We hear silences. But what are silences? And what analysis should we give of the auditory experience of silence? Part of what makes this second question interesting is a phenomenal contrast between experiences of silence and those experiences of absence that occur in other sense modalities, most notably experiences of darkness in vision. In the latter case, it seems plausible enough to claim that our experiences have a qualitative character with qualitative similarities to visual experiences of phenomena that are not absences. For example, the blackness that is introspectively salient when sighted subjects normally enjoy experiences of darkness can be modelled as located relative to other colors within a qualitative space, as in the Munsell solid. For auditory experiences of silence, by contrast, the claim that there are qualities characterizing those experiences seems not to be supported in so obvious or compelling fashion by introspective reflection on what it is like to have such experiences. In experiences of silence, there is no feature of the experience comparably salient to blackness in experiences of darkness. Sorensen (2008, 272), in his influential discussion of this issue, puts the point like this: “… silence does not auditorily resemble sound. Silence has no loudness, timbre, or pitch.” But these phenomenological points about the introspectable features of such experiences of absence raise the following question: How exactly should we account for this phenomenal contrast? Particularly pressing is a concern over how one should fit any answer to this question within one’s preferred account of experience of silence.
Sorensen summarizes the phenomenal contrast by claiming that experience of silence is “the most negative of perceptions.”
In what follows, I will call this claim the thesis of negativity. He articulates the thesis of negativity in the following way, “There is nothing positive being sensed and no positive sensation representing that absence” (2008, 272). The first conjunct of this claim is really part of an answer to the first question above: What are silences? For Sorensen, they are no positive entity; they are instead a “negative” entity—an absence. The second conjunct serves to explain the phenomenal contrast between experiences of hearing silence and of seeing darkness. When seeing darkness, our visual experiences have sensations; when hearing silence, our auditory experiences do not. This claim has been quietly influential: subsequent debate about experience of silence has been shaped by Sorensen’s thesis of negativity. In particular, two further ways have been suggested for how to capture the thesis of negativity and its motivating experiential contrast. Phillips (2013) and Soteriou (2011) have each put forward an account of experience of silence in keeping with the thesis, and which explains the phenomenal contrast between experience of silence and experience of darkness. However, owing to a range of shortcomings of Sorensen’s preferred account of silent auditory experience, the underlying analyses of the experience of silence offered by Phillips and Soteriou emerge as competitors.
In what follows, I will argue that both these competitors face significant problems of their own when we keep in view the central constraints of the debate accepted by these recent contributors. However, I will urge, this offers no support for Sorensen’s account of the experience of silence. Rather, it forces us to scrutinize more carefully the thesis of negativity. The structure of the discussion is as follows. In the opening sections of the paper, I situate Phillips and Soteriou’s claims about experience of silence within a framework of possible accounts of the experience of silence. In sections 4 and 5, I present my objections to their views and defend against possible replies. In the final section, I return to the question of the plausibility of the thesis of negativity.
2. Theories of Silence
We can distinguish between cognitive and noncognitive theories of experience of silence. According to cognitive theories, defended in detail by O’Shaughnessy, to hear silence requires “… a cognitive attitude, with silence figuring in its content …” (O’Shaughnessy 2000, 329). Experiencing silence amounts to this: you fail to have an auditory experience and consequently judge that there is no sound. One putative attraction of this approach is that it captures the thesis of negativity; it does this by denying that experience of silence involves an auditory experience. This position faces significant difficulties, however. As Phillips has urged, an hallucination of silence might not involve knowing, believing, or even being inclined to believe that it is silent. But then the cognitive theory appears unable to offer us an account of what distinguishes hallucinating silence from deafness. I will not dwell further here on the merits or otherwise of cognitive theories because the targets of the present discussion are all noncognitive theories of experience of silence. However, for further elaboration of his arguments against such approaches, see Phillips (2013, section 2).
Noncognitive theories of experience of silence fall into at least three kinds. The first of these is Sorensen’s target: the claim that, just as with experience of darkness, experience of silence involves a sensation. It is the denial of this claim that is operative in Sorensen’s formulation of the thesis of negativity. “Sensation” is, however, a highly contested theoretical term in the analysis of perceptual experience. What Sorensen seems to have his sights on is the following idea: (i) that experiences of silence instantiate, or are characterized by, some qualitative features. Recall that, according to Sorensen, the contrast between experience of silence and experience of darkness arises because experience of silence is qualitatively unlike experience of sound. In contrast, experience of darkness is qualitatively similar to other chromatic experiences. For this reason, Sorensen (2008, 270) rejects this conception of experience of silence: in his view, to construe experience of silence as a sensory experience having a “negative” object conflicts with the thesis of negativity.
