Problems of Medieval Armenian and Muslim Historiography: The Mxit'ar of Ani Fragment Author(s): Dickran K. Kouymjian Source: International Journal of Middle East Studies, Vol. 4, No. 4 (Oct., 1973), pp. 465-475 Published by: Cambridge University Press Stable URL: http://www.jstor.org/stable/162315 . Accessed: 22/06/2013 14:57 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp . JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org. . Cambridge University Press is collaborating with JSTOR to digitize, preserve and extend access to International Journal of Middle East Studies. http://www.jstor.org This content downloaded from 71.172.220.217 on Sat, 22 Jun 2013 14:57:27 PM All use subject to JSTOR Terms and Conditions Int. J. MiddleEast Stud. 4 (I973), 465-475 Printedin GreatBritain 465 Dickran K. Kouymjian OF MEDIEVAL PROBLEMS ARMENIAN AND MUSLIM HISTORIOGRAPHY: THE MXIT'AR OF ANI FRAGMENTI Only the introductory portion of the History written by the Armenian Mxit'ar of Ani has come down to us.2 However, an extensive passage on the Ghaznavids and Seljuks from the lost part of the work is quoted in the Universal History composed c. I268 by Vardan Vardapet.3 The content and the sources used for the compilation of this much-neglected narrative is the subject of this study. From biographical data and the final date (I I93)4 in the surviving introduction, The article was originally presented in a much abbreviated form as a paper at the 4th Annual Meeting of the Middle East Studies Association, 6 November 1970, in Columbus, Ohio. The author would like to acknowledge, with thanks, the travel grant provided by the Research and Conference Grant Program of the American University in Cairo for this purpose. The Mxit'ar of Ani fragment has already been treated in a related article: D. K. Kouymjian, 'Mxit'ar of Ani on the Rise of the Seljuqs', Revue des lStudes Armeniennes,new series, vol. vi (I969), pp. 331-53, which will henceforth be referred to as R. E. Arm. For the Armenian translation with Russian resume of the original paper, see now the journal of the Armenian Academy of Sciences, Erevan, Armenian S.S.R., Lraber, no. 4 (352), 2 (I972), pp. 74-84. Published by K. Patkanean and appended to his edition of the History attributed to Sebeos (St Petersburg, 1879), but with separate pagination; the text exists in a unique manuscript now no. 2678 of the collection of the Matenadaran (i.e. Manuscript Library) in Erevan, Armenian S.S.R. 3 Ed. and Russian trans. by M. Emin, 2 vols. (Moscow, I86I); by L. Alisan (Venice, 1862). Despite the conclusions of J. Muyldermans to the contrary (La dominationarabe en Armenie [Louvain/Paris, 1927], pp. 29 and 37; cf. R. E. Arm, p. 333, n. io), Aligan's edition, at least for the section containing the Mxit'ar fragment, is superior to Emin's, not only because he had access to much older manuscripts, but also because he had a better knowledge of oriental languages. To settle the debate on editions of Vardan one need only consider that when Patkanean appended the excerpt in Vardan at the end of his edition of Mxit'ar (pp. 49-52) he used Alisan's edition and not that of his fellow Russian Armenian, Emin. As for the manuscript tradition of Vardan, the four oldest - Library of the Mechitarist Congregation in Venice, nos. 516 (dated c. I300) and 1244 (dated I307), the Matenadaran, Erevan, no. 3074 (also dated 1307), and Museo Borgano, The Vatican, Arm. MS no. 30 (dated 1630, but probably from an original of 1274) - display no major divergencies one from the other or from the printed texts. The writer would like to take this opportunity to thank again the Research and Conference Grant Program of the American University in Cairo for a grant during the summer of 1970 which allowed him to examine these manuscripts (as well as many others) in person. 4 Ed. Patkanean, pp. 2, 46-8, also reproduced in G. Zarbhanalean, The History of Ancient Armenian Literature [in Arm.], vol. I (Venice, I897), p. 725. For biographical material on Mxit'ar, in addition to works already cited in R. E. Arm, p. 332, n. 6, one This content downloaded from 71.172.220.217 on Sat, 22 Jun 2013 14:57:27 PM All use subject to JSTOR Terms and Conditions 466 Dickran K. Kouymjian the history can be safely ascribed to the last years of the twelfth century. Near the beginning of the same extant section Mxit'ar provides a list of sources he used, which for the eleventh and twelfth centuries are exclusively Armenian.I Yet we know he was able to utilize Islamic, or at least Persian, sources, for both Vardan and Mxit'ar of Ayrivank' (ChronologicalHistory to 1289) report that he translated a book on eclipses from the latter language into Armenian.2 As will be seen later, Mxit'ar seemingly consulted both Armenian and Muslim works for the section of his history under study.3 An examination of the works of the writers of the period who are named by Mxit'ar - Aristakes (History to I070), Kozein (History of the Bagratids, written c. 1050), and Samuel of Ani (Chronology to 79) - reveals that the detailed information on the Ghaznavids and Seljuks presented by him did not originate from these works.4 There is the further possibility that he used other contemporary Armenian authorities but neglected to cite them; these might include the may consult P. S. Somal, Quadre della Storia Letteraria di Armenia (Venice, 1829), p. 0o6; Alisan, Snorhali and His Time [in Arm.] (Venice, 1873), pp. 126-8; A. Lazikean, New Armenian Bibliography and Encyclopaedia of Armenian Life [in Arm.] (Venice, 1909-I2), vol. I, cols. 2013-I4; H. Acarean, Dictionary of Armenian First Names [in Arm.], vol. III (Erevan, 1946), pp. 369-70; G. Hovsep'ean, 'Mxit'ar of Ani, Scribe and Miniaturist' [in Arm.], Hask Yearbook, vol. I (Antelias, 1948), esp. pp. 192-4, which discusses four different Mxit'ars of Ani living during our period, the one of the title not being the Mxit'ar of this paper. , Ed. Patkanean, p. 15, with details in R. E. Arm, p. 334; the list will be found in the next paragraph of the text below. 