Hudud al-'Alam, The Regions of the World
V. Minorsky

§ 22. The Majgharī

Chwolson, Izvestiya . . . Ibn Dasta (read: Ibn Rusta), pp. 101-23; Marquart, Streifzüge, pp. 27-74 and pasim; Dietrich, Byzantinische Quellen, Index sub verbis: Mazarer, Ungarn; B. Munkácsi, Die Urheimat der Ungarn, in Keleti Szemle, vi, 1905, pp. 185-222; Barthold, Basdjird, in EI; J. Németh, Magna Hungaria, in Mžik, Beiträge, pp. 92-6; Németh Gyula,

318   Commentary   § 22

A honfoglaló Magyarság kialakulása, Budapest, 1930 (a short résumé of this important work is La Préhistoire hongroise, in Nouvelle Revue de Hongrie, Budapest, June 1932, pp. 460-8, the communication of which I owe to the kindness of the author); C. A. Macartney, The Magyars in the Ninth Century, Cambridge, 1930 (a painstaking revision of Byzantine and Oriental sources, the latter being used in translations; Gardīzī's text accepted in Marquart's earlier interpretation); J. Moravcsik, Zur Geschichte der Onoguren, in Ungar. Jahrbücher, x, Heft 1-2, 1930, pp. 53-90. See Map xii.

The question of the remote Hungarian (Magyar) origins depends chiefly on linguistic evidence and more especially on that of loan words in Magyar and its cognate idioms. As the nearest of kin to the Magyar are the Voguls (on both slopes of the Northern Ural) and the Ostiaks (in the Obi basin), it was formerly admitted that the original home of the Magyars must be sought in Siberia. So Marquart, Streifzüge, 53, located the "Ursitze" of the Magyars in "southern Yugria, in the neighbourhood of the Ishim and in the Baraba [steppe east of Omsk]". More usually, following the indications of the Muslim authors (v.i.), the seats of the early Magyars were placed in the neighbourhood of the Volga Bulghārs, i.e. near the present-day Bashqir territory. Munkácsi in his Urheimat der Ungarn, p. 212, while criticizing these theories took an entirely different view, to wit that the region where the Magyar language underwent the influence of the [older] Turkish and Caucasian languages [1] lay in the northern Caucasus and that accordingly this was "das Urgebiet des Bildungsprocesses des Magyarentums"; and if some Magyars were found near the Volga this must be explained by some emigration from the Caucasian home in the northward direction.

Turning now to Muslim sources we must recognize that under Majgharī, Basjirt, and other similar names [2] Arab and Persian authors speak of two distinct groups, vis. the Uralian "Bashqirs" (whether Turks or Finno-Ugrians) and the Magyars (Hungarians) in their earlier country north of the Black Sea.

According to Prof. Németh's latest researches, the Bashqirs are originally a Hungarian tribe, which probably together with the Volga Bulghārs had migrated from the northern Caucasus northwards, cf. Munkácsi, o.c., 221. [3] The name of the Bashghirs [4] mixed with that of the Hungarians living near

1. For traces of former contact of the Magyars with the Ossets see now Hannes Sköld, Die ossetischen Lehnwörter im Ungarischen, in Lund Universitets Årsskrift, N.F., Avd. I, Bd. 20, No. 4, 1925 (where the Magyar-Osset contacts are placed circa a.d. 600— 800). In principle it is hazardous to associate the Iranian (i.e. Alān > Osset) elements in Hungarian exclusively with the Caucasus for the Alāns once stretched well to the neighbourhood of the Aral Sea. [The theories on the earliest home and migrations of the Magyars are necessarily very controversial.] [Cf. Appendix B.]

2. See their enumeration in Chwolson, o.c., 112, and Marquart, Streifzüge, 68-9.

3. Moravcsik, o.c., 89, thinks that this migration took place simultaneously with the westward trek of the Onoghundurs (§ 53) about the middle of the 7th century.

4. Németh explains it as *bäsh-ghur "Five tribes" [?].

§ 22   The Majghari   319

the Black Sea (Mod'eri) resulted in the form: Mojgher. This, together with the common origins of the two peoples, led to a situation under which the two were indiscriminately called now Bashghird, and now Mojgher. Those Hungarians who had travelled from the Caucasus to the north carried along with them some Turks, and later became turkicized by other Turks coming from Western Siberia. Kāshgharī considers the Bashqirs as Turks speaking a dialect akin to that of the Kimäk, but the Dominican Julian who, in search of the lost Hungarian tribes, visited the region of the Volga in 1235 found a "Magna Hungaria" near the "Magna Bulgaria" (i.e. the Volga Bulghārs). Moreover, some of the clan names of the Hungarians mentioned by Constantine Porphyrogenitus (a.d. 948) coincide with those of the present-day divisions of the Bashqirs ( = Hung. Kürt + Gyarmat = Bashq. Yurmatu = Hung. Jenő = Bashq. Yeney). See Németh, Magna Hungaria.

