The Science of Listening in Bioacoustics Research: Sensing the Animals' Sounds

Introduction: Animals' Sounds

The purpose of this article is to explore animals, their sounds, and the technologies through which they are recorded and analysed in the context of bioacoustics research. Animals’ sounds, either vocal or produced by other means, are recognized in Darwinian terms as intensive expressions, ‘associated with the anticipation of the strongest pleasure [and anger] which animals are capable of feeling’ (Darwin, 1886: 84). Other contemporary perspectives on animals’ sounds oblige a behaviourist model, which constructs an enormous taxonomy of variations in animals’ voices as individuated response mechanisms to specific contexts and situations (Barrows, 2011: 19–26). ¹ By contrast, in recent studies involving early 20th-century biologist and philosopher Jakob von Uexküll (2010), a ‘posthuman’ argument poses that animals perceive life through the systems of signs that constitute worlds as rich and multifarious as human worlds, and who construct meaningful connections in ‘tones’, at once the physical manifestation of sound as well as any formation that carries meaning between organisms. The purpose of this article is to locate and theorize upon the shared yet disparate spaces between human and non-human organisms, as mediated through new sounding technologies, within the context of contemporary bioacoustics research. In line with the posthuman project of de-anthropocentrism (see Wolfe, 2010), the article argues that contemporary scientific research re-centres the human, widening the sensorium beyond human sensation, while diminishing anthropocentric modes of reasoning by coding machines to ‘listen’ at a rate more complex than humans are capable of. While grounded in empirics, the intention is not to diminish the importance of theory: theorizations here, however, are conceived within the current article as empirical ‘abductions’ (see Swedberg, 2014), insofar as an abduction is a framework of improvisation within which researchers conjecture spontaneously derived research questions to their research subjects themselves; researchers conversing with an orangutan or placing small microphones on the backs of nighthawks, as discussed below, stand as examples of these empirical abductions.

The article also intends to re-tangle some of the relations between animals, their sounds, and researchers through emerging sound technologies, and to account for new motivations for research, new questions, new methods of sonic analysis and new increasingly imperceptible ways of listening to animals’ sounds. These relations together justify the need for developing a framework for understanding the actions and interactions of organisms, human and otherwise. And such connections can be thought of as ‘agential’, in the sense of the mediation of two or more worlds that constitute an Umwelt (Uexküll's [1982] definition of a ‘dwelling place’ that arises from such interactions). The agent in this case refers to the new recording devices, such as Autonomous Recording Units (ARUs), that digitally capture, store, and analyse sounds for the purposes of species identification, capturing details as small as individual sonic emissions and as large as entire populations – sub-organismic, organismic, and super-organismic agencies (see Tønnessen, 2015).

Much of the article is concerned with how research spaces ‘open up’ through the use of new instruments and new contexts that bear an ethics of care. The ‘closed’ lab culture of experts who isolate and observe animal behaviour – and who view the death and injury of animals as a generalizable sacrifice that serves a greater good (Lynch, 1988) – has given way to a more open, living lab made of empathic encounters and, especially, opportunities for new theorizing. Much of the ‘expert’ modality has been problematized as an ‘epistemic scaffolding’ that structures a myopically anthropocentric set of scientific conclusions (Nelson, 2013; Friese and Clarke, 2012), which reify isolated components of animal behaviour by stripping them of their social and historical contexts (Candea, 2013). But, as the laboratory opens up to public discourse, this includes folding public concerns into the ‘core sets’ that guide research procedures (Michael and Birke, 1994), incorporating a co-emergence of research, researcher, and researched so that they hold comparable status within the research endeavour (Michael, 2012). Accordingly, this opening has led to an interest in how laboratories condition encounters between human and non-human actants. Such examinations are shaped by the vocabulary and methodology of Science and Technology Studies (Crist, 1996, 2004; Friese and Clarke, 2012; Buller, 2013), and the political discourses which inform how a ‘wild’ animal becomes an object of knowledge (Asdal, 2008). On the other hand, some of this interest can be equally attributed to Haraway's (2008) ‘contact zones’, which has pushed for multi-species orientations towards knowledge production, and which lies beyond the closed model of research. This perspective underlies such recent empathy-centred concepts as ‘embodied empathy’ (Despret, 2013), ‘trans-species empathy’ (Chiew, 2014), and ‘care’ (Giraud and Hollin, 2016).

