Once accept the ontological priority of the subject - or of experienced being ( like the experience of understanding a theorem ) - over unexperienced being ( like the theorem itself , or the system of which it forms part , or some physical machine like the brain in which it might be incorporated ) - and there is no dialectic by which you can wriggle out into the real physical world again .sx Is there an alternative position ?sx Well , there was one proposed during the High Structuralist period ; and I want to repropose it , even though it requires , in the sense given above , a logocentric metaphysics .sx I think we can regard mathematics as the study of the infinite set of all possible structures ; science as the investigation of which of them are real ; and subjectivity as a side-effect of objective structures .sx It was probably this objectivist vision that drove structuralism when it was at its height .sx Notoriously , it is Derrida more than anyone who is responsible for undermining this vision .sx 4 MISREADING SAUSSURE .sx The greatest philosophical achievement of Jacque Derrida is supposed to have been to show that the same ubiquitous metaphysics , the metaphysics of presence , that underlies phenomenology and existential philosophy , also underlies structuralism ; and thus to undermine the scientific pretensions of structuralism .sx For this purpose the founding text of structuralism , Saussure's Cours de Linguistique G e n e rale , was analysed and what is usually taken as its most 'scientific' aspect - the phonology - was subjected to particular scrutiny .sx A critical step here was to show the existence of a metaphysical tradition of phonocentrism , alongside of or united with logocentrism , and governing all Western thought .sx Phonocentrism means something much stronger than the commonplace observation that some people , at some time - Romantic poets , Hebrew prophets , for example - have thought that an inner voice gives contact with God ; just as others have thought that an inner light gives such contact , and others that it is to be found in sacred books , or rituals , or dances , or peculiar physical exercises .sx Phonocentrism means the universal metaphysical privileging of speech over writing as the authentic vehicle of meaning and truth ; it is supposed to dominate the entire conceptual tradition of the West , and to be present even in our use of alphabetic writing .sx Derrida convicts Saussure of phonocentrism , and attempts to show from Saussure's own text that a kind of writing - an 'arche-writing' is presupposed by both speech and writing .sx ( This kind of reversal is known as 'deconstruction' .sx ) .sx Derrida has to face two problems here - a philosophical problem in mapping the phenomenological conception of the sign on to the linguist's conception ; and a historical one .sx The philosophical problem is that while it is true that Saussure's 'sign' is a combination of sound-image and concept-about-the-world , Saussure , with his fundamental principle of the arbitrariness of the sign , explicitly denies that the meaning is inherent in , or in any sense constituted by , the sound .sx Moreover , the sign for Saussure belongs to langue ; it is a part of a pre-existing social store of sound-image/ concept combinations and it doesn't represent any intention or idea in any subjectivity until it is used in parole .sx Husserl , however , doesn't have a concept of langue - few philosophers did - and seems to think of writing and speech alike as processes in some subjectivity .sx Finally , both sound-image and concept are for Saussure determined by purely conventional systems of oppositions with other signs ; and not the same system of oppositions either ; the principle later known as double articulation is already clear enough in the text .sx There is hardly any logical space for inserting presence into this theory .sx Really , Saussure's signs are intended to do a different job from Husserl's ; something like helping us to write grammars of Modern French ; they have to be forcibly conscripted into the philosophy of the subject .sx The historical problem that Derrida has to face - and his followers often magnificently ignore - is that there is no evidence whatever that phonocentrism in this strong sense exists , or even has existed .sx What the historical record shows is that in civilizations - at least since Ancient Egypt , from which we have an eloquent document about the wonderful privileges of being a scribe - writing , that is , ordinary , empirical , worldly , everyday writing , has been privileged in every possible way over speech .sx In every possible way :sx books have always been thought of as more authoritative than speeches , even to the point of having magic powers ; literate people have had political privileges over illiterate ones , even to the point of escaping with a penance when their fellow men were hanged .sx Compared with the mass of evidence against phonocentrism , there is not much for it .sx Voice has never been systematically privileged over writing ; the case is , rather , that a few important thinkers have protested at the privileging of writing over speech .sx Derrida trawls through the whole of intellectual history and picks up such items as Plato grumbling that you can't ask questions of a book .sx ( This is a slightly unfair summary , but much less unfair than Derrida's account .sx ) .sx What makes the point even more striking is that logocentrism and phallocentrism certainly do exist .sx If logocentrism means the desire to talk consistently about , and act on consistent assumptions about , the real world - i.e. rationality - then it is surely true that the whole of Western thought has evolved under the partial control of this metaphysical category .sx And a very good thing too ; the alternative is irrationality .sx And if you believe in any version of Freudian theory , phallocentrism is a condition of being socialized ; the only alternative is to be a schizophrenic , and only Deleuze and Guattari are in favour of that .sx But there is no evidence for phonocentrism ; it has to be manufactured , by applying deconstructive arguments to Saussure , to an essay by L e vi-Strauss which says that writing is a device for political dominance , and to an essay by Rousseau , On the Origin of Languages , which takes no very decisive stand either way .sx The crucial text in which these arguments are