University of New South Wales


Stephanie Bishop

Silent reading: The read voice



This essay examines the sonic experience of silent reading. How are we to understand the peculiar nature of this sound experience that occurs in silence? The essay considers the ‘unsounded’ quality of this phenomenon and examines the contradictions implied by the notion of the read voice. This voice is the voice that is read – a voice, I argue, that occurs as the sub-vocalisation of the written word and which involves the perception of the written word’s inner sonority. Such a voice, however, is closely aligned with the inner voice. In investigating the complex dynamics of this voice that is read I draw on the work of Denise Riley, Mladen Dolar and Giorgio Agamben in order to argue that the read voice is proximate to the inner voice associated with conscience and that our experience of ‘silent’ reading marks a re-shaping of this vital force of conscience.
Keywords: silent reading, inner voice, conscience



What voice do I hear when I read silently? Clearly my perception of the sound of a written word, recited inwardly, differs from one that is spoken aloud, but how? In this essay, I consider the contradictions implied by the notion of the read voice – a voice that suggests both silence (the written word) and sonorous expression (voice). What is the nature of this voice that appears to be without external sound; what do we hear – and how do we hear – when we read in silence? In what follows I will be addressing the experience of silent reading where sub-vocalisation (perception of an inner voice) is consciously present, and acknowledge that not all readers experience inner vocalization/sonority in silent reading.

Reading was not always a silent practice. As Alberto Manguel points out, ‘not until the tenth century does this manner of reading become usual in the West’ (Manguel 1996: 43). It was Augustine, in the Confessions, who first described this new phenomenon, observing Ambrose whose eyes, ‘when he read … scanned the page and his heart explored the meaning, but his voice was silent and his tongue was still… When we came to see him, we found him reading like this in silence, for he never read aloud’ (114). Ambrose’s silent reading was contrary to the normal practice of reading aloud, which occurred even if one was reading to oneself. Reading aloud involves speaking the words on the page out loud, and in contemporary practice is commonly used for the communal benefit of reading to someone else. In Augustine’s time reading aloud was also the means of reading for private purposes. Naturally, Augustine was baffled as to why Ambrose chose to read in this way – silently, without speaking the words aloud – and wondered if ‘perhaps … he needed to spare his voice, which quite easily became hoarse’ (114).

As Manguel notes, Augustine recognized ‘that letters … were “signs of sounds”’ (45). We see the words as they are written and what we ‘hear’ occurs in response to those words the eye perceives. But for Augustine these sounds were taken to be oral, externally heard phenomena. What happens to these sounds when they are heard inwardly? They are no longer the kinds of sounds that Augustine usually associated with written letters. Instead, in the act of silent reading that Augustine witnessed, ‘unheard’ words ‘flash upon the inward ear’ (Scott 2006: 127).

When I read silently I am struck by the inward perception of a voice. Unlike Manguel’s suggestion that our experience of inner hearing that accompanies the act of reading is little more than a historical slip of the tongue – a return of the repressed archaic practices of oral reading – I would argue otherwise [1]. Silent reading is fundamentally marked by the hearing of an inner voice: by the perception of inner vocality. Examining the ambiguity of sound that characterizes the voice that we hear when we read, this essay will explore how it is that we think we hear ‘something which is not actually sounded’ (Riley 2004: 9). The difficulty in defining the nature of the read voice rests in this sonorous liminality: silent reading involves an experience of sound that does not depend on aural ‘hearing’. While it is not a sound I hear externally, produced by a mouth, neither is it a purely phantasmatic sound. In this undecidability the read voice occupies a liminal zone, a threshold between various conflicting experiences: between an inner and an outer voice, silence and sound, muteness and speech. It represents none of these polarities entirely, and yet shares in each of them.

In what follows I consider the nature of this experience from three angles. Firstly, I consider the aural nature of my reading: what do I hear and how does this inner hearing differ from my perception of speech? What, I ask, am I listening to when I claim to read a ‘voice’? Secondly, I examine the sensory nature of this experience and the origin of this sound. Finally, I consider the embodied experience of the read voice in light of Giorgio Agamben’s description of the animal voice and its role in poetic apprehension.


