Interchange II: Closing Remarks

A) Walter Benn Michaels (WM) offers a closing response.

B) An exchange between Henry Staten (HS), Joshua Kates (JK), and Walter Benn Michaels that picks up an earlier part of their discussion.

C) Kates’s discursive response to Michaels’s response.

D) Michaels, in closing, answers both Kates and Staten.

A) Walter Benn Michaels (who begins by quoting Derrida from “Signature Event Context” [SEC])

“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 … when I say ‘my future disappearance’ … 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 ….”1

So, my piece argues that this cannot be true. One way that Steve [Knapp] and I have argued that it (or similar claims) can’t be true is an argument from identity—the idea that a text can have many (or any) meanings that are not the intended one requires some account of what the identity of the text consists in (such that, having many meanings it still counts as the same text). The point about the sign here is that if you think of the sign as signifier/signified, then (to take a classic example) “vegetable” understood to mean “growing” and understood to mean “vegetable” would be two different signs, not two different meaning of the same sign or two different meanings of the same text. Which is, in my view, a reason Derrida wants to get rid of the whole apparatus of the sign. So, the identity of the text (if you wanted to claim the same text had these different meanings) would consist in identity of signifier, not identity of signs. But while it’s obviously true that the same signifiers can have lots of different signifieds, it’s not obviously true that identity of signifier should count as identity of text (it’s closer to obviously false—see Pierre Menard).

But in this paper, I wanted to come at the problem through the account of the production of the text that is required to imagine the author as producing marks, not signs—the basic idea being that imagining the writer as producing a mark that can mean different things (produced and abandoned) instead of understanding the writer as meaning something (i.e., of imagining that the mark is the object of interpretation instead of the writer’s act being the object of interpretation) commits you to exactly the theory of action in which the intention is located outside the act. The problem with this is not exactly that it’s “mentalist” (in Henry’s terms) but that it treats the act on the model of event causality. Which is a mistake not because the act is removed from the world of cause and effect but because the intention cannot (or cannot only) be understood as the cause of the act but as its description. And the obvious way to get at the difference is by noting that every event has lots of causes, but every intention has only one (correct) description. So, it could be right to say that the heat caused Henry to open the window and maybe that his restlessness also caused him to open the window, but the correct description of what he did was he opened the window.

In the various things you guys have written, I don’t see a straightforward defense of the original Derridean position on “the nonpresence of my intention,” although what I take to be Henry’s defense of the idea that intention (when you identify it with “consciousness?”) plays some role, but not a dispositive one, seems closest. And I think Henry ends up committing himself to exactly the same view of intention as something outside the act that Derrida does—this is the mentalism that you think plays some role but a much diminished one. But (pace Josh’s nervousness about how to read Anscombe on mistakes), the relevance of intention here is built into any description of any act. It’s not like it matters what’s going on inside your head when you do anything—that is, there’s not some separate mental event going on when you open the window or when you say I’m opening the window. But it’s impossible to describe what you’re doing without the reference to what you’re intending to do. As you can see right away in the case of failure—I was trying to open the window is a description of what your intention was. So, if intention requires something mental, it’s a mentalism that no one can possibly avoid and that is not even sidestepped for a second by, say, not using the word “meaning.” Without the account of your intention, all you could say about a failed act of opening the window is that Henry pushed on the window for a while and then stopped. So, yeah, the possibility of failure is important, but precisely (as I say in the paper) because it makes clear the hegemony of intention. And a word like meaning only has the mentalist problems Henry associates with it if you’re committed to the idea of the intention as disarticulable from the text (which, of course, I argue that Derrida is and, indeed, is what he says in the quote at the head of this note and in the very idea of the abandoned text).

