# Against Hearing Sounds ## Introduction The idea that we hear sounds is thoroughly ingrained in our commonsense understanding of audition. It just seems obvious that we hear bangs, clangs, crashes, and splashes. Philosophical inquiries into auditory perception take sounds as their starting point: there are various accounts of what sounds are and how they allow for the hearing of sources: material objects, involved in movements and interactions such as collisions, breakings, and scrapings, which lead to the production of sound waves. I shall suggest here that common sense and philosophical orthodoxy are incorrect. We do not hear any of the aspects of the world that philosophers have identified as sounds. Rather, source events are heard directly without any type of sonic intermediary. To hear a source, I shall argue, is to hear a material object, meaning that audition is very much like seeing. The phenomenological differences between hearing objects and seeing them can be explained as the result of the way they are perceived in each modality, rather than . . .. Once this approach to source hearing has been laid out, I use it to respond to various objections that could be made to the idea that audition is soundless. ## 1. Hearing Sources, Hearing Sounds ### 1.1 Hearing Sources Let's begin with what everyone agrees: we hear sources. Sources are material objects, which can move or interact in such a way as to vibrate and cause sound waves to propagate through the surrounding air. Providing that those sound waves are above a certain amplitude and within a certain frequency range, they elicit auditory experiences in any listener in the vicinity: if I slam a door, drop a glass, and sweep up the pieces, you will have an auditory experience corresponding to each activity once the sound waves they create reach your eardrums. You will, thereby, hear the door, the glass, and the pieces. While [[George Berkeley|Berkeley]] thought that "in truth and strictness nothing can be heard but sound" (1713/2012, p. 67), it is now very uncommon to endorse such an austere approach; all contemporary philosophers interested in audition accept that source hearing occurs. Sources themselves – the door, the glass, and so on – are included in, rather than inferred from, auditory experience. We experience sources directly. There are several reasons for thinking that sources are heard, not inferred. First, audition provides quite precise details about sources, immediately and without any conscious effort. If I slam the door, you know straight away that two wooden objects have collided forcefully with each other, and if I drop a glass, you know immediately that something fragile has not survived its impact against something more sturdy. The immediacy of this awareness is comparable to vision: the weight and woodenness of a door are revealed as soon as the door is slammed, just as its colour and shape are revealed by a quick glance. Second, sources stand in straightforward causal relations to auditory experiences. Plausibly, one requirement on our perceiving a given thing in our environment is for that thing to have caused our perception of it in 'the right sort of way'. If my visual experience of a tabletop with a certain colour and shape is elicited by the light reflecting from a tabletop with those properties, then it seems correct to say I am seeing that tabletop. If on the other hand, my experience of a tabletop is produced by some other means, say by my taking a hallucinogen, then this will not count as my seeing that tabletop, even if it is right in front of me. Of course there may be difficult cases in which it is unclear as to whether a source is causing an experience in the right sort of way, but this does not detract from the idea that perceivability is constrained by causation. That which stands as the proper cause of a perceptual experience is properly the object of that experience; since sources are proper causes of auditory experiences, sources are proper objects of that experience. Third, the auditory system groups sound waves together based on the sources that have produced them, and in doing so produces auditory experiences which correspond to these sources. In a noisy environment, sound waves from multiple source events will hit our eardrums at the same time, but rather than a single cacophonic mix of all of them, we are in most cases able separate what we hear into distinct streams and choose to attend to one rather than the others: you might focus on me sweeping the floor, but could also decide to eavesdrop on the conversation taking place behind you instead. This grouping is achieved by the auditory system making use of various spectral and temporal features of sound waves. Objects tend to vibrate at multiple, harmonically related frequencies simultaneously, and so the auditory system will tend to group together frequency components which are related in this way; similarly, the frequency components produced by a single source event will be temporally related, likely to begin or change in amplitude synchronously, and so this can also serve as cue for grouping. This suggests that sounds are heard, not inferred, because [spell it out] Fourth, auditory grouping based on source is likely to occur because the purpose of auditory perception is to provide us with information about sources. Our ancestors would have had an interest in those aspects of the environment that they would desire to flee from, mate with, or feed on; that is, with material things, not the air that they disturb. This suggests that sounds are heard, not inferred, because [spell it out] These considerations do not show conclusively that sources must be perceived, but they do show that there are phenomenological, causal, and functional reasons for thinking that an account of audition which allows for source hearing is desirable. My account meets that desideratum, because I hold that we only hear sources; we do not hear sounds as well. I will address cases where we do not seem to hear source in section 3, but I first wish to set out both my view, and the orthodoxy it opposes. ### 1.2 Hearing Sounds Commonplace as it is to think that we hear sources, it seems very natural to say that we also hear sounds produced by these sources. Slamming the door makes a 'thud', smashing the glass makes a 'crunch', sweeping the broom across the floor makes a 'swish'. Many philosophers are convinced that we hear sounds, although they differ over what kind of thing sounds actually are. If sounds are thought of as products of sources, they are most obviously identified with sound waves. This is the default, non-philosophical way of thinking about sounds, and it has some philosophical proponents (O'Shaughnessey 2002; Sorensen 2008, p. 280–283). Similarly, while [[Matthew Nudds|Nudds]] does not argue that sounds are identical to sound waves he still maintains that the two are intimately connected: sounds are "structures or patterns of frequency components ... instantiated by sound waves, that would normally be interpreted by the auditory system as having been produced by a single source event (2010, p. 290). However, we should not assume that the proximal cause of a perceptual experience is necessarily what is perceived. Consider vision: light hitting our retinas is the immediate cause of our perceptual experiences but few would say that we see light; rather, we say that we see the objects from which the light has been