1 Hearing: Voice, conscience, speech

In his essay ‘On Reading’, Marcel Proust describes the peculiar sonic transgression that occurs in the movement between our hearing the read voice and the act of speaking with the physical voice. Emphasizing the paradoxical nature of the ‘soundless’ voice as it is read, and its difference from the voice with which one speaks, Proust describes the uncomfortable process of having his reading interrupted by the cook, an interruption which forces him to surface from the silence that had, until that moment, encased him. For instead of going about her tasks in silence, whilst Proust read, the cook

thought she had to say: “You’re not comfortable like that; what if I brought you a table?” And just to answer: “No, thank you,” you had to stop short and bring back from afar your voice which from within your lips was repeating noiselessly, hurriedly, all the words your eyes had read; you had to stop it, make it be heard, and, in order to say properly: “No, thank you,” give it an appearance of ordinary life, the intonation of an answer, which it had lost. (Proust 1987: 100-101, emphasis added)

How does a voice repeat words noiselessly? For Proust, the read voice remains one associated with the mouth, with lips and utterance, although it is also a submerged, inner voice that has had the capacity for external sound drowned out of it and must rise slowly, ‘back from afar’ in order to recover a semblance of its ordinary oral life which it forfeited in the act of reading. Resuming its oral mode, the muted voice that existed in the act of reading is forced into the shape of a verbal reply. But what, or how, was Proust hearing as he read? Is Proust even hearing when he reads? At first, the notion of voice as the carrier of the particular sonority associated with silent reading seems both inaccurate and insufficient: clearly writing is not vocalic in the way that I am when I speak to you, and yet when I read I am convinced that I can sense (if not exactly hear) a voice as it thinks or speaks.

One way of explaining the sonic nature of Proust’s reading would be to claim, as Clive Scott does, that when we read we hear the read voice via the ‘inward’ ear (Scott 2006: 127). The trouble with such a distinction is that all hearing occurs as something ‘inner’, whether as the infiltration of sound from the external to the internal body (sound moving into the interior chambers of the ear), the reverberations of the body’s own inner noises, or as resonances of sounds and words in the mind (Dolar 2006: 79). What differs in these instances is not so much the location of perception, but the ‘point of audition’ (Chion 1994: 89). Sounds are not received in different places (the outer or the inner ear), but are perceived to be coming from different regions. All acts of hearing, in this capacity, inevitably arise as an inner sensation. In order for the inner hearing of the read voice to be distinguished from any other act of hearing, the read voice that we hear when we read silently must therefore occur within us, and not simply be received through internal channels.

For Susan Stewart, the inwardly created voice that we hear when we read is one that we imagine. Rejecting the notion that lyric is synonymous with music, Susan Stewart argues that lyric, as ‘a practice of writing’, has ‘no sound’. Instead, what we think we are hearing when we read is only the recollected ‘sound of human speech’ (Stewart 2002: 68). Our apprehension of a poem through the hearing of the written voice thus depends, in part, upon the imagination: hearing a poem by ‘calling it to mind’ (69). Such a manner of imaginatively hearing the written voice is reliant upon our having memories of ‘speech experience’ that we draw on in the act of reading (68). In this instance the read voice is an imported voice, whose sound quality we imaginatively and reflectively impose on the text: it is dependent upon the recreation of our collective speech-memory.

We might be wary of the fact, however, that for Stewart, speech and voice are conflated categories: our speech-experience of the text is apprehended as a ‘voicing’ (69). This suggests that what we hear when we read is a quality of voice, and that the voice is the vehicle for speech. Yet Stewart’s emphasis on imaginary speech deflects our attention away from a consideration of the material origins of the read voice. We are asked to focus on the content of utterance rather than the nature of the sonic material that includes the inward reception of the text’s prosodic features. It is these prosodic features, I would argue, that make possible the recreated and imagined sound of a speaker’s voice, and that allow us to distinguish a mental hearing of vocality from the re-constituted sounds of physical speech. To recollect a sound in reading is to hear it within the mind, but this does not necessarily render it an imaginary and therefore unreal sound in the manner that Stewart suggests, nor does it render it synonymous with speech.

