Introduction to premodern war and religions: comparison, issues and results

ABSTRACT

Building and elaborating on the dossier’s five articles on East Africa, the Maghreb and Islamic Spain, Western Europe, the Aztec Mesoamerica, and the native American Northeast (plus Japanese and Byzantine history), this introduction discusses quandaries about comparison in the intertwined disciplines of History and Anthropology and suggests some hypotheses as to the relation between premodern warfare and religions. Side-switching was demonized (and punished as a quasi-religious sin) in Western Christianity, not so as a rule in the other societies here compared. It was ‘treason’. Sexual violence and rape was inhibited by religious conceptions in the same society, and among the natives of the American Northeast. Non-human powers might help or intervene in warfare, but there is no general pattern. As for the presence or absence of holy war, there may be correlation with the type of polity concerned. Established empires may be averse to the emergence of charismatic figures and sacral practices, as one sees with China and Byzantium. Central imperial elites may also dislike miracles, especially in offensive warfare. Evidently, while religion might shape this or that aspect of warfare, it was not the sole provider of ‘conditions of possibility’.

KEYWORDS:

• Comparison
• religion
• war
• rape
• treason
• miracles

The concept of ‘condition of possibility’, borrowed from No-Kantianism, allows scholarship to move away from the blunter notion of ‘causality’, which assumes that a phenomenon is the mechanical consequence of another phenomenon (or several other). One phenomenon renders (or several together render) possible that another phenomenon will have the shape or form that it takes. One can do here no better than adapt Keith Baker’s musing about a Foucauldian account of the French Revolution ‘that could not proceed by reference to causal necessity. It could only identify conditions making the Revolution possible without absolutely determining its occurrence or’ making necessary ‘its specific character’ (1994, 192). Applied to religion and warfare, it allows to simply avoid the banal problematic of religion as a cause for violence. Furthermore, it replaces the necessity inherent in causality with a possibility: given this or that aspect of religion, it is possible (but not necessary) that this or that aspect of armed violence will take this or that form. This dossier of articles examines the extent to which religions in the pre-modern world served as conditions of possibility for warfare, that is, shaped it.¹ Did all human societies regardless of their religions display the same patterns when waging, imagining, or remembering war? Alternatively, can scholarship identify an uneven distribution of similarities distributed? That is, for a given aspect of warfare (for example, the treatment of prisoners), were there societies that behaved in the same manner, and other societies that diverged from this pattern? Did a given religion (for example, Christianity) have the same formative impact across all the societies in which it was present? In aggregate, the five articles in this dossier provide materials to delve comparatively into two practices and notions connected to them: sexual violence in armed conflict and side-switching. This introduction will therefore first discuss the comparative method (part 1), then by way of background present data about roles and gender at war (part 2), before moving on comparatively to treason (part 3) and rape (part 4). It ends with a discussion of non-religious factors (part 5) and with the question of belief (part 6).

The articles are devoted respectively to the native Americans of the Northeast (Wolfgang Gabbert), East Africa (Richard Reid), the Mexica Aztecs (Caroline Dodds Pennock), early imperial China (Peter Lorge), Maghribi and Andalusian Islam (Pascal Buresi), and the Catholic West (Philippe Buc). Arguably, one of the dossier’s strengths is the breadth of the units considered, from empires (stable as imperial China or conquest-driven as with the Aztecs) to smaller polities (Western European, Japanese, sub-Saharan African, Western Islamicate) and polycentric aggregates (the American Northeast). These societies were knit together diversely: by family and kin, by lord-retainer bonds, or by institutions all the way to central administration. Some knew organized religion, with institutions and religious specialists. In other units, such as in the American Northeast, contact with extra-human powers was not a profession and did not make up a status group.

Comparison in the disciplines of History and Anthropology

Reflections about comparison and its limits or even legitimacy have since long involved the disciplines of Anthropology and History. A century ago, historical comparison was quite popular, and its legitimacy, common sense. No wonder. In an age in which Euro-American empires still held sway, comparison was often beholden to, or served directly master narratives opposing the West to the Rest. As is well known, Max Weber (d. 1920) and Émile Durkheim (d. 1917) inscribed their foundational approaches within the agenda – explicit with the German, implicit for the Frenchman – of explaining European specificity and European modernity. Durkheimian Sociology was hegemonic in France from the 1920s to 1950s and influenced history writing. It legitimized comparison and even called for it. Famously, one of the founders of the Annales School, Marc Bloch, drew on then current knowledge about Japanese history in his magnum opus, La société féodale (1939). More dogmatically, Frédéric Joüon des Longrais, a lone French pioneer in research on Japan, yet representative of the trend across area studies, could confidently trumpet in his L'Est et l'Ouest (1958, 1): ‘Il n'y a de science que du général, il n'y a donc qu'une méthode permettant une synthèse, la méthode comparative’ (The sole object of science is general patterns; there is therefore only one method allowing synthesis, the comparative method).

This happy confidence among Historians was shattered in the twentieth century’s last quarter, in great part owing to a ‘turn’ in the neighbouring discipline of Anthropology. Comparison fell out of grace starting with the 1970s and into the 1990s in au courant quarters owing to attacks on the method, most famously by Edward W. Said in his Orientalism (1978) and by contributors to the book edited by James Clifford and George E. Marcus, Writing Culture: The Poetics and Politics of Ethnography (1986). This volume’s most famous chapter may be Talal Asad’s ‘The Concept of Cultural Translation’ (1986).² A fundamental issue was the asymmetry in authority between the observer (usually from the West) and the observed (in the Rest). Another hurdle was scepticism as to the possibility of cultural translation. Was the object of observation so transparent that with a proper investment in time the Anthropologist could understand it fully? While many historians happily insulated from Anthropology and literary criticism kept comparing, such obliviousness could be professionally costly.

