Routes of Translation: Connected Book Histories and al-Jazari’s Robotic Wonders from the Mamluks to Mandu

Abstract

Over the course of the long fifteenth century, scholars and books moved across regions and spurred transcreations of numerous Islamicate manuscripts in South Asia. This essay undertakes a close reading of an early sixteenth-century Persian transcreation—that is, a translation in both form and content—of a twelfth-century Arabic compendium on mechanical devices. I examine what the historical event of translation in the South Asian region of Malwa and town of Mandu meant as it is read amidst cultural flows between Mamluk Egypt, Yemen, Mecca, and Hindustan. By analyzing the colophons of early Arabic copies of al-Jazari’s Compendium of Theory and Useful Practice for the Fabrication of Machines along with Da’ud Shadiyabadi’s Wonders of Mechanics in Persian, this study demonstrates how Shadiyabadi’s translation distances itself from al-Jazari’s book. As Shadiyabadi’s Wonders of Mechanics becomes the standard Persian translation of al-Jazari—appearing in subsequent Mughal and Iranian copies—the work of a scholar from a small North-Central Indian court links scholars, sultanates, and regions.

Keywords:

• Translation
• manuscript studies
• transregional
• Malwa
• connected histories
• al-Jazari
• automata

Introduction: Transregional, Translation, Transcreation

The Wonders of Mechanics (‘Ajā’ib al-Ṣanā‘ī) in Persian by Dā’ūd Shādiyābādī was not the first text to transmit knowledge of automata in the North-Central Indian region of Malwa. In the early eleventh century, the Paramāra king Bhoja (ca. 1000–1055) described many kinds of mechanical devices in Sanskrit in his Stories for the Bouquet of Love (Śṛṅgāramañjārīkathā).¹ Bhoja’s capital was in Dhār, the original seat of the Malwa sultanate before its capital moved to Mandu.² Bhoja’s Stories for the Bouquet of Love included automata (yantras) and water clocks apparently adapted from Greek sources.³ Some 150 years later, around the same time of the emergence of a sultanate in Delhi, a scholar thousands of miles away in Jazira (Mesopotamia) named Ibn al-Razzāz al-Jazarī composed the Compendium of Theory and Useful Practice for the Fabrication of Machines in Arabic (al-Jāmi‘ bayn al-‘ilm w’al-‘amal al-nāfi‘ fī ṣinā‘at al-ḥiyal, henceforth referred to as the Compendium).⁴ As a standard for scientific knowledge, this Arabic work was steadily copied for several centuries, which enabled its translation in sixteenth-century Malwa. In Shadiyabadi’s Mandu, al-Jazari arrived in the land of King Bhoja.

The first known copy of the Wonders of Mechanics is dated 1508 (British Library Or 13718). Shadiyabadi made this translation for his patron Sultan Nasir Shah Khalji (r. 1500–1510) of Malwa. For historians, the sultanate of Malwa, established by Dilavar Khan Ghuri in 1401 has held an esteemed position in the list of so-called regional sultanates that flourished between the fourteenth and sixteenth centuries. The reigns of two sultans—Hushang Shah (r. 1406–1435) and Ghiyas al-Din (r. 1469–1500)—have especially commanded the attention of scholarship for both their military and cultural achievements, while the reign of Sultan Nasir Shah has been relatively neglected.

In contrast to political histories that understand the ambitions and activities of these rulers as expressions of regional power, this essay argues that from the perspective of intellectual and book histories that the court culture of Mandu, even as it was distinctive, saw itself as part of a broader transregional network, one that straddled the vast spaces of the Islamicate world. I demonstrate this connection by a method I call “connected book histories,” extending Sanjay Subrahmanyam’s “connected histories” to the field of manuscript studies.⁵ Reflecting on the impact of this method on the “early modern in South Asia,” Rosalind O’Hanlon states that “the result [of the connected histories approach] has been a suggestive new framework for understanding mobility and transcultural possibilities that characterized early modern Eurasia’s royal courts.”⁶ This essay traces the movement of people, books, and manuscripts reflected in Shadiyabadi’s transcreation, the Wonders of Mechanics. In doing so, I demonstrate that a close analysis of the processes of transcreation reveals how Shadiyabadi sought to make al-Jazari legible to his immediate region of Malwa, all the while participating in a wider circulation of manuscripts.

Even though a ruler commissioned Shadiyabadi’s translation, its earliest copy is much humbler in scale and finish than some of the other royal commissions at Mandu, such as the famed Book of Delights or Ni‘matnāmah. Other imperial objects such as the Mughal album with its inclusion and responses to European prints and Persian calligraphy is a natural candidate for a connected book histories approach.⁷ While courtly albums exemplify some of the fanciest objects and the one percent of bookmaking, a work like the Wonders of Mechanics draws a different picture. It gives us a sense of bookish scholars and manuscript makers in collaboration, albeit with court patronage lurking in the background.

I demonstrate how connected book histories fruitfully illuminates the development of a transregional scholarly culture of Islamicate sciences in South Asia through a micro-level analysis of colophons, diagrams, images, notations, and language. Unlike books that may have been committed to memory and transferred through orality such as poetry or the Qur’an, the corpus of books under consideration here illustrates how one book led directly to another in this period. Scribes of al-Jazari manuscripts often insisted that the book was transferred from the hand of the author, not only in its text but also in its diagrams. This kept the author alive for them through a direct connection spanning long distances and centuries.

Previous scholarship has contributed to the study of connected book histories of the contexts studied here and reminds us of points of disconnection and reinterpretation. Elizabeth Lambourn has examined the material forms of writing and containment of a late thirteenth-century illegible letter from a Sri Lankan ruler to the Mamluk court of sultan al-Mansur al-Qalawun (r. 1280–90) in Cairo.⁸ Lambourn urges us to consider whether a medium of long-distance communication would be legible as a letter in the first place, and the probability of oral transmission when a text does not transfer the message. The Sri Lankan letter also serves as a salient example of disconnection despite the centuries of diplomatic exchange across the Indian Ocean and Red Sea. Regarding manuscript evidence, Christopher Bahl has mapped the transmission of Arabic texts, especially grammars, across the Western Indian Ocean (ca. 1400–1700).⁹ Bahl shows that the text need not be translated as Shadiyabadi translates al-Jazari for it to gain entirely new meanings in a South Asian context.

The preponderance of original works linked with the patronage of the Malwa sultanate—the Wonders of Mechanics, a multilingual illustrated dictionary, and the Ni‘matnāmah—suggests how this small regional, but well-networked court, shaped its world through multilingual books. The book culture of Malwa and its capital Mandu are not necessarily unique for the long fifteenth century. Rather, their survival enables us to reimagine the world of Shadiyabadi, as he was just one of many of his kind.

Shadiyabadi’s Wonders of Mechanics gives us a sense of his care for the subtleties of language and how he deployed it as a translator. His scholarly propensities provide insights into what Francesca Orsini has called a “local cosmopolitan”: a multilingual intellectual who mediates a larger cosmopolitan world for his local context. His works strategically incorporated new communities in Shadiyabad (the “City of Joy,” or Mandu) into adaptive knowledge systems. I have elsewhere analyzed his Key of the Learned (Miftāḥ al-Fużalā’), a multilingual Persian dictionary written in 1468, that survives in the form of an illustrated manuscript attributable to the 1490s.¹⁰ The Key of the Learned’s use of Hindavi words, often as clarifying synonyms for Persian entries, reveals an ease with moving between Hindavi and Persian. Here, I call attention to how Shadiyabadi sprinkled occasional Hindavi words in his Persian to gloss terms for local readers.

It is also worth noting that a close look at al-Jazari’s original Arabic work reveals him to be a local cosmopolitan as well. Al-Jazari freely calques Persian words into Arabic and marshals a range of Greek and Arabic sources to produce his compendium.¹¹ His Persian within Arabic speaks of the diversity of the intellectual milieu of twelfth-century Jazira.

This essay maps the circuitous routes of Shadiyabadi’s translation, the Wonders of Mechanics, at its moment of transcreation—that is, a translation in both form and content—from Arabic into Persian at the beginning of the sixteenth century.¹² Through an evaluation of this book’s transcreation in Hindustan, I examine its implications for connected book histories during the long fifteenth century, from the moment of Timur’s invasions of Delhi in 1398 until the rise of the Mughal dynasty in the mid-sixteenth century.¹³ I then situate the transmission of this work against the historical evidence of transregional currents that connected the Mamluk and Malwa sultanates, if the Wonders of Mechanics attests to a much earlier Arabic source, rather than another, more proximate intermediary.

Locating the Wonders of Mechanics in the Manuscript Tradition of al-Jazari’s Compendium

The British Library’s manuscript of the Wonders of Mechanics contains 191 folios with 159 technical diagrams of clocks, drinking vessels, liquid containers, fountains, and machines for raising water. It is modestly sized, measuring 22 × 14.5 cm, with roughly 21 lines of naskh text per folio. It entered the British Library’s collection in 1976 purchased via sale at Sotheby’s and was first published in J.P. Losty’s 1982 catalogue of the British Library’s Indian manuscripts.¹⁴ Its pages were pruned when it was rebound at His Majesty’s Stationary Office Bindery in 1978, leaving several diagrams trimmed. Because of the rough surface of its paper many drawings have rubbed onto opposite sides of the page. The diagrams of the manuscript reinterpret Artuqid and Mamluk models, although they often differ in scale and proportions (Figures 1 and 2).

Figure 1. Main drawing of figures and water mechanisms; from the category (naw‘) of the Arbiters of Drinking Sessions, Compendium of Theory and Useful Practice for the Fabrication of Machines of al-Jazari, Probably Mamluk Syria, 1315, Folio: 30.9 x 20.4, National Museum of Asian Art, Smithsonian Institution, Freer Collection, Purchase—Charles Lang Freer Endowment, F1930.77.

Figure 2. Main drawing of figures and water mechanisms; from the category (naw‘) of the Arbiters of Drinking Sessions Wonders of Mechanics of Shadiyabadi, Mandu, 1509, Folio: 23.2 × 15.7 cm, British Library Or 13718, f. 120b. © British Library Board.