2 Vardan, ed. Venice, p. I37, ed. Moscow, p. i8o; Mxit'ar of Ayrivank', ed. Emin (Moscow, i86o), p. 64. It is also mentioned by Arakel of Tabriz, History (seventeenth century), ed. (Valarsapat, I 884), p. 48, and probably copied from the latter in an eighteenthcentury chronology which has only recently been published, N. Polarean, 'Chronology' [in Arm.], Banber Matenadarani, vol. 9 (I969), p. 259. The name of the author is given as Ocie in older manuscripts and as Ocik'e in later ones as well as Afakel and the eighteenthcentury chronicle. Acarean, loc. cit., suggests the possibility that Mxit'ar of Ayrivank' understood it as a date, i.e. RCII = 1127, and that later it was mistakenly entered in Arakel under 1187, but this should probably be rejected, for all of the oldest manuscripts of Vardan agree exactly in understanding it as a name; Mxit'ar of Ayrivank' places the chronicle event just before I 191 and after i 181, and Afakel and the eighteenth-century also clearly regard it as a name. On the other hand, Brosset's more reasonable, yet still problematic, suggestion that it is a poor Armenian rendering for Persian zij, a book of astronomy, deserves further investigation, Mdm. de l'Acad. vol. iv (I862), no. 9, pp. 5-6; see R. E. Arm, p. 333, n. io for full citation. Alisan expresses the same opinion, but without reference to Brosset, Snorhali, p. 127. 3 Mxit'ar actually says as a preface to the short history in the Vardan fragment, ' A great deal of effort was exerted in discovering [the history of] the sultans who were Turks, and by the grace of God I found it [to be] as follows'; ed. Venice, p. 94; ed. Moscow, p. 127. The great effort was probably research into non-Armenian as well as Armenian sources; Ali?an comments much to the same effect, ibid. 4 Aristakes of Lastivert, critical ed. (Erevan, I963) and Russian trans. (Leningrad, I968), both by K. N. Yuzbasyan; Samuel of Ani, critical ed., A. Ter-Mik'elean (Valarsapat, I893); for Kozern and a discussion about the surviving parts of his lost history see below, pp. 467-8, n. 4. It is most interesting that Alisan felt the lost history of Kozern contained an account of the Seljuk invasions, but unfortunately he does not say on what grounds he bases this (Hayapatum [in Arm.], vol. i (Venice, 1901), p. 90). This content downloaded from 71.172.220.217 on Sat, 22 Jun 2013 14:57:27 PM All use subject to JSTOR Terms and Conditions The Mxit'ar of Ani fragment 467 History of Hovahannes Sarkawag (d. I129), the Chronicle (952-I 136) of Matthew of Edessa, the Chronicle to i 62 of Mxit'ar Gos, and that of Vahram (type and date of work uncertain).' But once again, none of these have any important material on the Ghaznavids or Seljuks with the possible exception of the history of Sarkawag. However, it has not come down to us. From excerpts of this work quoted by his pupil Samuel of Ani and later by Kirakos of Ganjak (History to I265)2 we have some idea of its general content. The first of its two parts dealt with the 'Scythians', i.e. the Turks, more probably the Seljuks, while the second part was specifically about Malikshah son of Alp Arslan.3 In a previous study (see p. 465, n. I), by incorrectly associating the name Kozein, in the oblique form 'Koziann', with Hovhannes Sarkawag, it was conjectured that Mxit'ar of Ani used this important history. Even now that 'Koziann' is clearly identified as Hovhannes of Tar6n, called 'Kozern',4 it is still held by this writer on the basis of contextual I On Hovhannes Sarkawag see below in the text; Matthew of Edessa, ed. (Jerusalem, 1869) and a later edition based on a more complete text (Valarsapat, 1898), Fr. trans., E. Dulaurier (Paris, 1858); Michael the Syrian, ed. (Jerusalem, I871), Fr. trans., V. Langlois (Venice, I868), but on the question of the various and divergent Armenian versions see J.-B. Chabot, Chroniquede Michel le Syrien, vol. I (Paris, 1899 [actually pub. I924]), pp. L-LI; Mxit'ar Gos, Albanian Chronicle, trans. and commentary, C. J. F. Dowsett, BSOAS, vol. xxI (1958), pp. 472-90; on Vahram see Alisan's comment in Vardan, ed. Venice, p. 94, n. 3, and Hayapatum, vol. I, p. 92, cf. R. E. Arm, p. 333, n. 12. This latter Vahram is probably not to be confused with Vahram of Edessa who wrote a rhymed history of the Kings of Cilician Armenia despite comments in some eighteenthcentury manuscripts, for which see H. A. Anasyan, Armenian Bibliology [in Arm.], vol. I (Erevan, 1959), p. LV. 2 Samuel, pp. 96-8, quoted in Alisan, Hayapatum, vol. II, pp. 336, 358-9, and Zarbhanalean, p. 609. Kirakos of Ganjak, critical ed. K. A. Melik'-Ohanjanyan (Erevan, I 96), p. 84; Kirakos probably took his excerpts from Samuel rather than from Sarkawag's text directly, for which see H. Oskean, Literary Researches [in Arm.] (Vienna, I926), P. 39. 