Hostname: page-component-745bb68f8f-s22k5 Total loading time: 0 Render date: 2025-02-11T10:48:14.279Z Has data issue: false hasContentIssue false
  • English
  • Français

The Islamization of Central Asia in the Sāmānid era and the reshaping of the Muslim world*

Published online by Cambridge University Press:  28 May 2009

D. G. Tor
D. G. Tor*
Affiliation:
Bar-Ilan University, Israel
*
tor@post.harvard.edu
Rights & Permissions [Opens in a new window]

Abstract

The Sāmānid-era drive to Islamize Central Asia led not only to increased Islamic influence within the steppes, but, concomitantly, to the transformation of internal Muslim political life. Developments within the Muslim oecumene that were shaped or influenced by this Drang nach Osten range from the legitimizing of the political fragmentation of the Persianate Dynastic period to changes in Muslim military culture and practice, the successful religious conversion of the Turkic steppe; and growing Turkic influence inside the Sāmānid realms, culminating not only in the downfall of the Sāmānids, but in the end of the era of Iranian political and military dominance and the beginning of a millennium of Turkic political hegemony.

Type
Articles
Information
Bulletin of SOAS , Volume 72 , Issue 2 , June 2009 , pp. 279 - 299
Copyright
Copyright © School of Oriental and African Studies 2009

The Muslim penetration into Central Asia began with the great wars of imperial conquest in the late seventh and early eighth centuries, which resulted in the establishment of the Muslim-ruled sub-province of Transoxiana.Footnote 1 After the major battles between the new Muslim empire and the adjacent Central Asian powers had ended in the early 750s, though, Transoxiana – indeed, Central Asia itself – sank into obscurity as a remote and neglected corner of the empire.Footnote 2 Transoxiana was a mere appanage of Khurāsān; and events from beyond the Muslim–Central Asian limes do not even appear on the horizon of Muslim historical consciousness for the next century, except on the infrequent occasions when large-scale Turkish incursions obtrude themselves upon the notice of the Muslim chroniclers.Footnote 3

The slackening of Muslim interest, and the Muslim offensive, in Central Asia changed dramatically, however, in the pivotal transition era of the autonomous Persianate dynasties. This era began in the latter half of the ninth century with the rise of the Ṣaffārids in the wake of the political implosion of the ʿAbbāsid caliphate, and lasted until the first actual irruption of untamed Central Asia south and west of the Oxus River in 1040, with the triumph of the Saljuqs and their Oghuz Turkmen.

In fact, the Persianate dynastic period constitutes the turning point in the balance of power or mutual influence between the two different cultural worlds, the Perso-Islamic and the Turkic Central Asian. Essentially, it is characterized by the ever-growing importance – military, political, cultural and demographic – of Central Asia inside the Islamic world itself. While the great religious and military headway made by Islam into the heart of Central Asia in the ninth and tenth centuries has long been noted, what has been subject to far less scrutiny is the inevitable concomitant of this phenomenon: namely, the direct and indirect influence of Central Asia within the Islamic realms themselves.

This paper will focus on the Sāmānid period, the most crucial part of this era for our theme, in order to suggest that the new impetus given to the Islamization drive in Central Asia during the Persianate Dynastic era generally, and the Sāmānid period in particular, had significant and lasting ramifications within the Islamic world itself. The Islamizing drive of the Sāmānid era not only changed and shaped some of the most fundamental Muslim political, religious, military and cultural phenomena and institutions – ranging from the process of acquiring political legitimacy to the focus and practice of jihād, the shift in practice of divergent methods of Islamic proselytizing, and the influence, import, and centrality of Turkic Central Asia on internal Muslim political life – but also resulted in an actual mass Islamization of large swathes of the Turkic steppe that had major long-term ramifications which lasted well beyond the Sāmānid era itself.

Jihād in Central Asia as a legitimizing strategy

First, on the political and legal–theological front, the battleground with infidelity in Central Asia gave legitimation to what was a wholly new and unprecedented political phenomenon in the Mashriq: the breaking of Muslim political unity by the establishment of usurping dynasties, for which no place yet existed in proto-Sunni theology.Footnote 4 Until the breakdown of the post-Ma'mūnid period, the political existence of the Muslim umma was in harmony with Islamic theology: to all intents and purposes there was one unitary polity headed by the caliph.Footnote 5 When this caliphal polity ceased functioning, however, its failure posed both a political question and a theological predicament. Eventually, someone, somewhere, would have to take upon himself to defend and rule the Dār al-Islām (or at least part of it) – and the most successful person to do so would be the one who provided not only brute strength but also a legitimizing purpose.Footnote 6 Thus was born a succession of Persianate dynasties in the eastern part of the Empire, which derived their political and theological legitimation from the espousal of devout and militant proto-Sunnism.Footnote 7

The Ṣaffārids were the first of the Persianate dynasties to arise from the remains of the politically moribund ʿAbbāsid caliphate. The Ṣaffārids, however, were short-lived as the dominant power in the Eastern Islamic world: partially, no doubt, because they were the trail-blazers in such an attempt to establish a politically legitimate successor state by carrying the banner of militant proto-Sunnism, primarily through jihād in Central Asia;Footnote 8 and partially because they soon ran afoul with their attempt to restore the earlier strength and unity of the Islamic world by meddling directly with the fainéant ʿAbbāsids themselves.Footnote 9

The Sāmānids adopted the Ṣaffārid model, and also shared the same sort of rise to power by force majeure.Footnote 10 Like the Ṣaffārids, the Sāmānids were upstarts. Starting out as local dihqāns, they began their political climb under al-Ma'mūn's government in the early ninth century: it is presumed that al-Ma'mūn's governor Ghassān b. ʿAbbād appointed the four Sāmānid brothers to local governorships in Khurāsān and Transoxiana: Nuḥ b. Asad in Samarqand; Aḥmad b. Asad to Farghāna; Shāsh and Ushrūsana to Yaḥyā b. Asad; and Herāt to Ilyās b. Asad.Footnote 11 During the subsequent decades of Sāmānid creeping expansion in Transoxiana, Sāmānid officials still recognized the rulers of Khurāsān, be they the Tahirids, the Ṣaffārids, or others, as their overlords.Footnote 12

The founder of the Sāmānid realm as the major ruling power of the Mashriq was Ismāʿīl b. Aḥmad, who replaced his brother Naṣr in 275/888 as the ruler of the realm. The campaign for the Islamization of Central Asia formed the cornerstone upon which his political policy and ambition were built. Ismāʿīl predicated his power on piety and, above all, holy warfare on the infidel Turkic frontier. In the words of one source: “Ismāʿīl was good; he loved Ahl al-'ilm wa'l-dīn, and honoured them, and through their blessing his kingship and that of his descendants endured and their days were long”.Footnote 13

Consequently, as part of this jihādī legitimizing strategy, the prelude to Ismāʿīl's showdown with his Ṣaffārid overlord was the largest-scale raid to date in infidel Central Asia.Footnote 14 Ismāʿīl carried out this major raid in the year 280/893:

In this year the news arrived that Ismāʿīl b. Aḥmad al-Sāmānī raided the Land of the Turks and, so it was reported, captured the city of their king, and his father and the Lady his wife. Upwards of 10,000 others were taken prisoner, a great many of [the infidels] were killed, and an innumerable quantity of animals was plundered…”.Footnote 15

Ismāʿīl, again like the Ṣaffārids before him, took care to publicize both this jihād activity and his allegiance to the caliph by sending extravagant gifts to the caliph to mark the occasion.Footnote 16

Ismāʿīl's pursuit of the jihādī path paid off: when the final showdown between Ismāʿīl and his Ṣaffārid overlord came about in the year 287/900, it was, according to explicit statements in the sources, Ismāʿīl's greater reputation as a holy warrior which tipped the scale between the two men. It is stated specifically in various sources that Ismāʿīl caused the defection of ʿAmr b. al-Layth's generals by representing, to a far greater extent than ʿAmr, the proto-Sunni ghāzī legitimating principle.Footnote 17 Consequently, when the moment of truth arrived, and the Ṣaffārid and Sāmānid forces met in battle:

Ismāʿīl converted [to his cause] the heads of the troops, from among ʿAmr's commanders, by making them afraid of God, saying: “We are ghāzīs and do not possess wealth; while this man [ʿAmr] continually seeks this world, we [seek] the Next. What does he want from us?”Footnote 18

ʿAmr's generals clearly found this argument convincing, since they abandoned their (less ghāzī) ruler in the field.

Thus, we see that the jihādī aspect of the Islamization of Central Asia played a crucial internal role in the Islamic world by supplying a legitimizing rationale to aspiring rulers, and even facilitating changes of ruler. According to the explicit statements in our sources, it was actually the prime factor behind the successful Sāmānid usurpation of the Ṣaffārids. With this transfer of power, the Sāmānid dynasty became the major political force of the eastern Islamic world.Footnote 19 Their realm also, and not coincidentally, became the major new centre for Muslim holy warfare for the faith.

The influence of Islamization in Central Asia on internal Muslim military culture and practice

Here on the military front we see the second major area in which the Islamization drive in Turkic Central Asia of the Sāmānid period possessed internal Islamic significance: the growing importance of Central Asia profoundly influenced military culture and practice inside the Islamic world, changing the Jihād itself. While Jihād, obviously, has always occupied a vital place in the religious, political, cultural and military life of Islam, the characteristic manifestations of Jihād have differed over the years, especially qualitatively and geographically.

