con-cessions or to recognize his rights as a member of the ruling dynasty. Be thishistorical truth or sensational fiction, the fact is that, in his presentation of An-dronikos’s defection to the Turks, Choniates draws clear parallels with his fa-ther Isaac and his brother John, adding to the emotional conditions that moti-vated the behavior of the latter, that is, extreme sensitivity and youthful arrogance,a combination of sexual passion and craftiness. Andronikos possesses the worstof a whole series of negatively connoted characteristics. Going against the prin-ciples of righteous rule, they exclude a person exhibiting them from a share inimperial power.

Andronikos’s return fourteen years later, once more, was motivated by emo-tions and realized through a skillfully choreographed sequence of ritual acts. An-dronikos’s love for his mistress and their children drove him to seek the emper-or’s pardon after Theodora had been kidnapped and brought to Constantinoplewith the aid of the governor of Trebizond. An embassy asking the emperor for

θηλυµαν�ς τα&ις παραλ γ�ις 1πιθ�ρν�µεν�ς µ��εσιν 1γ�ρ�πτει σαρκικ&ως κα3 1νασελγα�νει �κ�λ�στωςτ&+ η τ�&υ πρ#τ�υ 1�αδ!λ��υ θυγατρ3 Θε�δ#ρ+α.

104 Choniates, ed. van Dieten, p. 141, lines 1–2, trans. Magoulias, p. 80: 1ρυθρ�δ�ν +ω 1νσηµανθ2ν��&ηκε γραµµ�τι�ν. See also Dölger and Wirth, Regesten, no. 1471 (ca. 1167). Lilie, Byzantium andthe Crusader States, p. 195, correctly in my opinion, suggests 1168, given that Andronikos arrivedin Jerusalem shortly before Amalric’s marriage to Maria in August 1167 and that the love affair neededsome time to develop. Choniates somewhat inadequately calls the recipients of the letter τ�Oς κατ.Κ��λην Συρ�αν 6σ��ντας, but logically one has to think of King Amalric and the royal entourage.That the letter might have been handed over by the embassy of Alexander of Gravina (Dölger andWirth, Regesten, no. 1481 [arrival in summer 1168]), as Lilie, Byzantium and the Crusader States,p. 194 n. 218, assumes, seems quite convincing. William’s words, according to which Andronikosfled “favente Noradino” via Damascus to Persia (a region farther east or Seljuk territory?), do notnecessarily mean that he was actually received by Nur al-Dın or collaborated with him to the detri-ment of the Crusader States (Lilie, Byzantium and the Crusader States, p. 195).

105 Choniates, ed. van Dieten, p. 142, lines 16–22, trans. Magoulias, p. 81: σ��ιστ�ς δ2 Wν κα3�πατε@ν κα3 α�τ�ν �π�πλαν&+α Θε�δ#ραν . . . συν!κδηµ�ν κα3 συµπλαν&ητιν >κ�&υσαν 1πεσπ�σατ�Jκ�υσαν.

106 Kinnamos 6.1, ed. Meineke, p. 251, trans. Brand, pp. 188–89.

623Defection


amnesty prepared his restoration to his former position.107 Afterwards, Androni-kos appeared before Manuel carrying a heavy iron chain around his neck thatreached down to his feet, stretched himself out on the floor, and, shedding tears,begged forgiveness. In the end he even insisted on being dragged by the chainbefore the throne.108 Andronikos, drawing on gestures and symbols pertainingto the sphere of slaves and prisoners of war, demonstrated his unconditional sub-jugation and extreme humility before the imperial court. This enabled the em-peror to show mercy and grant a pardon without undermining his prestige ashead of the dynasty.109

For Choniates, who constantly interprets Andronikos’s deeds in the light of hissubsequent insurrection against the dynasty’s legitimate heirs, the whole scene isnothing but another example of Andronikos’s cunning and hypocrisy.110 Given thatin the summer of 1180, perhaps apart from the emperor’s imminent death, the fu-ture state of affairs was not yet foreseeable, this opinion seems highly doubtful. Inall likelihood, the sequence of ritual acts described, in all its details, was agreedupon in the course of the preliminary talks as a precondition for the defector’s re-integration into the court and the imperial family. While in 1139 the restorationof dynastic peace could be publicly celebrated through the emperor’s triumphalentrance into the capital, in 1180 Manuel I had no political success to commem-orate. An honorable reconciliation could be achieved only through a gesture ofhumiliation, projecting the emperor’s clemency toward his penitent relative.111

107 Choniates, ed. van Dieten, p. 226, trans. Magoulias, p. 128.108 Choniates, ed. van Dieten, pp. 226–27, trans. Magoulias, pp. 128–29.109 For Western parallels of public demonstrations of humility from the ninth century onwards see

Geoffrey Koziol, Begging Pardon and Favor: Ritual and Political Order in Early Medieval France(Ithaca, N.Y., 1992); Gerd Althoff, “Das Privileg der deditio: Formen gütlicher Konfliktbeendigungin der mittelalterlichen Adelsgesellschaft,” in Spielregeln der Politik im Mittelalter (above, n. 17),pp. 99–125. The rope around the neck was a commonly understood symbol of captivity used duringtortures and executions; see, for example, an eleventh-century depiction of Christ on the way to Cal-vary preserved in Elmali Kilise in Cappadocian Göreme (Catherine Jolivet-Lévy, La Cappadocemédiévale: Images et spiritualité [Paris, 2001], plate 141) and the portrayal of the flagellation andblinding of the rebel Lampros in Skylitzes Matritenses, fol. 225v, miniature 550 (Vasiliki Tsamakda,The Illustrated Chronicle of Ioannes Skylitzes in Madrid [Leiden, 2002], p. 250, fig. 534). I am es-pecially grateful to Dr. Maria Parani (University of Cyprus) for these references to the field of Byz-antine iconography.

110 Choniates, ed. van Dieten, p. 226, line 86, trans. Magoulias, p. 128: Π�λυµ��αν�ς δ< Wν<Ανδρ νικ�ς κα3 παντ��ις δ λ�ις κα' µεν�ς. Varzos, Γενεαλ�γ�α, 1:528–29, does not understandthe meaning of the described ceremony.

111 Very similar model patterns of conflict resolution can be observed in the Latin West from thetenth century onwards: Gerd Althoff, “Königsherrschaft und Konfliktbewältigung im 10. und 11.Jahrhundert,” in Spielregeln der Politik im Mittelalter, pp. 21–56, esp. p. 36: “Die Gegner des Königsaus den Führungsschichten und seiner eigenen Familie können auf weitestgehende Schonung rech-nen. . . . Voraussetzung für diese Schonung war jedoch, daß sie sich zur bedingungslosen Unterwer-fung bereiterklärten. . . . Der sich bedingungslos Unterwerfende brachte ja einen beträchtlichen Teilseiner dignitas in die Konfliktbereinigung ein.” For the crucial significance of intermediaries and pre-liminary negotiations preceding public reconciliation ceremonies see Hermann Kamp, Friedensstifterund Vermittler im Mittelalter (Darmstadt, 2001), esp. pp. 64–82 (“Das Bitten um Nachsicht undMilde”). For the idea of showing pity to captives in the framework of twelfth-century chivalrousmoral codes see Yvonne Friedman, Encounter between Enemies: Captivity and Ransom in the LatinKingdom of Jerusalem, Cultures, Beliefs and Traditions 10 (Leiden, 2002), pp. 65–72.

624 Defection


Two other cases of defection among members of the ruling dynasty are at-tested in the years of the Angeloi emperors. Isaac Komnenos, grandson of Man-uel’s brother Isaac and ex-ruler of Cyprus,112 after his release from Frankish cap-tivity in Palestine, probably in late 1195, fled to the court of Sultan Ghiyathal-Dın Kaykhusrau of Konya.113 Likewise, in about 1200 Michael Angelos-Doukas, dux of the thema Mylassa and bastard son of the sebastokrator JohnDoukas and, therefore, Alexios III’s cousin, fled to Kaykhusrau’s brother Ruknal-Dın after his rebellion had failed and imperial troops defeated him in battle.114

In Choniates’s narrative the behavior of the two rebels is molded on exactlythe same model as that of the sebastokrator Isaac and his sons: Isaac Komnenos’sdecision resulted from his excessive lust for power. We are told that he becamevery annoyed at letters of Emperor Alexios III recalling him to Constantinople:“I have learned to rule,” he angrily responded, according to Choniates, “not tobe ruled, and to lead, not to obey others.”115 His attempts failed, however, forneither the sultan nor other Turkish lords in Anatolia were prepared to supporthim. Likewise, the arrogance of his youth led Michael to sedition and defec-tion.116 Yet, since his flight coincided with a serious estrangement between Con-stantinople and Konya, including an attempt on Rukn al-Dın’s life, Michael wasmuch more successful in gaining the sultan’s military support and undertook aseries of raids against the towns in the Meander valley, “proving himself an evenmore pitiless murderer than the barbarians.”117 There is a clear parallel betweenIsaac’s elder son John and Michael, who were both accused of excessive arro-gance resulting from their youthful impetuosity.

Before proceeding to a further interpretation of the aforementioned incidents,a second group of Byzantine apostates, consisting of military commanders andcourt officials who fell from grace, should be considered. The most prominentcases are the protostrator Alexios Axouch, whom Emperor Manuel dischargedfrom his post in 1167,118 and the megas kontostablos Michael Palaiologos, who

112 He gained fame in crusader chronicles because of his confrontation with King Richard I in April1191; see John Gillingham, Richard I (New Haven, Conn., 1999), pp. 140–54; and A. Nicolaou-Konnari, “The Conquest of Cyprus by Richard the Lionheart and Its Aftermath: A Study of Sourcesand Legend, Politics and Attitudes in the Years 1191–1192,” Επετηρ�δα τ�υ Κ!ντρ�υ Επιστηµ�νικ#νΕρευν#ν 26 (2000), 25–123.

113 Choniates, ed. van Dieten, pp. 463–64, trans. Magoulias, pp. 254–55; Cheynet, Pouvoir, pp. 130–31, no. 183.

114 Choniates, ed. van Dieten, p. 529, trans. Magoulias, p. 290; Cheynet, Pouvoir, p. 134, no. 190.115 Choniates, ed. van Dieten, p. 464, lines 92–93, trans. Magoulias, p. 254: �ασιλε�ειν, �� µ�ν

�ασιλε�εσθαι λ!γων µαθε&ιν, κα3 Jλλων 1�ηγε&ισθαι, µ� γ�&υν >τ!ρ�ις πε�θεσθαι. For the bad imageof Isaac Komnenos in Byzantine and Western sources as a cruel tyrant see Nicolaou-Konnari, “Con-quest of Cyprus,” pp. 77–88.

116 Choniates, ed. van Dieten, p. 529, line 19, trans. Magoulias, p. 290: ν!�ς Wν κα3 α�θ�δης.117 Choniates, ed. van Dieten, p. 529, lines 23–24, trans. Magoulias, p. 290: �ε�ρων τ&ων �λλ��λων

κα3 νηλε!στερ�ς �νδρ�� ν�ς δεικν�µεν�ς.118 Alexios was the son of the Great Domestic John Axouch, a former Turkish prisoner of war,

who grew up at the imperial court and held the supreme military command under John II. Alexioswas married to Maria, a daughter of Manuel’s elder brother Isaac. For Alexios Axouch’s role at Man-uel’s court see Brand, “Turkish Element” (above, n. 9), pp. 5–6, 8–10; Magdalino, Manuel Kom-nenos, pp. 6, 7, 19, 61–62, 107, 218–19, 224; Cheynet, Pouvoir, pp. 109, no. 148, and 415–16;

625Defection


in 1256 came into conflict with Emperor Theodore II Laskaris.119 In spite of thechronological distance of nearly a century, there are some remarkable similari-ties between the two incidents. Both Alexios and Michael were members of highlyesteemed families, enjoyed an important position among the military leaders ofthe empire, and were related to the ruling dynasty through marriage. Both weresuspected of seditious plans against the emperor. As a result they lost his favorand were seriously threatened in their social and political position.

The basic motif of the historiographical narrative in both cases is the breachof allegiance and the question of guilt resulting from it: does the responsibilitylie with the emperor or with his official? As for Alexios Axouch, it is John Kin-namos who talks about secret contacts with and a sort of ideological defectionto the Seljuk sultan,120 whereas Niketas Choniates suggests that the reason forthe conflict was the emperor’s envy of individuals who distinguished themselvesby their personal abilities. Manuel could not stand Alexios’s respectability andthe high esteem he enjoyed among the soldiers and military commanders. Be-sides, he had an eye on Alexios’s wealth.121 Therefore the emperor gave the or-

Balivet, “Entre Byzance et Konya” (above, n. 10), pp. 54–55; and Necipoglu, “Turks and Byzan-tines” (above, n. 11), p. 257.