Sorensen prefers (ii): a theory construing experiences of silence as not characterized by any auditory qualities. This view is consistent with the thesis of negativity but has itself been the target of compelling criticism. As Phillips argues, it struggles to capture the distinction between hearing silence, being deaf, and hallucinating silence. If being deaf, hearing silence, and hallucinating silence are states which are each to be analyzed in terms of failing to have auditory qualities, then they will be subjectively indiscriminable. A newly deaf person might be unable to tell on introspection alone whether they are deaf or hearing silence. But, in this case, the deaf person is simply hallucinating silence: having an experience subjectively indiscriminable from hearing silence, but which is not a perception of silence. The problem is: deaf people are not perpetually hallucinating silence.
Departing from Sorensen, Phillips (2013, 338–40) articulates and defends a version of the third construal of experience of silence now on the market: (iii) experiences of silence are auditory experiences, but states of sheer auditory awareness. This strategy was also independently developed by Soteriou in response to aspects of Sorensen’s discussion. The suggestion is, in fact, a kind of spiritual successor to (ii), because it is a corollary of the view that experience of silence is awareness that is not characterized by any qualitative properties. One advantage of this formulation over Sorensen’s, though, is that it enables us to distinguish deaf subjects from subjects hallucinating silence by means of the claim that deaf subjects do not have auditory experiences at all, whereas hallucinating subjects have auditory experiences with no qualities. Clearly, this account of experience of silence is also consistent with the thesis of negativity: experiences of silence are “negative” relative to most other sensory experiences insofar as most other experiences involve awareness plus qualities. For ease of reference, I will call the core claim of this view of experience of silence the thesis of objectless experience.
There are two articulations of the thesis of objectless experience that have been developed in recent years, one due to Phillips and the other due to Soteriou. Phillips develops a version of the thesis along Moorean lines, according to which the structure of all experience involves two components: an awareness and some object to which that awareness is directed. Phillips defends the idea that experience of silence can be construed as merely involving an auditory awareness with no corresponding object to which the awareness is directed. Call this the pure-awareness theory. This offers a novel way to account for the thesis of negativity: experiences of silence are the most negative of all experiences because they alone are objectless. One of the major challenges for this view, as Phillips sees it, is to articulate a justification for the claim that such awareness would be auditory in character. Phillips urges that such pure awareness is auditory in character because it enables the subject to listen. The second way of articulating the thesis of objectless experience, also mooted by Phillips but defended at greater length by Soteriou, is to claim that experience of silence is objectless but has a temporal field. The idea is that auditory awareness has, in addition to an object of awareness, a form or structure of that awareness. In the case of audition, the form of awareness is temporal. On this view, the thesis of negativity is captured in the same manner as the foregoing version—there is no object of awareness—but what is distinctive about auditory experience is captured in terms of its temporal structure. Call this the structured-awareness theory. In what follows, my aim is to articulate difficulties facing both of these versions of the thesis of objectless experience, arising from other core commitments of the approach.
3. Denying causal dependence
The debate about how to understand experience of silence is also shaped, in part, by a denial of a further claim; one that, in spite of reflecting common-sense thinking about hearing silence, presents some widely acknowledged philosophical difficulties. This is the claim that we hear silences because those silences cause our experiences of silence. Call this the thesis of causal dependence. Sorensen is committed to this claim insofar as it is a corollary of his commitment to a further, stronger claim about causation and perception: that causal conditions are built into the concept of a perceptual experience. In Seeing Dark Things, Sorensen remarks on the predictive success of the causal theory of perception: “Indeed, I think the causal theory of perception, when unencumbered by a metaphysical outlook skewed in favour of positive things, makes bold, correct predictions about our perception of dark things” (2008, 6). And, “I think these acausal theories exchange a penetrating, empirical picture of perception for paper-thin conceptual cartography. This becomes evident when ‘conceptual truths‘ about vision are refuted by surprising consequences welling up from the causal theory of perception” (7). Whether or not this stronger claim is true, the weaker causal dependency claim will be sufficient here. That claim, however, runs into philosophical difficulties on several grounds: it is ontologically inflationary insofar as it implies the existence of absences. It might also be viewed with suspicion insofar as it generates the usual range of puzzles about absence causation. For these reasons, it might be thought that it is preferable to dispense with the idea that experience of silence depends causally on silence. This is an explicit motivation for Soteriou’s defense of the structured-awareness theory. According to Soteriou (2011, 198–99), we can say that an experience of silence counts as perceptual for the following reason: the form of the experience of silence is temporal, auditory experience has a temporal field. It is in virtue of this temporal field that you hear the temporal region your experience occupies. So, in hearing silence, one hears the temporal region as empty of sound. In contrast, Phillips’s pure-awareness theory is not explicitly motivated by the causal considerations Soteriou addresses, but it seems clear that Phillips’s view would gain the same putative theoretical advantages if a similar claim were advanced. Phillips need not go so far as to acknowledge that auditory experience has a temporal field. Instead, the pure-awareness theory could be coupled with the following claim: in having an objectless experience taking up a temporal interval, one thereby hears that temporal region as empty of sound.
