Chapter 1 Cultural Contexts of al-Bīrūnī’s Work and Writings

In: The Books Sānk and Pātanğal
Noémie Verdon
Search for other papers by Noémie Verdon in
Current site
Google Scholar
Open Access

1.1 Persian and Islamic Spheres of Influence

1.1.1 Kāṯ, Ğūrğān, Ray and Ğūrğānīya

Al-Bīrūnī spent his youth in the region known as Khwarezm. He was born in the capital city of the region, Kāṯ,1 also referred to as Kāṯ-Kala, and lived there from 973 to 995.2 Khwarezm (a region extending between the territories of present-day Uzbekistan and Turkmenistan) was at the time governed by the Afrighid dynasty (ca. from the early 4th to the late 10th c. CE).3 Qutayba b. Muslim al-Bāhilī had conquered the region in 712.4 The Afrighids most probably only became dependent on the Islamic Samanid dynasty (819–1005)5 by the end of the ninth century CE, that is, nearly two hundred years after Qutayba b. Muslim had penetrated the region.6 By the tenth century, the Arabic language, as well as Islamic culture, was relatively well-established in Khwarezm.7

Situated on important trade roads, Kāṯ was an emporium in the tenth century. Arab sources report that the region benefited from great prosperity in terms of commerce.8 For instance, the Ḥudūd al-ʿālam (982/983) describes Kāṯ as “the capital of Khwārazm and the Gate of the Ghūz Turkistān [… and as] the emporium of the Turks, Turkistān, Transoxiana and the Khazar” (Bosworth 1970: 121).9 Khwarezm, due to its location on the south-eastern side of the Aral Sea, moreover constituted a fertile oasis in the middle of arid steppes and deserts. The region was also considered an important intellectual centre.

In al-Bīrūnī’s time, two dynasties, the Afrighids in Kāṯ and the Maʾmūnids in Ğūrğānīya (Kunya-Urgench, located in present-day Turkmenistan), were competing to rule Khwarezm,10 a situation that culminated in a war in 995 between the two dynasties. These conflicts eventually caused al-Bīrūnī to flee from Khwarezm. The exact length of his sojourn outside Khwarezm is unknown, but it is known that he lived in Ray (now a south-eastern suburb of Tehran) probably some time between the years 995 and 997.11 In Al-āṯār al-bāqiya, al-Bīrūnī mentions his visit to Ray, where he met other scholars and led several research projects.12

Followers of Islam first conquered the city between the years 639 and 644, more than three centuries prior to al-Bīrūnī’s time. In the tenth century, the Buyids (r. from the mid-10th to the mid-11th c.) incorporated the city, which served as the seat of governing bodies, into their kingdom.13 In addition to the role of the city as an administrative and trade centre,14 Ray’s reputation as a centre of knowledge made it an essential destination for scholars. For instance, the physician and philosopher Ibn Sīnā (980–1037) visited Ray around the years 1014 and 1015.15

From approximately 1000 to 1004, al-Bīrūnī dwelt in ancient Gorgan, referred to as Ğūrğān in Arabic and located at the south-eastern corner of the Caspian Sea. In 1000, al-Bīrūnī dedicated Al-āṯār al-bāqiya to Prince Qābūs bin Wušmagīr bin Ziyār (r. 977 to 981 and 998 to 1012/1013) who governed the region at the time. Prince Qābūs was known to be redoubtable and cruel, and at the same time an important patron of science and art.16 The Arab Muslims came to the region in 650/651, but it appears that Islam only established there in the early eighth century.17 In the ninth and tenth centuries, the town was wealthy and comfortable,18 known for its silk, and strategically positioned for commerce.19 Although only a few main roads passed through the city, Ğūrğān was an important station on the axis between the North and the South. Southward, the road led to Ray, and to the North the route reached Khwarezm.

In 1004, al-Bīrūnī returned to Khwarezm. The Maʾmūnid dynasty had won the war against the Afrighids, and a new capital was established at Ğūrğānīya (modern Kunya-Urgench). Al-Bīrūnī lived there until the year 1017.20 Before becoming the capital city of Khwarezm, Ğūrğānīya was an emporium, linking the regions of Ghūz and Khurasan, in the same manner as Kāṯ had been.21 Further, during the eighth century, several institutions known as Bayt al-Ḥikma or Dār al-Ḥikma (House of Wisdom) were flourishing in the Islamic territory.22 These institutions generally housed large libraries and welcomed thinkers. One such establishment, the Maʾmūn Academy, was founded in Ğūrğānīya.23 It was an important centre of knowledge, which hosted numerous scholars. In addition to al-Bīrūnī, other renowned scholars worked there, including the mathematician and astronomer Abū Naṣr ʿIrāq,24 the Christian physician Abū Sahl al-Masīḥī al-Ğurğānī25 and Ibn Sīnā.26

1.1.2 Ghazna and Kābul, the Gateways to Early Medieval India

In the mid-seventh century, Arab Muslims made inroads on Sistan, from which they reached Ghazna and Kābul. Two centuries later, Alptigīn, a commander of the Samanid dynasty, took over Kābul probably around year 961, prior to seize Ghazna in 962. From the first incursions in the seventh century up to the tenth century, the political situation of these regions fluctuated between attempts of the Muslim governors to establish their authority in the area and upheavals of local rulers.27 The process through which Islamic caliphate annexed these eastern regions to its territory was thus relatively protracted, lasting from the mid-seventh to latter half of the tenth century.28

In 977, Sebüktigīn founded the Ghaznavid Empire with Ghazna (in present-day eastern Afghanistan) as its capital.29 His son Maḥmūd (r. 997–1030) considerably expanded the empire and in 1017 annexed to it the kingdom of Khwarezm.30 During his reign, Ghazna was also the administrative centre of his vast empire. From this year onward, al-Bīrūnī accompanied Maḥmūd at his court. Several of al-Bīrūnī’s astronomical calculations point to his sojourns in Kābul and Ghazna.31 He also composed numerous works in Ghazna, which include the Taḥdīd al-amākin and the Taḥqīq mā li-l-Hind, during Maḥmūd’s rule. Al-Bīrūnī may have passed away a few years later than 1048, perhaps around 1050.32 He spent thirteen years of his life, from 1017 to 1030, working for sultan Maḥmūd, and the remaining years for al-Masʿūd his son.

Kābul and Ghazna were positioned at the crossroads of two different cultural spheres. As Minoru Inaba points out, both cities were located in a network of roads leading to various cities and regions of Central Asia, by way of the Oxus River, and of the South Asian subcontinent,33 through the Khyber Pass.34 Geographically close to India, Kābul and Ghazna were places of important cultural and economic exchanges. The role of eastern Afghanistan at this crossroad remained crucial into the tenth and eleventh centuries, if for commercial reasons at the very least. According to Arabic sources, trade with India was prosperous.35 Bust, a city located to the south-west of Ghazna, has been considered the “gateway to Hind.”36 The geographical location of these Ghaznavid sites conferred to them a crucial role in cross-cultural exchanges which were taking place with the South Asian subcontinent and through which goods, art and ideas travelled.37

Historically, eastern Afghanistan and Gandhāra were incorporated into a number of successive empires, including the Achaemenid Empire (ca. 8th–4th c. BC), that of Alexander the Great, the Greco-Bactrian kingdom and that of the Kuṣāṇas (ca. 1st–4th c. CE). After the Kuṣāṇas, during the period of Sasanian rule (224–651 CE) in the region, several dynasties were ruling over the whole territory. Numismatics tends to show that these kings had been ethnically and/or politically related to each other and that they governed side by side or successively some areas of Central and South Asia, such as the regions north of the Hindukush, Kābulistān, Zābulistān and the Gandhāra region. Broadly speaking, these dynasties include the Kidarites, the Alkhans, the Nezaks, the Rutbils, the Early Shahis,38 the Rutbils and the Late Shahis.39 While scholars have yet to comprehensively detail the socio-political and religious situations of pre-Islamic Gandhāra, Kāpiśī and Panjab, archaeological data reveals mixed influences in terms of art and architecture.40 This is testimony to the specific location of this region, including Kābul and Ghazna, at the cross-road of different cultures.

Al-Bīrūnī mentions the Early Shahis (ca. mid-7th to early 9th c.) and the Late Shahis (early 9th to early 11th c.) in a well-known passage of the Taḥqīq mā li-l-Hind:

And the [Indians] had kings in Kābul, Turks who were said to be of Tibetan origin. The first of them who came was Barhatigīn (‮برهتكين‬‎).41 He entered a cave in Kābul, which it was only possible to enter by lying down. […] He was dressed in Turkish clothes, such as the qabaʾ,42 a hat, boots and weapons. […] He took possession of these places, assuming the title of Šāhi of Kābul (‮شاهية كابل‬‎). The kingship remained in the hand of his descendants for around sixty generations.

The Indians are careless about succession of things and negligent in the proper arrangement of the continuous chronicles of their kings. [When they] are at loss, they invariably resort to speculation. For this, we shall convey what some of their people mentioned. According to what I have heard, [the history of] such a lineage, [written] on a piece of silk, is found in the fortress of Nagarkot.43 I desired to find it, but [I] was prevented from doing so for [different] reasons.

In their group, there was Kaniṣka (‮كنك‬‎), who is at the origin of the vihāra (‮بهار‬‎) which is in Peshawar (‮برشاور‬‎) […] called the Kaniṣka-Caitya (‮كنك جيت‬‎).44

The last of them was Lagatūrmān (‮لگتورمان‬‎) and his minister, a Brahmin, was Kallara (‮کلّر‬‎). Times were auspicious to the latter, and, by accident, [Kallar] found hidden treasures, by which he gained the upper hand and became powerful. Then, the government turned away from his master because it had been with the members of his house for a long time.

Then, the manners of Lagatūrmān became wicked, and his deeds were disgusting, in such a way that complaints to his minister [about that] increased. Therefore, [Kallar] tied him (i.e., Lagaturman) and imprisoned him as a punishment. He enjoyed being the sole master of the kingship. He had wealth [at his disposal], as resources, and he made himself master of [the kingdom]. The Brahmin kings succeeding him were Sāmanta (‮سامند‬‎), Kamalū (‮کملو‬‎), Bhīma (‮بهیم‬‎), Jayapāla (‮جیپال‬‎), Ānandapāla (‮انندپال‬‎), Trilocanapāla (‮تروجنپال‬‎), killed in the year 412 of Hegira (i.e., 1021/1022),45 and his son, Bhīmapāla (‮بهیمپال‬‎), five years later. The [dynasty of the] Indian Šāhis (‮الشاهيّة الهنديّة‬‎) ended and not even a spark from the people of this House remains [today].46

These kings originally based in Kābul were among those whom the Muslims encountered in their successive waves of incursion into Kāpiśī, Gandhāra and Panjab. Therefore, it is worth discussing this passage. It is, however, likely that this excerpt only conveys few historical facts. The statement that the Early Shahis were of Tibetan origin is, for instance, probably not a historical reality. The generic term Turk (‮الترك‬‎; pl. ‮الاتراك‬‎) in Arabic can refer to several different and distinct tribes or clans, originally coming from Western Eurasia, while the term Tibetan (‮التبّت‬‎) generally designates the populations living beyond the Oxus River. Evidence indicates that Tibetan people never travelled beyond Gilgit in the upper Indus Valley, and that they were pushed from there in the beginning of the eighth century.47 Thus, it is unlikely that the Early Shahis were of Tibetan origin. It is more probable that they belonged to a Turkish tribe, the Khalajs, as Inaba convincingly suggests. During the seventh century, a branch of the Khalajs migrated from the northern Hindukush to Kāpiśī and ruled the area until the ninth century.48

According to the above report, the founder of the Early Shahis lineage was named Barhatigīn.49 Al-Bīrūnī connects this dynasty with that of the Kuṣāṇas and traces the lineage of the Early Shahis to a king named Kanik (‮كنك‬‎) who is to be identified with Kaniṣka (early 2nd c.).50 This connection is, however, not supported by historical evidence.51 Al-Bīrūnī’s report about this genealogy either indicates that the dynastic account located in Nagarkot traced back the lineage of the Early Shahis to the Kuṣāṇas or reflects the complexity of this history as revealed by recent research. It may be added that evidence from several archaeological sites of eastern Afghanistan, such as Mes Aynak, Tepe Naranj, Tepe Sardar or Khair Khana, indicates a long occupation, that is, from the Kuṣāṇa period up to that of the Late Shahis.52 The long period during which these sites were inhabited by different political groups may account for the narrative which connects the Early Shahis to the Kuṣāṇas and which al-Bīrūnī retransmitted in his account.

Nevertheless, the Early Shahis have long been considered as patrons of Buddhism,53 current scholarship based on archaeological research also shows that some sites associated with them also hosted Hindu cults. Whoever the Early Shahis might have been, the reign of these kings ended in the early ninth century, when the Late Shahis succeeded them.54

The second part of the above passage dealing with the Late Shahis appears to be relatively reliable historically, although the list of kings provided by al-Bīrūnī also diverges from the information provided by other sources, such as numismatics, epigraphy and the Rājataraṅgiṇī.55 Nevertheless, the Late Shahis are of particular interest for the present investigation: being the first opponents of the Ghaznavids in Gandhāra and Panjab, they were at the forefront as subjects of al-Bīrūnī’s research on India.

Thus, in Section 1.2, I put forth positive evidence of al-Bīrūnī’s visits to al-Hind. The material presented below show that the territory visited by him mostly belonged to the Late Shahi kingdom. Further, in order to connect al-Bīrūnī’s fieldwork with its socio-cultural context, I present, in Section 1.3, a few elements of this context and of the Late Shahi religion, based on a preliminary survey of archaeological and literary sources.56

Lastly, the towns where al-Bīrūnī resided and travelled in the Islamic cultural sphere were all prosperous in terms of material wealth and of intellectual developments, likely due to their locations on important roads. This situation, together with him travelling through several regions in the early part of his life, provided him with favourable conditions to have his mind open to novelty and to meet a variety of scholars and sciences. These conditions somehow prepared al-Bīrūnī’s work on al-Hind and his intellectual exchanges with local thinkers.

