1. 

According to Derrida, “To write is to produce a mark that will constitute a sort of machine which is productive in turn, and which my future disappearance will not, in principle, hinder in its functioning, offering things and itself to be read and to be rewritten.”1 If there is something a little strange about the beginning of this definition—to draw is equally to produce a mark; why isn’t to write to produce, say, a sign? And if drawing might also be understood as producing a sign, why isn’t to write to produce a linguistic sign?—Derrida’s turn to the mark instead of the sign is already being motivated by the second half of the sentence: the description of the mark as offering itself to be rewritten in a way that the sign (its meaning fixed by the signifier’s particular relation to the signified) could not. The sign, if there were such a thing, would just mean what it meant; the mark never just means what it means.

But it’s not just the deployment of the mark that’s striking here, it’s also the relation between this description and some of the other accounts of writing he gives in both “Signature Event Context” (SEC) and Limited Inc. For example, a page later he describes the “sign” as “readable” (that’s part of its “functioning”) “even if I do not know what its alleged author-scriptor consciously intended to say at the moment he wrote it, i.e., abandoned it to its essential drift” (LI, 9). And this equation of producing a mark with abandoning it is anticipated a few pages earlier when he describes the writer’s “absence” “from the mark that he abandons, and which cuts itself off from him and continues to produce effects independently of his presence and of the present actuality of his intentions” (LI, 5). Finally, this description of writing as producing a mark and of producing a mark as abandoning a mark is reprised in the reference to Nietzsche’s “‘I forgot my umbrella,’” “abandoned like an island among [his] unpublished writings” (LI, 63).

The description of writing as abandoning is not, of course, a psychological one, but it does bring out the degree to which Derrida’s account of writing involves (maybe assumes) a theory of action or, at least, an account of the kind of action writing is. And that theory is made explicit in the simultaneous insistence on the importance of the writer’s “intentions” in producing the mark and on the fact that those intentions are compromised by its “functioning,” that is, by its producing “effects” that may have nothing to do with whatever the writer’s intentions are. The “category of intention,” Derrida says, cannot “disappear”; “it will have its place, but from that place it will no longer be able to govern the entire scene and system of utterance” (LI, 18).

Why can’t it disappear? “Let’s be serious” (LI, 65). Is Derrida being serious? “No criterion that is simply inherent in the manifest utterance is capable of distinguishing an utterance when it is serious from the same utterance when it is not. Solely intention can decide this …” (LI, 68).”2 So, all ironic utterances are ironic in virtue of being intended ironically, and all serious utterances are serious in virtue of being intended seriously. This for Derrida marks the inadequacy of “code”; on the one hand, no utterance can even count for him as an utterance (as “readable”) unless it deploys some set of syntactic and semantic rules (his argument for the importance of code); on the other hand, no set of syntactic and semantic rules can ever determine whether the utterance is serious. And since an utterance like “let’s be serious” must be either serious or not (or some complicated version of the two), intention will have its place.

Then why can’t it govern? This question is in a way easy to answer—governing is very different from (almost the opposite of) abandoning. But govern (commander) is in its own way as striking a description as abandon is since it’s not clear what exactly it would mean for someone’s intention to govern the scene of utterance. Or maybe what’s not clear is what exactly the scene of utterance is. For example, suppose someone takes my exhortation to “be serious” seriously when I meant it ironically. Insofar as what I was trying to do was communicate with this person, I have failed, and if this is what it means for my intention to be unable to govern the scene of utterance, it seems pretty clear that Derrida is right. I said something designed to produce something like a knowing skepticism; instead, it produced an earnest credulity, and perhaps the hearer who was supposed to be amused instead comes to feel embarrassed. In other words, the effect of my utterance was very different from the effect intended. Which is at least part of what Derrida means when he says on numerous occasions that for the mark to be able to function as a mark, it must be able to produce effects in the absence of the intentions of the author and thus unintended by the author, as, for example, when he says that Searle’s “evaluative decrees” “tend to produce determinate effects, often quite different from those apparently intended” (LI, 39).