The clearest and simplest presentation of the case in Muslim sources is found in Iṣṭ., 2253, who says: "there are two classes of Basjirt (). The one is found at the farther end (ākhir) of the Ghūzz behind the Bulghār (alāẓahr B.) and they are said to be about 2,000 men [1] and to be protected by impassable thickets (mashājir); they obey the Bulghār. [2] The other class of them borders on the Pechenegs; both they and the Pechenegs are Turks and they border on Rūm." [3]

Much more entangled is the group of sources represented by I.R., 142, Gardīzī, 98, and Bakrī, ed. Rosen, 45, who under the name Majgharī mechanically string together the information referring to two different territories and most probably derived from different sources (Muslim al-Jarmī, Hārūn b. Yaḥyā, &c., cf. Marquart, Streifzüge, 28) as if the Uralian territory stretched without interruption down to the Black Sea. [4] The introductory paragraph (A) of these authors places the Majgharī in the north

1. In the curious legend on the formation of the Khirkhīz, Gardīzī, 85, says that their ancestor after having been obliged to leave the court of the Khazar-khāqān joined Bashjirt who "was one of the Khazar nobles and with 2,000 men lived between the Khazars and the Kimäks".

2. It is possible that I.Kh.'s (p. 31)  quoted in the series Toghuzghuz - Kharlukh - Kimäk- Ghūzz - J.f.r - Bajanāk - Türgish stands for  rather than for  supposed by de Goeje and Marquart to represent Chigil. The same consideration may apply to Mas'ūdi's (Murūj, i, 288):  (?). On other passages in Mas'ūdī relative to the Magyars see Marquart, Streifzüge, and our notes to § 53. [In principle I. Kh. 31, could hardly mention the little-known čaqir, or render čigil by *čghr].

3. In another passage Iṣṭ., 227, reckons from the Pechenegs to the Inner Basjirt 10 days, and from the latter to Bulghār 25 days. This last distance could only refer to the Magyars living north of the Black Sea. A parallel term to Basjirt al-dākhil is Bulghār al-dākhil (i.e. the Danubian Bulghār) mentioned in Iṣṭ., 226 (v.i. § 45). In Mongol times the Magyars occupying their present seats in Hungary were still called , cf. Juvaynī, GMS, i, 325.

4. In a convenient form the texts are synoptically presented in Macartney, o.c., pp. 30 and 42. There are, however, some misprints in the translations and Gardīzī's text is given without the final sentences.

320   Commentary   § 22

between the *Pecheneg [our "Turkish Pecheneg"] country and the Bulghār tribe of Asgil/Ashkil (see § 51). In a later part (B) they describe an extensive Majgharī territory reaching down to the Black Sea. However, in a more detailed description of this southern country the three authors disagree. I.R. and Gardīzī (B I) place the Majgharī between two large rivers disemboguing into the Rūm Sea, and in connexion with this land Gardīzī particularly names the peoples N.nd.r and M.rdat. On the other hand, Bakrī (B 2) says nothing about the rivers and as the neighbours of the Majgharī quotes the  and  undoubtedly connected with the Caucasus, cf. notes to § 50, 4. [1] Contrary to Marquart [2] I am inclined to think that, even supposing that I.R. (B 1 a) has in view the  home of the Magyars near the Azov see, Gardīzī (B 1 6) refers to the  stage of Magyar peregrinations when, expelled by the Pechenegs (a.d. 889), they spent some years in the region of the five great rivers emptying themselves into the north-western corner of the Black Sea, cf. Const. Porph., chap. 38, v.s., p. 313, note 3.

[Additional note. Only in Gardīzī and in the Ḥ.-'Ā. we find traces of the additional source (B 1 b) to which we can assign our details on the southern (*western) frontier of the Magyars, as well as on the V.n.nd.r (§53), Mirvāt  (§46), and perhaps the "Christianized Slavs" (§42, 17.). The source must originally belong to the very last years of the ninth century. It has nothing to do with Muslim b. Abī Muslim al-Jarmī (see notes to § 42) and one particular detail is in favour of its association with the name of Hārūn b. Yaḥyā (see note to § 42, 17.).]