1.2 Al-Bīrūnī’s Visits to al-Hind

1.2.1 The Geographical Delimitation of al-Hind

The elements considered so far have highlighted the strategic position of Zābulistān, Kābulistān and Gandhāra in terms of economy, politics and intercultural exchanges, and at the same time pointed to eastern Afghanistan as a border zone with al-Hind. Boundaries between cultural spheres, often represented by political groups, had fluctuated in the region. Before and at al-Bīrūnī’s time, the frontier zones between the Islamic sphere and the Indian world in the north-western subcontinent were particularly subject to changes. Such changes are, for instance, apparent from the Arabic and Persian geographical accounts written between the eighth and the early eleventh centuries. Several of these early works, such as the Šašnāma (end of the 9th c.), the Kitāb futūḥ al-buldān (The Book of the Conquest of the Countries) by al-Balāḏurī, the Kitāb al-masālik wa l-mamālik (The Book of the Roads and the Realms, 9th c.) by Ibn Ḵurdāḏbah, the homonymous work by al-Iṣṭaḵrī (mid-10th c.) and the Ṣūrat al-arḍ (The Shape of the Earth) by Ibn Ḥawqal (mid-10th c.), do not describe at length the inland areas of al-Hind, but rather regions of Sind, Gujarat, or coastal areas of al-Hind, as well as their islands.57

The focus on these territories in the mentioned accounts reflects the progress of the Islamic military conquests to the East, as well as the state of exchanges between the Islamic world and early medieval India. For instance, while commercial contacts between Arab merchants and the subcontinent existed since an early time through maritime network, Arab Muslims first arrived in Sind in the early eighth century. At first a border zone with al-Hind, Sind, became then part of the Islamic world. Further, up to the tenth century, that is, up to the Ghaznavid incursions in Gandhāra, Panjab and the Gangetic valley, north-western and central India remained mostly unknown to Muslim geographers.58 By the time of the composition of the Taḥqīq mā li-l-Hind in approximately 1030,59 these regions, however, were rather well known to Muslims. In addition, the Indus Valley, including Sind and Panjab, as the first place of contact between Muslims and non-Muslim locals via the land route, constituted a zone of economic and political interactions between the two cultural spheres at the time, witnessing important cultural and social transitions.60

In view of the above, the question of how these borders were regarded by al-Bīrūnī when he composed the Taḥqīq mā li-l-Hind must be addressed. It constitutes the first step in contextualizing his research on al-Hind in order to delineate which of his travels eastward were located in the actual territory of al-Hind. Answering this question also helps determine the places where he observed local customs and traditions. Al-Bīrūnī defines the frontiers of the territory of al-Hind as outlined by mountains and the sea, as shown in several passages of the Taḥqīq mā li-l-Hind:

This sea (i.e., the Indian Ocean) is generally named after what it comprises (i.e., islands), or what is opposite to it (i.e., coastal areas). [Here] we [only] need [to discuss] the part [of this sea] which borders the land of the Indians (‮الهند‬‎) and which therefore is named after them. Furthermore, imagine in the inhabitable world a high and uninterrupted mountain [range], as if it had a vertebral spine extending in the middle of its width in its length from East to West.61 […] The [mountain range] has on its surface a wide flat land with a [large] stretch (i.e., a plateau?) and [has] curvings that surround inhabited plains. Rivers flow from it in both directions (i.e., North and South). The land of al-Hind is one of these plains, surrounded by the aforementioned sea in the South and by the high mountains on the other sides.62

In this excerpt, the delimitation is confined to a topographical description of the frontiers. Further passages provide the names of places located on these borders:

Fort Rājagirī (‮راجكرى‬‎) lies to the south of the [mountain Kulārğak]63 and Fort Lahūr (‮لهور‬‎) to its west.64 I never saw stronger [forts] than these two; and at three farsaḵs from the [mountain Kulārğak],65 is the town of Rājāwūri (‮راجاورى‬‎).66 Our merchants trade with it, but do not go beyond it. This is the frontier of al-Hind from the northern side. In the mountains to the west of the [land of al-Hind] are several groups of Afghan tribes, [whose settlement] ends near the land of Sind (‮السند‬‎). The southern side of the [land of al-Hind] is [delineated by] the sea.67

As for the eastern islands in this sea, they are closer to the border of China [than of al-Hind]. They are the islands Zābağ (‮الزابج‬‎). The Indians call them Suvarṇadvīpa (‮سورن ديب‬‎)68 i.e., the islands of gold.69

You must imagine that the borders of the land of al-Hind are surrounded by mountains. To its North is the snowy Himavant (‮هممنت‬‎), whose centre is the land of Kashmir (‮كشمير‬‎) and which is adjacent to the land of the Turks.70

As was common at the time, border zones were conceptualized as rather wide regions. The territory was also delineated by natural boundaries, and at al-Bīrūnī’s time the western frontier of al-Hind corresponded to the mountains west of present-day Pakistan.


Figure 1

South Asian Subcontinent with visible natural borders

map prepared by the author

1.2.2 Evidence from His Writings

Al-Bīrūnī’s Taḥqīq mā li-l-Hind is in many respects a valuable source for studying Indian culture and history. Scholars consider his research methodology highly innovative for his time, and the data he provided is rich in historical information. Whereas the composition date of his work, namely around 1030, is known, his field of investigation, that is to say the territory covered by his research, is still uncertain, as are his actual sources of information. The following sections assess the question of al-Bīrūnī’s visits to al-Hind and observations there.

Toward the end of the nineteenth century, Carl Edward Sachau edited and translated two of al-Bīrūnī’s works, Al-āṯār al-bāqiya (Sachau 1878 and 1879) and the Taḥqīq mā li-l-Hind (Sachau 1887 and 1910). He also published a study on al-Bīrūnī’s transliterations of Indic words into Arabic (Sachau 1888). Sachau’s works significantly contributed to our knowledge of al-Bīrūnī’s life, his research on India and his sources of information about India. In the preface to the translation of al-Bīrūnī’s Taḥqīq mā li-l-Hind, Sachau writes that al-Bīrūnī stayed “at Multan, Peshawar, &c.”71 He also admits the “absence of positive information” but infers “with a tolerable degree of certainty, that our author […] stayed in different parts of India […].”72 Al-Bīrūnī’s life can indeed only be reconstructed by compiling and analysing passing comments scattered throughout his works, and his astronomical observations conducted at different places that are found within his writings. The lack of direct evidence on al-Bīrūnī’s travels in al-Hind thus constitutes a fundamental difficulty in answering the above question.

Further, Sachau only supports his statements about the places which al-Bīrūnī visited with evidence and arguments in two of his works, which have been insufficiently read and acknowledged in the scholarly world. They are the English preface to his Arabic edition of the Taḥqīq mā li-l-Hind (Sachau 1887)73 and his thorough study on al-Bīrūnī’s Arabic transliterations of Indic words (Sachau 1888). Based on al-Bīrūnī’s calculation of latitudes of places in al-Hind and on linguistic observations, Sachau writes in these two works that al-Bīrūnī only stayed in the Kābul Valley and Panjab.74 Later scholars, although generally coming to similar conclusions, merely refer to these two works. Furthermore, when Alberuni’s India was published by Sachau, British India included present-day Pakistan, as well as Bangladesh. Thus, when Sachau used the term “India” to describe al-Bīrūnī’s visits to al-Hind, this toponym included Pakistan. As I argue below, however, al-Bīrūnī travelled mostly to present-day Pakistan. The impacts of partition on place names, due to the division of the subcontinent into several nation states, have to be taken into account when considering al-Bīrūnī’s visits to al-Hind.

Finally, so much has been written on al-Bīrūnī and his life that difficulties arise when one tries to distinguish history from legend. Numerous authors assumed that al-Bīrūnī visited many places in India, whereas only a few modern scholars have provided some details of their arguments when assessing al-Bīrūnī’s travels in al-Hind. The objective of the paragraphs below is to summarize their statements about the geographical delimitation of al-Bīrūnī’s visits to the territory of al-Hind.

Suniti Kumar Chatterji locates some of al-Bīrūnī’s visits in western Panjab, adding that he must “have stayed for some time in Multan.”75 V. Courtois, for his part, maintains that “al-Biruni stayed in India several years and spent most of his time in the North West, within the limits of pre-partition Panjab.”76 Bimala Churn Law (1955) discusses the fact that al-Bīrūnī observed the forts of Rājagirī and Lahūr situated south to the Kashmir Valley.77 The well-documented biography of al-Bīrūnī by E.S. Kennedy briefly touches upon his travel in al-Hind. Kennedy notes that al-Bīrūnī’s “travel and residences in various parts of India […] were confined to the Punjab and the borders of Kashmir” and mentions al-Bīrūnī’s visit to Nandana, as recorded in his Taḥdīd al-amākin.78 M.S. Khan writes that “[i]t seems unlikely that al-Bīrūnī visited South India, but this question must remain open for investigation.”79 Baloch states that al-Bīrūnī must have visited Nandana in 1017, refers to places for which latitudes were calculated by him, and concludes that he “had visited parts of the Peshawar region, of Kashmir, Western Panjab, and of the Multan region of Sind.”80 Ahmad Hasan Dani casts doubt on the common view that al-Bīrūnī stayed in the city of Lahore, the capital of modern Pakistani Panjab.81

The view that he visited Lahore originates from the fact that Lahore became the second capital of the Ghaznavids under al-Masʿūd, Maḥmūd’s son and successor,82 and thus hosted scholars of the dynasty’s court. Dani, however, stresses that it “would not be unreasonable to say that al-Biruni’s account is more pertinent to the areas that fall within the Indus region, i.e., within the present territorial limits of Pakistan” and that al-Bīrūnī’s observations made in al-Hind “can hardly be perfectly true of the Ganges Valley much less of South India.”83 More recently, M.A. Saleem Khan notes that “[a]l-Biruni [stayed] in India—and present modern Afghanistan was […] part of India—and [visited] other places in the rest of India, [learned] its most important and difficult language i.e., Sanskrit, meeting with the learned pundits, [and] studying books.”84 Mohammed Hassan Syed (2003) argues that al-Bīrūnī stayed for a short period of time in today’s Pakistan.85 Most of the above statements are partially accurate and, as aforementioned, the above authors only partly substantiate their claims. Furthermore, only two scholars, Courtois and Saleem Khan, underline the fact that the boundaries of eleventh-century India were different from today’s, yet at the same time they continue to use the concept of India in a rather vague manner.

Mohammed Hakim Said and Ansar Zahid Khan (1981) provide a relatively detailed account of al-Bīrūnī’s life. According to these two authors, al-Bīrūnī travelled across some regions of al-Hind during three possible time periods, between the years 1020 and 1021, 1023 and 1024, or 1028 and 1029. These periods coincide with years during which the scholar’s presence in Ghazna is not attested by his astronomical observations or writings. The two authors further conclude that the two time periods between the years 1020 and 1024 constitute the most likely periods of al-Bīrūnī’s journeys to the East.

They also write that al-Bīrūnī visited Multan, Sialkot, near today’s Lahore, Nandana, Fort Rājagirī and Fort Lahūr. Furthermore, they refute the assumption that al-Bīrūnī accompanied Maḥmūd on all of his military expeditions. Ultimately, they state that “al-Bīrūnī seems to have travelled along Kābul and the Panjab’s routes.”86

Jai Shankar Mishra’s account similarly constitutes one of the most detailed analyses of al-Bīrūnī’s travels, including their possible duration and geographical limits. Mishra refutes “the view that he travelled in many provinces of India,”87 asserting that al-Bīrūnī only visited western Panjab. He bases his argument mainly on the study of the Taḥqīq mā li-l-Hind, mentioning the forts of Rājagirī and Lahūr as places that al-Bīrūnī actually saw, as well as the locales for which the scholar calculated the latitudes.88 Said, Khan and Mishra are among the rare modern authors who attempted to determine specific places of al-Bīrūnī’s visits to al-Hind, as well as dates for them, based on historical evidence.

Thus, the existing scholarship on al-Bīrūnī’s visits to al-Hind tends to show that he chiefly travelled in Gandhāra, western Panjab and to some extent Sind. These surveys, however, remain largely speculative and tentative. In addition, and more importantly, despite these conclusions, the common view that al-Bīrūnī visited many regions of al-Hind seems to have persisted in numerous studies that mention his life and work on Indian culture. If his work is to be used in order to draw information on the historical, religious or intellectual context of the early eleventh century, it is, however, necessary to determine what could have been his field of investigation before he composed the Taḥqīq mā li-l-Hind. The difficulty of the task is explained by the lack of positive evidence, as Sachau already noted. Al-Bīrūnī did not indeed inform his readers about the places he himself had visited, nor did, for instance, his contemporaries at the Ghaznavid court al-ʿUtbī and Bayhaqī.89

Limited evidence is available and thus arises the need to examine the past through several types of evidence, philological, archaeological and circumstantial. Works by Said and Khan (1981) and Mishra (1985) stand among the rare studies that consider al-Bīrūnī’s visits to al-Hind from a multidisciplinary perspective. The present book takes into account these two studies with the aim of supplementing them. Although absolute certitude might never be reached, several factors help clarify the question of where al-Bīrūnī’s travelled in al-Hind. They are: the historical circumstances related to his life, such as Maḥmūd’s military interests, the social and political contexts at the sultan’s court, and the various ways al-Bīrūnī collected data on India—be it through his direct observations, his readings of Sanskrit literature or his interactions with Indians. In addition, observations about al-Bīrūnī’s sources of information help contextualize his interpretations of Indian philosophy, which are the subjects of Chapters 4 through 6 of this book.

Al-Bīrūnī’s works particularly conducive to understanding the extent of his actual travels in al-Hind are the Taḥdīd al-amākin, composed in 1025,90 and the Taḥqīq mā li-l-Hind, written in approximately 1030.91 In the Taḥqīq mā li-l-Hind, a few passages refer to al-Bīrūnī’s direct observations.92 Yet, they are of no help here, as they do not specify where they did take place. The only positive evidence naming places that al-Bīrūnī visited in the territory of al-Hind lies in a passage where he provides the latitudes of a few locales and in four portions of text pointing to five toponyms.

In the Taḥqīq mā li-l-Hind, al-Bīrūnī states that he calculated the latitudes of the following sites: Fort Lahūr (34°10′), Ghazna (33°35′), Kābul (33°47′), Kindī, the stronghold of the Prince (‮كندى رباط الأمير‬‎ ‮‬‎; 33°55′),93 Dunpūr (34°20′), Laghmān (34°43′), Peshawar (34°44′), Wayhind (34°30′), that is, Hund or Udabhāṇḍa, Jhelum (33°20′), Fort Nandana (32°0′), Sialkot (32°58′), Mandahūkūr (31°50′) or modern Lahore,94 and Multan (29°40′).95 Longitudes and latitudes of some of these places are also provided by al-Bīrūnī in his mathematical treatise Al-qānūn al-Masʿūdī.96 These sites are located in Kāpiśī, Gandhāra and Panjab with the exception of Multan located in Upper Sind. Following this list of latitudes in the Taḥqīq mā li-l-Hind, al-Bīrūnī states:

We have not passed beyond these aforementioned places in the land of the [Indians], nor have we learned about [other] longitudes and latitudes in their books.97

Thus, according to al-Bīrūnī’s own words he did not travel beyond these regions. Another passage in the Taḥqīq mā li-l-Hind in which al-Bīrūnī writes that he could not go to faraway places such as the Valley of Kashmir and Varanasi98 lends support to the validity of this statement. Furthermore, in the Taḥdīd al-amākin al-Bīrūnī explains that he visited the area of Laghmān and Fort Nandana, while in the Taḥqīq mā li-l-Hind he says that he saw Peshawar, Fort Rājagirī and Fort Lahūr. These are the only explicit references by al-Bīrūnī to places he himself visited in al-Hind in these two works.