Yet after an eclipse, Comparative History returned. It returned with less totalizing ambitions. It now aims less at finding general rules than at parsing differences and explaining similarities when they occurred (Olstein 2015). It is understood to have the potential to (1) subvert assumptions about one or several of the societies being compared, (2) destabilize master narratives, and (3) complicate universalizing assumptions about human nature or culture (Haupt and Kocka 1996; Mahoney and Rueschemeyer 2002; Brekke 2006; Morris and Scheidel 2009; Bynum 2014). More recently (4), it has returned to proposing and explaining patterns in global history (Duindam 2018; Strathern 2019). Comparison also allows to ask new questions of familiar terrains. This renewed acceptance in History is likely related to new positions in the neighbouring discipline of Social and Cultural Anthropology (see Gingrich and Fox 2002; Gingrich 2012, 215–216) and to the overcoming of the postmodernist reflexes of the Clifford and Marcus generation. One will mention in particular Marilyn Strathern. Early on in The Gender of the Gift (1988, esp. 1–21), she took stock of translation question hegemony of Western concepts as problems, but turned these problems into an opportunity to refine both her analysis of New Guinea and of these very same Western concepts.

Anthropologists and Historians, however, must keep pondering the legitimate scope of generalization. The heated debate between Marshall Sahlins and Gananath Obeyesekere (see Borofsky 1997) over Captain Cook highlights this. We have on the one hand Sahlins’ reconstruction of Cook’s trajectory in Hawai’i, pitching two cultural scripts (one of which was ‘religious’, the islanders’ script). We have on the other hand Obeyesekere’s insistence that the rationalities at play on both sides, British and Hawai’ian, were empirical, instrumental, and tactical, and therefore in the last analysis very similar. In the controversy’s wake, Sahlins explored comparatively two disconnected wars, the Ancient Greek Peloponnesian War (431–404 B.C.E.) and the nineteenth-century conflict opposing two Fijian kingdoms. Greece’s Athens and Fiji’s Bau followed the same modes of maritime warfare, with a logic leading to imperialist expansion; the land-based polities, Sparta and Fijian Rewa, followed another logic, logics which reacted against one another and thus differentiated themselves further, logics that alas the book does not pursue in depth (Sahlins 2004). Sahlins is comfortable finding consistency within a culture (and thus speaking of culture); Obeyesekere would rather start from a shared human nature, which means that all human groups are prone to empirical, instrumental, and tactical action, and not determined by culture.

Countering worries about generalizing, Candea (2019, 205–219) has explained how it necessarily belongs to Anthropology. Members of his discipline regularly practice ‘caveated generalization’. Alan Strathern (2019, 2022) will readily agree. The Lilliputian conflict between lumpers and splitters, also classic in the discipline of History, can be overcome. André Gingrich (2015), in a short survey of the encounter between History and Anthropology aiming at elucidating issues of comparison, ends with the suggestion that there should be mutual tolerance between approaches tending to look for common traits across cultures (that is more prone to lumping) and those tending to relish a culture’s particularities (that is more prone to splitting).

Still, are societies, religions, or cultures proper objects? Are concepts like ‘religion’ extendable from the West to all of the Rest? Can one assume that a society or a ‘culture’ is a bounded and coherent, homogenous entity? Can one compare big and small human groupings together regardless of size? As for one of this dossier’s key terms, religion, Jonathan Z. Smith has since long underscored how starting in the sixteenth-century ‘religion’ was constructed on materials and assumptions coming from Protestant and Catholic Christianities. This invention of religion came with a systematic double bias: on the one hand, it affirmed a continuum between Christianity and alien conceptions and practices; on the other hand, it hierarchized them, with Christianity on top (Smith 1998). The learned concept of ‘religion’ that matured in the nineteenth century was extended with great ease to the neighbouring Abrahamic monotheisms, Islam and Judaism, and accessorily to Buddhism. It was projected also on practices and conceptions that far from constituted systems involving dogmas and scriptural canons. Further, the concept’s hegemony during the colonial era impelled non-western religious specialist entrepreneurs to refashion existing native conceptions and practices in order to fit them in the category of religion. This sometimes allowed these social actors to self-fashion as authorities on the religion they had de facto invented, and seek power and monopolies (see Campany 2003; and for example Oberoi 1994 for Sikhism; but see the nuanced assessment of Ānanda 2012 for Japan).

Given his critical genealogy of ‘religions’, their comparison was in Smith’s opinion highly problematic. Early on (1982, 20–36), he ironically assimilated comparison in religious studies to magic as opposed to science, thus dismissing it.³ In the wake of an earlier book comparing the current relation of China and India’s politics to religious ‘traditions’ over the watershed of their respective imperial experiences (Van der Veer 2014), Peter van der Veer has recently levied a more constructive critique. He endorses against the totalizing ambitions of Émile Durkheim Marcel Mauss’ paradoxically named ‘fait total social’ approach – the focus on a single item or practice that engages all the members of a society in all its dimensions. It allows, van der Veer submits, the ‘(…) elucidation of complex phenomena through comparison without “generalization” or modelling of “social systems”’ while steering ‘clear of both universality and endless particularity’ (these opposite shoals correspond to the Lilliputian extremes of radical lumping and radical splitting). As I read it, van der Veer proposes that ‘Anthropology’s focus on the sociocultural fragment’ permits to side-step what twentieth-century historical sociology made into essences, for example nations, civilizations, and religions, ‘without denying their social power’ (2016, 148). Will one, however, go so far in refusal of generalizations as to compare only ‘sociocultural fragments’? Does the refusal of essences mean to refuse to compare entities larger than fragments? Or looking at Van der Veer’s glass as half-full, shall one brush aside the ‘social power’ of nations, civilizations, and religions?