Geographically, one can see that certain areas served successively in different eras as loci of Jihād. Generally speaking, the Persianate dynastic era marked a significant shift of the focus of Jihād eastward. Whereas from the time of the Prophet until the ʿAbbāsid collapse in the mid-ninth century the most significant and important arena for Jihād was on the Byzantine frontier, from the late ninth century onward, from the time the Ṣaffārids and Sāmānids raised the languishing banner of Jihād anew under their own auspices, the Central Asian frontier became paramount.Footnote 20

This shift in the locus of Jihād is nicely illustrated by two very different historical anecdotes. In the first, the famous volunteer warrior Abū Isḥāq al-Fazārī asks his fellow holy warrior ʿAbdallāh b. al-Mubārak why he had to come all the way to the Byzantine border to battle Infidels when there were plenty of Turkish ones close at hand in Eastern Iran. Ibn al-Mubārak answered that while the Turks were only fighting about worldly power, the Byzantines were battling the Muslims over their faith, “So which is the more worthy of defense: our world or our faith?”Footnote 21

In contrast, the second anecdote purports to date from the latter part of the ninth century. It is supposedly the eyewitness account of a man who had been present when Aḥmad b. Ḥanbal was being consulted by an aspiring Khurāsāni holy warrior regarding the latter's desire to fight for Islam:

I was with Abū ʿAbdallāh Aḥmad b. Ḥanbal when a Khurāsāni man asked him: “My mother has permitted me to go on a ghazw, and I want to go out to Tarsus. What do you think?” [Ibn Ḥanbal] replied: “Raid the Turks.” I reckon that Abū ʿAbdallāh went to the words of God, may He be honoured and exalted: “Fight those of the infidels who are near you”.Footnote 22

This anecdote proclaims the new ideological and historical reality of the Sāmānid era: the stream of holy warrior volunteers that previously flowed north-westwards to the Byzantine border was diverted, to an ever greater extent, north-eastwards into Central Asia.Footnote 23

More proof of this shift can be found in the fact that the major cities of Transoxiana in the Sāmānid period begin to boast an official position of head of the Jihādi volunteer warriors: in Bukhārā, one reads of a ra'īs al-muṭṭawwiʿa;Footnote 24 in both Bayhaq and also in Bukhārā, of a sālār-i ghāziyān;Footnote 25 and in Samarqand, of a sālār-i muṭṭawwiʿa.Footnote 26 One also sees this on the individual level, in the number of men described in the biographical literature as jihādists of one kind or another who, beginning in the Sāmānid period and continuing throughout the Persian dynastic era, now concentrate on the Central Asian frontier instead of universally flocking to Tarsus.Footnote 27

Yet a third indication of the change in the primary world locus of Jihād to the Central Asian border is the unusually large presence of jihādic warriors attested in the Sāmānid armies, particularly in the early period, when the Sāmānids were still vigorously attempting to establish their legitimacy through Central Asian Jihād.Footnote 28 Thus, the sources repeatedly mention, as a significant military component of the Sāmānid armies, groups such as “the ghāzīs”;Footnote 29 “the ghāzīs of Bukhārā”;Footnote 30 and “the muṭṭawwiʿa”.Footnote 31 Likewise, the sources attest to the proliferation in Transoxiana at this time of the ribāṭ system which had been practised on the Byzantine border.Footnote 32

The Sāmānid era also witnessed what one might call a qualitative change in the practice of Jihād. Jihād from the beginning of Islam until the end of the Umayyad era was a state enterprise.Footnote 33 This is not true for the late Umayyad era and the ʿAbbāsid heyday: then it was conducted largely by volunteer warriors, with caliphal involvement limited to sporadic show campaigns by them and their relatives.Footnote 34 The establishment of the autonomous Persianate dynasties in the ninth century ushered in a new phase: since Jihād was the best method by which the new autonomous Persianate dynasties could establish their legitimacy and religious credentials, one sees a true blending of the two different patterns of religious warfare; Jihād in this period harnessed the manpower of the volunteer warriors together with a revival of the centralized governmental practice of Jihād.Footnote 35 Due to the location of these dynasties, beginning with the ephemeral Ṣaffārid realm, it was natural that they should make into the cornerstone of their rule fighting against the infidels in Central Asia, rather than the far-away Byzantine enemy. This new policy was to have truly significant consequences, both short- and long-term.

The Islamization breakthrough in Turkic Central Asia

As a result of this politico-religious strategy, the Sāmānid era marked a turning point in the process of Central Asian Islamization. The Sāmānid era witnessed, for the first time, the establishment of one of the greatest cultural, political and military centres of the Islamic world in Central Asia itself, on the border of the eastern non-Islamic lands. Transoxiana was transformed from a remote provincial backwater into the most powerful and important centre of the Islamic Empire, in virtually every field of endeavour. Thus, or so it has been conjectured, it was due to a combination of Sāmānid weapons; the efflorescence of the ribāṭ system which combined warfare and daʿwa; the political, economic, and cultural blossoming and prestige of the Sāmānid realms; and the constant missionizing and contact of all kinds between the Muslim world of Central Asia and the non-Muslim one, that the limes shifted steadily to the North and East.

It was probably due in large part to the enormous Sāmānid vitality and prestige, together with the incessant military campaigns and a sustained colonization programme within the pagan areas lying near the Sāmānid borders, that during this time the tipping point was reached in the Islamization of the contiguous Turkic peoples, culminating in the mass conversion of the Qarakhanids in the mid-tenth century.Footnote 36 For the first time in Islamic history, large-scale Islamization transpired not under military conquest, but apparently through various forms of engagement – military, cultural, commercial, social and religious.

Curiously, there is a remarkable and almost complete dearth of specific information regarding any actual missionary activities or individuals in Turkic Central Asia during Sāmānid times, despite the widespread view that the conversion of the Turks was due to Sufi proselytizing. The only concrete figures who have ever been mentioned in this connection are the two tenth-century men noted by Barthold, Abū'l-Ḥasan Muḥammad b. Sufyān al-Kalamātī of Nishapur, and Abū'l-Ḥasan Saʿīd b. Ḥātim al-Usbānīkathī.Footnote 37 Regarding the first figure, al-Kalamātī, Barthold seems to place upon him an interpretive construction that his biography does not justify; whereas Barthold construes Samʿānī's biography as stating that al-Kalamātī “passed into the service of the ‘Khān of Khāns’”, all that the biography actually says is that “He arrived in Jūzjānān, became attached to its rulers (Ittaṣala bi-awlaʼika al-salāṭīn), and died there before 350/961”.Footnote 38 The use of the plural form suggests that al-Kalamātī's connections were with local lords rather than with the supreme overlord. In fact, it is not even clear that al-Kalamātī was in any way associated with pagans or missionary activity; and he is nowhere described as a Sufi.

Regarding Abū'l-Ḥasan Saʿīd b. Ḥātim al-Usbānīkathī the evidence is a bit more promising. First, it is certainly suggestive that all five biographical entries for this nisba,Footnote 39 without exception, prove to be men who lived in the fourth/tenth century, at the very time when the nearby Turkic areas were being Islamized; and four of those five religious figures died in the decade between 370 and 380/980–990.Footnote 40 Abū'l-Ḥasan himself is stated to have lived in Samarqand, until “He went out to the land of the Turks sometime before [the year] 380/990, and returned from there to Usbānīkath and died there in those days”.Footnote 41 Once again, there is no textual evidence that any of these figures was a Sufi, nor that they engaged in missionary activity. In short, one must regard the assertions found in the scholarly literature regarding the actual specifics of Islamic proselytizing in the Turkic steppe during the critical ninth and tenth centuries as conjectural and speculative.Footnote 42

Another possible non-military facilitator of the Islamization of Central Asia, noted long since by Barthold, but not sufficiently appreciated or emphasized since, was Sāmānid-era Muslim mercantile settlement in the non-Muslim Central Asian steppe as an agent of Islamization:

… Alongside of the victorious Sāmānid expeditions and independently of them, peaceful colonization of towns [des villes] was pursued by the emigrants from Transoxiana … the Muslims continued the past colonizing activity of the Soghdians. Thus three Muslim towns were founded along the lower course of the Syr-Darya: Jand, Khuvara, and Yangikent … According to what the Arabic geographers say, these towns were inhabited by Muslims, but were under the domination of non-Muslim Oghuz Turks. This fact proves that these towns were not created in the conquered region by the Sāmānids, but that they were colonies founded with the consent of the indigenous Turks by emigrants from Transoxiana. The town of Talas that had been conquered by the Sāmānids, as well as the new colony of Yangikent, continued their commercial expansion into Central Asia.Footnote 43

Once again, this is an aspect of the Islamization process that has been largely neglected by researchers (who have tended to discuss mercantile contacts rather than actual colonization),Footnote 44 and regarding which there is also tantalizingly little documentary evidence.

The growing Turkic influence inside the Sāmānid lands

The growing Muslim engagement with Central Asia that was attendant upon the aggressive Sāmānid Islamization effort had another far-reaching effect, and that was in the corresponding heightened interest of the Central Asian Turkic peoples in, and their influence upon, the internal affairs of the Sāmānid lands. In other words, the growing involvement of the Sāmānids became a two-way street – it entailed the Turkic polities of Central Asia playing a role in internal Sāmānid affairs. This was true for the periods both before and after the official conversion to Islam of the Qarakhanids in the mid-tenth century, albeit in different ways.

The Turkic response to the heightened Sāmānid Jihād campaigns took the form both of actual large-scale military campaigns and, more commonly in the time of the flowering of Sāmānid strength, of the exploitation of internal factionalism and disgruntled rebels within the Sāmānid domains. With regard to the military response of massive campaigns, the new Sāmānid power was first tested shortly after having defeated the Ṣaffārids, while they were engaged both in trying to bring under control the former Ṣaffārid lands of Khurāsān, and in battling ʿAlid rebels in ṬabaristānFootnote 45 (thus showing, incidentally, that the Turks were well aware of developments inside the Islamic empire):

In [the year 291] the Turks set out in a great, innumerable company to Transoxiana; there were in their army 700 Turkish tents [qubba turkiyya], and that was only of the chiefs among them. Ismāʿīl b. Aḥmad sent to them a great army, and a numerous company of volunteer holy warriors [muṭṭawwiʿa] followed them.