set out is a very curious one ; it is called Of Grammatology and , as I have already said , it appears by title to be a satire on proposals for a science of semiology :sx a vast book on a pseudo-science of marks to mock the efforts of those who thought they were working on a real science of signs .sx But most of the arguments in Of Grammatology are serious enough , once one has adjusted to the phenomenological perspective Derrida takes for granted as the only possible philosophy .sx The method known as 'deconstruction' is particularly powerful within this framework .sx Phenomenological and existential accounts of the world are often heavily descriptive and metaphorical and rely on this for much of their force .sx It is characteristic of the deconstructive method to pick out ways in which the presuppositions of an argument undermine the argument ; or in which rhetorical procedures , which are essential to the presentation of some doctrine , undermine the doctrine .sx It will be seen that this is a looser version of the standard argument form called reductio ad absurdum , in which a theory is refuted by drawing self contradictory-conclusions from it .sx I personally think that deconstruction is valid only when it does entail a reductio ; and it often does , when the argument is phenomenological and metaphysical .sx But deconstruction is never very convincing when it is applied to science or engineering .sx Expositions of these are full of dead metaphors inconsistent with current theory , and nobody cares much .sx The atomic theory wasn't refuted when somebody split the atom , even though 'atom' means 'unsplittable' .sx The three texts considered in Of Grammatology are of rather different kinds .sx The Rousseau is a speculative philosophical essay comparable to the Husserl essay mentioned earlier .sx It is safe to say that Rousseau knew no more about the origin of languages than did Husserl about the origin of geometry .sx Deconstruction as a method bites very well on speculative philosophy of this kind .sx And the L e vi-Strauss is more political polemic than anthropology .sx The Saussure text , however , is a very different matter .sx It is a series of university lectures giving an elementary introduction to a science about which a great deal is already known .sx It contains an immense amount of factual detail .sx Its philosophy consists largely of a set of methodological proposals for reconstituting and developing that science - proposals which had been very successfully followed out by the time Derrida was writing :sx We have here the basic ingredients of the metaphysician's nightmare :sx that his philosophy will come to contradict the findings of an established science on some matter of fact .sx And - although he is careful to say that he is not questioning the right of a scientist , on the empirical level , to say what he needs to say - this in the end , in my view , is exactly what Derrida does .sx What Derrida is entitled to do , by his method , is to consider the categorial foundations of the proposed science of linguistics in order to establish what metaphysical commitments they involve .sx There are several of these he might look at .sx There is , for example , the definition of the sign as an arbitrary pair of signifier and signified .sx There is the sharp distinction between language , as a collection of signs , and speaking or writing as the use of them .sx There is the distinction between the synchronic study of a language as a system , and the diachronic study of its history .sx There is even the division within the synchronic study of language between the syntagmatic and the associative axes of connection .sx Each of these technical concepts has a genuine metaphysical dimension which would repay analysis .sx But rather than any of these , Derrida picked out for stress the one major distinction that is not part of the founding apparatus of linguistics as a science , being as familiar to laymen , or philosophers , as to linguists :sx the distinction between speaking and writing .sx The reason that he thought this distinction of importance presumably lies in the world-constituting functions that speech and writing , under very different and philosophically essentialist definitions , have in phenomenology .sx They have no particular philosophical significance in linguistics ; linguistics as a science , set up on the basis of the categorial distinctions above , is interested not in the putative world-creating functions of speaking or writing , and not even in the relative proximity of the signs used in them to the intentions of a putative speaker/writer - for the linguistics of langue is not concerned with intentions at all - but in the internal grammatical and other structures that speech and writing have .sx For this purpose linguists can study either speech , or writing , or both ; nothing in the categorial constitution of the science raises speech above writing in the way that langue is raised above parole .sx But linguists do have a lot to say about the relation between the two ; and one of the things they have to say is that the spoken language is the primary object of study in linguistics , writing being a merely derivative and secondary form .sx Their basis for saying this is a factual one .sx Languages have been spoken for perhaps a quarter of a million years ( that is my guess based on physical evidence - an analysis of the evolution of the vocal tract , Lieberman 1975 ) ; but writing has been around for only a few thousand .sx Most languages don't have writing systems .sx Even where they do , people always learn the spoken language first , and sometimes don't learn to write at all .sx Like everyone else who has taught linguistics , I have made these points to my first-year students .sx Saussure made them at great length and with vigour ; and every incautious word was paraded by Derrida as solid evidence for phonocentrism .sx For if one thing seemed certain to Derrida , it was that Saussure couldn't be getting that excited about a matter of fact .sx " The tone counts , " he said .sx Everyone knows people get excited only over metaphysics .sx If you put the text of Saussure back into its original context you can see a rather better reason to get excited .sx Saussure is giving a course of lectures to students of philology .sx Philologists are people who study old texts in order to describe the history of languages .sx