Whichever way one may choose to develop it, there are two distinct claims involved in this strategy of avoiding a causal condition on the perception of silence by appealing to the temporality of auditory experience of silence. The first is a claim about how it is that what is heard is determined. According to this strategy, when we experience silence, what is heard is not determined causally: there is no causal dependency of our experience of silence on some absence. Instead, what is heard is determined by the sheer temporality of auditory experience. The second claim concerns what is heard when we hear silence: a claim about what silence is. On Sorensen’s view, silences are entities that supplement an ontology of “positive” entities. By contrast, both Phillips and Soteriou claim that silence is a temporal region devoid of sound. That is, silences are simply certain kinds of temporal intervals, and not entities that supplement whatever is one’s preferred ontology of time and temporal regions. Call this the thesis of temporal regions. To hear silence is to hear a temporal region. So, the proposal is: to hear a silence—to hear a temporal region devoid of sound—is to have an objectless auditory experience coextensive with that temporal region.
The discussion thus far has turned on three claims that shape the debate about experience of silence:
The thesis of negativity.
The thesis of causal dependence.
The thesis of temporal regions.
Phillips and Soteriou’s respective articulations of the thesis of objectless experience agree on how to respond to these three theses. They endorse the thesis of negativity; and the thesis of temporal regions provides an account of what is heard when one successfully perceives silence. The thesis of objectless experience appears to fit comfortably with a denial of the thesis of causal dependence, which there is prima facie reason to be wary of endorsing in cases of absence perception.
4. Problematic cases
With regards to hearing silence, I will urge, there are three plausible claims that, prima facie, are at odds with the denial of the thesis of causal dependence. So, we are at least owed an explanation of why the thesis of objectless experience is, in fact, compatible with these claims, or, alternatively, why these claims are, contrary to appearances, false.
The first claim is:
(i) Subjects hearing silence in causally insulated surroundings hear numerically distinct, particular silences.
Consider two people A and B, each in neighbouring concert venues (concert halls 1 and 2) attending exactly synchronous, but distinct, performances of John Cage’s 4′ 33″. Concert halls 1 and 2 are causally insulated in the following sense: the sounds and silences in each hall are the causal product of the performers in that hall, not the other one. Notwithstanding the fact that Cage’s 4′ 33″ was intended to encourage listening to ambient sound, imagine that in this situation the performance in each hall is, in fact, perfectly silent: there are no ambient sounds to be heard. Subjects A and B are hearing distinct performances, and they are hearing those performances by hearing distinct silences in each hall. Subject A hears the silence in concert hall 1, subject B hears the silence in concert hall 2; moreover, the silence in concert hall 1 is not numerically identical to the silence in concert hall 2.
The particularity of the silences we hear can also be brought out as follows. The walls of the concert halls are causal blockers: they block, or prevent, each subject from hearing the silence in the other concert hall. This implies that the silences are causes, or at least potential causes. Preventers, after all, prevent causes from causing. Recognizing that subjects can hear causally insulated silences, then, leads very naturally to the claim that the auditory experiences of such subjects causally depend on the particular silences they are not causally insulated from.
The two further claims in tension with the denial of causal dependence are:
(ii) We sometimes hear silences indirectly, by hearing other silences.
(iii) Which silences we hear varies according to causal considerations.
Claims (ii) and (iii) are relevant to determining, in causally insulated situations where subjects are having silent auditory experiences, which silences are heard. In support of claim (ii), consider this modification of the previous example: subjects A and B both sit in a third room hearing distinct John Cage performances by means of over-ear headphones. There is a silence between the ear and the headphone, which causally depends on the silence in the concert halls. This case is just the auditory analogue of seeing football players in virtue of seeing their images on television. The latter is well explained by the causal dependency of the latter on the former. By analogy, what we should say in this case is that A and B hear distinct silences in the different concert halls in virtue of hearing silences at their ears.
Turning to (iii), consider the following permutation of the Cage example: A is in concert room 1 and B is in concert room 2. However, this time A wears headphones causally sensitive to changes in room 2, and B wears headphones causally sensitive to changes in room 1. Given (i) and (ii), we should say that A hears the silence in room 2, B hears the silence in room 1. This is exactly what we would say in the case where what is heard is not a silent performance.