Remarkably – given that she did so in the very same decade as the somewhat nihilistic Writing Culture volume was published – Marilyn Strathern proposed a dialogical ‘strategy of negation’ that overcomes Smith’s objections. According to Strathern (1988, 17), ‘the strategy of negation’ that is, to say that those other humans the scholar observes

do not have this or that is a statement highly dependent on the character of what this or that is for those who do have it. It can thus be seen simultaneously as a displacement of meaning and as an extension of it. I displace what ‘we’ think society is by a set of different constructs, promoted in opposition to order to suggest an analogy with ‘their’ view. At the same time, that very analogy grasped as a comparison (…) then extends for us the original meaning of the concept.⁴

Applied to the term ‘religion’ so vociferously attacked by Jonathan Z. Smith, Strathern’s reasoning empowers scholars to start from initially Western notions, use them while keeping them at bay, and end up with an extended concept. The dossier’s five articles employ a broad etic definition of religion, product of our initial discussions during the 2017 Workshop:

Religion is a set of practices and discourses shared with greater or lesser intensity by the members of a human group, structured by a community, institution or both. These practices and discourses deal with non-human entities or non-living humans to whom are attributed powers not normally available to human beings.⁵

Marshall Sahlins (2017; see also Graeber and Sahlins 2017, 2–3), followed by Alan Strathern (2019, 29–31 and passim) recently proposed the term ‘metaperson’ for such beings (an extension of religious studies’ notion of divinity).

As for war, while its deep history, reaching back into prehistory, has animated heated discussions, especially among environmental psychologists and other materialists, it is less controversial a concept. War is here defined as a situation in which ‘defined groups engage in purposeful, organized and socially sanctioned combat involving killing each other (…)’. We adopt two of Margaret Mead’s (1968) three definitional criteria. War implies, first, the ‘organization for the purpose of a combat involving the intention to kill (…), [and, second,] social sanctions for this behaviour, which distinguishes it from murder of members of its own group’ (see Kortüm 2001).⁶

If it is legitimate to use the concepts of ‘religion’ and ‘war’, is it legitimate to study war and religion in the Maghreb, Western Europe, the New World’s Northeast, and imperial China and Islamicate polities insofar as this means treating these geopolitical entities as coherent, somewhat uniform items? Is the proper approach, rather, not much more processual, leaving boundaries hazy and focusing on relations? Histoire croisée has its limits, despite excellent works (for example Subrahmanyam 2012). Taking a swipe at Global History’s love for entanglements, Jürgen Osterhammel (2016, 42) underscored how ‘(…) connections alone do not suffice to get to grips with the major institutions of the modern era, especially states and the structures of global capitalism. States and, by extension, empires are more than mere networks’.⁷ While there was not as much institutionalization in the premodern world as in the modern age (with which Osterhammel is concerned), it remains that one can hardly study vectors (connections) without understanding the nature of points (societies). And to work with points, one has to allocate to them, at least heuristically, a degree of internal consistency. Anthropologist Matei Candea (2019, 298) underscores that one has to posit the existence of such items possessing ‘some kind of unity (and difference) across time and space’ (in our case societies and religions) when one seeks to operate in a disciplined fashion what he calls frontal comparison, that is, an examination of the other from the background of our Western disciplinary knowledge. This approach (as Marilyn Strathern theorized it in the 1980s) can refine or problematize the categories and concepts constitutive of this very disciplinary knowledge.

The issues at the intersection of History and Anthropology are thus clear. Was premodern warfare primarily conditioned by pragmatic, practical rationality, a rationality that while incarnate in time and place, and influenced by social and political structures, is common to all human cultures? If so, was the conduct, experience, and commemoration of warfare determined by local context and conjunctures? Or was each society’s conduct, experience, and commemoration of warfare culturally scripted, in particular by religion? If so, was religion the sole or dominant local script? Answers pass through comparison, not only frontal ones but also lateral ones, putting side-by-side several societies to which the scholar is alien. This dossier’s editor, while taking due note of Peter van der Veer’s position, sides with Marshall Sahlins’ statement in Apologies to Thucydides (2004, 1): ‘It takes another culture to know another culture’.

When all is said and done, Comparative History, like Anthropology, will move forward only via ‘engagement with research problems, rather than from elaborating methodologies’ (Gingrich and Fox 2002, 21). What matters are the research questions. Juxtaposing the articles in this dossier, two such questions suggest themselves. They do not involve comparing essentialized items like religions, but aspects (in conformity with van der Veer’s creed) of culture that may owe something to religion: sexual violence (rape) and side-switching. We must first, however, make clear a few facts.

Gender at war, specialization

The objects studied in these five articles range across a wide spectrum from principally civilian societies to principally warrior societies. In the Tenochca Aztecs’ belief system, everything was war; everybody was a warrior; even activities on the ‘home front’ of female labour were, spiritually seen, warfare (Dodds Pennock 2022). In Catholic Europe, with the exception of the famous Joan of Arc (either hallowed as miraculous exception or demonized as a bloodthirsty witch and heretic), there are no female holy warriors, and very rarely fighting women. Some Muslim authors did document (or perhaps merely imagined as a perversion) the active participation in crusade fighting of European females (Nicholson 1997).⁸ In European sources, women appear in war only as companions and helpers. They do not wield weapons. Better documented, the early modern period shows that there were many women in armies, for logistical help (Lynn 2008), which was possibly the case in the Middle Ages. However, Christian concerns with purity likely inhibited women’s direct participation in combat. As in the Aztec case, Western societies had a home front, especially for holy wars: all were to mobilize themselves, via prayer and financial offerings, plus moral reform (Buc 2015, 52). Women fighters are also absent in the Islamic Maghreb. However, the Almohads feminized their religious opponents, the Almoravids. This was facilitated by the fact that the males among these desert Lamtūna Berbers wore a face veil, and by the fact that their women went about unveiled (Buresi 2022; Fromherz 2010). For Sub-Saharan Africa, one rare example of African women warriors is provided by the West African Kingdom of Dahomey’s female battalions (see Edgerton 2000). While Northeast America’s Iroquoian females did not fight, they participated in decision-making, triggering raids to avenge or replace their male dead, and played a role in the war-related torture of prisoners (Gabbert 2022).

