Maarten Steenhagen (University of Antwerp)
[Jump to Maarten’s response to invited comments]
Abstract: Sounds enable us to hear things. But in what way? Source representationalism, as I will call it, is a view about how sounds can make other things audible. It says that sounds can fulfil a representational function in perception. In this paper I defend this position by reflecting on the possibility of sound reproduction. That possibility lies at the heart of all sound recording and playback technology. Because of this possibility, the sounds that reach our ears can act as perceptual representations of the sources that emit them, making those sources audible despite their absence from the perceptual situation.
1. Introduction
Sounds enable us to hear things. But in what way? Here I will argue that sounds can make things audible by acting as representations in perception. My starting point is the observation that the sounds that reach our ears play a specific role in our perception of the world. Sounds, at least when they have the right audible characteristics, put you in a position to hear other things as well. I argue that this implies that sounds introduce into our experience an opportunity for perceptual representation. My argument exploits the possibility of sound reproduction—a possibility at the heart of sound recording and playback technology. The upshot is a novel and appealing version of representationalism about perception.
The position I defend accepts the following four theses:
- we can hear sounds
- we can hear sources of sounds
- sounds function as perceptual mediators
- sounds can function as perceptual representations
The fourth of these theses is the central claim I defend in this paper. Note, that is not a universal claim. I will not defend the claim that the way we hear things by means of hearing sounds is always representational—i.e. that sounds always function as perceptual representations. I only establish that they can so function, and sometimes do. Nonetheless, the universal claim that sounds always function as perceptual representations is interesting to consider. That is why, in the final section of this paper, I show that such a stronger conclusion does follow when one makes two widely held additional assumptions about the nature of perception.
But before all that, I begin by outlining how I think we should conceive of the way sound figures in human experience.
2. The Role Of Sound
Quite generally, we may observe that sounds that reach our ears can make other things audible. Sounds are objects of perception fulfilling the function of making other objects audible. It is by virtue of our perception of these objects of perception that we are able to perceive other objects. Hence, your hearing sounds of the right kind explains your ability to hear other things. In this way sounds act as perceptual mediators, as we may call them.[1]
Imagine that, at night, you are woken up by a series of sounds coming from the living room. Reluctant to get out of bed, you listen more carefully to what goes on. Was a window left open so that the wind could come in? Is someone down there? Suddenly you hear a set of nailed feet sliding over the wooden floor. You realise it is the cat. The animal must have managed to open the kitchen door again. What is it doing? Again, you listen more carefully to what your cat is doing, and are able to hear it playing with a small object.
In this example, just on the basis of the various recognisable sounds that reach you upstairs, such as a rustle of newspapers, or the clattering of some pens falling on the floor, you are able to hear nails, footsteps, a cat, and eventually are able quite reliably to track an animal through the downstairs room. You are able to listen to your cat and to what it is doing. I take this possibility to be obvious. The example brings out clearly how the sounds that reach your ears have a part to play in the explanation of your ability to hear other entities in your distal environment. That these sounds are audible to you in your bedroom explains your nocturnal ability to perceive a cat in an entirely different part of the house. And not only is the presence of sounds of the right kind sufficient for making the cat audible to you, it is necessary as well. Your perceptual attention, as it goes out to an animal downstairs, remains at all times utterly dependent on the sounds that reach you upstairs. It is only because of their audible characteristics that you are able to hear your cat. This doesn’t mean you have to listen to those sounds, of course. No doubt, they themselves may grab your attention for a brief moment, but in the example you do not pay much heed to these sounds; you listen to the cat, and to what it is doing. Nonetheless, whatever happens, you have to hear sounds of the right kind in order hear the animal downstairs. It is in this sense that sounds act as perceptual mediators.
This picture of the mediating role of sounds in our perception of the world presupposes both that we can hear sounds and that we can hear their sources. I take both assumptions to be uncontroversial. About the first I agree with Robert Pasnau. He suggests that “surely it is part of the meaning of the term ‘sound’ that sounds are somehow the subject of our auditory experience” (1999, p. 317). That we can hear sound is brought out too by the occasions in which perceiving sound becomes an end in itself. Think of listening to music. The music played on a lute or piano is an harmonious organisation of sound. We are only able to appreciate their melodies because we are able to hear the sounds that make them up, just as we are only able to appreciate a sculpture, say, because we are able to perceive the bronze it is made of.
The second assumption I am making is that we can hear more than sounds alone. Also this is generally accepted. Some material objects or bodies produce sounds when someone or something plays, hits, or otherwise moves them. It is in them and in their activity that sound originates. When this happens these bodies become a source of sound. Authors writing on perception today typically agree that at least some of the sources of sounds are themselves audible. Take an example. On the 24th of January 24, 1975, Keith Jarrett gave a solo piano concert in Cologne. Jarrett played that night on a badly tuned Bösendorfer piano. The instrument he played was the origin of the sounds that made up his improvised composition: it was one of the sound sources in the concert venue, and, arguably, Jarrett himself was another. I am assuming that the audience of Jarrett’s concert was not only able to hear the sound of his piano, but could hear a piano as well, and could hear Jarrett as he played its keyboard.
Just as we can see a magpie, or taste vinegar in our salad, we can hear things such as pianos, people in the corridor, aeroplanes overhead, and a pub fight down the street. And as the above examples bring out, we can hear all those things because we can hear the sounds they produce. Sounds, at least when they have the right audible characteristics, put you in a position to hear other things as well. It is this mediated structure that sets auditory perception apart from vision.[2] In the human sense of hearing one kind of object of perception plays a necessary role in enabling perception of an entirely different class of objects. This makes it philosophically rewarding to single out auditory perception, and reflect more carefully on the conditions that give rise to particular abilities to hear things in specific circumstances.
3. Recorded sound
My central claim is that, because sounds act as perceptual mediators, they introduce into our experience an opportunity for perceptual representation. By this I mean that sounds, as perceptual mediators, make possible a form of perception that does not require all of its objects to be present to a perceiver. Sounds can represent sources of sound in perception. My defence of this claim begins with a reflection on the technology to record and reproduce sound—a real and familiar phenomenon.
Many people will have in their possession the album recording of Jarrett’s performance, released as The Köln Concert. The record or AAC-file enables one to reproduce the sounds that an ECM Records engineer captured back in 1975 using a pair of condenser microphones and a portable tape machine. What is interesting about such audio recordings is that they allow sounds that sounded on an earlier occasion to sound once more. To see how this technology opens up the possibility for an auditory form of perceptual representation, we only need to reflect on what we can hear when we listen to the playback of a sound recording, and on the conditions that are required for hearing it.
It is quite clear, I presume, that playing back the recording of Jarrett’s 1975 performance enables you to hear the sound of a piano. This particular sound becomes audible as soon as it fills the room when playback starts. It may subtly decompose into individual sounds of piano keys pressed by Jarrett’s hands. But for convenience we may think of these individual sounds too as instances of the sound of a piano. So, despite the fact that there is no piano around, the sound of such an instrument is genuinely there in your room.
The technology for making this possible has only become available after Édouard-Léon Scott de Martinville, and later Thomas Edison, did pioneering work on sound recording and reproduction. The late nineteenth century invention of phonograph cylinders first enabled sounds produced at a specific moment in history to be captured, stored, and played back at a point in later time (cf. Sterne 2003). Scott de Martinville’s amateur rendition of the French song ‘Au Clair de la Lune’, recorded on 9 April 1860, is said to be the earliest sound sequence in history that was recorded in a way that allows it to be replayed.[3] This discovery, that sounds could be captured and reproduced, was revolutionary.
Somewhat frivolously, we may in imagination contrast listening to an ordinary sound recording with a situation in which a device is used that causes us to have auditory hallucinations. Perhaps by stimulating your auditory cortex, such a strange device could cause you to have an hallucinatory episode that would in some subjectively salient respects resemble listening to Jarrett’s solo improvisations. But compared to the experience of Jarrett’s original audience, your experience would have little more in common than just that resemblance. Strapped to such a device, not only would your experience require no actual piano, it would not require any actual sound either. It would at best merely seem to you that you are hearing the sound of a piano, just as people suffering from tinnitus merely seem to hear a buzzing or ringing.
Why is this contrast relevant? It brings out that people who enjoy a recording of Jarrett’s concert are in conditions that partially overlap with those of the original audience in Cologne in 1975. There is a commonality here. When you play back the recording of Jarrett’s improvisations, something that the audience heard back in 1975 also becomes part of your current perceptual situation: the sound of a piano. We could say that the sounds of a piano that were produced in 1975 sound again each time we play back the recording. The sound reproduced through your stereo speakers is a ‘common factor’ present both in listening to Jarrett’s improvisations in real life, and in listening to them via a recording. If someone were merely hallucinating hearing Jarrett’s music, it would not at all be obvious that any such common factor would be present.
If this is right, then sound recording technology offers you the opportunity to hear at least some of the things Jarrett’s original audience could hear. In particular, it allows you to hear precisely those things that enabled them to hear a piano. And that in turn makes it that audio recording technology offers us a way to make material objects and events audible at an earlier point in time audible again.
What Recordings Afford
It is natural to conceive of the situation of listening to a recording of a piano concert as an opportunity to hear a piano. Playing back the Köln Concert album seems to put you in a position to listen to a piano and to hear Jarrett playing its keys, just as the audience was able to back in 1975. We do not think of this situation as an illusion, but conceive of it as just a technologically mediated opportunity for having auditory perception of a concert as it took place in the past. The previous discussion of sound recording technology legitimises this natural understanding of what sound recordings afford. Audio recording technology makes it possible to reproduce the sounds that made a sound source audible in an earlier situation. And so we are right to think that they put you in a position to hear that source.
This implies that we can generalise the earlier observation, that we can successfully attend and hence listen what a sound is of, over both cases of actual listening and cases in which we listen to recorded sounds. In both cases we simply exploit the sounds that reach our ears to listen to something else. In both cases sounds fulfil their ordinary perceptual function.