119 For matters of chronology see Albert Failler, “Chronologie et composition dans l’histoire deGeorges Pachymère,” Revue des études byzantines 38 (1980), 5–103, at pp. 16–17, esp. n. 46. Forthe circumstances of Michael’s flight to Sultan ¨Izz al-Dın II, his sojourn at the Seljuk court, and hisparticipation in the battle of Akseray against the Mongols in the fall of 1256 see Deno John Geana-koplos, Emperor Michael Palaeologus and the West, 1258–1282: A Study in Byzantine-Latin Rela-tions (Cambridge, Mass., 1959), pp. 28–30; Günter Prinzing, “Ein Mann τυρανν�δ�ς J�ι�ς: ZurDarstellung der rebellischen Vergangenheit Michaels VIII. Palaiologos,” in Lesarten: Festschrift fürAthanasios Kambylis zum 70. Geburtstag dargebracht von Schülern, Kollegen und Freunden, ed. Io-annis Vassis (Berlin, 1998), pp. 180–97; and the comments in Macrides’s Akropolites translation,pp. 314–15, 317–19. The significance of the event in the broader context of Byzantine-Mongol re-lations is carefully analyzed by John S. Langdon, “Byzantium’s Initial Encounter with the Ching-gisids: An Introduction to the Byzantino-Mongolica,” Viator: Medieval and Renaissance Studies 29(1998), 95–140, at pp. 132–33 and esp. n. 218; idem, “Twilight of the Byzantine Lascarid Basileiain Anatolian Exile, 1254–1258: Continuity and Change in Imperial Geopolitical Strategy,” Viator:Medieval and Renaissance Studies 34 (2003), 187–207, at pp. 188–94, esp. p. 193 n. 65, where theauthor stresses the importance of the marriage between a daughter of John III Vatatzes and Sultan¨Izz al-Dın II, which made Emperor Theodore II the sultan’s brother-in-law. For a new examinationof Michael VIII’s relations with the sultanate of Konya see Dmitri A. Korobeinikov, “Michail VIIIPaleolog v Rumskom Sultanate,” Vizantiiskii Vremennik 64 (2005), 77–98, and his forthcoming mono-graph based on his 2004 Oxford D.Phil. thesis, Byzantium and the Turks in the Thirteenth Century.

120 Kinnamos 6.6, ed. Meineke, pp. 265–66, 267, trans. Brand, pp. 199–201.121 Choniates, ed. van Dieten, pp. 143–44, lines 70–72, trans. Magoulias, pp. 81–82: 1κ τ�&υ τFν

Jνδρα Qρ&αν κα3 στρατηγ�&ις κα3 στρατι#ταις 1ρ#µεν�ν κα3 �ιλ�τ�µ +ω γν#µ+η κα3 �ειρ3 �ιλ�δ#ρ +ωπρFς Kπαντας �ρ#µεν�ν, 5σως δ2 κα3 τ�&υ πρ�σ ντ�ς α�τ&+ω πλ��τ�υ Bπ�τ��ων :�εσιν. For the roleof envy in the strictly hierarchically organized court culture of Byzantium, mainly on the basis ofhistoriographical and autobiographical texts, see Martin Hinterberger, “� �θ ν�ς ανθρ#πινηαδυναµ�α και κινητ�ρια δ�ναµη” [Envy: Human weakness or moving force?], in Byzantium Ma-tures: Choices, Sensitivities, and Modes of Expression (Eleventh to Fifteenth Centuries), ed. Chris-tine Angelidi, Institute for Byzantine Research, International Symposia 13 (Athens, 2004), pp. 299–312.

626 Defection


der to take Alexios into custody, charging him with sorcery, although Alexioshad always maintained his deference and loyalty toward Manuel.122 On the con-trary, Kinnamos suggests that for a long time Alexios had been thinking of apos-tasy.123 Therefore, when he was entrusted with an expedition against Cilicia, hedeliberately visited the sultan of Konya in order to win him over to his conspir-acy.124 Moreover, in one of his palaces outside Constantinople Alexios orderedsome frescoes to be painted that, instead of depicting hunting scenes or the em-peror’s military exploits, showed victorious campaigns of the sultan, “foolishlymaking public in painting in his residence what should have been concealed indarkness.”125 Regardless of whether the episode can claim any historical authen-ticity, it is clearly based on the idea that Alexios, though a faithful member ofthe imperial entourage and a man of Christian Roman identity in the second gen-eration, still could be suspected of having a certain proclivity to the world of hisbarbarian ancestors and thus of being susceptible to any kind of temptation of-fered by the Turkish-Muslim enemy to the detriment of the empire.126

One can observe a very similar argumentation with respect to Michael Palaio-logos. All sources agree that Theodore II Lascaris constantly threatened him withhis anger and severity. In his autobiographical account Michael gives us his owninterpretation of the incident: “When the empire of the Romans passed to hisson [Theodore II], along with many others who tried them, we, too, had to en-dure the arrows of envy. But how did God keep us undamaged in this situationand how did he lead us from this grief to success? In summary, he offered usrefuge with the Persians. Here, too, he holds my right hand and gloriously caresabout me.”127 Michael appears as the victim of envy (�θ ν�ς), which, in turn,forms part of a divine plan, on the account of which Michael is led to the impe-rial throne.128 The historiographical sources on Michael’s flight unanimously stresshis great fear of the emperor’s imminent punishment. Yet they give different ex-

122 Choniates, ed. van Dieten, p. 144, trans. Magoulias, p. 82.123 Kinnamos 6.6, ed. Meineke, p. 265, line 16, trans. Brand, p. 199: 1κ π�λλ�&υ µ2ν �π�στασ�αν

Tδινε.124 Kinnamos 6.6, ed. Meineke, pp. 265, line 23–266, line 2, trans. Brand, p. 199: 1π�τηδες 1ς

<Ικ νι�ν συγγεγ�ν@ς τ&+ω σ�υλτ.ν ε5ς τε �ιλ�αν α�τFν Bπηγ�γετ� κα3 τ&ων ε6ς τυρανν�δα τειν ντωνδιελ!�ατ� τ��τ +ω π�λλ�.

125 Kinnamos 6.6, ed. Meineke, p. 267, lines 14–16, trans. Brand, p. 200: τ.ς τ�&υ σ�υλτ.ν�νεστ�λ�υ στρατηγ�ας, ν�πι�ς Kπερ 1ν σκ τ +ω �υλ�σσειν 1�ρ&ην τα&υτα 1π3 δωµ�των α�τFςδηµ�σιε�ων κατ. τ�ν γρα��ν.

126 Magdalino, Manuel Komnenos, p. 219, quite imprecisely talks about an “ethnic factor” in Alex-ios Axouch’s party.

127 H. Grégoire, “Imperatoris Michaelis Palaeologi de vita sua opusculum necnon Regulae quamipse monasterio S. Demetrii praescripsit fragmentum,” Byzantion 29–30 (1959–60), 447–76, atpp. 452–53.

128 For Michael VIII’s “autobiography,” a sort of introduction to a monastic foundation docu-ment, see Martin Hinterberger, Autobiographische Traditionen in Byzanz, Wiener byzantinistischeStudien 22 (Vienna, 1999), pp. 46, 60–61, 127–29, 134–37, 188–90, 266–76 (in general Michaelperceives his role as a mere instrument of the wonderful deeds of God). For the motif of envy andemotion-based interpretative models in Byzantine literature see the studies of Martin Hinterberger,cited above, nn. 16 and 121.

627Defection


planations for Theodore’s anger. According to George Akropolites,129 who em-bedded the discussion of the reasons of Michael’s flight into a dialogue betweenhimself and the emperor, Theodore II had frequently threatened Michael andopenly expressed his wish to blind him. Nikephoros Gregoras, like Michael him-self, makes use of the motif of envy: “When he realized how envy everywheregot stronger and stronger, how insidious words secretly came to his ears, andhow the emperor started to think of and faintly talk about punishments that nor-mally would be imposed only on barbarian enemies, he could no longer get anypeace.”130 The severity and irascibility of the emperor along with the maliciousconspiracies of insidious people caused Michael to set off for the Seljuk court.131

Only George Pachymeres hints at the possibility of disloyal behavior on the partof the ambitious commander, since suspicions openly spoken against him put himin a state of permanent fear. When his uncle, the megas chartularios Michael,was taken into custody for a careless word, he decided to leave the country.132

Byzantine authors juxtapose rulers’ envy and uncontrollable anger with theirseditious subjects’ irrationality and greed. Unhinged emotions can dominate ac-tions and forms of behavior on either side. Drawing a comparison between thetwo categories of apostates, members of the dynasty versus high-ranking offi-cials, one comes across very similar motifs, such as the destructive force of envy(the sebastokrator Isaac, Michael Palaiologos), fear of the emperor’s punishment(Andronikos Komnenos, again Michael Palaiologos), and lust for power (IsaacKomnenos, Alexios Axouch). There are, however, remarkable differences withrespect to their function. The members of the imperial dynasty formed the back-bone of the state’s stability. This holds true especially for the Comnenian systemof government, which was based on the practice of conferring the most impor-tant offices and military posts on the emperor’s relatives and in-laws, thereby giv-ing them a significant share in imperial power.133

The defection of persons pertaining to the empire’s inner circle was inevitablyviewed as an extremely dangerous threat to the very existence of the state. As aresult, these individuals were portrayed in a manner that actually disqualified themfrom imperial ranks. They had lost their minds and turned out to be completely

129 Akropolites 64, ed. Heisenberg, 1:134–35, trans. Makrides, pp. 312–13.130 Nikephoros Gregoras, Byzantina historia, ed. Ludwig Schopen, 1, Corpus Scriptorum Histo-

riae Byzantinae 19 (Bonn, 1829), p. 58; trans. Jan Louis van Dieten, Nikephoros Gregoras: RhomäischeGeschichte (Historia Rhomaïke), Bibliothek der griechischen Literatur 4 (Stuttgart, 1973), pp. 91–92.

131 Gregoras, ed. Schopen, p. 58, lines 5–15, trans. van Dieten, p. 92: 1δεδ�ει γ.ρ τF τ�&υ �ασιλ!ωςπρFς τ.ς τιµωρ�ας �π τ�µ ν τε κα3 �πην2ς κα3 M��[\�π�ν . . . κα3 µεγ�λας δια��λ.ς . . . , ]ς �8�θ�ν�&υντες κατ< α�τ�&υ συνεκ�κησαν κα3 τ.ς τ�&υ �ασιλ!ως �κ�.ς 1µπεπλ�κασιν . . . �υγ&+ η τ�νσωτηρ�αν π�ρ�'εσθαι �!λτι�ν π�σης 1τ�θει ��υλ&ης.

132 George Pachymeres, Relations historiques, ed. Albert Failler, trans. Vitalien Laurent, 1, CorpusFontium Historiae Byzantinae 24/1 (Paris, 1984), 1.9, pp. 43–45, esp. p. 43, lines 13–14: τFν �ε3περ3 >αυτ&+ω δεδι τα δι. τ�ν Bπ�τρ!��υσαν τ&ης �ρ�&ης Bπ�ψ�αν.

133 For the Comnenian system and the court hierarchy from Alexios I’s reign onwards see Lilie,“Macht und Ohnmacht” (above, n. 83), pp. 11–58; Magdalino, Manuel Komnenos, pp. 180–201;Cheynet, Pouvoir, pp. 249–301; and Michael Angold, Church and Society in Byzantium under theComneni, 1081–1261, 2nd ed. (Cambridge, Eng., 2000), pp. 15–41 (concerning the changing rela-tionship between the emperor and the church until Alexios I’s rise to power).

628 Defection


incapable of participating in the imperial government. Only genuine penitenceexpressed through publicly performed gestures of humiliation and the emperor’sclemency could restore them to their former status. On the contrary, high-rankingofficials, in a system that depended largely on personal loyalty, had to live underthe constant threat of opponents who were able to influence the emperor andwin him over for their own ambitions. Byzantine society did not provide the se-curities of the Western feudal system, where the lord was obliged to protect hisvassal and the latter, in case of a breach of allegiance, had a legal right to resisthis lord. In Byzantium, oaths of allegiance were normally one-sided, from theofficial to the emperor.134 It comes as no surprise that the idea of the emperor’swrath was especially horrifying and could even be employed as a reasonable jus-tification for apostasy. In this case, defection seems to have been, if not a legal,then at least an understandable form of reaction to a threat to one’s position. Ofcourse, this model worked only if a rebel’s undertaking was crowned by success,as was the case with Michael Palaiologos, the future emperor and “New Con-stantine.” If he failed, as did Alexios Axouch, who ended his life as a monk, therebel’s deeds were either ascribed to his insanity or deliberately concealed.

To judge from Akropolites’s account, Michael’s attempt to seek the sultan’s pro-tection was by no means an isolated case in the years of the so-called empire inexile but forms part of a general pattern of behavior among Byzantine aristo-crats fleeing the emperor’s wrath. Andronikos Nestongos, a member of an influ-ential family at the court of Nicaea and head of a failed conspiracy against Em-peror John III Vatatzes in 1224/25, managed to escape from the fortress ofMagnesia by running off to Muslim territory, where he spent the rest of his life.135

When in 1237 the blinded ex-emperor Theodoros Angelos returned from his Bul-garian captivity to Thessalonica, he exiled his brother, the despotes Manuel An-gelos, to Attaleia, where he was treated with clemency by the local Seljuk au-thorities and, according to his wish, sent to Emperor John III.136 In 1259 anopponent of Michael VIII’s ascent to the imperial throne, the protovestiarites The-odore Karyanites, also tried to take flight to the sultanate. Yet he was less fortu-nate than his predecessors, for on his way he was captured, robbed, and killedby a band of Turcomans.137

134 For oaths in the Western Latin sphere see, for instance, Charles E. Odegaard, “Carolingian Oathsof Fidelity,” Speculum 16 (1941), 284–96; and idem, “The Concept of Royal Power in CarolingianOaths of Fidelity,” Speculum 20 (1945), 279–89. For Byzantine oaths see Nikos Svoronos, “Le ser-ment de fidelité à l’empereur byzantin et sa significance constitutionelle,” Revue byzantine 9 (1951–52), 106–42; Angeliki Laiou, “The Emperor’s Word: Chrysobulls, Oaths and Synallagmatic Rela-tions in Byzantium (11th–12th C.),” Travaux et mémoires 14 (2002), 347–62; the article of AlexanderKazhdan, “Eid, B I. Byzantinisches Reich,” in Lexikon des Mittelalters, 9 vols. (Munich, 2002),3:1689–90, esp. col. 1690: “Der Eid des Kaisers gegenüber einem Untertanen, einem Privatmann,war dagegen selten, aber nicht unmöglich.”