Josh—even more than Henry—seems to think that a more complicated account of intentionality will alter the basic Against Theory (AT) claim—the text means only and always what its author meant by it. But the uncontroversial fact that virtually everything we do (and therefore virtually everything we can possibly call an intention) is only possible because of a million things that exists independent of our intention is just the condition of possibility of our acting—I couldn’t open the window unless there were windows and a whole set of beliefs and desires related to windows, etc., but all that doesn’t complicate my opening a window—it just helps to explain what opening a window is (and therefore what I am doing). Same thing would be true for knowing, what, say a “novel” is—insofar as I understand myself as writing one, it will help in understanding what I’m doing. So, yes, genre is relevant if the author is trying to write in one. The invocation of “context” in general seems to me otiose since everything you want from context is built into our idea of action in the first place. And, according to me, the same goes for rules, etc. The question of the kind of relevance rules have is only a question about whether someone is trying to follow the rules. If you’re trying to use a code and you screw up, the code is as important to understanding what you were doing as the window you were trying to raise is. Indeed, code only enters my discussion at all because Derrida assigns to it a crucial role in the original and necessary break with intention, the readability that survives the nonpresence of my intention. The way to put this in relation to Derrida would be that the aporetic relation between code and context isn’t really aporetic, but also that context adds nothing to the general claim that the object of our interest is what the speaker is doing and that the relevance of the rules is also entirely a function of the speaker’s following (or not) those rules.

B) Henry Staten

HS: I wasn’t imputing a theory of consciousness to him [WM], I was arguing that he needs to explain how he does without such a theory; or, alternately, in your helpful formulation, he needs to confront the tautologous nature of his claim.2 The tautology, as I read it, is “the meaning of ‘meaning’ is simply what the utterer means.” But I’m arguing that positing this tautology requires an explanation of how this can be so. I think the vast majority of people would agree with this tautology, but if you asked them to explain why this was so, they would resort to mentalist talk (if they were articulate enough to even be able to explain). So, in a statistical sense, Walter could argue that this is indeed the meaning of “meaning.” But Wittgenstein’s argument that meaning is use isn’t about current statistics; it’s about how the concept of meaning had to have evolved, and what must be the case for there to be language: it’s about how classical mentalism is simply wrong.

WM (now commenting on an example HS had earlier offered): But I do think you’re not feeling the force of your own example. If someone were to say to your happy guy, “actually, I happen to know the late whoever was just screwing with you and actually meant the opposite of what he said,” the happy guy would spend his life worrying about it—that’s all the theory of “mental activity” anyone needs for this discussion.

HS: I assume you mean that all you have to do is say “he meant” and that constitutes all the theory we need. This is a perfect example of how you beg the question of “the meaning of ‘meaning.’” All you have to do is use the word, and the magic act is fulfilled: meaning originating in the utterer, somehow that doesn’t need to be explained, is transcendentally glued to the utterance. You’re like the magician who has his assistant stand on a chair, and then removes the chair and leaves the assistant floating in mid-air. (If we carried this discussion much further we’d be brought up against the Derridean question of the “pristine origin.”) Transcendentality means: no merely contingent factor, nothing empirical, no chance can befall THIS. That’s also what an “essence” is: a something/nothing that always is what it is, regardless of context. You want all the advantages of metaphysics while shunning the association.

WM: If you were to say to happy (now less happy) guy, “Dude, it doesn’t matter what was going on inside his head, what matters is that the scene of whatever makes what he said count as sincere (the sincerity being just as ‘mental’ as the lack of it) praise,” he would think you had entirely missed the point.

HS: Of course! Because what matters here is indeed what went on inside his head. That happens sometimes; when we’re talking about sincerity of course the crucial bit is what went on inside his head. I have no problem with that; I’m entirely comfortable with the notion of meaning-in-the head, and I know just how important that is. What I deny is that meaning-in-the-head can perform the transcendental role that you assign it because meaning-in-the-head is one thing, and meaning in the actualized sign is another, and the MEANING of “meaning” is derived from words in the world, not from words in the head; as Volosinov says, word is two-sided and derives all its form from outside. You want to collapse the empirical phenomenon of language and language use into the mere functioning of cause and effect and reserve the phenomenon of actual meaning for the mental part, which is a very respectable point of view with an ancient lineage, but you want to do without all of the metaphysical baggage that that point of view brings with it. Here, however, it sounds like you are saying that what matters is what went on inside is head, but you don’t want to explicitly commit to saying that.

