Audializing migrant bodies: Sound and security at the border

Introduction

Recent scholarship in security studies and critical international relations theory has drawn attention to the dematerialization of borders, homing in on new technologies and modes of control that point to a paradigm shift in the ways in which we govern populations and define the external limits of sovereign territory (Amilhat-Szary and Giraut, 2015; Côté-Boucher, 2008; Squire, 2011; Walters, 2006). In this line of research, the telos of border-making shifts from concrete walls, checkpoints, and fences to abstract institutions such as passports, visas, and deportation regimes (De Genova and Peutz, 2010; Dijstelbloem and Broeders, 2015; McKeown, 2008). Moving beyond transnational legislation or intelligence sharing, the site of the border shifts again, to an intimate scale, in scholarly treatments of bordering practices undertaken by security agencies and the bordering behaviors of migrants or tourists. In this research, bodies and disaggregated body parts are reimagined as spaces upon which borders may be read and inscribed. Individual behaviors are intercepted at points where they interface with networked technologies and fed as data into pattern-finding algorithms, where they are assessed for their riskiness against an imaginary behavioral norm. This ‘corporeal turn’ has been studied in terms of emergent border technologies ranging from face recognition and body scanning, to DNA and fingerprint databases, to new forms of visuality made possible by drones and satellites (Adey, 2009; Amoore, 2006, 2009; Dillon and Lobo-Guerrero, 2008; Marlin-Bennett, 2013; Salter, 2006).

It is surprising, then, that this phased shift from ‘line in the sand’ (Parker and Vaughan-Williams, 2012) conceptions of borders toward corporeality has not more fully engaged with security practices as a domain that may be syncretically sensed, or with the ways in which human senses have been dissected alongside bodies. This is all the more the case when we consider the omnipresence of surveillance – itself a form of sensing – in logics undergirding the construction of securitized boundaries. Such lacunae stem in part from the uncritical reliance security scholars (and political scientists more broadly) place on visual epistemologies. Sight, in this mode of thinking, is intricately bound up with conceptions of truth and knowledge. This structure of thought is so naturalized that it is embedded in language – nouns such as ‘insights’, ‘observations’, ‘perspective’, and ‘transparency’ sit alongside such verbs as to ‘enlighten’, ‘examine’, ‘unveil’, ‘shed light’, ‘reveal’, ‘target’, and ‘zoom in’. All bespeak the close relationship between seeing and the production of knowledge, leading some to go so far as to categorize sight as primary and relegate hearing, smell, touch, and taste to a category of the lesser senses. ¹

Within literature on biopolitics and the corporeality of security, this tendency is perhaps best encapsulated in Louise Amoore and Alexandra Hall’s (2009) work on bodies at the border, in which the authors demonstrate how contemporary security practices of biometrics, data mining, and profiling trace their origins to Renaissance and Enlightenment efforts to visualize unknown certainties in the fields of aesthetics and medical science. Amoore and Hall reveal how today’s technologies, such as the Backscatter X-ray machine found in airports across the United States, the European Union, and the United Kingdom, dissect the body into component parts that are reassembled and sanitized for viewing by airport security officials. It is these processes of abstraction and recomposition, the authors argue, that move contemporary security practices beyond the domain of surveillance into a future-oriented domain of ‘visualization’. Such practices of visualization exceed straightforward surveillance, they posit, because of the underlying presumption that visualization will uncover knowledge beyond the material body, piercing ‘the coverings of dress, status, or feigned identity in order to reveal something of the unknown future hidden within’ (Amoore and Hall, 2009: 448).

The claims that Amoore and Hall make regarding the security establishment’s intention to suss out man’s true nature might be more readily analyzed in a framework that takes sound and the sonic body into consideration. The strong connection between voice, agency, and human subjectivity (consider the political implications of ‘having a voice’, for example) renders a sonic awareness critical to any investigation into predictive behaviors and the unknown certainties presumed to lie latent in the human body. Yet, even as critical surveillance studies have given us insight into what is at stake when visual techniques of control are applied vis-a-vis privacy and liberty, we have less experience thinking about the potential repercussions of mechanisms that target the auditory realm because we fail to perceive banal, ambient sound as inherently political. We have yet to develop a clear understanding of the ways in which our sonic footprint and practices interact with the complex, networked security mechanisms and data repositories that enframe material sound in the contemporary period.

Although ephemeral sound, already invisible, would seem to escape many security practices, today’s audio recording and playback technologies make it possible to objectify voice and the hearing ear, much as one would an iris or a fingerprint. It is not only recent technological developments that drive the need to examine sound. Sound scholar Jonathan Sterne (2003: 2–3) has pointed out that ‘as there was an Enlightenment, so too was there an Ensoniment’. Sterne’s description of how hearing and sound were constructed as objects and domains of thought between 1740 and 1925 demonstrates how hearing came to be imagined as a physiological process based on physics, biology, and mechanics. He argues that people harnessed their powers of auditory perception in the service of rationality and control, as they measured, isolated, and simulated sound and hearing. Yet, in literature on security today, this process has gone largely unexamined.