The best introduction to our text is Gardīzī's passage which is not only illustrative for the tradition (B 1) but which also contains details (B 1 b) on the neighbours of the Majgharī found nowhere else except in the Ḥ.-'Ā. Our literal translation follows the text as edited by Barthold, p. 98 (after the Oxford MS.) with the addition of some insignificant variants found in the Cambridge copy (marked C.):

"Between the Bulkār [read as in I.R. and Bakrī: *Pecheneg, cf. also § 6, 45.] country and that of the Asgil who are also of the Bulkār lie the frontiers of the Majgharī. [3] They are a class of Turks and their sālār (has) 20,000 horse. They call this sālār k.nda and this is the name of their greater king, (whereas) the sālār who makes the appointments (shughlhā khwānad) is called jula and the Majgharī do whatever he orders them. They possess a wide plain all covered with grass. Their country is 100 farsakhs by 100 farsakhs. Their country adjoins the Rūm Sea into which flow two large rivers [instead of  read: ] and they live between these two streams () and when (C. ) winter comes those who had gone far from the river (jayḥūn) come

1. Cf. also Mas'ūdī, v.i., notes to § 53.

2. Marquart's attempt to identify these two pairs of names (Streifzüge, pp. 176 and 496) has been followed by the later writers though Marquart himself finally changed his opinion (see notes to § 53).

3. This definition of the territory has in view the northern Majgharī, i.e. the Bashqirs (item A). The rest of the passage seems all to refer to the real Magyars (item B).

§ 22   The Majghari   321

near to it and stay there in winter. They catch fish and live on them. And [with regard to] the river (jayḥūn) which is to their left [we must add that] towards the Saqlāb (country) there is a tribe of Rūm who are all Christians. They are called N.nd.r. They are more numerous than the Majgharī but weaker than they. And of these two jayḥūns the one is called Atil () and the other Dūbā () and when the Majgharī are on the bank of the river they see the N.nd.rians. Above (zabar; C. zīr: 'below') these N.nd.rians on the bank of the river stands a large mountain and a water rises (from it) and flows on its side. Behind this mountain a nation of Christians is found whom they call M.rdāt. Between them and the N.nd.r there is a distance of 10 days. They are a numerous nation. Their clothes resemble those of the Arabs and consist of a turban, a shirt, and a coat (jubba). They have cultivation and possess vines (razān; in C. the text is slightly disturbed). Their water flows on the surface and they have no underground canals (kārīz). And it is reported that they are more in number than the Rum. They are a separate nation. Most of their commerce is with the Arabs. And that (other) river which is on the right of the Majgharī flows to the Saqlāb and thence to the Khazar lands and that river is the largest of the two (va ān rūd az īn har du rūd buzurgtar-ast). The country of the Majgharī is all trees and marshes (ābgīr 'lakes'?) and the soil is damp. They always vanquish the Saqlāb and constantly impose tribute on them and treat them as their slaves. The Majgharī are fire-worshippers and raid the Saqlāb and Rūs and bring captives (barda) from them. They take them to Rūm for sale. [1] These Majgharī are handsome and pleasant looking. They dress in satin (dībā). Their arms are embellished with silver and gold (instead of  read: *). They constantly go to sack the Saqlāb and from the Majgharī to the Saqlāb there is a distance of ten days." [2]

The crucial point is the identification of the two rivers which Gardīzī, perhaps misunderstanding the Arabic original (cf. I.R., 142) but following a regular Persian usage, calls jayh
̣ūn in the sense of "a large river". The author distinctly starts on his location of the N.nd.r from the river flowing "on the left" of the Majgharī, i.e. evidently on their west, because the peoples living beyond it lived in the direction of the Saqlāb, one of the westernmost peoples of Eastern Europe (§ 43). This makes it evident that the river  is one of the rivers of the north-western corner of the Black Sea, and probably Barthold was right in restoring in his text * (Dūnā, "Danube") instead of . [3] As regards the river flowing "to the

1. I.R., 143, mentions  as the point where the Majgharī slave-traders were met by the Byzantine merchants. If this place (cf. § 3, 6. and 8. and § 42. 15.) is Kerch (at the entrance of the Azov sea) there is an indirect indication that I.R. still referred to the Lebedia home of the Magyars.

2. I.R., 143, quotes the distance of 10 days between the Pechenegs and Slavs and, ibid., 142, records the Magyar attacks only on the Slavs. Gardīzī's variants may reflect an influence of his special "N.nd.r-M.rdāt" source (B 1 b).