WM: Which you would have. The question “was he complimenting me or not” does not require any special theory of mental activity, but it’s not about anything other than what the speaker meant.

HS: Here’s the collapse. No need to tie it to mental activity because “what the speaker meant” just means THIS (the infrangible what-you-mean-is-what-you-mean, and no need to worry about how this can be, or how it articulates with all the philosophical debates about meaning). Which I admire as a bold and noble enterprise that has held my interest for a long time—comparable to Wittgenstein’s positing “meaning is use” in a way that simply sidestepped classical theory of mind. But that’s a big bite to chew off.

C) Joshua Kates

Walter, your latest account of where you’re at gives a powerful restatement of your own position in respect to what you call “the identity of the text” and the role that intention must play in it. Your arguments go awry, however, in your discussion of the two authors you treat: Derrida and Anscombe—branding Derrida’s position wrong and Anscombe’s right. Neither, however, say what you say they say and these incongruities, in turn, let be identified the central problems with your own stance.

1) Derrida

Your criticism of Derrida culminates in the provocative portmanteau claim: a) that Derrida locates “intention … outside the act” of making a mark; and b) that Derrida’s so placing of intention outside the act gives a causal efficacy (like billiard balls) to the marks themselves. Neither of these can be right.

The first amounts to the kind of linguisticism all too common in poor interpretations of Derrida over the years. The principal problem any interpretation of the mark and writing faces is how they relate, not just to language but to its employment, to a message as well as to a code. Derrida’s notion of writing (and also the mark), whatever else it is, refers to a discursive instance, to language put to use; it is something done, something that happens, not a stand-alone structure or rule or so on. Many early readers of Derrida overlooked this and took Derrida’s writing basically as equivalent to language on its own—something like a signifier, or a supersignifier.

This same error appears in your submission, Walter, when you approach the mark specifically by way of code. You say of Derrida that “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, 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.’”

You here assert that Derrida denies “intended meaning” any weight since the supervening part played by code negates the role of intention “right from the start.” But were this Derrida’s conception, why talk of a mark or writing at all? The code, you say, makes the mark subject to multiple interpretations; it governs the text immediately and breaks with speakers (and presumably receivers), so what then is Derrida’s specific aim in talking of intentions and authors and their marks and inscriptions, even if the former two are somehow abandoned? Similarly, why—not to mention how—go on to discuss context and eventually signature?

Of course, your idea, the one underlying your title, Walter, is that, on Derrida’s scheme, an intention and a producer are needed in order to turn over the engine, as it were, to generate the marks or inscriptions in the first place (hence produced and abandoned). But what causes all the ruckus, what governs, on this scenario is not the mark, writing, or the inscription, but, again, the code, which amounts to saying that language gives the text its identity and everything else falls by the wayside.

Writing, the mark, however, are different from a code or even its instantiation. (Derrida later in SEC calls code their “condition of possibility and impossibility.”) Not the intentionless mark, some piece of code or language, operates (potentially in absolute absence), but rather what has been itself intentionally invested—hence Derrida’s continued talk of writing, the mark, and of the “written syntagma.” (Note in the omitted portions of the passage from SEC that you quote that Derrida always speaks of the “present” intention as what is missing—including cases of inattention—not the intention simpliciter, as he used to like to say.) The point of the mark (and of early deconstruction more generally) is that the intentional field, as Derrida was sometimes wont to put this, ultimately exceeds any individual’s intentional act and even all consciousness. A certain transcendental writing leads to a rethinking and a transumption of intention as a field—thanks to the former’s alliance with absolute otherness (“ … the ‘effect’ of transcendentality is linked necessarily to the possibility of writing” [SEC 8]). Here and in all his early work, Derrida thus combines the force of speech with the force of language, in part to account for the peculiar ontology of language (which requires a sort of transcendental difference, the rethinking of which writing also allows), none of which is captured in your talk of the mark.

