Pure and impure risk: Integrating and adapting to landmine danger in Western Sahara’s pasturelands

Abstract

This article explores the multivalent relationship between nomadic pastoralists and unexploded ordinance (UXO) in Western Sahara. There are more than 7 million explosive remnants of war in the disputed territory, largely placed by the Royal Moroccan Army during the 1975–91 war between Morocco and Western Sahara’s pro-independence Polisario Front. Drawing on ethnographic fieldwork carried out in and around Western Sahara, this article considers UXO through the concept of ‘matter out of place’ to argue that the normalisation of UXO risk by herders represents an opportunity to think through debates in environmental anthropology, particularly with regard to the construction of risk. The article argues that, as UXO become integrated into Saharan lifeworlds, the risk associated with them becomes assimilated into locally normative understandings of danger and belonging. It considers ‘purity’ and ‘contamination’ through a temporal lens, arguing for a processual understanding of how UXO risk becomes naturalised as part of desert life.

Introduction

On 27 February 2016, a herder named Shmad Jouly ¹ crossed into, and died in, a restricted military zone in the vicinity of the Berm, a Moroccan defensive wall, in Western Sahara. ² Shmad was a Sahrawi, a member of Western Sahara’s indigenous population, and he was accompanied by – and, according to some accounts, attempting to retrieve – several of his camels. The Berm, arguably the longest military barrier in the world at 2,700 kilometres, is for most of its length a low wall made of sand, punctuated by occasional manned military stations. The Berm bisects the territory of Western Sahara, separating the Moroccan-occupied zone (perhaps about three-quarters of the territory, although exact estimates of territorial control are difficult to assess) from the remaining area, controlled by the Polisario Front, a Sahrawi independence movement. Extending out from the Berm, into the Polisario zone, are extensive minefields, largely planted by the Royal Moroccan Army during the 1975–91 war between Morocco and Polisario. The landmines ³ in the territory probably number at least in the millions, although there is no comprehensive data available on either this or the placement of the mines. For the most part, these minefields are unmarked, or partially marked. In the area where Shmad died, some explosives are indicated by small circles of stones placed around known UXO, but the vast majority of UXO come with no warning at all.

Attempts to reconstruct what happened to Shmad were complicated by a number of factors, not least among them Morocco’s refusal to assist the subsequent United Nations (UN) investigation. Based on a series of interviews with Shmad’s extended family and others with knowledge of the incident, I was able to construct a rough sequence of events, corroborated where possible by UN documentation and contemporaneous Algerian journalism. In February 2016, Shmad and two hired shepherding assistants were herding camels in the vicinity of the Berm, roughly in the area of Mijek, Western Sahara. Grazing camels are not monitored continuously and, while Shmad had gone to visit friends, his camels had apparently approached the Berm, or possibly crossed it. By the time Shmad was able to locate them, several of his camels had been shot. Approaching the Berm in an attempt to recover his camels, Shmad disregarded warnings from nearby Moroccan sentries, who opened fire.

According to Shmad’s two hired assistants, the initial shots were non-fatal, and Shmad, who was shot in the legs, remained alive near the Berm, in an area known to be heavily mined. Attempts by other Sahrawi herders to reach him were hindered by the fact that the mines were unmarked; lacking mine-detection equipment, the herders asked UN ceasefire monitors in Mijek – part of MINURSO, the UN mission in Western Sahara – to assist them in accessing Shmad’s body. The MINURSO team initially refused the request and, by the time a UN mine-cleaning team had reached Shmad’s body, more than 48 hours later, he was long dead of blood loss and exposure.

Later that year, I conducted interviews with several people who had been involved in the incident, and published an article about it. After my article had been published, I visited the Jouly family once again in Smara. Shmad’s father had since passed away, but the rest of the family was still there. The family presented me with a poster to take back with me, with the suggestion that I should hang it in my office. It showed his smiling face, somewhat candid and caught off guard, on which were superimposed the words, in Arabic, ‘The Moroccan occupation kills with bullets and mines’. In Shmad’s case, his death had indeed been the result of a combination of both bullets, the direct cause of his death, and UXO, which had impeded the rescue attempts and therefore played an indirect, but determinative, role in the tragedy. The claim on the poster could therefore be read as asserting that the Moroccan occupation was lethal as a result not simply bullets as well as UXO, but that it could also be understood as a layered interaction of technologies and processes which produced emergent, unpredictable results through their encounters. In fact one could also argue that ‘occupation’ was the key conceptual step in the poster’s assertion, since Shmad had been killed by bullets fired from the Moroccan-occupied zone of Western Sahara, but he had died in the Polisario-controlled area, and that the landmines were therefore not simply a phenomenon of occupation but an aspect of desert life that existed and acted outside of the Moroccan occupation, in both spatial and temporal senses. And ultimately, of course, although Shmad happened to not trigger any landmines at all, their attendant and ambient risk could be said to have played a large role in his death by limiting access to his body by medical personnel.

