Jonathan Rée (1999) _I See A Voice: language, Deafness & the Senses - A Philosophical History_, London: HarperCollins.

[For notes from Rée on aesthetics, see 'The Aesthetics of Generative Code']

In the beginning was the 'voice'. Jonathan Rée's _I See A Voice_ begins with some fundamental observations on the intimacy of the voice (1999). Shyness about speaking in public demonstrates how the voice is linked to self-consciousness and privacy. Yet the voice is 'destined for other people: you speak, primarily, to be heard' (1999: 1). Hence it is inextricably linked to civil rights - to have a voice like a vote in civil society - voices and politics go together. On a more metaphysical level, the voice, not least because of its connection to breath, seems to demonstrate the very stuff of living being: a messenger of the soul and the breath of life. In summary, Rée's book attempts to undermine a series of false presumptions on the voice, what he calls 'shadowy metaphysical prejudices': including that the voice is connected to the soul or inward subjectivity; and that language takes two distinct forms - audible speech that occupies time not space, and visible writing that occupies space but not time (1999: 6). The voice occupies the dual territory of inward subjectivity and outward sociality, and despite all attempts to understand its vagaries is ultimately characterised by its indeterminacy.  

The idea that speech is a precondition of consciousness pervades thinking. It was Johann Gottfried Herder writing at the end of the eighteenth century that helped establish that voice and speech were distinct. Whereas the voice is simply animal exuberance to Herder, speech happens when the voice takes on the inflection of the human institution of language, in turn marking a shift from unreason to reason (Rée 1999: 67). Thus the human voice 'speaks' - and is understood as the source of human society through communication - traveling from the inner _natural_ realm of expression and sound to the outer world of _artifice_ and signification. (Note: Such logic informs the historical maltreatment of those that cannot speak (the 'deaf and dumb') as animals outside the human world of normal language.)

Babble: In the Bible, and following the word of God, the world contained one language - the single language of Adam who first named objects in the world. The Tower of Babel, designed to reach into heaven, displeased God such that 'he' decided to confound the language so that people would not understand each other's speech. According to _Genesis_ (2:19 & 11:1-9), 'therefore is the name of it called Babel; because the Lord did thee confound the language of all the earth' (in Rée 1999: 75-6). Subsequently everyone is left to babble, in a diversity of languages and confusion of tongues. In this way, the original sounds are also associated with God-like creative virtue. The soul is the source of speech perverted through metaphors evident in spoken language. 

The relationship between speech and writing remains unstable. Historically writing taken to be a picture of speaking, with many fascinating attempts to capture this and move beyond the limitations of the Latin alphabet. For instance (in 1775), Jonathan Steele proposed using five line musical notation for speech rather than standard alphabetic script. Phonography, developed in the early 19th century was another technique, where each language sound was given a shorthand mark, Isaac Pitman even claiming (in 1854) his phonetic shorthand to be an 'exact picture of speech itself'. In parallel, others concentrated on the physiology of speech such as Alexander Melville Bell (in 1849) describing the 'actual movements of the organs of speech' - calling this 'visible speech' (in Rée 1999: 255, 258). Visible speech was based on a universal notation system able to reproduce every dialect and language. Its symbols were not alphabetical but 'physiological, and therefore musical and arithmetical too' (Rée 1999: 258). It was algorithmic.

Amongst an even earlier history of speech synthesis and the many attempts to reproduce human speech by machines in the 18th century, Wolfgang Von Kempelen's 'machine parlante' was the first that allowed to produce not only some speech sounds, but also whole words and short sentences. In his book _Mechanismus der menschlichen Sprache nebst Beschreibung einer sprechenden Maschine_ (1791) he described the speaking machine so that others might reconstruct and improve it. It comprised of bellows to pump air into the voice box, where a reed controlled the release of air alongside movable parts corresponding to lips, palate, tongue and nostrils. (http://en.wikipedia.org/wiki/Wolfgang_von_Kempelen%27s_Speaking_Machine)
Note: Given the use of his chess-playing machine by Benjamin, it is interesting to speculate on the role of artificial speech in terms of its relation to speech as the source of human society and human history. There would be an interesting project of media archaeology to be undertaken here. It is also worth noting that the analysis of speech and associated facial movements informed the development of cinema through what was called 'speech photography' or 'photography of the voice' (Rée 1999: 253-4). 

But such examples also served to emphasise the inadequacies of the latin alphabet to capture dialects and phonetic diversity: making 'the art of recording speech almost impossible' according to George Bernard Shaw; later to write _Pygmalion_ that privileges phonetics (Rée 1999: 262). By the twentieth century, there was a general recognition that speech was not a continuous stream of sound but a series of discreet abstract sound-types, a phoneme: 'not an actual noise in the mouth or ear; rather it was its "mental equivalent" in the mind of a speaker or listener' (Rée 1999: 266; referring to the work of Baudouin de Courtenay in 1894). As part of this, speech could not be considered simply a natural and physical phenomena but was thoroughly cultural. It is these principles that Ferdinand de Saussure developed in his lectures, later published as _Course in General Linguistics_ (1916) to establish that a language is a 'system' based on differences between elements. The many phonetic systems that went before (including the few mentioned thus far) simply served to demonstrate their limitations. Writing systems are simply 'theoretical conjectures about the structures of languages they apply to, rather than mechanical records of their audible features' (Rée 1999: 269). 

To de Saussure, 'every element of language was "auditory in character" and therefore "unfolds in time alone"' (Rée 1999: 320) But it is indeed possible to pronounce two things at the same time according to Roman Jakobson, who describes 'simultaneous bundles of distinctive features' (in Rée 1999: 321). To Jacques Derrida, the mistake stems from a problem in the conceptualisation of the relation between speech and writing, based on a false assumption that vocality and time lie in contrast to writing and space, and in turn that writing corrupts the purity of language as 'a kind of pollution and above all a sin' - the 'phoncentrism' of Western culture (in Rée 1999: 322; however, Rée argues there is no real evidence for the claim against de Saussure). Together, and in general terms, the principle is set that languages are systematic but nondeterministic. Meanings are produced through the relations amongst the signs and is hence a situation full of misunderstandings. 

All the same, the reader still yearns for the authentic and legitimising voice behind the text, even if, or perhaps especially because it doesn't exist. In oral story-telling, Rée explains that the teller uses many voices and character whether telling a traditional story or simply in casual chatter (1999: 369). Once written the fullness is somewhat lost but for rudimentary techniques like punctuation and quotations. Quoting is particularly interesting in this connection, what became the common practice of using 'inverted commas', distinguishing the voice of the character in the text from that of the writer (or in code commenting out is used to distinguish the voice of the program from the voice of the programmer). The practice of attribution in this sense runs in parallel to the 'enclosure of common lands', and the idea that writers might try to 'reclaim some of the freedom of the oral teller'; 'sometimes theorized as "Free Indirect Speech"' (Rée 1999: 371). But in a general sense all writing is an invitation for speech - a script to be executed as vocal performance. Whether it is free is another problem of course.