Consider a real-life, non-musical example. To find out what had happened in the Boeing 737 that crashed near Pittsburgh on 8 September 1994, a team of audio forensic investigators played back the cockpit voice recorder that was retrieved after the accident. The team was particularly interested in a malfunctioning of the plane’s rudders . This is why they turned to playing back the flight recorder. It would be mischievous to suggest that they merely set out to listen to the sounds reproduced by their playback equipment. They clearly were paying attention to the plane itself—and with some success. They found out that the rudder’s eventual jamming was not due to inept operation of controls, but occurred because of mechanical obstruction. An observation like this requires a careful, well-informed ‘mining’ of a complex auditory scene, a scene here rendered audible by replaying the recording. The investigators’ finding was made possible because they were able to single out and track through time and space specific items audible by means of the recorded sound: the plane’s rudder, as it locked in its extreme position, and the members of the flight crew as they manipulated the controls.
Although they exercised their capacity to hear parts of a passenger jet with expert precision, the forensic team did not engage in anything beyond the pale of ordinary auditory perception and discrimination.[4] In listening to a recording it may become possible to focus one’s attention on something that goes beyond the heard sound itself. This is because the playback produces conditions that are sufficient for the potential exercise of such capacities. This is why aircraft are standardly equipped with flight-recorders. The sounds captured by such a device matter to us, but purely instrumentally: as a means of making audible events as they unfolded minutes before potential incidents.
This last example presents a project of enquiry that is on a par with the earlier night-time investigation into what your cat was doing. Whether one can engage in this mode of attentive listening to sound sources just does not depend on whether one listens to a recording or not. If this is correct, then that we are merely dealing with a recording is simply irrelevant to a source’s audibility. Instead, the audibility of anything whatsoever just depends on the audibility of the right sounds.
5. Hearing particulars
There is one way to misunderstand what I have been saying to far that would lead to an underestimation of the conclusions just reached. I am talking about ‘object hearing’, and not hearing of facts. Both in ordinary listening and in listening to a recording we can listen to particular objects or events. When you hear a cat or an airplane rudder, you do not just engage in some kind of epistemic hearing or ‘hearing that…’. Such epistemic hearing—coming to learn about facts about objects in the world through hearing—surely occurs. Yet we can hear more than just that a sound is of an airplane rudder, or what a cat is doing in your living room. We can hear a cat, a particular animal, a piano, a physical musical instrument, and a rudder, a mechanical airplane part.
Object perception is form of perception that is arguably more fundamental than epistemic perception. It is more fundamental in part because, I presume, the knowledge you come to have seems to be explained by your hearing particular objects and events. Your hearing the cat’s running through the room leads you to think that it is playing, and you believe that your vase fell because you heard it when it landed and shattered into fragments on the wooden floor. In other words, hearing that the cat broke the vase, or that the plane’s rudder suffered from mechanical obstruction, is rooted in an experience of hearing cats or rudders, audible particulars that are not themselves sounds.
6. Sounds As Representations
Let me sum up what I have shown so far. By playing back a sound recording at some moment in time a sound produced earlier becomes audible once more. This is possible because the playback reproduces a sound and makes it part of your current perceptual situation. Such reproduced sounds may enable you to hear an original sound source, at least if they are faithful reproductions of the sounds that made that source audible on an earlier occasion. This makes it that objects or events audible in the past can become audible again just by replaying a recording. This is why we can listen to a sound source by means of the played back sounds, and can legitimately regard that experience, as we do, as exemplifying the exercise of an ordinary capacity to hear and listen to a wide variety of things.
Now why would this kind of perception be representational, as I maintain? The answer turns on the way recording technology revives conditions of audibility.
It is typically assumed that a representational dimension of perception would manifest itself if things appear to a perceiver that are not present to their senses, or if we have reason to think that a perceptual state a perceiver is in does not depend on such presence (Rowlands forthcoming, p. 6; Siegel 2010, ch. 6). That is precisely the perceptual situation we find ourselves in when we listen to a sound source via a recording. You can now listen to Jarrett improvise without having him or his piano present precisely because his performance was recorded. For the same reason the malfunctioning aeroplane parts and crew members were audible even months after the craft had crashed. To be sure, auditory perceptions of absent objects have not always been possible, but nowadays they are. Because we have mastered the technology to record and reproduce sounds in our environment faithfully, we are now all able to hear things that are not present in our current perceptual situation, and in fact do so on a regular basis. In other words, reflection on the real and familiar phenomena of sound recording and sound reproduction brings out that you can successfully listen to an object even if that object itself is absent from your current perceptual situation.
Wherever a kind of perception does not require some of its objects to be present to the senses, we may take it to involve a form of representation of those objects. It is standard to take this kind of object-independence to be a defining mark of representation in perception. The perceptual situation we may find ourselves in when we listen to a sound source via a recording bears that mark. When we listen to a sound source via the played back sounds of a sound recording, our perceptual state does not depend on the presence of the source we hear. That source, as object of perception, is to that extent represented in the experience.
If this is right, then hearing a sound source via a recording is, as such, representational. The sounds reproduced by the playback device we use represent the source in perception. This is because our perception of the piano when listening to the recording does not depend on the piano’s being present. Conceiving of our perception of playbacks in this way, we are able to approach the concept of perceptual representation in a novel and surprisingly minimal way.
Note however that the representational dimension to auditory experience I have here identified only becomes obvious when we reflect on sound recordings. This is because, without the involvement of recording technology, the audibility of sound sources surely does seem to depend on their presence. Keith Jarrett and his piano needed to there in order for the audience in 1975 to hear them. Their audibility on that occasion depended on their presence on stage. That we conceive of such situations as live performances, contrasting with playback, reflects this. Therefore, although the sounds audible while at a live performance still fulfil a mediating role in one’s perception, this does not yet make it attractive to think of such perception as representational. It is only unproblematic to conclude that our hearing is representational in cases where the source we hear is quite clearly not present to us. And this raises the question if there is a more precise criterion for deciding whether a sound source is present to a perceiver or not. There is.
7. Recurring States Of Audibility
Was hearing the rudder not sufficient to make the rudder present to the audio forensic investigators, even after the plane had crashed? It is at least natural to think that the objects or events you perceive are, by virtue of being perceived, at least in a sense ‘present’ to you. That point is well taken. I am not denying that the rudder was present to the investigators in this sense.
But just as there is a sense in which the rudder was present, there also is undeniably a sense in which it was not present to the investigators. It couldn’t have been, as it and the sound it produced were obliterated in a deafening crash. The tension here is reflected in the somewhat see-saw way philosophers conceive of perceptual representation, as an experience of something that is or need not be there to be experienced.[5] We find it in the very catchphrase authors of various stripes use: presence in absence. Hearing a piano by means of a sound recording exemplifies this ‘present-in-absence’ structure. The point is that, though we indeed hear the piano when we listen to it via a recording, the instrument itself is not there to be perceived—it is not constitutively involved in our perceptual situation, nor does it need to be. It is in that sense not present to us.
Some may think that the piano’s absence from this perceptual situation has to do with the fact that we hear something that existed only in the past. That is not quite right. That something existed only in the past does not obviously entail that it is absent from a current perceptual situation. Think of the visual perception of extremely distant astronomical objects. The star Nova Delphinus exploded and vanished more than 10.000 years ago, yet its explosion was clearly visible from Earth during a 27-day period in 2013. The explosion could be seen in the night’s sky by the naked eye, and the event drew great public interest. May we now infer that the perception of earthlings who saw the explosion was representational because they saw something not present to them? This is at least not obvious. The star and its explosion, though in the past, needed to have been there to sustain their visibility in 2013. The state of visibility people were able to exploit in 2013 had existed for thousands of years, and constitutively involved Nova Delphinus and its explosion. It is reasonable to think that the star was present to them, both in that they saw the star and because the star was constitutively involved in their perceptual situation.
If there were no significant differences between hearing Jarrett’s piano when one listens to a playback of the recording and perceiving a distant star, then we should conclude that listening to a piano via a recording is not representational, or at least not obviously so. The analogy between these two cases may seem revealing at first blush. But it is mistake to take the parallel too seriously. There clearly is a relevant difference. This difference between our perception of Jarrett’s piano when we listen to a playback of the album and our perception of the distant super nova lies in the fact that the audibility of the piano depends on a re-introduction of something that was part of an earlier situation into our current one, whereas our perception of the star does not depend on any such redoing. In 2013 the star simply was still visible, whereas Jarrett’s piano is not still audible today. Playing back a recording makes the piano audible again. And it does so without requiring the piano itself.
This is what I mean when I say that, when hearing a piano via a recording, our perceptual state does not depend on the presence of the piano. When we listen to a recording our perception depends on a revived state of audibility, a state that in and of itself does not constitutively involve the sources audible because of it. Sound reproduction allows an original state of audibility to obtain once more—it does not extend an earlier state into the future, but brings the earlier state about again. And although this works by reproducing the original sound, it does not require us to reproduce the original source. Playing back the sound of a piano doesn’t require a piano. That is obvious. It is in this sense that we are able to hear the piano in its absence: in order to make that piano audible once more, only sounds need to be generated while the piano itself can remain absent.[6]
When we listen to a sound source via a recording our perceptual state does not depend on the presence of the source we hear. That source is to that extent represented. We now see why the parallel with the star case was misplaced. Although the two cases have in common that what we can hear existed in the past, they differ in that in the star case the past object must be still visible, whereas in the recording case the past object need not be still audible. What matters is that it is audible once more.[7] There’s a point in calling it re-presentation.
8. Source Representationalism
Assuming that there is nothing peculiar about the several examples I have been considering, we may accept that whenever we hear a source merely by playing back a sound recording, that source is represented in experience. I have explained why: it is because in such a situation a source’s audibility does not depend on the presence of the source. As a revived state of audibility, it is no longer constituted by the source. That is the invention we owe to Scott de Martinville and Edison. Played back sounds can act as perceptual representations.
Given the ubiquity of sound recordings, sounds often do fulfil such a representational function in perception. We can hear our friends on our voicemail, a rooster when the alarm clock wakes us, and panting athletes in an Olympic broadcast on TV. This establishes my main claim that, because sounds act as perceptual mediators, they introduce into our experience an opportunity for perceptual representation. Sounds, as perceptual mediators, make possible a form of perception that does not require all of its objects to be present to a perceiver. They can represent sources of sound in perception. Call this view ‘source representationalism’.