reflected. [[Casey O'Callaghan|O'Callaghan]] has made various criticisms of the idea that sounds are sound waves (e.g. 2007, ch. 2), and suggests instead that sounds should be identified with things one step earlier in the causal chain: events in which the air (or some other medium) surrounding the object is "disturbed, or changed, or set into motion" by its vibrations (2010, p. 36). On this conception, sounds do not propagate through the air, but are events happening at the surfaces of vibrating objects. Similarly, in a recent paper, Casati et al. present what they call an "Ockhamization" of source hearing: > The sound is ... identical with what we called the event source. The collision you hear is the sound you hear; the vibration of the stuff the tuning fork is made of is the sound. There is no difference between hearing the sound and hearing the event source, because there is no difference between the sound and the event source ... in point of the fact the identity view clears the metaphysical landscape. As there is no need of preserving the distinction between sound and event source, talk of 'event source' is redundant. You may indeed read the identity in reverse, as somewhat explaining away sounds. According to the reverse reading, what you hear are distal events. We take this to be a terminological matter. Call the items you hear 'distal audible events', call them 'sounds'. (2013, p. 2–3) While this is presented as a way of doing away with sounds in favour of sources, the view retains the basic idea of a sound as something heard, distinct from the actual things making the sound: the sound is the event in which the sources are involved. [perhaps briefly compare a view of perception on which we see both objects, and events in which objects are involved, and say that the events are things separate from the objects]. Although the sound wave theory and the event theory identify sounds with different kinds of entities, they are united in conceiving of sounds as individuals in their own right, distinct from source events. This type of approach I shall refer to as the separate sound view. The main rival to the separate sound view is the property view of sounds. One version of this approach is put forward by Pasnau (1999), who argues that sounds should be thought of not as individuals that objects produce, but rather a feature that they are heard to take on when they vibrate. We might think of this as comparable to seeing lumps of coal adopt an orange glow when they are heated. [[John Kulvicki|Kulvicki]] (2008; 2014) suggests another: sounds are permanent dispositions of objects to vibrate at certain frequencies, which are revealed to listeners when they stimulated similar to how we might think that the colours of objects are revealed to viewers when light is shone upon them. Leddington (2014), takes a slightly different tack, arguing that sounds are not properties of objects, but of source events themselves: we hear the door slamming event as bearing a particular property, rather than the door itself. On all of these views sounds are borne by sources, rather than distinct from them. ### 1.3 Hearing Sources through Hearing Sounds So, many philosophers think we hear both sounds and sources. Those who think so face a question: how does hearing one relate to hearing the other? There are three views in the literature; each has a prima facie problem. Proponents of the separate sound approach tend to suggest that we hear sources through the hearing of sounds. Nudds puts this idea as follows: > what is present in auditory experience—the things of which I am auditorily aware—are the sounds produced by them. I hear concrete material things by hearing the sounds they produce. (Nudds 2016, p. 100) and has suggested elsewhere that perception of sources is achieved by listeners hearing sounds as having been caused by their source events (2010 p. 18). The difficulty with this view is that, even if sounds are heard as having been produced by their sources, it is not obvious why this would allow for source hearing: if we see that a cake has been baked in a tin of a certain shape, we do not thereby see that tin; seeing that smoke has been produced by a fire does not allow for the seeing of that fire (see O'Callaghan 2011 for similar criticisms). The second view, owing to O'Callaghan (2011), is that we hear sources by hearing sounds as as parts of sources, in a similar way to how some have argued that seeing the front facing surfaces of an object allows for the perception of the whole object. This mereological approach can be challenged by pointing to an obvious disanalogy between hearing and seeing: by changing our viewpoint on an object we are able to see different parts of it, but we are not able to hear different parts, or different aspects, of a source by changing our location in relation to it, nor by any other means. Moreover, it is plausible to suppose that when you see the front facing parts of an object, you see them as parts, as belonging to a thing that has other parts you could potentially bring into view. O'Callaghan's response to these difficulties is appeal to the multimodality of perception: hearing a sound is to hear it as being part of an event with components that you could potentially perceive through other perceptual modalities (ibid. p. 24). Although there are very good reasons for thinking that perception is multimodal, the disanalogy with seeing parts of objects remains. It seems we are able to see the front facing surfaces of objects as parts of larger wholes without appealing to the multimodality of perception; arguably someone who could only see (not hear, feel etc.) would still see parts as parts, yet if O'Callaghan is correct about audition, a blind person would be unable to hear sounds as parts of sources, and so would never be able to hear sources. Beyond this, there is a general problem faced by all separate sound views. If what we are auditorily aware of in the first instance are individual sounds, then sources are, as Kulvicki puts it, at "one remove" from listeners (2008, p. 15, see also Pasnau 1999, p. 317), and so those who endorse the separate sound approach will inevitably have to complicate any account of how we hear what we hear. Consider, for example, the view that might seem closest to my own: Casati et al's Ockamization of sound. The issue with this approach is that, while the activities of objects are heard, the material objects which are participating in the activities are not – at least not directly: > The immediate object of auditory perception is just the event that is the sound, and then, in virtue of the perception of it we may recognize the thing source. 