But the claim that a text can produce effects utterly unintended by the author—or, to put the point in relation to governing—that the intention of the author cannot control the effect of his text on his readers—is utterly uncontroversial and is not really the point Derrida is making. Or, anyway, not only the point he is making. For the difference between the meaning of an utterance and its effect (at least the difference between the intended meaning of the utterance and its effect, a difference that, as we’ll see, is relevant whether or not the actual effect is the intended one) that underlies my previous paragraph is not one that Derrida accepts. Rather, he thinks that because the speaker’s intended meaning only counts as an utterance insofar as it is structured in “conformity to [a] code” and, since the code counts as a code only insofar as it is readable without any reference to (in the absolute absence of) the writer,3 the intended meaning is itself a kind of idealization that the utterance breaks with right from the start: “understanding the intended meaning” of the utterance is itself “problematical” (LI, 76). More strongly, the necessity of the utterance to be in “conformity to the code” is “incommensurate with the adequate understanding of intended meaning.” On this account, the meaning of the utterance and its effect cannot be separated. So, the apparently anodyne fact that the utterance can always have unintended effects gets its bite from the idea that meaning itself is a kind of effect which, precisely because the writer cannot control the effects, means the utterance can always have unintended meanings. It might make sense from this standpoint to talk about an author’s intended meaning as his preferred meaning—but even to produce the effect you sought to produce would not be the same as being understood. Thus, for Derrida, there’s a difficulty with the very idea of understanding the intended meaning, and that difficulty is related to the way that understanding (like misunderstanding) is itself a kind of effect; hence as Derrida puts it, his “substitution” of “intentional effect for intention” and of “‘mark’ for ‘sign,’” and, more generally, of “the functioning of the mark” for the “‘understanding’” of the “‘written utterance’” (LI, 61).

Here we begin to see why in Derrida the production of the mark is at the same time its abandonment. Writing is the production of a mark that the writer intends (let us say) seriously but that because he cannot control how the reader understands it (or, say, how others will cite it) may be taken to be unserious. It’s in this sense that the production of the mark is at the same time its abandonment. Because the mark is produced with the intention to communicate, the writer’s intention is fundamental. But because, to function even in the way in which it was intended, the mark must be able to produce effects that are unintended, the place of intention is always displaced. In other words, because it is relevant only insofar as it produces the cause of all those effects (and indeed could do so unintentionally), it can’t be said to fix the meaning of the utterance which, in Derrida, would be the same thing as controlling the effect. Thus, to produce a mark that you intend to produce one effect is at the same time to abandon the mark to the infinite number of different effects it might also produce.

But here we also begin to see the problems raised by writing as abandoning. Derrida indignantly denies Searle’s claim that he relies on “a notion of intention as something separable, intrinsic, and ‘behind’ the ‘expression’” (LI, 66). But his point that nothing “simply inherent” in an utterance can tell us whether that utterance is ironic or not, and that “Nothing can distinguish a serious or sincere promise from the same ‘promise’ that is nonserious or insincere except for the intention which informs and animates it” (LI, 68–69) complicates that denial. On the one hand, the idea that the intention animates the utterance supports his denial that it’s somehow behind it, separable from it. On the other hand, if the intention animates the utterance (if it’s not behind it but in it), how can the ironic and non-ironic utterance be the “same utterance” (LI, 68)? What’s the same is just “let’s be serious,” with the question of whether it’s serious or ironic determinable by an intention that is not inherent in and therefore must be somewhere outside the utterance.

It was the New Critical version of this scenario that, in a predeconstructive essay, Paul de Man criticized for misunderstanding the nature of intentionality.4 In “The Intentional Fallacy,” Wimsatt and Beardsley described the text as “detached from the author at birth” and going “about the world beyond his power to intend about it or control it.”5 Of course, deconstruction was not New Criticism, and “detached from” is a little more upbeat than “abandoned by.”6 But Wimsatt’s and Beardsley’s identification of intending with controlling and their denial of intention’s “power” to control does have a relation to Derrida’s claim that, properly understood, intention “will no longer be able to govern.” And, although de Man would himself become committed to the separability of the text from the speech act, what he criticizes in “Form and Intent in the American New Criticism” is precisely the way in which, changing “the literary act into a literary object,” the New Critics turn intention into “an outside force” (BI, 25).