The British Library’s copy of the Wonders of Mechanics is the only known copy likely contemporary with Shadiyabadi. A scribe copied the codex of British Library Or 13718 in November/December 1508 (914 A.H.) and it was authenticated on 26 January 1509 (915 A.H.).¹⁵ The verification note, possibly by Shadiyabadi himself, appears adjacent to the manuscript’s colophon.¹⁶ I have identified five later undated copies, which suggest that the Wonders of Mechanics became one of the standard Persian translations of al-Jazari’s Compendium.¹⁷ Even though it is listed in most surveys of sultanate manuscripts, the Wonders of Mechanics has received the least attention of Mandu’s books to date.¹⁸ Scholars of al-Jazari’s Compendium have overlooked it, although they cite other, later Persian translations.¹⁹

The devices described in Wonders of Mechanics were probably never realized in Mandu, but that is not to say that its knowledge did not help Mandu’s craftsmen understand how to harness mechanical skills. Al-Jazari adapted his text from the Abbasid Banū Mūsā who relied on ancient Greek sources such as the treatise on automata by Philo of Byzantium of the third century B.C.²⁰ A clock similar to that described by al-Jazari was possibly made in the twelfth century at the Eastern entrance of the Great Mosque of Damascus.²¹ This clock and an earlier tenth-century one, neither of which are extant, both interested Mamluk historians, which may suggest some of the Mamluks’ motivations for commissioning so many copies of al-Jazari’s work.²² My study of the early Arabic illustrated manuscripts of al-Jazari and of the Wonders of Mechanics thus negotiates the transmission of classical knowledge of automata in a world where its physical manifestations are far more challenging to trace.²³

Popular knowledge of automata also circulated widely within Persian wonders-of-creation (‘ajā’ib al-makhlūqāt) books by Zakariyya’ al-Qazwini (d. 1283). The earliest known Persian copy of al-Qazwini dates to 1300 and contains descriptions of crafts and automata.²⁴ Only in the fifteenth century did artists start illustrating the sections on the automata in these manuscripts.²⁵ Such compendia of wonders too circulated directly between Iran and Hindustan, and Mandu received such books.²⁶

The traffic between Iran and Hindustan is the cause we would most likely suspect for how the Wonders of Mechanics surfaced. The historical evidence for Iran and sultanate relations is vast. But, given the visual qualities of the Wonders of Mechanics, al-Jazari’s circulation, and the scattered Mamluk-Mandu connections, we cannot rule out direct cultural flows from the Mamluk world. It is improbable that the single copy of established Artuqid provenance was the source of the Wonders of Mechanics, when there are far more copies of Mamluk origins.²⁷

The earliest manuscript of al-Jazari’s Compendium was composed in 1206 shortly after the work was written in Diyār Bakr/Āmid under the patronage of the Artuqid sultan Nāṣir al-Dīn (r. 1200–39).²⁸ Thereafter it was copied steadily in Mamluk domains²⁹ until its first extant Indian version and translation, namely the Wonders of Mechanics.

The lasting interest in al-Jazari’s book, despite its age, is attested by the fact that Nasir al-Din Khalji (r. 1500–10) commanded Shadiyabadi to produce a translation of the Banu Musa’s Kitāb al-Ḥiyal, but Shadiyabadi chose al-Jazari’s work instead.³⁰ His decision to produce a Persianised Compendium may relate to its availability and the Kitāb al-Ḥiyal’s absence. It also may express Shadiyabadi’s high regard for al-Jazari as a scholar.

Concurrently, at the Bahmani sultanate just south of Malwa in the Deccan, Maḥmūd Shāh II Bahmanī (r. 1482–1512) commanded his court secretary Nīmdihī to translate the Kitāb al-Ḥiyal as well.³¹ One wonders whether Nimdihi too chose al-Jazari instead of the Banu Musa, or al-Jazari and the Banu Musa’s works were simply thought of as the same thing. To my knowledge, Nimdihi’s translation does not survive leaving these questions unanswerable. Nevertheless, Persian translations of Arabic works on automata appear as parallel cultural impulses in at least two regions of sultanate India, the Deccan and Malwa.

Around the turn of the sixteenth century, two other key manuscripts of al-Jazari’s Compendium were copied, dated 1485 and 1486, the latter of which is the primary source for Donald Hill’s English translation.³² These two copies shed light on how al-Jazari’s work was treated outside of India at a time roughly contemporary with the Wonders of Mechanics. After the Wonders of Mechanics, there remain nine extant al-Jazari-related manuscripts produced in India, three of which were al-Jazari’s work itself.³³ The circumstances under which Shadiyabadi accessed a manuscript of al-Jazari’s Arabic text to produce his work—whether he travelled or read the text in India itself—is a matter that can only entertain informed speculation. As I propose here, long-distance circulation and a network of madrasahs between Mandu, Egypt, and Mecca, are possible conduits by which Shadiyabadi came across al-Jazari’s Compendium.

The extant evidence of the al-Jazari manuscript tradition reveals the desire to authenticate the manuscript’s lineage (silsilah) tracing back to the original author’s hand. Such overt allusions to transfer found in colophons provide granular evidence of connected book histories. An al-Jazari manuscript dated 1485, now housed in the Bibliothèque nationale de France, illustrates such concerns. Its colophon states,

This book was completed, blessings for its owner, at sunset on Friday, 4 April 1485. The copy from which I transferred its diagrams was completed 20 July 1341. That manuscript was transmitted from the best copy, also from which I transmitted this copy. Those that are like it, and this copy, derive their text from the copy that was in the hand of the author. The letters, variants, and drawings of the devices are as he drew them with a ruler and freehand, blessings to the greatest God. That copy was completed on 16 January 1206.³⁴

At first glance, this colophon implicates a long practice of copying al-Jazari’s Compendium, because the scribe writes, “this copy and those that are like it/al-nuskhah mā mithālahi.” He identifies two key manuscripts: first, the copy from which he based his work upon created in 1341, and second, the original, “mother,” manuscript made in 1206. From the 1206 original manuscript to the 1341 copy, to the BnF copy made in 1485, the scribe inscribes his copy within a good genealogy tracing back to the hand of the original author and scribe.³⁵

That every detail of the original manuscript—its “letters, variants, and drawings of the devices/al-ḥurūf wa’l-ibdāl wa rasūm ṣūrat al-ashkāl”—are as al-Jazari drew them merits our attention for understanding processes of transmission.³⁶ And yet, the finish of the manuscripts in the al-Jazari corpus varies drastically. The 1485 copy, the colophon from which I quote above, for instance, contains wholly unfinished diagrams. Some spaces where the diagrams should occur are left blank.³⁷ In other cases, the image is completely pared down to the constitutive outlines, lacking any details or figures (Figure 3).³⁸ The image that corresponds to a main drawing of a water clock, typically one of the liveliest images in the entirety of early manuscripts of al-Jazari’s Compendium, appears plainly as the architectural frames of the alcove and reception chamber in which musicians would have been drawn. The drawing’s tentative red horizontal drafting lines, matched with those of the gridlines (misṭarah) from the opposite facing folio, suggest that this composition may have been transferred from another copy using a pounce. The spacing of the lines in these frames corresponds to the gridlines of the text page on the verso side of the folio. The artist has painted ochre and blue to demarcate the structures of certain architectural features such as the dome and pillars. The bare spaces are where figures would have been added at a later stage. Despite the unfinished state of these diagrams, the colophon still claims that the manuscript is finished, perhaps only from the perspective of its text. This suggests that the manuscript’s scribe and artist were probably different individuals, and that the processes of writing and illustrating occurred in different phases.

Figure 3. Incomplete drawing of zodiac clock, Compendium of Theory and Useful Practice for the Fabrication of Machines of al-Jazari, Dated 1485, Folio: 27 × 18.3 cm, Bibliothèque nationale de France, Arabe 2477, f. 8b.

Ample physical evidence of transfer mechanisms in the al-Jazari corpus signifies an intent to adhere to a prescribed programme of diagrams within these manuscripts. Two largely incomplete books on automata that postdate the Wonders of Mechanics by several centuries bear such evidence of copying. For example, we find tracings of diagrams in an eighteenth-century edited selection of different treatises on automata including al-Jazari’s.³⁹ Comparing the tracing with the diagram of an image of a vessel we can deduce that these are copies of each other, however it is not certain which came first.⁴⁰ A later Iranian Persian translation dated 1874 contains tracing papers with French color codes.⁴¹ This manuscript has several blank spaces where illustrations would have been appropriate. A French copyist has inserted tracings from an earlier Arabic manuscript and added these tracing papers within the manuscript. Taken together, these much later copies and the 1485 manuscript attest to an effort to preserve an “original” programme of the diagrams, which may have much to do with establishing the credibility of a manuscript’s transmission.

Regarding the text, the Wonders of Mechanics makes the translator’s agency clear, and its composition marks a rupture in the transmission of the Arabic Compendium. Despite its close adherence to al-Jazari’s work, Shadiyabadi’s act of translation distances it from the Arabic text—connection begets disconnection. In his introduction, Shadiyabadi identifies himself as the author, and his patron as Sultan Nasir Shah Khalji of Malwa, one of the patrons of Mandu’s Ni‘matnāmah.⁴² Shadiyabadi’s short introduction contains preface identifies him as the author of the text and explains the six larger typologies (anwā‘) of mechanical arts within which numerous smaller sub-sections of forms (ashkāl) and sub-sub-sections (afṣāl) are nested. While the frontispiece of BL Or 13718 is modern and the later scribe wrote a new text for that page, Shadiyabadi’s original prefatory text can be recovered from the manuscript’s later Mughal copy.⁴³ The similarity of the diction of the Mughal copy’s first few lines and that of Mandu’s Key of the Learned’s gives further reason to this claim.⁴⁴ After its colophon, the text also bears a standard table of the letters used to label parts of devices within the diagrams, and a glossary of technical terms with vernacular equivalents, which is not in al-Jazari’s Compendium.⁴⁵

For the most part, Shadiyabadi adheres to al-Jazari’s book as he calques Arabic technical vocabulary into Persian with ease. Like al-Jazari, he instructs his readers as to how the automata are made in his own first-person voice. Al-Jazari, for instance, describes his experiments with some of his devices. When devising the very first water clock in the Compendium he states, “I made an instrument according to this pattern, but it did not work correctly./fa‘amaltu ālah ‘alā hādhahi al-ṣūrah falam yasaḥḥa bihā ‘l-‘amal.”⁴⁶ Shadiyabadi mimics this in Persian: “I made this instrument in the described format, but it did not work correctly/bas sākhtam ‘amal-i ālat barīn ṣūrat kih namūdah shud ammā badān ‘amal ṣaḥīḥ nashud.”⁴⁷ In comparison to the Compendium dated 1486, Shadiyabadi elides only one full section towards the end of al-Jazari’s text, and appends two devices not included in the Compendium after the Wonders of Mechanics concludes.⁴⁸ Overall, it seems as if Shadiyabadi diligently translates al-Jazari into Persian although he never mentions him by name.⁴⁹

Regarding the British Library’s manuscript of the Wonders of Mechanics, we can safely state that Shadiyabadi, another scholar, or a scribe verified it the year after it was copied in 1508. Throughout the book, several glosses and corrections appear in the margins and on the text itself. These textual edits occur more frequently towards the beginning of the manuscript than at the end. The similarity in the handwriting between the word ṣaḥīḥ (correct) in the verification note and in a note where the scholar corrects his own redaction, or one of a previous editor’s, stating, “it is correct, it is not incorrect/ṣaḥīḥ ast ghalaṭ nīst,” suggests that the same individual responsible for authenticating the text also attended to its body text.⁵⁰ Whether this individual was Shadiyabadi himself is unknown, but the dating of the note and corrections reasonably fits within his lifespan.