Mark Salter’s (2006) investigation into borders, bodies, and biopolitics, which explores the interface between government and the individual body, seems, at first blush, to indicate an attention to the auditory. Drawing on Michel Foucault, it promisingly speaks of an international order driven by the globalization of a documentary, biometric, and ‘confessionary regime’. In Salter’s formulation, oral declarations of intent upon entering a country serve as confessions – the first acts of obedience establishing the fundamental relationship between sovereign and subject. Salter describes the security negotiations that take place at borders as a ‘dialogue’ between body and body politic, requiring the confession of all manner of corporeal, economic, and social information (Salter, 2006: 170). He invokes Jacques Derrida on language, and cites the difficulty a supplicant faces in requesting, in a language other than his own, rights and asylum. In Salter’s conception, language is multivalent, serving as a mode of cultural expression, a system of bureaucratic thought, and a vocal, performed exchange between two parties. Yet, despite the patently audial rites of passage that form the basis of his theoretical framework on borders, his subsequent argument about the construction of an international biopolitical order and the emergence of a global visa regime stems overwhelmingly from sight-specific or even non-sensory evidence. Thus, even when words themselves (confession, dialogue, language) seem to entreat a specifically sounded engagement with security and bordering, visual modes of thought remain difficult to evade. The hegemonic epistemology of the visual that permeates critical security studies has tended to collapse sensory data into a framework of visuality, occluding, paradoxically, alternative possibilities for knowing. This article suggests that a case for audializing the body is overdue because of the ways in which sound is already harnessed by powerful actors in the security sphere.

The sonic reimagining I propose serves as an important corrective to several theoretical frameworks: thinking about voice and hearing as bodily processes reminds us that political and economic systems, no matter how global and institutionalized they may be, ultimately target flesh-and-blood bodies. International relations ‘happens to’ the sensory body, even when political science prefers to pay attention to geopolitics at the scale of the state. As an object of analysis, sound is situated contemporaneously in individuated bodies, in capitalist technologies, in global communication networks, and in security paradigms, uniquely positioning it to draw attention to the ‘multiscalar’ production of borders (Laine, 2016) and security. Attending to the way in which material sound traverses infrastructures and bodies also attunes us to the linkages that make it possible to jump scales, bringing these elements together in recombinable assemblages. Inherently spatial in that its waves propagate omnidirectionally through matter, sound draws attention to place, lived experience, affect, temporality, and the tangible ways in which these elements interact and connect with global flows. With sound, we move Chris Rumsford’s (2012) call for the ‘multiperspectival’ study of borders into a multisensory register. Finally, at the theoretical level, sound’s immersive, multidirectional nature invites inquiry into workings of power that overflow extant paradigms and stretch beyond Foucaultian linear visions of panoptic discipline, reimagining space as a four-dimensional panaudicon instead. Thus, if we understand the panopticon as a system of control in which subjects discipline themselves in anticipation of being observed, the panaudicon, which both listens and speaks, creates a cartography of sounded influence that at times collects bodily sounds, and at times penetrates bodies with external sound in the interest of power. Elements of the panaudicon are recombinable, and thus the power it effects assumes multiple forms – disciplinary, but also productive, structural, coercive, institutional, and affective.

Sound, of course, has its own perspectival limitations, and my emphasis on aurality should not suggest that I am inattentive to the deeply synesthetic nature of knowing and perceiving. The object of this article, however, is to demonstrate how a sustained attention to the auditory realm will uncover security dynamics that might otherwise be left unheard. In doing so, I lay the groundwork for a research agenda that understands sound (and by extension the sensory body more broadly) as a mechanism of political power. The case study below speaks to the intellectual payoff of such an intervention for ongoing work on border security and human migration and draws attention to how a politics of sound recognizes human voice and hearing as important sites of critical inquiry and moral/ethical negotiation.

In what follows, I begin by laying out a general theoretical framework for sound-power, setting out core concepts by which the theory can be deployed across a variety of contexts. The framework I propose exceeds the empirical material that I treat in the case study. In my focus on security and migration, I tune in to a particular sound-power frequency, attending to the sonic–digital assemblage that emerges at the nexus of voice, hearing, and technologized sonic practice. This treatment represents only a small portion of the range of research that my framework makes possible, with the trade-off that it allows for a deeper focus on migration and bordering in a specific context. My discussion of sonic habits among migrants derives from fieldwork in 2015 and 2016 where I conducted formal and informal interviews with undocumented sub-Saharan African migrants in Takadoum, a slum on the outskirts of Rabat, Morocco, and in unofficial forest encampments outside of Nador, along Morocco’s north coast. Individuals in my data pool were all trying to reach Europe – either via boat or by scaling the treacherous and highly guarded fences separating the Spanish enclaves of Ceuta and Melilla from Morocco. I interweave their anonymized testimonies with discussions of ongoing state and commercial efforts to recapture and isolate voice and the sonic body. I conclude with a discussion of the ways in which security practices audialize the human body.

Toward a theory of sound-power

Sound enfolds us, saturates our bodies, and constitutes a critical part of our sensory map of the world, yet its ubiquity and ephemerality make it a particularly challenging object of analysis. In thinking inductively about what might constitute data in an investigation into sound-as-power, I sought a means by which to organize the diversity of mundane, everyday sound-events that comprise the empirical soundscape of migration at diverse junctures. In order to address these systematically, and building on Steve Goodman’s (2012) notion of sound as ‘vibrational force’, I break sound in into four classes based on frequency and anthropocentric perceptibility. Depending on age, health, and genetic disposition, human ears are tuned to pick up vibrations ranging from approximately 20 to 20,000 hertz (Hz), or cycles per second. Thus positioned at the interface of physiological reception and vibratory signal, this range constitutes the frequency of audible sound.