Laghmān99 was situated to the north of the Kābul river, between modern Kābul and Jalalabad.100 It was located on one of the roads possibly taken by Maḥmūd, which leads from Ghazna to Peshawar via Kābul.101 Al-Bīrūnī must have visited the region of Laghmān some time after he had followed Maḥmūd as a member of his court and prior to the composition of the Taḥdīd al-amākin in which he mentions his observation, namely between the years 1017 and 1025. More specifically, he states that he was there during a solar eclipse.102 An almost total (97 %) solar eclipse occurred on April 8, 1019, in this region. Therefore, it is possible to infer that al-Bīrūnī was in Laghmān on this specific date, as already suggested by Kennedy.103

The exact locations of Fort Lahūr and Fort Rājagirī are problematic. The two toponyms refer to several places on the subcontinent, and scholars made various hypotheses regarding their locations. A passage of the Taḥqīq mā li-l-Hind, however, describes an itinerary starting from Kanauj to Kashmir, via Rājagirī.104 Although not all places located on this route have been identified, it is clear that the itinerary goes north-north-west from Kanauj and then appears to run alongside the southern foothills of the Himalayan range from east to west. At Rājagirī, the road turns to the north and directly leads to the Kashmir Valley. Accordingly, Rājagirī was most probably located to the south of the Pir Panjal Range, at the entrance of one of its passes, such as the Pir Panjal pass, or alongside the Jhelum Valley.105 The localisation by al-Bīrūnī of Rājagirī to the south the mountain Kulārğak in the Pir Panjal Range also points to the same location. Fort Lahūr, which is often mistaken for present-day Lahore, the capital of Pakistani Panjab, was probably located to the south-west of the Pir Panjal Range, and may well correspond to Loharakoṭṭa referred to in the Rājataraṅgiṇī.106

As aforementioned, al-Bīrūnī calculated the longitudes and latitudes of the two forts in his mathematical treatise Al-qānūn al-Masʿūdī,107 while he provides the latitude of Fort Lahūr in the Taḥqīq mā li-l-Hind.108 In general, however, some problems arise to determine places based on the coordinates provided in these two works, as the figures are not always exact, need some revisions and diverge in the two works.109 Nevertheless, a study by the present author shows that one possible localisation based on these longitudes and latitudes concur with the localisation of the two places south to the Pir Panjal as proposed above.110 Fort Nandana is located in modern Pakistani Panjab, on a hilltop belonging to a series of mountains called the Salt Range.111 Based on the above, Figure 2 is a map depicting places where al-Bīrūnī lived and travelled, with known dates or time range.


Figure 2

Al-Bīrūnī’s places of residence

map prepared by the author

Thus, it can be established that al-Bīrūnī surely travelled in Gandhāra, northern Panjab and to a lesser extent in Multan in Upper Sind.112 Although the possibility that he visited other provinces of al-Hind after the composition of the Taḥqīq mā li-l-Hind cannot be completely discarded, evidence for such a hypothesis may never come forth. These observations concur with the assumptions made in earlier studies discussed above.113 Hereafter, I offer circumstantial evidence, discussed in a systematic manner, that corroborate these preliminary remarks.

1.2.3 Maḥmūd’s Interests and His Court

Beyond al-Bīrūnī’s list of latitudes and observations made on five places of al-Hind, other pieces of evidence indicate that his direct observations chiefly applied to Gandhāra and Panjab. First, al-Bīrūnī’s descriptions correspond to the interests of the Ghaznavids, particularly that of Maḥmūd. Maḥmūd focused much of his political and military attention to Gandhāra and Panjab. Thus, I argue that this interest, also motivated by economic concerns, is one reason for why al-Bīrūnī’s direct observations and research occurred in these two regions.

Prior to the Ghaznavids’ conquests, Gandhāra and Panjab chiefly remained a terra incognita to Muslim authors and rulers.114 Sebüktigīn, Maḥmūd’s predecessor, was the first known Muslim ruler to attack the cities of Laghmān and Peshawar. As for Maḥmūd, he concentrated many of his raids on the area by launching assaults on Laghmān in 1000, on Peshawar in 1001, on Wayhind in 1001 and 1008/1009, on Bhātinda (southern Panjab) in 1004/1005, on Fort Nandana in 1014, on Taneshwar in 1014, and on Fort Lahūr in 1015/1016 and 1021/1022, while in Sind he led expeditions to Multan several times in 1006, 1008 and in 1010/1011. In further southern and eastern regions, he also assailed Nārāyaṇapura (modern Rajasthan), also designated as Nārīn, in 1009, Kanauj and Mathura in 1018/1019, the forts of Gwalior and Kalinjar (in central India) in 1022, as well as the Somnāth temple in 1025/26.115 During his military excursions, he confronted several political and religious groups, the Late Shahis ruling over Gandhāra and Panjab, the Ismāʿīlīs in Multan, the Pratihāras occupying Somnāth, some parts of Rajasthan and of Mālava, and the Chandelas established in Kanauj and Mathura. Among those, the sultan’s conquests caused the Late Shahis’ dominion over Gandhāra and Panjab to end.116

Maḥmūd aimed to access the fertile land of Panjab, as well as to control the important routes leading to the Gangetic Plain.117 With the possession of Sind, the sultan could reach the seaport at Debal, near modern Karachi, enabling him to benefit from an important trade network through the Indian Ocean. As Inaba has shown on a map representing the territory of the Ghaznavids at Maḥmūd’s death, the sultan borrowed two main roads leading to the East, a southern one through the Sind and a northern one through Gandhāra, illustrating these two complementary advantages.118 Maḥmūd also proceeded on the northern route leading from Ghazna to Kanauj, via Peshawar and Lahore. Incidentally, al-Bīrūnī describes this road, when he deals with a network of routes starting from Kanauj and leading into various directions of the subcontinent.119 A parallel can thus be drawn between Maḥmūd’s territorial conquests and al-Bīrūnī’s intellectual exploration of Gandhāra and Panjab.


Figure 3

Example of Maḥmūd’s bilingual coins

Cappelletti 2015: 150, Ills 4B

Furthermore, the Ghaznavids attempted to establish their authority in these two regions while they chiefly conducted raids into other provinces of al-Hind. The minting of bilingual silver dirhams bearing their marginal and central legends in both Arabic (in Kufic script) and Sanskrit (in Śāradā script) (Figure 3) in Lahore located in modern Pakistani Panjab, reflects this attempt. A series of these coins was minted between the years 1027 and 1028 (418/419 AH).120

The legends in the margins of the coins in Arabic and Sanskrit provide dates of the coins as well as the place where they were minted, that is, Maḥmūdpur, possibly a name given after Maḥmūd to the capital of the region of Lahore (Lawhāwūr [‮لوهاور‬‎] according to al-Bīrūnī).121 The Sanskrit legend in the margin is found in two variants. Variant 2 of this marginal legend reads: ayaṃ ṭaṅkaṃ mahamūdapura ghaṭita tājikīyena saṃvatī 418/419.122 As V.S. Agrawala and Sara Cappelletti have shown, this reading stands as a translation of the Arabic original “In the name of Allah, this dirham was struck at Maḥmūdpur in the year 418/419.” The Sanskrit term ṭaṅka in the masculine gender means “stamped coin” and renders the Arabic dirham (‮درهم‬‎). The Sanskrit past participle passive ghaṭita, meaning “made” or “produced,” translates the Arabic passive verbal form ḍuriba (‮ضرب‬‎) translated as “was struck.”123 The meanings of the terms correspond in both languages.124

As for the central legend in Arabic, one finds the Islamic declaration of faith (šahāda), as well as a formula expressing the legitimization of Maḥmūd’s power on the obverse, as follows:

There is no God but Allah, Muhammed is the messenger of Allah. Maḥmūd, the right hand of the state, the custodian of religion.125

Before considering Maḥmūd’s political intentions behind the striking of the bilingual coins, it is worth discussing the underlying process of translation, because it resonates with the reflections in Chapter 4, 5 and 6 of the present book.126 The central Sanskrit legend on the reverse is puzzling. Several scholars have attempted to decipher it.127 The direct transliteration of the legend runs as follows: abyaktameka (1) muhammada avatāra (2) nṛpati mahamūda (3).128 The legend is written in an approximate Sanskrit and shall be emended in order to be grammatically correct, and thus interpreted. One possible emendation is the following: avyaktam ekam muhammado ’vatāro nṛpatir mahmūdaḥ. If one accepts this reading, the meaning of the legend is: “the unmanifest is one, Muhammed is the incarnation, Maḥmūd is the king.” This emendation facilitates drawing parallels between the Arabic and the Sanskrit legends in their communicative aim. The Sanskrit text here constitutes an attempt to translate the Arabic šahāda on the obverse, as is the case for the marginal legend discussed above.

The message communicated by the three Sanskrit sentences of this legend (see below Table 1) indeed matches that of the Arabic šahāda. In this interpretation, the Sanskrit term avyakta (unmanifest) was chosen by the translator(s) to render “Allah.” As shown in Section 2.4 below, al-Bīrūnī and his informants who possibly lived in Gandhāra and Panjab, namely in the same geographical context from which these legends originate, knew Yoga and Sāṅkhya philosophies. Further, al-Bīrūnī defined the concept of avyakta in the Taḥqīq mā li-l-Hind in a way similar to the Sāṅkhya philosophy as exposed in the Sāṅkhyakārikā and its commentaries.129 In this philosophy, avyakta is a key concept. It refers to the original cause (prakṛti), also known as the primary matter (pradhāna), which is the unique active origin of the phenomenal world, whereas God (īśvara) does not play an active role in the creation of existents.130

The Sanskrit word deva (deity) that one could have expected here is in general not used to refer to Allah by Persian and Arabic medieval authors, who rather relate deva to the concept of angels (‮ملك‬‎).131 Further, while the devas and the angels are several, avyakta and Allah are unique. The translator(s) of the legend had considered this feature when using the Sanskrit word eka (one) to transmit the idea of God’s uniqueness.132 The use of the Sanskrit term avyakta to translate “Allah” is also supported by variant 1 of the Sanskrit marginal legend, which has the expression avyaktīya nāme as a probable rendering of the Arabic bi-smi-llāh (‮بسم اللّه‬‎), meaning “in the name of Allah.”133

Thus, the concept of avyakta is the Sāṅkhya notion that best renders the concept of the Islamic God, as the unique creator of the phenomenal world. Another parallel between avyakta, that is, prakṛti, and Allah lies in the fact that they are conceived to be invisible to the common senses of perception. The term avyakta appears in the Bhagavadgītā, and its use in the Sanskrit legend of the bilingual coins could also be drawn from this work.134 Al-Bīrūnī abundantly quoted both the Kitāb Sānk and the Kitāb Gītā in the Taḥqīq mā li-l-Hind. These two books were thus presumably scriptures well known among his informants. Therefore, it is possible that principles of Sāṅkhya or of the Bhagavadgītā lie behind the use of the term avyakta in the Sanskrit legend of this coin.

Furthermore, the Sanskrit term avatāra (descent of a deity upon earth) appears to translate the Arabic rasūl (‮رسول‬‎; messenger), referring to Muhammed. The word avatāra does not constitute a literal translation of rasūl. Each of the terms bears distinct meanings according to the two religious contexts to which they belong, namely Hinduism and Islam, and at the same time shares the common point of referring to a figure who connects the divine sphere to the human world.135 Table 1 below shows the parallels between the two legends and illustrates the above discussion.

This method of interpretation that consists of substituting the source-concept with a target-concept whose meanings only partly overlap, such as in the cases of avyakta translating Allah and avatāra rendering rasūl, reminds one of al-Bīrūnī’s methods when he prepared his Kitāb Sānk and Kitāb Pātanğal.136 Did al-Bīrūnī himself translate—or help the translation of—the Arabic legend of these bilingual coins? The similarity between the way the šahāda was translated into Sanskrit on these bilingual coins and the way in which al-Bīrūnī translated Sanskrit literature into Arabic may suggest so, although this cannot be demonstrated. At any rate, if he had, he must have been helped by Indian scholars well acquainted with a form of Sāṅkhya philosophy and/or the Bhagavadgītā who conveyed to him the original concepts and were able to connect them with Islamic concepts and culture. Lastly, the above observations indicate that the method of cultural translation was common at the time.137

Table 1

Synthesis of the text of the bilingual dirhams of Lahore, based on Cappelletti138




There is no God but Allah

‮(لا اله الا اللّٰه).‬‎

The unmanifest is one

(avyaktam ekaṃ).

God / the creative cause is unique.

Muhammed is the messenger of Allah

‮(محمد رسول اللّٰه).‬‎

Muhammed is an incarnation [of God]

(muhammado ’vatāraḥ).

Muhammed is an intermediary figure between the human and divine worlds.

Maḥmūd, the right hand of the state, the custodian of religion

‮(يمين الدولة و أمين الملة محمود).‬‎

Maḥmūd is the king


Maḥmūd possesses the authority associated with his status.

Nevertheless, to come back the main discussion of this section, the bilingual coins were intended to legitimize the Ghaznavid power in the region.139 The transmission of the common Islamic šahāda in the form of a Sanskrit legend likely served as a means to enhance Maḥmūd’s authority through the appropriation of Indian concepts referred to by avyakta and avatāra. The Sanskrit legend addresses a non-Muslim Indian audience. Maḥmūd, expecting that illiterate Muslims would at least recognize the šahāda as a symbol, may have assumed the same for an Indian audience. However, as there exists no such thing as the šahāda in the Sanskrit tradition, local inhabitants of the region of Lahore would not have been able to recognize the Sanskrit legend as a symbol. Thus, despite the ruler’s attempt to integrate local traditions by way of the text of these coins, the Sanskrit legend probably had less impact on the population than desired by those who minted the coins.140 Even so, the example of these coins clearly illustrates the ruler’s efforts to integrate local concepts for the sake of establishing his authority in the region and reveals that intercultural exchange took place between Maḥmūd’s administration and local communities in Panjab at the beginning of the eleventh century.

