While this essay will speak primarily from a historian's perspective, I want first to acknowledge that the idea of “societies” has had a peculiarly interdisciplinary hold on scholarship, bridging the humanities and social sciences. For sociologists it is a disciplinary foundation, for anthropologists it is often an escape hatch from the problematic notion of “cultures,” and for historians it is an organizational scheme that has stealthily metamorphosed into an unconsidered worldview. While some disciplines, especially sociology, have invested more effort than others in critically examining the idea of societies, each has produced critiques that perhaps should have killed off “societies” altogether. And yet enormous numbers of new works continue to take “societies” as their objects of study (e.g., Thum Reference Thum2014), with even transnational and global studies framed as the interaction of supposedly distinct societies.

Engseng Ho is not the first to critically examine the notion of societies, nor does he claim to be. And yet despite the wealth of critiques of “societies,” these seductive products of collective imagination persist. What Ho offers us are new tools to navigate an intellectual landscape that may never let go of “societies.” His particular critique of societies grows from a fresh and deeper genealogy of the “internalist” notion of societies, locating its roots earlier than the high modern history of sociology. Ho's tracing of “societies” to Adam Smith and early European imperialism frames the “internalist” approach not simply as a misreading of the world, but a product of that world. This in turn means that societies are also categories in use by the historical actors we study—one of the reasons that “societies” have been so tenacious. Ho's toolkit, then, is not dependent on a wholesale discarding of the notion of societies, but functions especially well in discourses that are heavily conditioned by the internalist view.

This essay applies some of Ho's tools to a case in which the identification of societies is particularly thorny, and in which essentialized reifications of societies are seemingly implacable: the interpenetration of India and China. Thanks to the efficiency of maritime travel, discussions of mobility tend to focus mainly on examples that involve the sea. But as Ho notes, land-based mobility has an equally important role to play. Particularly before the arrival of the telegram and steam-powered travel, we might even argue that land-based networks, precisely because of their inefficiency, throw some of the advantages of Ho's contribution into particularly high relief. I want to look a little more closely at the peculiarities of land-based circulations by pulling at the various threads that have interwoven India and China over the last 2,000 years, especially Islamic threads.

When we think of longstanding exchanges between the subcontinent and East Asia, perhaps the biggest one, the kind of thing that makes it into world history textbooks, is the transmission of Buddhist traditions in the first millennium CE. The usual story of Buddhism's spread from India to China shares a great deal with the approaches Ho is critiquing: it is a story of unidirectional movement, not so much of people but ideas. It is, in Ho's words, a story of “marvelous things batted back and forth between separate societies, cultures, and civilizations.” What has not made it into our understanding of Asian history is that land-based connections between the subcontinent and East Asia did not end with the fading of Buddhist practice in the subcontinent. It turns out that these land-based connections have continued into the present. After the waning of Buddhism, the “mobile societies” that nurtured these links were made up of Muslims. And if we look at the body of Islamic texts consumed all over China today, we see the imprint of those South Asian connections on the shape of China's Islamic canons. Interestingly, Chinese Islams made very little impact on Indian Muslims, making this look initially very much like the traditional story of Buddhism's spread. That is, this seems to be a one-way movement. I want to address two questions about this later Sino-Indian connective pattern. The first is Ho's question: “Were these marvelous things batted back and forth between separate societies, cultures, and civilizations or did they spread and find new homes on the backs of societies that were mobile and interactive, interlaced, interpenetrated, and intermarried with others?” The second question is, why has this linkage gone unnoticed until now?

My answer to the first question is of course yes, it was indeed mobile and interpenetrated societies that enabled the movement of Indian Islams into East Asia. And this particular case throws some light on the relationship between circulation and transmission. On the far western end of the Qing Empire, the Karakoram and Pamir mountains threw up barriers that made trade profitable. The very inefficiency of transportation across these ranges elevated the prices of goods that could only be obtained by crossing them. And since high-priced, mobile luxuries, in this case pashm wool and opium, beckon with high profit margins, the imposing natural barriers ironically resulted in concentrated, regular, and professionalized channels of merchant travel. So it was that the famous Dabistān-i Maẕāhib, a description of various religions written in India in the seventeenth century, used informants from what is now Xinjiang for its description of Islam (Dabistān-i Maẕāhib 1877, 261, 266). A similar phenomenon played out in the forests of Burma and southwest China, where dengue fever, dense vegetation, and mountains colluded to create the deeply rutted roads of mule caravans (Ma and Ma Reference Ma and Ma2014). Both the mule caravaneers and the Karakoram merchants were predominantly Muslims: Panthays or Sino-Muslims in the case of the Burma link and Turki people over the Karakoram. The economics of barriers led to repetition and thus circulation.