This army managed to defeat the pagan Turks in a surprise attack: “[The Muslims] killed a number of them, so large as to be uncountable; the rest were routed, and their camp was plundered; and the Muslims returned in peace, pillaging”.Footnote 46 Ismāʿīl, apparently deciding that his deterrence needed bolstering,Footnote 47 set out less than two years later on a successful large-scale military counter-campaign in non-Muslim Turkestan.Footnote 48 After this second stinging defeat, the Sāmānids' Turkic neighbours apparently did not, at least according to the Muslim sources, repeat such a large-scale frontal assault until the very close of the Sāmānid era, a point to which we shall return presently.

Even more intriguing are the direct and indirect ways in which the originally non-Muslim Turkic powers contiguous to the Sāmānid domains affected internal politics in the Sāmānid realms. Before examining this issue, however, we must first address one of the oddest mysteries of this period: who precisely were the Turkic rulers and political groupings against whom the Sāmānids were campaigning? At some point, obviously, it was the Qarakhanids or Īlig-Khāns, yet we know next to nothing about when and how during the course of the ninth or possibly even tenth century the Qarakhanid confederation arose and came to dominate Turkestan.Footnote 49 The most widely accepted educated conjecture is Pritsak's, according to which the Qarakhanids are to be identified with the Qarluq rulers who by 840 ruled a territory which “extended down to Ferghana and the Syr Darya”.Footnote 50

Even if this Qarluq kingdom was indeed that polity which modern scholars identify with the tenth-century Qarakhanids, exactly which areas they ruled, at what point they ruled them, and what type of rule this was, all remain exceedingly hazy.Footnote 51 This strange silence in our sources regarding Turkic movements and the coming of the Qarakhanids has been noted and remarked upon; in Pritsak's words, “It is curious that these events of so much importance in the history of Eurasia should have found so little echo in the Islamic sources”.Footnote 52

The most probable explanation for this silence is that the Muslims – and their Sāmānid governors – were suffering losses on their borders; losses which the Sāmānid historiographers preferred to gloss over.Footnote 53 This conjecture has circumstantial evidence to support it: namely, the uncertainty surrounding the question of the borders between the Sāmānid and Qarakhanid realms;Footnote 54 the puzzling lack of much Muslim territorial gain to show for all the reported signs indicating incessant and sustained border warfare between the two sides; and, finally, statements that would seem to indicate that the Sāmānids ruled at some point areas that we suddenly discover, in the tenth century, to be under Qarakhanid rule.Footnote 55 Apparently, the Sāmānids – or at least the court historians being paid by them – did not wish to memorialize embarrassing Sāmānid setbacks.

Regardless of what the border situation may have been, though, the influence of Turkic Central Asia upon internal Sāmānid affairs can be felt almost from the beginning of the Sāmānid rise to power. This is apparent first and foremost in the very nature of the power and the kind of rule that the Sāmānids were able to achieve. It has long been noted that much of the Sāmānid territory, particularly its northern and easterly reaches, was not under direct Sāmānid rule, but rather functioned on a vassalage basis.Footnote 56 What has been less commonly noted is that it was surely at least in part the proximity of the opposing Qarakhanid power that prevented the Sāmānids from strengthening their hold on peripheral areas; the local rulers were able to play off the two sides against one another, to their own distinct advantage.

An excellent example of this tactic can be found in the revolt of Abū ʿAlī Muḥtāj, head of the clan functioning as local rulers of marchland Chaghāniyān and an important political figure in the mid-tenth century Sāmānid realms.Footnote 57 When Abū ʿAlī was dismissed from his post for abuse of power, he decided to revolt; what is revealing about this revolt is that when things went badly for Abū ʿAlī he was able to take refuge, either in what is described variously as “Turkestan”Footnote 58 or with someone described as “the Amīr of Khuttal”.Footnote 59 That is, because Abū ‘Alī was able to fall back upon either the Qarakhanid lands or those of an at least semi-independent marcher lord, he was not only beyond Sāmānid reach, but able to worry the central authorities to the extent that he could negotiate a settlement quite favourable to himself.Footnote 60

The involvement of Central Asian forces from beyond the limes is also apparent in internal Sāmānid family affairs. Even at the very founding of the dynasty's real power, in the time of Ismaʿīl b. Aḥmad, when Ismāʿīl ceased paying tribute to his brother, an important force in the fighting that erupted between the two was “the Turks of Isfījāb”.Footnote 61

Another example of Central Asian abetting of unrest in the Sāmānid dominions can be seen in the events that took place during the reign of the Amīr Naṣr b. Aḥmad. Shortly after Naṣr's accession, his great-uncle Isḥāq b. Aḥmad b. Asad and one of the latter's sons had revolted.Footnote 62 Although this revolt was suppressed, a decade later another of the sons of the erstwhile rebel, Ilyās b. Isḥāq, raised in turn the standard of rebellion.

Using Farghāna as his base, Ilyās “recruited among the Turks”, supposedly gathered 30,000 horsemen, and advanced on Samarqand. Eventually, after being defeated, Ilyās fled to Kāshghar, where he was hosted by one “Tughān-tegīn the Dihqān of Kāshghar”.Footnote 63 From Kāshghar Ilyās b. Isḥāq attempted to launch an irredentist attack on Farghana, but after being “beat[en] … yet again, he returned to Kāshghar”. This particular drama came to a close when Ilyās finally gave up, returned to the Sāmānid dominions, and was reconciled with Naṣr.Footnote 64

The significance of the outside Turkic involvement in this episode lies in the fact that these largely unidentified outsiders: a) supplied actual military manpower to aspiring rebels against Sāmānid authority; b) gave repeated refuge, apparently with impunity, to such rebels when events did not go their way;Footnote 65 c) provided a political insurance policy to rebels; as we saw in both Ilyās's case and in the earlier one of Abū ʿAlī Muḥtāj, rebels who had outside Turkic backing were able to negotiate excellent terms for themselves when they wished to return to the fold – apparently, there were no negative repercussions because the Sāmānid Amīrs did not dare punish them, since they could always turn once again to the Turkic enemy on the other side of the border.

Pressure could work both ways, however. In 332/944 a rebel Amīr, ʿAbdallāh b. Ashkām, “quarrelled with the Amīr Nūḥ, and withdrew to Khwarizm”. Once there, Ibn Ashkām “entered into correspondence with the king of the Turks and sought protection from him”. According to this story, the “king of the Turks” – presumably the Īlig-Khān – had a son who was being held hostage in Bukhārā. Nuḥ, accordingly, wrote to the Īlig-Khān, promising to free the Īlig-Khān's son if the latter would seize the rebel. The strategy worked; “The king of the Turks responded affirmatively to this, and when Ibn Ashkām realized what the situation was he returned to Nuḥ's obedience and left Khwarizm. Nuḥ was good to him, honoured him and forgave him”.Footnote 66 Note that even in this case, Nuḥ still found it wiser to conciliate the erstwhile rebel than to punish him; the marcher lords were too vital and unpredictable an element for the Sāmānid rulers to treat them with anything less than the utmost circumspection.

Mass conversion in Central Asia and the downfall of the Samanids

The great culmination of the entire Islamization process in Central Asia was undoubtedly the conversion of the Qarakhanids, which, as previously noted, took place en masse sometime around the middle of the tenth century.Footnote 67 Ironically, one of the most significant consequences of the outstanding success of the Islamizing drive during the Sāmānid era was the heightened interference by the erstwhile pagans in Sāmānid affairs it facilitated; the Islamizing of the Īlig-Khāns, by bringing them into the pale of the Muslim political world, thereby enabled them to pose a threat to the Sāmānids in a manner which would have been difficult if not impossible before their conversion.Footnote 68 In Frye's words: “So the Muslim Turks accomplished what the pagan Turks could not have done – the conquest of the Sāmānid kingdom”.Footnote 69

The process whereby the Sāmānid realm was overwhelmed by those very Turks whom it had laboured so long to bring into the fold was a gradual one, and there is much confusion in the sources regarding the exact course of events.Footnote 70 What is clear from all the versions is that after the conversion of the Qarakhanids, the Turkish rulers began for the first time intervening directly rather than by proxy in the Sāmānid dominions. Thus, in the early 990s, we see the first Qarakhanid direct military intervention. The episode began when, tempted by internal Sāmānid political turmoil,Footnote 71 “Abū Mūsā Hārūn b. Īlig Khān came from Turkestan”, and attacked first Iṣfījāb;Footnote 72 then, his appetite whetted, steadily advanced, until in 382/992 he conquered Bukhārā. Falling ill, the Khan retreated to Kāshghar, where he died soon thereafter.Footnote 73

There exists an alternative version of these events.Footnote 74 According to Ibn al-Athīr, there were in fact two separate, successive attempts made by the Qarakhanids in the early 990s on Bukhārā: one in 382/992, which ended in military defeat; and the second, more successful, attempt the following year:

Bughrā-Khān Īlig,Footnote 75 King of the Turks, went with his armies to Bukhārā. The Amīr Nuḥ b. Manṣūr sent against him a great army, but Īlig met them and defeated them, and they returned, broken, to Bukhārā, with him in their footsteps. Then Nuḥ went out himself with the remnant of his army and met him; they fought a fierce battle, which resulted in the rout of Īlig, who returned in defeat to Balāsāghūn, which was the seat of his rule.Footnote 76

This first foray ended in failure, but it was the harbinger of more such events to come. The very next year, 383/993, the aforementioned Bughrā-Khān renewed his onslaught upon the Sāmānids, this time with the fifth-column assistance of various Sāmānid magnates. The outcome of this attempt was the Qarakhanid defeat of the Sāmānids and the occupation of Bukhārā by Bughrā Khān.Footnote 77 It was during this second attempt, according to Ibn al-Athīr, that Bughrā Khān fell mortally ill, retreated from Bukhārā to “the land of the Turks”, and died.Footnote 78 What is important about Ibn al-Athīr's account is that, by contradicting the familiar picture of Sāmānid decay (Nuḥ b. Manṣūr is actually able to defeat Bughrā-Khān in battle), it provides an indication just how critical the issue of baronial loyalty was for the Sāmānids, and how skilfully the Qarakhanids were able to exploit the self-interest of various Sāmānid amirs.