'In virtue of' means, roughly, that one has indirect awareness of x when one's awareness of x is caused by one's awareness of something else (Huemer 2001: 55). Therefore, we say that our awareness of the thing source is caused by our awareness of its sound. (Casati et al, 2013, p. 4) Even if we can make sense of the idea that we can hear collisions or vibrations without hearing that which collides or vibrates, the idea that sources are heard indirectly seems problematic when we consider the reasons for thinking that sources are heard in the first place. Audition provides immediate apprehension of object properties such as size or material constitution: my auditory awareness of a collision seems no less instantaneous than my auditory awareness that the collision involved two heavy, wooden objects. Furthermore, these same properties determine in part the frequencies at which objects vibrate and so the frequencies of the sound waves that they produce, and can therefore be thought of as standing in straightforward causal relations to auditory experiences. Finally, our ancestors would have had more of an interest in the material objects participating in source events than in the activities themselves: it is important to know that it is the heavy feet of a predator that are hitting the ground, rather than the light ones of potential prey. It is therefore plausible to think that our auditory systems function to provide information about objects, more than they do about events. Given these considerations, I suggest that we take the exact opposite approach to Casati et al. and place material objects centre stage: to hear a source is to hear a material object. We might think that property views of sound allow for this. They postulate only one type of audible individual, the source event itself. They might thereby allow listeners to go, as it were, straight to the source. However, on further consideration, we can see that property views do not get us that much closer to sources after all. If we hear sources to bear only sound properties, then much of what is of interest to listeners remains, in fact, at one remove. First, notice that on property views, audition involves the attribution of only one type of property. This contrasts with vision and touch which we think of as allowing for the perception of objects as bearing multiple types of property simultaneously: we see objects to have a colour, texture and shape (e.g Nanay 2013, p. 38, Burge 2010, p. 24); we feel objects as having a texture, temperature, and weight (e.g. Fulkerson 2011). This difference alone might make us suspicious of property views, but there is a greater problem. If we hear sources as having only sound properties, then the other features of the object, features that we have an interest in, that are immediately apparent from auditory experience, and that determine the characteristics of of the sound waves produced, are still at one remove. Slamming a door reveals it to be wooden and heavy, smashing a glass that it is small and flimsy; we are able to tell immediately that the door was slammed rather than scraped or that the glass was smashed rather than rolled, but if sources are heard as barren of everything but sound properties, awareness of all of these aspects remains indirect. If this is correct, then the idea that sounds are properties in fact faces a very similar challenge to the idea that sounds are distinct individuals. That is, how do we become aware of these interesting aspects of sources? The options that the proponent of properties has are exactly the same as those available to those who endorse separate sounds: interesting properties could be inferred, or they could be heard as standing in some sort of causal or mereological relation. However, if this type of work still needs to be done, the advantage that property views have over separate sounds is negligible. ### 1.4 Soundless Audition Here are three ways that the debate about the relation between sounds and sources might continue. First, proponents of various views could give more reasons as to why their approach is preferable. Second, we could give up on source hearing and endorse an austere, Berkeleian, approach: we hear sounds and nothing more. I suggest we choose a third option: give up on sound hearing. On a soundless account of audition we hear source events their properties directly, not via the hearing of any special sonic individual or property. We have seen that there are various reasons for wanting an account of auditory perception which allows for the perception of sources, but we have also seen that complications arise when we try to make sense of the idea that sources are heard through sounds. Such complications could be avoided if sounds are discarded. --- ## 2. The What and the How of Hearing Sources As mentioned above, the idea that we hear sounds – however they are conceived – is thoroughly ingrained in philosophical and folk conceptions of audition. The ingrained idea leads to the thought that sounds must feature in an account of auditory perception. This necessity cannot be insisted upon as a matter of brute intuition derived from ordinary language. Consider this statement concerning vision: > Whatever we see, we definitely see appearances. Anything else we might see –material objects, rainbows, smoke, etc.– we must either see via the seeing of appearance, or in seeing appearances. It is important therefore to investigate not only how seeing appearances allows for seeing of other types of thing, but also the nature of appearances themselves: are they properties, individuals in their own right, or something else? It is certainly true that we can, and do, speak of seeing appearances. But to specify, up front, that appearances exist and are the primary explanandum of a theory of vision is evidently unjustified; to do so is to hypostaise hastily from ordinary language, and almost certainly to hinder the development of a proper account of vision. The proponent of appearances owes us more by way of explication of why we must include appearances in our theory. Similarly, the proponent of sounds owes us explication of why we need to include them in our theory of audition. My strategy here is to show that once we have a robust account of what source hearing is, the reasons for thinking that sounds must figure in audition become much less compelling. I develop the account in three stages. First, I lay out the key claim: audition is like vision, in that both modalities allow for the direct perception of material objects. I then consider the claim that we need to invoke sounds to explain the dissimilarities between hearing and seeing, and argue that these stem from structural differences between the two modalities. Third, I consider the claim that we need to invoke sounds to explain how the same source can be heard in different ways, and argue that this can be explained by appeal to perspectives. The theory thus developed provides the means with which to respond to various objections to soundless audition that I will examine in Section 3. ### 2.1 Hearing Material Objects *[Section to be developed: audition is like vision, in that both modalities allow for the direct perception of material objects.]