<< With Apologies, a Little More Foucault | Front Page | The Specter of the Name (For Jacques Derrida) >>
Thursday, April 05, 2007
Sign of a Sign?
Chapter 2 of Derrida’s Of Grammatology opens with an epigraph from Rousseau—a “fragment inédit d’un essai sur les langues.” And one hopes, had he gotten around to editing, he would have noticed it does not make a great deal of sense: “Writing is nothing but the representation of speech; it is bizarre that one gives more care to the determining of the image than to the object.” There is a kernel of an obvious right idea here; namely, phonetic alphabets piggyback on phonetic stuff—on sound. There is a (tolerably) clear sense in which, in typical natural language cases, written language is parasitic on the functions and mechanisms of spoken language. A phonetically-based writing system sets up correspondences between shapes and sounds, by way of conveniently plugging into a pre-existent syntax and semantics. But it hardly follows that writing is nothing but representation of speech. If that were true, the only thing we could write about would be sounds. But obviously we can write about all sorts of things—dogs and cats, numbers, the sun. The list goes on.
Rousseau has conflated ‘written language functions by x’ with ‘written language functions to x’. To say that ‘writing is nothing but the representation of speech’ is as silly as saying ‘speech is nothing but the representation of recursive rules of syntax, etc.’ In both cases, you mistakenly take a mechanism by which language manages to represent for the represented object.
In Rousseau’s defense, this seems to be an error other major thinkers have also committed. Derrida quotes Saussure: “Language and writing are two distinct systems of signs; the second exists for the sole purpose of representing the first.” He alleges that Aristotle and Hegel say the same. I’ll take his word for it. Certainly Augustine says something of the sort. Writing represents speech. Speech represents ideas. Ideas represent Ideas in the Mind of God. Something like that. Great chain of meaning; writing is low man on the totem.
Saussure’s error is the same as Rousseau’s. Just to be clear: suppose we have, on the one hand, a congenitally deaf person who has learned to read written English with perfect understanding; on the other hand, let there be a non-English speaker who has been trained to read out English writing (perhaps for the benefit of a blind person), who doesn’t understand a word of what he is pronouncing (see those passages on ‘reading’ in Wittgenstein’s Philosophical Investigations.) Now, in a perfectly good—but quite ambiguous sense—both of these individuals can ‘read’ English just fine. But if we now ask which of the two reads with understanding of what the written words represent, I think we would be strongly inclined to say: the deaf person. He understands what the language he is reading means. Saussure and Rousseau are (quite absurdly) committed to maintaining the opposite: namely, that the non-English speaker understands the meaning of the English words he is pronouncing, namely the sounds he produces. That is, Rousseau and Saussure are committed to maintaining that a person who, by hypothesis, doesn’t understand a word of English, understands English. That won’t fly.
Actually, there are some interesting complications to explore. Let me canvas a couple, without really digging in as deeply as one might. No one would ever say that spoken language exists only to represent the rules of syntax. Here the conflation of the way language functions, qua representational system, with what it functions to represent, is too obvious to miss. But you might say the following (I take it Saussure and Rousseau and Augustine and others would): speech serves to represent concepts/ideas. So when I say ‘the cat is on the mat’, I am representing, not the cat on the mat, but my idea of the cat on the mat. An alternative view would be: what is represented is the cat on the mat, not my idea. Because it’s not as though I’m only talking about my ideas. I’m taking about the world. This is not an easy case to sort out (to say the least), but one potentially sound objection to the concept view is: just because language functions by means of concepts doesn’t necessarily imply that it exists to represent concepts – as opposed to cats. Complicated.
Second point: we who can both speak and write English find it a bit strange to think of someone reading with no sense whatsoever about the regular correspondences between letter sequences and sounds. We are inclined to say there are some elements of what is being represented by the writing that the deaf reader is missing: namely, the sounds. This is not without practical implications. He isn’t likely to appreciate written poetry, for instance, even if he can be made to appreciate, intellectually, that there is a thing called ‘rhyming’, which ‘make’ and ‘bake’ and ‘sake’ and ‘lake’ do. The orthographic echo is sure to give him at least a hint of what it is that the sound does for people. But it won’t be enough. It’s funny. We may stand to him in very much the relationship that synaesthetics, who associate sounds or shapes with colors, stand to us. A synaesthetic, who finds certain words to be more ‘yellow’ than others, probably can’t help but feel that ordinary speakers are missing something, in not appreciating this association. All the same, it really has nothing to do with the meaning of English that a certain words seem yellow to someone (maybe even to several someones. You could have a community of synaesthetics.) Likewise, it really has nothing to do with the meaning of the written word ‘dog’ that it is associated with a certain sound. The word doesn’t mean any sort of sound. It means a kind of animal.
Take a case like Chinese: a speaker of Mandarin and a speaker of Hokkien can read the same newspaper. Each associates the characters with different sounds. But they don’t disagree about what the characters mean. Multiply examples: Latin, as pronounced at different stages of its career as a language. Any language where sound has drifted, while orthography has remained fixed. Or take the case of homonyms. A single sound corresponding to two different spellings, corresponding to two different words. I doubt we would want to say that it turns out, in these cases, that speech exists just in order to represent written language—that the true face of the word was written. The truth is: a word is neither essentially spoken or written. It’s more abstract than that. So Saussure is just confused: “the linguistic object is not defined by the combination of the written word and the spoken word: the spoken form alone constitutes the object.” No. It clearly doesn’t.
Part of the temptation to say that written words represent sounds is that writing is, historically, and also for every ordinary individual, developmental case, preceded by speech. You learn to ‘sound out’ written words. That is, it’s a kind of code for making sounds. When you are at the learning stage, it is more plausible to say that writing represents sounds. Later, of course, you learn to read so fast that you no longer even process words left-to-right, let alone sound them out. You just take in the gestalt of the whole word, even the whole sentence. That is, you become functionally indistinguishable from the deaf person who never associated the written shapes with sounds to begin with. The relationship between the shape and the system of syntax and semantic no longer runs through the sound system. So now the written words not only do not function to represent sounds. They plausibly don’t even function by means of sounds.
This is all pretty interesting, I think. I’m probably missing something, because I just started thinking about it. But getting back to Derrida: what interests him about his Rousseau epigraph is precisely this (very mistaken) notion that writing is a sign representing a sign. Derrida presents the view like so: on the one hand, there is an alleged unity of sound and sense—phonocentrism. On the other hand, “with regard to this unity, writing would always be derivative, accidental, particular, exterior, doubling the signifier: phonetic “Sign of a sign,” said Aristotle, Rousseau, and Hegel” (p. 29). I don’t want to get into the whole ‘Western thought has sought to repress writing’ thing. That’s pretty obviously not true. And I don’t want to hear about how Derrida means something new by ‘speech’ and ‘writing’—which he obviously does. But then he shouldn’t confuse the issue by presenting all these quotes from classical philosophers, because they were obviously using ‘speech’ and ‘writing’ in the ordinary, not Derridean sense. What I’m curious about is whether Derrida is, in effect, compounding Rousseau’s error. Derrida is at pains to say that he doesn’t want to just reverse the alleged valorization of speech, over writing. Rather, he wants to say that speech turns out to have the same qualities that writing has (which were felt to be inferior.) But the quality in question is precisely this ‘sign of a sign’ thing, which, although serious thinkers seem to have gone for it, is just a simple mistake. Writing is not a ‘sign of a sign’—that is, writing, even in phonetic alphabetic systems, does not exist solely (or even) to represent speech. So even if it were true, as Derrida alleges, that speech is secretly like (despised) writing, it would hardly follow that speech is always a ‘sign of a sign’.