Did all males participate in warfare? As the Native American shaman was not a religious specialist (in the sense that managing the spirits would have been his only calling), like Aztec priests, he fought. Configurations, as Buresi underlines, varied within the Islamic Far-West (the Maghrib plus Al-Andalus). The Maghribi Almohads made each male a fighter for Jihad, transgressing the existing norm that holy war was a collective duty, understand, that this duty was fulfilled if there was enough collective effort; the Taifa Muslims of Spain preferred to wage the ‘greater’ or spiritual Jihad of the pen and virtue; the Maghribi Almoravids waged war conventionally, mobilizing some but not all, under an Amir (Buresi 2022). Chinese thought insisted on the distinction between military and civilian, but practically speaking in some periods a sizable number of elite men partook of the virtues of both fields, and consumed symbolic assets belonging to the field they were not ascribed to (Wyatt 2009; Ryor 2009).

Side-switching – what the West calls ‘treason’

The religious imprint on side-switching, where it exists, can be highlighted comparatively. It is present in Western Christianity (Buc 2022), absent in Japan until the sixteenth century’s paroxysm in the archipelago’s secular civil wars (Buc 2020). Longue-durée Japanese tolerance for turncoats is exemplified by the Heike Monogatari ([ca. 1371] 2014, VIII.8, 432–437) in its description of the death of the Taira partisan Seno-o no Kaneyasu. Captured by Minamoto Yoshinaka, Kaneyasu convinced his guard Nariuji that he would switch to Yoshinaka’s side, and further give Nariuji his eponymous domain (Seno-o). On the way, Kaneyasu got drunk Nariuji with sake; killed him and his thirty retainers; and then mobilized against Yoshinaka, calling forth in the region ‘every man loyal to the Heike’. Yoshinaka could well afterwards lament: ‘I don't like this at all (…) I should have executed him’. Yoshinaka’s retainer Kanehira had warned him indeed that Kaneyasu ‘was no ordinary man. I urge you a thousand times to do it, but no, you had to spare him’. Kaneyasu died heroically; two retainers fought alongside and died of their wounds. Yoshinaka inspected their heads and said, ‘Ah, (…) these were true warriors, each worthy to face a thousand. What a shame I could not spare them’. The lack of a systematic critique or even demonization of side-switching in Japan lasted until perhaps the mid-sixteenth century (Buc 2021, 274–281). In East Africa, side-switchers were punished (or pardoned) when captured, but there is no discourse on treason. In Africa in general, many warlords had double ethnicity or family loyalties (maternal and paternal), which might have made side-switching less of an object of critique.⁹ The non-demonization of turncoats seems to have been a feature of most Islamicate societies (Buresi, personal conversation, 22 October 2018). On the one hand, integration of the defeated and others may be a tendency of nomadic societies, or simply of societies where labour is scarce, as in Sub-Saharan Africa; on the other hand, religious norms also came into play. Sunni legal discourse prioritized the reintegration of rebels into the Community, even at a cost to the Islamic ruler (Kraemer 1980; Abu El Fadl 1999, 2001). Bent on purge rather than on reconciliation, the North African Almohads, focus of Buresi’s article, may be an exception, explainable by the initially eschatological nature of the movement, founded by a Māhdi. Or the Almohads may represent a potential or virtuality within Sunni Islam, seldom actualized, but still existent as a potential or virtuality (see the West African jihadi Caliphates evoked by Reid, 2022; Buc forthcoming). They parallel the often actualized potential for purge inherent in Western Christianity (Buc 2015, ch. 5). The American Northeast natives considered intestine warfare a serious issue (one reason why the Iroquois League was formed was to limit blood feuds). Finally, the Aztecs stood at the opposite pole. Their foundational myths glorified civil war, and thus they likely did not ‘demonize’ the side-switching practices that the West calls treason and (thanks to its initial belief in demons) demonizes.¹⁰

Rape and sexual violence

Women’s fate as victims at war has been a hot theme since at least former Yugoslavia’s civil war (1991–1995). It is embedded in wider controversies about the naturalness of rape. Whether, that is, males are pushed by nature to rape females (and inhibited to do so by culture) or whether to the contrary they are motivated by so-called ‘rape cultures’ (Goldstein 2001, ch. 6; Gottschall 2004; Clark 2014). The issue is a minefield. Recent studies cannot avoid memory politics, and rape memories have often been politicized insofar as the choice to testify or remain silent can be motivated by the desire to avoid shame or by the will to shame (Hayden 2000, 33).¹¹ Furthermore, like violence in general (see Clastres [1977] 1997), sexual violence oftentimes considered to be constitutive of other, primitive thus implicitly inferior cultures (Baxi 2014, 145). This said, our dossier does provide materials for the topic, and suggests the differential impact of culture and religious culture on sexual violence.