Circulation, as Ho points out, creates the infrastructure, social and physical, upon which one-way migrations, mobilities, and transmissions are possible: guest houses, roads, markets for fresh animals, fields reserved for caravan fodder, kinship networks, etc. The case of Islam in China seems at first to be a case of one-way migration and dissemination. But the India-China movement hitched a ride on the backs of circulatory travelers and through circulatory channels. The repeating mule trains of the Panthays and the trans-Karakoram/Pamir paths made it possible for scholars and shaykhs to bring Sufi traditions from Sirhind via Afghanistan to Qinghai (Papas and Ma Reference Papas, Wei, Hille, Horlemann and Nietupski2015; Thum, Reference Thum, Takehiro, Brophy and Yasushiforthcoming) and to print Yunnanese texts in Lucknow (Murata Reference Murata2000).

I would also like to suggest, however, that in some cases the repetition that is circulation can develop into highly regularized patterns that are deeply ingrained in intermediary societies, and that, ironically, such deep entrenchment can also insulate mobile populations from those around them. Indian texts made a big splash in the mosque libraries of China, but texts from the intermediate territory did not. Our muleteers’ networks were well enough established to make it unnecessary for a traveling scholar to engage with the intellectual traditions of Burma. The effect is something like the insulation on electrical wiring. Along the length of the wire, it suppresses the local effects of a passing signal, but by doing so manages to transmit a signal over a longer distance. Unfortunately this effect makes it easier for us to imagine separate civilizations batting around marvelous things, and to ignore the mobile, entangled societies that are not just doing the work of the batting around, but must be considered partly constitutive of those supposedly separate civilizations. The illusion makes sense because we are steeped in the one-way migrations that have displaced circulation over the last century.

Here Ho's historiography is particularly useful for reminding us of the dominance of trans-hemispheric (or old world–new world) migration in discussions of mobility and the effects of this dominance on our approaches to inter-Asian connection. There is something quite special about movements between the Americas and the Eastern Hemisphere: they are very new and they were built out of nothing in a very short period of time. Within the Americas and within the Eastern Hemisphere, any new migration movement had to take into account older patterns of circulation. But with the sudden linking of the Americas to Asia and Europe in the fifteenth century, the routes, economies, and cultures of mobility had to be built from scratch. Perhaps it is the drama of this rupture that has planted the trans-hemispheric model, dominated by one-way migrations, so firmly in our minds. What inter-Asian connections, and particularly land-based connections, suggest is that ancient patterns of migration and circulation demand a special attention to the effects of ancientness, to the fact that any change in the movement of people is, in the inter-Asian context, an adaptation of something preexisting, rather than an innovation.

These dynamics of adaptation need more study. They are important, because, as Ho notes, old circulations reemerge in surprising places. When the Chinese Communist Party moved into Yunnan, many Panthays took to the rutted mule roads and moved to Burma (Forbes Reference Forbes1986). Some continued onward to Thailand, where they established the Attaqwa mosque and school (Attaqwa Mosque 1988). Today, in Yunnan, the Xiaoweigang mosque keeps a copy of a nineteenth-century Yunnanese book printed in Lucknow, India, but it bears the stamps of the 1960s Attaqwa mosque in Thailand, as well as a Chinese inscription from the 1980s, testifying to the revived circulation of people and ideas along old routes. On the other end of the Sino-Indian frontier, when the Communist Party took control of Xinjiang, Uyghur Muslims hitched a ride over the Karakorum mountains along the old merchant paths, bought debts owed by merchants in British India, and used them to fund their travel to Mecca, where they are now Saudi citizens known as Turkestanis (Balci Reference Balci2007; Thum Reference Thum2016).