Properly understood, he argues (which is to say, in order to be the kind of thing that demands to be understood), a poem must be recognized as “an intentional object” rather than a “natural object” (BI, 24). Natural objects (his example is a stone) have no “‘meaning’’” beyond “the totality of their sensory appearance,” and the scarequotes he puts around meaning are meant to suggest that they don’t really mean at all. By contrast, an intentional object (the example is a chair) seems “meaningless” only if you leave out the crucial (“constitutive”) “part”—that it is “destined to be sat on” (BI, 25). The mistake the New Critics make is to see “‘intent’”

by analogy with a physical model, as a transfer from a psychic or mental content that exists in the mind of the poet to the mind of a reader, somewhat as one would pour wine from a jar into a glass. A certain content has to be transferred elsewhere, and the energy necessary to effect the transfer has to come from an outside source called intention. (BI, 25)

In reality, however, the “concept of intentionality is neither physical nor psychological but structural … The structure of the chair is determined in all its components by the fact that it is destined to be sat on, but this structure in no way depends on the state of mind of the carpenter who is in the process of assembling its parts” (BI, 25).

The problem he sees here is the identification of the intention as a mental state and thus as something outside the object, which is the picture that Derrida explicitly repudiates but is nonetheless required by the description of the sincere and the ironic “let’s be serious” as the same utterance. Above we treated Derrida’s assertion that the question of whether the utterance was ironic could not be determined by anything “inherent” in it as a critique of the power of code, which it is. There is no sincerity convention, and if there were, it would be the gateway to its own ironization (That’s the point of “let’s be serious”).7 But now we can also see that Derrida’s idea of what is inherent in the text is what’s in it in virtue of its participation in the code. So, what’s not inherent (what the code can’t provide, the intention that makes the difference between serious and ironic) can only be where de Man criticizes the New Critics for putting it—outside.8

The point of inside and outside here is not to deny that the author and the text are separate; it’s to deny that the way to understand the intention is as the cause of the text. A thought experiment like the wave poem in “Against Theory” was meant to bring out the necessity of intention to meaning by removing it altogether, but Wimsatt and Beardsley make a version of the same point in their simultaneous acknowledgment that “a poem comes out of a head not a hat” and their insistence on the irrelevance of that fact. Indeed, following Jennifer Ashton’s revisionary reading of “The Intentional Fallacy” as the effort to argue above all the irrelevance of the cause of the text to its meaning, they could be understood as making the same argument as de Man but just failing to see that their target was not intention as such but intention as cause.9 For the point of de Man’s calling the chair an intentional object is not to acknowledge its causal history (that it was made by someone who intended something). That’s just “a poem comes out of a head not a hat.” De Man’s point (what he means by calling the intention structural) is that to understand what the chair is—to see it as a chair and not as an assemblage of wood and metal—is to understand what it was made to do. The intention is relevant not as the cause of the act but as its description, its meaning. Because the chair is indeed separate from the person who made it, we can, with Derrida, imagine a sense in which it’s abandoned. We could, for example, not worry about it being a chair and use it as a ladder, or a desk. But the carpenter hasn’t failed to make a chair because the chair gets used as a desk. The question of how we use it is different from the question of how we understand it.

But maybe it’s different for texts; de Man acknowledges that “the case of the work of literature is of course more complex” than the chair but maybe, according to Derrida, it’s different in kind. For, on Derrida’s account, the displacement of intention is internal to any text in a way that it isn’t to de Man’s chair. It’s perhaps true to say that the chair could only function as a chair if it could function without our understanding it as intended by the carpenter to be a chair, but it isn’t at all true to say that it could be identified as a chair without our understanding it as made to be a chair. With texts, however, Derrida argues that the very existence of the text as text is dependent upon the possibility of it being disconnected from its author. Indeed, what he describes as the “point of departure of Sec’s argumentation” (LI, 63) is “[t]he fact that even [in Searle’s words] if one breaks ‘with the strategy of understanding the sentence as the utterance of a man’ … who meant something by it … as one always can” then (in his own words) “the mark still does not cease functioning …. A minimum of legibility or intelligibility remains” (LI, 63).

The idea here is that if we break with the strategy of understanding “let’s be serious” as the utterance of someone who was either being or not being ironic (since both involve somebody meaning something) we are still left with something—the “minimum of legibility or intelligibility”—that enables the marks to keep functioning and, moreover, to keep functioning as the same utterance. It’s trivially true that the same marks can be used for different utterances—I say “let’s be serious” seriously and you quote me ironically—but that’s only because we’re not breaking at all with understanding it as the utterance of someone who meant something by it. What Derrida is interested in (more generally, what everyone committed to the idea that a text can mean something more than or other than what its author means by it) is the idea that “let’s be serious” is legible as the same text (same utterance) when it’s serious and when it’s not, that is, legible as the same text independent of its being used by someone to mean something. So, what makes it the same?