* I have already shown that everyone thinks we hear source, but since I want say we only hear sources, I should be precise. My claim is that the primary objects of audition are material objects, heard directly. [stuff on hearing material objects] This view can be illuminatingly contrasted with some stuff Kulvicki says. In recent work, he has moved away from the idea that sounds are dispositional vibratory properties, and suggests that they should instead be thought of as "audible profiles": > Sounds are ... features of ordinary objects and happenings, in addition to perspectival features that abstract over them. Since audible profiles are abstractions over features of objects, events, and environs, those features are also natural candidates for being heard. These include dispositions to vibrate in response to being struck, aspects of the events that cause vibrations, such as intensity, and properties of the spaces in which one hears. The cymbal sounds differently when struck with force than when lightly tapped, for example, in a large room, or close-quarters. In some contexts, it is natural to identify what is heard with a feature of the source (the bell has a nice sound), while in others the audible profile is a better candidate—the bell's sound is quiet and washed out from so far off. (REF) One could read Kulvicki here as proposing an ontological novelty, an auditory profile, as identical to sounds and as the genuine object of audition—much as Mark Johnston proposes sensory profiles (Johnston 2004). If this is so, the proposal faces similar objections to Johnston's, concerning the obscurity of the novelty, its dubious suitability as an object of perception, and the whiff of indirect realism hanging around the theory (see Dunn 2008). However, this seems a bad reading, Since he goes on to say: "Such a catholic view of sounds renders them theoretically uninteresting. Or, to put it differently, it renders them just as interesting as sights are to vision". (ibid. p. 91) This suggests that the term 'sound' should be understood as a means of describing the totality of features, properties and objects that are perceived rather than a unique object of audition". This suggests that his view is actually that "sounds" are rather like "appearances". It is grammatically fine to say that the ghost has my mother's appearance, but ontologically suspect (theoretically uninteresting) to say that there is a thing, the appearance, that I see when I see both the ghost and my mother. On this JK and I agree, more or less. The disagreement is over what the genuine objects of audition are. JK seems to suggest they are features or properties of objects. I don't want to deny that we are sometimes best described as hearing properties, but I hold that, usually and paradigmatically, we hear objects. Compare: we see colours; we see objects that have colours; perhaps even we see objects by seeing their colours; but the primary things we see are objects. Another proposal contrasting with mine is that we do not hear sources, or objects; rather, we hear events in which they are involved. Indeed, a lot of the literature talks of source events, not sources. And one might think this because of the following: objects are not always making noises. But they are always being visible. So in the vision case, it makes sense to talk of seeing objects, but in the audition case, it makes more sense to talk of hearing events. This is silly. (I have some thoughts about why . . . though does JK discuss this in his paper?) So my proposal is that we hear material objects; these are our primary auditory objects; and we hear them directly, not by hearing sounds. I will now nuance this view by discussing two reasons why one might think we need sounds as well as objects. ### 2.2 Hearing the same object differently One reason you might have for thinking that we need to invoke sounds is that, without them, we can't explain how it is that auditory experiences of the same object vary depending on the activity in which they are participating. A cymbal hit forcefully sounds different than does a cymbal brushed lightly. So, surely, there are two different sounds here; we hear the sounds, not just the source, since the source is the same in both cases. We can solve this problem by borrowing Kulvicki's idea of perspectival features. While Kulvicki couches these ideas in terms of what sounds are, and seems to allow that sometimes it is profiles rather than sources that are heard, the idea that sources are heard perspectivally provides one way of explaining experiential variation. It is sometimes suggested that the reason why the visual appearance of objects differs depending on viewpoint or how they are illuminated is that these types of viewing conditions affect the perspectival properties the object is seen as bearing. A plate looks different depending on whether it is viewed from above or to the side, or in bright light or dim light, because it is attributed different perspectival properties in each case. Kulvicki suggests that auditory variation can be accounted for in a similar way: the difference between hearing a cymbal struck forcefully or lightly need not be explained in terms of a different separate sound being produced in each case, but as a result of the same object being heard with different perspectival properties. Note also that we need not invoke sounds to explain auditory constancy (REF). Objects are often immediately identifiable regardless despite variations in how they sound: a cymbal still sounds like a cymbal regardless of whether it is struck gently or forcefully. Kulvicki thinks that we need to invoke sounds, and identify them with vibratory dispositions, in order to explain this. (REF). However, if, as I have suggested, we hear objects as bearing properties such as size or material constitution, we have an alternative way of accounting for auditory constancy: a cymbal still sounds like a cymbal regardless of how it is hit, not because we hear its dispositional vibratory property in both cases, but because in both cases we hear it to be a metal object of a certain size. All in all, then, we need not invoke sounds to explain differences (or constancies) in experiences of the same object; we can do so by arguing that sources are experienced perspectivally. ### 2.3 Why are hearing and seeing different? Although perspectival properties provide an explanation as to how auditory experiences of the same object can differ, one which could be incorporated into the soundless approach they do not provide an explanation of what makes audition unique phenomenologically. If we see and hear objects, their properties, and their activities, what makes hearing so different from seeing? It is important for a soundless approach to be able to meet this challenge, as we might think that one motivation for postulating sounds in the first place is that they help explain what makes audition distinctive: hearing differs phenomenologically from seeing because we hear special types of individuals or special types of property. Many properties or features are ones that we could just as well see as we hear. Why then is the experience of hearing a cymbal being struck hard so different from seeing the same event? Similarly, Kulvicki notes that seeing a collision involves seeing distinct objects moving through space, but that hearing one involves hearing a "amalgamated mess that privileges none [...] as the primary object heard" (ibid. p. 86), or a "blast" (ibid. p.87). I suggest that the differences between auditory and visual phenomenology are best explained by appealing to structural differences between the two modalities rather than differences in . . .. That is, we should think of hearing objects as differing from seeing objects in the same sort of way that touching objects differs from seeing objects. Visual experiences are – obviously – quite different from tactile ones, yet we do not normally take this as evidence that we are seeing but not touching material objects, or vice versa. Rather, the difference can be explained in terms of the way objects are perceived in each modality. For example, it is often said that vision possesses a certain structural feature which touch does not: a spatial field. When my eyes are open I take in a portion of space, some parts of which are empty, some parts of which are filled by material objects. As Martin puts it: > One is aware of the location