This aspect of landmines, and UXO in general, is a recurring theme in the academic literature. When unmarked by signs, fences, or clearly identified ‘minefields’, UXO act upon the behaviour of people who live alongside them, often in indirect ways. In this sense, I consider such ‘lost explosives’ as having an imaginative dimension as well as a physical one. In a landmark study of UXO in the Korean Demilitarised Zone (DMZ), the anthropologist Elena Kim refers to landmines as ‘rogue infrastructure’ that ‘lie in wait and ambush their victims’, which, through their ‘volatility as explosives produce psychic effects that render them effective area-denial weapons’ (Kim, 2016). In fact, the uncertainty of where the UXO might be located, particularly in the Sahara, renders them ‘effective area-denial weapons’ in areas well beyond their initial emplacement. In the Sahara, heavy winds continuously reshape sand dunes, sometimes rendering UXO themselves mobile. In recent years, the Sahara’s unpredictable and sometimes extremely heavy rainfall, a result of a changing climate (Lecat, 2016) further moves or otherwise destabilises emplaced UXO. In this sense, UXO in the Sahara are less likely to cordon off territory (though, of course, some areas, particularly around the Berm, are known to be specifically mined) than they are to simply create a risk over a large area that is unpredictable and cannot be mitigated.

In this article, I explore how UXO risk in Western Sahara is constructed and imagined by locals engaged in multiple forms of mobility: among them nomadic herding, long-distance trade, and circulation between the Tindouf refugee camps and northern Mauritania. Although I do not adhere to a singular theoretical frame, I use a phenomenological lens to show how the ambient risk of UXO acts upon desert life, and to explore how such long-term, low-intensity, ambient risks are folded into the experience of everyday mobility via a process of temporal integration in the life of the desert. Where UXO removal technicians often speak of an area as ‘contaminated’ by explosives (a turn of phrase that has led some scholars to note its resonance with Mary Douglas’s ‘matter out of place’), I argue that it is precisely the quotidian, ‘background’ nature of UXO that allows them to become a part of the desert environment, ‘acceding’ as one of many risks faced by those who move through Western Sahara. UXO risk in Western Sahara itself, I argue, follows roughly this phenomenological pattern of division into ‘pure’ risks – those that can be tolerated, ignored, or accepted – and ‘impure’ risks – those that are experienced as novel, intrusive, and unacceptable. The phenomenological construction of UXO risk varies over time and tracks generational change, such that the experience of risk coheres with emic understandings of ‘naturalness’ in the Sahara. In this way, UXO in Western Sahara allow us to think through the boundary between ‘natural’ and ‘unnatural’, and to consider which risks are considered internal or external to the environment over time. This article therefore aims to contribute to the anthropology of UXO, risk, and imaginaries of environmental belonging.

A personal phenomenology of risk

The ‘ambient’ nature of UXO, and their attendant risk, implies to some extent a quotidian phenomenological or experiential dimension to the relation between humans and ‘lost’ explosives. I was (from a certain vantage point) lucky enough to have a direct experience of this ambience during my fieldwork. On a particular day, I was invited to join a Polisario convoy on a trip to visit the Berm. We were to drive from Smara to the nearest Moroccan military station along the Berm, escorting a documentary crew who were planning to make a short film about Sahrawi resistance to the Moroccan occupation. I had spent time in the Liberated Territories before, although I had never actually seen the Berm itself, so I eagerly accepted the invitation.

We drove for several hours in a combination of Hiluxes, military jeeps, and Spanish-built Land Rover-copy Santanas, the latter of which (universally referred to simply as Land Rovers, or by more specific local names; see Berkson and Sulaiman, 2017) have a unique place in Sahrawi society, including a Land Rover dance (Deubel, 2011). We stopped occasionally, notably at a hut where numerous examples of Moroccan UXO were displayed alongside fragments of a Moroccan F-5 jet fighter, presumably shot down during the 1980s. Here the personal, even intimate nature of war was on display: in addition to shreds of the ejection seat were the pilot’s flight boots, torn and weathered. Yet when the convoy was moving, members of the documentary crew took turns to stand up in the back of the Hiluxes, exhilarated by the desert wind and the excitement of an unfamiliar landscape.

Eventually, the convoy slowed, pulled into a semi-circle formation, and stopped. We disembarked, shaking out our legs and elaborately unfolding our head-coverings (lithām) as protection against the sun and dust. As a relative newcomer to the desert, I experienced it variously as startlingly verdant and filled with life, or barren, illegible, and latently hostile.

On this occasion the latter perception prevailed. A few hundred metres away, was the Berm. From this distance it was hardly discernible; I thought that, if I were alone, I probably would not have been able to tell it apart from any other minor geomorphological feature of the desert. As we walked slowly towards it, our group fanned out, leaving sometimes tens of metres between each other. This was not done out of concern for landmines, which we had been warned proliferated around the Berm itself, but rather, I think, as a response to being crammed together in the Land Rovers and Hiluxes for hours. For my part, I was in a contemplative mood, and drifted further from the others, fiddling with my lithām, which I always struggled to tie correctly.

As we came closer to the Berm, it took on a definite shape. A few watchtowers and blocky military buildings were visible. Circles of rocks began to appear in greater concentration. Some of these had been placed around obvious UXO sitting on the surface: a delicately placed, rusted mortar round. Others were empty, and presumably denoted buried landmines, or UXO which had later been removed. Although these circles could be found anywhere in the desert, they existed here in much greater numbers, and reminded of me of ancient sailors interpreting floating logs and leaves as evidence of nearby land. A certain distance from the Berm, our group halted, and a brief discussion was held. We then proceeded more cautiously, and, it seemed to me, more sombrely. Ahead of us, tyre tracks curved towards the Berm; these were left by young Sahrawi men who occasionally, at night, drove close to the Berm and screamed at it – a rite of passage, an act of resistance, and a deliberate self-exposure to landmine risk.

