From Reference to Sense

Modern formal linguistics was born of a strange reversal and irony. From Locke to Leibniz, Kant to Frege, the modern philosophy of language—following in certain of the footsteps of ancient Greek philosophy—has largely concerned itself with the question of how best to formulate true propositions and certain knowledge. Natural language, for our Enlightenment forefathers, would simply not do. Post-Babel, post-Flood, that truest human language—the language of Adam—was lost, a language that performatively named that which it already knew, the essence of things as they really were (Reference AarsleffAarsleff 1982; Reference Eco and FentressEco 1995; Reference OkrentOkrent 2009). Over the years, much effort has gone into getting closer to that more basic language, that fantasy moment when word and referent were one, when representation and truth were the same. Some of the better-known projects within this imaginary bear familiar names: the real characters of George Dalgarno and Bishop John Wilkins (Reference Wilkins1668), the rational calculus of Gottfried Leibniz (Reference Leibniz and Loemker1976), and the conceptual notation of Gottlob Frege ([1879] Reference Frege and Bynum1972), a self-declared continuation of Leibniz’s rational calculus, not to mention contemporary schemes like Steven Pinker’s mentalese (Reference Pinker1994) and Jerry Fodor’s language of thought (Reference Fodor1980).

A significant part of the genealogy of modern formal linguistics lies here, in the attempts to develop a truly logical structure out, and instead, of natural language. With Frege’s discoveries in quantification, his application of functional analysis to intrapropositional content, and, most importantly, his differentiating of sense (Sinn) from reference (Bedeutung), this once Adamic project took its modern, secular form. With regard to sense and reference in particular, an order of symbolic regularity independent of the world, which need not bother itself with actual extension, was “discovered.” Propositionality could be described as purely internal to a semiotic system, without any necessary appeal to what Charles Sanders Peirce—another inventor of the logic of relatives—called indexicality (see Reference LeeLee 1997, 16–39). By eschewing indexicality, Frege’s conceptual notation eschewed the hope for a language that could directly encode the essence of the world, which could reveal it in its immediated, iconic form. Rather, but no less ambitious, the notation would encode universal and necessary logical relations and intensional content, revealing the internal rationality of symbolic thought.

While Frege, like Leibniz, Wilkins, Bacon, and others, advocated for the replacement of natural language with its artificial supplement (in at least certain domains), the ironic reversal of the history that follows Frege is that it is precisely Frege’s logical notation that came to be applied to model natural language. That which was meant to replace language was rediscovered as its essence. No longer the supplementary corrective to that imperfect medium, logic came to be seen as at the heart of natural language. Logical positivists of the early twentieth century, influenced by Frege, took up his separation of language from world as the basis for their correspondence view of language—that propositions encoded and reflected states of affairs, a world always exterior to language—enabling the modern semantico-syntactic approach of formal linguistics that dominates much of the field of linguistics today.

But Frege’s work in logic didn’t simply clarify what aspects of language contributed to propositional content and right thought. It also reformulated the relationship of propositionality to truth and reference as such. As is well known, Frege’s ([1879] Reference Frege and Bynum1972) early conceptual notation works off of a principle of compositionality, namely, that the reference of a proposition is a function of the reference of its components. Hence, by identity and substitution, a proposition should maintain its reference if the referents of its parts remain identical across substitutions. Consider, for example,

1. (1a) If and , then and are equivalent propositions.

2. (1b) Similarly, and both compute equally if .

But, as Frege noted in his landmark essay “Über Sinn und Bedeutung” (Reference Frege, Geach and Black[1892] 1980), the same is not true for all propositions (see Reference LeeLee 1997, 35–36). Consider, for example,

1. (2a) Frege believes/says that the Morning Star is Venus.

2. (2b) Frege believes/says that the Evening Star is Venus.