The culmination of the process of reverse influence that we have noted was the actual overthrow of the Sāmānids a few years after Bughrā Khān's dry run, at the hands of the Qarakhanids in 389/999.Footnote 79 In effect, the ever-increasing Islamization of Central Asia brought Central Asia itself, en masse, inside the Islamic world, until the last ethnically Persian dynasty to hold sway as the regional power of the eastern Islamic world was finally overwhelmed.

Longer-term influence of the Sāmānid-era Islamization drive

It should be noted that the significance of the Islamization drive that took place in Central Asia during the Sāmānid period did not end with the downfall of the Sāmānids; rather, its ramifications continued to be felt well into the eleventh century. Thus, at the downfall of the Sāmānids in 999, their realms were split between two Muslim polities representing Central Asian rulership of different kinds inside the Islamic world:Footnote 80 the Turkic Qarakhanids, who apart from their religion were a product of Turkic Central Asian rather than Perso-Islamic culture;Footnote 81 and the Ghaznavids, whose founder, Sebuktegin, was a Turkic mamluk of the old assimilationist model.Footnote 82 In short, the unforeseen outcome of the success of the Sāmānid-era drive to bring Islam to Central Asia was the end of the era of Iranian political and military dominance, and the beginning of the Turkic political dominance that was to hold sway in the Islamic world for the next millennium.

The second longer-term ramification of the Sāmānid-era Islamization drive in Central Asia is the practical effect it had on the successor Ghaznavid polity, heir to both the old Islamic lands inside the limes, as well as to newly Islamized areas in a very different part of Central Asia: the south-eastern area. Maḥmūd of Ghazna was himself a product of the interaction between the Islamic world and Central Asia: he had a double Central Asian heritage as, on the one hand, the son of a Sāmānid mamluk military commander – the old safely Perso-Islamized slave model, representing Central Asia tamed by Muslims;Footnote 83 and on the other, his mother's heritage, and the area where he was raised, was Zābulistān, an only very recently Islamized area on the remotest border of the Islamic oecumene in south-eastern Central Asia.Footnote 84 Maḥmūd had already demonstrated a radical break with the old model of Central Asian assimilation when he overthrew his own masters; for the first time in the area between the Mediterranean and India, we see the son of a Turkic slave not merely contenting himself with controlling the ruling dynasty, but with replacing it.Footnote 85 By this act, Maḥmūd passed from the old subservient model of his father to the new one of the Turkic assumption of political control of the Islamic world.Footnote 86

Equally significantly, the developments of the Sāmānid period affected the policy and direction of the Ghaznavid state. Like their Ṣaffārid and Sāmānid predecessors, the Ghaznavids needed to establish their Jihādi credentials.Footnote 87 In fact, given the Sāmānid success at having established their religio-political credentials, Maḥmūd found himself in even greater need of the Jihādi legitimizing principle when he usurped the Sāmānid lands and broke his obedience to his erstwhile overlords. This is probably at least part of the explanation for his almost single-minded focus on Jihād, to the point where he became one of the most famous jihādic warriors in Islamic history, known by the sobriquet Maḥmūd Ghāzī.Footnote 88

Again as a result of Sāmānid-era success, Maḥmūd had to turn his eye to new infidel pastures; the Sāmānid Islamization of Inner Asia deflected Maḥmūd of Ghazna's attention towards the Indian subcontinent and areas of eastern Central Asia adjoining it. While much attention has been paid by scholars to Maḥmūd of Ghazna's conquests in India, far less has been paid to the continuing importance of Central Asia in the Ghaznavid realms and military policies.Footnote 89 Yet one of the prime areas where this Jihād was directed were the last remaining pockets of infidelity in the areas that became Afghanistan and Kashmir; that is, there was a direct continuation of the Islamization drive, merely focused on a different part of Central Asia.Footnote 90 This is surely one of the reasons why it was also Maḥmūd, rather than the Qarakhanids, who inherited the Sāmānids' mantle as prime magnet for aspiring jihādists.Footnote 91

The Sāmānid-era Islamization successes had another, paradoxical, effect on the Jihād in Maḥmūd's time: tensions during the earlier part of his reign with the now-Muslim Qarakhanids, who had their eye on the old Sāmānid lands in Khurāsān, hindered the Jihād against, and conquest of, India; the Īlig-Khān's invasion of Khurāsān in 396/1006, when Maḥmūd was busy on campaign in Mūltān, for example, forced Maḥmūd to rush back home.Footnote 92

In conclusion, the Islamization process in Central Asia during the Persianate dynastic period, culminating under the Sāmānids, had an effect inside the Islamic world that was profound and lasting. First and foremost, the Jihād in Central Asia provided a legitimizing rationale for the new proto-Sunni dynasties that broke the political unity of the caliphate. In fact, this legitimizing rationale was quite possibly the deciding factor in facilitating the successful establishment of the new paradigm and norms of rulership that subsequently took form during the ninth through eleventh centuries.

Second, the drive for Islamization in Central Asia during this period changed the nature of Jihād itself, both qualitatively and geographically, by restoring intensive governmental involvement in the fighting and by changing its geographic focus. Third, growing engagement with the Turkic infidels outside the Islamic border led, in turn, to the increasing role of the neighbouring Turks in internal Sāmānid affairs. Fourth, the massive investment in and fostering of Islamization by the Ṣaffārids and Sāmānids finally caused a breakthrough in the Islamization process, during which most of the Turkic peoples contiguous to the Muslim world adopted Islam. Finally, this mass Islamization led in turn directly to the takeover of the Islamic world by the Turks; the immediate Turkic successors to the Sāmānids, the Qarakhanids and the Ghaznavids inaugurated a period of Turkic political dominion in the Mashriq that was to last a thousand years.

Footnotes

*

This article was first presented at the conference “L'Islamisation de l'Asie centrale: Pratiques sociales et acculturation dans le monde turco-sogdien”, at the Collège de France, November 2007. The author is grateful to Étienne de la Vaissière for having elicited the writing of it, and to all those who commented upon it, particularly Jürgen Paul. Additionally, the author thanks David Morgan for his comments upon the article draft.

References

1 On the establishment of Muslim rule in Transoxiana, see Barthold, W., Turkestan Down to the Mongol Invasion, third edition, tr. Minorsky, T., ed. Bosworth, C. E. (Taipei: Southern Materials Center, 1968), 180–93Google Scholar.

2 In the words of Golden, Peter: “The Arabs, however, did not seek to establish themselves deep in the steppe. Instead, they retired to the ribāṭs (border forts), oasis city-states and rich, urban trappings of the Khorasanian towns”, “The Karakhanids and early Islam”, The Cambridge History of Early Inner Asia, ed. Sinor, Denis (Cambridge: Cambridge University Press, 1990), 344Google Scholar. On the course and significance of the clashes of the 730s with the disintegrating Türgesh kingdom see Gibb, H. A. R., The Arab Conquests in Central Asia (London: The Royal Asiatic Society, 1923), 59–85Google Scholar, especially 84–5. On the earlier and later decades see Barthold, Turkestan, 187–96. For a succinct discussion of the significance of the Battle of Talas, see Golden, “The Karakhanids and early Islam”.

3 E.g. the laconic remark in Abū Jaʿfar Muḥammad b. Jarīr al-Ṭabarī, Ta'rīkh al-umam wa'l-mulūk, ed. Ibrahim (Beirut: Rawā'iʾ al-Turāth al-ʿArabī, n.d.), 8: 580: “The Tokuz Ghuzz came to Ushrūsana”.

4 The term proto-Sunni is borrowed from Juynboll, G. H. A., “An excursus on ahl al-sunna in connection with Van Ess, Theologie und Gesellschaft, vol. IV”, Der Islam 75, 1998, 318–30Google Scholar.

5 On the religio-political significance of the caliphate until al-Ma'mūn's time, see Crone, Patricia, God's Rule: Government and Islam (New York: Columbia University Press, 2004), 21–3Google Scholar. On the loss of the early religious authority of the caliphate, ibid., 130–33. For the consequent political crumbling of the caliphate resulting from this process see Tor, D. G., “Privatized Jihād and public order in the pre-Seljūq era: the role of the Mutaṭawwiʿa”, Iranian Studies 38/4, 2005, 555–73CrossRefGoogle Scholar.

6 On this problem see Paul, Jürgen, The State and the Military: The Sāmānid Case (Papers on Inner Asia 26. Bloomington, IN: Indiana University Research Institute for Inner Asian Studies, 1994), 6–7Google Scholar; and Mottahedeh, Roy P., Loyalty and Leadership in an Early Islamic Society, revised ed. (London: I. B. Tauris, 2001), 180–81Google Scholar.

7 Scholars such as Julie Meisami and Elton Daniel have correctly noted the Sāmānid need to legitimize their rule, while paradoxically not attributing their stringent adherence to an “ideologically ‘correct’ version of Islamic history and doctrine” to this need, but rather to what they describe as a wholly unrelated aim: “to counter the teachings of various heterodox and sectarian groups”, Meisami, Julie, Persian Historiography to the End of the Twelfth Century (Edinburgh: Edinburgh University Press, 1999), 24Google Scholar. The present author concurs with their analysis, but believes that Daniel overlooked how integral a part of the legitimizing project the Persianate dynastic militant proto-Sunni stance was; this is why we find it so conspicuously displayed by all three of the major Persianate dynasties – Ṣaffārids, Sāmānids and Ghaznavids.

