Peoples and settlement in Anatolia and the Caucasus, 800-1900 9780860782223

Citation preview

Peoples and Settlement in Anatolia and the Caucasus

awd rn, aRenee rcmao rr “4 PEE ay ee aws? . ;' ie re .° nowy : apJoan aar. : oo .eoa ,ayi,asHon eyae Bete oo‘4 “Ee # bF, La a : oo ee ag i, oat! 42 we uy —_ ant. a oa os . mm eae oh i os. rn x ak, a cows a a “ahs tet a ae at J . ae pe . ; ° ‘ 7s a apts ».'™ :“oksg 2 a » a . rot oh ‘} tye : ; ies mh. ; se ARNT . oF 4“oo.Raw, Sigs, Pie oes Moe ii: ot, eee ee, ter 2s i oe,7 iae.. " i:tk sal ifi 42oy etvt"gheyeet - a4x_ vs,: :wee. ‘ ES

oe ‘ pn ft St Me Sa, , “— | we Tae Vaal 5 ee | hed uN H " °

rire cree °Fae_ .Bee is geateyy - , -*: *. . Aaeersoe Se

Be re? ee em BO oy iC ne:fsohaos ce: oebe ee 0 Retires eet ekePe CsEM et Rene... | neSRM | en ome en cy i. 5 a SE a OPN, pleas Ae OS a ebigen Oe geut Fa mS a oe ae TR ghee s

. Fe ae ad . ” * : “et oe ot 3 Oj aae Feae onsOS ee a- abeep tN ee Nis sy ak ad+“Sas ?+o Mee, “ egoe nyae Baaey by ae iatPy oeee Chea akeCy pe sas Fa SAR a val, 2on” e cre ;ea- ‘thit"yy y 3Brig aS 3“he Te Mey okrcMe ai ae aeesFH 2igay. PE My ‘fs eh. tf pen :aenoo ot

ae ~ 2 weet. od ne ce" a itv) OA ; a . . * “SM ad RN : oe ahs "a ae Laas a epee

‘ r *. aad et.*oygee : ES ae, .m > Sea ae ae a a aee * ay) aeors vey ty fmm ce a LTON ot ur .a * o ©SRT :,rs °*.;Sy pera RRS Sm re satay tee a Aad ne athe aeNeh Sg gt oa a.ety! 4q's F :4ce +.int aGra eeTe* ba Neat oa wae ees Oe, :ith asfoo Wena act ayhe hy .eo.POE, Diy i : ye . , ne Boba te “pu ae Bs ™ ty ; . oe Pd Pe ne ak: iad 4 mu ee a) * ‘ wat Gq i, apes e a a ag ae Sd . fr Ao, od ie ty ‘ger moe i Se eae ae ja es cee RMB Be | Fee os * S ee ace ee i ae 5 a a me has va cae an a ere re? ok Oia ar,Tepe seer: .a£en ee porte poe x °ae. .Si: Meaiest , Roe so .SEO se asag zm; om uF OTSpeST oyeecooe“.‘ ee ee *+3

ae; DeSertswlFe; +ame aeaT UF 6B PR aebat Cie - 27 Ewa 34ie pone >Ety t |}4& : , “7s ter .ceo oemu - ,*ssae Ie>+, ow oyRee FR.2ae *.wegen a» was ee oy Rew 7” oe 3: Rost:

Ce Bie ee a2 eebewo, od: 0)Ring eye ‘ " a4otigyt a i ,be lhhaeal ot ioe nr ae ogee Bea oe *pe is ee Ly SRa +, aede® 1? hot .eek, rh iia ee%, Aety ee ay Ne Raeaehea ayeCeman RCTS

Li ele Ree ee a. eo ; s it alate ‘ . i eg SE SN rer ‘ Sow "\ ae CREE

rr ey) . -al . ena . i»ory * SRat . * |>an 7 oePo, f ye:; 4200 ; # Foo $ = Baer , ‘ ? I‘ “ay é.ee na:he ar ’et hyeca ual: ;, Lona »Se BR aRw aye73eee RS byie par nlp set at nm avai shes NE ROM SO es ol ether , - a ésee 4 Lael ts i rd -:xso : 7 we ' ganA may ‘ a ™ pas oe APhans 4 P\ }:; .ees a hors A gery iE Om icsAES ‘ "aby enan a- ahaa rf, ;_

oe- ,Bye" ae uBed . woy - a ges fee. ~. ~ uo Do sails, . Pe we Ee AY ‘ae Lant*. Oty uh doe ee Es) Tp a eee, ‘hid “2 od ee “eq eee Sra a ey a ae ~ ime ous, PA Fm a rewe q von o te ie .RODE: OP DARTNe So on a ie cae ca re -tiiak fe "pi 1 SRMLo AEede Ws noe mere e. i re aNoeeePe,PMR . PePops age8sTE one

Grand K Alexios IIIIIIand E nd Komnenos Alexios and Empress Theodora Kantakouzene f the chrysobull for Soumela monastery, 1364.

See Introduction. PhSinasi nasi Basegmez. e Introduction. Photo: °

Anthony Bryer

Peoples and Settlement in Anatolia and the Caucasus, 800-1900

aS Rs |

(3 =—s VARIORUM REPRINTS London 1988

British Library CIP data Bryer, Anthony

| 4 ) 4 Peoples and settlement in Anatolia and the Caucasus, 800-1900. — (Collected studies series; CS274).

1. Turkey— History 2. Black Sea Region — History I. Title —_IT. Series

956.1 DR440 ISBN 0-86078-—222-0

Copyright © 1988 by Variorum Reprints

Published in Great Britain by Variorum Reprints 20 Pembridge Mews London W11 3EQ

Printed in Great Britain by Galliard (Printers )Ltd

Great Yarmouth Norfolk

VARIORUM REPRINT CS274

Ge

f 9-04 77 oh pts

EAE FS CONTENTS

Introduction Vli—x I Achthamar and Digenis Akrites 295-297 Antiquity 34. London, 1960

II Han Turali rides again , 193-206 Byzantine and Modern Greek Studies 11. Birmingham, 1987

Ill A seal of Epiphanios, | Archbishop of Cyprus 19-24 Kypriakai Spoudai 34. Nicosia, 1970

IV Greek historians on the Turks: the case

of the first Byzantine-Ottoman marriage 471-493 The Writing of History in the Middle Ages. Essays presented to Richard William Southern, ed. R. H. C. Davis and J. M. Wallace-Hadrill. Oxford: Clarendon Press, 1981

a Trebizonde 216-234

V Une eglise “a la demande du client” Proche-Orient Chrétien 32. Jerusalem, 1982

VI Eclipses and epithalamy

in fourteenth-century Trebizond , 347-352 Byzantium. A tribute to Andreas N. Stratos, ed. Nia Stratos, II. Athens, 1986

Vil “The faithless Kabazitai and Scholarioi” 309-327 Maistor. Classical, Byzantine and Renaissance | Studies for Robert Browning, ed. Ann Moffatt (Byzantina Australiensia 5). Canberra, 1984

VIII The treatment of Byzantine place-names 209-214 Byzantine and Modern Greek Studies 9. Birmingham, 1984/85

IX Nicaea, a Byzantine city 22-31 History Today XX1 1.

London, January 197] ,

vi

X The structure of the late Byzantine town:

Dioikismos and the Mesoi 263-279

Continuity and change in late Byzantine and early Ottoman society, ed. A. Bryer and H. Lowry. Birmingham and Washington D.C., 1986

XI The question of Byzantine mines in the Pontos: Chalybian iron, Chaldian silver, Koloneian alum

and the mummy of Cheriana 133-150 Anatolian Studies 32. London, 1982

XII Rural society in Matzouka 53-95 Continuity and change in late Byzantine and early Ottoman Society, ed. A. Bryer and H. Lowry. Birmingham and Washington D.C., 1986

XIII Byzantine porridge 1-6 Studies in Medieval History presented to R. H. C. Davis, ed. H. Mayr-Harting and R. I. Moore. London: Hambledon Press, 1985

XIVa Some notes on the Laz and Tzan (I) 174-195 Bedi Kartlisa. Revue de Kartvélologie

XXI-XXII (50-51). Paris, 1966

XIVb Some notes on the Laz and Tzan (II) 161-168 } Bedi Kartlisa. Revue de Kartvélologie XXITI-X XIV (52-53). Paris, 1967

XV ___ The last Laz risings and the downfall

of the Pontic Derebeys, 1812-1840 191-210

Bedi Kartlisa. Revue de Kartvélologie XXVI. Paris, 1969

XVI ‘The three Cyrnils 155-158 Continuity and change in late Byzantine and early Ottoman society, ed. A. Bryer and H. Lowry. Birmingham and Washington D.C., 1986

XVII The crypto-Christians of the Pontos

of Trebizond 363-365

and Consul William Gifford Palgrave 13-68

Index 1-11 Deltio Kentrou Mikrasiatikon Spoudon 4. Athens, 1983

This volume contains xii + 322 pages.

INTRODUCTION These articles develop the themes and extend the scope of my first Variorum Collected Studies, The Empire of Trebizond and the Pontos (London, 1980). A major theme is the survival and development of the medieval cultures of those Christian Anatolian and Caucasian peoples who became, until this century, subjects of the Ottoman Empire. Lesser themes are that material for research on them is still coming to light, but that it may not always be what it appears at first sight. These points are

nicely illustrated in the frontispiece, which happily breaks Variorum

tradition in being a portrait not of the author, but of the Grand Komnenos Alexios III of Trebizond (1349-90) and of his Empress Theodora Kantakouzene, at the head of the chrysobull which Alexios granted the great pilgrim monastery of Soumela in December 1364. One

of only two ‘original’ Trapezuntine chrysobulls which survive, this majestic document is 30cm wide and unrolls to 3-35 m. All editions of it

derive from a transcription of 1775; the last published sighting of the bull in Soumela itself comes in 1879. Now, over a century later, Sinasi

Basegmez has ‘rediscovered’ it, and the archive and treasury of Soumela, safe in the Ayasofya Museum, Istanbul: see his “Trabzon Rum Imperatoru Komnenos’un Fermanlani”, Antika, 22 (1987), 10-29.

We plan a proper edition of the archive. However, it does not take a second glance to notice that the portraits of Alexios and Theodora do not belong to the opening lines of the

chrysobull proper, engrossed in 1364. Patched into the original document, Alexios and Theodora have the air of being not Byzantine, but later Romanian, rulers of Moldovlachia, where Soumelan monks and scholars went fund-raising, with relics such as this bull. Perhaps a link is the Trapezuntine pundit Sebastos Kymunites (1630~—1702), Frontisterion professor in Trebizond and Bucharest and author of a

demotic version of this very charter, the curious history of which demonstrates the longevity of an Anatolian cult and culture which flourished as much under Ottoman as Byzantine rule. On Digenes and like heroes, Study I was the first published (in 1960), and Study II the last (in 1987). They are inconsistent in transliteration,

but the reader will have to put up with that: in 1960 I was an undergraduate. But they are consistent in a theme that the cultures of Anatolia and the Caucasus cannot be taken in isolation, particularly in their common wealth of heroic poetry and /aographia.

Vill

The context of Study III is the first Orthodox/Muslim condominium of Cyprus, and establishes by his seal the existence of an archbishop of the island. In his subsequent Le Corpus des Sceaux de l’ Empire Byzantin,

V (3), L’Eglise. Supplement (Paris, 1972), p. 287, no. 2015, the late Fr Vitalien Laurent was more daring than I in suggesting that this grubby piece of lead, acquired in Istanbul, once sealed a letter from Archbishop Epiphanios to Patriarch Ignatios in 870. Study IV is the first of several on the symbiosis of Greeks.and Turks, starting at what should have been its most intimate point: the ‘marriage’ between Sultan Orhan and (another) Theodora Kantakouzene in 1346, which set off a whole chain of such alliances. The sequence got off to an ominous start, for the bridegroom does not seem to have attended his

own wedding. Relations between Orthodox rulers could be just as tricky. In Study V, I argue that the Chrysokephalos cathedral in Trebizond was custom rebuilt in 1214-35 specifically to house the

requirements of an antique liturgy which the Grand Komnenoi recognised could not be performed for them in the Hagia Sophia, Constantinople: that of coronation. Reviewers of David Winfield’s and my subsequent Dumbarton Oaks Study on The Byzantine monuments and topography of the Pontos (Washington D.C., 1985), I, p. 120; II, pls. 161, 162, complained that we offered no plan or useful photographs of the building: they are here. An egregious scholar of Christian minorities

in the Middle East, Jean-Michel Hornus prepared this article for publication the day before his tragic death in 1982. Studies VI and VII start with the evidence of two Bodleian MSS for the beginning and end of dynastic and social strife in Trebizond, which coincided with similar, if obscurely related, upheavals in the West as well as in the Ottoman and Byzantine Empires. The first, of 1335, 1s an epithalamy for Basil, father of the Grand Komnenos Alexios III, and his Palaiologan bride, whose marriage was darkened by a midday eclipse. The second sheds light on the midnight assassination in 1429 of the Grand Komnenos Alexios IV, grandson of Alexios III, who was buried outside the Chrysokephalos cathedral. Unlike the Soumela archive, his remains were transferred to Greece by Pontic refugees in 1923, but we may never know if they are the only known skeleton of a Byzantine emperor, for they were firmly reinterred at New Soumela in 1980.

Studies VIII—X explore the themes of topography, settlement, and the texture of Byzantine towns in which communities coexisted 1n their own quarters even before the Ottomans called them mahalles. Study IX takes the example of Nicaea, for which I used the MS diary of John Covel’s visit of 1677, which has since been published by Julian Raby, “A seventeenth-century description of Iznik-Nicaea”, Istanbuler Mitteilun-

1X

gen, 26 (1976), 149-88. My own article may be the last record of the church of Nicaea’s patron, St Tryphon, destroyed soon after 1967; but

the photographs of Iznik at the turn of this century, taken by Sir Benjamin Stone and now in the Birmingham Public Library, remain to be published. Studies XI-XIII are on the continuity of material culture: mining, agriculture, even dietary differences between the peoples of Anatolia; the context is Study VII of my first Variorum collection. To this group I

would add (but for copyright reasons not here), my “Byzantine agricultural implements: the evidence of medieval illustrations of Hesiod’s Works and Days”, Annual of the British School at Athens, 81

(1986), 45-80. Like Studies X and XVI, Study XII derives from the Birmingham-Dumbarton Oaks ‘Demography’ project of 1978-82. The

whole field needs further exploration. |

Studies XIVa-b and XV concern the Laz, a Caucasian people who

from Antiquity to this century have flourished under (and used) - Anatolian.rulers without losing their identity. They appeared in Bedi Kartlisa, a journal which died with its editor, Kalistrat Salia, in 1985. Although Study XIVa in particular is largely a heap of information, I include this group because interest in the Laz is evidently reviving, not least among the Laz themselves who are now scattered from Istanbul to Hopa, but also among anthropologists of the north-eastern corner of Turkey: cf. M. E. Meeker, “The Black Sea Turks: some aspects of the ethnic and cultural background”, International Journal of Middle East

Studies, 2 (1971), 318-45; and the work of David Braund and Christopher Hann. The Laz, whether at home in their own summer pastures, or (like the Balkan Vlachs) enterprising entrepreneurs elsewhere, need further research. Study XIVb discusses the great castle of Tzanicha (Canca), which Winfield and I subsequently published in our Dumbarton Oaks Study of 1985: I, pp. 309-10, figs. 106-8; IT, pls. 244a-248b. Its painted churches may be related to a nearby one, dated 1265, which I am publishing; and maybe to others 1n the Kanis (Harsit)

valley, reported by G. Uslu, Gdimiishane ve ¢evresinin tarihi-sanat eserleri (Istanbul, 1980), 23—27.

Whether the Tzanichitai were Laz or not, or even held Tzanicha, they represented a perennial local phenomenon of great clan chiefs. Their

rivals, the Kabazitai and Scholarioi, had been faithless to the Grand Komnenos Alexios IV in 1429 (Study VII). Ostensibly the Ottomans only got to grips with their successors, under the guise of Derebeys, in 1812-40 (Study XV), when for a moment it looked as if the peoples of the Pontos, like those of other parts of the Ottoman Empire in the Balkans, Egypt or the Lebanon, might have gone their own centrifugal way under local dynasts.

xX

But the Patriarchate offered Byzantine and Ottoman rulers of Constantinople centripetal control of their more wayward Orthodox subjects. How this worked in detail in the 1620s is illustrated in Study XVI. Even so, until this century both the Porte and the Patriarchate had difficulty in defining who a Pontic Orthodox was. The fact that many were miners (Study XI) put them into a peculiar civil category which, in

Study XVII, I argue drove them into a twilight world as cryptoChristians. In “On fifteenth-century Crypto-Christianity”, Byzantine and Modern Greek Studies, 9 (1984/85), Anna Frangedaki cites an earlier precedent, and the whole question is explored further in a doctoral dissertation presented by a Pontic Greek to the University of

Tiibingen: K. Fotiadis, Die JIslamierung Kleinasiens und die KryptoChristen des Pontos (1985). In the Pontos the matter came to a head in 1857, and was investigated for the British Foreign Office by Consul William Gifford Palgrave in 1867. But Palgrave’s own faith was

more enigmatic than that of any ‘Crypto-Christian’, on which further light will be thrown by B. Braude’s “The Heine-Disraeli syndrome among the Palgraves of Victorian England”, in Jewish apostasy in the modern world: missionaries and converts in historical perspective, ed. T. Edelman (New York, forthcoming).

Study XVII, my inaugural lecture of 1980, was dedicated to my parents, to whom it is proper to extend the dedication of this book. Not least, they have kept the offprints which are collected here.

ANTHONY BRYER Centre for Byzantine Studies and Modern Greek, University of Birmingham

) PUBLISHER’S NOTE The articles in this volume, as in all others in the Collected Studies Series, have not been given a new, continuous pagination. In order to avoid confusion, and to

facilitate their use where these same studies have been referred to elsewhere, the original pagination has been maintained wherever possible. Each article has been given a Roman number in order of appearance, as listed in the Contents. This number is repeated on each page and quoted in the index entries.

2%.

, SES

=“ysz Ca Be“rstyalae: Rat x ; “5 “pdb: sami Ragan or

y 2 ey Bese. -

. a ene Rays : oer PES,

ery Ren Sa . ty °;i 25 > aeSees. Oo aes

yn . ° - aoe ae tpestaes: esintge RN Poe, Es: weet: : - Ges S8oxs, Mee bet. Res

» >. . ¢ . pe i FS on al a ee ae oa i ae Saag

ft ees BAS Sieet GeAS ds aed.2 a agg Sgi Be gt Zoe” _ OES OS — .Ba ORE ieee. : 7. if gar. ae 3OY, & o, 2.oy ,au, aPan ty -.csae eae aoe 2 lls aoa Sed UNE ae 5p 4 x SMe aes Pe oe2S a I RR ON. aPa ikcg ag rm Re, .“oe 3 mt id le oeraeaYVAC - Won Rae SF Med ; is oa 7, 8s FP S ‘ RAR ane * "Mee, ln Bg BEE sag all.

ee.a—— I ae ae. ..s yet -" NS BOR, . wee 2s aaPly f :.xs SR Beso ie a , Sea eg a Tt oes . ae’ £ae . ¥ 7 a 1 po? '4”rm. a. io ‘2 RR ao F ee SY . i “a rm ,.. a *: . ana LotPees ueeentae 6. BE seer ye gyOR a Re: me * ‘ a a RS TR ae Te. Ye kes Hes, . 4 _# EW $5, %8 ee 2 , BD 9,ERRORS 7 $. .*iREE BEE Resnor mk aSR . ~ ee a.4. Sg og Sap 2.AS wa" ‘BE 4 Po ieaera orn “ae Ai ..an aig. . .:égf rye |“weBe .Soy :2 OOS ot Pe SAR wo es as ae “ey? ae! ere oe 4 iw at > "come 3 tig * aa wee NES. Ee ataass 5

ae . :arr’ :atay Lowe rewes »SON, wee Se aeON ee git OT OR een Be Oa SES rae Lee €,a..;os :os 5 Pfs Toa 5:.“a:.pee b>fe .:a"!Pn ace aaes ea ED re Ea en SP DA ao FU geen Jr ae ee ;.; ;a:_y» LF «niga ag, oe ee en Soe aaie aeh Bee tae % sage RET rie atoh, ¥eed 2Lt eee: Wi cee Sees LAS INR. Soe ape weg Bi te ale ee on Migek, fa ” “e : i “ ; Let cs a ee Ba N oR AES By ‘ i re ; ow IB Bee ee ne ea! : s, : . + 7 2 aed : 8 PO RR Oe we 3 apt SR nee net . . rs “1 Baseo ere igs o's ie Sa © ARR h A Pe ce eT “wo . 3 4 ° | Fg ae om _* Ne gle GRE. ire: Ea 2 gS cae “NS ca le hal

rn. 4. Gini EA . as : werk he eT eee i: An oe ere

bg rar 1:Oe igs:Yo TAY pare ar Rear angiao EM wove bh Tae~woe aaa wot. reRe . led eS© : °cCe aeSo en CloeeE ae Pec aeLEY ‘Aes SRR Bes OE Sete woee: Ha Bh ROA. atetRL, Pte etSe ete OR ae eedP ie %y. wv aeae etealte . $+PY Pet, 72OS? ¢SN ae- 4 :neOE eae Be a$y GA pee CAS ee: ake OE eeDAR eSbag Cee ot Ot - ERE 4ae ak ooh Seon" Oe ee ~ Geen Bete ‘tye yk eee Saeer Ss SA rs ai ee Cee De Beg Sos fe Ye nt ok. 4 . . : i ee, he ust es, Fagen i SS as an Be OEE se 95 a ee ate Bigs ing GAR A A Boss ae a 0 8 aed Oe Sirah geet: pes iis Ree TY “a: Wu pe eee aS a4? BAS ig Lcd 8S & Ge TREE eg BS wea MEM Eee firs wf AO ENS he eg ar ; Ba: “ 5 am oe aS TEE as UE acs shades ge % he: 3 ete PRE: eM, Ramer MOE SogP ih,Rgahin, GF CO th DN SEES OSE Beg anie: ; ‘ Ct hy nsoe sant: eesee ESgee gest RU i, Ew sere os : %as RRAeeePe Se NE Sel aTSaas TG SED Os SS gg Es ee a ee ee nen i pO no gee hs nn SS Re esPe i ee OES ei gC Beam FREE OCA Bee ~ ee3am a a4coee pede BEE Be Pe .Pb Rae ere arc SERS po Nae Seoe Shy a a lee LE VERE we NSE Sager Sete EE etSo SS SG My Moet, ee woe phe Pac eraess ARETE . :4‘Roy REZ EDR OR, BARE BE CS goIRE SS Sei Se esBienes SEIS gkPes Cookie SG": Lssent oie ~ss gm geok StgeBS: Seeee Were Be en, xOS Leae ye oesey. ‘eee

RIB, Gan 2 SRO fete pigs ee TS Hee PB ee Eee ae rae 7 BB ns By er acne ye fa Prepare SE SS ake eS CORTES, ea BS % pn isis LER oes GEER Fae i eeraeeo eAConey reBIoe Re SO eeoo2Sec 737:geek ‘a BLES OR SEES i EES ea CLO SSMes ORE es oh©.Zn eeOE te: we ligtgaRe Gadee BR 8 § BC 0 Ses ae ye ees iLiteohh EGe No wee OS a tg ae ee SS We BORON? ys igi OTR) BKPA eee Pie ee ark a a ee gen : a;SERS 4al . : wo;ee gr oo egGeer OOS” ERNIE: io’ 3 PROPOR AY

ae Osweg! CPt NESS see 8 eSSe CGE OETA. a io ,a:nn = ry Te PL §en2a See: 7BEM tO OO Lhe Rg Kadyeee Pe Eb Gas SR? GBS ERGEN SO De ae Rig ta ae Ress ME OS RR: ote iaaoer,_.,. are ROMER E85 4. BAS | ;7CRESS IK vn EE Pk40"arPe NVERR Seon ae LS Res a aag¢aare me RE ae " ERO EE ASE GRO Ee SES aeHie ae RES eeMPO aes GS52DIS ERI OL SO OE SBMS he RS, Rey wate ne re ; * wee EER SEER SaBELO Mee pe gem. Ses aS HE SR YES RE ok aes SE: ,GOREN, é eo “hs .A mf " - ae Ce Sey

ates eA ee we RED: > S=ER - —, tee Pe EE anoe Fe>¢enced ;,4g ih,z:3es * .-@ 1gag eSae: Bye ‘he 3atieye.ae~‘,SRN OF4; *.ess OR 25 ; tit potRimes a Ses ¥. {“er: GeO Gin Seewoes «FNoa , ‘.-who welOU seni ZF aad ee alk ,£ >, Pet

ae ¥ Ss 3 oe > oe es

ae 8&EIR Saceeoe &of POO eee Nese “6 ,, mo *$rds.». 4Peg 5“ z *7e0lU bis 5 ~aie .aia ah * o na AStPF Res . oaOL 4aw. wea .ka . .Me, °ae _ :. mE “ . “hee : ‘he nny, & Gee Mite og a oe cme 2 fg .;.:EYoan he a nn ‘ t= te ? ; Pj Lon “4 3 t . Sone , eR . . x Sew a -s ; _¢ eee I i : : . mmgli 5 a eres ~ . e r ve a AE 2M. Ee “et as . Lg x e BP! ~ % . ve k. ~ 7) gine = . 5 . OR ee 3oe. ‘oe Fs rn ;gg“=Piave. : a%. en ievee |Be Mg , asof aef © oat CE©ae: coal ‘ soSo 3a;ys men”. ge aw wm gLLgO™ Es ,ipSBS ae . "4Se’ i 4 ne43 :bin Boe, Bue a oo,er FJ oF . A. o% £ 6MS > YeeAin ¢ : - —inten ow: umtat *. f éOa WE as: +X we ; Sakae * dive ra ee©aefF, g en e ee, eee eee b Py fy. Bd pews o£. oeeeee

; , «3 .‘ Kd - NG Pe IEG calig gon? hehe BeI ED, a ;eewe ee. Roy tog : Ee ae a. ¥ ae Susear,” x We, AGEEEE GE age, ON Sea oS ie A 1 OSS oe Om me§. rer = Lot Be Sebe kM Lazty“alaebas “ +OLE: BG. SSaeome

@RE Boh | ASoaBRAK: | ahen SEBEL GIREeDE. LLNS Ree the: 4 . ieOSBE Ge“an al -.TSae1 Ce ga45Paa 5 E» P.we ‘ye na:oa. UEP toe SRC ae a Ot Seg Peaks STM Seek © pteteen ak. Tees. Peet ROS ae ORY - CoE ; mS z‘3 See aGAS RG fegaatratg DMR aces Te Se Ait23. wertronn og oaBREE 2Sere RRORE ot AA4. . ie, sone $a tceeness ee oe Fees ter eM ye GIGI ETERS ig ee Ea. ek, ieue 45 eeeogee aes ER yehoe 2tae ie ":; rane ae TESA %Tg ey RS: 4. ane . eeewan Se otBSS WENGE SeBe 5ees: iCelie: CreSn ane oe |eeEE ey Ee ‘orei Samee. AK’, BO “ wae hie ipos pera SEBEL. we AES, Beige Od a a ~4

ONE: we te ie OOF es ae om. gg Fs ae oT wl 4&5 see ek ey fe a La: Soe 38 nS “¢ * ’ ‘ott NR ee . : . a ae: | ra 3 * 0g, TPF E | i ee j : 7 Ee: 7 ge a 5 ai ow ot as , me Doves ee a 4 PAK neo. eg, er Fp FF: Mee” 2 SO ., 5 : j Re, oo Ve a. ” & . O™ : ee ’ MG ee2Ss™ LM eee, : rr ae ot: ; Rg Ce Be ae a. 2 « T . .:“Sage tte ”°: te :2:,_ eee oo 4 Pgh aps, ego ° ~ ‘Pee part cate ; ahs wn St 4Fsa:.&7os . .aoo. - 1 My. re nd ae ' Silane seen a chee A Cal te fir oF ~~. we 5 . Fb Me Sa it Se ny EO cm. £8 ee pg ee ~, :

ys staal aceon ie MSs ME OE PEE ee , .MM a ae ons ‘ ™. ag oe GE Sead bg tng gee t aes mn ee a a as RAO RE. ae IO Fe ; SS a we ? ad PRG RE a 2 OE Bee Ea Se . +, _ teas son, ante pages ; ME, a Ps :.ha ‘aeor 5:aBe ”Bas adsOER rcs oe Big AE3ot pogo “ge an eT OF Sak ce wis dl . OI “a Th ~- oe on Mp . : rs Bas 4. bb ERE TA Zee oF OR eR os og a an “ #-~ A : os : . Gy ge SBE eo citi oe MR ES SOR: . Fe Ly, $8 he: ie of he -~ — © ae !ie . Po PS6OE PR my Bees TREae GREG : Gi: ke:eo ed,” onBogie wn RE EET AS.nee: OS ORS Be$7pe . nan ; we & a -2 ee ogy. tySouk. (ele CaS lShiiONES pao iyi BY CS ee gee : . a2 mn ov Ooi sad men — ae we a Dee | ONS 4+ re BR OI Pe i Yee ies Gee # aa x ee y ,&¥ ca ¥ os *§ fe* ng ae .rrabi ws ‘ BS : £ ema MOS faeoom 3S ae aTEE fet Sass

eye i on SA ee Uy; Es -¥ s Eee. ey Spe 8G Ea Ce ee Baa anae :PRS ; Symi. hslon ‘ —s € % .éRY, . .“A ,2id & ees’ ghee _* , rg ee. Ge IY § OES Wee Gn OE dS te car . ~ —_—* :I2.~* _Doe / ° aSme OA a5EO es eS Be fiLe Me hee Leia eae ‘: ko: ae 5. gh *Me 9NT Bs “..3ue 5aan :‘...,°; “— ne oS nlyhey. oer SoS, aay 2gphe TORE ey ee Sees OS Fey EES Ss Pay? 4, "ER ny »Ee 4ie oe ~“os .5 rd{ SSO SN ON! ee gion: Bie AE RG Ges se Cag NER eG tue. ay. eT Ee LYRE SEF RGSS: i4a’“Spy Wee a 3 BMG w_— .$,oj SERGE aior TESS Bago ar is icau ® wg aden at Te boy . ee ae i,AS bah Lig *EES Sa ve 7co €6% ;a‘§ 1é i “ vf MI ASAP BIS ee. ie ¥k .py es Jte* NY ie wee ‘ ye . a 2*.Be 4~4 ei 4 2 of ky é BES Re ogee . a } * oo * paw. Ree re: $ : rd Fog id

“* ee a oe ha had — , ‘ ‘ tet >

ee ek Gane): aosee ie ‘ :Me cae! *ye ~ dee gM “ oye fua ~, a.—.ys Cshaut. -_ ~ ..a: yo . eae| ;Re Fy3ls . Mee . ug in oe. oe * ron mine ae: ; . ~ ™, . i : Sha ‘é . la “ y 5 re . . ae an . « . na ws, as ; ; a : we to, oa i tt Palld Lat os a . es oe ¥ ~. 5 . ° * 1 ‘. ¥ Ye mo ; 7 : , + ’ . * & , * ies be : . . “ eo aaa i

3 “ue ~~ x NY pee ol “, ’ caren “ be - . ie yee ey * pow Mee ON, i i ae. ns tyr OO ae ae : : yx ,4“I ee, Ki

‘a ‘. ow h * oF Sent mT s * ‘ . 5 . Mean aa ooo x sp oo ~ Na. : 4 La i © fy, so nd] eee aah hr. Ehoe aRo ty NO :Pom ee ee 2: Fr prs . -Ft ae -aa’ bie . ,“a 1ee yas eet . ~*te iyoh . “ANAS. a ES oy->oo As:~ 4tet ean_— xesa, bl oe pe iasgl. LAO ° ‘, « Dial vaice 22 : .7 a }% ft i % aol a -eo . ranaMe i 4 7 ~i 4- ‘3idt%ette .it.a/~ae .. ‘; «

Baw Sitae PB rye 4—4,*t‘.Soe ey me Many iad ne Lis ' am wad 4,Se? .\RY 7 Seo 7Ge4 M ~ hk ~"“ha«ys 4oor ayLa 1 ghee teoh “Lo : ‘aae ,. a= ~“-~“oe ET Na anne ran onNy SNe oe oa Te eo way, ne : ope Re 5 2 :, an Wigg sts ie “oy a “ye . ; aaa . a . * 7. tps ‘me a thes 4 a ae ; .

OT Rata Eee oe. a BA cee ee _.:” } a ek N ie \ sy 4 \ i Wa a “ her _ 4 v. 5 ej . .>Me. ein f) at * tame ig ~, * oy we . a . “ + ener a. ; . ah ae F a. .at: . . al Boistice | 9 ce ae _*. . “y “y a Se al er an La we! .. ., a . a ; oth \ : —_ > * =. . ~ . + . "eo 7 * 7 me, “ ao om . . :, 4ee .a ..¥ 4, ;ge “% we \ywe . va Pe Neaeeehy > .~. peaa aw, m af : ie ; wo ag~« ¥.eeee + oa :ia.oe .. .oo .t gai a re: sh anaes x 4 ole: a =r oo + %, ano reung art ie: ~ ae : me . . eo Aer | ' . fay ihe F ae “Se a 2 5 td +3 irl St a nn a. “a ‘ iol‘See ¥s 2 ree *ayeaoseat a” * +nore iy um ayaa ‘Wyoe un La “agspe ee carte 1, eae .ba.. Coon ; . a : d . “a . ee Shy fer % eat} we *: P. cP -~ ‘ a8 ! oe aren af “ 'are mye ;te. Pe :~. ey vf th: oe . a: zBR hefeSzYeme Bee " oR eye a. > 7". -\ -* mt¥"gienfeck i + ': .\ .Fy ‘hea:. fi ag*:mT x . 'ot ; : 4:ni hg?| ie oefs; ;— at Ja

ue bin : .ad fteer ‘ tsf.+" A oo ’ JS "AL : %aes a oy E ~i.. Ly Xo 31 . ihyee ne ‘a ’ . .}a :}atiAN Pee . ili eesey ~s.Ne , eaah «8,F4% yyAp‘AEE 4afaa4 hy 7 ioe On,: ’See hy Nk“ora: i7aR + Bg CE ite Stylesk yA CREE: ia |

Be ee \ o Shee SRE he BS j ete ee SP, mo, _ ° ay: aFis Seat gl aeH x=k - TOA . :tT bksO~. aee. aw Ay ia‘:4:a:-: . ar a “he! 3 NG: riage een: Sd . ie, ECA ens. i AL ee me Re oe i EER ee. ; : ol aBo oa A ay CA A a gtoeRE v y3 i oar ee.avs. gee eee eeae7 a‘ wo ; .*‘° ‘ igh BS : * Ag eis, ie . cae tir ™ : . 3 a ~~ “ * a . Me , tg ro a NE % td o ‘ . ey ’ ™” ofa y ‘3 4 * +t ais wae : a » a : ae - oo 3 . a ae oe a _ . ° H nA . WE, %eo 47one oe TS, a.,oy we i Te ys Bo et.ei:etaCorea thytie 7% ys,aS=nag Me en+os a. : ae aie: scanBaa " ~“*og eeari ‘nes, yea : cers Raga US: .Pe . oaA .EY ee hc eae ne. nt were a. al ae om a ..“os BB iw aes ine Pa. eee anf oe . :_oS > oie imam me VO ia“hy wR. 4 } am eAe Le ‘;.oo .

a ty a ae / tata o4 a Bs 2mae we,:a1 :te; :2S ahe! . Sas nhRes, . a, TS : aean:

yo 5‘ ihaan a ean ‘Sh ers ai ne Se io ata }~oot wre Pee. oaue nyoe a ae— i vekh ae at2. val Me . t.3 *hone o “e Nps :“Ba 4 Rta os aeead 44 as t“S tN , :~ Oe ni &. -* _.. “-9ok ~ ae ,.y ° Me wo owe J 4 . Witte oy % " 7v as yas Sy \y ry. . % ? aac ae ™ net : a Sale dea 1, ’ re ie as, by fy pos .

PR 7riswee ides Ce ee bo .Apa wee « : 4 ee NN A . . : . pee ayyy ‘ot oeTENE a, ee .;2ar’ Ry ets, rs % a. % . ee ot 48 , ; 2 . Sih area we * 4 # ak, se ge KG ; . OY FN ee: 2~~ Ny De ER EU ROR neé .okey ot*1No : “4etny oanpence om My . e. ie - 4Oy .ek ay. . ;a. wee On Mena ye | EES- oy ; ae ze S , swe? a: ao4. “Raper. vee ms Bet+4,es4 raafoo,ee

. se ee Ales ; * ar’ . . . ed 0 ne PA on ae Se oe ;

or . . *. »igfere i, Mee . ; : . aon ‘ie a 7hs. fo ef i .ctt. C . **,Fd4 ;iaes

Cat fom, ey 7 nie ne “ati. a a ‘hf , rn, i ae en ¥ » UTES 8 > i ee ‘ Loe :he Nie a sy eS Sake ; ay aaa a. ae . ‘ a a a Me tg Fo ; a . og pie are ”7 or) wex:ee, Wee oe,. etaaa2~vo * an Ea ; _# a ; eo ao — ; . Mes Mh, amots .i2 "it te i pasmnaa aes. :ncay : ee .Pret,

| |mg. a age -tae “i s

Han Turali rides again

Digenes Akrites was ‘discovered’ in May 1868 when a monk of Soumela lent a manuscript of the poem to Sabbas Ioannides, who

sent it on to Paris where C. Sathas and E. Legrand published it in 1875. Ioannides himself re-edited this version in Constantinople in 1887. David of Sassoun was ‘discovered’ in 1873 when Karekin Servantstian (1840-92) took the story cycle down from

the lips of one Gurbo on the plain of Mus and published it in Constantinople in 1874.' Thus the two great medieval Christian 1. This Note has a vast and quite turbulent bibliographical background, which would dwarf the text to little purpose, so I am limiting references to items which are relevant to its point (apart from a funerary excursus in n.6). I am grateful to David Ricks, Bruce Lippard and Michael Ursinus for advice. For literature on Digenes, start with H. -G. Beck, Geschichte der byzantinischen Volksliteratur (Munich 1971) 48-63. I fancy that the ‘discovery’ of the poem in 1868 may have been the result of a rearrangement of the Soumela MSS inevitable on the

building of a new library for them after 1864: see A. Bryer and D. Winfield, ‘Nineteenth-century monuments in the city and vilayet of Trebizond: architectural and historical notes: part 3’, ‘Apyeiov IIdvtov 30 (1970) 276. After 1887 Ioannides deposited the MS in the library of the Constantinople ‘EAAnuxds ®uUWodoyiKxdc¢ LvAAOYOG; the frequent assertion that it is now ‘lost’ seems to be derived from an

opinion expressed by S. Kyriakides in 1936: J. Mavrogordato, Digenes Akrites (Oxford ; 1956) xi. In fact, after 1923 most of the Soumela library was transferred (eventually) to the library of the Tiirk Tarih Kurumu, Ankara, including other Pontic MSS such as a cartulary of Vazelon monastery, which the TTK preserves. I see no reason why the Digenes MS should be lost: has anyone tried asking for it at the TTK library recently? For English readers, David is most accessible through two complementary translations: by A.K. Shalian, David of Sassoun: The Armenian Folk Epic in four cycles (Athens, Ohio 1964), which is based on an ‘official’ composite text published in Erevan in 1939; and by an Armenian native of Trebizond, L. Surmelian, Daredevils of Sassoun

(Denver 1964), based on a single text.

‘epics’ of the Anatolian borders, Greek and Armenian, the geographical and historical backgrounds of which at least overlap,

were ‘discovered’ and published almost simultaneously. There is a more tantalising coincidence. Ioannides was &s5d0KaAoc and first serious Greek historian of Trebizond when Servantstian was Gregorian Armenian bishop of that city. Yet I do not think that either refers to the other. Such was the mutual exclusiveness of even the Christian communities which co-existed cheek-by-jowl in a place like Ottoman Trabzon that it seems quite possible that

the Greek didact and Armenian prelate never met, let alone discussed their great and comparable findings. It is more surprising

that, with notable exceptions, this mutual exclusiveness lingers a century later among students, not only of Digenes and David, but of other Arabic and Turkish ballad cycles, heroic poetry, epics and romances of Asia Minor. It is true that approaches to the genesis and transmission from memory to writing of such material has been greatly refined — for example, in the case of Digenes, by Roddy Beaton in this

journal.? (We are now faced with problems of transmission from writing to memory, for Dede Korkut has been reduced to a Turkish school textbook; in the same way it was publication of his deeds which finally killed the Serbian Prince Marko Kraljevic where the Turks had failed in 1395). Many of the Anatolian stories reveal more than one layer of ‘historical’ context in which later names may be attached to earlier tales before they were edited

in the middle ages, recorded from the last century, or both. The whole corpus is heterogeneous in genre, but shares in common the interactions of the Christian and Muslim peoples of Anatolia as its matter. Examination of interconnections within the corpus as a whole has been hesitant — if only because it would take a polymath such as Henri Grégoire to master it all. But sightings of the willo’-the-wisp of topicality and ‘historicity’ were at first bolder, when

Grégoire himself was not the only one to burn his fingers in that 2. ‘Was Digenes Akrites an oral poem?’, BMGS 7 (1981) 7-28; but see now S. Alexiou,

Baciiswsc Atyevic "Axpitns (Athens 1985). If Byzantinists wish to test the versions of Digenes against a more-or-less genuine oral poem, they should read a single version of David of Sassoun. 194

ignis fatuus.> It is out of fashion. For example one can learn rather more about medieval Georgian feudal institutions from the works of Wakhtang VI (d.1737, a sort of latter-day Constan-

tine VII Porphyrogenitus), than from the romance of Shota Rustaveli (Supposedly b.1166, whose Knight in a Panther’s Skin

passed, like the Cretan Erotokritos, from writing to memory). But on the whole the Muslim material, from Seyyid Battal through

the Melikdanismendname to the gesta of Umur Pasa (which pretends to be ‘history’ anyway) is more susceptible to pegging to ‘actual’ events than the Christian.* Yet who can deny that,

in the same way that there is something called ‘Homeric archaeology’, such texts as Servantstian’s David and even the Escorial version of Digenes offer through the distorting lenses of their genesis and transmission a glimpse of a frontier ethos and its realia which is at least more vivid, and archaeologically recognisable, than any Escorial Taktikon?>? The question is 3. E.g. in papers in H. Grégoire, Autour de l’épopée byzantine (London 1975). There

are further hazards. I do not know whether to be gratified or alarmed at the way my first undergraduate (but still wary) foray into this field has been taken up: ‘Akhtamar and Digenis Akritas’, Antiquity 34 (1960) 295-97; H.M. Bartikian, “Notes sur l’épopée byzantine ‘‘Digénis Akritas’’ ’, Revue des Etudes Arméniennes 3 (1966) 166; and G. Huxley, ‘Antecedents and context of Digenes Akrites’, GRBS 15 (1974) 332-33.