Yet sound as vibration extends beyond the capacity of human ears. Occurring at frequencies below 20 Hz, vibratory infrasound falls below the typical spectrum of human audibility. Humans often perceive infrasound as a physical rumbling that accompanies powerful noises such as thunder or jet engines. In a military context, infrasound limns the shape and extent of inaccessible spaces such as caves or tunnels, reducing human risk in the information-gathering process. Ultrasound, at the high end of the spectrum, is equally inaudible to humans. Like infrasonics, ultrasound is used to map space – in this case by radiologists and sonographers in the production of images of the human body. Together, these ‘inaudible’ vibrations constitute important sources of sound-power, with broad application in military, medical, and other technologies. Scholars disagree on whether silence exists in any absolute sense (assuming the presence of a gravitational atmosphere in which sound waves may propagate); yet, like inaudible sound, it remains useful, beyond metaphor, as an analytical category. Politics tends to conceive of silence as an absence of agency, but it may equally constitute a mechanism of power; consider the affective impact of the ‘silent’ Standing Man political protests that swept parts of the Middle East in 2013, or the symbolic and actual silencing that occurred in Calais in 2016 when migrants sewed their lips shut to protest eviction from a refugee camp.

Frequency represents one way of classing sound; a more quotidian mode of differentiating among sounds is on the basis of loudness. In terms of the physiology of ears, the human threshold for pain and the attendant damage to the aural mechanism begins at approximately 110 decibels (dB). This threshold and the politics that surround it in relation to the use of music as torture, sonic weaponry in war, or urban noise deleterious to health mark important and always evolving sites of political contention. Thus, sounds demarcated as ‘hazardous’ in this way stand apart from others in the sense that they are not delimited by frequency per se, but rather by an anatomical threshold triggered by decibel levels.

Thus disaggregated and conceptualized as material force, sound may be grouped into four classes as audible, inaudible, silent, and hazardous so as to strengthen and isolate its analytical purchase. Understanding sound as a vibrational force that affects matter around it highlights not only sound’s function as an actant in its own right, ² but also sound’s positionality within the network of power relationships that envelop it, produce it, and are impacted by it. Sound may effect power at a coercive level, as a tool of brute force, as when security personnel rely on sonic weaponry such as the Long Range Acoustic Device, or sound cannon, to create distance between the producer and the intended target. It may physically delimit mobility and territory at the ‘sensory border’ when infrasonic listening devices pick up movement and sound alarms to alert guards. Isolating sound in this way, and tuning in to its specific effects, enables us to identify a causal agent that is often missed. It also allows us to question how, precisely, sound functions as power. ‘Power’ is famously a contested concept; yet, however power is conceived, and whichever face it shows, sound is likely to be a part of the sociopolitical relationship under examination. This theorization offers a roadmap by which to begin to assess its relevance.

This framework for conceptualizing sound as power has application for the study of politics that I can only gesture toward here and that I treat more fully in other work (Weitzel, forthcoming). My fieldwork in Morocco centered on audible sounds experienced by migrants’ during their journey or at the point of the securitized border. Thus, many of the examples and sound-events that follow in my case study originate in voice. Voice requires additional theorization as a distinct subset of audible sound because of the way it enfolds communication technologies and language into its sounded presence. As a unique subsection of audial sound, voice represents a step away from ‘banal’ sounds that may be perceived in a given soundscape. Theorizing voice as a form of sound-power, however, enables us to move beyond discourse analysis or communications studies approaches, and to open new ways of thinking about the role spoken language plays in migration and security. Linguistic imperialism has often subjugated the sonic to the semiotic register (Goodman, 2012: 82), with the result that voice’s specifically sonic attributes, as well as the entanglement of material, biological sound in the infrastructures that convey voice and enable hearing, are often disregarded. Approaching voice with a sonic sensibility rectifies that without diminishing the obvious importance of linguistic communication.

In addition to being entangled in language, voice is always mediated, even as mediating communication technologies – while not exclusively sound specific – are always already implicated in the sonic body. Sound reproduction, transmission, broadcasting, and amplification all draw attention to the powers and possibilities of voice separated from its ‘original’ body, produced at least partially through non-human sources such as fiber-optic IT cables or cell-phone speakers. A rigorous tracing of the sonic in these networks allows us to understand how sound ‘audializes’ the human body, working analogously to sight in Amoore and Hall’s (2009) reading. Voice, hearing, and sonic practices, alongside the mediating technologies that enable them in today’s globalized world, come together as a man–machine assemblage that, at its core, is highly political, not only in the way in which it is constructed and accessed, but also in how this assemblage is harnessed in support of mobility and security. Conceptualizing voice, hearing, and technologized sonic practice as a distinct unit – a Deleuzian assemblage – creates space for rethinking the relationship between migrant subjectivity and identity claims, as well as the sound-based security politics that strive to penetrate them. In the process, we identify new actants in the political sphere and strive to develop the critical listening skills and vocabulary necessary to combat sound-power when it is wielded against us. Using recent Western Mediterranean migration as a case study and springboard for discussion, the article proceeds by examining this sonic–digital assemblage and the modes of knowledge it produces.