Akin to Shadiyabadi’s rendering of al-Jazari in text, the Wonders of Mechanics follows the technical drawings of al-Jazari’s Compendium closely. When compared with the 1486 copy of al-Jazari, the artist completed all but two corresponding drawings.⁵¹ In both instances, empty space is allocated for the illustration that may raise the possibility that the scribe and artist were different individuals, or that writing and drawing occurred in successive phases. The finished illustrations are executed in red and brown ink with fillings of ochre, green, blue, pink, brown and red watercolors (Figure 2). Except for the modern frontispiece, there is no illumination in the manuscript.⁵² In comparison to the 1486 al-Jazari manuscript, figures and animals are frequently omitted in the Wonders of Mechanics and even further redacted in the later Mughal version. In such diagrams—made for the expressed purpose of communicating technical knowledge—the non-mechanical figures would have likely been added during the final stage of the drawing process.⁵³

Throughout the Compendium, al-Jazari frequently comments on the purpose of his diagrams. At the end of his description of a technical mechanism operating one of his water clocks that includes representations of musicians he writes, “no picture is needed for the trumpeter, since he makes no movement, and the sound of the trumpet comes from another source.”⁵⁴ The impetus to clarify how the device (ḥiyal) works motivates al-Jazari to provide images (ṣuwwar, mithāl). Even if al-Jazari did not make the technical drawings himself, this rationale provides insight into why an image would be extraneous. Hence, the diagrams are instrumental for transmitting functional information that text does not necessarily convey.

The style of the diagrams in the Wonders of Mechanics strikes one as the product of a sultanate artist trained in a Persian practice who was looking at a much earlier Arabic manuscript of al-Jazari, possibly of Mamluk provenance, for inspiration. Let us take some of the basic elements of the Key of the Learned’s paintings as comparisons to the Wonders of Mechanics’ diagrams. First, consider two corresponding representations of musicians. In Shadiyabadi’s illustrated definition of the stringed instrument (chang) from the Key of the Learned we see two musicians, one holding a drum and the other the chang (Figure 4).⁵⁵ They are entertaining a courtier who offers them some wine. The figures sit kneeling and are shown in three-quarter profile view. They wear lightly adorned flowing robes (jāmahs) that cross at the neck, and are painted teal, indigo, and red. The artist carefully depicts the individual white folds of their turbans with a shaft at the centre.⁵⁶ Another band appears in the Wonders of Mechanics’ illustration of an automatic boat situated on a pool for a drinking party (Figure 5).⁵⁷ Here, the figures are not human, but are life-like representations attached to a large-scale ship. Elements familiar from Key of the Learned appear. Figures are shown in three-quarter profiles and the folds of their turbans part in the middle. The tambourine and chang players hold their instruments in similar ways. Unlike in the Key of the Learned these figures sit cross-legged and not on bended knees. Other illustrations from the Wonders of Mechanics show figures in a similar position.⁵⁸ The artists repeatedly, albeit subtly, evoke the draping fabric adorning figures from earlier Arabic manuscripts, a feature not nearly as pronounced in the line-drawings of a copy of al-Jazari made a few decades earlier.⁵⁹ The shapes of figures’ faces differ, which may be the result of the artist’s desire to evoke the Mamluk elements of an earlier manuscript. Or, this could simply be the result of the different mediums—one being a more finished painting and the other being diagrammatic.

Figure 4. Stringed instrument (chang), Key of the Learned of Shadiyabadi, Mandu, ca. 1490, Folio: 33 × 25.4 cm; Painted surface: 8.1 × 11.9 cm, British Library Or 3299, f. 98b. © British Library Board.

Figure 5. Main drawing of the boat (kushtī), Wonders of Mechanics of Shadiyabadi, Mandu, 1509, 23.2 × 15.7 cm, British Library Or 13718, f. 104b. © British Library Board.

If the artist’s instincts tended towards a Persian style, the diagrams also suggest that he emulated an earlier Mamluk manuscript. Pinning down the exact Mamluk source is impossible here—whether it is as early as 1315, or much later—but, many visual elements point to a fourteenth-century source. In general, the manuscript’s pages lack colored grounds and the bare paper serves as the background for the diagrams. This emphasis on the illustrations’ diagrammatic elements recalls the overall appearance of an earlier, Mamluk model. Because there are no extant al-Jazari scientific manuscripts in India prior to Mandu’s Wonders of Mechanics and the only known Artuqid copy is the earliest one of 1206, a fourteenth-century Mamluk inspiration, if not an unknown Persian intermediary, would be the likeliest of all. One difficulty in visualizing the similarities between a Mamluk Compendium and Mandu’s Wonders of Mechanics is the incongruent scales of these manuscripts as the latter measures roughly half the size of the earlier models.

Other subtle details demonstrate how the artist, scholar, or scribe responsible for the Wonders of Mechanics attempted to transcreate a Mamluk manuscript. Take the concordant representations of the figures from a drinking session in a manuscript made in 1315, probably in Mamluk Syria (Figures 1 and 2).⁶⁰ The mechanical elements of the diagrams, particularly those inside of the dome, align well, although there is some variation in the coded letter markings. One intriguing detail is how the Mamluk artist has handled the pink robe with flows that allude to draping.⁶¹ The drapes are formed with U-shaped flares in a darker pink with fillings of a lighter tone. This does not appear on the ochre and blue textiles of the sultanate drinkers’ robes in the Mandu manuscript, which seem to reference animal skins. Rather, similar folds appear on another figure in the Wonders of Mechanics, namely, one that functions as a drink dispenser (Figure 6).⁶² The folds here are black outlines highlighted with green following the contours of the curving robe of the figure. These folds are not the same as the bursts of the Mamluk figure’s textile, but their particularity to this reworking of an al-Jazari manuscript gives us an idea of the Wonders of Mechanics’ earlier source. In addition, similar flares of the figure’s headdress appear on other figural images in the Wonders of Mechanics and recall the headdresses of earlier Arab painting (Figure 7).⁶³ These flares also appear in earlier Qur’anic ornament, notably, they occur in the borders of early fifteenth-century Indian Qur’ans that are inspired by exchange with fourteenth-century Iranian and Mamluk manuscript cultures.⁶⁴ It is also worth noting that the corresponding illustration of the drinker in a fifteenth-century manuscript made a few decades before Shadiyabadi’s work lacks these folds and flares and comes across plainly depicted.⁶⁵ If the artist of Mandu’s Wonders of Mechanics was looking at a later Iranian manuscript instead of a Mamluk one, then perhaps they too would have opted for a much plainer decorative programme as in the copies made in 1485 and 1486.

Figure 6. Main drawing of figure Wonders of Mechanics of Shadiyabadi Mandu, 1509, British Library Or 13718, f. 116a. © British Library Board.

Figure 7. Semi-sectionalized drawing of figure, showing mechanisms, Wonders of Mechanics of Shadiyabadi, Mandu, 1509, British Library Or 13718, f. 114a. © British Library Board.

All this evidence suggests that within the corpus of the al-Jazari tradition, the Wonders of Mechanics exists on the border between South Asia and the world of Mamluk manuscripts. Where it adheres to al-Jazari’s model, it implies Shadiyabadi’s deep understanding of the original Arabic text. Where it departs from the al-Jazari tradition, it becomes adapted within the interregional world of the Hindustani sultanates and throws light on the innovations occurring in the multilingual local of Mandu itself.

A Translator’s Local Ambitions

Shadiyabadi’s translation negotiates frontiers between India, the Persian cosmopolis, and the Islamicate world. Travis Zadeh’s study of Abbasid period geographies elucidates how the act of translation (tarjumah) both established and dissolved frontiers.⁶⁶ In the Wonders of Mechanics, the boundaries between the larger Islamicate world and South Asia became more porous. At the same time, bringing al-Jazari into a new episteme—constituted both by language (Persian) and place (South Asia)—further separated these two worlds. While Persian language practices, behaviors, and material cultures constituted a Persian cosmopolis, I show here how the Persian of the Wonders of Mechanics illustrates a concern for the local intellectual culture of Mandu. The introduction of the text does not name Shadiyabadi as the “translator” (tarjumān, mutarjim, nāqil) of al-Jazari. Rather, it uses sartorial metaphors common for characterising the process of translation: “Nasir Shah ordered Shadiyabadi to outfit the Kitāb al-Ḥiyal of the Banu Musa, which is a bride covered in dazzling and fresh clothing, in the dress of the fairies of Persian.”⁶⁷

In a wider focus, one must bear in mind the strong culture of translation in the Wonders of Mechanics’ world. The Persian translation movement of South Asia sponsored in large part by the Islamicate courts beginning in the twelfth century continued steadily for several centuries. Scholars have emphasised the translation movement from Sanskrit and Indic knowledge systems into Persian and how Sanskrit writers also adapted elements from their sultanate milieu.⁶⁸ However, the many instances of translation from Arabic and Turkic into Persian affirm that the Hindustani sultanates were not only interested in Indic knowledge. All knowledge preoccupied them.⁶⁹ Like al-Jazari’s Compendium, al-Qazwini’s Wonders of Creation in Arabic was also adapted into Persian, a few decades later in 1547 Bijapur.⁷⁰ The classics of Islamicate sciences in Arabic thus were brought into a Persian episteme of Hindustan.⁷¹ With regard to sultanate material culture, Finbarr Barry Flood has explored how objects such as ivories, metalwork and coins, and architecture underwent processes of transcreation in South Asia from Islamicate or Indic models to Indo-Islamicate works.⁷² Whereas Flood uses the term “translation” for the objects and monuments as a metaphor, manuscript culture experienced both a transcreation in material terms and a true translation in the linguistic sense in Hindustan. It is thus not only one single element—text or manuscript form—that Shadiyabadi and Mandu’s book atelier translated from one form to another. Rather, these manuscripts were transcreated as wholes.