4. On Wakhtang VI see D.M. Lang, The last years of the Georgian monarchy, 1658-1832 (New York 1957) 32-48, 118. The latest English version of Shota Rustaveli

is Katharine Vivian, The Knight in Panther Skin (London 1977); and French S. Tsouladzé, Le Chevalier a la Peau de Tigre (Paris 1964). Whatever its date of composition or even authorship, which are now, not before time, in question, the romance is of such a literary nature that social conclusions drawn from it must be treated with the reserve one would put on the Erotokritos as a ‘source’ for Venetocratic feudalism: an example is Nino Salia, ‘Le poéme médiéval Géorgien’, Bedi Kartlisa 19-20 (1965) 15-30. On Seyyid Battal, see S.P. Kyriakidés, ‘Eléments historiques byzantins dans le roman épique turc de Sayyid Battal, martyr musulman du Ville siécle, est-il devenu, dans le légende, le contemporain d’ Amer ( + 863)?’, B 11 (1936) 563-70, 571-75; and (for a local interpretation) M. Aslanbay, Seyyid Battal Gazi’nin, hayati ve bazi menkibeleri (Eskisehir 1953). Admirably, Iréne Mélikoff (-Sayar) has edited, translated and commented upon both La geste de Melik Danismend (Paris 1960); and Le destaén

d’Umir pacha (Paris 1954). '

5. The question is faced squarely and sensibly by N. Oikonomideés, himself ‘discoverer’

of the Escorial Taktikon. Compare his Les listes de préséance byzantines des [Xe et Xe siécles (Paris 1972); and ‘L’ ‘‘épopée’’ de Digénis et la frontiére orientale de Byzance aux Xe et XlIe siécles’, Travaux et Mémoires 7 (1979) 375-97. For ethos, cf. J.F. Haldon — H. Kennedy, ‘The Arab-Byzantine frontier in the eighth and ninth centuries: military organisation and society in the borderlands’, ZR VI 19 (1980) 79-116. 195

where to draw the line, which Grégoire so gamely overstepped, in seeking ‘historicity’. It is hardly a new problem and is exacerbated from the start by the random way in which history sticks to places. For example, when the Hospitallers began building their castle at Bodrum in 1407, local Greeks and Turks apparently had no name or legend to account for their colossal quarry: unknown to them it happened to be the Mausoleum at Halicarnassus, which was still being listed as a Wonder of the World.® But less then five years 6. A. Luttrell, “The later history of the Mausolleion and its utilization in the Hospitalier castle at Bodrum’, in The Maussolleion at Halikarnassos, ed. K. Jeppesen, II (Aarhus, 1986) (= Jutland Archaeological Society Publications, XV:2), 133-35. His P1.111, of MS Holy Sepulchre, Jerusalem, 14, f.312v, dateable 1066-81, shows the Mausoleum in the form of a ciborium baldachin (four free-standing columns on a square marble enclosure carry a dome). This Byzantine type of tomb was evidently shared by that of Patriarch St. Athanasios I of Constantinople (d. before 1323): AliceMary M. Talbot, Faith healing in late Byzantium (Brookline, Mass. 1983) 14, 56-57, 126-27. Another example of this type, the tomb of the Grand Komnenos IV (d. 1429 and incidentally father-in-law of han Turali’s great-grandson) survived until 1917 and is illustrated in my ‘The faithless Kabazitai and Scholarioi’, in Maistor. Classical, Byzantine and Renaissance Studies for Robert Browning, ed. Ann Moffatt (Canberra 1984) pl.3. The last built of the type known to me are the canopied tombs of King Solomon II Bagration of Imereti (d. 1815) — see Bedi Kartlisa 25 (1968) 214-17 — and of Metropolitan Konstantios of Trebizond, who was buried sitting down in his surviving tomb in 1879: see ‘Apyeiov ITdvtov 29 (1968) 103-5. The practice of enthroning dead kings and prelates may have given rise to this sort of tomb and persisted until this century: see the photograph of the funeral procession of Patriarch Ioakeim HII (d. 1912) in H.G. Dwight, Constantinople. Settings and Traits (New York and London 1926) 507. A.L. Yakobson reports another type of tomb from the Crimea, of stone-cut models of triple-apsed domed churches, or simply of their open apsides in a sort of shrine,

which may have Armenian and Georgian parallels, and for which late Byzantine antecedents have been proposed: see A.L. Yakobson, ‘Model hrama iz raskopok Eskikermen v Krimu i problema novogo architekturnogo stilya v Vizantii’, Zograf 8 (1977) 30-33; A. Grabar, ‘Observations sur |’ Arc de Triomphe de la Croix dit Arc d’Eginhard et sur d’autres bases de la Croix’, Cahiers Archéologiques 27 (1978) 77-83 (for which references I am grateful to Dr. Zaga Gavrilovic); and A.L. Jakobson, ‘A propos des relations entre les régions littorales au Nord et au Sud de la Mer Noire’, BS 42 (1981) 43-51 and figs. 1-Sb. However, all examples Yakobson has so far published look to me to belong to a common nineteenth-century (and I think no earlier) Pontic type, with which some have declared epigraphic links. A similar one was labelled Byzantine in the Council of Europe exhibition in St. Eirene, Istanbul, in 1985. Comparable examples in their homeland are illustrated in S. Ballance, A. Bryer and D. Winfield, ‘Nineteenth-century monuments in the city and vilayet of Trebizond: architectural and historical notes: Part 1’, ‘Apyeiov [1dvtou 28 (1966) 264, pl.34. Nevertheless

it is clear that late Byzantine mausolea and funerary practice await their student. 196

before a German mercenary prisoner of Timur’s war identified Caleoglu Kale (a minor if older Anatolian monument) on the spot as the distant Templar ‘Castle of the Sparrowhawk’ of fourteenthcentury Provencal fée romance. Who knows? Schiltberger said he heard (but may even have inspired) a local Greek legend about the castle, which was held by a beautiful but bewitched maiden who, like all good Pontic princesses, lured suitors to their fate; Sabbas Ioannides himself recorded the same story there five centuries later.’ Here the snake of myth had by 1402 already bitten its own tail to give subsequent twists to the tale. But there are too many distressed damsels and black brides in Anatolian castles

as it is, and no version of the legends in which Caleoglu Kale is enswathed contributes anything to its archaeology, any more than one can reconstruct, say, a revolutionary Paulician manifesto from Digenes, or date the life of its hero (as Ioannides did) to precisely 936-69. Yet we must push on — warily. First: it seems to me that the most useful way to revive the problem of ‘historicity’ is to stand it on its head and ask why apparently ‘historical’ events stuck in the memory before trying to peg them to ‘actual’ happenings. Second: we ought to keep in mind, now that it is more widely accessible, the whole corpus of material, in a debate which Sabbas Ioannides and Karekin Servantstian failed to start in Trebizond in 1868-74. These are large injunctions and this is a modest Note. But it may illustrate the problem anew by pulling together, and adding material to, a famous event: the first ‘marriage’ between a Byzantine 6€ox01va and a Tiirkmen (as opposed to a common, or garden, Turk) han, which was probably consummated somewhere between Trebizond and Sinir in August 1352. The ‘actual’ event is recorded in Trapezuntine chronicle, to which can be added a A final and understandable misconception is that the sculpted rams found outside such Armenian churches as the Twelve Apostles at Kars, or at Varzahan (midway between Sinir and Bayburt) are Akkoyunlu monuments: in fact all appear to be Armenian tombs. (The frontispiece of R. Curzon, Armenia {London 1854] may be iden-

tified as of the now lost example at Varzahan; the original drawing is now in the collection of Francis Witts Esq., of Upper Slaughter). 7. For the large bibliography of this small site, see A. Bryer and D. Winfield, The

ale» pl.29. monuments and topography of the Pontos (Washington D.C. 1985) I, 103-6; 197

reference in the Aan’s biography. A similar event is reflected in an Oguz ballad cycle, to which I offer what may be an echo in

an Armenian folk tale. Some, but not all, of this Greek, Persian, Turkish and Armenian material and background has been discussed widely in print already, which will therefore be summarised only for new readers who start here. The Four ‘Sources’ I. The twelve or thirteen Turkish ballads, legends or stories,

associated with Dede Korkut were probably edited, perhaps around Tabriz, perhaps around Erzurum, perhaps in the fifteenth

century. The two sixteenth-century MSS, in Dresden and the Vatican respectively, have only aroused serious study in this cen-

tury, particularly after the publication of the Vatican MS in 1952.8 The immediate context is Tiirkmen life in Anatolia in the

fourteenth century, but there is a substratum reaching back to a time before Dede Korkut’s Oguz heroes crossed west over the wandering Oxus. On the whole, ballads based on a perhaps eleventh-century past, and those (like Ballad 6) concerned with fourteenth-century encounters with Anatolian Christians, seem to be discrete. Ballad 6 is well known. Han Turali, son of Hanli hoca, our Akkoyunlu (White Sheep) hero, seeks a bride. Only the tekfur of Trabzon has a paragon princess: her name ts Salcan hatun and she can draw two bows at once. Turali rejects the warnings of his father against venturing down from the windy yayla, where Tiirkmen graze their flocks freely, to the forested castle-lands of the Euxine coast, where infidels farm. Han Turali rides into the meydan of Trabzon to claim his bride. But first the tekfur sets him three tasks: to down a royal lion, a black bull and a raging camel-stallion, three monsters who have already accounted for thirty-two previous suitors; their heads hang around the square where Turali warms up in front of the tekfur on his throne and 8. F. Iz, s.v. ‘Dede Korkut’ in EF has a bibliography up to 1958, to which may be added (besides items mentioned below): O.S. Gékyay, Bugtnkii dille Dede Korkut massallari (istanbul 1943); K.M. Fahrettin, Dede Korkut, Oguzndmelevi (Istanbul 1952); V.M. Zhirmunskiy and A.N. Kononov, Kniga Moego Deda Korkuta (Moscow and Leningrad 1962); and H. Korogly, Oguskiy geroicheskiy epos (Moscow 1976). 198

the princess in her palace. Our hero is stripped mother-naked; he even unveils himself. At this point the Aatun ‘went weak at the knees, her cat miaowed, she slavered like a sick calf. To the maidens by her said she said: ‘‘If only God Most High would put mercy into my father’s heart, if only he would fix a brideprice and give me to this man! Alas that such a man should perish

at the hands of a monster!”’ ’. Turali disemboweled the bull; there were alarming scenes in the meydan when ‘The lion roared and every single horse in the square pissed blood’; and the Trapezuntines tried to nobble the camel. Turali won his princess. Folding his tent, they rode off for seven days to the pastures of the Oguz. But the tekfur of Trab-

zon repented his loss and sent six hundred of his warriors to reclaim his daughter. She, but not Turali, was ready for him: the hatun cut up and routed her father’s own army. His manhood thus slighted, Turali could only challenge his bride to single combat. She shot a headless arrow ‘‘that sent the lice in his hair scuttling down to his feet.”’ They were reconciled, married by Hanh

hoca (who was also reconciled), and lived happily ever after.’ II. The Armenian folk tale of The Fox, the Wolf, the Bear and

the Emerald-Bird was taken down for the first time by T. Nawasardeanc‘ in Vagarsapat (Etchmiadzin), probably between 1876 and 1882.!° There are two themes, perhaps two separate stories. The first relates how an Armenian prince left his father and, during a famine, befriended in turn the creatures of the title. At the end of the famine, the Fox said: ‘This man has kept us all these years. Let’s do him a good turn . . . He is still single. So let’s go find him a wife. I have heard the fagavor of Trapizon 9. Cf. G. Lewis, The Book of Dede Korkut (London 1974) 119; and F. Siimer, A.E. Uysal and W.S. Walker; The-Book of Dede Korkut. A Turkish epic (Austin, Texas and London 1972) 101. | summarise because I have gone over this before in ‘Greeks and Tiirkmens: the Pontic exception’, DOP 29 (1975) = The Empire of Trebizond and the Pontos (London 1980) V, 119, 134-35 (where the question of dowry/brideprice is also raised). 10. | am most grateful to Professor Charles Dowsett of Oxford for tracking down the published origins of the story translated by L. Surmelian, Apples of Immortality (London 1968) 221-24 as T. Nawasardeanc’, Hay zogovrdakan hek‘iaf ner (Vagarsapat 1882) II, 29-37, reprinted in H. Orbeli, same title (Erevan 1959) II, 254-59. 199

has a very lovely daughter, a praiseworthy maiden in every way.’

The others cried: ‘Yes, let’s go to Trapizon!’ In Trapizon the king’s ploughman was enjoying his siesta. The Fox yoked the Bear and the Wolf to his plough. The tagavor and his daughter came out on the palace balcony to see what was going on, when the Emerald-Bird snatched the bride. It flew her back to the Armenian prince. The king sent his army to retrieve the princess. The four creatures routed it and the tagavor of Trapizon relented. The Armenian prince was reconciled with his father: he and his Greek bride lived happily ever after. III. Michael Panaretos (d. after 1390), the laconic Greek court chronicler of Trebizond, reports that on 29 June 1348 a Tiirkmen coalition, including Mayyat Eixkentapic (Mehmed the Rikabdar, ‘stirrup-holder’) of Bayburt and TovpadAinex (Turali beg) of the “Auiti@tat (Akkoyunlu) attacked Trebizond, fought for three days, and fled injured and crestfallen, losing many Turks on the way. Next year the Grand Komnenos Alexios III (1349-90) succeeded to the throne. In 1352: ‘the Emperor’s sister, lady Maria

the Grand Komnene, went away to marry XovtAovunéKnc |Fahreddin Kutlu beg], son of TovpaAf [Turali] who was emir of the "Apitwtor |[Akkoyunlu]|, in August.’ They appear to have lived happily ever after. Maria revisited Trebizond in August 1358 as Seonoiwayat (despoina hatun); Alexios III built the massive Kovxoe castle (Kog Kale) on his Akkoyunlu borders in 1360 and

was prepared to help Kutlu beg during the summer grazing of 1363, but joint Greco-Ttirkmen exercises were called off because

of the bubonic plague. But the emir paid a state visit to the Trapezuntine diplomatic hospitality compound above the city in July 1365, and the emperor returned the visit in 1367.!!

IV. Abu Bakr-i Tihrani wrote his account of Kutlu beg in Persian in about 1478, the year of the death of his patron, Uzun Hasan. Uzun Hasan, the greatest of the Akkoyunlu emirs, was

great-great-grandson of Turali and husband of the Grand Komnene Theodora the despoina hatun, who was great-great11. Michael Panaretos, ‘Tlepi t@v MeyaGAwv Kopvyvev’, ed. O. Lampsides, ‘Apyeiov

Iidvtov 22 (1958) 70-76.

200

granddaughter of the emperor Alexios III and was.buried in Diyarbekir, where Abu Bakr wrote his account. His information on early Akkoyunlu leaders appears to be derived from oral tradition (in which case he had access to both sides of the story), but it has been suggested that his biography of Turali’s son Kutlu beg may contain material from the life of Turali himself.'? At any rate Abu Bakr’s official revision of events has Kutlu capture Tisbina (5€onowwa), daughter of the tekfur of Trabzon, as his bride. Later sources endow Kutlu with great piety and zeal: he was a very perfect gazi against Trapezuntine and Georgian infidels, went on pilgrimage to Mecca 39 times over, and founded the surviving mosque at Sinir, where he was buried in 1389, the year before his Christian brother-in-law died. Sinir (now Suntir!) is Luvopia, as its name suggests, a border place, only 33km. south-

east of Alexios III’s new border castle of Koukos.” Discussion

It might be wiser to leave these four ‘sources’ at that: but let us ride on with Turali — warily. Opinion is so far about evenly

divided about linking ‘sources’ I and III: Dede Korkut and Panaretos. Geoffrey Lewis, author of the most felicitous English

translation of Dede Korkut, puts the strongest and most commonsensical case against seeking topicality and ‘historicity’ in these quicksands, so I will take it up. For example, he points out that in Ballad 3, Bamsi Beyrek of the Grey Horse cannot have been incarcerated in the ‘infidel’ (i.e. Christian) castle of Bayburt for sixteen years, because the place in fact fell under Muslim rule from the late eleventh century, just when Dede Korkut’s heroes were arriving on the scene. This is literally true. It was the melik 12. Abii Bakr-i Tihranj, Kitab-i Diyarbakriyya, edd. N. Lugal, F. Siimer, I (Ankara 1962) 12-15 and 90 (a memory that the then sultan of Trabzon — perhaps Alexios IV of n.6 above — supplied siege equipment to Kara Osman [1403-35] son-in-law of Alexios HI, at Erzincan); J.E. Woods, The Aqquyunlu. Clan, confederation, empire (Minneapolis and Chicago 1976) 46-49. 13. On the importance Sinir, see Elizabeth A. Zachariadou, ‘Trebizond and the Turks (1352-1402)’, ‘Apyeiov [Id6vrov 35 (1979) 335, 339, 349; and A. Bryer, ‘The question of Byzantine mines in the Pontos: Chalybian iron, Chaldian silver, Koloneian alum and the mummy of Cheriana’, Anatolian Studies 32 (1982) 136, 144-45: it may have been Alexios’s silver rather than his mummy that Turali was after. Woods, 48, 239; Bryer and Winfield, Pontos, II, x-xi, 308-10; II, pls. 249-51. 201

Mugith al-Din Tugrilsah. (d. 1225), son of the Seljuk Kilic Arslan

II, who largely built the awesome walls of Bayburt in 1213 and we have seen that Panaretos records that Turali was himself allied

: against Christian Trebizond with a Muslim ruler of Bayburt in 1348. But Ballad 3 may well be right in reflecting a higher truth of popular belief that Bayburt was then still ‘infidel’. The melik, who had the misfortune of being prisoner and sort of vassal of first a Cilician Armenian king and then a Trapezuntine emperor, and whose son was baptised to marry a Georgian queen, evidently allowed (or even sponsored) the building of a surviving Orthodox church within his new citadel on whose walls he is still proclaimed a gazi. Here Armenian colophons attest Christian life (including Gospel repair) in the 1340s; as late as 1530 Ottoman Bayburt was 77% Christian in households. Its last recorded Greek bishop was appointed in 1633, but its Armenian bishopric (for it was largely an Armenian place) flourished until this century.'* Whatever the

period of Ballad 3, Bayburt was irremediably ‘infidel’ to Turkmens in its surrounding pastures. It is thus unnecessary to argue with Lewis that this section was either composed before the twelfth century, or that ‘Bayburt’ replaces the name of some earlier city — which is as well, because these Turkmens would have encountered few ‘infidel’ citadels before they reached Bayburt. Before tackling the topicality of Ballad 6, its folk motifs must be unravelled and disposed of. This is more than usually tricky in the Dede Korkut cycle for, by coincidence or inheritance, they are unusually wide. Some are commonplace: for example Delu Dumrul’s struggle with Azrail, Angel of Death, in Ballad 5 follows the lines of Hercules’s or Digenis’s (in the ballads rather than poem) encounter with Thanatos — more interesting is that the story in-

corporates that of Admetus and Alcestis. As Grimm himself noted, most striking is Ballad 8, where Basat and Tepegoz reenact the tale of Ulysses and Polyphemus in Odyssey, ix, with a Byzantine twist: Tepeg6z is the 40-cubit Lapavtannyzoc, a Cyclops who reappears in Armenian folk tales.!*> C.S. Mundy 14. Lewis, 18; Bryer and Winfield, Pontos, 1, 352-55. Woods, 47, 238 sits on the fence, which Siimer, Uysal and Walker avoid altogether. 15. It was, of course, a TecoapaKxovtannyus who sowed the seeds of Iconoclasm: on such Greek giantry see S. Kyriakides, ‘EAAnvixn Aaoypagia (Athens 1922). 202

favours an enduring repertoire of Anatolian tales from which both borrowed rather than a direct Homeric link with Dede Korkut.'®

Whatever the answer, the appearance of an Amazon in Ballad 6 should hardly surprise in such company — it would be surprising if Amazons did not in any encounter between Pontic Greeks and Turkmens. The Amazons whom the Argonauts met there were

native, and endemic, to the Pontos.!'’ They had already figured (naturally on the Greek side) in the Melikdanismendnaéme."® The motif of han Turali’s Herculean labours needs no comment either, nor would that of the pastoralist’s romantic and actual addiction to bride-snatching, found in Digenes and many other Oriental and Western traditions — were it not for the fact that it is inherently tricky to snatch an Amazon bride.'? The themes contradict. David of Sassoun came across the same problem when he encountered Hantut Aafun in the Armenian epic: “Two or three years ago,

They took me to a prince, the son of a king. They put me in a room, The prince came in, We wrestled playfully,

I grabbed his arm, the arm came off... Than I reached for the other arm, That arm came off, too; his back was broken,

He gasped and died... I made a vow that I would (only) marry the man who could down me. Today we fought; you threw me down. From now on I am your wife, you are my husband, Take me wherever you wish.’ 0 16. ‘Polyphemus and Tepeg6z’, Bulletin of the School of Oriental and African Studies 18 (1956) 279-302. 17. S.A. Nikolaides, “ApaCovec kai Mdévtoc’, "Apxzeiov évtov 26 (1964) 250-56.

18. Irene Mélikoft, ‘Les Géorgiens et les Arméniens dans la littérature épique des Tures d’Anatolie’, Bedi Kartlisa 11-12 (1961) 27-35; and the same’s ‘Géorgiens, Turcomans et Trébizonde: Notes sur le ‘‘Livre de Dede Korkut’’ ’, Bedi Kartlisa 17-18 (1964) 18-27. 19. Cf. W.J. Entwistle, ‘Bride-snatchtng and the ‘Deeds of Digenis’, Oxford Slavonic

Papers 4 (1953) 1-17. S. Thompson, Motif-Index of Folk-Literature (Copenhagen 1955-58), Motifs H310-359 (especially H332.1 and H345), H901.1 and H1161 are most

relevant to Dede Korkut Ballad 6. 20. Trans. Shalian, 323-24. 203

Geographically, Dede Korkut illustrates the Tiirkmen trauma of having to descend from the upland yayla to make contact with the ‘infidels’ of the coast more vividly than Panaretos’s catalogue of summer skirmishes.?! What is interesting here is the placing of the story in a Christian agricultural context in the Armenian ‘source’ II. If the story in Dede Korkut and this tale are related (I venture to each other rather than through a Greek intermediary),

this is its most telling alteration. The Armenian animals know how to use a Greek plough, which han Turali would not have deigned to touch. The Armenian tale and Dede Korkut share the notion that the ruler of Trebizond tried to regain his princess from the clutches of her suitor, and failed. This is more serious. Even Panaretos supplies evidence that not only emperors, but their subjects, found

such diplomatic alliances distasteful. He records that ‘some people came almost to rebelling against the Emperor’ Alexios III in 1362

when the celebi Taceddin sought a marriage alliance: he had to wait seventeen years for his Greek princess, and then after much parleying.2* Doukas expresses such disgust at the first great Byzantine-Ottoman marriage, between Orhan and Theodora Kantakouzene, in 1346.77 From the Orthodox point of view it was no marriage at all, while from the Turkish it could be interpreted as a form of vassalage: such hostage wives could not become an ulu hatun. Orhan and Theodora broke the barrier in 1346; in his Dustur-

name of about 1465 Enveri records the earlier courtship of Theodora by Umur, a very gentle gazi.*4 It made an impression. Trebizond took the plunge six years later when Kutlu beg married Maria Komnene: his son, grandson and great-grandson took wives from the same family in their generations. It was a deeper 21. X. de Planhol, ‘La signification géographique du livre de Dede Korkut’, Jour-

nal Asiatique 254 (1966) 225-44. | 22. Panaretos, ed. Lampsides, 74, 78.

23. Doukas, Istoria Turco-Bizantina 1341-1462, ed. V. Grecu (Bucarest 1958) 59; cf. A. Bryer, ‘Greek historians on the Turks: the case of the first Byzantine-Ottoman marriage’, in The Writing of History in the Middle Ages. Essays presented to Richard William Southern, edd.R.H.C. Davis and J.M. Wallace-Hadrill (Oxford 1981) 471-93. ney mar pacha (in n.4); P. Lemerle, L ’Emirat d’Aydin, Byzance et l’Occident (Paris

204

plunge, because, whoever the early Ottomans claimed to be, Turkmens like han Turali were not a smart match in 1352. Yet the alliance was the first of eleven in which princesses of Trebizond married Muslims, eight of them Tiirkmens. Who did it impress? I propose that it stuck hard in the Oguz imagination, and is maybe remembered in Armenian folklore. But perhaps more significant is that the substantial body of Greek Pontic balladry and tales has no obvious recorded memory of it. This suggests

that Turks were more pleased than Greeks. Finally comes the least important question: do all our ‘sources’ refer to the same event of 1352? Geoffrey Lewis argues against:

it was not Turali but his son who actually married the princess of Trebizond and anyway her name was Maria, not Salcan.?> I am not too worried about this. The Akkoyunlu were widely known -

as Turalids after their heroic ancester and Abu Bakr, their genealogist of ‘source’ IV, seems to have confused han Turali with his son too.”° I do not know where Salcan got her name from (it does not sound Turkish), but Abu Bakr got it right by calling her Tisbina (S€onowwa), just as Panaretos got it right by calling her -yat (hatun). In any case Georgian princesses with

Turkish epithets changed their names on marrying into the Trapezuntine imperial family, and presumably the reverse also occurred. Those seeking realia will indeed find that the surviving meydan square of Trabzon, which even Panaretos called the paitaviy,

was indeed where the emperor and his princesses displayed themselves at Easter npdxvyic before Tiirkmens, Armenians and Frankish traders, and that the heads of Turkish (rather than ‘in-

ftdel’) miscreants were brought there.2’ Did the emperor ever regret the alliance and rescue his sister? I do not know, but moving south to the borderlands between Alexios III’s new castle at Koukos and Kutlu’s base at Sinir, Panaretos cannot conceal small and unseasonable Greek raids on Tirkmen-controlled Armenian lands which came to grief in the January cold: in 1369 when the Turks broke their treaty and Greeks died in the treacherous cave 25. Lewis, 17-18.

26. Abu Bakr, I, 15; Woods, 46-47, 238-39. , 27. Panaretos, ed. Lampsides, 75, 77, 78.

205

of Golacha (Colasana); and in 1373 when 140 Christians were killed, some by Turkish sword, but most frozen in the snow.” In both cases the tek fur of Trabzon’s warriors may have been routed by his own sister’s husband. Otherwise, all four ‘sources’ are agreed on one thing. Whoever they were, han Turali and Salcan hatun, Kutlu beg and despoina

Maria, lived happily ever after. Centre for Byzantine Studies & Modern Greek University of Birmingham

28. Panaretos, ed. Lampsides, 77.

206

A SEAL OF EPIPHANIOS, ARCHBISHOP OF CYPRUS Study of the archiepiscopal succession in the medieval Church of Cyprus presents peculiar difficulties, especially during the three centuries of Arab attack and influence from 648 until 965, and has not, so far, been greatly helped by the rather meagre sigillographic evidence!. In particular no Cypriot ecclesiastical seal between the early eighth and late eleventh centuries has yet been recorded. The lead bull described below and illustrated in Plate oo fills part of this very wide gap.

eitig VO t V3 oy . ed _ h ‘ pon a % yen gd “ h at | © ; go ; ’ e aN >

oe iy lg

rdatiy olan, ey. 167) 4

“gal sie & ¥ a ye)

Plate OO. A seal of Epiphanios, Archbishop of Cyprus. Photographs 1 1/2 times actual size.

DESCRIPTION. The seal is struck slightly off centre to the right. Parts of both sides are in the process of oxidation. ‘The upper part of the cord channel has erupted on the obverse, taking away most of the face of the bust. Condition:

poor. Diameter: 20 mm. Coll.: author. Obverse. Bust of the Mother of God, facing, holding a medallion of the Child in both hands, between the sigla: M—P and @Y. Legend round the upper circumference:

CEO... sere WA = Q@eo[téxe Boner TH a So]bAw 1. Twelve seals of six medieval archbishops of Cyprus (none an Epiphanios) are published or noticed by V. Laurent in Le Corpus des Sceaux de l’Empire Byzantin, V (2), Paris 1965, pp. 302-316, nos. 1480-1487, and Album. The present seal will be included in the forthcoming volume V (3).

20

Reverse. Neatly engraved legend in four lines with flourishes above and below:

+-IMAPXIE ETDs +’ Exupa[vJiw apyte-

MICKOTIC@ TLOXKOTTO)

KVITPOV | Kuzpou =’Enipaviw apytertoxonm Kuzpov

The seal thus reads: ““O Theotokos, save thy servant/Epiphanios, Archbishop of Cyprus’’.

DATE. ‘The Reverend Father Vitalien Laurent, A.A., has not only gone far to establish the unusually tricky fast: of the medieval archbishops of Cyprus2, but has an unrivalled experience in the field of Byzantine sigillography. I am therefore most grateful to him for an opinion on the date of the seal, based on the photograph reproduced in Plate oo. He writes: “Bien qu’en mauvaise condition, la piéce garde des signes manifestes qui permettent de |’ assigner a la seconde moitié du [Xe siécle: les caractéres épigraphiques du revers, le manteau cételé sur les épaules a |’ avers, avec la disposition du feuillage au revers, pour moi ne font pas de doute’’3. Father Laurent also makes it clear that the seal cannot be assigned, stylistically, to the eleventh century.

ATTRIBUTION. Four archbishops of Cyprus are known to have taken the name Epiphanios: the celebrated St. Epiphanios (1) of Constantia-Salamis (368-403/406); Epiphanios (II) (present at the Sixth Ecumenical Council of 680-1); Epiphanios (III) (fl. 870); and Epiphanios (IV) (Late eleventh century).

Father Laurent’s dating of the seal, by style, to the second half of the ninth century makes its attribution to Epiphanios (III) virtually certain. Epiphanios (III), who is also listed in the twelfth-century synodikon of the Church of Cyprus for the Sunday of Orthodoxy‘, is the author of a letter of sup-

port to Ignatios, Patriarch of Constantinople (847-858, 867-877), evidently written very soon after 28 February 870, the conclusion of the Council held to condemn Ignatios’ great rival, Photios5. The letter is not informative but is 2. V. Laurent, “Les fastes épiscopaux de |’Eglise de Chypre”, Revue des Etudes Byzantines, VI (1948), pp. 153-166.

3. Letter of 5 August 1970. 4. N. Cappuyns, ‘Le synodicon de Chypre au XIle siécle’”’, Byzantion, X(1935), pp. 489-504.

5s. The letter is in Joannes Domintcus Mansi, Sacrorum Conciliorum nova et amplisstma collectio, Graz 1960 (reprint), XVI, coll. 307-8. Cff. K. Ch. Hadjipsaltes (Chatzipsaltes), «"H Kirrpos xata td Sevtepov futov tod évétou xal tac d&pydc tod Bexdktov w.X. alavocg»n, ITenpaypéva

tot & dteOvotc Bulartwodoyixoé Lvuvedgiov Oeccadovixnc, 1, Athens 1956, p. 335; Ma-

A SEAL OF EPIPHANIOS, ARCHBISHOP OF CYPRUS 21 written in terms more than conventionally self-abasing, so much so that an eighteenth-century Archbishop of Cyprus, stung by the implied slight on the autocephaly of his Church, maintained that Epiphanios cannot have been an arch-

bishop at all and was, perhaps, only a subordinate bishop®. It would not, of course, have been beyond the resources of the chancery of the controversial Ignatios to ‘improve’ upon the wording of the letter, but if any further proof is required that Epiphanios was Archbishop of Cyprus, the seal provides it. Although many seals of the Church of Cyprus derive from Syria, suggesting links between the Churches of Cyprus and Antioch at other periods and during the years of Arab influence’, the provenance of the present bull is, significantly and appropriately, Constantinople.

DISCUSSION. The seal comes at a peculiarly interesting crux in the history of Cyprus. ‘The position of the island as a demilitarised territory, administered as a sort of Byzantine-Arab condominium and liable to taxation by both empires, between 688/g and its reconquest by Nikephoros II Phokas in 965, is well known8.

One may wonder if this arrangement did not mask a reality that, like most noman’s lands, Cyprus had been so ravaged and depopulated that it had ceased to be of much economic use to either the Byzantines or Arabs, who therefore largely abandoned the island to its own devices?. But it retained, on occasion, a nouel I. Gedeon, /7atgtapyixoi [Tivaxec, Constantinople 1890, pp. 287-9; Charles Joseph Hefele, Histoire des Conciles, tr. H. Leclercq, IV (1), Paris 1911, pp. 486-537; and F. Dvornik, The Photian Schism, Cambridge 1948, pp. 145-158. 6. Cited by J. Hackett in A History of the Orthodox Church of Cyprus, London 1901, pp. 258-9.

~, Cf. Laurent, ‘‘Fastes’’, p. 160, n.l. 8. Cf. R.J.H. Jenkins, ‘“Cyprus between Byzantium and Islam, A.D. 688-965’, in Studtes presented to David Moore Robinson, ed. George Mylonas and Doris Raymond, II, St. Louis, Miss.,

1953, pp. 1006-1014; K. Ch. Hadjipsaltes, Tenoaypéva, II, pp. 327-341 and the same’s «H Kubrpos, t6 BuSavtov xat 1o “IoAau», Kunoraxal Lnovdai, XX (1956), pp. 13-29. 9. It is, for example, difficult to see how Cyprus could have raised much tax in coin. The Cypriot mint closed down after 628/9 and Mr. Philip Whitting kindly informs me that, although bronze folleis of Constans II (641-668), overstruck — perhaps specifically for Cyprus as in the case of the overstrikings for Sicily — by Constantine IV (668-685) are plentiful and must have

circulated for a very long time, the island had no evident major influx of coin or mint of its own again until the twelfth century. (See Michael F. Hendy, ‘“‘On the administrative basis of the Byzantine coinage, c.400-c.g00, and the reforms of Heraclius” in the “‘Byzantina-Metabyzantina’”’ issue of the University of Birmingham Historical Journal, XI1 (2) (1970) (forthcoming); and the same’s Coinage and Money in the Byzantine Empire, 1081-1261, Washington 1969, pp. 136-142, 360, 371-2). There is no obvious supporting numismatic evidence for a Byzantine re-occupation under Michael III, Basil I or Leo VI. The answer may simply lie in the severe depopulation of the island. None of the great classical cities survived the condominium, when “‘the island presents a sad vacuum in the sphere of Byzantine architecture and art” (Andreas Stylianou and Judith

22

certain strategic usefulness for Byzantine and Arab fleets. That these occasions were transient and left little mark on the internal affairs of the island is surely the clue to why the history of the island in the late ninth and early tenth centuries is so baffling.

Constantine VII Porphyrogennetos (913-959) states that his grandfather Basil I (867-886) incorporated Cyprus as an imperial theme for seven years, when

it was given an Armenian governor. Thereafter it reverted to the Arabs!9, The implication is that there was a temporary break in the condominium. Critics have been wary of this statement. Constantine naturally wished to enhance the reputation of Basil and Cyprus is not otherwise known as a regular theme

during the period!!. The theory that Cyprus was not in fact re-occupied by Basil, but by his son Leo VI (886-912), from 906 to 915, has been discounted !2. The theory that it was not in fact re-occupied by Basil, but by the last Amorian

emperor Michael III (842-867), is more widely accepted. It depends partly upon the rehabilitation of Michael III, largely proposed by the late Professor Henri Grégoire and vigorously opposed by the late Professor Romilly Jenkins! 3. A. Stylianou, The Patnted Churches of Cyprus, n.p. 1964, p. 15). The story of the temporary transfer of the Cypriot population to Nea Justinianopolis in the Kyzikos peninsula in 691-8 may suggest that, even then, it cannot have been very large. (Cf. Constantine Porphyrogenitus, De Administrando Imperio, ed. Gy. Moravcsik and R.J.H. Jenkins, I, Budapest 1949, pp. 224-5 and the commentary in II, London 1962, pp. 180-1). 10. Constantine Porphyrogenitus, De Thematibus, ed. B.G. Niebuhr (C.S.H.B.), Bonn 1840, p. 40: «6 88 paxdprog xal repradvuyoc xal dolduoc Ev BactAstion Bactlretog sic Oépatog taEw tadtyy nxarttotyoe, xal Stentpacev év abt otpatnyov “AdrgEtov Exetvov tov meptBdytov, tO yévos ’Apuséwov, bg xal érexpdtycev abtig ypdvoug éemt& maAw 58 dpypéOy bxd TOV Lapaxynvaer, xal tadtTHY Popo-

Aoyovaw do xal mpdtepov». Cf. (for an even more explicit version) Hadjipsaltes, [Iengaypéva, II, p. 334; George Hill, A History of Cyprus, I, Cambridge 1940, p. 294; and Héléne Ahrweiler, Byzance et la Mer, Paris 1966, p. 442. 11. See Héléne Antoniadis-Bibicou, Etudes d’histoire maritime de Byzance, Paris 1966, pp. 89-90.

12. Proposed by R.H. Dolley in “‘A forgotten Byzantine conquest of Kypros’’, Bulletin de l’ Académie de Belgique, XXXIV (1948), p. 2009ff., and rejected (on the evidence of Arab sources)

by Jenkins (loc. cit.) ; Hadjipsaltes (Tengaypéva, II, p. 333f.); and George Ostrogorsky (History of the Byzantine State, tr. Joan Hussey, Oxford 1968, p. 238, n. 1). 13. Henri Grégoire affirmed that ‘““The oft-repeated assertion that it was Basil who recaptured Cyprus from the Arabs is erroneous; the reoccupation in fact took place in 853, at the start of the attacks against Damietta and the Egyptian sea-board”’ (The Cambridge Medieval History, IV (1),

ed. J.M. Hussey, Cambridge 1966, p. 107, n. 1). Romilly Jenkins maintained that: ‘Some scholars, without a tittle of evidence and on pure inference only, have stated that the incorporation of Cyprus as a theme must have been the work of Michael III and Theodora, which Constantine VII is here trying to claim for his ancestor Basil. In my view, we cannot treat affirmative state-

ments in this cavalier fashion. If we are going to reject plain testimony of this sort, we may as well make up our minds that we can know nothing whatever about the period, and rewrite its

A SEAL OF EPIPHANIOS, ARCHBISHOP OF CYPRUS 23

One fact ts certain: the Uspenskiy Takttkon (apparently ignored by Jenkins) makes it clear that the archontes of Crete and Cyprus were officials in the Byzan-

tine naval system, but the document cannot be dated more closely than to the period 842-856!4. One may suspect that both islands were advance bases for the great naval campaigns of 853-9- and no more!5. Jenkins frequently maintained that Byzantine authors did not, on the whole, falsify unpalatable information (in this case the achievements of the discredited Michael III), but simply omitted it!6,. This seems to be what has happened here. The firm evidence of our administrative document for an earlier, and perhaps temporary, Byzantine naval base in Cyprus does not invalidate the statement of Constantine Porphyrogennetos

that Basil made the island a theme (albeit an irregular one) for seven years. The letter of Ephiphanios to Constantinople does not necessarily prove that one of these seven years was 870; indeed the fact that the archbishop is not one of the 103 or more Greek prelates who signed the Canons of the Council could suggest that he was unable to reach the capital in 869-870, at a time when Ignatios was making great efforts to attract a large and sympathetic attendance. Bréhier suggested that the seven years were from 874 to 881; Hadjipsaltes argues otherwise.!7 It 1s in fact impossible to be sure which the seven years were — or even if they were in fact seven years. Constantine Porphyrogennetos names seven years as the period of sojourn of the Cypriots at New Justinianopolis during the reign of Justianian II (685-695, 705-711),18 and it is possible that in the mind of the emperor, or of the compiler of his works, seven years had simply become a period associated with Cypriot affairs. To seek to prove that Cyprus was outright Byzantine or Arab during this period is to pursue a will-o’-the-wisp. For example independent sources indicate that the island was in Byzantine hands in c. go2, that Paphos (at least) was in Muslim hands in go04 and that Cyprus was being used as a Byzantine base soon after 905.19 ‘There is no reason to question any of these indications, for in history according to our own taste and fancy” (Byzantium, the imperial centuries, A.D. 610-1071, London 1966, p. 157). That the ‘plain testimony’ (cited in note 10) may not be the full story is argued by Hadyjipsaltes (locc. citt.) and made virtually certain by the evidence of the Uspenskij Taktikon.

14. Ahrweiler, Byzance et la Mer, pp. 71-92; Hadjipsaltes, enxpaypéva, II, p. 331f. 15. .A.A. Vasiliev, Byzance et les Arabes, I, Brussels 1935, pp. 212-8. 16. E.g. Jenkins, Byzantium, p. 156. 17. Louis Bréhier, Le Monde Byzantin, I, Paris 1948, p. 132; Hadjipsaltes, [Tengayyéva, II, Pp. 336.

18. Constantine Porphyrogenitus, D.A.J., I, pp. 224-5; II, pp. 180-1. 19. Hackett, Church, p. 49; Ioannes Cameniata (who was carried off from Thessalonike via Paphos as an Arab captive), De Excidio Thessalonicensi, ed. I. Bekker (C.S.H.B.), Bonn 1838, P. 596; and Ostrogorsky, History, p. 258.

24

reality Cyprus did not have a complete Byzantine administration until after 965. The island’s ports would be used by either or both major powers for their naval requirements almost annually. In the hapless island itself few things are certain.

But one is the fact that Archbishop Epiphanios (III) wrote to the Patriarch in 870, to which can now be added the tangible evidence of his seal.

IV

Greek historians on the Turks: the case of the first Byzantine-Ottoman marriage

SIR RICHARD SOUTHERN divided Western medieval scholarly views

of Islam into an ‘Age of Ignorance’, the twelfth century of ‘Reason and Hope’, and a cold-blooded ‘Moment of Vision’ in the decade 1450-60.1 The same sequence can be discerned among Byzantine

intellectuals too, but their vision was enlarged by experience of Westerners (whose behaviour they could not help comparing with that of Muslims),? and limited by the fact that they had to live not so much with Islam, as with Muslims. Symbiosis did not necessarily

make Byzantines more aware of the nature of Islam than were Westerners. Indeed it deprived them of the focus of distance. Some

later Byzantines, faced with the facts of Turkish conquest and Western superiority, did not evade the question of why their Godprotected Empire was in palpable decline.? But none of the four Westerners who shared Sir Richard’s ‘Moment of Vision’ in 1450-60

were faced with the realities of the problem that the anti-Unionist Patriarch Gennadios Scholarios was left to explain to his faithful after 1453. In fact he did not waste much time in speculating on the 1R. W. Southern, Western Views of Islam in the Middle Ages (Cambridge, Mass., 1962).

2 The contrast between the Norman rampage in Thessalonike in 1185 and Saladin’s relatively well-ordered conquest of Jerusalem two years later did not escape Niketas Choniates, Historia, ed. van Dieten (Berlin, 1975), pp. 297-306, 575-7. See also S. Vryonis Jr, ‘Byzantine attitudes towards Islam during the late Middle Ages’, Greek Roman and Byzantine Studies, 12 (1971), pp. 263-86.

21, Sevéenko, ‘The Decline of Byzantium seen through the eyes of its intellectuals’, DOP, 15 (1961), pp. 167-86.

Greek historians on the Turks

waywardness of Fortune, and few, if any, Byzantines had enjoyed the luxury of formal debate with Islam since before 1430.1 For Gennadios, the question of marriage was more pressing, and the application of the principle of ‘Economy’ to relax canon law to protect what he could of Byzantine family structure which the Turk had long broken and assimilated in Anatolia, but which the Ottoman conquest of Constantinople came by Fortune just in time to attempt to salvage in the Balkans.?

So Byzantines’ very intimacy with Muslims, especially Turks, translates Sir Richard’s three stages into, first, a honeymoon; second

a marriage; and third, not divorce, but the almost simultaneous realization that there had been no marriage at all. These stages overlapped, all three in 1346.

The first stage belongs to the heroic age of Byzantine-Muslim courtship celebrated in epic poetry which stretched into the fifteenth century. The Armenian David of Sassoun and the Byzantine Digenes

were both “‘twyborn’ heroes, effective brothers of their Muslim antagonists. In the ballads and romances of the ghazi Sayyid Battal, the Melikdanishmendndme and Dede Korkut, Orthodox and Muslims fight and inter-marry like clean-limbed heroes, who are not the sort of people who are much bothered by religious or cultural obstacles. The Tiirkmen Danishmendids were not above tacking an Armenian genealogy on to their own (a useful claim to rule in that dynastically

minded land), while their own epic boasts of alliance with the Byzantine Gabras family.? In Sir Richard’s ‘Century of Reason and 1See Manuel II Paléologue, Entretiens avec un Musulman, ed. T. Khoury (Paris, 1966); and J. D. G. Waardenburg, “The two lights perceived: Medieval Islam and Christianity’, Nederlands Theologisch Tijdschrift, 31 (1978), p. 276. 4 Ch. G. Patrineles, Ho Theodoros Agallianos tautizomenos pros ton Theophanen medeias kai hoi anekdotoi Logoi tou (Athens, 1966), 68-71; S. Vryonis Jr, “The conditions and cultural significance of the Ottoman conquest in the Balkans’, IIe Congrés international des études du sud-est européen (Athens, 1970), pp. I-10; and the same’s ‘Religious change and continuity in the Balkans and Anatolia from the fourteenth through the sixteenth century’, in Islam and cultural change in the Middle Ages (Wiesbaden, 1975), pp. 127-40. 3 Iréne Mélikoff, La geste de Melik Danismend, i (Paris, 1960), passim; P. Wittek, The rise of the Ottoman Empire (London, 1963), p. 20. Among growing

[472]

Hope’ when in Anatolia some marriages pass from epic and romance into reality, there at first seemed no good reason why it should not be hoped that both Seljuks and Tiirkmens should be baptised into the Empire, as had the Slavs in the past. Of nine known members of the

Gabras family in that century, between 1098 and 1192, one was canonized, four (including the saint) served the Byzantines as dukes and generals, three served both Byzantines and Seljuks in the same

capacities, and two served the Seljuks only (the last, perhaps inheritor of St. Theodore Gabras’s estates, as a vizir).1 A Gabras was

naturally involved when in the 1140s the daughter of the Seljuk Sultan Masud married the Byzantine John Komnenos to become (according to a later Byzantine memorialist) grandparents of Osman and great-grandparents of Orhan, the first Ottoman ruler to marry a Byzantine princess.? It is this marriage, between the ghazi emir Orhan and Theodora Kantakouzene in 1346, which I want to examine in the chronicles.

It symbolizes a turning-point in Byzantine-Turkish relations and illuminates all other imperial alliances. It tells less of the fate of countless lesser Christian, Muslim, and mixobarbarian Anatolians when they found that it was too late to speculate on whether Islam was St. John of Damascus’s roist heresy, or even the authenticity of the Gospel of Barnabas, for they were now wedlocked.® literature on Dede Korkut, the most recent English translation is The book of Dede Korkut, trans. F. Siimer, A. E. Uysal, and W. S. Walker (Austin and London, 1972); and the most recent study is Kh. Korogly, Oguzskiy geroicheskiy epos (Moscow, 1976).

1A. Bryer, ‘A Byzantine family: the Gabrades, c. 979-c. 1653’, University of Birmingham Historical Journal, 12 (1970), pp. 175-81; C. Cahen, Pre-Ottoman Turkey (London, 1968), p. 210.

2G. Sphrantzes, Memorii 1401-77, ed. V. Grecu (Bucharest, 1966); pp. 208-12; Wittek, Rise, p. 9; Bryer, UBHJ, 12 (1970), p. 177. 3 Nikephoros Gregoras, Byzantina Historia, ed. L. Schopen, III (Bonn, 1855),

p. 509 (on the mixobarbarians of Bithynia) and 555 (on Constantinopolitans with Latin heads and ‘Persian’—i.e. Turkish—bodies); Doukas, Istoria TurcoBizantina 1341-1462, ed. V. Grecu (Bucarest, 1958), p. 61 (on mixobarbarians of Constantinople); S. Vryonis Jr, The decline of medieval Hellenism in Asia Minor and the process of Islamization from the eleventh through the fifteenth century

(Berkeley, Los Angeles, and London, 1971), pp. 176, 182, 228-9, 440, 446; [473]

Greek historians on the Turks

The background is the contest between the Byzantine usurper and former regent, John VI Kantakouzenos (1341-55), in Thrace, and the legitimate Emperor John V Palaiologos (1341-91) and his mother, Anna of Savoy, who, with their chief minister Alexios Apokaukos, were holed up in Constantinople. All Byzantine rulers were competing for the support of a nest of four west Anatolian emirates which had emerged on the marches of the old Seljuk state now that the Mongol J/hans of Iran had lost control of it. The two leading emirates were of Ottoman Bithynia (which had the advantage of being nearest Constantinople) under the ghazi emir Orhan (1324-62), and of Aydin (which had the advantage of controlling the

port of Smyrna) under the ghazi emir Umur (1334-48). Kantakouzenos, and Palaiologos and his mother, were seeking Turkish support as much against each other as against the Serbian ruler Stefan IV Dushan (1331-55). The evidence

Kantakouzenos wrote his great apologia for his rule before about 1369, when he was the monk Joasaph. In it he states that he betrothed

one daughter to the Despot Nikephoros II of Epiros in 1342, when

he noted that he had two others. He first met Umur of Aydin in 1335, the year after the emir had taken Smyrna and had embarked on a piratical career in the Aegean. Umur came over to Thrace to assist Kantakouzenos in 1342, 1343, and 1344; they were comrades in arms.}

Enveri composed his Diisturndme in about 1465. It is of substantial

historical value, although he was not shy of interpolating his own H. I. Cotsonis, ‘Aus der Endzeit von Byzanz: Burkludsche Mustafa, ein Martyrer fiir koexistenz zwischen Islam and Christentum’, Byz. Zeits., 50 (1957), pp. 397~404; P. Charanis, ‘Internal strife in Byzantium during the fourteenth century’, Byzantion, 15 (1941), 230 (on religious syncretism) ; John of Damascus, PG 94, cols 764-73. A ghazi is a dedicated warrior for Islam. 1 John Kantakouzenos, Historiarum libri IV, ed. L. Schopen, i (Bonn, 1828), Pp. 476-95; ii (Bonn, 1831), pp. 195, 336-48, 390-403, 529-34; D. M. Nicol, The Byzantine family of Kantakouzenos (Cantacuzenus) ca. 1100-1460 (Washing-

ton, D.C., 1968), pp. 42, 45, 53, 55, 59, 60, 66, 100.

[474]

experiences into his poetic account.1 Book 18 comprises the gesta of Umur, a very perfect ghazi. Lemerle dates an encounter between

Umur and one of Kantakouzenos’s daughters in Didymoteichos (between Selymbria and Adrianople) to 1344, his last campaign with

him.? Enveri versifies what seems to be two versions of the same incident. In the first the tekfur Kantakouzenos feasts Umur in his palace. The tekfur’s three daughters are displayed: they are as beautiful

as huris. One, simply entitled the Despoina,’ is even more beautiful than the others. But Umur, shah of the ghazis, would not look upon them. The tekfur says: “O Lord of the World, Iam your poor slave,

1 Le destan d’Umur pacha, ed. and trans. Iréne Mélikoff-Sayar (Paris, 1954),

to be read in conjunction with the commentary by P. Lemerle, L’Emirat d’Aydin, Byzance et l’Occident (Paris, 1957). As Lemerle, Aydin, pp. 139-41, points out, Enveri’s account of Umur’s improbable raid on Kilia in the Black

Sea, supposedly of 1341, in Destdn,, ll. 1209-1306, appears to be based on Enveri’s experience of following Mehmed II on his attack on Kilia in 1462, when it was in the hands of the voivode Vlad III Dracul (1456-76), called the Impaler. But Lemerle’s explanation that Enveri’s story, of how the Turks terrified local prisoners by pretending to be cannibals, might be a reference to Vlad’s infamy is unconvincing: Vlad impaled his victims; he is known to have roasted only one and never to have eaten them; see N. Stoicescu, Vlad Tepes (Bucarest, 1978), pp. 68-95; Doukas, ed. Grecu, 431. On the other hand, S. Vryonis Jr, “Evidence on human sacrifice among the early Ottoman Turks’, Journal of Asian History, 5 (1971), pp. 140-6, offers evidence of the survival of (among others) the custom of offering human sacrifice for the distinguished dead among Turks until the fifteenth century, while there seems to be a case of cannibalism in Anatolia under the ilhans in 1277; see A. Bryer, “The fate of George Komnenosg, ruler of Trebizond (1266-1280)’, Byz Zeits. 66 (1973), p. 346 n. 68 (where for ‘beaten’ read ‘eaten’). According to George Pachymeres, De Michaele et Andronico Palaeologis, ed. I. Bekker, 1 (Bonn, 1835), p. 134, Anatolian Greeks believed that nomads practised cannibalism and in the context of their other cultural survivals, the fear that the Turks also practised cannibalism may have been real enough, if actually unfounded, in 1341 or 1462, among the inhabitants of Kilia. 1 am most grateful to Mr M. E. Martin for discussion. 4 Lemerle, Aydin, pp. 175-9. 3 A tekfur is usually a Christian, and so subordinate, ruler. Despoina is usually a title, ‘lady’, the equivalent of hatun, rather than a personal name. Princesses of Trebizond who married Tiirkmens took the composite title of despoina-hat (un).

[475]

Greek historians on the Turks

you are my han. My goods, myself, my son, my daughter: every-

thing I have is yours. Take one of my three daughters.’ Umur blushes and retreats to his tent, where an evil voice urges him to take

the girl. He replies: “Is he right to give a daughter to his brother? The tekfur is my brother; his daughter is my daughter. In our faith that cannot be done.’ Enveri then describes a second encounter (or version), heading it ‘The purity of the pasha’. The lovely Despoina goes on a secret hunt ‘in the forest of faith’. Clothed in robes encrusted with pearls, rubies,

and other jewels, she unveils herself before Umur, protesting her love for him and begs the ghazi to take her. Umur asks her name. The Despoina asks him to accept her first. Umur refuses; such an alliance is haram in his faith. The Despoina pleads. The emir sends her

away. Glancing back through her tears she sees Umur covering his eyes: a ghazi’s duty, day and night, is to wage holy war and resist worldly desires.®

On 28 October of that year, 1344, Umur lost Smyrna to the Latins who kept it until Timur’s invasion of 1402. His emirate was now largely landlocked and he was of no further strategic use either to Kantakouzenos, or to Anna of Savoy who was already negotiat-

ing with the Ottoman emir Orhan, who thus became the most useful (or dangerous) Byzantine neighbour, for he was now the

emir with most room for manoeuvre. So early in 1345 Kantakouzenos was in the process of changing sides from Umur to Orhan. He subsequently defended himself at painful length against the charge of having invited the Ottomans into Europe.® It is quite

true that he was not the first to introduce the Turks, but Kantakouzenos’s decisions of 1345-6 did much to decide which Turks would eventually take Constantinople. His decisions were the

more painful because while Orhan was a relatively unknown quantity, it is evident that Kantakouzenos and Umur had been warm

friends. Writing before 1360, Nikephoros Gregoras emphasized 1 Destan, ll. 1739-74. * Destan, ll. 1775-1820. Haram (opposite to halal) is that which is forbidden. * Kantakouzenos, ii. pp. 502-24.