As I have presented the view, source representationalism can be accepted by philosophers of many stripes, as it does not make any specific assumptions about the nature of perceptual experience, or the way such experience comes about. Someone sympathetic to the view is able to approach the notion of perceptual representation in a surprisingly minimal way. All they need to do is take serious what is entailed by our mundane capacity successfully to listen to and hear sound sources via the sounds they produce, and bring out that we can just as well exercise this capacity when we listen to played back sound recordings.
Nowadays the claim that perception can be representational is mainstream. Yet, the specific way in which I have advanced this claim is interesting for at least two reasons. First, many philosophers defend or assume representationalism as a universal thesis. They typically conceive of it as the view that perception as such is representational. Implicit in this is that representationalism about perception is an all or nothing affair: either all of perception is representational, or none of it is. My argument brings out why this is misleading. I have defended a weaker claim about perceptual representation, and have done so coherently. Sounds can represent material objects or bodies in perception. This implies no more than that, potentially, there are circumstances in which perception is representational. The logical availability of source representationalism proves how one can coherently maintain that merely some perception is representational. In this way my argument disrupts a tidy way of outlining the shape of the debate about the nature of perception. Already this makes it worth considering the position.
There is another reason why source representationalism should be of interest to those thinking about perception. It embodies a story about what perceptual representation is, or at least about what form it can take. In this regard I have not just broken lines with the mainstream, but moved well beyond it. Several authors have objected to standard representationalist theories of perception that they remain largely silent on what perceptual representations are. It is typically assumed that perception is representational, yet never elucidated what perceptual representation consists in, or what does the representing. I have here advanced a representationalism that provides a template for answering what perceptual representations are. Sounds stand out as a paradigm of perceptual representation. They are entirely mundane, public phenomena that can confer audibility on their sources in a way that does not depend on the presence of those sources. In this way, my defence of source representationalism offers an informative and productive model of thinking about perceptual representation more generally.
In the remainder of this paper I will do three things. First, I will argue that my argument does not rest on the assumption that sounds are abstract individuals. Second, I will make more explicit how in my view we must conceive of auditory representations. Finally, I will identify what further assumptions could move someone to accept source representationalism as a general principle governing how sound sources figure in perceptual experience.
9. Sounds As Repeatables
I claim that sounds can act as perceptual representations. My case turns on the idea that playing back an audio recording reproduces part of the perceptual conditions of an earlier situation. That part, it should now be clear, consists in the sounds that were audible when the recording was made. In this way, my argument draws out an implication of the real and familiar phenomenon of sound reproduction. It could now be thought that my argument presupposes quite a weighty view of what sound reproduction comes down to. Do I not assume that to reproduce a sound is make one and the same sound recur, thereby presupposing that sounds are repeatable or abstract individuals? The idea that sounds are repeatable is controversial. But I will explain why my argument makes no such commitments.
Some authors think that our prephilosophical conception of sounds pictures them already as repeatables. Something is a repeatable only if by its very nature it can recur in history, i.e. if it has the potential to exist more than once. Repeatables may be classified as abstract entities. To clarify, this would not make sounds ahistorical. Even as abstract entities sounds could find their origin at one moment in time, and could go out of existence for good at another.[8] Photographic images are like that too: they come into being at a specific moment in history, yet can be re-instantiated many times in news papers and on web logs. If sounds are abstract like that, then we may suppose that your speakers are able to repeat the particular sounds that were produced by a Bösendorfer piano in January 1975. The very sounds audible back then would sound once more, now reproduced by your speakers. Call this the repetition theory of sound reproduction.
To my mind it seems that in the current age people conceive of sound reproduction in this way.[9] Yet some authors wish to resist the assumption that sounds are repeatable and abstract. And so, to conceive of sound reproduction in terms of a repetition of a particular sound would require an assumption about sound not all parties wish to make. But my argument does not rest on this assumption. Those who do not wish to claim that sounds are repeatables may conceive of sounds as being more like people. Just as it is impossible for the individual Napoleon to live twice, it is impossible for the individual sounds produced by Jarrett’s piano to occur again. On such a picture, the sounds your speakers broadcast are newcomers to the world and originate just in the coppery coils of your tweeters and woofers. Those sounds bear no ontological relation to sounds that died for good against the walls of the Cologne opera building in 1975. Call this the re-creation theory of sound reproduction. On this conception, and unless it has lasted for decades, a sound heard in 1975 and a sound heard in 2016 must be ontologically distinct.[10]
The claim that we can exploit sound recordings to listen to recorded sources does not depend on the metaphysical dispute between repetition theory and re-creation theory. What matters is that a well-calibrated stereo set can broadcast sounds sufficiently like the sounds that Jarrett’s off-key piano produced in 1975. For all that is needed is the revival of an initial state of audibility. And for that, it suffices to produce a sound that shares the relevant audible characteristics with the recorded sound, and moreover tracks the latter’s characteristics. The ties required between original and reproduced sounds may remain merely be qualitative and causal, not ontological. In other words, all that may be needed for a sound to be reproduced faithfully is that the played back sound sounds sufficiently like the original sound, and moreover sounds the way it does because the original sound sounded the way it did. Both parties can accept this characterisation of sound reproduction.
That playing back the recording of Jarrett’s 1975 performance enables you to hear the sound of a piano may be understood in this neutral light. The playback results in a sound that is sufficiently like the sound Jarrett’s piano produced when the recording was made, and moreover it sounds the way it does because the recorded sound sounded the way it did. That would be enough to justify calling the reproduced sound ‘the sound of a piano’, just as we quite naturally talk about a faithful reproduction of the Nightwatch as if it was itself that work of group portraiture. But furthermore this is no more than a linguistic point. What matters is that in listening to a recording of Jarrett’s piano concert you find yourself in a very similar auditory situation as the original audience back in 1975, similar enough for Jarrett’s piano and performance to become audible to you.
If this is right, then regardless of whether we think sounds are abstract individuals or not, the possibility of sound reproduction can be accepted. Accordingly, on either the repetition theory or the re-creation theory of sound reproduction, one may accept that playing back a sound recording can revive an original condition of audibility, making a source audible in the past audible once more.
10. What Is Representation?
Once we accept that sources can be represented in perception in the way I have argued they can, we may wonder, what makes perceptual representation of this kind possible?
Few people able to hear have difficulty listening to sound sources via sound recordings. And so it seems that we possess either a natural or acquired capacity for representational perception that is identical to or continuous with our natural capacity to hear.[11] Just as other perceptual capacities, the capacity for representational hearing is exercised in response to something.[12] And just as our capacity to hear in general, our capacity to hear things representationally is responsive first and foremost to the audible sounds in our environment. Hence, the sound we can hear plays a determining role in the exercise of this capacity. It determines what is audible in our environment and what is not. If that is right, then first and foremost the sounds you hear make perceptual representation in the auditory modality possible.
Here is how this works. We can safely assume that, quite generally, what we can hear is at least in part determined by the audible character of the sounds we perceive. Things can sound the way they do to us in part because the sounds that reach our ears have the qualities they have. If the sound reproduced by the speakers in your room did not have audible qualities of the right kind, then hearing that sound would not enable you to hear a piano at all. On occasion sounds might come out all garbled—pitched-shifted, say, or reversed or contracted. When that happens, perhaps you are able to hear only some noises. Yet when your equipment is functioning well, the sounds you hear and the qualities they have can explain why, in the absence of a piano, you are able to hear such an instrument.
And so audible sounds figure as an essential cog in the machinery of representation in auditory perception. They can do so because the world of hearing includes both perceptual mediators and the objects and events they mediate. This fact about our auditory world makes it possible to record and reproduce a mediator quite independently of the material objects it enables one to perceive. Once we are capable of exploiting this fact, sounds can confer audibility on things that are not present to us. It is in this sense that the sounds themselves can come to function as auditory representations. A sound is an objects of perception that can fulfil a representative function in perception, making objects and events audible in a way that does no longer involve them.
So conceived, sounds, as auditory representations, do not depend on the object we perceive. Neither do they depend on the perceiving subject. They are ‘representational vehicles’ that exist quite independently of any particular act of hearing. Charles Travis has called such representations ‘allorepresentations’, and he contrasts them with acts of representing, which clearly are not independent of a subject’s particular states.[13] We may contrast this conception of auditory representation with two alternative pictures. The first focuses on the sound recording. It identifies the record or audio file itself, as a store of auditory information, as what represents the sound source, and makes it possible for us to hear it in its absence. The second picture psychologises the conditions of representation, and conceives of auditory representation as something essentially mental. What represents the sound source, it maintains, is a specific mental state, such as an experience or sensation. I need not deny that either of these phenomena are representational. What matters however is that the position I have been sketching reveals both of these pictures to be at best incomplete. They overlook the potential representational function of sounds themselves.
11. Globalisation
Finally, someone could wonder if my argument has not captured in general what it is to hear things. All I have claimed is that hearing a piano by listening to a recording of its sound is representational. But an interlocutor might be tempted to try to distill from that local possibility a global principle. ‘Surely…’, they might say,
‘…if sounds are perceptual mediators, then also in everyday experience they do the heavy lifting. It is those sounds that make the world audible to you. So when a coach drives down your street, you can hear it only because its sound manages to invade your study. Right? Then at least to the auditory experience the coach itself seems to make no contribution. Does this not suggest that, strictly speaking, things like coaches can never be present in our perceptual experience? Does it not imply that, even when we hear something without the intervention of recording technologies our perception must be representational as well?’
If this imagined interlocutor were right, source representationalism would offer a general answer to the question how we hear things. But that general answer does not follow from what I have argued. I do not accept the interlocutor’s line of reasoning, at least not as it stands. Their conclusion would require further premises. Let me begin by explaining more clearly why their conclusion does not follow from my argument.