Each of the three foregoing examples suggest that claim (i) is true. But once we reflect further on (i), it turns out that this generates a significant problem for the ways Phillips and Soteriou each develop the thesis of objectless experience. Why is this? The reason is that (i) is incompatible with the claim that what is heard in perception of silence is a temporal region. If the latter were true, then in the first scenario described, subjects A and B would both hear temporal regions. But then, because the two performances are exactly synchronous, A and B would hear the very same temporal region. But this claim is, by itself, implausible: subjects A and B, it seems right to say, hear numerically distinct silences. What has gone wrong? Well, clearly the denial of the thesis of causal dependence, with its unpopularity amongst contemporary theorists of causation, provides theoretical cover for this implausible commitment. The challenge of making sense of absence causation at least lends some credence to it. Nonetheless, this is a significant—in my view prohibitive—cost to accepting Phillips and Soteriou’s development of the thesis of objectless experience.
Restricting the discussion for the present to a firm commitment to the thesis of temporal regions, it might be thought that Phillips and Soteriou could escape this difficulty by denying that this thesis has the implication that subjects A and B would hear one and the same temporal region. This could be done by acknowledging exactly overlapping but distinct regions of time. It would then be possible to claim that subjects A and B each hear one, but not the other of these. Such a commitment faces several serious problems, however. First, temporal regions are plausibly individuated by their locations in the temporal dimension. So, if temporal regions x and y possess all the same temporal relations to all other regions of time, they will simply be identical. Note the further problematic implication this has for the thesis of temporal regions: it would simply be impossible for there to be more than one co-occurring silence.
Second, acknowledging overlapping regions of time faces a problem analogous to one faced by a solution to the ship of Theseus puzzle that postulates multiple, exactly overlapping ships that are subsequently separated by the processes of renovating and reconstructing the original ship. With respect to the ship of Theseus puzzle, we are owed an explanation of why there are exactly two overlapping ships, rather than some other number. After all, one could simultaneously renovate and reconstruct either of the newly separated ships. In the present case, we are owed a reason for thinking that there are only two exactly overlapping temporal regions: Why not some other, larger number? After all, if there had been three performances occurring, then there would have to have been three overlapping temporal regions, etc. It seems right to conclude, then, that if the thesis of temporal regions is true, causally insulated subjects will not hear numerically distinct silences.
If this is correct, then the following exclusive disjunction is true: subjects hearing silence in causally insulated surroundings hear numerically distinct silences, or, when hearing silence, subjects hear temporal regions. So, which disjunct is true? First, a dialectical point: the claim that we hear temporal regions is part of a philosophical theory, whereas the claim that subjects in causally insulated environments hear distinct silences is data to be explained, or explained away, by a philosophical theory. So, it appears that the onus is on Phillips and Soteriou to explain why, contrary to appearances, subjects do not hear numerically distinct silences in situations where they are causally isolated. Moreover, we are owed an explanation of why this is compatible with the cases appealed to here in support of (i).
One strategy might be to question the following assumption of the foregoing argument: that subjects hear the performance by hearing the silence. Instead it might be urged that hearing a silent performance is independent of hearing silence. How might this be done? The following is a possibility: one account of hearing silence maintains that subjects hear silence by hearing the ending and beginning of sounds. This is sometimes known as the contrast account of hearing silence. Although, as a theory of hearing silence, this is a competitor to the thesis of objectless experience, it might be modified to explain why subjects A and B hear distinct performances. The idea would be that despite hearing one and the same silence, subjects A and B hear distinct performances because they hear distinct sounds ending. Subject A hears the performance in hall A because she hears the ending of the sounds in hall A that mark the start of the performance; the same goes, mutatis mutandis, for B.
The principal difficulty facing this objection is that it is only incidental to the significance of the example that the silences in question are performances. The same difficulty would arise if the concert halls were being used by subjects A and B for contemplative silent meditation or no purpose at all, rather than listening to an avant-garde musical performance. These would be cases where one is simply hearing the silence and nothing else. So, it’s not clear how this line of objection could possibly offer an adequate reply to the problem.
Moreover, the view about hearing silence, from which the forgoing suggestion is drawn, faces significant problems noted by both Phillips and Soteriou. Crucially, these difficulties arise also for the derived account of hearing performances just proposed. With respect to hearing silence, it is possible to hear silence without hearing the ending of a sound. For instance, imagine subjects A and B sleep through the beginning of the performance and wake up halfway through. Then they would hear silence just as surely as if they were awake and listening from the beginning. Partly owing to this kind of example, Phillips and Soteriou themselves reject the contrast account of hearing silence. But the same problem affects the proposal that subjects hear silent performances by hearing the ending of sounds. Just as the subjects would hear the particular silences they awake into, so too would they hear the particular performances they awaken in the middle of. The contrast account offers no explanation of how this can be.
It might be thought that there is an available explanation of how subjects A and B hear the performances they do, but which does not require hearing distinct silences. Perhaps subjects hear the particular performances they do by being in the particular hall they are in. The idea is that the location of the subject determines which performance is heard. But this will not do for two reasons. Firstly, in the absence of causal considerations, it leaves it entirely mysterious why it is that being in a particular place should determine that a subject hears a performance occurring at that location. Where a subject is located is, in general, relevant to what she hears, but this seems to be because of causal features of its local environment.