[476]

how close the emperor had been to the emir, a barbarian who had

behaved almost like a Hellene. The two were like Orestes and Pylades, and Kantakouzenos grieved when the Latins killed Umur

in 1348. But Gregoras also makes clear the lively and vehement passion which Orhan had developed for Theodora (whom he calls Maria) Kantakouzene, a tradition developed by Doukas, a century

later. Doukas was what would later be described as a ‘Levantine’. He was the sort of Greek whom Latins chose to represent their interests at the Ottoman court. He was a Uniate by confession, but his rhetorical lament on the Fall of Constantinople reveals him to have been deeply Byzantine by instinct. His chronicle breaks off in mid-sentence as he witnessed the Ottoman conquest of Genoese Lesbos in 1462. His grandfather, Michael Doukas, had been among two hundred opponents of Alexios Apokaukos, Anna of Savoy's

minister and former protégé of Kantakouzenos, who had been rounded up in Constantinople in summer 1345. The captives assassinated their captor when he came to inspect their prison, but only six, including Michael, escaped execution. It is significant that Michael Doukas did not join Kantakouzenos thereafter. He had had enough, and the emirate of Aydin offered a kind of solution. Here Umur’s brother Isa installed him in Ephesus. Doukas adopted his foreign residence for his homeland, and esteemed and honoured the foreigner and barbarian as one crowned by God, recalling to mind the wicked deeds of the Romans. My grandfather foresaw that shortly all the lands from Thrace to the Danube would fall into the hands of the Turks who would soon become absolute masters just as it happened not long before in Phrygia and Asia and in the provinces beyond.?

Doukas maintained that it was Kantakouzenos who unexpectedly 1 Gregoras, ii (Bonn, 1830), pp. 648-9, 762-3, 835; Doukas, ed. Grecu, pp. 58-61. Laonikos Chalkokondyles, Historiarum libri decem, ed. I. Bekker (Bonn, 1843), p. 24, is neutral. 2 Doukas, ed. Grecu, 47; trans. H. J. Magoulias, Decline and fall of Byzantium

to the Ottoman Turks by Doukas (Detroit, 1975), p. 65; D. I. Polemis, The Doukai. A contribution to Byzantine prosopography (London, 1968), p. 196.

[477]

Greek historians on the Turks

made Orhan the offer of a marriage contract for Theodora, with

lavish dowry, in January 1346. Understandably enough Kantakouzenos's story is that it was Orhan who first sent an embassy to ask for his daughter's hand. But Kantakouzenos goes on to say that before he agreed to Orhan’s proposal he sent a legation to Umur (a symmachos whose moderation and loyalty he respected)

to ask for his approval, and by implication permission, for the marriage. This should have been in late 1345 or early 1346, depending upon whether one accepts Doukas’s or Kantakouzenos’s

version of who made the first move. But Kantakouzenos’s story makes it highly probable that it was the Despoina Theodora who was the princess whom Enveri made Umur meet in 1344.

On 16 April 1346 Stefan IV Dushan had himself crowned emperor at Skoplje. On 21 May 1346 Kantakouzenos had himself crowned emperor at Adrianople (Edirne). The coronations are surely linked, but their sequence would be more complicated if there were any grounds tojustify Alderson’s statement that Dushan had married

his daughter (supposedly another Theodora) to Orhan in 1345.” During these events, Kantakouzenos records that he invited Orhan to send a numerous army to escort his bride from Selymbria (Silivri). Orhan sent thirty ships, cavalry, and the most eminent of his followers there in early summer (Doukas’s date). Kantakouzenos then describes how he, his empress Eirene, and their three daughters (Maria, Theodora, and Helena) proceeded with his army to Selymbria. The empress and her daughters spent the night in the imperial

tent; Kantakouzenos with the army. On the next day the empress, Maria, and Helena remained in the tent. Kantakouzenos was the only person on horseback; apparently even Orhan’s cavalry was dismounted. Kantakouzenos had ordered a prokypsis platform to be built of wood outside the town. Theodora mounted the prokypsis. It was surrounded by silk and cloth-of-gold curtains. They were 1 Kantakouzenos, ii. p. 586. 2 A.D. Alderson, The structure of the Ottoman dynasty (Oxford, 1956), p. 165,

table xxii; M. A. Purkovi¢, ‘Byzantinoserbica’, Byz. Zeits. 45 (1952), pp. 47-9. Iam grateful to Mrs Zaga Gavrilovié for this reference. [478]

drawn to disclose the bride encircled by lights carried by kneeling eunuchs. Trumpets, flutes, and all manner of musical instruments

were sounded. When they fell silent, encoomia were recited in honour of the bride. Then Kantakouzenos held a feast for Greeks and Turks which lasted for several days.

The Turkish wedding guests at Selymbria apparently took their time to disperse; they were being a nuisance to all parties in Thrace in late summer. But it was Kantakouzenos who now won the race to Constantinople, which he entered on 2 February 1347, for his second coronation on 21 May. On 28 May 1347 he married his third daughter, Helena, to John V Palaiologos; their own daughter was to marry Orhan’s son, Murad I, in 1389.? An unusual feature of Theodora’s marriage was that it took place on Byzantine, not Turkish, territory: hence there was no call for nymphostoloi, bridal escorts.2 But the price Kantakouzenos paid for the alliance seems to have been the absence of the bridegroom himself. If this is so, Doukas should be preferred to Kantakouzenos as to who made the first move over the marriage. Kantakouzenos had been helped by Orhan in 1345 and would surely have mentioned

his presence at his own wedding in 1346. As it was, the first (and perhaps only) time he seems to have met his son-in-law was when Orhan came to Turkish Skutari (Uskiidar, Chrysopolis), opposite Constantinople, in the summer of 1347. Kantakouzenos describes how he came over to see Orhan for a few days. He and the emir ate together, while Orhan’s four sons ate at a nearby table. They were

not Theodora’s children (she probably mothered Halil), but Ibrahim (b. 1310), Siilayman (b. 1316), Murad I (b. 13262), and Sultan (b. 1324). Even then Orhan would not leave Turkish territory, but Kantakouzenos, Theodora, and the four Ottoman princes 1 Kantakouzenos, ii. pp. §86—8. Orhan was then about 60. He seems to have

married another Christian, Asporcha, at about the same time; Alderson, Structure, loc. cit.

2 Kantakouzenos, il. 591-6; Nicol, Kantakouzenos, pp. 62-5, 135-58; Alderson, Structure, p. 166, table xxiii. * Michael Panaretos, Peri ton Megalon Komnenon, ed. O. Lampsides (Athens, 1958), p. 72.

[479]

Greek historians on the Turks

crossed to Byzantium for further festivities, where Theodora stayed with her mother and sisters for three days before returning to Bithynia.}

Theodora Kantakouzene was widowed in 1362, when she may have returned to Constantinople to recount her affairs to her father before 1369; at any rate Kantakouzenos, his empress Eirene, and

their three daughters are last found reunited as hostages of Andronikos IV in Galata in 1378-81.2 Kantakouzenos concluded

his account of the marriage of 1346 with a description of the dangers Theodora faced among the barbarians of Bithynia in defence

of her Orthodoxy in 1346-62. She not only resisted religious conversion, but managed to persuade many Christian converts to Islam to return to their faith, assisted the poor, gave gold to ransom

slaves from the barbarians, and was a haven of salvation to those Romaioi who, by God's will, suffered. According to her father, her life among the barbarians had been a shining example of virtue to all.8 The context

Among late medieval dynasties, the Byzantine were pivotal: they eventually linked the ruling families of Western Europe with the kiriltays of Central Asia and understood (or compromised with) both series of systems. They also brought their new Turkish and Tiirkmen neighbours into the network. By their alliances from 1352 with local Tiirkmen chiefs the Grand Komnenoi of Trebizond had a hand in creating territorial principalities out of wandering clans.‘ By his alliance with Orhan in 1346 Kantakouzenos paved the way to the supplanting of one extended ruling family, the Palaiologos-

Kantakouzenos, by another, the Ottoman, in 1453, by when 1 Kantakouzenos, ii. 28; T. S. Miller, “The history of John Cantacuzenus (Book IV): text, translation, and commentary’ (The Catholic University of America, unpublished Ph.D. thesis, 1975), pp. 165, 267-8. 4 Nicol, Kantakouzenos, p. 135. § Kantakouzenos, ii. pp. 588-9. “ A. Bryer, ‘Greeks and Tiirkmens: the Pontic exception’, DOP, 29 (1975),

pp. 130-2.

[480]

Mehmed II had a rather better hereditary claim to the Byzantine throne than had Kantakouzenos in 1347. Between 1297 and 1461, thirty-four or more Byzantine, Trapezuntine, and Serbian princesses married (in order) Mongol hans and

ilhans, Turkish emirs, and Tiirkmen begs. During this period no

Orthodox is known to have married a Muslim princess.1 In Kantakouzenos’s youth, when the most important alliances were still Mongol, there was what amounted to an imperial finishing school for such brides: Realising that they could not defeat the barbarians in battle, the emperors of the Romaioi escaped destruction by conciliating them with gifts and courtes-

ies. They were made particularly tractable and friendly towards the Romaioi by being given brides from among the ladies of the imperial family. ... So, maidens of exceptional beauty, not only of the aristocracy but also of lowly origin, were brought up in the imperial palace like the princesses and, as the need arose, were betrothed to the satraps of the Mongols.”

Kantakouzenos’s initiative in marrying Theodora to a Turk in 1346 was surely in the mind of Alexios III Grand Komnenos of Trebizond when he took the plunge and married the first Byzantine princess to a Tiirkmen in 1352, beginning eleven such alliances. As Chalko-

kondyles remarked, the Trapezuntines preserved their Hellenic speech and customs, but intermarried with neighbouring barbarians, ‘so as not to have trouble owing to the ravaging of their land by the latter’.® 1 Marriages listed in G. E. Rakintzakis, ‘Orthodox-Muslim mixed marriages, ca. 1297-1453’ (University of Birmingham, unpublished MA thesis, 1975). The list of Byzantine marriages in M. Izeddin, ‘Notes sur les mariages princiers en Orient au moyen 4ge’, Journal asiatique, 257 (1969), pp. 139-56, is less reliable. 2 Kantakouzenos, i. p. 188. 8 Chalkokondyles, 461-2; Bryer, DOP, 29 (1975), pp. 135, 149 Appendix 11.

Cf. the supposed reply of Prince Nicholas I of Montenegro (1860-1918) to a criticism of the meagreness of that country’s exports: “What about my daughters?” W. Miller, Trebizond, the last Greek empire (London, 1926), p. 69; H. de Windt, Through savage Europe (London, n.d.), pp. 49-50. Defending the Kantakouzene marriage of 1346, D. M. Nicol, The last centuries of Byzantium,

[481]

Greek historians on the Turks The ceremony

Apart from Theodora’s betrothal and marriage contract, the only element of her marriage ceremony recognizable from late Byzantine

imperial wedding ritual is the prokypsis, a sort of illuminated imperial tableau vivant, of which Kantakouzenos’s description of the

sensational performance of 1346 is our best account. The late Byzantine antiquarian spirit encouraged the revival or elaboration of at least three features of imperial ceremony from the twelfth or

thirteenth centuries: the raising of the emperor on a shield, his anointing, and his prokypsis.1 Prokypsis is in fact, though probably

not etymologically, derived from the epiphany of the imperial family in the imperial box, or kathisma, of the hippodrome. It was 1261-1453 (London, 1972), p. 210, notes that ‘things in Constantinople could still compare favourably with the standards in Trebizond, where the Emperor Alexios II, who died in 1330, had married no less than three infidel wives, and where the Emperor Basil who succeeded him gave two of his daughters as brides to local emirs.’ It should be made clear that the sequence of marriages of Trapezuntine princesses to local begs (not emirs) began six years after the

Kantakouzenos alliance (and twelve years after the death of the Grand Komnenos Basil in 1340). Furthermore, Alexios II’s wife was the daughter of

an Orthodox Georgian, Beka Jageli. The honorific epithets applied to his children by other alliances (Ana-Kutlu, Aza-Kutlu, and Ak-Bogha) are certainly

Turkic or Mongol, but that is no evidence that their mothers (probably concubines) were Muslim, for Orthodox Georgian princesses (and even one Palaiologine) assumed Turkish epithets, such as Giilhan (Eudokia), wife of the Grand Komnenos Manuel III. The only Grand Komnenos known to have taken a Turkish wife (or probably concubine, for he was also married to a Georgian) was John IV, who, rather unconvincingly, explained that he had

done so in the interests of piecemeal proselytization. See Panaretos, ed. Lampsides, pp. 63, 64, 66, 67, 70, 78, 81; Pero Tafur, Travels and adventures,

1435-1439, trans. M. Letts (London, 1926), p. 131; J. Sauvaget, ‘Noms et surnoms de Mamelouks’, Journal asiatique, 238 (1950), p. 37; Gy. Moravesik, Byzantinoturcica, 11 (Berlin, 1958), pp. $7, 60, 92. 1D. M. Nicol, ‘Kaisersalbung, The Unction of emperors in late Byzantine coronation ritual’, Byzantine and Modern Greek Studies, 2 (1976), pp. 37-523

C. Walter, ‘The significance of Unction in Byzantine iconography’, ibid. Pp. 53-74; and the same’s ‘Raising on a shield in Byzantine iconography’, Revue des études byzantines, 33 (1975), pp. 315-56. Dr Michael Jeffrey, of the

University of Sydney, is to publish a convincing demonstration that the prokypsis ceremony can be traced back to the marriage of Theodora, neice of Manuel I Komnenos, with duke Henry II of Austria in 1148.

[482]

performed on state and religious feasts, such as weddings. By the thirteenth century the term meant not only the display itself, but the platform on which it was held. In Constantinople and Nicaea it was

staged near, or in the gallery of, a church. In Trebizond a gallery seems to have been specially built for it in the coronation church of the Chrysokephalos after 1223, but the prokypsis may also have been held in the Epiphaneia court of the palace and in the main square, the meidan.1 Spatharakis has recently demonstrated that the scene in Cod. Vat. Gr. 1851, fo. 7v, long regarded as an illustration of an actual prokypsis, cannot be so identified.? But the convention

of late Byzantine portrait groups which he studied is surely the pictorial equivalent of prokypsis.* Such illustrations and actual displays of prokypsis served to demonstrate the exact status of each member of a swiftly changing hierarchy of collegiate sovereignty

as definitively as any Politburo grouping at a May Day parade: emperors, co-emperors, empresses, princesses, and despots were carefully disposed according to rank. For marriages the prokypsis 1 See the lemma to the Planudean Anthology, XVI, 380 in W. R. Paton, The Greek Anthology, v (London, 1970), p. 385; A. Cameron, Porphyrius the charioteer (Oxford, 1973), pp. 200-1; Pseudo-Kodinos, Traité des offices, ed.

and trans. J. Verpeaux (Paris, 1966), pp. 171, 181, 183, 195-8, 202-3, 208-9, | 226-7, 286-7; A. Heisenberg, Aus der Geschichte und Literatur der Palaiologenzeit (Sitzungsberichte der bayerischen Akademie der Wissenschaften, philos.-

philol. und hist. Klasse, Abh. 10, Munich, 1920), pp. 85 ff.; O. Treitinger, Die ostromische Kaiser und Reichsidee nach ihrer Gestaltung in hofischen Zeremoniell

(Jena, 1938), pp. 112 ff.; Ph. Koukoules, Byzantinon bios kai politismos, iv (Athens, 1951), pp. 134-5; E. M. Kantorowicz, ‘Oriens Augusti—Lever du Roi’, DOP, 17 (1963), pp. 159-62; H. W. Haussig, A history of Byzantine civilization,

trans. J. M. Hussey (London, 1971), pp. 38, $5, 193, 398, 401, and pl. 32; Panaretos, ed. Lampsides, p. 78; and on the whereabouts of prokypsis in Trebizond, A. Bryer and D. Winfield, The Byzantine monuments and topography of the Pontos (Washington, D.C., forthcoming), chap. 20.

2 See Koukoules, Bios, iv, pl. 5; J. Strzygowski, ‘Das Epithalamion des Palaologen Andronikos II’, Byz. Zeits. 10 (1901), pp. 553-4, and pl. vi, 2; S. P. Lampros, Leukoma Byzantinon Autokratoron (Athens, n.d.), pl. 80; and I. Spatharakis, The portrait in Byzantine illuminated Manuscripts (Leiden, 1976),

: pp. 210-30 and pl. 162. 8 e.g. the portrait group of Manuel II Palaiologos’s family in Lampros, Leukoma, pl. 84, and Spatharakis, Portrait, pl. 93.

[483]

Greek historians on the Turks

served the function of a modern royal wedding group photograph. But in 1346 Theodora had to stand alone with her eunuchs. She was, however, awarded the epithalamia recited at a prokypsis, such as had

been sung for Theodora Komnene, bride of Henry, brother of Conrad II, in 1147-8, and for Constance, daughter of Frederick II, bride of John Vatatzes in 1244, at their own displays.! One can only speculate as to how Kantakouzenos’s poets handled their theme in 1346; Gibbon is scornful as to their likely quality. Rubriquis and Ibn Battuta report on similar Mongol displays, and the Seljuks and Danishmendids had their audience chambers. But there is no evidence for prokypsis in the simple early Ottoman custom and Kantakouzenos evidently made two concessions to Orthodox and Muslim sentiment in 1346. First he ordered a secular wooden prokypsis stage outside Selymbria. The town and its citadel was then substantial and well-walled; it was probably governed by

Phakrases. Its churches included one recently built by Alexios Apokaukos, assassinated in the year before. But even if common sense dictated that Turkish troops should not be allowed into the town, clerical sense surely denied a church as the theatre for the prokypsis.® Second, as in a Muslim, but not Orthodox, marriage, the mother and sisters of the bride were excluded from the show, even as spectators. The marriage in custom

Historians of the early Ottomans regularly comment on the debilitating effects on the first emirs of their encounter with imperial 1Kantorowicz, DOP, 17 (1963), p. 161; G. Schlumberger, Byzance et croisades (Paris, 1927), pp. 63-4. 2 E. Gibbon, The history of the decline and fall of the Roman Empire, vi (London, 1788), p. 317.

3 William of Rubruck in Contemporaries of Marco Polo, ed. M. Komroft (London, 1928), p. 99; The travels of Ibn Battiita A.D. 1325-54, trans. H. A. R. Gibb, ii (London, 1962), p. 483; Alderson, Structure, p. 93; what appears to

be a Danishmendid audience chamber at Neocaesareia (Niksar) is described . in Bryer and Winfield, Monuments, chap. 7; S. Eyice, ‘Alexis Apocauque et l’église Byzantine de Sélymbria (Silvri)’, Byzantion, 34 (1964), pp. 77-104; and Démétrius Cydonés, Correspondance, ed. G. Camelli (Paris, 1930), p. 214.

[484]

Byzantine family ceremonial, beginning with the wedding of 1346."

This is hard to credit. If Orhan had been a witness of his wife’s prokypsis, it is difficult to know what he would have made of it. Despite Enveri’s conventional account of the splendour of the Despoina’s robes in 1344, Kantakouzenos did not have access to the

Byzantine imperial regalia in Constantinople (which had anyway been pawned to Venice in 1343) and Gregoras was scornful of the clay utensils and sham vestments which were to be used at his second

coronation in 1347.2 Others would wish to lay the whole question

of the status of Ottoman women, and the harem, at a Byzantine door, for which there is marginally more justification than the view

that the so-called Ottoman ‘Law of Fratricide’ has Byzantine antecedents. More important is the question of dowry and _bride-price. P. Magdalino, ‘Byzantine churches of Selymbria’, DOP, 32 (1979), pp- 347; 350, notes two extramural churches, St. Agathonikos and St. Alexander.

1 Alderson, Structure, p. 77 and n. 3, on the ‘luxurious ceremonial still

surviving at the Byzantine court’, which Orhan, for one, ‘must have witnessed...’ 2 Gregoras. ii. pp. 788-9; T. Bertelé, ‘I gioielli della corona bizantina dati in pegno alla repubblica Veneta nel sec. XIV e Mastino II della Scala’, Studi in onore di Amintore Fanfani, ii (Milan, 1962), pp. 91-177.

3 See N. M. Penzer, The Harém (London, 1967), p. 288; a popular view 1s expressed by Selma Ekrem, Turkey old and new (New York, 1947), p. 73; ‘How, then did the Turkish women lose their ancient freedom? When the

Turks conquered Istanbul they discovered veiled women and apartments where they were segregated, for the Byzantine Greeks did not allow their women much liberty. The Turkish men were much impressed by what they saw...’ In fact urban, rather than ethnic, influences seem to have been more important; see Gertrude H. Stern, Marriage in early Islam (London, 1939), and ‘Harim’, Encyclopaedia of Islam, 2nd ed. As for fratricide, it is true that Andronikos III Grand Komnenos executed his brothers Michael and George on his accession to the throne of Trebizond in 1332, before any known cases of Ottoman fratricide. But it was hardly a Byzantine custom, or even a regular Ottoman one before Mehmed II institutionalized the practice, and it would be claiming much for a petty Grand Komnenos to influence a Sultan over a century later. See Doukas, ed. Grecu, pp. 37, 39, 71-3, 235, 287; Panaretos, ed. Lampsides, p. 64; Alderson, Structure, 27; but Miller, Trebizond, p. 43, reverses the claim and makes Andronikos III follow ‘Turkish practice’.

[485]

Greek historians on the Turks

Orthodox and Islamic canons recognize the dowry (mahr) but are in

direct opposition to Mongol and Tiirkmen custom, which is a bride-price. When in 1452 Sphrantzes led an embassy to find a bride for the last emperor of Constantinople, he found that the Grand Komnenos John IV of Trebizond (whose family had been marrying Tiirkmens for a century) still offered a dowry, but George VIII of Georgia (where Orthodox had been marrying Muslims for three centuries) demanded a bride-price. Where, as in the Caucasus, dowries and bride-prices existed side by side, the bride-price tends

to drive out the dowry.! It was clearly essential to the status of a Byzantine bride in a Muslim land that she have a dowry; equally it was a mark of ownership to pay a bride-price, which Muslim rulers would not wish to forgo. There is evidence that the Grand Komnenoi and their Tiirkmen allies in the Pontos solved the problem by

paying both dowries and bride-prices, leaving each party to interpret the results as it wished.? In this ambiguous field Kantakouzenos would naturally offer a dowry: the brideprice was Orhan’s military support. But, like the Tiirkmens of the Pontos,

Orhan may well have viewed the marriage of 1346 as a mark of vassalage. His ulu hatun (chief wife and mother of Murad I) was Niliifer, whom he married in 1299. Evidently a convert to Islam, she was the daughter of the Byzantine governor of Yarhisar. Alderson 1 Sphrantzes, pp. 74, 362; Makarios Melissenos makes George VIII say: ‘It is not our custom for women to give money to men when they are going

| to marry, but men to give money to women.’ Sphrantzes was surprised and replied: ‘I have never heard such a custom and law, as your highness mentions.’ The Georgian ruler laughed and went on to speak of other wonders. In 1452

a Byzantine emperor could hardly afford to accept a bride on such terms. But that the Georgian may not simply have been bargaining is shown by the fact that the Orthodox Ossetes abandoned the dowry system in favour of the bride-price (kalim) of their Sunni Chechen-Ingush neighbours: see J. F. Baddeley, The rugged flanks of the Caucasus, i (Oxford, 1940), pp. 207, 267. Cf. Doukas, trans. Magoulias, p. 272 and n. 51. Dede Korkut, ballad vi, confirms

that fourteenth-century Tiirkmens offered a bride-price to Byzantine fathersin-law: see Bryer, DOP, 29 (1975), pp. 135-6. But that the dowry also sur-

vived among the Ottomans is evidenced by The Turkish letters of Ogier Ghiselin de Busbecq, trans. E. S. Forster (Oxford, 1968), pp. 28, 118-19. * Bryer, DOP, 29 (1975), pp. 135-6.

[486]

had no doubt that the marriage ‘symbolised the annexation of her father’s estates’. The status of the bride

So far as Byzantines were concerned, the most satisfactory diplomatic marriage was perhaps their last. Theodora Grand Komnene became hatun of the Akkoyunlu ruler Uzun Hasan (1449-78) in 1458. Dowry and bride-price had been paid; she had been taken to her husband by a nymphostolos. High Commissioner for an empire which expired in 1461, she nevertheless maintained a Christian household with chaplain, and succoured her husband’s Christian subjects. Most remarkably, she conducted personal negotiations on behalf of her husband with Western powers.? More typical was perhaps the case of a certain still unidentified product of the palace

finishing school for Mongol brides. This Palaiologan princess became the third hatun of the Kipchak han Muhammed Ozbeg (1312-41). Like his other hatuns she had an impressively large household, and freedom enough to obtain permission to return home for the birth of her child in 1332. Ibn Battuta joined her caravan from Astrakhan to Constantinople. At the Byzantine border in Bulgaria he was disturbed to see that she abandoned her portable mosque and started eating pork and drinking wine: ‘Inner sentiments concealed [hitherto] suffered a change through our entry into the land of infidelity.’ The Palaiologan princess was evidently a closet Christian. 1 Alderson, Structure, p. 85; Niliifer (or Liiliifer), apparently the first Greek to enter the Ottoman dynasty, gave her name to a river near Bursa and to the Niliifer hatan imareti in Nicaea, built in 1388. Ashikpashazade in Vom Hirtenzelt zur Hohen Pforte, trans. R. Kreutel (Graz, Vienna, Cologne, 1959), p. 39. 2 An ulu hatun was the senior wife, a sort of Valide Sultan, probably by right

of mothering a designated heir. It is notable that no Byzantine princess is known to have reached this status of considerable authority. But a puzzling Greek ceramic epitaph of 28 September 1342 from Erzincan mentions an unnamed megale chatouna who apparently died as a presbyterissa: F. Cumont, ‘Inscription de |’époque des Comnénes de Trébizonde’, Mélanges Pirenne, 1

(Paris, 1926), pp. 67-72; J. E. Woods, The Aqquyunlu. Clan confederation, empire (Minneapolis and Chicago, 1976), pp. 101, 135, 158; Bryer, DOP, 29 (1975), pp. 137, 149 n. 146 to appendix ii. 2 Ibn Battuta, trans. Gibb, ii. p. sor. [487]

Greek historians on the Turks

The most surprising feature of her father’s account of Theodora’s

married life is not that she kept her faith and made no secret of it but that she encouraged Orhan’s subjects to revert to Orthodoxy. Kantakouzenos cannot be accused of ignorance of either Orhan’s Bithynia or of Islam—or rather anti-Islamic polemic, with which he

was to fill over 320 columns of Migne’s Patrologia, where one naturally looks in vain for any hint that the monk Joasaph had a Muslim son-in-law. The penalty for reversion to Christianity was death, last known to have been paid in that area as late in 1819. One of Kantakouzenos’s own protégés, the monk Meletios, could have told him that with some feeling, for Meletios had begun his career

as Muslim ulema in Orhan’s own capital of Bursa.1 By his own epigraphy and Ottoman historiography, Orhan was a ghazi: he could do no other than convert. The question is whether Kantakouzenos was disingenuously trying to justify his alliance with Orhan, or whether he was reporting actual conditions in Bithynia, where it is possible that things were still so fluid that reversion may not have been such a reckless act. Bithynia was the last Anatolian province,

apart from the Pontos, to experience wholesale conversion. As a

good ghazi, Orhan had urged his soldiers to take its Christian women as wives when he captured Nicaea in 1331.2 Anatolian Greeks commonly lost first their daughters, then their faith, and finally their language. But, aside from marriage, conversion was only the first stage in the process of turning a Greek into a Turk, which elsewhere took three generations or more. Orhan’s Bithynia was in the throes of the first stage for a generation after 1331, which (as Kantakouzenos knew) is unusually well documented. By 1338 large numbers of Christians of Nicaea had evidently converted, for Patriarch John XIV Kalekas (1334-47) sent them, then and in 1340, 1John VI Kantakouzenos, Contra Sectam Mahometicam Apologiae IV and Mahometem Orationes Quatuor, PG 154 cols 371-692, probably written in 1360;

R. Clogg, ‘A little-known Orthodox neo-martyr, Athanasios of Smyrna (1819)’, Eastern Churches Review, § (1973), pp. 28-36; Vryonis, “Human sacrifice’, p. 144.

2 Vryonis, Decline, p. 392 n. 126; Ashikpashazade, trans. Kreutel, pp. 67-8.

[488]

letters which are commonly taken as the first evidence of widespread

crypto-Christianity (public profession of Islam and private confession of Orthodoxy). He condoned it as the only alternative for those public converts to Islam who did not seek martyrdom. In fact it is not clear from Kalekas’s letters whether they are a statement of an existing situation or an incitement to crypto-Christianity—which, elsewhere, appears as a widespread phenomenon much later and not primarily through Muslim pressure.1 But it is clear that Kalekas knew the risks of reversion. Kalekas was no friend of Kantakouzenos,

or of another of the emperor’s protégés, (St.) Gregory Palamas, archbishop of Thessalonike (1347-59). But an ill wind drove Palamas on to Orhan’s shores and he was able to report on the situation at first hand. He visited the emir’s summer palace and Nicaea in 1354 as a prisoner before he was ransomed from Orhan (probably by Kantakouzenos). Palamas had a mixed reception, but was impressed by the theological tolerance of Orhan’s grandson and the friendliness of a crowd before which he debated with a mullah

in Nicaea: “One of them said, the time will come when we shall understand each other; and I am glad, and pray that the time may come soon. Palamas and the mullah were perhaps able to reason and hope, but others had evidently entered an unreasonable twilight between the two faiths in which there was no hope. Local Chiones, apparently Christians who had adopted Judaism as the best of both worlds, found themselves denounced as retrograde by both Palamas

1H. W. Lowry, ‘The Ottoman Tahrir Defters as a source for urban demographic history: the case study of Trabzon (ca. 1486-1583)’ (unpublished

ULCA Ph.D. dissertation, 1977) for the process of conversion; Vryonis, Decline, pp. 341-3; N. E. Meliores, Hoi Kryptochristianoi (Athens, 1962), pp. 37-9. The supposed continuity of these ‘crypto-Christians’ with the later phenomenon described, for example, in M. E. Durham, The Burden of the Balkans

(London, 1905), pp. 205-8, 291-2, is questioned in A. A. M. Bryer, “The Pontic revival and the new Greece’, Hellenism and the first Greek war of liberation

(1821-1830): continuity and change, ed. N. P. Diamandouros and others (Thessaloniki, 1976), p. 176.

4J. Meyendorff, A study of Gregory Palamas (London, 1964), p. 197; Sevéenko, ‘Decline’, p. 179.

[489]

Greek historians on the Turks

and the Muslims.! Palamas’s and Kalekas’s letters reveal a nebulous area between the two faiths (perhaps bridged by local cults) created by the strain of swift conversion, which had yet to be healed by a

generation or two of assimilation. But they do not reveal any evidence of reversion to Orthodoxy: the drift to Islam was inexorable. Even Kalekas had not advocated reversion, for it would have been simply irresponsible. By claiming that Theodora had

encouraged reversion, was Kantakouzenos doing no more than putting the best light he could upon her marriage? The political ends had at least justified the means, for Theodora brought the emir to see

his father-in-law at Skutari in 1347 (the forerunner of at least six similar Tiirkmen state visits to Trebizond after 1352) and Orhan remained an ally.? The marriage in canon law

So far as Orhan was concerned the events of 1346 constituted a marriage whether he was there or not. Theodora lay outside the degrees of affinity and her people had received the Book before the Prophet, while Orhan himself did not exceed the canonical limit of four wives. This makes Umur’s previous rejection of the Despoina all the more puzzling. In the second version of his tale, Enveri was following convention in making his ghazi hero reject the blandishments of women, particularly Christian. He had already made him reject the advances of the marchioness of Bodonitsa in somewhat similar circumstances.‘ But this is not Umur’s argument in the first

version of the story, which Lemerle found inexplicable.> It was that he was Kantakouzenos’s brother and so could not marry his own daughter. Kantakouzenos describes Umur as symmachos, but 1 Anna Philippides-Braak, ‘La captivité de Palamas chez les Turcs’, Travaux

et Mémoires, 7 (1979), pp. 204-5, 214-18. A satisfactory etyomology for *‘Chiones’ has yet to be found.

* Bryer, DOP, p. 29 (1975), p. 135.

>A, J. Arberry, The Koran interpreted, i (New York, 1955), pp. 58-9, 100-1, 103, 128; N. Daniel, Islam and the West. The Making of an Image (Edinburgh, 1960), pp. 135-46. “ Destan, Il. pp. 531-64; Lemerle, Aydin, p. 77. 5 Lemerle, Aydin, p. 175.

[490]

Enveri's term for the relationship meant more than that: the tekfur was Umur’s kardesh. The situation is understandable in the surviving Turkish institutions of kardesh, not just a brotherhood but a blood-

brotherhood. In fact Gregoras had got it right by comparing Kantakouzenos and Umur to Orestes and Pylades. A kardesh has an entrée into his adoptive family’s household which cannot be abused.

In late 1345 Kantakouzenos may quixotically have asked his symmachos Umutr’s permission to marry Theodora to Orhan because

he thought he had offered her to him first. But Umur, as kardesh, would have interpreted it as a fraternal discussion over the future of a daughter (or niece) who had once been inexplicably offered to him. So far as Kantakouzenos the emperor was concerned, he bravely describes the prokypsis of 1346 as a marriage. Theodora was at least outside the canonical age of twelve years for a bride and Gregoras was enthusiastic about the alliance, whereas he had condemned the notorious case of Simonis, whom Androikos II had given at the age

of five years to the mercy of the thrice-married barbarian, but nevertheless Orthodox, Milutin of Serbia.4 So far as the Church was concerned, the age limit was perhaps the only canon that the ‘marriage’ of 1346 did not break: they are too many to list here.? 1 Gregoras, i. p. 243; Nicol, The last centuries, pp. 126-7; there is a wallpainting of the unhappy bride at Grafanica. 2 Byzantine civil and canon lawyers would have found it difficult to know

where to begin, if they had wished to condemn imperial marriages with Muslims, because neither law embraced the matter. But impediments to marriage with heretics, infidels, and Jews, and to polygamous marriages, are abundant enough and even betrothal would have been illegal, for that had to be blessed by a priest. The problem had long risen, on a less scandalous Jevel, with the question of Orthodox marriages with Roman Catholics. See (for civil legal objections), K. Harmenopoulos, Procheiron Nomon to legomenon He

Hexabiblos (Athens, 1935), pp. 319, 348; and (for canon legal objections), G. A. Rhalles and M. Potles, Syntagma ton theion kai hieron kanonon (Athens,

1852-59), ii. pp. 251-4, 471, 498 (baptism, sometimes practised by Turks, was invalid because they regarded it as a sort of prophylactic only), pp. 500-1,

505-6; iil. 173, 180, 198-9, 364; iv. 337, 439, 476; D. M. Nicol, ‘Mixed marriages in Byzantium in the thirteenth century’, Studies in Church History, | (1964), pp. 160~72; Daniel, Islam and the West, pp. 146-8.

[491]

Greek historians on the Turks

So far as ordinary Byzantines were concerned, some had come to the ‘Moment of Vision’. Obscure but apparently popular feelings were to make Alexios III, Grand Komnenos of Trebizond, hold off

a full seventeen years before he gave Tajeddin beg an imperial bride in 1379—although there seems to have been no objection to her marrying an Orthodox Serb later. So far as Kantakouzenos, monk and Orthodox theologian, was concerned, the “Moment of Vision’ must have been painfully clear,

as his defensive account of the events of 1346 hints. He was in a unique position to know that the honeymoon was over before it had begun, and that the alliance was unequal. But it took Doukas to admit that the marriage between Theodora Kantakouzene and the ghazi emir had been no marriage at all. Doukas was writing in the same decade as Enveri. His terminology is as conventional as the

Ottoman poet's. Whereas Enveri had insisted on the chastity of Umur, Doukas makes Orhan insatiable, following traditional Christian medieval’ rumour of Muslim behaviour—although he knew Orhan’s successors as well as any Greek.? But Doukas, unlike Kantakouzenos, had nothing to gloss over, and his later account of

the sequence of events in 1346 is probably to be preferred to the

emperors own. Above all, his disgust with ‘this abominable betrothal’ is authentic enough: When Orhan heard [Kantakouzenos’s] ambassadors proposing this unexpected marriage contract and making promises of infinite treasures, he was like a bull which had been parched by the burning heat of summer, and was

with mouth agape drinking at a hole filled from the coldest water, but unable to get his fill; thus was he transformed as he listened because of barbarian incontinence. This nation is intemperate and lustful as no other people, incontinent beyond all races and insatiate in licentiousness. It is so inflamed by passion that it never ceases unscrupulously and dissolutely from

having intercourse by both natural and unnatural means with females, males and dumb animals. The people of this shameless and savage nation, moreover, do the following: if they seize a Hellene or Italian woman, or a 1 Panaretos, ed. Lampsides, pp. 74, 78-80.

* Daniel, pp. 141-6, 320-2. For Doukas’s other accusations of Ottoman immorality, see ed. Grecu, pp. 87, 123, 201, 211, 381-5. [492]

woman of another nation, or a captive or deserter, they embrace her as an Aphrodite or Semele, but a woman of their own nation or of their own tongue they loath as though she were a bear or a hyena. Orhan, therefore, when he heard of the proposed marriage with Kantakouzenos’s daughter (for she was beautiful in form and her countenance was not without grace) and the size of her dowry and the betrothal gifts sent by Kantakouzenos, quickly gave his consent.

1 Doukas, ed. Grecu, p. 59; trans. Magoulias, p. 73.

[493]

|V Une église “‘a la demande du client” a Trébizonde (!)

Notre point de départ est la méthodologie utilisée par Thomas F. Mathews dans son ouvrage The Early Churches of Constantinople: Architecture and Liturgy (Pennsylvania, 1971). L’idée fondamentale de Mathews est que les églises anciennes de Constantinople ont été construites pour étre le lieu de célébration de la liturgie. Puisque nous ne possédons plus la liturgie de cette époque, |’architecture de ces anciennes églises devrait nous servir de guide pour reconstituer la liturgie en question. Mathews était bien sir conscient des embiiches inévitables dans un tel raisonnement du genre: «de la poule ou de |’ceuf,

lequel est premier?» Par exemple, depuis la publication de son livre, des fouilles dans les fondations du skevophylakion préjustinien adjacent

a Sainte-Sophie de Constantinople ont eu lieu en 1980. Elles auraient révélé une communication avec |’église proprement dite, fait qu’ignorait Mathews, et l’on peut se demander en quoi cela risque d’affecter sa reconstitution de la Premiére Entrée de la liturgie des Sanctifiés (2). Mais il est trop t6t pour en discuter tant que les résultats de ces derni¢res foutlles n’ont pas été publiés.

Nous appliquerons donc pour Il’instant la méthode de Mathews a l’église de la Théotokos Chrysocéphale, maintenant appelée Fatih,

Orta du Biyiik Camii, a Trébizonde. Mon hypothése est que cette église a été batie, ou plutét rebatie, au début du treiziéme siécle, dans une intention bien particuliére: celle d’y célébrer une liturgie de couronnement. La description qui suit sera publiée également, sous une forme différente, dans le volume a paraitre: Byzantine Monuments and Topo-

graphy of the Pontos, comme Dumbarton Oak Study. Cette église y figurera sous le n° 120, dans le chapitre 20 consacré a la ville de Trébizonde elle-méme.

(1) Cette étude a été présentée ala journée Conférence annuelle sur les Eglises

orientales organisée par les Selly Oak Colleges et le Département de Formation continue de l’Université de Birmingham, le 16 mai 1981, sur le théme: «Architecture des Eglises orientales». C’est avec une émotion profonde que je me permets de dédier cet article a la mémoire du Professeur Jean-Michel Hornus: sa mort tragique, survenue en mars 1982, ne m’a pas permis de lui exprimer toute la reconnaissance que je lui dois pour la conception, la rédaction, et la traduction des idées ici exposées. (2) Mathews, pp. 156-62.

217

Situation

La cathédrale conventuelle de Trébizonde se trouve au milieu de la ville fortifiée moyenne, originellement entourée de constructions monastiques a arcades et, sans doute, du palais épiscopal. Cet environnement a ensuite été remplacé par un foisonnement de magasins et de boutiques qui ont été enlevés par les Russes en 1917. Inscriptions

I. L’inscription du linteau nord, sur une dalle de marbre probablement en provenance d’une autre construction et se rapportant | a l’empereur Hadrien, était recouverte en 1970 et semble avoir disparu avant 1973 (3).

II. Une inscription datée de 913/14, se rapportant a la restauration de son tréne par le métropolite Basile et en formant, semble-t-1l, la partie postérieure, trouvée sur le sol en 1877 et mentionnée en 1879 (4),

peut étre considérée comme perdue, ainsi qu’une autre version non datée de l’inscription qui avait été portée a l’église St. Georges de Kourtza (5).

III. Une inscription AAEEIOX KOMNHNOE sur un bloc de marbre de Im 80 de long, portant en son centre une téte de taureau en relief doré avec une palmette de chaque cété — sans doute une pierre de l’époque classique réutilisée par la suite — trouvée également sur le sol en 1877 et mentionnée en 1879, est aussi considérée comme perdue (6).

IV. L’épitaphe du grand oikonomos Zacharie incorporée dans le Kulakli Cesme, a l’extérieur de la cathédrale, a disparu avec la destruction de la fontaine aprés 1918 (7). V. Un bloc dans le mur supérieur nord de la baie C du plan C et de la planche 2) portant en lettres trés légerement gravées, de 11 cm de

haut: ... ION EIII BACIA ..., associé 4 une série de blocs réutilisés portant un relief aussi trés léger caractérisé par un motif en corde sinueuse et fleurs qui se continue autour de la fenétre de la galerie orientale au-dessus de la baie C (pl. 3) (8). (3) M. Paranikas, «’Emypagai xai vopiopata Tpanelobvtog», CPSyll, 29 (1907), p. 301. (4) Marengo, Missions Catholiques, 11 (1879), p. 302.

(5) Millet, BCH, 19 (1895), p. 422 et fig. 1. (6) Marengo, loc. cit. (7) Millet, BCH, 19 (1895), pp. 424-5 (fac-similé et transcription), a préférer a@ Paranikas, CPSyil, 29 (1907), p. 300.

(8) Pratiquement invisible du sol (comme sur la pl. 3). J’ai été autorisé a accéder au toit de la travée D du plan le 22 septembre 1979 grace a la bienveillance des responsables de l’Ayasofya Miizesi et du Fatih Camii.

218 Une église «a la demande du client» 4 Trébizonde Architecture

Une basilique 4 déme 4 trois nefs avec (a l’origine) une seule abside pentagonale, le narthex, l’exonarthex, les galeries au-dessus des nefs et des narthex, une piéce (indiquée D sur le plan) au nordouest et enfin les porches nord et sud. Ce dernier, le porche sud (B), a disparu depuis longtemps, ses portes ont été bloquées et l’une d’elles utilisée pour servir de mihrab (9). Comme on pouvait s’y attendre concernant |’église la plus impor-

tante de l’Empire de Trébizonde, les structures de la Chrysocéphale ont été remaniées plusieurs fois. Le changement de plan le plus important dans la présente structure a été que ce qui était probablement a l’origine une basilique a galeries s’est vu ajouter un déme et une croisée,

longitudinale au-dessus de la nef et transversale au-dessus des bascétés. Cela sera discuté plus loin. Décoration Des panneaux de bois cachent le plancher opus sectile et recouvrent

aussi les peintures murales et la mosaique de l’abside, qui représente peut-étre la Mére de Dieu. Les dalles de marbre des murs de |’abside centrale demeurent visibles (pl. 4 et 1). Les «restes considérables d’une brillante mosaique sur les murs extérieurs» (10) notés par Finlay sont aussi recouverts. Ils avaient été décrits comme représentant soit l’Annonciation soit les saints Constantin et Héléne. Un enduit d’époque turque recouvre probablement des peintures murales byzantines dans l’ensemble de l’église. Le porche nord emploie des colonnes ioniques réutilisées, en marbre cannelé et avec chapiteaux. Il ne demeure plus aucune trace de ce qui a dd étre la plus splendide collection de monuments funéraires et d’éléments pour la liturgie impériale trabizontine. Au voisinage de |’église, le Kulakli Cesme, la tombe d’Alexis IV et la fontaine a téte de dragon d’Alexis II ont tous disparu depuis 1918, et la fontaine 4 téte de dragon méme peut-étre déja en 1877. Il y avait, (9) La bibliographie de la Chrysocéphale est abondante. En plus de Millet, Talbot, Rice et Ballance, une étude particuliére lui a été consacrée dans N. Baklanov, B, 4 (1928), pp. 377-391. Voir aussi Bzhshkean, tr. Andreasyan, pp. 46, 47; Chrysanthos, AP, 4-5 (1933), pp. 221-2, 229, 240, 246, 373, 375, 379-386, 389-393, 531; Janssens, Trébizonde, pp. 24, 74, 76, 84, 91, 108, 132, 159, 220-1, fig. 35; Diehl, Ma-

nuel, pp. 769, 785-6; I. E. Kalphoglou «‘O év TpaxeCodvti vads tic GeotdKov XpvooKxegdrov», Argonautes, 5 (1916), pp. 9-10 (que je n’ai pas pu consulter); Millet,

BCH, 19 (1895), pp. 420-5, 441, 445-8, 451-3, 458; Millet et Talbot Rice, Painting, p. 112; Papamichalopoulos (1901), pp. 190-3; Ritter, Erdkunde, XVIII, p. 881; Rottiers (1829), p. 205; Succi, Trebisonda, pp. 216-220; Talbot Rice, B, 5 (1930), pp. 51-4, Pls. 2-6; Charles Texier et R. Popplewell Pullan, Byzantine Architecture (Londres, 1864), pp. 198-9. (10) Finlay (MS, 1850), f. 43a; Marengo, Joc. cit.

219

peut-€tre fixée a une colonne, une Panagia Chrysocéphale portant l’Enfant, a laquelle Andronic le Gidon offrit des pierres précieuses et des perles provenant du butin conquis lors de sa victoire de 1223 sur un certain Mélik. Un centre d’intérét ultérieur pour les pélerins doit avoir été la tombe thaumaturgique de Denys, fondateur (avec Alexis III) du monastére «de Denys» au Mont Athos. Ce Denys fut enseveli 4 la Chrysocéphale entre 1382 et 1389 (11). Histoire

La Chrysocéphale fut l’église métropolitaine (12), de couronnement (13) et d’inhumation (14) des Grands Comnénes, un refuge en cas de danger (15) et l’église conventuelle de ce qui doit avoir été l’un des plus riches monasteres de l’Empire. Lazaropoulos lui applique toujours des périphrases ronflantes. Comme la Sainte-Sophie de Constantinople c’est simplement la peyaAn éxxAnoia. Dans la bulle du 26 octobre 1314 c’est l’ecclesia beata Maria Crisocofola sita in Castro Trapezonde, Et méme le Continuateur habituellement sobre de Panaréte va jusqu’a l’appeler «la trés sacrée église de la Plus-que-sainte Théotokos

Chrysocéphale» (16). L’épithéte «a téte d’or» a d’habitude été comprise comme faisant allusion aux tuiles de bronze doré dont on pense qu’elles recouvraient le ddme. Mais Lazaropoulos donne des indicatons, rétrospectives montrant que l’église était déja désignée par ce qualificatif sous le régne de Constantin IX Monomaque (1042-55). La correspondance de Théodore Prodromos avec le métropolite Etienne Skylitzes (1120-40), puis l’oraison funébre qu’il prononca pour celui-ci sont aussi témoins de |’utilisation de cette épithéte. Lazaropoulos pense

qu’elle s’explique par la présence dans |’église d’une icéne a dorures de la Théotokos (17). (11) Lazaropoulos in éd. Papadopoulos-Kerameus, FHIT, p. 131; Janin, EMGCB, p. 277; Oikonomides, Dionysiou, p. 12. Je remercie le professeur Spéros Vryonis Jr., pour ses remarques sur ce point. (12) En plus du tréne métropolitain de 914, la Chrysocéphale contenait la tombe du métropolite Barnabas (1333), dans laquelle le métropolite Basile fut enseveli ensuite (1364); Panaretos, éd. Lampsides, p. 15. (13) Le Grand Comnéne Jean III fut couronné a l’ambon de la Chrysocéphale en 1342; Panaretos, éd. Lampsides, p. 67. (14) La Chrysocéphale contenait les tombes du Grand Comnéne Andronic [et Gidos (1235), de Jean II (1297), Théodora Cantacuzéne (1426) et (a l’extérieur) probablement d’Alexis IV; Panaretos, éd. Lampsides, pp. 63, 65. (15) Lazaropoulos in éd. Papadopoulos-Kerameus, FHIT, p. 118. (16) ASL, xiii (1884), p. 517; Lazaropoulos in éd. Papadopoulos-Kerameus, FHIT, p. 115; Panaretos, éd. Lampsides, p. 81: «év t@ navoént@ va® tific dnepayiag Geotdékov tfi¢ XpvcoKxepdAov».