Hearing a piano my means of a recording and hearing a piano in real life have much in common. They at least share a common factor: a sound that makes a musical instrument audible. I argued that when that sound is played back, it can make a piano audible in a representational way. The fact that this representational situation is made possible by something both situations have in common may make it tempting to infer that also non-technologically mediated cases of hearing sound sources are representational.
That inference would be fallacious. Despite everything I have said so far, hearing a piano in real life may still be a fundamentally different kind of experience from hearing a piano via a recording. It may still be thought that, in ordinary cases of auditory perception, the fact that the audibility of the sound sources we hear constitutively depends on them does make a difference to the way we come to hear those sources. In other words, my argument does not rule out the possibility that our perception remains not unaffected by a source’s presence, regardless of whether this is manifest in experience or not. This is why the general conclusion envisaged by the interlocutor does not follow from what I have argued, and so would require further premises.
The stronger conclusion requires an assumption that implies that the presence of a sound source could never make a difference to the perception we have of it. Such an assumption would justify someone to accept that all hearing of sources is representational. In what follows I will identify two assumptions that would be of the right kind.
The first of these is about the phenomenology of perception. Several authors writing about perception assume that what it is like to undergo an experience fully determines the phenomenal character of the experience. If that were right, then it would follow that also ordinary listening is representational. This is because hearing in real life can be indistinguishable from hearing via recordings. Think of those talented musicians in the bowels of London’s Underground. At some distance, you may doubt whether they are actually strumming their guitar or are using a covertly played recording instead. In such a state of doubt, reflecting on what you hear may not be sufficient to distinguish between the two possible scenarios. For any original experience of a sound produced live, we can specify a conceivable, indiscriminable recording counterpart in which you would hear that source through a recording instead. By hypothesis, the audible qualities heard in each of these scenarios are indistinguishable. If what it is like to undergo an experience fully determines the phenomenal character of the experience, then for every original experience there is a recording counterpart that does not differ in phenomenal character. And we may take the lack of difference in phenomenology to entail a lack of difference in representational properties. This would mean that also the original experience is representational given that its indiscriminable recording counterpart would be, despite the fact that in the one case the audibility of the source depends on it constitutively.
Perhaps some people find this line of reasoning at least suggestive. Yet the argument is conditional on a questionable assumption about perceptual experience. It presupposes that subjectively indistinguishable experiences must be of the same kind. Many will not be impressed. Subjective indistinguishability is generally no adequate criterion of sameness, and there seems no good reason to take the domain of perceptual psychology to be an exception. Why couldn’t two entirely different sorts of experiences nonetheless be indistinguishable for someone who is having them? If they could, then we simply have not established that all hearing of sources is representational.
There is another assumption that could motivate someone to take my conclusions to imply a general thesis about our auditory perception of sources of sounds. It is often suggested that among the objects of perception some are special because they, given the qualities they possess, determine our experience. The idea is that we are only perceptually sensitive to a limited range of qualities, and that there is some well-defined set of objects that have the exclusive privilege to be their bearers. These are the ‘proper’ objects of perception. To someone attracted by this picture, it may seem natural to limit what can be present to us in experience to these proper objects. For consider, it would only be such objects that are actually able to make a difference to what we experience.
The earlier interlocutor could now assume that sounds and only sounds are the proper objects of audition. From this it would follow that sound sources are always heard in their absence, given that, on this model of the objects of perception, only sounds possess the qualities our ears are sensitive to. In order to revive its earlier state of audibility, one doesn’t need to drag along that piano, not just because we can reproduce a sound independent of its source, but simply because that instrument was never doing any real perceptual work.
I think that there is more to say in favour of the second assumption than there is to defend the first. Historically, the notion of a proper object has been influential in modelling sense perception. However, also this second assumption is a substantive claim about audition. Aren’t trained ears sensitive to more than the qualities a sound can possess? Can’t we learn how to perceive qualities that are possessed by material objects such as piano’s and cats? All I need to say here is that both of the assumptions can be resisted without damaging the central claim I have defended in this paper.
12. Conclusion
My argument for source representationalism can be accepted by philosophers of many stripes. I have clarified that it only shows that some auditory perception is representational in virtue of the way sounds can figure in experience. The availability of this local form of representationalism shows that an assumption frequently made by philosophers of perception is too simple. The assumption is that perception as such either is a presentation of its objects, or it is an experience in which those objects are represented. Charles Travis exploits this assumption to motivate a wholesale rejection of representationalism about perception. He agrees that perceptual representation does not depend on the presence of their objects, yet takes it to be undeniable that, as he writes,
…perception, as such, simply places our surroundings in view, affords us awareness of them…. It confronts us with what is there… (Travis 2013, p. 65)
But surely hearing is a form of perception. And surely hearing does indeed ‘confront us’ with what is there: the sounds in our environment. Yet if what I have argued is right, then it simply doesn’t follow that hearing requires the presence of all of its objects. The explanation is not primarily a psychological one. It has to do with our auditory world. Because sounds act as perceptual mediators and because we have invented the technology to record and reproduce them, sounds introduce into our experience an opportunity for perceptual representation. It is because of these two facts that even long after it ceased to be audible, a complex auditory scene can be made audible to us once more. This enables you to perceive objects and events that need not be present to you at all. It is for this reason that perception, as such, can do more than Travis thinks.
References
Caston, Victor (1998). “Aristotle and the problem of intentionality”. Philosophy and Phenomenological Research 58 (2):249-298.
Dodd, Julian (2000). “Musical works as eternal types”. British Journal of Aesthetics 40 (4):424-440.
Kalderon, Mark Eli (2015). Form Without Matter: Empedocles and Aristotle on Color Perception. Oxford: Oxford University Press.
Martin, M.G.F. (2012). “Sounds and Images.” British Journal of Aesthetics 52 (4): 331–351.
Nudds, Matthew (2001). “Experiencing the production of sounds”. European Journal of Philosophy 9 (2):210-229.
Nudds, Matthew (2010). “What Sounds Are”. In Dean Zimmerman (ed.), Oxford Studies in Metaphysics: Volume 5. Oxford: Oxford University Press.
Nudds, Matthew & O’Callaghan, Casey (eds.) (2010). Sounds and Perception: New Philosophical Essays. Oxford: Oxford University Press.
O’Callaghan, Casey (2007). Sounds: A Philosophical Theory. Oxford: Oxford University Press.
Pasnau, Robert (1999). “What Is Sound?” Philosophical Quarterly 49 (196): 309–324.
Rowlands, Mark (forthcoming). “Arguing About Representation.” Synthese: 1–18. doi:10.1007/s11229-014-0646-4.
Scruton, Roger (2010). “Hearing Sounds.” In Oxford Studies in Metaphysics : Volume 5, edited by Dean W. Zimmerman, 271–278. Oxford: Oxford University Press.
Siegel, Greg (2014). Forensic Media: Reconstructing Accidents in Accelerated Modernity. Durham: Duke University Press.
Siegel, Susanna (2010). The Contents of Visual Experience. Oxford: Oxford University Press.
Sterne, Jonathan (2003). The Audible Past: Cultural Origins of Sound Reproduction. Durham: Duke University Press.
Travis, Charles (2004). “The Silence of the Senses.” Mind 113 (449): 57–94.
Travis, Charles (2013). “The Silences of the Senses.” In: Perception: Essays After Frege. Oxford: Oxford University Press:
Walters, Lee (2013). “Repeatable Artworks as Created Types”. British Journal of Aesthetics 53 (4):461-477.
Notes
[1] There is considerable disagreement over the nature of sounds and the way auditory perception is able to afford perception of the material world. But at least many authors currently writing on the topic seem to endorse something like my characterisation of the enabling role of sounds in perception. See O’Callaghan 2007, Nudds 2010, and O’Callaghan and Nudds 2009 for seminal contributions to these ongoing discussions.
[2] Arguably, olfactory perception shows a similar structure. We can smell the Roquefort only because we can smell the scent it gives off. Vision as such lacks that structure. However, I argue in other work that perceptual mediators can be (and have been) introduced, contingently, also in the visual modality; see my ‘Pictures as Perceptual Mediators,’ ms..
[3] Interestingly, playing back the song was not yet possible at the time of recording. Only much later, scientists at the Lawrence Berkeley National Laboratory succeeded in converting the original paper-based recording to a digital audio file. It can be played back here: https://goo.gl/QbdK4f
[4] See Siegel 2014 for a nice discussion of the listening culture of these forensic contexts.
[5] We already find this in Aristotle’s On Memory and Recollection: “How does one remember what is not present in the first place, given that, while the affect is present, the object is absent?” (I 450a25–27). See also Caston 1998, 257ff.
[6] Note that I am talking here of reviving audibility, not reviving sounds. To some it may be attractive to assume that we are able to revive states of audibility because we are able to revive sounds. Yet, the idea that sounds, as particulars, can live again is controversial. Below (§9) I will show that I am not committed to the assumption that sounds are repeatable individuals.
[7] Perhaps a better parallel to draw would be between listening to a recording of a piano concert that took place in the past, and being faced with a photograph of a star currently no longer visible in the Earth’s sky. Just as the payed back sound, it seems, the photograph may be said to represent the star in perception. It puts us in a position to look at it once more, despite the fact that the star is now absent from our perceptual situation.
[8] The most explicit proponent I am aware of is Mike Martin (2012), but a conception of sounds as abstract individuals is at least implicit in Nudds (2001) and Scruton (2010). That we have no reason to think having the potential to exist more than once, or in multiple location, entails being ‘outside space and time’ is brought out well by Lee Walters (2013). A failure to recognise this point is likely to make quite implausible views seem inevitable (see e.g. Dodd 2000).
[9] The point is brought out well in Jonathan Sterne’s (2003) cultural history of sound recording techniques and their reception.
[10] Perhaps some sounds have indeed lasted for decades. A nice instance of such an ‘old’ sound might be the sound of the Niagara Falls. The longest note in John Cage’s Organ²/ASLSP, as it is currently executed in Halberstadt, Germany, would be a musical example. The note is expected to sound for thirty-five years.