3 On the author and the work see, Alisan, Souvenirs of the Armenian Fatherland, vol. ii (Venice, 19212), pp. 248-73; idem, Hayapatum, loc. cit.; Acarean, op. cit. vol. III, pp. 571-2; Lazikean, op. cit. vol. II, cols. 107-8; A. Abrahamean, Eight Lectures [in Arm.] (Antelias, I955), pp. 79-96; Zarbhanalean, pp. 609-10; Samuel, pp. 96-8; Oskean, op. cit. pp. 1-64, esp. 37-41. 4 An additional note attached to offprint copies only of the writers article in R. E. Arm, vol. vi, correctly identified Kozein/Koziann with Hovhannes of Tar6n, an eleventhcentury author of several works including a History of the Bagratids, believed to be lost. He is mentioned by a large number of contemporary and later authors. For fuller details see Acarean, vol. III, pp. 566-7; Zarbhanalean, pp. 570-I, 788, is not aware that Hovhannes of Tar6n and Kozefn are one and the same person, probably repeating the mistake of Mxit'ar of Ayrivank', ed. Emin, p. 23, and an eighteenth-century manuscript (Anasyan, op. cit. p. LIV) where the two names are listed separately. MS. no. I775 of the Matenadaran in Erevan contains the first pages of the beginning of Kozern's history (folios 8v-I6r), which according to the heading in the manuscript was a history of the house of the Bagratids (f. 8v). G. Abgaryan quotes verbatim the opening paragraphs of this history, Banber Matenadarani, vol. 6 (I962), pp. 5o-I, and again in his 'The History of Sebeos' and the Enigma of the Anonymous (Erevan, I965), pp. I28-30, and maintains that the first chapter of the pseudo-Sebeos (the so-called Primary History of Armenia) belongs to the lost part of the History of Kozefn. An examination of the surviving pages of this This content downloaded from 71.172.220.217 on Sat, 22 Jun 2013 14:57:27 PM All use subject to JSTOR Terms and Conditions 468 Dickran K. Kouymjian similarities that Mxit'ar probablyutilized Sarkawag's History and that conclusions based on this premise are valid. The work seems to have existed in Mxit'ar's time, for his contemporary Samuel of Ani used it, and later writers - Kirakos, Vardan, Mxit'ar of Ayrivank' - mention it. More factually, sections of Sarkawag's work quoted by Samuel and Kirakos contain the unusual spelling 'Sarc'uk' for Seljuk, a form also used by Mxit'ar.I (On the other hand, the Mxit'ar passage uses hijri dates, while normal Armenian era dating is used in the fragment of Sarkawag's work preserved by Samuel.) Unfortunately, a definitive statement on the relationship between the lost History of Hovhannes Sarkawag and the passage in Vardan from the lost History of Mxit'ar cannot be made. Yet it is clear that either Mxit'ar or possibly an Armenian predecessor (i.e. Sarkawag) relied on Islamic works, not only because hijri dates are used exclusively in the fragment (while in other extant parts of Mxit'ar's work they are not2), but also because of (i) the relative accuracy in the rendering of Muslim names and titles, (2) the use of many Arabic and Persian words, and (3) the detailed account of certain incidents totally unrelated to Armenian history.3 What then were Mxit'ar's Muslim sources? Since the fragment, which, being written before the close of the 6th/I2th century, is comparatively early, deals with details of Bfyid, Ghaznavid and Seljuk history, the answer to this question is of special interest to the historiography of these dynasties as well as that of Armenia. Before entering into a discussion of this matter, a paraphased summary of the Mxit'ar fragment on the history of the sultans of the Turks preserved in Vardan is herein presented.4 Mahmud son of Sebuktegin(Sbk't'anay),like Ardashirthe Sasanian,became great in the city of Balkh (Bahl) in the land of the Kushans. Word of his strength reachedthe Caliph (Xalifay) who, becoming frightened, sent him presents, an insignia of office (alam, Arabic 'alam),the laqab(lalap) Amin al-'aidil(Amin-adl,read Yamin al-Dawla), and called him Sultan. With these honors Mahmud'sreputationgrew even greater.He historydoes not revealmuch of interest;beginningwith Adam and Eve in Paradiseand continuingaboutthe Hebrewprophetsandkings,it goes up to the birth of Christ(f. i6r). The historywas to be in two parts: (i) from the beginningof the world to 887 when the first Bagratid King Asot was crowned, (2) from the reign of Asot to the author's own days (c. 1050). It is possible, indeed very likely, that the second part discussed the Seljuk invasions, which occurred during the author's floruit, as Alisan had himself surmised (see p. 466, n. 4 above). The author would like to thank L. Xac'ikyan, Director of the Matenadaran, and B. C'ugaszyan, Assistant Director, for kindly making available photographs of the manuscript text of the Kozein fragment. I Samuel, p. 98; Kirakos, p. 84, in the variants at the foot of the page; for Mxit'ar, ed. Venice, p. 96, ed. Moscow, p. i29, the passage to be translated shortly in the text. The final section of the extant introduction of Mxit'ar's work also uses the form azgn Turk'ac', Sarc'ukik'n, 'the nation of the Turks, the Seljuks', ed. Patkanean, p. 48. Alisan felt very strongly that Sarkawag's history began with the origin and development of the Seljuks 2 Ed. Patkanean, passim; Zarbhanalean, pp. 725-6. Hayapatum, vol. I, p. 95, col. I. 3 See the passage given in the text, below, and the discussion in R. E. Arm, p. 351. 