It can’t be that the signs “let’s be serious” are the same since the sign consists of signifier and signified, and the whole point is that they don’t have the same signifieds—one means something like “let’s not be serious.” This is why Derrida wants to move from the sign to the mark. And indeed, all the way through SEC/Limited Inc, it’s this repetition or identity of the mark (i.e., whatever makes the marks rather than the signs the same) that underwrites the idea that there’s a minimum of legibility belonging to the utterance irrespective of what someone meant by it, the minimum that survives no one meaning anything by it.10

But how is the mark (as opposed to the sign11) even minimally legible? The answer for Searle is simpler than for Derrida—for Searle, the mark in itself is a sign or at least a word. Indeed, for Searle, if the marks look like words in English, then they are words in English no matter how or by whom they were produced. Derrida’s claim is narrower—they’re “legible” as words in English. But this serves more to highlight the problem than to solve it. Suppose we invent a language where the words look like they do in English, call it Schmenglish, but have different meanings. The marks are now legible in two languages and, of course, we could invent more. So, what language are they in? At this point the force of insisting that they really are words in English or Schmenglish becomes unclear, since any marks will really be words in any language we might invent. In other words, if the language the writer was using (what the writer meant by the words she produced) doesn’t determine the language the text is in (which it can’t if we suspend the question of her intention), nothing does. This is what it means to imagine the act of writing as a kind of abandoning. To produce marks in English is to abandon them to the possibility of being treated as marks in Schmenglish.

On the one hand, the writer has her place—there’s no mark at all without her. On the other hand, for the mark to be able to function, it must be able to produce effects that are unintended by her. And because the intention is relevant only insofar as it plays a role in producing the mark that causes all those effects (intended or not), it can’t be said to fix the meaning of the utterance, which Derrida treats as controlling the effect. Hence even though Derrida invokes “legibility” and “intelligibility,” he is suspicious of “understanding,” which he wants to replace with “force.” What’s crucial for him here is the difference between understanding what someone has done and being affected (he sometimes calls it “touched”) by what someone has done. Even when they are the same, experiencing the intended effect is not the same as understanding the intended meaning, and what Derrida wants is not exactly to conflate meaning with effect but to insist that what we think of as the meaning of a text is better thought of as the effects it can produce. Those effects are what can’t be fully controlled by the writer. But it’s this redescription of writing as causing an effect that makes the intention into some kind of mental entity. Thus, although he understands very well that it’s a mistake to treat the intention as if it were outside or behind the utterance, his treating it as the cause of the utterance—the production of a mark rather than the writing of a text—nonetheless consigns it to the outside (to what’s not “inherent” in the utterance) and builds abandonment in. And thus—turning the meaning of the utterance into its more or less intended effect—he understands the inability to control the effect of the utterance as the inability to control its meaning.

2.

But it seems pretty clear that the failure to communicate what you mean can’t be identified with the failure to mean it, and we can perhaps get at both the problem with thinking of the intention as a form of control and with thinking of the failure to control as a problem about the intention by this route. Suppose I send a note asking two people to meet me at the bank in the afternoon. When I get there, one of them is waiting; the other never shows up. When I see him the next day, he says he did in fact go to the bank to meet me. But where I meant for him to meet me at the First National Bank, he understood me to be setting up a meeting by the side of the river.12 So, one of my readers understood my note; the other didn’t.

We can bracket the question of why the river-goer misunderstood me, especially since the first reader didn’t. The point is just that I successfully communicated with one reader and failed to communicate with the other (since one reader understood what I meant while the other didn’t) and that the simultaneous success and failure gives us two different descriptions of what I did—I wrote a text that meant meet me at the (First National) bank and I wrote a text that communicated that meaning (to at least one interpreter). The difference between these two descriptions is brought out by a third. I can also say of the same act that I wrote a text that meant meet me at the (First National Bank) and that I failed to communicate it to a reader.