Speaking of marks having a causal force, as Henry powerfully points out in some of his contributions, is, then, wholly out of line with, yes, Derrida’s intentions. The latter only holds if the mark and writing have been emptied of any act-character whatsoever, again raising the question of why talk about the mark or writing at all, not to mention contexts, events, and signatures. The text so understood becomes entirely foreign to any instance of genuine understanding, rather than being what includes and accounts for such understanding (as well as being something more, a certain excess)—Derrida wanting both.

You want to establish the identity of the text, Walter, because without such an identity, you fear whether an interpretation is right or wrong will be impossible, meaningless. Given the foregoing, it is far from clear, however, that marks as conceived by Derrida do comprise a text’s identity in your sense. This would be not because the text has no identity but because Derrida frames it as having both an identity and something other than an identity, and the latter—not because intention is taken out of it, but because intention is taken out of itself—is different from what it is usually thought to be. That the text should be both identical and not may be foolish—it may lead too easily to the notion that a text is mere marks about which one can say anything (as Henry also points out); it may even be impossible, but these are different judgements than yours that marks (based on a code) for Derrida simply are the text’s identity and that they function causally.

2) Anscombe

Once more, let’s ask along with you Walter, in what, then, does the identity of a text consist? What is a text? The three of us—you, Henry, and I—all agree a text is not a signifier, a sign, a piece of language or code. Is it thing-like at all? Perhaps a tool (or conglomeration of tools) of some kind informed by a techne handed down across generations, as Henry, I believe, suggests?

I say no: it is not a thing of any sort. You, Walter, agree, I think: the ontology of text for you consists in it being an intention. In your work with Knapp and for a while after, your recourse to intention, however, merely pushed the problem back a step; you and Knapp systematically avoided the question posed by intention’s own ontology. More recently, you have turned to a theory of action to account for intention’s ontology. The ontology of action now remains unplumbed, and as is evident in your response (see below), you reify action and hence reify the non-reifed text under new terms.

Of course, you specifically appeal to Anscombe for your theory of action. But just as with Derrida, so with Anscombe, you frame her teaching on terms counter to her own account.

More specifically, you say that there can only be one answer and thus one intention, on which basis you wish to affirm the self-same text: “every intention,” you write “has only one (correct) description.” For Anscombe however, actions are events under a description, a description that indeed involves why what happens happens, which is where the intention part comes in. This does not, however, prevent there being multiple descriptions of a given action, multiple intentions on the part of those doing them, nor on the part of those describing, which are not always or even usually the ones doing the actions.

Anscombe indeed allows that there can be multiple descriptions of someone doing what is otherwise the same thing. She speaks of a person crossing a road and says that most of what people do falls under the intentional, just by dint of being identified or described, by us—“to call an action intentional is to say it is intentional under some description that we give (or could give) of it.”3 Such intentionality clearly allows for multiple imputations and multiple descriptions of someone’s act: he is crossing the road; he is going to the store; he is avoiding his neighbor; and so on.

This conflicts not with your claim, Walter, that it is impossible to describe what someone is doing without speaking of their intention, without granting them first person authority—the agent in question may say I wasn’t trying to avoid so-and-so, and that description might be discredited, though, as Henry points out, this could also lead to a longer conversation if the person nevertheless crossed the road repeatedly when so-and-so came into view; intention is not the exclusive provenance of the agent.

Anscombe’s own account denies that the author/speaker always and necessarily has a privileged role in offering a description of their action; it also and equally denies that there is always a single such description. In respect to her most central and discussed example, Anscombe indeed has something more complex to say than you about what could be called the univocity of the act.