3. The Khazar king's letter (which also mentions the name V.n.nt.r, cf. § 46) positively applies the name Rwnā/Dwnā to the Danube, cf. Kokovtsov, o.c., pp. 75 and 92, but this document is suspect.

322   Commentary   § 22

right" of the Majgharī, the mention of the Khazars shows that it must be sought in the eastern part of the southern Russian plain. Marquart, Streifzüge, 32, quotes the Hungarian chronicler Simon de Keza according to whom the Hungarians called the Don Etul. [1] This may be a hint for the identification of Gardīzī's Atil which at this place cannot apply to the Volga held at that time by the Khazars. More than this, the Khazars at the zenith of their power controlled the steppes up to Kiev, and so historically even the Dniepr would suit the condition of flowing from the Slavs to the Khazar lands. The name atil was certainly employed in a general sense as is shown by the term  explained as "(the land) between the rivers", see Marquart, Streifzüge, 33. If the element  corresponds to Magyar köz, köze "terra intermedia", the first element is undoubtedly atil taken in the sense of a river (cf. jayḥūn). [2] As Const. Porph., chap. 38, enumerates the five rivers of Atelkuzu we know that the latter comprised the space between the Dniepr and the Sereth. Might Gardīzī's Atil perhaps be an echo of the term Atelkuzu? [3]

Coming now to the Ḥ.-'Ā. we see that its author with regard to the Majgharīterritory followed exclusively the tradition A and entirely disregarded the tradition B. He places the Majgharī near the Ural mountains as the last territory in the series of the northern Turkish lands (§§ 18-22, east to west: Kimäk, Ghūz, Turkish Pechenegs, Khifchākh, Majgharī). This disposition of chapters is still more significant in view of the fact that the southern territories of Eastern Europe (§§ 43-9) are described in an opposite direction (west to east: Saqlāb, Rūs, Inner Bulghārs, Mirvāt , Khazarian Pechenegs, Alān, Sarīr) and that the two series of countries are even separated by an intermediary zone of countries (§§ 50-3) enumerated in a sort of bustrophedon east to west: Khazar, Burṭās, Barādhās, and V.n.nd.r. Cf. Map xii.

Our author undoubtedly represents the same tradition as I.R., Gardīzī, and Bakrī, and in his sources certainly the two different Majgharī homes were found. As in § 22 he proposes to describe the Bashqir country (A), the question is what he has done with the residue of information relative to the Magyars (B)? In the immediate neighbourhood of the Magyar territory Gardīzī mentions the people N.nd.r screened by a mountain from another people M.rdāt. These peoples are also described in our text: the Majgharī

1. Cf. I.Kh., 54, on the Tanais, "the river of the Saqāliba" which the Rūs merchants follow before reaching the Khazar capital.

2. See in Volga-Turkish dialects, Yayq-itili, Vätkä-itili, Aq-idil, &c. Cf. Marquart in Ungar. Jahrb., ix/1, 1929, p. 96. [The word is said to be of Chuvash (< Bulghār) origin.]

3. If Dūbā is the Danube and Atil the Dniepr (or even the Don) it is difficult to call the eastern river the larger of the two. One could perhaps imagine that in the original Muslim report based on Byzantine sources *Dūnā as a more familiar name stood for its less know affluent Sereth, cf. a similar confusion of an affluent with the principal river in § 6, 13. [I.R., 142, only says that "ont of the two rivers is larger than the Jayḥūn", which gives a better sense. Cf. 'Aufī, v.i., p. 324.]

§ 22   The Majghari   323

are the northern neighbours of the V.n.nd.r (§ 53) and the Mirvāt (§ 46) live south of the V.n.nd.r mountain. Consequently the order of enumeration of the peoples is maintained, but the starting-point being different, the Majgharī, V.n.nd.r, and Mirvāt  are disposed in a north-to-south direction, so that, instead of the Majgharī, the Mirvāt  come to be the maritime people on the northern coast of the Black Sea. This basic error [1] will be especially considered in the notes to §§ 53 and 46. See sketch on p. 440.

Having ignored the southern Magyars our author transferred to the inhabitants of the northern territory all the characteristics found in the sources with regard to the "Majghari" and as a matter of fact belonging mostly to the southern Magyars.