Although we may never know what Shadiyabadi’s source manuscript(s) looked like, his translation is not a standard, word-for-word Persian translation. Rather, it shows a marked ambition for local comprehension. The Wonders of Mechanics’ appendix of technical terms is the work’s most obvious effort for vernacular didacticism. Unlike the prefatory glossary of Shāhnāmah of 1438 attributed to the Deccan city of Bidar, the two-folio-long glossary of the Wonders of Mechanics contains Indic vocabulary.⁷³ Indic vocabulary does not appear as lemmas, but as the synonyms that are equivalent to the Persian glossary entry. For example, in his definition of the sabādah, or whetstone, he writes, “and the people of Hind call it a kahrasān.”⁷⁴ Cross-referencing Shadiyabadi’s own lexicon, the Key of the Learned defines the whetstone, or fasān, with a Persian definition and the phrase, “wheel of the whetstone that in Hindavi is called the kahrasān/charkh-i sabādah kih hindavī kahrsan gūyand.”⁷⁵ The Indic vernacular word kahrasān here entered a sultanate context.⁷⁶ That it appears in both Shadiyabadi’s Key of the Learned and Wonders of Mechanics suggests its ubiquity to his personal lexicon. It was a vernacular word in two senses—first, it was Hindavi, and second, it was a word associated with everyday life. Only a few years after he wrote these works the sultan of Malwa started issuing public decrees in Hindavi.⁷⁷

Let us dwell on one popular vernacular term found in the Wonders of Mechanics that prompts further visual and philological analysis.⁷⁸ One of architectural significance is the word for a protruding or slanting eave known as the chajja. Chajjas are popular in sultanate architecture and their brackets are often where Islamicate building types (mosque, tomb) commonly adapt from Indic ornament.⁷⁹ In his chapter on the elephant water clock, al-Jazari takes pains to detail every component of this device. The elephant’s carriage has a balcony upon it, which Shadiyabadi translates into Persian as such, “Chapter 12, On the Functioning of the Balcony … that Indians call the chajah, and a man sits upon it./faṣl-i davāzdiham dar bayān-i kayfiyyat-i ‘amal-i rawshan … kih ānrā ahl-i hind chajah gūyand va mardī bar ān rawshan nashistih bāshad.”⁸⁰ Here, and at least once more, Shadiyabadi equates the term rawshan (balcony) with chajja.⁸¹ The Wonders of Mechanics’ illustration of the balcony depicts a figure facing forward, seated upon a throne framed by red and yellow cusped arches topped by a dome (Figure 8).⁸² Below him a network of curved lines and arches form ornamental brackets that support the balcony. The corresponding diagram from a Mamluk Compendium dated 1354 displays a simply delineated red frame that may be unfinished (Figure 9).⁸³ Within the frame a figure wearing a stylised red robe sits beneath a dome. Considering the simplicity of this architectural frame, the only notable features are the two support brackets formed of concave and convex cusps. The visual focus on these brackets highlights how the rawshan or chajah protrudes from a building.

Figure 8. Figure on balcony, and semicircle of roundels Wonders of Mechanics of Shadiyabadi, Mandu, 1509, Folio: 23.2 × 15.7 cm, British Library Or 13718, f. 60b. © British Library Board.

Figure 9. Figure on balcony, and semicircle of roundels, Compendium of Theory and Useful Practice for the Fabrication of Machines of al-Jazari, Mamluk, 1354, Folio: 40 x 27.7 cm, National Museum of Asian Art, Smithsonian Institution, Freer Collection, Purchase—Charles Lang Freer Endowment, F1932.19.

Given its ubiquity in architecture and occurrence in the Wonders of Mechanics a fuller philology of chajja is in order. A. Y. Hassan defines rawshan as a balcony in his glossary of technical terms in al-Jazari’s Compendium.⁸⁴ Although it appears in al-Jazari’s Arabic text, the word rawshan carries a Persian etymology and can also mean “window” or more commonly “light.”⁸⁵ It marks one of the many Persian words that the “local cosmopolitan” al-Jazari calques into Arabic. Likewise, chajja in common Hindi parlance means “balcony” as well. It originates from the Sanskrit word chādya that means “eave” and is a feature of western Indian temple architecture from the end of the tenth century and into the eleventh century.⁸⁶ Chādya is related to the root chad that means to cover over something.⁸⁷ A Classical Hindi dictionary of the vernaculars of Avadhi and Brajbhasha contains the lemma chajjā with the definition, “the part of the roof’s wall that protrudes,” suggesting that the word was possibly used in early modern Hindavi.⁸⁸ Scholars of sultanate and Mughal architecture have also taken note of the chajja attesting to its ubiquity for these buildings.⁸⁹ Ebba Koch, for instance, provides the definition of chhajjā as a “sloping or horizontal projection from the top of a wall supported by brackets to protect from rain or sun.”⁹⁰ If Koch bases this definition on Mughal Persian sources, as she states, this would continue the term’s genealogy from the sultanate Wonders of Mechanics, further underscoring continuity in the artificial rupture of the Mughals.

That Shadiyabadi clarifies the Persian word rawshan with chajja relates to his aspirations for his text to be understood locally. Surely, the built environment Shadiyabadi inhabited was full of chajjas that displayed Indic ornament. The tomb of Sultan Hushang in Mandu finished around 1439 has a wide chajja with curving brackets (Figure 10).⁹¹ The brackets show lotus designs and geometric ornament common to temple architecture of late medieval Western India.

Figure 10. Tomb of Sultan Hushang, Mandu, with marble chajja and brackets, Completed ca. 1439, Photograph: Vivek Gupta, August 2015.

Based on these examples, traces of the vernacular in the Wonders of Mechanics convey the widespread use of Indic vocabulary for sultanate material culture and Shadiyabadi’s local sensibilities.⁹² His use of such vernacular words may have anticipated an immediate interregional circulation of his work for Persian readers within South Asia, and eventually beyond. And yet, his Persian translation did not need to be of a Sanskrit epic such as the Rāmāyaṇa for it to contain hints of the local context. The few Indic terms in his glossary further substantiate this claim. His ambitions are Persianate insofar as he translates al-Jazari’s Compendium into Persian, but vernacular elements—in both text and image—of the Wonders of Mechanics make it inextricable from its Hindustani context. For an author whose oeuvre contains a poetic commentary on Khaqani, a multilingual Persian lexicon, and this Persian translation of al-Jazari’s Compendium, these few vernacular elements demonstrate Shadiyabadi’s attention to his immediate region. Here, the process of transcreation inflects the kind of text Shadiyabadi brings into Persian from Arabic. Shadiyabadi transcreates al-Jazari’s Compendium by making it more legible to audiences who know Persian but would likely appreciate some Indic equivalents. He reaches a cosmopolitan world literate in Persian, but also the local, interregional patchwork of the Hindustani sultanates.

Transregional Exchanges between Malwa and the Mamluk World

While the Wonders of Mechanics was made in India, let us now zoom out to the wider Indian Ocean and Red Sea networks that likely enabled this translation to occur. Shadiyabadi clearly encountered an earlier manuscript of al-Jazari’s Compendium. And though we do not know which manuscript(s) this was, some evidence suggests a Mamluk path of exchange. Scholars have previously perceived Mamluk characteristics of sultanate manuscripts, but these arguments are primarily supported by stylistic similarities between scripts, ornament, and figural representations.⁹³ In absence of documentary evidence that elucidates the circumstances of Shadiyabadi’s education or access to manuscripts, I propose some pathways of intellectual exchange and the evidence of gifting between Mandu and Cairo. This evidence allows us to further conjecture the Wonders of Mechanics’ Mamluk source.

In several ways, the Wonders of Mechanics attests to a scholastic exchange that may have occurred through the institution of the madrasah or religious school. Mandu’s Madrasah-i Bām-i Bihisht, or the School of the Heavenly Vault, anchored Malwa’s intellectual life. Active around the time the Wonders of Mechanics’ composition, it contributed to a network of madrasahs including the madrasah of Maḥmūd Gavān (completed 1472) to its south in Bidar, and the madrasah of Chanderi (dated around 1425) to its north. Madrasahs, however, were not the sole sites of learning in the fifteenth century. Sufi lodges and informal milieus disseminated and produced knowledge as well. Madrasahs, though, by their design serve as an index of knowledge consumption and production. The architecture of the Nāṣiriyyah Madrasah in Cairo completed in 1303 enables Elias Muhanna to evaluate distinct intellectual strands flourishing in Mamluk Egypt.⁹⁴ Muhanna also draws attention to the rise in the patronage of hundreds of madrasahs in Egypt and Syria during the fifteenth century, a phenomenon which may have inspired the sponsorship of similar institutions in places such as Mandu, Bidar, and Chanderi.