[11] I speak of a capacity advisedly. Merely hearing a recorded sound is clearly not sufficient for hearing its source. It took the audio forensic experts some effort to discern in the sounds they could hear perfectly well the sources they were interested in. This suggests that our perception of sources by means of recorded sounds is an exercise of an ability or skill, something that points in the direction of a perceptual capacity of sorts.
[12] We may think of it as a reactive capacity as described in Kalderon 2015, p. 24. He borrows the term from Nietzsche, but traces the idea back at least as far as Aristotle.
[13] Travis 2004; 2013.
Steenhagen’s paper uses audio recordings, along with a few claims that are commonplace in the philosophy of audition, to argue that sounds can play a representational role. We hear sounds, he says, and we sometimes hear their sources, but only in virtue of hearing sounds. Sometimes, we can hear a sound’s source even when that source is absent. In those cases, exemplified by audio recordings, the sounds function representationally. In such cases you hear sounds, and you might hear the loudspeakers which are the proximal sources of those sounds. But because of how audio recording works, you also hear through their proximal sources to the original events recorded.
According to Steenhagen, “It is this mediated structure [sounds and sources] that sets auditory perception apart from vision.” (4) And it is also this mediated structure that enables sounds to serve as representations when recordings are played back. I’ll try to push back against, or at least try to complicate, this very interesting proposal by considering an analogy with vision.
According to a familiar story, the perception of some photographs provides for mediated visual contact with their subjects (Barthes 1982, Walton 1984). We see the photos, the story goes, but because the photos are pictures, whose form depends counterfactually in the right way on the objects photographed, we see those objects too. This is controversial for all sorts of reasons that are not central to the present discussion. For now, notice that playbacks of audio recordings fit a similar pattern. We might not want to call them pictures (see Martin 2012, e.g.), though I think we should for the very same reasons we call certain visible patterns pictures (Kulvicki 2006, Ch 5). At a minimum, they are the closest thing we’ve got to auditory pictures. And audio recordings depend counterfactually on sources in a robust manner reminiscent of photographs. So, you might think that audio recordings provide mediated perceptual access to their sources. We hear the playback and, thereby, hear the source of the recording.
The foregoing sounds quite supportive of the Steenhagen’s claims, and it is, but I think it also raises some worries.
First, the thesis that photos enable indirect seeing is highly controversial, for a number of well-known reasons. That controversy applies here as well, so Steenhagen’s example might offer new insight into the visual case or be threatened by it. Do the standard objections to Walton’s thesis also apply in the auditory case?
Second Steenhagen suggests that audition works quite differently than vision. It is mediated in a way vision is not, which enables a kind of source representationalism not available to sight. But if vision is really so unlike audition it’s a bit of a surprise that we have found a visual phenomenon deeply analogous to the auditory case. Furthermore, Steenhagen suggests that audition is more like olfaction than vision (4, n2). But it’s interesting to note that we have not developed a rich (ripe?) practice of olfactory representation. Similar remarks apply to touch and taste. In that same footnote, Steenhagen says that he elsewhere defends the claim that pictures can be perceptual mediators (4, n2). But this should be surprising, given the way he defends his view in the auditory case.
To be fair, Steenhagen does not claim that audio recording demonstrates that audition is different from vision. He asks us to take it for granted that they are different, in the sense that sounds differ from their sources and things seen don’t. The claim he takes to be distinctive is that “because sounds act as perceptual mediators, they introduce into our experience an opportunity for perceptual representation.” (5) And “the representational dimension to auditory experience I have here identified only becomes obvious when we reflect on sound recordings.” (11) But such mediated perceptual contact seems possible with some photographs even though vision is not mediated in the way audition is (on Steenhagen’s view). So, perhaps we have at least a couple of ways things perceived can serve as mediators.
That brings us to the third point. We need to explain why both vision and audition seem to allow for this kind of displaced perceiving. It’s possible, as I just suggested that there are two ways, or more, for this to happen. In that case, audition and vision turn out to be quite different, but alike in allowing this to happen. But one might worry that these interesting phenomena suggest vision and audition are much more alike than Steenhagen suggests. Perhaps vision is mediated in the manner of audition, or perhaps audition is just as unmediated as vision is. Both of these options go against assumptions Steenhagen makes for the purposes of this paper, but I hope the discussion so far has shown that you can engage with these phenomena without making those assumptions. In fact, I don’t think the considerations Steenhagen raises can help us decide between any of the extant theories of sounds. That’s not a big problem, given the paper’s stated goals, but I’m curious whether Steenhagen thinks these displaced perception phenomena can help us more than I think they can.
Fourth, let me quibble with how Steenhagen describes playbacks of audio recordings. Listening to Jarrett’s Köln Concert, he suggests, “allows you to hear precisely those things that enabled [the original audience] to hear a piano.” (7) That is, you hear the same sounds the audience heard, since that is what enabled them to hear the piano. You can thus also hear that piano. But he also says “We do not think of this situation as an illusion….” (7) I think the latter claim is true even if we leave out consideration of all our other senses. Playbacks of audio recordings sound quite different from the original scenes and they even sound quite different from one another.
The fifth point is a bit more substantial, but it builds on the fourth. The way in which audio playback differs from the original suggests a fairly tight analogy between photography and audio recordings. That is, the playback seems to perspectivally like, but non-perspectivally different from its source. When played back, the sounds seem to come from someplace in the room where one sits, not a concert hall, even though there is some sense in which it sounds like something in a concert hall. The volume of playback can be adjusted in many ways, but it does not seem to affect the represented volume of the recorded scene. One can make the playback bright (lots of high-frequency energy) or dull (not much of that) without making the recorded concert seem that way. In the visual case, photos can be Kodachrome or black and white, hi contrast or low. They are all flat, even though the scene depicted is deep. So photos lack many features of depicted scenes, but are, perspectivally at least, much like them. The same seems true for audio recordings and it suggests, to me at least, that one might want to give both a similar treatment. Doing so, as I mentioned above, threatens the deep distinction Steenhagen sees between audition and vision.
One final quibble. Stenhagen says “When we listen to a recording our perception depends on a revived state of audibility, a state that in and of itself does not constitutively involve the sources audible because of it.” (13) What he means, I think, is that the source becomes audible even though the source is not in any interesting sense present. That poorly tuned baby grand that Jarrett used has likely met its end, but we hear it nonetheless. Perceiving, however, is a relation, and it’s worth pointing out that recording, in Steenhagen’s sense, needs to be as well if it’s to “revive audibility”. An exactly similar recording, but of a different concert, would not put perceivers in touch with the piano Jarrett played. Similarly, an invented recording, even if it’s identical to Jarrett’s album, affords no perceptual contact. Thus, while the source need not be a constituent of the state one hears, it needs to be a proper source of it. Here too it seems we can take our lead from the literature on photography in trying to sort out the (im)proper ways of being a source.
References
Barthes, R. 1982. Camera Lucida. Trans. R. Howard. New York: Hill and Wang.
Kulvicki, J. 2006. On Images: Their Structure and Content. Oxford: Clarendon Press.
Martin, M. 2012. Sounds and images. British Journal of Aesthetics 52(4): 331-51.
Walton, K. 1984. Transparent pictures: On the nature of photographic realism. Critical Inquiry 11(2): 246-277.
1. In his rich and provocative paper, Maarten Steenhagen argues for a novel view about the role of sounds in our perceptual economy:
Much of the argument hinges on the idea of a perceptual mediator, and my comments will focus on this.
2. Here’s how Steenhagen introduces the idea:
I think he’s right about this much: hearing anything at all requires hearing sounds; and what you hear depends on the sounds you hear. In order to hear the cat, you have to hear the right sounds. However, this is a very weak notion of mediation. After all, it also applies to the role of color and/or shape in vision. Whatever happens, you have to see colors and/or shapes of the right kind in order to see the cat; in particular, you must see the color(s) and/or shape(s) of the cat (or some part of it). So, are color and shape also perceptual mediators? We don’t usually think of them this way, and it’s not a consequence that Steenhagen wants; he writes: “It is this mediated structure that sets auditory perception apart from vision” (pp. 3–4). Unfortunately, however, Steenhagen’s explicit account of perceptual mediation doesn’t secure this separation.
So, what’s going on here? I think that what’s doing the heavy lifting in Steenhagen’s argument is the idea of sounds as particulars. He says that audition has a “mediated structure” because, in hearing, “one kind of object of perception plays a necessary role in enabling perception of an entirely different class of objects” (p. 4). Now, in one sense, colors and shapes are “objects of perception,” for they are among those “things” that we can see. However, this isn’t the notion of “object” I think Steenhagen has in mind here. I think he means ‘object’ in the sense of particular or individual rather than universal or property. In other words, I think he means that audition is mediated precisely in the sense that, in order to hear one sort of particular (piano, person, collision), we must hear another sort of particular (a sound). Nothing like this is true of vision.
What are the consequences for Steenhagen’s argument? Well, if there’s more built into the idea of perceptual mediation than he suggests, then source representationalism isn’t as metaphysically neutral as he might like:
Suffice it to say that I think that accepting source representationalism requires more than this, since there are coherent accounts of audition that take this mundane capacity (and what it entails) seriously, but reject the idea that sounds are particulars, and so, perceptual mediators in any but the trivial sense described above. (See, e.g., Pasnau (1999), Leddington (2014), and Leddington (ms.).)
3. But whatever we think of sounds as perceptual mediators, if we believe, with Steenhagen, that listening to an audio recording of a past event (such as Jarrett’s performance in Köln) puts us in perceptual contact with the original event, then such audio recordings—or better: their playbacks—will count as perceptual mediators. In other words, if the playback of such an audio recording is “transparent” (Walton, 1984), then perceptual contact with one event will enable perceptual contact with another (otherwise absent) event.
Two comments here. First, this preserves the structure of “presence in absence” that Steenhagen takes to be sufficient for a certain kind of representation in perception. Second, this means that even if we reject the idea that sounds are particulars, and so, themselves candidates for being perceptual mediators, we can accept that the playback of a sound recording is a perceptual mediator, and so, introduces a kind of representation into perceptual experience.