Arguably, medieval Christianity had an inhibiting effect, evidently far from total (Buc 2022). The absence of rape among the natives of the New World’s Northeast is clear (Gabbert 2022).¹² This documented avoidance finds a non-politically correct counterpart in the Islamic law of war (Buc 2021). Across the world of Islam, it was established norm that the females of a defeated infidel enemy would not be massacred but enslaved, by contrast with the women of legitimate Muslim enemies.¹³ Females distributed as booty might be quickly subjected to sexual abuse (Ali 2016, ch. 3). After Saladin’s crushing victory at Hattin over the Christian crusader states (1187), his armies conquered, among other places, Bethlehem, then Jerusalem. For both cities, the Sultan’s secretary and biographer Imad ad-Din al-Iṣfahānī (1972, d. 1201) unabashedly devoted long gloating descriptions to the violation of females (Buc 2021, 5–6). For the Western Islamic polities studied by Buresi (2022), Maghribi sources indicate that sexual violence against recognized Muslim enemies was frowned upon. The early Almohads did sell the women of tribes considered ‘hypocrites’ into slavery, but there is no mention of rape. I know of only one document suggesting Almohad sexual abuse of Christian women in battleground Spain: a poem composed to celebrate Caliph al-Mansor’s crushing victory at Alarcos (1195), which describes ‘patrician’ Christian females as ‘tamed antelopes’ that sob at night, disheveled (Al-Marrākushī 1893, 254–256). Possibly, Almohad rigorism limited sexual violence as a whole – even against the Almoravids, whom they considered false Muslims.

Encounter, that is, inter-cultural contact, did not necessarily modify these differential propensities. In the Holy Land, Latin Christians complained of Muslim sexual violence, but did not mention their own camp’s abuses, be it positively or negatively; Muslim authors both complained about Frankish abuses and mentioned positively their own sides (Friedman 2002, 169, 171). Both sides shared in sexualized allegories of conquest, but only one claimed actual sexual conquest.¹⁴ As for the Northeast America examined by Gabbert, the Europeans might rape, but the Amerindian natives did not approve of it, and bad European example did not lead them to adopt it (Abler 1992; White 1991).

Other societies occupy a place in between, strengthening the hypothesis of a diversity influenced by religious culture. Karl Friday (2009, 73–74) has listed a number of instances of sexual violence and sexual shaming for the period of 1000–1200 in Japan.¹⁵ The only brake to rape at war seems to have been the fear of retaliation or the desire to spare the honour of a clan one might want to ally with. All in all, rape does not appear often in Japanese war tales. It rears its ugly head at the very end of the long section devoted to the destruction of the Kamakura Shogunate and of the Hōjō family of shogunal regents (1333) in the Taiheiki (compiled circa 1360/70, but likely with materials contemporary to the events for the 1330s). There the author(s) pile up spectacular vignettes, positive and negative: acts of sublime and spectacular bravery in hopeless combat or suicide, mournful reflections on husband-wife affection in death, or shameful betrayal of one’s lord or relatives to save one’s hide or get some gain. The elegy for the allegedly corrupt regime in Kamakura recognizes both vice and grandeur in defeat. Among the vignettes come intimations of massive violence done to women: ‘(…) there came uncertain tidings of Kamakura: of chaste wives, pledged to eternal fidelity, who were carried off by wretched countrymen to taste the sorrow of Wang Chao-Chūn (…)’ (Taiheiki [ca. 1360/70] 1959, 335). The reference to a Chinese lady’s forced marriage to a barbarian is not to rape on the spot, but the implications of sexual violence, and by low-class perpetrators to boot, are clear.¹⁶ Significantly, sexual violence is in no way one of the causes of the downfall of the Kamakura shogunal regime (and for that matter of the hegemonic Ise Taira house in the earlier wars of 1180–1185). The gods and Buddhas punish with defeat many transgressions (including, as one might expect, destruction of shrines and temples), but rape is not one of these sins.

As to Africa, explains Richard Reid (2022), there is no particular evidence of rape at war. Women were captured, and many of them ended up as concubines, likely in greater proportions in the household of leaders. There is scant evidence also for rape during Aztec conquests, although one may cautiously assume a tribute in women on the basis of what happened to the Spanish conquistadores (Pennock, email of 16 November 2018). Claude Lévi-Strauss (1967, ch. 5) theorized the exchange of women as the pinnacle of gift-giving, itself the prime vector of amity and alliance. And conversely, for Lévi-Strauss (1967, 78) war meant the failure of bonding: There is a link and a continuum between hostile relationships and the providing of reciprocal services: Exchanges correspond to peacefully resolved wars, and wars are the product of failed transactions.¹⁷ Normal exchange of women is what Marshall Sahlins, in dialogue with Thomas Hobbes and Marcel Mauss (2017 [1972, 2003], 155–164, 202–205) identifies as balanced reciprocity, usually asymmetrical. It is as in Lévi-Strauss the prime means of peace and alliance. In many cultures, one did indeed offer women to other groups as a sign of peace or to establish an alliance (how much say these females had in these bargains is open to debate). Marilyn Strathern (1988, 374) notes that ‘Homicide compensation for injury to the clan [in Hagen, New Guinea] replaces the person of the [killed] man much as bridewealth replaces the persons of clan sisters’. This further incites to consider the American Northeast practice of forcibly taking females (and males) into a clan as standing on a continuum with negotiated marital alliance and conflict resolution through wergild (Sahlins 2017 [1972, 2003], 204), but denying all bonds with the other party. So considered, abduction represents an extreme, disjunctive hypertrophy of what Sahlins calls ‘negative reciprocity’ (2017 [1972, 2003], 173; see as well in this direction Clastres 1997, 54–55). Thus if one takes the logic of this tradition in Anthropology further, the capture of women, and worse, their sexual shaming, is a demonstrative refusal of peace or at the very least the signal that a renewal of peace would be steeply asymmetrical. This seems verified by studies on the Yugoslav, Rwandan, and Congo conflicts (Hayden 2000; Clark 2014), where sexual violence is a means to prevent and deny, demonstratively and in public, any form of togetherness between the ethnic groups concerned. By stealing women, the native Amerindians did not aim at establishing bonds, even asymmetrical, with English settlers or other tribes. Their interest was solely in replenishing the clan. However – and here I come to the point – the refusal to rape attested in Jerusalem’s 1099 conquest during the Catholic First Crusade, in the 1573 Bartholomew’s Day Massacre of Huguenots, and other religious war episodes (Buc 2015) does not fit what one would come to expect given the Lévi-Strauss and Sahlins model. Catholic holy warriors between ca. 1100 and ca. 1600 definitely wanted to break off all relations with religious antagonists, be they pagans or heretics. However, their religious beliefs inhibited sexual violence. Not all violence (for in 1099 crusaders and in 1573 Catholic guerriers de Dieu murdered and tortured cruelly, this so with glee) but this specific type of violence, rape. Here History interrogates what Anthropology seems to suggest: If sexual violence is a prime way to refuse reciprocity, why then do some cultures refuse rape and chose other means to deny relationship with the enemy?