Not only did the Indian sultanates have madrasahs in Hindustan, but they patronized them in Mecca as well. Meccan historians report that these madrasahs were built in coordination with the Mamluks.⁹⁵ These building projects marked the sultanates’ efforts to participate in a transregional Islamicate world in which the Mamluks played a central role. Of the 23 known madrasahs founded in medieval Mecca, the madrasahs of Bengal (date constructed 1410–11), Gulbarga (1427–28), and Cambay (1461–62) are some of the earliest Indian examples.⁹⁶ A letter sent from Sultan Mahmud Shah Khalji of Mandu (r. 1436–69) in early 1467 reports that he had sent an emissary on behalf of the Malwa sultanate to initiate two madrasahs.⁹⁷ These madrasahs were called Bayt Umm Hāni’ (House of Umm Hāni’) and Dār al-Malā‘ibah (House of al-Malā‘ibah). These structures were situated to the south of the Ḥaram mosque, which was a desirable location for a religious institution of the period. We know that the Mamluks kept a watchful eye on the activities and of these madrasahs, because Malwa’s madrasahs were seized and destroyed by an exploitative Mamluk representative in 1462 for an unknown reason.⁹⁸ The Abbasid caliph al-Mustanjid of Cairo also cites Mandu’s madrasahs in Mecca in his letter to Mahmud Shah Khalji dated 5 April 1465, after they were demolished.⁹⁹

There is also reason to believe that the curricula of these Indian madrasahs in Mecca were not restricted to religious instruction. The Gulbargiyya Madrasah associated with the Bahmani Sultanate of Gulbarga and Bidar (1347–1527) accepted students regardless of their affiliation with a particular school of Islamic law.¹⁰⁰ In this case, we can imagine al-Jazari’s Compendium listed on some madrasah syllabi. One of Shadiyabadi’s professors who saw value in the knowledge of automata may have brought back a Mamluk al-Jazari manuscript from Mandu’s outpost in Mecca.

Long-distance contacts with Yemen could have also served as a channel through which Shadiyabadi accessed a Mamluk manuscript of al-Jazari’s Compendium. During the late thirteenth to mid-fifteenth centuries, the nearest port to Mandu, the city of Cambay, had mercantile ties with Yemen. It operated a robust export market of marble stone carving, evidence of which survives in Yemeni centres of Juban and Aden.¹⁰¹ Ornamental decoration from the Manṣūriyyah Madrasah built by al-Malik al-Mansur (r. 1478–89) in Juban offers close comparisons to those on Mandu’s monuments.¹⁰² Furthermore, roughly 50 kilometres away from Juban lies the Yemeni city of Rada‘, where the grand ‘Āmiriyyah Madrasah stands. Scholars have drawn parallels between the architecture of the ‘Āmiriyyah built in 1504 and Mandu’s fifteenth-century monuments. The Congregational Mosque (1454) and the School of the Heavenly Vault (early fifteenth century) of Mandu bear close resemblances to the ‘Āmiriyyah in their portal and horseshoe-shaped archway respectively.¹⁰³ The roof pavilion or chatrī crowning the ‘Āmiriyyah is a common Indian structure that also appears on Mandu’s Mosque of Mālik Mughīs (1452).¹⁰⁴

We may thus envision an adventurous Shadiyabadi entangled within this network between India and Yemen. He could have journeyed to Yemen to study at the Manṣūriyyah madrasah or travellers may have carried manuscripts on their voyage crossing the Indian Ocean. Yet, the direction of artistic exchange from India to Yemen remains to be seen. It is also not certain if the source of the sultanate features on Yemeni madrasahs was in fact Mandu. Gujarat could be a likely source given the similarities between dyed cotton textiles from that region and the wall decorations in the ‘Āmiriyyah.¹⁰⁵ The cities of Gulbarga and Bidar of the Bahmani sultanate would also make likely candidate especially because of the similar kinds of wall painting at these sites.¹⁰⁶ The problem of identifying a sultanate source with precision is significant, suggesting that courts were indeed “networks of people” in this period.

Aside from madrasah networks, the historical record contains an extraordinary report of objects arriving in Hindustan from Mamluk Egypt. The Abbasid caliph of Cairo, al-Mustanjid (1455–79), sent a number of gifts including holy books, a ring, a sword, robes of honour, a flag, and a letter to Mahmud Shah Khalji at the beginning of year 1466.¹⁰⁷ This report is contained in a section of the Traditions of Mahmud Shah entitled, “A Chapter on the Arrival of Messengers from the Abode of Perpetuity, Cairo, and the gifts from Commander of the Faithful/guftār dar zikr-i āmadan-i rasūlān az dār al-khuld miṣr va tuḥaf az hażrat amīr al-mu’minīn.”¹⁰⁸ This gift exchange occurred towards the end of Mahmud Shah’s reign in January 1465 (870 A.H.).¹⁰⁹ The Traditions of Mahmud Shah praises these gifts profusely: “the first gift was like the quarry of jewels before mankind and the source of the elixir of life of the worlds, namely this gift was a Qur’an and blessed books [muṣḥaf-i mujīd va furqān-i ḥamīd]. The sight of [it, i.e. this gift] was like the sights of the brilliant of the most brilliant.”¹¹⁰ “Second, a ring that was the seal of Solomon set with his stone and the stone of the golden sun; this was the ring of the good hand.”¹¹¹ “Third, a sword whose sharpness blunts the sword of the sun, and in the name of the horizon the sun’s sword is concealed.”¹¹² This list of three gifts including two robes of honour (khila‘tayn sharīfayn) is corroborated with the contents of the caliph’s letter fully quoted in Arabic in the Traditions of Mahmud Shah.¹¹³ Through their gifts, the Abbasid caliphs endowed Mahmud Shah with spiritual power. Apparently, these royal gifts and letter from Egypt enamoured Mahmud Shah so much that he dreamt he wore one of the robes of honour, mounted a horse in procession, and appeared before the eighth-century early Abbasid caliphs al-Mansur and Harun al-Rashid.¹¹⁴ The fact that these gifts penetrated the psychology of Mahmud Shah signifies the importance of having the prestige of the broader Islamicate world for the ruler of Mandu, an attitude that perhaps motivated the Wonders of Mechanics’ commission.

The fragments of evidence of Mandu and Mamluk relations allows us to imagine al-Jazari’s Compendium changing hands between scholars behind the scenes of transregional diplomatic affairs. The Egyptian embassy’s gift of a Qur’an and holy books does not mean they presented a copy of al-Jazari’s Compendium to the court of Mandu. Here, the most direct implication is for fifteenth-century Indian Qur’ans.¹¹⁵ Mandu’s initiation of madrasahs in Mecca and Mahmud Shah’s dream cast a portrait of Malwa as an ambitious sultanate. Because of these transregional aims the later ruler Nasir Shah Khalji commanded Shadiyabadi to translate the Kitāb al-Ḥiyal, which implies Shadiyabadi’s access to an original Arabic text.

Nevertheless, I do not want to overstate the case of Malwa-Mamluk exchange, because a Persian intermediary may have facilitated Shadiyabadi’s access to the Compendium. The large numbers of Iranian émigré intellectuals to Hindustan during this period certainly animated Mandu’s republic of letters.¹¹⁶ At the neighboring court of Bidar, the governor Mahmud Gavan forged epistolary friendships with the Timurid-Sufi poet Jami as well as the Mamluks in a distinctive idiom of Arabicised Persian.¹¹⁷ After all, Shadiyabadi was a commentator on Persian poets and the Wonders of Mechanics is in Persian with illustrations showing regional Persian elements.

Conclusion: Transregional, Translation, Transcreation

This study has shown how the transcreation of an Islamicate manuscript at the turn of the sixteenth century in South Asia both connected and distanced locales between Cairo and Mandu. Shadiyabadi’s intellectual horizons were both immediate and distant. Through a glossary and vernacular clarifications, he sought to make al-Jazari legible to his immediate region of Malwa, and quite likely to the world of other neighboring sultanates. In translating al-Jazari’s instructions on how to make certain devices in first-person, he localized al-Jazari for his audiences. Processes of both vast geographical reach and intensive localisation, palpable in the connected book history approach to the Wonders of Mechanics, were important elements of intellectual culture of scholars writing in fifteenth-century Hindustan.

Many constellations emerge through this study of Shadiyabadi’s translation of al-Jazari in Mandu. These networks are as far-reaching as Egypt, Mecca, Cairo, and Yemen, and are as close to home as the Deccan. Such attempts to nativize knowledge of automata in Jazira and South Asia mirror cases in the medieval Latin West as studied by Elly Truitt, and in early modern Northern Europe as analysed by Angela Vanhaelen.¹¹⁸ Yet, every point of connection also came with disconnection; the distancing from a source through reinterpretation, translation, transcreation, and vernacularisation. Here, one cannot help but return to the fact that Shadiyabadi, a scholar whose very name linked him to Shadiyabad or Mandu, was not the first to localise transregional knowledge of automata in Malwa—king Bhoja of Dhar had undertaken a similar task centuries earlier. Likely unbeknownst to him, Shadiyabadi continued this tradition by translating wondrous robots for his community.

Acknowledgements

I am grateful to Daud Ali and Ayesha Sheth for inviting me to contribute to this volume and to the workshop Courts of North India & The Deccan (c. 1347–1562) at the University of Pennsylvania. I am also grateful to Elly Truitt, Lynn Ransom, and the participants of the Translating Science symposium hosted by the Schoenberg Institute for Manuscripts Studies, also at Penn, where I presented parts of this research in November 2022. All errors are my own.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Notes

1. Daud Ali, “Bhoja’s Mechanical Garden: Translating Wonder Across the Indian Ocean, circa 800—1100 CE.” History of Religions 55 no. 4 (2016): 464-465.

2. Upendra Nath Day, Medieval Malwa: a Political and Cultural History 1401-1562 (New Delhi: Munshiram Manoharlal, 1965), p. 23. See also Michael Willis, “Dhār, Bhoja and Sarasvatī: from Indology to Political Mythology and Back”, Journal of the Royal Asiatic Society 22, 1 (2012): 135-36.

3. Ali, “Bhoja’s Mechanical Garden.”

4. Regarding the title of al-Jazari’s see George Saliba, “The Function of Mechanical Devices in Medieval Islamic Society,” Annals of the New York Academy of Sciences (1985): 142. I follow recent work by Lamia Balafrej situating al-Jazari’s work in scholarship on slavery, which has used the titling “A Compendium on the Theory and Useful Practice for the Fabrication of Machines.” See Lamia Balafrej, “Automated Slaves, Ambivalent Images and Noneffective Machines in al-Jazari’s Compendium of Mechanical Arts, 1206,” 21: Inquiries into Art History and the Visual 4 (2022): 737-774. See also Meekyung MacMurdie, “The Manuscript Machine: Assemblages and Divisions in Jazari’s Compendium,” in Destroyed—Disappeared—Lost—Never Were, ed. Beate Fricke and Aden Kumler, (University Park: Penn State University Press, 2022) 113-128, for a recent study of al-Jazari’s Compendium.