135 Akropolites 23, ed. Heisenberg, 1:37, lines 23–25, trans. Macrides, pp. 169–70, with com-ments on pp. 170–71: Sς κα3 νυκτFς �υγ@ν περ3 τ. τ&ων Μ�υσ�υλµ�νων 1�#ρµησε, κα3 δι&ηγεν1κε&ισε µ!�ρι τ&ης α�τ�&υ τελευτ&ης.

136 Akropolites 38, ed. Heisenberg, 1:61, lines 16–17, trans. Macrides, pp. 206–7, with commentson pp. 208–9: �ιλανθρωπ�+α µ&αλλ�ν πρFς α�τFν �ρησαµ!ν�υς.

137 Akropolites 77, ed. Heisenberg, 1:160, line 1, trans. Macrides, pp. 346–47, with comments onp. 349 n. 7: Q δ2 �π�δρ.ς περ3 τ. τ&ων Περσ&ων +T�ετ�.

629Defection


Akropolites’s judgment on these cases follows the same principles as the inci-dents discussed above, clearly reflecting the author’s personal stance toward po-litical developments and contesting factions. While the rebel Nestongos, havingdisregarded kinship and bonds of friendship with the emperor, simply lost hisright to be considered a Byzantine aristocrat, the despotes Manuel appears as avictim of his brother’s wrath, being saved by the Turks’ and John III’s clemency.The author accuses Theodore Karyanites of having murdered the Mouzalon broth-ers, so that his own death at the Turks’ hand is but just punishment. We willnever learn whether the true reasons for Karyanites’s violent death lay else-where, perhaps in the especially cordial relations between Michael VIII and Sul-tan ¨Izz al-Dın Kaykawus II, which a few years later would prepare the sultan’sown flight to the imperial court.138

Local Rebels Collaborating with the Enemy (1118–1260)

Another category of defectors consists of local lords in the Anatolian borderregions of the empire, mainly in the Pontus and in the Upper Meander valley.Their vicinity to Muslim areas brought about increased contacts with Turkishpopulations and tendencies of mutual acculturation. Naturally, this situation some-times resulted in various forms of political cooperation and even defection. “Thus,custom, reinforced by time, is stronger than race and religion,” Choniates as-serts with respect to the Greek inhabitants of the Lake Pousgouse region (Beyse-hir Gölü), who by mingling with the Turks of Konya had strengthened their mu-tual friendship and their commercial ties to such an extent that they regardedthe Byzantines as their enemies.139 Similar phenomena can be observed amongthe lords of the Pontus region, such as Constantine Gabras of Trebizond, who inthe years 1126–40 governed the region as an independent ruler, simultaneouslymaintaining close contacts with the Danismend emir Gümüstekin Ghazı, or thelocal ruler Kassianos, who in 1130 submitted to the same emir, delivering to himhis fortresses in the coastal region of the Pontus.140 These phenomena becamemore frequent in the years between Andronikos I’s seizure of power (1182/83)and the fall of Constantinople in 1204. As is generally known, this period is char-acterized by the decay of central authority and the appearance of independentrulers in the provinces.141 Individuals who, under these sociopolitical circum-

138 See below, pp. 643–47.139 Choniates, ed. van Dieten, p. 37, lines 92–93, trans. Magoulias, p. 22: �7τω �ρ ν +ω κρατυνθ2ν

:θ�ς γ!ν�υς κα3 θρησκε�ας 1στ3ν 6σ�υρ τερ�ν.140 Cahen, Pre-Ottoman Turkey, p. 94; Anthony Bryer, “A Byzantine Family: The Gabrades, c. 979–c.

1653,” University of Birmingham Historical Journal 12 (1970), 164–87, at p. 170; idem, with Ste-phanos Fassoulakis and Donald M. Nicol, “A Byzantine Family: The Gabrades, an Additional Note,”Byzantinoslavica 36 (1975), 38–45, at p. 39, both articles reprinted in idem, The Empire of Trebi-zond and the Pontos, Collected Studies Series 117 (London, 1980), nos. 3a, 3b; Cheynet, Pouvoir,pp. 104–5, nos. 137–38, and 404–5, 417. See also Choniates, ed. van Dieten, p. 34, trans. Magou-lias, p. 20; and Michael the Syrian 16.3, ed. Chabot, 3:227 (trans.), 4:610, col. B, lines 25–27 (Syr-iac text).

141 For this process see Jürgen Hoffmann, Rudimente von Territorialstaaten im byzantinischen Reich(1071–1210): Untersuchungen über Unabhängigkeitsbestrebungen und ihr Verhältnis zu Kaiser und

630 Defection


stances, defected to the Turks either were members of the provincial aristocracyor attempted to obtain the backing of the provincial population. Flight to theSeljuk neighbors could be motivated by the desire to muster military support inorder to initiate an insurrection against the central government. In other casesapostates were compelled to seek refuge with the Turks because they had lostthe support of the local Byzantine population.

That was the case with Manuel and Alexios, the sons of the megas domestikosJohn Vatatzes, who in Phrygian Philadelphia in 1182 resisted the usurpation ofAndronikos I.142 After John’s sudden death, the inhabitants of the city decided torecognize the new regime in Constantinople, so that the two brothers, facing thepressure of the townspeople, had no choice other than to flee to the court of Sul-tan Kılıc Arslan.143 The Vatatzes brothers, however, did not pursue any plans forfurther resistance. In contrast, in the following years, some new leaders of localinsurrections aimed at military cooperation with the neighboring Turks.

The “ex-emperor” Theodore Mangaphas of Philadelphia in 1191, despite theguarantees of Isaac II, was forced by the dux of Thrakesion, Basil Vatatzes, toseek refuge in Konya.144 The two Pseudo-Alexioi, in 1191 and 1195 respective-ly,145 pretended to be Manuel’s murdered son and legitimate emperor Alexios II.All three of them convinced Turkish lords to support their cause, if not with reg-ular troops, then at least with the permission to recruit warriors from the Tur-coman nomads who were settling along the frontier regions of the Seljuk sultan-

Reich, Miscellanea Byzantina Monacensia 17 (Munich, 1974); Lilie, “Macht und Ohnmacht,” passim;and Cheynet, Pouvoir, pp. 427–40, 446–58.

142 Choniates, ed. van Dieten, pp. 262–64, trans. Magoulias, pp. 146–47; Cheynet, Pouvoir, p. 113,no. 154; John S. Langdon, “Backgrounds to the Rise of the Vatatzai to Prominence in the ByzantineOikoumene, 997–1222,” in ΤF /Ελληνικ ν: Studies in Honor of Speros Vryonis, Jr., 1: Hellenic An-tiquity and Byzantium, ed. John S. Langdon et al. (New Rochelle, N.Y., 1993), pp. 179–212, atpp. 197–98 n. 35.

143 Choniates, ed. van Dieten, p. 264, lines 55–57, trans. Magoulias, p. 147: �8 δ2 τ�&υ µεγ�λ�υδ�µεστ�κ�υ υ8�3 δε�σαντες, µ� συσ�εθε&ιεν κα3 καταπρ�δ�θε&ιεν <Ανδρ�ν�κ +ω, µεθ�στανται 1κε&ιθενκα3 τ&+ω <Ικ�νιε&ι σ�υλτ.ν πρ�σρ��νται. In addition, see August Heisenberg, “Kaiser Johannes der Barm-herzige: Eine mittelgriechische Legende,” Byzantinische Zeitschrift 14 (1905), 160–233, edition ofthe vita at pp. 193–233 (Β��ς τ�&υ Pγ��υ <Ιω�νν�υ �ασιλ!ως τ�&υ <Ελε�µ�ν�ς), esp. p. 205, lines 12–14: �π�δρ�ντε τ��νυν ε6ς <Ικ νι�ν ��ικ!σθην, �σµ!νως σ�&ωε τ�&υ τ&ης �#ρας =γεµ ν�ς δε�αµ!ν�υκα3 �ιλ��ρ νως (in this version the Vatatzes brothers are called Nikephoros and Theodore).

144 Choniates, ed. van Dieten, pp. 399–401, esp. p. 400, lines 81–82, trans. Magoulias, pp. 219–20: <Αλλ< �&Bτ�ς τ&+ω τ�&υ <Ικ�ν��υ σ�υλτ.ν πρ�σρυε�ς, &+4 = κλ&ησις Καϊ��σρ ης. See also Cheynet,Pouvoir, p. 123, no. 168. Michael the Syrian 21.3, ed. Chabot, 3:394–95 (trans.), 4:728, col. C, lines1–44 (Syriac text), quotes a letter of Kılıc Arslan II mentioning a “nephew of the king of the Ro-mans” (“bar ah� a d� -malka d� -Romaye”) who had arrived from Philadelphia and had declared his sub-mission before the sultan’s throne. The sultan considered the event as a sign of God’s favor, to whichMichael had significantly contributed through his prayers. In spite of some inconsistencies with re-spect to the details mentioned by Choniates, the “nephew” might be identified with Theodore Man-gaphas. For Kılıc Arslan II’s contacts with Michael the Syrian see Dorothea Weltecke, Die “Beschrei-bung der Zeiten” von Mor Michael dem Großen (1126–1199): Eine Studie zu ihrem historischenund historiographiegeschichtlichen Kontext, Corpus Scriptorum Christianorum Orientalium, Sub-sidia 110 (Louvain, 2003), pp. 107–9.

145 Choniates, ed. van Dieten, pp. 420–22, 461–63, trans. Magoulias, pp. 231–32, 253–54. See alsoCheynet, Pouvoir, pp. 123–24, no. 169, and p. 130, no. 182.

631Defection


ate.146 While Mangaphas can be regarded as an exponent of the provincialaristocracy, aiming at political autonomy on a local scale, the two Alexioi weredescendants of obscure origin from Constantinople and Cilicia respectively.147 Theparticularity of their undertaking lies in the fact that they presented themselvesbefore both the Byzantine provincial population and the Turkish sovereigns withinsidiously obtained claims for legitimate rights to the imperial throne. “He [thefirst of the two Pseudo-Alexioi] embodied the hero of the drama so perfectly,”Choniates remarks, “and imitated the appearance of Emperor Alexios so skill-fully that he even dressed his blond hair in the same way as he did and simu-lated the lisp of the imperial child’s tongue.”148 Apart from creating local strong-holds in the Meander valley or the border regions around Ankyra,149 it remainsunclear whether the two pretenders were actually trying to seize power in Con-stantinople as well. At least there were some court officials who were attractedby the rebels.150

Be that as it may, Choniates judges these defectors quite differently from theaforementioned categories of apostates. They are not blamed for their disloyaltyto the ruling dynasty or the central government, for it was the excessive crueltyof Andronikos I and the incapacity of the Angeloi emperors that had provokedthe disintegration of the empire. What discredits them is their malicious misbe-havior against their fellow countrymen and coreligionists. Outbursts of crueltyduring their raids on Byzantine territory make them even worse than their Turk-ish allies. With respect to Theodore Mangaphas, Choniates talks about the un-restrained burning and pillaging that destroyed the livelihood of the peasantry inthe Meander valley. The rebel insulted Turks who took pity on Christian people,delivered Christian captives into the hands of barbarians, and even allowed thefamous church of the Archangel Michael in Chonai to be burned to the groundby Turkish raiders.151 It is no longer the act of defection itself that is criticizedbut rather the behavior ensuing from it. With his actions, this type of apostategoes against basic principles of what in Byzantine eyes constitute divinely or-dained harmony and social order, that is, philanthropy and piety, and, therefore,he has severed his ties with the Christian-Roman moral sphere.152

146 Choniates, ed. van Dieten, p. 420, lines 24–26, trans. Magoulias, p. 231: πρFς τF <Ικ νι�ν Jπεισικα3 τ&+ω γ!ρ�ντι M�θε3ς σ�υλτ.ν . . . 4ς υ8Fς �ληθινFς τ�&υ α�τ�κρ�τ�ρ�ς Μαν�υ�λ; Choniates, ed.van Dieten, p. 421, lines 44–48, trans. Magoulias, p. 232: Τ&+ η συ�ν&+ η δ< Dµως τ�&υ νεαν��υ 1κλιπαρ�σειQ Περσ�ρ�ης καθυπενδ�Oς . . . συνε�#ρησε δ< �&�ν Dµως δι. σ�υλτανικ�&υ γρ�µµατ�ς, D �ασιν �8Τ�&υρκ�ι µ�υσ��ρι�ν, Dσ�υς Eν σ��η τ&ων BπF τ�ν 1κε�ν�υ �ρ��ν Bπ�π�ι�σασθαι �δε&ως; Choni-ates, ed. van Dieten, p. 461, lines 16–18, trans. Magoulias, p. 253: τ &+ω τ&ης π λεως <Αγκ�ραςσατραπε��ντι πρ�σρυ&ηναι κα3 Bπ< α�τ�&υ πρ�σδε�θ&ηναι 4ς υ8Fν �ληθ&η τ�&υ α�τ�κρ�τ�ρ�ς Μαν�υ�λ.