Since I am lighter than a Hilux, I reasoned that it was safest to proceed in the tyre tracks – if they had not triggered the mines, surely neither would I. After a few more metres, lost in thought, I heard a shout from behind me. One of our Polisario guides was gesturing at me to stop, which I did. The documentary crew, roughly parallel to me relative to the Berm, had set up their cameras and were filming. I called back to our Polisario guide that I was in the tyre tracks, and hopefully therefore safe. He disagreed. The mines move, he shouted – blown by wind and circulated by rain, of which there had lately been an unseasonably large amount. It was impossible to count on the mines remaining in place over time, and in any case the trigger mechanisms were largely degraded, meaning they exploded unpredictably. Nervously, knowing that even this was not enough to guarantee safety, I retraced my steps in reverse, backing up until I was parallel with the guide.

I was lucky, of course, not to have triggered any UXO, and grateful for the intervention of the Polisario guide, whose long experience of navigating the risk gave him a much greater sense than I had of where the danger truly began. Yet at the same time, fundamental to the experience was the basic unknowability of UXO. A general guideline – more mines closer to the Berm, fewer mines further away from it – yielded to heuristics and personal comfort. Though it was possible to use detection equipment to identify and mark ‘minefields’, there was no guarantee that these would remain static over time and, in any case, detection equipment was of limited use when it came to scattered, degraded ammunition. The guide’s concern for our safety probably meant that he was likely to err on the side of caution, and the tyre tracks evidenced the fact that Sahrawis do often take much larger risks with UXO, but however low the risk might be, it was never absent, and never quite possible to forget.

This sort of unknowability – of landmines, but also of other desert phenomena – was a recurring theme for my interlocutors. Uncertainty of outcome can be positive or negative; either a good or a bad outcome might come to pass with respect to the desert’s many unpredictable phenomena, most obviously and crucially rainfall (it is for this reason that Sahrawis sometimes refer to themselves as a ‘children of the clouds’, in the sense of following, and being dependent upon, the rain). The precarity of desert life was emphasised by my interlocutors through the phenomenon of resilience, as something that could be inculcated through exposure: my host father, for instance, boasted to me about sending his older son into the deep desert alone in his early twenties, where he was forced to develop self-reliance and problem-solving skills in response to the arbitrary challenges associated with desert life. His son, he told me, later applied these skills to his job as a truck driver in Spain.

Yet resilience was of limited value in the face of mobile UXO, and some risks were simply best avoided. Throughout the Sahara, rock formations and small hills can be found, cliffs that ripple and undulate, outcroppings emerging like craggy islands from the flat gravel of the desert around them. Prior to starting my fieldwork, I was an avid hiker, always eager to climb hills for a view. In the Sahara, initially, I was no different, taking every opportunity to ascend any elevation I could find. Yet as I grew closer to my interlocutors, they repeatedly and consistently expressed astonishment that I would want to climb; my host brother, in particular, could not fathom why I would want to explore the desert on foot. Small hillocks that I would have regarded as barely noteworthy at home were referred to by my interlocutors as ‘death sentences’. And, of course, my interlocutors were ultimately correct. Far from medical care, with no possibility of an evacuation if I fell and injured a leg or ankle, these hills were indeed potentially quite dangerous. I might be able to climb any hill safely, but equally there was a small but tangible chance that I might not, and so I had to treat every hill as a serious risk.

On the other hand, my interlocutors could be quite cavalier about other risks that I considered very serious. Life in the desert involved constant contact with camels, a crucial symbol of nomadic life, economically vital to the desert economy, and a vector for Polisario’s post-war reconstruction of Sahrawi nomadism (Volpato and Howard, 2014). A key part of my fieldwork involved attempting to learn the skills of livestock management alongside my interlocutors, and, while I had a limited amount of success learning to manage sheep and goats, I never made much progress with camels. Realistically, I was afraid of them: camels are huge and unpredictable, and riding them, or even being near them, felt to me like a constant voluntary assumption of massive risk. I thought of myself as probably able to climb dozens of hills without incident, but each interaction with a camel was fraught. Every minute spent on the back of a camel felt like needlessly assumed peril. Yet my small and ageing host father, for instance, raised in the desert but now very much of an urban environment, thought nothing of managing camels much larger than himself, and I witnessed herders roughhousing with, or jumping on camels, and I was told that some men even apparently attempt to ‘surf’ the sand by grabbing a camel’s tail and allowing themselves to be dragged along. All of these felt to me like extraordinary risks, but at the same time, camels were crucial to desert life; there was no way to live in the desert without normalising and accepting the risks of association with camels. ‘Camel risk’, like UXO risk, was very real, and herders frequently suffered injuries in the course of their duties; yet the everyday experience of camel risk was far more casual, and naturalised, than the risk associated with UXO – matter out of place – or hills, which seemed to me the quintessential matter in place, rendered solid in their places by vast stretches of time.

The imagined landmine

The anthropological literature on UXO has tended to focus on precisely this experiential dimension, of ‘living with’ or ‘being alongside’ explosives, or to stress the integration of UXO into the landscapes of human life (Chapman, 2010; Pardo Pedraza and Morales Fontanilla, 2023). DeAngelo and Jones (2019) speak of ‘explosive landscapes’, while Arensen (2022) shows how the language of ‘mined land’ distinguishes between categories of risk in Cambodia. Henig (2012, 2019) explores life alongside ‘military waste’ in Bosnia and Herzegovina, while Zani (2018) develops the concept of ‘bomb ecologies’. For Zani’s interlocutors, the abandoned bombs are explicated with reference to biological phenomena, nicknamed ‘pineapple bombs’ or ‘cucumber bombs’ according to their shape. Buscemi (2021) further distinguishes between ‘living’ and ‘dead’ landmines, while Stanbridge (2011) documents how people living alongside UXO can incorporate them into new forms of agency, for instance by fashioning tools out of the scrap metal left after explosions.