(17) Lazaropoulos, dans éd. Papadopoulos-Kerameus, FHIT, pp. 76, 115, 131; Uspenskij, Ocherki, pp. 15-16; Baklanov, B, 4 (1928), p. 389; Migne, PG,

220 Une église «a la demande du client» a Trébizonde

Aprés 1461 la Chrysocéphale devint la mosquée principale de Trébizonde et les propriétés du Hrisokefal Manastiri passérent au Fatih Sultan Evkaf, qui existe encore et en détient sans doute toujours une partie. Le monastére jouissait alors d’un revenu mensuel de 1.208 aspres et possédait des terrains 4 Chortokopion dans la circonscription de Matzouka (Ortahortakop et Yukarihortokop, dont les noms ont maintenant été transformés, avec un manque complet d’imagination, en Ortakéy et Yukarikéy, 4 Macka) (18). Comme Balkanov, Sélina Ballance a souligné que beaucoup des secrets de l’architecture de la cathédrale chrysocéphale et de son katho-

licon demeurent cachés sous |l’enduit qui revét ses murs. Mais elle a proposé une séquence architecturale qui entraine |’adhésion et elle a en outre proposé des hypothéses de datation (voir le plan) (19). Ce sont, successivement:

1) Une ou plusieurs église(s) ancienne(s) située(s) sur le site et de date inconnue, ow le tréne du métropolite a été restauré en 913/14; 2) une basilique entiérement nouvelle, construite au dixi¢éme ou onziéme siécle et constituant la présente structure de l’abside jusqu’au narthex, avec les six travées de la nef et des bas-cétés de |’église;

3) un remaniement majeur au douziéme siécle, comprenant Pélévation de la voite et l’adjonction de la croisée et du déme, ainsi peut-étre que l’addition de l’exonarthex. Ce remaniement n/’aurait entrainé aucune modification de |’organisation du béma ni du plan au sol de l’étape précédente;

4) addition au treiziéme siécle de la piéce D au nord-est; 5) addition a la fin du treiziéme ou au quatorziéme siécle du portail nord (sans prendre en considération le portail sud, B) accompagnée peut-étre d’un élargissement des fenétres de l’abside; 6) addition de l’absidiole sud, E, au quatorziéme ou au quinzieme siécle.

Cette succession d’étapes architecturales est logique et doit étre acceptée, sous réserve que de futures investigations derriére l’enduit de revétement et sous le plancher ne découvrent des indications conCXXXIII, col. 1257; L. Petit, «Monodie de Théodore Prodrome sur Etienne Skylitzes, métropolitain de Trébizonde», Bulletin de l'Institut archéologique Russe a Constantinople, 7 (1903), p. 12.

(18) Gdékbilgin, TTKB, 26 (1962), pp. 309, 315; cf. Vryonis, Decline, pp. 324-5. Il semble y avoir la une confusion avec Sainte-Sophie.

(19) Ballance, AS, 10 (1960), p. 151; suivie par Demetrokalles, MC, 13 (1967), pp. 112-4, figs. 21, 22; et par R. Krautheimer, Early Christian and Byzantine Architecture, pp. 280-1, 353 n. 55, qui, cependant, date la troisiéme étape du treizi¢me si¢cle, sans en donner de raison.

221

traires. Mais la datation ne tient aucun compte des indications historiques, qui ne peuvent pourtant pas étre ignorées. Du point de vue historique, en effet, il est clair que a) l’église a été remodelée en vue de certaines fonctions particuliéres aprés 1204, b) ce remodelage peut probablement étre daté de la période 1223-35, c)

’église a été entiérement reconstruite durant la période 1341-51, enfin, d) certaines parties du batiment n’ont été touchées ni en 1364 ni en 1426. Ces faits correspondent aux étapes de l’architecture mais non aux dates proposées par Sélina Ballance. La premiére étape est la premiére église (ou les premiéres églises)

existant sur le site. La tradition d’une fondation remontant au quatrieme si¢cle doit étre rejetée (20). Mais on peut raisonnablement supposer que la situation privilégiée, au coeur de la Ville moyenne, a amené la construction d’une église a cet endroit bien avant les témoignages épigraphiques les plus anciens révélant qu’elle était la cathédrale de Trébizonde avant 913/14: le premier évéque (par opposition au futur métropolite) attesté a Trébizonde est mentionné dés 253/4 (21).

Cette église, ou une autre existait 4 la date 913/14 puis aux périodes 1042-55 et 1126-40.

Le second élément connu est que de_ riches dotations furent affectées a l’église durant la période 1223-35. Cela a tenu une place importante dans l’attaque de Trébizonde par le Mélik en 1223. Le butin en a fourni a la Chrysocéphale un splendide évangéliaire, «des plerres précieuses et des perles resplendissantes» (22). Le principal

héros du siége, Andronic I& le Gidon (1222-35), passa la nuit la plus tragique de ce siége a prier a |’église avec un moine de la Chrysocéphale

appelé Gérasime. Andronic le Gidon, second Grand Comnéne de Trébizonde, fut enseveli a la Chrysocéphale en 1255. Sa tombe était dans le parabéma que j’identifie avec la travée sud-est (E) du plan. Cette tombe se trouvait encore la, intacte, quand Théodora Cantacuzene y fut ensevelie 4 son tour en 1426 (23). On peut en conclure que

la disposition de ce qui se trouvait du cété oriental était restée (20) La tradition attribuant la fondation 4 Hannibalien, neveu de Cons‘tantin le Grand, date du dix-septiéme siécle: voir Fallmerayer, OF, I, p. 140; Chrysanthos, AP, 4-5 (1933), p. 380. Lebeau (Brosset), Bas-Empire, XX, p. 487, n. 3, semblant citer Bzhshkean (1819), p. 66 (et non 106), a une version aberrante, selon laquelle la Chrysocéphale aurait été construite comme monastére de femmes par Flavius Julien Constantin, «roi du Pont». L’idée que Justinien chargea les Génois de construire sur le site une cathédrale a St André, avancée dans Succi, Trebisonda, p. 216, est curieuse. (21) Chrysanthos, AP, iv-v (1933), pp. 182-191.

(22) Lazaropoulos, dans éd. Papadopoulos-Kerameus, FHIT, p. 131; Chrysanthos, AP, iv-v (1933), pp. 245-6.

(23) Lazaropoulos, dans éd. Papadopoulos-Kerameus, FHIT, p. 120 ;Panaretos, éd. Lampsides, p. 81, «év t@ koyntnpia tob Tidwvos sig td napapnyuo.

222 Une église «a la demande du client» a Trébizonde pratiquement identique de 1235 4 1426. Une déduction moins certaine peut étre tirée du fait que l’inscription III, d’Alexis Comneéne, fut incorporée dans une dalle de |’a4ge classique réutilisée (sans doute téte en bas)

pour le plancher opus sectile de l’église. Il est peu vraisemblable que cette inscription se rapporte a l’empereur byzantin Alexis I¢* Comnéne

(1081-1128), car Trébizonde fut en quasi rébellion pendant la plus grande partie de son régne(24). Le Comnéne le plus ancien a qui elle puisse

se rapporter est alors Alexis I¢* Comnéne (1204-22), le premier Grand Comnéne. Ceci peut signifier que le plancher en mosaique de la Chrysocéphale fut posé aprés 1222. Une réalisation semblable, provenant de Trébizonde, a Sainte-Sophie, date de 1250-70, a St-Eugene de 1291,

a St-Michel de Platana de 1300 environ. A Constantinople le plancher de l’église de Stoudion peut dater d’aprés 1261. Le plancher de la Chrysocéphale peut donc étre le premier d’une série réalisée au treizieme siécle 4 Trébizonde. L’inscription V peut aussi étre mise en relation avec l’inscription III car, bien que déja réutilisée a ce stade, la forme de ses lettres est tardive, datant peut-étre du treizi¢me siécle et en tous cas sans doute pas antérieure a celui-ci. Si tel est le cas, l’em-

pereur a qui elle semble faire allusion pourrait étre aussi Alexis [¢'. La troisitéme étape de l’histoire de la Chrysocéphale peut sans doute étre confondue avec la seconde mais doit cependant étre étudiée a part. Les stratéges locaux et les ducs de Chaldie avaient exercé un pouvoir pratiquement autonome sur Trébizonde avant 1204. Mais tous tenaient leur autorité de l?empereur de Constantinople. C’était donc au tr6ne du métropolite que revenait la place d’honneur dans la cathédrale de 913/14. De 1204 a 1282 le souverain Comnéne de Trébizonde gouverna au contraire le Pont sur une base toute différente. Il ne se considérait plus comme un empereur byzantin mais bien comme /’empereur byzantin: le seul Empereur des Romains. Dans les circonstances tout a fait nouvelles qui prévalaient, il ne pouvait plus bénéficier de son droit légitime a étre couronné a Sainte-Sophie de Constantinople. Mais son prestige dynastique était bien plus grand que celui des Lascarides

et des Anges d’Occident, qui commengait juste a se développer. Et c’était 4 peine si on entendait déja parler des Paléologues. Jusqu’en 1214 la prétention d’Alexis et de David Comnéne de se faire couronner,

ou recouronner a Constantinople apparait comme sérieuse et raisonnable. Seule leur défaite par les Lascarides et les Seljoucides en 1214 leur enleva toute chance réelle d’atteindre la Capitale. Mais la prétention demeura. C’est seulement quand Michel Paléologue eut conquis Constantinople en 1261 qu’il devint évident que Manuel 1¢™ Comnéne n’y serait jamais couronné. Lorsque ce méme Michel signa son accord

avec Jean II de Trébizonde en 1282, les Grands Comnénes en la personne de ce dernier renoncérent a leur prétention d’étre les seuls em(24) W. Fisher, «Trapezus im 11 und 12 Jahrhundert», Mitteilung des Instituts far Osterreichische Geschichtsforschung, x (1889), pp. 177-207; Bryer, Gabrades, pp. 175-7.

223

pereurs des Romains. Ainsi la période durant laquelle il fut important que les Comnénes soient couronnés comme empereurs des Romains mais durant laquelle aussi il était évident que ce couronnement ne pouvait avoir lieu qu’a la cathédrale de Trébizonde, et non a Constantinople, s’étend de 1214 a 1282. A partir de 1282 cependant, le principe fut admis qu’il pouvait y avoir simultanément plusieurs empereurs byzantins dans des régions différentes. Mais, en tant qu’empereurs de Tout |’Orient, les Grands Comnénes continuérent a exiger jusqu’a la fin de leur dynastie un cérémonial de cour, des habits, une administration et une étiquette extrémement stricts (25). Un empereur byzantin, et particuliérement quelqu’un soucieux

d’affirmer son droit a ce titre, avait besoin d’une cathédrale, et de préférence une cathédrale imposante, comportant certains éléments qui ne faisaient plus partie de l’architecture habituelle des douziéme et treiziéme sié¢cles et que, pour différentes raisons, on ne trouve normale-

ment plus aprés |’iconoclasme. L’église du couronnement devait au moins comporter un métatorion (vestiaire ou piéce pour |’habillement), un ambon spacieux (chaire probablement située au centre-est et assez

vaste pour contenir plusieurs personnes lors du couronnement proprement dit), l’évangéliaire le plus splendide que 1’on puisse trouver afin de prononcer sur lui le serment du couronnement, enfin des catéchouména (galeries desquelles on pouvait voir l’empereur en prokypsis et l’acclamer et dans lesquelles il pourrait en temps normal recevoir la communion). La Chrysocéphale est, de toutes les églises de Trébizonde, la seule a avoir recu ces différents éléments. A Sainte-Sophie de Constantinople il semble que le métatorion, ou un métatorion, se trouvait dans la travée sud-est de l’église qui n’avait qu’une seule abside centrale (26). A la Chrysocéphale le métatorion se trouvait probablement dans la travée nord-est (C). Elle est la seule église de Trébizonde que nous connaissions et qui ait trois (25) Miller, Trebizond, pp. 18, 28-9. Aprés 1282 les Grands Comnénes sont remarquables par la correction du costume dans lequel ils apparaissent sur les peintures murales, les piéces et la chrysobulle pour Dionysiou (qui est elle-méme un cas extréme de diplomatique archalsante, revenant aux onzi¢me et douziéme siécles). On leur connait quinze offices attribués a leurs courtisans (y compris celui de tatas et ceux de parakoimomenoi) et ils acceptaient les acclamations de Paques comme faisant partie du cycle liturgique impérial: Panaretos, éd. Lampsides, p. 75; Oikono-

mides, Dionysiou, pp. 50-61; du méme, «The chancery of the Grand Komenoi: imperial tradition and political realitys, AP, 35 (1978), pp. 299-332. (26) Cyril Mango, «The Brazen House. A study of the vestibule of the imperial palace of Constantinople». Arkaeol. Kunsthist. Medd. Dan. Vid. Selsk., IV (4) (1959), pp. 64, 72, 89 n. 82, 90 n. 86, 91; Mathews, Early Churches, pp. 96, 132, 134; J. B. Papadopoulos, «Le mutatorion des églises byzantines», Mémorial Louis Petit (Bucarest, 1948), pp. 366-372; H. Kahler et C. Mango, Hagia Sophia (London, 1967), pp. 64-65.

224 Une église «a la demande du client» a Trébizonde nefs avec une seule abside, caractéristique que l’on avait jusqu’a présent considérée comme impliquant une date ancienne (27). Les bascétés de la Chrysocéphale se terminent a l’est par des piéces carrées a voute en déme. Plus tard la piéce du sud-est fut agrandie par |’ajout d’une abside arrondie (E). La piece du nord-est ne recut jamais d’abside mais contient le seul accés aux escaliers montant aux galeries. On lui ajouta ensuite une salle bien plus vaste sur le cété nord (pl. 4). Les piéces pour l’habillement impérial exigent un accés direct au béma et aux galeries, sans que |’on ait a traverser le corps principal de l’église ni méme 4a y entrer.

Mais elles n’ont pas besoin d’abside. C’est probablement parcé que la travée nord-est de la Chrysocéphale était le métatorion que |’église dans sa seconde facture eut une seule abside, centrale. Le métatorion est une piéce relativement petite et il n’est guére étonnant qu’il ait fallu le compléter par d’autres constructions au cours de l’histoire de 1’?Empire de Trébizonde. Sélina Ballance écrit: «la fonction des adjonctions au nord-est demeure incertaine» (28). En fait elles peuvent avoir rempli toute une série de fonctions: comme voie d’entrée du Grand Comnéne lui-méme et comme vestibule de |’église; comme salle 4 manger (ces deux fonctions trouvent leur équivalent 4 Sainte-Sophie de Constantinople); comme skévophylakion afin d’y conserver les évangiles et les regalia impériaux pour lesquels aucune autre disposition n’avait été prévue; ou tout simplement comme un métatorion (D) plus spacieux. Le second élément que les nouveaux empereurs byzantins auraient di ajouter a leur cathédrale de Trébizonde au cas ou ils ne |’y auraient pas déja trouvé aurait été un ambon. Les ambons, avec des balustrades solea les reliant au béma, jouaient un réle important pour |’ancienne célébration liturgique byzantine. Mais ils disparurent avec la transformation du rite de déambulation liturgique de l’évangile. A l’origine l’évangile avait été lu depuis l’ambon. Mais ensuite la procession fut réduite 4 une simple apparition avec lecture aux portes du béma, comme cela se pratique encore dans la liturgie orthodoxe actuelle. Il n’y a pas de trace matérielle d’ambons aprés la période iconoclaste: les planchers de la basilique du Stoudion a Constantinople et de celle du Pantokrator (douziéme siécle) ne révélent aucune trace de fixation pour l’ambon (29). Mais aprés que l’ambon eut perdu ses fonctions dans la liturgie de Constantinople, il demeurait nécessaire d’en conserver au moins un pour une autre fonction: le couronnement de |’empereur. Ce n’est pas dire qu’un empereur ne pouvait étre couronné qu’a |’ambon. Certains ne le furent pas. Mais au quatorziéme siécle, et jusqu’au (27) Ballance, AS, 10 (1960), p. 149; «L’abside unique dans une église a trois nefs était assez courante dans les églises les plus anciennes, disons jusqu’au huiti¢me siécle, mais est extrémement rare ensuite (sauf peut-tre en Lycaonie).» (28) Ballance, AS, 10 (1960), p. 148. (29) Mathews, Early Churches, pp. 179, 180 n. 5.

225

rituel de couronnement de Manuel II Paléologue en 1391, l’ambon est expressément mentionné (30). Peut-étre un ambon subsista-t-il a la Chrysocéphale (comme c’est le cas a la cathédrale de Kalambaka ou a Sainte-Sophie de Thessalonique) en tant que survivance encombrante du passé. Mais on se demande si un tel ambon aurait été assez spacieux pour les couronnements une fois que les Grands Comnénes se posérent en prétendants au tréne byzantin, a partir de 1240 (31). De toute maniére Panaréte témoigne de l’existence de cet élément archaique de mobilier liturgique puisqu’il écrit que, le 9 septembre 1342, le Grand Comnéne Jean II] «fut couronné a l’ambon de la Chrysocéphale» (32) par le métropolite Acace. Le secret de l’endroit exact ot l’ambon était érigé a des chances de se trouver dans le dessin du sol en mosaique, recouvert aujourd’hui par le plancher de bois de la mosquée.

Mais, organiquement et traditionnellement, l’emplacement le plus appropri€é pour celui-ci serait le centre-est de l’église (A sur le plan). Le troisiéme élément nécessaire a un couronnement était un évangéliaire assez magnifique pour la circonstance. La cathédrale en possédait certainement un. Mais, apres la victoire sur le Mélik en 1223,

Lazaropoulos rapporte qu’elle regut en don un manuscrit des quatre évangiles particuliérement splendide. Le 30 juillet 1341, le Grand Com-

néne Michel arriva par mer pour réclamer le tréne. Panaréte écrit: «Dans la soirée les archontes descendirent avec un formulaire de serment ainsi que le métropolite Acace avec l’évangéliaire, et ils le reconnurent comme souverain légitime ...». On n’eut toutefois pas le loisir de le couronner car il fut déposé dés le lendemain. En 1858 le métropolite Constant de Trébizonde offrit au Tsar Alexandre II un bel évan-

géliaire enluminé supposé provenir de la Chrysocéphale et qui est maintenant le MS Léningrad Gr. 21-21a (33). Savoir s’il est le méme évangéliaire que ceux mentionnés en 1223 et en 1341 demeure matiére a conjecture.

Le quatrieme élément nécessaire pour un couronnement était les galeries qui, tout comme |’ambon, sont bien plus fréquentes dans

les églises byzantines anciennes que dans les plus tardives. Baklanov | tient pour assuré que les galeries existaient sur trois cétés de la Chry-

(30) Pseudo-Kodinos, éd. Verpeaux, pp. 257-261, 353-6. Cf. B. de Khitrowo, Itinéraires russes en Orient, I (Genéve, 1889), pp. 143-47; Kahler and Mango, Hagia

Sophia, pp. 66-67. L’ambon de Sainte-Sophie de Constantinople fut reconstruit aprés qu’une partie du déme se fut effondrée sur lui en 1346. (31) C. Mango, Byzantine architecture (New York, 1976), 132; J. Darrouzés, «Sainte-Sophia de Thessalonique d’aprés un rituel», REB, 34 (1976), pp. 45-78. (32) Panaretos, éd. Lampsides, p. 67: «ei¢g tiv XpvcoKxégadov Ev 16 Gu Bavv».

(33) Lazaropoulos, dans éd. Papadopoulos-Kerameus, FHIT, pp. 131-32; Panaretos, éd. Lampsides, p. 66; Chrysanthos, AP, 4-5 (1933), pp. 390-92, 622-73; Janin, EMGCB, p. 279; V. D. Likhachova, Byzantine Miniature (Moscou, 1977), pl. 5 a 10.

226 Une église «a la demande du client» a Trébizonde socéphale, c’est-a-dire 4 l’emplacement destiné aux femmes. L’utilisation exacte, et méme |’existence tout court, de telles galeries pour les femmes a été mise en doute depuis. Dans les églises byzantines anciennes un bas-cété semble avoir été parfois réservé aux femmes. Mais il y a peu d’indications permettant de supposer que les sexes aient été séparés dans les églises byzantines postérieures. Presque toujours les galeries des églises byzantines anciennes sont désignées non pour les femmes mais pour les catéchuménes. Ces catéchouména commencérent 4 disparaitre en méme temps que la ségrégation des catéchuménes en classe séparée. Au séptiéme siécle, certains de ces catéchouména abandonnés qui n’avaient pas été transformés en oratoires étaient sujets de scandale, car certains s’en servaient comme lieu de rendez-vous galants. Pourtant quelques galeries apparaissent a nouveau, souvent comme éléments architecturaux sans signification précise, sur les franges

du monde byzantin aprés l’iconoclasme. On les retrouve du neuviéme au douziéme siécle en Bulgarie, en Italie, 4 Kastoria, au Lac Prespa et en Nubie. Elle apparaissent aussi a Mistra sous les Paléologues (34).

Comme pour ce qui concerne |’ambon, seul le rituel byzantin tardif exigeait encore des catéchouména. I1 utilisait explicitement ce terme archaique pour désigner les galeries dans lesquelles l’empereur devait monter — au cours de la cérémonie de présentation appelée prokypsis — pour étre présenté a son peuple et acclameé par lul, aprés avoir été couronné 4 l’ambon et avoir prété serment sur les évangiles (35). Ces gale-

ries n’étaient donc en rien semblables aux oratoires supérieurs surplombant le narthex de Sainte-Sophie de Trébizonde ni aux autres galeries byzantines tardives, mais ouvraient au contraire du sol vers le haut. Il y a aussi une fenétre plongeant sur le béma depuis la galerie au-dessus de la piéce E au sud-est. Une galerie trilatérale de cette sorte,

avec de grands arcs supérieurs, fut construite par Yaroslav de Kiev pour son église royale de Sainte-Sophie 4 partir de 1037 (36). Les arcs ouverts de la galerie de la Chrysocéphale (pl. 1) sont semblables, quoi(34) Mathews, Early Churches, pp. 129-133; Krautheimer, Architecture, pp. 191, 204-5, 210, 212, 220, 226, 241, 256, 270, 286; C. Delvoye, s. v. ‘Empore’, Real-

lexikon zur byzantinischen Kunst (Stuttgart, 1963-). Les femmes furent pourtant placées derri¢re des tentures de soie dans les galeries de Sainte-Sophie de Constantinople lors du couronnement de Manuel II en 1391; voir Khitrowo, Itinéraires, I, p. 143.

(35) Pseudo-Kodinos, éd. Verpeaux, pp. 32-3, 269; Mathews, Early Churches, pp. 128-9. Les Grands Comnénes semblent avoir paradé au square Meidan a Paques et le palais possédait une tribune. C’est peut-étre de celle-ci qu’il s’agit pour Giilhan-Eudokia, épouse du Grand Comnéne Manuel III, qui fut «couronnée dans le prokypsis impérial» en 1377; voir Panaretos, éd. Lampsides, pp. 75, 78; Bessarion, éd. Lambros, NE, 13 (1916), p. 189; et A. Bryer, «Byzantine historians on the Turks: the case of the first Byzantine-Ottoman marriage», Essays presented to Sir Richard Southern (Oxford, 1981), pp. 482-84.

(36) V.I. Lazarev, Mozaiki Sofii Kievskoy (Moscou, 1960), pp. 39-42.

227

qu’a une échelle plus réduite, et paraissent avoir joué le méme réle. Plus tét, et peut-étre aussi 4 la Chrysocéphale, la cour et le sénat recevaient la communion dans la galerie, ot se tenaient aussi les synodes. Pour résumer ces seconde et troisiéme étapes, on peut dire qu’aprés

1214 la construction d’une église contenant certains éléments archaiques: un métatorion, un ambon et des galeries dites catéchouména, doit avoir été une nécessité urgente pour les Grands Comnénes. Lazaropoulos et Panaréte écrivaient plus d’un siécle et demi aprés. Mais le

premier se souvient de donations massives faites durant la période 1223-35 et tous deux laissent entendre que le plan au sol refléte une organisation demeurée identique aprés cette période. Non seulement la tombe de Gidon, datant de 1235, était encore la dans le parabéma en 1426, mais celle du métropolite Barnabas (1333) était toujours intacte a la Chrysocéphale quand le métropolite Basile y fut enseveli en 1364 (37). Nous pensons donc que |’église (ou les églises) qui se trouvait(aient) la du dixiéme au douziéme siécle fut (ou furent) completement reconstruite(s) avant 1235 et certainement aprés 1204, probablement méme aprés 1234, sous la forme d’une basilique englobant le batiment actuel, de l’abside au narthex, avec six travées dans la nef et les bas-cétés. Elle comprenait un métatorion au nord-est, des catéchouména — accessibles depuis le métatorion — au-dessus du narthex et des bas-cétés et, probablement le plancher opus sectile et un ambon. Cette nouvelle construction doit avoir incorporé la moulure réutilisée

et l’inscription réutilisée V sur les murs supérieurs nord et oriental de la galerie surplombant le métatorion (pl. 2 et 3). Cela correspond au second état architectural de Sélina Ballance, qu’elle date du dixiéme ou du onziéme siecle. Il faut relever quelques points. La nouvelle église avait 4 peine le minimum de caractéristiques et de mobilier liturgiques nécessaires pour une église de couronnement. Ce sont les catéchouména pour les

acclamations qui exigérent un narthex au-dessous d’elles plutét que "inverse. Le métatorion et les galeries étaient réduits au minimum. Il n’y avait ni déme ni croisée, c’était un batiment trapu. Peut-étre n’y avait-il aucune église 4 d6me a Trébizonde au début du treizi¢me siécle. De méme nous venons de démontrer que |’abside pentagonale assez lourde est le premier exemple que 1’on puisse dater de ce qui devint ensuite une caractéristique typique de 1l’architecture des Grands

Comnénes. Le modéle de la nouvelle église fut peut-étre St-Eugéne qui est, aprés la Chrysocéphale, 1’église monastique la plus importante de la ville. Elle est dédiée au saint patron de celle-ci et fait face a la cathédrale depuis le faubourg oriental de Trébizonde. St-Eugéne était alors une basilique a voiites en berceau, avec nef centrale et deux bas-cétés, comprenant cing travées. Mais, a la différence de la Chrysocéphale, elle n’avait besoin ni des galeries ni du métatorion qui, pour (37) Panaretos, éd. Lampsides, pp. 75, 81.

228 Une église «a la demande du client» a Trébizonde

cette derniére, rend impossible que le plan original ait comporté des absidioles latérales. La quatriéme étape se situe dans la période 1341-51. Le temoignage

s’en trouve dans un hymne a |’occasion d’une nouvelle inauguration de l’église par le métropolite Acace aprés une restructuration profonde mais mal définie du batiment. Cet hymne fut composé pour la féte patronale de |’église, qui était le jour de l’Annonciation (25 mars) (38). L’auteur de l’hymne était André Libadénos, que 1|’on retrouve occasionnellement composant a Trébizonde entre 1335 et 1361. Le métropolite Acace régna de 1339 a 1351. On peut peut-étre raccourcir encore la période durant laquelle cette nouvelle fondation a di se produire. Panaréte déclare en effet que les Turcomans attaquérent la ville le 4 juillet

1341 et que «tout Trébizonde fut réduit complétement en cendres». Il ajoute que Jean III fut couronné a la Chrysocéphale le 9 septembre 1342. Oikonomidés en déduit que, si l’église fut brilée le 4 juillet 1341 et était 4 nouveau en état pour servir 4 un couronnement le 9 septembre 1342, elle a pu seulement étre remise en état pour une nouvelle inauguration le 25 mars 1342 (39). Le raisonnement est précis. Mais il présuppose qu’elle ait été reconstruite en moins de neuf mois durant une période de troubles civils. Dans un autre passage Libadénos décrit incendie de 1341 avec de nombreux détails. Or il faut noter qu’il ne dit pas explicitement que la Chrysocéphale en ait été victime (40). Panareéte ne la mentionne pas non plus (alors qu’il parle de |’incendie de St-Eugéne) et Lazaropoulos pas davantage (alors qu’il précise que le St-Eugene de son époque n’est pas le méme que celui antérieur 4a 1340). Il est dangereux de s’appuyer sur un argument a silentio mais il est aussi prudent d’examiner s’il n’existe pas d’autre date possible. Contrairement a la coutume, Alexis III fut couronné le 29 janvier 1349 non a l’ambon de la Chrysocéphale mais a St-Eugene (41). Cette derniere église ne présentait pas les facilités convenables pour un couronnement. Si l’empereur y fut couronné parce que la reconstruction de la Chrysocéphale n’était pas encore terminée, cela nous donnerait comme date de la nouvelle inauguration de cette derniére les 25 mars 1349, 1350 ou 1351, c’est-a-dire jusqu’a la fin du régne d’Acace. Mais 11 peut y avoir eu d’autres motifs au changement d’édifice: Acace avait eu partie liée avec les empereurs précédents, rivaux (38) N. Banescu, «Quelques morceaux inédits d’Andréas Libadenus», Byzantis, 2 (1913), pp. 364-5; Old. Lampsides, «3ad Ss og RS sf . TRS : SLoee PERE NSTI aoe : p RR Oy SEER 2 SS TT ga Shao ve FEN cS. S as SE So RR “ate ~ wae es FNN BES.RENEE rane ee, ACE RSROSS os ar are

- tS ISS ee sence Oa eg - a .

. :ayy gsaeae:Be&. cei.. g: APSE: aes OE RS SET S wee . BRS WSRIAT, SeRE Pe28g ae ges cE PPAR PORSCNS Pe soe .Ry ee ee BCE. ROE Reg SE” we Set 4See ~OE ES s. BEAL ES Seat Cg ES . RBS BK, aaa BRS SLE GA ee fo eh eae ny te BSS CS RS SS Lee MERE tS v YS yy os PRIS UR RRS ee gg eS . an he

‘ . 2 oe “ : * =| -i.hegey ; .etSR yLY . ond rr ne wm : Te : f4 ‘ ae wos, “ eT E _ ’ on . : : ee

QO

CO vag Boy, ; Parr on “" .SAM aee. OT BOE cdhe . UM. os. :. Hoy »EA am . to ee tFE oroe res soe aaan Sari: .ag :gt. ery Fhe ea ra . Tal Cl i.Poe ae eea: Moe le on ane *. Po Tt ar

. . . Leos cove wa :to eee SEL “Sof - me . “* ‘ : ; : re care oe an soon, :o PyRe ae se “ , An . .: on Ria See ee aoe : toms ThAR Fe ;ts. ee My.Ghee : 6 " oe Peg OBaN . vo .. Le— ne aRs we whe We wy A, foSee TP: ;ptr

PERE . fy .wos ‘ oe ar *> eR Hine oeeee TORN. fo . ak. AY BP Sy te ““let oanRa :et eeRa a. lea Merge OS SES SRen OSES EE 8 moaa . i Ps Soe De Re oyteeeRe RS eo Stowe :.. et . SNE CARES ELS os ee cand : .

. : . Sos ARES ewe “

.

Py

ae ——

al oi ; . ; : ed ; Go, ,.we ae : . we A oo .,4idfPy ‘.. *.of«.SO ~ : t (2) .oF “* Fi rss ...we % «es a : . ‘ s:anest ;7a:“ue.3“3 we: aa wegee we* Ne Pam 4 as.

#4 oo ” 2? : A i

*~ 3 “4 ’ . - . a “s . mo - . . 7aa iex3y2y*¥:7f °; a° . . ee g :54 roo _. of goose Ss : < he : :rE,, >aowaezs.¢#os; i aegs

”ie . ; ra ' i fgt awae » (SE og Be .x. leA a+wee . ‘.: Ly “A. ¥awe} * ord . Pa SPaf.ona.. Magen

Boy LA ok : .

“% P “ 2 : “ . ; os ‘,a wt ; . Aw eePe: ee :oe Bo > °Fi: . ., Laer

. Ps . whe! . one ne — ~ : we es ah:

y: ; - _ ; : me a we . _ . CALMED NS Ak. SP Se ON Rae“os, eeeKP I Se SPes.=OL ee Be ae acee

: nS 5 ae BR on SUA eo PEG a age ERR E IR MRE EE

FON e eran PINE IEEE ., “ . ° PRM: eM oe Re Sate eS a Ky BREE Lo

SR BU oe Fe

Sie

ae a 4 os a ae ee rn nN f. ., cet

be at ati< 2,

. i. +ieae bendak fae us fadae Lg

ok

\ 4s i*

ti

ie tag Hsxis pan tsa. ce EL TM aeee t*.J ee ae iesem oe RS SNE ftaa Bakes Da ok Die ePd:ah a oak nd Py

Bhi ict SL Ed . )

Ai ul i)Bath, aaa vcan | ind Mam 5s Han ue ‘ffRap whe “.pei ’oe oe st ee |— . «. ‘F] %> % he Ps in! wey Pr. :.:ee ripe, ae Oe et ae . Ra a xGah 3 Fa Ne he ota re alee ,

Beee eetategty Ay RAtoi,egAg ae. om poe Qiu Tae “ihe vent aicataraiag SS" 2” t,t ae Ey a ee ;

- wer anett tte ne my Parts

SR es ae Us Tk Gt

scat rE he a, 0 Bee a at pos ’Praad Etage om | hd+A A AN . Te yonifs . Fg Oe ee ag td rie“aoa peEE VGee ey. OE adA$.PERE

peo 2 ae a ims Re Afni ce Rec ae . AC aPye eae aot ahss fae *a 6 te ‘ 4 spe ‘ xMas ate D ae ® hpRY ea perk bi Beaks BME og oS Pome Ree i ie an site Par ' . a eT aa Be: ae Aes Ke ze fan eget a fi, my . ’ aoe a+ #., a Tat Te ae oe penal . —_ TAT TA tate oe "

ee eS + gtk an arr Pes te OS a BB . of | Se Se al Sia SRE an Pw! . Loo . 1 pale ee a aa a pogo ‘ ET a a A aot. J ney te nT rr . 1 va [ee Syn F ee 27. : ete wt ® . aa Paar ' ' ate ves + tmp Seog saa na a ES the agg ta» ree, "A ve eT “et Me tan et ae ' ' “eo, . 5 a ean Sea Le OD ess ea Fn re Mae ow‘ . |". SOs ye aay toe , Se “9 Poms . 8ated ee vs ee reony Pe 3 ”aos ::AS eepost .te CrMy oe weet root,“;oPR aaeae pee hs Hoe AS, ” etc ee ae ; egPe OR‘ ore. fe . vee SOB, PORE , ' ates a: Res Ceny Met a) ea oi)1h, a, ye eeae Lo a,.Pe GUD ok aOL ann 9: .I ate aae Ge, he? SRke ae

VI

ECLIPSES AND EPITHALAMY IN FOURTEENTH-CENTURY TREBIZOND Andreas Stratos followed the Anatolian campaigns of Heraclius with his

customary and happy combination of care and zeal. They took both emperor and historian to the Pontos, where the Byzantine army thwarted a night attack by the Persians during the lunar eclipse of 28 July 622, and where in 626 the unfortunate empress Marina gave birth to Heraclius (II), called Heraclonas?. This offering in memory of Andreas Stratos, on eclipses which darkened the Pontos, invaders from the East, and an ill-used empress from Constantinople, would therefore be familiar to

him. But it belongs not to the seventh, but to the fourteenth, century.

The fourteenth-century Byzantine world shared Western social and political upheavals which the Pontos, then represented by the Empire of the Grand Komnenoi of Trebizond, did not escape. Indeed the Trapezuntine ‘time of troubles’ lasted longer than most: from 1332 - 63, under external pressure of Tiirkmens, Constantinopolitans, Italians and, not least, the Black Death. A text which reflects the dynastic origins of the Pontic civil wars and revolutions of the period has not yet been brought into discussion

of it and is offered here. It is a poem added to the final folio, 158v, of MS Laud. Gr. 3, in the Bodleian Library, Oxford, being Four Gospels written in 1285/86 (Fig. 1). First transcribed by Kirsopp Lake, it is now published by A. Turyn, Dated Greek manuscripts of the thirteenth and fourteenth centuries in the libraries of Great Britain (Washington, D.C., 1980),

p. 46, thus: 1. A.N. Stratos, trans. M. Ogilvie-Grant, Byzantium in the Seventh Century, I. 602 - 634 (Amsterdam, 1968); N. Oikonomidés, “A chronological note on the first Persian campaign of Heraclius (622), “Byzantine and Modern Greek Studies, I (1975), 1 - 10; T. S. Brown, A. Bryer and D. W infield “Cities of Heraclius”, Byzantine and Modern Greek Studies, 4 (1978), 16-30 = A. A.M. Bryer, The Empire of Trebizond and the Pontos, Collected Studies (London, 1980), I.

348 Katvov @padn onpepov év TparnelCobvtt KAE0C

Kai KadAwniler | tac woxds pardpbver tac Spdcetc: — | ék yap éoac fiAtoc &Kk 5é SvoLaHV CEATvH: — |

Baoiieioc 6 BaotAebds 6 Kopvynvoc 6 péyac: — |

5 Kai Baotric f Aaunpavyic Eipnvn kai Koonia: — | ovverodpapovtes Ev tavt® matiCovor Ev | KOKAQ: tod KOOLOD TO TETPALEPES ApLlotatc Sadovyiatc:

Oo && avtvyov ov(pa)vod tod | patvopévonv digpov: Koinov ayGAAOv TpaneCove | kai yopeBe kai oxnpta:

10 Kai d0Eacpotc avdlnevte: odv Kpotacs enna’ The emperor who is “the sun from the Orient” in line 3 can only be the Grand Komnenos Basil of Trebizond (1332 - 40).But both his wives were called Eirene. This lustrous empress Eirene can only be his first wife, natural daughter of Andronikos III Palaiologos (1328-41), for she came from Constantinople as “the moon from the West”, in line 3, while the second Eirene was a native despoina of Trebizond. Eirene Palaiologina arrived in Trebizond on 12 September 1335 and presumably married Basil when he received ritual acclamations as emperor on 16 September, for, or soon after, which the poem would have been composed. The author mentions such acclamations in the last line. The date of composition can probably be narrowed down to months rather than years after 16 September 1335, for epithalamies on the marriage of Basil and Eirene Palaiologina would soon have become inappropriate. How soon may be judged by the fact that although Basil delayed marrying

Eirene ‘of Trebizond’ until 8 July 1338, she had already borne him two children by 5 October 1337. She would have been his mistress by the early weeks of 1336 at the latest. According to Gregoras, Eirene Palaiologina lost no time in arousing high and low and making cries of protest to heaven and earth, but gives no date for them. Confidence in Basil’s régime would have been further shaken when on 5 July 1336 the GCobanli Seyh Hasan-i Kiiciik’s (1327 - 41) attack on Trebizond was only washed out by one of the summer rainfalls for which the Pontos is notorious — a surviving Trapezuntine horoscope for 1336 reveals that showers had in

fact been predicted for July, but that forecast is surely more a tribute to the astronomer-astrologer’s local experience than to his prescience. The convenient identity of names of the rival Eirenes gave the clergy a chance

ECLIPSES AND EPITHALAMY IN FOURTEENTH-CENTURY TREBIZOND 349

nm

- =PGa Pe. als NEI a ios. NIE Cae. se ee . a Py, a , , ao” op» ‘Naa 6B law. UU OU a a MeO,’ "Pa . 3 ~~ ao he eg RS pee me

. . si wy1aes He - .4*a,A aaa: Pins : . “ a2Pe bE. . . Nn

“ a, : ; 4 RO ‘ - “a on .

; ae _ ae noi %y _: . a _ 4 a , a oe 4° Pe «! aan 4. ; -

. 3 é * x a 4 4 : P » . te i j é a ts ede ry “E. : fs, pee

4, pate “ , #-r ! on ‘ /~ - boo , as : “aoe a wePt, . . a. .| ae * 4ved. . ] seg. r :*oo vo *y_ |& ne: © & $fj .,ia

ae) ee Tey a es. a

Se AT OM TA ee ee Oe ” a ,, ‘ i : ' , oe : ae a” :, | : “ q a ze i . — . :. ‘ay pth “4s 3 on thee70 - ye as eae? ad: “. .. ,Sy a %* +>. 4,, 2. -a *Sa , + atee * a i: cy

ao”? oak: ne fig app ters pte eevg| a ae Fe et gat TN Bhar a. Coes £ o Ra — > of Vas x. oa eee cae + ©: : “ me rae if ee . SS tee ~~» we. ‘ iears a the: ers Kw ie all oe OE OP ot an ro . ~ as ry? Mm. eee oh Rg Ree eo & lg 110: See 7 Ke ak **.% . hs > ” Bi " a re’ ’ ¥ ta 34 a \ a: ; , i *

. a Sa 4 ¢ oat =e ‘ + = rc? wv *. i i ‘ ,” wa # tf | } r > -

Pn ae a eer. Wat Ree see pee 2B te ee fe (tee eee ae 397h *BW cae. wy . ‘ ; oe . . : . a wae ge Cue 2 oe 1, 8 me OS r +g ee* “Fe «eebsaU wilwa ‘ b= ‘i ,a ~we “re>:%ard . i °° rho eo "We ; ey_ 5 dk =~" a a. . o “ Le , . ; . 4 -7” . ieh - ,‘: e.bd * ™, ; ;aee> ae Ba, * «*..Ps Rep ea 2ate 2** ae ‘og A

co. , So a

on y * fd & . ae _ "3 te a “a ‘ rf ; ‘Row ‘ mere wet " ie € rs ” * de

i*Ee aa: .2‘. .:w.a..at . ;&*4 ;iryaf 4 .i? ¢ -® .4nA rme *. * és oy * > re .;oe . a > ow w ‘nid+_ay a ww _@ ' Bs ... :. . a ”— « s 29 . 7 a. hs if: ‘ 'i q,

. - meg e'*re, ue eya* 2 - - Ng Feil v4 mye eet ; .|oe7t*ey 2f| "0 er oe _~ co 2. | N ma vo } i ‘a a ‘ : =

“Me % i*deo a . Ty=:3 a- =ae:. ra.i.’.s:we f& j“ SS ,peers : ba * .. “~s get anw ae re -m 7 . ‘iJ .as .“oe eee ed toy

“PO, ST ae ae Oe Se ie ies an me Oe ee he

| “mee om Mise es re t - ee es bes wepe Ses ‘ . Pi. ‘s ; 4 2 Pe: ae aesh=ies eed 7 = i‘yo_ i; aJ =a ~. =- +Pp . ’: ft: " i "" o7stes. MA

Fig. 1. Oxford Bodleian MS Laud. Gr. 3, f. 158b.

350

to hedge their bets: no one knew which Eirene they prayed for with Basil on public feast days. Metropolitan Gregory of Trebizond (1333 - 39) seems to have been equally evasive about the affair when he saw Patriarch John XIV Kalekas (1334 - 47) in Constantinople, but held off solemnising Basil’s second marriage (and so agreeing to dissolve the first wedding which

the Oxford poem celebrates) until 8 July 1339. The event brought down spirited patriarchal letters of condemnation on both Basil and Gregory’. Caught between his local emperor and his patriarch, no metropolitan of Trebizond could win — particularly if a daughter of his patriarch’s own

emperor was the injured party. |

In Trebizond things had come to a head in a peculiarly awesome way between the fourth and seventh hours of Monday 3 March 1337 when a full eclipse of the sun followed a partial eclipse of the moon only a fortnight

before. The eclipse demonstrated two facts. First, although the famous astrologer-astronomers of Trebizond may have been capable of forecasting summer rains, local students of Ptolemy’s Almagest (like Chioniades, whose

very name indicates his reputation as a pundit) were evidently unable to allay fears of eclipses: indeed that of 5 May 1361 when the stars shone in the noonday sky for an hour and a half, was to catch the imperial family itself by surprise and full of prayers and supplications. Second, and more important, the solar eclipse of 3 March 1337 revealed that Basil’s adultery

was already both widely known and disliked, for during the terrifying midday darkness, his “people rose against the emperor, so that they gathered together outside the citadel and hurled stones at him”’. So 3 March 2. Michael Panaretos, ed. O. Lampsides, ITegi tév Meyddswv Kopurnydy (Athens, 1958), pp. 62, 64 - 65, 73; F. Miklosich and J. Miiller, Acta et diplomata graeca medii aevi, I (Vienna, 1860) pp. 199 201 and J. Darrouzés, Les regestes des actes du patriarchat de Constantinople, { (v) (Paris, 1977), nos. 2193 - 94, pp. 148-50 (for Kalekas’s letters); Nikephoros Gregoras, ed. L. Schopen, Byzantina Historia, I (Bonn, 1829), pp. 536, 548-51 (who mentions both eclipses of 1337). Panaretos calls Eirene ‘of Trebizond’ “‘despoina”, the Patriarch “motchalida” and Gregoras “hetairida.” Eirene Palaiologina must not be confused with a third Eirene, her half-sister and namesake, who married Michael Asen: D. M. Nicol, The Byzantine Family of Kantakouzenos (Cantacuzenus ) ca. 1100 - 1460 (Washington,

D.C., 1968), pp. 229 - 30. S.P. Lambros, “Tpaxelovvriaxdv dpooKdéniov tod Etovs 1336”, Néoc ‘Elanvouripeov, 13 (1916), 44. 3. Panaretos, ed. Lampsides, p. 65. The Ptolemaic astronomical school of Trebizond has attracted a large bibliography which has been sometimes uncritical, but see most recently: D. Pingree, “Gregory Chioniades and Palaiologan Astronomy”, Dumbarton Oaks Papers, 18 (1965), pp. 130 - 60, and (on Chioniades’s name = akhond or, ‘sage’) Anna Philippidis - Braak, “La captivité de Palamas

chez les Turcs: dossier et commentaire, “Travaux et Mémoires, 7 (1979), p. 218; O. Lampsides, “Georges Chrysoccis, le médecin, et son oeuvre”, Byzantinische Zeitschrift, 38 (1938), pp. 312 - 33;

ECLIPSES AND EPITHALAMY IN FOURTEENTH-CENTURY TREBIZOND 351

1337 is our most certain terminus for the date of the poem, for after it had been blotted out on that day, no eulogist could liken Basil to “the sun from the Orient.” Who is the author of the poem? The script in Fig. 1 fits the period around 1335. An obvious candidate is Andreas Libadenos, later chartophylax and protonotary of Trebizond, who in, or very soon after, 1335 followed his friend the Grand Komnenos Basil there from Constantinople.

He was caught up in what he so graphically described as the “pwooc EuMvAOV” which Basil’s marriages engendered and is last heard of in 1361,

still as public orator, at the court of Basil’s second son by Eirene ‘of Trebizond’, the Grand Komnenos Alexios III (1349 - 90). The poem is not in Libadenos’s hand: his autograph MS Monac. Gr. 525 differs from the note in MS Bodl. Laud. Gr. 3 in at least sixteen regular letter forms. But this hardly disproves Libadenos’s authorship: indeed it would be surprising

if the note were autographed, for, as Turyn points out, slips in the last line of the poem suggest that it was not copied by its author, and two phonetic peculiarities hint that the copyist was a Pontic Greek, which Libadenos, fresh from Constantinople in 1335, was not. Although the rather H. Hunger, Die Hochsprachliche Profane Literatur der Byzantiner, II (Munich, 1978), pp. 251 - 52; and A. Bryer, “The estates of the Empire of Trebizond”, ’Agpyeiov Idvrtov, 35 (1978), p. 329, n1 = The Empire of Trebizond and the Pontos (London, 1980), Collected Studies, VIII. Although the Trapezuntine horoscope of 12 March 1336 was drawn up “according to the Egyptians” — NE, 13 (1916), p. 33 —I am grateful to Dr. T.S. Pattie of the British Library for confirming that there is nothing in it which demands the use of a Ptolemy’s Almagest, known to Chrysokokkes who is a candidate for the horoscope’s authorship. It is even more difficult to know how far Byzantines were able to predict eclipses. The classic point of departure is a solar eclipse in the 1080s which Anna Komnene claims was predicted by a mere hypogrammateus of Alexios I Komnenos, foreknowledge of which the emperor used to psychological advantage over the ignorant Patzinaks: Alextade, B. Leib, II (Paris, 1967), pp. 92 - 93 (=VII, ii, 8). F. Chalandon identified the eclipse with that of 1 August 1087: Les Comnéne I (Paris, 1900), p. 105 n. 1: cf. Georgina Buckler, Anna Comnena (Oxford, 1929), pp. 211, 435. But the 1087 eclipse was barely visible in Constantinople, nor do any others of the period fit the chronology of the Patzinak wars. R. R. Newton tends to reject the tale as a “literary eclipse”: Medieval chronicles and the rotation of the earth (Baltimore and London, 1972),

pp. 550 - 55. The eclipses of 1337 and 1361 actually passed over Trebizond and I am indebted to Miss Mary Cunningham for checking them in T. von Oppolzer, Canon der Finsternisse (Vienna, 1887) pp. 244, 246 and Charts 122 - 23. But even if Trapezuntine astronomer-astrologers had been able to predict them, forewarning does not necessarily allay panic or the ritual seeking of a scape goat, who was Basil himself on 3 March 1337. On 12 August 1971 I witnessed a lunar eclipse in Pontic Oinaion (Unye). Open-air cinemas broke up in disorder, citizens poured into the meidan where a sheep was sacrificed and the ramadan cannon fired, while those with rifles and small arms shot them at the lost moon from their roofs. Mass hysteria continued until the moon re-emerged fully, although the eclipse had been widely predicted and explained on radio and in the press for some days before.