147 Choniates, ed. van Dieten, pp. 420, 461, trans. Magoulias, pp. 231, 253.148 Choniates, ed. van Dieten, p. 420, trans. Magoulias, p. 231.149 Choniates, ed. van Dieten, pp. 421, 461, trans. Magoulias, pp. 231, 253.150 Choniates, ed. van Dieten, p. 421, trans. Magoulias, p. 231.151 Choniates, ed. van Dieten, p. 400, trans. Magoulias, p. 220. See further Necipoglu, “Turks and

Byzantines” (above, n. 11), p. 260.152 For the conceptual framework see Demetrios J. Constantelos, Byzantine Philanthropy and So-

cial Welfare (New Brunswick, N.J., 1968); for philanthropy and piety in Byzantine imperial ideologysee Herbert Hunger, Elemente der byzantinischen Kaiseridee in den Arengen der Urkunden, Wienerbyzantinistische Studien 1 (Vienna, 1964).

632 Defection


The Seljuk Sultans of KonyaBegging for the Emperor’s Protection (1118–1260)

Let us now pick up the thread we left with Shahanshah’s reception by AlexiosI in 1116 and follow the sequence of Seljuk defectors who in the course of thetwelfth century sought refuge at the Byzantine imperial court. Interestingly, whileByzantine aristocrats were increasingly tempted to try their luck with powerfulAnatolian Turkish lords, Sultan Shahanshah, too, seems to have set an enticingprecedent for his successors. Michael the Syrian—the main source for the inter-nal situation of many Turkish emirates during the twelfth century—tells us thatthis development was mainly due to the civil strife among Kılıc Arslan I’s surviv-ing sons, Masüd, Ärab, and Togrıl Arslan, a conflict that also involved Masüd’sfather-in-law, Gümüstekin Ghazı, and other local lords in the Upper Euphratesregion.153 “All these things happened among the Turks,” Michael concludes, “whobecause of the anger they nurtured against each other fled to the Christians.”154

Ghazı’s seizure of Melitene from Togrıl Arslan’s and his mother’s hands (De-cember 10, 1124) triggered the fraternal discord. On the pretext of taking re-venge against his brother’s treacherous behavior toward the family, Ärab at-tacked Masüd with a huge military force, compelling him to flee to EmperorJohn. With the support of Byzantine money, Masüd, along with his allyGümüstekin Ghazı, took up the war against his brother Ärab. In the summer of1127, together with Turkish and Armenian allies, Ärab fought a series of fiercebattles against Ghazı and his sons and, after having suffered several defeats, wasforced to take flight to Byzantium. Michael the Syrian, mainly focusing on thesituation in eastern Asia Minor, devotes no more than a few words to the Seljuklord’s stay in Constantinople. That “Emperor John received Masüd joyfully”155

might hint at an official ceremony celebrating the guest’s arrival in the frame-work of imperial propaganda, as had been the case during the meeting with hisdeceased brother Shahanshah, but there is no way to make further inferences aboutthe ceremony’s form and contents. In Ärab’s case we cannot even be sure thathe was actually received at the imperial court, for Michael mentions only his flight“to the Greeks” and his subsequent death.156 In any event, the fact that the Byz-antine emperor granted sanctuary to three of Kılıc Arslan’s four sons within adecade of internal disorder demonstrates how important an instrument of polit-ical practice defection had become for both the refugees and the hosts. Perhapsthe frequent arrival of Seljuk opponents in Constantinople was an additional fac-tor inducing the sebastokrator Isaac to set off in the opposite direction three yearslater.

153 For the political situation see Chalandon, Les Comnène, pp. 42–46, 77–81; Cahen, Pre-OttomanTurkey, pp. 92–94; idem, Formation, pp. 15–20; and Turan, Türkiye, pp. 167–70.

154 Michael the Syrian 16.2, ed. Chabot, 3:224 (trans.), 4:609, col. A, lines 13–14 (Syriac text):“hallen kullhen hway baynat Turkaye d�a-b� -rugzhon d-äl h�d� ad� e b�a-K� rist�yane met�gawwasın-(h)waw.”

155 Michael the Syrian 16.2, ed. Chabot, 3:223 (trans.), 4:608, col. B, lines 36–37 (Syriac text):“Iwanı malka h�addıya©ıt� qabbel l-Masüd.”

156 Michael the Syrian 16.2, ed. Chabot, 3:224 (trans.), 4:609, col. A, line 12 (Syriac text): “wa-¨raq lwat� Yunaye w-eb�ad� .”

633Defection


Almost three decades passed before another Seljuk ruler—this time SultanMas¨ud’s son and successor Kılıc Arslan II (1155–92)—showed up for a visit inthe Byzantine capital. The circumstances of his flight resemble the state of af-fairs prevailing in 1125–27 in Asia Minor, but with the difference that in about1160 the empire’s expansionist policy in the East had reached its apogee. Man-uel’s winter–spring campaign of 1158–59 to Cilicia and Antioch resulted in therecognition of his supremacy by Reynald of Châtillon, King Baldwin III of Jeru-salem, and the Armenian lord Thoros, as well as in a peace treaty with Nur al-Dın of Mosul and Aleppo.157 Since September 1158 Baldwin III had been mar-ried to Manuel’s niece Theodora. The negotiations over Manuel’s marriage withConstance of Antioch’s daughter Maria were well under way.158 In short, the em-peror had managed to establish a network of alliances with the most powerfullordships on Anatolia’s southeastern fringe, and his position was impressivelystrong when he got involved in Seljuk affairs. Fortunately, in this case the avail-able source material is relatively rich and detailed, so that we get a rather clearand manifold insight into the political setting, ideological dimensions, and con-temporary perceptions of Kılıc Arslan’s reception at the imperial court.

The most thoughtful observer, Niketas Choniates, presents the sultan’s arrivalas an event of major significance in the framework of a detailed report onByzantine-Seljuk affairs covering the period from Sultan Mas¨ud’s death in 1155until the eve of the battle of Myriokephalon in 1176.159 After referring to thedivisions of Anatolian territories among Seljuk and Danismend princes ensuingfrom the death of Gümüstekin Ghazı’s son and successor Muh�ammad (ca. 1134–ca. 1141), the author mainly dwells on the conflict between Kılıc Arslan andMuh�ammad’s brother Yagı-basan (<Ιαγ�υπασ�ν; d. 1164), who is described aslord of Amaseia, Ankyra, and the province of Cappadocia.160 Emperor Manuelappears as a cunning diplomat striving for the downfall of both potentates andinstigating them through secret messages to wage war against each other. As aresult of this conflict the sultan undertook his trip to Constantinople.161 Despitethe peace treaty that was concluded during his stay and the enormous amountsof money and gifts he received from the emperor, the sultan turned out to be atraitor in that he continued to pursue his expansionist plans against his relatives

157 For the political background see Chalandon, Les Comnène, pp. 442–55; Lilie, Byzantium andthe Crusader States, pp. 176–83; Magdalino, Manuel Komnenos, pp. 70–71; and Necipoglu, “Turksand Byzantines,” pp. 261–62. For the diplomatic contacts and agreements concluded at this occa-sion see Dölger and Wirth, Regesten, nos. 1428–32 (dating the treaty with Nur al-Dın to May 1159)and nos. 1435–36.

158 Lilie, Byzantium and the Crusader States, pp. 175–76, 184–87; Magdalino, Manuel Komnenos,pp. 69–70, 72.

159 Choniates, ed. van Dieten, pp. 116–25, trans. Magoulias, pp. 66–71.160 Choniates, ed. van Dieten, pp. 116, 117–18, trans. Magoulias, pp. 66–67. For the details see

Chalandon, Les Comnène, pp. 431–35, 456–66; Cahen, Pre-Ottoman Turkey, pp. 100–101; idem,Formation, pp. 21–25; Turan, Türkiye, pp. 197–201; and Magdalino, Manuel Komnenos, pp. 76–78.

161 Choniates, ed. van Dieten, p. 118, lines 29–31, trans. Magoulias, p. 67: κα3 πρFς καιρFν τ.Dπλα καταθ!µεν�ι, Q µ2ν 1π!µενε κατ. �#ραν, Q δ2 σ�υλτ.ν τ&+ω �ασιλε&ι πρ σεισιν.

634 Defection


and the Danismends. He managed to extend his power significantly, expellingmighty lords such as Yagı-basan, Muh�ammad’s son Dhu l-Nun (∆αδ��νης), andhis own brother Shahanshah and seizing important cities in eastern Asia Minorsuch as Amaseia (1171), Sebasteia (1174), and Melitene (1177).162 Very soon, healso neglected the commitments made in his treaty with the empire, so that theByzantine-Seljuk border region underwent a new period of incessant raids andcounterattacks.163

Interestingly, Choniates, apart from a faint allusion to the sultanate’s internalinstability, does not explain why the sultan actually decided to go to Constanti-nople. Nor does the second main witness, John Kinnamos, who restricts himselfto the laconic observation that “the sultan came to Byzantium as a defector inorder to petition the emperor regarding matters advantageous to him.”164 OnlyMichael the Syrian explicitly refers to a conspiracy of Yagı-basan and other emirsin favor of the sultan’s brother Shahanshah, so that the sultan was compelled toseek the emperor’s backing.165

In accordance with the event’s highly official character, the narratives on KılıcArslan’s visit are structured along the sequence of publicly performed ceremo-nial acts projecting various ideological and propagandistic aspects, such as bothrulers’ status and honor, the newly established hierarchical relationship betweenthem, gestures of devotion and benevolence, and other current practices of Byz-antine diplomacy.166 We get additional insights via a panegyric composed by thecourt rhetorician Euthymios Malakes on the occasion of Manuel’s triumphal en-trance in 1162,167 an intriguing text that embeds the Seljuk sultan’s arrival into

162 For the details see Cahen, Pre-Ottoman Turkey, pp. 101–4; idem, Formation, pp. 26–32; andTuran, Türkiye, pp. 202–5.

163 Choniates, ed. van Dieten, pp. 121–25, esp. p. 121, lines 27–28, trans. Magoulias, pp. 68–71:Q δ2 �πατε@ν Wν κα3 µηδ2 τ� 1στιν Dλως �λ�θεια 1γνωκ#ς π�τε τ&ων τε συνθηκ&ων 1πελ�θετ�; Choni-ates, ed. van Dieten, p. 123, lines 68–69, trans. Magoulias, p. 70: τ τε δ2 π�λυδ�ναµ�ς γεγ�ν@ς τFπρFς τFν �ασιλ!α α6δ&ηµ�ν �πε��σατ�.

164 Kinnamos 5.3, ed. Meineke, pp. 204, line 22–205, line 1, trans. Brand, p. 156: Κλιτ'ιεσθλ.ν Qσ�υλτ.ν 1ς Βυ'�ντι�ν α�τ µ�λ�ς &;λθε περ3 τ&ων α�τ&+ω συµ� ρων �ασιλ!ως δεησ µεν�ς. The trans-lation of Brand, Deeds, p. 156, renders α�τ µ�λ�ς as “voluntarily,” but since the term is constantlyused for persons who shift to the enemy’s side, it is clear that the Byzantine authors use the term inits modern sense: “defector.”

165 Because of some missing folios in the manuscript, Chabot’s edition completes the passage usingBar Hebraeus’s chronicle: Michael the Syrian 18.8, ed. Chabot, 3:319 (trans.); compare with Bar He-braeus, Chronicon, p. 328, lines 15–26. The original text is more accurately preserved in the Arme-nian version of Michael’s chronicle partly published as Extrait de la chronique de Michel le Syrien,Recueil des Historiens de Croisades, Documents Arméniens 1 (Paris, 1869), pp. 309–409, at pp. 355–56.

166 Kazhdan, “Byzantine Diplomacy” (above, n. 28). For a useful summary of Byzantine theoreti-cal concepts of the empire’s relationship with the outside world see Evangelos Chrysos, “Τ� Βυ'�ντι�και η διεθν�ς κ�ινων�α τ�υ µεσα�ωνα” [Byzantium and the medieval international community], inByzantium as Oecumene, ed. idem, Institute for Byzantine Research, International Symposia 16 (Ath-ens, 2005), pp. 59–78.

167 The text is edited by Anastasios Papadopoulos-Kerameus, Noctes Petropolitanae (Saint Peters-burg, 1913; repr., Leipzig, 1976), pp. 162–87: Τ�&υ α�τ�&υ ε6ς τFν α�τ�κρ�τ�ρα κ�ρι�ν Μαν�υ�λτFν Κ�µνην ν, 1κ�ωνηθε3ς Dτε ε6σ&ηλθεν ε6ς Κωνσταντιν��π�λιν Q σ�υλτ&αν�ς πρ�σελθ@ν α�τ&+ω.