Perhaps the richest theoretical treatment of landmines, as mentioned, is the work of Elena Kim (2016, 2022). Drawing on her work in the Korean DMZ, Kim develops the concept of ‘rogue infrastructure’ to capture the social and ecological entanglements of landmines that outlive the original purposes and contexts of their emplacement. Alert to the risk of ‘anthropomorphic projection, by taking into consideration the mine as an actant’, she calls for ‘ethnographic attention to mines and humans as cohabiting and coconstitutive of a shared ecology’ (2016). Crucially for my argument, although it is not central to her analysis, Kim also notes the ‘background’ nature of landmine risk:

In the three years that I had been traveling to the CCZ [Civilian Control Zone] and neighboring areas with ecological researchers and others, mine warning signs and information notices had rarely warranted a second glance. Especially as I became accustomed to the scenery, they rapidly receded into the background, reduced to visual static that had become as unremarkable to me as they were to my colleagues, who, when I asked them about landmines, would say, ‘oh, yes, there are so many mines around here,’ without further elaboration. ⁴

Mines, in this analysis, are ‘coconstitutive’ of a ‘shared ecology’, but have also, for Kim, ‘receded into the background’. That is, they exist in a temporal dimension as well as a physical one, and the perception of landmine risk can vary over time. And while Kim’s ‘rogue infrastructure’ may scale imprecisely onto the risks and social lives of non-landmine UXO (as in the Saharan context), it is the simultaneously ‘unnatural’ and ‘background’ nature of the attendant risk that holds in tension the landmine’s ambivalent place within the shared ecology.

The perception of landmine risk has been explored fruitfully by Schwenkel (2013). Noting that the boundaries of risk ‘have never been stable … but are culturally contingent’, Schwenkel connects this to the work of Mary Douglas, whose work on ‘risk cultures’ involves ‘shared constructs of danger and threat that reinforce moral boundaries and regulate social conduct’ (2013: 136). Schwenkel further draws out the parallel to Douglas’s Purity and Danger (1966): ‘As abject matter out of place’, Schwenkel writes, ‘explosive remnants of war are an enduring form of “pollution”; legally ambiguous, disruptive, and dangerous, they are the cause and consequence of political, economic, moral, and social disorder’ (2013: para. 22). According to the UNMAS (UN Mine Action Service) (2023), Western Sahara is one of the most UXO-contaminated territories on earth. In fact, such passing references to Douglas’s terminology – UXO as ‘contaminants’, ‘pollutants’, or as the quintessential ‘matter out of place’ (e.g. Kim, 2016) occur frequently in the wider literature, even when Douglas’s work is not explicitly invoked.

If landmines do indeed ‘contaminate’ land, if they are indeed ‘out of place’, it is worth recalling that this ‘out of place-ness’ has always been conceived of as relative, a matter of perspective. Whose place is contaminated, and in what way? The desert is always imagined multiply, by multiple individuals, communities, and actors. In the Saharan context, where, very approximately, spatial usage is gendered – with ‘outdoor’ spaces being coded as more masculine, and ‘indoor’ spaces as more feminine (Porges, 2019) – the individuals overwhelmingly most likely to come into contact with UXO are adult men. A caveat, and an analytic point of entry, therefore, is that desert imaginaries are various, overlapping, and mutually coconstitutive; one person’s contaminant is another’s fact of life. ⁵ As other scholars have pointed out – notably Volpato and Howard (2014) – Polisario’s post-1991 reconstitution of the pastoralist economy of its ‘Liberated Territories’ paralleled other aspects of its nationalist project, most notably its emphasis on ‘de-tribalization’ (Wilson, 2016). Alongside other infrastructural projects, such as the construction of wells and other water access points, the development of community gardens, and a system of buses to facilitate access to education for pastoralist children in Polisario territory, a large-scale process of demining has been undertaken. This is essential to the economic viability of the pasturelands, but also constitutes an assertion, on the part of the Sahrawi nationalist project, of who and what the desert is for. In this sense, the invocation of Douglas is narrowly correct: UXO represent a novel, unfamiliar feature of desert life, and the ‘decontamination’ of the desert represents a process of purification – a return to a time before UXO.

To an extent, this demining work has been enormously successful. Though reliable figures are hard to come by, given the plurality of actors involved, at least tens of thousands of UXO (in this context, a number that is both fairly large and ultimately negligible) have certainly been removed from the Liberated Territories. Insofar as Morocco waged an essentially ecological war against Polisario from 1975 to 1991, deliberately mining pastures, poisoning wells, and massacring livestock, similarly the recovery process has been ecological work, arguably parallel to the Western practice of rewilding; it gestures, too, at the shared imaginary of Saharan landscapes. In the course of this massive clearance effort, the character of the UXO involved have changed; uncertainty increases rather than decreases over time. Although there are fewer and fewer mines (via clearance operations, or explosions, or degradation into non-functionality), the mines that remain are more and more geographically and behaviourally unpredictable.