332

commonplace style of the poem is in line with that of Libadenos’s other writings, I can find no specific link between them. So the matter must remain open. What these events opened was dynastic and social strife for more than a generation in what was perhaps the most prosperous and certainly the most densely populated part of the late Byzantine world, the Pontos.

4. Turyn, Dated Manuscripts, p. 46 and notes 108a and 108b; O. Lampsides, ’Avdpéov ArBadnvov Blog xal “Eoya (Athens, 1975), passim, especially pp. 288, 273 - 74, 28. I am most grateful to Dr, Odysseus Lampsides for sending me the originals of his photographs of Cod. Monac. Gr. 525 on pp. 25 - 37 of his work, for comparison with MS Bodl. Laud. Gr. 3 in Fig. 1 here.

VII

“The Faithless Kabazitai and Scholario1” I. The obit for 26 April 1429 Robert Browning does not insist that his pupils share his passion for Byzantine manuscripts and those who wrote them. But they end

up with it. It is done casually: a muttered aside, or postcard

reminder, that there might be something of interest in, say, a midfourteenth-century synaxarion from Trebizond, now convenient to

Balliol. Such means brought me to Bodleian MS Gr. Mt. d. 6, formerly Peristerota MS 12. In 1976 I published 37 obits and

other marginal notes from this synaxarion, in a relay of 23 hands, running from 1368-1563.! 1 proposed that it came from the cave monastery of St Sabbas, above Trebizond, where in twists of the civil wars the Grand Kommenos Michael (1341, 1344-49) banished his depraved son the Grand Komnenos John III (1342-44) in 1344-45 and was in turn banished in 1349-50 by the Grand Komnenos Alexios III (1349-90), whose own death is memorialised in the manuscript.

I overlooked a 38th entry and 24th hand in the synaxarion, and am most grateful to Dr Irmgard Hutter for drawing my attention to it.2 It is no consolation that others who have examined the manu-

script since 1898 did not notice this scribble either, for it is

the most interesting of the series. It is an obit on fol. 62', supra, referring to 26 April (pl. 1). The first line is heavily

cropped and the name commemorated is lost entirely. The second

line is erased, but can be read with difficulty under ultraviolet light. It is in two hands. The first line and second half of the third line are in what I had already identified as Hand N, who wrote an obit for 22 March 1430; in the relay of annotators, he succeeded Hand M after 1418 and was succeeded by Hand O before

1434. This fits in with an obit date here of Tuesday A.M. 6937 = A.D. 1429, referring to 26 April (St Basil of Amaseia). The second

(erased) line and first half of the third line are in an otherwise

unidentified hand, say Hand N2. The obit reads:

1. Ae Bryer, "Some Trapezuntine monastic obits (1368-1563),” REB, 34 (1976), 125-38 = tdem, The Bnpire of Trebizond and the Pontos (London, 1980), Collected Studies, IX.

2. It will appear, with her discussion, in Irmgard Afutter, Corpus der Byzantinitsachen Mintaturhandschriften, 3, Oxford Bodleian

Library, III (Stuttgart, forthcoming).

[Hand €] + [™ avr meeps exocimene «| TW

[Hand N2, erased] nd tdv dntotov tiv dpy6vmyv adtod Kapaxt {eat xat (dv) Zyovdaprf&twv:- | [Hand N2] (xat) pvno®(n)t(c) adt(dv) K(bpr)ye ev tH Baotheta OOv; ~

{Hand N] Exous SANE.” tata) Y:-| Line 2, by Hand N2, presumes a first one, either beneath what

can now be seen of Hand N's first line (where there is no sign of over-writing), or above it in the top of the margin, which has now been cropped for binding. This first line should have followed the formula of all the other obits by recording the name of someone who

had died, or perhaps suffered, under “his archontes the faithless Kabazitai and Scholarioi” (line 2) and was to be memorialised (by inference) on 26 April. The Kabazitai present no problem. This great archontic family was (with the Scholarioi) prominent in the civil wars of 1340-63. They dominated major offices of state in the capital and were dukes of Chaldia, with (according to the interpolator of Chalkokondyles) their home-castle at Mesochaldia (Kegi) Kale, or perhaps Golacha-Colagana). In 1461 some Kabazitai followed their Grand Komnenos to exile, some held out in their old

fief in Chaldia (Torul) until 1479, and others were still holding coastal lands in the Ottoman defter of 1486.° The Scholarioi present a problem, however, which must be disposed of.

Il. Who were the Scholarioi?

It either says much for the persuasiveness of Fallmerayer, who in

1827 first analysed the causes of internal strife in Trebizond, or for the gullibility of all subsequent historians of the subject who have unquestioningly followed his analysis, sometimes word for

word, that discussion of Pontic social and political troubles mist still start with Fallmerayer's explanation of who the Scholarioi 3. Michael Panaretos, Pert ton Megalon Komnenon, ed. 0. Lampsides

(Athens, 1958), 65-73; L. Chalcocondil, Excpunert tstorice, ed. V. Grecu (Bucharest, 1958), 265-66; tdem, Histortaruwn demonstrattones, ed. Darkd (Budapest, 1922), II, 220, 256; tdem, Histortarun demonstrattones, ed. I. Bekker (Bonn, 1843), 463, 494, 496; and N. Beldiceanu, “Les Qavazid/Kabazités & la lumiére d'un registre ottoman de Trébizonde,” Studia Turcologica memortae Alexit Bombact

dicata (Naples, 1982), 41-54. For Mesochaldia, see A. Bryer and

D. Winfield, The Byzantine Monuments and Topography of the Pontos (Washington, D.C., forthcoming), section XXII, no. 20. 310

“THE FAITHLESS KABAZITAI AND SCHOLARIOIL”

were. He wrote: “Die Faktion der Einheimischen k6nnen wir nach einem ihrer vornehmsten Sitze die Mesochaldier, oder nach ihrer Grundlage die trapezuntisch-kolchische Partei nennen, wahrend die der Eingewanderten den Namen Scholarter trug, und im Gegensatze zur

vorigen die trapezuntisch-byzantinische heissen mag. Diese Scholarier waren urspringlich Palastsoldaten der Kaiser von Kon-

stantinopel. Ihre Entstehung fallt in das vierte Jahrhundert zuruck, nach Procopius, der in den Geheimen Denkwiirdigkeiten hier-

uber folgende Nachricht gibt:...” going on to quote Procopius, Anecdota, xxiv, 15-23 on the Scholarioi palace guards under Zeno,

Justin I and Justinian. A century later Miller (hardly a social historian) was content to summarise Fallmerayer: “The principal factions were known as the Mesochaldaioi and the Scholarioi, the former representing the original Pontic aristocracy, the ‘back

woodsmen" of the Trapezuntine territory, the latter being the descendants of the palace guards of Byzantium. The strength of the

former lay in the provinces, that of the latter in the capital.” A virtue of Charanis's study of “Internal strife in Byzantium during the fourteenth century” of 1941 is that it attempted to

place the Trapezuntine troubles in their contemporary Byzantine context, but he did not advance Fallmerayer's analysis.° Finally in 1969 Janssens reverted to the sixth century and before, and

added embroidery to Fallmerayer which Miller, at least, would have enjoyed: “Ces transfuges de Byzance de plus ou moins haut parage sont désignés sous le nom de Scholartot et pouraient étre comparés dans ume certaine mesure a ces nobles normands qui accompagnaient

Guillaume le Conquérant et se heurtérent plus d'une fois a l'ariStocratie saxonne au cours de l'histoire médiévale de l'Angleterre.

Cette comparaison appelle naturellement des réserves. En effect, on

ne saurait mettre sur le méme plan les rudes compagnons de Guillaume et leur attitude en pays conquis avec l‘entourage byzan-

tin des premiers Comnénes. Le nom méme de Scholarioi mérite quelque

commentaire... going on to quote Procopius, Anecdota at length again. Janssens concluded: “On peut imaginer que leur présence dut

assez t&t indisposer l'aristocratie locale, et que celle-ci oppose a la faction des Scholarioi un parti des Mesochaldaioi représentant ses intéréts."/ First: the “Mesochaldaioi” are not mentioned in any Trapezuntine source. Fallmerayer presumably coined the term from mentions

of Mesochaldia as a toponym in Panaretos and the interpolator of

4. J.P. Fallmerayer, Geschichte des Katsertunas Trapesunt (Munich, 1827; rp. Hildesheim, 1964), 169.

5. W. Miller, Trebizond (London, 1926; rp. Chicago, 1969), 44.

6. P. Charanis, “Internal strife in Byzantium during the four teenth century,” Byzantton, 15 (1941) = idem, Soctal, Economic

and Poltttcal Life tn the Bysanttne Bnpire (London, 1973),

Collected Studies, VI, 218. ,

7. #£E. Janssens, Trébizonde en Colchide (Brussels, 1969), 100. 311

Chalkokondyles.®

Second: the “Scholarioi”™, even in Procopius's time, were ceasing to be palace guards - as he noted. By the tenth century the term seems to have indicated, rather loosely, a tagmatic soldier or officer and is not found thereafter.” Yet they seem to echo in Lazaropoulos's analysis of the parties which emerged in the 1330's: of xat ddd\(rov Sug ppedevtec - of péev *Aucvtyvavtapdvtae bxéyovto, of 6€ LyoAapavtar. A later and probably Trapezuntine

order of precedence explains that the protospathartos was locally called the anyrtzantartos or emir candar, an officer first attested by Panaretos under its Turkish name in 1345.11 Members of the Meizomates, Doranites, Mavrokostas and Sampson families held the office: despite the Prosopographtseches Lextkon der Palatologenzeit,

the Amyrtzantarioi were not a family.! The Scholarioi were not Scholarantai, but by the fourteenth century were, as Gregoras

specially states, no more or less than a family, which is not attested in Trebizond until 1340 but survived the fall of the empire in 1461. As the PLP will show, most members of the

LxOAGptoc, LyoAaGpnc and Lyokaprc family were Trapezuntine, but the

name is also found in the Balkans in the thirteenth and fourteenth centuries and was borne most notably by patriarch Gennadios II (George) Scholarios of Constantinople (1454-56, 1462-63, 146465).

Whatever the origins of their name, and late Byzantine

occupation names quickly became family ones, the Scholarioi of Trebizond were not palace guards, still less the hereditary palace 8. Panaretos, ed. Lampsides, 74; Chalkokondyles, Bonn ed., 494, 496.

9. Héléne Glykatzi-Ahrweiler, Recherches sur 1l’admintistration de L’Enpire bysantin aux IX-XI® siécles (Paris, 1960), 5 note 7, 29 note 6, 30; P. Lemerle, The Agrartan History of Byzantium from the ortgins to the twelfth century (Galway, 1979), 98, 116 note 1, 118

note l. 10. Fontes Trapeasuntint, I, ed. A. Papadop(o)ulos-Kerameus (St

Petersburg, 1897; rp. Amsterdam, 1965), 134.

ll. Pseudo-Kodinos, Tratté des Offices, ed. J. Verpeaux (Paris,

1966), 341-42, 438; Panaretos, ed. Lampsides, 68.

12. Panaretos, ed. Lampsides, 68; E. L(obel), "A chrysobull of Alexios III Grand Komnenos," The Bodletan Quarterly, 3 (1921), 140-43; V. Laurent, “Deux chrysobulles inédits des empereurs de Trébizonde Alexis IV/Jean IV et David II", Archeton Pontou, 18 (1953), 26, 267, 269; N.A. Bees, note in BZ, 17 (1908), 487; Chalkokondyles, Bonn ed., 465; PLP, ed. E. Trapp et al. (Vienna, 1976), I, 80 no. 830. 13. Nikephoros Gregoras, Byzantinae histortae litbrit xxxviit, edd. L. Schopen and I. Bekker (Bonn, 1830-45), II, 681: genous d’esan ton Scholarton; Panaretos, ed. Lampsides, 66. I am most grateful to Dr H.-V. Beyer of the PLP for making its unpublished entries on the family available to me. 312

“THE FAITHLESS KABIZITAI AND SCHOLARIOI"” | guards of Byzantium stretching back to the “"Scholarioi” of Justinian and even Constantine the Great. Like the “Mesochaldaioi”, the “Scholarioi”, as guards of Trebizond, were not conceived until

1827, when Fallmerayer named then. The Scholarioi, like the Kabazitai, were no more than archontic dynasties, in the Pontic mould of great ruling families.

That said, Fallmerayer's equation has its uses. While both Kabazitai and Scholarioi held coastal lands and offices at court, only the Kabazitai had a Mesochaldian power base in the castles and highlands of Inner Chaldia. Further, both Kabazitai and Scholarioi were acting as imperial guards, or “Scholarioi”, on 26 April 1429, when Hand N2 of the Bodleian synaxarion ventured that they had been faithless, before Hand N thought better of the opinion. Whom had they betrayed? Both Kabazitai and Scholarioi were zealous patrons of churches and defenders of the faith against Trebizond's Hagarene neighbours.

It mist be taken that they had been faithless to the person or

persons memorialised rather than to their God.

Trapezuntine archontic families held faith with their retinues as well as with their Grand Komnenos: had they been faithless to their own followers? Fallmerayer pointed out the basic difference between the Kabazitai and Scholarioi: they would have been conflicting masters to share. They were opposed in 1345 and 1355, but found on the same side in 1340, 1344 and 1426. We do not know if they were on the same side in the crises of 1332-40, 1341-42, 1344, 1350-52, 1363 and 1395-1404, but the rule seems to have been that

all archontic families sank their differences when there was a chance to dominate the Grand Komnenos (and that the people of

Trebizond itself sank differences in opposing them). Such a view may lead to a trap as deep as that set by Fallmerayer: that members of late Trapezuntine (and for that matter Caucasian or Constantino-

politan) dynasties acted as coherent social and political forces.

Panaretos, Lazaropoulos and Hand N2 suggested that they did. The internal history of the Grand Komnenoi themselves, and Ottoman evidence of how members of the Kabazites, Scholarios or Amiroutzes

families went their separate ways after 1461, suggests that they

did not when under stress.4 At all events, it is hardly likely that it was to their common followers that both Kabazitai and Scholarioi were faithless in 1429. So it must be taken that they

betrayed a common lord. He can only be a Grand Komnenos whose name

has been cropped in line 1 of pl. 1. I propose that he is Alexios IV.

14. N. Beldiceanu and Iréne Beldiceanu-Steinherr, “Biens des Amiroutzés d'‘aprés un registre ottoman de 1487", TM, 8 (1981), 63-78; A. Bryer, “Rural society in Matzouka,” in Continuity and Change in Late Byzantine and Karly Ottoman Society, edd. A. Bryer and H. Lowry (Birmingham, forthcoming). 313

III. The assassination of the Grand Komnenos Alexios IV

Pero Tafur visited Trebizond in 1438. He picked up the outline of the story.!> “The emperor there is a Christian and a Greek, and

they say that the father of the present emperor, in order to

disinherit his elder brother, approached the Grand Turk, asking him to support him, and killed his father, and he had two sons and the younger son killed his father, whereby the words of the Evangelist

were fulfilled: ‘For with the same measure that ye mete withal it shall be measured to you again.'/6 The elder brother was he whom I had seen in Constantinople, living in exile with his sister, the

Greek empress, and they say that his relations with her are dishonest. “!/ He also found that the reigning emperor, John IV (1429-1458/60), only knew by rumour that his brother Alexander had

married Maria, daughter of Dorino I Gattilusio of Mitylene, and that he himself had married a Turkish bride in the interests of piecemeal proselytization./® Although a little garbled, Tafur's two-stage Oedipal tale can be checked in outline, but Panaretos's court chronicle of Trebizond closes before these events and our remaining sources, Western and Byzantine, are of varying degrees of

reliability. What is clear, however, is that John IV had not been loyal to his father Alexios IV (1417-29), nor had Alexios's sons and wife been faithful to him any more than he had been faithful to

his father Manuel III (1390-1417). Members of the three generations of Grand Komnenoi were variously accused of adultery,

apostasy, incest, parricide, piracy!? and rebellion (the order is merely alphabetical). These charges are probably justified. Alexios IV was son of the Grand Komnenos Manuel III by Eudokia, daughter of David VIII of Georgia (1346-60). Clavijo

15. Pero Tafur, Travels and Adventures 1435-1439, ed. and trans. M. Letts (London, 1926), 116, 130, 138, 150. Cf. A.A.

Vasiliev, “Pero Tafur, a Spanish traveller of the fifteenth century, and his visit to Constantinople, Trebizond and Italy,” Byzantion, 7 (1932), 75-122; and tdem, “A note on Pero Tafur", Byzantton, 10 (1935), 65-66.

16. Mark, IV, 24. 17. Maria Komnene, sister of John IV, Alexander and David Grand

Komnenoi, wife of John VIII Palaiologos (1425-48); she may be represented by Pisanello: A. Bryer, “Pisanello and the princess of Trebizond,” Apollo, 76 (1962), 601-3; she died in 1440.

18. Tafur, ed. Letts, 131-32: “I said: ‘My Lord, they say, rather, that you gave her to you so that she could turn you into a Moor, by

reason of your expectations from her and the little that you

have.*”

19. N. lorga, Motes et extraite pour servir & l’histotre dee

Croisades au XV@ siecle, I (Paris, 1899), 272-75, 304-5; F. Thiriet, Régestes des délibérations du Sénat de Véntse concernant la Romante, II (Paris, 1958), no. 1781. 314

“THE FAITHLESS KABAZITAI AND SCHOLARIOIL"”

reckoned him as being about 25 years old in 1404 (i.e. born about 1381) .29 However Panaretos records not his birth, but that of his brother Basil in 1382, who has been presumed to have been the elder brother. 7! That there may be something in Tafur's story is

that while Basil is not heard of again, Alexios IV was already described as co-emperor in 1395. But although the Grand Turk, Bayezid I (1389-1402), campaigned perilously close to the Pontos, there is no evidence of any direct Ottoman-Trapezuntine contact;

aged about thirteen in about 1395, Alexios IV would have been precocious in making the contact himself. Eudokia, his mother, died on 2 May 1395. Manuel acted fast by sending his sister to Constantinople to seek not only a new bride for himself, but a wife for his young son. The two brides, Anna Philanthropene for Manuel and Theodora Kantakouzene for Alexios IV, landed at St Phokas, Kordyle (now Akgakale, about 25 km. west of Trebizond) on 4 September 1395. Next day Admiral the Grand Duke Scholari(o)s escorted them to the capital. It poured with rain.?2 Alexios IV would have been co-emperor for over 22 years before he finally succeeded Manuel III in 1417; in turn John IV would have been co-emperor for a dozen years before Alexios IV was assassina-— ted in 1429. Late Byzantine (and Georgian) collegiate sovereignty,

far from satisfying hankerings for power, caused further difficulties in Trebizond when crown princes were manipulated by archontes. Twenty-two years is a long time to be in waiting. Tafur states outright that Alexios killed Manuel; Finlay ventured the unsubstantiated opinion that “Alexios was suspected of hastening

his father's death.”22 There may be something in this too, for Manuel seems to have faded away. The Continuator of Panaretos affirms that he died on 5 March 1412 after reigning 27 years (i.e. until 1417).24 An inscription on an Armenian altar refers to both Kyr Manawli and Kyr Alexi as tekfure in 863 = 1413/14, which is the

last known reference to Manuel. 2? Venetian sources refer to a

20. Ruy Gonzales de Clavijo, Mmnbajada a Tamorlén, ed. F.L. Estrada (Madrid, 1943), 75; idem, Bnbassy to Tamerlane 1403-1406

(London, 1928), trans. G. le Strange, 111; cf. F.L. Estrada, “Viajeros espanoles en Asia: la embajada de Enrique III a Tamorl4n

(1403-1406)," Revtsta de la Universidad Complutense, 3 (1981),

227-46.

21. Panaretos, ed. Lampsides, 80. 22. Panaretos, ed. Lampsides, 81; D.M. Nicol, The Byzantine Family of Kantakouzenos (Cantacuzenus) ca 1100-1460 (Washington, D.C., 1968), 168-70. 23. G. Finlay, A History of Greece, IV (Oxford, 1877), 392, 398.

24. Panaretos, ed. Lampsides, 81. 25. J. Fallmerayer, “Original-Fragmente, Chroniken, Inschriften und anderes Material zur Geschichte des Kaiserthums Trapezunt,”

AbhBay, Phil.-hist.Kl., 4 (1844), 106-7. This is apparently the inscription in the Armenian church of the Mother of God (Asduadzad-

315

single emperor of Trebizond in February and March 1416, and Alexios

IV issued a prostagm in favour of Dionysiou without reference to Manuel in September 1416.7© ig it was not until Manuel died on 5 March 1417 that Alexios IV became senior, or sole, emperor, he was not one of the five new emperors who had been proclaimed in the East of which the Venetian Senate was informed in October 1417.27

Clavijo's account of Alexios IV's relations with his father

Manuel III, and the interpolator of Chalkokondyles's story of John IV's relations with Ate father Alexios IV, have a certain sameness, which Tafur observed. The key figure in both is a protovestiartos, after the panhypersebastos the second non-imperial dignitary in the Trapezuntine taxie.28 The office was customarily held by members

of such archontic families as the Doranites, lLoukites and

Kabazites: George Amiroutzes himself was last protovestiarios.*? Clavijo's account is that at some time before 1404 (and presumably after 1395) Manuel had appointed, and fallen under the spell of, a

protovestitartos who did not come from the archontic circle. He Looked noble enough, but was said to be the son of a baker. Young Alexios and the archontes found common cause against Manuel and his confidant, whom the emperor refused to give up.

zin), given with the omission of reference to Manuel and Alexios by Menas Vadapet Bzhshkean, dAtstory of the Pontos whtch ts on the Black Sea (in Armenian, which I am grateful to Dr Sebastian Brock

for translating) (Venice, 1819), 79; cf. P. Minas Bijigkyan,

Karadeniz ktytlarit Tarth Cografyast, tr. H.D. Andreasyan (Istanbul,

1969), 54. Fallmerayer, and Miller, Trebtzond, 79, mistakenly computed the year 863 of the Armenian Great Era to 1417: it is 8 December 1413 - 7 December 1414.

26. GM. Thomas and R. Predelli, Dtplomatartun Veneto-Levanttinun,

II (Venice, 1899), 169; Thiriet, Régestes, II, no. 1602; N.

Oikonomidés, Actes de Dionystou (Paris, 1968), I, 10-15, 50-61, 97-101, 155-57.

27. Iorga, Notes et Extraite, I, 264.

28. Pseudo-Kodinos, ed. Verpeaux, 344, 347. 29. Papadopoulos~Kerameus, Fontes, I, 1; Panaretos, ed. Lampsides, 67, 69, 70, 71, 78; Chalkokondyles, Bonn ed., 494, 498; ed. Darkd,

II, 219-220; Beldiceanu, JM, 8 (1981), 63-78; Iorga, Notes et Extraits, III, 246 (Jeronimo de Nigro, “megavistiarius” who is probably the Genoese protovesttartos who saved Geoffroi de Thoisy from Mingrelian hands in 1445; see Wavrin in Anctennes Crontcques

d’Engleterre, ed. Mlle Dupont (Paris, 1863), I, 95- 97; III, 151-59). The identity of the protovesttaritos in 1404 and 1426 respectively is unknown and candidates are dubious. For the

former, S. Ioannides, Atstoria kat statistitke Trapezountos (Constantinople, 1870), 138, has a protovestiarios John Marouzos,

who married Anna, daughter of Alexios III and founder of St Philip's, Trebizond; cf. Un Phanariote (i.e. E. Rizo-Rangabes), Livre d’or de la nobleese Phanartote (Athens, 1892), 87. Although

316

“THE FAITHLESS KABAZITAI AND SCHOLARIOI”

“The son making war on his father, defeated him, besieging him during three months in the capital, for all the great nobles of the state ('los mayores omnes del ynperio') were his partisans. "2 The latter presumably included both Kabazitai and Scholarioi and the scene repeated that of the 1340's, when an archontic quarter around St Eugenios faced the imperial citadel in siege across the eastern

ravine of Trebizond. This time, according to Clavijo, the akolouthos, or chourtzes (variously dHortchi, Vrehi and Koros),>! 52nd in the Trapezuntine taxis, patched things up between all parties: “In what followed the emperor his father suffered much

in honour, and in reputation, but he insisted on keeping about his person that base~born officer the” protovesttartos. 2 Like Tafur, Clavijo picked up local gossip, but in this case it must have been more recent. When did the co-emperor Alexios revolt against his father? In 1395 there had been two emperors; in 1404 Clavijo is at pains to explain why both Manuel and Alexios were emperors. But only one emperor of Trebizond is mentioned in Timur's letter to the Kegent John VII Palaiologos of 15 May 1402. The letter. greatly exaggerates the military and naval potential of the Pontos. But embassies which led to it had passed through Trebizond on their way to and from Palaiologos and Timur on 29 August and 10 September 1401, where they may be expected to have

noticed how many emperors, if not soldiers and ships, the place had. 2? Was it then that Alexios besieged Manuel and could claim to

the Mourouzai had a Pontic background, perhaps in Chaldian Mourouzanton (Muruzli), I would be inclined to dismiss John Mourouzos and Anna Komnene, who are otherwise unattested, as later

Phanariot fantasy, were it not for the fact that a “Moruz” figures

in the 1486 defter. At all events, the son-in-law of Alexios

III was unlikely to have also been son of a baker. For the protovesttartos of 1426, I am equally dubious of George, proposed by

A.K. Hypsilantes, Za meta ten halostn (1453-1789) (Constantinople, 1870), 10 - a mine of misinformation.

30. Clavijo, ed. Estrada, 76; trans. Le Strange, 112.

31. Pseudo-Kodinos, ed. Verpeaux, 344, 347; Professor V.L. Ménage kindly offers the etymology of kor = belt.

32. Clavijo, ed. Estrada, 75, where proconestart should surely be read as protovestart; trans. Le Strange, 112, has protovestattr

(again, better protovestart); at any rate, it is not protovesttarites.

33. M.M. Alexandrescu-Dersca, la campagne de Timur en Anatolte (Bucharest, 1942; rp. London, 1977), 123-24 (letter to be re-edited by A. Luttrell); O. Lampsides, “To axiopiston tes epistoles Dabid A' Megalou Komnenou kai to egkomion Trapezountos tou Bessarionos,™

Byzanttaka, 2 (1982), 53-60 (defending the military estimates of the 1402 letter); Hilda Hookham, Tamburlatne the Conqueror (London, 1962), 243-44; lorga, Notes et S&rtraits, I, 113; M. Sanudo, Vitae ducun Venetorun (Milan, 1730), cols. 797-98. 317

be sole emperor?

The interpolator of Chalkokondyles has a similar story of over twenty years later. He states that John IV accused Theodora Kanta-

kouzene, his mother and Alexios IV's bride of 1395, of adultery with the then protovestiartos. He murdered the protovesttarios and proceeded to lock his parents in a room with the intention of killing them also. But the archontes of the city, who appear to have been on John's side against his father, as they had been on Alexios's side against hts father a generation before, dissuaded him from further murder, and he escaped to Georgia. In Trebizond

Alexios IV named his son and John's brother Skantarios (i.e.

Alexander) co-emperor in his place. In Georgia Alexander the Great (1412-43) befriended the renegade Alexios and gave him his daughter as wife. John went on to Caffa seeking someone who had a ship in order to invade Trebizond and make war upon his father Alexios. He found a Genoese captain called (name missing) with a large caravel,

furnished with all kinds of arms, and set out against his father Alexios in Trebizond. They landed at St Phokas, Kordyle, and captured the monastery and set up camp within the place. Then the

Kabazitai came secretly to him to betray their emperor. For they formed the imperial bodyguard in the proasteton of Achantos (now Ahanda, near Platana~Akcaabat, midway between Trebizond and Kordyle) ,°* with armaments and engines of war. The Kabazitai allowed the enemy to reach Alexios safely and let John's archontes enter Alexios's war tent unsuspected, where thy murdered him at midnight. But the Emperor John had ordered them not to kill his father, but only to capture and bring Alexios to him. But they killed Alexios, rather than capturing him, hoping that John would show himself the better pleased. But afterwards the Emperor John

had one of them blinded and cut off the hand of the other, to

demonstrate that he had not wanted to have his father killed, only to have Alexios brought to him. In this way John entered into the rule of the empire; and he showed considerable reverence towards

his father, burying him first in a tomb in the Theoskepastos

monastery and later moving him to the metropolitan church (of the

34. There are in fact two candidates for Achantos. One is the “Achantakan” ~ i.e. chandax or palisade - of St Kerykos, which the Tiirkmens attacked in 1336: Panaretos, ed. Lampsides, 64. Bryer

and Winfield, Pontos, section XX, no. 39 (Church L), have tentatively identified this with the site and well-built church at

Kryoneri (Soguksu) above, and to the south-west of, Trebizond. The other is Ahanda, a group of settlements along the stream north-west of Platana (Akcgaabat), where there are signs of a chapel and other medieval remains: section XVII, no. 7. I incline to this Ahanda as

that of Alexios IV's assassination at Achantos, on grounds of its modern name and because it is midway on the highroad between

Trebizond and Kordyle. 318

“THE FAITHLESS KABAZITAI AND SCHOLARIOI”

Chrysokephalos).>>

The interpolator of Chalkokondyles, who supplies this story,

is also the sole source for other details of fifteenth-century

Trapezuntine events. Spirited and gossipy, he reveals in this and other passages that he knew the geography of the Trikomian coast between Trebizond and Kordyle intimately - for example his is the only literary reference to the medieval site of Ahanda. But he is quite innocent of the passion for dates shown by local annalists such as the continuator of Panaretos and is difficult to control.

The sheer colour of his accounts is against him. As Laurent observed “son origine et sa valeur sont encore 4 6tablir”.2° ‘the

Bodleian obit should now make him be taken seriously as a source.

As Kurganskis pointed out, the interpolator garbles his

account of Trapezuntine and Karakoyunlu and Akkoyunlu alliances,

and dates the marriage of the Grand Komnenos Alexander to the

period between the flight of John IV and the assassination of

Alexios IV, whereas Tafur testifies that it must have been almost a

decade later. 37 This sort of error is understandable in one

writing well after 1461. In other details he can now be checked further. John must have escaped to Georgia before the mother he had abused died on 12 November 1426, and was buried inside the Chryso-

kephalos cathedral. 28 The interpolator's statement that he married

a Georgian princess soon after is not in necessary conflict with

Tafur's that he was married to an infidel in 1438; Grand Komnenoi had successive wives as well as mistresses and Tafur may not have been too clear either as to whether Georgian princesses, who took

Turkish epithets, were Turkish too. 2? Tafur also stated that

35. Chalkokondyles, Bonn ed., 462-64; ed. Darks, II, 220. 36. V. Laurent, “L'assassinat d'Alexis IV, empereur de Trébizonde

(+ 1429). Date et circonstances,” Archeton Pontou, 20 (1955),

140 note 3. Cf. V. Grecu, °Zu den Interpolationen im Geschichtswerke des Laonikos Chalkokondyles”", BSHAcRoun, 27 (1946), 92-94

(the Trapezuntine interpolations in the Bonn ed. are in pp. 462-66 and 494-98); H. Ditten, Der Russland-Exkurs des Laonikos Chalkokondyles (Berlin, 1968); and tdem, “Die Korruptel Chorobton und die Unechtheit der Trapezunt und Georgien betreffenden Partien in Laonikos Chalkokondyles' Geschichtswerk,” Studta Byzantina, ed.

J. Irmscher (Halle, 1966), 57-70 - a supposed confusion between Rostov-on-Don and the obscure village of Chorobe (Horovi) in (significantly) Trapezuntine Trikomia.

37. M. KurSanskis, “La descendance d'Alexis IV, empereur de Trébizonde,” REB, 37 (1979), 239-47.

38. Panaretos, ed. Lampsides, 81. 39. M. KurSanskis, “Relations matrimoniales entre Grands Comnénes

de Trébizonde et princes Géorgiens,” Bedt Kartlisa, 34 (1976), 116-17 (on Turkish Caucasian epithets), and 122-25 (on John IV's

marriages). His reading of Tafur is that it was John IV, not 319

Alexander, brother of John IV, was Alexios IV's elder son; albeit in exile, he was the first emperor of Trebizond he met, and he may have assumed that he was older than John IV. The interpolator of Chalkokondyles (who calls Alexander “Skantarios” in the Turkish

style) does not state the order of birth, but explains why

Alexander had a certain pre-eminence. On the flight of John to

Georgia before 12 November 1426, Alexios made Alexander co-emperor.

The rest of the interpolator's story can be checked. On 8 November

1427 the Genoese government wrote to its Crimean officials

restraining its officials from making war on Alexios IV because his son had fled to Caffa - presumably from Georgia. 49 Genoa was on poor terms with Trebizond, though it may have been with John IV,

rather than Alexios IV, that the Republic was on poor terms. It may be that the Genoese kept John in Caffa until they established which emperor would be more amenable to them. At any rate Antonio de Allegro was sent from Caffa to Trebizond to discuss peace terms on 22 August 1428. 41 He would have seen the coemperors Alexios IV and Alexander and, if the way that Tafur reported to and from the then exiled Alexander and John IV ten years later is anything to go by, would have informed them of the

movements of the exiled John IV and told the rebel of conditions in Trebizond. I venture that Antonio de Allegro's relative Domenico de Allegro was the Genoese captain whom the interpolator of Chalko-

kondyles states John IV made his protostrator for his invasion;

in December 1448 John's “prothocapitaneus” and “prothostrator” was identified in a Genoese document as Domenico de Allegro. 42 This

high office in the Trapezuntine taxts, second only to the Grand Domestic, had never before been given to a foreigner. Was it the deal which was worked out with John after Antonio had seen the situation in the Pontos? It would have assured a Genoese presence

in Trebizond as the price for assistance in furthering John's

candidature.

Two articles, respectively by Miller and Laurent, have

narrowed down the date of the assassination of Alexios IV, hitherto Alexios IV, who had approached the "Grand Turk", who was in fact

the Crim-Tatar khan of Solkhat, giving John the wife Tafur was to see in succession to his Georgian one, during his sojourn in the Crimea in 1427-290; the alliance of John's brother, the last

Grand Komnenos David, with Maria of Gotthia would also have been

arranged then. I think these propositions are too ingenious, for they suppose that John IV would have both been allowed by the Genoese of Caffa to negotiate with two rival local powers, the Gotthic and Crim-Tatar, and that both would at that time have

considered it worth making an alliance with a refugee pretender to

the Trapezuntine throne.

40. lorga, Notes et Extratts, I, 463-64. 41. lLorga, Notes et Extraits, I, 467-77.

42. Iorga, Notes et Extraits, III, 234. 320

“THE FAITHLESS KABAZITAI AND SCHOLARIOI”

thought to have been in 1447, to between 8 November 1427 and 28 October 1429.43 The latter published a decision of the Venetian Senate on 28 October 1429 to subvene for the refortification of its base in Trebizond, requested by its batli, “quod propter novitates in Trapesunda inter Imperatorem [i.e. Alexios IV] et Cheloiant [i-e. Kalotoannes or John IV] filium suum, qui interfeci fecit Imperatorem patrem suum et fecit se imperatorem...."44 The Saptentes carried the motion with two against and one abstention. From this Miller, Vasiliev, and others, have taken the assassination to have taken place in September 1429; so far as the batlt’s report is concerned there was no doubt as to who was the murderer.

But autumn communications between the eastern Black Sea and Venice

do not justify a September date. The fall of Trebizond itself,

probably on 15 August 1461, was not heard of in Venice before 20 October. * The point is important because it has been assumed that a second deputation sent by the Genoese consul at Caffa to find out about peace with the emperor on 2 August 1429, was sent to Alexios IV who, if the Bodleian obit is of his death, was over three months dead. 46 In fact, however, no emperor is named; he would have been

John IV. It is likely, therefore, that the Genoese delegation was following up John's Genoese alliance after his coup, so alarming

the local Venetian batlt and precipitating his letter to the

Senate: here the dates fit.

IV. The evidence of the Bodleian erasure (pl. 1) and of the Marciana omission (pl. 2)

I propose that the erasure in our Bodleian obit of 26 April 1429 is

related to an ostentatious gap of ten lines in the final folio of the unique manuscript of Panaretos's chronicle of the Grand Komnenoi, in Venice Marc. Cod. 608, fol. 312° (pl. 2), and ask why and when the information they once provided was suppressed. The Marcian manuscript is not’a holograph, for Panaretos died soon

after 1390; a continuator or continuators added five (or more)

entries thereafter. The copyist of Panaretos and his continuator(s) added a probably Trapezuntine taxis immediately after the chronicle

in the manuscript, suggesting a court provenance, and use, for 43. W. Miller, “The chronology of Trebizond,” EHR, 38 (1923), 408-10; Laurent, Archeton Pontou, 20 (1955), 138-43; cf. Vasiliev, Bysantton, 10 (1935), 65-66.

44. Laurent, Archeton Pontou, 20 (1955), 139-40; cf. Iorga, Notes

et kxtratts, I, 505.

45. There is no parallel for the issue of a bull in the name of a dead emperor by and with his effective parricide. Second (and so far unremarked) is the evidence of a graffito low on the base of the bell-tower of the Hagia Sophia monastery church, just west of the Pharos. Its building began in 1426/27,

during Alexios's last, unguarded years. It was John IV who

embellished the church and completed the tower, which was finally

51. Beldiceanu, “Les Qavazid/Kabazités,” 42 and note 4; PLP, I/5 (Vienna, 1981), nos. 10007-11. 52. Beldiceanu, "Les Qavazid/Kabazités,” 43. On St Phokas, Kordyle,

see Bryer and Winfield, Pontos, section XVII, no. 7. On its

trading monks, see S. Karpov, Trapeszundskay imperty t sapadnoevropeiskie gosydarstva v XIII-XV vv. (Moscow, 1981), 37, note 162.

53. Chrysanthos (Philippides), “He ekklesia Trapezountos,”~

Archeion Pontou, 4-5 (1933), 500-3; Laurent, Archeton Pontou, 18 (1953), 250-56.

323

painted in 1442-44. He added two undated imperial portraits on an exterior panel of the tower. The panel remains prominent, but the paintings have weathered. They depicted Alexios IV in public assordation with John IV, which an inscription identified as his gon.” The third, most spectacular and open display John made of his remorse was the splendid free-standing tomb which he built for his father outside the main apse of the Chrysokephalos cathedral. The period between when the obituarist N2 of the Bodleian syn-

axarion, and the continuator of Panaretos could have told the truth

about the assassination, and then had to suppress it, is therefore limited. One limit is 26 April 1429. Another is probably August 1432, when the Pharos bull was issued. Falling within those brackets would be the sending of Alexios IV's last co~emperor, Alexander, on his long exile, and the elevation (and marriage) of David as John IV's co-emperor. In this connection it may, or may not, be significant that the Italians were under the impression that only John IV was emperor on 6 March 1431;°° if this is really so, it postpones the marriage of David to November 1431 or even later. Perhaps falling outside our range of dates was the superseding of Hand N by Hand O in the relay of annotators of the Bodleian synaxarion, between 22 March 1430 and some time before 1434. And at some stage soon before or after August 1432 John IV

made the most public expiation possible, in displaying open

portraits of himself and his father on the Hagia Sophia bell-tower, and by moving the remains of his father from the Theoskepastos, to be buried not, like all other Grand Komnenoi, within the confines of a monastery, but outeide the cathedral in the most public place

in Trebizond. That done, John IV would have been at last in a position to rehabilitate the Kabazitai and Scholarioi and silence local gossip about his own role in the murder. At this point Hand

N would have found it prudent to erase Hand N2's entry, and in the court itself the copyist of Panaretos would have skipped an awkward passage, though not without leaving a tell-tale gap.

54. Fallmerayer, “Original-Fragmente,” II, 95; G. Millet,

“Inscriptions byzantines de Trébizonde,” BCH, 20 (1896), 432. G. Finlay's unpublished diary for 1850, in the British School at Athens, is mistaken in noting the date A.M. 6912 = A.D. 1403/04 in

the paintings. The wallpaintings of the tower will be published in

full for the first time in Bryer and Winfield, Pontos, section

XX, no. 112. While the John IV as patron is obvious enough in the exterior paintings, one seeks in vain for any sign of his patronage in the decorative programme of the interior which seems to have

been made by and for monks. But the Hagia Sophia itself was extensively redecorated in John IV's reign.

55. Bryer and Winfield, Pontos, section XX, no. 25; N. Baklanov, “Deux monuments byzantins de Trébizonde,” Byszantton, 4 (1928), 373~91; Chrysanthos, Archeion Pontou, 4-5 (1933), 388 and pl. 11.

56. Iorga, Notes et Ectraitse, I, 538-39.

324

"THE FAITHLESS KABAZITAI AND SCHOLARIOI”

V. The evidence of the tomb (pl. 3) and of the skeleton (pl. 4) Although what must have been a splendid array of tombs of other Grand Komnenoi within the parabema of the Chrysokephalos was destroyed when the cathedral became the Fatih Camii on the Ottoman conquest of 1461, that of Alexios IV outside was spared until 1918.

It owed its survival to an Ottoman tradition that it had been re~ used for the burial of a Turkish hero of the siege of 1461. He was

a young shepherd called Hogoglan, perhaps from the nearby village of that name, who was supposed to have been either the first to get into the city on 15 August, or to have made the entry possible by shooting away the chain of the Tabakhane drawbridge. He was then

said to have been either killed in the assault, or by Mehmed II

himself in disbelief of his story. After 1461 the bars of the crosses in the roundels of the tomb were removed (pl. 3), an inscription referring to Hogoglan was added, and until 1914 the cannon ball and chain of the tale were hung in the tomb. This hand-

some building served as the model for the tomb of Solomon, last king of Imereti, outside the then Greek cathedral of St Gregory of Nyssa in 1815, and for theabove surviving tomb of the Metropolitan Constantios of Trebizond the Theoskepastos in 1879. U In 1916 Feodor Uspenskij excavated the tomb of Alexios IV and Hosoglan, after the Russian occupation of Trebizond (pl. 3). He found that a sarcophagus made of re-used marble slabs and contain-~

ing a skeleton (pl. 4) facing west, towards the apse of the Chrysokephalos, had been subsequently broken into, disturbed, and probab-

ly robbed of ornament, for the insertion of a second skeleton.

There are some puzzling questions about Uspenskij's account which cannot now be answered: why did he believe that the sarcophagus had

been broken into from below; why, despite this fracture and the fact that the tomb stands high above the ravines of the Middle City, did he find it waterlogged; and why did he conclude from the

fact that the bones of the first skeleton were saturated in fat,

that the body had been embalmed? These questions cannot now be answered, but an explanation for the third may be that, following Orthodox practice, the earlier skeleton represents a reinterment of the bones alone (now often washed in wine), from a first burial (in this case in the Theoskepastos). This is now often done three or more years after death, but such parallels are too insecure to draw 57. 1I.P. Meliopoulos, "Peri tou mnemeiou Alexiou G”,” (sic) Hot Komnenot (1916), 205, 234-5; H.F.B. Lynch, Armenta, fravele and Studtesa (London, 1901; rp. Beirut, 1965), I, 22; Selina Ballance, A. Bryer and D. Winfield, “Nineteenth=century monuments in the city

and vilayet of Trebizond: architectural and historical notes,” Part 1, Archeton Pontou, 28 (1966), 246-51; Part 2, Aroheton Pontou 29 (1968), 103-108. 325

any conclusion about when our tomb was built. It is also now too late to question the obvious conclusions which Uspenskij drew as to the identity of the two skeletons, but they remain logical: that the earlier interment was of Alexios (pl. 4). On the Russian withdrawal from Trebizond this skeleton was entrusted to Chrysanthos Philippides, metropolitan and historian of Trebizond, who in 1918 reinterred it in the parabema of St Gregory of Nyssa. I cannot trace what happened to the remains ascribed to Hogoglan, but the

Turks had their revenge by destroying the tomb outside the

Chrysokephalos. During the Exchange of Populations of 1923 George

Kandilaptes, the Chaldian antiquary, salvaged the remains of

Alexios IV and brought them to Greece, where they found their way

to the Byzantine Museum in Athens. They were boxed in a mich smaller lLarnax than the extent of Uspenskij's skeleton warranted. In 1980 the remains of the Grand Komnenos were taken with much ceremony by a number of Greek Pontic organisations to a fifth and final resting place in New Soumela, near Kastania of Beroia (pll. 5 and 6). Here the wonder-working icon of the Panagia, attributed to St Luke, now attracts thousands of Pontic pilgrims every 15 August, as it had to Old Soumela monastery, near Trebizond, to which it had

been given by Manuel III, father of Alexios IV. Alexios IV's is probably the only surviving skeleton of a Byzantine emperor, and it

is unfortunate that it was not examined by an osteologist during

its sojourn in the Byzantine Museum. But Uspenskij's report

provides the final confirmation of the midnight assassination of Tuesday 26 April 1429, when the Kabazitai and Scholarioi families proved faithless to Alexios IV, rather than to his son John IV, for he found the Grand Komnenos decapitated.°°

58. F. Uspenskij, “Usbypal'nitsa tsarya Aleksiya IV v Trapezun-

tini,” VtzVrem, 23 (1917-22), 1-14 and fig. 3 (my pl. 4);

Chrysanthos, Archeton Pontou, 4-5 (1933), 386-89 and fig. 10 (my pl. 3); “To historiko ton oston tou Alexiou D” (1417-1429),” Ponttake Hestia, II, vi, 32 (March~April 1980), 141-42; a special report on the reinterment of the remains in what appears to be a new reliquary measuring 46 x 28 x 33 cm., in Ponttake Hestta, II,

vi, 33 (May-June 1980), 233-49, with my pl. 6 on p. 242; cf.

Makedontke Zoe, XV, 170 (July 1980), 9. These reports are contradictory as to the exact date of the reinterment; however midnight of 26 (rather than 25/26 or 26/27) April 1429 is the exact time of the assassination as the Byzantine day began at sunset. In stating that the skeleton is probably the only known surviving one of a Byzantine emperor, I overlook relics of Sts Constantine and Helena (scattered), St Theodora (in the cathedral of Corfu, from Constantinople), perhaps parts of John Tzimiskes (in the Great Lavra, Athos), as well as possible remains of thirteenth-century Angeloi

(in Arta): see A.K. Orlandos, Byzantine Mnemeta tes Artes = Archeton ton Byzantinon Mnemeton tes Hellados, II (Athens,

1937), ii, 47-49; cf. D.M. Nicol, The Despotate of Eptros (Oxford, 326

“THE FAITHLESS KABAZITAI AND SCHOLARIOIL”

1957), 198. On New Soumela, see the annual Hemerologton Leukoma

Panagtas Sounela (Thessalonike, 1965- ). In my pursuit of the

fate of the remains of Alexios IV, I am most grateful to Mrs Athena Kalliga of the Epitrope Pontiakon Meleton, Athens, Mr David Turner of this Centre, Dr Myrtali Acheimastou~Potamianou, Director of the Byzantine Museum, Athens, and Professor R.E.F. Smith; and to Mrs Gaye Bye for typing this. STOP PRESS. Bay Cumhur Odabagioglu of Trabzon kindly tells me that

Sakir Sevket, Trabzon Tarthi (Istanbul, A.H. 1294/1877 A.D.), a work unknown to Uspenskij and unavailable to me, states that the tomb of Hosoglan (and of Alexios IV) was reopened in 1842. “Only a skull of a young man” was found. This excavation may account for the evidence of disturbance found by Uspenskij 75 years later, as much as Hogoglan's own interment in 1461. Either the excavators of 1842 were not very thorough or they are responsible for adding the bones now lying at New Soumela. Who can now tell? 327

O89 ée wee Eee 7rolan 7a aahet% ae”oN! re uh gee a vm 4. 8 A vq Py 3 OOH i: aa a eee hat ¥ toe save OG o PAA: Fe mf wrgey CA,

a ae -. an 4 | Lo OTN eke tna lk Oe lcSe

& a P * | ‘ , eae & aes , nA o . PX 4 “2 OQ, & fF . »@ . | AM dad yFuatees; - asl + P:-

SOD_—A ate cab uv ew evo SGparwespary(ey ny od et Bd rmgeheus Ainipiou ° Gut rep Aw piép 6 -wHee.

7OF ar o~ ole? . of ~ > P 2 ~ ~ 7 o2 . X ® —, r AY a , a”

p Ov -re awrnauriuewmy rat GG, TOU Te ayo

ABE PS TEU Ct TOG > £4.6 -yabn ON AtauTuwY anes

—s~ 2 ¢ » “op 7 ~ a _ , J « - 6 Nee NG 4 WTS Levres ayaspeAnas gt Dp Grane - 1 0979 SHOU EK TO

om engad . gre oylas mgamon oy Levyuprren - Joy

a An a | . ’ 7

qo -KYUOoc sa byoe CAILs pice - Kenoon oraGm Sb Ge rs

a ae o> ~ 7 .—~* » Ss ~~ 8 tet rer swrae arn xeveroms enpored vaenp Ene ie, comet emis Loonie mn Le a! as Mune. onan tu - ee th 0h ieee

Se Fed °”SRERe

Plate 1 The erasure. Oxford Bodleian MS Gr. Lit. d. 6, fol. 62° supra.

a

It ee 2

mp p unl tal YPCONe SF

daien C5 nn he OS

a ters ye ees a

ert ee

TSS sand Pe KT peel o

(ME aber -greterchd cs 2g re ape FR —

Plate2 The omission. Venice Marciana MS 608, fol. 312".

rs. *ye-_ . ,aaa. ;ae wo *,Pati # ,*, al ee yo ni ar . Wage eo rn * a Mh y os‘aoea ; ™ | , . : . ss - .