8 In Bosworth's words, “Campaigns against the Khārijites and the infidels of eastern Afghanistan gave the Ṣaffārids prestige in the eyes of the orthodox …”, Bosworth, C. E., “The Ṭāhirids and Ṣaffārids”, The Cambridge History of Iran. Volume 4: The Period from the Arab Invasion to the Saljuqs, ed. Frye, R. N. (Cambridge: Cambridge University Press, 1975), 112Google Scholar.

9 See Tor, D. G., Violent Order: Religious Warfare, Chivalry, and the ʿAyyar Phenomenon in the Medieval Islamic World (Istanbuler Texte und Studien 11. Würzburg: Orient-Institut Istanbul, 2007), ch. 5Google Scholar, “The ʿAyyār and the Caliph”, 159–83.

10 Obviously, it is the Sāmānids rather than the Ṣaffārids who managed to appropriate the status of militant Sunni Persianate dynastic paradigm to themselves. While it is outside the scope of this paper to delve into this question, we can note briefly that the reason the Sāmānid jihādī dynastic state became the paradigm rather than the Ṣaffārid one on which they were modelled was due to three main factors: 1) greater Sāmānid dynastic longevity, coupled with the fact that they were the victors over the Ṣaffārids, and could therefore perform a thorough damnatio memoriae; 2) Their far lower level of tension with the ʿAbbāsids, which in turn was the outcome of the twin happy facts that: a) they shared no common border with the caliphs; and b) the latter were most satisfactorily neutralized during the course of the early tenth century by their own precipitous political decline; 3) above all, the Sāmānid mobilization of historical writing in service of their own legitimating project. On this last point, see Meisami, Julie, “Why write history in Persian? Historical writing in the Sāmānid period”, Studies in Honour of Clifford Edmund Bosworth. Volume II: The Sultan's Turret: Studies in Persian and Turkish Culture, ed. Hillenbrand, Carole (Leiden: Brill, 2000), 348–74Google Scholar.

11 Abū Saʿīd ʿAbd al-Ḥayy b. Ḍaḥāk b. Gardīzī, Maḥmūd, Tārīkh-i Gardīzī, ed. Ḥabībī, ʿAbd al-Ḥayy (Tihrān: Dunyā-yi Kitāb, 1363/1984), 322Google Scholar; ʿIzz al-Dīn Abū'l-Ḥasan ʿAlī b. al-Athīr, Muḥammad Ibn, al-Kāmil fī'l-tā'rīkh, ed. Tornberg, (Beirut: Dār Ṣādir [reprint of the Leiden 1863 edition], 1399/1979), 7: 279Google Scholar; Abū'l-Ḥasan ʿAlī b. Bayhaqī, Zayd (Ibn Funduq), Tārīkh-i Bayhaq, ed. Bahmanyār, Aḥmad (Tihrān: n.p., 1361/1942), 68Google Scholar; Ḥamdallāh b. Qavīnī, Abī Bakr Mustawfī, Tārīkh-i Guzida, ed. Navā'ī, ʿAbd al-Ḥusayn (Tihrān: Amīr-i Kabīr, 1362/1943), 377Google Scholar; Mīrkhwānd Muḥammad b. Khavāndshāh, Sayyid Burhān al-Dīn, Tārīkh Rawḍat al-ṣafā' (Tihrān: Markazī Khayyām Pīrūz, 1339/1959), 4: 30Google Scholar, although it grossly aggrandizes the Sāmānid position after the death of Ṭāhir b. al-Ḥusayn.

12 According to al-Narshakhī, Tārīkh-i Bukhārā, ed. Mudarris Raḍavī (Tihrān: Sanā'ī, 1351/1973), 93 (repeated again on 94), Yaʿqūb b. al-Layth's name was recognized in the khuṭba until the latter's break with the caliph in 874–875; that is, the Sāmānids, even in the early period of Ismāʿīl's rule, recognized the governor of Khurāsān as their overlord.

13 Ibn al-Athīr, al-Kāmil fī'l-tā'rīkh 7: 282. On the same page Ibn al-Athīr also bolsters Ismaʿīl's religious credentials by describing how he was honoured by seeing the Prophet in a dream, and by depicting the great respect Ismāʿīl showed towards a Shafiʿite ʿālim.

14 Barthold notes this change in policy without, however, drawing the present author's conclusions regarding its purpose and function: “As we have already seen, the Sāmānids renounced the defensive policy of the previous governors of Khurāsān and Transoxiana. They ceased maintaining the walls that served to defend the cultivated lands from the nomads, and commenced … military expeditions into the steppe regions”. Barthold, , Histoire des Turcs d'Asie Centrale, French, tr. by Donskis, M. (Paris: Librairie d'Amérique et d'Orient, Adrien-Maisonneuve, 1945, 48)Google Scholar. Serguey Kliachtorniy (“Les Samanides et les Karakhanides: une étape initiale de la géopolitique impériale”, Cahiers D'Asie Centrale 9, 2001, Études Karakhanides, 39), however, comes to the same conclusion as the present author regarding this campaign: “Aux yeux du monde musulman, Ismâʿîl devint le grand combattant pour la foi; celui qui propagea l'islam loin à l'est. En fait, cette campagne prépara politiquement et matériellement la guerre du Khorassan qui commença cinq ans plus tard”.

15 Al-Ṭabarī, Ta'rīkh al- Ṭabarī, ed. Muhammad Ibrahim (Beirut: Dār al-Turāth, n.d.), 10: 34, repeated almost verbatim in Ibn al-Athīr, al-Kāmil, 7: 464–5, and Mīrkhwānd, Rawḍat al-ṣafā' 4: 32; described independently in al-Narshakhī, Tārīkh-i Bukhārā, 102, mentioning the Islamization of Ṭarāz as its objective. This raid was described by Bosworth (s.v. “Sāmānids”, EI 2 8: 1026) as follows: “One role which Ismāʿīl inherited as ruler of Transoxiana was the defence of its northern frontiers against pressure from the nomads of Inner Asia, and in 280/893 he led an expedition into the steppes against the Qarluq Turks, capturing Talas and bringing back a great booty of slaves and beasts”.

16 Anon, . (attr. al-Qāḍī al-Rashīd b. al-Zubayr), Kitāb al-dhakhā'ir wa'l-tuḥaf, ed. Ḥamīdallāh, Muḥammad (Kuwait: al-Turāth al-ʿArabī, 1959), 42Google Scholar. On comparable Ṣāffārid behaviour, see D. G. Tor, Violent Order, chapters 3, 4 and 6.

17 Thus, for instance, Ismāʿīl took care to write to ʿAmr and his court beforehand, publicizing his efforts in pious warfare: “God is between you and me. I am a border man, drawn up in battle array against the Turk; my clothing is coarse, my men are rabble without pay, thus you wrong me [by not leaving me to my pious occupations]”. Shams al-Dīn Muḥammad b. Aḥmad b. al-Dhahabī, ʿUthmān, Siyar a’lām al-nubalā', ed. al-Arna'ūṭ, Shuʿayb et al. (Beirut: Mu'assasat al-Risāla, 1419/1998), 12: 517Google Scholar.

18 Anon, ., Tārīkh-i Sīstān, ed. Bahār, M. (Tihrān: Kitābkhānah-i Zuwwār, 1314/1935), 256Google Scholar.

19 In the words of Mustawfī Qavīnī, Tārīkh-i Guzida, 347, “pādshāhān-i jihān”.

20 Soucek has also noted this phenomenon: “The Arabs subsequently transmitted this zeal to the converts of the newly conquered Central Asian territories, so that when the caliphate began to lose its youthful vigor, the jihād was no longer led by them but by a new Iranian dynasty of Transoxania, the Sāmānids”. Soucek, Svat, A History of Inner Asia (Cambridge: Cambridge University Press, 2000), 70CrossRefGoogle Scholar.

21 Found in Ibn al-ʿAdīm's Bughyat al-talab, cited and translated in Cook, D., “Muslim apocalyptic and Jihād”, Jerusalem Studies in Arabic and Islam 20, 1996, 98Google Scholar. For similar traditions on the superiority of fighting the Byzantines rather than pagans see Abū Dā'ūd Sulaymān b. al-Sijistānī, al-Ashʿath, Kitāb al-Sunan: Sunan Abī Dā'ūd, ed. ʿAwwāma, Muḥammad (Beirut: Mu'assasat al-Rayyān, 1998), 3, 205Google Scholar, in the section “Kitab al-Jihād”, chapter 8, “In praise of fighting the Byzantines above all other nations”, traditions #2482–83.

22 Quran 9: 123. The source of the anecdote is Abū'l-Ḥusayn Muḥammad ibn al-Baghdādī al-Ḥanbalī, Abī Yaʿla, Ṭabaqāt al-fuqahā' al-Ḥanabila, ed. ʿUmar, ʿAlī Muḥammad (Cairo: Maktabat al-Thaqāfa al-Dīniyya, 1419/1998), 1: 87Google Scholar.