However, there remains a word of caution against this ‘species turn’. For example, Matthew C Watson (2016) makes a timely rebuttal to the root metaphoric assumptions that are unquestioned within the field: namely, that poststructural terminological designations for animal behaviour in the service of opening up science in fact place an unchecked ‘mytheme’ around the animal. This removes species from their interconnected phylogenesis and its adaptability. Important scientific advancements in extinction intervention (part of what, some argue, makes conservationist biology a ‘crisis discipline’ [Soulé, 1985]) ² are diminished, in place of metaphors that paint the impending disaster with a ‘cruel optimism’ (Berlant, 2011, quoted in Watson, 2016). Watson notes, ‘Those of us in the human sciences must be willing to take seriously the statements that biological scientists have about their objects of knowledge’ (p. 166). I argue instead, based on the empirics of bioacoustics research and researchers, that sound is an extractive measurement for the purposes of data analytics. But within these data analytic processes there remains the embodied, the aesthetic, the affective, and the political. Emerging technologies thus contribute to the de-centring of anthropocentric hearing, but they re-centre the relations that humans may have with other species, insofar as the sense of technological hearing occurs at a rate far in excess of human hearing.

Scientific studies that use sound to gather data are not likely to expand the experience of listening, but instead to project their findings onto public digital archives for the purposes of community engagement and knowledge mobilization. Joeri Bruyninckx (2015), for example, points to the growing collaborations between ‘citizen scientists’ and the Cornell Library of Natural Sounds, which is in need of the public's hobby recordings in order to expand its growing catalogue of bird songs. The collaborative and empirical use of sound reinforces Julian Henriques's (2011) thesis that ‘sounding’ binds together everyone and everything involved in the production of sound: human (researchers), non-human (birds, monkeys, fish, plants), and para-human (including the technologies discussed below, such as ARUs and pattern-identifying software).

This article simultaneously contributes to new cultural theories of sound, which do not reduce sound to the status of wave, event, object, phenomenon, or transduction. Instead, newer sound theories embrace a more multifarious, expanding, and evolving explication of the intertwining between sound, body, place, sensation, and, generally, the virtual, and combined advancements in cultural theory along with the natural sciences. The virtual, the haptic, the affective – in short, that which vibrates beneath or above the surfaces of perception – are making their appearance within disciplinary streams beyond sound studies, where they explore a disparate set of concerns. Simply, such sound theorists as Mark Grimshaw and Tom Garner (2015) are more interested in the question of sound as a method. For them, sound emerges from within exosonic and endosonic vibrations, ³ enabling us to map a field, milieu, territory, or place, generating a ‘sonic aggregate’. In the sonic aggregate, when exosonic elements engage an endosonic series of associations, the emergent perception unfolds as though it were an aural Kanizsa Triangle, where form and contour may appear where none actually exist. None of this would be possible, however, without a spatio-temporal point of reference, a continuing framework of variation within the endosonic and the exosonic, and the edges that contain them. Bioacoustics is such a field of containment.

What follows is an intentionally broad discussion of bioacoustics as an example of how emerging technologies used for scientific research can act as an empirical event against which theorizing is possible. This is less a ‘theory of science’ (or field of science) than a call to ask questions with the sciences that vitalize new theorizations. Bioacoustics, thus, should be expedited not only into the scientific body (which it has) but into the language of cultural theory as well, since its tools and technologies expand the borders, boundaries, and flows of knowledge and transdisciplinary research innovations. The article explores: the necessity and usages of bioacoustics research; its co-emergence with technologies of sonic storage and dissemination; its correlative discoveries; the implicit retangling of hearing and listening; the places where animals are recorded; data compression; expanding edges and boundaries of conceptions of ecosystems; and finally the new soundscape ecology. The scientific perspective helps us understand how we hear just a slice of the sonic spectrum, and that sound is a method for elucidating how researchers are implicated in their research, including the animal/human/sound/species/techno-/listening semiospheres that contain and expand their capacities and potentials.