4 Italicized words in parentheses indicate the Armenian spelling found in the text; when not so indicated the Armenian either has already been given or is identical or very close to the accepted form. This content downloaded from 71.172.220.217 on Sat, 22 Jun 2013 14:57:27 PM All use subject to JSTOR Terms and Conditions The Mxit'ar of Ani fragment 469 marched to India to the famous idol of Somnath (Mat'an),' which he destroyed, and aftertakingmuch booty returnedhome, leavinghis son Muhammadthere.Then he took Gurgan and from its ruler 400,000 gold pieces (tahekank')and a treasurein the year 420/[I029] of their (i.e. the Muslim) era. He then passed to Rayy (Re); the ruler of the city, Majd al-Dawla Rustam (Maja-dawla Rstom), who had gathered Daylamite (Dlmikk')troops three days before, advancedto meet him. Mahmud arrivedat Sarav (Sraw) with a formidableforce and 250 elephants,and seeing Majd al-Dawla, he said, 'Have you read the Shdhndma(Sah namaz)?' And the other answered, 'Yes.' Now Mahmud asked, 'Have you ever played chess?' The answer was again, 'Yes.' 'Then does a king enter another king's square (tun, lit. 'house')?' And the ruler of Rayy remained silent. Mahmud seized him and sent him captive to Khurasan (Xorasan) and took all his possessions in Rayy. Then Mahmudwent to Tabaristan(Taparastan) and Tafabi (?)2 and Sari; he gave his son Mas'ud (Masxut) Rayy, Qazwin (Lazuin), and the whole of Kuhistan(K'ohastan).3Leaving Mas'udin Rayy he went to Saravand took Ioo,ooo dinars (denari) and passed to Nishapur (Nasawur) in 42I/[I030] of their era. Mas'ud then took Hamadan (Hamian) and Isfahan (Aspahan) and returned to Rayy,wherehe heardthat his shahna(sahnays)had been killed in Isfahan.He went there again, killed 4,000 men and returnedto Rayy. Just at that time they brought him the sad news of his father's (i.e. Mahmud's) death and that his brother Muhammadwas made Sultan. Rising in revolt, Mas'ud marched [toward Ghazna], seized his brother, blinded him, and took his sovereigntyand throne. Sometime before, Mas'ud's father, while on his way to help Qadir-KhanBoghraKhan (Xtrlan-P6lrlan)[the Qarakhanid],met an army of Turks; later, he returnedon the same road and with a greatvictory,4he seized their amirYabghu (Ap'alu)and sent him captive to Khurasan.The amir's people begged Mahmiud,and later Mas'ud, to release him, but neither would. Thus angered, the Turks passed across the Oxus (Jahan) with all their troops and fighting with all their might seized Nishapufr.Afterwards they destroyed the armies of Sultan Mas'ufdat Dandanqan (Dadanlan) near Marv (Mrmn).The sultan fled to Ghazna ('i Lazinn) and then India, but on the way he was killed and his blind brother Muhammadwas enthronedin Ghazna ('i -azuin [sic]) where until today his sons rule. As for the leader of the Turks, whose name was MuisaYabghu (Museap'alu)son of Seljuk (Sarc'uk'),he had five [sic, read two] nephews whose names were Abui Salim (Abusalim,read AbufSulayman) Da'fud(Dawut') Chaghri Beg (('alrbek), [and] Abuf Talib (Abutalip)Tughril Beg (T6lril-Bgk).It was Tughril who received the authority I Tentatively it would seem that ArmenianMat'an is a badly transcribedform for Arabicmandt,' idol', in its genericsense, but originallyManat,one of the paganidols of the Ka'bah,which was closely, though incorrectly,associatedby Muslim writers with Somnath,especiallywith regardto Mahmfid'scampaignthere in 416/1025-6. 2 Other localitiesmentionedin the text are easily identifiable;Taiabi, however,presents some problems. It would seem too far removedto be either Tarab near Bukhara (V. Barthold, Turkestan,2nd ed. [London, i968], p. i 5 n.) or Darab-jird in Fars (G. Le Strange, Lands of the EasternCaliphate[Cambridge,I905], pp. 248, 288-9), though Alisan identifieswith the latter (Vardan,ed. Venice, p. 95, n. 4). 3 Kuhistan in Syria, i.e. al-Jibal,V. Minorsky, .Hudud al-'Alam (London, 1937), p. I50. The authorwould like to thank ProfessorsWilliam Gohlmanof Baldwin-Wallace College and William Hanawayof the University of Pennsylvanniafor independently suggestingthe more correctreadingin place of the author'sKhufzistan. 4 The words 'with a great victory' were inadvertentlyomitted from the more exact translation(of the Seljuk portion only) given in R. E. Arm, p. 339, but were properly includedon page 342. In general,for specificquestionson the Seljukssuggestedby the Mxit'arfragmentone should consult the translationand commentaryin R. E. Arm. This content downloaded from 71.172.220.217 on Sat, 22 Jun 2013 14:57:27 PM All use subject to JSTOR Terms and Conditions 470 Dickran K. Kouymjian of the sultanateand dividingthe lands of Khurasan,he extended his realm for fifteen years. Later Tughril came to Rayy and discovered two treasures filled with gold, seized them, and sent to the Caliphaskingfor his blessing. The latterhonoredhim with ambassadors,an insignia of office (alam), presents, he read his name from the minbar (mambar),and