Stewart’s privileging of speech within the fabric of writing is liable to provoke Derridian rage, for it is precisely this devaluation of writing that he is so critical of, the independent import of writing being repressed when ‘read as speech or the surrogate of speech’ (Spivak 1976: lxx). At first glance it might seem that the argument of this essay stands in opposition to Derrida, and closer to Stewart, but this is not so.

Derrida opposed the superiority of speech and its privileged relationship to truth, rejecting the idea that writing, a form of gross ‘material inscriptions’, interferes with, or blocks, the ‘ideal transparency’ between truth and speech (Norris 1987: 67). While I am preoccupied with the ‘sounds’ perceived through the act of silent reading, these sounds are not identical to the supposed superiority of human speech so savaged by Derrida. What I want to suggest is that we might take Derrida’s theorisation of the relationship between speech/vocality and writing a step further. In contrast to Stewart’s perceptions, what I hear when I read silently is not identical to speech. Even dialogue is something that, when read, is experienced as speech that is written, and thus as speech that is subsumed and coloured by the inward and polymorphous vocality that ultimately communicates the written dialogue to me – a vocality conditioned by the graphic marks of the writing along with its formal and musical tectonics. In the process of listening to the ‘sounds’ that are produced through silent reading, writing takes centre stage and it is the writing that actively creates, and shapes a particular kind of sonic experience, one that is related to our experience of voice and speech, but not identical. Writing, in this context, is neither a supplement, nor a subordinate (98). Writing is not, as Derrida says of Rosseau’s literary ambitions, an attempt to ‘recapture speech’ (Derrida 1992: 79).

Rather, a passage read silently generates a sonic-kinetic response – a certain kind of inner voice – yoked to the text but which grows within and on behalf of the reader. It could be argued that writing gives us access to voice whilst sidestepping the intermediary of speech, thus trumping speech as the carrier of voiced truth. In silent reading the phonic substance of writing is no longer disguised, but perceived internally (Derrida 1976: 7) [2]. Silent reading is thus marked by the experience of my reading a voice, but it is not the same as ‘hearing (understanding)-oneself-speak’ (7). But how is this so, and how might this read voice be differentiated from my perception of speech?

The difference between our perception of voice and speech is most troubling when considered in relation to the notion of conscience. As Mladen Dolar argues, in the ethical tradition of the voice it is our listening to an inner voice speaking that becomes the hallmark of human consciousness (Dolar 2006: 83-86). Voice and speech are entangled in our experience of consciousness. And yet the voice is different from speech. While speech presupposes the existence of a voice, it is the primary experience of a voice that signals the presence of consciousness.

Denise Riley, in Inner Speech’, suggests similarly, arguing that the perception of an inner voice is distinguished from the notion of inner speech precisely due to the voice’s historical alignment with the experience of conscience. While inner speech suggests the unfiltered ‘talk’ that carries on incessantly inside our heads, the ‘inner voice’ is traditionally represented as a ‘materialisation of conscience’ (Riley 2004: 15). This ‘inner voice’ is what speaks to me – it represents the identity of my conscience, not simply the words that are said. For both Dolar and Riley, the notion of an inner voice masquerading as our conscience is entangled with the experience of speech: I recognize my inner voice by its sonic force, while my inner speech remains the linguistic content that is carried by the current of that voice. As the voice of conscience, or as the voice that is my conscience, an authoritative distance is implied despite the sense that this separated voice seems to speak from inside me (Riley 2004: 15). The inner voice thus represents two things. First, it is the sonority through which we hear our inner speech. Secondly, as my conscience that speaks to me, my inner voice is an aspect of my more generic experience of inner speech. The speaking of my inner voice, moreover, is distinguished from my more mundane and ongoing inner speech by its forceful sound: our conscience (as a voice) is an authority speaking to us and it is the sonic quality of this voice that confers an identity upon my conscience that speaks. Nonetheless, a blurring of categories remains, and it is one that is historically grounded. Dolar argues that the tradition of voice as conscience was initiated by the Socratic limiting of communicated thought to speech. In committing ‘nothing to writing’, the Socratic theme of the voice became synonymous with the act of speech (Dolar 2006: 85). This conflation of terms remained throughout the voice’s later ethical manifestations, as the voice of conscience, the divine voice, or the voice of reason – in each instance the voice remaining associated with the form of an inner speaking.