This is in part because known and marked minefields are easier to clear than hidden ones, but there are characteristics unique to the Saharan context. Because Saharan UXO are of such a variety of types, decay unpredictably, and are themselves mobile, there is no one-size-fits-all clearance procedure. Known and marked UXO, or those placed and then removed in semi-predictable patterns, are more likely to have been cleared, resulting in an extremely unpredictable ‘population’ of remaining UXO, ‘lying in wait’, just as Kim writes. The unique feature of Saharan UXO – their mobility – also changes the experience of their risk. If the UXO are mobile, then in some sense there is no set of precautions at all that one could take in response to them.

This, in turn, complicates arguably the most salient feature of long-lasting UXO: their propensity for spatial closure. The landmines around the Berm were functioning as a Foucauldian panopticon, closing off space via their imaginative dimension. They transferred the burden of governance from state to citizen, in the classical manner of govermentality: since the landmines might be anywhere, they could be everywhere, and therefore the circulating herders were bound to act as if they were everywhere. But beyond the most obvious minefields, such as the area around the Berm, the effect was scattered, distributed, and frayed. And, in turn, if there was nothing to be done, no set of actions one might take to alleviate the risk associated with them, then they were simply part of desert life, a risk to be tolerated, the less thought about the better.

Consider this in light of the death of Shmad Jouly recounted in the introduction. Although Shmad’s direct cause of death was blood loss from a bullet wound, the general opinion in his family was that the wound would have been survivable if appropriate medical treatment could have been promptly accessed. The issue was simply that the presence – or potential presence – of UXO caused a significant delay as the MINURSO team had to clear a route to his body. While the question of whether his death resulted, in part, from UXO-related delays in the MINURSO team’s access to his body is unknowable, it seems clear that the role played by UXO, one way or another, was basically phenomenological. The MINURSO team, unaccustomed to the ambient risk associated with UXO and constrained by operating procedures that did not apply to Sahrawi herders, was bound to act as if the ground might be dangerous, in a way that Shmad, who had walked on that ground in pursuit of his lost camels, was not. For Shmad, and herders like him, UXO had receded into the background and become simply an accepted risk of doing business in the deep desert.

Pure and impure risk

Given the ubiquity of ‘purity and danger’ as an allegory (if not necessarily a theoretical frame) in studies of UXO, it is worth briefly revisiting what Douglas was attempting to do in Purity and Danger: An Analysis of the Concepts of Pollution and Taboo (1966). For Douglas, ‘dirt’ as a categorisation is basically taxonomic, a classificatory heuristic. ‘Dirt’, she writes, ‘is never a unique, isolated event. Where there is dirt there is system’ (1966: 35). Elsewhere, ‘Dirt was created by the differentiating act of mind … a by-product of the creation of order’ (1966: 16). Dirt is a way of describing or adjudicating which things fit into a given world-schematic and which do not. It is also relative: ‘There is no such thing as absolute dirt’, as dirt fundamentally exists ‘only in the eye of the beholder’ (1966: 2). Dirt is not an inherent property of an object, or an objective classification, but rather something relative, contextual, and mutable. ‘What is clean in relation to one thing may be unclean in relation to another’ (1966: 10). Dirt, in this sense, is ambiguous along multiple registers: as a matter of perception, vantage point, and physical or temporal context.

Dirt, and the discursive maintenance of a particular thing as dirty, is socially maintained and reiterated. ‘Purity/impurity discourses act’, writes Duschinsky (2013: 63), ‘as a homeostatic system which ensures preservation of the social whole.’ Dirt is that which transgresses a given social schema, a collective imaginary; it represents that which cannot be integrated into a particular community’s view of the world:

Purity and impurity do more than judge self-identity, however. They can play a fundamental role in its performative construction; they are well adapted for smuggling assumptions into our discourses regarding the essence of particular phenomena and forms of subjectivity, simplifying a complex world into a stark contrast between the dangerous and the innocent, the valuable and the valueless, the necessary and the contingent, the originary and the prosthetic, the real and the apparent, and the unitary and the fragmented. (2013: 63).

If purity and dirt are socially imagined and temporally contingent, we may then consider how the things coded as pure or dirty might have changed over time. As discussed earlier, acculturation to the desert by Sahrawi youth is in part a matter of inculcating skillsets (as in the case of my host father leaving his son in a particular valley for several months to learn self-sufficiency). It is for this reason that the desert is sometimes referred to by Sahrawis as the ‘school of the nomad’. It follows, then, that the nature of those ‘nomadic skillsets’ can fluctuate over time as the desert ecology fluctuates.

In fact, the temporality of the desert ecology, and the nature of its perpetual, if uncertain, change, was a matter of some contestation among my interlocutors. Some considered the desert’s rhythms inviolable – many reported any and all ecological variation (such as rainfall) to be something that would even out in the end, the desert itself being capable – in a literal sense of having agency – of integrating all anthropogenic change into itself, given sufficient time. And if the risk associated with UXO had indeed been phenomenologically rendered ‘pure’, then this facet of the desert imaginary was once again apparent. Simultaneously (and not necessarily in contradiction) the desert was clearly changing in substantial ways, and virtually all the herders I spoke with reported these changes. Asked about climate change, herders from older generations (that is, those with a direct experience of the 1975–91 war with Morocco, and the colonial period before it) tended to express disinterest, or active disbelief, in climate change, regarding it as a future phenomenon, or something confined to the West. Yet many of those same individuals reported ecological changes that I would have been comfortable describing as climate change: declining (or erratic) rainfall over the past twenty years, for instance.