But now, whence the temptation to say, with Steenhagen, that the the representational work here is done by the sounds rather than the playback of the recording? Of course, when you play the recording, sounds are produced, and it’s thanks in part to the production of those sounds that listening to the playback puts us in perceptual contact with the original event, but why think the sounds themselves are representational? Here are two analogies. Suppose, first, that playbacks of audio recordings of past events are like spoken accounts of past events; in this case, we might think that, just as spoken accounts represent past events partly in virtue of the representational powers of the individual utterances or statements that constitute those accounts, so playbacks of audio recordings represent past events partly thanks to the representational properties of the sounds those playbacks make. But suppose instead that playbacks of audio recordings of past events are like photographs of past events; in this case, we might think that, just as photographs represent past events partly in virtue of the patterns of color on their surfaces, so playbacks of audio recordings represent past events partly in virtue of their patterns of sound; and just as we’re not tempted to say that the colors themselves have representational properties, we wouldn’t, here, say that of the sounds, either.
So, I want to suggest, again, that what’s doing the deep work in Steenhagen’s account is not merely reflection on “our mundane capacity to listen to and hear sound sources via the sounds they produce,” nor simple analysis of the role of audio recordings in our perceptual economy, but, critically, the idea that sounds are particulars (so, like utterances or statements) rather than properties (so, like colors).
4. There’s much more to be said about this paper. It includes a number of interesting arguments worth discussing. However, I’m going to conclude with a comment that does not bear directly on the paper’s main claims. Steenhagen takes for granted the widely-accepted idea that “we can hear more than sounds alone” (p. 3). He writes:
I’m happy with the idea that we hear sound sources; still, to my ear, Steenhagen’s description of the situation seems off-key. Pianos, as such, don’t make noise. Mere objects are silent. Only in activity is their noise. In other words, sound sources are always events. This is part of our ordinary, everyday conception of sound. Of course, in hearing events we are sometimes said to hear objects that participate in them. So, in hearing the event Jarrett playing the piano, I may be said to hear Jarrett and the badly-tuned piano, but Jarrett and the piano are not (therefore) distinct sound sources. (That said, if you listen closely to the recording, you can hear Jarrett humming; so now we have two distinct sound sources: Jarrett’s playing and Jarrett’s humming.) Anyway, while this point does not bear directly on the argument of the paper, it’s neither trivial nor irrelevant. That’s because the tendency to treat mere objects as sound sources is part of what leads philosophers to think that sounds must be particulars, not properties. After all, if sounds are properties, then presumably they are properties of their sources. But while the idea that sounds are properties of objects is deeply problematic, the idea that sounds are properties of (noisy) events is not only defensible, but intuitively plausible. (For full discussion of this matter, see Leddington ms.)
Works cited
Leddington, J. (2014). What We Hear. In R. Brown (Ed.), Consciousness Inside and Out: Phenomenology, Neuroscience, and the Nature of Experience (pp. 321–334). Dordrecht: Springer.
Leddington, J. (ms.). In Defense of Sounds as Properties. Available at http://www.jasonleddington.net/work-in-progress.html.
Pasnau, R. (1999). What Is Sound? The Philosophical Quarterly, 49(196), 309–324.
Walton, K. L. (1984). Transparent Pictures: On the Nature of Photographic Realism. Critical Inquiry, 11(2), 246–277.
We hear a meow; we hear a cat. We hear sounds; we hear things that makes sounds. What is the relationship between these two sorts of auditory objects? What is the relationship between the two acts?
In his path-breaking paper, Maarten Steenhagen does (at least) two different things. He argues convincingly, and against conventional wisdom, first that we hear soundmakers, and second that we do so by hearing sounds. I largely agree with him about this. I’ll take up this important theme in section II.
I. Quibbling with Maarten
First, however, some quibbling about matters that urgently need clarification. Maarten argues that:
Because an audio-reproduction is the same as the original, the same sound can occur in the absence of the soundmaker. So, Maarten says, sounds stand to soundmakers as “perceptual representations” stand to external objects. (That is, we hear soundmakers/external objects by hearing sounds/perceptual representations, and we could hear the same sound/pr in the absence of the soundmaker/eo.)
So, he concludes:
I disagree with these claims, most vehemently with 3.
Re 1: Maarten is quite concessive about 1. He writes: “all that may be needed for a sound to be reproduced faithfully is that the played back sound sounds sufficiently like the original sound.” Perhaps, then, there is no need to argue about 1.
But just for the record, let me rehearse some well-known difficulties about identity through time. Consider a person, Justin, in 2005. Consider the Prime Minister of Canada in 2016. These are one and the same. But how can this be? The PM is PM; Justin in 2005 was not. The solution to this riddle is to distinguish between individuals that continue in time (or “continuants”), and temporal stages of those individuals (or “time slices”). Considered as a continuant, Justin in 2005 and the PM in 2016 are the same and both are PM in 2016. Considered as time slices, Justin in 2005 is not PM, but this does not create a problem of logic, because the 2005 slice is different from the 2016 slice.
By analogy, consider Jarrett’s original performance and the sound produced when I play the recording. In one sense, these cannot logically be the same: one began and ended in December 1975; the other began and ended in July 2016. But it might make sense to say that they might be temporal stages of the same continuant. This continuant “became audible” on several different occasions. Each of these occasions is a distinct spatiotemporal slice of the continuant. (I acknowledge that there are other logical models that capture the distinction that I am attempting in this paragraph; I don’t worry about this since they all yield a distinction that is homologous with that between continuant and slice.)
So here’s the thing. If, when I play a recording of Jarrett’s performance, I am playing the same sound, then that same sound is a continuation rather than a repetition. Here’s a continuant: [—1 —2]. In it, the second dash continues the first in the sense that it is another part of the whole, which occurs later in reading order. (The superscripts are inserted for labeling purposes only; the two dashes are meant to be qualitatively identical.) The second em-dash, —2 is a slice of the spatial continuant. —2 is, however, distinct from —1 and from the composite consisting of the two together, [—1 —2]. There is no repetition of the composite; it’s one thing. The second dash continues the first, but does not repeat it in the required sense of being the same thing.
It’s not clear which path Maarten wants to take with regard to the audio-playback. I think that what he ought to say is that it continues the original performance because it “sounds sufficiently like the original” and is causally continuous with it. The sound is a continuant, and each replaying is a (distinct) slice.
Re 2: Let’s put to one side the question about continuants and slices. Quite independently, I have reservations about the “sounds sufficiently alike” stipulation.
First of all, we know that all recordings distort the original. I play the recording on my iPhone headphones; I play it on my stereo speakers. These sound different, and presumably both sound different from the original. So what does Maarten have in mind when he says that the audio playback is the same? Maybe we should imagine a near “perfect” system. Its amplitude curve is more or less flat for all frequencies.
Somewhat less plausibly, at least as far as current technology goes, stipulate that the imaging is three-dimensionally stable even when I move. Not only do I hear exactly the same thing as I would have if I had been in seat 42F of the original hall, but also what I hear shifts in the same way when I squirm in my seat, move my head, or leave in a hurry to use the toilet. Let’s call this a 3D sound recording. Since the original sound was 3D in this way, and since hearing soundmakers ultimately depends on this, the example requires that the reproduction be 3D as well.
Even with the 3D recording, there is an additional point frequently made about photographs. They never merely reproduce because the photographer always imposes her view—the exposure, the colour balance, the depth of field, the shutter speed, the pan are all ways of manipulating what the ultimate product looks like. (This point is often made against Kendall Walton’s 1984 view that we literally see our dead ancestors when we look at photographs of them.)
Similarly, a recording never reproduces because the producer and the sound engineer impose their idea of how the recording should sound. It does not merely re-audibilize the original.
Re 3: Sounds are external to the mind. What was the sound of Jarrett’s out-of-tune piano? Some have held that it is a source event: the hammers of the piano striking the improperly tuned strings. Others think it is a disturbance of the air. This is plausible too. (I’ll come back to this controversy in a moment.) Either way, it is an external entity, not something in the hearer’s brain or mind. The sound of the piano is not an event in the hearer’s auditory system or in auditory consciousness.
How then can it be a perceptual representation? In his Conclusion, Maarten draws out what he takes to be the implication of his argument. He writes: “some auditory perception is representational in virtue of the way sounds can figure in experience.” He seems to say that sounds are like auditory perceptions—constituents of experience that represent external things—or possibly something like a sense-datum. This is wrong: a sound is not like an auditory perception or datum; it is an external thing. A few sentences later, Maarten acknowledges this. Concurring with an aspect of Charles Travis’s direct realism, he writes: “hearing does indeed ‘confront us’ with what is there: the sounds in our environment.” Unfortunately, he doesn’t notice that this implies that sounds do not perceptually represent anything beyond themselves. They are things we hear; they are not themselves auditory perceptions.
II. Sounds vs Soundmakers
Let me return to the beginning. I hear a meow; I hear a cat. What are the grounds on which Maarten says this? Some think that one hears only sounds, never soundmakers. What are the grounds on which he would oppose these people? Unfortunately, he does not tell us. He says only that it is an “assumption,” and that “Authors writing on perception today typically agree that at least some of the sources of sounds are themselves audible.”
Here’s the problem. Some think we hear soundmakers because of the following argument:
(a) We often say things like “I hear a cat,”
(b) It is generally pretty clear what we mean when we say things like this.
(c) Utterances like the one in (a) are often true in the sense attached to them in (b)—whatever that sense might be.
(d) Therefore, we hear things that make sounds, like cats—in some sense or other.
Call this the argument from ordinary language. I find this argument completely worthless because it tells us nothing about audition as such; it doesn’t tell us anything about how cats impose themselves on auditory experience. Maarten himself gives us a rich and nuanced account of how we hear soundmakers. The argument from ordinary language has nothing interesting to say about this (or about anything else, for that matter, aside from the vagaries of linguistic usage). He really shouldn’t rely on it.