Secondly, this suggestion forces a counterintuitive and implausible interpretation of the third permutation of the Cage performance scenario, in which subject A is in concert hall 1 but wearing over-ear headphones that are causally sensitive to changes in concert hall 2. The present suggestion implies that subject A, if hearing any performance at all, must be hearing the performance in hall 1, where she is located. This seems to be mistaken precisely because our intuitions about what is heard track the causal sensitivity of the headphones. Moreover, consider this case: imagine that the performance in hall 2 runs on longer than the performance in hall 1: people start moving around in hall 1 at exactly 4′ 33″, whereas the performers and audience lose track of time in hall 2 and remain silent for a further minute. In this case, subject A sitting in hall 1 cannot be hearing the performance in hall 1 because this has ended. If she is not hearing the performance in hall 2, then she is not hearing any performance. So, we will be forced to say that she hears no performance at all for the final minute. This interpretation, though not incoherent, is much less plausible than the claim that she hears, for the whole duration, the performance in hall 2. It seems, then, that we should not deny that we hear the particular silent performances we do by hearing the particular silences we do. But this returns us full circle to the problem with which we began: subjects hearing silence in causally insulated surroundings hear numerically distinct, particular silences, but this cannot be accounted for if hearing silence is a matter of hearing a temporal region.
5. Spatiotemporal fields
The discussion so far has taken as its target two proposals concerning experience of silence: Phillips’s pure-awareness proposal and Soteriou’s structured-awareness proposal. The argument developed here has assumed that, in both cases, experience of silence is nonspatial and that, consequently, silence should be understood as a temporal region unqualified by sound, rather than a spatiotemporal region. In the case of Phillips’s pure-awareness proposal, this commitment is required: Phillips’s proposal differs from Soteriou’s precisely with respect to whether awareness is structured. According to Phillips, experience of silence is neither spatially nor temporally structured in the sense that it has neither a spatial nor a temporal field: it simply occurs over a period of time. For the structured-awareness proposal, however, aspatiality is not an immediate consequence of the core commitments of the view.
As a consequence, we can distinguish between two flavors of the structured-awareness theory corresponding to whether experience of silence is taken to have only a temporal field or structure, or both a spatial and a temporal structure. The first, an aspatial version, will clearly face the problem outlined above that, in the initial Cage example, subjects hear one and the same silence. However, a version of the structured-awareness theory that construes experiences of silence as having a spatial field would avoid this problem because the silences heard will be individuated spatially as well as temporally. We hear the spatiotemporal region where the performance of Cage is occurring, not some other spatiotemporal region. We do this in virtue of the spatiotemporal field of our auditory experiences of silence.
At this point our topic intersects with the related question of whether auditory experience of sound is spatial or aspatial. This question has been the subject of considerable recent controversy (see, for example, O’Callaghan 2007, Casati and Dokic 2009, O’Shaughnessy 1957 and 2009, Nudds 2009, and Meadows 2018). According to distal theories of sound, sounds are entities that are located where, as they claim, mature auditory experience presents sounds as being: at a distance from the hearing subject. Part of the motivation for this view is the claim that mature auditory experience is spatial in character. Those who defend the view that sounds are either sound waves (or properties of these) or that they are sensations commonly resist the claim that auditory experience is spatial. Instead, they urge, any spatial content is a contribution of some cognitive component.
Clearly, then, the debate over competing accounts of sound is of some relevance here, but it doesn’t straightforwardly decide the issue of the spatiality of auditory experience of silence. This is because the most that can be concluded from the phenomenology of mature auditory experience of sound is that experiences of sound—“positive” auditory experiences—are spatial. To warrant a conclusion about experience of silence, what would be needed is some reason for thinking that the phenomenology of mature auditory experience of sound reveals something about auditory experience at a level sufficiently general to apply also to auditory experience of silence. But in the current debates on the metaphysics of sound, no such reason is given. So, even if positive auditory experience is spatial, it does not immediately follow that negative auditory experience—experience of silence—is spatial.
So, we need to ask if it is independently plausible that experience of silence is spatial? There is at least one kind of experience of silence where it is: contrast cases, where one hears silence following some sound. For example, sitting in a perfectly dark anechoic chamber, a subject might hear a bell ring and hear it to be in front and over to the left and, as the sound dies away, subsequently hear silence where the sound was heard to be. This, it might be thought, speaks in favour of the spatiality of experience of silence. The problem is, this kind of example doesn’t show unequivocally that experience of silence itself is spatial. It is possible and perfectly plausible that this spatial content is derivative from the spatiality of the contrastive positive experience of the sound.