For younger herders (those who began herding after the 1991 ceasefire, largely within Polisario-controlled territories), the desert was temporally disjointed, lacking the sense of steadiness and permanence attributed to it by older herders (Isidoros, 2018; Volpato and Howard, 2014). Accordingly, I develop the concepts of pure risk versus impure risk to denote, on the one hand, risks that are acceptable in the context of knowable, repeatable processes, incurred in the course of regular social and economic life, and therefore unavoidable; and, on the other hand, risks that are constructed as extraordinary, external to regular life, and therefore avoidable. My attempts to climb hills, and the horror with which my interlocutors often reacted, thereby denote an impure risk, since climbing was not a regular part of Saharan life. In the context of this article, UXO in the Sahara shift over time, via their normalisation and integration into desert life, from an impure to a pure risk. Pure risks are those which derive from matter in place; impure risks are those which derive from physical or behavioural contaminants.

The demarcation between these categories is of course not absolute, but it does track an emic reality in which certain risks are perceived as risky, while others are not. While it may be impossible to avoid UXO in general, both older and younger herders avoid obviously mined places along the Berm (though, again, younger Sahrawi men sometimes test minefields along the Berm as feats of daring). Elsewhere, it has been argued that the key factor differentiating normal and abnormal risk is iteration: risks which happen rarely can be accepted, while those repeated thousands of times in the course of daily life require some sort of mitigation. Diamond (2012) reports a similar argument in Papua New Guinea; while he happily camped under dead or fragile trees in the course of fieldwork on the grounds that the odds of any given tree collapsing were low, his interlocutors scrupulously avoided doing so because, over a lifetime of camping, the risk would elevate substantially. The perception of a risk that is ambiguous but normalised differs from that of a risk that is ambiguous but intrusive, and the ambiguous ‘background hum’ of which Elena Kim writes is audible or not audible partly as a function of this process of risk normalisation. Ambiguity is therefore not quite sufficient to describe what was actually happening when my interlocutors cared, or did not care, about the presence of UXO in their pastures.

The culturally contingent construction of ‘purity’ and ‘impurity’ in the context of risk naturally calls to mind other of Douglas’s core works on risk and contigency (Douglas, 1992; Douglas and Wildavsky, 1982). Douglas’s cultural theory of risk builds on her thinking around taboo in order to elaborate how an accretion of social norms can, over time, condition which of a given set of risks are prioritised and which are relegated to ‘background’ status. A genealogy of this formulation is present in the borrowing of concepts from spatial anthropology – in the work of Bourdieu (1977) and Lefebvre (1974) in particular – in more recent studies of living with UXO risk. Operative here is the idea that habituation is itself culturally conditioned – or that, if the passage of time is key in when people become habituated to risk, this is not in itself a sufficient explanation of how this process happens. To bridge this gap, a mechanism of action is required.

Habituation is akin to naturalisation, or the purification, over time, of that which was previously constructed as a contaminant. It can also happen in reverse. Out in the desert one day with a number of Sahrawis, I helped to make a lunch of fried eggs on a portable camping stove. As we were cooking in the shade of an acacia tree, the wind picked up rapidly, carrying with it the usual wave of Saharan sand. Immediately our food was coated in a thin layer of greyish-orange dust. The older Sahrawis implacably continued preparing the meal, and ate it without apparent discomfort; but several younger people in our group refused the food, unwilling to eat the sand, considering it to be contaminated. Teasing quickly broke out, with the older men calling the younger ones soft, urbanised, unaccustomed to desert life. Whether or not the sand was a contaminant depended upon one’s point of view, the regularity with which one had been exposed to sandy food.

‘Pure’ risks, those necessarily entailed in the course of nomadic life, were in some sense not coded as risks at all. As Sahrawi herding practices change over time (Volpato and Howard, 2014; Wilson, 2016), the social world of herding changes as well, and older and younger herders – approximately speaking, those socialised into herding life before and after the war with Morocco – construct different purities and different contaminants. Younger herders, acculturated in the postwar environment, experienced different risks; my host brother’s friends often drove their trucks perilously close to the Berm as a dare, speeding through minefields older herders would avoid even on foot. Equally, they were less fluid and fluent around camels. My host brother, whose knowledge of camels far exceeded my own, tended to refer to them by a single noun, while older Sahrawis, such as my host father, used a much wider and subtler array of nouns to differentiate camels by age, sex, colour, and so forth.

More recently, with the rekindling of active hostilities with Morocco in late 2020, a new threat emerged in the Liberated Territories: Moroccan military drones began firing occasional missiles towards Sahrawi vehicles. Ostensibly, these missiles were aimed at Polisario military targets, but Sahrawi social media and WhatsApp groups quickly filled with reports of civilian vehicles (including sometimes Mauritanian civilian vehicles) destroyed by drone-launched explosives, often accompanied with grisly photos. Like UXO, hidden beneath the ground, the drones were largely invisible, and therefore impervious to most mitigating strategies. As with UXO, the drones sometimes targeted livestock, the economic basis of nomadic life. Anonymous, sky-bound, hyper-modern: they were, in a sense, beyond agency and, above all, they were a novel feature of desert life. The drones therefore represented a deeply ‘impure’ category of risk.