What I really want to know is this: does Derrida really commit himself to the view that language always represents language. That is, the reason there is (famously) nothing outside the text, boils down to this Rousseau thing, plus the thought that speech is secretly like writing? I’m serious. I really really want to know whether, and if so why, Derrida thinks that language always represents language.
Comments
This post misunderstands Of Grammatology in so many ways, I don’t even know where to start. I’m not going to have time to follow up on any discussion here, so I’ll just point you in the right direction.
“I don’t want to hear about how Derrida means something new by ‘speech’ and ‘writing’—which he obviously does. But then he shouldn’t confuse the issue by presenting all these quotes from classical philosophers, because they were obviously using ‘speech’ and ‘writing’ in the ordinary, not Derridean sense.”
Well, it’s unfortunate that you don’t want to hear about this, but the distinction between writing and arche-writing, between speech and “speech” is indispensable, and not so easily parried. First off, the “classical philosophers” Derrida cites are not “obviously using ‘speech’ and ‘writing’ in the ordinary, not Derridean sense.” Each of the texts Derrida addresses operate on the basic of a structured opposition between two types of signs, one which signifies directly and one which signifies indirectly. Derrida uses the words speech and writing (without quotes) to describe this structured opposition, and he uses these terms because they are the terms on the basis of which the opposition is usually articulated. He uses the words, in short, because the texts he’s citing use those words. And the texts, in turn, use those words in contradictory ways. Thus for Aristotle, Rousseau, Saussure, etc. speech is used to describe both an immediate sign and an audible sign, because these thinkers (erroneously, as your initial analysis suggests) equate the two defintions. Our contemporary understanding of speech, in other words, is the result of this equation that both you and Derrida are trying to undo.
So, if you object to the use of the terms speech and writing, let’s put them out of play for a bit and talk and the opposite between the immediate sign and the sign of a sign. The fact that neither “speech” (Derrida’s quotes here - speech in the sense of the audible sign) nor “writing” (the graphic sign) inherently corresponds to either side of the conceptual opposition rather confirms than overturns Derrida’s analysis.
Thus, the fact that writing, in practice, may not actually first signify speech on the way to signifying thoughts is irrelevant (and not inconsistent with Derrida’s analysis)...because if you follow Derrida’s analysis of speech, speech too turns out to be already writing (arche-writing, in the sense of a sign of a sign, not in the sense of graphic signs). Speech, as a supposed signifier of “thoughts” is already structured more like how we thought writing was structured than like how we thought speech was structured. And thus even if actual writing (graphic signs) does not work by first signifying speech, and we try to theorize how it directly signifies thought, we should end up in the same place as Derrida did when he analyzed speech, i.e. the conclusion that writing too is a signifier of a signifier. All your initial analysis established is the graphic and audible signs work in the same way, but Derrida’s analysis insists that this way is in both cases as a sign of a sign. You don’t need the empirical fact of graphic signs signifying audible signs for this to follow logically.
As for your final question, “does Derrida really commit himself to the view that language always represents language,” I would direct you to Derrida’s analysis of the “trace,” the “imprint”
and the “hinge” in Of Grammatology, whereby even something like what we would call “experience” is structured as a sign of a sign, i.e. structured like language, in a certain sense. Thus the “signs” of which writing and speech are “signs” are not necessarily limited to graphic and audible signs. “Il n’y a pas de hors-texte” thus doesn’t mean that language has nothing to do with what we call “the world,” except in the sense that what we call “the world” is no longer assumed to work the way we thought it did.
Sorry, surlacarte, I think you misunderstood the post. But perhaps my brusque sidelining of certain issues misled you. I really only meant that these issues are irrelevant to the present question.
I also don’t object to the use of the terms ‘speech’ and ‘writing’. I like ‘em fine. (See my post. I use ‘em.) The question is really a simple one: does Derrida respond to Rousseau by arguing that the latter’s error was not to realize that what he said was true of writing, but not speech - namely, that it is a sign of a sign - was actually true as well of speech. If so, that’s a bad argument. The correct view is not that, because writing is not so different from speech, therefore Rousseau’s points apply to writing and speech alike. Rather, the correct view is that what Rousseau says is true of writing is true of neither writing nor speech. Neither writing or speech is ‘a sign of a sign’.
Putting it more simply (and just for starters): does Derrida think that Rousseau is right about writing, i.e. it is a ‘sign of a sign’? I think the answer is yes, he does think Rousseau is right. But Rousseau seems to be wrong. So Derrida is wrong. Right?
Now I expect you will reinterate that “if you follow Derrida’s analysis of speech, speech too turns out to be already writing (arche-writing, in the sense of a sign of a sign, not in the sense of graphic signs).” And that if I am just going to wave away all that arche-writing stuff, as irrelevant, I am begging the question. But I think you’ve seized the argument wrong way round. I am in the process of following Derrida’s analysis of speech, to the alleged conclusion that speech is already like writing, ergo ‘a sign of a sign’. I am noticing that the argument seems to be problematically tangled up in a false premise about the nature of writing - namely, it is a ‘sign of a sign’. The conclusion that speech is always (already) a sign of a sign in some sense follows the from two premises: 1) speech is like writing; 2) writing is always a ‘sign of a sign’. I am actually pretty willing to grant 1. But 2 is false, so it seems. (Unless you’ve got some argument.)
I simply disagree about the relevance of the discussion you so brusquely sideline. It’s necessary to clarify the premises of Derrida’s argument that speech is a sign of a sign, which I think you have gotten wrong. To explain:
You haven’t disproven that writing is a sign of a sign, or that Rousseau is wrong about this point. Writing can still be a sign of a sign without being a sign of speech. In fact, the model of writing as a sign of speech, as Derrida reads it, is just that, a model for the concept of the “signs of a sign.” It is not, at least for Derrida, an empirical claim about the relationship between graphic and audible signs.
As a result, you get the syllogism backwards. It’s not, as you put it:
1) speech is like writing
2) writing is always a ‘sign of a sign’
Therefore 3) Speech is a ‘sign of a sign’
It’s actually more like
1) Writing is always a ‘sign of a sign’
2) Speech is always a ‘sign of a sign’
Therefore 3) Speech is like writing
In other words, the argument that speech is a ‘sign of a sign’ does not follow from any empirical claims about writing. It follows from an analysis of the way in which “classical philosophers” talk about speech: their own discourse, despite its best efforts, reveals that the model they set up through their (empirically inaccurate) account of writing fits all signs, including speech.
