Navigation – Plan du site

Postures of listening

An ontology of sonic percepts from an anthropological perspective
Victor A. Stoichita et Bernd Brabec de Mori


This essay identifies and describes three ways of listening that are available to all human beings. Beforehand, we argue that the concept of “sound”, as borrowed from acoustics and commonly used in anthropology, is too vague and too limited. In order to be able to understand the full range of human auditory experiences as found in ethnography, as well as the social interactions which they afford, we propose a distinction of at least three postures of listening. We define these as “indexical”, “structural” and “enchanted”, by contrasting their interactional salience in various settings. The auditory “things” that exist for each of the three stances (their ontologies) are also shown to be different. This trichotomy provides a promising theoretical framework for some longstanding problems in anthropology. After discussing some critical questions and possible shortcomings of our model, we conclude by looking closely at one of these issues: the definition of “music” and its ethnographic relevance throughout the world.

Haut de page

Texte intégral

Listening as an anthropological issue

In his argument “against soundscape”, Ingold (2007) highlights the problematic status of sound in anthropology. Most researchers (Ingold included) agree with the physicalist definition of sound as pressure waves propagating through the air. Sticking to this definition, sound should be the auditory equivalent of light . However, Ingold remarks, it is more often, and apparently unproblematically, compared to sight (leading to concepts such as “soundscape”, which Ingold criticizes). In Ingold’s terms, this incoherence “reveals much about our implicit assumptions regarding vision and hearing, which rest on the curious idea that the eyes are screens which let no light through, leaving us to reconstruct the world inside our heads, whereas the ears are holes in the skull which let the sound right in so that it can mingle with the soul” (Ingold 2007: 10).

The confusion between sounds as waves and sounds as auditory objects is partly anglocentric. In German, for instance, a distinction between Schall (sonic waves) and Klang (perceived sound) is well established and largely reflected upon. But Ingold’s critique also reveals a deeper gap between the physicalist description and the way people experience their auditory realms. It is not just the incorrigible anthropologists who “misunderstand” the nature of sound. If you ask someone to locate the sound of a distant waterfall and that person answers by pointing towards the waterfall, he already departs from the wave definition. To follow it, he would need to point to a vast volume of air around the waterfall and encompass himself within it. The fact that many people do not locate sounds like this could indicate that they simply do not experience them as air waves. Casati and Dokic (2005) summarize several arguments in this respect, leading to the conclusion that “if sounds were sound waves, we would be almost always mistaken in our aural perceptions on important aspects, which fact … amounts to accounting for auditory perception in terms of a massive error theory”. Massive error theories may not be a problem in physics, but they are hardly ever satisfactory in anthropology.

If we are to understand what people experience through their ears, we then need to leave aside the fact that ears sense pressure waves. Just as with the eyes perceiving light rays, this is not untrue from a biological point of view, but it is not of immediate interest for understanding cultural representations and social interactions. What we need is a more specific description of the kind of things people sense with their ears, and the general interactions these things afford.

  • 1 We are particularly grateful to the participants in the workshop “Sonic beings? The ontologies of m (...)

After providing more background to auditory perception in the following section, we will present the main thesis of this paper, which is a model of three alternative modes of listening that can be adopted by any listener. The modes are akin to postures a listener occupies when paying attention to a sonic event. We call these postures of listening: (A) indexical listening; (B) structural listening; and (C) enchanted listening. These will be described in detail. During discussions with colleagues, we found that some specific questions were repeatedly raised. 1 After proposing the model of listening postures, we include a brief section for answering these frequently asked questions (FAQs). Finally, we conclude with a section dealing with some of the consequences of our model, specifically for the case of “music” as a form of enchanted listening.

From hearing to listening

  • 2 Attention plays a crucial role at later stages of auditory scene analysis. It actually also modulat (...)

Some of the cognitive processes involved in the treatment of auditory information are considered universal in cognitive sciences. Such is the case of “auditory scene analysis” (Bregman 1994), which describes the segregation and processing of auditory streams. The segregation is a prerequisite for scene analysis and relies on low-level “bottom-up” mechanisms (Alain et al. 2000). Although it has been shown that auditory stream segregation must be learned, it is considered universal among humans as well as many non-human animals (Bregman 2013; Faragó et al. 2014 for emotional valence in dog’s streams; Shamma et al. 2011). 2 An auditory stream receiving focus of attention appears in the foreground and is perceived as more salient than other streams that are blended into the background or blocked out entirely.

From a cognitive point of view, “attention” can be defined as the “mechanism that allows certain information to be more thoroughly processed in the cortex than non-selected information” (Cohen et al. 2012: 411). It is tempting to compare auditory attention to a kind of “bottleneck” (e. g. Pilcher et al. 2016: 1039), although in detail it is more likely a network of several subsystems which sometimes operate in parallel (Caporello Bluvas & Gentner 2013: 12; Cohen et al. 2012: 412; Demany et al. 2015; Irsik et al. 2016). Attention should be distinguished from “awareness”, by which “we often become conscious of [an item’s] attributes at the expense of unattended items” (Cohen et al. 2012: 411). People are able to control their attention and focus selectively on specific parts of an auditory scene. The unattended parts receive less cognitive processing, do not reach awareness, and can induce spectacular effects of “change deafness” (e.g. Cherry 1953; Fenn et al. 2011; Irsik et al. 2016; Vitevitch 2003). On the other hand, it has also been shown that unexpected auditory events in unattended streams may receive increased cortical processing (i.e. attention) while remaining below the level of conscious awareness (reviews in Cohen et al. 2012 and Snyder et al. 2012: 9 sqq.). Finally, some sounds should ideally always capture attention and make their way “bottom up” into awareness (alarms, for instance, are meant to bypass any “change deafness”). The distinction between attention and awareness helps to understand where auditory processes stop being universally predictable. In fieldwork, anthropologists and ethnomusicologists gather data relevant to people’s aware perceptions. But the allocation of attention is probably the earliest stage where perception can be modulated by cultural preferences.

Beyond low-level processes of auditory stream segregation, most cognitive operations involved in hearing arguably depend on learned systems of knowledge and meaning. This was shown in Steven Feld’s work (1990 2000) about hearing and meaning among the Papuan Kaluli, and even more specifically by Rafael José de Menezes Bastos (1999, 2013). This author shows how the analysis of native terminologies and axionomies leads us to observe that, for example, among the Xinguano indigenous people, “sound is as material as stones are for us” (de Menezes Bastos 2013: 292), or that in the same society, sound is not merely perceived, but “sound is actively sought and captured by the ears” (de Menezes Bastos 2013: 292). The very base of auditory perception acquires a different range of meaning and agentive power when conceptualized in such a way. Building upon this, Menezes Bastos’s former students Mello and de C. Piedade (2005) also demonstrate that the construction of space based on auditory perception among the Central Brazilian Wauja shows particularities that are different from an alleged Western classical tradition. They assume that the ontological characteristics of sound depend on culture, so that although psychophysical basic processes may be similar, the characteristics of the things heard are particular.

  • 3 This may seem similar to (and is probably inspired by) Pierre Schaeffer’s discussion of preobjectiv (...)

Ethnography shows that anthropologists cannot really presume what kind of things people hear. The internals of the biological ear (an organ which clearly senses pressure waves) do not tell much about the things that actually constitute people’s auditory realms. This applies to “sounds” but also to “language” or “music”. Firstly, these concepts are rather cumbersome in cross-cultural comparisons. “Language”, for example, can have different extensions in various societies, including or not the sounds of animals, rivers and so on. Likewise, “music” does not even exist as a category in most non-European languages (sounds for healing, hunting or having fun are not necessarily linked by an overarching concept). More importantly perhaps, “language”, “music”, or “the acoustic environment” refer us to phenomena. We propose not to take phenomena for granted, and to start by describing the modes of awareness which allow for auditory phenomena to occur. This is one significant difference with most existing theories of listening: we do not ask how many ways there are to hear “music”, or how listening to “music” compares with listening to “language”. We propose to start our inquiry before, at a stage where we do not know at all what people listen to. 3 From there, we can simply observe how humans interact with their surroundings through their ears: what they do, how they react, what they say about it. On this basis, we try to give “functional specifications to the structures that must be present” (Hutchins 1995: 131) beyond the reach of our methods of investigation.