After Maḥmūd’s death, the city of Lahore, Mandahūkūr in al-Bīrūnī’s writings, became the eastern capital of the Ghaznavid Empire and the outpost for the administration of the subjugated provinces. Governors were appointed in Lahore by the Ghaznavids and a Muslim community established there at an early date.141 In addition to the Ghaznavid attempt to control Lahore, literary sources reveal that the dynasty posted governors in Nagarkot142 and Nandana,143 and that the principles of Islam were being practiced in Bhātinda.144 Arabic and Persian primary sources also record that Maḥmūd appointed a certain Sukhpāl to administer Multan.145 No extant account suggest similar endeavours in cities such as Taneshwar, Kanauj or Somnāth, located further east and south in present-day India.

Further, the relative proximity of Ghazna to Gandhāra, to Panjab and to Sind suggests that these areas were more accessible to al-Bīrūnī than other provinces of al-Hind. These three regions were rather close to the centre of the Ghaznavid Empire compared to other places further east. Stabilization following Muslim incursions was often a long process, especially in regions distant from the Islamic centre. For instance, an entire century was necessary for the official establishment of Islam in Khwarezm (from the early 8th to early 9th c.) and approximately three centuries were required in Kābul (between the end of the 7th and 10th c.).146 The Ghaznavids had to repeat attacks on territory in al-Hind, including Laghmān, Wayhind, Multan, Nandana, Kanauj and Fort Lahūr, in order to establish and maintain control.147

If, in the areas of Panjab and Sind that were geographically close to Ghazna, political trouble between the Ghaznavids and the local rulers existed, it is likely that additional tension also occurred in the regions farther east. The region of Khwarezm and some parts of al-Hind were particularly far from Ghazna. Such remoteness prevented the Ghaznavids from holding them under their rule.148 Thus, I posit that due to the greater distance between Ghazna and some territories conquered by Maḥmūd, the sultan only conducted intermittent raids there, rather than establishing his authority through a governor or other officials. Al-Bīrūnī, however, needed a long-term cooperation with Indian scholars in order to gather his material on India and to pursue his translations of Sanskrit literature into Arabic. Consequently, I suggest that he did not visit far-away places in central and southern India, let alone stay for an extended period of time there.

Lastly, scholars largely presumed that al-Bīrūnī always accompanied Maḥmūd in his conquests of the East, and thus visited every place attacked by the sultan. This led some to conclude that al-Bīrūnī could observe the culture and customs of many regions of al-Hind. The subsequent paragraphs pose the question of his position at the Ghaznavid court and the extent to which he accompanied the sultan in his military campaigns in the East.

Al-Bīrūnī’s contemporary, Bayhaqī, reveals that conditions of the officials at the Ghaznavid court changed in accordance with the plots being orchestrated at the court, the sultan’s dispositions and other officials’ behaviours.149 To start with, the following anecdote related to the poet Firdawsī exemplifies the insecurity and volatility of official positions at Maḥmūd’s court. Having presented his epics to the sultan, the poet was not satisfied with his reward. After expressing his discontent, he was forced to go into exile in order to survive.150 While the details of this story vary from author to author and may not all be historically accurate, it offers a portrayal of Maḥmūd’s reputation and attitudes toward members of his court.

Al-Bīrūnī’s status at the court was also likely precarious. In the following passage drawn from the Taḥqīq mā li-l-Hind, al-Bīrūnī states:

I have found it very hard to work my way into this [subject] (i.e., al-Hind), despite my desire which I alone possess in my time, and although I sacrificed myself generously and as much as possible in collecting their books from places where they were likely to be found and in gathering those who were on the right way to [find] them from places where they were hidden. Who else besides me has the same [opportunity to learn this subject],151 unless he were endowed with Allah’s help, which I was deprived of in my capacity of movement, in which I was unable to come and go completely freely and independently. Thank Allah for what he granted me.152

Al-Bīrūnī was clearly aware of his reliance on the sultan’s benevolence, and, simultaneously, his precarious position at Maḥmūd’s court. On the one hand, he benefited from some support for his research, and, on the other, he had to subordinate himself to the ruler’s will. This passage does not, however, reveal the extent to which al-Bīrūnī was dependent upon Maḥmūd and his court. In addition, in the postface to the Kitāb Pātanğal, he explains:

As for the impossible [things] which are [referred to] in this book (i.e., the Kitāb Pātanğal), they can be accounted for in two ways. […] The second way is that the Indians have a greater propensity [for recounting absurd things] and a lesser one for reflection and study to such an extent that I could only compare their books on astronomical calculations—with respect to the meaning and with respect to the order and the arrangement. When pearls are mixed with dung, and jewels with clay, the [Indians] are not rightly guided to distinguish between these and they make no effort to study them and refine them […].153

In this passage, al-Bīrūnī explicitly rejects the very doctrines that he had detailed in the Kitāb Pātanğal. It is possible that, as a Muslim scholar, he disagreed with some philosophical principles presented in this book. He may also have included these comments to guard himself against a censorship by Maḥmūd, who would have considered the Kitāb Pātanğal unorthodox.

The Persian version of al-ʿUtbī’s work, the Kitāb-i-Yamīnī, records that Maḥmūd kept many captives from his military campaign in Khwarezm at the court, without specifying their identity or social rank. This work further comments that the captives were held in Ghazna and later sent to various regions throughout al-Hind.154 Al-Bīrūnī may have been among these men who were held captive in Ghazna or al-Hind, although no definitive evidence supports this hypothesis.

The opinion of modern scholars on the issue of al-Bīrūnī’s freedom and position during Maḥmūd’s reign is divided.155 The above remark by al-Bīrūnī does not necessarily indicate that he was a prisoner, but simply dependent on the royal court. However, in my view, he benefited at least from the space, resources and time necessary to pursue his work. In any case, al-Bīrūnī stayed for approximately thirty years (from 1017 to ca. 1050) at the Ghaznavid court, thirteen of which (from 1017 to 1030) were under Maḥmūd’s patronage. Therefore, whatever problems occurred between the scholar and the sultan, the two did collaborate for quite some time.

Evidence exists that some members of Maḥmūd’s court accompanied the sultan when he travelled. For instance, Farruḵī, a poet at the Ghaznavid court, states that he accompanied Maḥmūd on some of his conquests in al-Hind, notably to Somnāth, Kathiawar, Bulandshar, Kanauj and Taneshwar, as well as at the time of the sultan’s battle with the king Trilocanapāla.156 Bosworth notes that Bayhaqī and Gardīzī accompanied Maḥmūd during some of his campaigns157 and explains that the Dīwāns, that is, governmental bodies, generally followed the royal courts.158

Bayhaqī explains how a court official was required to organize and equip the sultan’s quarters, which included providing herds of sheep to allow the sultan to welcome guests wherever he was.159 Furthermore, Maḥmūd’s army required the contributions of engineers, prospectors, blacksmiths and others in order to enable the army to proceed, for instance by building roads and strongholds on the way to the foreign lands or by providing the facilities required by Maḥmūd’s military campaigns. In addition, numerous soldiers belonging to the army, elephants and other military equipment were part of Maḥmūd’s expeditions.160 These were large-scale excursions, and their organization necessitated the participation of a variety of specialists. Experts in various domains thus escorted Maḥmūd during his travels.

While al-Bīrūnī may have accompanied Maḥmūd during some of his military campaigns to the East, at times the scholar also certainly resided in places secured by the Ghaznavids. In addition, the sultan consulted with al-Bīrūnī as an astronomer and adviser to the court. Since early times, astronomers and astrologers assisted Muslim rulers in planning their military campaigns or other political matters. During the Delhi Sultanate, historians also counselled the rulers on political matters.161 At least two known examples indicate that al-Bīrūnī may have held such a position under Maḥmūd. In 1024 in Ghazna, the scholar met with a delegation sent by the Volga Turks and in 1026 with another one attached to the dynasty Kʾitan which ruled over southern Manchuria and northern China at the time.162

Further, due to his knowledge of Indian science and language, al-Bīrūnī certainly assisted the sultan as an interpreter and mediator in his interactions with Indians.163 One story narrates how Maḥmūd, returning from his raids against Mathura and Kanauj in 1018/1019, had met al-Bīrūnī somewhere between Kābul and Ghazna. Maḥmūd then showed him a precious stone stolen from a temple situated in Mathura.164 While it remains uncertain whether this incident really occurred, this narrative suggests that al-Bīrūnī did not travel with Maḥmūd’s army and court to Kanauj and Mathura. More convincingly perhaps, in The History of Khwarazm as handed down by Bayhaqī, a reference is made to al-Bīrūnī who narrates an episode during which Maḥmūd sent him an envoy from al-Hind while he was in Kābul. Thus, both pieces of evidence indicate that al-Bīrūnī did not always accompany the sultan on his military campaigns.165

If the scholar had accompanied Maḥmūd on some of his campaigns, it is, however, difficult to know exactly at which dates. Al-Bīrūnī visited India after 1017, more than fifteen years after Maḥmūd’s first raids in Laghmān and Peshawar, in 1000 and 1001 respectively. Thus, based on the above, al-Bīrūnī may have travelled during the sultan’s campaign of Fort Lahūr (1021/1022), Gwalior/Kalinjar (1022) and Somnāth (1025/1026). However, as shown in the present chapter, al-Bīrūnī did not venture eastward beyond Panjab during his visits to al-Hind. Thus, I conclude that the scholar accompanied Maḥmūd to the East during his journeys but did not necessarily travel as far as the actual battlefields. He indeed probably remained in regions where Maḥmūd had secured a certain level of political stability through his earlier raids.166 It is also likely that al-Bīrūnī at times travelled with a military escort independent from Maḥmūd and thus spent time in places where the Ghaznavids had already established some authority, enabling him to interact with Indian Brahmins.

1.2.4 The Various Sources of Information

In the preceding sections, I presented evidence from al-Bīrūnī’s writings and from his socio-political context showing that he chiefly travelled in Gandhāra and Panjab. Al-Bīrūnī’s descriptions in the Taḥqīq mā li-l-Hind of numerous places in India, however, led scholars to believe that he personally visited many regions of al-Hind. Yet, as I argue below, many of his descriptions are not based on direct experience, but on oral and written sources.

For instance, in the following passage al-Bīrūnī suggests that he did not personally see the regions of Kashmir and Varanasi:

This is the reason,167 too, why their (i.e., the Indians) sciences have disappeared beyond the limits [of the world] conquered [by the Muslims] and have fled to places where our hands cannot reach, namely Kashmir [‮كشمير‬‎], Varanasi [‮بانارسى‬‎] and other similar [places].168

In all likelihood, the toponym Kashmir here stands for the Kashmir Valley, as al-Bīrūnī describes it as “a plain that high and inaccessible mountains surround.”169 This extract is unique in that it explicitly indicates places where al-Bīrūnī did not visit at least prior to his composition of the Taḥqīq mā li-l-Hind: the Valley of Kashmir and Varanasi.170 This passage also suggests that the scholar was not able to cross the frontiers of the world conquered by the Ghaznavids, and rules out the possibility of him having travelled to South India. Another passage concurs with this latter observation:

Before, one or two foreigners could enter [Kashmir], especially Jews. Now, they do not let any Indians whom they do not know [enter it], let alone the others.171

Even so, al-Bīrūnī abundantly refers to Kashmir in the Taḥqīq mā li-l-Hind. He describes geographical, ethnic and social features at length, names cities and mountains, adduces itineraries leading to the Kashmir Valley, mentions Kashmiri customs,172 reports on the alphabets and scripts in use there,173 and presents detailed accounts of religious and astronomical practices.174 He portrays the Kashmir Valley in more detail than any other region discussed in the Taḥqīq mā li-l-Hind. Since al-Bīrūnī did not visit it, he must have based his account of Kashmir on sources of information other than his direct observation.

A few passages in al-Bīrūnī’s writings confirm this statement. In 1036, the scholar compiled a catalogue (‮فهرس‬‎) of the works of the physician and philosopher Muḥammad Ibn Zakariyyāʾ Rāzī (ca. 854–925/935). This catalogue also includes a list of his own works.175 This auto-bibliography notably provides the following two titles: Answers to the Questions of the Astronomers of al-Hind (‮الجوابات عن المسائل الواردة من منجمى الهند‬‎)176 and Answers to the Ten Kashmiri Questions (‮الجوابات عن المسائل العشر الكشميرية‬‎).177 These works are no longer extant, but their titles indicate that al-Bīrūnī interacted in some manner with Indian astronomers and with residents of Kashmir.

Several passages found in the Taḥqīq mā li-l-Hind also indicate that al-Bīrūnī had some contact with Kashmiri thinkers. For instance, regarding a festival celebrated in Kashmir al-Bīrūnī draws his information from the account of a certain Jīvaśarman. On this festival, he states: “The people of Kashmir whom I have seen do not agree with this [account] regarding the place and the time [of the festival].”178 In another passage, al-Bīrūnī mentions calendars of the year 951 of Śakakāla179 which had been brought from Kashmir.180 In addition, a further statement by the scholar testifies to intellectual interaction between him and certain Kashmiris. He declares: “We have verified these [methods] in the Zīğ (i.e., astronomical handbook) which we have composed for Syāvapala (?) (‮سياوپل‬‎) the Kashmiri.”181

These passages show that al-Bīrūnī met Kashmiris and exchanged books with them. M.S. Khan, noting the comprehensiveness and accuracy of al-Bīrūnī’s account of Kashmir, declares that Kashmiri scholars probably helped him in gathering information;182 this remark finds support in the above observations. Thus, although the Kashmir Valley was unreachable for al-Bīrūnī and his peers, intellectual exchange were taking place between the territory of Gandhāra and Panjab on the one hand, and that of Kashmir on the other. These interactions enabled al-Bīrūnī to describe at length a region that he had not visited himself at the time of the composition of the Taḥqīq mā li-l-Hind.

As for other regions of India, the Taḥqīq mā li-l-Hind also reveals that al-Bīrūnī communicated with people from Kanauj, Multan and Somnāth.183 As a result, the scholar could convey pieces of information about Kanauj, the alphabet and the calendar in use there, and its history.184 As for Somnāth, al-Bīrūnī informs his readers about the year when Maḥmūd attacked its temple, namely 416 AH (1025/1026),185 provides a detailed account of its idol and reports some myths associated with the temple.

Further, in Chapter 7 of the Taḥqīq mā li-l-Hind, al-Bīrūnī discusses the views on cosmography in the Ādityapurāṇa, the Vāyupurāṇa, the Viṣṇupurāṇa and the Matsyapurāṇa. He furnishes comparative tables with the different names of the regions of the world presented in these texts. In two tables, he provides the names of netherworlds (pātāla), oceans and islands (dvīpa) which he had heard (‮مسموع من الألسنة‬‎).186 Not only had he recourse to an interlocutor for information related to subjects such as cosmography; he was also able to supplement his knowledge of such topics based on Sanskrit literature with information provided by this oral source. However, al-Bīrūnī did not identify his interlocutor or mentioned his place of origin.