Dolar’s own investigation of this Socratic theme of voice rests upon a shaky foundation for the very reason that, as with Susan Stewart’s analysis, voice and speech remain largely interchangeable terms. Precisely where the difference lies between voice and speech is not always clear, and one could often be taken for the other. Yet should we attend to the difference between the two it would seem that in contrast to the sonority of the voice, speech emphasises the content or message carried by language. Despite the possibility of ‘writing down’ speech – transcribing the spoken word or conversation – speech is invariably concerned with the act and content of ‘talk’. It may well be the case that it is the voice that speaks – that the voice is in some ways the medium of speech. In this instance the voice remains the sonorous matter of the utterance and while this sonority contributes to the meaning of the utterance, it is distinct from the linguistic content of the words spoken.


2 Sound: Inner voice – read voice

Giorgio Agamben describes the voice in somewhat different terms. Firstly, it is the sound quality that is emitted by the vocalic organs. Secondly, it is a crucial identifying feature of a species or individual: the unique sound that is emitted by the voice – I might not see you, but I know who you are by sound [3]. The voice is a means of recognition: it is the personified sound that we can recognize despite the absence of a body. And while the voice is tied to a particular physical self – generated by a physical self and associated with such a body, it can travel where the body cannot, and can persist when the body is absent.

Agamben’s description of the voice as an identifying phenomenon corroborates with Riley’s suggestion that we recognize our conscience by its vocalic quality; by the force of its sound. This sound experience of my inner voice is closely linked to the voice I perceive when I read. Both are ‘heard’ in a similar inner region and with a similar force. Moreover, when I read I do not feel myself to be channelling a foreign voice. Nor am I inhabited by a phantasmal voice associated with the text. Reading is not an act of spirit possession or aural hallucination. In contrast to my experience of silent reading, the pathological hearing of voices is distinguished by the fact that such voices ‘operate in an authorial mode of address, running a continual commentary on the person’s own behaviour and conduct; insulting, judging, commanding or directly addressing’ (Blackman 2001: 21). Unlike this hallucinated voice, the voice I hear when I read lacks the capacity to make a judgment upon me, and is not talking to me about myself. Instead, I willingly participate in its existence, and often desire to prolong it, to listen to it attentively and with pleasure. This is a very different voice-hearing from a genuine hallucination, defined by Blackman as ‘systematized, all-powerful, all-pervading ‘events’ which engulf a person’s general cognitive capacities’ (22). Blackman goes on to argue, however, that there are certain kinds of voice-hearing that are not pathological, and may well be a ‘normal variant of behaviour’ (189). But even these normal, or normalized voices remain foreign voices, voices that the hearer learns to accommodate and accept: the voices become characters that routinely inhabit and interrupt the hearer’s psychic life. Blackman argues that voice-hearers can foster a sense of amicability with the invading voice, yet the interlocutory nature of the hallucinated voice allows it to retain the capacity to intercept the individual’s inner speech. The heard voice, whether a hallucination proper, or a non-pathological and befriended voice, retains an agency that the read voice lacks.

Instead, the kind of hearing that is associated with silent reading has more in common with the musical experience of audiation. Different from the notion of audition – which is the simple perception of sound – audiation defines a musical thinking, a thinking in music. Audiation refers to a cognitive process whereby sounds are attributed meaning. It involves the mental hearing of sound, even in the absence of external sound and implies the retention of sound in the mind – a musical inner hearing – that does not correlate to, or depend on, the presence of something being sounded. Reading a musical score, for instance, and inwardly hearing the notes, constitutes an act of audiation. While a text does not present itself as a score for performance in the same way, there is a correlation in that the prosodic features of the writing shape how that writing is inwardly heard. One audiates, rather than hallucinates, in the act of silent reading. Extending this notion of audiation to silent reading requires us take up and modify Roland Barthes’ concerns in ‘The Grain of the Voice’. Barthes’ question, ‘How does language manage when it has to interpret music?’, is now aimed at something both more specific and more amorphous: how does language manage when it has to interpret the music of sub-vocalisation as it occurs in silent reading, when it has to interpret the mental hearing of sound? (Barthes 1978: 179). In the case of silent reading one is listening to the voice that seems to sing in the head as one reads. We are thus shifting Barthes’ ‘musical object’, further ‘displac[ing] the fringe of contact between music and language’ (181). The musical object is the voice heard in the head when one reads silently. At issue is the encounter between language and the music of this internal voice. When one audiates in the act of silent reading what one hears is the grain of an internal voice, the friction between the music of words and language (185). Just as we hear the grain of the singing voice – Barthes’ Russian bass – so we also here the pitch, tone and sonority of the read voice in our heads.