This approach is not meant to reveal especially new things about how people listen to “sounds”, “music” or “language” (although it might serve to clarify a few points along the way). The expected benefit would rather be to understand how they listen to things like “gods-as-vibrations”. Beings such as these are not unusual encounters in ethnography, but they remain critically absent from typologies of listening (when they are allowed entrance, it is only as “sounds” or as “music”; see for instance Becker 2004). Another objective of our inquiry is to determine whether some listening modes are found in all human societies. On the one hand, ethnographies show great variability in the way systems of auditory knowledge are built. On the other hand, humans engage in remarkably cross-cultural activities: they listen for prey when they go hunting, they talk to each other (sometimes to animals and rivers too), they dance … Beyond the general ability to pay attention to sounds, are all modes of listening culturally variable? Or are there some modes which are shared by all human beings? By investigating this question, we might ultimately even be able to say something about temptingly universal phenomena such as “language” and “music”.

  • 4 “Theaffordances of the environment are what itoffersthe animal, what itprovidesor (...)

Here we will deal exclusively with listening, leaving aside sound production. Listening, as opposed to hearing, implies awareness. There is, however, more to listening than awareness. In many cases, the sound one hears affords a certain reaction to the hearing individual. Gibson suggested that the perception of affordances is embedded at a very low level in our appraisal of the world (Gibson 1977, 1986). 4 As Martin Clayton puts it for sounds, “we do not passively perceive and subsequently decode sonic information, so much as actively scan sound energy for patterns of which we can make sense” (Clayton 2001: 11). Cognitive and acoustic experiments have shown that this “scanning” requires a mixture of short-term memory and anticipation (for a recent overview see Castellengo 2015: 146 sqq.). This means that auditory objects like tunes, words or familiar noises are never “sensed” literally. The present, through our eardrums, brings us only a short glimpse of them. In Husserl’s terms, “the objectivity of the sound that lasts is constituted in the ‘continuum’ of an action that is in part remembrance, in a very short, punctual part, perception, and in a larger part, expectation” (Husserl 1964: 36–37, quoted in Castellengo 2015: 487, our translation). By “listening” we refer to this complex of action, memory and imagination.

We define alternatives of listening as distinct ways of using a given item of sensory information by the same being to construct different kinds of listening objects. These objects are of different kinds when their affordances for that being differ. In a given acoustic environment, the listener adopts one of these alternative listening postures, and thereby perceives a specific set of entities, which opens the way to specific relational possibilities. As their name implies, alternatives yield incompatible results: the same auditory field can be apprehended in one way or another, but probably not in two ways at the same time (see FAQ section for discussion).

We propose that three alternative postures can be identified in all human societies. We do not claim that these listening modes are the only ones which humans have developed. We do think, however, that these modes of listening are used and applied among all humans, and they are probably the only ones with such universal character. Such an assertion can of course not be proved. The best one can do to address something that might be common to all humans is to make one’s claim refutable (in the sense of Popper 1959). To that effect, we try to describe each alternative and the kind of auditory world it constructs in a precise way, while also setting out its most obvious empirical consequences.

Alternative A: Indexical listening

We sometimes infer from the sounds that we hear the physical state of the world. Animals can do this too. It is a matter of forming a hypothesis about the cause of the pressure wave.

Birds in Rainforest
Bird calls and drops of water falling on leaves
Ambient recording of July 2003 during a walk through a small rainforest area in Sian Ka´an Natural Park in southern Yucatan, Mexico
Crédits : Recorded and posted under CC 3.0 license by Reinsamba on

For such a hypothesis to be obtained, sounds are treated as indexes in the sense of Peirce (see also Peirce 1992; 1998: 13). Through this way of listening – moving upstream to the physical cause of the sound heard – animals endowed with hearing attain three different kinds of knowledge.

Firstly, they discover and locate other entities acting in their surroundings (“something is walking there”). From the sounds heard, they assume the mere existence of “something”.

Secondly, they can form a hypothesis about the thing or being which they hear. Acoustic signatures allow us, for example, to infer that what is there in the branches is a fruit dove, or that the person on the phone is Aïcha. This is a hypothesis about the entity’s identity. Although it may occur simultaneously with the first inference, existence and identity are distinct. We can indeed be wrong about the species of the bird in the tree or the identity of the person on the phone, but something or someone is definitely there. Both inferences are indexical, because they are built on causal associations. These associations are grounded in observed recurrences: this kind of sound is usually produced by a fruit dove, that one by Aïcha, etc. (for a perspective from cognitive psychology, see e.g. Keller & Stevens 2004).

There is a third kind of indexical inference in sound. It enables listeners to build a hypothesis about the interior state of other beings. This is what occurs, for example, when we think that a person’s voice “betrays” the speaker’s thoughts or feelings. The feelings may not be apparent in the words uttered by that person, but we still think that we sense fear, happiness or anger in the tone of the voice. We rely for this on an indexical interpretation of vocal prosodic features (intonation, rhythm, timbre, etc.) that we understand causally as being produced by specific interior states. Note how this is different from intentional communication: when we say that a voice “betrays” the speaker’s inner feelings, the inference cannot be said to rely on the deliberate use of a shared symbolic system. The “betraying” indexes are in fact not symbols at all in the sense of Peirce. There is no convention stating how they should be interpreted. The listener can only guess Aïcha’s interior state through the observance of previous regularities (e.g. usually, when Aïcha sounds like this, she is in an angry state of mind), and the inference is affected by gradual variations in sound (Aïcha can sound more or less angry depending on how salient the corresponding prosodic features are in her voice).

Of course, indexical interpretations can be wrong, and indexes can also be faked. Let’s consider the first inference about existence. It is usually accurate in ecological environments: if you hear something there, something is probably there. But illusions can be built on this inferential process with special tools like recording and playback equipment. Listen to a recording of your favourite string quartet for example.

Allegro of 4th String Quartet by Gustavo Becerra-Schmidt
Performbed by Philadelphia String Quartet
Crédits : Released under CC 3.0 license on

  • 5 This is true as long as the extension is within the range of possibility: compare the string quarte (...)

How many things do you hear? If you listen in stereo, the direct indexical answer should be two. Most Europeans, however, will feel that they are listening to “the quartet” – two violins, a viola and a cello, that is four – not to the loudspeakers. In such a case, the recording and playback apparatus function as an extension of the situation in which four actual sound sources were recorded and mapped onto the stereo panorama. 5

The second kind of indexical inference – identification – can also be faked. The fruit dove, for example, can be tricked into believing that it hears the calls of a potential mate, when in fact a hunter lies hidden in the bushes. It is worth noting that tricks of this kind work precisely because they are exceptional. Humans too rely on voice recognition in many daily interactions, because voices are difficult to imitate.

Lastly, interior states can also be falsely attributed. By using vocal “icons of crying” (Urban 1988), professional mourners can convey feelings of sadness while mourning for a family they hardly know (cf. Amy de la Bretèque 2013: 89). Here again, it takes some special skills to get the imitation right and make it convincing. Stage actors also need to train their voices in order to acquire the special ability to embody various states of mind on demand. Because such skills remain uncommon, listeners tend to infer the interior states of other beings from the way they sound, even when they are well aware that the feelings are enacted and not spontaneously felt.

One of the reasons why acoustic indexes are hard to fake is because the listener’s inferences are very sensitive to infinitesimal variations. In indexical listening, it makes a difference whether the voice we hear is a little higher or lower pitched, whether consonants receive a bit more or less stress, whether the steps we hear sound a bit further or closer. As shall be seen, many of these infinitesimal variations are stripped away in structural listening, which will be outlined hereafter. Some of them are relevant, again, in enchanted listening, but there they pertain to a totally different ontology.

Another specific characteristic of indexical listening is the way it maps sounds in space. When we hear them as indexes, sounds share the same location as their physical source. If we hear the fruit dove on that tree, for example, we expect to also find it there physically. In indexical listening, hearing, sight and touch should map a similar world. We will see that this contrasts with the two other kinds of listening to be described.

Alternative B: Structural listening

Contrary to indexical listening, structural listening does not strive to reach the physical cause of the sounds heard. Instead, its main scope is to abstract relevant patterns from auditory data. Relevance can take on different forms, depending on the cognitive task at hand.

This mode of listening is used, for example, in understanding language. Saussure (1916) and later Martinet (1965: 15; 1991: 20) have insisted on the functional importance of double articulation in all human languages: utterances can be broken into minimal semantic units (signs, or monemes in Martinet’s terminology), and the latter can themselves be broken down into a limited number of non-semantic units (phonemes). The core of this theory is that meaning is not achieved through the “positive” content of the units used to construct the message, but rather through their oppositional “value” in a given system (Saussure’s term, see 1916, Chap. 4). This approach has been exported from linguistics to various other research domains, including anthropology and (ethno)musicology.