Listening-In: Bioacoustics as a Field of Sound, Life, and Loss

Within the vocal folds of an orangutan, scientists have discovered that, under controlled conditions, the primate has the capacity to ‘voice’, evidenced by its ability to (1) mimic grunts emitted by researchers as a mode of ‘primal communication’, and (2) manipulate the tones and timbre of its own grunts free from external stimuli (Lameira et al., 2016). ⁴ This capacity, which some neurologists have described as a ‘freedom from immediacy’ (Shadlen and Gold, 2004), is central to the concept of volition: the capacity to simultaneously anticipate one's own actions (intent) and to reflect on those actions (agency) (Haggard, 2008). The use of an ethnoprimatological methodology, whereby a researcher engages in playful communication with the researched subject (Malone et al. 2014), not only contributes to knowledge about shared ancestry, but introduces a practical, shared space between researcher and researched that encourages biosociality (Meloni, 2014). The discovery of its voice builds upon decades of study on primate communication in controlled environments, which largely held that these animals are incapable of ‘voicing’ in a manner similar to humans (Bolhuis and Wynne, 2009). ⁵ In the orangutan experiment, however, researchers would lead a game of ‘do as I do’ vocalizations (called ‘wookies’), which helped to establish a common ground in human and non-human communication. ‘Being able to socially learn new voiceless and voiced calls’, they explain, ‘would have, thus, effectively set the evolution of an ancestral hominid articulatory system on a course towards a vocal system fundamentally similar to modern speech’ (Lameira et al., 2016: 2). The research laboratory in this case becomes a kind of ‘semiotic playground’ (Upton, 2015: 177–9), wherein a set of restraints allows researchers to interpret a field (vocalization) based on a series of restraints (controlled conditions).

The meanings and significances around animal vocalization have occupied scientific discourse as proof for a variety of conclusions that support or refute the animals’ volition in the making of these sounds: René Descartes defended his notorious animal vivisections, for instance, claiming in court – against the complaints raised by neighbours who could not bear the screams bleeding into the streets any longer – that any of an animal's vocalizations were simply an automatic response, and that equating screams of pain with suffering was a category error. He asserted that animals were incapable of experiencing anything like an emotion or physical agony (see Bernstein, 1998: 133). Such an exclusionary practice, involving the bifurcation of the body as mechanical automata distinct from the capacity for abstract speech, was also applied to human populations. As Jacques Rancière notes, Aristotle followed his own famous ‘man is a political animal’ argument with a comparison to slaves as those who understand language but do not possess it (Rancière, 2004: 4–5).

It was Richard L. Garner's 1890s pioneering playback technique in Washington, DC, that inaugurated the interest in animal vocalizations (Bruyninckx, 2015: 349). Garner recorded on a graphophone the sounds emitted by captured monkeys when poked with sticks by their handlers. He took particular interest in one monkey that ‘chattered’ at its situation: It had been recorded onto a graphophone and later played back to another monkey in the hopes of eliciting an artificial conversation and tapping a ‘monkey language’ (Radick, 2007: 82–3). Garner linked his use of recording and playback technology with his beliefs that evolution offered insights into human experience, and also with his view that sound technology exposed a language beneath the ‘lowest’ forms of human life and deep within the evolutionary tract of primate life (Radick, 2007: 6). Bernard Stiegler (1994) has termed this union between technological advances and advancements in evolutionary thinking ‘epiphylogenesis,’ describing it as ‘the pursuit of the evolution of the living by other means than life’ (p. 135).

It was the development of the spectrograph recorder across the mid-20th century that helped to improve the study of animal vocalization through the visualization of sound, and inaugurated the field of bioacoustics (Mundy, 2009). While throughout much of the 20th century bioacoustics was used as a method for species identification, the field is currently aligned with crisis-driven data collection and environmental monitoring, which has become more popular in scientific and amateur circles in recent decades, and offers the opportunity to conduct large-scale research projects that monitor biodiversity and its increasing depletion. Thus, conceived as a broad-range monitoring tool, bioacoustics has potential applicability and utility in a range of contexts beyond the ‘monkey language’:

It is a low-cost means of data collection: the transductors are possibly the cheapest of any technological monitoring devices on the market. In the face of decreasing funding opportunities for researchers, bioacoustics is not only reliable and accurate, but relatively inexpensive to supply to a research team (Servick, 2014).

Bioacoustics researchers have examined the role of sound and communication in the composition of an ecosystem (Lorimer, 2012), across a variety of organisms (Gogala, 2014), and in rural, suburban, and urban habitats (Hinchliffe et al., 2005; Warren et al., 2006). Vertebrate and invertebrate animals (Altringham, 1996; Walters et al., 2013; Berry et al., 2004; O'Connell-Rodwell et al., 2012; Barlow et al., 2015), insects (Gogala, 2014), and plants (Gagliano et al., 2012) all emit sound to communicate, and are capable of sensing changes in patterns of sound (Gagliano et al., 2012), which implies that communication between organisms is intentionally encoded by senders and intentionally decoded by receivers (Hopp et al., 2012).