We might think that these pairs of statements compute equivalently because the referents of all their parts are the same. But, as Frege argued, they are not. One might be true and the other false, for example, if Frege didn’t know that the Morning Star and the Evening Star have the same reference. What Frege realized is that in such so-called oblique contexts the subordinate clause cannot be treated by its reference; it can’t be computed relative to the object or state of affairs out there in the world to which it otherwise corresponds. Rather, it has to be treated by its sense; that is, to its propositional content as it is subordinated, or made relative, to the sense and reference of the matrix clause, namely, what Frege believes or says. From this separation of sense from reference, Frege concluded that the referent of a proposition is not a state of affairs in the world as such, but a truth value. Such truth was distinct from the sense of the proposition. And so goes the standard history of formal semantics and syntax passed down to students today (Reference GamutGamut 1991, 1–15).

But we might read Frege’s “discovery” of the disjuncture of sense and reference in another way. What Frege discovered was that reference—that indexical relation of sign form to referent—can be shifted, decentered, and suspended through reflexive semiotic forms such as reported thought and speech constructions, so-called oratio obliqua (2a–b). In such constructions, propositions in the subordinate clause are made contingent on the matrix clause verbs that describe, report on, and typify them. In such oblique contexts, the subordinate clause is shifted, its indexical ground calibrated not to the world as such but to the world invoked by the citing event, that is, the referential world denoted by the matrix clause.

By discovering a world-independent realm of propositional sense, Frege identified a set of metalinguistic constructions that draw on the indexical and interdiscursive power of language to act on itself, to suspend and transpose referentiality by making it relative to other discursive events. But Frege did more than recognize this semiotic potential in language. By drawing on it he brought something new into the world, the very concept of sense as distinct from reference, a reflexive apprehension of language as a purely propositional system. By citing the capacity of language to cite, he entailed into being a whole new epistemic and metaphysical enclosure. And yet, the very constructions Frege identified harbor the decentering of that very enclosure, for, just as reference can be bracketed through oblique contexts, so too can the very distinction of sense and reference. This was, as I discuss below, J. L. Austin’s (Reference Austin1962) discovery of “performativity.”

In this article I call this reflexive interdiscursive capacity of semiotic practice citationality. As explicitly manifest in canonical citational acts like quotations, citationality refers to a more general property of (meta)semiosis: the ability to re-present an event of discourse while reflexively marking that representation as not(-quite) that which the citational act presences. While citationality is a concept developed by Jacques Derrida (Reference Derrida1988a) in his critique of Austin’s discussion of performativity, my usage is critically different from Derrida’s. For me, citationality is a reflexive quality of certain kinds of acts, acts whose citationality is, in one way or another, foregrounded as constitutive of their status as the particular kind of act they are construed as being. This means, as I argue, that citationality is not a transcendental feature of all semiotic activity, as Derrida would have it, but is an aspect and achievement of particular (kinds of) discursive acts in context.

As I show, citationality is one powerful means through which language acts, through which repetition begets difference, through which newness comes into the world. In what follows I explore the (meta)semiotics of citation, focusing on the ways in which the citation’s reflexive organization enables and constrains certain kinds of performative potency. I pay particular attention to Derrida’s (Reference Derrida1988a) argument that at the heart of the explicit performative is citationality, an iterability of the sign which is also the explicit performative’s limit, what undoes it even as it gives it its “umph.” While I show that Derrida is correct, I suggest that it is for the wrong reasons. I then turn to the limits and remainders of citationality, focusing on Luke Fleming’s (Reference Fleming2011) discussion of taboo speech as an example of the recalcitrance of certain kinds of performative signs to being cited.

From Quotation to Citation

1. (3) Austin (Reference Austin1962, 163) said: “nowhere could, to me, be a nicer place to lecture in than Harvard.”