~ ae ae ere Be ; |q 4 j ”4 _age $a Se..§ | OVI

4 a : , * ; ee it ~ we $ 43! way | 47". t,, ai ld |¢. ( 9a eee.+» | uae £& ad ;-“”tia o, . ao FRR: 12 weae ; a “ae i’. foo . we all =. +

go aa f, int _ “¢ * t y « . — “— .

«eNP t " t $ Ro: » . . Ba q ~"Re‘ - t .

r) . ee ba ; ] ) , _ * F 4 . . a “oe o ~” :

sO . een St nee, ’ ne, Se ae om aaa\aewy Ware *-*am, -; anmmm . on Plate3 The tomb of the Grand Komnenos Alexios IV from the south at the time of its excavation in 1916. The metatorion of the Chrysokephalos cathedral is to the west (left). ~ Mier eee.

REE * a. eee . nee ib mee

a reCeSome, 2)wwe CE SS heos DA>Sees RRS aiee eee “6 225i os ape > SS > = ¥ aa ' =e! a - . on a= ‘s.Be ‘i

% Cf ~\ f Se3 BL

R—— t_—. «th ‘ we me=f of a.Ee ‘Lad 4APy Wigs =o> =aSS ey ae 5 alee pteor Woe. re a in On > Te, ~ ; ae: ; =r

‘ . i . oy . t é ol ee ae. Say ee AS eorw e e Sy ye, | «oe k ; Wy,

Byte sy oy a, © 2 vu = beet Bits *s ‘4 “{ Fsae odee aeBas oevan Wi{ te LEY

— a ‘ Wiig 25:

cA an folie 04 | Semen Pak A ee. *

"e/a, 4 —see, 4, ?eal & ai?OY oh. ,MCE isk eee

A Fy... teen Ne mye af.QR ; ® r 0gFA ees ~ ie wrx ey as ae A Me Pe, Ae SAE ae eee, Sh See. : + “=p >» ay ae . 0 t = 7 1, “{ ‘ ; ate 5

+ ~" ras Be ‘ ai fifi: " + ’ ° “Ad i o? . 7 : ie of ‘. _ Patek i: 7 _ . 7 =if and th al in th Y meagre remains seemed to be in imminent danger apustery and there are several in ue museum. of being cleared away in 1967. Evidence survives

Dated 857, 858 and 859, each reads: “The Tower

of another church on the southern street and

of Michael, the Great Emperor and Autocrat in th unl d € Tornik

. ; line . e Dormition of thetheMother God, second of defence against Arabs inof case oe

the most

Christ’. Michael had repaired the walls asa 7) yo an unlocated monastery Of 1 OFNIKEs.

of retreat, but it proved unnecessary. Its early famous church of the city, was all but destroyed prosperity was never recovered, but Nicaea was m 1922. Clearly one of the most important of all still important; the Continuator of Theophanes Byzantine monuments, its appearance and dec-

noted that it was a ‘city of ancient wealth and °Fation can only be reconstructed from earlier

large population’. descriptions and photographs. The drawings of The new enemies were internal. Soon after 863 John Covel (1677) of the church are published

the Paulicians, heretic rebels from the east, here for the first time. | reached Nicaea and in 978 Bardas Sclerus took The Dormition was not a large church; the city in his revolt. On September 23rd, 1063, originally founded by, or belonging to a monastery

came a serious earthquake, but stretches of the f, Hyacinthus, it was centrally planned and walls, including twenty towers, and two churches domed. Its interest lies in how early it can be were speedily rebuilt. Then in 1071 the Seljuk dated — perhaps even to the sixth century. It was Turks defeated the Byzantines far to the east at certainly redecorated by the iconoclasts in the Manzikert and found most of Anatolia unex- eighth century, who replaced what was probably

pectedly open to them. In the ensuing chaos, 4 figure of the Virgin in the apse mosaic with a Nicephorus III Botaneiates proclaimed himself simple cross. Soon after the lifting of the iconoEmperor in Nicaea on March 25th, 1078. He _ clast ban in 843 (or maybe even in 787), Nauobtained the throne in Constantinople, but left | cratius replaced the cross with another austere his Turkish mercenaries in Nicaea. They first and striking figure of the Virgin and Child who declared for another rival, but by 1081 Siileyman, | were blessed (in a then surviving part of the the Seljuk leader, had made it his capital. Like earliest mosaic) through rays of heaven which 26

om »- ; ee a a a. aa AB= OR DA ibeet tA IS 8 awes : a — . : ) : os eS + es " ’ *# -— ‘ae “8 & St aa ik hi

REA « Ste OF@oeShwe Ee ee: St veOS. Cibsie ints

a .ee it: ae , aS Pepe z+‘, et bs a “ut ne .. ORS _ eo th OY si ‘ ia a an mS. Ly LL al:

|aah aa le ee3uM : ae vvve * ; ie fo ke , . > a.ey* BiTe ’ adh oe2. cis o. *\ eo: ‘ .a . 4".i .é° cars . ee, ie. , .ty oat an ‘, / a, a me 2 sg +e i x . a a ‘ . ¥ jg 5 “ : ri : _ he ae mr: : a. r) . ; ¥ e- * ee an Ao : ' a Le a : ics : OM coef ° a an : tS A Ru : r 4 me. : *. ; . a aa — ok + 4 TT ‘a. i , : » gas [4 a oe . on . | . ; . ar fx, # ae & ae . Fi a See te *2 :« e“ws“

ne7wee f ** a x) :: en Ts bad &a =aaof: | Le BS:ae . O ;-f. ...|eS “a : “ al ae | as + : . » ok Zz ; a. a N a RG o < “. ~ Po, (2) : ‘n rn . i + an i ae ; *, . | mo LR 1 “ aaa ee ~, a sa ~a Y ae Mp | a > | SE te OS Z x a7 . T" a. ? be Cae | ttoe ‘} —_ | ¢ , = ~. oa . : " mR . cM FS i eee. ah Se , 0 a. rn a, = ‘Pa. ¥ 4“,*ee “)s.x. :. i we” . ,. ©F, . . we ra 4 f ie ~ a . > > Je ee

é :2 . OP- ae i) .* .:.4t+ :’

eS .:” .

%

Such insulae may originally have comprised about eight back-to-back private houses. More important are the gaps in the domestic grid, for public places which few late Byzantine towns chose to inherit. Two or four oblongs create a square agora or temple area; a sequence of insulae allow a stadium or hippodrome. These features

are evident in Thessalonike, as well as in some Milesian colonies on the Euxine.

The modern street plan of Sinope (an isthmus that lends itself even less to Hippodamian principles than did Miletos itself), reveals a cardo or decumanus, and

sites for an agora, a temple area and an acropolis, all within a wall; what was probably an extramural gymnasium became a possible Byzantine granary and is now ‘Balat kilise’. Trebizond shows up its colonial origins less well, but distinctively (fig.3). The classical vicus was never walled and lies in the eastern suburb by the commercial harbour. It is in this area, where the modern street plan traces a grid of insulae, that the rather meagre total of classical discoveries, and all but one of the classical inscriptions, have been located.® The plan reveals no sign of a threatre, hippodrome or gymnasium. for which scholars have naturally looked in vain, but the ancient agora is obvious enough. It is the familiar cleared block of four insulae, called the Meydan even in late Byzantine times, and still the main

square of the town.’ But Trebizond’s distinction from the first is that it was a dispersed settlement. Fig. 3 shows that the ancient walled acropolis stood inland, quite apart from its own port, 800 m. to the north, and from the vicus, up to 1

km. to the east...

Constantinople was the last great city to be inaugurated and endowed with civic conveniences on the old lines. These should have included a street plan, which one would expect to find signs of between the Golden Horn and Great Palace areas, and along the Mese, especially. True, the milion, and some high streets and fora survive, but despite Justinian’s legal restrictions on tampering with street lines, any regular grid has been submerged beneath a warren of streets. 81 suspect that infilling settlement was too slow and haphazard to take regular root, while monastic sprawls can play havoc with any Hippodamian order. Ancient planners took no account of which way future churches were to face. Inherited street grids are all the more precious because most places did not actually

start with one. Apart from their public buildings, older foundations grew more Hampartumian, “The archaeology of Histria (4th-7th c. A.D.),’ AP, 35 (1979), 134-36; and J. Smedley, ‘Archaeology and the history of Cherson: a survey of some results and problems,’ AP, 35 (1979), 172-92. Fig. 1, an aerial view of Nicaea taken by the author in 1971, shows its grid within a pentagon of walls. More interesting are what seem to be signs of a projection of the grid into a larger pattern in the field system to the north and east of the city, looking very much like the imposition of Roman regimentation on local terrain and land use: centuriation. Yet while centuriation is found in Dalmatia and as far east as Acre (Akko), I do not know that it is found elsewhere in Anatolia. See J.J. Wilkes, Dalmatia (London, 1969), 225,228 and pl.30; I am grateful to Professor A. Birley for discussion.

6. Zosimos, Nea Historia 1, 33. Bryer and Winfield, Pontos, section XX. 7. Michael tou Panaretou, Peri ton Megalon Komnenon, ed. O. Lampsides (Athens, 1958), 75. 8. Justinian, Digest, xliii, 8, 10 and 11. CF. Vickers, JHS, 92, (1972), 161 and n.24. 266

a ” o ; Fg “a. oe . * , . 7 ms *. P of a 3 », 4 * s id » " eA

% ar, aa “ , ‘ ae 7 a - ° ia? - ° ao%eeihe te a# "7 4,. ..cm i A ., ‘a ’“4 *§ aq7” ; a .. ; eee ee ™ : 6y"i i g»>‘. iRe .;"s a ;"-Bas “a. 4 >Par asd4SMES, Kte; a: 7) 1 as 2 hee, me ;mel mi “a . pe %ie’ *aoe . e. ,e>_isYaoe . Fe :age °Ba *™; oF waon Mi

@ ¢'ce » 4... . , a +“se iS cade PA PJ. Re a. . ” @. ééna a .ifLd eee, ge &:Sie ereeadae - &dip : ‘+h Ww. i as a.

."a4, ‘, aSad 4SR ae ee “.. - ca ius eg tea SE7" ol“..aaa .“e .gS“ yey 2‘. a4 fWy rey xg” od b.1OT *3 ie as. 7";5 *% ' o™ Se x, " Loan ; oat . e a, a, ‘, ey 4%" ne ‘ ma A) *, u& aywar : -.oo ‘.,} a) a)bd “: -af ers , oO “.a..an. .~‘.an . 37Wi ;. -° om x 3 td ie he ow e 5. ‘7* aa> +h. if 7 », Fag he ee TRE, +. 4 ye , eae . Fe ay ek ; . Pte: . _ F »~*~ *‘ we . Ye : »Sa” mh: eS‘ff r64‘ 8 4 |yop Te%.. &, .ne. , “a "4 ' "“4aa r 4x-ery 7Sawe Etdhee . .5“. y.:

| ey A Se OA eo) t ithe .,8‘~% aa? a. 4 * =. aa *eoeo, °& " . Lo a & *. « . °* ee . — > . ‘ ” * pr , . 7 Pt ‘ ‘s ; . ‘ B 7 “%’ 7 > » + . s * * . i* 4

"Ss & a Res! . ae re. n | [a A cy Cae, ee, e. iaee(ee eo ‘gee Hf .4 “3 & hoe Lg on ~-a ‘9 OE “iy 4a¥Ee aE Oe Bye ; i: ie aah . bd . r “ . BS . bal . 3 F . ose . wae 2 y PY . b .: aan ~., * a ra ‘ « vs .= 1g. ri 4 Le . ow ¥ . ‘s”: i hs i's . ¥ oe) a ~*y “4” a4*ww ‘ . #>3[du i"ae . Fate " .Es a,i‘°Y \ ,"Pa Lo’4Ae. abet ae . % = ; “a/ ae one Vane hot©n ‘ an SeZ 4aeae °3 :‘yoa \. xPat tte Py . +

.; ;ae i; FLtd!~?a €«C = ee> 4=. 4, a oO pan< . Pe N Ss . ‘ ‘ sc, “Ya x dae By os is

. +" _ » . ° ee

*? P‘eo > > 4°d~ 4 9yy | “aaart7. ” pee o. * .are a eh

‘J * \ ~ “ fe ; nd L | S : 4 ww ®: es | ¥ o ns o ai * b Sane} Ek + nea .aoe t~reA\) . af"‘ce & 4, . y _ fs , oe “he. nO wv» i ‘ ; + x @ wa7 . * X * . = , 4, "g . : f* ‘: “ ow ae “SY . ae . ‘ poe ond

¥ . XY % "~~» >. Wea oy .;| :.woe a ge8: ;ay - ice 7me 3» ie! 2t Nte Eo iy ° ea *aNee? a Po ¥3 .‘iV La..we ra ety PES x oe © nt rns oe a's — . ans ' ~y ‘ *ve Fi 2 ‘ , . Rs > Ps wi a . J * z * x € s $ Meson E "e a ibe! we. ~ . ; , , ’ ve b ;a>.’.o$«>» a). “ am — -% . ; — ¥o ‘ a “ £ . » he . ; rs ‘ ; . _ e* « ,. =, an ‘s oe * ea, « : * ae! “a wee 28 te we ; ra ' ¢ . 4 ¥ eg a . a / . ass . +] * : eo = ma St . ; % ® .. & x es ne " . ~. : 4 * _ es ~ A Sa t a “ . . .. #_ i. ; . . Put * (ES « s : . OF . ~ igre neal Heng, ee sy ao

+i.+estx‘eYs My «A 4 iw PNG Pe 7 ?. .bn Foal al.¥: +) . ”eeoi|: og3“,;..»:, ”ms aPe7tate 7°,A,agal 2‘. sk +s vt Ay : j ; Se ‘ ifREE 3 ~We. cs i. (Le, * . .os. faCg ; aa .Aeeee ter * o£ - aee Ps ° :} :Se . Ff vg. hs ,SI %. Boes ee oe US*: gs oeaad pn rr arBh ae 8. Pe fo ai wr Re ge . © In 14. In all cases | have made my own calculations of the size of cities. Constantinople I estimate at 1,830 ha. within the Theodosian walls from the map in W. Miiller-Wiener, Bildlexikon zur topographie Istanbuls (Tiibingen, 1977), but Russell, TAPS, N.S. 48/3 (1958), 77 gives it 640-1 ,200 ha. in the period between the building of the Constantinian and Theodosian walls.

I estimate 37.5 ha. for Galata (Pera) from A.M. Schneider and M. Is. Monidis, Galata. Topographische-archdologischer Plan (Istanbul, 1944). For Thessalonike I estimate 270 ha. for the walled city and 15 ha. for the acropolis from Vickers, JRS, 92 (1972), 161, fig. 4 and other plans; but Russell, TAPS, N.S. 48/3 (1958), 77 (‘Thessalonia’ in table 80) gives it 234 ha. The estimates for Trebizond derive from the plans attached to Bryer and Winfield, Pontos, section XX: 1.9 ha. for the acropolis, 20.2 ha. for the middle and lower cities walled in the later medieval period, and 92 ha. for extramural areas known to have been built up, totalling 114.1 ha.; but Russell, TAPS N.S. 48/3 (1958), 81, table 84 gives it 105 ha., based on the supposedly Byzantine city in H.F.B. Lynch, Armenia. Travels and Studies, | (London, 1901; reprinted Beirut, 1965), 30, which is a sketch the size of four thumbnails. These rough hectarages may be divided into the following population estimates. For Constantinople in 1477, figures are taken from H. Inalcik, ‘Istanbul’ EF using the conventional multipher of five; they are abbreviated rather misleadingly in the same’s “The policy of Mehmed II toward the Greek population of Istanbul and the Byzantine buildings of the city,’ DOP 23-24 (1969-70), 247,249, where a still unpublished register of Galata in 1455 is noted. The figures for 1477 total 14,803 households in Istanbul and 1,521 in Galata. On the population in Thessalonike, see O. Tafrali, Thessalonique au Quatorziéme siécle (Paris 1913), 15-17; Russell, TAPS , N.S. 48/3 (1958), 77-78, 99-100; and especially now the Ottoman figures published and analysed

by H.W. Lowry, ‘Portrait of a city: the population and topography of Ottoman Selanik (Thessaloniki) in the year 1478),’ Diptycha, 2 (1980-81), 254-93. The 1430 figure is depressed

because it is on the eve of conquest, but an Italian estimate of 40,000 for 1423 seems even more distorted in the other direction, while Russell’s figure of 30,000 is simply based on the size of the city. Lowry’s totals are 10,414 for 1478 (58.5% Christian); more than 20,331 in 1500 (about 42% Muslim); and 29,220 in 1519 (53.8% Jewish). The estimate of 1438 of 4,000 for the population of Trebizond is by Pero Tafur, Travels and Adventures 1435-1439, tr. M. Letts (London, 1926), 50, 131 (he gives Coron and Pera each populations of 2,000, which is equally credible). The figures for Trebizond thereafter are published and analysed by H.W. Lowry, The Ottoman tahrir defters as a source for uban demographic history: the case study of Trabzon (ca. 1486-1583) University of California, Los Angeles, Ph. D., 1977). Lowry’s total are 6,711 for ca. 1486 (80.78% Christian); 7,017 in ca. 1523 (85.68% Christian); 6,100 in 1553 (53.28% Christian); and 10,575 in 1583 (53.62% Muslim). 15. Figures kindly supplied by Mr. A.E. Dunn of the University of Birmingham Centre for Byzantine Studies, who is surveying Chrysoupolis: 6.5 ha. walled area, 2.5 hs. citadel, 7 ha. extramural settlement

270

LATE BYZANTINE TOWN

Trebizond the bulk of the population lay outside both the classical and medieval walls in 92 out of the 114 inhabited hectares. More intriguing is that it is only in the register of ca. 1486, when the walled city housed the then entire and largely intrusive Muslim population, that Trebizond’s suburban population density was lower than its intramural one: the defters of 1523, 1553 and 1583 show a lower density within the walls.!® So the presence or absence of walls may not be so significant. What is more important is that the late Byzantine townsman expected to have about two or three times as much /ebensraum as his ancestor or contemporary.!’ Even more significant is that his Western contemporary seems to have brought his own high scale to Byzantium. In striking contrast to Constantinople in 1477, Galata had a density of over 200 per hectare, and even the tiny Genoese and Venetian quarters of Trebizond seem to have had a density of about 130 per hectare in ca.1486.!° Even Ottoman statistics may lie, and the fifteenth is an admittedly unrepresentative century in the career of the late Byzantine town. In the end it ts still literature, rather than even archaeology (which, they say, cannot lie) that gives the most vivid picture of the late Byzantine town. Literary sources strongly confirm the impression of the dispersed nature of the late Byzantine town: it was a series of garden suburbs. Nicaea (fig.1) looks green enough from the air: its quality of spaciousness was praised in the thirteenth century, but may also be termed emptiness. Ecomiastic and travellers’ description of Trebizond all harp on its greenesss: the place was one great leafy orchard.!? Manuel II recommended early fifteenth-century Thessalonike to Demetrios Chrysolaris: ‘Everything is within the walls’ there you may walk about if you desire without hurting your feet, whether you wish to relax, play, take a breath

of fresh air, or enjoy the beauty of the flowers; and so there is no need for you to ride about on a noble horse.’° But a horse was essential to getting around Constantinople, where the Blachernai palace lies 6.5km. from the Hagia Sophia. In the 1380s Demetrios Kydones rhapsodised spring in Constantinople, where he rode beneath the trees: ‘The fields and gardens of the city are coming into flower and the leaves put the tracks into such shadow that those who traverse them think these

tunnels run not through a city but up in the mountains.’”! All this is unsurprising. The area between the Constantinian and Theodosian walls in Constantinople was espectially slow to build up and still encloses market gardens:

the Blachernai to the north, and-the fishing village of Psamathia to the south, of 16. Lowry, Trabzon, 70,95,162,204, tables IV, IX, XVIII, XXV. The relative densities per hectare are: ca.1486 intramural 80; extramural 54; 1523 50:64; 1563 40:57; 1583 63:100. I have not taken into acount the emergence of a southern suburb in 1553, but it accounted for only 0.49% of the population then, and 1.51% in 1583.

17. See n.14. 18. That is, assuming that all 214 Roman Catholics shown as living exclusively in the eastern suburb of Trebizond in ca. 1486 in Lowry, Trabzon 70, table IV, were in fact quartered in the Vene-

tian base of 1367 (0.45 ha) and Genoese Leontokastron (1.2ha). 19. M. Angold, A Byzantine government in exile. Government and society under the Laskarids of Nicaea (1204-1261) (Oxford, 1974), 111; A. Bryer, ‘Nicaea, History Today, 21 (1971), 22-31;

and the eXphrasis in Bryer and Winfield, Pontos section XX. 20. Manuel, Letters, ed. Dennis, 112-13, letter 43. 21. Démétrius Cydonés, Correspondence, ed. G. Camelli*(Paris, 1930), 112-13, letter 43. 271

the strip were more populous. Psamathia boasted two great monasteries, but the monks of the Stoudios were just as much encouraged to farm as they were to copy

the manuscripts for which they are better known, while the future patriarch Euthymios described the monastery which he founded nearby in about 892 as ‘the quietest, most pleasant of places.’?2

St. Gregory Palamas found the churches and shrines of fourteenth-century Thessalonike were deserted for three months of the year, for the citizens were engrossed in the harvest and vintage: he implies without-the city walls.7> It is no surprise to find that 113 of a recorded adult make population of 163 in the Hellespontic town of Lampsakos (now Lapseki) were agriculturalists in 1219.4 What is more significant is that places like Thessalonike and Constantinople may have enjoyed the same bosky atmosphere in more prosperous days too. The major urban industry of even Comnene Constantinople may well have been agriculture and that not all extramural. In about 1200 Nikolaos Mesarites described what lay around the church of the Holy Apostles, in the heart of Constantinople itself: The land which lies about the Church is not only fitted for the growing of plants, and for the sowing of seed, but you may see in it trees growing to a great height and laden with fruit and with the vines which climb up in them, and crops growing under the trees; for all the land around this Church is full of strength and rich in wheat. People from whom this church is far distant can see from afar the wheat being brought in; for those who live near it, the wheat alone which grows in the land about their houses is sufficient for their nourishment, and they need have no care for invasions of barbarians, for the mighty waves of the sea, for the dangers from pirates, for the laborious drawing up of ships on the shore, for the troublesome handing of the grain by shippers, or for any of the other things which the mischievous minds of sailors can devise.*° But the late Byzantine polis and polichnion was not simply an agrotown, or intensification of its own hinterland. Two travellers sharpen the characteristics of Constantinople. In 1332 Ibn Battuta described Constantinople as Istanbul and Frankish

Galata, which had good but squalid bazaars. Istanbul consisted of a citadel with palace (the Blachernai), a paved city and market (round the Hagia Sophia) and thirteen inhabited villages within a wall.2° By 1403 things had deteriorated, but 22. Vita Euthymii patriarchae CP, ed. P. Karlin-Hayter (Brussels, 1970), 24-35; cf. A. Bryer, ‘The late Byzantine monastery in town and countryside,’ in The church in town and countryside, ed. D. Baker (= Studies in Church History, 16) (Oxford, 1979) = The Empire of Trebizond and the Pontos (London, 1980), Collected Study, VI, 222-23. 23. Gregory Palamas, PG 151, col. 333; cf. Tafrali, Thessalonique au XIVe siécle, 29 and n.2. 24. Angold, Nicaea, 110. 25. G. Downey, ‘Nikolaos Mesarites: Description of the church of the Holy Apostles at Constantinople’, TAPS, N.S. 47/6 (1957), 863, 897-98. The passages owe something to Libanius’s encomion on Antioch, but A.M. Schneider, Die Bevélkerung Konstantinopels im.XV. Jahrhundert, NachrGott (1949), 233-44 points to cultivated intramural areas and thinks that this passage may refer to the Lykos valley. 26. H.A.R. Gibb, The travels of Ibn Battuta, If (Cambridge, 1962), 508. 272

| LATE BYZANTINE TOWN Clavijo was struck by the same pattern: Galata ‘is but a small township, but very populous,’ and in Istanbul was the Blachernai citadel and a trading quarter between the Hagia Sophia and the Golden Horn: otherwise, ‘though the circuit of the walls is thus very great and the area spacious, the city is not throughout very densely populated. There are within its compass many hills and valleys where corn fields and orchards are found, and among the orchard lands there are hamlets and suburbs

which are all included within the city limits.’?’ |

Modern urban densities exhibit a regular decline as settlement recedes from a single

centre.” The late Byzantine town was rather a series of centres. But Ibn Battuta’s Constantinople of a citadel, a city and thirteen villages was itself not new. The fifthcentury Notitia urbis Constantinopolitanae is an inventory of the place under fourteen regiones, most of which were self-sufficient in baths, stoas, bakeries, churches, night watchmen, firemen and their curators. The Ottomans gave new life to the old pattern by organising the resettled city by thirteen nahiyes, each with its own jurisdiction.”? Later Byzantine domestic settlement was certainly more introverted and abhorred the open public spaces of the past, but otherwise the urban tissue of the dozen or more nexuses of both fifth- and fifteenth-century Constantinople was unexpectedly similar. Against my last suggestion lies one of the most remarkable reversals in Byzantine historical thinking over the last three decades: that far from having similarities, the early and late Byzantine town are basically incomparable. The now traditional view is that a major watershed between the Late Antique and later Byzantine worlds can be located in the demise of the ancient city. The post mortem debate has gone

on to discuss how far the city was pushed (by various invaders from the fifth century) and how far it was already dying (as the central government bailed out city fathers and the church replaced some municipal functions).*° But much of the

importance of this debate is assumed by external historical approaches. For the classicist it is important that cities of the Milesian inheritance died, but it should not be over-looked that a settlement pattern even older and more central to Anatolia,

the chiliakomai, survived.*! To the Western and Islamic historian, the growth of 27. F.L. Estrada, Embajada a Tamorldn (Madrid, 1943), 56-57; R.G. de Clavijo, Embassy to Tamerlane 1403-1406, tr. G. Le Strange (London, 1928), 88-89. 28. C. Clark, ‘Urban population densities,’ Journal of the Royal Statistical Society, Series A, 114 (1951), 490-96.

29. Notitia dignitatum, ed. O. Seeck (Berlin, 1876), 229-43; J. Ball, A description of the city of Constantinople, as it stood in the reigns of Arcadius and Honorius (London, 1729); H. Inalcik, ‘Istanbul,’ E/?, the fifth-century regiones (two of which lay outside the city) are not, of course, identical with the nahiyes — for example Regio XI, with the Holy Apostles, did not, like the Nahiye VII, with the Fatih complex, reach the Golden Horn. 30. C. Mango, Byzantium. The Empire of New Rome (London, 1980), 60-87 puts the matter elegantly. On the debate, see A. Kazhdan and G. Constable, People and power in Byzantium: an introduction to modern Byzantine studies (Washington D.C., 1982), 15, 38-43. 31. Strabo, Geography, XII, iii, 39; D. Magie, Roman rule in Asia Minor, | (Princeton, 1950), 179-80; ‘‘It is not surprising that inland Pontus, remote and mountain-girt as it was, should have remained unaffected by Hellenism . . . Down to the time of the Roman conquest there prevailed the old Asianic system of domain-land . . . Both king and nobles owned fortified 273

distinctive varieties of the city are vital, but neither is the kind of settlement that should necessarily be sought in Byzantium. To literary Byzantines, what happened to their cities between the fifth and seventh centuries was perhaps less of a watershed than may now appear to us. The Greeks had a word, or rather two words, for what they had alrady identified as a cyclical phenomenon, which I reintroduce into a discussion begun by Strabo. The first term is common enough: synoikismos, which describes the sometimes artificial creation of a city from a number of settlements. Strabo explains how many

Peloponnesian chorai which had contributed to the Homeric Catalogue of Ships where subsequently synoikicised into poleis: ‘For example Mantineia in Arkadia was synoikicised from five demoi by the Argives.’>? The phenomenon is well known to anyone who has watched the creation of ‘new’ towns such as Telford and Milton Keynes in modern England. What is more interesting is what happened to synoiki-

cised Mantineia in 385 B.C. The new polis was dioikicised.** To Aristotle, Xenophon, Plutarch and Strabo, dioikismos indicates the reverse process: the disintegration of a city into its component parts.°4 This was not necessarily (as probably happened in the Mantinike) a retreat to the original founding villages, but the dismemberment of a city into physically and functionally distinct, and sometimes antagonistic, quarters on the same site — like the Constantinopolitan regiones and nahiyes. In about 1200, Sardis (now Sart) had, for example, a manned citadel, below which were found knots of settlement, totalling 10 ha. in extent, which lay respectively 700m., 200m., and 1,200m apart.*° Even the isolated peak of the citadel shared two settlements in 1304, when Tiirkmens joined the Byzantine garrison in mutual fear of the Mongols, with only a party wall to divide them.*° The late Byzantine town was dioikicised. It was not a haphazard process, for some knots of population now had specialised functions (monastic, maritime, military &c.), or their own religious identity, duplicating for themselves civic conveniences which had once been provided by and for the whole city. At its simplest, a Byzantine settlement was knotted round a plateia and what amounted to a parish church. But any intrusive Latin quarter had to establish anew such once shared buildings as baths, churches, courts, ovens, markets, fondacos, warehouses, loggias, caravansarays and walls.*’ strongholds which they used as residences, and around these were villages which served as

kine centres .. . In such a system cities, in the Greek sense of the word, were entirely 32. Strabo, Geography, VIII, iii, 2. 33. S. and Hilary Hodkinson, ‘‘Mantineia and the Mantinike: settlement and society in a Greek polis,’’ BSA, 76 (1981), 261-65. 34. Aristotle, Politica, 1311; Xenophon, Historia Graeca, V, ii, 7; Plutarch, Camillus, 9. 35. Wy Foss, Byzantine and Turkish Sardis (Cambridge Mass. and London, 1976), 76-89, 180 plan

36. Foss, Sardis, 121-24, anthologising Pachymeres. 37. Cf. J. Prawer, ‘Crusader cities,”’ in The medieval city, ed. H. A. Mishimin (essays in honour of R. S. Lopez) (Yale, 1977), 187. 2714

LATE BYZANTINE TOWN

There are a number of late Byzantine ‘new’ towns of various sorts. The castle-

and-vicus town, such as Mistra, can be, like Palaiochora on Aegina, a refuge settlement.°® This type does not always lend itself conveniently to a dominant minority, Frankish or Ottoman, who naturally secure an acropolis first. Thus on Naxos (fig.2), the circular Sanudo kastro of 1207-1383 still embraces the Latin cathedral and townhouses of the Catholic squirearchy who also have pirgoi in the country, but is throttled by an encircling Greek burgo — the Orthodox cathedral is almost suburban.°? Similarly the Ottomans commanded the castle on top of the great rock of Monemvasia after 1540, but surely found it irksome that the only effective access to it lies through the walled Greek lower town.” Less clear are the functions of the new fortified towns of the Strymon, some of which showed signs

of self-determination in the fourteenth-century, if only to keep out of Kantakouzenos’s wars. A place like Chrysoupolis grew by stages. Were its walls to keep strangers out or livestock in? Did local archontes keep country houses on their pro-

asteia, or (like their Nicaean predecessors) live in town; or both (as did their Trapezuntine contemporaries)?*! From the town planner’s point of view, Bouras is quite right in regarding such

places as haphazard. Compared with the breathtaking urban engineering of Pergamon (Bergama), or even Pergamene Assos (now Behram Kale), similar late Byzantine terraced sites, such as Mistra and Palaiochora, are largely innocent of planning. But socially and functionally, the late Byzantine town is anything but haphazard. To begin with, some Pontic towns, such as classical Trebizond, had never really been synoikicised. There the acropolis lay inland, but late Byzantine Koralla (Gorele) and Herakleia (Arakli) were turned inside out with a seaside citadel and inland walled settlement. Places like Kenchrina (Kalecik-Hisartistii) and Rhizaion (Rize) consisted

of a single (Greek) population, but performed four functions: castle, market, port and settlement, which lay quite apart.** The complication comes when more than one group pursues the same function. The Tiirkmen and Greek garrisons who were uneasy co-tenants of Sardis acropolis are paralleled in Smyrna, where emiral and Christian castle eyed each other across the bay. In fourteenth-century Samsun there seem to have been four distinct walled settlements each strung a bowshot apart across 38. N.K. Moutsopoulos, He Paleachora tes Aigines (Athens, 1962). 39. Fig. 3 is taken from A. Polychoniadis and C. Chadjimichalis, ‘‘Structure and elements of the physical environments in Naxos,’’ in Shelter in Greece, ed. O. B. Domanis and P. Oliver (Athens, 1974), 83-97. Aglaia Kasdagli, of the University of Birmingham Centte for Byzantine Studies, is preparing a study of late medieval Naxos. 40. Charis Kalliga, of University of London King’s College, is preparing a study of Monemvasia.

41. Angold, Nicaea, 108; F. Tinnefeld, ‘‘John Kantakouzenos and the towns of Thrace and Macedonia during the civil war, 1341-47,’ unpublished paper given at the University of Birmingham Byzantine Seminar, 8 March 1979. On fourteenth-century Macedonian towns, see e.g. A. Xyngopoulos, Ereunai eis ta Byzantina mnemeia ton Serron (Thessalonike, 1965); and Catherine Asdracha, La région des Rhodopes aux XIIle et XIVe siécles (Athens, 1976); Cantacuzenus, Bonn ed., 541].

42. Bryer and Winfield, Pontos, sections XIV, XV, XXIV, XXV. 275

the dreary shore, three with the same commercial function: Turkish Samsun, Italian Simisso, Greek Aminsos and finally a settlement in the old acropolis of Amisos.” But whereas in Macedonia the new walled towns seem to have offered refuge, the Pontic instinct was to abandon such places in times of trouble and to melt into the valleys. Faced with a Turkmen threat in the 1450s, Chalkokondyles claims that the Grand Komnenos John IV found Trebizond reduced overnight to a population of

50; the remainder, ‘women, cowards and traitors’ all, had prudently gone to ground.” It was the same in 1832 when Osman pasa Haznedaroglu and an army of 7,000 concluded the ‘War of Siirmene’ by taking that place in style and carrying off the entire population they found there: three old women and an infant.* In some respects, Trebizond was but the market quarter for the subtended Trikomian, Matzoukan and Gemoran valleys. The 1461 Morean defter reveals an ‘urban’

| proportion of about 14%; by the same definition, the Pontic town population was then about 10% of the total, but that is a more notional figure.“ By the 1450s there was a residue of inhabitants of some Pontic towns like Trebizond who could not melt away in times of trouble, for they had nowhere to go. They included Latins and Armenians who were quite unrepresented in the valleys — the latter largely arrived after Timur’s sack of Sivas in 1400. By ca. 1486 they numbered 1,048, or almost 19% of the urban population.*’ But this may not conflict with Chalkokondyles’s figure of 50 brave citizens of the Grand Komnenos who stayed to face the Turkmens, for there were several Trebizonds. Late Byzantine Trebizond had quite distinct social and functional regiones, which translated fairly well into an Ottoman scheme of mahalles. These quarters lived cheek-by-jowl under a number of different franchises and jurisdictions: imperial, canon, Armenian, Venetian and Genoese (fig. 4). Tafur’s estimate of a total popula-

tion of 4,000 in 1438 does not look a bad guess and is line with later Ottoman figures.*® Of these a number were in imperial service: there were 63 known (as opposed to theoretical) offices of state, involving families totalling maybe 300 souls, before any garrison is considered. There were 16 or 17 active Orthodox monasteries,

two Latin houses and one or two Armenian ones, together with the staff of the cathedral and the administration of their lands — some within the city. There were 43. Bryer and Winfield, Pontos, section IV. 44. Laonicus Chalcocondyles, Bonn ed. (1843), 462-66. 45. Public Record Office, London, FO 524/1; cf. A. Bryer, ‘“‘The last Laz risings and the downfall of the Pontic derebeys,’’ Bedi Kartlisa, 26 (1969), 203. 46. N. Beldiceanu and Iréne Beldiceanu-Steinherr, ‘‘Recherches sur la Morée (1461-1512),”’ Soforsch, 39 (1980), 42; M.T. Gokbilgin, ‘‘xvi. ytizyil baslarinda Trabzon livasi ve Dogu Karadeniz Bolgesi,’? BT7TK, 26 (1962), 295.

47. Lowry, Trabzon, 48; Bryer and Winfield, Pontos, section XX, s.v. 48 Kaymakli. 48. See n.14. In both Trebizond and Thessalonike the Greek quarters at least kept the identity of their old names as Ottoman mahalles: see Lowry, Diptycha, 2 (1980-81), 260, 264, and the table of Byzantine and Ottoman quarters of Trebizond in Bryer and Winfield, Pontos, section XX. In Istanbul nearly 90% of the 219 sixteenth-century mahalles were named after the founder of the local mosque: Inalcik, ‘‘Instanbul,’’ E/. In the Morea, however, quarters seem to have been named after the first name inscribed in the defter: Beldiceanu and Beldiceanu-

Steinherr, SOforsch, 39 (1980), 42.

276

LATE BYZANTINE TOWN

Greek archontes who had town houses — of which there are hints in one unknown

and two identifiable ‘quarters. They would have been based primarily on hospitokastra on estates — some, such as that of the Tzanichitai of Tzanicha, up to 70 km. inland. But that their town houses were also fortified, or could at least be locked up, is suggested by the fact that they were themselves gated in ‘ta tén archonton hospitia’ of Trebizond during the uprisings of 1350. They needed such town houses in order to have a say in the central government. The fortifications of other quarters emphasised their automony: both Italian bases, three (perhaps five) monasteries, and the imperial citadel itself were walled.*° Besides Latins, Armenians, administrators, clerics, monks and archontes, there remains a residue in Tafur’s 4,000 who are called the mesoi and, simply, the ‘people’. The mesoi are a bourgeoisie, dependent upon commerce as much as land. Oikonomides has argued that the Italian merchant stations replaced, rather than stimulated, the growth of the late Byzantine bourgeoisie; that Italian capitalism in the Black Sea did little to arouse Byzantine mercantile instincts before it was too late, but that the emergence of a small but identifiable Byzantine commercial class, which had solme financial competence and political awareness, must be reckoned with before it was aborted in the fifteenth century. Neverthless, no one has ever been able to make a theme out of ‘Orthodoxy and the Rise of Capitalism’, for it is hard to imagine a fifteenth-century Byzantine beginning his Constantinopolitan account book, as did Giacomo Badoer, with the invocation ‘In the name of God

and of good profit.’° Was there a native merchant and middle class to reckon with in a place like Trebi-

zond? Italian sources, such as Badoer, and Greek, such as Bessarion, give startlingly different perspectives, like looking through either end of the same telescope. But Bessarion and Clavijo confirm the evidence of the Italian treaties, that the Trapezuntine bazaar and cardo — the ‘via imperiale’ — were indeed impressive and bustling, something on the lines of a description of the Constantinopolitan

bazaar in the 1430s. The St. Eugenios fair rivalled that of Thessalonike’s St. Demetrios. Trapezuntines of all conditions were aware of market forces and had their own currency: witness the extraordinary range of agricultural produce and manufactured goods that the English were able to buy in the Trapezuntine market in 1292, and the fact that a notary such as Sgouropoulos referred to the mechanics 49. Panaretos, ed. Lampsides, 70, 73; A. Bryer, ‘“‘The faithless Kabazitai and Scholarioi.’’ Maistor. Classical, Byzantine and Renaissance Studies for Robert Browning, ed. Ann Moffatt (Canberra, 1964), 309-28. Of archontic quarters, the court or nobility seem to have held the-St. Eugenios area during part of the civil wars of the fourteenth century, while there were two Chaldian archontic quarters: around St. John-Without-the-Walls (endowed by a group of Chaldians

in 1306); and in Kanitou, an area in the eastern suburb known in the tenth and fourteenth centuries and presumably named after the Kanis (Harsit) river and Kanites family, which | cannot now identify. See fig. 4, and O. Lampsides, ‘‘Topographika semeiomata. Kanitou,”’ Pontiaka Phylla, 2 (1937), 58. Walled monasteries were the H. Sophia, the Pharos and the Chrysokephalos, and (perhaps) St. Eugenios arid the Theoskepastos. 50. Oikonomidés, Hommes d’affaires, 130-31; Il libro dei conti di Giacomo Badoer, ed. V. Dorini and T. Bertelé (Rome, 1956). Cf. now, P. Schreiner, ‘‘Kupcy i tovary Priternomorjja. Fragment vizantijskoj kontorskoj knigi,’’ Byzantinobulgarica, 7 (1981), 215-19.

277

of maritime investment and insurance in the early fourteenth century: there is even a reference to price fluctuations in a horoscope of 1336.7! Nevertheless, it looks as if the Trapezuntine mesoi were relatively few in number and limited in scope. The defter of ca. 1486 identifies the occupations of only 40 Christian heads of household. Mostly priests, this admittedly tiny sample reveals only a butcher, bakers, candlemakers and the like (including, significantly, a drover and shepherds): no merchants of consequence are listed. The overland and overseas

trade had required formidable capital outlay and credit facilities which only the Italian cities could offer and in which only the Grand Komnenoi themselves could enter on an equal footing with Latins (by disadvantaging their own subjects). I can find only four examples of local Trapezuntine shipmasters, besides the emperors themselves. The paradigm of the local Trapezuntine merchant, such as Bessarion saw cross-legged in the bazaar stalls, was a certain John. A stout anti-Latin, John caught a Genoese ship to the Crimean in about 1340, exhibited firm anti-semitic and anti-Muslim feelings too, and ended up by hazard as a patron saint of Moldavia. But he was no more than a pedlar. What skimpy evidence we have makes it difficult to compare the bourgeoisie of Trebizond even with the evidently more agressive

mesoi of Thessalonike.° , ,

Nevertheless, there was more to a place like Trebizond than a collection of

disparate mahalles, and certainly a collective identity which made people take its name as their own. It was derived not so much from the elusive mesoi, as its ‘people’

who figure much more frequently. As in Thrace, the fourteenth-century ‘people’ of Trebizond exhibited an almost reckless spirit and successfully overturned every emperor, archon or Latin of whom they disapproved. There are signs that, like the Matzoukans, the ‘people’ of Trebizond had a communal instinct and perhaps even organisational distinction among the other jurisdictions amid which they lived. In hagiography, at least, there was a demarch, or mayor (once played by St. Eugenios).

They had no known guild, like Thessalonike’s Mariners, but did have a corps of nightwatchmen. We know of them from inscriptions on part of the lower walls which

seems to have served as a sort of fourteenth-century notice-board: they include maledictions. These nyktotalaloi — or night dellals — patrolled the walls of the 51. Clavijo, ed. Estrada, 76; Bessarion, ed. S. Lambros, Egkomion eis ten Trapezounta (Athens, 1916), 45-47, 56; Oikonomidés, Hommes d’affaires, 59, 97-100; Stephen Sgouropoulos on Alexios II in A. Papadopoulos-Kerameus, Analekta Hierosolymitikes Stachyologias, 1 (St. Petersburg, 432-33; the horoscope of 1336 in ed. S. Lambros, Neos Hellenomnemon, 13 (1916), 41.

$2. Lowry, Trabzon, 72; Oikonomidés, Hommes d'affaires, 85; S. P. Karpov, Trapezundskaja Imperija i zapadnoevropejskie gosudarstva v XIII-X V vv. (Moscow, 1981), 173-74; Ibn Battuta,

ed. Gibb, II, 486; F. Thiriet, Régestes des délibérations du Sénat de Venise concernant la Romanie, II] (Paris, 1960), No. 2532 of 30 July 1440; N. lorga, Notes et extraits pour servir a l’histoire des Crisades au XVe siéle, 1 (Paris, 1899), 539; Th. Mpougioukes, Akoluthia tou hagiou martyros loannou tou Trapezountiou (Jassy, 1819); A. Bryer, ‘‘The Latins in the Euxine,’’ XVe Congres International d’Etudes Byzantines, Rapports et co-rapports (Athens, 1976), 10-12; Lj}. Maksimovic, ‘‘Charakter der sozial-wirtschaftlichen Struktur der spatbyzan-

tinischen Stadt (13.-15. Jh.) ”, JOB, 31-32 (1981) (= XVI. Internationaler Byzantinistenkongress, Akten, 1/2), 184-85.

278

LATE BYZANTINE TOWN

pourtzes — or burg — in 1302. By ca. 1486 there were five recorded delldis, three Christian and two Muslim: nights in Trabzon are still made hideous by their plaintive whistles. Such institutions confirm that what Gregoras describes as the demos of Trebizond was a recognisable community, with citizens able to fend for themselves as a body.>? A nascent late Byzantine bourgeoisie is certainly there, but was to some extent the child and victim of the Italians and in a place like Trebizond does not appear to number great merchants. More striking, perhaps, is the sense of municipality, or demos, among the ‘people’ of the city. But the city remained largely for the convenience of its hinterland, manned on out-of-market-days by people who had no business in the country. It was essentially dioikicised and, like most Pontic towns, was really a mahalle of its own hinterland. As elsewhere, the Ottomans inheritied

the street plans and the regiones, strengthening the mahalle instinct. But they decapitated the imperial, archonic and clerical classes within the cities, presumably with the mesoi too. Then, by deportation of their own people, the Ottomans reimposed synoikismos, giving late Byzantine cities those civic conveniences, municipal buildings and public spaces once found in cities of the Milesian inheritance. They eventually even recreated a Greek bourgeoisie — this time with access to capital and markets which the Byzantine could never provide. But that is another story.

53. Fontes Trapezuntini, 1, ed. A. Papadopoulos-Kerameus (St. Petersburg, 1897; reprinted Amsterdam, 1965), 126; H. Grégoire, ‘‘Les veilleurs de nuit a Trébizonde,’’ BZ, 18 (1909), 490-99;

Lowry, Trabzon, 72; Nicephorus Gregoras, Bonn ed. (1829), II, 679-80. 54. In, e.g., R. Jennings, ‘Urban population in Anatolia in the sixteenth century: a study of Kayseri, Karaman, Amasya, Trabzon and Erzurum,”’ /nternational Journal of Middle East Studies, 7 (1976), 21-57; Leila T. Erder and Suraiya Faroghi, ‘“The development of the Anatolian urban network during the sixteenth century,’’ Journal of the Econonmic and Social History of the Orient, 23 (1980), 265-303; the same’s Towns and townsmen of Ottoman Anatolia (Cambridge,

1984); N. Svoronos, Le commerce de Salonique au XVIIIe siécle (Paris, 1956); and B.Demetriades, Topographia tes Thessalonikes kata ten epoche tes Tourkokratias, 1430-1912

(Thessalonike, 1983). , |

279

XI

THE QUESTION OF BYZANTINE MINES IN THE PONTOS: CHALYBIAN IRON, CHALDIAN SILVER, KOLONEIAN ALUM AND THE MUMMY OF CHERIANA

This article began as a Note, offering literary evidence for the working of the glittering black sands found on some Euxine beaches to relate to the analyses of two samples made by R. F. Tylecote in his “Iron sands from the Black Sea,” Anatolian

Studies, 31 (1981), 137-9.' But in preparing it I read Prentiss S. de Jesus’s lucid

study of The development of prehistoric mining and metallurgy in Anatola (Oxford, 1980),* and re-read Speros Vryonis Jr.’s pioneering article on “The question of the Byzantine mines,” in Speculum, 37 (1962), 1-17.% This explains why, while the Note has grown into a modest article, its conclusions have shrunk. For sources for Byzantine mining from the seventh to twelfth centuries, Vryonis had largely to fall back upon prospective and retrospective literary evidence: mostly legal before the seventh, and Turkish after the thirteenth century. He noted how very few direct references there are to mining in the central period and in Byzantine Anatolia thereafter, and while I can add a handful of further indications for Pontic iron, alum and mummy, I must discount his most important one for silver. Vryonis offered a choice of answers to his question: either that “Byzantium had no access to mines and therefore they did not appear in the sources;” or that Byzantines did mine, but “Byzantine sources simply do not mention this type of ordinary or common

matter.”* Vryonis favoured the latter solution, but the question remains largely where he left it twenty years ago, and his explanation was cited again recently. I think it questionable. Even leaving aside charters, which are obliged to treat banausic matters, Byzantine literary sources also do so. Within the conventions of Byzantine epistolography Niketas Magistros, for example, was quite as capable of giving a detailed description of iron mining and smelting in the tenth century as had

Apollonius of Rhodes in his own genre.° In fact the difficulties of relating Byzantine literary to archaeological evidence can be overrated. So the pessimist can set aside the prospective and retrospective evidence and argue that there is virtually no firm evidence that the Byzantines mined noble metals in particular (or, for that

‘Cf. his A Hustory of Metallurgy (London, 1976). For discussion and references I am grateful to

M. R. Broome, Archibald Dunn, Dr John Haldon, Bruce Lippard, Professor Victor Ménage, Dr Michael Metcalf, Susan Mossman, Professor R. E. F. Smith, David Winfield, and a meeting of the Royal Numismatic Society, London, 16 March 1982. I thank Mrs Gaye Bye for typing. 2B. A. R. International Series 74. Cf. his “Metal resources in ancient Anatolia,” Anatolian Studzes, 28 (1978), 97-102. 3Cf. his Byzantium: tts internal history and relatrons with the Muslim world (London, 1971), Collected Studtes, VI. “Vryonis, “Mines,” 17.