Now, to prove this would actually require reconstructing the basis on which Derrida equates speech with the “sign of a sign.” And that would require a much more careful reading of Of Grammatology (you know, with quoting and stuff) than either of us has offered yet, much more careful than I have time to get into at this point. But I do think it would reveal that the argument for speech as a sign of a sign is a lot more complex than simply:
1) Accepting Rousseau’s account of writing as an empirically valid description of graphic signs
2) Assuming that any properties Rousseau ascribes to graphic signs must also apply to audible signs
But, I await evidence to the contrary.
Derrida uses the words speech and writing (without quotes)
You know, when John wrote
I don’t want to hear about how Derrida means something new by ‘speech’ and ‘writing’—which he obviously does. But then he shouldn’t confuse the issue by presenting all these quotes from classical philosophers, because they were obviously using ‘speech’ and ‘writing’ in the ordinary, not Derridean sense.
he wasn’t asserting that Derrida uses quotation marks when he uses the terms. Rather, he was quoting the terms, because if he had written “Derrida means something new by speech and writing”, that would mean something like, “Derrida’s acts of speaking and writing mean something new”.
Not all quotation is done by means of quotation marks, of course. I just quoted something by means of enclosing it in <blockquote> tags, causing it to be indented. And I quoted you by italicizing what you had written. And you quoted the very words John enclosed in quotation marks by italicizing them and preceding the italicaized words with the phrase “the words” (that’s how we know you didn’t mean the words “speech”, “and”, and “writing”, see).
All this is, I suppose, perfectly obvious. I just thought it curious that you found it necessary to make clear that Derrida uses the terms under discussion “without quotes”, since no one would have gotten the impression that he did from what John had written, and you had to do the exact same thing that John did to make himself understood, albeit by different means, in order to make yourself understood.
Yours in triviality,
You might say that you were referring to the words as symbols (probably the wrong, schemey term) by italicizing them, but, given that he used ‘ and not “, that’s what John was doing too.
Ben,
My distinction between “with quotes” and “without quotes” was not meant as a criticism of John’s use of quotes, but rather a commentary on the (at least semi-rigorous) distinction between writing and “writing” IN Of Grammatology. Sometimes in order to distinguish between arche-writing (i.e. the “sign of a sign” whether graphic, audible or otherwise) and writing as graphic sign, Derrida uses quotation marks on the latter. I was merely trying to indicate the same distinction in my own text. My criticism of John, as I hopefully clarified in my follow-up post, is that by failing to insist upon the distinction between writing (with quotes) and writing (without quotes), he misstates Derrida’s argument.
In this particular case, I blame Derrida. Not that the distinction writing/"writing" doesn’t have a certain elegance. But it definitely causes problems whenever anyone attempts to quote Of Grammatology.
Thank you, ben, that’s a bit of a relief. (I was thinking of writing something similar myself, but it seemed trivial.)
surlacarte, you gloss Derrida’s argument like so:
1) Writing is always a ‘sign of a sign’
2) Speech is always a ‘sign of a sign’
Therefore 3) Speech is like writing
In my post, I have given an argument for why 1 is false (if it is believed on Rousseau-ish grounds.) Is there any reason, besides the bad, Rousseau-ish one, why I should buy 1? I am certainly not going to simply take it as a premise. As to 2, let us suppose you are right that he doesn’t argue for it as I say? Then how DOES he argue for it. It is just as implausible as 1, on its face.
It seems to me - and I have included only a few quotes and stuff, I admit - that the chapter on “Linguistics and Grammatology” is presented as if it contains an argument for the crucial claim that language is necessarily ‘a sign for a sign’. It seems to me that Derrida argues that what Rousseau says about writing must actually be true about language generally, ergo all language is a sign of a sign. If you do not think that is how the argument goes, in broad outline, then can you at least give a hint as to how you read the chapter?
What role do you take the Rousseau position to be playing in Derrida’s presentation, if it is not the role I am proposing for it?
Also, do you concede that my argument against Rousseau’s claim is a good one? That is, it is wrong that written language represents the sound of spoken language. So whatever the argument may be that writing is ‘a sign of a sign’, it isn’t Rousseau’s.
[Warning in advance: I’ve gotten a bit worn out and haven’t finished this response. In fact, the part I left out at the end is probably the most crucial question. The end will follow shortly, tonight or tomorrow, so take this as a to be continued]
Well, despite my best intentions to work on other things, I opened up Of Grammatology. But I actually do really appreciate the chance to brush up on my Derrida, and your latest response, John, presents your objections very clearly and systematically, so let me do my best to, first, support my claims with textual evidence, and second respond to your points one by one. [Apologies, by the way, for starting my first comment off in such a polemical manner.]
First, the crucial passage from Of Grammatology (p. 41 in Spivak’s translation):
Writing is not a sign of a sign, except if one says it of all signs, which would be more profoundly true. If every sign refers to a sign, and if “sign of a sign” signifies writing, certain conclusions--which I shall consider at the appropriate moment--will become inevitable. What Saussure saw without seeing, knew without being able to take into account, following in that entire metaphysical tradition, is that a certain model of writing was necessarily but provisionally imposed (but for the inaccuracy in principle, insufficiency of fact, and the permanent usurpation) as instrument and technique of representation of a system of language.
I take the first sentence to mean that the attribution of the descriptor “sign of a sign” to writing can never be made on the basis of any property specific to writing. It must be a universal characteristic of the sign in general, or not a characteristic of the sign at all. Such a universal claim could never be founded (or undermined) on empirical arguments about the way that people who can hear (or who can’t hear) use speech and/or writing, because such arguments would necessarily be contingent on the specificity of the medium. The reason, in short, that writing is a sign of a sign must be the same as the reason that speech is a sign of a sign, and this is certainly not the case if we say that the reason writing is a sign of a sign is that writing is a sign of speech. The same logic clearly would not apply reciprocally to speech.
What, then, is the significance of Saussure’s treatment of writing as the exclusive domain of the sign of a sign, despite Derrida’s objection that either all signs are signs of signs, or none are? Derrida makes this clear in the final sentence of the quoted passage - “a certain model of writing was necessarily but provisionally imposed ... as instrument and technique of representation of a system of language.” Saussure uses the model of writing as a sign of a sign as a means of explaining how language, as a system of differences, works. The significance of Saussure’s account of writing thus does not lie in the accuracy in fact or principle of its statements about graphic signs (note the portion of the passage I just omitted: “but for the inaccuracy in principle, insufficiency of fact, and the permanent usurpation"). Instead, it lies in the fact that, drawing on this erroneous account of graphic signs, Saussure is able to arrive at a conceptualization of the sign in general (whether graphic, audible, or otherwise) which is even more profound than he could have imagined (i.e. which he “saw without seeing"). The value of the account is conceptual rather than propositional: it reveals nothing about actual writing, and thus isn’t falsifiable on the basis of what it reveals about actual writing.
Now, it still doesn’t follow that the concept of the sign of the sign actually corresponds in any real way to any real system of signs. On the one hand, even if we could not prove that any signs were signs of signs, Of Grammatology would remain an invaluable explication of Saussure and Rousseau, whose texts are clearly structured by the logic of the sign of the sign in suprisingly complex ways that exceed empirical claims about graphic signs. But that’s still a bit of a cop-out, and it is a legitimate question to ask whether and how and why actual writing and actual speech do correspond to the logic of the sign of a sign which Saussure and Rousseau thought to be a unique property of graphic signs.