Birdsong has been of particular interest to bioacoustics researchers (Kronenberg, 2014; Borker et al., 2015; Merchant et al., 2015; Rempel et al., 2005), in part because they change noticeably with the introduction of new stimulants (anthropogenic noise, for instance; Warren et al., 2006), but also because they are readily identifiable by the naked ear and are species-specific so that identification requires minimal (if any) professional training (Cottman-Fields et al., 2013). For bioacoustics researchers interested in biodiversity loss, birdsong is of particular interest because ecological gradients have been shown to have effects on sensitive migratory acoustic communication patterns (Bayne et al., 2008). Again, bioacoustics in general, and birdsong in particular, reframes communication in such a way that includes human (Katti and Warren, 2004), non-human (Warren et al., 2006), and more-than-human (i.e. sound recording technologies; Snaddon et al., 2012) into new non-hierarchical relationships.

To return to the example of the orangutan, researchers are free to play by virtue of the fact that they are not required to listen out for the subtleties of vocalization themselves, as the experiment uses inexpensive recording technologies capable of storing large high-definition datasets with high-accuracy analytic capacities. These technologies are integral to bioacoustics, and the datasets are at the heart of the discourse of conservationist interventions (Towsey et al., 2014), such as when an animal population's collective calls begin to weaken – in such a context, animals’ sounds are read as signs regarding the terms and conditions of their environment. Michael Gallagher (2015) thus includes bioacoustics recording as a type of ‘nature style’ field recording, since the hand-held recorders and ARUs are non-intrusively positioned in environments where they pick up otherwise imperceptible vibrations, which can be visualized on spectrographic displays for scientific measurement. These visualizations make apparent the many animal calls that occur below or above normal human hearing: mice emit ultrasonic mating calls (Portfors, 2007), plants emit infrasonic and ultrasonic vibrations when distressed (Gagliano et al., 2012), and elephants communicate over great distances using infrasonic vibrations to locate migrating herds (Herbst et al., 2012). Whereas earlier research in bioacoustics was intended to identify animal species, as though sonic emissions were mechanical and physiological byproducts of an animal's property for the purposes of enumeration, contemporary bioacoustics research reads animals’ sounds as messages: as agents in the construction and maintenance of their environments. The technologies involved in this research have migrated from instrumental (species identification) to being part of a more complex network and co-constitution (conservationist intervention, species protection, and the identification of shared ancestry).

Concluding Remarks

To return once more to the orangutan, emerging sound technologies open up new spatialized intimacies between researcher, researched, and research. In alignment with the ‘species turn’ generally, the article has intended to identify how animals’ sounds, within such spatialized intimacies, are always meaningful. It is relatively obvious that these spaces of meaning were opened through early sound recording technologies (e.g. Garner's graphophone playback techniques), which clearly inaugurated a co-constitutive techno-sphere and semio-sphere. Similarly, the new technologies allowed for new perceptions of those communicative affects beneath human senses, those which were ‘living by other means than life’ (Stiegler, 1994: 135). Indeed, the spectrograph contributed further to the visualization of the ‘nether-sonic’ worlds in animal utterances, which today are used to structure predication about biodiversity along with how animals vocalize to respond to and manipulate their immediate environments. It is within this nether-sonic realm that has been theorized a shared ancestry, and has also been an event for new theorizations of new technologies – in this case, sound technologies.

Contemporary bioacoustics research expands outward to involve global research teams, including professional and volunteer scientists, and dissemination that reassembles and reimagines global populations and ecosystems, while coding listening technologies to map species according to their vocalizations, such as the AI experiment, Bird Sounds. Demonstrably, long-range analysis of animals' sounds in ecosystems requires new listening tools for storing an expanded duration of digital data. The pragmatic usages of emerging sound technologies are obvious: early detection of biodiversity loss, faster rates of conservationist interventions, a deeper sense of shared ancestry. But many of these benefits arrive at the expense of de-anthropocentrizing the listening experience, which tends to go against the grain of cultural theories of sound and sounding, which themselves tend to privilege the experiential, the perspectival, and the aesthetic configurations of sonic immersion. Indeed, if a tree falls in the forest, it should be for more than a philosopher to wonder whether it makes a sound. Emerging technologies, insofar as this study has demonstrated, thus de-centre anthropocentric hearing, but they re-centre the relations that humans might have with other species, insofar as the sense of hearing expands beyond that which any human can hear alone. The scientific perspective helps us understand how we hear just a slice of the sonic spectrum, and that sound is a method for elucidating how researchers are implicated in their research, including the animal / human / sound / species / techno- / listening worlds that contain and expand their capacities and potentials.