gave him the title Rukn al-Dawla (Ruk'nadovla).And from that day he was proclaimedconqueror.' The passage, even in its abridged form, contains a wealth of diverse information which naturally lends itself to a detailed commentary. As mentioned above (p. 465, n. i), the Seljuk portion has already been treated in a monograph, and the section on the Ghaznavids will be studied thoroughly in a forthcoming article. For the present discussion, only two episodes will be singled out to help determine the sources used by Mxit'ar. They are (I) from Ghaznavid history, the interrogation of Majd al-Dawla by Mahmud of Ghazna, and (2) from Seljuk history, the occurrence and use of the Turkic title yabghu. The dialogue between Mahmud of Ghazna and Majd al-Dawla Rustamb. Fakhr al-Dawla of Rayy is best known in the version preserved in Ibn al-Athir's al-Kdmilfi al-ta'rikh (finished c. 61 9/I222). After the death of his capable mother Sayyida (the defacto ruler of Rayy) in 419/1028, Majd al-Dawla assumed the full responsibility of governing the city. We are told that being unable to control his Daylamite troops, he called on Mahmud for help. The latter was waiting for he took Rayy. Under that year Ibn alsuch an opportunity and in 420/I029 Athir records the following conversation. Having had Majd brought before him Mahmud asks, 'Have you read the Shdhndma, the history of the Persians,2 and the Ta'rzkh of al-Tabari, the history of the Muslims?' Majd answered, 'Yes.' 'But your conduct was not like one who had read them,' said Mahmud. He continued, 'But do you not play chess?' 'Yes,' replied the other. 'And did you ever see one king approach another?' questioned Mahmud. 'No,' said the ruler of Rayy. 'Why then,' admonished Mahmud, ' did you call to your kingdom one who is stronger than yourself?' And saying that, Mahmud sent him captive to Khurasan.3 1 The passage has been published in Armenian as follows: ed. Venice, pp. 93-7; ed. the Venice version again by Patkanean (op. cit. n. 2), pp. 49-52; by Alisan again, Souvenirs, vol. II, pp. 353-4; and partially (the initial section on Mahmud of Moscow, pp. 127-3I; Ghazna) by A. Alboyadjian, History of the Armenian Emigrations [in Arm.], vol. ii (Cairo, 1955), p. 24, n. i. It has been translated into French, M. F. Brosset, Additions et eclaircissementsa l'histoire de la Georgie (St Petersburg, I851), pp. 220-2; Russian, M. Emin (Moscow, i86I), pp. 118-21; Turkish, H. D. Adreasyan, 'Muiverrih Vardan Turk Fiituhati Tarihi (889-I269)', Istanbul Jniversitesi Edebiyat Fakultesi Tarih Semineri Dergisi, vol. I/2 (I937), pp. I69-72; English (partial), R. E. Arm, vol. vi, pp. 339-4.1 2 There is an added poignancy and irony to the story. Firdausi completed the Shahndma at the commission of Mahmud (c. 400/1009-10), but disappointed by the terms of the payment, he fled from the Ghaznavid court and took refuge with the Bufyids, specifically, accordingto some authorities(e.g. Ethe), at the court of Majd al-Dawla in Rayy; see E. G. Browne, A Literary History of Persia (Cambridge, I906), vol. ii, pp. I4I, 13I n. 3 Ibn al-Athir, ed. Tornberg (Leyden, I851-76), vol. IX, pp. 261-3 (reprint, Beirut, vol. IX, pp. 371-2); the dialogue has been translated at least twice - Browne, vol. II, p. This content downloaded from 71.172.220.217 on Sat, 22 Jun 2013 14:57:27 PM All use subject to JSTOR Terms and Conditions The Mxit'ar of Ani fragment 47I Though Mxit'ar's account neglects to mention that fact which makes sense out of the chess episode, namely Majd al-Dawla's invitation to Mahmud, his version and Ibn al-Athir's are nearly the same. Though the famous Arab historian wrote some thirty years after Mxit'ar of Ani, there is absolutely no conceivable reason to assume he could have used the Armenian as an authority. Where then did Ibn al-Athir get his story? Of the existing Ghaznavid sources written before his time - 'Utbi, Bayhaqi, Gardizi, and the anonymous Mujmal al-tawdrzkh wa'lqisas - none mention the incident.I Perhaps it may have been included in the lost parts of Bayhaqi's Mujalladdt which dealt with Mahmud, for the surviving section, the Ta'rikh-i Mas'udi, only begins in 421/1030, just after the events related above, but there are no references to such a story in the works of later writers who had access to these lost parts.2 Yet should we some day find the missing parts of Bayhaqi, or even a new Persian source, we might conjecture its use by Mxit'ar, who knew Persian, but there would still be a problem since we are not sure whether or not Ibn al-Athir could use that language.3 Two further possibilities may help solve this historiographical question. The story may have been transmitted by the Persian 'Ali b. Zayd Bayhaqi, known as Ibn Funduq, in his Mashdrib al-tajdrib, which was written not in Persian but in Arabic, during the second half of the 6th/I2th century. Though the work has not come down to us, there is reason to suppose that Ibn al-Athir may have used it.4 But once again the problem of Ibn Funduq's source would arise as well as that I6o, and M. Nazim, The Life and Times of Sultan Mahmud of Ghazna (Cambridge, 193 ), pp. 82-3. For further details on Majd al-Dawla, see K. V. Zettersteen, 'Majd alDawla', Encyclopaedia of Islam (ist ed.), or Islam Ansiklopedisi, vol. vii, pp. 