So, that brings me to your latest comments.
First, you state that, in your post, you have given an argument for why the Rousseau-ish grounds for believing that “writing is a sign of a sign” is false. I don’t refute that. Based on the passage from Derrida I quoted above, I doubt that, in principle, he would either. I say “in principle” because I suspect that there’s more going on in your particular choice of examples (the deaf man, and the Chinese Box) than either of us quite grasps [the resemblance to the deaf man in Nietzsche’s “On Truth and Lie,” for example, is interesting]. Nonetheless, this is certainly not where I would want to focus my commentary, so I will grant, wholesale, your claim that the Rousseau-ish grounds for calling writing a sign of a sign are inadequate.
Continuing, you ask if, absent these bad reasons for calling writing a sign of a sign, there are any good reasons. My answer is yes, but I will acknowledge that Derrida may not directly answer the question. It seems, though I may be missing something, that the burden of Derrida’s argument is on proving that speech is a sign of a sign, since the writers he addresses already treat writing as a sign of a sign.
This is not a problem, however, unless your original account of Derrida’s premises is correct. As long as Derrida does use the ready-made equations “writing=sign of sign” and “writing=speech” to prove “speech=sign of sign,” Derrida’s claims about speech still need to be evaluated independently. And this ultimately will turn back upon the question of whether writing is or is not a sign of a sign, because the arguments about why speech is the sign of a sign will hopefully tell us something about writing (graphic signs) too. So in fact we may need to reverse the order of the premises:
1) Speech is a sign of a sign because of properties x, y, and z
2) Writing shares properties x, y and z
Therefore, 3) Writing is also a sign of a sign
[Or treat premises 1 & 2 independently; in either case, the point is that the premise about speech is never dependent on the premise about writing, so the accuracy of the latter is not relevant to the accuracy of the former, though the reverse may be true]
This takes us, precisely to “Linguistics and Grammatology,” which, I’m quite glad you observe, “is presented as if it contains an argument for the crucial claim that language is necessarily ‘a sign for a sign.’” The burden of our either of our arguments is to reconstruct the logic of this claim…
[I’m going to pause here. I know I haven’t defended the part of my argument you’re most skeptical about, and I will, but this comment is getting long and I’d like to submit it and take a break. The answer, when I return, will likely have something to do with “articulation,” “the absence of the referent” and the “transcendental signifier.” To be continued...]
Well, I’ll await part II, surlacarte, thank you for part I. Let me just sign off for tonight myself with a clarification. I said I was setting certain things to one side and I could have been clearer about exactly what and why. I should have said: I’m setting to one side Derrida’s historical claims about how phonocentrism has dominated Western thought/philosophy. I don’t buy it, but, at least for blog purposes, I set to one side any attempt to discuss all of Western civilization. It’s quite enough to worry about (nevermind what everyone has allegedly thought since Plato) what Derrida thinks we should think. This thing I am setting aside may creep back in, because I think the problem with claims like ‘speech is like writing’ is that they are actually rather trivial, whereas Derrida thinks they are radical. So, for example, my problem with surlacartes’ first formulation was really not just that I doubted the premises, but that I would have been willing to freely grant the conclusion AS a premise. But I really don’t care about that. I care, specifically, about arguments for ‘all language is a sign of a sign’. Which is what surlacartes is now focusing on, so fine.
Moving right along, it does make a bit of sense to me - in the abstract - to reverse the order of argument, as surlacartes suggests, thereby making the Rousseau point an ironic curlicue of an afterthought, rather than a dubious premise, but I really am not seeing the x, y and z that will then get us ‘all speech is a sign of a sign’.
OK, one last point: surlacarte characterizes my arguments as ‘empirical’ - as opposed to the purely conceptual arguments that, allegedly, we will be getting from Derrida. But, just for the record, I don’t think my argument is empirical. It hasn’t actually been empirically tested, nor need it be (I hope). It’s a pure, conceptual thought-experiment. It is somewhat medium-specific, yes, since it’s a thought-experiment about writing. But that doesn’t make it empirical. I take it that general reflections about language are likely to be no more nor less empirical than the arguments I offer, then.
John, there is an important way in which speech (or writing) represent ideas. They do so in the same way that we could say a book of sheet music represents music. Peter Elbow, for example, suggests that we view a piece of writing as a set of directions, transmitted from person #1 to person #2, for reconstructing in person #2’s mind the mind of person #1. One of the things in person #1’s mind might be assertions about the state of the world, but not everything in his piece of writing might be. For example, most of what’s in a novel is not an assertion about the world but a construction of a world that might be held up as a model or exemplum of certain aspects of the real world.
The question I have is this: are we using “representation” properly when we say that sheet music represents music? Sheet music certainly represents *ideas* about how music might be made, ideas transmitted from one person to all people in a socially conventional manner.
(One other point: there is an important sense in which writing can be or can avoid representing speech. Mina Shaughnessy investigated this in depth in her brilliant *Error and Expectation*, in which she demonstrated that most of the errors made by “basic writers” are errors of the translation of speech into writing—that is, basic writing errors occur when the writer writes a transcription of how s/he speaks.)
What I think about this is here, via the method of poetastic analysis.
Seriously, my problem with “the sign of a sign” is that sometimes writing or speech appear to be play acts. (I know that this is heavily covered Theoretical territory.) Adam has been into this area recently with music vs noise and weird saturnian hexagons, but in short, the words in a poem are only partly signs of signs.
Not that I know anything about Derrida’s argument in particular.
Luther, I grant that there has to be some sense in which speech represents ideas, not just objects. What I object to is the suggestion that language only represents our ideas - our idea of the dog, as opposed to the dog itself. I don’t think that turns out to work.
The sheet music analogy has some interesting features. I thought about using it, comparing my hypothetical non-English speaker who can pronounce English sentence with a musician who performs a score. I don’t think it is really right to say that sheet music represents ideas about how music might be made. If you want that sort of thing, read some music criticism. Sheet music, on the other hand, represents series of notes, which are sounds and not ideas (although of course if you were brain dead you couldn’t read music). If sheet music were about ideas, then I think it would have to assert something - or at least entertain some claims - and I don’t think it does. Sheet music consists of a set of imperatives, which are obeyed if a set of notes are generated; not a set of claims which could turn out to be true or false.
Oops, my second link above should have been to here.
It is, of course, perfectly possible that viewing sheet music may cause ideas about how music might be made. But that is different than saying the sheet music represents those ideas. (My sentence ‘the cat is on the mat’ may cause you to think all sorts of cat-related thoughts. But my sentence doesn’t represent those thoughts.)
For example, the thought most reliably caused by utterances of ‘the cat is on the mat’ is ‘why do analytic philosophers always use that sentence as an example?’ All the same, the sentence does not REPRESENT this thought.
Alright, folks, I have fallen asleep with the lights on and woken back up, which means part of me isn’t entirely sure that I slept. But I’m back to finish what I started.
First off, some quick comments on recent comments.