Digital–sonic assemblages: Crossing the Sahara

The sonic body, the technologies and infrastructures that enable it, and the quotidian human practices that animate the relationship between these component parts constitute a man–machine digital assemblage and useful site of inquiry. For migrants traversing the western Sahara, this sonic assemblage structures the politics of the desert crossing as well as individual decisionmaking along the journey. It is telling that the stretch of greatest physical danger and emotional trauma for the migrants I interviewed was the vast sandscape that constitutes the Malian border with Algeria – the section of the trip in which cell-phone coverage was non-existent and migrants could not project beyond their immediate environment for help. In Mali, licensed mobile network coverage was limited to urban spaces and extended to only 35% of the land area in 2015. ³ The paucity of IT infrastructure effectively puts most of Mali out of touch, literally, with the global and economic assemblages that serve to bring voice communication into a single operating system connected to digital networks and surveillance regimes. Whereas migrants could reach out to request emergency financial assistance, advice, or emotional support along other legs of the journey, during this segment of the trip their voices were plunged into a structural silence, in which ambient noises and near voices were their only source of audial knowledge, and immediate contacts their only recourse for support.

When one component of the tripartite assemblage goes faulty or is missing, as in Mali’s border zone, relationships of power that are otherwise naturalized are thrown into high relief, and the resultant flux creates possibilities for a transformed politics. A lack of connectivity does not mean these spaces are diminished as sites of complex global politics. On the contrary, such ‘silent’ zones are often uniquely significant in terms of global security – precisely because they are so illegible. In these spaces, voice and hearing flow into divergent technologies, and unorthodox sonic practices emerge. In particular, migrants cited the role that Thuraya satellite phones played in their overland journey. Bypassing conventional terrestrial conduits, Thurayas link to a constellation of Low Earth Orbit satellites to serve as surrogate technology where large-scale infrastructural development is deemed cost-ineffective. Structural factors constraining access to this alternative communications system confer power upon a distinct set of actors. With the company’s shareholders including private US and UK entities, Thuraya Corporation’s satellite-based phones effectively enable communication outside of cell-tower-based network systems, turning off-the-grid spaces of the Sahara into an information-rich zone for those few that can afford it.

None of the migrants reported having access to a satellite phone. Instead, they claimed that ‘only Touaregs have Thurayas’, in reference to the semi-nomadic peoples who live in the region and have parlayed the trafficking of sub-Saharan Africans into a cottage industry. These traffickers capitalize on the absence of conventional telecom infrastructure to gain leverage vis-a-vis the migrants who cannot afford satellite phones. The border under consideration is particularly thorny for many West African migrants because it often marks their first illegal crossing, taking them out of the Economic Community of West African States, where the protocol of free movement allows citizens to travel without a visa within the region. Migrants generally make the passage in crowded and unreliable trucks. Human suffering is legendary; at the mercy of the drivers and extorted for money along the way, migrants are injured, starved, violently raped, and abandoned in the desert to die. The other set of actors in this region with sufficient wherewithal to afford Thurayas is the Algerian military and border police, as the narrative below poignantly reveals.

Abubakr, from Cameroon, felt exceptionally lucky to have crossed the desert in a single 13-hour trip (Interview K, 2016). He joined 30 migrants of mixed nationalities in a new truck, which meant the engine and wheels were in good condition and the Touareg driver could travel more or less continually after departing from Gao in Eastern Mali. Abubakr reported the driver used a Thuraya to touch base with other smugglers about the location of police in the desert. This enabled him to alter routes and avoid confrontations, which minimized driving time. Yet, after encountering what appeared to be unmarked ‘checkpoints’ in the middle of nowhere, Abubakr realized that the driver was also in touch with Algerian police themselves. At intervals, the man would call policemen to the site of ‘the border’; these would extract payment from the migrants, but also would allow the truck to pass without extended delay. Here, audial pre-knowledge and the ability to recognize a voice without semantic identifiers served as a sonic ‘handshake’ – a practice that confirmed identity, allowing the driver to discriminate between collaborating and unknown policemen, with important results vis-a-vis trip duration and the extortion of migrant money. The entire journey was organized in real time, verbally, in a zone where regular cell phones could not operate.

Ownership of the material trappings of communication begets control over access and makes possible the privileging of some population subsets and the silencing of others. At this border, sonic practices, coupled with the irregular infrastructures that suppress certain voices, produce a mode of ‘live governance’ and flexibility that traditional, territorial states aspire to achieve. This sonic assemblage produces an alternative kind of sovereignty – one that is based on speed and flexibility and has little in common with the sovereignty of the international order. This model of control does not require visualization or territory to proliferate – only sound waves and satellite links – making possible the imagination of alternative spatio-temporal regimes.

Whether enabled by technology or by performance, sonic practices that exist outside of networked regimes neutralize conventional attempts to harness sound as a site of knowledge production. ⁴ In more connected milieus, wiretapping and electronic eavesdropping are perhaps the most obvious examples of sound-power in a security context. Here, evolving technologies make it possible to listen clandestinely to live conversations for the purpose of gaining intelligence. From the perspective of the eavesdropper, the success of a wiretapping operation depends on a thoroughgoing control of the technological infrastructure that supports the conversation; on pre-knowledge, or the predictive capacity to identify conversations that have potential intelligence value; and on the ability to interpret the aural dialogue accurately. Interpretive capacity, in turn, relies on linguistic and cultural expertise, as well as on the sonic quality of the wiretap itself.

With such factors in place, disembodied voice produces a stolen knowledge that could not otherwise be acquired, and distant listeners supervise speech in the maintenance of power. This eavesdropping, marked by the coercive capture of a bodily product, complicates the idea that having a voice equals representation. It also complicates the connection between the voice and the self, drawing attention to the ways in which voice may be severed from its body and intellectual author and forced to betray her. Migration authorities do not generally have the authorization, linguists, or staff to conduct such eavesdropping; yet the state capacity is there, and as voice recognition technology improves, and if migration continues to be discursively structured through a lens of security, these listening practices are likely to expand.