How I hear my inner voice, and what I hear when I read silently are, in this way, proximate experiences. It is not another voice that I hear when I read, but my own inner voice heard back to me. My inner voice is subject to the architecture of the author’s voice: the voice that I hear when I read is my inner voice re-shaped by the features of the text [4]. The grain of the voice, Barthes’ argues, is the material residue of ‘the body in the voice as it sings’ (1978: 188). In the case of silent reading the material residue is triple: it is the body of the text - the physical features that shape and infect the reading voice - and the particular sonority of my own inner voice that is in part a reside of the sonority that I associate with my own external speaking voice. The written words ‘sculpt’ my inward hearing of myself [5]. It is not a fictional voice that I hear when I read but a new and adapted experience of my own inner voice. This inner voice is the medium of my consciousness. It is also the medium of my silent reading. Thus, when I read I am hearing the sounds through which I normally identify myself – the sounds through which I possess self-consciousness – suspended and re-formed.

In this way the vocalic quality that normally constitutes my conscience is temporarily reconfigured by the prosodic peculiarities of the text. What I might think of as the writer’s voice is my own inner voice shaped by the writer’s style. The writer’s voice is my inner voice transformed by reading writing. This neurological entanglement of our modes of inner speaking with the activity of silent reading is compellingly revealed in a particular instance of frontotemporal dementia. In this case, the first signs of the onset of dementia were marked by a loss of silent reading: the patient began to ‘read aloud at bedtime… And it was difficult for him to stop. Moreover, he often spoke aloud to himself during the day or suddenly began singing, all features which were very unusual for him’ (Vercueil & Klinger 2001: 705). The loss of the ability to read silently was caused by the ‘inability to inhibit the vocalisation of inner speech’ (705). The case demonstrates that the read voice is a variant of the inner voice: the difference being that silent reading is an ‘externally triggered inner speech, as opposed to an internal production of inner speech linked to the thinking process’ (705). This non-vocalized inner voice that I hear when I read silently is my own, but it is my own inner voice provoked and mediated by an external text.

Such findings prompt us to ask what is silent in the experience of private, inner reading. Perhaps it is an experience that is precisely this: inner reading, but not explicitly silent. Yet there is clearly an aspect of this reading practice that happens without external sound, and in the absence of clearly definable inner sound. The inner sounds I hear when I read, for example, are not the real sounds of my heart beating or the reverberations of my breath that echo within me when I press my fingers to my ears. When Augustine looked upon Ambrose as he read, he heard none of the familiar murmurs that he associated with the act of reading. But for the reader, this silence is far from an ‘acoustic blank’ and is instead coloured by ‘specific resonances’. What is silent when we read in this inner way is the ‘reading organism’: while we read in silence, silence is not the hallmark of what we experience when we read (Scott 2006: 121, 123).

Indeed, this activation of a sub-vocal ‘inner voice’ through attentive silent reading is supported by recent neuroscience research. ‘Basic introspection’ one study claims, ‘suggests that adults continue to pronounce written text covertly using auditory verbal imagery’ (Perrone-Bertolotti et al 2012: 17554). Auditory verbal imagery (AVI) occurs when we inwardly talk to ourselves. Uncontrollable AVI is a symptom of mental illness such as schizophrenia. This study sought to test the neural basis of AVI by taking ‘intracranial recordings from four neurosurgical patients with intractable epilepsy’, the aim being to measuring the level of neural activity when reading silently and when hearing a story read aloud (17555).