1) John, my primary interest here was to “reverse the order of the argument” rather than to necessarily defend the reversed argument, so while I’ll still take a stab at the latter (I think I’ve committed myself to doing so), I’m content with your concession either way, since I think it leaves us right where we should be - pouring over Derrida’s words, instead of battling over the logical consistency of misunderstood excerpted sentences, which is the mainstay of facile criticisms of deconstruction (not that I’m necessarily accusing you of doing this exactly, but think about the most common responses to “il n’y a pas de hors-texte")
2) On the word “empirical,” this may not be the best choice of words on my part. The reason for the word choice is that medium-specific arguments are always empirical in a certain sense insofar as they assume a historically, culturally, and biologically specific model of how the medium works that is somewhat contingent - specifically, phonetic writing, as well as the human apparatus of sensory perception and sign production that happened to evolve. The point, in any case, is to distinguish between arche-writing and the particular form of graphic signs (really should be using the grapheme) that “predominates” in “our culture.”
3) I won’t get into the sheet music thing, since I’m unclear what that helps clarify. It seems like a substitute for going back to the text?
Now, for going back to the text.
Here are 3 ways in to speech as sign of a sign:
1) Articulation and the phoneme
A great portion of Derrida’s analysis of speech in “Linguistics and Grammatology” takes place on the level of the phoneme, which is somewhat difficult to explicate for a variety of reasons, not the least of which is that the primary position it is intended to undermine (the natural relationship of speech and sense) is not something we even think we believe in any more. Whereas as Heidegger talks about the “voice of Being” as if their were still some direct relationship between speech and being, post-Saussure, we’re already somewhat out of this position anyway, which is somewhat the point. So, to an extent, my argument about the phoneme will still have to be something more like a model than a proposition.
Anyway, part of the presumed difference between speech and writing that makes speech seem like natural or direct signification has to do with articulation, i.e. the breaking up or spacing of the plenitude of experience into discrete elements. To say that writing is always already articulated is standard fare: when you look at this post on your computer screen, you don’t see a series of pixels or a large mass of black lines and curves. You see letters. You can’t not see letters. There is no direct sensory experience in writing, since the world already makes itself present in the form of the articulated trace, the division into discrete elements. Derrida’s analysis of phonetic writing pays a great deal of attention to elements like the period, the space, and the comma which can be said either, on the one hand, to space or articulate speech, breaking up the fullness of the speech stream into units like the word and the sentence, or, on the other hand, reveals that speech is always already articulated, which is the final conclusion. Thus the phoneme as sign of a sign: not the sign of a presence, like the speech stream, but the sign of a trace of that presence, the presence itself never really making itself “present” at all.
2) Duration, langue/parole, absence/presence
I’m making a bit of a leap by connecting Derrida’s analysis of duration to his analysis of presence and absence, but I think it’s justified. One argument commonly made in privileging speech over writing is that writing, by giving duration to speech (or to discourse, if we’ve rejected the claim that writing is first a sign of speech) removed it from the context in which it was articulated. Thus writing is at the root of Saussure’s distinction between langue and parole, i.e. between language and utterance. Writing is a sign of a sign insofar as we attempt to ground its meaning in the moment of its production (or reception, for that matter): we imagine the written text as a sign for the full presence of the speech act, which grounds its meaning. It matters not whether the speech act actually involved speech (phonemes). Speech is the category in which we conceptualize a fully present discourse in which langue and parole co-occur. Of course, the trajectory of Derrida’s analysis is to deny speech this full presence. Partially this is accomplished through the idea of articulation - even in the context of articulation, there is no direct experience of the speech stream, which is always already articulated into phonemes. There are other ways to make this argument which are probably more familiar, so I won’t pursue this much further. Regardless, speech would be a sign of a sign if its meaning was always guaranteed by a moment of the speech act and by a full presence of the voice which could never in fact be made present, and which never in fact appeared as fully present in the first place.
3) The absence of the referent
In the model where writing signifies speech which signifies thought, what is thought anyway? Certainly not a full presence. Most often, it ends up being conceptualized as something like an inner monologue. So even if writing doesn’t signify outer speech, speech and writing are both thought of as signifying an internal discourse, which signifies what? This is in a sense the limit of the hermeneutic model - even if you can reproduce the “original thought,” the original thought is not a full presence, but something that in turn must be interpretted. Unless there is a direct relationship between thought and experience or, as you suggest in your original post, some other transcendental signifier like the “mind of God,” then the “great chain of meaning” never comes to an end. It matters not whether writing is the “low man on the totem pole” or on par with speech - in either case, the logic of the chain of meaning (which is the logic of the supplement) is essential to the way speech and writing work.
That’s my best for now. The text itself is much more complex than this, and there are a variety of arguments that fit into and don’t fit into the question of speech as sign of sign to varying degrees. The point is to work through each of these claims carefully without jumping on the validity of the supposed end point in the abstract.
"Use” vs. “reference”!!!1!!1!
Wheeeeee!
But wait—is it really “use versus reference”?
No, I think it’s “"use" vs. “reference"” all right.
???
Doesn’t this comment pretty much wrap up the ridiculous discussion of quotation marks?
I wasn’t discussing quotation marks, I was referring the discussion of quotation marks.
oh
"Unless there is a direct relationship between thought and experience or, as you suggest in your original post, some other transcendental signifier like the “mind of God,” then the “great chain of meaning” never comes to an end. It matters not whether writing is the “low man on the totem pole” or on par with speech - in either case, the logic of the chain of meaning (which is the logic of the supplement) is essential to the way speech and writing work.”
Laying my cards on the table somewhat: I don’t buy this stuff about the chain of meaning, for more or less Wittgensteinian reasons. It’s a mistake to assume it must be ‘interpretation all the way down’. Wittgnestein would say this merely stands a bad metaphysical explanation on its head, rather than standing back from bad metaphysical explanations (which is what you should do.) The whole chain of meaning thing seems to me an unwarranted speculative leap, on Derrida’s part. And my original sense was that the Rousseau argument was somewhat papering over the lack of a better argument at this juncture.
But I’m just telling you what I think, not why I think it. I’ll try to follow up more later.
I haven’t digested surlacarte’s lengthy posts, but let me try something on you [John] directly.
The way you present it, the original problem is that Rousseau thinks that writing represents speech, which itself represents the world (or maybe our ideas, which represent the world). In any case writing is supposed (as surlacarte puts it in his/her first post) to signify indirectly where speech signifies directly. I think we all agree that this is wrong. Writing and speech are on the same footing as far as degree of “directness” is concerned. The question is how to put the point. John says neither speech not writing is a “sign of a sign” while surlacarte’s Derrida says they both are. We have traded an agreement for two disagreements.
Here’s what I think is happening. John, you seem to be using both “is a sign of” and “represents” to mean something like “denotes”. But there are more kinds of “representation” (and, perhaps, of “being a sign of” something) than semantic denotation. (Goodman’s The Languages of Art is an interesting attempt at an exhaustive catalogue).
Why does this matter? Let’s look at your example of the non-English speaker and the congenitally deaf person: “But if we now ask which of the two reads with understanding of what the written words represent, I think we would be strongly inclined to say: the deaf person. He understands what the language he is reading means.” Naturally I see your point. But I’m not sure this helps.