Our own proposal is much simpler than most structuralist theories. It states merely that listening in a structural way is possible . In other words, we claim that acoustic data can be parsed for oppositional units, according to a system. We do not presume to say that this is the most relevant way of describing a specific activity, not even linguistic communication. Like the other alternatives described here, it is simply an available option for the listener.

  • 6 This competence probably needs learning. Babies, for example, are initially more sensitive to vocal (...)

The interesting fact, for our purpose, is that if the listener adopts the structural hearing posture, the cognitive result is, by definition, entirely abstract. Structural elements do not depend on the media which conveys the information. For example, the same words can be uttered by various voices, with various pitches, various intonations, louder or softer, or at various speeds. 6 And they can also be written down, flashed over the sea in Morse code, or transmitted as binary data over computer networks. Such operations imply an abstract level, where the “value” of the different phenomena, notwithstanding their different material forms, remains the same.

The English Language in 67 Accents & Random Voices
User Truseney92 on Youtube demonstrating a variety of English accents

This is how Saussure reached his conclusion that “it is impossible for sound, which is material, to belong by itself to language” (de Saussure 1916: 164, our translation). In a broader definition of language, which would include notably prosodic factors, gestures and contextual implications, structural listening is only a small part of what we do in linguistic communication. But when we listen for structures, we set aside most acoustic features of sound, and the resulting auditory object is hardly acoustic at all. Whereas infinitesimal variations are important in both indexical and enchanted listening postures, here only a few oppositions matter.

Structural listening is not limited to linguistic communication. It can be triggered at will by the listener, on any kind of sound. For example, bird calls can be “understood” in a structural sense, when the way in which the bird vocalizes transmits an omen (cf. Walker 2010). In entirely different contexts, data sonification is often used to convey information that is clearly abstracted from the actual sounds (see e.g. Supper 2014). Musical notations are yet another achievement of structural listening. In order to write music down (whether we compose or transcribe it from what we hear), we need an abstraction layer, where sound and graphic signs share some common structure. The mundane ability to consider that, say, a xylophone and a flute play “the same” melody points to a similar abstraction. To assert their equivalence, the listener must retain only the structural relations of the pitches, discarding the obvious differences in the sound spectra of these instruments.

Tortoise gallop
Audio collage of two structurally similar tunes
Part 1: Theme of "Gallop infornal" (aka "French cancan") from J. Offenbach, Orphée aux Enfers.
Part 2: Theme of "Les Tortues" from Le Carnaval des Animaux by Camille Saint-Saens
Crédits : Koninklijk Filharmonisch Orkest van Vlaanderen (part 1) / London Sinfonietta (part 2)

Structural listening contrasts with both the indexical and enchanted postures in several ways. We have already mentioned its insensibility to many variations of sound. This extends to spatial location. Physical space is important in indexical listening. As we shall see, enchanted listening constructs a space of its own. But space is abolished altogether in structural listening. Whether the person telling you Cinderella’s story is sitting to your right or your left does not change the representation you form of the narrative. A conventional musical transcription will be similarly unaffected by the position of the sound source being transcribed.

Alternative C: Enchanted listening

We have distinguished indexical listening, which relates the sounds to their physical causes, and structural listening, which searches them for abstract patterns. Humans have at least one more way of directing their consciousness to sounds. It is characterized by a split between sounds and their physical causes. When listening in this way, sounds seem to form an autonomous realm “which stands apart from the ordinary workings of cause and effect, and which is irreductible to any physical organization. At the same time, it contains a virtual causality of its own, which animates the elements that are joined by it” (Scruton 1997: 39).

  • 7 Sacks (2007: 182) relates the following discussion with Michael Torke, a “true synaesthete” (who ha (...)

The elements mentioned by Scruton are objects of listening, hence “sounds” in a broad sense. But they have properties that sounds lack in the other two ways of listening. “Virtual causality” and “animation”, for example. A listener may sense entities which “move”, relate to each other in various ways, and possibly embody an agency of their own. She might also perceive them as “coloured”, “shaped” or “textured”. Our listener can be well aware that colours cannot normally be heard (even true synaesthetes do not actually “see” colours in sound 7). She hears them nevertheless, and they have nothing to do with the colours of the sound sources. She may also check that these properties vanish away when she adopts another manner of listening. She can “step back” and listen indexically, for example, to the physical source: where is it, what is it, does it move, etc. In that world, there are no more lines, movements or textures, just some sound coming from some being or object. She can then step in once more, whereupon only the same original movements, colours and textures will exist again.

If she uses a European language, our listener will probably label what she perceives in this listening as “music”. Ethnomusicology has shown, however, that this concept is not found in many other cultures. In the Amazon, for example, indigenous people often use terms that refer to specific ways of singing, at the same time excluding the sounds of instruments, which, on the other hand, are understood as voices of spirits (Brabec de Mori 2012; de Menezes Bastos 2013; Piedade 2013). Even in those societies where the concept of “music” is used, its extension is ambiguous. In many Muslim societies, for example, what is called musīqī (with this loanword from Greek) explicitly excludes calls to prayer ( adhān ) and recitation of the Qur’an, both of which are nevertheless definitely perceived as “music” by the average European, and as peculiarly agentive vocal productions by the pious Muslim (Hirschkind 2004; Shiloah 1997). For this reason, it is better to stick to the listening ability itself. We will refer to it as “enchanted listening”.

  • 8 To illustrate : “a flute, no less than an axe, is a tool, an element in a technical sequence; but i (...)
  • 9 We adopt Gell’s definition of agency: “whereas chains of physical/material cause-and-effect consist (...)

We understand enchantment in reference to the “technologies of enchantment” discussed by Alfred Gell (1988, 1992, 1996, 2006). Gell’s proposal stemmed initially from an analysis of the concept of technology (Gell 1988). He observed that some technologies are meant to modify the world, while others target instead the way the world is perceived. 8 The latter he named “technologies of enchantment”. Enchantment remained, however, an obscure concept throughout his work, which focused primarily on the “technological” side of the proposal. Some years later, on a different path, Philippe Descola pointed out that the beings and things which constitute people’s experiential worlds are first recognized and categorized through a set of low-level cognitive processes. He called these processes “schemas”, in the general sense of “abstract structures that organize understanding and practical action without mobilizing mental images or any knowledge conveyed in declarative statements” (Descola 2005: 149, translation from 2013: 59). Descola proposed a cross-cultural survey of collective schemas which govern identification and relationships. He compared in particular the distribution of “interiorities” and “physicalities” in different ethnographies, and arrived at the conclusion that “these principles of identification define four major types of ontology, that is to say systems of the properties of existing beings” (Descola 2005: 176, translation from 2013: 64). Now auditory experience is normally not an autonomous ontological realm. Most of the time human beings do not consider it under a distinct “system of the properties of existing beings”. Indexical listening points to non-auditory causes. Structural listening points to non-sonic structures. On occasion, however, human audition can materialize specific systems of sonic beings which then display particular sets of sensory and relational properties. People typically describe them in terms of unhearable dimensions which are not linked to any physical causes. Quite often too, sonic beings are endowed with autonomous agencies, meaning the capacity to initiate actions by themselves. 9 We call enchanted listening the fact of experiencing a properly auditory ontology.

This ability appears to be shared by all human beings. In all human societies there exist interactions that rely on sounds being apprehended as a world of their own. In such a world, sound events are related to each other rather than to their physical sources. They obey specific intrasonic causalities. For example, people may feel “tensions” and “releases” in sound, and react to them emotionally, as well as bodily through dance. This is a classic way of experiencing tonal music. The building of “tension” and “release” is taught in composition classes, and is also used as a basic analytical framework by many musicologists.

Tension and release
Demonstrtation of chord "resolution" in tonal music
[Description at]: "Leading tone repeats four times over dominant (V) chord which then moves to the tonic (I) as the leading tone resolves upwards to the tonic"
Crédits : Hyacinth at Wikipedia

  • 10 An in-depth exploration of intra-sound causal principles in Western tonal music can be found in Ock (...)