More interesting arguments try to show how we can, by using only hearing, without the aid of inference and judgement etc., identify soundmakers, or (more commonly) why we cannot. Now, this depends substantially on what one means by “using only hearing.” On the approach that I espouse, one can use hearing actively, i.e., one can use it while moving around and exploring one’s surroundings. (Remember what I said about 3D sounds earlier.) Without going into details, one can (as blind people also do) learn to move around one’s environment following sounds to their sources. This is how we hear soundmakers—we identify the sources of sounds by actively “following” the sounds they make. This is an attitude that fits well with what Maarten says (though it’s also possible that his approach is intended to be more “ordinary language”). This is where I found his approach very valuable.
Let me conclude with a thought about sounds—about whether they are events that disturb the air, as Casati and Dokic (1994), Pasnau (1999), O’Callaghan (2007), and Matthen (2010) have argued, or disturbances in the air (sound waves), as it was generally thought before 1994. Are sounds events like a hammer hitting a string, or are they the compression waves that emanate from such events? An argument for the compression wave theory is that sounds are commonly thought to have properties such as pitch and volume. Sound producing events don’t have these properties. An argument for thinking that they are air-disturbing events is that audition is locational: we hear sounds occurring at a distance—in the confined places from which sound waves emanate. This contradicts the sound wave theory, because sound waves are all around us; they are not confined to the place that we hear a sound coming from. On the face of it, both theories have merit; yet each implies that the other is wrong.
On Maarten’s theory, it is possible to reconcile the two conflicting considerations. We hear soundmakers. This accommodates the locational aspect of auditory perception. We hear the Jarrett concert as occurring somewhere in front of us, but the sound waves are all around us. It is the soundmaker that we hear at a distance. We hear the Jarrett music as being out of tune. Yet, neither Jarrett nor the event of the hammer striking a string was out of tune. It is the sound waves that are out of tune. What to do?
On Maarten’s theory, there are two distinct things we hear: a sound and a soundmaker. On this theory, it is possible to hold that sounds are disturbances of the air. But on his theory, we also hear soundmakers. The locational function of auditory perception is adequately accommodated in terms of these. We can revert to the theory of sounds that was held before 1994. Sounds are disturbances of the air, but they are not the only things we hear.
Maybe we need to include a third object: we also hear a hammer striking a stretched string, or a loudspeaker membrane vibrating. I see no reason to exclude these events. We hear Jarrett (soundmaker); we hear the hammer hitting the string (sound event), and we hear disturbances of the air (sound).
REFERENCES
Casati, R., and Dokic, J. (1994) La philosophie du son Nîmes: Chambon.
Matthen, M. (2010) “On the Diversity of Auditory Objects,” Review of Philosophy and Psychology 1 (2010): 63-89.
O’Callaghan, C. (2007) Sounds New York: Oxford University Press.
Pasnau, R. (2007) “The Event of Color,” Philosophical Studies 142: 353–369.
Walton, Kendall (1984) “Transparent Pictures: On the Nature of Photographic Realism,” Critical Inquiry 11: 246-77.
Maarten Steenhagen’s paper on sounds explores the fascinating topic of sounds reproduction and its implications for a philosophical theory of audition. It offers a stimulating discussion about auditory intermediaries and their special role in the debate opposing representationalist and naive realist theories of perception.
In his paper, Maarten proceeds roughly as follows. He first argues that sounds are perceptual mediators in the sense that sounds are objects of perception fulfilling the function of making other objects, their sources, audible. He then argues that, as perceptual mediators, sounds can function as perceptual representations in the sense that they enable the subject to hear the source of the sound even when the source is absent from the current perceptual situation. The representational aspect of sounds is defended by discussing the case of recorded sounds and the auditory experiences we can have when listening to them. In the last parts of the paper, Maarten discusses the notion of perceptual representation and consider in what way we can say that listening to a recorded sound reproduces the perceptual conditions that were present when that sound was originally produced. His conclusion, if I understand correctly, is that philosophers have often rely on a misleading dichotomy between direct awareness of objects and a representationalist view of perception. According to Maarten this dichotomy is misleading since in audition, at least, both are involved.
My main concern about Maarten’s paper is the lack of explicit account regarding the ontology of auditory perception and the nature of sounds. As I will try to show, the absence of explicit exposition regarding the ontology of auditory perception clouds considerably Maarten’s position and makes it vulnerable to many objections.
1. The nature of sounds
Maarten distinguishes explicitly sounds from their sources, but he unfortunately tells us very little about their nature.
Different clues let me think that Maarten considers sounds to be proximal entities. He tells us for example that sounds travel from their original position to the listener’s ears and that the sounds are in the vicinity of the subject whereas their source can be distal. But the proximal view of the nature of sounds is not the only possible approach to the nature of sounds. And I believe that the distal approach to sounds presupposed by Maarten greatly affects what his paper tries to show.
It seems for example that the main reason Maarten considers sounds to be mediators in the perception of their source is that sounds give access to objects that are not in the hearer’s immediate vicinity. He says :
It seems therefore that if sounds are not proximal but distal entities the role of sounds as perceptual mediators vanishes. If sounds function as a kind of causal and spatial relay for the perception of entities located at a distance they must be located between the subject and the sources. But unfortunately Maarten gives us no reason to endorse a proximal theory of sounds.
2. The nature of sources
But maybe the location of sounds is not relevant for the role played by sounds as mediators. Maybe Maarten’s idea is to show that objects, like cats, can be heard only indirectly. They can be heard only by hearing directly some other entities like sounds.
But here again, we have too little information to appreciate Maarten’s view. What is the nature of the sources? Are they objects (like cats) or events (like explosions)? What is the ontological relationship between the sounds and their sources ? Does this relation explain why hearing sounds give access to their sources? And why should we accept the general claim that we can hear the source of a sound? We surely can hear the cat playing and meowing but I don’t find it so obvious that we can hear the cat itself. By listening a sound we certainly acquire some knowledge about the object that has produced the sound, but it is far from evident that this knowledge rests on a direct or indirect perception of the source itself. Sounds, like footprints, enable to recover information about the objects involved in their production. But in the same way that we don’t see an animal by seeing its footprints, it seems that we don’t need to hear the source of the sound to acquire information about its size, material constitution, structure, etc. and eventually be able to recognize the source of a sound.
3. The nature of recorded sounds
The lack of commitment regarding the nature of audition and sounds also affects the way recorded sounds are presented. Consider Maarten’s claims that ” such audio recordings is that they allow sounds that sounded on an earlier occasion to sound once more” and ” So, despite the fact that there is no piano around, the sound of such an instrument is genuinely there in your room.” The original sounds have been produced by Keith Jarrett playing on a piano at the Opera House in Cologne. When listening to the record, Jarrett, the piano and the Opera House are not where the record is played. So what does it mean to say that the original sounds “sound once more” and that the piano “is genuinely there in your room”? I think that Maarten owes us a more detailed explanation of the relation between the original sound and the recorded sound. Maarten seems to consider that his approach to audition can accommodate diverse ontological theories of sounds. He writes :
I don’t agree with Maarten’s view on this matter, since I believe like Brentano, that perception, like the other mental states, is essentially intentional, which means that mental states are essentially of or about things. Understanding perception involves therefore to capture the nature of this aboutness or directedness of mind and therefore to account for the ontology of perceptual experiences, their intentional objects and the relation between them.
In the case of recorded sounds, the nature of the intentional object seems crucial since it determines very different theories of audition.
Like Maarten, we can believe that recorded sounds mimic in a certain way the sounds that were heard in Cologne while assisting to Jarrett’s performance. In that case, the sounds you hear in your living room when listening to the record and the sounds emerging from Jarrett’s piano must have a singular relation that explains the fact that the sounds heard in your living room can make you enjoy a concert given in Cologne more than 40 years ago.
But there is another option which makes this particular relation unnecessary. According to this option, there is no reproduction of sounds, since the sounds you hear in your living room are identical with the sounds that were produced by Jarrett’s piano in Cologne. According to this view, the greatness of Scott de Martinville and Edison is not to have invented devices able to reproduce sounds but rather to have created instruments literally able to store and transmit them to future music lovers.
The fact that the sounds where produced miles away and years ago is not really a problem for those who defend a distal theory of sounds. It is certainly more inconvenient for those who hold a proximal theory of sounds.
Certainly, much more need to be said about what it means to store and to transmit sounds, but I don’t see why a transmission theory of recorded sounds should not be a viable approach to explain what is involved by hearing recordings. Anyway, it seems that whatever the nature of audition of recorded sounds one defends, there is no way to dispense with a satisfactory ontology of sounds and audition.
4. Naive realism vs representationalism
If we take the transmission theory of recorded sounds seriously, there is no representation involved in hearing something through a recording device since there is no intermediary or mediator between the subject and the original sound. The fact that the event heard happened many years ago illustrates the fact that artifacts, like phonograph or telescope, give us access to regions of reality that are not accessible with our natural perceptual equipments. In the same way, telescopes extend our visual perception to objects distant in space and time, phonographs extend our auditory perception to objects or events distant in space and time.
Contrarily to Maarten’s conclusion, I therefore think there is no need to introduce intermediaries into the relation between the perceivers and the intentional objects of perception. Auditory perception like all perception gives us a direct access to reality.
From a more general standpoint, I don’t believe like Maarten that representationalism and naive realism are somewhat compatible. Like black coffee, which ceases to be black after adding only a drop of milk, there is no naive realist approach to perception with a little pinch of representationalism.
Hearing a distant body, such as hearing the bell of Bow Church in East London, intrinsically has a mediated structure. We hear the bell by (or in virtue of) hearing some sound that reaches our ears. The sound of the bell functions as a perceptual mediator. In his comments, Leddington rightly suggests that I take such mediators to be particulars, not properties. As I conceive of it, a perceptual mediator is some perceptible particular (x) that makes something wholly distinct from it (y) perceptible, insofar x is itself perceived. If mediators are particulars wholly distinct from some more distal object or scene, then this suggests a contrast between those senses that for their proper functioning depend on the availability of mediators in our environment (audition, olfaction), and those senses that do not (vision, touch). In a way, my paper builds on that contrast.