The results of the research show that the auditory areas of the brain are highly responsive to written language that is perceived silently.  Silent reading, they claim, generates AVI. Recent fMRI’s also show that silent reading produces high activity in the auditory areas of the brain, proof of the presence of an ‘inner voice’ heard whilst reading. Reading, the study shows, ‘spontaneously elicits auditory processing in the absence of any auditory stimulation … [c]ontradict[ing] a recent claim that readers are unlikely to experience an auditory image of a written text, unless the text corresponds to sentences spoken by people whose voice is familiar’. An ‘inner voice’ is produced they argue, ‘even when reading narrative with no identified speaker.’ The inner voice that is consciously experienced when reading is intensified by attention, and is more present, they claim, when readers are silently reading a difficult text (17560).

Another recent study examines the neural activity in normative silent reading and shows that voice-sensitive areas of the brain are activated when silently reading a ‘quotation … from a fictional person’. Further, the quality of the voice activated by the fictional person is described as a general ‘inner voice’ because it belongs to an imagined entity ‘rather than [belonging to] the voice of a specific individual’ (Petkov & Belin 2013: 155). 

An inner voice experienced when reading silently is thus assumed. Is it possible then, in the act of reading, to isolate the particular sonic experience of this inner reading voice? The subjection of the sonorous inner voice to the architecture of language suggests that an intertwining of sound and sense constitutes the experience of silent, inner reading. One might therefore be wary of Giorgio Agamben’s interpretation of how we hear and cognize a poem, given that Agamben suggests that our cognition of a poem is marked by a linear progression from sound to sense, isolating what we might call the ‘bare voice’ from meaningful language.

Agamben’s linearity stems from the distinction he makes between the human voice and the voice of the animal. The voice is regarded as ‘pure natural sound’. Pre-grammatical and unarticulated, the voice, when it exists as pure sound, represents the ‘“confused voice” of animals’ (Agamben 1999: 68). The human voice, Agamben claims, differs from this animal voice in that it ‘can be written … and articulated’ (68). Of greater interest to Agamben, however, are the voices which fall between these two categories. They are what Agamben terms the inarticulate voices – the voices whose sense we can’t determine – but which can nonetheless be written. Like the animal voice, we cannot understand their meaning, yet like the human voice, their sounds can be transcribed. This voice category is constituted by the human imitation of animals, ‘these brekekeks and koi’ that are inarticulate in the sense that their meaning is hidden from us, but which are capable of being articulated in the form of writing, despite their ‘sound’ remaining indecipherable (68-69).

Yet when this confused animal voice enters into writing, the voice of the animal can no longer belong to pure sound. Consequently, when we read the voice that has been written we no longer hear it as an animal voice, but as an animal voice ‘separated from nature’: it becomes a voice that ‘shows itself in letters as pure intention to signify whose signification is unknown’ (69). It becomes a read voice. What shifts in our understanding of ‘voice’ when it is cut off from its animal sound-source? For Agamben, imitated animal sound, as it is transcribed, attempts to capture the pure sonorous voice before it has ‘become signifying discourse’, and yet in this process the pure sound, now moulded into writing, no longer exists (69). In Agamben’s view, the read voice represents a category of sound in which the animal voice presupposes writing, but which is also subsumed and negated in the act of writing. Marking the removal of the sonorous, oral voice, the voice we hear when we read depends on the negation of the vocal utterance. What is heard in silent reading thus resembles a poor copy, or faded imitation of transient live sound.

Agamben’s analysis must be considered with caution for the reason that it assumes the cognition involved in reading to be marked by an ordered sequentiality, following an evolutionary line through which animal sound culminates in meaningful human sense. In Agamben’s theory, sound does not itself possess cognitive value. Certainly, Agamben seems to suggest, we experience sound in the act of reading or hearing poetry, but it confuses us. And so we look beyond it towards the articulations of a conceptual language that will explain these troubling animal noises.