Look at what I emboldened. As we use it, the word “word” is ambiguous. Does it mean “inscription-type” (or “sound-type"), or does it mean “concept (i.e. as reliably associated, in the relevant contexts, with a particular inscription- or sound-type)”? I think this ambiguity is significant and indeed useful (if that’s the word I want). That is, I think one of our problems arises when we see it and attempt, as it seems Saussure does, to disambiguate by distinguishing sharply between the physical type and the concept and sticking a referential relation in between them to keep them at arm’s length. On this view, the physical sign “dog” refers to the concept “dog”, which itself has Fido and Ren and Santa’s Little Helper as referents. Saussure thinks he needs to say this in order to capture the arbitrariness of the sign (but I’m no Saussure expert). We’ve agreed that we don’t want this.
But the way you use the concepts (the words?) “written words” and “represent” here, John, makes me think that’s what you too are doing, albeit unintentionally, when you demand that it is the deaf person and not the non-English speaker who understands what “the written words” “represent”. We would do better to say, as I had us say above, that the two cases are on a par: the d.p. understands the words/concepts (could recognize their referents when presented, etc.), while the n.-E. s. understands the (in his case merely sonic) referents of the words/inscriptions, i.e., what they (in this sense) represent (otherwise he wouldn’t be able to produce the right sounds when presented with the corresponding inscriptions).
Normally these two sorts of representation happen at the same time, and we feel no need to pick one as the primary one. We miss this if we pull them apart as you do (that is, and don’t put them back together again). Naturally we want to see ourselves, and our words/concepts, as (pace Rorty) “representing”, or at least referring to, the world rather than other signs. But that doesn’t mean we should pull the two senses apart and crown one as primary.
In other words, the settled ambiguity of “word” (and the corresponding one of “represent") means that the distinction you insist upon (not, again, that it’s not real) has been, if you like, pre-aufgehoben. I like it that way; let’s leave it (as long as we don’t get confused). It can’t be irrelevant, for example, that I don’t think that guy would read aloud very well if he didn’t understand the meaning of the words/concepts as well as the (in his case merely sonic) reference of the words/inscriptions. After all, that’s one reason computers don’t “read” aloud very well.
I also think this is the same thing which confuses me about your (manner of) resistance to Knapp & Michaels. They insist, like Saussure, on the arbitrariness of the sign (the sand-poem qua inscription), telling us that it for it to be meaningful it requires interpretation, where that means something like divination of authorial intention (which, ex hypothesi, there isn’t here). You counter by finding a (literally) unintended meaning in the sand-poem qua meaningful sign, however manifested, making the subsequent train-wreck inevitable.
But let’s not get into that again. My point is just that your rejection of the “ubiquity of interpretation” is unnecessary (for our purposes, including Wittgensteinian rule-following ones as in PI §201) in just the same way as your rejection of both (i.e., even when presented as a correction to Rousseau, putting speech and writing on a par, as we demand) speech and writing as “signs of signs”. To say that each is a “sign of a sign” (or, say, that there is no “hors-texte") need not mean anything like that language can’t represent the world, but (maybe) simply that it doesn’t represent either the Cartesian or the platonic world-in-itself. Which it doesn’t.
Similarly, again, the “ubiquity of interpretation” need not be feared if we understand what it means. It’s not the necessity of giving life to dead signs (as in K/M, or Kripkenstein); it’s that of aligning our meanings with those already there in the living ones (and, not coincidentally, of aligning our beliefs with the facts in inquiry). Or in other words, in my sense the later Davidson is just as “hermeneutic” as are Gadamer et al.
Now that I know what I think (I think), let me have another crack at what surlacarte said.
I’ve never actually read Derrida, but since this is just a blog, and on the basis of a rough aquaintence with Hiedegger and Wittgenstein, I’ll take an amateur stab anyway. Speech and writing are “signs of signs” by virtue of the fact that both are necessarily encoded, i.e. differentially articulated. But that means that they pick out features of experience/the world precisely on the basis of their being encoded/articulated. Which means that those features that are picked out, including the means by which we interact and thereby communicate with one another, themselves partake of a sign-character: hence, speech/writing is always a “sign of a sign”. (In structuralist jargon, this is called “double inscription").
Derrida is by no means denying the existence of an independent world, nor of any distended experience/thought. That “there is no inset-plate” simply means that reference is always a more complex matter than we would take it to be. That is, that even the most complex account of reference is not quite complex enough. The “joke” is that he is not denying that reference occurs: it’s not a matter of “lingustic idealism”, nor of any “pure” constructivism. Rather, the “point” is that any reference is itself subject to re-contextualization/redifferentiation, such that the process begins all over again. And that de-stabilizes any purely semantic account of “meaning” and any idealization of the modal-relational or illocutionary “conditions” which would suffice to guarantee or authenticate it. Derrida is at once renewing and problematizing the Heideggerian account of Being as what “gives” beings and of the “ent-framing” by which that would occur and be received. The upshot is that rather than attempting to secure “meaning” ontologically and thereby certify knowledge epistemologically, Derrida is engaged in questioning the conditions of communication by which knowledge is at once set into the world and subject to the same encoding processes/practices, and further questioning the relations of inclusion and exclusion by which who, what, and where are raised into discourses claiming the “authority” of knowledge.
But basically, I don’t think that there’s a lot new or original with Derrida. The criticism of a representational account of meaning and language as the basis of knowledge and the foundation of social relations goes all to way back to Hegel’s criticism of Kant. To put it in a rough quasi-Wittgensteinian way, what is it about our words/meanings that represents “things” other than the ways in which we use them? Would there be a special property of words and their “grammar” that could guarantee their correspondance to a unique world, independent of the social exchanges of those who use them to communicate? How would reference function independently from the faillible understandings of those who use words to communicate with and understand one another?
"Inédit” means “unpublished.” Indeed, the Collins Robert includes no mention of “unedited” as a meaning of this word—it either means “[hitherto] unpublished” or “new” (in the context of publishing).
Curse those faux amis! Of course you realize that this error—even if you were conscious of it and did it in jest—completely and absolutely destroys, undercuts, and invalidates everything else you say in this post.
Suppose I invented another writing system for English. A picture of a banana was the word “the.” And so on… every separate word had a picture. Then I would say, my banana-picture is the sign of a sign. It is a sign of the word “the.” The printed word “the” is also a sign of the word “the.” And the pronounced word “the” in speech is ALSO a sign of the word “the.” Phonocentrism is the mistaken believe that somehow the pronounced word “the” is a more direct sign than the banana or the written word “the” is. That’s pretty much the gist of what Derrida is saying. Whether the metaphysical consequences follow from this is another question… That is, the leap from phonocentrism to logocentrism.
The fact that phonetically based alphabets seem to are as transcriptions of spoken language is a confusing factor. What happens is that then the written system can function all by itsef, as Holbo’s “deaf person reading” example shows. Suppose I learned banana language as a purely visual medium and never thought about the sounds behind those pictures. The characteristic of the banana language is that it can function as a sign system without any sound. The fact that spoken language uses the medium of sound is an accident. Sign systems as a whole--the argument runs-- are more like writing than speech in the sense that the sign can function in the absence of a speaker sound the words. Take all the conceivable ways of representing English, signs of signs--only the spoken language needs sounds.