I take sounds to be perceptual mediators. Understanding sounds to function as perceptual mediators in auditory experience does not commit one to a robust view of what sounds are, apart from that sounds are themselves audible particulars. That sounds are audible particulars produced by their sources seems to me a more natural view than the view that sounds are audible properties of their sources––a view Leddington thinks I should take seriously. Consider the following. When we end the sound of a bell by removing the air around it, we can do so without changing the bell in any way, proving that the sound is not a property of its source. According to Leddington, who realises that I am moved by such considerations, this shows that our disagreement lies at a deeper level. Ultimately, it may be that he and I disagree about the admissible sources of sounds. Mizrahi puts the question to me in clear terms. She seems sceptical about my understanding of what sound sources are. “Are they”, she asks, “objects (like cats) or events (like explosions)?” In my view both could be a source of sound. Just as both a person and an earthquake can produce damage or rubbish, they can both be producers of sound. In fact, I think it is more common to identify bodies as the sources of sounds, and I think this is nicely reflected in a term suggested by Matthen: many bodies are ‘soundmakers’. This of course does not take away that, as Leddington puts it, “Mere objects are silent” and that “[o]nly in activity is their noise”. Yet I do not think that from these facts it follows that “sound sources are always events”, as Leddington urges. I regard it as more accurate to say that a piano is only a source of sound if it participates in some event, such as its being played or thrown off a stage.
Mizrahi suggests that I need to say more about the nature of sound. In particular, she thinks I need to affirm explicitly that sounds are proximal entities, as opposed to distal ones. A distal conception of sounds, she thinks, would have no room for sounds’ playing a mediating function. In principle, I am happy to say a bit more. I prefer to think of sounds as extended, and as both distal and proximal: the sound originates at a source, and hence is located over there; but over the course of its life the sound spreads out, like the concentric disturbance on a pond. In this way a sound, while being over there, can reach your ears here, eventually. But I don’t think this really matters for my argument. You can accept the conclusions of my paper even if you do not share my specific conception of sound. Ultimately what matters is only whether we need to hear a sound in order to hear anything else, and whether a sound is wholly distinct from what it makes heard. If you affirm both, you have paved the way for sounds as perceptual mediators. Further issues can be left open, contrary to what Mizrahi suggests.
The central argument in my paper makes the step from mediation to representation. I bring out that because auditory perception depends on perceptual mediators, it harbours the possibility of a form of perceptual representation. The rough idea is that because mediators are particulars independent of the things they enable us to hear, we might be able to capture or record a mediator, and make it available for perception at a later point in time. The technology of sound recording has actualised this possibility, and as a consequence we can now exploit the representational possibilities of the auditory sense on a daily basis using stereo sets, iPhones, and other devices.
Both Mizrahi and Matthen suggest a way to conceive of sound recording that is different from mine (as they acknowledge). They suggest we can think of the recorded sound as a continuation of the original sound, instead of as a repetition. This is really quite nice, and worth exploring as a possible form recording technology can take. Yet (as both Matthen and Mizrahi acknowledge) it differs from how I am thinking about what recording technology does. When we try to capture the nature of sound recording and playback, a decisive reason against the ‘continuation’ or ‘transmission’ view of sound recording is that, when playing back a recording, we can repeat it. In experiencing such repetitions we seem to hear the same sound over and over again, instead of hearing it continue with intervals of silence. Because playback is repetition, we can should conceive of it as re-introducing a sound (a perceptual mediator) at a later point in time. But, I argue, it is a re-introduction that liberates the sound from its original source. Because it is the second (or third,…) instalment of its sounding, so to say, the played back sound no longer depends on the original source. And it is because this sound none the less suffices for making audible a source, the state of audibility of this source generated by the play-back is now representational. Kulvicki rightly points out that for other sensory modalities that I conceive of as having a mediated structure, such as olfaction, we do not have comparably ripe practices of technological representation. This is right. But such a development can equally not be ruled out. In fact, several start-ups are well on their way to refine these technologies.
Sound recording techniques, on my conception, supply the sounds in our environment with a status that is even more special than that of being a mediator. Because of how played back sounds no longer depend on the presence of the source they can, besides functioning as perceptual mediators, become perceptual representations as well. Matthen worries that this is confused, because sounds are external things, and external things cannot be perceptual representations. Admittedly, my claim is provocative. It is right that presently perceptual representations are theorised in exclusively mental terms: they are conceived of as brain states, perceptions, or mental imagery. But remember, some of the early sense-datum theorists were able coherently to maintain that the sense-data present in sensory experience were mind-independent, physical things such as facing surfaces. Likewise, it is not incoherent to maintain, as I do, that at least some perceptual representations are external to the mind. It is in this sense that sounds can both be perceptual representations and public objects of perception. (The analogy with photographs, which I will address next, is of use here. Despite being external to the mind, most of us are willing to count photographs as representations as well.)
Because my argument makes so much of these auditory technologies, it is tempting to wonder whether vision could not allow for similar technologies as well. The obvious candidate here is the technology of photography. In their comments, Kulvicki, Matthen, and Leddington pick up on this analogy, each using it to make different observations. Kulvicki sees a problem with the parallel. He observes that photographs can be conceived of as perceptual mediators in the way played back sounds can. I think this is right. But according to Kulvicki, this fact shows that the contrast I want to draw between vision and audition is exaggerated. He writes that “if vision is really so unlike audition it’s a bit of a surprise that we have found a visual phenomenon deeply analogous to the auditory case.” Perhaps it is surprising that mediation is possible in visual experience, I agree. But this does not undermine the contrast between audition and vision. The point is that visual experience of the material world does not depend on perception of photographs or images, whereas auditory experience of the material world does depend on the perception of sounds. We can see the material world immediately, placing vision on a par with touch. If we tried to make sense of the idea that photographs make other things visible using the model ordinary vision suggests, the only reasonable route would be to look at optical technologies such as mirrors or telescopes, and to argue that a photograph is a natural extension of what these technologies afford. This naturally leads to the idea that photographs are ‘transparent’: they make things visible insofar as they are themselves not perceived, acting more or less like media such as glass, lenses, or (perhaps) mirrors. This is Kendall Walton’s strategy (Kulvicki, Matthen, and Leddington mention his work). I do not take Walton’s strategy. I do not, because I think ordinary vision provides a poor and too restrictive model for thinking about perceptual representation. In my view, we can better approach photography on the model of auditory perception, and claim that photographs fulfil a role in visual experience that parallels the mediating role of sounds in auditory experience.
In the end, I am happy to acknowledge that my claim that sounds can act as perceptual representations is easily extended to photography. But this does not undermine the fundamental difference between auditory perception and vision; between those senses that depend for their proper functioning on the availability of mediators in our environment, and those senses that do not. To put it boldly, to appreciate the representational nature of photography requires you to acknowledge that human experience can encompass both immediate and mediated forms of awareness. More generally, a corollary of this pluralism of forms of perceiving is that, possibly, some but not all forms of perception allow for perceptual representation. To think that the nature of perception is either exclusively representational or exclusively relational is in my view too simplistic.
Thanks, Maarten, for this illuminating reply. I think I’m getting clearer about your view and about where our disagreement lies.
Here’s just one point for the time being. You write:
First, a quibble: remove the air from around the bell and you will “change the bell” in the sense that you’ll change how it vibrates.
But the main point I want to make is that the argument from vacuums is not helpful in deciding the issue between sound-particularists and sound-generalists. Here’s why: if sounds are properties of their event sources, it’s still an open question whether those event sources are best individuated relationally or non-relationally vis-á-vis the relevant media. This means that it is an open question whether evacuating the air around the ringing bell destroys or preserves the sound source. (If the sound source is medium-relative, then introducing a vacuum destroys it.) Therefore, your argument goes through only if you can show that the property theorist is committed to treating sound sources non-relationally. (And even if you can do this, you still have to provide independent reason to think that introducing a vacuum destroys rather than merely hides sounds.)
Thanks for these remarks, Jason. I think you are absolutely right that the argument from vacuums is not helpful when we want to settle the debate between sound-particularists and sound-generalists. In my response to you I used it together with the specific assumption that sound sources are bodies, such as copper bells, and the assumption that sounds are either particulars or properties of their sources. I think the argument then becomes more relevant, though I admit that more needs to be said, as one might say.
Hi Maarten,
Thanks for an interesting paper and discussion! I want to ask you about your claim that sounds are PERCEPTUAL mediators. Like Mohan, I wasn’t sure why you think that the fact (if it is one) that sounds represent their sources makes sounds PERCEPTUAL mediators. I think this is important because you take your argument to show that auditory perception is (at least partly) representational. But, I take the ‘representation vs relation’ debate to be about the relation we stand in to the objects of perception (so to sounds, on your view of them as particulars), not about the relations which hold between different objects. It would really help me to understand your view if you could answer this question: 1. are sounds PERCEPTUAL mediators in a world with no sentient life (so, no perceivers)? I think it would be strange to answer ‘yes’, but if you answer ‘no’ I think more metaphysical work needs to be done. (If sounds require perceivers to represent sources, then this will complicate your metaphysics of sounds as objective particulars.)
Thanks!
This is a good point to raise, Laura. It also pushes the issue into difficult areas. Here is what I think now. In a world without sentient life, sounds would not be able to fulfil a function of perceptual mediators. This is because something can only fulfil that function if it actually increases what some perceiver or other can perceive. But by hypothesis, in the world you ask us to imagine there are no perceivers. So nothing can fulfil the mediating function. This means that there is a sense in which perceptual mediation, and (a fortiori) perceptual representation of the form I discuss in my paper, do require perceivers.
You suggest that this conclusion will complicate my metaphysics of sounds as objective particulars. I am not sure if it does complicate this metaphysics. But even if it does, I do not think the complication is problematic. Sounds are natural phenomena that fulfil a mediating function in our experience. In a world without sentient life, though sounds would not be able ever to fulfil this mediating function, they could nonetheless exist as objective particulars. We can imagine that in such a world trees would still produce a sound when falling down in a storm, and that the sound of thunder would still echo in a canyon. That sounds are mind-independent particulars is essential to them, whereas the fact (and I think it is one) that sounds are mediators for their sources is a contingent fact about them. And because this is a fact that (in part) depends on a functional role, a change in the setting in which sounds occur may very well undo their having this special status.