Acceeding to the natural

Back in Tindouf, in the central administrative town of Rabouni, I spent several days with ASAVIM (Asociación Saharaui de Víctimas de Minas). This Sahrawi-run organisation is unique among UXO-focused efforts insofar as its focus is not the explosives, but those who have survived explosions; it aims to rehabilitate victims of UXO incidents, often by providing support, new employment, and other opportunities Headquartered in a low building just off the main road, ASAVIM’s offices present a somewhat jarring introduction to UXO risk. Its entranceway is lined with flowerpots in which artificial limbs are placed, appearing almost to sprout. One room is filled with pictures of post-explosion scenes: maimed livestock, burned-out vehicles, gruesome human injuries.

Daha Bulahi, one of ASAVIM’s leaders, walked me through the process. Demining in Western Sahara falls into two broad categories. First, there are ‘expert’ deminers, often funded by external non-governmental organisations (NGOs) or UN agencies, who engage in a highly technical process of explosive removal, often conducted remotely or in heavy armour and with the use of advanced demining technology – this was the process used by the UN team that carved a safe path through the minefield to access Shmad’s body. Then there is ‘vernacular’ demining, of the type in which Daha was experienced. Upon locating explosive material in the desert, Daha and his team – and, the implication went, Sahrawi herders on their own, all the time – would dig, by hand, under the material, so as to avoid trigging the explosive plate on top. Grabbing the material from underneath, the deminer would stand and throw it backwards over the shoulder, letting it explode in the air or on the ground, hopefully a safe distance away. I was sceptical of this technique, not least because Daha was missing one eye and part of his hand. Then again, out in the desert, far from the access of formalised demining teams, there were really no other options.

Daha was full of stories of nomads and UXO, the implications of which all seemed to carry a valence of regrettable, unavoidable tragedy, and often of human ingenuity in the face of it – the bravery of the vernacular deminers, for instance, or the fortitude of nomads who, having stepped on a mine, lit fires to cauterise their own wounds. It was hard not to notice that these stories were framed not only as injustices perpetrated by Morocco (as was usual for stories of interactions with the Berm or Moroccan security forces) but also, on some level, as things that simply happened to happen, and to which the proper response was pride in the resourcefulness of Sahrawis living and working in the deep desert. No longer able to actively engage in demining due to his injuries, Daha carried out a different kind of work with ASAVIM. Daha drove me around Rabouni and one of the refugee camps, showing me ASAVIM’s projects. Often these involved planting gardens or donating livestock to survivors of UXO incidents, a new way of extracting agency from, and connecting desert lifeways to, the consequences of the explosions. In Daha’s telling, the work represented a kind of full circle: the contaminated desert, the wronged herders, and now the creation of new ways of living in, relating to, and making a living from, the desert.

The connection between UXO and livestock, too, was unavoidable, and provided another way of understanding the integration of UXO into desert lifeworlds. Alongside the ubiquitous imagery of burned-out trucks – blackened, the windows blown out and the doors blown open – a consistent visual signifier of UXO risk on both social media and on ASAVIM’s walls was the maimed camel, either dead and prostrate on the ground, or injured, hobbled, bloodied. Camels represent more than a livelihood – they are a symbol and vector of Sahrawi identity in the desert. They also present a unique challenge to ASAVIM and other UXO-resilience programming, since, as I was repeatedly assured by herders in Smara, ‘Camels can’t read’ (Porges, 2018). Insofar as nomadic land-use is itself an assertion of Sahrawi nationalism, and insofar as camels represent the most concrete usage to instantiate that nationalist vision, the ongoing UXO threat to livestock is itself a mechanism by which the Moroccan state complicates and obstructs Polisario’s national project.

Yet this challenge can itself become normalised, moving from a political to a ‘merely’ ecological landscape of risk. For younger herders in Western Sahara, the presence of UXO need no longer seem so unusual. Unlike older generations, who remember a time both prior to UXO proliferation and the immediate aftermath of adaptation to the novel risk, younger herders have grown up incorporating UXO risk into their understanding of the desert. Stories of maimed livestock, burned-out vehicles, and occasional fatal incidents, lose their power to shock, and simply become part of the inherited reality of using the desert. Again and again, my younger interlocutors spoke of UXO risk as simply another phenomenon to account for in considering grazing routes and travel patterns. In this sense, habituation transfers risk not only from ‘impure’ to ‘pure’, but also from ‘political’ to ‘natural’.

Missing from some past studies of risk is an explication of the mechanism by which a risk can become normalised. While it is obviously the case that risk normalisation takes place over time, the passage of time itself is not a sufficient explanation. Instead, we might borrow from spatial anthropology to understand the ways in which a particular habitus consists of precisely those vectors of environmental presence which can be materially ‘used’ or engaged with (following e.g. Ingold, 2011). As the set of tasks required to sustain life in the desert itself changes over time, so too will the language (as in the example of camel vocabulary given above) and behaviours which come to seem familiar.

This, in turn, problematises the question of what it means to ‘belong’ in the desert at all. Camels, as potent a symbol of Saharan indigeneity and belonging as they have become, have become normalised and integrated into the fabric of the desert ecosystem through centuries of reiterated life and death. The integration of new technologies into Sahrawi society are often accompanied through reference to earlier ecological phenomena – so Land Rovers, sometimes referred to as the ‘sister of the Sahrawi people’, are categorised by their resemblance to different types of livestock (Berkson and Sulaiman, 2017), while the names of mountains themselves encode the names of earlier, extirpated desert species. The landscape, and the set of ecological phenomena which ‘belong’ in it, are constantly in the process of being reimagined, along an iterated, processual set of pathways, dynamic rather than static, fluid rather than ordained.