The Taḥqīq mā li-l-Hind contains other elements pointing to the importance of al-Bīrūnī’s interactions with Indians. Al-Bīrūnī refers to a traveller who communicated with him on the area located to the north-east of Varanasi and the realm of Nepal.187 It is furthermore possible to deduce from the Taḥqīq mā li-l-Hind that pilgrims also transmitted him information.188 In addition, al-Bīrūnī’s account of the Early and Late Shahis originates from an oral report. He states: “According to what I have heard, [the history of] such a lineage, [written] on a piece of silk, is found in the fortress of Nagarkot.”189 In the same passage, al-Bīrūnī explains that he could not find this silk document. The person(s) who informed him of its existence may have been the very narrator(s) of this royal chronicle. The scholar mentions the titles of several grammar books about which he came to know through an oral account.190 Although he did not visit many provinces of al-Hind himself, al-Bīrūnī met people, such as merchants, ascetics and pilgrims from various parts of India.191

Furthermore, al-Bīrūnī interacted with Indian scholars—Brahmins in general, and specifically astronomers and philosophers—some of whom must have belonged to the court of the Late Shahis, as I show below.192 Other passages indicate that al-Bīrūnī drew on oral sources concerning the custom of eating beef and the status of low-caste people vis-à-vis Brahmins.193 Al-Bīrūnī also devoted a complete chapter of the Taḥqīq mā li-l-Hind to the description of the lifestyle and duties of the Brahmins.194

Al-Bīrūnī provided references and descriptions of certain places located in al-Hind. These include Mandahūkūr, namely modern Lahore in Panjab, Mathura and Taneshwar. The scholar sparsely refers to Gujarat, Prayāga (Allahabad), the Kannara region, Varanasi and some places in present-day north-eastern India. His description of various itineraries starting from Kanauj also suggests that this information was orally transmitted to him, because these routes link many cities or regions of India he could not possibly have experienced first-hand. His account includes places on the eastern coast of present-day India (modern West Bengal), in the North (modern Nepal, Kashmir), in the North-East (modern Assam), in the centre (modern Bihar, Madhya Pradesh and Uttar Pradesh), in the South-West (modern Sind and Gujarat) and in the South (Kannara). It cannot be concluded, however, that al-Bīrūnī travelled to all of these places. Even though he demonstrates extensive knowledge of al-Hind, this knowledge was not necessarily based on personal visits. Instead, it wasoften accounted for by his interactions with Indians.

In addition, al-Bīrūnī accessed a large number of written sources, which built up his knowledge on the culture and sciences of al-Hind.195 He was acquainted with the Vedas, the Smṛti of Manu, the Viṣṇudharmottarapurāṇa, which he entitles Viṣṇudharma, and the Viṣṇupurāṇa, through oral accounts of Brahmins or quotations found in other books he read.196 He also consulted the Ādityapurāṇa, the Matsyapurāṇa, the Vāyupurāṇa,197 the Bhagavadgītā—referred to as the Kitāb Gītā by him—and the Mahābhārata. Al-Bīrūnī rendered into Arabic the two foundational texts related to the Sāṅkhya and Yoga schools of thought, respectively, the Kitāb Sānk and Kitāb Pātanğal. He also translated—or began translations of—astronomical works, such as Brahmagupta’s Brāhmasphuṭasiddhānta, referred to by him as the Brāhmasiddhānta, the Pauliśasiddhānta by Puliśa, and the Bṛhatsaṃhitā and the Laghujātaka by Varāhamihira.198 Lastly, he quotes works whose authors were known to him, but have yet to be identified: the Srūdhava by Utpala from Kashmir, the Karaṇatilaka by Vijayanandin from Varanasi, and certain works by Vaṭeśvara, who hailed from Nāgarapura, by Durlabha, a native of Multan, Śrīpāla and Jīvaśarman.199

Thus, thanks to informants coming from many regions of al-Hind and to accessing a large amount of Sanskrit literature, al-Bīrūnī gathered data that exceeded the scope of his travels in al-Hind. In contrast, I have presented evidence above that delineates the territory of al-Hind visited by al-Bīrūnī, a territory mostly confined to Gandhāra and Panjab. As a result, I hypothesise that some of his interlocuters belonged to this region and to the society living there, and that they constituted the primary layer of people he interacted with. Therefore, in the subsequent section, I discuss some preliminary socio-cultural traits of this area and, in Chapter 2, I further investigate this hypothesis.

1.3 Elements of Culture of Gandhāra and Panjab

1.3.1 Laghmān, Peshawar, Fort Rājagirī, Fort Lahūr and Fort Nandana

The present section focuses on the historical and social contexts of the five locales that al-Bīrūnī certainly visited, namely Laghmān, Peshawar, Fort Rājagirī, Fort Lahūr and Fort Nandana, all located in today’s eastern Afghanistan and Pakistani Gandhāra and Panjab.200 These five sites belonged to the kingdom of the Late Shahi dynasty, immediately prior to al-Bīrūnī’s journeys there. Arabic and Persian sources detail the encounters between Alptigīn, Sebüktigīn and Maḥmūd, and these local rulers.201 Sebüktigīn (977–997)202 launched several raids against the Late Shahis in Kābul as well as in the regions of Laghmān203 and Peshawar. Maḥmūd continued the attacks against the Late Shahis, mainly in regions stretching from present-day eastern Afghanistan to Pakistan. He defeated four kings of this dynasty: Jayapāla (ca. r. 964–1002),204 Ānandapāla (ca. 1002–1010),205 Trilocanapāla (ca. 1010–1021/1022)206 and Bhīmapāla (ca. 1021–1026/1027).207 These kings, originally established in Kābul, shifted their capital city most probably at first to Wayhind/Udabhāṇḍa, and later on to Nandana in the Salt Range and to Lahore, ultimately taking shelter in Kashmir.208

A preliminary examination of archaeological data and literary sources allows for a reconstruction of some details regarding these five locales. Laghmān is the first place which will be dealt with here, as it is the city closest to Ghazna. When al-Bīrūnī visited this region, he observed a solar eclipse, which he describes in the Taḥdīd al-amākin:

Again, though they (i.e., the Khurasanian calculators) had not discussed the solar eclipse that took place in Dhū al-Qaʿda, year four hundred nine of the Hijra, the reserved amongst them said that it would occur below the horizon of Ghazna, and that it would not be seen there. However, it happened that we were near Lamghān, between Qandahār (i.e., Gandhāra)209 and Kābul, in a valley surrounded by mountains, where the sun could not be seen unless it was at an appreciable altitude above the horizon. At sunrise, we saw that approximately one third of the sun was eclipsed and that the eclipse was waning.210

Archaeological excavations have not yet been conducted in Laghmān, which makes it difficult to reach conclusions on the site or on the type of society that lived there. However, in 1960 the head of a statue, probably dating to the second half of the first millennium, was found by accident in the region on a mound named Qalʾa Amir Muhammad (Tagao). According to Klaus Fischer who examined it, the head may be associated with the Early or Late Shahi dynasties.211 It represents a female goddess, Durgā Mahiṣāsuramardinī or Pārvatī.212 In addition, eight (Proto-)Śāradā inscriptions are found in caves at Laghmān.213 They, however, deserve examination to either decipher them or to update their existing decipherment in order to provide any historical information.

Literary sources indicate that the city was an important site during the last centuries of the first millennium CE. The account by Xuanzang, who visited Laghmān in the early seventh century, bears witness to the importance and prosperity of the region located on a trade road. Xuanzang also reports that Laghmān belongs to a country of Brahmins.214 In 982/983, the Ḥudūd al-ʿālam describes Laghmān as “an emporium of Hindūstān and a residence of merchants […] [which] possesses idol-temples.”215 Similarly, the Persian translation of al-ʿUtbī’s Kitāb-i-Yamīnī portrays the region of Laghmān as one of the most prosperous of the time and as belonging to the land of the Late Shahi king Jayapāla.216

Al-Bīrūnī also mentions the city of Laghmān in the Taḥqīq mā li-l-Hind, in which he gives its latitude217 and locates it across the stream of the River Sāwa (today’s Alishing River ?).218 In addition, when he discusses different calendars of al-Hind, he states that the people of Laghmān start the year with the month Mārgaśīrṣa (November–December).219 He also provides an alternative name for the city of Laghmān: Lanbaga (‮لنبَگا‬‎).220 Since al-Bīrūnī most probably visited the region during the year 1019,221 namely nineteen years after Maḥmūd’s takeover of Laghmān, his descriptions of this region show that local people kept their calendrical systems, based on ancient Hindu traditions, even after Muslims reached the region.

Further, in the Taḥqīq mā li-l-Hind, al-Bīrūnī conveys his observations made in Peshawar. There, he witnessed local practices:

After seven and a half gaṭī (‮گهری‬‎; pl. ‮گهریات‬‎) have elapsed, they beat the drum and blow a winding conch, named śaṅkha (‮شنگ‬‎) [in the Indian language] and spīd-muhra (‮سپيد مهره‬‎) in Persian. I have seen this in the land of Peshawar (‮بلد پرشور‬‎).222

The modern city of Peshawar lies in present-day northern Pakistan, east of Laghmān.223 According to Xuanzang, the population and the wealth of the city, designated then as Puruṣapura, were declining in his time, as was the practice of Buddhism.224 Al-Bīrūnī describes a timekeeping ritual involving the use of a conch (śaṅkha), a symbol of the Hindu god Viṣṇu. Nevertheless, in the absence of additional pieces of contextual information, it is difficult to determine which specific Indian religious group celebrated this ritual in the region of Peshawar. In addition to this passage, al-Bīrūnī mentions Peshawar at a few other places. He explains that it lies opposite of the Ghorvand River, to be identified with today’s Kābul River,225 and provides its latitude.226 He also recalls that Kaniṣka had a vihāra built there.227 His visit(s) to the region of Peshawar took place between the years 1017 and 1030. Lastly, al-Bīrūnī’s observation of the above ritual shows that locals continued their religious practices several years after the Ghaznavids annexed their territory to their empire, as was observed above in the case of Laghmān.

The Persian Kitāb-i-Yamīnī based on the Arabic work Al-Yamīnī by al-ʿUtbī, in the description of Maḥmūd’s attack on Peshawar, refers to the city as being located “in the midst of the land of Hindustan,”228 thereby suggesting that this territory was outside the frontiers of the Islamic boundaries and possible inhabited by Hindus. According to a recent study, numerous archaeological sites of the Valley of Peshawar could be associated with the Late Shahi dynasty, most of which have not yet been studied.229 Among those, the ancient site of Hund, that is, Wayhind/Udabhāṇḍa, was an important city for the Late Shahi kings, before the Ghaznavids definitively pushed them eastward, and most probably served as their main capital until the end of Jayapāla’s rule in the year 1002. It is thus possible that the former eminence of Peshawar waned with Wayhind/Udabhāṇḍa emerging as a new centre for the Shahi rulers from the mid-seventh century onward.230

In another passage of the Taḥqīq mā li-l-Hind, al-Bīrūnī describes two forts as strong places situated to the south of the Kashmir Valley:

Fort Rājagirī (‮راجكرى‬‎) lies to the south of [the mountain Kulārğak], and Fort Lahūr (‮لهور‬‎) to its west. I never saw stronger [forts] than these two; and three farsakhs from it (i.e., the mountain Kulārğak), is the town of Rājāwūri (‮راجاورى‬‎). Our merchants trade with it, but do not go beyond it. This is the frontier of al-Hind from the northern side.231

Al-Bīrūnī mentions the two forts a few times in the Taḥqīq mā li-l-Hind. In one passage, he describes Fort Rājagirī as being situated on the road from Kanauj to the Kashmir Valley, via Taneshwar, to the south of the Pir Panjal Range, before bifurcating north to the Valley.232 Fort Lahūr was most probably located to the south-west of the Pir Panjal Range233 Maḥmūd attempted to seize the fortress of Lohkot (i.e., Lahūr) twice. This would have facilitated his access to Kashmir.234 However, the sultan was never able to take it. When he attempted to attack Fort Lahūr, Trilocanapāla (ca. 1010–1021/1022), the Shahi ruler of the time, asked assistance to the Kashmiri king Saṅgrāmarāja (r. ca. 1003–1028).235 Al-Bīrūnī visited these regions between the years 1017 and 1030. He does not, however, provide any details that may suggest which religious traditions were followed in this region.

Farther east lies Fort Nandana where al-Bīrūnī calculated the circumference of the earth. He states:

When I happened to be living in the fort of Nandana in the land of India, I observed from an adjacent high mountain standing west of the fort, a large plain lying south of the mountain. It occurred to me that I should examine this method there. So, from the top of the mountain, I made an empirical measurement of the contact between the earth and the blue sky.236

The remains of two temples were found there in a rather impaired state which does not allow for thorough qualitative archaeological interpretations (see Temple A in Figure 4). However, these two edifices belong to a larger group of temples located in the Salt Range that have been associated with the Late Shahis. The discovery of various coins at the sites has made it possible to date this group of structures between the sixth/seventh and the eleventh centuries.237 According to some scholars, Nandana became the capital of the Late Shahis after they had been defeated in Wayhind/Udabhāṇḍa and shortly before they were attacked by Maḥmūd.

To my knowledge, al-Bīrūnī is the only Arabic author from this period who mentions Nandana, most probably because this site located much farther east than the four others was not known to Arabic and Persian authors who preceded him. He may have spent some time in Fort Nandana between 1017 and 1025.


Figure 4

Temple A, Nandana

Meister 2010: fig. 52

1.3.2 The Late Shahis and Their Religion

As shown above, all locales dealt in the preceding section were connected with the Late Shahi kingdom. Therefore, I set forth here pieces of archaeological material related to the Late Shahis, in order to paint a preliminary picture of their religion and society. The present section also illustrates the difficulties of dealing with the historical material associated with the Late Shahis due to the lack of systematic and comprehensive research ever since Rehman’s study was published in 1979. Before the Late Shahis were pushed eastward by Alptigīn and subsequently by the Ghaznavids, their kingdom extended from the north-west in Kābul and Wayhind/Udabhāṇḍa to the north-east in some areas of northern Panjab.238

Xuanzang (early 7th c.) explains how the ruler of Kābul region followed a specific pattern of seasonal migrations, residing in Kābul in summer and in the territory of al-Hind in winter.239 The succeeding rulers, the Early Shahis (ca. mid-7th to early 9th c.) and the Late Shahis (early 9th to early 11th c.), also probably originally maintained their capital in Kābul during summer and in al-Hind during winter. The Late Shahis established several administrative and political centres in the territory of al-Hind, namely Wayhind/Udabhāṇḍa, Fort Nandana and Lahore, successively, depending on the wars taking place between the Muslim rulers and these kings. When Alptigīn attacked Kābul, followed by the Ghaznavid army to Gandhāra, in the second half of the tenth century, the local rulers withdrew further east and expanded their kingdom to western Panjab.