The first point of the example is just to make clear this difference and the fact that in characterizing it we are completely reliant on a distinction between meaning and communicating, a distinction that we can then see would survive even if neither of my readers had understood me. Suppose they both knew I liked to go swimming on summer afternoons and both (very reasonably) took bank to be river. They were wrong. I meant First National; the meaning of what I wrote is just as unaltered by the fact that two people didn’t understand it as it was when only one didn’t. And suppose the text survives and becomes part of my published correspondence (“I forgot my umbrella”) and no one ever understands it. Just as its meaning was unaffected by the first misreader, it will be equally unaffected by all his successors.

The second point of this example is to raise the question of what consequences (if any) my failure to communicate should have for an understanding of my success in communicating. Derrida says that failure is a necessary possibility of every speech act. The question here is what the fact that someone can always misunderstand a text’s meaning tells us about our idea of what textual meaning is.

We can begin by asking why Derrida insists on this necessity—what’s at stake in denying that there could be a speech act or text that not only didn’t but couldn’t fail? The object of his critique is a certain ideal scenario of how meaning/communication works, an ideal that, whether or not it is ever achieved, is nonetheless at work, he thinks, in the accounts of communication he means to problematize. This “philosophical ‘ideal,’” he says, is “the presence to self of a total context, the transparency of intentions, the presence of meaning (vouloir-dire) to the absolutely singular uniqueness of a speech act” (LI, 17). It’s this ideal—of the transparency of the intention to the writer and to the reader—that Derrida denies, and he denies it not because it’s unrealistic or even impossible but because, were it somehow to be possible, there would be no such thing as either an utterance or a text. (Actually, there would be no such thing as meaning, since what the text means is treated here as if meaning could be independent of the text, but there can be no question of what a text means if there is no text.)

Rather, in order for my “written communication” to retain its function as writing, i.e., its readability,”—that is, for it to be readable in the first place—“it must remain readable despite the absolute disappearance of any receiver …. A writing that is not structurally readable—iterable—beyond the death of the addressee would not be writing. … Imagine a writing whose code would be so idiomatic as to be established and known, as secret cipher, by only two ‘subjects.’ Could we maintain that, following the death of the receiver, or even of both partners, the mark left by one of them is still writing? Yes, to the extent that, organized by a code… it is constituted in its identity as mark by its iterability, in the absence of … every empirically determined ‘subject.’ This implies that there is no such thing as a code—organon of iterability—which could be structurally secret” (LI, 7–8). And if the code implies “the radical absence of every empirically determined receiver in general,” it does the same, and “for the same reasons, for the sender or the producer.” So, to “write is to produce a mark that will constitute a sort of machine which is productive in turn, and which my future disappearance will not, in principle, hinder in its functioning … I ought to be able to say my disappearance, pure and simple, my nonpresence in general, for instance the nonpresence of my intention of saying something meaningful … of my wish to communicate” (LI, 8).13

How does the mark continue to function in the absence of both producer and receiver? Take Derrida’s case of the code known only to two subjects. They invented a code; one sent the other a message in the code; they’re both gone, and we have the text. The text can function as a text only because its participation in a code makes it at least potentially readable to us regardless of the fact that we cannot know what it was intended by its producer and understood by its recipient to mean or even if was intended to mean something. That is, the text follows some set of rules or conventions, and we can, at least in principle, figure out what those rules are without ever knowing anything about or even interesting ourselves in the text’s author. This is what’s meant by the nonpresence of my intention of saying something meaningful.

But how, if we’re trying to decipher the code, can what the author meant actually be irrelevant? How can it not be (just the opposite) completely dispositive? Suppose, just to keep things simple, that the code she invented was a Caesar shift—every letter in the text stood for the letter two places before it in the Latin alphabet. So, e.g., “o” means “m,” “g” is “e.” The text reads oggv og cv vjg dcpm; when we decipher it, what we get is “meet me at the bank.” Is the intention of the writer irrelevant to or superseded by this functioning of the code? After all, it’s the code, not the writer, that requires those transformations. But the code only requires those transformations because the writer is using the code—that is, trying to decipher it just is trying to figure out what code the writer was using. The fact that it’s in a code makes this kind of procedure possible—we’re not being asked to figure out from a bunch of marks on a page what the producer of those marks was thinking when she made them (we’re not being asked to approximate the “ideal” scene of communication); we’re being asked to read what she wrote. But even though the fact that it’s in a code detaches it from what the producer was thinking, that fact in no way detaches it from what she was doing since although she may be dead and gone, the question of what code she was using is as relevant as it would be if she were alive and waiting for us at the bank.