23 Pace Golden's, Peter assertion, An Introduction to the History of the Turkic Peoples: Ethnogenesis and State-Formation in Medieval and Early Modern Eurasia and the Middle East (Wiesbaden: Harrassowitz, 1992), 211–12Google Scholar: “Muslim sources often depict the entire frontier as the scene of Jihād. Strictly speaking, this was undoubtedly an exaggeration … Jihād, in Central Asia as in Southeast Asia and Africa, when practiced was more often the domain of the newly Islamicized local populations pursuing political goals than of foreign Muslims”. Golden adduces no evidence, however, to back this assertion. Cf. the contrasting analyses of Paul, Jürgen, Herrscher, Gemeinwesen, Vermittler: Ostiran und Transoxanien in Vormongolischer Zeit (Beiruter Texte und Studien 59. Beirut: Orient-Institut, 1996), 103–05, 108–13Google Scholar, and Bosworth, C. E., The Ghaznavids: Their Empire in Afghanistan and Eastern Iran, 994–1040 (Beirut: Librairie du Liban, 1973), 31–2Google Scholar.

24 Shams al-Dīn Muḥammad b. Aḥmad b. al-Dhahabi, ʿUthmān, Ta'rīkh al-Islām, ed. Tadmurī, ʿUmar ʿAbd al-Raḥmān (Beirut: Dār al-Kitāb al-ʿArabī, 1412/1992), 18: 33Google Scholar; al-Amīr al-Ḥāfiẓ Abū Naṣr ʿAlī b. Mākūlā, Hibat Allāh Ibn, al-Ikmāl fī raf'il-irtiyāb, ed. Muʿallimī, ʿAbd al-Raḥmān b. Yaḥyā (Hyderabad: Dā'irat al-Maʿārif al-ʿUthmāniyya, 1967), 1: 21Google Scholar, referring to the late third/ninth century.

25 In Bayhaq, Ibn Funduq, Tārīkh-i Bayhaq, 51, referring to the year 378/988; in Bukhārā, Gardīzī, Tārīkh-i Gardīzī, 361.

26 Najm al-Dīn ʿUmar b. Muḥammad b. al-Nasafī, Aḥmad, al-Qand fī dhikr-i ʿulamā'-i Samarqand, ed. al-Hādī, Yūsuf (Tihrān: Mir'āt al-Turāth, 1999), 329Google Scholar.

27 E.g. al-Nasafī, ibid., 65, 281, 329, 386, 400 (referring to an organized group of muṭṭawwiʿa), 569; Abū Saʿd ʿAbd al-Karīm b. Muḥammad b. al-Samʿānī, Manṣūr, al-Ansāb. Ed. ʿAṭā, M. ʿA. (Beirut: Dār al-Kutub al-ʿIlmiyya, 1998), 5: 213Google Scholar. On the flow of eastern Iranian warriors to Tarsus in order to engage in volunteer Jihād activities against the Byzantines, see Bosworth, C. E., “The City of Tarsus and the Arab-Byzantine Frontiers”, reprinted in The Arabs, Byzantium and Iran: Studies in Early Islamic History and Culture (Variorum Collected Studies Series. Aldershot: Ashgate, 1996)Google Scholar, article XIV, and D. G. Tor, Violent Order, chapter 2. The present author is not trying to assert that the stream of eastern Iranian volunteers ceased; rather, that the eastern limes began to compete in importance with the western one, and probably to surpass it.

28 Jürgen Paul, The State and the Military, 20–22, has noted the close relations between the earlier Sāmānid rulers and the religiously motivated warriors and religious leaders, and even that the loss of this support “was instrumental in [the Sāmānid] downfall”, without, however, drawing the present author's conclusions regarding the central legitimizing role of these groups in the actual establishment of Sāmānid rule. For a closer examination of the composition of the Sāmānid amies see Tor, D. G., “The Mamluks in the military of the autonomous Persianate dynasties”, IRAN: Journal of the British Institute of Persian Studies, 46, 2008, 213–25CrossRefGoogle Scholar.

29 Al-Narshakhī, Tārīkh-i Bukhārā, 98. The anonymous Sāmānid-era geographical work, Ḥudūd al-ʿālam, tr. V. Minorsky (Cambridge: E. J. W. Gibb Memorial Series, 1982 reprint), 112, also notes the ubiquity of ghāzīs in the Transoxianan Sāmānid dominions.

30 Al-Narshakhī, Tārīkh-i Bukhārā, 100.

31 Attr. Ibn al-Zubayr, Kitāb al-dhakhā'ir, 141, 142. Bosworth correctly notes in his evaluation of the historicity of this reported incident, “Here, then, lies the main value of … [this] story. We see the importance of the ghāzī and volunteer contingents of the Sāmānid army, stationed in various strategically-placed towns and regions of the empire's northern fringes, where they could always be sure of opportunities for jihād against the pagans”. Bosworth, C. E., “An alleged embassy from the Emperor of China to the Amir Naṣr b. Aḥmad: a contribution to Sāmānid military history”, Yād-nāme-ye Irāni-ye Minorsky, ed. Mīnuvī, M. and Afshār, I. (Tihrān: Tihrān University, 1969), 26Google Scholar. Sāmānid muṭṭawwiʿa are also found e.g. in Allāh, Rashīd al-Dīn Faḍl, Jāmiʿ al-tawārīkh, ed. Ātesh, Aḥmad (Tihrān: Dunyā-yi Kitāb), 1362, 1: 17Google Scholar.

32 Jürgen Paul, The State and the Military, 16–17.

33 Blankinship, Khalid Yahya, The End of the Jihād State: The Reign of Hishām Ibn ʿAbd al-Malik and the Collapse of the Umayyads (Albany: SUNY Press, 1994), chapter 1Google Scholar, “Jihād and the caliphate before Hishām”, 11–36.

34 D. G. Tor, “Privatized Jihād and public order”.

35 Thus, as Frye notes (R. Frye, “The Sāmānids”, Cambridge History of Iran. Volume IV: The Period from the Arab Invasions to the Saljuqs, 150), the jihādic warriors constituted a significant factor in Sāmānid military forces: “On [military] expeditions the ghāzīs or warriors for the faith were an important factor in Sāmānid successes”. This was true of the Sāffārid armies as well, although that fact has been obscured by the uncertainty and controversy surrounding the meaning of the term ʿayyār; see e.g. Bosworth, , “The armies of the Sāffārids”, BSOAS XXXI/3, 1968, 538–9Google Scholar; idem., The History of the Saffards of Sistan and Maliks of Nimruz (247/861 to 949/1542–3) (Costa Mesa: Mazda, 1994), 340–5; and Tor, Violent Order.

36 According to Abū ʿAlī Aḥmad b. Miskawayh, Muḥammad, Tajārib al-umam, ed. Amedroz, (Baghdad: al-Muthanna, n.d. [reprint of the 1915 Egyptian edition]), 2: 181Google Scholar:“In [the year 344/955] around 200,000 tents of Turkmen converted to Islam”. Ibn al-Athīr, al-Kāmil, 8: 396, places this event six years later, in 349/960. The issue of dates and catalysts is extremely problematic; it has been discussed most recently by Paul, Jürgen, “Nouvelles pistes pour la recherche sur l'histoire de l'Asie centrale à l'époque karakhanide (Xe–début XIIIe siècle)”, Cahiers D'Asie Centrale 9, 2001, Études Karakhanides, 19–22Google Scholar. This kind of national conversion, mandated by political rulers is, of course, well-known in the Christian context from both the Anglo-Saxon and the Frankish examples; on Clovis's conversion and its attendant political considerations see Wood, Ian, The Merovingian Kingdoms, 450–751 (London: Longman, 1997), 41–8Google Scholar; for the famous conversion of the Northumbrians under King Edwin at a royal council of 627, see Bede, , A History of the English Church and People, tr. Sherley-Price, Leo, rev. R.E. Latham (London: Penguin Books, 1968), 126–9Google Scholar, and also Stenton, F. M., Anglo-Saxon England, third edition (Oxford: Oxford University Press, 2001), 114–5Google Scholar. The question always arises in the case of mass conversions of the depth of the actual religious commitment of the converts. While this is to a certain extent unanswerable – particularly in the present case, in light of the extreme paucity of information on the Qarakhanids – at least in the Frankish context there have been several convincing attempts to argue for genuine “barbarian” intellectual sophistication and religious conviction; see Moorhead, John, “Clovis' motive for becoming a Catholic Christian”, Journal of Religious History 13/4, 1985, 329–39CrossRefGoogle Scholar, and Shanzer, Danuta, “Dating the baptism of Clovis: the bishop of Vienne vs the bishop of Tours”, Early Medieval Europe 7/1, 1998, 29–57CrossRefGoogle Scholar.

37 Barthold, Turkestan, 255–6. Note also the example of an alleged itinerant Muslim preacher cited by de la Vaissière, Étienne, Sogdian Traders: A History, tr. Ward, James (Leiden: Brill, 2005), 315–6Google Scholar.

38 al-Samʿānī, al-Ansāb, 4: 641. Paul, at the aforementioned conference, pointed out that this text is corrupt – but that in any case none of the manuscripts supports Barthold's reading.

39 Usbānīkath or Subānīkath was in the tenth century a heavily-fortified border town lying north-east of the Jaxartes river; see Strange, G. Le, Lands of the Eastern Caliphate (London: Frank Cass, 1966), 485Google Scholar.

40 al-Samʿānī, al-Ansāb, 1: 131–2.

41 al-Samʿānī, al-Ansāb, 1: 132.

42 Paul, “Nouvelles pistes”, 22, does not fall into the trap of the assumed Sufi missionaries; rather, he notes the various empirically verifiable factors that went into aiding Islamization: “Car on distingue très nettement les vecteurs culturels qui aident à l'Islamisation de la region de Kachgar: le commerce d'abord, la supériorité culturelle (supposée) de Mavarannahr sur le Turkestan oriental et l'intensité des contacts et échanges entre les deux regions…, [et] les guerriers de la guerre sainte…”.

43 Barthold, Histoire des Turcs, p. 49. La Vaissière, Sogdian Traders, 314 has also drawn attention, in a different context, to the description in Ibn al-Nadīm's Fihrist of the numerous Muslim colonies in pagan Turkestan during the Sāmānid era.