Anscombe’s most famous example is that of a man pumping poisoned water into a house to kill its (evil) inhabitants at the behest and pay of someone else. In discussing it, she offers a hierarchy of descriptions, acts, and thus intentions, that integrates some, and not other, acts into a single act, while all are otherwise intentional: pumping the water, arranging the hose, beating out a rhythm while doing so, etc. Many intentions and acts can thus describe a single event, here one which in part takes place simultaneously and also over time (the poisoning).

Taking Anscombe at her word, then, how many more intentional acts and possible descriptions of same, Walter, would be furnished by a novel or a poem, which unfolds over time both in being written and being read by multiple readers on multiple occasions? Literature just is more complicated than opening a window. And though some such acts as described may be folded into an overarching one—Shakespeare gives us Edmund to contrast with Edgar—it is unlikely all are, and this is especially the case, since, again, these descriptions (of the author’s intentional acts, of their text) are ones, though not exclusively, that “we could offer.”

Your main pushback against me falls on this possibility of multiple readings, each answering, to various degrees, to what an author did and meant. It centers on context, which for you further takes in (linguistic) rules and conventions, all being subordinated to a univocal intention. For you, rules, code, conventions, language are all built into the action and only relevant insofar as this is the case. (“The question of the kind of relevance rules have,” you write, “is only a question about whether someone is trying to follow the rules. If you’re trying to use a code and you screw up, the code is as important to understanding what you were doing as the window you were trying to raise is.”)

You let yourself too easily off the hook here, however. I have long thought and still think you have a real problem in the manner in which you take into account codes, conventions, and languages in AT and in all your subsequent writings. Unlike my own view, which denies them any relevance at all, since not just texts but all discourse transpires entirely as an event, you acknowledge codes, conventions, and rules—here as a “condition of possibility” alongside “a million other things.”

Your position in respect to language and code so qualified, however, is precisely the reverse of the one that you impute to Derrida: for you codes and the such-like only count insofar as they are governed, not abandoned by the intentional act of the penperson (the writer/author); for Derrida, you claim, codes and conventions are the broader category that supersedes the intentional act and survives its abandonment. Either scenario, however, makes no difference, and you yourself necessarily fail into the same problem that you accuse Derrida of succumbing to.

Granting your stipulation about intention’s governing, the problem remains of what to say about code, language, rules themselves—indeed of how to speak about them at all. Even as a condition of possibility, you must grant that conventions, rules, codes exist and are identifiable in their own right. Yet this flies in the face of what you have long denied: namely that someone could treat a text simply as a piece of language, an instance of code or convention. Though you would insist that they are not interpreting if they do this—they only interpret when they interpret what the author intended by using the conventions, codes, and so on—as long as there are codes and so forth, readers could take the texts to be governed by the codes or language and dispense with the intention (even if this may mean choosing among languages). But if, to this extent, interpreting the text as language or code is even merely possible, intention must be outside, or at least not inside, the text.

Recall in AT signifiers in the wave poem only resembled signifiers. Should not, then, instances of code, convention, or language, divorced from their intention not be code, convention, or language at all—only resemble same? But then how to talk about them in their own right or grant them any currency, including as a condition of possibility?

In my submission, I unfold further how to eliminate any role for languages or code and so on when conceiving texts and discourse more generally, a position which baldly stated appears counter-intuitive. (In other recent essays, I also suggest that the manner in which something like the force of language and its signifiers can be registered is in the always necessarily retroactive fashion sketched by Shannon’s information entropy.)

My account requires context; it makes the sentence and larger wholes the units of expression, and it requires context in both other things said, other instances of discourse, but also in the world in its familiarity, as well as its unfamiliarity. The idea is that discourse, writing as well as speech, are of a sort that they can never stand alone (are not things of any sort, not even ones standing in some identifiable techne) but are always coeval with something other than themselves, the worldly affairs, circumstances, aims that they are about and express. For this reason, discourse is not reifiable, not objectifiable as a thing or system or structure—only in more speech, in more writing with their own contexts, in a citationality without reserve, as I suggest in my submission’s second part, can it be charted.