The further point here would be that the code only functions with reference to a language: e.g., dcpm encrypted with a two-letter shift is “bank,” with a 24-letter shift it’s “fero,” which, in Esperanto, means “iron.” If our text had just this one word, deciphering would not just involve figuring what code the writer was using but whether she was speaking English or Esperanto. If we think of the language as also a kind of code, then maybe we can make more sense out of the claim that because the message is (and must be) in a code (that is, written according to a structurally readable set of semantic and syntactic rules), it is right from the start disconnected from the writer’s intentions and necessarily able to function in the radical absence of those intentions? After all, unlike the Caesar cipher, which the writer has constructed, the language she’s speaking is English, and she didn’t construct that. English has a certain autonomy from the speaker; when she writes “meet me at the bank,” those marks already have meanings not determined by her in a code not invented by her.

But why should the fact that bank has several meanings in English make the sentence “meet me at the bank” function in the absence of anybody meaning anything by it, in the radical absence of the author’s intention? Supposedly, it’s because the “sign” (in this case “bank”) “possesses the characteristic of being readable even if the moment of its production is irrevocably lost and even if I do not know what its alleged author-scriptor consciously intended to say at the moment he wrote it, i.e. abandoned it to its essential drift” (LI, 9). But even to think of it as a sign—or, given Derrida’s suspicion of the sign—to think of it as an iterable mark (i.e., one that’s organized according to a code)—is to refuse the idea that the moment of its production is irrevocably lost. Just the opposite; any account of what it is is at the same time an account of what it was meant to be. Because insofar as a code is a set of rules (x means y, which is to say, when I say x, I mean y), to treat the mark as belonging to a code is to treat it as having been produced by a writer who was following those rules. If you think of the mark as coded, not only do you not think of the moment of its production as irrevocably lost, you are already producing at least a partial account of that moment. The supposedly aporetic relation between intention and code is undone; any account of the code is already the account of an action.14

Here we again see the entailments of writing as abandoning. Derrida treats the writer of the sign the way Donald Davidson famously treats (among many examples) the driver braking her car: what she does is not brake the car but press her foot on the pedal—whether the car actually slows down is not up to her. Her action functions as the cause (she hopes) of the car slowing down. Which is the way the intention matters for Derrida. That’s why it makes sense for him to insist that “the category of intention will not disappear; it will have its place, but … it will no longer be able to govern the entire scene and system of utterance” (LI, 18). Davidson puts his version of the point more epigrammatically: “We never do more than move our bodies: the rest is up to nature.”15 Maybe we could say for Derrida, all we ever do is produce/abandon the mark; the rest is up to the essential drift.

Which is a not implausible way to think about writing if we’re thinking of it as nothing but a form of communication: I tell you to meet me at the bank, and I hope you understand what I’ve meant. But, as we’ve already seen, it’s not a very plausible way to think about writing if we’re interested in the question of what the text means rather than of what it does or doesn’t communicate. Because the question of whether I’ve succeeded in meaning something has nothing to do with the question of whether you or anyone else has succeeded in understanding it. Or, to put the point the other way around, the fact that it must be possible for you to understand it (that I must have actually produced a speech act and done so following some set of practices that you could in principle figure out—no structural secrecy) has nothing to do with the question of whether you actually do understand it.

I write “I am going to the bank” with the intention to get my readers to believe I will be at the bank by getting them to recognize my intention to tell them I will be at the bank. The self-reflexivity is crucial because there are other ways they might come to believe I was going to the bank (I might be packing my bathing suit), but the speech act does not just involve their getting information about where I’ll be; it gives them information by getting them to recognize that I am giving them this information. How might this fail? It might fail because I don’t succeed in getting them to believe what I told them. It hasn’t produced what Austin called the relevant perlocutionary effect. We could say here that I have successfully communicated with my readers but have failed to get the desired uptake—their belief in, say, my sincerity.