44 See e.g. Kovalev, Roman K., “Dirham mint output of Samanid Samarqand and its connection to the beginnings of trade with northern Europe (10th century)”, Histoire et Mesure 17/3–4, 2002, Monnaie et Espace, 197–216CrossRefGoogle Scholar; and idem., “The mint of al-Shash: the vehicle for the origins and continuation of trade relations between Viking-age northern Europe and Samanid Central Asia”, Archivum Eurasiae Medii Aevi 12, 2002–2003, 47–79.

45 Ibn al-Athīr, al-Kāmil, 7: 522.

46 Ibn al-Athīr, al-Kāmil, 7: 533.

47 Another possible motive may have been the threat of a Ṣaffārid revanche in eastern Khurāsān at this time, and his consequent need to strengthen his religious credentials; Ismāʿīl finally had to conquer the Bust area again from the Ṣaffārids soon after this campaign took place (see Anon., Tārīkh-i Sīstān, 291–3) – particularly if the later date for this raid given by Mīrkhwānd in the following footnote is correct.

48 Mīrkhwānd, Rawḍat al-ṣafā', 4: 36 (who, however, places this raid in 295); Ibn al-Athīr, al-Kāmil, 7: 547. Although, as Barthold, Histoire des Turcs, 33, has correctly noted, these campaigns, for all the fanfare that accompanied them, contained more public relations than substance: “The Sāmānids decided to adopt an offensive policy. But their attacks for the most part bore the character of incursions, and the conquests made in the name of Islam under these sovereigns were insignificant: nothing was added to the Muslim possessions except the regions extending from the valley of the river Tchirtchiq up until the river Talas”.

49 On this problem see Golden, An Introduction, 198–9 and 214–16.

50 Pritsak, O., “Von den Karluk zu den Karachaniden”, Zeitschrift der Deutschen Morgenländischen Gesellschaft 101, 1951, 277Google Scholar; also “Die Karachaniden”, Der Islam, 31, 1953, 22–4.

51 In the delightful understatement of one scholar: “Thus, by the mid-9th century, these Eastern neighbors of the Sāmānids claimed a hegemony over the steppe peoples. The extent to which this was translated into real power beyond the confines of the Karluk confederation and its allies is open to question” (Peter Golden, “The Karakhanids and Early Islam”, 351). Neither Golden nor Bosworth, s.v. “Īlek-Khāns or Ḳarakhānids”, EI2, ed. B. Lewis et al., III: 1113–17, addresses the even thornier question of borders. Kotchinev, Boris D., “Les frontières du royaume des Karakhanides”, Cahiers D'Asie Centrale 9, 2001, Études Karakhanides, 41–8Google Scholar, identifies this problem but, unfortunately, proceeds to address the issue only from the very end of the Sāmānid period.

52 Pritsak, O., “Two migratory movements in the Eurasian steppe in the 9th–11th centuries”, reprinted in Studies in Medieval Eurasian History (London: Variorum Reprints, 1991), Article VI: 157Google Scholar. In another work he states that “Despite [the Qarakhanids' being near neighbors to Muslim territory] the information in the Islamic historiography regarding their eastern neighbors is very incomplete, meager, and confused”, O. Pritsak, “Von den Karluk zu den Karachaniden”, 278. Similarly, Kliachtorniy, “Les Samanides et les Karakhanides”, 39, writes: “Nous sommes confrontés à toute une série de questions embarrassantes. Premièrement, on ne sait rien sur les diverses invasions des Turcs dans le Mavarannahr au IXe siècle, bien que les sources témoignent des confrontations armies avec les Turks aux frontières d'Isfijab, de Chach, et du Ferghana. Du même, il reste à élucider qui furent les initiateurs de ces hostilités, des ghâzîs ou des pillards Turcs”.

53 We have empirical evidence that the omitting of inconvenient historical facts was indeed practised in Sāmānid historiography from works such as Balʿamī's “translation” of Ṭabarī, where, for instance, he discreetly cuts Ṭabarī’s entire section on the devout holy warrior background of the Sāffārids; see D.G. Tor, Violent Order, 90–91.

54 The cartographers of the Tübinger Atlas des Vorderen Orients (Wiesbaden: Reichert, 1977–c.1993), Map B VII 8, for instance, dealt with the problem by simply straddling the fence: they marked the relevant debatable areas as both Sāmānid and Qarakhanid.

55 Thus, for instance, Bayhaqī, Tārīkh-i Bayhaq, 68, states, referring to the period around 891, that “[Sāmānid] rule stretched/spread from Kāshghar all the way to Rayy”. Obviously, by the tenth century, Sāmānid rule did not extend anywhere near to Kāshghar.

56 What Bosworth calls “the assemblage of territories making up the Sāmānid empire”, Bosworth, , “The rulers of Chaghāniyān in Early Islamic times”, Iran: Journal of the British Institute of Persian Studies 19, 1981, 3CrossRefGoogle Scholar. This was true of the border areas even from the earliest times; thus, for instance, after the conquest of Isfījāb in 840, “Significantly, [the city] remained a largely independent possession of the local Turkish dynasty, which owed only three obligations to the Samanids: military service, the presentation of symbolic gifts, and the inscription of the name of the Samanid suzerain on their coinage”. Davidovich, E.A., “The Karakhanids”, History of Civilizations of Central Asia, Volume IV: The Age of Achievement, AD 750 to the End of the Fifteenth Century. Part I: The Historical, Social, and Economic Setting, ed. Asimov, M. S. and Bosworth, C. E. (Paris: UNESCO, 1998), 120–21.Google Scholar

57 This episode is summarized in Bosworth, “The rulers of Chaghāniyān”, 5–8, without drawing the present author's conclusions.

58 Ibn al-Athīr, al-Kāmil 8: 459–60; Mīrkhwānd, Rawḍat al-ṣafā’, 4: 47.

59 Gardīzī, Tārīkh-i Gardīzī, 344.

60 The strikingly good terms accorded to Abū ʿAlī have been noted by Bosworth, “The rulers of Chaghāniyān”, 8.

61 “Naṣr gathered an army, and sent a letter to Farghāna, to his brother Abūl-Ashʿath, asking him to come with a large army. He sent another letter to Shāsh, to another brother, Abū Yusuf Ya'qub b. Aḥmad, that he come with his army, and that he also bring the Turks of Isfījāb…” al-Narshakhī, Tārīkh-i Bukhārā, 97.

62 Al-Narshakhī, Tārīkh-i Bukhārā, 111–12; Gardīzī, Tārīkh-i Gardīzī, 330–31; and Mīrkhwānd, Rawḍat al-ṣafā', 4: 40.

63 As Barthold notes, “We have no information by which we may solve the question whether the ‘dihqān’ of Kāshghar, Tughān-tagīn, had any connexion with the Qarā-Khānid dynasty” Barthold, Turkestan, 256. On the other hand, Grenard, M. F., “La légende de Satok Bughra Khān”, Journal Asiatique 15/1, 1900, 34Google Scholar, has no doubt that this was a Qarakhanid governor, and even draws the conclusion that “From this name and title we see that the khān did not reside in Kāshghar; and that, in effect, the capital of the Turco-Qarluq was Balāsāghūn …”.

64 Ibn al-Athīr, al-Kāmil, 8: 133–4. This intriguing revolt has been noted only by M. F. Grenard, “La légende de Satok Bughra Khān”, 34–6, and Barthold, Turkestan, 241, albeit only very briefly in passing. Perhaps equally revealing is Jūzjānī's (surely deliberate) omission of any mention of the Turkish role in these revolts, which he chronicles in some detail; he attributes the pardoning of the rebels to Naṣr's clemency alone; Jūzjānī, Minhāj-i Sirāj, Ṭabaqāt-i Nāṣirī, ed. Ḥabībī, ʿAbd al-Ḥayy (Tihrān: Dunyā-ye Kitāb, 1363/1944), 208Google Scholar.

65 Note that this sanctuary policy is apparent as early as the year 301/914, when the slaves who murdered the Sāmānid amīr, Aḥmad b. Ismaʿīl, fled to Turkestan; Narshakhī, Tārīkh-i Bukhārā, 111.

66 Ibn al-Athīr, al-Kāmil, 8: 415.

67 On the discrepancy in dates see Barthold, Turkestan, 254–5. For a strange and poetic account of the circumstances surrounding this conversion, see Grenard, “La légende de Satok Bughra Khān”, 5–79. The story itself is found on pp. 6–10.

68 Although the Mongols succeeded in accomplishing such a feat several centuries later, they had a great deal more strength at their disposal than did the Qarakhanids; the latter failed throughout their pre-Islamic phase to overcome the Sāmānids. Moreover, when the Sāmānids tried to rally the military support of the Transoxanian populace when faced with the eventual Qarakhanid conquest, their failure to do so was directly attributable to the fact that the latter were Muslims; as Frye remarks, “The Sāmānid amir tried to rouse the people of his domains against the invaders but he failed. The people of Bukhārā would not listen … especially when their religious leaders assured them that the Qarakhānids were good Muslims like themselves and there was no need to fight for the discredited Sāmānids against them” (“The Sāmānids”, 159).

69 Frye, Richard, Bukhara: The Medieval Achievement (Costa Mesa: Mazda, 1997), 147Google Scholar. Or, in Soucek's words, “Once they entered the community of the Dār al-Islām as Muslims, these Turks reversed the trend of actual conquest and themselves conquered Transoxania” (Inner Asia, 76).

70 Despite Frye's admirable attempt to harmonize the conflicting versions into one narrative; R. Frye, Bukhara: The Medieval Achievement, 141–7. Frye similarly avoids entering into the issue of the divergent accounts in his “The Sāmānids”, 157, where he notes merely that “The course of events is unclear…”.