Comparatively, we have thus the following configuration: a culture of war where rape is normatively and explicitly forbidden (Christian Europe); one where rape is indifferent (Japan); one where rape would be at odd with a religious imperative of integration of females into the victors’ families, yet involves forced unions, among the American Northwest Iroquoians (Gabbert 2022) and Sub-Saharan Africa; one where the martial elite's desire for labour in an underpopulated continent pushed to the forcible incorporation as slaves of male and female captives (Sub-Saharan Africa, see Reid 2022); and one where the enslavement of females from different religious groups is normatively fostered and can involve rape (Muslim North Africa and Syria). All the same, rape is not the sole war atrocity. By contrast, Islamic jurisprudence forbade the killing of non-combatants, including women – these very non-combatants who were slaughtered without problem in Japan and during Christian holy wars (Buc 2015), and in North America as well (sometimes with atrocious tortures) when they were not spared to be integrated in the victorious group to compensate tribal deaths.

It is not all religion(s)

While the workshop’s starting hypothesis was that religions shape some dimensions of armed violence, religions are only one factor among those giving form to warfare – its conduct, its experience, its commemoration. Other forces or cleavages (Gorski and Türkmen-Dervisoglu 2013) do intervene. Imperial China and East Rome a.k.a. Byzantium (Stouraitis 2015) exhibit commonalities across religious difference. The imperial centre of respectively China (often but not always dominated by a Confucian court elite) and Christian Orthodox East Rome manifested the same reluctance to allow generals to obtain a religious aura. This charisma might be used to take over the imperial centre. This may be one of the reasons why East Rome refused to grant the status of martyrs to soldiers fallen in battle (Haldon 1999, 27–33). As shown by Lorge, under the Tang and the Song, Chinese Confucian elites regulated carefully who would receive a cult in the capital’s Military Temple. This showcased discipline and loyalty to the emperor over and above spectacular successes. This Chinese peculiarity is worth commenting upon.

In the imperial Chinese case, a key discursive binary is the duality wen (文) – wu (武), civil – martial, corresponding (among other correspondences) to a contrast between, on the one hand, enticing rewards and proper rites, and on the other hand, warfare, fear-inducing punishments and coercive social control. Looking at the manoeuvres around the imperial capital’s Martial Temple discussed by Lorge (2022), it would seem that Confucian elites in negotiation with the emperor during the Song (960–1279) and Tang (617–907) dynasties controlled martiality (and thus warfare). One can see why they would have wanted to. Contextually – first with the opening of the Empire to Central Asia and its nomadic militancy under the Tang (Skaff 2009), then with the catastrophic Song loss of North China to the semi-nomadic Jurchen – but also structurally, given the systemic pairing between wu and wei, which expressed itself in rites (Lagerwey 2010, 9–10), the imperial capital was almost forced to negotiate again and again, at the level of cult, the relationship between martial and civic virtues. Structurally, that is, since rather than being in simple opposition, wen and wu were in a dialectical relationship. They had to be distinct, but both had to exist; in the words of Nicola Di Cosmo (2009, 18), elites viewed too sharp an opposition between the two ‘as a sign of decay’. The capital had to address martiality because of the historical context (to discipline present and future generals) and because of this central Confucian paradigm.

The Sunni Abbasid Khalifate in its heyday shared the Byzantine and Song reluctance to let warlords acquire a charismatic aura (Heck 2004, 109–111). Such is likely the logic of mature empires: beyond the initial phase of imperial foundation or expansion, the ruling centre and its religious institutions will try to inhibit the emergence of holy war charisma. Contrariwise, it is telling that the much more decentralized early medieval Frankish and Armenian polities generated a pattern different from the Byzantine one. Independently of one another, Christian Frankland and Christian Armenia elaborated in an analogous sequence first the figure of the saint who fights spiritually Christendom’s enemies, then, second, the figure of the holy warrior who fights these enemies both spiritually and with material weapons – a warrior who is willing for God’s sake to suffer martyrdom and to slaughter enemies. Without being in contact, the Franks and the Armenians drew on the same biblical fundus scripts for holy war and the annihilation of the religious opponent (Buc 2015; Preiser-Kapeller 2017). Further, the Christian Ethiopian highland kingdom waged war in ways reminiscent of Armenia and Francia. There also warfare was linked to religious identity, combining physical and miraculous violence (Reid 2022). One empires and three smaller, more segmentary polities thus related in opposite ways to sanctified warfare, even though they shared the same Christian religion.¹⁸ A moment of centralization in Catholic Europe further supports this thesis. The late eleventh and twelfth century, arguably the apex of the crusading movement (Flori 1991), witnessed the emergence of the monarchical papacy. By the last third of the twelfth century, the popes monopolized the right to declare men or women saints. Throughout the High and Late Middle Ages, martyrdom at war existed, was frequently mentioned and hoped for, but was seldom recognized by Catholic Christendom’s papal centre. The monarchical papacy, a clerical state, managed sanctity with as much finesse as Imperial China. Dead holy warriors were not a threat to the popes, but the popes chose to canonize other profiles of saints, in the main men and women with proven loyalty to the papacy or its key lay allies (Vauchez 1997).