SNicétas Magistros, Lettres d'un Exilé (928-946), ed. L. G. Westerink (Paris, 1973), 65; Argonautica, I1, 1002-8; A. A. Gordus and D. M. Metcalf, “The alloy of the Byzantine miliaresion and the question of the reminting of Islamic silver,” Hamburger Bettrage zur Numzsmatzk, 24/26 (1970-72), 17.

134

matter, marble) after the seventh century.° The optimist can point to the abundance of Byzantine gold, and to a lesser extent silver, coin. But, so far, numismatic metallurgists have been unable to tell us with certainty where Byzantine gold and silver came from, and how far it was new, imported, or inherited recycled

stock.’ The sigillographer can point to the widespread and abundant use of lead seals in the central period, but cannot answer whether the rapid tailing off of the striking of molybdobulls in the thirteenth century was prompted by a dearth of lead or the spread of the Western fashion for sealing with wax.® The findings of scholars such as Tylecote and de Jesus cannot be used to answer Vryonis’s question either. But ores do not shift and de Jesus at least provides a guide to where in Anatolia Byzantines could have worked them most easily, as well as a clue to why it may take a while before Vryonis’s question can be answered. Mining

and smelting commonly take place in parts that archaeology does not commonly

reach. Even then, where excavation of a late Byzantine site, in Macedonian Siderokavsia, has encountered evidence of iron-working, it must be confessed that it

has not added much to Athonite charters for our knowledge of it. But mining in ®Vryonis’s evidence for Byzantine mines has been restated by R.-J. Lilie, Dre Byzantinzsche Reaktion auf die Ausbrettung der Araber (Munich, Mescellanea Byzantina Monacensta, 22, 1976), 258-62, who tabulates references to gold, silver, copper, iron and lead in the Balkans and Anatolia. The list uses secondary sources and is incomplete for late medieval mining in the Balkans, but shows that direct contemporary evidence for mines worked by Byzantines after the seventh century may be reduced to: GOLD. Anatolia: none. Balkans: Transylvania/Bulgaria in r1th c. SILVER. Anatolia: Amaseia (Amasya) in 12th c.; Argyropolis-Gimishane in 1gthc. Balkans: none. COPPER. Anatolia: Cyprus in roth c.; Eastern Anatolia in roth-reth c. Balkans: none. IRON. Anatolia: none. Balkans: Chalkidike in roth c. LEAD. Anatolia: none. Balkans: none. Infuriatingly, Lilie gives inadequate references; if that for Amaseian silver is to Abul Feda, it must be moved out of the Byzantine period. See Géographie d’A boulfeda, edd. M. Reinaud and M. de Slane (Paris, 1840), 383, cf. Vryonis, “Mines”, 7 and n. 35. This may refer to the galena deposit at Gumus: de Jesus, Anatolia, 263. The evidence for the silver of Argyropolis-Gimtshane is discounted below. I have also eliminated evidence for silver mining in Armenia, the.Caucasus and the Tauros under Arab domination. In his “The circulation of silver in the Moslem East down to the Mongol epoch,” Harvard Journal of Astatic Studies, 2 (1937), 291-310, R. P. Blake postulated that a cause for a Muslim silver famine from the tenth to thirteenth century was that these eastern Anatolian silver mines had been lost to the Byzantines, but there is no evidence that they actually worked them then: indeed the indications are against, for while Basil II’s conquests in Armenia, the Caucasus and Taron (where mines had been attested) come largely after c. 1000, P. Grierson finds that the bulk, if not all, of his silver coinage comes before that date: “The gold and silver coinage of Basil II,” American Numtsmattc Society Museum Notes, 13 (1967), 167-87. ’For the state of research, see C. Morrisson, ‘Projets de recherche et nouveaux aspects dans les

sciences auxiliaires,” XVI. Internationaler Byzantinistenkongress, Akten, I, i = Jahrbuch der Osterretchischen Byzantinistik, 31/1 (1981), 106; and D. M. Metcalf, “Analyses of the metal contents of medieval coins,” in Methods of chemical and metallurgical investigation of anctent coinage, edd. E. T. Hall and D. M. Metcalf (London, Royal Numismatic Society Special Publication 8, 1972), 383-434. According to Erica C. Dodd, Byzantine silver stamps (Washington D.C., Dumbarton Oaks Studies 7, 1961), 21, 32-33, the practice of five-stamp hallmarking Byzantine silver began breaking down from 621 and ended in 641-51. ®Cf. N. Oikonomidés, “Quelques remarques sur le scellement 4 la cire des actes impériaux byzantins (VIIle-IVe s.),” Zborntk of the Philosophical Faculty of Belgrade, 14 (1979), 123-28. There is a cluster of lead deposits in north-west Bithynia which would have been lost about 1300: see de Jesus, Anatolia, map 12; cf. Tylecote, Metallurgy, 76. ®°A_ Guillou, Siderokausta I. Recherches d'anthropologie Byzantine. La ctuilisatton matérielle a Siderokausta. Les fouilles de 1976. Rapport préliminazre (Paris, E.P.H.E.S.S., 1977), the same’s Stderokausta II. . . Les fouslles de 1978 (Paris, 1979); E. D. Clark, Travels in vartous countries of Europe Asta and Africa, Part 2, Section 3, III (London, 1816), 411-13; and R. Kocaer, Osmanle Altmlart. Gold cotns of the Ottoman empire (Istanbul, 1967), 70-4 nos 49-65, 82-3 nos. 97-8, g1-2 nos. 131-2 for gold coins of Sidrekapsi 1520-74.

THE QUESTION OF BYZANTINE MINES IN THE PONTOS 135 the late medieval Balkans entered the orbit of Bohemian and Saxon techniques and regulations, which are a case apart. '° It is time, however, to re-open Vryonis’s question in Anatolia, choosing its area which is richest in minerals and remained Byzantine — or rather Trapezuntine — to the last. The metallurgical wealth of the Pontos, from the stinking arsenic mines of

Mt. Sandarakourgion"’ in the west to the still active copper workings of the Mourgoule (Murgul) River to the east, is well known. Copper illustrates our problem: it is abundant, but medieval evidence for its mining and working falls to pieces in the hands.'? So I have concentrated on minerals for which there is more substantial literary evidence: Chalybian iron, Chaldian silver, Koloneian alum — and the mummy of Cheriana, for which we have just a toponym.

I Chalybian Iron Tylecote’s first sample of iron sands from the Black Sea comes from the mouth

of what is now named the Kirazlik River and nearby village, 42 km. west of Trebizond (Trabzon). The river and place are victims of the wholesale destruction

of ancient names in these parts in the last generation, which the Roberts have deplored most eloquently.’* Older inhabitants and maps know Kirazlik as Kiregon, Cf. N. Beldiceanu, Les actes des premiers sultans conservés dans les manuscrits turcs de la Bibliothéque Nationale de Paris. 11: Réglements miniers, 1390-1512 (Paris-La Haye, Documents et Recherches VII, 1964). "Mt. ‘Realgar’: “in addition to the painfulness of the work, they say that the air in the mines is both deadly and hard to endure on account of the grievous odour of the ore, so that the workmen are doomed to a quick death”: Strabo, Geography, XII, iii, 40; tr. H. L. Jones, (London — Cambridge Mass., 1969), V, 450-1. Mt. Sandarakourgion lay in the district of Pompeiopolis (Tagképrii — but the early Byzantine site is evidently 8 km. east-north-east at Kizlar Kale, Zimballi, which would repay further investigation). Mt. Sandarakourgion appears to correspond better with Ovacik-Hasandegin than Giimiis which de Jesus, Anatolia, go-7 and map 11, prefers.

If we were limited to documentary sources, medieval evidence for Pontic copper might be limited to an old copper cauldron worth 8 aspers which was willed in Vazelon Act 79 of 1260, in F. Uspenskij and V. V. Beneshevich, Actes de Bazélon (Leningrad, 1927); 47; and to a couple of stewpots bought by Walter and Jak the cook for the English embassy which passed through Trebizond in 1292: C. Desimoni, “I conti dell’ ambasciata al Chan di Persia nel 1292,” Attz della Socteta Ligure di Storia Patria, 13, (1884), 598-669. De Jesus offers copper mines “said” to have been worked by Genoese (who are locally credited with much, most improbably mining) around Kerasous (Giresun) and Rhizaion (Rize) and medieval workings at Mourgoule (where a wooden shovel is dated by C-14 to 316 B.C.+). Mourgoule lay on the Byzantine and Trapezuntine eastern borders from the tenth century until the

1390s, when it passed to the Saatabago, which took furious beatings from the Mongols and Akkoyunlular before the Ottomans reached it by 1552. See, e.g., Constantine VII Porphyrogenitus, De

Administrando Imperio, ed. Gy. Moravcsik, tr. R. J. H. Jenkins (Washington, D.C., 1967), 220; N. Oikonomidés, Les listes de préséance Byzantines des [Xe et Xe stécles (Paris, 1972), 260, 269, 362;

J. F. Woods, The Aqquyunlu. Clan, confederation, emptre. A study in 15th/gth-century TurkoIranian politics (Minneapolis-Chicago, 1976), 48, 100-4, 151. In our forthcoming Dumbarton Oaks Study of The Byzantine monuments and topography of the Pontos, section XVII, David Winfield and I publish a late Byzantine church at Fol Maden, 22 km. due south of Kireson (Kirazlik), which appears to correspond with an unnamed site no. 55 of possible prehistoric or ancient copper workings in the Pontos in de Jesus, Anatolia, map 9. At nearby Yortun (Kurtin), in the Philabonites (Harsit) valley, which has a medieval castle, Hamilton found 6,700 men slaving in a copper mine, 34 of which are reported as being worked in the sancak of Trabzon later in the century: W. J. Hamilton, Researches in Asta Minor, Pontus and Armenia (London, 1842), I, 259; V. Cuinet, La Turqute d’Aste (Paris, 1890), I, 56-8, 68. Does all this add up to more than a possibility that Byzantines mined Pontic copper? '3“Cette éradication, cette manie de changement abolit un passé national, tout comme un bulldozer qui détruit une nécropole ou un édifice”: Jeanne and L. Robert, “La persistance de la toponymie Antique dans I’Anatolie,” in La Toponymie Antique (Actes du Colloque de Strasbourg, 12-14 juin 1975, travaux du Centre de recherche sur le Proche-Orient et la Gréce Antiques, Université des Sciences Humaines de Strasbourg, 4 (Leyden, n.d.), 62.

136

PONTIC SITES 2s

athe SEA

Rene Lee ee eee - kourgion: oe a re..... AMt.Sandara . i eg. woemB . Kiregon...'., Singion "Gam B o-oo uegon |

“Ne st Amasya p

' Kot yore | Argyria N CHALYBIA Kerasous *\Trebizond Fole Onn e orul

Pe Gumus$hane/Tzanicha

bt TFercan CHAL Oo. e inir *®Baybdburt

Koloneiae eMumya

) 50 100 km.

CHERIANA

G oe Erzincan Sivas umugakar Boanrate’ iL. Map of Pontic Sites

Hamilton’s “Kerasoun”, which probably reflects one of the most ancient survivals on

this shore: a memory of the Kerasous, three days west of Trebizond, where Xenophon reviewed his Ten Thousand and found that they actually numbered 8,600."* Were it not for the fact that the toponym also appears in the anonymous pertplus, geographers would perhaps ignore Xenophon’s statement of distance and ascribe this “Old”, or “Second”, Kerasous to the better-known Kerasous of Giresun, 115 km. further west.'° But it seems to have been Xenophon’s Kerasous, unmentioned in any medieval source, that was remembered at Kireson, near Vakfikebir, until it became Kirazhk. Tylecote’s second batch of sands comes from beaches scattered from just west of Oinaion (Unye) to just west of Kotyora (Ordu), including one opposite Arrian’s

“Island of the Cilicians” (Hoynat) on Cape Jason (Yasun): there may be hagiographical evidence for iron-working in the area.'® This is, of course, the shore of Chalybia: as its very name proclaims, so synonymous with iron that some have doubted whether the Chalybians were an ethnic unit at all. Most recently de Planhol has maintained the non-geographical nature of Chalybia, but ignores the medieval evidence which reveals that Chalybia survived by name as a thirteenth-

century district, and a fourteenth-century client emirate, of the Empire of Trebizond, precisely where Jason and the Argonauts were supposed to have found ‘*Xenophon, Anabasis, V, iii, 2; J. M. Kinneir, Journey through Asta Minor, Armenta and Koordistan in the years 1813 and 1814 (London, 1818), 333-4; Hamilton, Researches, I, 250-1; Chrysanthos (Philippides, Metropolitan of Trebizond), “He ekklesia Trapezountos,” Archeton Pontou, 4-5 (1938), 85. ** Anonymous periplus, 36, in A. Baschmakoff, La synthése des périples Pontiques (Paris, 1948), 122-3; Bryer and Winfield, Pontos, section XVII. ‘SArrian, 23, in Baschmakoff, Synthése, 96-7; Bryer and Winfield, Pontos, section XI. The monastery of St. Phokas held an iron-rich estate at Tzampourou (= ? Cam Burunu on Cape Jason): see F. Halkin, Auctartum Biblothecae Hagtographicae Graecae (Brussels, Subsidia Hagiographica 47, 1969), 194 no. 2047t; A. Bryer, “Greeks and Tiirkmens: the Pontic exception,” Dumbarton Oaks Papers, 29 (1975) = The empire of Trebtzond and the Pontos (London, 1980), Collected Studies, V, 124 N. 82.

THE QUESTION OF BYZANTINE MINES IN THE PONTOS 137 it.” In a famous passage, Apollonius of Rhodes reported that the Chalybians knew nothing of agriculture or cattle-breeding, “but they cleave the iron-bearing land and exchange their wages for daily sustenance; never does the morn rise for them without toil, but amid black sooty flames and smoke they endure heavy labour.”"® Strabo’s remains a recognizable description of Chalybia: “Upon the whole, the seaboard in this region is extremely narrow, for the mountains, full of mines and forests, are situated directly above it, and not much of it is tilled. But there remains for the miners their livelihood from the mines, and for those who busy themselves on

the sea their livelihood from their fishing . . .”" Tylecote found that about half the Chalybian, and only about 26% of the Kireson (Kirazlik) sands sampled were magnetic, from which iron can be simply wrought, and indeed I have no literary evidence for iron-making in the meaner sands east of Chalybia. Is iron ore easier to sift from sands than to dig? Niketas Magistros, exiled somewhere between Granikos (Kocabas) and the Sea of Marmara in 928-46, describes how local iron-workers preferred to take the nodules from the sands of a river mouth on the Hellespont.” But Tylecote’s operator was unable to

concentrate the magnetite of his first sample by panning. The Chalybians had evidently mastered the knack when the Castilian ambassador Ruy Gonzalez de Clavijo found them toiling on the beach at Oinaion on 7th April, 1404. The place was then in the hands of a Trapezuntine archon, a Melissenos, who paid tribute to Timur, but was overlooked by a Turkmen settlement in Galeoglu Kale commanding

the route south, which probably belonged to a fourth party, the Chalybian emir. Nevertheless, “down in the port, standing by the sea-shore we noticed a number of houses that were smithies, for in these parts the sea casts up on the strand a certain fine black sand, and from this they sift out nodules that are worked up into iron bars.”?!

When Hamilton arrived at Oinaion in 1836, the Chalybians were still hard at

work, but had shifted a little inland, leaving the beach to the fishers. Their methods do not seem to have advanced much since Apollonius of Rhodes and Strabo knew them there. They dug the ore “by scraping up the soil. . . with a mattock, and collecting small nodular masses, which I understand is the form in which it is universally found in this district . . . The mode of extracting the tron is however very slow and laborious, the ore is smelted in a common blacksmith’s forge,” in which (to use English measures) 405 Ib. of the rude material, heated by 675 lb. of charcoal, yielded only 134 lb. of iron. “The blast of the furnace is kept up for twenty-four hours, during which time the mass must be constantly stirred, and "7X. de Planhol, “Geographica Pontica: (11) Les Khalybes. Nom de peuple ou qualicatif professionel?” Journal Asiatique, 251 (1963), 300-9; Michael Panaretos, Pert ton Megalon Komnenon, ed. O. Lampsides (Athens, 1958), 63, 73, 80; Joseph Lazaropoulos in Sborntk istochnikov po istorit Trapezundskoy imperit, ed. A. Papadop(o)ulo(s)-Kerameus, I (all published) (St. Petersburg, 1897; reprinted Amsterdam, 1965), 61; A. Bryer, “The estates of the Empire of Trebizond. Evidence for

their resources, products, agriculture, ownership and location,” Archeton Pontou, 35 (1979) = Collected Studtes VII, 395, n. 1.

'® Argonautica, II, 1002-8; cf. I, 1323, Il, 375, IV, 1475; tr. R. C. Seaton (Cambridge Mass.-London, 1961), 171. '®Geography, XI, iii, 19; tr. Jones, V, 402-3; D. Magie, Roman rule tn Asta Minor to the end of the third century after Christ (Princeton, 1950), I, 179, 186; II, 1068-9, 1213. 20Fd. Westerink, 31, 65. 21“F en el puerto junto conla mar auia vnas pocas de casas de ferrerfas; e en aquel derecho dela

caua, echaua el mar vna arena negra, menuda, e allegauan la, e fazian della fierro”: Embajada a Tamorldn, ed. F. L. Estrada (Madrid, 1948), 73; tr. G. Le Strange, Ruy Gonzales Clavijo, His embassy from Henry III of Castille to Tamburlaine the Great at Samarkand, 1403-1406 (London, 1938), 108.

138

the scum and scoria raked off, after which the melted iron is found at the bottom, which, from the specimen I saw, appeared of very good quality.” The tenth-century Marmaran iron-workers noted by Niketas Magistros had collected their iron sand, “then washed it in plenty of water; they roasted it dry by

fire, and, having smelted it in a furnace, heated it anew to be worked by the smiths.”** Tylecote’s Chalybian sample was separated and heated to 1100-1200°C

for 0-5-3 hours over a period of about eight hours. These treatments go far to explain why the mid-Byzantine “Farmers’ Law” and other sources indicate how precious iron-shod tools then were.** The problem was not finding the ore so much as fuelling it to an adequate heat. Iron-working and forests are, as in the Pontos, co-essential. The ore is too scattered and its working too spasmodic for the state to

control it, but mining and charcoal duties are necessarily linked in the late Byzantine and early Ottoman Empires.” Charcoal-burners in the English New Forest today reckon to reduce their timber to about one quarter. Such factors explain the dismay with which the early seventh-century villagers of Mossyna, north-

west of Ankara, greeted St. Theodore of Sykeon’s decision to take away the iron cage, which they had constructed for his mortification by pooling their iron-shod agricultural implements, and why they insisted that he accept a wooden substitute. By Hamilton’s Chalybian standards, even the 161 |b. of iron belts with which the Holy Man was corseted, would have called for 2-2 tons of rude material, 3-6 tons of charcoal and 14°4 tons of timber to produce: Theodore’s epithet of “sederophagos”’ was hard earned—by his villagers.2° Of course medieval smiths, like those of Mossyna, reworked existing stock, but there is a limit to how many times ploughshares and pruning-hooks can be turned into ascetics’ cages and back again. All this is little indication of the availability of 1ron in the Byzantine Empire. The timber and beaches of Chalybia could always provide it, but villages in less fortunate areas may hardly have qualified for the Iron Age. On the other hand the armouries of Constantinople itself were capable of producing numbers of complex bronze, iron and steel weapons at short notice — for example for the Cretan expedition of 949. ?’

II Chaldian Silver If in 1836 it took 50 Ib. of charcoal and go lb. of rude material to yield a pound of Chalybian iron, Hamilton’s equivalent figures from Chaldia show that it took, very roughly, 260 tons of timber to produce about 65, tons of charcoal to roast about 1°8 tons of argentiferous (galena) lead to yield 34 lb. of silver and, on the seventh day of poisonous firing, one pound of Chaldian gold. At 0°84% of silver the ore was exceptionally rich, but this was at a time when the Chaldian mines were failing and yielded only about 68 |b. of silver and 2 lb. of gold a year. In 1711, when they were 22Hamilton, Researches, I, 274-5; Cuinet, Turquie, I, 17, 88. 23Ed. Westerink, 65. 24Cf. the “Farmers’ Law” in W. Ashburner, Journal of Hellenic Studies, 30 (1910), 85-108; 32

(1912), 86-g5, articles 22, 62; J. W. Nesbitt, Mechanisms of agricultural production on estates of Byzantine prakizka (University of Wisconsin, unpublished Ph.D. dissertation, 1972), 112-14; M. Kaplan, “Quelques remarques sur les paysages agraires Byzantins (Vléme siécle—milieu Xléme siécle),” Revue du Nord, 62 (1980), 155-76; H. Magoulias, “Trades and crafts as viewed in the lives of

the Saints,” Byzantinoslavica, 37 (1976), 16-17, 22-3 (for blacksmiths); T. Tueoteoi, “Le travail manuel dans les typtka Byzantins des XIe—XIlle siécles,” Revue des Etudes Sud-Est Européennes, 17 (1979), 455-62; and my forthcoming “Digging in Byzantium,” in a Festschrift for Thor Shevchenko. 28See J. F. Haldon, Tagmata (forthcoming), chapter V, n. 197; de Jesus, Anatolza, 100; Tylecote, Metallurgy, 41. 26A. J. Festugiére, Vie de Théodore de Sykéon (Paris, 1970), 25-26, 31. 27Constantine VII Porphyrogenitus, De Caerimonits aulae Byzantine, 1, C.S.H.B., 664-78.

THE QUESTION OF BYZANTINE MINES IN THE PONTOS 139 reaching their peak, they produced about 1,000 Ib. of silver and 120 lb. of gold; by contrast the Gold Coast was then exporting about 3,500 lb. of gold in a good year. It is a moot point whether Chaldian silver or charcoal ran out first; not surprisingly Inner Chaldia, where a nest of Greek villages provided both charcoal and mining services in lieu of harac for almost two centuries, remains largely denuded of trees today. *°

There is no need to rehearse the history of later Greek Chaldian mining under the Ottomans here, but the assumption that it antedates the Ottoman conquest of Ardasa (Torul) in 1479, must be questioned. Even today the town of Giimiishane has the agreeable air of a Turkish Dawson

City. It reminds travellers, and those who should know better, of the earlier Argyropolis and of boom-days under Greek “archzmetallourgor’ when it was the

mining capital of the Ottoman Empire. But this Gimushane was built only a century ago at the point where the road and telegraph line ran most conveniently through the Kanis (Harsit) valley for the export, not of silver, but, of apples. We must begin again. Strabo mentions Chaldian silver mines, closed before his time, but seems to

confuse them with Chalybian iron-workings.* Thereafter the alleged literary evidence for medieval Giimiishane (Argyropolis) and its silver mines rests on four

accounts. In 1293 Edward I of England's envoy, Nicholas of Chartres, visited ‘“Dymesho” (which has been taken to be Gimiishane). In 1294 Marco Polo noted a silver mine on the Trebizond-Tabriz route in, or near, Paperth (Bayburt), and in at least one version, an abundant silver mine at “Argiron” (which has been taken to be Argyropolis). In 1332 come two references. On his way from Amasya to Erzincan, Ibn Battutah visited the then Turkish silver-mining town of “Kumish” which was in the hands of Eretna, while the Moroccan geographer, al Umari, reported mines

near Badhert (Bayburt) and at “Gumush Saray” (which has been taken to be Giimishane). That there were late medieval silver mines near Bayburt is not in question; Maden Bayburt and Bayburt were Seljuk mints in 1289-91 and 1288-go respectively, but this largely Armenian town was continuously in various Muslim hands until 1829. But Marco Polo’s “Argiron” is in fact Erzincan, and “Kumish” or “Gumush Saray” is probably Giimtisakar, on Ibn Battutah’s route. This place may

in turn correspond to one or more of the Seljuk mints respectively named Gimiisbazar, Yamisbazar, Maden Masbazar, Giimiis and Maden Bazar, which were

active in the period 1257-88. I do not know where “Dymesho” is, but it cannot

28Hamilton, Researches, 1, 234-38; “Journal du pére Monier d’Erzeron a Trébizonde,” in Lettres

édifiantes et curteuses, II (Lyon, 1819), 373-80; A. Bryer and D. Winfield, “Nineteenth-century monuments in the city and vilayet of Trebizond: architectural and historical notes, Part 5,” Archeton Pontou, 30 (1970), 324-50. 29Cf. W. W. Smyth, A year with the Turks, or sketches of travel in the European and Astatic dominions of the Sultan (New York, 1854), 105; A. Bryer, “The Pontic revival and the New Greece,” in Hellenism and the First Greek War of Independence (1821-1839), ed. N. P. Diamandouros et. al. (Thessaloniki, 1976) = Collected Studies, XII, 177-78. But charcoal-burning for localised smelting can be no more the scapegoat for the deforestation of Anatolia than the goat itself: see W. C. Brice, “The Turkish colonization of Anatolia,” Bulletin of the John Rylands Lebrary, 38 (1955-56), 27-35; G. H. Willcox, “A history of deforestation as indicated by charcoal analysis of four sites in eastern Anatolia,” Anatolian Studies, 24 (1974), 117-343°Geography, XII, iii, 19.

140

represent Giimiishane either, for Nicholas of Chartres hired a ship there.*' The identification of any of these places with Argyropolis or Gimushane would in any

case have been anomalous, for I propose the following sequence: that what Byzantines and Trapezuntines called Tzanicha fell to the Ottomans in 1479, that silver and gold mines were not opened there until 1546, that a mint of Tzanicha

operated thereafter until after 1574 and before 1595, that a new town called Gumiishane was built to serve the mines before 1598, and that the name Argyropolis was not concocted by local Greek schoolmasters as a Hellenisation of Gumtshane until about 1846, when the silver mines were in their last gasp. There is no ancient

or medieval Giimushane or Argyropolis. This should clear away the surviving literary evidence that Byzantines or Trapezuntines mined noble metals in Anatolia after the seventh century.

Tzanicha is the stupendous rock and castle topped by two, probably fourteenth-century, painted churches, which can be glimpsed 2 km. north-west of modern Giimiishane. It overlooks the river Kanis and the secret, wrecked, town now called Sulemaniye or Eski Gumushane, briefly Argyropolis, where for over two centuries the air was never clear of the fumes of roasting galena.

As de Jesus shows, silver has been mined at unknown periods in the area between two of the three late medieval centres of Chaldia: Tzanicha and Torul, north-west of Kovans. Torul is an area name which has now settled on the defile and town of Ardasa, but which was in the fourteenth century probably localised at Panaretos’s Golacha (Colagana Kale), Clavijo’s Dorileh: the Spaniard turned south before reaching Tzanicha.** The Fatih prudently skirted this mountainous corner in 1461 and only sent his son to settle matters with the last vestige of the empire of

Trebizond, which had become the pocket principality of Torul, in 1479. Torul would have embraced Tzanicha, the ancient and second medieval stronghold of 3*Desimoni, “Chan,” 552-3, 572, 594-5; The book of Ser Marco Polo the Venetian concerning the kingdoms and marvels of the East, tr. and ed. H. Yule and H. Cordier, I (London, 1929), 46, 49 n. 3; Marco Polo, The description of the world ed. A. C. Moule and P. Pelliot, I (London, 1938), 21: “The others are Argiron which zs great, [and] a very great quantity of stluer is dug there... .”, 22; II, vi: ‘“. , . & in quodam castro quod uocatur paperth est maxima argentera & inuenitur hoc castrum eundo de trapesunda in thauris.”” It is presumably one of the three Turkish silver mines in Simon de Saint Quentin, Hestozre des Tartares, ed. J. Richard (Paris, 1965), 68. Voyages d'Ibn Batoutah, ed. C. Defrémery and B. R. Sanguinetti, II (Paris, 1949), 293; The travels of Ibn Battuta, tr. H. A. R. Gibb, II (Cambridge, 1962), 436-37 (in fact a visit to the future site of Gimughane would have involved Ibn Battutah in a considerable detour, while the ‘“Kumish” he did visit lay not in, but two days away from what are evidently the Pontic Alps). Al-Umart’s Bericht uber Anatolien in seinem Werke Masalzk alabsar ft mamaltk al-amsar, ed. F. Taeschner (Leipzig, 1929), 20 (referring to A.H.733). Vryonis, ‘“Mines,”’ 8-9; A. Bryer, “The fate of George Komnenos, ruler of Trebizond (1266-1280),” Byzantenische Zettschrift, 66 (1973) = Collected Studies, IV, 347 and n. 76; Lilie, Reaktzon, 259; R. Drews, “The earliest Greek settlements on the Black Sea,” Journal of Hellenic Studies, 96 (1976), 28, most recently repeats that “In Marco Polo's time the town was called Argyropolis."” Giimtigkar (Maden) will be discussed in David Winfield’s chapter on “Routes” in Bryer and Winfield, Pontos. It may be the Ma-adin an-nuhas referred to by al-Muqaddasi in E. Honigmann, “Une itinéraire arabe a travers le Pont,” Annuaire de l’Instztut de Philologte et d'Histozre Ortentales, 4 (Brussels, 1936), 263, 266; cf. J.

von Hammer, Hustotre de empire Ottoman, tr. J. J. Hellert, IV (Paris, 1836), 211, 435; H. van Lennep, Travels in little-known parts of Asta Minor: with illustrations of Biblical literature and researches tn archaeology (London, 1870), 216-17; E. de Zambaur, Die Munzpragungen des Islams. Zettlich und ortlich geordnet. (Wiesbaden, 1968), s.v. Hacci Kéy. The place seems to have escaped de Jesus, Anatolia. 1 am grateful to Mr M. R. Broome for information on Seljuk mints. On Bayburt, the conclusions of D. Winfield, ‘A note on the south-eastern borders of the empire of Trebizond in the thirteenth century,” Anatolian Studies, 12 (1962), 163-72 are emended in Bryer and Winfield, Pontos, section XXVIII. 32De Jesus, Anatolza, maps 13, 15; Panaretos, ed. Lampsides, 74, 77, 78; Clavijo, ed. Estrada, 80; Clavijo, tr. Le Strange, 117.

THE QUESTION OF BYZANTINE MINES IN THE PONTOS 14! Inner Chaldia: it is probably garbled as Thia in the Jtzneraria and is Procopius’s and Panaretos’s Tzanicha, fourteenth-century seat of the Tzanichites family. * After 1479 Canca (to use Tzanicha’s Turkish spelling) entered the ndahiye of

Kovans (Kouazi, Lerin), sancak of Bayburt, pashaltk of Erzurum; the Murathanoglu family ran it from Keci Kale (?Mesochaldia), the third centre of Chaldia. Neither Canca nor Gumushane figure in the defters of 1516, 1530 or 1591, but Canca may be represented in them by a place called Kaledere. At all events a firman of Sileyman I (1520-66), dated Saban 953/October 1546 is evidence for a new mint established at Canca, when gold and silver had been found there: then a small place, it presumably adopted the name of Siileymaniye. Gold and silver coins

bearing the mint name of Canca (not to be confused with the Bosnian mint of Canica) are dated 1520, 1566 and 1574: the anomaly is explained by the fact that almost all Ottoman coins of the period seem to bear the date of a sultan’s accession. Finally, writing in 1598, Saraf al-Din states that a populous town called Gimtshane had been built in his day to serve silver mines near “Vourla” (Canca?), in Kovanis (Kanis, Kovans), which was a Greek-speaking area.* This appears to be the earliest reference to the newly-coined name of Gumushane, which Evliya equates with Canca in the seventeenth century. But Gimushane was good enough for even the Greek priests of the town in the eighteenth century until a new breed of schoolmasters taught them better. * But disposing of the literary evidence for ancient and medieval silver mining in

Chaldia leaves one with a formidable residue of actual silver in late medieval Trebizond. For example in 1292 the English spent 708 aspers in Trebizond on

33Panaretos, ed. Lampsides, 70; K. Miller, Jtznerarza Romana (Stuttgart, 1916), col. 681; Procopius, Buzldzngs, III, vi, 26. On the campaign of 1479 (missed by F. Babinger, Mehmed the conqueror and his time, who did not use Tursun, even in the most recent edition [Princeton, 1978]), see H. Inalcik, “Mehmed the conqueror (1432-81) and his time,” Speculum, 35 (1960), 408-9; but not (paradoxically) Tursun Beg, The history of Mehmed the conqueror, tr. H. Inalcik and R. Murphey (Minneapolis-Chicago, 1978). There is an echo of it in A. K. Sanjian, Colophons of Armenian manuscripts, 1301-1480 (Cambridge, Mass. 1969), 325 of 1480, colophon 1: Mehmed “marched into the canton of Xaxt’ik’ (Chaldia, not Chaneti, as identified in 427], and set fire the city called Babert’”

(Bayburt).

34]. Miro$lu, Bayburt Sancag: (Istanbul, 1975), 86-93, esp. 89 and maps possibly identifying Kaledere with Canca, but it is difficult to see how a place with a population of 15 Muslims and 12 Christians, producing wheat, barley and melons in the defter of 1591 could become §araf al-Din's great town producing silver by 1598: see Chéref-nameh ou Fastes de la nation Kourde par Chéref-ou'ddine, prince de Brdlis, 1, i (St. Petersburg, 1868; reprinted Farnborough, 1969), 187, 189 and 549 n. 365. Bryer and Winfield, Pontos, section XXII, follow M. P. Borit (szc: his real name was P. Briot), “Identification of Mount Théchés of Xenophon,” Journal of the Royal Geographical Soctety of London, 40 (1870), 464, in ascribing the site on the Kanis between Tzanicha and Kovans at Zindanlar Arazi, still called Murathanofgullari, to Murat IV (1611-40), but Saraf al-Din reveals that it must be named after the sixteenth-century family. Panaretos, ed. Lampsides, 74; H. Inalcik, s.v. “Dar alDarb, EI?, II, 118b, using (he kindly tells me) the Basvekalet Arsivi, istanbul, Fakete no. 288, for the unpublished firman; for the coins, see $. Sevket, Trabzon tarthz, I (Istanbul, 1294/1875,-76), 113;

Kocaer, Osmanl: Altinlart, 47 (where Canica and Hance, but not Canca, are equated with Giimiishane), 68 no. 41 (gold of 1520), 80 no. 8g (gold of 1568), 88 nos. 116 and 117 (gold of 1574); H. Etem, Meskukat: Osmantye (Istanbul, 1334/1915-16), no. 872 (gold of 1520), no. 1110 (gold of 1566), nos. 115-16 (silver of 1566), nos. 1256-66 (gold of 1574), nos. 1267-78 (silver of 1574). The mint does not appear to be commonly found. 35Evliya Celebi, Seyahatndmes:, 1 (Istanbul, 1314/1895-6), 566: “The Ottoman mints in order

are... Erzurum, Sivas, Canica, Giimiishane . . . These are in Anatolia.” Eighteenth-century Greek pilgrim transliterations of Giimiighane in colophons in E. C. Colwell, The four gospels of Karahssar, I

(Chicago, 1936), 33-93. On the place thereafter, see D. Oikonomides, “Argyropolis,” Archeton Pontou, 3 (1981), 145-94; Bryer and Winfield, “Nineteenth-century monuments. Part 3," Archeron Pontou, 30 (1970), 334.

142

silverware and repair, 565 aspers on two silver basins for the Ilhan, while Italian references to silver bullion are common enough from the 1290s and Italians were demanding hundreds of thousands of aspers of the Grand Komnenoi a few decades

later, which were presumably paid in bullion, if at all. Uncoined silver was circulated by the ingot of about 0:2 kg. (Mongol saumah, which conveniently became the Greek soma and Italian summo), found all over the Euxine with variants of the asper which was coined from it. One of the most notable achievements of the

Grand Komnenoi was their mint. They minted no known gold coins, although nomismata figure in charters, perhaps as monies of account. From 1222 they minted copper trachea and at some time between 1238 (the accession of the “Most

Fortunate” emperor Manuel I) and 1245 (the earliest apparent documentary reference to them) they introduced their own ‘“God-protected” silver aspers. Manuel’s early aspers appear to have been trachea too; KurSanskis suggests that the flat coins first described as kyrmanouelata in 1250 were introduced in the middle of

Manuel’s reign and continued to be minted after his death in 1263 until the accession of the Grand Komnenos John II (1282-97): aspers were minted thereafter until 1458. The Imerethian kings imitated the coins of Manuel and John II, even after 1297, calling both types k¢ermaneulz. * Why and where did the Grand Komnenoi obtain silver for coin? In 1243 the Seljuk state and its Trapezuntine ally went down to the Mongols at Kése Dag on the

Tzanicha-Erzincan road; in 1258 Hiilegii, first I/han, took Baghdad. Mongols demanded tribute in silver. Trebizond was certainly a Mongol tributary by 1253 and again probably in 1280, when his archontes delivered their hapless ruler George to the mercy of the second Ilhan, Abaga, at Tabriz: Ragid al-Din seems to have regarded Trebizond as (an albeit rebellious) tributary of the fourth I/han, Argun, in 1290—but he was also under the impression that Abaga’s wife Maria “Mougliotissa’’, daughter of Michael VIII Palaiologos, was in fact a Grand Komnene. It is tempting to regard the upsurge of Levantine aspers of various sorts—Cypriote, Cilician Armenian, Seljuk, Trapezuntine, Georgian and Euxine baracats—as a response to Mongol tributary demands and KurSanskis suggests that the kyrmanouelata might fit the I/hans’ bill. But one would not pay tribute in Christian coin and

those Georgian coins which, from 1247, bear the tamgha of the Ilhan are not kirmaneult. All the Levantine states had gone over to a silver system before the Mongols, except the Russian area which poses another problem. Whereas we have no evidence that Trapezuntine tributaries mined silver, Russian tributaries of the

3®Desimoni, “Chan,” 590-608; cf. Bryer, Collected Studies, VI, 389 n. 1, 391; G. I. Bratianu, Actes des notatres génots de Pera et de Caffa de la fin du tretzteme stécle (1281-1290) (Bucarest, 1927), nos. 186, 219; M. Balard, Genes et l’Outre-Mer 1, Les Actes de Caffa (Paris-The Hague, 1973), nos.

178, 175-77, 221; F. B. Pegolotti, La practica della Mercatura, ed. A. Evans (Cambridge, Mass., 1936), 25, 40-1; D. A. Zakythinos, Le chrysobulle d’Alexts III Comnéne empereur de Trébizonde en faveur des Vénitiens (Paris, 1932), 32, 62-4; N. lorga, Notes et extraits pour serutr @ l’histotre des crowsades au XV stécle, 1 (Paris, 1899), 272-5, 304; Blake, “Circulation,” 315 n. 74; N. Oikonomidés, Actes de Dtionystou (Paris, 1968), 60; Vazelon Acts 49 of 1245 and 64 of 1259; D. M. Metcalf and I. T.

Roper, “A hoard of copper trachea of Andronicus I of Trebizond (1222-35), Numtsmatic Circular (June 1975), 237-9; M. KurSanskis, “The coinage of the Grand Komnenos Manuel I,” Archeton Pontou, 35 (1979), 23-37; D. M. Lang, Studzes in the numismatic history of Georgia in Transcaucasia (New York, A.N.S. 130, 1955); the same’s “Georgia in the reign of Giorgi the Brilliant,” Bulletin of the School of Ortental and African Studies, 17 (1955), 85-90; and E. Schilbach, Byzantinische Metrologte (Munich, 1970), 192 (on the Trapezuntine soma). As an example of the very large quantities of aspers in circulation, M. Balard, La Romanie génotse (XIle-début de XVe stécle), 1 (Rome, 1978), 136-7 cites imperial indemnities to the Genoese totalling 133,649 aspers in 1315 and 274,279 aspers in 1338 (Basil actually paid 25,000).

THE QUESTION OF BYZANTINE MINES IN THE PONTOS 143 Golden Horde did not even have access to silver in their lands west of the Urals. * To this problem R. P. Blake had the ingenious answer that it was the Mongols themselves who provided the silver, which flowed West in great quantities from

China; Chinese silver contains antimony, giving the asper its white name and appearance. But Chinese silver comes largely from the Sung south, which the Mongols did not conquer until after 1250, and there are signs that much of that comes from Japan, while Metcalf and Gordus have found that Trapezuntine aspers are exceptionally short of antimony. * Perhaps one can get further by examining the aspers themselves. The size, and apparent quantity, of Trapezuntine aspers fell steadily between 1238-1458, in a career comparable to that of the silver coins of the I/hans. About two-thirds of all

catalogued asper types come before 1297, when their average weight, hitherto pegging the dzrhem of the Ilhans, falls faster than Mongol debasement to below 2°6 grammes. Fourteenth-century aspers (and kzrmaneult imitations) average about 1°9 grammes, eventually dipping below 1°3 grammes in the fifteenth century. Until 1297 too, the exchange rate between the asper and the hyperper of Constantinople seems to reflect the decline of the hyperper: thereafter the face value

of the asper responds only sluggishly to the drastic slide of the hyperper. The Venetian ducat: asper ratio tells the same story. Gordus and Metcalf confirm that

there was indeed a hiatus in the career of the asper after 1297. A sample of 36 which they analysed suggests that thirteenth-century aspers were of very high purity

(one achieving 100% silver), but sank in fineness after 1297, to, in one case, less than 50% in the fifteenth century. This tells us nothing about mines, but one factor may be significant. While the silver element of the alloy of the asper shrinks after 1297, the gold element actually grows at first —in one early fourteenth-century case to an extraordinary 1°66%. It is to be presumed that the Grand Komnenoi did not wittingly use gold to adulterate their silver. Elsewhere Gordus and Metcalf suggest,

with relation to the Byzantine milzaresion, that “high-gold” silver may indicate

97H. F. Schurmann, “Mongolian tributary practices of the thirteenth century,” Harvard Journal of Asiatic Studies, 19 (1956), 371-4; J. M. Smith, Jr., “Mongol and nomad taxation,” Harvard Journal of Astatic Studies, 30 (1970), 46-85; A. Evans, “Some coinage systems of the fourteenth century,” Journal of economic and bustness history, 3 (1931), 481-96; R. P. Lindner, “The challenge of Qilich Arslan IV,” Near Eastern Numismatics, Iconography, Epigraphy and History. Essays in honor of G. C. Miles, ed. D. K. Kouymjian (Beirut, 1974), 411-17; Ragid al-Din, Djamt-at-Tavarth, ed. A. A. Alizade, tr. A. K. Arends, III (Baku, 1957), 66, 297-78, 351 n. 24 (actually 23); A. Bryer, “The fate of George Komnenos ruler of Trebizond (1266-80),” Byzantinische Zettschrift, 66 (1973) = Collected Studies, IV, 346-7; S. de Vajay, “Essai chronologique a propos de la famille du Grand-Comnéne Manuel 1238-63,” Byzantinische Forschungen, 6 (1979), 281-90; KurSanskis, “Manuel I,” 29-32, suggesting that the Mongols witheld from Andronikos II (1263-6) and George (1266-80) the right to issue aspers in their own names. 3®Blake, “Circulation,” 328; A. A. Gordus and D. M. Metcalf, “Non-destructive chemical analysis of the Byzantine silver coinage of Trebizond,” Archeton Pontou, 33 (1976), 32; W. S. Atwell, “International bullion flows and the Chinese economy, ca. 1530-1650," Past & Present, 95 (1982), 68-go.

144

fresh-mined metal, and “low-gold” silver may point to bullion or coin that has been

recycled.” If this is so, I speculate that the Grand Komnenoi may have mined silver in the

fourteenth, rather than thirteenth, century. There are two possible sites for their mines: the Tzanicha area, and Argyria. Tzanicha is the most obvious candidate, but fourteenth-century Inner Chaldia lay perilously close to Muslim borders and Tzanicha itself was the fief of the Tzanichites family. From 1313 Trebizond suffered raids from rulers of Bayburt, which may or may not have bypassed Inner Chaldia, but from about 1340 an Akkoyunlu group emerged under Turali, immediately to the south, which was not pacified until the famous wedding of Kutlu beg, son of Turali, with Maria, sister of the Grand Komnenos Alexios III (1349-90) in 1352. His centre seems to have been Sinir (Sinor, Sinur, now Suntri)— as its name suggests, a border place — 30 km. west of Bayburt and 15 km. south of the Vavuk Pass on the Torul-Tzanicha-Kovans-

Bayburt road. Here he is said to have built the mosque and was buried in 1389. The Akkoyunlu alliance of 1352 gave Alexios III some respite, when he was able to deal with the Tzanichitai, but pressure from other Turkmen groups resumed from

1360. The obvious approach to Chaldia from Sinir and the south is not over the Vavuk, but further west on the Erzincan road over Kose Dag. Here Alexios III built the still impressive castle of Koukos (Kog) on the Kose Dag-Tzanicha road in 1360. The place 1s 15 km. south-east of Tzanicha and, with two other castles, Golacha (Colasana Kale, 10 km. west of Tzanicha) and Sorogaina (Soruyana Kale, 18 km. west), is the key to what was to be the Ottoman mining area. In 1361 the emir of Erzincan promptly besieged “Kiig-i Trabizon” (in a Turkish source, presumably Koukos) and Golacha (in Panaretos). Golacha was lost in 1369, regained and lost in

3° Average weights and exchange rates are summarized in graphs in A. Bryer, The society and institutions of the empire of Trebizond (Oxford, unpublished D.Phil thesis, 1967), figs. 27-31, drawing on the Goodacre Collection (Heberden Coin Room, Ashmolean Museum), R. Ratto, Monnaies byzantines et d autres pays contemporains a l’époque Byzantine (Amsterdam, 1959), 123-4; O. Retowski, Dre Munzen der Komnenen von Trapezunt (Moscow, 1910); J. Sabatier, Description

générale des monnates Byzantines, II (Paris, 1862), 306-37; the Whitting Collection (now Barber Institute Collection, University of Birmingham, where my own aspers have joined it); and W. Wroth, Catalogue of the coins of the Vandals... andof... Trebizond (London, 1911), 230-310. It would be meaningless to offer more precise figures until M. KurSanskis’s eagerly awaited Corpus of Trapezuntine coins is completed. Cf. M. D. H. Larsen, “The metralogy of the coins of Trebizond,” Seaby’s Coin and

Medal Bulletzn (July 1955), 270-1. On Mongol money, see M. A. Seifeddini, Monety Il’khanov XIV veta (Baku, 1968), and J. M. Smith Jr., “The silver currency of Mongol Iran,” Journal of the Economic and Social History of the Ortent, 12 (1969), 16-41. Exchange rates in Schilbach, Metrologie, 125; D. A. Zakythinos, Crise monétazre et crise economique @ Byzance du XIIle et XIVe stécle (Athens, 1948). On analysis, see Gordus and Metcalf, “Miliaresion,” 20; and the sames’ “Trebizond,” 33-35, which may be summarized thus:

Average 1238-63/80 1280-97 1297-1330 1349-90 1390-1412 1412-29

alloy 7 coins 10 cotns 5 coins 4 coins 4 coins 6 coms Silver 97°37% 88-76% 73°22% 68:67% 56°00% 51° 38% Gold 0°19% 0° 33% 0°70% 0° 42% 0°55 % 0°45 % Other 2°44% 10°91% 26°08% 30°91% 43°45% 48:°17%

THE QUESTION OF BYZANTINE MINES IN THE PONTOS 145 1374 and regained by 1404; Sorogaina was recovered in 1355. There seems no way of determining whether the struggle for Inner Chaldia hints that the Grand Komnenoi had mines there and wished to keep them, or that border conditions were too insecure for mining. Certainly it would be difficult to operate mines around Tzanicha if Koukos, Golacha and Sorogaina were in hostile hands, but there were also times of peace after the Akkoyunlu alliance when, for

example, the imperial family preferred to retire to Mesochaldia (?Keci Kale, Kovans, 25 km. east of Tzanicha) to escape the Black Death of 1362 and painted churches of the period such as that at Dipotamos (Ikisu), testify that a quiet life was possible. But Chaldia is not the only possibility: silver could also be found more conveniently and securely on the coast at Argyria. Argyria (Halkavala) lies between Kireson (Tylecote’s Kirazlik) and Tripolis (Tirebolu) at the mouth of the Philabonites (which the Harsit becomes in its lower reaches). It is guarded by Petroma (Bedrama) castle, which was held by an imperial Count of the Tent in 1268; Alexios III took trouble in clearing the hinterland of Tziapnides (Gepni Turkmens) in 1380. The site does not figure in de Jesus, but is in Arrian and could conceivably be the evidently Pontic Argyria in the Iliad, 11, 857; the indefatigable Hamilton found that its silver mine had been worked until about 1800. *"

One could therefore speculate that the Grand Komnenoi possibly recycled silver, perhaps of Turkish origin before, say 1297 and possibly mined some thereafter, possibly around either, or both, Tzanicha and Argyria. One could postulate that the introduction of the asper was probably not prompted by Mongol tributary demands (on which we know next to nothing), still less fuelled by Chinese silver, but

that its spread (in hoards from the Crimea to the Caucasus and Tabriz) was probably stimulated by the commercial requirements of a Levantine and then Mongol silver monetary zone which had emerged as the West was going over to a

gold standard. But while analysis of coin alloy is clearly a valuable step, the historian cannot allow himself the luxury of deducing probable or possible answers until results are more conclusive, or “finger-printing” by lead isotopes makes it

possible to relate an alloy to a specific ore.** So as things stand, there is no conclusive evidence that the Grand Komnenoi mined in Chaldia or elsewhere, although there is conclusive evidence that they defended, and largely controlled Chaldia, which itself destroys Ibn Battutah’s evidence for late medieval Turkish silver-mining there, and in turn disposes of any firm contemporary literary evidence

*°Panaretos, ed. Lampsides, 63-79; Papadopoulos-Kerameus, Sbornik, 12-14, 31. For Sinir, see: Saraf al-Din, Cheref-nameh, 1, 1, 187-88; Miroglu, Bayburt, 63-64; Woods, Aqquyunlu, 46, 48, 65, 67, 239 n. 35. Elizabeth A. Zachariadou, “Trebizond and the Turks (1352-1402),"" Archeton Pontou, 35 (1979), 335, 339, 349 nn. 1 and 2g, took me to task over identifications of Golacha and Koukos in Bryer, Collected Studzes, V, 143-6, after I had come to the conclusions now presented, which will be developed in Bryer and Winfield, Pontos, section X XII; I am grateful to her for drawing my attention to Sinir, which does not figure in section X XII; Koukos does, but not the identification of it with “Kugi Trabizon,” which she proposes “one can safely assume” is Golacha, citing O. Turan, Istanbul’un fethinden Once yazlmes Tartht Takuimler (Ankara, (1954), 81. Simir is a candidate for Sinoria in Strabo, Geography, XII, iii, 28; see O. Blau, “Aphorismen alter und neuer Ortskunde Klein-Asiens,” in Petermann, Mithe:lungen (Gotha, 1865), 252. *‘Arrian, 24, in Baschmakoff, Synthése, 96-97; Hamilton, Researches, I, 258-60; Kinneir,

journey, 332; G. Sakkas, He historia ton Hellenon tes Tripoleos tou Pontou (Nikaia, 1957), 2; Panaretos, ed. Lampsides, 79; Vazelon Act 50 of 1268. “‘aFor this method and its potential, see N. H. Gale and Z. A. Stos-Gale, “Cycladic lead and silver metallurgy,” Annual of the Brittsh School at Athens, 76 (1981), 169-224.