of visual objects not only relative to other visually experienced objects, but also to other regions of the spatial array—regions where nothing is experienced, but where something potentially could be. (REF) Touch, however, is not field-like in the same way, in that it does not present a space with objects extending across parts of it; rather objects are felt as pressing on the skin and so are tied to our perception of the body and bodily sensations (Martin 1992, Richardson 2010). This structural difference leads to the shapes of objects being perceived in different ways: spherical objects are seen as spherical, and bell-shaped objects are seen as bell-shaped, in virtue of the spatial field allowing them to be perceived as extending through space; perceiving an object as a sphere or as bell-shaped is quite different: an object's shape can be felt in the way it moulds our grip as we hold it, or through the sensations at our fingertips as we run them across its surface. On the account of source perception which I am proposing here, the differences between hearing an object and seeing it can also be accounted for by appealing to structural differences. Material objects are heard directly, but in a characteristically auditory way. What way is this? Elsewhere, I have tried to bring out how hearing objects move differs from seeing objects move, as well as how hearing space differs from seeing space (withheld to facilitate blind review). Here, I will focus on an aspect of hearing that will be important when considering various potential objections to soundless audition in the next section: audition's presentation of shape, or lack thereof. We have just seen that, in different ways, material objects are seen and felt as extending in space. In both cases, the table is perceived as having spatial boundaries, edges where the object ends and the rest of the world begins. Audition, however, does not present material objects as spatially extended in the same way. While we often hear what we hear as located in relation to us (to the left, behind, etc.), and audition often provides awareness of the approximate size of objects, we do not hear objects as filling out a portion of space: one cannot hear the contours of the door as it closes, or the exact shape of the fragments of a shattering vase. O'Callaghan (REF), drawing on arguments that have been made in work on visual perception (e.g. Siegel 2006, p. 434; Martin 2007, p. 706), has argued that this inability to hear the shape of objects gives us reason to doubt that they are heard at all. Regardless of how plausible such an idea is for vision there is a straightforward reason why the same requirement cannot be made of audition: whichever aspects of the world we do hear, we do not hear as having spatial boundaries. Using spatial extension as a means of deciding what is heard would therefore exclude any and all aspects of the world, including medium disturbances (O'Callaghan's preferred candidate for what sounds are), sound waves, object-less source events, and spaces. Rather than being understood as evidence that we do not hear material objects, a lack of spatial extension can be taken as revealing how we hear them. In vision and in touch objects are presented as extending over portions of space; audition presents objects as having properties and as occupying certain locations, but not as filling out space in one way or another. --- ## 3. Sourceless Audition? So far, I have suggested that a soundless account of auditory perception is both desirable and plausible, and that source hearing is best understood in terms of material objects being perceived in a characteristically auditory way. How might a defender of sounds respond? A promising strategy would be to show that while sources can be heard, and perhaps are even usually heard, they are not always heard. If it can be shown that we occasionally have auditory experiences which are not of material objects then we might presume that what is heard in these cases are sounds. Showing that sounds are sometimes heard, even if not in all cases, would be enough to defeat soundless audition, but a defender of sounds might go further and argue that a neater and more parsimonious account of audition is to say that sounds are, in fact, heard in all cases, but that they only sometimes allow for the mediate perception of sources. Either way, inquiries into auditory perception would return to questions concerning the nature of sounds, and the best way to explain mediate perception of sources. In this section I will examine three types of case which could be put forward as examples of sourceless audition, and argue that the defender of soundless audition has a plausible response available for all of them. In doing so, I hope to chip away at the intuition that sounds must figure in auditory experience, as well as highlighting the versatility of the soundless approach. ### 3.1 Unknown Sources One way of arguing that we sometimes hear sounds but not their sources is to point to cases in which we are unable to identify a source from the auditory experience it elicits. If we really do hear the activities of material objects, why are we sometimes unaware of what those objects and interactions are? The idea that being unable to identify a source counts against hearing it, can be found in remarks from Campbell: > Suppose that outside in the street you hear a bulldog and a pekingese start to fight; you hear the whole thing from the initial growls and yaps to the final triumphant or defiant yowls, and can tell exactly when the one dog leaves off and the other begins. Well, you might say, that bulldog put up a good fight, but it had no chance. Here the demonstrative 'that bulldog' is an auditory demonstrative—you may at no point have looked at the scene—but it refers to an ordinary physical object. In contrast, if you simply hear a grinding noise in the street outside and have no idea what the source is, you may simply refer to the noise itself, you may say that you wish that noise would stop (1997, p.65–66) Here it seems we have a clear contrast between sounds allowing for the hearing of sources in one case, the yaps and growls let us hear the bulldog and the pekingese, but not in the other, the grinding noise is heard as simply a grinding noise, detached from whatever produced it. Experiences similar to this latter case are not unusual, we often hear something but must turn our heads, or switch on the light, to learn exactly what is happening to what. However, a proponent of soundless audition can deal with such cases quite easily. There are innumerable situations in which we see something yet cannot identify it. One might look out of the window in the middle of the night, see something lurking out there in the gloom, but not be able to tell what that something is. Say that something is a fox. The fact that we cannot identify it as such, does not mean that we are not seeing that fox, or seeing something other than that fox: we simply do not see it clearly enough for it to be identifiable. Notice also that in both the visual and auditory cases, while full identification is not possible, we are still presented with some properties of what is being perceived. The lurking thing might be seen as having a certain size and shape, we can be sure that there is not an elephant or a tank out there; similarly, by describing whatever is heard as a grinding noise Campbell shows that he has auditory