For the record: I knew, even before Adam left his comment, that ‘inédit’ means unpublished.
Dave, I don’t want to argue too strenuously about the deaf person vs uncomprehending pronouncer cases. I actually meant to emphasize that ‘read’ is quite ambiguous, hence there will be a certain tendency for it to make all associated terms, like ‘about’, ‘understands’, even ‘refers to’ ambiguous as well. It is possible to say that the written words mean sounds (like notes in sheet music). The mere complexity of the case is enough to scotch the Saussure/Rousseau view as unworkable. I’ll settle for that.
The real question that I genuinely am interested in is whether Derrida is, in some sense, a ‘linguistic idealist’ as you put it. I think he is. And that this is his problem. I think he is commmited to denying that reference to extra-linguistic reality occurs. (I realize he actually doesn’t want to be committed to this. But it seems to me he is, and that’s the trouble.) The short version is that it is possible to speak loosely about everything as a text, because everything is interpretable - the dog as well as ‘the dog’. But there is an important difference between what the dog means and what ‘the dog’ means. Namely, the arbitrariness of the sign applies to the two cases rather differently, and Derrida misses the difference. I’ll try to write more later.
Saussure is rather tricky on the linguistic sign. Here’s a passage from the Course in General Linguistics (Baskin translation, McGraw-Hill, 1966, p. 66)
The linguistic sign unites, not a thing and a name, but a concept and a sound-image. The latter is not the material sound, a purely physical thing, but the psychological imprint of the sound, the impression it makes on our senses.
He then undertakes a brief discussion, introducing a simple diagram illustrating his point that “the linguistic sign is then a two-sided psychological entity” in which “the two elements are intimately united, and each recalls the other.” He then introduces his example—“tree” and “arbor”—some variants on the two-sided diagram, and finally introduces his famous trio of terms, sign (for the whole thing), signified (for the concept), and signifier (for the sound image) (p. 67). He does not, so far as I can tell from this translation (not of Saussure himself, as we know, but of lecture notes taken down by students), talk of the signifier as representing the signified. Rather, the two are associated, and it is through their association that they constitute a sign representing some object or state of affairs in the world.
As far as I can tell, it is quite common for discussions of these matters simply to forget that the Saussurian linguistic sign includes the concept and to talk as though the sign were the sound itself (not even the sound image of Saussure’s signifier). Much of the preceding discussion seems to go that way, though I think Dave Maier was getting at this sort of thing in pointing out that there are various kinds of representation. Whether this has much bearing on Derrida’s argument is something I’ll leave to others.
But I want to make a few more remarks. Linguists distinguish between phonetics and phonemics. The former is about the acoustic properties of the speech stream; it is about physics. The later is about the linguistically relevant properties of the speech stream; in a sense, it is about Saussure’s psychological imprint. As an acoustic entity, the speech stream is fairly continuous; in particular, it is all but impossible to identify discrete phoneme boundaries on a sound spectrogram. You can’t take a tape recording of speech and cut it into discrete pieces such that each piece contains the sound of a single phoneme. They tried that after WWII and, much to their surprise and displeasure, discovered that the snippits could not be reassmbled in a different order so as to sound different words. The discontinuity, the phonemes, has to do with how we hear the speech stream.
Which implies that there is more than one way to hear a speech stream. If you don’t speak, e.g. Japanese, it will, as the saying goes, sound like Greek to you. It is not only that you don’t know how to associate the signifieds of Japanese with the appropriate signifiers, but that you cannot even properly hear the signifieds. (Note that linguistic and non-linguistic sounds are analyzed by different regions of auditory tissue.) And if a speaker of Japanese confuses /l/ and /r/ that is because the difference between those sounds is not phonetically significant in Japanese and so the speaker has trouble hearing any difference.
There is thus something of a “gap” between the actual sound of the speech stream and our subjective apprehension (perception) of it. But that is true on the concept side as well—perhaps this is what John Holbo had in mind when pointing out that both dog and ‘the dog’ are interpretable. How is it that we can recognize and distinguish trees and dogs? For that matter, how is it that dogs and monkeys can do it? For that problem has nothing to do with language itself; it’s a problem about perception.
We don’t have some immediate apprehension of the external world. We perceive it with sensors that have various limitations and analysers that work in specific ways. Why, for example, is it relatively easy to recognize photographs of faces when they are presented in normal orientation but somewhat difficult to do so when the images are presented up-side down? The images are the same. Why should our perceptual and analytic equipment be so sensitive to orientation?
So what? On the one hand, perception is such a very difficult and subtle business that I don’t think complacency about it is warranted. Nor is it simply about perception. We talk about all sorts of things that can’t be apprehended by the senses—e.g. truth, justice, and the American way. How is it we can do that at all? But I don’t think the notion of an infinite regression of signs representing signs is a terribly useful way of thinking about these difficulties.
Thanks for the Saussure lesson, Bill. I have that book somewhere - I should read it before shooting off my virtual mouth. Clearly I was wrong to conflate the writing vs. speech and sound (inscription) vs. concept issues in Saussure. I like the idea of a “two-sided psychological entity,” though maybe I would say “intentional” rather than “psychological.”
I still await John’s explanation of how to interpret dogs.
Thanks, Dave, though I note that I didn’t make the comment specifically for you or for anyone else. Given that we’re dealing with signs of signs—perhaps indefinitely in both directions—I just wanted to provide a little clarification about Saussure. Although it’s been some time since I traveled among people professing semiotics and structuralism, etc. it’s been my experience that even very sophisticated people think that the Saussurian sign is only the signifier so that one might then think of the signifier as being the sign of the signified (which, in turn, represents something in the external world).
Whether or not Saussure’s actual idea makes any difference in the context of this discussion, that I’ll leave to those who’ve read Derrida more closely than I have.
I apologise in advance for the incivility of my presuming both to contribute to this discussion in advance of any formal introductions and to sense some ulterior motives — prejudices, even — animating some of the remarks made throughout the course of this discussion.
(I apologise, too, for the convoluted nature of my opening apology; I’m re-reading Austen’s Pride and Prejudice at the moment, and I just love the way they spoke back then—or should I say, the way such a manner of speaking is represented by Austen’s written-down version of her idea of how her characters’ would speak if they were to speak in the manner of contemporary, real-world speech?)
I’ll begin by recounting my affective response to the relevant parts of the discussion. First up, I enjoyed the original post, especially its requests for guidance and gestures of openness to feedback, which is what immediately prompted me to consider writing a response. Then I was quite taken aback by the confrontational and seemingly uncharitable tone of surlacarte’s first reply. Then I was impressed by John Holbo’s relatively civil, non-inflammatory response to that reply. This oscillation of sensation continued — paralleling the turns taken by each of these two contributors — until surlacarte modified his/her tone and offered an excellent (if necessarily insufficient) account of D’s Grammatology.
In other words, in this discussion John Holbo seemed the picture of academic civility and tolerance, whilst surlacarte seemed unwarrantedly aggressive and opposed to the spirit of questioning evinced by John Holbo’s initial post.
Then I read this: “The real question that I genuinely am interested in is whether Derrida is, in some sense, a ‘linguistic idealist’ as you put it. I think he is. And that this is his problem. I think he is commmited to denying that reference to extra-linguistic reality occurs. (I realize he actually doesn’t want to be committed to this. But it seems to me he is, and that’s the trouble.)”.