“Tension” and “release” refer to a world where sounds have the ability to build up in an equilibristic pile, oppose “their own” antagonistic energies, lose “their own” momentum, leading “by themselves” to new sound events which resolve the instability and release the tensions accumulated. None of this happens in a world where sounds occur because a musician pressed some buttons on her instrument. It happens in a world where sounds obey their own causal rules, a world where sounds occur because of other sounds. 10

“Tension” and “release” are merely examples that apply to specific kinds of music. They are by no means universal concepts, and other representations can be shown to fulfil similar roles elsewhere. In Papua New Guinea, for instance, the Kaluli use (or at least used in the 1980s) a wide range of water-derived terms to describe positions and movement tendencies in the sound realm. According to Feld (1981), in the Kaluli language, sa is a standalone term for a waterfall. It can also be used as a prefix for many things related to water, and also to song. For example, a sa-we:l refers to “the ledge or upper place from which the waterfall drops”, which in song corresponds to “the leading pitch in a line or phrase from which the melody descends”. Hence one can correct someone’s singing using such a sentence as “the waterfall ledge is too long before the fall”. Feld gives numerous other examples outlining a consistent use of hydraulic representations by the Kaluli people in commenting their songs.

In Kaluli aesthetics, references to water flows play a role similar to the mechanics of “tension” and “release” in Western tonal music. Listening to something like a water flow or an equilibrium of energies demanding resolution are both enchanted experiences, because sounds then interact with each other in a suspended realm, according to rules of their own.

Enchanted sounds or verbal metaphors?

  • 11 See the distinction between “true” and “metaphorical” synaesthesia in psychiatry: “For most of us t (...)

The phenomenal reality of such worlds has often been overlooked in anthropology and musicology, possibly because the corresponding verbalizations have been categorized as “metaphors” by the concerned scholars. The usual assumption is that “metaphors” are not “for real”. With a few exceptions (Gay 1998; Rice 2001), they tend to be treated as mere linguistic devices with only a vague relation to the speaker’s experiences. 11 Ethnomusicologists often strive to map these metaphors to the “real” features of sound. Even though acousticians have generally acknowledged Schaeffer’s early warnings that “signals”, as measured by instruments, should not be confused with “sounds”, as experienced by humans (see e.g. Schaeffer 1966: 269; Castellengo 2015: 34, 139), frequency, amplitude and spectral composition are treated as actual parameters of “sound” perception throughout anthropological and ethnomusicological literature. Tension, colour or “waterfallness” are not, although they are probably closer to the actual percepts experienced by the listener.

In their classic study of metaphorical thinking, Lakoff and Johnson (1980) have shown that metaphors are not merely ways of speaking, but actual cognitive devices which enable humans to grasp the world through analogies, and which influence in return the way humans experience the world. In their general definition, “the essence of metaphor is understanding and experiencing one kind of thing in terms of another” (Lakoff & Johnson 1980: 5). “Tension” and “release” are metaphors mapping mechanical processes onto the sound realm. “Flowing like a waterfall” is a metaphor mapping hydraulic processes and geographical features onto the voice. In the early twentieth century, musicians in Cairo used the Arabic word tawriq , meaning “covering with a thin layer of paper or plaster”, to designate a subtle heterophonic outline of the vocal part added by the instruments to a singer’s voice (Racy 1988). Ethnography is replete with such examples where sound is understood in terms of something else.

Metaphors bridge ontological gaps. It does not follow that they are arbitrary poetic figures contrasting with more “literal” uses of language. For one thing, the concept of literal meaning – as opposed to “figures” of speech – remains a puzzle in experimental linguistics (Gibbs & Colston 2007: 582). Secondly, Lakoff and Johnson have shown that conventional metaphors can be used to say true or false things in just the same way as non-metaphorical propositions (Lakoff & Johnson 1980: 159–184). It follows that even wordings like “covering with a thin layer of paper or plastic”, which are obviously not grounded in auditory perception, can be used by listeners to understand and structure their actual sound experiences (a fact demonstrated ad absurdum when metaphors are used at wrong occasions and contradicted by the listeners as being false). The problem, then, is to determine which aspects of the source domain are mapped onto the target.

Enchanted listenings often behave as collective schemes of identification and relationality (Descola 2005: 149 sqq.). Let us illustrate this through another example from Mount Bosavi. Aside from the water flow vocabulary, in some contexts Kaluli people also use a completely different set of representations which describe the “hardening” or “softening” of sound (Feld 1986). Through the “hardening” ( halaido ) of its sound (a complex process involving both woodcraft and magical operations), the ilib drum, Feld explains, can leave the listeners with “the sensation of hearing the voice of a bird ‘inside’ the sound of the drum, and then hearing a further reflection, the voice of a spirit child coming through the voice of the bird to call ‘father’” (Feld 1986: 94).

Ilib (hand drum) duo at Mount Bosavi
Recorded at Sululib longhouse (Bosavi area), Papua New Guinea, 1977
Crédits : Collected by Steven Feld. Published on the LP Music of Oceania, Musicaphon BM30SL2702. Side B track 12. Reproduced with author’s permission.

The ilib drum “‘talks’ ( tolan ), and actually says dowo , ‘father’” (Feld 1986: 94). For the Kaluli, this drum ideally embodies a ghostlike being endowed with its own agency. A complex net of non-verbal interactions depends on this being’s existence. There is, for example, the reaction of extreme sadness, which prompts some listeners to “thrust their way out onto the dance floor, sobbing loudly, and, brandishing a resin torch, strike and burn the drum whose sound moved them so deeply” (Feld 1986: 93). For this to happen, the listeners need to really experience the sensation of a bird reflecting the spirit of a dead child calling for his father from inside the drum . Referring to Boyer’s analysis of “supernatural” beings, one could say that this experience is “counterintuitive” or “counterontological” for the Kaluli (Boyer 2001: 64 sqq.). They are perfectly aware that there is no actual bird inside the drum. The bird at which they aim their torch in despair is a purely auditory object, a sonic being. The Kaluli experience of ilib drumming bears the mark of enchanted listening, because it causes, or results from, ontological shifts and the reframing of human/non-human agencies.

Enchanted listening does not actually require articulate cosmologies. “Movement”, for instance, is a basic metaphor for sound processes throughout Western music (e.g. “rise” and “fall” in a tune, rhythmic “swing”, “walking” bass).

“Buildup” and “breakdown” in a trance music track, analyzed in Butler 2006 : 315.

“Buildup” and “breakdown” in a trance music track, analyzed in Butler 2006 : 315.

Abbreviations: bass drum (BD), riff 1 (R1), snare drum2 (SD2), snare drum 3b (SD3b), riff 2 (R2). Musical excerpt from Communication (Somebody Answer The Phone) by Mario Più, 1999, Incentive ‎– CENT2T.

"Buildup" and "breakdown" in a trance music track
Analyzed in Butler (2006 : 225-227, 311-317).
DJ Shiva critically comments on this : "This is also [one of] those obvious builds… buildups […] we’re all about just bringing you up and smashing you down." (quoted by Butler 2006 : 226)
Crédits : Mario Più, Communication (Somebody Answer The Phone). Avex AVTCDS-236 starting at 4:40 [Integral tune on the companion CD of Butler 2006]

It is nevertheless an enchanted experience. The loudspeakers stand still, the band members on stage stand still, but something moves in the music. Clarke (2001) has argued that musical “movement” is best understood as a truly perceptual phenomenon, rather than as a way of saying (see also Becker 2010). Clear signs indicate that people experience it, even without their verbalizing it. The “movement” in what they hear sets them in motion; they dance, bounce their heads, tap their feet … Like with ilib drumming, the ontological shift in sound is manifested by a vocabulary, but also by a complex net of interactions, which would be absurd without it. It should be noted that one cannot dance to indexes or to structures: they have no “movement” to communicate. Contagious auditory movements operate from a time-space of their own.

They can be personified too. The bourgeois salon is in this respect no less a stage for supernatural encounters than more “exotic” rituals. Here is a famous introspection by Marcel Proust into his fictional Swann’s mind:

“With a slow and rhythmical movement she led him here, there, everywhere, towards a state of happiness, noble, unintelligible, yet clearly indicated. And then, suddenly having reached a certain point from which he was prepared to follow her, after pausing for a moment, abruptly she changed her direction, and in a fresh movement, more rapid, multiform, melancholy, incessant, sweet, she bore him off with her towards a vista of joys unknown. Then she vanished. He hoped, with a passionate longing, that he might find her again, a third time . (Proust 1987: 327, adapted from the translation by Scott Moncrieff 1922)

This is about a musical phrase, or more precisely, about its relation to Swann. Although the translation by Moncrieff uses the inanimate “it”, which seems logical in English, this pronoun does not exist in French. A “little phrase” ( petite phrase ) is feminine, and Proust originally writes about elle (her), which makes the description even more ambiguous. Out of context, the reader could well believe that Swann encounters an actual female person. This is intentional, and over the course of the novel, the petite phrase , which always appears personified, becomes a sort of emblem for Swann’s relation with Odette de Crécy.