access to some of the source's properties: metallic objects moving against each other grind, dogs, rubber balls, and falling raindrops do not. The idea that difficulties identifying a source counts against our hearing it can also be found in the following argument by Matthiessen: > knowledge about which sound we hear is sometimes not very informative. Often we want to single out the thing that made the sound. The sentence "I'm listening to the chirping of one of those five goldfinches, but I can't tell which one I hear" can describe an everyday situation, whereas "I'm looking at one of those five goldfinches, but I can't tell which one I see" sounds a bit stupid. Obviously, the sentence "I hear a chirping" does not go as far as the sentence "I hear this bird ... chirping". Talk about hearing things is richer than talk about hearing sounds. It includes a thesis about which thing is the source of the sound ... Sometimes it not so easy to find out which thing we hear, and then we have to turn our head or get closer to single it out. (2010, pp. 96-97) Whereas in the Campbell example the suggestion is that we hear a sound and not a source because we are not able to identify the type of source we are hearing, in this case the idea is that we must hear a sound rather than a source, because we are unable to match up what we hear to one of the things that we see. What Matthiessen's argument reveals, I suggest, is a certain bias towards visual perception, one which we have no reason to accept. To see why this is let us again consider touch. Imagine you and a friend are sat across from each other at a dinner table. You shift your foot so that it rests against the table leg ... until it moves. Your foot was not actually touching the table but rather your friend's foot. In this case you have perceived your friend tactilely and visually, and failed to match up what you are seeing with what you are touching, but you are still seeing and touching the same thing. Your failure to match your experiences up can be accounted for by the fact that you are seeing and touching different parts of your friend (and that you are not seeing and touching the same part), and also the fact that, as mentioned earlier, touch and vision are structured differently and so present the world in different ways. With this in mind, imagine you are now sitting in a forest and that five birds are singing: one is cheeping, one is chirping, one is cooing, one is warbling, and one is twittering. A single bird is perched in your field of view, but you cannot tell whether she is the cheeper, the chirper, the cooer etc. In this situation, it does not seem at all stupid to say "I am looking at one of those five birds, but I can't tell which one I see". We saw in the previous section that audition does not present individuals as spatially extended in the same way that vision does, and, on occasion, this impedes our ability to connect what we hear with what we see: we do not hear the five birds as extending over a particular portions of space, and so it is difficult to match one of these individuals with the one that we do see as taking up a particular location in space. This, however, is not a reason for thinking that we hear material objects any less directly than how we see them. The proponent of soundless audition can say that, just as in the tactile case, we can perceive the same individual via two perceptual modalities simultaneously yet not be able to match up one experience with the other. By assuming that the demonstrative 'those' must be cashed out in visual rather than auditory terms, Matthiessen begs the question. ### 3.2 Audio Playback as Sound Production So far I have focused on cases of what we might call natural source events, those which involve real-life things, moving or interacting with one another. However, in the 21st century, many of the auditory experiences that we enjoy have technological causes: the sound waves which reach our ears often do not come from objects bashing, breaking, or scraping together, but from stereo speakers, televisions, and mobile phones. Auditory experiences elicited by devices such as these present another way in which the case against soundless audition could be made. Indeed, Despite a recent move towards thinking that sources can be heard directly, rather than through sounds, Soteriou thinks that hearing recordings speaks against ubiquitous soundless audition: > when we hear recorded sounds, the bearers of acoustic properties that we hear are pure audibilia – pure audibilia that we can hear without hearing (or misrepresenting) their 'source events' (ibid. p. 5) If you use your mobile phone to record me slamming the door and dropping a glass and then play it back later, anyone listening will have auditory experiences similar to those that would have been produced by the actual door slamming or glass smashing. Similar, but not exactly the same: it would be obvious in this case, and in most cases of auditory playback, that you are listening to a recording, rather than the real thing. Given that auditory experiences are occurring in the absence of actual slammings or smashings, this types of case seems like a very good candidate for sourceless audition. Moreover, invoking separate sounds allows for a straightforward explanation of what is going on here: the phone's speakers are producing imperfect copies of those sounds made by the original source events, which are similar enough for the original source be determined but dissimilar enough to reveal that they are not the originals. Audio technology therefore seems to allow, as [[Roger Scruton|Scruton]] puts it, sounds to be "detached completely from their source ... and listened to in isolation" (2009, p.58). If we always hear sources, as the soundless account of audition specifies, how do we account for the fact that recordings allow auditory experiences as of source events which are not presently occurring? How do we account for the fact that most recordings are nevertheless recognisable as recordings? Soteriou mentions, and dismisses, a tempting strategy for a soundless account of recordings. Perhaps pressing play leads to illusory experiences of doors slamming and glasses smashing? Like Soteriou, I do not think this approach is especially promising. First, given the ubiquity of recordings, it would mean that auditory illusions are rampant in everyday life, and philosophical theories on which we are constantly subject to illusions are generally considered undesirable. Second, such an account does not explain why we are almost never taken in by these illusions. That is, why recordings are so obviously recordings? Despite these challenges I think that a plausible, informative, soundless account of recorded audio is available, one which explains how we can have auditory experiences as of sources which are not presently occurring, accounts for the experiential differences between hearing recordings and hearing real life events, and does not rely on illusion. We can begin by noting that there is a type of visual experience which is commonplace, which perceptually presents objects which are in actual fact absent, and which is not normally considered illusory. Pictures surround us, are readily identifiable as such, and elicit non-illusory visual perceptions of absent objects. These, I suggest, should be soundless audition's model for audio playback: pressing play leads to the perception of an auditory