5. Sanjay Subrahmanyam, “Connected Histories: Notes towards a Reconfiguration of Early Modern Eurasia,” Modern Asian Studies 31 no. 3 (Jul. 1997): 735-762.

6. Rosalind O’Hanlon, “The Early Modern in South Asia,” Journal of Early Modern History 27 (2023): 159.

7. See Yael Rice, “The Global Aspirations of the Mughal Album,” in Rembrandt and the Inspiration of India, ed. Stephanie Schrader, (Los Angeles: Getty Publications, 2018) 61-77.

8. Elizabeth Lambourn, “Material Cultures of Writing in the Indian Ocean World: A palm-leaf letter at the Mamluk Court,” in The Archaeology of Knowledge Traditions of the Indian Ocean World, ed. Himanshu Prabha Ray (London: Routledge, 2021) 146–68.

9. Chistopher Bahl, Histories of Circulation: Sharing Arabic Manuscripts across the Western Indian Ocean, 1400—1700 (PhD thesis, SOAS, University of London, 2018).

10. Vivek Gupta, “Images for Instruction: A Multilingual Illustrated Dictionary in Fifteenth-Century Sultanate India,” Muqarnas: An Annual on the Visual Cultures of the Islamic World 38 (2021): 77–112.

11. See my discussion of rawshan later in this article.

12. For a study that demonstrates “transcreation” as a concept within manuscript studies, see Vivek Gupta, “How Persianate Is It? A World-Making Book Transcreated from Iraq to India,” Persian Cultures of Power and the Entanglement of the Afro-Eurasian World, ed. Matthew P. Canepa (Los Angeles: Getty Publications, 2024) 238-56.

13. An important study of this period is Francesca Orsini and Samira Sheikh, eds. After Timur Left: culture and circulation in fifteenth-century North India (New Delhi: Oxford University Press, 2014).

14. Jeremiah P. Losty, The Art of the Book in India, London: British Library, 1982, p. 68. See the unpublished notes of Muhammad Isa Waley.

15. BL Or 13718, f. 188b.

16. Éloïse Brac de la Perrière was the first to speculate the Mamluk features of this manuscript. See Brac de la Perrière, “Du Caire à Mandu: La Transmission des Modèles dans l’Inde des Sultanats,” in Ecrits et culture en Asie centrale et dans le monde turco-iranien, Xe-XIXe siècles, eds. Francis Richard and Maria Szuppe, (Paris: Association pour l’avancement des études iraniennes, 2008) 354.

17. The first manuscript is of Mughal provenance and is CBL In 24. Linda Leach dates this manuscript to the 1590s based on its illustrations, but this has yet to be closely studied. See Mughal and other Indian paintings from the Chester Beatty Library, vol. 2, 534. The second manuscript is Lot 146 of Sale 1557 at Christie’s (9 October 2014), attributed to the Northern Deccan, 1600. The manuscript is fragmentary and survives in 72 folios. The third is Occult Sciences Persian MS 301 in the Telangana Oriental Manuscripts Library and Research Institute, Hyderabad, which I tentatively attribute to Mughal India at the beginning of the seventeenth century. Its opening folios and final folios may belong to another manuscript as they do not match BL Or 13718’s text. The colophon page gives the name of the author Mullā Shāh Fatḥullāh Shīrāzī, an intellectual at Akbar’s court, and date 9 July 1585. The fourth manuscript is Misc. Persian MS 219 also located in the Telangana Oriental Manuscripts Library and Research Institute, Hyderabad. This book is incomplete and undated and it is incorrectly titled Ṣinā‘at-i Ablnīvas (Apollonius). The fifth manuscript is MS 708 in the Shahīd Muṭharī Sipāhsālār Library, Tehran, entitled Jar al-athqāl. I thank Daina Buseckaite for drawing my attention to this manuscript. MS 707 of the same collection bears the same title, but its text does not corroborate with Shadiyabadi’s. The Jar al-athqāl is a work on engineering, clocks, and mechanics, and shares many similarities with the Wonders of Mechanics. The Rampur Raza Library holds two manuscripts by this title, RRL A.3689 and A.3691.

18. Losty, The Art of the Book in India, 68; Brac de la Perrière, L’art du livre dans l’Inde des sultanats, 71; Ibid, “Du Caire à Mandu,” 354; and Yves Porter, Z. Vesel and S. Tourkin, eds., Images of Islamic Science: Illustrated manuscripts from the Iranian World (Tehran: IFRI, 2009), 62.

19. Donald Hill, The Book of Knowledge of Ingenious Mechanical Devices (Kitāb fī ma‘rifat al-ḥiyal al-handasiyya) by Ibn al-Razzāz al-Jazarī (Dordrecth: D. Reidel Publishing Company, 1974), 5; al-Jazarī, Al-Jāmi‘ bayn al-‘ilm w’al-‘amal al-nāfi‘ fī ṣinā‘at al-ḥiyal, 15-16.

20. Hill, The Book of Knowledge of Ingenious Mechanical Devices, 9; Saliba, “The Function of Mechanical Devices in Medieval Islamic Society,” 143.

21. Finbarr Barry Flood, The Great Mosque of Damascus: Studies on the Makings of an Umayyad Visual Culture (Leiden: Brill, 2001), 115–16; See also, Elly Truitt, Medieval Robots: Mechanism, Magic, Nature, and Art (Philadelphia: University of Pennsylvania Press, 2015).

22. Flood, The Great Mosque of Damascus, 119.

23. See Persis Berlekamp, “Symmetry, Sympathy, and Sensation: Talismanic Efficacy and Slippery Iconographies in Early Thirteenth-Century Iraq, Syria, and Anatolia,” Representations 133 (2016): 91, for an illustration of the doorknockers in the first al-Jazari manuscript juxtaposed with contemporaneous physical evidence from the Jazira (Mesopotamia).

24. Süleymaniye Fatih 4147, ff. 187-197. Karin Rührdanz, “Zakariyyā al-Qazwīnī on the Inhabitants of the Supralunar World: From the First Persian Version (659/1260-61) to the Second Arabic Redaction (678/1279- 80),” in The Intermediate Worlds of Angels: Islamic Representations of Celestial Beings in Transcultural Contexts, eds. Sara Kuehn, Stefan Leder, and Hans-Peter Pokel (Beirut, 2019), 384–402, and Hannah Hyden, The Ethics of Wonder: The Persian Qazvini Corpus (1300—1632) PhD dissertation (Harvard University, forthcoming).

25. See Anna Caiozzo, “Entre prouesse technique, cosmologie et magie. L’automate dans l’imaginaire de l’Orient medieval,” in La fabrique du corps humain. La machine modèle du vivant, ed. Véronique Adam and Anna Caiozzo (Grenoble, 2010) 43–79.

26. See Vivek Gupta “Remapping the World in a Fifteenth-Century Cosmography: Genres and Networks between Deccan India and Iran,” Iran: Journal of the British Institute of Persian Studies 59/2 (2021): 151–168. For a discussion of how the artists of Mandu’s Bustan (terminus ante quem, 1502-3) may have been indebted to the Cairo Bustan made a few years earlier in Timurid Herat, see Lamia Balafrej, The Making of the Artist in Late Timurid Painting (Edinburgh: Edinburgh University Press, 2019) 222–225.

27. Topkapı Ahmet III, No. 3472 is the Artuqid copy.

28. Fuat Sezgin, ed., Al-Jāmi‘ bayn al-‘ilm w’al-‘amal al-nāfi‘ fī ṣinā‘at al-ḥiyal (Frankfurt: Institute for the History of Arabic-Islamic Science, 2002), vi. It is preserved in the Topkapı Ahmet III, No. 3472. For an analysis see Rachel Ward, “Evidence for a School of Painting at the Artuqid Court,” in The Art of Syria and the Jazira, 1100—1250, Oxford Studies in Islamic Art 1 (Oxford: Oxford University Press, 1985), 69-83. The early studies of this manuscript include Kurt Holter, “Die islamische Miniaturhandschriften vor 1350,” Zentralblatt für Bibliothekswesen 54, no. 2 (1937a): 6, and “Die Galen-Handschrift und die Makamen der Hariri der Wiener Nationalbibliothek,” Jahrbuch der Kunsthistorischen Sammlungen in Wien XI (1937b): 37, that show twelve of its folios are later additions. J.M. Rogers, The Topkapı Sarayı Museum—The Albums and Illustrated Manuscripts, trans. and ed. J.M. Rogers (London: Thames and Hudson), 30-31, accurately dates the manuscript to 1206.

29. These include MS Topkapı Sarayı, Collection Hazine 414 (dated thirteenth century?); a dispersed manuscript dated 1315 with folios in American collections; CBL Arab 4187 (fourteenth century); MS Aya Sofia 3606 (dated 1354); and, Topkapı Sarayı Collection Ahmet III, No. 3350 (dated 1459). Hassan’s introduction to his critical edition of the Compendium, 12-14, and Sezgin’s facsimile of MS Aya Sofia 3606, v-vi are most instructive for reconstructing this manuscript history.

30. BL Or 13718, f. 3a. The introduction also contains the statement where Nāṣir Shāh commands Shadiyabadi to translate the Kitāb al-Ḥiyal of the Banū Mūsā into Persian. “farmān shud kih kitāb-i ḥiyal-i banī mūsā kih ‘arūsī-yi maḥrūs bilibās-i istabraq tāzī-yi malbūs ast, agar ūrā bikiswat-i parīnān-i bārsī muktafī kardānī.”

31. Jean Aubin, “Indo-Islamica I La Vie et L’oeuvre de Nīmdihī,” Revue des études islamiques 34 (1966): 71, who cites the manuscript of the Kanz al-mā‘ānī of ‘Abd al-Karīm Nīmdihī, Istanbul ms. Esad Efendi no. 884, ff. 194a-196b. I thank Emma Flatt for providing me with this reference. See Flatt, The Courts of the Deccan Sultanates: Living Well in the Persian Cosmopolis (Cambridge: Cambridge University Press, 2019) 85.