Aligned with the experience of onomatopoeia and glossolalia, Agamben defines poetry in relation to a sound intending signification (1999: 71). The language of poetry is marked by this difficulty of the negative: as an ‘experience of the letter’ the voice of the poem intends to signify, being more than animal sound, and yet conveys a set of sounds whose sense is not yet known. The vocality that we hear in the enunciation of a poem is thus something of a stalled phenomenon, existing for Agamben as a negative event in two ways: located in the death of the animal’s pure sound, and in the negation of meaningful speech (1999: 71).

Due to this process of the animal’s voice-removal, Agamben associates the voice that is enunciated through reading with a voice infused by negativity, linking the peculiar sonority of poetry with the ‘idea of a “dead language”’ (Agamben 1991: 33). Like the animal sound that the human imitates in writing, the dead letter is defined by a sound not formed into a speech I understand. As an ‘experience of an unknown word’, my encounter with a dead word is composed of a suspension between sound and sense: on the way from one to the other, I grasp the initial sonority of the word, but without yet knowing the word’s meaning (33). Uncomprehended, the dead word incites my ‘desire to know’ (Agamben 1999: 64). In my perception of such a word I become suspended in ‘the no-man’s land between sound and signification’ (64). No longer immersed in the world of decipherable language, I find that my thought ‘dwells in “the voice alone”’: the vocality of the read voice becomes the inarticulate and negated sound of the animal (65).

But the logic of my cognition suggests that when I read I do not experience a text in such a linear fashion. I do not move from the confusion of sound to the logic of sense. I do not make sense of what I read through a series of temporally successive stages, hearing the text before I know its significance. Instead, when I encounter a poem I do so through a series of layers that can be touched upon simultaneously. Sound and sense reflect and merge with one another, and do not strike me in any set order. My cognition of a poem is spatialized in a way that Agamben does not account for. In contrast, Agamben’s temporal ordering of our linguistic cognition echoes Kant’s argument in the Transcendental Aesthetic that our inner experience (of which inner reading may be a part) is governed by the condition of time and hence, for Kant, by successivity. The notion of silent reading as a potentially spatial perception of language is readily aligned with the experience of language encountered through use of Sign, where the ‘linguistic use of space’ means that ‘what occurs linearly, sequentially, temporally in speech, becomes simultaneous, concurrent, multiveiled in Sign’ (Sacks 1989: 87).


3 The voice without a body

Against Agamben’s suggestions, the voice that I hear when I read is not a poor substitute for an original oral performance. Because the read voice is not yoked to its writer’s vocalic organs in the way that the animal voice is sounded by a guttural cry, what I hear when I read cannot be denigrated as a mere after-effect of speech. This difference between the oral-animal voice and the silent inner voice of reading can be all too evident in a writer’s oral reading of what they have written. What we hear when the poet reads out their work can be vastly different to what we hear when we read the same work to ourselves. The discrepancy that may result between the public performance of a work and its private inner recitation is complex. Often at the mercy of the occasion, the writer’s oral rendition of what they have written can be affected by a variety of environmental factors – the acoustics of the room, for instance, or the attentiveness of the audience. To grant that the poet may not be the best reader of their own work, and to recognize the difference between what we hear in performance and what we hear to ourselves, is to appreciate the peculiarities of the read voice: one that is not equivalent to the speaking, physical voice and that cannot be readily externalized without causing an alteration in its sound qualities.

Because of this difference, the sound of a poem, once it has been read out, can alter my inner perception of the same text. No longer do I ‘hear’ the writing as I once did. The sonorous quality of the physical voice can come to plague my reading like an irritating jingle that has entered me without permission and which I cannot expel [6]. What is at issue is nothing so individuated as the writer’s own vocal possessions and possessiveness, but rather the resonances of a particular work, freed from an authorial body. The read voice is the voice of the writing, the voice of the poem or the novel for instance, and it is for this reason that it can ‘sound’ incommensurate with, and independent from, the vital life of its writer. The read voice is not subservient to oral performance and the perception of an oral voice. Unlike the argument posited by Francis Berry that what we hear when we read a poem silently is the poet’s own voice – and that the sounds of poetry are thus governed by the properties of the physical body – the body that is suggested by the written voice is not represented by the hovering presence of a living authorial body presiding over the words on the page, but exists as a missing body, one that has passed us by, or that lingers back stage, a phantom presence (Berry 1962: 3-4, 8).