431-2, and, G. C. Miles, The Numismatic History of Rayy (New York, I938), pp. 171-97 passim. I 'Utbl, al-Ta'rlkhal-Yamzin,written in Arabic prior to 43I/1039-40, the year of the author's death, ed. (Cairo, I869), with a Persian version by Jurbadhqani (c. 602/20o6), ed. A. Qawim (Teheran, 1955). The Ta'rikh-i Mas'udz of Abu'l-Fadl Bayhaql (385/995 to 470/1077) comprises only some five of the supposed thirty-volume history, the Mujalladdt; the section preserved covers the years 421/1030 to 433/I041. The Zayn al-akhbdr of The anonymous Gardlzi was written before 444/1053 and contains events to 432/104I. according to the unique manuscript in the Bibliotheque Mujmal was written in 520/1126 Nationale, Paris. The author would like to thank Prof. C. E. Bosworth of the University of Manchester for sending photocopies of the sections from Gardizil, ed. Nazim (Berlin, I928), pp. 90-I, and the Mujmal, ed. Bahar (Tehran, 1939), pp. 403-4, used to check this statement. 2 An examination of the second and revised edition of the Russian translation of Bayhaqi, A. K. Arends (Moscow, I969), which contains in the appendix 19 excerpts from the lost parts of the Mujalladdt quoted by later authors, shows no mention of the episode; of course this does not absolutely exclude the possibility of its being preserved in the still-missing parts of the work. 3 'For, much as we may admire on the one hand the breadth of his documentary researches, nothing, on the other hand, indicates that he knew Persian' (Cl. Cahen, 'The Historiography of the Seljuqid Period', Historians of the Middle East, ed. B. Lewis and P. Holt [London, I962], pp. 65-6). Contrariwise, his brother Diya al-Din ibn al-Athir seemingly knew Persian (see G. Von Grunebaum, Islam, Essays in the Nature and Growth of a Cultural Tradition [London, I96I], pp. I09, n. 9 and 178), and therefore one might conjecture that Ibn al-Athir himself could in fact use Persian sources. The reference to 4 Cahen, ibid. p. 66. Diya al-Din was supplied by Prof. Gohlman. This content downloaded from 71.172.220.217 on Sat, 22 Jun 2013 14:57:27 PM All use subject to JSTOR Terms and Conditions 472 Dickran K. Kouymjian of Mxit'ar's access to it.' A second alternative might be that some caliphal or Baghdadi chronicle, in this case probably written in Arabic, may have included the tale. A likely candidate is the History (a narrative to 447/I055) of Hilal al-Sabi, of which only the years 388/998 to 393/I003 have been preserved, but which was used and partially transmitted in Ibn al-Jawzi's al-Muntazam (6th/ I2th century), Sibt Ibn al-Jawzi's Mir'dt al-zamdn (7th/i3th century), and an anonymous chronicle in the Munich State Library dated 644/I246-7, all three of which expressedly used Hilal's work.2 In the Muntazam we do in fact have preserved the Fathndma actually issued by Mahmud of Ghazna and sent to the Caliph after his conquest of Rayy; but again, though Ibn al-Jawzi's work is quite detailed, it has no mention of the confrontation with Majd al-Dawla.3 Furthermore, it would seem that the absence of our story from the rather full account of Ibn al-Jawzi, who relies on al-Sabi, would also exclude its existence in Sibt Ibn al-Jawzi or the Munich anonymous, though a check through the manuscripts of both authors must be made before a definitive negative statement is possible. A final supposition that the story may have been totally fabricated around the turn of the 6th/ 2th to 7th/I3th century seems improbable in view of its existence in two completely unrelated works. Therefore, we are back where we started without a very encouraging probability as to its origin. At the end of this study I Of course the lost part of the Mujalladdt comes to mind: however, another possibility i. the work of the Buyid vizier Abu Sa'd Mansur b. al-Husayn al-Abl, which is named by the anonymous Mujmal as one of its sources. On Abu Sa'd see C. E. Bosworth, 'On the Chronology of the Ziyarids in Gurgan and Tabaristan', Der Islam, vol. 40/I (I964), p. 30, n. 10; the author again expresses his thanks to Prof. Bosworth for this suggestion and a copy of the article. As for the question of Mxit'ar': use of Ibn Funduq or other Arabic sources, we have no definite evidence that he did not know and use Arabic, only positive proof that he was able to use Persian (p. 466, n. 2 above). Alisan's statement that the translation of the astronomical work already cited (p. 466, n. 2 above) was made from an Arabic book, yArab dprut'ene (girak [in Arm.] [Venice, i88i], p. 95, col. 2), must be considered an unintentional slip, for in the same work (p. o00, col. 2) he quotes the source of our information directly from his edition of Vardan (p. I 37), 'i parsik lezue', 'from the Persian language'. Nevertheless, Arabic as well as Persian, Georgian and other languages, Eastern and Western, were common in the city of Ani at that time (Alisan, ibid. p. 96). A detailed linguistic analysis of the non-Armenian vocabulary in both the surviving introduction and the Vardan fragment needs to be made to determine if such words were borrowed from Arabic rather than Persian texts. 