According to Rehman, epigraphy and numismatics indicate that the Late Shahis were worshippers of the Hindu god Śiva.240 A stone found at the site of Wayhind/Udabhāṇḍa bears a (Proto-)Śāradā votive inscription that can be roughly dated from the second half of the tenth century, that is, during the reign of Jayapāla (ca. r. 964–1002).241 The inscription, mostly written in ślokas, is dedicated to Śiva. It begins with a formula to praise Śarva, which refers to one of the eight forms (aṣṭamūrti) of Śiva, known under the collective name of mūrtyaṣṭaka.242 In this inscription, Śiva is referred to in his form of Pinākin (lit. the one armed with a bow), which figures in a myth in which Śiva destroys three cities of demons (tripura). The text of this inscription also praises Śaṅkara, another name of Śiva,243 and Umā.244 The inscription mentions Jayapāladeva and his predecessor Bhīmadeva, gives a description of the town of Udabhāṇḍa and indicates the occasion on which it was written, that is, the construction of a temple devoted to Śiva.

There are many inscriptions associated with the territory and/or the period of the Late Shahis. However, their state of study and decipherment is poor, while in many cases their texts only consist of a few (partial) lines. Rodziadi Khaw, who recently catalogued and described numerous (Proto-)Śāradā inscriptions found in Gandhāra, classifies them into five categories, ranging from “unpublished” to “published and satisfactorily deciphered” (2015: 94). These labels in his categorization highlight the poor degree to which these inscriptions have been studied so far.245 All of the inscriptions associated to the Late Shahis reflect a society involving Brahmins in its official activities and celebrating Hindu deities, including Śiva.246 Some of these inscriptions bearing the mention of a date, with or without a year, show that the persons who had the texts engraved followed the traditional luni-solar system commonly in use in ancient India.

Coins associated with the Late Shahis are difficult to interpret. The following description of these coins is mostly based on Rehman’s work and aims at offering an impression of their types.247 Two common types of coins are associated with the Late Shahis. The first type portrays a bull and a horseman (gold, billon, and silver).248 This type, which is already seen with coins minted by the Indo-Scythian and Indo-Greek kings, was also widespread later on in Gandhāra. The second common type of coin (made of copper) linked to the dynasty depicts an elephant and a lion. The motifs seen on both types of coins are recurrent not only in early Indian coinage, but also in Hindu iconography. Originally the lion figures as the vehicle (vāhana) of the goddess Durgā, and the bull often represents Nandin, Śiva’s vehicle, both deities being commonly depicted together with these two animals in Indian art and coinage. Even so, the occurrence of these motifs does not necessarily point to a possible cult of these deities.249

The coins of king Sāmantadeva (late 9th c.), referred to as Sāmanta by al-Bīrūnī, display a trident (triśūla), Śiva’s attribute, and a star-shaped pendant as a decorative feature of a horse. A coin, issued by Bhīmadeva (ca. r. 921–964), referred to as Bhīma in al-Bīrūnī’s report, represents a king seated on a throne and a woman on its obverse. The two persons display the clothing and hairstyles of the time. Above their heads are a triśūla and a diamond-shaped object. On the reverse, a king, similar to the one shown on the obverse, is seen beside Lakṣmī, the consort of Viṣṇu and the goddess of wealth and prosperity. This representation on Indian coinage reveals a close connection between kingship and this goddess.250 However, this representation cannot suggest that the Late Shahis were specifically devoted to Lakṣmī and Viṣṇu. This coin may have been minted for a special occasion on which celebrating Lakṣmī as a provider of wealth was deemed necessary, or for the celebration of Dīpāvali, the Hindu festival dedicated to her.

As for architecture, several temples belonging to the territory of the Late Shahis, including the aforementioned Fort Nandana in the Salt Range, display similar features.251 These temples dating from the sixth to the eleventh century, have, for instance, conical nāgara roofs, a type of śikhara construction.252 In Barikot in the Swāt Valley, an area also belonging to the Late Shahi territory before the arrival of the Ghaznavids, fragments of marble sculptures, perhaps of Viṣṇu and of figures representing his attribute and associated with the ruins of a monumental temple, has been uncovered. According to archaeological surveys, this temple was in use from the late seventh to the late tenth century.253 Moreover, as Rehman discusses, a few sculptures representing Hindu deities, such as Viṣṇu, Śiva, Kārttikeya and Durgā, were found at different sites in the kingdom of the two Shahi dynasties. However, further information, such as on their exact dates and archaeological contexts, is often missing.254 Lastly, a passage of the Rājataraṅgiṇī refers to a temple dedicated to Viṣṇu and built by king Bhīmadeva.255

This material demonstrates that populations living in the Swāt Valley, Gandhāra and Panjab celebrated various Hindu gods from approximately the seventh to the early eleventh century, a geographical area and chronological period which correspond to that of the kingdoms and rules of the Early and Late Shahis. Furthermore, this material suggests a certain continuity in ritual practices of each of the two ruling dynasties, rather than a fracture between the practices of both dynasties. Yet, a systematic and thorough investigation is necessary in order to reach conclusions about how these rulers lived their religion and how the dynastic change took place.

Furthermore, in the Taḥqīq mā li-l-Hind, al-Bīrūnī describes thirty-seven Hindu festivals, most of which were likely taking place in the regions he visited in al-Hind and celebrated thus within the society ruled over by the Late Shahis. Among these festivals, ten were dedicated to women or to female deities, one to Śiva, four to Kṛṣṇa, the avatāra of Viṣṇu, and six to Brahmins.256 If the celebrations of these festivals are indeed to be connected to the society of the Late Shahis, then they indicate that this society did not worship one specific Hindu god at the exclusion of another one and rather venerated several Hindu deities concurrently. However, at the current state of research on the Late Shahis’ material culture, and without an extensive and multidisciplinary examination of historical data related to them, any further conclusions would remain conjectural. Lastly, the above overview shows that these rulers used Sanskrit as their literary religious language and (Proto-)Śāradā as their script in inscriptions and coinage.257

1.4 Concluding Remarks

This chapter highlighted the importance of determining the historical and social contexts of al-Bīrūnī’s life in order to better understand his journeys in al-Hind. It showed that he spent his life in three major geographical zones: 1) western Central Asia (Kāṯ, Ray, Ğūrğān and Ğūrğānīya), 2) eastern Central Asia (Kābul, Ghazna and Gandhāra), and 3) the western Panjab (Fort Nandana). The cultural and political contexts diverged in these three zones. In the western part of Central Asia, Islam became well-established by the tenth century. In its eastern areas, Islamic authority had been continuously challenged by the Early and Late Shahis, and the region could be subdued only at the end of the tenth century. Islam, with the Ghaznavids, entered western Panjab in the early eleventh century, that is, at al-Bīrūnī’s time. The territory of the above three geographical zones was divided into various political entities: the Afrighids ruled in Kāṯ, the Buyids in Ray, the Ziyārids in Ğūrğān, the Maʾmūnids in Ğūrğānīya, and the Ghaznavids in Kābul, Ghazna and western Panjab.

The present chapter further pointed out the economic, cultural and intellectual prosperity of each of these regions, which fostered communities of literates with whom al-Bīrūnī could interact. Ray and Ğūrğānīya in particular were influential and respected intellectual centres where he could access important libraries. In addition, the regional ruler of Ğūrğān, Qābūs, who supported him in his efforts, and Maḥmūd, whose court included many scholars, most probably facilitated al-Bīrūnī’s research. Kābul, Ghazna and western Panjab, far from being isolated or sterile areas, were at the centre of various exchanges between the West and East. Located in a frontier zone, but connecting different cultural areas, these regions witnessed important cultural changes and exchanges.

As al-Bīrūnī crossed this cultural frontier, he discovered Indian religion, sciences and literature chiefly in Gandhāra and Panjab. Several elements suggest that he only visited a confined area of the South Asian subcontinent: al-Bīrūnī’s direct and explicit observations made in Laghmān, Peshawar, Fort Rājagirī, Fort Lahūr and Fort Nandana located in Gandhāra and Panjab, Maḥmūd’s interest in these areas and control over them, and the variety of al-Bīrūnī’s sources of information for the Taḥqīq mā li-l-Hind. Thus, whereas his writings indicate sojourns in the mentioned regions, there is no positive evidence of him travelling to cities like Taneshwar, Kanauj, Somnāth or Mathura, which were, however, conquered by Maḥmūd. I therefore argue that al-Bīrūnī chiefly observed cultural traditions of the above five places that are Laghmān, Peshawar, Fort Rājagirī, Fort Lahūr and Fort Nandana.

The above discussion of the cultural history of the five places visited by al-Bīrūnī and the religion of the Late Shahis is a preliminary one. In order to grasp the complex history of the Late Shahis and the practices of their society, a thorough and deeper investigation into texts, as well as archaeology of their sites, epigraphy and numismatics, is necessary. The little material presented above merely suggests that the Late Shahi adopted a form of Hinduism as their religion and the structure of a Brahminical society. However, for the purpose of the present book it is sufficient to note that most of al-Bīrūnī’s travels in al-Hind actually took place in the territory of the Late Shahis.

Lastly, al-Bīrūnī observed local and living traditions in Laghmān and Peshawar, two sites defeated by the Ghaznavid long before his visits there. These examples point to the survival of pre-Islamic practices in territories conquered by the Ghaznavids. In addition, these traditions are also to be connected to customs adopted by the society of the Late Shahis. These observations are supported by the analyses presented in Sections 2.3 and 2.4 that show that al-Bīrūnī observed Hindu traditions and mainly interacted with Brahmins.


Well-grounded studies situate al-Bīrūnī’s native town in Kāṯ (Kennedy 1970: 147–148; Shamsi 1979: 261–265; Yano 2013).


His observations recorded in the Taḥdīd al-amākin indicate that he stayed in the region of Khwarezm until the year 995 (Ali 1967: 77 and 211; Kennedy 1970: 148; Shamsi 1979: 268–269; Said & Khan 1981: 125). See the map picturing al-Bīrūnī’s places of residence on Figure 2 of the present book.


Bosworth 2011a and b.


Le Strange 1930: 447; Bosworth 2012a.


Bosworth & Crowe 2012.


Bosworth 1976: 25 and 2011a.


Bosworth 1976: 21–22.


Bosworth 2012a.


See also Bosworth 2012b.


Ali 1967: 78; Bosworth 2012a. On the rulers of Ğūrğānīya, see Debarnot 1985: 67–70 and Bosworth 2011c. Ğūrğānīya should not be mistaken for Ğūrğān, which is the ancient name of modern Gorgan, in present-day Iran.


Al-Bīrūnī was in the region of Kāṯ in the year 997, as he observed there a lunar eclipse conjointly with the mathematician Abū l-Wafāʾ, who was based in Baghdad. However, during the two the periods extending from 995 to 997 and from 997 to 1000, the exact events of his life remain obscure (Ali 1967: 214–215; Kennedy 1970: 148–149).


Al-āṯār (2001), p. 433.18–19; Sachau 1879: 338.


Nagel 1990. See Guy Le Strange (1930: 186), referring to the tenth-century geographer Ibn Ḥawqal.


Le Strange 1930: 227.


Gutas 2011.


Bosworth 2012d.


Bosworth 2012c.


Referred to by Le Strange (1930: 377).


Hartmann & Boyle 2012.


Al-Bīrūnī calculated the latitude of Ğūrğānīya and made other astronomical observations there up to the year 1016. See Ali 1967: 46–49, 50, 87, 96 and 113.


Le Strange 1930: 448.


Balty-Guesdon: 1992; Sourdel 2012a; 2012b.


Today, an institution of the same name is located in modern Khiva (Uzbekistan), approximately 170 km south-east of Kunya-Urgench.


Goldstein 2012.


Said & Khan 1981: 66–69; Dietrich 2012.


Gutas 2011.


Gibb 2012. See also Kuwayama (2002: 181–182) for a summarized table of Islamic incursions in the region in the seventh century. See Inaba (2015: 112–114 and 118–121) who provides a detailed account of the Muslims’ attacks on the two cities, their interactions with regional dynasties and the foundation of the Ghaznavid dynasty.


On Islamization in Central Asia as a process, see, for instance, De La Vaissière 2008 and Arezou 2017.


Nazim 1931: 24–26; Inaba 2015a: 119.


Nazim 1931: 56–60. See Bosworth 2011e: 372–374 for an account of the relationship between the Khwarezm Šāh Abūʾ l-ʿAbbās Maʾmūn bin Maʾmūn and sultan Maḥmūd.


Ali 1967: 86 and 271; Shamsi 1979: 270–274.


Hermelink (1977) places the date of al-Bīrūnī’s death in the year 1048 CE (440 A.H.), while Kennedy (1970: 151) argues that it occurred a few years later. See also Sami 1973: 27.


Inaba 2013: 81, figs. 2–3, and 85–87.


Dagens et al. 1964: 52.


For instance, Ibn Ḥawqal (10th c.), referred to in Bosworth 2012e. See also Le Strange 1930: 348–351.


As stated by Rehman (1979: 8) with reference to the Ḥudūd al-ʿālam. See also Inaba 2015a: 110–112.


Hallade 1968: 33; Schlumberger 1978; Sourdel-Thomine 1978; Elverskog 2010: 26; Dietz 2007: 49–50; Inaba 2015b; Alram 2016: 57–58. For the early connections between Kāpiśī-Gandhāra regions with Kashmir on the one hand, and with al-Hind on the other, see Grenet 2002: 212–214.


See fn. of the introduction to this book.


Rehman 1979; Kuwayama 2002; Inaba 2004, 2005, 2010 and 2015a; Alram & Pfisterer 2010; Alram 2016: 45–151; Elverskog 2010: 27; Vondrovec 2010.


Rehman 1979: 289–292; Rowland & Rice 1971: 32–33. Other examples of mixed influences in the art and architecture of Central Asia are displayed in Dagens et al. 1964.


-tigīn is a usual ending of Turkish names, e.g., Alptigīn, Sebüktigīn.


On the qabaʾ see Flood 2009: 65–67.


The ruins of Nagarkot, also referred to as Bhīmnagar (present-day Kangra Kot or Kangra Fort), are in today’s Himachal Pradesh on the foothills of the Himalayan range. See also Dey 1927: 135; Nazim 1931: 89–91; Bhattacharyya 1991: 227. On Nagarkot, and its possible role in al-Bīrūnī’s knowledge of India, see below pp. –.


This is a transcription from the Arabic kanika ğit. On the vihāra of Peshawar, see Dani 1969: 37–39 and Salomon 2018: 44 and 47.