image. As we shall see, we do not think of photos, videos, or any other type of images as producing a novel type of individual which is perceived in lieu of ordinary material objects. Rather, we think of images as material objects which elicit a unique type of experience. If recordings are images, they should not be thought of as producing sounds, but as a special type of source. What makes experiences of images special is that they exhibit what we can refer to as twofoldness. That is, when we look at a picture we experience two aspects simultaneously: that which is depicted, and that which does the depicting. According to Martin to see an image as an image, is to see an object as having "an appearance which it does not exemplify" (2012, p. 343); [[Robert Hopkins|Hopkins]] says that the distinctive phenomenology of picture perception arises from having "an experience whose content somehow includes the pictures' object" yet one which "remains a way of seeing the picture" (1998, p. 16). Looking at a photo, you have a visual experience as of the picture's subject, the face of your beloved, for example, but you also, simultaneously, experience a shiny piece of paper covered in coloured blotches. Hopkins also points out that our experience of these two aspects is integrated: looking at the photo is not like having two separate experiences of different objects, or of looking at one object and imagining another; rather, to see what is depicted is inextricably tied up with seeing the depictor. On occasion we fail to see a picture as a picture, and duality is not present in our experience. One way this can happen is when we fail to detect the picture surface, and thereby mistake a picture of an object for the actual object, as when looking at a trompe l'oeil. The reverse is also possible: you can be presented with a white canvas covered in irregularly shaped black splodges and see it as simply that. However, if I tell you that you are actually looking at a picture of a dalmatian, your experience, often immediately, changes, you suddenly see the dalmation in the canvas, and thereby the picture as a picture. That twofoldness is a feature of picture experience is fairly uncontroversial, although there are disagreements as to how exactly it should be characterised. It is also important to realise that twofoldness is an attempt to characterise how it is phenomenologically, to look at a picture and not an account of what picture perception consists in. Rather, duality is a feature of our experiences of pictures that theories of picture perception seek to explain. Some accounts, for example, argue that duality is best accounted by viewers experiencing marks on on the picture surface as resembling real world objects (e.g. Hopkins 1998), others think that pictures engage our recognitional capacities (e.g. Lopes 1996), or that looking at a picture involves a special type of pretence that the marks on the picture surface are actually real world object (e.g. [[Kendall Walton|Walton]] 1990). If what I will suggest now is correct, then these controversies can be transposed to audition. Is auditory image perception best explained in terms of resemblance, recognition, or pretence? Here, however, my intention is simply to show that duality is also a feature of our experiences of recordings, that they can therefore be thought of as a type of image experience, and thereby defend soundless audition. That hearing a recording involves an auditory equivalent of one of pictorial experience's folds I take to be fairly obvious. Just as the photo elicits a perceptual experience as of your beloved, the recording elicits an experience as of a door being slammed and a glass being smashed. Regardless of whether we interpret this 'as of' as referring to an actual perceptual experience of a smashing, or just 'seeming to hear a smashing, it is clear that hearing a recording of the event has a great deal in common, phenomenologically, with hearing the real life smashing. That there is an auditory analog of seeing the surface of a picture is less obvious and will require some justification. First, we should be careful to specify exactly which aspect of the physical environment serves as the surface of the auditory image. Soteriou, drawing on Scruton, points out that: > The mechanical behaviour of the radio is not revealed to one in auditory perception ... there is a respect in which the radio itself is not an object of our interest in auditory perception. That doesn't quite entail that we can listen to the sounds produced by the radio without listening to the radio ... However, it does offer some support for the idea that we can focus on sounds in auditory perception without taking ourselves to be latching on to events we can see, and while having no auditory interest in the visually perceivable object producing such sounds. (forthcoming, p. 39) This is, to some degree, correct. We do not hear the mechanical workings of playback devices, at least in most cases. It is often impossible to tell just through listening whether the recording you hear comes from a needle working along a record's groove, or a magnetised tape passing over a tape head, or the data contained on an mp3 file being converted into electrical signals. The same is true of cinematic images: watching a film in the cinema does not reveal the workings of the projector, and watching an (old fashioned) television does not reveal the activities of the cathode ray tube. But this does not count against the idea that watching a film or television series is to see a type of image. This is because the second aspect of duality, the real world object that features in our experience, is not the contraption used to produce depictions on the screen, but the screen itself. Similarly, I suggest, the real world source that we hear when we hear a recording is the playback device's speaker, not the mechanisms which caused the speaker to vibrate. In Section 2 I made some suggestions about how we should understand the hearing of material objects. Exactly the same considerations apply in the case of speakers. First of all, I argued that while it is perfectly plausible to think that we hear material objects as located in space, we do not hear them as bounded individuals which extend through space. Hearing a speaker, therefore, differs from seeing the surface of a picture because we do not hear speakers as spread out in three dimensions, but this should not detract from the idea that speakers feature in our experience of recordings in the same way that the picture surfaces do when we view images. Second, in seeing a picture surface we see some of the properties of that picture surface: its size and shape, whether it is matte canvas or shiny photo paper. The same is true for auditory images. As with hearing real life source events, listening to a recording provides the listener with immediate perceptual awareness of some of the material properties of the speaker. We can tell, for example, whether the recording we are listening to is being played back on a tiny, tinny radio speaker, or medium sized stereo speaker, or an enormous public address system. Perhaps less obviously, there is also some reason to think that we hear what materials speakers are made from. Cheaper speaker cones tend to be made out of paper or plastic, and more expensive ones from fibreglass or carbon fibre. Although only audiophiles will be able to say exactly what a speaker is