This is already evident in the ways in which cell-phone metadata is being used. While not itself auditory, this second-order process is part of the sonic–digital assemblage in which voice is entangled. In addition to producing spoken intelligence, phone calls, text messages, and the activated cell phone itself produce a host of digital byproducts that are fed into datasets to be mined for commercial and security use. These practices have spawned a burgeoning line of academic research characterized as the ‘datafication turn’ (Broeders, 2009; Mayer-Schönberger and Cukier, 2013; Taylor, 2015). The breakdown of anonymity for irregular migrants that this datafication spawns contributes, ironically, to a silencing of biological voice and self-authorship. There is no reason to ask a migrant her identity, or to inquire into her story, if her sonic practices have already been exploited to render her knowable.

Sophisticated wiretapping at the national level has a low-tech counterpart at the level of the individual migrant. Benjamin, who lost his parents to ethnic fighting in Guinea Conakry, worked his way north to Morocco, halting at each major hub along the route in order to earn money to pay for the next segment of his trip. He lingered at the gare routière in Gao for two months, collecting tips in exchange for handling luggage and performing odd jobs. During the day, he would pay heed to the coxeurs who negotiated onward passage across the desert, solicit news from home from newly arrived Guineans, and listen to the reports of migrants renewing their efforts to reach Europe, having been deported or otherwise thwarted on previous attempts (Interview E, 2015).

Varied acts of listening in, such as these, served as an important source of intelligence for migrants. While a minority reported having a clear sense of the journey upon their departure, learning from older brothers who preceded them, most gathered information and made logistical choices en route in real time, based on immediate circumstances and chance encounters. Ambient sounds and listening practices also manifested in migrant accounts as an important source of knowledge and power along the journey, worthy of their own adaptive practices and habits. During desert crossings, for example, migrants reported taking evasive action when they heard the sound of helicopters or airplanes overhead. At these sounds, truck drivers would pull over and hide among sand dunes or rocky outcroppings in an attempt to escape aerial surveillance. ⁵ Here, sound signals presaged sight, alerting the migrants to the presence of visual surveillance and potential capture. This was equally true in the forested migrant camps of Nador, along Morocco’s Mediterranean coastline. Migrant scouts listened at the edges of camps for the approach of Moroccan police and then shouted a call-and-response alarm, passing their voice from person to person through the forest so the larger community might shift deeper into the hills as evasive action.

This live listening practice, and the knowledge it purveys, represents, as in the Thuraya example, a distinct informational circuit located outside of mainstream communication systems and their attendant disciplining modes. Whether this alternative sonic circuit functions to the migrants’ advantage depends on the circumstances, but understanding this oral sphere as a separate and legitimate realm of knowledge production in its own right has important repercussions for actors attempting to access or influence this community. It was often via this informal dialogue that migrants made fulcral decisions, for example, about whether to follow the western or central migration routes out of Tamanrasset.

Moreover, the exigencies of precarious travel result in cell-phone and communication habits that frequently undermine security logics embedded in global technological assemblages, making everyday sonic practices unique (if sometimes unintentional) sites of resistance to the biopolitical exploitation of big data. Julian Reid (2009) has drawn our attention to the biopolitics of connectivity and the double-edged nature of infrastructures and technologies that enable connection while at the same time enforcing specific tendencies and habits, thereby enfolding populations into particular security regimes. While, at the far end of the spectrum, this biopolitics of connectivity feeds into networked war (Dillon and Reid, 2009: 106–126), we can find everyday examples of this phenomenon that appear, at face value at least, less violent.

Phone sharing, for example, was commonplace among the migrants I interviewed. Hassan, a migrant from Guinea Conakry, did not own a phone, but borrowed that of his traveling companion (a friend from his home village) whenever he needed to call home or to ask his brother to send money via Western Union (Interview A, 2015). While almost all migrants professed a preference for having their own cell phone, they often pooled resources and agreed to fluid sharing arrangements. Sometimes there was a primary owner who ‘rented’ use of the device to friends; at other times, if one person couldn’t gather enough capital on the spot, phones were co-purchased and migrants would sell their shares when paths diverged.

Having regular access to a mobile phone went beyond convenience or privacy concerns. Kofi, from Guinea Bissau, described how owning a phone in Tamanrasset, a well-established migration hub in the southern Algerian desert, enabled him to make travel and financial arrangements without leaving his safe house, as opposed to others who relied, out of necessity, on public phone kiosks in the center of town. As fixed sites that were of critical importance to migrants needing to contact support networks for assistance, teleboutiques were notorious traps. Kofi had learned from other Guineans in town that police would wait until the kiosks were full before raiding them, capturing undocumented migrants and loading them into trucks for deportation (Interview I, 2015). In this way, the need to call home is conscripted into systems of control; transnational IT networks performed a spatialized, disciplining function that forced migrants into contact with local security regimes.

Beyond communicating their need for help in an emergency, migrants spoke of ways in which verbal exchanges with family provided affective sustenance that gave them the courage to continue. As Kofi relayed, ‘you have to speak to hear their voices. You have to ask about their health; your mother, father, brother, sisters. Never text, always voice – it makes you feel them close and gives you strength’ (Interview I, 2015). Others relayed the outpouring of emotion that they experienced simply listening to the ‘regular’ ambient sounds of their village while they waited for a family member to come to the phone. Sounds may produce particular affective states or moods, or trigger memories that collapse temporal distances. Migrants relied on music to cheer them in their homesickness, or give them the courage to approach an obstacle. These straightforward examples of mood manipulation might seem inconsequential in a discussion on security. Yet the visceral power of sound (‘[it] gives you strength’) generated in the hearing–voicing technology nexus of the call home is not exclusively the purview of the subaltern; sound’s affective power is similarly of interest to the state, and its utility is being exploited at borders.