146

that the Byzantines mined silver or gold after the seventh century. Vryonis's question remains wide open.

III Kolonetan Alum The ancient uses of alum were limited to medicinal purposes (it is Greek styptic), to silver-and gold-polishing, and as a pigment; it is locally used today for tanning. These purposes do not require great quantities of alum, and indeed there does not seem to be any specific literary evidence for Byzantine alum mining. Demand for, and use of, alum changed dramatically from the twelfth century, when the European textile industry began to require it on a very large scale, mostly for dyeing and fulling— Turkish and Egyptian dyestuffs apparently did not call for

alum.*? Egypt was tried first. A charter of 1236 seems to be the first definite indication of the mining and export of Anatolian alum —in this case to Cyprus.“ Vincent of Beauvais noted that it was found near Sebasteia (Sivas) and Rubriquis found that Franks were running the alum trade in Ikonion (Konya).“ In 1275 the Zaccaria company obtained its famous alum concession at Phokaia (Yeni Foca) which the Genoese held until the Ottoman conquest of 1458. In 1461 Cardinal Bessarion learnt of the fall of his native city of Trebizond; Pope Pius II launched his

crusade to recover eastern ‘Christendom and found the means of financing it: the discovery of an even better source of alunite at Tolfa, in the Papal States. By 1463 some 8,oo0 alum miners were toiling at Tolfa. The crusade died with Pius the following year, but the abrupt and providential switch of a major source of alum from Anatolia to a papal monopoly lasted longer — until the Genoese began running

Tolfa too in 1527. 42C. Cahen, “L’alun avant Phocée. Un chapitre d'histoire économique islamo-chrétienne au temps des Croisades,” Revue d'Histoire Economique et Soctale, 41 (1963), 433-47; the same’s PreOttoman Turkey (London, 1968), 160-1; W. Heyd, Hestotre du commerce du Levant au Moyen Age, II (Leipzig, 1886; Amsterdam, 1967), 565-71. “3The Marseilles-Cyprus treaty in Regesta Regni Hterosolymitant ed. R. Rohricht (Innsbruck, 1893), 280 no. 1071; and H. E. Mayer, Marsezlles Levantehandel und ein akkonenstsches Falscheratelier des 13. Jahrhunderts (Tiibingen, 1971), appendix 10. Cahen, “L’alun,” 440-41 writes: “A vrai dire, j'ignore si’il peut se trouver aucune attestation de l’expoitation byzantine [sc. of alum] a l’époque macédonienne, ou antérieurement . . . il est peu vraisemblable que se soient les Génois des Zaccharia qui l’aient découvert [alum at Phokaia], bien qu’on ne semble pas avoir de preuve d'une exploitation antérieure.”’

**Vincent of Beauvais is cited by J. H. Mordtmann, s.v. ‘“Karahisar”, EJ':“. . . aluminis minera iuxta Sebastiam quae valet unam argentariam.” Cf. Saint Quentin, Tartares, ed. Richard, 69. William of Rubriquis, ed. C. Dawson, The Mongol mission (London, 1955), 218: “At Iconium I came across several Frenchmen and a Genoese merchant from Acre, by name of Nicholas of Santo Siro, who, together with his partner, a Venetian called Boniface of Molendino, has the monopoly of alum from Turkey, so that the Sultan cannot sell to anyone except these two, and they have rendered it so dear that what used to be sold for fifteen besants is now sold for fifty.” “®See C. Singer, The earliest chemical industry. An essay in the historical relations of economics and technology illustrated from the alum trade (London, 1948), a magnificent book, esp. 89-95, 139-65; M.-L. Heers, “Les Génois et le commerce de !'alum a fin du Moyen Age,” Revue d'Histotre Economique et Sociale, 32 (1954), 31-53; J. Delumeau, L'alun de Rome, XVe-XIXe stécles (Paris, 1962), esp. 15-17; Angeliki E. Laiou, Constantinople and the Latins (Cambridge Mass., 1972), 152-3; A. Bryer, “Ludovico da Bologna and the Georgian and Anatolian embassy of 1460-1,” Bed: Kartlsa, 19-20 (1965) = Collected Studzes, IX, 188-89; Memotrrs of a Renatssance Pope. The commentaries of Pius II. An abridgement, tr. Florence A. Gragg, ed. Leona C. Gabel (New York, 1962), 233: Giovanni de Castro detected alum at Tolfa because “similar herbs grew on the mountains of Asia.” Then he went to Pius II and said “Today I bring you victory over the Turk. Every year they wring from the Christians more than 300,000 ducats for the alum with which we dye wool . . . But I have found seven mountains so rich in this material that they could supply seven worlds . . . and the Turk will lose all his profits.”

THE QUESTION OF BYZANTINE MINES IN THE PONTOS 147 From the mid-thirteenth century until 1458 the burgeoning European textile industry was heavily dependent upon massive imports of Anatolian alum. Where did it come from? The obvious answer is the alum mountain of Phokaia, north of Smyrna (Izmir), where Doukas gives a clear description of how the Genoese processed it.“ But the

first hint of a Euxine rival to Phokaia comes in Michael VIII Palaiologos’s prostagma in favour of the Zaccaria company in c. 1275. It restrains the Genoese from importing alum from the Black Sea also; a Greek alum ship was seized soon

after and the embargo was not lifted until 1304.*7 During the period of the embargo, however, there is evidence that the Genoese were shipping alum within

the Euxine, that it was handled through Trebizond, and that it originated in Koloneia. In 128g is the first mention of a Genoese shipment of “allume di rocca di Colonna” —a cargo of 500 cantara. In 1290 the export of alum through Trebizond

to Crimean Caffa (Feodosiya) is twice recorded—one a consignment of 50

cantara.

The alum came from three Anatolian regions, of which “Colonna” produced

the finest quality, called “rocca”. “Colonna” is, as has long been recognised, Koloneia. The Roman name was still in use in the fourteenth century, together with variants of the more vulgar “Black Castle”, for Koloneia was a Mavrokastron in the eleventh century before the Turks made it a Karahisar —the reverse of the process whereby Giimtishane became Argyropolis. By ecclesiastical legerdemain Koloneia was also to become Nikopolis, but Greeks and Turks were more often concerned with distinguishing it from other Karahisars: as the oriental Black Castle, Sarkikarahisar; but usually by its product, as Sap-, or Sebinkarahisar; in 1340 Mustawfi identifies it as Karahisar Limuniyah.*® But Heyd’s identification of Koloneia simply as “Karahisar” has led to its sad confusion among recent commentators with the “6Ducas, Historia Turcobyzantina (1341-1462), ed. B. (V.) Grecu (Bucarest, 1958), 205; Doukas, Decline and fall of Byzantium to the Ottoman Turks, tr. H. J. Magoulias (Detroit, 1975), 148: “The alum is produced from the rocks of the mountain ridge which dissolve into sand when brought into contact with fire and then with water. Pouring this sand, derived from the rock, into a cauldron of water, it is brought to a light boil. The sand is further broken down and its dense mass remains in the solution like cheesy milk, while the hard and earthy elements are thrown out as worthless. The solution is emptied into vats to settle for four days; it solidifies around the edges of the receptacle and sparkles like crystal. The bottom of the receptacle is also covered with crystalline particles. After four days, the excess solution is drawn off and poured into the cauldron, adding more water. More sand is thrown in; the compound is brought to a boil, and once again poured into the vats in the manner described above. The alum is then removed and stored in warehouses. It is a necessary ingredient used by dyemakers. All ships sailing from the East to the West must carry a cargo of alum in their holds. Frenchmen, Germans, Englishmen, Italians, Spaniards, Arabs, Egyptians and Syrians in the dyemaker’s trade are

all provided with alum from this mountain.” |

“7F. Délger, Regesten der Katserurkunden des ostrémischen Retches, WI (Munich-Berlin, 1932) 65 no. 2016, 66 no. 2020; Heyd, Commerce, I, 438-9. *8Balard, Actes de Caffa, nos. 574, 813. “9 Artaliates, Corpus Scriptorum Historiae Byzantinae, 125; Skylitzes, ibid., 679; Anna Comnena, Alextade, ed. B.Leib (Paris, 1937~—76), III, 76; Die Seltschukengeschichte des Ibn Brbi, tr. H. W. Duda (Copenhagen 1960), 151-2, 306, 344; Colwell, Gospels of Karahtsar, 33-94; X. A. Sideropoulos, “Peri tes en Mikrai Armeniai Nikopoleos,” Ho en Konstantinoupolet Hellenthos Philologtkos Syllogos (Parartema), 17 (1886), 135; D. E. Oikonomides, Ho Pontos kat ta dtkata tou en autot Hellentsmou (Athens, 1920), 87-90; S. Ioannides, Historia kat statisttke Trapezountos hai tes peri tauten choras (Constantinople, 1870), 199; A. D. Mordtmann, Anatolien: Skizzen und Ressebriefe aus Kleinasten (1850-1859), ed. F. Babinger (Hanover, 1925), 437-42; R. Janin, s.v. “Colonia,” Dtctionnasre d'Histotre et de Géographie eccléstastique; B. Darkot, s.v. “Karahisar,” Islam Ansthlopedest; Saraf alDin, Chéref-nameh, I, i, 189-90; The geographical part of the Nuzhat-al-Qulib composed by HamdAllah Mustawft of Qazuin in 740 (1340), tr. G. Le Strange (London, 1919), 97.

148

other Karahisar, otherwise labelled by its own more sinister product: Afyonkarahisar (Akroenos, also a Nikopolis), south-west of Ankara and far from the sea. © In the 1290s Koloneian alum was evidently taken east through Cheriana and Chaldia and then north to Trebizond. By the 1340s after the Euxine embargo had

been lifted and Gepni and other Tirkmens endangered the Cherianan stretch, it was sent due north to the other Kerasous (Giresun), as Pegolotti’s handbook describes. Pegolotti’s information is that Koloneian alum was not only the finest in Anatolia, better even than the Phokaian, but the best in the world, just as Evliya

was later to eulogise it. Its annual yield was the same as that of Phokaia in the 1340s: 14,000 Genoese cantara (about 700,000 kg. or 684 English tons) were said to

be produced, and exported through Kerasous. The laborious porterage over the Pontic Alps took seven days.*' There are indeed remains of at least five hans spaced out along the 117 km. of the old Koloneia-Kerasous track. *? If correct, Pegolotti’s information raises problems. The supposed Koloneian

alum production is substantially greater than the total Italian tonnage out of Trebizond in any year—which can rarely have topped 200 tons. The contracts of 1289-90, when there was admittedly an embargo on exporting alum out of the Black Sea, amount to 26g tons at the most. The only interest the Genoese are known to have had in Kerasous comes in 1348 when they raided the place in a reprisal unconnected with alum; neither Venice nor Genoa had an officer there and alum is not mentioned in any Italian-Trapezuntine trading treaty.** Kerasous has an ancient and medieval history for the export of hazelnuts, still its chief product;

otherwise the only hint that it dealt in alum, other than Pegolotti’s, is that Kerasous—or perhaps nearby Geraprino (Kesap)— traded in scarlet kermes in the

ninth century and later. Kermes dye, inferior to “true” purple, is prepared with alum.™ If Koloneian alum was so fine and exported through Kerasous in such handsome quantities, it should have made the Grand Komnenoi a fortune to equal that of the Zaccaria brothers, or such as enabled Pius II to launch a crusade when Trebizond fell in 1461. Clearly it did not. An explanation may lie in the fact that, while the emperors of Trebizond controlled Kerasous, they never held the massive castle of Koloneia, which in Pegolotti’s time was, with the silver-mines of Gimusakar, in the hands of Eretna S°Heyd, Commerce, I, 438; followed by, e.g., P. Laven, Renazssance Italy, 1464-1534 (London, 1966), 63 and map. 5’ Pegolotti, ed. Evans: “Allume di rocca di Colonna ene il migliore allume che si lavori, e lavorasi in Turchia dentro al mare, e fae scala a Chisende di Turchia dentro al mare alla marina presso di Trabisonda, e viene 7 giornato infra terra; e fanne il detto luogo per anno in somma secondo dicesi da 14 mila cantara di genovesi” (in one incredible variant, the sum of “14 mila mila cantara” is named); it is not made explicit, but assumed, that the entire production went through Kerasous; see also 43, 293, go6; and Schilbach, Metrologre, 188. ®2Bryer and Winfield, Pontos, section XVI. °3Panaretos, ed. Lampsides, 68. **W. Tomaschek, Zur historischen Topographie von Kletnasten im Mittelalter (Vienna, 1891), 80-1; K. Kretschmer, Die Italienischen Portolane des Mittelalters (Berlin 1909; reprinted Hildesheim, 1962), 648; A. Delatte, Les portulans grecs, | (Li¢ge-Paris, 1947), 237-8; Ioannides, Hzstorza, 218; Cuinet, Turquse, I, 67-8; Constantine VII Porphyrogenitus, De Caertmonis (appendix), Corpus Scriptorum Htstorzae Byzanttnae I, 656 (the Prine of 911?); Héléne Ahrweiler, Byzance et la Mer (Paris, 1966), 107; Pliny, Natural History, vi, 3-4; Magie, Roman Rule, I, 182-83; II, 1974; I. Dujéev, The Book of the Eparch (London, 1970), 39, 166, 247, 273, 289; Pegolotti, ed. Evans, 119, 123, 144-5, 416, 420; F. Thiriet, Régestes des délibératzons du Sénat de Venise concernant la Romanie, III (Paris, 1961), no. 2349 (a commission to obtain kermes in Trebizond in 1434); W. Gell, Narrative of a journey tn the Morea (London, 1823), 183; E. Brown(e), A brief account of some travels .. . (London, 1685), 42.

THE QUESTION OF BYZANTINE MINES IN THE PONTOS 149 (1341-52). Pegolotti is probably correct in detailing the quantity and quality of Koloneian alum mined in the 1340s and in stating that there was a seven-day route

to Kerasous, but the northern track must have been only a minor outlet for it, depending upon Genoese embargoes and Trapezuntine-Turkish relations. Other indications suggest that the bulk of Anatolian alum was handled through the emporia of Sivas and Konya before some reached the Cyprus and European routes. Sivas especially was the magnet of Seljuk and later internal Anatolian commerce, and Eretna’s capital.** This makes one wonder whether some of the alum exported through Phokaia in fact came from Koloneia, for the Phokaian product was a mixture of 40% fine “rocca’, such as Koloneia produced, slaked with 60% inferior small alum.* In the same way that the label “Made in Hong Kong” may tell only part of the truth, we may have to distinguish in “alum of Phokaia” that which the place simply re-exported from the interior, adulterated with the alum actually mined at Yeni Foca. So, although alum mining around Koloneia probably existed from Pliny’s time, large-scale exploitation of it was developed not by Byzantines or Trapezuntines, but

by Turks for Franks. When in 1473 Mehmed II took Koloneia, after the battle of Baskent in Tercan (Derxene), he assigned its alum mines to his treasury.*’ But the Greeks had the last word, for they worked the mines in the alum-streaked mountains north of Koloneia until this century, washing it in dammed rivers and shipping it from villages such as Koinouk (G6yniik) to Kerasous. Miners laid off by the failing silver mines of Gumiishane revived, or maintained the Koloneian alum industry in

the nineteenth century. The principal archimetallourgot concessionaries of the silver mines had come from the extraordinary Phytianos family, of Phytiana (Beskilise, now Guzelomuk). They ran the Church too. The see of Chaldia had been refounded in c. 1624 and gravitated to Gimtshane in 1654, from when Phytianoi provided archbishops with only two breaks until after 1757. When the

silver mines failed they added “All Metal-bearing Lands” to their title. The Phytianoi followed their flock and when the alum miners of Koloneian Koinouk built their church in August 1841 it was blessed by a Christopher Phytianos of Chaldia, now archbishop of Nikopolis (i.e. Sebinkarahisar).™

IV The Mummy of Chertana Mumya Kale is a large walled early and late medieval settlement about 8 km.

north of Cheriana (Ulusiran), which David Winfield and I are publishing elsewhere.*®® Its name is singular.

Mummy is Greek moumza, Mustawfi's mumzya, Turkish mum and mumya. The curative properties of mummy were highly prized in the Arab world and, later,

in the medieval West; mummy was included in a diplomatic present to the Byzantine emperor Romanos II (959-63). Mummy is a mineral, a bitumenous pitch, resembling the resinous substance which was used as a substitute, as kermes is

a substitute for “true” purple. The substitute was obtained by pounding down 5§Zachariadou, “Turks,” 336, 355. 56 Pegolotti, ed. Evans, 43, 244, 293, 369. 57 Pliny, Natural History, xxxv, 184; Woods, Aqquyunlu, 134 Babinger, Mehmed, 313, 316. 58H. T. Okutan, Sebinkarahisar ve civart (?Kelkit, ?1948), 17; S. Vryonis, Jr., The decline of medieval Hellenism in Asta Minor and the process of Islam:zation from the eleventh through the

fifteenth century (Berkeley-Los Angeles-London, 1971), 449 and n. 14; Bryer, Jane Isaac and Winfield, “Nineteenth-century monuments, Part 4,” Archeton Pontou, 32 (1973), 219-27, 238-52, pls. 180-3, 191-5. 5°Bryer and Winfield, Pontos, section XIX.

150

Egyptian mummies whose medicinally recycled embalming fluids gave them their name. Early Turks seem, it is true, to have mummified their distinguished dead in this area: perhaps a princess at Mamahatun (Derxene, Tercan); and the curious visitor may encounter some dessicated Mongol princelings in the crypt of the Gok Medrese Camii of 1276/77 in Amasya and even a Mumya Dede in a cellar at old Harput. But true mummy has nothing to do with these, or Egyptian, mummies. True mummy was, and is, chiefly obtained from seepage in a cave near Shiraz in Iran, but a fourteenth-century medical treatise states that it was also obtainable from Byzantine places— where it was lighter in colour and rather scarce. ™ With no evidence other than its singular and specific name, I propose Mumya Kale as a candidate for one of the sources of Byzantine mummy.

*°Mustawfi, tr. Le Strange, 198; Heyd, Commerce, II, 635-36; Pegolotti, ed. Evans, 70, 124, 179, 295, 422-3; C. Ducange, Glossartum ad scriptores mediae et tnfimae Graecttatis, 11 (Lyons, 1688), Appendix, col. 148, s.v. “Momion”, citing the then unpublished former Cod. Reg. 2686, f.436 (on

Byzantine mummy); W. R. Dawson, “Mummy as a drug,” Proceedings of the Royal Soctety of Medictne (Section on the History of Medicene) 21 (November 1927), 34-9; Ibn Haugqal, Configuration de la Terre (Kitab Surat al-Ard), ed J. H. Kramers and G. Wiet, II (Paris, 1964), 294 (‘““mumiyay’”). Professor N. Panayotakis (to whom I am grateful for discussion) is to publish the letter of a Fatimid prince to Romanos II.

XII

e e Rural Society in Matzouka Pontic Chaldia, with Trebizond as its capital, was recognised as the north-eastern

frontier theme, or military province, of Anatolia by the 820’s.! It subdivided naturally into a series of banda, or military parishes. Their existence is only attested when the theme of Chaldia became the basis of the empire of Trebizond after 1204,

but it would not be surprising to find that they antedated as well as outlasted the theme. They certainly outlasted the empire, as Ottoman circumscriptions, and some banda survive as administrative units today. This is no evidence for millennial continuity, but simply of geographical imperative, for banda are coterminous with the

steep valleys which snake inland up to a watershed in the Pontic Alps SOkm. or more south of the Euxine shore. To reach one bandon from another, it is often easier to double down to the coast, which all but one reach, or to climb round the mountain pastures, which all penetrate, than to attempt the intervening mountain ridges which mutually isolate them. Until this century the result was that each valley nurtured a subculture with noticeably distinct dialects, customs and costume. As

explained in the Introduction, we chose the bandon ‘of the Matzoukans,’ which became the nahiye of Macuka, in which to study continuity and change, because it is the home of the monastery of Vazelon, the land and other charters of which concern the area from 1245-1704. It is distinguished from other banda by the facts that it has no coastal outlet (for Trebizond itself blocks it from the sea), that it embraces more than one valley system, and that most commercial and military highroads to the capital ran through Matzouka. Its most important distinction, that it retained its Christian population more tenaciously than any other bandon after the fall of Trebizond (Trabzon) in 1461, will be discussed later. I. Geography The bandon consists of the Pyxites-Prytanis river (Degirmen Dere) and its tributaries, which debouch just east of Trebizond. Flanked by the banda ‘of the Trikomians’ and ‘of the Gemorans’, the Matzoukan bandon began somewhere south

of Trebizond and Mount Minthrion (Boz Tepe), and follows the Pyxites up its valley.2 About 14 km. due south of Trebizond an eastern tributary, the Galiana 1. P.Lemerle, ‘Thomas le Slave,’ 7M, 1 (1965), 286 and note 119. 2. With the exception of Samarouxa which strays from Matzouka to Gemora in the Acts, returns to Macuka in the defters and has now returned to Yomra, the boundaries of Trapezuntine and early Ottoman Matzouka are not known to be other than those of the modern Macka Iicesi, mapped in the Trabzon 1973 Il Yilhgi (Ankara, n.d.), 73.

EYZEINOZL TONTOL

Bdv5ov ° VATNEZOYE

. Bavbov TPIKQMIA

Apvwva e rHMQPA

covud e TavapovEa YUASdoepa ° MaCeod

E(6tEa ,BENEVELA °

ell uRaTt

TAaga * outard

TIAPXAPIA | Zaven “"QAaca XatoadBEeopa e

Actualoipov ' *Meprotepa

Kootopé6c¢ ° zovutca Mavaioavot ° XaBa EnEAGraft:tava

Movuvtavto e \ ZaBota

wave taxa e Aoupent SWepcotepedta p.

MOUNT KC L Lax Koop foxudnrv

Kovoep& Xarugbofi Fig .p aKkotvAra Ldkvon eKovoncdtov BafedrAGv pp. $ MATZOYKA B&hySov

“AYOUPOA » ; TIAPXAPIA e Kovuvaxna Aaoaxay

— ee ne Kt - -Fig.3

4\coupedrd uy.

2 Ko€ vaca ) *TIAAAIOMATZOYKA Bdvéov Xauwlv

NAPXAPIA

XAAAIA S€ua oO 5 10m Cross sections 2 & 3

54

BYZANTINE MATZOUKA

(Galyana), reaches another 14 km. up to the monastery of Peristereota (Kustul), at about 750 km. above sea level. There is a second bifurcation near modern Macka,

22 km. south of Trebizond. Here another eastern tributary, the Panagia (Meryamana) reaches a further 18 km. south to the monastery of Soumela, at over 1000 m. Here too, the main highway, which has followed the floor of the Pyxites (henceforth the Prytanis) from its Galiana fork, also divides. A summer road climbs the ridge between the Panagia — and its own intermediate tributary, the Larachanes (Larhan) — and the Prytanis valleys, eventually achieving the upper air of the Pontic Gales (Pylai), some 2314 m. above and 60 km. south, of Trebizond. The winter

road climbs more sedately south-south west along the tumbling Prytanis. At Dianacha (Zanha), which faces Chortokopion (Hortokop) castle astride the summer road, from the west bank of the Prytanis, another tributary runs down from the west: the Moulaka (Meksila). It carried a track down from Cepni Turkmen summer pastures on its western flank. Overlooking the southern shoulder of the Moulaka is Vazelon (Ayana, St. John) monastery, at below 1200 m. and about 35 km. southsouth-west of Trebizond (fig.2). The winter road continues to climb up the Prytanis into Upper, Palaio- or Bas-Matzouka. Palaiomatzouka was itself a breakaway bandon from after 1384; its castle stands at 950 m. below modern Hamsik6y (fig.3).

The winter road finally leaves permanent settlement and the headwaters of the Prytanis above the tree line, and breasts the windswept Zigana Pass, at 2022 m., about 70 km. south-west of Trebizond. These winter and summer roads were always

busy in season, for they account for the last two, alternative, stages before Chortokopion of the caravan route from Trebizond to Tabriz, 32 days’ away. But they were also warpaths in spring. Invading or attacking forces, usually mounted, forced the Moulaka track in 1367 and 1461, the winter road as far as Chortokopion in 1332 and 1357, and the summer road in 1222, 1361 and 1367, not to say numerous lesser raids of which the Acts of Vazelon is the only record of summer insecurity.’

II. Population and Settlement The very position of Matzouka makes it central to Pontic history. But, apart from

the untoward distraction of events, it does not differ greatly from other banda. With an average coastal temperature range of 7°-23°C, all suffer a rainfall of well over 1000 mm. a year and are heavily wooded—almost rain forests. All were relatively densely populated and shared the same settlement pattern. The Acts of Vazelon do not offer global population figures. The Matzoukan settlements from the 1512-20 defter in fig. 1 are out first fair idea of total populations, adding up to about 14,000 souls, but may themselves be incomplete because

most of the holdings of the three main local monasteries of the bandon are not 3. Michael Panaretos, Peri ton Megalon Komnenon, ed. O. Lampsides (Athens, 1958), 72, 73, 80; G. Zerzelides, ‘Toponymiko tes Ano Matzoukas,’ AP, 23 (1959), 87-188; the same’s ‘Hermeneutika tou toponymikou tes Ano Matzoukas,’AP, (1961), 245-90; L. Maksimovic! This clear stateprovided a Protovestiarios for the Grand Komnenos Alexios III, and even a son-in-law for the Grand Komnenos David, as part of Phanariot mythology:see S. Ioannides, Historia kai statistike Trapezountos kai peri tauten choras (Constantinopole, 1870), 138; Un Phanariote (E. Rizo-Rangabé), Livre d’or de la noblesse Phanariote (Athens, 1892), 87. 45. Beldiceanu, REB, 35 (1977), 184-89. 46. Mamia was Gurieli of Guria in 1459 when he was part of David’s alliance against the Turk; he sent his wide Helen to refuge with him in 1461. He is a candidate for the Giirgi tekvur who lost Trapezuntine holdings after 1461. What appears to be an archon of Guria is buried beside the Hagia Sophia, Frebizond. See Chalkokondyles (Bonn ed.), 467, 4395; Bryer, CS V, 1215 n. 35 and X, 183, n.32; Beldiceanu, Byzantion, 49 (1979), 27. 47. Chrysanthos, AP, 4-5 (1933), 536; F. Babinger, Mehmed the Conqueror and his time (Princeton, 1978), 307; J.E. Woods, The Aqquyunlu. Clan, Confederation, Empire (Minneapolis-Chicago, 1976), 128-30 (for background). It is difficult to know who the ‘nephew’ of David was. 48. Woods, Aqquyunlu, 142; H. Inalcik, Speculum, 35 (1960), 425; Bryer and Winfield, Pontos, XXII, nos. 11,15. 49. Beldiceanu, AP, 35 (1979), 66-67. 50. R. Gonzales de Clavijo, Embajada a Tamorlan, ed. F.L. Estrada (Madrid, 1943), 80. 51. O.L. Barkan, ‘Osmanli Imparatorlugunda bir Iskan ve Kolonizasyon metodu olarak suriinler, ‘Istanbul Universitesi Iktisat Fakiiltesi Mecmuasi,15 (1953-54), 219 n. 88; Beldiceanu, AP, 35 (1979), 66; cf. Gdkbilgin, B7TK, 26 (1962), 336-37.

70

BYZANTINE MATZOUKA

ment of the transfer of military obligations in return for a land fief by the same family to a Muslim from a Christian power is muddled by the fact that Barkan, Beldiceanu and Vakalopoulos, who discuss it, assume that Torul transferred its allegiance from David Komnenos to Mehmed II (who in fact never went there) soon

after 1461, rather than eighteen years later and only seven years before our register.> The confusion 1s understandable, for it is not clear that Torul can have been wholly independent in 1461-79. During this period some of its Christian fiefholders held land in Ottoman Matzouka as well, but were apparently not dispossessed

in the Pontic coast. One may speculate that the independence of the Christian warlords of Torul of the coastal government in 1461-79 was similar to that described so vividly by Clavijo earlier, but that Chaldian families like the Tzanichitai

too were anxious to keep their lands in areas like Matzouka to wish to keep up resistance elsewhere. At all events, the direct transfer of military obligations, which

Beldiceanu is probably justified to refer to as those of pronoia, is stated baldly enough in the 1486 defter, where Barkan records that ‘it becomes evident that it belongs to nine non-Muslims, who where infidel sipahis following the Emperor, the master of the property, and who became treacherous to him and abandoned

him’. |

A military pronoia in Torul would have been substantial. In 1404 Leon II

Kabazites had a mounted retinue there, garrisoned several major castles (including

Torul itself), and defended the area against Cepni. In turn he exacted toll from caravans. Half the salary of the tenth-century strategoi of Chaldia had similarly been derived from customs.™ Such major fiefs, pronoia (a term not in fact found in Trapezuntine sources) or fimar, would not have been numerous and were most at risk in the deportations of the Pontic aristocracy (including a Chaldian Kabazites) after 1461. The problem comes with lesser military or other holdings, for they are largely known only by title. By western Byzantine analogy, some could have been quite small. Between 1254 and 1310, Theodore Sapouas, Theodore Zosimas, Thomas Sachas, Therianos Basakios and Constantine Kibousenopoulos describe themselves by the historic title of stratiotes, when there are polemarchs too. Maksimovic¢ notes

that most appear to be of some social standing. Between 1397 and 1449 a more shadowy sequence give their names as strategos. All presumably owed military service, but it is impossible to hazard its terms. An Andronikos the Stratégés appears

in 1482 in Gemora (Yomra), but cannot be linked with any specific Andronikos there of the defter four years later.> 52. Barkan and Beldiceanu, /occ.citt.; A.E. Vacalopoulos, Origins of the Greek nation (Brunswick, 1970), 229 and 355.

53. Barkan, 220. 54. Clavijo, ed. Estrada, 80; Constantine Porphyrogentius, De cerimoniis aulae byzantinae, |. (Bonn ed.), 696-97. 55. Vazelon Acts I of 1431, 7 or 1482, 11 of 1435, 25 of 1260, 28 of the thirteenth century, 30 of 1295, 35 of 1310, 44 of 1276, 81 of 1397, 82 of 1265, 103 of 1386, 106 and 108 of the 13th century, 111 of 1254, 123 and 134 of 1415, 138 of 1432, and 175 of 1449; Lazaropoulos in Papadopoulos-Kerameus, FHIT, I, 118-19; Miklosich and Miiller, A&D, V, 280; Laurent, 71

More distinctive is evidence for a Trapezuntine equivalent of the Palaiologan allagion, a sort of mounted sergeantry with attendants (knights and squires are too pretentious terms for what were simply those who could maintain a horse), who

served against unspecified holdings or immunities. In Byzantium the allagion indicated no more than a military unit, and in Trebizond allagatores are known by name and title alone. A probably fourteenth-century order of precedence places

four ideal and perhaps fictitious varieties of protallagatores low on the list, and an archon of the allagia near the bottom. That such mounted men actually existed as part of the local defence system is borne out by the signatures of Alexios and Eugenios Psomas, ‘allagatores of the bandon of the Matzoukans’ in 1278, and

there is a thirteenth-century somatophylax or attendant. In fourteenthcentury Byzantium the megala allagia were divisions of central troops stationed in

the provinces, much on the lines of later Ottoman spahis. The term is not found in Trebizond, but Panaretos describes how during the civil war of 1340 the basilika allagia of the capital divided their loyalties: some joined the Palaiologan party of Eirene (whence may have come such terms); and others the bandon magnates. The ‘imperial guards’ of 1340 do not add up to megala allagia in title, but provide parallel evidence that, as in Byzantium, the palace troops had country cousins.°° Humbler

and more numerous than the major fief-holders, some of whom demonstrably became timariots after 1461, the allagatores would have been less at risk if they were to survive at all. AP, 18 (1953), 262; L. Maksimovic, ‘Pronijari u Trapezuntskom carstvu,’ Zbornik Filozofskog Fakulteta, 12 (1974), 393-404. Stratiotai appear, with archontes, as an important element of Trapezuntine society in a horoscope of 1336: see S.P. Lampros, ‘Trapezountaikon horoskopion tou etous 1336,’ NE, 13 (1916), 40; cf. I. Sevéenko, ‘Society and intellectual life in the fourteenth century,’ Actes du X1Ve Congres international des Etudes byzantines, Bucarest 1971, 1 (Bucarest, 1974) = Social and intellectual life in late Byzantium (Collected Studies) (London, 1981), I, 75. For modest pronoias and conditions in western Byzantium, cf. e.g.: D. Jacoby, ‘Les archontes grecs et la féodalité en Morée franque,’ 7M 2 (1967), 421-82; M Angold, A Byzantine government in exile. Government and society under the Laskarids of Nicaea (1204-1251) (Oxford, 1974), 124-26; N. Oikonomidés, ‘Notes sur un prakitikon de pronoiaire (juin 1323),’ TM, 5 (1973) = Documents et études sur les institutions de Byzance VIleXVe s). (Collected Studies) (London, 1976), XXIII, 335-46; Laiou Thomadakis, Peasant Society,65-66; B. Ferjancic, ‘Quelques significations du mot stratiote dans les chartes de basse Byzance,’ ZVI, 21 (1982), 95-102. 56. Vazelon Acts 63 of 1278, 106 of the 13th century; Panaretos, ed. Lampsides, 65. That the list of precedence in Venet. Marc. Gr. 608, ff.376v-377r is of Trapezuntine origin is argued by Oikonomides, AP, 35 (1979), 311; and by Bryer, CS, V, 140 n. 109. But that some of its offices cannot actually have applied to Trebizond is shown by anachronistic titles of the

allagion: the protallagator ton monikaballanon, ton mourtaton (Turkish renegades), ton ‘zakonon (tzannoi would be more appropriate to the Pontos), and fon tzaggratoron (archers), as well as an archon ton allagion: Pseudo-Kodinos, Traité des Offices ed. J. Verpeaux (Paris, 1966), 346-47. On the western Byzantine allagion, see Angold, Nicaea, 186; and N. Oikonomidés ‘A propos des armées des premiers Paléologues et des compagnes de soldats,’7M8 (1981), 353-71. I have not seen Dr Mark Bartusis’s dissertation on The late Byzantine soldier: a social and administrative history (Rutgers University, 1984), but am most grateful to him for discussion. His ‘The cost of late Byzantine warfare and defence,’ will be published with the Proceedings of the Nineteenth Spring Symposium of Byzantine Studies, edd. A. Bryer and M. Ursinus (ByzFor, forthcoming). D.A. Zakythinos, Le despotat grec de le Morée Ii (Athens, 1953), 93.

72

BYZANTINE MATZOUKA

Beldiceanu first proposed that two unusual terms imported by the Ottomans to describe Christian holdings in the Pontos represent existing institutions.~’ First, the 1486 defter reveals numbers of a lesser category, in all 270 musellem, a mounted sergeantry, and their 338 yamak, or attendants. Like the Trapezuntine allagatores, Ottoman miisellem are also found on Lemnos and among Christian militias of the Balkans. Their equation cannot be demonstrated specifically, but the allagatores remain inescapable candidates for local origins which Beldiceanu

sought — he also suggested Trapezuntine origins for the Christian Ottoman derbencis, or pass-holders, of the Pontos, for which there is no specific evidence.” The second term, which Beldiceanu proposed described an existing and more commoner institution, is that of bastina a heritable Christian small-holding (which could pass through widows and daughters if need be), subject to a tithe and some obligations of defence. Save for the last feature, all these are qualities of the commonest and most characteristic of holdings found in the Acts of Vazelon: the gonikeion. This was a family croft, which could be freely inherited, mortgaged, sold, endowed, exploited but probably not abandoned. It would be held by women in the absence of menfolk, and in part, and on behalf of another family member. It was subject to dekatia. Gonikeia could be widely scattered (one in eleven or more places) and were not large. A substantial one, in 1260, consisted of a house, choraphion, garden, outhouse, pasture, threshing floor, swine, two bulls, and a copper cauldron worth 8 aspers — horses, cavalry or other, are not mentioned. Fiscal rights over a gonikeion could be transferred from the state to a monastery, but even when a widow grants her own gonikeion to the monastery it does not reduce her status; the term paroikos does not appear once in the Acts of Vazelon. Nevertheless there are signs that the gonikeion-holders of Matzouka were losing ground before 1461 and are in a minority thereafter.°’ If, as I propose, the imported term bastina may be equated with the gonikeion, we are left with the problem of military service. If it indeed existed after 1461, it was evidently at the humblest level, a home guard. Such a local mass levy corresponds to Bessarion’s description and to what we know of actual turnouts of Matzoukans in times of invasion. That local defence remained in continuous Christian hands before and after 1461 is at least not a matter of speculation. It can be demonstrated archaeologically, as well as through the defters. The Tzanichitai’s own home castle of Tzanicha (Canca) contains two probably fourteenth-century painted chapels, which have been aban-

doned but never slighted. Of Matzouka’s four manned castles before 1461, two preserve chapels. Doubera, for which Soumela monastery was enjoined to maintain a kastrophylax in 1364, houses a shrine which was venerated until this cen57. Beldiceanu, AP, 35 (1979), 59-69. 58. Beldiceanu, AP, 35: (1979), 68-69. 59. Vazelon Acts 79 of 1260, 80, 102, 105 and 106 of the thirteenth century (no mentions thereafter);

Act 39 for an example of a peasant abandoning his inheritance to the monastery as early as 1264, through misery and bankruptcy, after Vazelon foreclosed on his grandfather’s mortgage; Uspenskij and Beneshevich, Vazelon xcii-xciii; cf. Laiou-Thomadakis, Peasant society, 184.

73

tury, while Vazelon’s own watch-castle of the Lavra, at Gantopedin, has a church which was kept in repair until recently. The evidence of the 1486 defter that the active coastal castle of Kerasous (Giresun) was manned by Christian marines 1s backed by later Greek folksong that Christians there continued to bear arms; while Ottoman evidence of Christian garrisons at the seaside fortresses of Koralla (Gorele) and Tripolis (Tirebolu) is matched by medieval churches within their walls which were only abandoned in the nineteenth century.© Further east comes the intriguing evidence of Contarini that some Ottoman Christians were thought to have been no more loyal to the Sultan than they had been to their Emperor. A consequence of the fall of Trebizond was that Venice found access to Uzun Hasan was blocked. There remained a long way round, east of Trebizond, which is the route which Contarini was advised to take from Caffa to reach Uzun Hasan in 1473. The Land of Arhakel (Armenian Hemsin), then an Akkoyunlu satellite, could be approached by landing on the Ottoman shore at ‘Tina’ (Athenai, now Pazar, about 140 km. east of Trebizond). There is a notable small medieval fortress on the shore there (Kiz Kulesi) and a route to the staggering heights of Zil and Varos, castles in the interior, which reminded Purchas of the Cimmerian Gloom a century later. Contarini was told that Athenai was then ruled by the Turk but ‘there was only a castle belonging to the Greeks, in which I should certainly be placed in safety.’°! The 1486 defter makes his first two points abundantly clear in Athenai, while the third shows that,

as in 1472, Pontic Christians could still be expected to support the Akkoyunlu connection. After 1461 Christian timariots were doing no more for the Sultan than they had

for the Emperor, save that their southern and eastern borders had now shrunk. When they expanded again in the sixteenth century, the new Georgian and Armenian marches'still had to employ Christian castle-holders. East of Athenai, Soterioupolis (Borcka) has the remains of a church (perhaps even a cathedral) within the Ottoman fortress. To the south and east Miroglu reveals Christian timariots in early sixteenthcentury Bayburt and Ispir. Elsewhere I have argued that the surviving churches in each citadel were built under, if not by, Mugith al-Din Tugrilgsah between 1213 and

1225 — the hapless melik who before he got into Matzoukan hands had been in those of the Armenian Leon I of Sissouan — even his son had to turn Christian at the Georgian Queen Rusudan’s behest. Mugith al-Din did not confine himself to Christian works, as his inscriptions on Bayburt walls proclaim. He probably built 60. Beldiceanu, AP, 35 (1979), 71-73; Miller, Trebizond, 197; Bryer and Winfield, Pontos, XIII, nos. 1 and 6; XV, nos. 2 and 3; XVII, no.3; XXI, nos. 17, 18, 23, 24, and 52; XXII, no. 15. In 1609 Bordier was told that Giresun (Kerasous) island had held out against the Turks for seven years after 1461, and found that in Giresun itself, ‘les habitants de la ville presque tous les Grecs, font la plus part avoient connoissence et quelque amitié avec les soldats francois pour avoir esté deux mois en garnison en ce lieu.’ J. Bordier, ‘Relation d’un voyage en Orient,’ AP, 6 (1934), 114-15. 61. Barbaro and Contarini, Travels in Tana and Persia, ed. Lord Stanley of Alderley, trans. S.A. Roy (London, 1873), 116; Purchas, His Pilgrimage (London, 1614), 342 (cf. Odyssey, XI, 14); Gokbilgin, BTTK, 26 (1962), 321-25; Bryer and Winfield, Pontos XXVI, nos. 1, 5 and 6. 714

BYZANTINE MATZOUKA

the mosque in Ispir citadel too, but the place was registered as 100% Christian in 1520.

That Pontic defence remained in Christian hands until well into the sixteenth cen-

tury is unquestioned, and the defters confirm that many of the second-rank fiefholders of the Grand Komnenoi became timariots. It cannot be proved that the shadowy Trapezuntine allagatores were translated into Christian miisellem, but it would be difficult to explain away the presence of 270 of these Christian horsemen and their 338 yamak in 1486 without antecedent. The same argument goes for the lowest stage of free-holders, the bastina. | am not concerned with the etymology of the term save that it is imported, but there are enough parallels to show that it was probably applied to gonikeion-holders. The extent to which the Ottomans swallowed a local defence system in 1461 should

not be surprising, if two factors are considered. First, given the overwhelmingly Christian rural population, they could do not other, though Contarini and the abortive invasion of 1472 show that, even before the defter of 1486, some could not be trusted if given the choice between Uzun Hasan and the Sultan. In many respects the events of 1461 should have left the rural defence system largely unmoved. The

Pontos and the Ottoman state had never been neighbours; there had been no war of attrition; the Fatih’s army had come and gone largely by-passing opposition, leaving Matzoukans and others to continue to defend the rim of their world against the same Turkmen threats as before. Second, the defence system would have been all the easier to swallow because practical features of it were already to the Turkish taste. By the fourteenth century Trebizond had its emir candar, perhaps an emir dogan, and its horsemen were riding

@ la Turque. This adaptation of the fighting aspect of its neighbours, illustrated in the knights with their short stirrups, kefiye and matzouka (mace) in fig.4, may have been just as important a change as that of the later sixteenth century, when Christian timariots and miisellem began to lose their grip: by contrast 1461 was

less of a turning point.” VI. Public Justice and Private Bonds I do not seek to overlook the very real changes of 1461. For example, like other banda, Matzouka had its imperial officials: a doux and kephale, a primikerios and a krites. A few local landholders assumed the tital of kyr. All this was swept away, and local officials do not seem to have been transferred to the Ottoman payroll. While the last Acts of Vazelon adopt Ottoman terms like ‘tapu’, the deftérs were not compiled by local Greek speakers. The new aga was to be roughly equivalent to the old kyr, the new ayan to the.old archontic class, but they were of different 62. I. Miroglu, XVI. ytizyilda Bayburt Sancagi (Istanbul, 1975), 113-14, 117, 141-50, 184; cf. S. Vryonis, ‘Islamic sources for the history of the Greek people,’ Indiana Social Studies Quarterly, 32 (1979), 65. For monuments at Soterioupolis, Ispir and Bayburt, and the melik’s building see Bryer and Winfield, Pontos, XXVII, no. 3; XXVIH, nos. 1 and 3. 63. Pseudo-Kodinos, ed. Verpeaux, 341-42, 348; Laurent, AP, 18 (1953), 261, 267, 269; Vazelon Act 166 of the fifteenth century; Clavijo, ed. Estrada, 77. 75

.a

OO e t y. 4 { 4a fi A () x (\ am, 4 4 KD ' ys ‘ eo N /: : ~7 * Mee Tk —» ZA af St .e ?9t|I) . a; *e : :| .Ke Rad ' : ¥ ° = _ ' nO a \ AW, os Y) . | | |\ sD) URIT T : urgh . e FWinfield 1. TAXIARin ’ 0: FG. 5: tienne td*etéByzan :ae 9a ‘ _.‘ Gabrie | . ee . — > ae... ee4tito ° €as 7Bsioe ‘idge ¥ aisaeae Se2th eePAS :.EE ne . ae ex| BS vast .: , #24 ot JES Heee, MY x’ ft: hg", aegee, Be pA PRO me, “pin

3"ae: | as Sel” oS PkSy. “2 real Aho -w, Rk edie os ae “a-" aa’ Rat+ ae ae,x.ote . weOg *xa \R al “gg ge Oe : 43,+), SUES. P* oe % . :.4re Se, zLe. pes :Mek 1-aS ee beet Cn oe \Sone. ae OFS wg oe 4a‘;Te ; eae... > ee, mot eRe” 4 aBP a“ee Tar” aSagres ee ries. 3oe Kata —— se, “4aSRG Se F “4 ~. ¥% uetebX ra a b a oo as its > ee Tt. SAB . q -_ At % s, ‘ E. eee “he. ES te 9314 4 Lisae: ye a on TUE |, “Ree . : . . ‘i J " : hg ato. — 9 ene” Pt o¥ s ai : a at . % or ees a. wy re fi ° %.. . Sed “ : we ’ oe BY Teo eee meee eS ae “ay oo eS AR ay tty Mec me “es PS . pers :

.ry ‘Feo, Rit cece LSE Fee Sg~Ss, ti igh Ree .BSS. Re ’ aa te, 2.6, bo%ote “‘. \a\mh * Sy Sig? ro aes... i £¥-ERT oe Sg A geete aS .an ~5&va -ma. 7og Coane 7tikes Par) TAME 4 eS ss,at BE aawe’ ATS 4“% awe ‘‘oR osie SE: < Ay oe "Ng BRIM, 97-8 a es aa 8ocr

oe ".>, oo roS *”a.RC: ooo dad ian ae? 7érNSea Bax, : yo. - xsepe we IN g%: aSem a“fl Pyahs §thiam ws . aos aeMb ie ," ot" vi or”? aWEES Baten *fy ee D ieaS, “9 a. 3ai % .es Bo ‘ys tasfw, >: oo" 44e oe. coRE, nt re Ser vot ‘ay aa *Pia gy,id ee »Ps My the aos7 4¥. ‘ts ee oe is>gat SR § f7° ay rneed 7. a. %‘4% ee, PR Zi iBe} on’, aay. Lo crea aaEee Ee Be «> ne »“ RF a, ~wor weE~~ 8ae 3+e a"ie st, A , gee ? “