Or it might fail for what would be a more basic reason—they don’t understand what I’m saying. They think I’m telling them I’m going to the riverbank (presumably for a swim, a reasonable interpretation, given the bathing suit) but in fact the bank I mean is the First National. So, here too I have failed to produce the effect I was trying to produce—but in this case, the failure is not perlocutionary but illocutionary. It’s not that they don’t believe what I told them. It’s that they don’t understand what I told them. So, if the first failure is a failure to get them to believe me, the second is a failure to get them to understand me, a failure to communicate.

But is it a failure to mean? Here is where the original scenario—one reader understands, the other doesn’t—has its relevance, since its point is that the failure to communicate is a failure to get people to understand what I mean, not a failure to mean. And that the failure to produce the intended uptake (or, put differently, the fact that an utterly unintended uptake was produced instead) only counts as a failure because the meaning of my utterance is autonomous with respect to uptake of any kind. Indeed, the failures of uptake only count as failures because we take the meaning of my utterance to be whatever I meant by it, and we take me successfully to have meant whatever I meant.16

Failures to communicate—failures to produce the intended effect—are not as such failures to mean. Producing unintended effects is not producing unintended meanings. This is just one way of getting out what I’ve been trying to suggest is at the heart of Derrida’s argument in SEC/Limited Inc—his commitment to describing the act of writing on the model of event causality and his consequent insistence on the open-endedness built into the drift of cause and effect. It’s because he himself sees quite clearly what follows from that model that he is skeptical of the idea of understanding an utterance: he sees what we call understanding as just one possible effect among others. Which makes the normativity of understanding something, in the end, arbitrary, and which makes sense of his memorable account of the “arbitrariness” involved in his decision (“contingent, artificial, and external”) to divide the text of Limited Inc abc into the twenty-six sections determined by the letters of the alphabet. Once we abandon ourselves to the infinite drift of cause and effect, he argues, wherever we stop must be in some sense arbitrary: “Must the surface of the paper, the contents of the time at our disposal, etc. all be integrated into our calculations? If so, what about the ink remaining in my typewriter ribbon? And yet: why not? That is the question” (LI, 45).

The paper, the time, the ink are all (in relation to the project of writing) even more arbitrary and contingent than the decision to go by the alphabet, which is at least a decision. They’re just the conditions under which he happened to write. And, of course, they are meant to raise the question of why, once one’s started down that path, one should stop there. Reading a written utterance, why not take into account the quality of the paper as well as its surface, the spaces between written lines, or the font? And, of course, we can easily imagine the ways in which, say, the use of a hard-to-read font like this one might be something we did indeed need to take into account. But the ease of the example just makes obvious the answer to Derrida’s question, “why not?” The font matters when it’s supposed to matter. A material description of any act (what Davidson calls “nature,” what Anscombe calls “physics”) can go on for a very long time. But it’s precisely because the very identification of an act as an act makes some of those material conditions matter differently from others that the question of why you stop where you stop (as opposed to where to stop) is easy. If the only font I have is Blackmoor LET, the fact that I’m using it is a lot less likely to count as something we need to take into account just because we’re a lot less likely to believe that we’re meant to take it into account. Of course, we might be. The point here is not that refusing a causal theory of action helps us either write or read better; it’s that it explains why writing and reading are what they are. And that to refuse the normativity built into “is supposed to” is to redescribe writing as abandoning and disconnect reading from understanding.

This really is the point made by Derrida himself in his more extended discussion of the abandoned “I have forgotten my umbrella” in Spurs: Nietzsche’s Styles, a discussion all of which is pitched in opposition to the “hermeneut,” committed to the idea that “I have lost my umbrella” “must mean something” and searching for the “true sense of the text” (S, 131, 107). By contrast, under the “regime of quotation marks” (the marks that break with any given context), “[r]eading is freed from the horizon of the meaning or truth of being” (S, 107). And, as we began by noting, this isn’t just an epistemological point. The “hermeneut ontologist,” committed to the idea that “I have lost my umbrella” “must mean something,” searches for the “true sense of the text” (S, 131, 107). Because “a thousand possibilities remain open,” he can’t find it. It might, for example, be written in “some more or less secret code,” less because no code can be structurally secret, more because when the author dies, it will nonetheless be “undecipherable.” But this only “withdraw[s] it from any assured horizon of a hermeneutic question” (S, 127). That is, for the “hermeneut ontologist,” it poses an epistemological problem (how will we ever know?) but leaves the ontological commitment (to a “true sense of the text”) intact.