Scholarship on affect (Adey, 2009; Anderson and Adey, 2011; Goodman, 2012; Thrift, 2008) has demonstrated how mood, micro-expressions, and unconscious bodily functions increasingly come under surveillance and scrutiny at borders and airports. Bodily sounds that convey anxiety – quickness of breath, a vocal tremor, or the rush of blood (made audible with the aid of a stethoscope) all afford clues to particular affective states. Security practices that attend to such clues, like visualization technologies and data-mining operations, attempt to tap into an inner identity in order to determine intentionality. This, it is thought, provides security agents insight by which to preempt criminal action. Thus, both our hearing, as in Kofi’s description, and our emitted bodily sounds, as above, are jointly objectified in the audialization of the body to produce a real or imagined emotional state that may be formally judged in accordance with logics of preemption and security.

The flipside of listening and the technology that enables it is, of course, speaking. The next section begins with a discussion on voice understood as a subset of sound, moving beyond its purely communicative capacity, or semanticized intention, to attend to how the sonic aspects of voice are essential to the performance of identity and personal safety in ways that carry significance in migration and security spheres. I also discuss how material voice continues to be an object of intense scrutiny and a growing site of biometric cooptation as it is abstracted and divorced from its source.

Voice as resistance and a governmentality of sound

Voice’s particular sonic qualities make it an interesting case to examine within a sound-power framework. As mentioned above, voice conceptualized as sound demands justification because of the way in which semantics and language are entangled with the production of bodily sound. There is no denying the significance of linguistic meaning-making, but to ignore the vehicle, or medium, through which this meaning is transmitted is to miss part of the story. The tone of a church bell contributes to the sonic identity of a village, even as its ring indicates the hour. Affect and the physiological reactions that betray it change the sound of the words that escape our lips and imbue them with meaning that goes well beyond their linguistic content. Such a conceptualization of voice pays close attention to what Roland Barthes (1984) has called the ‘grain of the voice’, or the throaty, straining, mellifluous intonations, silences, stammers, rhythms, and coughs that transgress boundaries of diplomacy, deception, or ‘mere’ articulation. In understanding voice in this way, I draw attention to meaning derived from timbre, pitch, and accent, as well as the repository of knowledge produced and exploited when such sounds are collected.

Whether we characterize voice as enabling a weak or strong form of communicative empowerment, it continues largely to evade the international grid of surveillance information technology infrastructures have come to represent. As discussed in the previous section, unmediated voice and the ‘undisciplined’ everyday telephonic practices that migrants recount thwart would-be eavesdroppers and the quality of digital back-end data collected from cell-phone usage and mined by governments, corporations, and security agencies. From a technological perspective, voice also represents a uniquely complex sound – one difficult to reproduce with high fidelity and challenging for machines to read digitally. Voice, in this listening, represents a dynamic medium leveraged to resist state capture – thereby resuscitating power in what appear to be instances of powerlessness.

Voice enacts this resistance in three ways discussed here – through performance, withholding, and sonic particularity. Silencing, or withholding voice, represents perhaps the most straightforward way by which to resist efforts to ensnare the body in networks of power. In this connection, Efia, from Guinea, recalled sharing one phone among six travel companions. The group never used the phone in the vicinity of border crossings for fear the sound of their conversation would attract the attention of the police:

Anyway, [the Touaregs] did not permit it. They don’t want you to tell anyone what you are doing, and they don’t know who you are talking to. There are ears everywhere, and you don’t know who people are. ⁶ (Interview H, 2016)

Although traffickers were important enforcers of this censorship, migrants self-censored as well. In Efia’s group, they would shush someone speaking loudly or reprimand another for talking on the phone at an inopportune moment. Efia explained that within the group ‘you had to respect the rule of not speaking because it would be your fault if everyone was caught’. Here, silence served directly to avoid physical detection.

Another instance of voice as resistance was revealed when Efia described how she got past an Algerian interrogator at the border. By ‘listening in’ to conversations with migrants who preceded her at the border, she had learned that this interrogator spoke only French, and that he was mean-spirited and impatient. In collaboration with her travel companions, she devised a plan to feign incomprehension and lack of French as a way of evading his questioning. Those who could pretended English was their mother tongue. Efia, exclusively a Francophone, fell back on silence and pretended to be mute. She responded to the interrogator exclusively in gestures. Unlike in the previous example, where silence created a material, spatialized envelope of protection, here Efia’s silence functions to evade categorization, reclaiming bodily agency and recapturing, in the process, the freedom to act rationally in accordance with her shifting interests as she understands them.

This self-censorship, or self-muting, speaks to the relationship between voice, truth, and the performance of subjectivity. Efia performs muteness as a strategy for crossing a border when her vocalized self would be denied entry. She regains a modicum of control in a situation stacked against her by consciously manipulating the sounds she emits. In his Collège de France lectures, Foucault notes that ‘the subject who speaks commits himself’, suggesting that the speech act initiates ‘a pact between the subject of enunciation and the subject of the conduct’ (Foucault, 2015: 406). Yet this captures just one aspect of power that human voicing effects, and not the full range of sounded agency.