32. Hill, The Book of Knowledge of Ingenious Mechanical Devices. The Bodleian manuscript is MS Greaves 27.

33. These include CBL In 24 (1585-90), QNL HC.MS.02290 (Deccan, ca. 1600), Christie’s Lot 147 (Northern Deccan, 1600), Bodleian MS Fraser 186 (dated 1638), Rampur Raza Library MS III.34 (18th century), and NYPL Spencer Indo-Pers 02 (copied 1621, illustrated in the 19th century). BL Add MS 23391 relates to NYPL Spencer Indo-Pers 02 but was likely made in Iran. Bodleian MS Fraser 186 factors prominently into Hill’s study, and Barbara Schmitz recorded the latter two manuscripts in Islamic Manuscripts in the New York Public Library, 165-66; and, Ibid, Mughal and Persian Painting, 79.

34. BnF MS Arabe 2477, f. 112a. tamma al-kitāb wa’l-ḥamdu li’l-malik al-wahāb waqt ghurūb yawm al-juma ‘sāb‘rabī‘al-ākhir sanna tisa‘īn wa thamānīmā’a li-l-hijrah. wa wajadtu ‘alā’l-nuskhah alatī naqaltu minhā mā ṣawwartuhi nujaz hādhā’l-kitāb yawm al-juma‘rāb‘ṣafar sanna 742 li’l-hijrah. wa kānna maktūban ‘alā’l-nuskhah alatī naqaltu minhā hādhahi al-nuskhah mā mithāluhi wa hādhahi al-nuskhah manqūlahi min nuskhah naqaltu min khaṭṭ al-mawlif wa ammā al-ḥurūf wa’l-ibdāl wa rasūm ṣūrat al-ashkāl fa-mimmā rasmahi bi-ḍabṭ wa rasmahi bi-khaṭṭahi raḥmat allāhu ta‘ālā. al-farrāgh min nuskhahā rāb‘jumādā al-ākhira sannah ithnayn wa sittamā’ah.

35. These are CBL Arab 4187 (fourteenth century) and MS Aya Sofia 3606 (dated 1354). The codex of CBL Arab 4187 is currently not accessible and therefore its colophon cannot be verified. A survey of all extant copies shows that there were two copies made some time in the fourteenth century and 1354 respectively. This makes the existence of a 1341 copy very probable.

36. BnF Arabe 2477, f. 112a. This exact formula of “letters, variants, and drawings of the devices,” appears in several manuscripts. Notably, it appears in the dispersed manuscript dated 1315, and the colophon of Bodleian MS Greaves 27, f. 113b. For the former, see Duncan Haldane, Mamluk Painting (Warminster: Aris & Phillips Ltd, 1978), 36, and for the latter see Hill, The Book of Knowledge of Ingenious Mechanical Devices, 206.

37. BnF Arabe 2477, f. 26a.

38. Ibid, f. 8b.

39. NYPL Spencer Indo-Pers 02.

40. NYPL Spencer Indo-Pers 02, f. 23a (copy); f. 34a (drawing).

41. BnF Suppl. Persan 1145 and 1145a.

42. Ibid, f. 3a.

43. CBL In 24, f. 1b, lines 1-8 would be the original text that should appear on BL Or 13718, f. 2a.

44. The opening words of BL Or 3299, f. 2b and CBL In 24, f. 1b both read, “ḥamd-i mutavāfir va sanā’-i mutakāsir.”

45. BL Or 13718, f. 188b.

46. Hill, The Book of Knowledge of Ingenious Mechanical Devices, 17; al-Jazarī, Al-Jāmi‘ bayn al-‘ilm w’al-‘amal al-nāfi‘ fī ṣinā‘at al-ḥiyal, 11.

47. BL Or 13718, f. 5a.

48. The elided section concerns the cast brass door made for a king’s palace. This section may have been on folios that were possibly removed during the process of rebinding the manuscript. See Hill, The Book of Knowledge of Ingenious Mechanical Devices, 191-5. The extra devices are found on BL Or 13718, f. 189a-b. The first is clearly labeled as ṭasht or basin. The second device’s illustration appears as a swirling wheel.

49. Flatt, The Courts of the Deccan Sultanates, 85.

50. BL Or 13718, f. 7a.

51. There are two cases where there is space left for an illustration that does not appear. The first is on f. 32a where the illustration would show one of the six bows (i.e. bracing rods) for holding spheres together, although a crude drawing has been added in the margins. The second appears on f. 109a where the diagram would have been of the matrix of entry holes into the valve (buzāl).

52. This differs from CBL In 24, the 1585-90, Mughal Wonders of Mechanics, that contains illumination in several of its diagrams.

53. Representative diagrams where figures or animals are omitted appear on ff. 41b, 44a, 64a, 67a, 79b, 83b, 85b, 88a, 89a, 91b, 98b, 105b, 138b, 144a, 147b, 150b, and 173a of BL Or 13718. BnF Arabe 2477 exemplifies the phenomenon of adding the figures last in the drawing process.

54. Hill, The Book of Knowledge of Ingenious Mechanical Devices, 27. In Arabic, this corresponds to “wa’l-bawāq la ḥājjah illā ṣūratahi idh laysa lahu ḥarakah wa ṣawt al-būq yakhraj min juhat ukhrā.” See al-Jazarī, Al-Jāmi‘ bayn al-‘ilm w’al-‘amal al-nāfi‘ fī ṣinā‘at al-ḥiyal, 37. In Persian Shadiyabadi writes, “va yik shakhṣ az jumlah-i ṣūrat-hā dar īnjā namūdah shud bas fahm kun ẓāhir-rā va bā taṣvīr-i bāqī-yi ālāt-i ḥājjat nīst az ānkih ānrā ḥarakat va āvāz nīst. [f. 17a] va āvāz-i būq az duvvam ṭarf-i ū bīrūn ayad va ṣurat īnast,” See BL Or 13718, f. 17a-b.

55. BL Or 3299, f. 98.

56. CBL In 02, f. 36b, the Nujūm al-‘Ulūm (Bijapur, 1570), isolates this turban decoration as a motif and defines it as sunbalah, or ear of corn.

57. BL Or 13718, f. 104b.

58. A closer match of a figure on bended knee appears in the Wonders of Mechanics’ depiction of figures in a drinking session. See, Ibid, f. 120b.

59. Bodleian MS Greaves 27 dated 1486.

60. MS Aya Sofia 3606, f. 345a and BL Or 13718, f. 120b.

61. This kind of draping on robes derives a long genealogy in Islamicate arts of the book. Many examples can be found in thirteenth-century Arab painting.

62. BL Or 13718, f. 116a.

63. Ibid, 114a. For another early example see the pupil’s headdress in a depiction of Aristotle and his pupil in the thirteenth-century Kitāb na‘t al-ḥayawān, BL Or 2784, f. 94a.

64. See the medallion on BnF Arabe 2324, f. 137a, made in 1307-10, Tabriz, that has these flares. They also appear on the far outer borders of the Qur’an, Keir Collection VII.42, attributed to the end of the fourteenth century. For this second example, see Brac de la Perrière, “Manuscripts in Bihari Calligraphy,” 68, fig. 4.

65. Bodleian MS Greaves 27, f. 67b. They also do not appear on fourteenth-century concordant illustrations, but these figures do have highly intricate textile garments. See the figure in Boston, Museum of Fine Arts, 31.125, from the dispersed manuscript dated 1315. Here, the automaton wears a luminous flower-patterned green and gold textile robe. Also, it has been previously argued that the Mughal Wonders of Mechanics clearly derived from Mamluk model, however the absence of folds such as these seem to indicate that it took its cue through an intermediary manuscript such as the original Mandu manuscript under discussion (cf. CBL In 24, f. 48b). Linda Leach makes this claim about a Mamluk model in her catalogue of the Chester Beatty Library’s Indian manuscripts. See Leach, Mughal and other Indian paintings from the Chester Beatty Library, vol. 2, 534.

66. See Travis Zadeh, Mapping Frontiers across medieval Islam: Geography, Translation, and the ‘Abbāsid Empire (London: I.B. Tauris, 2011), 1-12. For the Persian cosmopolis, see Emma Flatt.

67. BL Or 13718, f. 3a. “farmān shud kih kitāb-i ḥiyal-i banī mūsā kih ‘arūsī-yi maḥrūs bi-libās-i istabraq tāzī-yi malbūs ast agar ū-rā bi-kiswat-i parīnān-i bārsī muktafī kardānī.” For a discussion of sartorial metaphors used for translation, see Finbarr Barry Flood, Objects of Translation: material culture and medieval Hindu-Muslim encounter (Princeton: Princeton University Press, 2009), 61-88, and Vivek Gupta, “Arabic in Hindustan: Comparative Poetics in the Eighteenth Century and Azad Bilgrami’s The Coral Rosary,” Journal of South Asian Intellectual History 4/2 (2021): 192-93.

68. Emblematic of this is Audrey Truschke, Culture of Encounters: Sanskrit at the Mughal Court (New York: Columbia University Press, 2016). Obrock, “Muslim Mahākāvyas,” identifies ways in which sultanate Sanskrit texts respond to Persianate culture.

69. A good example of Turkic or Chaghtai translation into Persian is the Bāburnāmah. This totalizing tendency is well represented by the Nujūm al-‘Ulūm (Stars of the Sciences), CBL In 02.

70. Rieu, Catalogue of Persian Manuscripts in the British Museum, vol. 2, 464.

71. Here, it is worth recalling an earlier instance of the widespread translation of scientific texts, namely the early emergence of the Arabic illustrated manuscript in the tenth century that occurred through several adaptations of Byzantine culture and translations from Greek works on the cosmos and sciences. See Eva Hoffman, “The Beginnings of the Illustrated Book: An Intersection Between Art and Scholarship,” Muqarnas 17 (2000): 37-52.

72. Flood, Objects of Translation.

73. For the Shahnamah’s glossary see Peyvand Firouzeh, “Convention and Reinvention: The British Library Shahnamah of 1438 (Or. 1403),” Iran: Journal of the British Institute of Persian Studies 57 (2019): 12. The Wonders of Mechanics’ glossary appears on BL Or 13718, ff. 189b-190b.

74. Ibid, f. 190a.

75. BL Or 3299, f. 214b. The related definition of mortar, or javāz, is included in the Key of the Learned and illustrated. See, Ibid, f. 89a.