635Defection


the conceptual matrix of classical imperial encomia.168 On the basis of this ma-terial the main steps of the ceremonial procedure can be summarized as follows:1. Preliminary negotiations conducted by the sultan’s chancellor Christopher, which re-

sulted in an oath sworn by the emperor on the sultan’s safe-conduct (Michael the Syr-ian, Armenian version, p. 355)

2. Planned triumphal entrance with the emperor and the sultan standing together on theimperial chariot—the plan failed because of an earthquake (Kinnamos, pp. 206, line12–207, line 2; Choniates, pp. 118, line 38–119, line 54)

3. Solemn audience granted to the sultan in the imperial palace (Kinnamos, pp. 205, line5–206, line 11):• self-presentation of the imperial family in a prokypsis ceremony• the sultan’s prostration before the emperor (Malakes, p. 165, lines 15–23)• the sultan’s taking a seat on a lower throne beside the emperor (Malakes, p. 167,

lines 18–20; Kinnamos, p. 206, lines 9–10)• the sultan’s adoption as the emperor’s spiritual son (Malakes, p. 167, lines 9–12;

Kinnamos, p. 208, lines 7–8; Choniates, p. 123, lines 76–78)4. Opulent meals, gifts, and every sort of kind treatment (Kinnamos, p. 207, lines 9–11;

Michael the Syrian, Armenian version, p. 355)5. Games in the hippodrome (Kinnamos, p. 207, lines 11–14; Choniates, pp. 119, line

55–120, line 81)6. Conclusion of a peace treaty (Kinnamos, pp. 207, line 15–208, line 5)7. The sultan’s reconciliation with emissaries of other Eastern potentates (Kinnamos,

p. 208, lines 6–16)8. Farewell reception, on the occasion of which the sultan received huge amounts of

money and gifts (Choniates, pp. 120, line 90–121, line 22)A close examination of the three versions written from the Constantinopoli-

tan perspective reveals that the event underwent a successive reinterpretation, dis-torting the idealized image of imperial rhetoric into a harsh criticism of the em-peror’s attitudes and decisions. Euthymios Malakes’s panegyric shows manysimilarities with Attaleiates’s presentation of Nikephoros III’s encounter with theKutlumus brothers. In order to depict Kılıc Arslan’s arrival as a major success ofManuel’s policy, Malakes exalts the sultan’s royal prestige, placing him in onelineage with the ancient kings of Persia Darius and Xerxes.169 Accordingly,Malakes especially highlights the symbolic value of the sultan’s prostration as asign of his voluntary humiliation and reduction to the status of the emperor’sfaithful servant.170 In return, Manuel is described as having excelled even Alex-ander the Great and King Solomon, for the emperor not only defeated his foe on

168 For rhetorical rules and models of Byzantine imperial panegyrics see now Dimiter Angelov,Imperial Ideology and Political Thought in Byzantium, 1204–1330 (Cambridge, Eng., 2007), pp. 51–64.

169 Malakes, ed. Papadopoulos-Kerameus, p. 165, lines 15–16: �ρ�ηγFς �&Bτ ς 1στι Περσ�δ�ς,µ&αλλ�ν δ2 τ&ων �ρ�ηγ��ντων 1ν α�τ&+ η τF κε��λαι�ν . . . �&Bτ�ς 1κε&ιν�ς Q ∆αρε��υ κα3 ^!ρ��υ τ&ωντ&ης Περσ�δ�ς α��ηµ�των �π γ�ν�ς.

170 Malakes, ed. Papadopoulos-Kerameus, p. 167, lines 9–10: Q µ2ν γ.ρ δ�&υλ�ν >αυτFν καλε&ι κα3τ. τ&ων δ��λων θ!λει π�ιε&ιν κα3 πρ�σκυν�σας ε&6τα παρ�σταται.

636 Defection


the battlefield but adopted the sultan, thus including him in the circle of his rel-atives and granting him a seat below his own.171

Those details refer to the festive setting, which publicly visualized the sultan’shierarchical elevation resulting from his symbolic act of deference to the emper-or’s supremacy. In this way, Kılıc Arslan, just like Shahanshah in 1116, was cer-emoniously acknowledged as a member of the imperial entourage, but with thedifference that as the emperor’s spiritual son he had gained one of the most hon-orable positions a foreign sovereign could obtain in Byzantium.172 Kinnamos in-forms us that the sultan’s reception formed part of a prokypsis ceremony, a newcourt ritual introduced by Manuel I, displaying the imperial family in an es-pecially magnificent public appearance on a lit platform accompanied by musicand encomiastic poems.173 The sharp contrast with the 1116 meeting, where theemperor and the sultan expressed their proximity by riding side by side, revealsthe high degree of refinement that Comnenian court ceremonies regarding rep-resentatives of the Seljuk sultanate had reached since the early twelfth century.A comparison with the nearly contemporary meetings with King Baldwin III ofJerusalem in Manuel’s camp in Cilicia (early 1159) and with King Amalric inConstantinople (1171)174 clearly demonstrates that the lower throne on whichthe sultan came to sit was a common feature of reception ceremonies for foreignrulers, symbolizing the emperor’s ascendancy within the universal hierarchical sys-tem of political authority.

In contrast, Kılıc Arslan’s adoption as Manuel’s spiritual son, with respect tothe Muslim orbit, constitutes a remarkable novelty in the sphere of Byzantinecourt rituals. In the ninth and tenth centuries this distinction used to be con-ferred upon the Bulgarian king, while in Alexios I’s reign there are some in-stances of adoptions of crusader lords such as Godfrey of Bouillon.175 The onlyknown precedent as far as Byzantium’s eastern neighbors are concerned is theadoption of the Persian king Khusrau II Parwiz by Emperor Maurice in 590.176

171 Malakes, ed. Papadopoulos-Kerameus, p. 167, lines 10–12: Q δ2 κα3 πα&ιδα π�ιε&ιται κα3 τ�&ις�6κε��ις 1γγρ��ει κα3 ��ι�&ι κα3 τ&ης παρ< α�τ&+ω καθ!δρας, &=ς κα3 µεγ�λ�ι �ασιλε&ις 1πεθ�µησαν;ibid., p. 167, lines 18–20: τFν >αυτ&ων �ασιλ!α λαµπρFν λαµπρ&ως 1νθ�δε πρ�θρ�νι' µεν�ν κα3 τFνµ!γαν τFν Περσ&ων �ρ�ηγ!την α�τ&+ω καλ&ως Bπ�π�δι' µεν�ν.

172 For this usage see Franz Dölger, “Die Familie der Könige im Mittelalter,” Historische Zeitschrift60 (1940), 397–420, repr. in idem, Byzanz und die europäische Staatenwelt: Ausgewählte Aufsätzeund Vorträge (Darmstadt, 1964), pp. 34–69, esp. pp. 41–42; Evangelos Chrysos, “Byzantine Diplo-macy, A.D. 300–800: Means and Ends,” in Byzantine Diplomacy (above, n. 28), pp. 25–39, at pp. 25–28; and Jasmin Moysidou, Τ� Βυ'�ντι� και �ι � ρει�ι γε�τ�ν!ς τ�υ τ�ν 10� αι#να [Byzantium andits northern neighbors during the tenth century], Historical Monographs 15 (Athens, 1995), pp. 51–63.

173 For prokypsis see Ernst H. Kantorowicz, “<Ανατ�λ� τ�&υ δεσπ τ�υ,” in Das byzantinische Herr-scherbild, ed. Herbert Hunger, Wege der Forschung 341 (Darmstadt, 1975), pp. 258–80; and An-gelov, Imperial Ideology, pp. 41–42 (with further bibliographical references).

174 For Baldwin III see Kinnamos 4.20, ed. Meineke, p. 185, lines 18–19, trans. Brand, p. 141:Xδραν τ! τινα �θαµαλ�ν καθι'&ησαι παρ!θετ�. For King Amalric see William of Tyre 18.24, ed. Huy-gens, p. 846, lines 23–24: “secus eum in sede honesta, humiliore tamen, locatus est.”

175 Dölger, “Familie der Könige,” pp. 48–49.176 Michael Whitby, The Emperor Maurice and His Historian: Theophylact Simocatta on Persian

and Balkan Warfare (Oxford, 1988), pp. 287–89.

637Defection


This late-antique model case of a royal refugee seeking the emperor’s protectionmight have inspired Manuel’s court to revive this practice.

The possibilities and techniques of Byzantine court propaganda are clearly il-lustrated by the panegyric of Malakes, who brings the Persian past and Seljuk-Turkish present into an unbroken line of continuity by drawing parallels be-tween the Seljuk sultan and the ancient Persian kings. Hence the sultan of Konya,from an ideological point of view, becomes the emperor’s most extolled partneramong the rulers of the Muslim world. Accordingly, Kinnamos characterizes thesultan’s arrival as “something sublime and wonderfully great, which, as far as Iknow, never happened to the Romans before.”177

That a barbarian ruler was granted the privilege of mounting the imperial char-iot along with the emperor apparently caused anger and indignation among thepeople of Constantinople. Kinnamos explicitly refers to the resistance of Patri-arch Luke Chrysoberges (1157–70), who declared that “impious people shouldnot be allowed to pass through divine objects and sacred decorations.”178 Thiswas indeed a well-established principle of Byzantine court ceremonies, repeat-edly confirmed by older sources such as the tenth-century treatise De adminis-trando imperio of Constantine VII.179 Hence Manuel, in his attempt to presenthimself publicly together with his newly adopted spiritual son, acted contrary tocustomary attitudes and religious sensibilities. The narratives argue that the vio-lation of this taboo provoked divine interference, causing a strong earthquake inConstantinople and thwarting the whole triumph. Kinnamos and Choniates agreethat the earthquake was a conspicuous sign of God’s wrath. Later events, culmi-nating in the emperor’s ignominious defeat in 1176, prove that the earthquakewas in fact a bad omen for imminent disasters.180 In this way the triumphalentrance—one of the most momentous state ceremonies in Byzantium181—turnedinto a symbol of the emperor’s personal failure; in other words, because of theemperor’s deviation from traditional rules and the ensuing lack of unanimity, rightorder was seriously disturbed. As a result historical memory inverted a “good”ritual into a “bad” one, thereby compromising both Manuel’s political purposesand the ideological context of the sultan’s visit to the Byzantine capital.182

Another important aspect of the sultan’s stay in Constantinople consisted ofamusements, gifts, and games. Anna Komnene interpreted these features as ef-

177 Kinnamos 5.3, ed. Meineke, p. 205, lines 1–2, trans. Brand, p. 156: πρ&αγµα Bψηλ ν τε κα3δαιµ�ν�ως Bπ!ρ�γκ�ν κα3 Dσα 1µ2 ε6δ!ναι �Cπ�τε Jλλ�τε /Ρωµα��ις ε�τυ�ηθ2ν πρ τερ�ν.

178 Kinnamos 5.3, ed. Meineke, p. 206, lines 15–18, trans. Brand, p. 157.179 See, for instance, Constantine’s famous prohibitive statement on the sending of imperial dia-

dems or state robes to barbarian recipients: Constantine Porphyrogenitus, De administrando impe-rio, ed. György Moravcsik, trans. R. J. H. Jenkins, Corpus Fontium Historiae Byzantinae 1 (Wash-ington, D.C., 1967), 13, pp. 66–68.

180 Kinnamos 5.3, ed. Meineke, pp. 206–7, trans. Brand, pp. 157–58; Choniates, ed. van Dieten,p. 119, lines 43–50, trans. Magoulias, p. 67: θεFς δ< ;κ�ρωσε τ. τ&ης =µ!ρας 1κε�νης λαµπρ� . . .µην�ειν τF θε&ι�ν κα3 µηδ< Dλως �ν!�εσθαι πρ�κ�ψειν Dλως ε6ς θρ�αµ��ν µ� θε�σε�ε�ας Jνδραµετεσ�ηκ τα.

181 McCormick, Eternal Victory.182 My interpretation relies on Buc, Dangers of Ritual, pp. 70, 72 (with similar examples from the

early-medieval West).

638 Defection


fective instruments through which Alexios played out his ruse against his guestAbu l-Qasim.183 Conversely, William of Tyre’s report on King Amalric’s visit in1171 highlights the positive effects of generous gifts honoring the king and hiscompanions, the display of relics stored in the imperial treasuries, and varioussorts of spectacles.184 Interestingly, in Choniates’s account these elements were usedwith a sarcastic intention. The unsuccessful “parachute jump” of an otherwiseunspecified “descendant of Hagar”—most probably a man of the sultan’sentourage—from a tower erected above the hippodrome’s starting line inducedthe silversmiths in the marketplaces to mock the sultan.185 What is more, Choni-ates reduces Manuel’s diplomatic strategy toward the sultan to absurdity by con-trasting the ritual of a farewell reception with a derisive comment allegedly ut-tered by the sultan at some later moment. All kinds of luxurious items and preciousobjects, piled up in a glamorous reception hall, formed the setting of the conclud-ing meeting between the two rulers, in which the generous emperor, displayinghis irresistible power over his enemies, is contrasted with the astonished and greedybarbarian, who promises to deliver the city of Sebasteia.186 Afterwards the au-thor turns the scene’s imagery into ridicule through ironic words put into the mouthof the sultan, who jeers at the empire’s strategy to buy off barbarian allies withhuge amounts of money: “The sultan ironically used to say to his intimates thatthe more injuries he inflicted on the Romans, the more treasures he received fromthe emperor: ‘for whoever is able to be superior,’ he says, ‘usually receives thegifts as well, so that his victories do not increase further,’ because festering dis-eases require more treatments for them to subside and to cease to spread.”187

In summary, Kılıc Arslan II’s flight to Constantinople appears, on the one hand,as the apogee of the Seljuk defection movement to the imperial court, which inthis case was willing to make hitherto inconceivable concessions to a non-Christianforeign ruler, and, on the other, as the culmination of Manuel’s misplaced policytoward the Seljuk Turks, which caused internal discord, violated basic principlesof Byzantine court rituals, and turned out to be to the advantage of Kılıc Ar-slan’s expansionist plans without having any positive results for the empire.

In the course of the subsequent rivalries among the Turkish lords in Anatolia,Constantinople continued to offer shelter to overthrown princes, such as the lastDanismend ruler Dhu l-Nun, who, after succeeding his assassinated nephew Isma¨ılin the lordship of Sebasteia in early 1173, had been expelled by Sultan Kılıc Ar-slan in the summer of 1175, and the sultan’s brother Shahanshah, who lost theremainder of his political influence after the failure of a large coalition of thesultan’s enemies under the leadership of Nur al-Dın in 1172.188 Apart from a Byz-

183 See above, p. 612.184 William of Tyre 20.23, ed. Huygens, pp. 944–45.185 Choniates, ed. van Dieten, pp. 119–20, trans. Magoulias, pp. 67–68.186 Choniates, ed. van Dieten, pp. 120–21, trans. Magoulias, pp. 68–69.187 Choniates, ed. van Dieten, pp. 124–25, trans. Magoulias, pp. 70–71.188 Choniates, ed. van Dieten, p. 123, trans. Magoulias, p. 70; Michael the Syrian 20.1, ed. Chabot,

3:357 (trans.), 4:708, col. A, lines 34–35 (Syriac text), and 3:368 (trans.), 4:714, col. B, lines 43–44(Syriac text). In addition, see Cahen, Pre-Ottoman Turkey, pp. 102–3; idem, Formation, pp. 29–30;and Turan, Türkiye, pp. 203–5.