Fearing that her own voice – her language, her history, her accent and nervousness – would betray her in an interrogation, Efia took refuge in her capacity to remain silent. Her plan was that the dumb futility of the exchange would frustrate the guard sufficiently that he would lose interest and let her pass. Writing about lip-sewing practices in migrant camps in Calais, Banu Bargu (2017) has suggested that this form of resistance expresses the desperate terms that migrants face along their journey and their radical rejection of the conditions imposed upon them. It also points to silence as an overlooked site of power. When Efia opts to engage solely via gestures, she substitutes one mode of communication for another, thereby dramatizing or miming cooperation while maintaining a sphere of privacy and power. Efia’s interrogator, unable to trick the migrants into speaking French and lacking other means by which to determine their identity, eventually waved them through.

Uma, a 19-year-old English speaker from Nigeria, planned to employ a similar strategy when she got to Spain. She learned through conversations with migrants in the Bolingo Forest encampment, where she waited for her chance at a Mediterranean crossing, that Spanish immigration officials deport Nigerians ‘because they think we come from a rich country’ (Interview L, 2016). Uma planned to pretend to be Cameroonian. When asked how she expected to conceal her identity and lack of French, she explained that English is spoken in parts of Cameroon, in little villages; she intended to say she is from one of those villages. Though precarious, Uma’s plan is not unheard of; migrants often arrive without any official documentation or formal modes of identification. Receiving officials at asylum-granting institutions are forced to assess country of origin and a host of other personal details on the basis of scarce or absent information – a process that Marco Jacquemet (2009) has referred to as the ‘entextualization’ of an asylum-speaker’s verbal performance. The oral interviews that migrants and refugees undergo, and the testimony they provide, conscript them into Salter’s confessionary complex while highlighting the salience of the specifically audial nature of the performance.

Vocal performances in this context depend on knowledge of international law, a masterful ear for the nuanced call-and-response that occurs between border-crossers and customs officers, and an athletic ability to control bodily sounds and utterances (not to mention sufficient cultural and linguistic wherewithal effectively to style oneself as something else). The grain of the voice may be harnessed to hide in plain sight, as when migrants rely on accent to engage in hidden dialogue (Interview F, 2016), or it may serve as a mode of inserting the agentive self into complex regulatory structures comprising transnational migration regimes, as with Efia and Uma. Weighing their options at the border, both migrants decided that vocalizations were the most deeply revealing, and therefore perilous, indicators of identity and thus took steps to silence or alter them.

In these examples, the power struggle takes place face to face, between two people, making it easier to imagine how vocalizations could play a role. Yet voice as a mechanism of sound-power scales up to national and international levels as well. Digital cell data, of course, is easily parsed by machines, allowing for the expeditious processing of millions of distinct inputs that bring cell-phone users worldwide into a domain of government surveillance as described above. Human speech, however, because of the myriad (embodied, granular) variations in voice, continues to represent a significant challenge to automated systems; computers, like the impatient Algerian border guard, are simply ill equipped for the task. Despite some prominent utilizations, speech and voice recognition technologies are still in their infancy and plagued by technical difficulties and inconsistencies.

Thus, while computers may sift through back-data to select among millions of phone calls according to identified risk factors, they cannot readily analyze voice itself. A wide range of activities can temporarily or permanently alter an individual’s voice: having a cold, for example, or an afternoon of screaming at a football match, or simply aging. Voice biometrics is further complicated by the fact that speech and voice recognition technologies are being developed primarily in English. Artificial intelligence thus remains inept at recognizing the population most targeted by migration authorities. It similarly has trouble with female voices. As an intricate and sophisticated biological sound, machines cannot ‘read’ voice effectively, rendering it a source of inadvertent subversion and a reclaiming of the vocal body as something that is personal, indivisible, and, for the time being, situated at the margins of the technological reach of the state.

Voice and speech recognition, however, are quickly evolving fields, and ones that have practical applicability across a variety of sectors, such as health and disabilities, science and technology, linguistics, and security. The global voice and speech recognition market is expected to be valued at US$6.19 million in 2017 and is likely to reach US$18.3 billion by 2030. Its growth is being driven by increased demand for speech-based biometric systems. ⁷ In the realm of migration security and asylum requests, forensic phonetics have been playing a role since the 1990s, ⁸ in that language experts are hired by governments and immigration agencies to analyze the speech of asylum-seekers in order to determine, ‘empirically’, the speaker’s country of origin. In a program known as Language Analysis for Determination of Origin, or LADO, linguists are hired by border control agencies to analyze particular aspects of speech – accent, vocabulary, and syntax – in order to verify a refugee’s claims about her home country. As of 2010, there were eight governments and private laboratories using LADO to assess the legitimacy of asylum claims (Cambier-Langeveld, 2010: 69). The language analysis centers are located in Belgium, Germany, the Netherlands, Sweden, and Switzerland; combined, they conduct analyses for clients worldwide.

These forensic investigations center squarely on questions of identity, inquiring into who the speaker is and how this might be proven. As intrastate bordering practices, these techniques are in place specifically to catch people like Uma seeking to gain entrance to Europe on her own terms. Using a combination of assessment tools ranging from computerized voice recognition technologies, trained linguists, and consultants with native language ability, these language labs evaluate biological voice, which is held up as distinct from its semantic content and, because of this, inherently more ‘truthful’. The techniques represent allied mechanisms by which sonic bodies are forced to conform to the categories and classifications that undergird the international order – in this case, national belonging.