The date of Trilocanapāla’s death is corroborated by other literary accounts (Nazim 1931: 95).


Taḥqīq (1958), p. 348.10–351.3; Sachau 1910: II/10–13.


On references to Turkish people, see the study by P.B. Golden in Bazin et al. 2012. On references to Tibet in early medieval Arabic and Persian geographical accounts, see Akasoy 2011.


Inaba 2004: 107–108, 2005 and 2010: 443–449. See also Rehman 1979: 37–47 and 294–297, Wink 1990: 114–128 and Kuwayama 2002: 262–265.


Rehman 1979: 45–47.


On the date of Kaniṣka, see Dietz (2007: 57, n. 51).


See studies by De La Vaissière (2003), Grenet (2002), Inaba (2005 and 2010), Vondrovec 2010, Alram & Pfisterer (2010). Louis de La Vallée Poussin (1935: 17–18) and Dinabandhu Pandey (1973: 63) have previously cast doubt on the historical value of this section of al-Bīrūnī’s account.


Verardi 2011: 308–309; Alram 2016: 106, 125 and 143. See the discussion on the relative chronologies of some of these sites, as well as on their religious affiliations, in Kuwayama 2002: 162–193, 200–207 and 222–259.


Rehman 1979: 285.


La Vallée Poussin 1935: 19; Pandey 1973: 67; Rehman 1979: 88 and 298–3002; Mishra 1983: 31–32.


Rehman 1979: 89.


The society of the Late Shahis remains relatively unknown despite their important role in the early encounters between Islamic and Hindu cultures between the mid-ninth and the early eleventh century CE. The outline in Section 1.3 only superficially broaches the subject. Interest in the Shahi kings and their material culture has lately increased among scholars. See for instance the international project Cultural Formation and Transformation: Shahi Buddhist Art and Architecture ( [accessed October 2023]). This project is led by Prof. Em. Deborah Klimburg-Salter (Dept. of Art History, Univ. of Vienna; Harvard Univ.) and is part of the research centre CIRDIS (Center for Interdisciplinary Research and Documentation of Inner South Asia). An edited volume presenting the most recent research on the two Shahi kingdoms is one of the outcome of this project. Recent publications include Filigenzi 2015: 36–40, on the Shahi period, and Khan 2017 on the Late Shahis. On the Late Shahi coinage, see Thomas 1846, MacDowall 1968 and Alram 2016: 151–153. For earlier references, see also Cunningham 1875: 82–83, Dani 1969: 54–56, Mishra 1972, Pandey 1973: 77–132, Rehman 1979: 89–167 and Wink 1990: 125.


On Islamic geography see Miquel 1967; for a global overview of the geographical accounts of al-Hind, see Wink 1990: 109–192. On the Šašnāma, see Elliot & Dowson 1867: 131–211, Ahmad 2005: 98, n. 1, and Friedmann 2012. On al-Balāḏurī, see Elliot & Dowson 1867: 113–130, Bosworth 1988 and Becker & Rosenthal 2012. On Ibn Ḵurdāḏbah, see Elliot & Dowson 1867: 12–17, Bosworth 2011d and Hadj-Sadok 2012. On al-Iṣṭaḵrī, see Elliot & Dowson 1867: 26–30, Bolshakov 2012 and Miquel 2012. On Ibn Ḥawqal, see Elliot & Dowson 1867: 31–40 and Khalidov 2011.


Verdon 2015: 34–46. An analogous example is that of the city of Ghazna which appears in geographical writings after being included in the Islamic territory (Inaba 2015a: 114–116 and 2015b: 105–107). See also observations by Rehman (1979: 1–3) on al-Hind and its evolving frontiers, and Brauer’s work (1995) on the conceptualization of frontiers in medieval Muslim geography. On the concept of frontiers in ancient times, see Thapar 2002: 47–48. Derryl N. Maclean (1989: 1–82) discusses at length the process by which Buddhist communities almost disappeared from Sind when Muslim settled in the region.


On the date of the compilation of the Taḥqīq mā li-l-Hind, see above fn.  of the introduction to this book.


See, for instance, Ralph W. Brauer (1995: 33–44) on the concept of boundaries in Arabo-Islamic geography.


The delimitation and description of the inhabitable world constituted a common topic among Arabic authors, much indebted to Ptolemy’s views (Zadeh 2011: 88–91). Early medieval Muslim geographers generally conceptualized the division of the world into climes. See al-Bīrūnī’s description of them in the Tafhīm (Tafhīm, pp. 143–145, no. 241). In the same work, al-Bīrūnī describes different regions of the world and provides a map of them (Tafhīm, pp. 121–125, nos. 211–212).


Taḥqīq (1958), p. 157.1–11; Sachau 1910: I/197–198.


This mountain (‮كُلارْجَك‬‎) is located south of the capital of Kashmir, i.e., Srinagar, according to al-Bīrūnī (Taḥqīq [1958], p. 167.1–2; Sachau 1910: I/207). It may be one of the mountains belonging to the Pir Panjal Range (Stein 1900: II/297–298).


Fort Lahūr (or Lawhūr) does not stand for present-day Lahore, in Pakistani Panjab. Al-Bīrūnī appears to refer to the latter city by the toponym Mandahūkūr. On the locations of Fort Lahūr and Fort Rājagirī, see below p. 30. On al-Bīrūnī’s mention of these places, see below pp. 51–52.


The farsaḵ is a historical unit of distance of Persian origin. Traditionally, it represented “the distance that men could march in an hour” (Bivar 2010) and equated to approximately 5 to 6 km.


Rājāwūri probably corresponds to the modern Rajauri district situated to the south-east of Punch in present-day Jammu and Kashmir. See also Sachau 1910: II/320, Dey 1927: 165, and Bhattacharyya 1991: 258.


Taḥqīq (1958), p. 167.5–9; Sachau 1910: I/208.


In Sanskrit, the compound suvarṇadvīpa means “golden island.” It was probably used as a name for the Indonesian island of Sumatra.


Taḥqīq (1958), p. 169.3–5; Sachau 1910: I/210.


Taḥqīq (1958), p. 214.3–5; Sachau 1910: I/258.


Sachau 1910: I/xv.


Sachau 1910: I/xvi.


Sachau’s comments in the preface to his edition of the Taḥqīq mā li-l-Hind were brought to my attention by Prof. Emeritus Michio Yano.


Sachau 1887: xiii and 1888: 6.


Chatterji 1951: 86.


Courtois 1952: 35.


Law 1955: 9–10.


Kennedy 1970: 150.


Khan 1976: 91, n. 24.


Baloch 1973: 39.


Dani 1979: 186–187.


Bosworth 1977: 64. On Lahore in the context of Maḥmūd’s rule, see below pp. 32–38.


Dani 1979: 187.


Khan 2001: 21.


Syed 2003: 36.


Said & Khan 1981: 82–86.


Mishra 1985: 11.


Mishra 1985: 11–13. On the latitudes, see below pp. 28–29.


On Bayhaqī, see Bosworth 2004: 20.


Taḥdīd [1992]; Ali 1967.


Taḥqīq (1958); Sachau 1910. Al-Bīrūnī’s mathematical treatise, Al-qānūn al-Masʿūdī, is only sparsely referred to in the present book.


Al-Bīrūnī witnessed the way in which the Indians catch gazelles (Sachau 1910: I/195) and a struggle between an elephant and an animal he calls gaṇḍa (Id. I/204). He informs the reader that he had seen Brahmins (Id. II/134.; on al-Bīrūnī’s meeting with Brahmins and astronomers, see below Section 2.3) and enumerates more than thirty Indian religious festivals, some of which he may have observed (Id. II/178–185. See Verdon 2019a: 68–71).


According to Sachau, the reading Kirī is also found. This stronghold was most probably located between Kābul and Peshawar near Jalalabad (Sachau 1910: II/341–342; Bivar 1979: 169–170 and 172–173).


In the Taḥqīq mā li-l-Hind, al-Bīrūnī most probably refers to Lahore as Mandahūkūr (‮مندهوكور‬‎), described as the capital of the region of Lawhāwūr (‮لوهاور‬‎; also Lawhāwar) and located to the east of the Ravi River (Īrāwah; ‮ايراوه‬‎; Taḥqīq (1958), p. 165.7; Sachau 1910: I/206). Mandahūkūr may be a corrupted form of Maḥmūdpur (see also below pp. –).


Taḥqīq (1958), p. 270.5–11; Sachau 1910: I/317. Al-Bīrūnī also provides the latitudes of Ujjain, Taneshwar and Kanauj based on Arabic works, by Yaʿqūb Ibn Ṭāriq and Abū Aḥmad Ibn Ğīlaġtakīn, and on Sanskrit sources, such as Balabhadra’s work and the Karaṇasāra by Vaṭeśvara (Taḥqīq (1958), pp. 269.10–270.5; Sachau 1910: I/316–317). On Vaṭeśvara, see Pingree 1991: 555–556.


Sachau 1887: xii–xiii and 1910: II/317. See also Al-qānūn (1955), pp. 561–562 and 573–574. The reader may refer to A.D.H. Bivar’s study (1979) on the coordinates of places located between Ghazna and Peshawar as found in the Taḥqīq mā li-l-Hind and in Al-qānūn al-Masʿūdī. Bivar, for instance, points out the discrepancies between the coordinates found in both works, as well as between al-Bīrūnī’s calculations and actual coordinates.


Taḥqīq (1958), p. 270.13–15; Sachau 1910: I/318.


Taḥqīq (1958), p. 16.17–18; Sachau 1910: I/22. The passage is translated below p. .


This place is also known as Muraṇḍa (Dey 1927: 113).


Rehman 1979: 13.


Taḥqīq (1958), p. 165.5–11; Sachau 1910: I/206.


Bivar 1979: 169. See below p. , for al-Bīrūnī’s description of this eclipse.


Kennedy 1970: 150.


Taḥqīq (1958), pp. 164.15–165.2; Sachau 1910: I/205.


Abdur Rehman (1979: 275–276 and 2003: 9) argued that Rājagirī should be identified with the so-called Rāja Gīrā’s castle, a site located on the Mount Rāja Gīrā near modern Uḍegrām in the Swāt Valley. The ruins of another fort, known as Rāja, are located around 8 km north-east of the modern city of Taxila. Considering al-Bīrūnī’s description of Rājagirī Fort’s location, these two other options are, however, unlikely.


According to Arabic and Persian literary sources, Maḥmūd of Ghazna attempted twice to capture a place situated on the foothills of Kashmir, known as Lohkot or Loharin (Nazim 1931: 104–105). Marc Aurel Stein identified Lohkot with the Castle of Lohara, referred to as Loharakoṭṭa in the Rājataraṅgiṇī, and located south-west of the Pir Panjal Range (RT IV.177, p. 50; Stein 1900: I/138 and II/293–300). This identification is more probable than with Chota Lahore (small Lahore) lying to the east of the Swāt Valley which has been at times identified with the Fort Lahūr of al-Bīrūnī. See also Khan 1979: 223–224.


Sachau 1887: xii–xiii; Al-qānūn (1955), p. 562; 574.


See above p. .


On al-Bīrūnī’s calculations of these coordinates, see Bivar 1979.


Verdon forthcoming.


Bhattacharyya 1991: 229.


Although al-Bīrūnī gives the latitude of Multan, the research of this book suggests that he did not conduct his research there. See below p.


See above pp. –.


Grover 2006: 44; Verdon 2015: 38–40.


Nazim 1931: 86–122. Inaba (2013: 77–79, table 1), provides a table listing Maḥmūd’s conquests in Central Asia and India, based on several primary sources.


Inaba 2016. See also Thapar 2002: 508: map 12. Several Indian dynasties that ruled over other parts of al-Hind, which correspond to present-day north-western India, and were attacked by Maḥmūd are enumerated in Mishra 1983: 69–70. On the encounter between the Late Shahis and the Ghaznavids, see Rehman 1979: 130–167.


Nazim 1931: 88–89.


Inaba 2013: 76, fig. 1. I am grateful to Professor Minoru Inaba for having drawn my attention to more specific military interests of Maḥmūd.


Taḥqīq (1958), pp. 155–170; Sachau 1910: I/196–212. See map 2.1 in Verdon 2015: 42. Kanauj being the starting point of al-Bīrūnī’s itineraries has to be seen in the light of the prestige this city enjoyed in the post-Gupta period. See Thapar 2002: 405–407 and Elverskog 2010: 45. Similar itineraries have been reconstructed in Schwartzberg 1978: 33 and Deloche 1968: planche VII. Grover (2006: 46–48) also describes routes passing through Panjab. For an account of the routes described by al-Bīrūnī, see Verdon 2015: 40–43. Indian cities are also mentioned by al-Bīrūnī in the Tafhīm (Tafhīm, pp. 143–144, no. 241).


Numbers 13–14 in Thomas 1859: 23–24. V.S. Agrawala (1943) and Sara Cappelletti (2015: 69–102) examine these coins in detail. Two other coins minted in 1021/1022 CE (412 AH) have been studied by Thomas under numbers 11 and 12 (1859: 22–23). Readers further interested in these bilingual dirhams may consult the studies by Thomas (1847 and 1859), Agrawala (1943), Bhattacharyya (1964) and Cappelletti (2015). See also Chatterji 1951: 96–100, Said & Khan 1981: 88, Khan 2001: 62–63 and Flood 2009: 41–42.


According to Thomas’s reading: maḥmūdsar (Thomas 1859: 23–24). See Khan 1979: 221–226 and fn. of Chapter 1 of the present book. Bosworth (2007: 299) doubts the identification of Lahore as Maḥmūdpur.


Agrawala 1943: 155 and 157–158; Bhattacharyya 1964: 54–56.


For a clear and synthetic exposition of the two variants, see Cappelletti 2015: 72 and 82. See also Agrawala 1943: 155–161 and Bhattacharyya 1954: 114.


The Sanskrit legend is, however, grammatically incorrect, as the agreement of the grammatical gender is not accurately observed. The correct form in the nominative case should be: ayaṃ ṭaṅkaḥ ghaṭitaḥ (or ayaṃ ṭaṅko ghaṭitaḥ), not ṭaṅkaṃ, since this word in the sense of “a stamped coin” is a masculine noun (I am grateful to Dr. Maas for having brought this to my attention). The Sanskrit mahamūdapura refers to the minting place. The word is, however, not found in the expected locative case. These features may indicate that the person(s) behind this translation lacked grammatical knowledge of Sanskrit.


Agrawala 1943: 156; Bhattacharyya 1964: 53; Cappelletti 2015: 87.


Two variants (A and B) of the Arabic central legend also exist (Cappelletti 2015: 72, 87 and 93). The differences between these variants are, however, relatively minor, as far as the present discussion is concerned.