Let us turn now to the second question: why has the interpenetration of Islamic China and Islamic India gone unnoticed? The dominant explanations of Islam's role in China focus on a presumed incompatibility between Islam and Chinese culture. In this understanding, syncretism is the most salient trait of Chinese Islam, an Islam characterized by its “surprising” accommodation of Chinese concepts such as the dao (“way”). This view is encouraged by highly nation-bound thinking, in which Islam is both naturally foreign to the Chinese national character and, when it takes root, isolated from the rest of the Islamic world by that distinctive character. In Ho's terms, Muslims in China pose a threat to the aggregation we know as China. They ask us to disaggregate China into the tangle of interconnections that constitute it. The common emphasis on syncretism, by contrast, serves to defend the aggregate of China; if the strange beasts of Chinese Islams are really the hybrid offspring of two separate civilizations, those civilizations can remain intact. In practical terms this has led, for example, to a focus on Chinese-language Islamic texts (e.g., Frankel Reference Frankel2011). Such texts have been attractive because Chinese-language Islamic terms borrow a great deal from Confucian and Daoist terminology and therefore appear to support the syncretism approach.

But by deemphasizing the Persian and Arabic texts, the Islamic interweaving of China and India becomes almost invisible. Such a view can only survive if we restrict ourselves to the body of Chinese Islamic texts written in Chinese, a canon known today as the Han kitab. In practical terms this focus is understandable, given that scholars of general Islamic history usually lack the linguistic tools to deal with China, and Sinologists rarely have training in Arabic-script languages (another effect of the “internalist” bent of most scholarship). But the Han kitab was only one fragment of the Islamic traditions that have flourished throughout China; Chinese Muslims have long engaged with texts in Persian, Arabic, and even Turkic, especially before the early twentieth century. Those texts portray a very different kind of Islamic China, one closely connected to the rest of the Islamic world through India, the Hijaz, and Central Asia, both consuming and creating texts firmly and obviously embedded in inter-Asian Islamic networks. It is also a highly variegated Islamic terrain, striated by Sufi orders, competing scholarly lineages, and overlapping textual cannons that knit together Muslims of multiple ethnic and linguistic groups within South and East Asia.

Ho's outside-in approach might encourage us to start with the Persian and Arabic sources and follow the links inward. In a step-by-step movement from Lucknow to Mandalay to Kunming to Shanghai, from Persian to Chinese, from Sufism to Confucian-inspired Islamic apologetics, “China” as a supposedly stable, continuous culture disappears among the cosmopolitan interconnections of moving texts and people. Mobilities, circulations, and partial societies force us to look below the narrative surface of aggregates like India and China, draped as they are in a puzzle-piece cloth of identity categories, at the tangle of human intercourse. Thanks to the transnational turn, such tangles are the subject of much scholarship, but they often feature as exceptions—the routinized exceptions that buttress category narratives, the batting around of marvelous things. Ho asks us to normalize the tangle itself (rather than a few exceptions) and, where possible, to do so without resorting to the categories that paper over interconnection.

And yet, we historians are not alone in our drive to parcel the world out into regionalized units. Our subjects did it too. This calls, to use Ho's terms, for a reexamination of the constitution of the internal, from the perspective of the external. Part of the story of inter-Asian connection is the vast investment humans make in hiding and eliding linkages that shape their lives. For example, Turki Muslims of Xinjiang under Qing rule built a shared historical vision out of the rich textual corpus that inter-Asian connections had brought to their region over the preceding centuries. But they stripped these texts of elements that made connection obvious (Thum Reference Thum2014). Reading through this canon of domesticated texts, or listening to its recitation, Turki people might have imagined something not too distant from a “society,” all the more so because they could not easily see the diverse origins of their shared textual corpus. Perhaps we may even take this as an auxiliary definition of society: the selective elision of interconnection, which parcels the world into territorialized (or de-territorialized, in the case of diaspora) units. As much as Ho's mobility toolkit dissolves societies, it also throws such elision into high relief and with it the historical forces that have imagined societies into existence.