Here I have quoted Austin, to the word. We know that item 3 is a direct quotation from a number of signs: the subject of the sentence is the proper name of a speaking agent (“Austin”) and the matrix clause verb is a verb of speech in the past tense (“said”). It has as its complement another proposition enveloped by double quotation marks, an American convention for direct quotation. Further, the colon that separates the matrix and the subordinate clauses, as well as the reference information “(1962, 163)” that provides the year of publication and the page number, also indicate that this is a direct quotation. This reference information supplies a set of directions—mediated by the bibliography of this article—by which to find out what the publication was, where and when this event of discourse took place, and whether the quote is accurate. Compare this direct quotation with the following:

1. (4) Austin (Reference Austin1962, 163) said that nowhere could, to him, be a nicer place to lecture in than Harvard.

The quotation marks are gone, and the complementizer “that” indexes that this is not a direct quotation but an indirect quotation. What Karl Bühler (Reference Bühler and Goodwin[1934] 1990) called the origo, or zero point of indexical reckoning, of the cited event (“that …”) has been shifted away from its original context to the context of the citing event. As Valentin Voloshinov (Reference Voloshinov1986) would put it, the point of view embodied through the matrix clause has “accented” the speech which it is reporting. Hence the deictics have shifted from “to me” to “to him.” While the wording has shifted slightly, the propositional content is, more or less, “the same.” Similarly, consider,

1. (5) Austin (Reference Austin1962, 163) claimed that Harvard was the best place to lecture.

The voice of the citing event has further colored, even biased, the cited event. The main verb has changed from “say” to “claim,” indicating a distancing, typifying Austin’s utterance as a “claim” rather than simply something he said. The citing event is metapragmatically commenting on the cited event (Reference Silverstein and LucySilverstein 1993), its epistemic status made relative to Austin’s own (suspect) predilections, and thus kept at an arm’s length.

All such citations are examples of what Michael Silverstein (Reference Silverstein2005, 7) has called interdiscursivity. An interdiscursive act is a discursive act that links two or more discursive events (minimally itself and another, or even itself and a figuration of itself) within the same semiotic frame, in this case, within the same sentence. By doing so, citations weave together different events into one complex act. The citation reanimates other events of discourse, presencing them in a context alien to their original utterance. And, crucially, it brings reflexive attention to this very operation. Citations are interdiscursive acts that are reflexive about that very fact. This reflexive interdiscursivity involves a play of sameness and difference, identity and alterity, what Charles Sanders Peirce (Reference Peirce, Hartshorne and WeissCP 2.276, 2.283) termed iconism and indexicality.

There is an iconic ground between the citing and cited event. For Peirce (Reference Peirce, Hartshorne and WeissCP 2.276), iconism is a relationship of a sign vehicle to its object based on their similarity. Citing something re-produces it in some manner and to some degree, whether this be propositionally (in indirect quotations) (Reference LeeLee 1997, 277–84), in linguistic form (in direct quotes), or in material quality—as in, for example, this digital scan of my print copy of Austin’s text (fig. 1), which, if I could, I might present to you face to face, wherever you happen to be reading this. Like Peirce’s example of the map of the island drawn on its beach (Reference Peirce, Hartshorne and WeissCP 2.230), the citation contains within it a replica of what it cites. This means that the citation is based on an irreducible sameness. Without this sameness, however it ends up being construed, there can be no citation.

Figure 1. J. L. Austin, How to Do Things with Words (1962, 163)

This sameness, however, is always marked by a difference. The citation is irreducibly indexical. Indexicality involves, as Peirce (Reference Peirce, Hartshorne and WeissCP 2.283) defined it, sign vehicles that, through time-space contiguity or causal relationship, point to their object of reference. In canonical citations like quotations, matrix–subordinate clause relations, punctuation marks, and reference information all indexically direct our attention, or point, to their object of reference, in this case, something Austin said. By indexing another act, citations coordinate distinct time-spaces, bringing them into a common relation of coevalness and contiguity, a coordination that makes their iconism possible (Reference SilversteinSilverstein 2005). This means that any citational act depends on inscribing a difference, or gap, between the very acts that are made iconic with each other (Reference DerridaDerrida 1988a; Briggs and Reference Briggs and BaumanBauman 1992). Without this gap there is no citation. Without the maintenance of the difference between Austin’s voice and mine, without distinguishing that some part of my act—the subordinate clause—is not, in fact, wholly mine, the quotation would simply be (figured as) my statement, and thus not a citation at all. I could have simply written, for example,

1. (6) nowhere could, to me, be a nicer place to lecture in than Harvard.