76. R. S. McGregor defines the word kharasān as a whetstone with a Brajbhasha etymology. See McGregor, ed., The Oxford Hindi-English Dictionary (Oxford: Oxford University Press, 1994) 230.

77. Richard Eaton, India in the Persianate Age (New Delhi: Allen Lane, 2019) 135.

78. Future research will unlock the significance of the other vernacular words included in this glossary.

79. For Merklinger’s reference to a chajja in Mandu’s architecture, see Sultanate Architecture of Pre-Mughal India, 160.

80. BL Or 13718, ff. 59b-60a.

81. The second example appears on Ibid, f. 92b, lines 8-9.

82. BL Or 13718, f. 60b.

83. This is a detached folio from Süleymaniye Aya Sofia MS 3606 now in the Freer Gallery of Art, F1932.19.

84. al-Jazarī, Al-Jāmi‘ bayn al-‘ilm w’al-‘amal al-nāfi‘ fī ṣinā‘at al-ḥiyal, 580.

85. Steingass, A Comprehensive Persian-English Dictionary, 595. If calling the balcony in Arabic a rawshan may have been perceived as a register shift from one language to another, this may explain why Shadiyabadi includes the word chajah.

86. M. A. Dhaky, ed., Encyclopaedia of Indian Temple Architecture: North India Beginnings of Medieval Idiom, ca. A.D. 900-1000, Vol. II, Part 3: Text (American Institute of Indian Studies: New Delhi, 1998) 147-85, 406, 410, plates 373 and 390, and “The Genesis and Development of Māru-Gurjara Temple Architecture,” in Studies in Indian Temple Architecture, ed. Pramod Chandra (New Delhi: American Institute of Indian Studies, 1975) 134–5, also refers to the khuracchādya, or curved and ribbed eave or awning.

87. Vaman Shivaram Apte, Revised and enlarged edition of Prin. V. S. Apte’s The practical Sanskrit-English dictionary (Poona: Prasād Prakāshan, 1957–59) 716.

88. Kalika Prasad, Bṛhat Hindī Kosh (Vārāṇāsī: Jñānamaṇḍala, 1965) 260. “chatkā dīvār ke bāhar nikalā hūā bhāg.”

89. Elizabeth Merklinger, Sultanate architecture of Pre-Mughal India (New Delhi: Munshiram Manoharlal, 2005) 160.

90. Ebba Koch, Mughal Architecture: An outline of its history and development, 1526-1858 (New Delhi: Oxford University Press, 2002) 137–8. As a preface to her glossary, Koch states, “the meaning of vernacular terms has, where possible, been derived from Mughal sources of the sixteenth and seventeenth centuries. Sanskrit-derived terms which were adopted by the Mughals are transliterated according to their spelling by Mughal authors.”

91. Merklinger, Sultanate architecture of Pre-Mughal India, 100.

92. My forthcoming book, Worldshaping Wonders: Manuscripts and Experience in Hindustan, treats the traces of the vernacular in the diagrams, especially the case of Shadiyabadi’s mahout.

93. Saryu Doshi, “Colour, Motif, and Arabesque,” in India and Egypt: Influences and Interactions, eds. Saryu Doshi and Mostafa El Abadi, Marg 45, no. 2 (1993): 112-35, and Brac de la Perrière, “Du Caire à Mandu.” Finbarr Barry Flood’s reading of the foundational inscription at the Congregational Mosque of Mandu (completed 1454) has stressed that it was called a nuskhah (copy) of the Great Mosque of Damascus. See Flood, “Idea and Idiom: Knowledge as Praxis in South Asian and Islamic Architecture,” Ars Orientalis 45 (2015): 153n8. See Zafar Hasan, “The inscriptions of Dhār and Māṇdū,” Epigraphia Indo-Moslemica (1909-10): pl. XI where the naskh inscription clearly reads, “nuskhah bayt al-ḥarām.”

94. Elias Muhanna, The World in A Book: Al-Nuwayri and the Islamic Encyclopedic Tradition (Princeton: Princeton University Press, 2018), 56-65.

95. Richard T. Mortel, “Madrasas in Mecca during the Medieval Period: A Descriptive Study based on Literary Sources,” Bulletin of the School of Oriental and African Studies, University of London 60 no. 2 (1997): 236–52, bases his survey of the Meccan madrasahs on the writings of Taqī al-Dīn Muḥammad ibn Aḥmad al-Fāsī (1323-1429) and Najm al-Dīn ‘Umar ibn Fahd (d. 1480).

96. Ibid, 244-49.

97. John L. Meloy, “ ‘Aggression in the Best of Lands’: Mecca in Egyptian-Indian Diplomacy in the Ninth/Fifteenth Century,” in Mamluk Cairo: Crossroads for Embassies, eds. Malika Dekkiche and Frédéric Bauden (Leiden: Brill, 2018); BnF MS Arabe 4400, f. 181a; Aḥmad Darrāj, “Risālatān bayna Sulṭān Mālwa wa’l-Ashraf Qāytbāy,” Majallat Ma‘had al-Makhṭūṭāt al-‘Arabiyyah 4 (1958): 112. The manuscript containing these letters, BnF MS Arabe 4400, has received attention from Malika Dekkiche, “The Letter and Its Response: The Exchanges between the Qara Qoyunlu and the Mamluk Sultan: MS Arabe 4440 (BnF, Paris),” Arabica 63 (2016): 579-626. Meia Walravens, “Arabic as a Language of the South Asian Chancery: Bahmani Communications to the Mamluk Sultanate,” Arabica 67 no. 4 (2020): 409–435, explores similar issues for the Bahmani sultanate.

98. Meloy, “Aggression in the Best of Lands.”

99. Bodleian MS Elliot 237, ff. 263a-264a.

100. Mortel, “Madrasas in Mecca during the Medieval Period,” 246.

101. Elizabeth Lambourn, “Carving and Communities: Marble Carving for Muslim Patrons at Khambhāt and around the Indian Ocean Rim, Late Thirteenth—Mid-Fifteenth Centuries,” Ars Orientalis 34 (2004): 107. Lambourn’s study of the list of luggage of an Indian Ocean merchant in the Geniza documents reconstructs an earlier, twelfth-century micro history of these pathways of portability. See Abraham’s Luggage: A Social Life of Things in the Medieval Indian Ocean World (Cambridge: Cambridge University Press, 2018).

102. Selma Al-Radi, The ‘Amiriya in Rada‘: the history and restoration of a sixteenth-century madrasa in the Yemen, ed. Robert Hillenbrand (Oxford: Oxford University Press, 1997), 86-7; Venetia Porter, “The Architectural Decoration,” in The ‘Amiriya in Rada‘: the history and restoration of a sixteenth-century madrasa in the Yemen, ed. Robert Hillenbrand (Oxford: Oxford University Press, 1997), 121.

103. Al-Radi, The ‘Amiriya of Rada‘, 90, 94; Brand, “The Khalji Complex in Shadiabad, Mandu (India),” 117-20.

104. Al-Radi, The ‘Amiriya of Rada‘, 90, 99.

105. Ruth Barnes, “The Painted Ceiling of the ‘Amiriya. An Influence from Indian Textiles,” in The ‘Amiriya in Rada‘: the history and restoration of a sixteenth-century madrasa in the Yemen, ed. Robert Hillenbrand (Oxford: Oxford University Press, 1997) 139–47.

106. Vivek Gupta, “Interpreting the Eye (‘ain): Poetry and Painting in the Shrine of Aḥmad Shāh al-Walī al-Bahmanī (r. 1422—1436),” Archives of Asian Art 67 no. 2 (2017): 189-208.

107. Bodleian MS Elliot 237, ff. 261b-264a; Day, Medieval Malwa, 213.

108. Bodleian MS Elliot 237, f. 261b.

109. This date is based on Upendra Nath Day’s reading of the manuscript. See Medieval Malwa, 213.

110. Bodleian MS Elliot 237 f. 261b. “avvalān tuḥfah kih ma‘dan-i javāhir bijānib-i ‘ālimiyān va manba‘-i āb-i ḥayāt jahāniyān ast ya‘nī muṣḥaf-i mujīd va furqān-i ḥamīd kih dīdah baṣīrat aw law al-abṣār az tajallī-yi mutajallī shavad.”

111. Ibid, “thāniyān khātimī kih khātim-i sulaymān zīr dast nigīn-i ū va nigīn-i zarīn-i khūrshīd dast-i khūsh angushtarīn ūst.”

112. Ibid, f. 262a. “thālithān shamshīrī kih az tīzī-yi ū tīgh-i āftāb bī-tāb shudah dar bi-nām-i ufuq mukhtafī shavad.”

113. Ibid, ff. 263a-264a. Interestingly, on folio 263b, the letter also mentions a majlis that the Mamluks hosted in Mecca in honor of a certain Shihāb al-Dīn Aḥmad al-Sindī al-Tarjumān (the Translator).

114. Ibid, f. 264a-b. This is corroborated by ‘Abdullāh Muḥammad ibn ‘Umar al-Makkī al-Āṣafī Ulughkhānī, Ẓafar al-wālih bi-Muẓaffar wa-ālih [An Arabic history of Gujarat], ed. E. Denison Ross, 3 vols. (London: John Murray, 1910-28), vol. 1, 204, written in the early seventeenth century. Also see, Meloy, “Aggression in the Best of Lands.”

115. Brac de la Perrière, “Du Caire à Mandu.”

116. The Timurid ruler, Abū Sa‘īd Mirzā, also sent an emissary to Mandu in 1468. Bodleian MS Elliot 237, f. 302a; Day, Medieval Malwa, 213.

117. Flatt, “Practicing Friendship,” 69; Ibid, The Courts of the Deccan Sultanates, 188-97; and, Walravens, “Arabic as a Language of the South Asian Chancery.” For an earlier study of Gāvān, see Richard Eaton, A Social History of the Deccan, 1300-1761: Eight Indian Lives (Cambridge: Cambridge University Press, 2006) 59–77.

118. See Truitt, Medieval Robots, and Angela Vanhaelen, “Strange Things for Strangers: Transcultural Automata in Early Modern Amsterdam,” The Art Bulletin 103 no. 3 (2021): 42-68.