639Defection


antine attempt to reconquer Neokaisareia on behalf of Dhu l-Nun in 1176,189

we do not know anything about the imperial government’s attitudes toward thesepeople and about the role they played henceforth at the Byzantine court. It comesas no surprise that, during the critical years of the central government’s decayunder Andronikos I and the Angeloi emperors, Constantinople to a certain ex-tent lost its attractiveness for Seljuk defectors. While, as discussed above, severalByzantine local rebels were eager to secure their Turkish neighbors’ support, weknow of only one anonymous dignitary from Konya who seemingly fled to IsaacII Angelus shortly after the conquest of Konya by Emperor Frederick I’s Germancrusaders in May 1190.190

The series of twelfth-century Seljuk sultans who came as refugees to Constan-tinople ends with Ghiyath al-Dın Kaykhusrau.191 One of the ten sons of KılıcArslan II, he had obtained the city of Sozopolis192 as a result of the old sultan’ssuccession arrangements, and, after his father’s death in 1192, he succeeded himon the throne of Konya, while the greatest part of the Seljuk Empire remained inthe hands of his elder brother Qut�b al-Dın Malikshah (d. 1195). A few yearslater Kaykhusrau was expelled by another brother, Rukn al-Dın Sulaymanshah(1196–1204), thus being forced to lead the life of a homeless prince wanderingfrom one potentate to another.193

Kaykhusrau’s stay in the Byzantine capital, which lasted until Alexios III’s over-throw in July 1203, was in many respects decisive for the subsequent develop-ments in Byzantine-Seljuk relations. Notably, Emperor Alexios III both adoptedand baptized the dethroned sultan in the years of his stay in the imperial pal-ace,194 so that his ideological and spiritual ties with the Byzantine court elite andthe ruling dynasty became even more intertwined than his father’s had been. Inaddition, Kaykhusrau married the daughter of Manuel Maurozomes, a high-ranking aristocrat,195 with whom he presumably was closely connected by a net-

189 Michael the Syrian 20.4, ed. Chabot, 3:368–69 (trans.), 4:714–15 (Syriac text); Turan, Türkiye,p. 207.

190 Choniates, ed. van Dieten, p. 414, trans. Magoulias, p. 228.191 For the event and its political background see Cahen, Pre-Ottoman Turkey, p. 116; Alexis

G. C. Savvides, Byzantium in the Near East: Its Relations with the Seljuk Sultanate of Rum in AsiaMinor, the Armenians of Cilicia and the Mongols, A.D. 1192–1237, Byzantine Texts and Studies 17(Thessalonica, 1981), p. 82; and T. Baykara, I. Gıyaseddin Keyhusrev (1164–1211) Gazi-Sehit, TürkTarih Kurumu Yayınları 24. Dizi, no. 20 (Ankara, 1997), pp. 22–24. The various interpretations ofKaykhusrau’s defection in Byzantine and Oriental narratives have been recently analyzed at length inAlexander Beihammer, “Der Vierte Kreuzzug und die Eroberung Konstantinopels im Spiegel orien-talischer Quellen,” in The Fourth Crusade Revisited: Atti della conferenza internazionale nell’ottavocentenario della IV Crociata, 1204–2004, ed. Piarantonio Piatti, Pontificio Comitato di ScienzeStoriche, Atti e Documenti 25 (Vatican City, 2008), pp. 244–74, at pp. 260–66.

192 For Sozopolis (modern Uluborlu) in Phrygia, which was not incorporated into the Seljuk sul-tanate until after 1180, see Klaus Belke and Norbert Mersich, Phrygien und Pisidien, Tabula ImperiiByzantini 7 (Vienna, 1990), pp. 387–88.

193 Cahen, Pre-Ottoman Turkey, pp. 111, 114–15; idem, Formation, pp. 39–42; Turan, Türkiye,pp. 237–41.

194 Akropolites 8, ed. Heisenberg, 1:14, lines 12–14, trans. Macrides, p. 124: �υγ.ς παρ. τ�νΚωνσταντ�ν�υ +T�ετ� κα3 τ&+ω �ασιλε&ι Bπ�δ!δεκται <Αλε�� +ω, �απτ�'ετα� τε παρ< α�τ�&υ κα3 υ8�θετε&ιται.

195 Choniates, ed. van Dieten, p. 626, trans. Magoulias, p. 343; Ibn al-Athır, Al-Kamil fı l-Tarıkh,9 vols. (Beirut, 1994), 7:470, lines 17–18; trans. Donald S. Richards, The Chronicle of Ibn al-Athır

640 Defection


work of common interests. Once Alexios III decided to leave Constantinople un-der the threat of the Fourth Crusade’s military forces, the Seljuk prince fled alongwith his father-in-law to a fortress on the latter’s landed estates.196 In March 1205,only a few months after the conquest of Constantinople and the ensuing down-fall of the Byzantine imperial government, Kaykhusrau managed to regain thethrone of Konya, replacing his nephew Kılıc Arslan III.197

On account of his bonds of spiritual kinship, marriage, and personal friend-ship with leading figures of the Angeloi regime, the sultan became immediatelyinvolved in the conflicts resulting from Theodore Laskaris’s attempts to createhis “Byzantine successor state” in Nicaea and to lay claim to the imperial title.In 1205 Kaykhusrau supported the plans of his father-in-law, Manuel Maurozo-mes, to get rid of his rival Theodore, launching attacks in the region of the Me-ander valley.198 Several years later, at the instigation of the ex-emperor Alexios,who after an adventurous odyssey had been offered shelter in Konya, Kaykhus-rau started a new war against the emperor of Nicaea, which came to an end in1211 with the sultan’s death on the battlefield near Pisidian Antioch (modernAlasehir).199

Most notably, for the first time we are able to derive abundant informationfrom Muslim sources, namely, the universal chronicle of Ibn al-Athır and Ibn Bı-bı’s Seljuk chronicle, so in this case we can judge the sultan’s stay at the imperialcourt from the Seljuk perspective as well. This is all the more important insofaras the details the Byzantine authors provide are quite limited and contradictory.Niketas Choniates basically rejects any successful outcome of Kaykhusrau’s visit,whereas George Akropolites merely notes the refugee’s adoption and baptism.200

Instead, Ibn al-Athır and especially Ibn Bıbı highlight the respectful reception theByzantine emperor granted to his guest, primarily dwelling on ritual and formalaspects of the encounter and the ensuing relationship between the two rulers.201

In the highly idealized and rhetorically elaborated Persian narrative of the Seljuk

for the Crusading Period from al-Kamil fı©l-ta©rıkh, 3: The Years 589–629/1193–1231: The Ayyubidsafter Saladin and the Mongol Menace (Aldershot, Eng., 2008), p. 83; Aksarayli, Musamarat al-akhbar (Mogollar zamanında Türkiye Selçukluları Tarihi), ed. Osman Turan, Türk Tarih KurumuYayınlarından, III. Seri, 1 (Ankara, 1944), p. 31. See further Necipoglu, “Turks and Byzantines” (above,n. 11), p. 261.

196 Ibn al-Athır, 7:470, lines 19–20, trans. Richards, p. 83.197 Ibn al-Athır, 7:470, lines 12–13, trans. Richards, p. 83 (dates to Rajab a.h. 601 [February 22–

March 23, 1205]); Ibn Bıbı, ed. Houtsma, pp. 26–29, trans. Duda, pp. 38–41; Aksarayli, ed. Turan,p. 32.

198 Choniates, ed. van Dieten, p. 626, trans. Magoulias, p. 343.199 Akropolites 10, ed. Heisenberg, 1:16–17, trans. Macrides, pp. 131–32. For Manuel Maurozo-

mes and the battle of Alasehir see now Ilias Giarenis, Η συγκρ τηση και η εδρα�ωση τηςΑυτ�κρατ�ρ�ας της Ν�καιας: � αυτ�κρ�τ�ρας Θε δωρ�ς Α& Κ�µνην ς Λ�σκαρις [Establishmentand consolidation of the Empire of Nicaea: The Emperor Theodoros I Komnenos Laskaris], Institutefor Byzantine Research 12 (Athens, 2008), pp. 62–82, 122–28.

200 Choniates, ed. van Dieten, p. 522, lines 7–8, trans. Magoulias, pp. 286–87: Καϊ��σρ ης δ2�ρα�!α Jττα �ιλ��ρ�νηθε3ς κα3 τ&ων 1ν ν&+ω ��δ< Dσ�ν ε6πε&ιν 1νδε!στερα. For Akropolites see above,n. 194.

201 Ibn al-Athır, 7:470, lines 17–18; Ibn Bıbı, ed. Houtsma, pp. 14–18, trans. Duda, pp. 27–31.

641Defection


chronicle, Kaykhusrau is portrayed as an exiled, but still legitimate, sovereignwho received recognition and honor as such from his brothers in Elbistan andMelitene, from King Leo I of Lesser Armenia, and from many Muslim lords innorthern Syria.202 Consequently, his arrival in Constantinople is described as thecoming, not of a refugee, but of a guest of the highest rank and significance: “Thebasileus ( fasilyus) of that time [Alexios III] deemed the sultan’s arrival a greatbenefit and considered his share or even his independence in his kingdom neces-sary. When they met they used to sit together on the throne and to show eachother friendship and kindness.”203

More specifically, the emperor’s benevolent attitude is explained by the sul-tan’s self-presentation as the immediate descendant of the great Seljuk sultans ofBaghdad and as heir to their political traditions and claims: “My forefathers hadconquered the world from East to West, and your forefathers used to send trib-ute and taxes to their treasuries. And you walked with me on the same path.”204

Hence, Ibn Bıbı constructs a bridge linking the ideological tradition of the Seljuk-Persian sultanate and the Byzantine imperial sphere, both of which came to forma part of Kaykhusrau’s twofold royal identity. He appears as the heir of Alp Ar-slan and Malik Shah and, at the same time, as a sovereign worthy of sharing thethrone with the Byzantine emperor.

After 1204 the Byzantine imperial concept underwent a deep crisis with theestablishment of a Latin empire in Constantinople and the emergence of numer-ous claimants to legitimate succession within the remnants of the Byzantine state.This constellation allowed the Seljuk sultan to take his share in the imperial her-itage. Greek documents originating from the sultanate’s diplomatic correspon-dence with the Frankish king of Cyprus and the Venetian podestà of Constanti-nople attest to the usage of Byzantine imperial honorifics for Kaykhusrau’s son¨Izz al-Dın Kaykawus (1211–20).205

From the Byzantine perspective, too, things had radically changed with theevents of 1204. As the old center had ceased to exist, fleeing to collaborate withthe sultan was no longer considered a seditious act deserving of condemnation.What mattered was, above all, survival amidst the fierce struggles with internaland external opponents. Theodore Laskaris took refuge with the sultan shortlybefore he seized power in Nicaea, and Theodore Angelos, the future ruler of Epi-rus and Thessalonica, probably did the same.206 Manuel Maurozomes, after his

202 Ibn Bıbı, ed. Houtsma, pp. 9–13, trans. Duda, pp. 23–27.203 Ibn Bıbı, ed. Houtsma, p. 14, lines 2–4, trans. Duda, p. 27.204 Ibn Bıbı, ed. Houtsma, p. 15, lines 1–5, trans. Duda, p. 28.205 For details see M. E. Martin, “The Venetian-Seljuk Treaty of 1220,” English Historical Review

95 (1980), 321–30; Alexander Beihammer, Griechische Briefe und Urkunden aus dem Zypern derKreuzfahrerzeit: Die Formularsammlung eines königlichen Sekretärs im Vaticanus Palatinus 367, Textsand Studies on the History of Cyprus 57 (Nicosia, 2007), p. 170; and idem, “Multilingual Literacyat the Lusignan Court: The Cypriot Royal Chancery and Its Byzantine Heritage,” Byzantine and Mod-ern Greek Studies 35 (2011; forthcoming).

206 Akropolites 6, ed. Heisenberg, 1:11, lines 2–3, trans. Macrides, p. 118: κα3 γ.ρ 1ν τ&+ω µετα�Oκα3 περ3 τFν περσ�ρ�ην ��κετ� συν�θη τ��τ +ω τελ�&υντα, κα3 συµµα��αν πρ�σε�λη�ε. For possiblereferences to the same event in orations of Niketas Choniates and for the chronological sequence seeMacrides, p. 119 n. 7; for Theodore Angelos’s relations with the Turks see Macrides, p. 209 n. 9.