Of course, the absence of explicit citational marks in item 6 is ultimately no impediment to construing it as a citation of, or more aptly, an allusion to, Austin. Simple familiarity with Austin’s text might be enough to “get it.” Or, as in the case here, the essay’s larger poetic text structure (Reference Silverstein and LucySilverstein 1993)—that is, the multiple repetitions of this text fragment in the preceding discussion as attributed to Austin—would suffice to fill in the missing quotation marks. The point is that to construe an act as a citation requires us to construe it as reiterating some other act, to reckon it as distinct from that other act, even if, or precisely because, its form is “the same” as that which it cites. Whether such citationality is indexed with quotation marks or through simply being “in the know,” this reflexivity is criterial. There is no citational act without the metacommunication of this play of indexicality and iconism, sameness and difference. A citation is an interdiscursive act that is reflexive about that very fact.

Every citational act suspends or brackets something of what it cites, even as it carries something over from it. The citation focuses our attention on something that is immanent in, but perhaps not otherwise manifest to, our experience of that which is cited. Consider, for example, Austin’s (Reference Austin1962, 94–107) distinction of locutionary, illocutionary, and perlocutionary acts. Silverstein (Reference Silverstein, Cline, Hanks and Hofbauer1979, 208–16) argues that Austin’s classification objectifies different ways of reporting “what happened” in English (see fig. 2).Footnote ¹ Each mode of reporting is a citational act, re-presenting some particular aspect of the cited (linguistic) action. In 7a, the citational act brackets everything about the cited event except for its sonic qualities, here mediated through a conventional representation of nagging “nja nja nja.” Propositional content or morphological form is not represented. In 7b the original signal form is bracketed, but propositional content is foregrounded and carried over. In 7c propositional content is (partially) carried over, and the rhetic act is now framed as a particular kind of act (what Austin called the illocutionary act), in this case decreeing or ordering. In 7d the sound qualia, the propositional content, and conventional illocutionary type of the cited utterance are bracketed, while only the resultant effect is reported, framed as caused by the implicitly cited event.

Figure 2. Some ways of reporting “what happened” in English

What is of interest here is how different citational forms in English are productive as a function of their citational (meta)semiotics. These forms bracket and re-present something (sonic qualia, reference, sense, etc.) and in doing so open up new discursive spaces. The citational affordances of English semantics and grammar are deployed by Austin to performatively bring something into the world: a set of distinctions which have come to constitute the field of so-called speech act theory. Below I focus on Austin’s discussion of “explicit performativity,” discussing the ways in which the explicit performative’s citational structure not only suspends sense and reference—and thus that metaphysical domain of truth functionality that Frege “discovered” through indirect quotations—but, in doing so, makes it possible for certain social facts to be brought into existence.

From Citations to Performatives

J. L. Austin (Reference Austin1962) saw his discussion of performatives as troubling the analytic philosophy of his time’s “correspondence” theory of “statements” (which he also called constatives). As Austin noted, performatives like 8a through 8c aren’t statements—that is, they aren’t to be evaluated as either true or false, corresponding or not corresponding to some state of affairs in the world—even if they “masquerade” as “statement(s) of fact” (Reference AustinAustin 1962, 4):

1. (8a) “I name this ship the Queen Elizabeth.” [Said in a naming ceremony]

2. (8b) “I bet you that it will rain tomorrow.”