Only in this way can a text be an action, if it is an event in this sense. Anscombe’s own burden when it comes to action (generally) in fact is showing how it can be at once an event (what happens, in her case involving causal relations, which I do not believe pertain directly to discourse) and also something I do, something subject to a description involving an answer to the question why, and thus intentional. You downplay this aspect of Anscombe’s thinking, including how for her, actions involve contexts, causes, and effects in order to adapt her theory to texts; you also dispense, of course, with the whole business of it being (often) we who offer descriptions that identify agents’ intentions. So proceeding, as Henry also suggests, you again make the act absolute, effectively reifying and mystifying the (almost wholly) non-reifiable. There is, however, no need when speaking of texts to pushback against what some, like Derrida, at least at moments, deem the absolutely ungovernable by insisting that intention can never be abandoned. The governed and the ungoverned, the abandoned and not, instead, are always at work together every time that there is discourse.

D) Walter Benn Michaels’s Closing Response

There are lots of questions raised in these closing comments but several seem to me to stand out: what kind of difference (if any) is made when the action in question is, say, writing a novel rather than opening a window; what the contribution of the reader is; what the relation between code and intention is and, as Henry vividly puts it, whether anyone can say “what you mean is what you mean” without producing an account of what they mean by meaning.

Josh makes the “Literature just is more complicated than opening a window” argument by adducing Anscombe’s example of the man pumping water where, in contrast to our man opening the window, there are at least four descriptions (moving his arm, operating the pump, replenishing the water supply, poisoning the inhabitants) of what he’s doing. But Anscombe’s point is not that there are four descriptions instead of just one, it’s that that “we can speak equally well of four corresponding intentions, or of one intention” because the first three are “swallowed up” by the fourth. So, to understand the first three is to understand them as the “means” to the fourth, the “end.” If you don’t understand the end, you haven’t understood the action. And insofar as there are also descriptions that don’t form part of the “series” of means toward that end—say, the man is “beating out the rhythm of God Save the Queen in the clicking of the pump”—the correct description of that is equally an account of what he means to be doing.

So, if we take this multiplicity as a model for writing a novel or a poem, we haven’t for a moment produced anything that takes us away from the intentionalist claim that understanding complicated and even conflicting acts is nothing but understanding (the complicated and conflicting) things the agent is doing. And we haven’t moved one step in the direction of deconstruction’s “essential drift.” To go there, we would have to trade in means and ends for causes and effects, which would include the act’s consequences in its description and make intention relevant in precisely the way I say Derrida does—not as an agent’s answer to the question “why” but as a cause.

Josh’s further concern is not just with how many descriptions of an action there can be but also with who gets to give them. He reminds us that it’s “we” who are coming up with our account of what the man is doing and reprimands me for dispensing with “the whole business of it being (often) we who offer descriptions that identify agent’s intentions.” But, when we’re reading a text—which is to say, explaining what someone else has done—who else could it be? And what difference does it make? If I understand Anscombe’s man clicking out God Save the Queen as proclaiming his allegiance to the Crown while he poisons the bad “party chiefs” (in hopes of putting in the “good” ones), that’s my description of his action. And if you understand him as making a joke about his dislike of the Queen, whom he considers actually to be one of the bad ones, that’s your description. But they’re both descriptions of his action. And whatever he was doing, it’s not altered by what we think he was doing. Which would be true even if one of our descriptions seemed to us (and maybe even to him) better than the description the man himself gave. Someone who explains my behavior better than I do does not thereby alter the behavior he explains.