See the discussion by Cappelletti (2015: 69–72).


Thomas first described one of these coins in 1847 (no. XLII) (1847: 269–270 and 323–324). After the finding of new exemplars of this type of coin, Thomas revised his reading (nos. 11 to 14) (1859: 22–24). I follow here Agrawala’s reconstruction (1943: 156, lines 1–3).


See below pp. , and .


See chapter 3, section 2 and chapter 6, section 3.3 below.


See Ernst 2003: 177.


See also Cappelletti 2015: 95–96 on the interpretation of this expression.


Cappelletti 2015: 72 and 82.


Cappelletti 2015: 94–95.


The term avatāra was also used in a Kālacakra text, the Vimalaprabhā, to qualify Muhammed (Cappelletti 2015: 98–102). See also Chatterji 1951: 96–97.


For a definition of the translational strategy of substitution, see below Section 4.4.2.


Bhattacharyya 1964: 56.


Variant B of the Arabic legend is represented in this table. See Cappelletti 2015: 87 and 93–102.


Bhattacharyya (1954: 115–116) provides other examples of attempts of legitimization by using pre-existing symbols on coinage.


This observation has been inspired by a discussion with Ms. Sara Cappelletti.


In 1163, the Ghaznavids lost Ghazna and established their government in Lahore (Bosworth 2007: 147; Jackson & Andrews 2012).


Nazim 1931: 90.


Id. 93.


Id. 101.


Rehman 1979: 149–150 and 326–328.


See above pp. –.


Nazim 1931: 29, 86–99 and 104–113.


Bosworth 1963: 73.


See for instance Bosworth 1963: 60–61 and 64.


Huart & Massé 2012.


Sachau’s Arabic edition (Taḥqīq [1887]) here offers a better reading of the text than the Hyderabad edition (Taḥqīq [1958]).


Taḥqīq (1958), p. 18.5–10. Sachau 1910: I/24. The last portion of the quotation chiefly consists of idiomatic Arabic expressions. Therefore, I also provide Sachau’s translation of the full passage: “I have found it very hard to work my way into the subject, although I have a great liking for it, in which respect I stand quite alone in my time, and although I do not spare either trouble or money in collecting Sanskrit books from places where I supposed they were likely to be found, and in procuring for myself, even from very remote places, Hindu scholars who understand them and are able to teach me. What scholar, however, has the same favourable opportunities of studying this subject as I have? That would be only the case with one to whom the grace of God accords, what it did not accord to me, a perfectly free disposal of his own doings and goings; for it has never fallen to my lot in my own doings and goings to be perfectly independent, nor to be invested with sufficient power to dispose and to order as I thought best. However, I thank God for that which He has bestowed upon me, and which must be considered as sufficient for the purpose.”


KP, pp. 199.7–200.4; Pines & Gelblum 1989: 272.


Reynolds 1858: 448.


Sachau 1910: I/ix–xvi; Shamsi 1979: 270; Said & Khan 1981: 70–82.


Bosworth 1991: 43.


Bosworth 1963: 127.


Bosworth 2004: 18.


Quoted in Bosworth 1963: 65.


Bosworth 1963: 118.


Auer 2012: 16.


Minorsky 1951: 234–235; Said & Khan 1981: 80–81; Tetley 2009: 65–66.


On intercultural exchanges between Indians and Muslims taking place at the Ghaznavid court, see below Sections and .


Said & Khan 1981: 76–77.


Bosworth 2011e: 376–377.


This was suggested by Said and Khan (1981: 84–86).


Al-Bīrūnī refers here to the animosity of the Indians towards Muslims due to Maḥmūd’s invasions.


Taḥqīq (1958), p. 16.17–18; Sachau 1910: I/22.


Taḥqīq (1958), p. 165.11; Sachau 1910: I/206.


In his Pharmacology (Kitāb al-ṣaydana fī l-ṭibb), compiled at the end of his life in approximately 1050, al-Bīrūnī asserts that he had seen apples in Kashmir (Said 1973: 91, under the entry tuffaḥ, no. 20).


Taḥqīq (1958), pp. 165.19–166.2; Sachau 1910: I/206.


Sachau 1910: I/206–208.


Sachau 1910: I/173–174.


Sachau 1910: I/393, I/116–117 and II/178.


Boilot 1955: 165–166.


Boilot 1955: 199, no. 71.


Boilot 1955 200, no. 72.


Taḥqīq (1958), p. 489.10–15; Sachau 1910: II/181.


The year 951 of Śakakāla corresponds to the year 1029 CE. See the online converter at [accessed October 2023].


Taḥqīq (1958), p. 328.9–10; Sachau 1910: I/391.


Taḥqīq (1958), p. 512.18–19; Sachau 1910: II/208. The name has not yet been identified. Sachau makes some assumptions about this figure, but with little certainty (Sachau 1910: II/400).


Khan 1976: 92, n. 28.


Taḥqīq (1958), p. 125.5–6, 129.3–4, 170.4–5, 347.15–18 and 451.4–5; Sachau 1910: I/161, 165, 211, II/9 and 129. The case of Multan, as a place where al-Bīrūnī may have been, is discussed below p. 77–78.


Sachau 1910: I/173, 199, II/5, 9 and130.


Taḥqīq (1958), pp. 347.20–348.2; Sachau 1910: II/9.


Taḥqīq (1958), pp. 187 and 193; Sachau 1910: I/230 and 235.


Taḥqīq (1958), p. 160.5–6; Sachau 1910: I/201.


Taḥqīq (1958), p. 466.5–6; Sachau 1910: II/148.


See the full excerpt from the Taḥqīq mā li-l-Hind presented above pp. –.


Taḥqīq (1958), pp. 104.14–105.1; Sachau 1910: I/135.


In the Taḥdīd al-amākin, al-Bīrūnī collects pieces of information about distances between cities from travellers’ accounts (Ali 1967: 14). On different sources of information related to the territory of al-Hind described in the Taḥqīq mā li-l-Hind, see Verdon 2015: 43–45.


Taḥqīq (1958), pp. 17.16–18.5, 456.12 and 475.14; Sachau 1910: I/23–24, II/134 and 163. On his interactions with Indian scholars, see further Section 2.3 below.


Taḥqīq (1958), p. 458.2–7; Sachau 1910: II/152–153.


Taḥqīq (1958), pp. 452.5–457.7; Sachau 1910: II/130–135.


See the list of al-Bīrūnī’s literary sources in Sachau 1910: I/xxxix–xl and Shastri 1975. See also Mishra 1985: 35–43.


Although the Viṣṇupurāṇa is quoted at length in the Taḥqīq mā li-l-Hind, al-Bīrūnī admits that he could not read it himself. See Taḥqīq (1958), p. 101.5; Sachau 1910: I/130–131.


Jan Gonda (1951) scrutinizes the way in which al-Bīrūnī transmits information drawn from the Purāṇas, solves several of Sachau’s doubts about the identification of Arabic transliterations of Sanskrit proper names, and states that some of al-Bīrūnī’s readings might be valuable for scholars interested in purāṇic studies.


Taḥqīq (1958), pp. 6.2, 119.8–9, 122.5–6 and 327.2; Sachau 1910: I/8, 154, 158 and 389.


Taḥqīq (1958), pp. 121.6–13, 122.4, 128.17, 198.5, 250.2, 281.19, 304.15, 309.2, 348.6, 388.11, 489.10 and 490.1; Sachau 1910: I/156–157, 164, 240, 298, 334, 361, 367, II/9, 54 and 181–182.


As for other places located in al-Hind and whose latitudes al-Bīrūnī calculated, such as Kindī, Dunpūr, Jhelum and Sialkot, they are not dealt with here because the Taḥqīq mā li-l-Hind contains too little information on them, despite their possible significance for al-Bīrūnī’s encounter with Indians.


See for instance Nazim 1931 and Rehman 1979: 125–167.


Nazim 1931: 28–33.


Reynolds 1858: 38.40. Laghmān, Laghman or Lamghan was situated in eastern Afghanistan, lying on the northern side of the Kābul River; see Rehman 1979: 13.


Id. 469. The dates of the reigns of Late Shahi kings are drawn from Rehman 1979: 89–167. For a comparative table of the different datings proposed in Pandey 1973: 80–114 and Mishra 1972: 9–223, see Khan 2017: 48.


Rehman 1979: 4, n. 17 and 2003: 3–4; See also Reynolds 1858: 327–328. For more references on the Late Shahis, see above fn. in Chapter 1 of the present book.


The Rājataraṅgiṇī describes a battle between Maḥmūd and Trilocanapāla (Majumdar 1957: 67). See also Rehman 1979: 4, n. 18.


Nazim 1931: 86–121.


Al-Bīrūnī describes Wayhind/Udabhāṇḍa as the capital of Gandhāra (Taḥqīq [1958], pp. 165.8–9 and 215.7–216.1; Sachau 1910: I/206 and 259; Qānūn (1955), p. 562; Ğamāhir, p. 236.8; Said 2001: 293), while al-Muqaddasī (ca. 945/46–991) relates that it is a provincial capital (referred to in Bosworth 1970: 254 and Rehman 1979: 17). On Udabhāṇḍa, as an administrative and political centre for the Late Shahis, see Verdon 2021, and for a discussion on the status of Lahore at the time of Jayapāladeva and his son Ānandapāla, see Rehman 1979: 139–141. See also Cunningham 1871: 52–54; Stein 1893: 198–200 and 1900: II/337; Rehman 1979: 4.


Arabic textual sources used the name Qandahār for two different places: a place located in south-eastern Afghanistan and a second one corresponding to Gandhāra in Peshawar region.


Translation by Jamil Ali (1967: 261). See also Taḥdīd [1992], p. 291.21–292.3.


Fischer 1964: 38. See also the brief description in Kuwayama 2002: 225–226.


Fischer 1964: 37–38. Whereas Durgā is honoured by herself, Pārvatī is almost exclusively worshipped as the spouse of Śiva.


See Foucher 1947: 386–387; Pl. 37–38 and Humbach 1986.


Watters 1904: I/181–182; Kuwayama 2002: 204.


Bosworth 1970: 92.


Reynolds 1858: 35–40. See also Pandey 1973: 35.


Taḥqīq (1958), p. 270.9; Sachau 1910: I/317.


Taḥqīq (1958), p. 215.3–4; Sachau 1910: I/259; Rehman 1979: 13.


Taḥqīq (1958), p. 347.12–15; Sachau 1910: II/8–9.


The inhabitants of Laghmān are referred to as lampāka in the Purāṇas (Bhattacharyya 1991: 202).


See above pp. –.


Taḥqīq (1958), p. 285.2–4; Sachau 1910: I/338.


Dey 1927: 162; Bhattacharyya 1991: 256.


Wriggings 2004: 60. See also Kuwayama 2002: 211.


Taḥqīq (1958), p. 215.5–6; Sachau 1910: I/260.


Taḥqīq (1958), p. 270.9; Sachau 1910: I/317.


Taḥqīq (1958), p. 349.8–9; Sachau 1910: II/11. See above p. 18.


Reynolds 1858: 280.


Khan 2017: 57–59.


Dani 1969: 56; Rehman 1979: 16.


Taḥqīq (1958), p. 167.5–7; Sachau 1910: I/208. This passage has already been quoted earlier in this book, see p. .


Quoting Jīvaśarman, al-Bīrūnī reports that the Swāt country is opposite the district of Girī, which may be the same district to which Fort Rājagirī belonged (Taḥqīq [1958], p. 390.1–2; Sachau 1910: II/182).


Taḥqīq (1958), pp. 164.15–165.2; Sachau 1910: I/205. See above pp. 30–31 for a discussion on the location of these two forts.


Nazim 1931: 104–105.


RT VII.47–53, pp. 23–24; Stein 1900: I/270–272. On other relationships between the Kashmiri kings and the Late Shahis, see below p. .


Translation by Ali (1967: 188). See also Taḥdīd [1992], p. 222.10–223.1.


Rehman 1979: 266–267 and 273–274; Meister 1996 and 2010: 32–38; Khan 2017: 70–71.


For further references on the Late Shahis, see above fn. of Chapter 1.


Inaba 2013: 89–90.


Rehman 1979: 33–34. Pandey (1973: 187) is of the same opinion.


See Pandey 1973: 135–137 and Rehman 1978 and 1979: 246–247, 308–318. For a transliteration and translation of this inscription, see Rehman 1979: 310–313. See also Rodziadi Khaw 2015: 119–121.


Rao 1997: II.2/398–406.


Rao 1997: II.1/40.


On Umā’s representation in sculptures see Rao (1997: 122, 124 and 137). Umā, Durgā and Devī can also be regarded as a feminine aspect of Viṣṇu (Rao 1997: 332).


See Rodziadi Khaw 2015 and 2016: 64–114 for a catalogue and description of the Śāradā inscriptions of Gandhāra. See also Rehman 1979: 218 and 241–248.


See for instance Rodziadi Khaw 2015: 97–98.


Rehman 1979: 194–217. See also Thomas 1846, MacDowall 1968 and Alram 2016: 151–153. New research is now conducted by Arturo Annucci on the Late Shahi coinage. See for instance Annucci 2023.


Billon is an alloy of silver and copper (Bhattacharyya 1954: 118).


Cappelletti (2015: 54–55 and 152, fig. 6B) also provides us with a generic description of the first type.


Singh 2017: 166 and 189.


A study by Ijaz Khan (2017) focuses on the political aspect of the Late Shahi society. As an up-to-date study of sites located north of the Valley of Peshawar, it highlights the importance of the defensive activities of the Late Shahi in the region.


Rehman 1979: 281–284; Meister 1996 and 2010.


Filigenzi 2015: 36–38.


Rehman 1979: 285. See also Pandey 1973: 233–236 and Kuwayama 2002: 222–248.


RT VI.178, p. 97; Stein 1900: I/249.


Verdon 2019a: 68–78.


Al-Bīrūnī does not mention Śāradā as one of the scripts used in al-Hind. His silence on this type of script confirms Slaje’s remark that this term was not used before the eleventh century (Slaje 1993: 15–16.). Observations by Rehman on the Śāradā script lead to the same conclusion (1979: 237–241). Al-Bīrūnī explains that the script siddhamātṛkā (siddamātrika; ‮سدّ ماترك‬‎) was in use in the regions between Kashmir and Kanauj (Taḥqīq [1958], p. 135.3–16; Sachau 1910: I/173). For al-Bīrūnī, Śāradā is the name of a Kashmiri idol (Taḥqīq [1958], p. 89.12–13; Sachau 1910: I/117), most probably referring to the Śāradā pīṭha, i.e., the ancient Hindu temple whose ruins are located to the east of the Kashmir Valley.

  • Collapse
  • Expand