2 al-Muntazam, partially published (including the years under discussion), vol.. v-x (Hyderabad, I938-I941, with a recent reprint); Mir'dt al-zamdn, still unpublished for the years under consideration, but a later section on the Great Seljuks has recently been published, A. Sevim, Mi'rdtii 'z-Zeman fi Tarih ii'l-Ayan 447/1056-479/I086 (Ankara, I968), on which see Cl. Cahen, 'A propos d'une edition...', Arabica, vol. xvI/i (1970), pp. 82-91; the Munich anonymous has not been published, but reference to the manuscript and its contents will be found in M. Kabir, The Buwayhid Dynasty of Baghdad 334/946447/1055 (Calcutta, I964), p. 216 and passim. A discussion of Hilal al-Sabi's history will be found in Cahen, 'Historiography', pp. 60-4. 3 This Fathndma (al-Muntazam, vol. viii, pp. 38-40) has been translated in extenso by Bosworth,' The Imperial Policy of the Early Ghaznavids', Islamic Studies, vol. I/3 (I962), pp. 70-2. This content downloaded from 71.172.220.217 on Sat, 22 Jun 2013 14:57:27 PM All use subject to JSTOR Terms and Conditions The Mxit'ar of Anifragment 473 something will be said about one final possibility, the work of a contemporary to the fall of Rayy, one Ibn Hassul. Our second example, taken from Mxit'ar's History, concerns the identity of the Seljuk Yabghu. The title goes back at least as early as the eighth-century Kok Turkic Confederation, being used in the Orkhon inscriptions to designate the office next after the Qaghan, the King. After the breakup of the K6k Empire, the Oghuz Turks, who were an integral part of it, kept the title yabghu and used it to designate their leader. By the end of the 4th/Ioth century, when the Seljuk Turks broke away from the Oghuz Yabghu, whose capital was at Yengi-Kent on the lower Syr Darya, or shortly after this, Arslan-Isra'il b. Seljuk took the. title yabghu as an act of defiance as well as a sign of his family's increasing power. The title was normally held by the eldest male member of the family and ArslanIsra'il was at the time senior to Musa, Seljuk's only other surviving son.' Whether or not 'Yabghu, leader of the Oghuz', mentioned by Gardizi as aiding a SamAnid ruler in 393/1003 against the Qarakhanids, was Arslan or the Oghuz Yabghu himself is still not clear,2 but certainly by 416-17/1025-6, when Arslan was captured and imprisoned by Mahmud of Ghazna, he definitely held the title yabgu, for not only does the Akhbdr al-dawlat al-Saljuqiyya (early 7th/l3th century) refer to him as 'Arslan Yabghu called Isra'il',3 but Mxit'ar of Ani confirms this by calling the captured leader (amir) of the Turks only by his title: Yabghu (Ap'alu). Since Mxit'ar wrote at least a quarter of a century before the Akhbdr was composed, his source is independent of it and provides a link with an earlier and now lost source, perhaps in this case the Malikndma.4 Further along in the Armenian text, around the events associated with the battle of DandAnqan (431/1040), we find that now it is Muisa b. Seljuk who has the title yabghu and even more that he is clearly identified as the leader of the Turks.5 Of the early sources which have anything to say about Mfisa, only Mxit'ar and Zahir al-Din Nishapuri in his Saljuqndma (second half of the 6th/ I2th century) correctly identify him as both the Seljuk Yabghu and the uncle of Tughril and Chaghri Begs.6 Thus, though Mxit'ar's subsequent testimony makes I A further discussion with full references to the literature will be found in R. E. Arm, pp. 337-8. 2 Gardizi, ed. M. Nazim, p. 64; in addition to the citations in R. E. Arm, pp. 338-9, nn. 42-5, see also Cahen, 'Arslan b. Saldjuk', EP. 3 Turkish trans., N. Liigal (Ankara, 1943), pp. 2-3 and R. E. Arm, p. 342, n. 75. 4 On which see the indispensable study by Cl. Cahen, 'Le Malik-nameh et l'histoire des origines seljukides', Oriens, vol. II (1949), pp. 31-65; cf. R. E. Arm, p. 332 and passim. 5 Further confirmation of this point is found in Mxit'ar of Ayrivank' (p. 2, n. 2 above; this point not presented in R. E. Arm), who, probably using Mxit'ar of Ani, places Musa Yabghu (Musep'ayloy) at the head of his dynastic list of Seljuks, ed. Emin, p. 22. He again mentions Mfisa in the chronological part of his work after the year A.D. 901 [sic!], '...the Turkman (T'urk'man) Seljuk (Salcuk') and Mfisa Yabghu (Muse P'aloy) and Tughril Beg (D6llabek)... etc.', p. 55. 6 There is an almost literal version of the Saljiqndma preserved in the Rdaiat al-sudur (end of 6th/i2th century) by al-Rawandi, quoted here in the ed. by M. Iqbal (London, I902), pp. 102, I04; cf. R. E. Arm, pp. 336, n. 34, p. 346, n. 91. This content downloaded from 71.172.220.217 on Sat, 22 Jun 2013 14:57:27 PM All use subject to JSTOR Terms and Conditions 474 Dickran K. Kouymjian it evident that Tughril was the most powerful member of the Seljuk family in the post-Dandanqan period - although he never took the then seemingly honorary title yabghu - it is also evident that we must not regard his uncle Musa as simply 'far from dynamic'.' It is obvious that in our second episode neither the yabghu question nor the many other details about the Seljuks contained in the fragment allow us to be as precise in discussing Mxit'ar's authorities for this section as we were in considering those for the Ghaznavid one. We have nothing as pronounced as the events centered around the capture of Rayy to pin-point the Islamic sources on the Seljuks which Mxit'ar may have used. Nevertheless, generalizations