‘Impure’ risks are often met with specific countermeasures. They demand action: moving a herd, marking a zone of danger, rallying communal efforts or institutional support to mitigate the threat. In the case of UXO, this was indeed the early response: an irregular, unpredictable hazard required a solution, and the UN, NGO, and community-led efforts to clear UXO, mark known minefields, and disseminate UXO education represented just such an effort to mitigate the threat. But if ‘impure’ risks are those against which specific countermeasures can be taken, ‘pure’ risks, inherent in and constitutive of the desert environment, are those which must simply be tolerated. Thus the emergence of ASAVIM, which seeks to support victims without taking action against the UXO, while younger herders simply learn to live with the risk. For the new generation of Sahrawis, UXO, though unpleasant and intrusive, are something to be endured, tolerated, adapted to, and, eventually, at least partly ignored. The initially extraordinary nature of UXO becomes, over time, mundane, quietly embedding itself in the rhythms of everyday life and in a sense moving beyond the realm of human agency – either to emplace or to resist.

The trajectory of UXO risk in the desert mirrors a broader phenomenon: the normalisation of crises in the face of climate change. Like UXO, the impacts of climate change in and around Western Sahara – rising temperatures, desertification, erratic rainfall – began as intrusive, destabilising events that demanded urgent attention. For Sahrawis in both the refugee camps and in the pastures of Tindouf and the Liberated Territories, these changes exacerbate existing vulnerabilities – including via increased UXO risk. Over time, however, the cumulative impact of climate change, like UXO, shifts from a disruptive force to a familiar backdrop, reclassified as an inescapable part of life in the desert. What begins as a crisis is absorbed into the everyday, no longer seen as something to be actively resisted but rather as another hazard to navigate in an already hostile environment. A specific event – the Western Sahara War – had evolved into a condition of life. During my later visits to Smara, heavy, unseasonable rainfall began to flood the camps on an annual basis. The first floodings, in 2016, were an acute crisis requiring an influx of aid (Lecat, 2016), as traditional mud-brick houses dissolved in the rain and thousands of Sahrawi refugees were rendered homeless. Yet adaptation quickly followed; the annual rainfall and frequent flooding could be partly mitigated through the use of newer, concrete structures. This creeping normalisation represents not solely a material adaptation, but an ideological one. More than any specific risk-management strategy, it was the erosion of precaution in general that tracked the change of a category of risk from ‘impure’ to ‘pure’, as acute crises acceded to the category of natural and, ultimately, normal.

Conclusion

The integration of UXO into the Sahara’s ecology draws attention to the shared imaginary of the ecology itself, and the ways in which ‘naturalness’ is a contingent, evolving set of assumptions about which phenomena ‘belong’ in the desert. While all boundaries between ‘human’ and ‘natural’ are, to an extent, contingent and culturally determined, such distinctions are not fixed, and can evolve over time. This article has argued that one way to track the naturalisation – or, as I put it, ‘accession to the natural’ – of desert phenomena is through an understanding of the relationship between these phenomena and subjectively (if collectively) adjudicated perceptions of risk. Following the terminology of Mary Douglas, and, to an extent, her theorizations of emplacement and risk, ‘pure’ risks might be considered as those which are necessarily adopted in the pursuit of a ‘natural’ lifeway – examples in the West might include the risk of vehicle accidents, or heart disease and cancer associated with moderate alcohol consumption. ‘Impure’ risks are those which are experienced as unusual, intrusive, or unnatural – those which relate to material or phenomena that are out of place. While I do not argue for a static or impermeable reification of this apparently binary contradistinction, I do argue that it holds value as a way of thinking through emic perceptions of risk variance over time.

This phrasing borrows deliberately from the work of Douglas (1966), which explores purity and impurity in the context of taboo. Douglas argues that the coding of cleanliness/uncleanliness is itself a contingently constructed boundary, much like the nature/culture boundary which has been so widely theorised and critiqued. In borrowing from Douglas, I argue that the purity or impurity of a given risk can evolve over time (as in the case of successive generations of Saharan herders experiencing similar risks in different ways) and determine which risks are taken for granted, ignored, or coded as unavoidable. Missing from past studies of UXO risk has been a theory of how, why, and when UXO risk becomes part of the background, fading from an inescapable intrusion to an ignorable fact of life. I have argued that this process is best thought of as a gradual, collective imaginative transition – manifesting in vocabulary change, land-use practices, and narratives of danger – in which iterated risks are rendered ‘natural’ and relegated to the background.

UXO in the Sahara appear to be the quintessential ‘matter out of place’, and describing them in this way is a classic construction in anthropological studies of UXO. I argue that a more fruitful avenue of inquiry would ask how matter which is out of place is gradually rendered in place, and therefore ‘fades into the background’ (another recurring construction) to become simply another latent, ultimately invisible risk. Risk therefore becomes a key to understanding how the nature/culture boundary (or any other such boundary) can evolve over time, how phenomena are ‘naturalised’, and how ostensibly ‘out of place’ phenomena – from the introduction of camels to the Sahara to the placement of UXO in the 1970s and 1980s – come to be an accepted, if ultimately risky, facet of an imaginary of ecological belonging that is neither timeless nor static, but rather constantly evolving via both human and non-human interventions in the broader desert ecosystem.