But what if God Save the Queen were in code, maybe a warning to the about-to-be poisoned? Derrida says that a code cannot be structurally secret, that it must be able to function in the “radical absence” of every addressee and of every producer. Josh thinks that when Derrida insists that the text breaks with the intention of the writer he means “the present intention,” not intention as such, that he leaves room for “what has been intentionally invested.” Presumably this means something like what people have meant or ordinarily mean when they utter the words in question? But in Derrida’s insistence on a “secret cypher” known only by “two ‘subjects’” and necessarily able to function in the radical absence of both, there’s no intentional investment to be found. Indeed, in Spurs, there’s only “one subject,” the point in both texts being that “I have forgotten my umbrella” is “structurally liberated from any living meaning,” that it must function without the invocation of any intention.4 And it’s only because he imagines this absolute separation of the text from any intention that he can also imagine the text as “indefinitely open.”5

But how then should we understand what Josh describes as Derrida’s own commitment to the (intentional) act, to “language,” as Josh says, “put to use.” That’s actually the question addressed by “Produced and Abandoned,” the question of what kind of act Derrida understands writing to be. And the answer I give is that he assigns intention a causal role, and that it’s in this causal role that intention has its place but is unable to govern, which is what it means to think of the utterance as “abandoned” or “structurally liberated” (same description, different affect).

That’s why I’m not saying that intention does govern—scenes of writing, codes, or anything else. Or that it can never be abandoned, or that we need to push back against the “absolutely ungovernable.” The whole melodrama of governance (with the cast of characters celebrated in Josh’s closing peroration: “the governed and the ungoverned, the abandoned and not … always at work all together every time”) is precisely what’s created by the imagination of intention as something either addable to or subtractable from the text, as governing or failing to govern. Once you think of intention as governing (or not) and of meaning as being governed (or not), you’re committed (whether you want to be or not) to what Anscombe called the imagination of intention as an “extra property” of actions rather than a “form of description of events.”

Henry’s persistent worry—that I ought to explain what I think meaning is and not just keep on saying something like “the meaning of ‘meaning’ is simply what the utterer means,” and that I need to be able to do so without resorting to “mentalist talk”—seems helpful to me here in a way that maybe my response to his irony example didn’t bring out. Not because I can say what he wants me to but because the identification of the meaning of a text with what its author means emerges from what we do regardless of what we say about it.

Part of the problem with the mean teacher is that it’s more plausible to think of him as just lying, in which case there’s no issue about what he meant, just about how he really felt. But in situations where our understanding of what someone said (as opposed to our belief in the truth of what someone said) is foregrounded, the question of intention is too. When, for example, we produce conflicting interpretations of a text, what are we disagreeing about? We might occasionally understand ourselves as disagreeing about the meaning of a word in the sense that we disagree about its definition. And we could end that disagreement by looking it up in the dictionary. This would at least put a little distance between our sense of a word’s meaning and our sense of what the speaker meant by it.

But not very often and not for long. Not for long because resolving our disagreement about what the word means in the language would leave open the question of whether the speaker was using it to mean what it means in the language. And not very often because our characteristic disagreements are precisely about how the speaker was using it. The dictionary doesn’t tell us whether Wordsworth’s “rocks and stones” are inanimate objects or living bits of nature. And the dictionary won’t tell us whether Milton was of the Devil’s party or whether the ghosts at Bly are figments of the governess’s imagination. Those disagreements are about how words are being used in a particular text. In other words, the identification of meaning with what the writer/speaker means is already in place whenever we disagree (or agree) about the meaning of a text, and it doesn’t matter what most people would say or whether they would resort to mentalist talk. Just as, in an only slightly different way, it doesn’t matter if Derrida isn’t talking about ordinary meaning but is denying the absolute fullness of meaning. Because if the absolutely full meaning he denies is the text that can’t be misunderstood, that’s the conflation of meaning and communicating that I criticize. And if it isn’t the text that can’t be misunderstood, then the denial is mistaken—no meaning is ever not full.

Another way to get at this is just to say that the identification of what the text means with what the author intends is already in place without our having any recourse to a theory of intention or to a positive account of intention as a mental state. While the work that it does in place—determining the act—makes clear the mistake in imagining that the text we write could be either controlled by or liberated from what we meant.