Population: 20,000, as in I.R. and Gardīzī.
Country: 150 x 110 farsakhs; Gardīzī and Bakri: 100 x 100 farsakhs; I.R.: "extensive country".
The King's name:  (read: ). I.R. and Gardīzī,  principal king, but  real administrative chief; Bakrī, title .
The Majgharī live on fish. Ditto in Gardīzī, but I.R. and Gardīzī more decisively say that they are fishermen [an important feature for the inhabitants of the region of great rivers].
Rich but vile (?) [not found elsewhere; does the last trait refer to the northern Majgharī?].
Trees and waters, as in I.R. and Gardīzī.
Good-looking, as in Gardīzī.
Victorious wars against "infidel" neighbours. I.R.: dominate over the Slavs (several details on slave trade); Gardīzī: raid the Slavs and Rūs.
Apart from the general epitomizing tendency of our author one seems to discover on his part a desire to smoothe the details not tallying with his general conception (cf. the point on enemies and perhaps fishing). [2] As regards the name of the king, the form  is explained by the confusion of the final  with . The name is certainly  *Jula, cf. Const. Porph., chap. 40, pp. 174-5:  and Hungarian Gyula. Our author omits the name of the chief of executive power k.nda for which Const. Porph. strangely gives  (perhaps: ). The title as it stands in Muslim sources may be connected with that of the dignitary who occupied the third place in the Khazar hierarchy:  ("k.nd.r khāqān" or "the

1. On its disturbing influence, cf. note to § 6, 45.

2. It is true that the Rūs are mentioned as the western and northern neighbours of the Majgharī. In § 6, 45. the river Rūtā rises strangely from a mountain situated between the Majgharī, the Rūs, and the Pechenegs (cf. notes to §§ 20, 47 and 52 on the supposed seats of this people on the right bank of the Volga). [This is a hint at some non-Uralian seats of the Magyars but our author, who does not say a word on the presence of this people near the Black sea, goes halfway in placing the Magyars somewhere near the Oka (?) and imagining that this territory was connected with the Urals. One of the western sources of the Oka is called Ugra (= Hungarian!). According to N. P. Barsov, Ocherk russkoy istoricheskoy geogr., Warsaw 1885, p. 241, Ugra lay on the road connecting the Dniepr with the Volga.] See Map xii.

324   Commentary   §§ 22-3

khāqān's k.nd.r"?), Yāqūt, ii, 436-40 (after Ibn Faḍlān ). Munkácsi, in Keleti Szemle, x, 1909, pp. 179-80, compares it with kündi/kündü which the Altai Turks in quite recent times used to give to their dignitary next in rank to their ruler (zaysan). [1]

[Additional note. In his Streifzüge, 161, 164, Marquart, misled by the idea that the two pairs of names "N.nd.r and M.rdāt" and "Ṭwlās and *Aughas" were identical (cf. § 50, 4.) came to the conclusion that the river  was "Kuban". In Komanen, 99, Marquart was less categorical and wrote with reference to our  (which he found in Toumansky's translation, Zap., x, 1897): "Auf die Frage, welcher Fluss unter dem  zu verstehen ist, gehe ich hier nicht ein. . . . Die Erörterung dieser Frage, welche bekanntlich für die Bestimmung der älteren Wohnsitze der Magyaren von grosser Wichtigkeit ist, ist zwecklos, so lange die Parallelberichte des Muḥammad-i 'Aufī und der Ḥudūd al-'Ālam nicht veröffentlicht sind."

We have commented on the identity of the names Dūbā/Rūtā/Rūthā (§ 6, 45.) as resulting from the comparison of the Ḥ.-'Ā. with the other sources and may add that 'Aufī does not contain any important new data on the subject. Here is the passage on the Magyars (misspelt ) according to Brit. Mus., Or. 2676, fol. 67v.). 'Aufī first quotes the well-known data on the vastness of the Magyars' country (100 x 100 farsakhs), on their 20,000 horse and on the ra'īs called K.nda, adding that the Magyars own tents (khargāh) and wander with their herds. Then he goes on:

"Their lands adjoins the Rūm [= Black] sea. The haunts of this people are on the banks of two rivers (daryā) of which the one is called W.fā and the other Atil both being larger than the Jayḥūn. Between them and the Saqlāb goes on a perpetual war about religion and they are constantly victorious over the (Slavs), and taking prisoners from them carry them to Rūm and sell them. They are continuously in possession of great wealth on account (of this) trade."]

1. In the Shāh-nāma, ed. Mohl, ī, 76, 179, 190, &c., K.nd.r is the name of a Saqlāb hero fighting in the Turanian army on the right hand of the khāqān. [The name of the mountain  (§ 5, 12.) may be connected with the same title. Under § 18 the name was tentatively restored as *Kändür-tagh in view of the name of the river Kängir. But should the analogy be sacrificed, the simplest restoration would be perhaps Kandā'ūr, v.s., p. 308, n. I.]

[Previous] [Next]
[Back to Index]