It probably takes Proust’s mastery to verbalize sound impressions with such delicacy. On the other hand, that most readers understand his description points to the fact that this kind of interaction is not completely foreign to them. Watt and Ash (1998) have shown that British listeners are prone to relate musical excerpts to traits of personality such as age, gender and emotional states. Once you admit that auditory objects can be “female”, “move” and have affects of “their own”, it is indeed easy to figure as well how they might “charm” a listener like Swann. Or, if you admit that a drum can “speak” like a bird, and that a bird can reflect a dead spirit calling “father” (which is admittedly more complicated for European minds), it becomes understandable that one could relate emotionally to the pulsating sound of the Kaluli ilib drum.

Enchanted listening implements collective auditory schemes pertaining to distinctly auditory ontologies. The ontologies are collective in the sense that they are shared by a significant number of people, allowing for interactions and mutual understandings. Only the listening mode however – the fact of switching to a specifically auditory ontology – can be considered universal.

To conclude this section, let us summarize how enchanted listening differs from the other two modes. Indexical listening is the “default” mode. Its objects are auditory aspects of non-auditory beings or events. The objects of structural listening are hardly auditory at all: the whole process is oriented towards abstract patterns. Enchanted listening is the only mode where auditory objects stand by themselves. This entitles them to a distinct mode of identification (neither indexes nor structures will do) and a distinct mode of relation (auditory objects relate directly to other auditory objects): in other words, an ontology which exists only in audition.

One of the interesting properties of such an ontology is that it brings new beings to social interactions. Contrary to indexes or abstract structures, the objects of enchanted listening do not refer us to something else. They instantiate other beings here and now. Such beings materialize into sound and literally “possess” the acoustic vibrations (see Sartre 1940: 165 for an equivalent in images). Western musical theory also has a concurrent interpretation of a process similar to what we call enchantment. Following Pierre Schaeffer, the severing of indexical ties, and the carelessness for abstract meanings, has been described as a “reduced” form of listening (Chion 1983: 32 sqq., 1994; Schaeffer 1966, Chap. 15 and 20). In Schaeffer’s terms, the listening is “reduced” because it puts the world into “parenthesis”. This is supposed to lead to an objectivation of sound “for its own sake” (Chion 1983: 33). We cannot contend that this might be what some listeners are seeking. But what Schaeffer seems to have missed is that when auditory experience is apprehended “for itself”, the outcome is often a transient irruption of new things and beings into human interactions. All over the world, ethnography shows the actual opposite of a “reduction”: enchanted listening practised as a path to augmented social realities.

Discussion of this distinction and its relevance for anthropology

After presenting this model in various seminars and conferences, the authors have gathered many insightful remarks and objections. Some of these were integrated above, in the proposal itself. Some other frequently asked questions are addressed here. The most important one – regarding the usefulness of such a model – is discussed extensively in the conclusion.

Q: How can you say that these ways of hearing are universal, without a systematic review of all available ethnographies?

→ A claim that something is “universal” can only be proven false (Popper 1959). Therefore, we propose a hypothesis, under a form which can be tested and falsified. If there exists at least one human society where at least one of our A/B/C alternatives is not an option, our hypothesis is false. Until now, we have been unable to find such a society across any of the ethnographies we are aware of. Posture A (indexical listening) is obviously available to all mammals endowed with hearing: all of them are able to infer the presence of something or the occurrence of an event from the sound it makes. Posture B (structural listening) is demonstrated each time humans use systems of sounds with double articulation to convey meaning. Such systems exist in all human societies, and we usually refer to them as “language”. We have shown that structural listening is not used only for “language”, and we are also aware that “language” can involve far more processes than structural listening. Nevertheless, the existence of “language” in any human society is a strong indicator that posture B (structural listening) is a shared human capacity. The real question is whether posture C (enchanted listening) is practised in all human societies. A society without posture C would be one where sounds cannot have unhearable properties (colour, texture or depth, for example), are never connected to each other by causal laws, move only if their source moves physically, and of course cannot be endowed with autonomous agencies. Does such a society exist? We have at least been unable to find it.

Q: Why are there only three alternatives?

→ Each of the three listening postures described here could be broken down further into subcategories. A fourth alternative would not be a subcategory, but rather a posture operating at the same level of generality, and incompatible with the three postures already described. We are not aware of such an alternative, although we cannot exclude that it exists. If it exists, it remains to be seen whether it is observable in all human societies. Our attempt here is not to address all the ways of listening, but only those which are available to all human beings.

Q: Isn’t listening to language distinct from structural listening? You cite Saussure and Martinet who insisted on the peculiarity of double articulation, compared to other sound systems. Wouldn’t double articulation deserve a fourth alternative?

→ Linguistic systems, in the sense of Saussure, are indeed special among other sound systems because they articulate two systems of value: an utterance can be broken into semantic units, which can further be broken down into non-semantic ones. Each kind of unit has value only through its opposition to the other units of its kind. This double articulation has important consequences for the system (Martinet 1965: 8 sqq.), which set human languages apart from other sound systems like musical scales. Nevertheless, this is a special property of an object. It doesn’t imply that the listening mode which apprehends it is special too. In our view, parsing acoustic data for oppositional units is a distinct listening mode, whether the underlying system is articulated at one or two levels.

Q: Are you sure that A/B/C modes of listening are alternative? Could they not be used in conjunction too?

→ The question is whether a person can focus on the same auditory streams in different ways at the same time. There are indeed sound activities which are defined, culturally, as having an interest for several modes of listening. "Songs", for instance, are vocal productions whereby Western listeners deem it interesting to focus on the voice and its melody just as much as on the lyrics. Cognitive experiments found contrasted evidence showing that, in the perception of sung words, pitch and semantic content were treated either independently of each other (e.g. Besson et al. 1998; Bonnel et al. 2001), or as an integrated percept (Gordon et al. 2010; discussion in Schön et al. 2005). The main problem here is that these studies assess only how listeners identify words, pitch patterns and semantic or syntactic incongruities. These are all instances of the same listening mode (i.e. structural). To answer our question, an experiment should investigate, for example, the interplay of structural listening with the particular ontologies of the enchanted mode. To the best of our knowledge this has not yet been tried. There are some indications, however, regarding indexical vs structural processings. Vitevitch (2003) asked people to repeat a list of words spoken by a recorded voice; the recorded speaker was changed midway through the list, but only half of the participants noticed the change. In a more realistic simulation of a phone conversation, Fenn et al. (2011) found even lower results (only 6% of their participants noticed that the interlocutors at the other end of the line had changed). People could of course reliably detect the changes if cued to do so, or if the differences between the voices were overtly salient (Fenn et al. 2011). These findings show that it is certainly possible to use indexical and structural listening in a conversation, but that the normal mode of attending to someone speaking is structural. This sends to the background significant acoustic features which would have been relevant in the indexical mode. Such clues prompted us to present the listening modes as alternatives. This is in line with our general definition of listening, which implies selective focus. However, since attentional processes can, to some extent, be allocated in parallel (Bonnel et al. 2001; Cohen et al. 2012; Demany et al. 2015), we do not make a strong claim about an essential incompatibility between the three modes.

Q: What about the sound producers? Where do they fit into your system?

→ We set aside the logic and postures of sound production as a different topic. In our view, sound producers can adopt any of the three listening postures mentioned above. For example, in a logic of “music” production, the producers are simultaneously primary listeners of their own sound. They can do it in an enchanted or a structural posture, depending on the moment and the kind of music (whether they improvise or play from a score, for example). In “speech”, on the other hand, people hardly ever listen to their own voice. Instead, they usually concentrate on the meanings they wish to convey. These examples illustrate that listening postures are only loosely correlated with sound production postures. The latter would deserve a study of their own.

Q: What exactly do we gain from this theory? After all, we already knew that daily sounds, language and music were different. Is your demonstration not just a sophisticated way of rediscovering the wheel?