So much for the gestures of openness and the requests for guidance. It seems John Holbo is already well guided by his prior assessments of Derrida’s character (sorry, philosophical commitments) and has no real interest in opening his reading to any form of guidance that seeks to disabuse him of such prejudices. (Is there a lesson here on the occasional non-coincidence of sincerity and civility?)
Thanks are owed to Dave Maier, john c. halasz and Bill Benzon (and surlacarte) for their respective illuminations on the issue. Dave Maier’s insistence that a great deal of the discussion has hinged on a very particular, unstated conception of “representation” is very much to the point. john c. halasz provides an excellent overview of the implications of D’s various arguments — despite never having “actually read Derrida” — for any “theory” of language and reference. And Bill Benzon’s point about the signifier being the “sound-image” is a cardinal one. (BTW, Bill Benzon, I wonder if you could confirm a suspicion for me? I’ve long suspected, but never had the patience to confirm, that Saussure never once in the Course uses the term “referent”; do you know if this is true? I don’t mean to suggest, of course, that Saussure denies the existence of reference or referents, etc. It just strikes me that many commentators on Saussure — Barthes amongst them — include the concept of “the referent” within his theory of the sign, when I don’t think he even introduces the term, let alone defines or elaborates on it....)
Back to the central questions:
does Derrida really commit himself to the view that language always represents language. That is, the reason there is (famously) nothing outside the text, boils down to this Rousseau thing, plus the thought that speech is secretly like writing? I’m serious. I really really want to know whether, and if so why, Derrida thinks that language always represents language.
It’s interesting, John Holbo, that you ask these questions in light of your reading of Chapter 2 of the Grammatology. Can one assume that in entering Chapter 2 you had first gone by way of Chapter 1?
However the topic is considered, the problem of language has never been simply one problem among others, But never as much as at present has it invaded, as such, the global horizon of the most diverse researches, diverse and heterogeneous in their intention, method, and ideology. The devaluation of the word “language” itself, and how, in the very hold it has on us, it betrays a loose vocabulary, the temptation of a cheap seduction, the passive yielding to fashion, the consciousness of the avant-garde, in other words — ignorance — are evidences of this effect. The inflation of the sign “language” is the inflation of the sign itself, absolute inflation, inflation itself. (p.6)
By a slow movement whose necessity is hardly perceptible, everything that for at least some twenty centuries tended toward and finally succeeded in being gathered under the name of language is beginning to let itself be transferred to, or at least summarized under the name of writing. By a hardly perceptible necessity, it seems as though the concept of writing — no longer indicating a particular, derivative, auxiliary form of language in general (whether understood as communication, relation, expression, signification, constitution of meaning or thought, etc.), no longer designating the exterior surface, the insubstantial double of a major signifier, the signifier of the signifier — is beginning to go beyond the extension of language. In all senses of the word, writing this comprehends language. Not that the word “writing” has ceased to designate the signifier of the signifier, but it appears, strange as it may seem, that “signifier of the signifier” no longer defines accidental doubling and fallen secondarity. “Signifier of the signifier” describes on the contrary the movement of language: in its origin, to be sure, but one can already suspect that an origin whose structure can be expressed as “signifier of the signifier” conceals and erases itself in its own production. There the signified always already functions as a signifier.... This, strictly speaking, amounts to destroying the concept of the “sign” and its entire logic. Undoubtedly it is not by chance that this overwhelming supervenes at the moment when the extension of the concept of language effaces all limits. (pp.6-7)
To affirm in this way that the concept of writing exceeds and comprehends that of language, presupposes of course a certain definition of language and of writing. If we do not attempt to justify it, we shall be giving in to the movement of inflation that we have just mentioned, which has also taken over the word “writing” and that not fortuitously. For some time now, as a matter of fact, here and there, by a gesture and for motives that are profoundly necessary, whose degradation it is easier to denounce than it is to disclose their origin, one says “language” for action, movement, thought, reflection, consciousness, unconsciousness, experience, affectivity, etc. Now we tend to say “writing” for all that and more: to designate not only the physical gestures of literal pictographic or ideographic inscription, but also the totality of what makes it possible; and also, beyond any signifying face, the signified face itself. And thus we say “writing” for all that gives rise to an inscription in general, whether it is literal or not and even if what it distributes in space is alien to the order of the voice: cinematography, choreography, of course, but also pictorial, musical, sculptural “writing”. One might also speak of athletic writing, and with even greater certainty of </b>military or political writing</b> in view of the techniques that govern those domains today. All this to describe not only the system of notation secondarily connected with these activities but the essence and the content of these activities themselves. (pp.8-9)
There is a great deal to note in these passages (and much in what I was forced to omit). I’ve emboldened a few key words to try to indicate some of the things that are at stake here. Note:
(1) that the “devaluation” of the word language via its “inflation” (which is an “inflation of the sign itself") “betrays” “ignorance”. Why might that be the case?
(2) that the concept of “writing” does not merely replace that of “language” but “is beginning to go beyond the extension of language”. Given that the inflation of “language” already amounts to an “absolute inflation” such that it includes the “totality of [a historico-metaphysical epoch’s] problematic horizon” (p.6), what does it mean to say that the concept of writing goes beyond such an inflation?
(3) that “signifier” and “sign” are part of only a series of terms (including “derivative”, “auxiliary”, “exterior”, “insubstantial”, “accidental”, “secondarity") which Derrida uses to recall the concept of writing; that Derrida suggests the comprehension of language by writing destroys the very “logic” of “the ‘sign’”, and he indicates a distance from that logic through the depiction of “signifier of the signifier” as describing the “movement” (not structure) of language and through a focus on the function (not meaning) of the signified within the “logic” of writing. Is it really the case that Derrida’s primary object here is to formulate a theory of language, (semio-linguistic) representation and reference, that his invocation of “the signifier of the signifier” is a proposition that “language always represents language”?
(4) that “writing” comes to encompass not only linguistic and graphic signs but also “action, movement, thought, reflection, consciousness, unconsciousness, experience, affectivity, etc.” and “the totality of what makes [inscription] possible”, “all that gives rise to an inscription in general”, where inscription is (roughly) equated to a process or practice or movement of “distribut[ing] in space”, and so “writing” also encompasses “military and political writing”, which is to say “the essence and content” of the “activities” of warfare and politics as such. Does the concept of “writing” that’s being invoked here seem to have much at all, by now, to do with signifiers and signifieds?
Every critic who reads Derrida’s Grammatology reads Chapter 2. Few critics seem to bother to read Chapter 1 and Chapter 3 (let alone Part II!), or if they do, they nevertheless see Chapter 2 as proposing a “theory of language and signification”. Consequently, they largely ignore the fact that the book is more a “critique” of Heidegger (see Chapter 1, esp. pp.18-24) than of Saussure, that it is in no way a theory of language but rather a genealogical tracing of the theme and the valuing of presence throughout the history of philosophy (where “presence” is shorthand for a whole chain of values including “origin” and “originality”, “interiority”, “essence”, “necessity”, “priority”, “nature”, “man”,