So, if contrast cases provide minimal support at best for the spatiality of experience of silence, what about noncontrast cases, such as when subjects awaken into silence? These cases are those where, such as in the Cage examples under discussion, the problems seem to arise for the objectless account of experience of silence. This is where the bodies lie. Moreover, if such experiences turn out to have no spatial content, then this would be further reason to think that in the contrast cases the spatial content is derivative from positive auditory experiences. So, would it right to claim that in these cases the silent auditory experience is spatial? If anything, there is even less reason to think so than in the contrast cases.
The principal reason concerns the subjective discriminability of experiences of silence. Consider subject A listening to the Cage performance in concert hall 1 from the first permutation of the example. The silence she hears extends to the inner edges of the concert hall. Imagine that, unbeknownst to her, halfway through the performance noise-cancelling over-ear headphones are placed over her ears (she has recently been to the dentist and a local anaesthetic still numbs her face). Owing to the causal considerations that determine which silences are heard (recall claim [iii] above), she will then be hearing a silence that extends only to the inner edge of the earphones. Moreover, the silence she is hearing could, in principle, undergo radical deformations of shape and extensity without subject A being even in principle able to discriminate these changes. With her eyes closed, four sound-insulating walls could be interposed between her and the performers and audience without her being able to tell. These could be (silently!) moved to accommodate a fifth, then a sixth, etc. But if this is so, what sense can be made of the claim that her experience of silence is spatial? It will do no good to urge that in all these cases she hears the silence at her ears, for this is just another way of conceding that her silent auditory experience is not spatial.
A further reason for thinking that noncontrastive silent experiences cannot be spatial arises from the fact that such experiences would be phenomenally indistinguishable from the experiences of the newly deaf. As Sorensen and Phillips point out, it may be impossible for a newly deaf subject to tell on the basis of introspection alone whether she is deaf or hearing silence. However, the newly deaf person, presumably, enjoys no spatially qualified auditory experience.
If so, then neither do those who wake into silence: if their experiences were spatially qualified, then this would be a subjectively introspectable difference between being deaf and hearing silence. It follows, then, that auditory experiences of silence are aspatial.
It seems, then, that there is no convincing argument for the spatiality of experiences of silence, and the balance of credit suggests such experiences are aspatial. This prevents an explanation of the first Cage example, according to which the subjects hear distinct spatiotemporal regions. So, the difficulty for the objectless thesis of experience articulated in section 4 stands.
6. Counterfactuals and perceptual sensitivities
Phillips and Soteriou might respond to the problem of specifying which silences are heard in the Cage examples, but without appealing to causal conditions, by appealing to the idea of perceptual sensitivities. Indeed, Phillips and Soteriou each independently suggest this as an alternative to causal dependence. The thought here is that for any subject, which silent spatiotemporal region is heard depends simply on whether, had a sound occurred in that region, the subject would have heard it. So, in the first Cage example, the account might go as follows: subject 1 in hall A hears the silence in that hall, not the silence in hall B, because, had a sound occurred in hall A, subject 1 would have heard it. By the same token, subject 1 doesn’t hear the sound in hall B because it’s false that, had a sound occurred in hall B, she would have heard it. The same applies, mutatis mutandis, with respect to subject 2. The general idea, then, is which silence is heard by a subject is determined by counterfactual claims, the truth value of which is related to the perceptual sensitivities of the subject: which sounds a subject would have heard, had they occurred.
Notice this proposal implies that silences should be properly understood to be spatiotemporal regions, rather than simply temporal intervals. This is more explicitly at odds with Soteriou’s stated position that sounds are purely temporal objects than it is at odds with Phillips’s stated position, which is less committal on this point. Nonetheless, it’s compatible with the spirit of both accounts. Moreover, the suggestion has the following advantage: it enables us to account for the numerical distinctness of the sounds heard by subjects 1 and 2 in the first Cage example without making a commitment on the spatiality of auditory experience. So, at first glance this looks like a promising response.
The problem with this proposal, though, is that in a wide range of cases it simply doesn’t correctly identify which silences are heard and which are not. Consider the following case: subject C is outside concert hall 3, in which a Cage performance is occurring; C is unaware of which piece is being performed and is straining to hear. However, the hall is very well sound insulated, such that even full symphonies at their loudest cannot be heard outside. Is C hearing the silence in hall 3? It seems right to say she is not, and what seems relevant here is the effectiveness of the wall as a causal blocker: had there been a symphony being played she couldn’t have heard it, so how can it be correct to say that she hears the silence? Nonetheless, imagine, as seems possible, it is true that some sounds, had they occurred in hall 3 would have been heard by C: the sound of a bomb exploding, for instance. In that case, according to the counterfactual account, C hears the silence inside hall 3. But this is very plausibly false.