Belief and credence

Finally, dealing with religion and warfare raises the issue of belief. What sort of a ‘belief’ did one have in non-human intervention in or for warfare? What was the credence attributed to these miraculous actions? Belief, credit, credence, faith denote different modalities. China was a case of its own. The ‘modality of belief’ (régime de croyance, Veyne [1982] 1989) fostered by the constituted Confucian philosophy-theology was one of the studied distance to the preternatural. In his discussion of Confucianism, Kiri Paramore (2017, 21) cites the Analects, ‘[The Master] sacrificed to the dead as if they were present. He sacrificed to the spirits, as if the spirits were present’, and goes on to submit that ‘over two thousand years, people in East Asia have thought about religious ritual in these “as if” terms’.

In the imperial centre, the wen-wu dialectic was managed by a Confucian negotiation, on Confucian terms. The authority enjoyed by Confucianism at court meant that whatever Buddhism and Daoism had to offer for warfare (they did have something to offer) would be inflected by Confucian dominance. Some Chinese Buddhist sources recount how Buddhist rites provoked the intervention of Buddhist deities and powers; interestingly, these metapersons did not kill directly, but scared the enemy, thereby contributing to victory.¹⁹ This might be owed to some Buddhist pacifism, which might have made the writer shy away from having the gods fight directly. Alternatively, and more likely, the Buddhist author accommodated himself or herself to a court Confucian audience: rites could produce only a ‘quasi’, illusions of divine intervention. When we go to Japan, however, sources betray no such idea that maledictions had merely illusionary effects. Divine powers did fight in human battles (Conlan 2003, 170–181). In Japan, Buddhism alloyed with kami worship was strong enough among the warrior class, and not only with it. Courtier diaries expressed laconically wonder at miracles, but did not question them; chronicles, more verbose, gave them credence. A stunning example comes from the diary of Hanazono tennō (1314). The emperor recalled how during the preceding century’s Mongol invasions a wounded snake had appeared and revealed via a medium that he and other gods had fought the enemy (Conlan 2001, 252). In the Catholic West and in Byzantium too, divine intervention and actual war miracles occurred, and the military saints might fight materially – defensively in Byzantium (in particular to protect cities), aggressively during the Catholic crusades (Lapina 2015) and during other wars conceived of as holy. An example of the latter is the 1264 battle of Lewes (which saw the victory of rebellious English barons fighting for a sacralised constitution of the realm, all the more sacred as the realm’s leading men had sworn oaths to uphold it).²⁰ As for Africa, Richard Reid warns us in this volume that one walks on eggs and treads difficult postcolonial political ground when employing missionary or colonial administrators’ reports that precolonial African warriors relied on magic or deities. The colonial cliché of the primitive and violent, superstitious ‘Dark Continent’ is never far away. Yet one can surmise cautiously that these men did turn to preternatural means and metapersons. On the other hand, there is no evidence even in this potentially biased evidence that in East Africa and Sub-Saharan West Africa metapersons intervened during the fight itself.²¹ Divine action in combat is also absent in sources documenting Aztec warfare, and among the natives of the American Northeast. There, metapersons, while contributing to success in combat, did not participate in it.

Conclusion

Comparison highlighted variations both across and between the societies considered as well as limited patterns. Some societies’ warfare and war cultures were more shaped by religion than others (most markedly the Aztecs, and a close second West and Central European Christianity). Religious factors did affect notions of side-switching as well as the valence of sexual violence. Yet the same ‘faith’ did not always shape warfare in the same manner, since other factors intervened. The Christian religious elites of the European West, Armenia, and Ethiopia shared notions of holy war that were in the main absent from Christian East Rome, an explanation being the contrast between segmentary polities (with a royal centre but politics determined by lord-follower or patron-client bonds) and empire. As for islamicate polities, the thrust of Buresi’s contribution is that given a putatively ever-identical ‘Islam’, one observes in the interconnected spaces of the Maghreb and al-Andalus three different relationships to warfare.²² Whether this is due to variations within this Islam, or to conjuncture, or to both, cannot be determined. Two neighbouring East Asian societies with comparable religious building-blocks also elicit significant differences. War miracles were routinely accepted and sought by Japanese warrior and courtier elites (Conlan 2003), an attitude not present at the Chinese imperial court. A factor may be the exact relationship obtaining in the two ensembles between Buddhism, Confucianism, local kami cults or Daoism. Or is it the nature or size of the polity? As the refrain goes, more studies are needed … And as another refrain goes, this volume meant to provoke questions, and not answer them all.

Acknowledgements

This dossier of articles is the derivative product of a workshop, based on pre-circulated papers, held on 19–20 May 2017 at the Österreichische Akademie der Wissenschaften in Vienna (Austria). It brought together a dozen scholars on the theme of religion and warfare. Vienna University’s Karin Jirik provided efficient organization. Not all workshop papers have come into the dossier, for reasons of space and/or individual participant decisions. However, some of the discussion in May 2017 on cultures are either not represented in this dossier (Japan with Mikael Adolphson, Armenia with Johannes Preiser-Kapeller, Byzantium with Iannis Stouraitis) or represented by several speakers at the workshop but by only one in this dossier (Africa with Richard Rathbone; China with James Benn and Georg Lehner; Europe with Hans-Henning Kortüm) have influenced the articles including this introduction. Finally, I thank these colleagues for data and input for the introduction, to which I cannot always give detailed credit, alas. Special mention must however go to Wolfgang Gabbert, who commented several versions of this text.