71 And according to some accounts, foolishly summoned by Sāmānid political players; Rashīd al-Dīn, Jāmiʿ al-tawārīkh, 1: 47.

72 Gardīzī, Tārīkh-i Gardīzī, 368.

73 Gardīzī, Tārīkh-i Gardīzī, 369. Rashīd al-Dīn's version, Jāmiʿ al-tawārīkh, 1: 45–51, though much more detailed and differing in a few points, for the most part corroborates Gardīzī's.

74 For present purposes, it is immaterial which version is correct: both show the same growing encroachment of the newly-Islamized Turks upon the Sāmānid realms.

75 “Shihāb al-Dawla Hārūn b. Sulaymān Īlig, known as Bughrā-Khān the Turk, ruler of Kāshghar and Balāsāghūn to the borders of China”, Ibn al-Athīr, al-Kāmil 9: 98. On this Qarakhanid title and other titulature see Pritsak, “Die Karachaniden”, 23–4, and Golden, “An Introduction”, 215.

76 Ibn al-Athīr, al-Kāmil, 9: 95. Cf., for instance, the account of unmitigated Sāmānid defeat, in only one campaign, in Ḥamdallāh b. Abī Bakr b. Aḥmad b. Qazvīnī, Naṣr Mustawfī, Tārīkh-i Guzīda, ed. Navā'ī, ʿAbd al-Ḥusayn (Tihrān: Intishārāt-i Amīr-i Kabīr, 1339/1960), 385Google Scholar.

77 Ibn al-Athīr, al-Kāmil, 9: 98–9.

78 Ibn al-Athīr, al-Kāmil, 9: 100. Yet a third version is offered by al-ʿUtbī (in Jarfādhqānī's Persian translation): Abū'l-Sharaf Nāṣir b. Jarfādhqānī, Ẓafar, Tarjamah-i tārīkh-i yamīnī, ed. Shiʿār, Jaʿfar (Tihrān: Bungāh-i Tarjamah va Nashr-i Kitāb), 1345, 94–98Google Scholar; it is closer to Gardīzī's than to Ibn al-Athīr's.

79 ʿUtbī/ Jarfādhqānī, Tarjamah-i tārīkh-i Yamīnī, 184–99; Ibn Funduq, Tārīkh-i Bayhaq, 69–70; Jūzjānī, Tabaqat-i Nāṣirī, 216; Gardīzī, Tārīkh-i Gardīzī, 376–8; Ibn al-Athīr, al-Kāmil, 9: 148–9; Rashīd al-Dīn, Jāmiʿ al-tawārīkh, 1: 108–19.

80 On the agreement of these two Turkic powers on dividing the spoils see Bosworth, The Ghaznavids, 39–41; ʿUtbī/ Jarfādhqānī, Tarjamah-i tārīkh-i Yamīnī, 249–50; Rashīd al-Dīn, Jāmiʿ al-tawārīkh, 1: 142–4. Maḥmūd also married the Īlig-Khān's daughter at this time; see Nāẓim, Muḥammad, The Life and Times of Sulṭān Maḥmūd of Ghazna (New Delhi: Munshiram Manoharlal, 1971), 47–8Google Scholar.

81 Thus, the Qarakhanids have been termed the “première dynastie turk musulmane d'Asie centrale à avoir gardé son caractère tribal”. Jürgen Paul, “Nouvelles pistes,” 13.

82 Note that the sources view the Ghaznavids as the true heirs of the Sāmānid mantle; Mustawfī Qazvīnī, for instance, Tārīkh-i Guzida, 351, declares that “In [the Caliph al-Qādir's] time Sāmānid rule [dawla] ended, and their dominion fell to the Ghaznavids”.

83 On Sebuktegin's background see e.g. ʿUtbī/ Jarfādhqānī, Tarjamah-i tārīkh-i Yamīnī, 19–20.

84 On Mahmud's mother see Yaḥyā b. al-Qazvīnī, ʿAbd al-Laṭīf, Lubb al-tawārīkh (Tihrān: Intishārāt-i Bundād ū Gūyā, 1363/1984), 142Google Scholar. Bosworth, Ghaznavids, 43, notes that a Hindu dynasty had ruled in the Kabul valley until Alptegin, Sebuktegin's master, had conquered the area.

85 As Bosworth notes, Ghaznavids, 44: “Judging by Sebuktegin's last wishes, he did not envisage that his family should set up an independent dynasty, despite the evident decay of the Sāmānids”. At least one chronicler clearly felt queasy about the Ghaznavid complicity in the downfall of their overlords: Ibn Funduq, uniquely, portrays Maḥmūd as having retained at least nominal allegiance to the Sāmānids until the end, and as having played no part in the Sāmānid downfall: see Tārīkh-i Bayhaq, 70.

86 Therefore the Ghaznavids in particular mark a transition between the old individual Mamlūk, assimilationist model and the new one of entire Turkic peoples invading large swathes of the Islamic world.

87 Although his son has drawn more attention as a holy warrior, Sebuktegin was occupied with ghāzī activities as well, and began the Ghaznavid drive in India; note that according to al-ʿUtbī, the very first thing that Sebuktegin takes care to do upon his ascent to power, after distributing fiefs, is to fight Infidels: “Sebuktegin took all [necessary measures] in preserving his rule, and attended to the affairs and prosperity of all … Then he turned his face to Jihād against the Infidels and the humbling of the enemies of the faith, and the country of Hindūstān, which was the dwelling of the enemies of Islam and the worshippers of idols he made into the abode of religious warfare for the faith [Dār al-Ghazw]”. ʿUtbī/ Jarfādhqānī, Tarjamah-i tārīkh-i Yamīnī, 20. See also Rashīd al-Dīn, Jāmiʿ al-tawārīkh, 1: 11; Mustawfī Qavīnī, Tārīkh-i Guzida, 389; Jūzjānī, Ṭabaqāt-i Nāṣirī, 225–6, terms Sebuktegin “al-amīr al-ghāzī”. In Rashīd al-Dīn's echoing of ʿUtbī/ Jarfādhqānī in Jamiʿ al-tawārīkh, 1:6: “He set his face toward Jihād against the Infidels and the enemies of religion”.

88 See e.g. Jūzjānī, Ṭabaqāt-i Nāṣirī, 228, where he is referred to repeatedly as “Maḥmūd al-Ghāzī” and “Sulṭān ghāzī”; also Niẓām al-Mulk, Abū ʿAlī Ḥasan Ṭūsī, Siyar al-mulūk, ed. Darke, Hubert (Tihrān: Intishārāt-i ‘Ilmī va Farhangī, 1378), 74, 75Google Scholar; and Qazvīnī, Lubb al-tawārīkh, 142: “Maḥmūd b. Sebuktekīn was a famous ruler; he continually performed ghazwas against the Infidels in India…”. According to Mustawfī Qavīnī, Tārīkh-i Guzida, 392, it was on the campaign in India in 392/1002 that “Yamīn al-Dawla … obtained the title of ‘Ghāzī’”.

89 An exception to this is Nāẓim, The Life and Times of Sulṭān Maḥmūd of Ghazna, 41–85. Note that this is an oversight to which most of the primary sources do not fall prey: the Indian and the Central Asian conquests and ghazi raids are given equal mention in e.g. Anon, ., Mujmal al-tawārīkh wa'l-qiṣaṣ, ed. Najmābādī, S. and Weber, Siegfried (Neckarhausen: Deux Mondes, 2000), 313Google Scholar. One explanation for this relative neglect of the Central Asian conquests has been noted by Paul, “Nouvelles pistes”, 17: all post-Barthold studies “portent l'empreinte de l'approche de Barthold. Cela est vrai surtout pour la délimitation du champ de recherche”. Although Paul was referring to the dynastic and political approach, his observation is equally valid regarding the geographical area under consideration; in Barthold's wake, most “Central Asianists” tend to limit their geographical purview of Central Asia to the area that accords with what Kotchnev calls “the traditional Soviet definition” of the term – in effect, to Inner Asia; Kotchnev, Boris D., “La chronologie et la généalogie des Karakhanides du point de vue de la numismatique”, Cahiers D'Asie Centrale 9, 2001, Études Karakhanides, 50Google Scholar.

90 See e.g. his campaigns in 391 in Peshawar, Gardīzī, Tārīkh-i Gardīzī, 385; 401/1011 in Ghūr: Ibn al-Athīr, al-Kāmil, 9: 221–2; mentioned briefly in Gardīzī, Tārīkh-i Gardīzī, 391; ʿUtbī/Jarfādhqānī, Tarjamah-i tārīkh-i Yamīnī, 312–4; Rashīd al-Dīn, Jāmiʿ al-tawārīkh, 1: 167–9; Mustawfī Qavīnī, Tārīkh-i Guzida, 393; and Bosworth, , “The early Islamic history of Ghūr”, Central Asiatic Journal 6/2, 1961, 122–8Google Scholar; in Kashmir in 406/1015, together with 20,000 volunteer warriors, Ibn al-Athīr, al-Kāmil, 9: 265–8.

91 Mention of large numbers of jihādic volunteer warriors accompanying Mahmud's armies can be found in e.g. Ibn al-Athīr, al-Kāmil, 9: 343; Gardīzī, Tārīkh-i Gardīzī, 385; and so forth.

92 E.g. Gardīzī, Tārīkh-i Gardīzī, 388; Rashīd al-Dīn, Jāmiʿ al-tawārīkh, 1: 146–54; implied obliquely in Qazvīnī, Tārīkh-i Guzida, 393. For an account of the history of Qarakhanid-Ghaznavid relations during Mahmud's time, see Nāẓim, The Life and Times of Sulṭān Maḥmūd of Ghazna, 48–56.