642 Defection


attempt to win the imperial crown had failed, paid homage to his son-in-law andbecame a loyal servant at the sultan’s court of Konya.207 Kaykhusrau’s godfatherand ex-emperor Alexios hoped to regain the imperial throne with the aid of thesultanate’s military forces.208 This altered situation is clearly illustrated in thechronicle of George Akropolites, who now describes the sultan as the exiled em-peror’s “friend” (συν�θης) and as the mighty protector of Alexios’s legitimateimperial rights: “And an embassy on the part of the sultan reached him [The-odore I] that announced the arrival of the emperor and father-in-law [AlexiosIII] and expressed the opinion that he [Theodore] illegally had usurped the sov-ereign rights of someone else. The emperor got excited about these words andhe was seized by rather great fear. For the sultan used the Emperor Alexios onlyas a pretext, in fact intending to attack and to plunder or to subjugate the wholestate of the Romans. Things were so to speak on the razor’s edge for the Em-peror Theodore. Thus he called together all his followers and checked if they wereon his side or on that of his father-in-law, the Emperor Alexios.”209

Likewise, Akropolites presents Michael Palaiologos’s flight to Sultan Ïzz al-Dın Kaykawus II (1246–61) in 1256 in a way that projects Michael’s claims tothe throne. Despite his defection Michael is still “well disposed toward the em-pire,”210 and the Turkish dignitaries in the sultan’s court consider him “worthyof being emperor,”211 thus putting a prediction of Michael’s future rise to powerinto the mouths of Muslim leaders. Highlighting the enemy’s admiration for thetalents of an outstanding Roman commander certainly is a widely used motif ofclassicizing Byzantine historiography. With respect to the historical situation inthe late 1250s, however, it also reflects the new level that Byzantine-Seljuk rela-tions had reached by that time in terms of both realpolitik and ideology.

All these aspects are embodied in the person of the last prominent defector tobe examined in this essay, Sultan Ïzz al-Dın Kaykawus II (1246–61), the first-born son of Ghiyath al-Dın Kaykhusraw II (1237–46).212 Because of the con-flicts with the Mongols in the period preceding Hülegü’s major onslaught against

207 Ibn Bıbı, ed. Houtsma, pp. 26, lines 1–4, and 29, line 4, trans. Duda, pp. 38, 41.208 Akropolites 8–9, ed. Heisenberg, 1:14–16, esp. p. 14, lines 8–10, trans. Macrides, pp. 124–30:

Q �ασιλεOς <Αλ!�ι�ς ��λεται πρFς τFν τ�&υ <Ικ�ν��υ σ�υλτ.ν �πελθε&ιν, Sν <Ιαθατ�νην `ν µα'�ν9&;ν γ.ρ τ��τ +ω συν�θης.

209 Akropolites 9, ed. Heisenberg, 1:15, trans. Macrides, p. 129.210 Akropolites 64, ed. Heisenberg, 1:134, lines 23–24: τF τ�&υ �ρ�ν�µατ�ς 1π�σταµαι �ιλ�ρ#µαι�ν.

Macrides, p. 312, translates: “I know him to be a friend of the Romans in his thinking.” My inter-pretation further specifies the meaning of the word philorhomaios in the sense of “being loyal to theRoman (i.e., Byzantine) state.” In the context of Michael’s defection to the sultan, this aspect obvi-ously was much more important for the imperial government than his general affinity to the Ro-mans, which for a Byzantine aristocrat seems more or less self-evident.

211 Akropolites 65, ed. Heisenberg, 1:136, line 27–1:137, line 1, trans. Macrides, p. 315: �8 µετ.τ�&υ περσ�ρ��υ µεγιστ&ανες τελ�&υντες . . . κα3 S �ησ� τις τ&ων παλαι&ων, τυρανν�δ�ς J�ι�ν :κριναν.For the use of Euripides’s phrase J�ι�ν τυρανν�δ�ς as an expression put into the mouth of enemiesrecognizing their opponent’s qualities see Prinzing, “Ein Mann τυρανν�δ�ς J�ι�ς” (above, n. 118),pp. 196–97 n. 46, and Macrides, p. 317 n. 5.

212 On this person see Cahen, Pre-Ottoman Turkey, pp. 271–79; idem, Formation, pp. 175–91;Turan, Türkiye, pp. 458–97; and Rustam Shukurov, “The Family of Ïzz al-Din Kay-Kawus in Byzan-tium” [in Russian], Vizantiiskii Vremenik 67 (2008), 89–116.

643Defection


the central Islamic lands (1258) and the internal struggle for power among theleading members of the Seljuk elite, Ïzz al-Dın II in the years 1257–61 strength-ened his precarious position by maintaining especially close ties with the Byzan-tine government. After his first overthrow following his army’s defeat in the bat-tle of Aksaray against the Mongols (October 1256), he escaped via Attaleia andLaodicea to Byzantine territory, where in January 1257 he was received by Em-peror Theodore II in Sardis.213 With the support of Byzantine auxiliary forcesoffered by the emperor in exchange for the city of Laodicea, Ïzz al-Dın in thespring of 1257 regained the throne of Konya and arrived at a new accommoda-tion with his younger brother Rukn al-Dın Kılıc Arslan IV (1248–65), providingfor shared rule over divided territories. The subsequent recognition of this statusby Hülegü, however, brought no long-term stability in the sultanate, and in 1261Ïzz al-Dın II fell victim to the intrigues of internal opponents combined with re-newed Mongol pressure.214 As a result, the sultan was compelled to flee a sec-ond time to the Byzantine court, from which he was never to return.215

While Ïzz al-Dın’s behavior at first glance may appear as just another link ina long chain of very similar patterns of action deeply embedded in the politicalculture of Anatolian Turkish principalities, a careful analysis of this sultan’s re-lations with Byzantium in the years 1257–61 reveals an unprecedented degree ofintensity and intimacy. First and foremost, Sultan Ïzz al-Dın was constantly sur-rounded by a strong Christian element within his entourage. His mother be-longed to a local Christian family and enjoyed an honorable position at the courtof Konya, which Pachymeres highlighted by referring to her as “the old moth-er.”216 Among the court dignitaries were influential relatives of hers, such as Kir

213 Ibn Bıbı, ed. Houtsma, pp. 288–90, trans. Duda, pp. 273–76 (mainly concentrating on the con-flicts within the sultanate); Aksarayli, ed. Turan, p. 49. The meeting with Theodore II is described indetail by Akropolites 69, ed. Heisenberg, 1:143–44, trans. Macrides, pp. 325–27, along with The-odori Scutariotae additamenta 39, 47, ed. Heisenberg, 1:292, 294–95. For the date see ibid. 47, 1:295,lines 3–6. For further details concerning the flight see Cahen, Pre-Ottoman Turkey, pp. 276–77; idem,Formation, pp. 186–87; Turan, Türkiye, p. 485 with n. 47; Langdon, “Byzantium’s Initial Encoun-ter” (above, n. 119), p. 134; and idem, “Twilight” (above, n. 119), pp. 192–93. For the battle ofAksaray see Turan, Türkiye, pp. 478–81; Langdon, “Byzantium’s Initial Encounter,” pp. 132–33 withn. 218; and idem, “Twilight,” pp. 188–89, 190–91.

214 Cahen, Pre-Ottoman Turkey, pp. 277–79; idem, Formation, pp. 187–90; Turan, Türkiye, pp. 485–87, 493–96.

215 Ibn Bıbı, ed. Houtsma, pp. 296–98, trans. Duda, pp. 282–84; Aksarayli, ed. Turan, p. 70.The most informative Byzantine source is Pachymeres 2.24, ed. and trans. Failler, pp. 182–85; inaddition, see Gregoras 4.1, ed. Schopen, p. 82, trans. van Dieten, pp. 103–4. For further details seeGeanakoplos, Michael Palaeologus, p. 81 with n. 28; Cahen, Pre-Ottoman Turkey, p. 279; idem,Formation, pp. 190–91; and Turan, Türkiye, pp. 497–98. For the background with respect toByzantine-Mongol relations in this period see David O. Morgan, “The Mongols and the Eastern Med-iterranean,” in Latins and Greeks in the Eastern Mediterranean after 1204, ed. Benjamin Arbel, Ber-nard Hamilton, and David Jacoby (London, 1989), pp. 198–211, at p. 204; and Langdon, “Byzan-tium’s Initial Encounter,” p. 135 with n. 227 and pp. 136–38 with n. 241. For Ïzz al-Dın’s furtherfate see Cahen, Pre-Ottoman Turkey, p. 279; idem, Formation, p. 191; and Langdon, “Byzantium’sInitial Encounter,” p. 135 n. 227.

216 Ibn Bıbı, ed. Houtsma, p. 213, trans. Duda, p. 204, calls her Bardulıya; Pachymeres 2.24, ed.and trans. Failler, p. 183, line 23: πρFς δ2 κα3 γηραι&+α µητρ�, �ριστιαν&+ η 1ς τ. µ�λιστα �Cσ+η. Forfurther details see Turan, Türkiye, p. 458; Cahen, Pre-Ottoman Turkey, p. 271; and idem, Forma-

644 Defection


Kadıd, the sultan’s sharabsalar (cellarer), who is designated as “maternal un-cle.”217 Ïzz al-Dın also patronized other persons of Byzantine-Christian pedi-gree, a case in point mentioned by Pachymeres being the so-called Basilikoi fromRhodes who are portrayed as rich megistanes and intimates of the sultan.218 More-over, the fact that in 1261 the metropolitan of Pisidia (Antioch) led Ïzz al-Dınto Byzantine territory219 shows that he also maintained close relations with theremnants of the Greek ecclesiastical hierarchy in the sultanate. Open to furtherdiscussion is the question whether Ïzz al-Dın was actually married to a Byzan-tine princess or even a daughter of John III Vatatzes, as a rather elusive piece ofevidence in an Armenian source might suggest.220

Be that as it may, the strong Christian element at Ïzz al-Dın’s court certainlycontributed decisively to the successful communication with the imperial courtduring the period in question.221 The cordial relations with Emperor Theodore IIpersisted without rupture after the takeover of Michael VIII, who shortly afterhis first coronation in January 1259 dispatched an embassy to Konya announc-ing his rise to power and discussing the state of affairs in the sultanate.222 Backin Nymphaion, Michael received Seljuk emissaries and promised full support toÏzz al-Dın if he were forced to take refuge with him.223 In relating these contactsPachymeres places special emphasis on the prevailing spirit of friendship,224 whichin all likelihood went back to Michael’s stay at the sultan’s court in 1256.

Against the background of the Mongol invasion of the Near East in 1258, Mi-chael VIII sought to secure his Anatolian territories through a policy of accom-

tion, p. 175, which refers to a different version according to which Ïzz al-Dın was the son of “adaughter of a Greek priest,” while his brother Kılıc Arslan IV was “the son of a Turkish woman ofKonya.” Ibn Bıbı, however, speaks of a jariya-yi Rumıya, i.e., “a Greek slave maid.”

217 Ibn Bıbı, ed. Houtsma, p. 297, trans. Duda, p. 284. Aksarayli, ed. Turan, p. 40, also mentionsa certain Kirkhaya Rumı-ra ki khal-i u bud (K. the Roman, who was his maternal uncle); see alsoVryonis, Decline, p. 203, who interprets the Persian form as possibly reflecting the Greek name “KyrGiannis.”

218 Pachymeres 2.24, ed. and trans. Failler, p. 183, lines 1–11.219 Pachymeres 2.24, ed. and trans. Failler, p. 185, line 3: τ�&υ Πισσιδ�ας πρ�αγωγ�&υντ�ς.220 Langdon, “Byzantium’s Initial Encounter,” pp. 120, 134 (referring to the chronicle of Kirakos

Gandzakets©i); and idem, “Twilight,” p. 188, argues for the reliability of the source and the politicalsignificance of this marriage. I would like to express my gratitude to Dr. Dmitri A. Korobeinikov(Russian Academy of Sciences, Center of Byzantine Studies) for providing me with his translation ofthe pertinent passage from Kirakos Gandzakets©i, Patmut©iwn Hayots© [History of the Armenians],ed. K. A. Melik©-Ohanjanyan (Erevan, 1961), p. 318. On the basis of philological arguments Ko-robeinikov concludes that the sultan probably married a member of the large Dukai-Vatatzai clan,thereby becoming the emperor’s son-in-law. This and many other questions related to Ïzz al-Dın’srelations with Byzantium will be thoroughly analyzed in the same author’s forthcoming monograph,Byzantium and the Turks (see above, n. 119).

221 See also Vryonis, Decline, pp. 377–78, 391.222 Pachymeres 2.6, ed. and trans. Failler, p. 141, lines 11–15. See further Franz Dölger and Peter

Wirth, Regesten der Kaiserurkunden des oströmischen Reiches, 3: Regesten von 1204–1282, 2nd rev.ed. (Munich, 1977), no. 1859.

223 Pachymeres 2.10, ed. and trans. Failler, p. 149, lines 15–18.224 Pachymeres 2.6, ed. and trans. Failler, p. 141, lines 14–15: Jλλ< 1ς τ. µ�λιστα ��λ�ις �&�σιν;

ibid., p. 149, lines 16–17: Bπισ�νε&ιτ� πρ�σ�ωρ�σαντα τ&+ η /Ρωµα�ων Bπτ�αις τε δ!�εσθαι �ερσ�ν.

645Defection


modation toward the Ilkhanids and the Golden Horde in combination with a cer-tain extent of control over the sultanate of Konya. These endeavors convergedwith Ïzz al-Dın’s need for protection against possible Mongol encroachments andhis internal opponents.225 This constellation resulted in an increasing tighteningof personal bonds on the Konya-Nicaea axis, which, apart from the sultan him-self, also encouraged other Seljuk dignitaries, both Christian and Muslim, to switchsides. The aforementioned Basilikoi fled to Michael VIII after his ascent to powerand were granted important ranks as parakoimomenos tou koitonos and megashetaireiarches respectively.226 Melik (Μελ�κ) or Ghiyath al-Dın Masüd, Ïzz al-Dın’s son, had already fled in 1258 to the imperial court, being viewed by hisfather for some time as a potential threat to his own ambitions.227 Älı Bahadur,a Seljuk emir who had played an important role in the conflicts with Kılıc Ar-slan IV,228 defected shortly after Ïzz al-Dın’s flight in 1261. The fact that by thattime many of the sultan’s former supporters had been executed by the new re-gime in Konya229 rendered the Byzantine court even more important as a placeof sanctuary.