made from just by listening, everyday listeners will still be able to hear a difference between different types of speakers, and we might presume that this is because they are hearing these different material properties. Finally, we can also note that we can fail to perceive recordings as recordings in similar ways to how we sometimes fail to perceive images as images. Although uncommon, expensive speakers can provide such high quality audio reproduction that, in the right conditions, they can fool a listener into thinking that they are hearing a real life event, rather than a recording of it. That is, we can sometimes have trompe l'oreille experiences. Arguably, auditory equivalents of the other type of duality failure can also be found. In the phenomena known as sine wave speech listeners are presented with recordings of people speaking that have been artificially manipulated so that, at first, they are entirely unrecognisable as speech, sounding instead like artificial bleeps and bloops. If, however, listeners are told which words are being said, then on second listening, they are able to hear the utterance in the artificial noises. That is, once a listener is suitably informed the words on the recording 'pop out' in the same way that the dalmation pops out of the black and white spots on the canvas for a suitably informed viewer. The preceding considerations, I suggest, show that, rather than being a problem for soundless audition, experiences of recordings demonstrate quite how rich it can be. We hear material objects not sounds, but can experience them in a variety of different ways. --- ## Conclusion Here I have argued that a soundless account of auditory perception is both desirable and plausible. We have seen that there are good reasons for wanting an account of audition which allows for source perception, but that sounds, conceived of as either individuals or properties get in the way, of our hearing sources. If this is correct, then the onus is on those who endorse sounds to show that they are necessary, rather than the defender of soundless audition to show that they are not. I have suggested that source hearing should be understood as a way of perceiving material objects, given some suggestions as to how hearing objects differs from seeing them, and argued that once these differences are taken into account, arguments purporting to show the necessity of sounds become much less compelling. *[Tiny bit more]* --- ## Bibliography Aristotle, "De Anima" in The Complete Works of Aristotle the Revised Oxford Translation Edited by Jonathan Barnes. Oxford: Oxford University Press. 1984. Berkeley, George, Three dialogues between Hylas and Philonous Peterborough: Broadview Press. 2012 Bregman, Albert S. Auditory scene analysis: The perceptual organization of sound. MIT press. 1994. Campbell, John, and Michael Martin. "Sense, reference and selective attention." Proceedings of the Aristotelian Society, Supplementary Volumes 71. 1997: 55-98. Currie, Gregory. Image and mind: Film, philosophy and cognitive science. Cambridge University Press, 1995. Fowler, Carol A., and Bruno Galantucci. "The relation of speech perception and speech production." in The handbook of speech perception. 2005: 632-652. Fulkerson, Matthew. "The unity of haptic touch." Philosophical Psychology 24, no. 4 (2011): 493-516. Gaver, William W. "How do we hear in the world? Explorations in Ecological Acoustics." Ecological Psychology 5, no. 4. 1993: 285-313. Gaver, William W. "What in the world do we hear?: An ecological approach to auditory event perception." Ecological Psychology 5, no. 1. 1993: 1-29. Hopkins, Robert. Picture, Image and Experience: A Philosophical Inquiry. Cambridge University Press, 1998. Kulvicki, John. "The nature of noise." Philosophers' Imprint 8, no. 11. 2008: 1-16. Kulvicki, John. "Sound stimulants: Defending the stable disposition view." in Perception and its Modalities edited by Stokes, Dustin, Mohan Matthen, and Stephen Biggs, Oxford University Press, USA, 2014. Kulvicki, John. "Auditory perspectives." in Current Controversies in Philosophy of Perception edited by Bence Nanay, Routledge, Taylor and Francis. 2016: 83. Leddington, Jason. "What we hear." in Consciousness inside and out: Phenomenology, neuroscience, and the nature of experience, Springer, Dordrecht, 2014: 321-334 Lopes, Dominic. Understanding pictures. Oxford: Clarendon Press, 1996. Martin, M. 'Sight and Touch', The Contents of Experience, edited by Crane, Tim. New York: Cambridge University Press. 1992 Matthiessen, Hannes Ole. "Seeing and hearing directly." Review of Philosophy and Psychology 1, no. 1 2010: 91-103. Nanay, Bence. Between Action and Perception Oxford University Press, 2013. Nudds, Matthew. "Sounds and space." in Sounds and Perception: New Philosophical Essays. edited by Nudds, Matthew, and Casey O'Callaghan Oxford University Press, 2009 Nudds, Matthew. "What are auditory objects?." Review of Philosophy and Psychology 1, no. 1 2010: 105-122. Nudds, Matthew. "The Distinctive Role of Sounds and Smells" in Current Controversies in Philosophy of Perception. edited by Nanay, Bence. Routledge, Taylor and Francis. 2016 O'Callaghan, Casey. Sounds: A philosophical theory. Oxford University Press, 2007. O'Callaghan, Casey. "Sounds and events." in Sounds and perception: New philosophical essays edited by Nudds, Matthew, and Casey O'Callaghan Oxford University Press, 2009. O'Callaghan, Casey. "XIII—Hearing Properties, Effects or Parts?." In Proceedings of the aristotelian society, vol. 111, no. 3 pt 3, pp. 375-405. Oxford, UK: Oxford University Press, 2011. Remez, Robert E. "Sine-wave speech." Scholarpedia 3, no. 7 2008: 2394. Richardson, Louise. "Seeing empty space." European Journal of Philosophy 18, no. 2. 2010: 227-243. Schellenberg, Susanna. "The situation-dependency of perception." The Journal of Philosophy 105, no. 2 2008: 55-84. Scholl, Brian J., and Patrice D. Tremoulet. "Perceptual causality and animacy." Trends in cognitive sciences 4, no. 8 (2000): 299-309. Scruton, Roger. Understanding music: Philosophy and interpretation. Bloomsbury Publishing, 2016. Scruton, Roger. "Sounds as secondary objects and pure events." in Sounds and perception: New philosophical essays. edited by Nudds, Matthew, and Casey O'Callaghan, Oxford University Press, 2009. Shoemaker, Sydney. "On the ways things appear." in Perceptual Experience. edited by Gendler, Tamar, and John P. Hawthorne, Oxford University Press on Demand, 2006: 461-480. Siegel, Susanna. "Which properties are represented in perception." in Perceptual Experience. edited by Gendler, Tamar, and John P. Hawthorne, Oxford University Press on Demand, 2006: 481-503. Soteriou, Matthew. The Mind's Construction: The Ontology of Mind and Mental Action. Oxford University Press, 2013. Soteriou, Matthew. "Sound and Illusion" in Perceptual Ephemera edited by Crowther, Thomas, and Clare Mac Cumhaill, Oxford University Press, Forthcoming 2018. O'Shaughnessy, Brian. Consciousness and the World. Oxford University Press, 2003. Pasnau, Robert. "What is sound?." The Philosophical Quarterly 49, no. 196, 1999: 309-324. Sorensen, Roy. Seeing Dark Things: The Philosophy of Shadows. Oxford University Press, 2008. Walton, Kendall L. "Transparent pictures: On the nature of photographic realism." Critical inquiry 11, no. 2 (1984): 246-277. Walton, Kendall L. Mimesis as make-believe: On the foundations of the representational arts. Harvard University Press, 1990.