→ Our search for listening postures started a few years ago, precisely from our repeated frustrations with concepts such as “sound”, “language” and “music”. We found that these concepts were used inconsistently in current anthropology and ethnomusicology alike. Definitions were seldom given, or, when given, seldom followed (see Ingold’s remark on the “sound” in “soundscape”, for example). We realized that adding yet another definition for these concepts was not the way to go. Instead, we looked for what people did with sounds, and what that revealed about their auditory experience. One crucial fact, for which we tried to maintain a central place throughout our proposal, is that human audition is never entirely constrained by the outside world. The same vibrations become alternative kinds of things in audition, depending on the posture adopted by the listener. The differences are ontological. They affect which things exist in audition and what properties they have, including their interactional affordances. We believe that this is an interesting and possibly new way of framing auditory experiences in anthropology. One of the problems our distinction could address is the relevance and the extension of the concept of “music”.

In place of a conclusion: Music, an old anthropological challenge

As mentioned above, many languages do not have a word to encompass the same activities as the English term “music” (Nettl 2001). It is therefore not an option to “let the informants decide” on what they consider “music”. Anthropologists have to choose between either following local vocabularies of sound activities, or seeking an overarching definition in order to bypass linguistic differences.

The first path (advocated most notably by Feld 1994, 2004, and de Menezes Bastos 2013) has the advantage of sticking closely to the relations drawn by indigenous people. Words, birds, waterfalls, rattles, chanting … all of these can be linked and traversed fluidly, as no prior concepts or distinctions are brought in by the researcher. A problem arises, however, if one attempts to compare such descriptions cross-culturally. “Sound” is no better a concept for these comparisons than “music”. In its acoustic definition (pressure waves propagated from a source and sensed by the ears), it is simply a medium, akin in this respect to light. Hence Ingold’s critique, quoted at the beginning of this article: to look at the way people relate to their surroundings in “sound” (as a medium) would be just as insipid as to look for their relations in “light”. We have shown that the phenomenal objects of auditory experience can be of very different kinds, and that their distinction has little to do with the pressure waves themselves. Moreover, we argue that universal alternatives exist in listening. If this is true, researchers can investigate and cross-culturally compare something somewhat more precise than relations in “sound”.

Many researchers actually already do this by following the second path mentioned above. Throughout anthropological and ethnomusicological literature, “music” is used to refer to a general human ability, comparisons are made between “musics” of various societies, researchers gather for conferences about “music”, and publish in journals about “music”. Outside the specific domain of ethnomusicology, in anthropology at large, the use of “music” and related terms like “singing” goes largely unquestioned. They appear commonly, for example, in ethnographic descriptions of rituals, parties and ceremonial gatherings, with no real discussion of their relevance to local categories of sound production. In ethnomusicology, Blacking’s definition is still pivotal in any discussion of the topic.

At the core of Blacking’s pioneering work stands the idea that music is “humanly organized sound” (Blacking 1973: 3). Blacking’s proposal constituted a powerful scientific proposal in its own time, compounded by a political stance: that music was indeed a shared human capacity (rather than a “gift”, which some people had and others not). But problems arise when attempting to specify how to understand “organized”. Indeed, the definition should automatically encompass not only music, but also language. Most researchers, however, consider these to be distinct, and an extensive bibliography strives to map the relations of music and language, therefore implying their heterogeneity (see Feld & Fox 1994 for a general overview, and Levman 1992 for a discussion specifically related to Blacking’s definition).

It is likewise debatable whether “humanity” binds together an interesting set of phenomena. Some animals are known to produce sounds which are neither predetermined genetically nor clearly functional from a semiotic point of view. On these grounds, some researchers suggest that “music” is a biological ability shared with other species (Keller 2012; Mâche 1992; Martinelli 2009).

Pied Butcherbird singing
Cracticus nigrogularis at Brisbane, Australia.
Pied Butcherbird songs are highly variable, and have been analyzed as musical improvisations (Taylor 2010)
Crédits : Recorded by Keith Blomerley

On yet another line, anthropologists have challenged the centrality of “humans” in the understanding of social interactions in general (Ingold 2011, and Latour 2005, among others; for ethnomusicology see Brabec de Mori & Seeger 2013). Considering the auditory realm in this perspective, the ontological coherence of “ humanly organized sounds” cannot be taken for granted. From an ethnographic point of view, it is actually quite common for humans to locate musical agency outside the human realm. Examples range from Amazonian shamanic cures to Melanesian medium performances, or to rock and hip-hop musicians driven by a “flow”. Computer music adds to this array, because algorithms organizing sounds can only ambiguously be understood as “human”. The “humanly organized sound” conception overlooks the variety of these interactions in an effort to trace back all agencies to human beings (see also Brabec de Mori in print).

Notwithstanding its shortcomings, Blacking’s elegant definition has grounded many ethnomusicological endeavours. Slight alternatives have been suggested, like “human sound communication outside the scope of spoken language” (Nettl 2005: 25), or “sounds produced and organized by a culture” (Nattiez 1987: 95). All such definitions frame music as an object, rather than a process. In a thorough critique of this conception, Hennion (2007) proposed to understand music as an emergent phenomenon. The epistemic consequence is to no longer study the music, but rather how it is fleshed out from a multitude of social interactions. This approach is, however, incompatible with the idea that music could be something universal (Hennion 2007: 351). There is indeed only little likelihood that the same social processes give rise to the same emergent phenomenon in different societies. Understanding music as a social process takes us back to following the variety of indigenous wordings, with no overarching concept to enable comparisons.

The approach we suggest bypasses many of the problems outlined above. To begin with, we do not need to define “music”. We identified three alternative ways of relating to sounds. One of them (enchanted listening) is often triggered by the kind of sound activities that are subsumed under the term “music”. We have shown above that “music” can also be apprehended through the other two listening modes. But, if there is anything specific to it, it is probably due to its privileged link with auditory enchantment.

We believe that the general properties we described for this way of listening – the ontological shift and the mapping of agency onto the sound realm – can account for many effects attributed to “music”. Specific agencies operate in “music” because the things it is made of have particular ontological properties. But we can intentionally switch back and forth between this and other ways of listening. In other words, the enchanted alternative is always (just) an option. As with the other alternatives, it is adopted by an individual, often according to culturally formed suggestions. One of these suggestions could be the ontological category of the sound producer – enchanted listening is linked to human sound productions in some societies – although it cannot be considered a universal condition. The same holds for criteria such as organization. We do not see these as constituting a specific kind of object or activity (“music”). We see them as cultural determinants that tend to orientate the listeners in given contexts towards specific ways of listening.

Music is not universal, in any sense of the word. But enchanted listening is, as a capacity to consider a distinct realm where sounds interact primarily with each other. If this is true, we should also question the implicit assumption that what people describe as colours, movements or beings in sound are “in the end” frequencies, amplitudes and spectral components of air waves. It should be possible to take people seriously and give a positive empirical status to the enchanted things and beings that appear at times in their auditory experiences.

Haut de page


“Selectively attending to auditory objects,” Frontiers in Bioscience no. 5, D202–D212.

Paroles mélodisées. Récits épiques et lamentations chez les Yézidis d’Arménie
, Paris, Classiques Garnier.

Deep listeners. Music, emotion and trancing , Bloomington, Indiana University Press.

“L’action-dans-le-monde. Émotion musicale, mouvement musical et neurones miroirs, ” Cahiers d’ethnomusicologie no. 23, pp. 29–52. Available online: [last accessed November 2017].

“Singing in the brain: Independence of lyrics and tunes,” Psychological Science no. 9/6, pp. 494–98.

How musical is man? , Washington, University of Washington Press.

“Divided attention between lyrics and tunes of operatic songs: Evidence for independent processing,” Perception & Psychophysics no. 63/7, pp. 1201–13.

Religion explained: The evolutionary origins of religious thought , New York, Basic Books.

“About magical singing, sonic perspectives, ambient multinatures, and the conscious experience,” Indiana , no. 29, pp. 73–101.

“Introduction: Considering Music, Humans, and Non-humans,” Ethnomusicology Forum no. 22, pp. 269–86.

Music and Non-human Agency. In Ethnomusicology. A Contemporary Reader , Vol. II, Jennifer Post (ed.), New York/London, Routledge.

Auditory Scene Analysis: The Perceptual Organization of Sound , Cambridge, Mass., MIT Press.

“Three directions in research on auditory scene analysis,” Proceedings of Meetings on Acoustics no. 19, 010021.