3. (8c) “I promise you that I will come tomorrow.”

Sentences like these may be appropriate or inappropriate to their context of utterance, successful or unsuccessful in their effects, sincere or insincere, and so on, but they aren’t true or false in the same way as a proposition like “All birds sing” is. Rather, such sentences are what Austin calls happy or unhappy.

Explicit performatives are different from constatives in another way as well. Even if a performative like “I bet you …” is grammatically similar to a statement (“He bet you …”), to say “I bet you …” is not, Austin argues, to make a statement about betting. It is to make a bet. In other words, the performative brackets some aspect about propositions—their truth functionality, their Fregean sense and reference—and by doing so, brings a social fact into the world: an act of betting has occurred by virtue of having felicitously said 8b. Below I argue that this double motion of suspension and creation is a function of the auto-citational organization of the explicit performative, the way in which it cites itself into being.

Before doing this, a brief comment. While initially Austin introduces the performative–constative distinction, by the end of How to Do Things with Words, all the various classifications and distinctions that he has constructed he deconstructs. Ultimately, performative and constative are not types of utterances but, rather, epistemological constructs. All utterances are part performative, part constative. What differs is what aspect we emphasize of them. But in making this move Austin goes back on perhaps his most important points: use is the most important arbiter of linguistic analysis and the performative is a material act. The issue is not that some acts can be epistemologically viewed by philosophers and linguists from the point of view of the constative or performative (or some blend of the two), but that contextually situated events of language use themselves draw on these various distinctions to do what they do. Different events of language use differently diagram this very distinction and inscribe it in their material form. Indeed, all language may be both constative and performative, but not all language metacommunicates this fact. There is a real material difference between, for example,

1. (9a) “Birds sing.”

2. (9b) “I hereby command you to sing right now!”

Items 9a and 9b differ in their indexical specificity and, relatedly, their reflexivity vis-à-vis their context of utterance. In contrast to 9a, 9b indexically grounds itself in its context of utterance and explicitly diagrams that very fact in its linguistic form (e.g., in its use of speech participant indexicals like “I” and “you,” adverbs like “hereby” and “right now,” and in the semantics of the matrix clause verb). By contrast, 9a is characterized by the total absence of any such reflexivity. While 9b seems to prefigure its own uptake as a performative act of “commanding,” 9a presents itself as a universal proposition.Footnote ²

But if the explicit performative is not absolutely distinguishable from the constative—as Austin’s failures demonstrate—what do we gain by rejecting Austin’s final move that the distinction is merely an epistemological one? What I argue below is that the explicit performative materializes a reflexive, citational (meta)semiotics, though of a particular sort. Explicit performatives, of course, are not the only acts to render such (meta)semiotics intelligible. Moreover, as linguistic anthropologists have argued, explicit performativity is a very special case of performativity more generally (Reference Silverstein, Cline, Hanks and HofbauerSilverstein 1979; Reference AghaAgha 2007; Reference FlemingFleming 2011). Explicit performatives do, however, given their highly transparent, conventionalized, and localized form, illuminate something of the semiotics of citationality (much as canonical citations illuminate other semiotic qualities). So what is this semiotics?

Explicit performatives are particular among citational acts in that, by their utterance, they cite themselves. Note that all explicit performative verbs (promise, command, order, bet, etc.) can also be used to cite other discursive events. As performatives, however, they seem to cite themselves. To say “I promise” is to describe what one is doing in saying “I promise” as a promise. It is to make a token utterance of the verb “promise” a token of the cultural type “promising.” Indeed, it is no coincidence that the so-called explicit performative verbs are also the names of the actions they instantiate (Reference AustinAustin 1962, 32). This is precisely why Austin, concerned with action more generally, was interested in them. As Austin realized, what makes the explicit performative explicit is that it marks itself as a performative. It makes clear what kind of speech act the utterance is: “To do or say these things is to make plain how the action is to be taken or understood, what action it is” (Reference AustinAustin 1962, 70). The central fact about the explicit performative, then, is that, as a token utterance, it cites itself as a token of the type of speech its (verbal) semantics metapragmatically denotes.Footnote ³