This paradoxical status of the read voice as an identifying voice that nonetheless lacks a body, shares something of the uncanniness that Michel Chion ascribes to acousmatic sound, a sound “‘that is heard without its cause or source being seen’” (Chion 1994: 18). For Chion, this acousmatic sound phenomenon suggests a listening practice inherent to technology – the telephone or the stereo for example – in relation to which we can listen to a sound that is detached from its origin (18). It reflects a mode of listening that is also evident in the practice of reading, whereby the voice that is written is rendered acousmatic – unable to associate it with a face or a source, its displaced ‘sound’ haunts the page as a ‘kind of … shadow’ (21).

Separated from a body, the acousmatic voice is free to roam across alternate stretches of time, dwelling in the past tense or eliciting temporal suspension. An intimate voice because of this absence of physical boundaries, the ‘disembodied’ acousmatic voice can inhabit the body of another. We identify with such a voice partly due to the way in which it exists by taking up residence inside us. It shares the uncanny quality of the voice of the dead, or of the almost dead: a voiceover that continues to speak despite the absence of a body and which wanders over the surface of the mind, just as the filmic acousmetre wanders ‘along the surface’ of the screen (23). It is a voice that speaks ‘off-screen’ due to the fact that the ‘place of vocal production … isn’t visible’ (127).

As Dolar argues, the nature of the acousmatic voice, in which sound is rendered errant by being displaced from its origin, marks the voice as inherently ventriloquistic. When Dolar refers to the acousmatic voice, however, he does so in light of the physical voice, not the filmic voice. In Dolar’s reading, the physical voice is acousmatic because of the way in which it emerges from a hidden place within the body – the organs that produce vocal sound not being visible (Dolar 2006: 69). A similar principle applies to our perception of the read voice in which the source of the voice is obscured, even while we feel it take up a sonorous residence within us. My experience of the read voice implies a process by which the voice that is inside my body, and that I know to be mine, allows itself to be shaped by the vocalic patterns suggested by the writing which I read. I donate the inner voice of my body and receive the organs of the text. I permit such organs to function inside me, as my own, and in this way the words and I become a symbiotic organism. I become complicit in the process of my being ventriloquized: the words are animated by my participation, and my inner voice is thus enlivened by a strange heartbeat. Reading becomes an act of ventriloquy: I am not so much ‘spoken from elsewhere’ in the sense of an automated possession, but instead actively cultivate the experience of a voice that is not naturally my own, adapting my inner voice to the patterns of the text (Riley 2004: 20).



[2] Above all, and quite unlike the speech that rouses Derrida’s suspicion, the ‘voice’ we experience when we read has the ability to convey forms of graphic inscription; to inhabit marks of punctuation, to feel the alterations of line and layout. This element of the read voice is expanded on in the latter part of the essay. return to text

[3] ‘Is there a human voice’, Agamben asks, ‘a voice that is the voice of man as the chirp is the voice if the cricket or the bray is the voice of the donkey?’ (Agamben 1993: 3). Just as we identify others by the sound of their voice, I identify myself as the self I suppose myself to be by way of this inner acoustic presence. return to text

[4] As Clive Scott argues, Kristeva’s notion of the semiotics of the text can be taken as that which re-shapes this voice (2006) (see Kristeva 1984: 69). return to text

[5] Clive Scott describes this phenomenon in relation to an oral reading of Shakespeare: ‘Shakespeare’s words sculpt this voice… Shakespeare becomes this voice’s temperament at the time’ (2006: 65). return to text

[6] Susan Stewart in Poetry and the Fate of the Senses also argues for the incommensurability of the poet reading with the poem written (2002: 104). return to text


Works cited

Agamben, G 1999 The End of the Poem: Studies in Poetics, trans D Heller-Roazen, Stanford University Press, Stanford, California return to text



Stephanie Bishop is a lecturer in Creative Writing at the University of New South Wales. Her first novel was The Singing.


Return to Contents Page
Return to Home Page

Vol 17 No 1 April 2013
General Editor: Nigel Krauth. Editors: Kevin Brophy & Enza Gandolfo