Unlocking the groove. Rhythm, Meter and Musical Design in Electronic Dance Music , Bloomington/Indianapolis, Indiana University Press.

“Attention to natural auditory signals,” Hearing Research no. 305, pp. 10–8.

“Sounds”, in Stanford Encyclopedia of Philosophy . Available online: .

Écoute musicale et acoustique: avec 420 sons et leurs sonagrammes décryptés , Paris, Eyrolles.

“Some experiments on the recognition of speech, with one and with two ears,” Journal of the Acoustical Society of America , no. 25/5, pp. 975–9.

Guide des objets sonores. Pierre Schaeffer et la recherche musicale , trans. C. North & J. Dack). Paris, Buchet Chastel/Institut national de la communication audiovisuelle. Available online:, last accessed November 2017.

“The three listening modes,” in Audio-Vision , C. Gorbman (trans.), New York, Columbia University Press, pp. 25–34.

“Meaning and the Specification of Motion in Music,” Musicae Scientiae , no. 5/2, pp. 213–34.

“Introduction: Towards a Theory of Musical Meaning (In India and Elsewhere),” British Journal of Ethnomusicology , no. 10, pp. 1–17.

“The attentional requirements of consciousness,” Trends in Cognitive Sciences no. 16/9, pp. 411–7.

“Auditory attention is divisible: Segregated tone streams can be tracked simultaneously,” Journal of Experimental Psychology: Human Perception and Performance , no. 41, pp. 356–63.

Beyond nature and culture , trans. Janet Lloyd, Chicago/London, The University of Chicago Press [French ed. 2005, Par-delà nature et culture , Paris, Gallimard].

“Humans rely on the same rules to assess emotional valence and intensity in conspecific and dog vocalizations,” Biology Letters no. 10/1, 20130926.

“‘Flow like a Waterfall’: The Metaphors of Kaluli Musical Theory,” Yearbook for Traditional Music no. 13, pp. 22–47.

“Sound as a Symbolic System: The Kaluli Drum,” in David P. McAllester & Charlotte Frisbie (eds.), Explorations in Ethnomusicology: Essays in Honor of David P. McAllester , Detroit, Information coordinators, vol. 9, pp. 147–58.

Sound and Sentiment. Birds, Weeping, Poetics, and Song in Kaluli Expression , Philadelphia, University of Pennsylvania Press [2nd ed.].

“From Ethnomusicology to Echo-Muse-Ecology,” The Soundscape Newsletter no. 8, pp. 9–13.

“Sound Worlds,” in Patricia Kruth & Henry Stobart (eds.), Sound , Cambridge/New York, Cambridge University Press, pp. 173–200.

“Doing anthropology in sound,” American Ethnologist , no. 31/4, pp. 461–74.

“Music and Language,” Annual Review of Anthropology , no. 23, pp. 25–53.

“When less is heard than meets the ear: Change deafness in a telephone conversation,” The Quarterly Journal of Experimental Psychology , no. 64/7, pp. 1442–56.

GAY LESLIE C., 1998.
“Acting up, Talking Tech: New York Rock Musicians and Their Metaphors of Technology,” Ethnomusicology , no. 42/1, pp. 81–98.

“Technology and Magic,” Anthropology Today no. 4/2, pp. 6–9.

“The Technology of Enchantment and the Enchantment of Technology” in J. Coote & A. Shelton (eds.), Anthropology, Art and Aesthetics , Oxford, Oxford, Oxford University Press, pp. 40–63.

“Vogel’s Net. Traps as Artworks and Artworks as Traps,” Journal of Material Culture , no. 1, pp. 15–38.

Art and Agency. An Anthropological Theory , Oxford, Clarendon Press.

“Parfum, symbolisme et enchantement. Terrain , no. 1, pp. 19–34.

“The Future of Irony Studies,” in R. W. J. Gibbs & H. L. Colston (ed.), Irony in language and thought: a cognitive science reader , New York/London, Taylor & Francis, pp. 581–93.

“The Theory of Affordances,” in Robert Shaw & John Bransford (eds.), Perceiving, Acting, and Knowing: Toward an Ecological Psychology , Hillsdale, Lawrence Erlbaum, pp. 67–82.

The Ecological Approach to Visual Perception , London, Taylor and Francis.

“Words and Melody Are Intertwined in Perception of Sung Words: EEG and Behavioral Evidence,” PLoS ONE , no. 5, e9889.

La passion musicale. Une sociologie de la médiation , Paris, Métailié [2nd ed.].

“Hearing Modernity: Egypt, Islam, and the Pious Ear,” in Veit Erlmann, Hearing Cultures. Essays on Sound, Listening and Modernity , Oxford/New York, Berg, pp. 131–52.

Leçons pour une phénoménologie de la conscience intime du temps , trans. Henri Dussort, Paris, PUF.

Cognition in the Wild , Cambridge/London, MIT press.

“Against soundscape,” in Angus Carlyle (ed.), Autumn leaves: Sound and the environment in artistic practice , Paris, Double Entendre, pp. 10–3. Available online:, last accessed November 2017.

Being alive: Essays on movement, knowledge and description , London/New York, Routledge.

“Broad attention to multiple individual objects may facilitate change detection with complex auditory scenes,” Journal of Experimental Psychology: Human Perception and Performance no. 42, pp. 1806–17.

“Zoomusicology and Ethnomusicology: A marriage to Celebrate in Heaven,” Yearbook for Traditional Music , no. 44, pp. 166–83.

“Meaning From Environmental Sounds: Types of Signal-Referent Relations and Their Effect on Recognizing Auditory Icons,” Journal of Experimental Psychology: Applied , no. 10/1, pp. 3–12.

Metaphors We Live By , Chicago/London, University Of Chicago Press.

Reassembling the social: An introduction to actor-network-theory , Oxford/New York, Oxford University Press, Clarendon lectures in management studies.

“The Genesis of Music and Language,” Ethnomusicology , no. 36/2, pp. 147–70.

Music, myth, and nature, or, The Dolphins of Arion , Chur, Switzerland/Philadelphia, Harwood Academic Publishers.

MARTINELLI DARIO, 2009. Of birds, whales, and other musicians: An introduction to zoomusicology , Scranton/Chicago, IL, University of Scranton Press.

La linguistique synchronique. Études et recherches , Paris, PUF.

Éléments de linguistique générale , Paris, Armand Colin.

“Diferentes escutas do espaço: hipóteses sobre o relativismo da percepção e o caráter espacial da audição,” Simpósio Internacional de Cognição e Artes Musicais. Curitiba: Editora do Departamento de Artes da UFPR. MENEZES, Maria Lúcia Pires , pp. 219–48.

A musicológica kamayurá: para uma antropologia da comunicação no Alto-Xingu , Florianópolis, Editora da UFSC.

“Apùap World Hearing Revisited: Talking with ‘Animals’, ‘Spirits’ and other Beings, and Listening to the Apparently Inaudible,” Ethnomusicology Forum , no. 22, pp. 287–305.

Musicologie générale et sémiologie , Paris, Christian Bourgeois.

“An Ethnomusicologist Contemplates Universals in Musical Sound and Music Cultures,” in Nils L. Wallin, Björn Merker & Steven Brown (eds.), The origins of music , Cambridge, MIT Press, pp. 463–72.

The Study of Ethnomusicology. Thirty-one Issues and Concepts , Urbana, University of Illinois Press.

“The Role of Repetition in Perceived Musical Structures,” in Peter Howell, Robert West & Ian Cross (eds.), Representing Musical Structure , London/San Diego/New York, Academic Press, pp. 129–159.

“On similarity, derivation and the cognition of musical structure,” Psychology of Music , no. 32/1, pp. 23–74.

“What Is a Sign?,” in Nathan Houser, André de Tienne, Jonathan R. Eller and others (eds.), The Essential Peirce: Selected Philosophical Writings. Volume 2 (1893–1913) , Bloomington, Indiana University Press, pp. 4–10.

“Flutes, Songs and Dreams: Cycles of Creation and Musical Performance among the Wauja of the Upper Xingu (Brazil),” Ethnomusicology Forum , no. 22/3, pp. 306–22.

“Auditory Attention and Comprehension During a Simulated Night Shift: Effects of Task Characteristics,” Human Factors , no. 58/7, pp. 1031–43.

The logic of scientific discovery , New York, Hutchinson & Co.

Un amour de Swann , Paris, Flammarion [American ed., 1922, Swann’s Way. Remembrance Of Things Past , C.K. Scott Moncrieff (trans.), New York, Henry Holt and Company].