about early Armenian historical works made with regard to Ghaznavid history are equally valid for Seljuk history (see above, pp. 465-8). All we can say is that the pattern which develops out of the Armenian account seems more related to the Zahir al-Din group of Persian sources (which are independent of the reconstructed Malikndma) than to the Malikndma itself, this despite certain similarities between the latter work and Mxit'ar's.2 The problem still remains very unsatisfactorily resolved. We are left to speculate that there was yet another early Seljuk source, perhaps now lost, which was used either by Mxit'ar or his predecessor Hovhannes Sarkawag as well as later Muslim authors. As has been suggested elsewhere (R. E. Arm, vol. vi, pp. 348, 352-3, and above, p. 473), a distant possibility is a supposed late 5th/I ith century Chronicleabout the early Seljuks by Abu'l-'Ala' Muhammad b. HIassuil(d. 450/1058), a vizier of Majd al-Dawla, who subsequently worked in Rayy as a bureaucrat for the Ghaznavids and later for the Seljuk conqueror of the city, Tughril Beg.3 Beside the fact that Ibn Hassul was employed by all the major figures in both the episodes cited above, what makes him particularly interesting from the point of view of the Armenian sources is that in another existing work of his, a Risala which was to have been a preface to the same Chronicle, he shows, like Mxit'ar but unlike the pro-Chaghnr Malikndma, a strong bias towards Tughril,4 and, more importantly, again like Mxit'ar and Sarkawag, he uses the rare form (at least for Muslim sources) of Sarjuk (s.r.j.k) for Seljuk (normally spelt s. ljfq).S To add further support to this supposition, a Ta'rzkh by one Abu-'l-'Ala-iC. Cahen, Pre-Ottoman Turkey(London, 968), p. 20; see the more detailed discussion in R. E. Arm, pp. 346-7. 2 For details, R. E. Arm, pp. 352-3. C. E. Bosworth, The Ghaznavids (Edinburgh, 1963), p. 59; Cahen, 'Le Maliknameh', pp. 36-7. 4 Risala, ed. 'A. 'Azzawi with Turkish trans. by $. Yaltkaya, Belleten, vol. Iv (I940), pp. 250-66 and 51 pages of Arabic text; a thorough discussion of the work and its author is given by Cahen, 'Le Malik-nameh', pp. 37-8. 5 Risala, text p. 49, trans. p. 265. This section of the Risala has been translated into English by D. M. Dunlop, The History of the Jewish Khazars (Princeton, 1954), p. 259; cf. Bosworth, Ghaznavids, p. 220, and R. E. Arm, p. 348, n. 103. For the Armenian usage see above, p. 468, n. i, and the text, and R. E. Arm, p. 347, nn. 98-9. 3 This content downloaded from 71.172.220.217 on Sat, 22 Jun 2013 14:57:27 PM All use subject to JSTOR Terms and Conditions The Mxit'ar of Ani fragment 475 Ahwal is in fact cited by the early 8th/I4th century Persian historian Hamdullah Mustawfi Qazwini in the introduction to the section on the Seljuks in his Ta'rzkh-iguzida.I Claude Cahen believes that by a simple orthographic correction this personnage is almost certainly to be identified with Abf'l-'Ala Muhammad b. 'All b. Hassil and one and the same as the author of the Risala.2 Furthermore, if we can believe a note in H. Raverty's translation of Jfzjani's Tabaqat-i Nasirt, this Ta'rzkh of Abfi-'l-'Al-i-Ahwal (now to be read Abu'l-'Ala ibn HIassil) states that Seljuk had four sons named 'Isra'il, Mika'il, Musa-i-Beghu (sic, to be read Musa-Yabghu)...and Yfinus'.3 Since Raverty claims to have actually used this Ta'rkkh,4a manuscript of Ibn Hassil's work may in fact still exist. Of course, like the vizier's other work, the Risala, it would probably have been written in Arabic and would raise again the question of Mxit'ar's use of it. Yet if we can take Raverty's statement referring to such a manuscript at face value, there would seem to be good reason for less caution about the existence of such a history, which, since Ibn Hassul worked for the Ghaznavids in Rayy, might have contained information on the capture of the city by Mahmud and, not unlikely, even the story about the author's former employer, the Buyid Majd al-Dawla. If this proves to be true, then Ibn Hassil's Ta'rlkh could have served very well as the foundation for the details on both the Ghaznavids and Seljuks found in Mxit'ar's report as well as the original source for the Muslim accounts.5 No matter how one may react to the myriad speculations presented above, it is clear that Mxit'ar of Ani's short history of the sultans of the Turks as preserved in Vardan is a well-informed and detailed account which, when more thoroughly examined, may help us to understand better some of the important aspects of Muslim and Armenian historiography of the 5th/IIth to the 7th/i3th century. THE AMERICAN BEIRUT, UNIVERSITY OF BEIRUT LEBANON Ed. (facsimile) and abridged trans., E. G. Browne, 2 vols., GMS (London, I9I0-13), facsimile p. 434. 2 'Le Malik-nameh', pp. 37-8, for a complete discussion. 3 H. G. Raverty, The Tabaakdt-iNdsirl (London, I881-99), pp. ii6-i8, n. 3. 4 Raverty, p. 117, n. 3. 5 Bosworth likewise comments (letter of io October I970): 'This leaves rather a mystery, and I can't, offhand, suggest any obvious solution, unless al-Abi [p. 472, n. i above] or Ibn Hassul are possible relaters of the anecdote.' This content downloaded from 71.172.220.217 on Sat, 22 Jun 2013 14:57:27 PM All use subject to JSTOR Terms and Conditions