Moreover, this line of argument also shows that the proposal misidentifies which silences are heard in the first Cage example. In the initial articulation of the example, it was left indeterminate how far apart the two concert venues are. Let’s now stipulate that halls 1 and 2 are contiguous with each other, but very well sound insulated. Imagine that subject A in hall 1 is straining to hear which piece is being performed in neighboring hall 2. Does A hear the silence in hall 2? Most naturally, we should say no. However, there are possible sounds—explosions and the like—which, had they occurred in hall 2, would have been heard by A. So, according to the counterfactual account, A hears the silence in hall 2. Consequently, the counterfactual account still misidentifies which silences are heard.
Does it matter to the success of the foregoing argument that it assumes that the halls are close to each other? I suggest not. As a contingent matter of fact, the closer the halls, the larger the set of possible sounds that, had they occurred, subject A would have heard; the further apart the halls, the smaller the set. But very distant concert halls would still be such that were loud enough sounds to occur in them—such as that generated by the eruption of Krakatoa in 1883—those sounds would be heard at very great distances away. So, a range of relevant counterfactuals would still be true of these cases even though it is false that subjects in those concert halls hear the silences in the others. And when one reflects on it, there should be no surprise at the difficulties facing the counterfactual account here: clearly the reason why distance is relevant to determining which and how many sounds can be heard by a particular subject, with their specific perceptual sensitivities, has to do with causal considerations.
There is a further difficulty which tells against the proposal, though independently less compelling than the foregoing argument. Consider a scenario in which two subjects C and D are in a concert hall listening to a Cage performance. However, C is a child with acute audition but who only listens to Taylor Swift; D is the child’s hard-of-hearing grandmother, who is trying to edify the child through exposure to avant-garde music. Do C and D hear numerically the same silence? There are reasons for thinking they do, but such a claim is incompatible with the counterfactual proposal.
Firstly, how many silences are there in the concert hall? A very natural answer is: just one; the silence that extends to the inner boundaries of the concert hall’s walls. This answer seems justified by causal considerations: the reason that the silence extends there and no further has to do with the presence of a causal blocker, the wall, that prevents (or at least frustrates) sounds getting in. Moreover, wherever else there might be causal blockers—such as at the boundaries of the chairs in the hall or the skin of C and of D—we would find the boundaries of the silence (e.g., were D to sit mouth closed, the silence in her mouth would be numerically distinct from the large silence filling most of the hall).
The problem for the counterfactual proposal is that if there is only one silence in the concert hall, then C and D must hear numerically the same silence. But this is not a claim compatible with the counterfactual account because C and D have different perceptual sensitivities: different counterfactuals will be true of each. Had a fly started buzzing on the far side of the concert hall, C would have heard this, but not D. The silence C hears extends to the far side of the hall but, crucially, D cannot be hearing a silence that extends to the far side of the hall. So, C and D must be hearing numerically distinct silences: C hears a larger silence than D does.
One natural enough reply to the foregoing argument might be to claim that C and D do not hear numerically distinct silences, but instead hear one and the same silence in virtue of hearing distinct spatiotemporal parts of that silence. There is, after all, nothing unusual about the idea that two subjects can perceive the same object by perceiving different parts of it, e.g., tasting the same block of cheese by tasting different parts of it or seeing the same elephant by seeing different parts of it. A difficulty facing this objection is that in the unproblematic cases like cheese and elephants, there is empirical license for acknowledging the distinct entities in virtue of which the composite object is supposed to be perceived: the many protein molecules of the cheese, or the rump and head of the elephant. This is not so with the proposed spatiotemporal regions of the silence. A more serious problem with the proposal, though, is the following. For the objection to be effective, the proposed spatiotemporal parts must not themselves be silences. But this is incompatible with the proposal it is pressed into service to defended, according to which silences just are silent spatiotemporal regions. Consider the region supposed to be heard by D: it is silent. So, it fits the bill: it’s a silence numerically distinct from the silence of which it’s a part. But then we have the result that the present reply was introduced to avoid: there’s more than one silence in the hall… .
The foregoing discussion shows that the burden lies with the counterfactual account to convincingly argue that there is more than one silence in the concert hall. One way to do this would be to defend an alternative answer to the question: what determines how many silences there are? However, this must be done in such a way that it avoids gross ontological inflation. An obvious suggestion would be to say that the number of silent subregions of space that the hall can be divided into fixes the number of silences. But this clearly won’t avoid gross ontological inflation. If space is infinitely divisible, then there are infinitely many silences; if not, then there will still be very, very many silences.