Voice is used against itself, as it were, creating a forcible break between intention and the medium by which it is conveyed. The right to determine yourself, to ‘voice’ yourself according to your own wishes, is wrested away as forensic speech science attempts to pinpoint a ‘scientifically’ determined identity without reference to individual subjectivity. This is all the more pronounced since analysis often probes beyond what we normally construe as voice to incorporate intonation and non-linguistic features such as tongue clicking or even coughing (Schilling and Marsters, 2015). Speech analysis, while purporting to detect geographic areas of primary socialization, has also been used in legal settings to make assessments about age, sex, and social characteristics encompassing such things as education and profession (Jessen, 2007). It has been harnessed to diagnose medical conditions and even psychological states such as depression (Lopez-Otero et al., 2014).

Thus far, the practice has not become globally standardized; it has been met with criticism and legal appeals in a number of cases, and represents one tool among many that governments are mobilizing to determine eligibility for asylum. Some linguists themselves have pushed back, drafting a set of guidelines for the deployment of LADO. The second item on their general guidelines states that:

Language analysis cannot be used reliably to determine national origin, nationality or citizenship. This is because national origin, nationality and citizenship are all political or bureaucratic characteristics, which have no necessary connection to language. In some cases, language analysis CAN be used to draw reasonable conclusions about the country of socialization of the speaker. (Language and National Origin Group, 2003)

Achieving accuracy outside a controlled experiment is extremely difficult in this genre of analysis. Plagued with poor results, practitioners have debated the relative aural skillsets of trained forensic linguists versus untrained native speakers in the context of authoritative identity claims based solely on speech analysis (Cambier-Langeveld, 2010). Dialects become muddled through travel and exposure, speakers adopt rhythms and words for stylistic purposes, and people adapt their language and intonation simply to match the needs of a situation, often making unique adjustments for each interlocutor (consider how you sound when you talk to a baby). The fallibility of the science only underscores the need to understand better the uses to which LADO is being assigned. The 1951 United Nations Convention Relating to the Status of Refugees, which lays out the rights of refugees at the international level, holds that ‘convention provisions are to be applied without discrimination as to race, religion, or country of origin’. This begs the question of the intention of government policies that support language analysis as a tool to determine, or verify, country of origin at borders.

Conclusion: The panaudicon

Self-muting, the awareness of one’s own voice as a source of power, the sense of having ‘ears everywhere’, and the role of sonic assemblages that entangle sound, body, and technology, all recall the mechanism of Foucault’s panopticon and the way in which power coopts both the surveillant and the surveilled into its domain of discipline. In the sound-power framework I have laid out in this article, the domain of disciplinary security so often understood in visual terms might just as easily be reimagined as a panaudicon. This is not as far-fetched as it might seem; Jeremy Bentham’s design of the ideal penitentiary (which Foucault’s writings in Discipline and Punish reference; see Foucault, 1979) included mechanisms for auditory surveillance that are elaborated in his seminal writings. Bentham envisioned a series of tin tubes connecting each cell to the inspector’s lodge at the core of the building, imagining that ‘by means of this implement, the slightest whisper of the one might be heard by the other’. The sound system was similarly expected to ‘prevent one prisoner from knowing that the inspector was occupied by another prisoner at a distance’ (Bentham, 1995: 31). Bentham’s technology was not sufficiently sophisticated to allow for asymmetrical or unidirectional listening, yet the groundwork for a regime of total audibility was nonetheless established.

Sound produces a genre of knowledge and power that enacts a self-disciplining familiar to Foucault scholars, but it overflows this particular conception of power, moving into Deleuzian assemblages and networks of control that may be more appropriate to the multitude of hybrid sovereignties and bordering practices scholars observe in the contemporary moment. Sonic infrastructures are multipurpose – they convey the sonic, of course, but they also produce it in accordance with a complex set of technologies that go into making communication take the form of sound. The constructed nature of voice has biological corollaries as well – we are socialized into our language, voice, and syntax. This socialization coincides with physiological features that morph over time – the shape of our throat, the size of our tongue, the angle of our teeth – making each voice unique. In this regard, human voice, much like human hearing, is always already mediated. Power relations infuse each point along this complex chain, impacting individuals differently and serving various functions contingent on circumstances.

From highly sophisticated transnational wiretapping, to self-healing in the form of affective support derived from phone calls home, to the scientific abstraction of sounds produced in speech, the sonic constitutes a form of knowledge distinct from other sensory modes. Along with the infrastructures that undergird sounded communication, these modalities are already deeply enmeshed in paradigms of control and resistance. It would be shortsighted to continue conceptualizing sound as politically neutral – or, worse, irrelevant. Arguments for data privacy and the protection of basic human rights meet with widespread approval, but protocols need to be rewritten to take the particulars of a sonic sensory body into account. The human ear, the larynx, and technologized sonic practices come together as a post-human geopolitical assemblage and distinct site of inquiry. In security and migration, these work in concert to reconstitute abstracted sound in the service of rationality and control, audializing the human body by ascertaining affect as an indication of intention; by pinpointing a body to a registered name or nationality, thereby concretizing identity and determining its status for inclusion or exclusion; by fixing bodies in space and time via their connectedness to global networks and the communication habits they effectuate; and, finally, by drawing attention to the hearing body as a target – a receptacle for induced sounds – and highlighting corporeal vulnerability when sound is wielded against us. This adds a valence to the concept of audialization that pushes beyond visualization as it is commonly understood; namely, it speaks to the weaponization, or capture, of an isolated sensory register for exploitation.