The explicit performative, however, reflexively typifies itself in a particular way: it conflates the event of its utterance and the event denoted by the matrix-clause verb by typifying them as one and the same. In this way, the explicit performative keeps in play and effaces the gap between citing and cited event, not by erasing its citational form (as in implicit allusions or nonexplicit performatives where the citational frame is absented from the utterance) but by describing the moment of utterance in, and as, the moment of utterance. Consider the following similarly reflexive linguistic forms:

1. (10a) “This is a sentence.”

2. (10b) “This sentence has (about) five words.”

3. (10c) (when spoken:) “I am speaking.”

4. (10d) (when read:) “You are reading this.”

Each of these sentences is “true,” or self-confirming, by virtue of the fact that it is spoken, or read, or both. As it turns out, any verbum dicendi in English has this quality when spoken out loud in the first person, present, indicative, active. As such, all explicit performatives describe what they do as they do it. This is why it is important for them to be uttered by the speaker in the present tense, along with adverbs like “here and now,” “hereby,” and the like. The performative sketches out the very context of its utterance, and through a circularity of self-reflexive reference and description, seems to magically bring that which it describes into existence. This circularity makes explicit performative formulae different from canonical citations. Consider, for example,

1. (11a) Speaker_k,t2: “He_i promised_t0 that he_i would pay her_j five dollars yesterday_t1.”

   Said on the day after t₁ (= t₂)

2. (11b) Speaker_i,t0: “I_i hereby_t0 promise_t0 that I_i will pay you_j five dollars tomorrow_t1.”

   Said right now (= t₀)

The reported speech construction 11a brings two events into alignment (an event of promising and an event of payment) and differentiates them from the moment of utterance: they are of a different time and place from the speaker’s utterance. The explicit performative in construction 11b, by contrast, sketches the event of utterance—a materially singular, token event of speech—as that very act of promising that it represents, bringing the token event of uttering “promise …” into alignment with the event type denoted by “promise … .” The speech participant indexicals “I” and “you” in subject and indirect object roles, the simple present tense, active voice, indicative mood, and the verb’s semantics all serve to suture together the citing and cited, token event and event type into a single performative act.

What is important is that the explicit performative doesn’t simply reflexively align its denotational content to its moment of utterance as a token event, but that it invokes itself as a token of a particular cultural type of action. In doing so the performative brings into being in the here and now that cultural type, a social fact made real. Through this self-reflexive circularity the performative simultaneously cites itself as a token of a type and effaces that very distinction. It brings the general to bear on the particular, equates the singular with the universal, such that this utterance of a promise is (seen to be) the same as the cultural institution of promising itself. This alignment, what Silverstein (Reference Silverstein and Lucy1993, 52) calls nomic calibration, is what makes explicit performatives different from canonical citations like quotations and more akin to ritual performances (Reference Silverstein, Schieffelin, Woolard and KroskritySilverstein 1998, 137–38). While canonical citations are token events of language that cite other token events within some temporal envelope, that is, that report some past, present, or future event of discourse,Footnote ⁴ explicit performatives invoke, and operate within, the seemingly timeless realm of the transcendent. What is key to see is how this ritual nature of performatives turns on the capability of token events of language to reflexively cite themselves as tokens of particular performative types while effacing that token–type relationship at that very moment.