Disclosure statement

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

Additional information

Funding

Funding for the 2017 Workshop was provided by the Österreischiche Akademie der Wissenschaften (Vienna) and by the University of Vienna’s Historisch-Kulturwissenschaftlich Fakultät, its Institut für Geschichte, its Institut für Österreichische Geschichtsforschung, as well as its research group (FSP) ‘Gemeinschaft, Konflikt und Integration’. The initial editorial work took place at the Central European University’s Institute for Advanced Studies (Budapest, Hungary), whose support is also herewith acknowledged.

Notes

1 Note however that Baker’s (1994) argument musters the concept in the course of a conceptualization of individual events understood as individual (Foucauld’s ‘eventality’), whereas we take it in a more structuralist direction. Furthermore, we focus on just one condition to explore it, whereas ‘no list of conditions of possibility can (…) be considered logically complete’ (Baker 1994, 195).

2 See Asad (1986).

3 That Smith replaced religion with ‘ritual’ leaves him open to his own form of critique, genealogy. See Buc (2001).

4 Strathern (1988, 348) explains: In my own account,

negation is meant to set up a relation between sets of ideas that are, on the one hand, the social constructs of others and, on the other, social constructs as specifically deployed in an analysis not reducible to a homology with these constructs.

5 Definition suggested by James Benn at the 2017 Vienna workshop.

6 Mead 1968 writes, ‘(…) the intention to kill and the willingness to die’ (emphasis mine). However, all warriors did not die willingly. Richter (1992, 38) underscores how in

contrast to European notions that to perish in combat was acceptable and even honourable, Iroquois beliefs as recorded in later eras made death in battle a frightful prospect, though one that must be faced bravely if necessary. The slain, like all who perished violently, were excluded from the villages of the dead, doomed to spend a roving eternity seeking vengeance. Both in capture and in the afterlife, a person taken in combat faced perpetual separation from his family and friends.

I owe this nuance and reference to Wolfgang Gabbert.

7 Peer Vries (2019), in an extended review, intensifies Osterhammel’s point, pleading for grand narratives.

8 One possible exception: Albert of Aachen (2007 vi.21, 430) reports that when Jerusalem was stormed on 15 July 1099, men and women allegedly broke through the gates, helping in the fight. This gender-bending may be owed to the apocalyptic atmosphere or simply to the capture of the Holy City being the climax of the expedition. Note however that while well-informed, Albert was not a participant in the First Crusade.

9 The first information from Richard Reid; the second insight from 2017 Workshop participant Richard Rathbone.

10 Wolfgang Gabbert comments that the definition of the relevant unit(s) has to be considered. In the Precolumbian case, primary loyalty was with the city state (altepetl). Thus, relations to other city-states or the integration into a larger entity, such as the Aztec empire, for example, were always tactical and pragmatic. Claims to loyalty (that is, the expectation of not being betrayed) may be related to the level of integration of the polity(ies) involved.

11 Not to mention the issue of rape perpetrated on men, for which see among a dense recent literature Miranda (2007).

12 Despite scholarly controversies, it seems that Great Plain natives did not rape either. The record for the incidents that most argue for sexual violence, in Minnesota in the 1860s, is (1) tainted, and (2) if there was rape it came as retaliation in a situation of extreme stress (categorical statement in Slotkin 1975, 457; references in Calloway 1983, 189, 202; but see Namias 1993). Heard (1973, 98–101) documents a telling asymmetry in testimonies: other women were raped, I was not.

13 Khadduri (1955, 129–132), for the jurist majority consensus against the enslavement of Muslims. But others have underlined the lack of uniformity in opinions, e.g. Landau-Tasseron (2006), following Abu El Fadl (1999).

14 My thanks to Jocelyn Wogan-Browne (Fordham), who reminded me of the dense presence of sexual metaphors in the West.

15 For Japan, I am grateful for guidance and advice to Mikael Adolphson and Tom Conlan. I take full responsibility for errors and idiocies.

16 Some information on rape (in general, the focus is not on warfare) among Japanese warrior elites in Tonomura (1999).

17 ‘Il y a un lien, une continuité, entre les relations hostiles et la fourniture de prestations réciproques: les échanges sont des guerres pacifiquement résolues, les guerres sont l’issue de transactions malheureuses’. See the comments in Sahlins (2017 [1972, 2003], 166, 281), and the critique in Clastres 1997).

18 The Franks were Chalcedonian Christians, the Ethiopians and Armenians Miaphysite Oriental Christians (who rejected the 451 Council of Chalcedon's definition of Christ’s nature). This divergence in Christology played no role in shaping warfare, as one might a priori expect.

19 As explained by James Benn at the 2017 Vienna Workshop.

20 Chronicle of Saint Martin, Dover, deciphered by Kingsford (1980) 85.

21 In Sub-Saharan West Africa, ‘Divination always preceded warfare as a matter of course as did a variety of forms of rituals that sought the approval and support of ancestors. When possible, one chose terrains and dates for combat because of their propitiousness’, Richard Rathbone, email of 9 December 2018.

22 Buresi is therefore not contrasting two areas of the Islamic world that are vastly distant from one another, as Clifford Geertz’s famous Islam observed (1968). See the critique by Varisco (2005).