RACY ALI J., 1988.
“Sound and society: The ‘takht’ music of early twentieth-century Cairo,” in James Porter & Ali J. Racy, Selected Reports in Ethnomusicology , Los Angeles, Dept. of Ethnomusicology, UCLA, no. 7, pp. 139–70.

“Reflections on Music and Meaning: Metaphor, Signification and Control in the Bulgarian Case,” British Journal of Ethnomusicology , no. 10/1, pp. 19–38.

Musicophilia: Tales of Music and the Brain, Revised and Expanded Edition , London, Picador.

“Absolute pitch in infant auditory learning: Evidence for developmental reorganization,” Developmental Psychology , no. 37/1, pp. 74–85.

L’imaginaire. Psychologie phénoménologique de l’imagination , Paris, Gallimard.

Cours de linguistique générale , Paris, Payot.

Traité des objets musicaux: Essai interdisciplines , Paris, Le Seuil.

“Musical and Linguistic Processing in Song Perception,” Annals of the New York Academy of Sciences , no. 1060, pp. 71–81.

The Aesthetics of Music . Oxford, Clarendon Press.

“Temporal coherence and attention in auditory scene analysis,” Trends in Neurosciences , no. 34/3, pp. 114–123.

“Music and Religion in Islam,” Acta Musicologica , no. 69/2, pp. 143–55.

“Attention, Awareness, and the Perception of Auditory Scenes,” Frontiers in Psychology , no. 3. Available online: , last accessed November 2017.

“Schaeffer phénoménologue,” in Ouïr, entendre, écouter, comprendre après Schaeffer , Paris, Buchet Chastel/INA-GRM, pp. 53–67. Available online: , last accessed 8 February 2017.

Les différents modes d’existence , Paris, PUF.

“Sublime frequencies: The construction of sublime listening experiences in the sonification of scientific data,” Social Studies of Science , no. 44, pp. 34–58.

“Ritual Wailing in Amerindian Brazil,” American Anthropologist , no. 90, pp. 385–400.

“Blowin’in Birdland: Improvisation and the Australian pied butcherbird,” Leonardo Music Journal , no. 20, pp. 79–83.

“Change deafness: The inability to detect changes between two voices,” Journal of Experimental Psychology: Human Perception and Performance , no. 29/2, pp. 333–342.

“Soulful voices: birds, language and prophecy in Amazonia,” Tipití , no. 8/1, article 1. Available online: , last accessed November 2017.

“A psychological investigation of meaning in music,” Musicae Scientiae , no. 2/1, pp. 33–53. Available online: , last accessed November 2017.

“Attention is critical for spatial auditory object formation,” Attention, Perception, & Psychophysics , no. 77/6, pp. 1998–2010.

Haut de page


1 We are particularly grateful to the participants in the workshop “Sonic beings? The ontologies of musical agency”, which we convened at the EASA Conference 2012. At the Research Centre for Ethnomusicology in Nanterre (CREM-LESC/CNRS/UMR 7186) and the Institute of Ethnomusicology in Graz (University of Music and Performing Arts), many colleagues and students helped us shape our argument over the years. We received important intellectual contributions from Estelle Amy de la Bretèque, Emmanuel de Vienne, Matei Candea, Malik Sharif and Thibaud Aimard-Kesraoui, who reviewed in detail and discussed with us preliminary versions of this text. We are also grateful to the anonymous reviewers who expressed helpful comments on our proposal.

2 Attention plays a crucial role at later stages of auditory scene analysis. It actually also modulates “from the top down” some very early processes of stream segregation (Caporello Bluvas & Gentner 2013; Zobel et al. 2015).

3 This may seem similar to (and is probably inspired by) Pierre Schaeffer’s discussion of preobjective modes of listening (Schaeffer 1966: 113 sqq.). Our study differs in method and ethnographic coverage, but the most important distinction will perhaps appear in relation to “enchanted listening”. In our analysis, the suspension of indexical and structural/semantic interpretations is not a “reduced” listening, as Schaeffer posits, but rather the “augmented” experience of a new auditory realm. We agree with Solomos (1999) in his argument that Schaeffer did not actually consider the dissolution of the “sound object” into distinct ontologies dependent on the listener’s system of knowledge and intentions.

4 “Theaffordances of the environment are what itoffersthe animal, what itprovidesorfurnishes, either for good or ill” (Gibson 1986: 127). As summarized by Gibson himself, the core of his thesis is that “the composition and layout of surfacesconstitutewhat they afford. If so to perceive them is to perceive what they afford.” In other words, appraising action possibilities does not occurafterperception but right within it. A liquid surface, for example, is “sink-into-able” for heavy mammals but “stand-on-able” for water bugs. Mammals and bugs never actually perceive the same surface.

5 This is true as long as the extension is within the range of possibility: compare the string quartet recording with a progressive rock album, where “stereo effects” are employed, so that, for example, the guitar solo circles around the listener or the drums jump from right to left: in that case, the upstream inference of indexical listening would invoke a space with flying guitarists and teleporting drums. This space is not possible, because it cannot exist without changing the ontological properties of reality (on possibility as a mode of existence, see Souriau 2009: 134 sqq.). If such ontological shifts in space occur, we are confronted with another kind of auditory space that is explored in detail in the section about enchanted listening.

6 This competence probably needs learning. Babies, for example, are initially more sensitive to vocal pitch than adults. They must learn to lose some of this sensitivity in order to acquire language (Sacks 2007: 138; Saffran & Griepentrog 2001).

7 Sacks (2007: 182) relates the following discussion with Michael Torke, a “true synaesthete” (who happens to be a composer). Torke explained to Sacks that he vividly saw the colour blue when he heard a D-major chord. Sacks asked Torke what would happen if he listened to D-major when looking at a yellow wall. Would he see green? Torke’s answer was negative: both the musical and the visual colors were “true” colours for him, but they would not mix together. This indicates that even for “true synaesthetes”, auditory colours remain distinct from optical ones.

8 To illustrate : “a flute, no less than an axe, is a tool, an element in a technical sequence; but its purpose is to control and modify human psychological responses in social settings, rather than to dismember the bodies of animals” (Gell 1988: 6).

9 We adopt Gell’s definition of agency: “whereas chains of physical/material cause-and-effect consists of ‘happenings’ which can be explained by physical laws which ultimately govern the universe as a whole, agents initiate ‘actions’ which are ‘caused’ by themselves, by their intentions, not by the physical laws of the cosmos. An agent is the source, the origin, of causal events, independently of the state of the physical universe” (Gell 1998: 16). In theory there are no limits to a chain of cause and effect. In practice, however, infinite causalities are unmanageable for finite brains. At any given time, it is a cognitive requirement to consider that some objects can just do things “by themselves”. At that particular moment, such an object is an “agent” in the subject’s cognition.

10 An in-depth exploration of intra-sound causal principles in Western tonal music can be found in Ockelford (1991, 2004).

11 See the distinction between “true” and “metaphorical” synaesthesia in psychiatry: “For most of us the association of color and music is at the level of metaphor. ‘Like’ and ‘as if’ are the hallmarks of such metaphors. But for some people one sensory experience may instantly and automatically provoke another. For a true synaesthete, there is no ‘as if’ – simply an instant conjoining of sensations” (Sacks 2007: 177).

Haut de page

Table des illustrations

Titre “Buildup” and “breakdown” in a trance music track, analyzed in Butler 2006 : 315.
Légende Abbreviations: bass drum (BD), riff 1 (R1), snare drum2 (SD2), snare drum 3b (SD3b), riff 2 (R2). Musical excerpt from Communication (Somebody Answer The Phone) by Mario Più, 1999, Incentive ‎– CENT2T.
Fichier image/png, 1,0M
Haut de page

Pour citer cet article

Référence électronique

Victor A. Stoichita et Bernd Brabec de Mori, « Postures of listening  », Terrain [En ligne], Lectures et débats, mis en ligne le 18 décembre 2017, consulté le 16 janvier 2018. URL :

Haut de page


Victor A. Stoichita

Articles du même auteur

Bernd Brabec de Mori

Articles du même auteur

Haut de page
  • Logo CNRS
  • Logo Université Paris Lumière
  • Logo Ministère de la Culture
  • Logo Université Paris Ouest
  • Logo MAE - Maison Archéologie Ethnologie
  • Logo FMSH-Diffusion
  • OpenEdition Journals