The real critique of the hermeneut is not that he can’t find the one “true sense” but that he’s mistaken to think there is one. Indeed, there aren’t even several. Rather than making the pluralist point that there can be many understandings of the sentence (that’s the polysemy he rejects for dissemination), Derrida is arguing instead that, strictly speaking, there can be no understanding of it, that whatever the reader makes of the text will not count as an understanding of it.17 Whether it’s the writer’s effort to “control” the meaning or the reader’s effort to “impose” it, the “limit [of] the will to mean” is built into meaning itself (S, 133).

The problem, as we’ve already begun to see, is that once you’re committed to the idea that meaning is something you try to control and/or that it’s something you impose, you’re committed also to an account of the text in which the intention must be (what Derrida rightly denies it can be) outside of it. Indeed, this is exactly what the reduction of the text to the mark requires since the act of writing is here conceived as the writer’s effort to impose a meaning on the mark she has made and since the act of reading involves the impositions of other meanings, which need not be the same as the writer’s. The reason the mark is abandoned is because that imposition necessarily fails, leaving a remainder that remains necessarily “open.”

But it’s not like the act of writing can be plausibly imagined as making some marks and somehow adding to them the meaning you supposedly have in your head. Or that reading can be understood as looking at some marks and adding the meaning to them. More fundamentally, there is no such thing as the imposition of meaning. Rather than impose a meaning on the marks she produces, the writer means something by them. One way to put this is to say she produces not marks but sentences. And what the reader either understands or misunderstands is those sentences. Indeed, if the differences between readers could best be described in terms of their imposing different meanings on the same marks, the very possibility of misunderstanding (the failure of communication that Derrida rightly insists is crucial to any account of its success) would actually be an impossibility. Which is a point that Derrida himself makes when, urging us not “to dissociate questions of ‘power relations’ or of ‘rhetorical coercion’ from questions of the determinacy or indeterminacy of ‘meaning,’” he insists that “[w]ithout play in and among these questions, there would be no space for conflicts of force,” and that the “imposition of a meaning supposes a certain play or latitude in its determination” (LI, 144–45). The scare quotes around meaning and the identification of its “imposition” with “conflicts of force” suggest the degree to which questions about what the text means have here been replaced by questions about what we make it mean and therefore the degree to which the normativity intrinsic to disagreements in interpretation has been replaced by the questions of force intrinsic to competing impositions. Indeed, this is precisely what Derrida admired in Austin, what he took to be Austin’s freeing his analysis from the “authority of the truth value” and “the true/false opposition” and substituting for it “the value of force, of difference of force” (LI, 13).

The problem here is not that Derrida has a cynical or nihilist account of interpretation, in which force replaces reason. It’s that, in the effort to put intention in its place (as the cause of the text rather than as its purpose and as the hope for some particular effect rather than as an explanation), he has had to produce an account of reading and writing in which there’s no place for interpretation—in which, because there’s no “true/false opposition,” there can be no such thing as understanding or misunderstanding. This is because (when it comes to action) understanding and misunderstanding go with intention—the intention is what’s understood or misunderstood. The fact that every speech act can be misunderstood is not a function of the limit of the speaker’s intention (what it can’t control) but of its hegemony—if the text didn’t mean what the author intended, it could never be misunderstood. It’s one thing not to feel the force of an act; it’s another not to understand its meaning.

“What if” Nietzsche “meant to say nothing,” Derrida asks, “or at least not much of anything, or anything whatever?” (S, 125–27). What if he “was only pretending to say something?” Or what if he wrote in “some more or less secret code?” Or what if the sentence, which is after all, in quotation marks, “is not his [n’est pas de part en part ‘de lui,’ comme on dit]”? “We will never know,” Derrida says, or, “At least it is possible that we will never know” (S, 127). But the relevant point here is not that we may never know (which of course is true); it’s what we may never know—what Nietzsche was doing. No account of what “I have forgotten my umbrella” means (including meaning nothing at all) is anything other than an account of what Nietzsche was doing: pretending, quoting, misquoting, etc. And even if, as Derrida suggests, we take the quotation marks as evidence of the way the lost umbrella’s “readability” makes it available to “expropriate,” the difference between what it means outside quotes and what it means inside quotes is just the difference between saying something and quoting someone saying something. Expropriating is also an act.