The events of 1261 also brought about a fundamental redefinition of the ideo-logical relationship between the emperor and the Seljuk sultan, which mani-fested itself in unparalleled forms of imperial self-representation granted for thefirst time to a Turkish-Muslim lord. Ibn Bıbı and Pachymeres agree as to the highlyhonorable treatment the Seljuk guest enjoyed at Michael VIII’s court. While IbnBıbı somewhat elusively notes that “the lord of the Romans made extraordinaryefforts in honoring him [the sultan], and they were all day long occupied withfestivities,”230 the Byzantine historian, most likely referring to a prokypsis cer-emony, relates that the sultan was sitting beside the emperor on an imperial po-dium, surrounded by terrifying bodyguards and making use of the symbols oflordship by wearing red sandals.231

By recalling the setting of Manuel I’s official appearance with Kılıc Arslan IIin 1162, the much-elevated status that the Seljuk sultan had been granted in thehierarchical thinking of Byzantine court ceremonial becomes obvious. Instead ofbeing placed below, he sat on the same level with the emperor, and, even moresignificant, he was entitled to present himself with one of the most distinctiveinsignia of imperial authority. The sultan’s privilege to appear publicly in red shoesis all the more remarkable in that the court of Nicaea unmistakably rebuked sim-

225 For details see Langdon, “Byzantium’s Initial Encounter,” pp. 135–37, and the bibliography citedthere.

226 Pachymeres 2.24, ed. and trans. Failler, p. 183, lines 11–19, who with respect to the Basilikoi’sintegration into the Byzantine court hierarchy uses the characteristic phrase κατ. /Ρωµα��υςµετασ�ηµατισθ!ντες (transformed into Romans).

227 Pachymeres 2.10, 24, ed. and trans. Failler, pp. 149, lines 18–21, and 183, lines 26–27; Grego-ras 4.1, ed. Schopen, p. 82, trans. van Dieten, p. 103. See further van Dieten’s comment ibid., p. 239n. 132, and Vitalien Laurent, “Une famille turque au service de Byzance: Les Mélikès,” Byzanti-nische Zeitschrift 49 (1956), 349–68, at pp. 361–62, 368.

228 Ibn Bıbı, ed. Houtsma, p. 297, trans. Duda, p. 284.229 Ibn Bıbı, ed. Houtsma, pp. 298–99, trans. Duda, p. 286.230 Ibn Bıbı, ed. Houtsma, p. 297, trans. Duda, p. 283: “malik al-Rum dar ta¨z� ım-i u mubalaghaha

kard wa-hame ruz bi-äysh mashghul budand.”231 Pachymeres 2.24, ed. and trans. Failler, p. 185, lines 7–9.

646 Defection


ilar acts of its rival claimants for the imperial office in Epirus and Thessaloni-ca.232 In the Palaiologan period, according to the treatise of Pseudo-Kodinos, eventhe despotes, who were mostly sons of the senior emperor and second in rank tothe basileis, wore two-tone shoes (δι� λεα π!διλα).233 This distinction, there-fore, cannot be explained as an extraordinary sign of courtesy or as recourse tothe old bonds of spiritual kinship. Rather, it expresses a sort of recognition onthe part of the Byzantine court of the general tendency of Konya to incorporateimperial elements into the sultan’s symbolism of authority, a trend that after 1204had significantly intensified, as the formulaic patterns of the sultans’ official cor-respondence with Christian potentates reveal.234 Through ¨Izz al-Dın’s arrival in1261 these features found their way into Byzantine court ceremony, celebratingthus a kind of ideological unification of Muslim and Roman-Christian traditionsof lordship. From Alexios I’s cape on Shahanshah’s shoulders until the red shoeson ¨Izz al-Dın’s feet, it was a long journey of nearly one and a half centuries.

This survey of two centuries of Byzantine-Seljuk apostasy and defection revealsa fascinating field of human behavior, where legal views, moral constraints, andethnic stereotypes stood in sharp contrast to day-to-day practices and survivalstrategies in a conflict-ridden world. As for the historical facts, the story beganin the 1050s with seditious Seljuk commanders arriving in Constantinople andoffering their services to the Byzantine army in exchange for incomes and titles.During the following decades, the increasing Turkish pressure on the empire’s east-ern borderlands caused Armenian lords and Byzantine local commanders to de-velop a modus vivendi with the invaders, ranging from pretense of deference toopen collaboration and even conversion to Islam. The recognition of the Seljuklordship of Nicaea as an independent political entity in 1081 prepared the groundfor a new bilateral relationship, enabling the Byzantine emperor to present him-self as the protector of threatened Seljuk lords who were coming as refugees toConstantinople and to establish bonds of spiritual kinship with them, so that thesultans of Rum gradually became integrated into the Byzantine Empire’s innercircle. This practice, among other factors, may have inspired discontented Com-nenian aristocrats from the 1130s onwards to set off in the opposite direction,taking refuge with powerful Turkish potentates, a practice that later on wasadopted by Byzantine generals and local rebels as well.

The events of 1204, once more, turned things upside down, with the sultan ofKonya becoming directly involved in the contest of several claimants to the impe-rial title. Contemporary narrators referring to these events swayed between extol-ling successful attempts to win over dignitaries from the foe’s camp and condemn-ing defectors abandoning their own ranks. Court rituals provided an appropriateframework for stressing concepts of superiority and projecting successful ruler-ship. On the other hand, emotional outbursts emanating from a general atmo-sphere of contest and insecurity provided a good basis for justifying morally un-

232 See, for instance, Akropolites 38, ed. Heisenberg, 1:61, trans. Macrides, p. 207, concerning theadoption of red shoes and red ink by John, son of Theodore Angelos, in Thessalonica.

233 Pseudo-Kodinos, Traité des offices, ed. and trans. Jean Verpeaux (Paris, 1966), p. 274, line 12.234 See above, n. 205.

647Defection


acceptable forms of behavior. In some cases, however, our informants deliberatelyreversed these interpretative patterns, downplaying the significance of conflictsor dwelling on the failure of rituals. When the emperor himself became the objectof harsh criticism, the Turkish barbarians, who in the eleventh century were re-garded as uncivilized “Huns” or “Scythians,” susceptible to all manifestations ofRoman cultural and moral superiority, could even be considered highly esteemedauthorities, able to judge a Byzantine general’s suitability for the imperial throne.Moreover, the gradual elevation of fugitive Seljuk sultans seeking the emperor’sprotection, who from the Comnenian court title of sebastos successively ascendedto the status of the emperor’s “spiritual son” and “godchild,” eventually resultedin a remarkable synthesis of Byzantine and Seljuk-Muslim ideas of lordship. Thefact that Michael Amiraslan, the Seljuk dignitary mentioned at the beginning ofthis essay, even on his coffin insisted on his family’s claims for a purple-born lin-eage is but one consequence of this ideological amalgamation.

Appendix

The following tables provide a synopsis of the prosopographical data of all individualsinvolved in Byzantine-Muslim defection. The entries are arranged in chronological orderand indicate each defector’s name and original status, the court he took refuge with, andthe results of the defector’s stay at the foreign court.

Table 1Turkish Warlords and Potentates (1050–1118)

Date Defector Sheltering court Outcome

1056/57 Ibn Khan al-Turkumanı,Turkish commander

Michael VI service in the imperial army

1070 Arısghı/Chrysoskoulos,Turkish commander,brother-in-law of SultanAlp Arslan

Romanos IV title of proedros, service inthe imperial army

1078 Sulayman b. Kutlumusand his brothers

Nikephoros III support for Nikephoros’srevolt, control over citiesin Bithynia

1086 Siaous, Seljuk dignitaryof mixed Georgian-Turkishorigin, emissary on behalfof Sultan Malik Shah

Alexios I baptism, support of Alexios’spolicy in Asia Minor, duxof Anchialos

1092 Elchanes, Turkish potentatein the region of Apolloniaand Kyzikos

Alexios I baptism, court titles

1092 relatives and companions ofElchanes, among them acertain Skaliarios

Alexios I baptism, court titles

1092 Abu l-Qasim, Seljuk lord ofNicaea

Alexios I peace treaty, title of sebastos

1116 Shahanshah, Seljuk sultan ofKonya

Alexios I reception ceremony project-ing proximity, peace treaty

648 Defection


Table 2Christian Frontier Lords (1050–1118)


1067/68 Aristakis, Armeniancommander

Alp Arslan, sultan of theGreat Seljuk Empire

temporary conversionto Islam

1085 Philaretos Brachamios,Armenian potentatein northern Syria

Malik Shah, sultan of theGreat Seljuk Empire

temporary conversionto Islam

1085 Philaretos’s son Sulayman b. Kutlumus,lord of Nicaea

agreement concerningan attack on Antioch(Anna Komnene)

Table 3Comnenian Aristocrats and Court Officials (1118–1260)


1130 Isaac, sebastokrator,brother of EmperorJohn II

Ghazı Gümüstekin,Danismend emir;Mas¨ud, sultan ofKonya

temporary stay1130–1138/39,conspiracies againstJohn II

1140 John, son of thesebastokrator Isaac

Ghazı Gümüstekin,Danismend emir

conversion to Islam,marriage with adaughter of thesultan of Konya

1167 Andronikos Komnenos,second son of thesebastokrator Isaac

Saltuk, emir ofErzurum (viaAntioch, Jerusalem,and Syria)

temporary stay1167–80, raids onByzantine territory(John Kinnamos)

1167 Alexios Axouch, proto-strator, son of JohnAxouch (a formerTurkish slave, megasdomestikos underJohn II)

Kılıc Arslan II, sultanof Konya

accusations ofconspiratorial plans(John Kinnamos)

1195 Isaac Komnenos, grand-son of Manuel I’sbrother Isaac, ex-rulerof Cyprus

Ghiyath al-DınKaykhusrau, sultanof Konya

failed attempt toconclude a militaryalliance

1200 Michael Angelos-Doukas,dux of Mylassa,bastard son of thesebastokrator Johnand cousin ofAlexios III

Rukn al-Dın, sultanof Konya

raids on Byzantineterritory

1204 Theodore Laskaris,despotes, son-in-lawof Alexios III


temporary stay beforehis takeover inNicaea

1204 Theodore Angelos-Doukas,brother of MichaelAngelos-Doukas


temporary stay beforehis flight towestern Greece

(continued )

649Defection


Table 3 (continued )


1224/25 Andronikos Nestongos,cousin of John IIIVatatzes and head of aconspiracy

¨Ala© al-Dın Kayku-badh, sultan ofKonya

lifelong stay

1237 Manuel Angelos, despotesof Thessalonica, brotherof Theodore Angelos

Seljuk authorities ofAttaleia

handed over toJohn III Vatatzes

1256 Michael Palaiologos, me-gas kontostablos underTheodore II

¨Izz al-Dın KaykawusII, sultan of Konya

protection againstTheodore II, par-ticipation in thebattle of Aksarayagainst the Mongols(October 1256)

1259 Theodore Karyanites,protovestiarites

Seljuk territory killed by Turcomanwarriors

Table 4Local Potentates and Rebels (1118–1260)


1126–40 Constantine Gabras, semi-independent governor inTrebizond


military alliance

1130 Kassianos, semi-independent governor inthe Pontus region


submission,annexation of histerritory

1182 Manuel and Alexios, sonsof the megas domestikosJohn Vatatzes based inPhiladelphia


protection againstAndronikos I

1191 Theodore Mangaphas,local potentate ofPhiladelphia


raids on Byzantineterritory, eventuallyhanded over tothe imperialgovernment

1191 Pseudo-Alexios“Kausalones,” localrebel in the region of theMeander valley



1195 Pseudo-Alexios, local rebelin the region of Ankyra

Muh�yı al-Dın, emir ofAnkyra


1204 Manuel Maurozomes,potentate in AsiaMinor


high offices at theSeljuk court

1211 Alexios III, ex-emperor Ghiyath al-DınKaykhusrau, sultanof Konya

military support,campaign againstTheodore ILaskaris

650 Defection


Alexander D. Beihammer is an Associate Professor in the Department of History and Ar-chaeology at the University of Cyprus (e-mail: [email protected]).

Table 5Turkish Potentates and Sultans (1118–1260)


1125/26 Masüd, sultan of Konya John II protection against internalopponents, temporary stay

1126/27 Ärab, brother of SultanMasüd

John II protection against internalopponents, died inConstantinople

1162 Kılıc Arslan II, sultan ofKonya

Manuel I protection against internalopponents, peace treaty,addressed as the emperor’s“spiritual son”

1175 Dhu l-Nun, Danismendruler

Manuel I protection against internalopponents

1196 Ghiyath al-DınKaykhusrau, sultan ofKonya

Alexios III protection against internalopponents, marriage with adaughter of ManuelMaurozomes, baptism

1257 Ïzz al-Dın Kaykawus II,sultan of Konya

Theodore II military support for regaininghis throne

1258 Ghiyath al-Dın Masüd II(Melik), son of Ïzzal-Dın Kaykawus II

Michael VIII protection against internalopponents, temporary stayuntil his flight to theGolden Horde 1278/79

1259 Basilikoi, dignitaries ofGreek descent at thecourt of Konya

Michael VIII offices of megas hetaireiarchesand parakoimomenos toukoitonos

1261 Ïzz al-Dın Kaykawus II,sultan of Konya

Michael VIII protection against internalopponents, bestowal ofimperial insignia, baptism,temporary stay until hisflight to the Golden Hordein 1278/79

1261 Älı Bahadur, Seljuk emir atthe court of Konya

Michael VIII protection against internalopponents

651Defection