An alternative suggestion might be to appeal to the number of possible perceptual sensitivities that subjects could have: there are as many silences as there are possible subjects who differ from each other with respect to their perceptual sensitivity to sounds. An immediate worry arises, though: for any realist about silences there will clearly be something unwelcome about making the number of silences in a space mind dependent in this way. But ignoring this worry, the view nonetheless makes no advance on avoiding the unwelcome ontological extravagance facing the previous suggestion. Clearly the number of possible perceptual sensitivities that subjects could have is far greater than one. So, we will be committed to many, many more silences than it seems right to acknowledge.
This second argument is of more limited force against the counterfactual proposal than the first argument offered in this section. It’s open to the defender of the counterfactual account to complain that the opening premise of the argument, that C and D hear the same silence, will be false if the counterfactual account is true. Consequently, the argument begs the question against the counterfactual account. This is certainly a legitimate complaint but, as I have urged, that premise enjoys independent justification from the claim that there is only one silence in the hall. The force of the argument I have just laid out depends in part on how plausibly this latter claim can be denied. I have argued that the two most obvious ways of doing this each face serious problems, but clearly this is not exhaustive of the possibilities. Nonetheless, combined with the first argument presented in this section, to the effect that the counterfactual proposal systematically misidentifies which silences are heard in a range of cases, there is a compelling case against the counterfactual proposal.
7. The thesis of negativity—redux
The arguments pressed thus far have identified a tension between two commitments incurred by Phillips and Soteriou in their development of the idea that experience of silence is “objectless.” The problems arise specifically because of a tension between denying the thesis of causal dependence and accepting the thesis of temporal regions. Clearly, one possible reply to this tension would be to retain the analysis of experience of silence proposed by Phillips and Soteriou—that it is objectless but possesses a temporal field—while accepting the thesis of causal dependence. Indeed, as Phillips’s and Soteriou’s analyses of experience of silence are spiritual successors to Sorensen’s, one could imagine Sorensen finding such a position attractive.
However, I suggest that before pursuing this strategy we should scrutinize more carefully the central claim about auditory experience of silence with which this discussion began: the thesis of negativity. In particular, it is important to distinguishing clearly between the following distinct claims about experience of silence:
(a) Experiences of silence instantiate no qualitative properties which can be modelled along dimensions of similarity alongside properties possessed by experiences of sound, unlike experiences of darkness and color.
(b) Experiences of darkness instantiate qualitative properties that are much more salient to introspection than any qualitative properties which may be possessed by experiences of silence.
The first, (a), is a metaphysical claim, which provides support for the analyses of experience of silence offered by Sorensen, Phillips, and Soteriou. The latter, (b), is an epistemic claim that provides much less compelling justification for the tilt and direction of those analyses. So, which does introspection entitle us to? It is hard to see how introspection could uncontroversially support (a) given well-established evidence of human fallibility with respect to introspecting the contents of our metal states. But there is further reason to be wary of endorsing (a).
As Sorensen (2008, 287–90) acknowledges, being in a position to hear silence—the complete absence of noise—is very hard to come by.
We are often so impressed by the contrast between a noisy buzzing café and the quiet of our noise cancelling headphones that we call the latter silence, even when we hear the sounds filtering through, or the sound of blood in our own blood vessels. Very few humans—and certainly even fewer with a sensitivity to the niceties of introspection—will have experienced silence, strictly understood, which alone is good reason to be conservative in the conclusions we draw about silence from reports of experiences of silence.
One could imagine someone objecting in the following way: look, when we subtract some ambient sound in a noisy environment, such as the clattering of mugs in the café, we may be left with some quieter sounds, such as the sound of blood in our neck arteries. These loud and quiet experiences may have qualitative similarities. But if, per impossibile, such sounds, indeed all sounds, could be removed without compromising our ability to have auditory experiences, then how would our experience be? Surely there would be no qualities for which questions about their similarity relations to the properties of sound-filled experiences could arise. This reply is unconvincing. To see this, observe that exactly analogous reasoning produces a false conclusion for visual experiences. The analogous thought experiment is: remove all nondark or nonblack objects from within the visual field. A congenitally blind philosopher who has also never encountered testimony by sighted subjects might reason that the removal of all such objects should result in a visual experience with no qualitative similarities to color experience. But this is simply false. Most subjects with normal hearing are in exactly the same epistemic position in respect of experience of silence as such congenitally blind philosophers are in respect of color experience. It seems, then, that an argument based on subtracting auditory elements can provide no good evidence for (a) or any of the proposed analyses of experiences of silence we have considered.
The upshot, then, is that the most we are entitled to believe by introspection about experience of silence is the following claim about the phenomenal contrast with experience of darkness: that blackness is salient to introspection, whereas there is no comparably salient property in experiences of silence. For all the evidence considered so far, it remains perfectly plausible that experiences of silence have qualities that can be located within a quality space, just as blackness is in the Munsell solid. That is: that the thesis of negativity, understood in a way that underwrites the arguments for the thesis of objectless experience, has not been convincingly established.