This is precisely why explicit performatives in English tend not to be in the continuous present aspect, for example, “I am promising you to come tomorrow.” The explicit performative’s self-reflexivity focuses on the generality of the utterance as a type of act, rather than the fact of utterance as a token event (as focalized by the continuous aspect which draws out the unfolding durational nature of utterance as a material act). Hence the explicit performative has the same tense structure as a universal statement like “Birds sing,” whose verb phrase is nomically calibrated to its denotatum (fig. 3). By its nomic formulation of the verb, the performative partakes of the universal even as, through its speech event indexing deictics, it staples that universality to its particular context of utterance.Footnote ⁵ The performative splits itself, playing on its doubled reflexive formulation as (utterance-)token and (action-)type. It simultaneously speaks, as it were, in two different temporalities and (meta)physical realms. While its indexical metapragmatic signs (personal pronouns, temporal adverbs) typify the performative as a material token event in context, its verbal indexicals—or lack thereof—typify that token event as a token of a universal type, the descriptive backing of what kind of action being provided by the semantics of the metapragmatic verb (“promise”) and its complement (“I will come tomorrow”).

Figure 3. Indexical (non)specificity, nomic statements, and explicit performatives. Adapted from Agha (2007, 43, 57).

Just as with the initial examples that we began with—“Frege believes/says that the Evening Star is Venus”—the citational form of the performative shifts the propositionality of its subordinate clause so that it is relative to the matrix clause. But unlike the indirect reported speech construction, the matrix clause of the explicit performative is calibrated, and conflated in fact, with the nomic realm of the general and with the singular token event of utterance itself, its self-reflexive circularity bringing the denotation of the bracketed proposition into existence in the here and now. The performative, as with all citational events, brackets the truth of what it cites through its reportive calibration. But by also being nomically and reflexively calibrated through its reportive frame, the explicit performative harnesses that which it cites, drawing on its power of description to make it so in the context of its utterance.

Citations shift the indexical origo of that which is cited, suspending, or at least displacing, its propositionality. Citing something can change its point of view, as Bakhtin (Reference Bakhtin, Emerson and Holquist1982) and Voloshinov (Reference Voloshinov1986) emphasize. The performative is that act where the origo/point of view of the citing and the cited event are made to be the same, where citing and cited are construed as a single act. The line that keeps apart the citing and the cited in canonical citations is seemingly erased in the explicit performative, even as it is, paradoxically, maintained and diagrammed. Hence its magical, ideological efficacy.

In citing itself, the performative, then, doesn’t simply bracket its own propositionality. It refunctions it. It inaugurates something new, and not simply particular social facts (e.g., acts of promising, betting, etc.). As Austin pointed out, performativity opens up spaces of linguistic intelligibility and philosophical practice that exist beyond truth and falsity as such: that is, the study of linguistic happiness, appropriateness, sincerity, justness, authoritativeness, seriousness, successfulness, tastefulness, and so on. The point is that every citational act opens up a space beyond that through which it emerges, an excess to its own conditions of possibility.

Citationality and Performativity

If the explicit performative brackets and refunctions the Fregean sense and reference which it cites and emerges out of, is the performative vulnerable to this very reflexive operation? Can a performative itself be cited, decentered, bracketed, denuded? This is the question which Jacques Derrida (Reference Derrida1988a) raises in his critique of speech act theory. Derrida charges Austin’s treatment of performativity with an inability to answer this question without normative fiat, without excluding the “necessary possibility” (Reference DerridaDerrida 1988a, 15) that any performative can be, and thus must be designed to be, cited.

This ability to cite the performative, Derrida argues, is a more general feature of all signs, what he calls iterability, the necessity that for every sign to be construed as a sign it must conform to some “code” (signs Peirce called types or legisigns), a principle of identity which regiments every iteration of it (signs Peirce called tokens or replicas) as “the same.” And yet, as repeatable, every sign replica is liable to new context-specific meanings. As both token and type, the sign “carries with it a force that breaks with its context” (Reference DerridaDerrida 1988a, 9), to be decontextualized and recontextualized, transported across time and space, accruing new meanings and values (Bauman and Reference Bauman and BriggsBriggs 1990; Briggs and Reference Briggs and BaumanBauman 1992; Silverstein and Reference Silverstein, Urban, Silverstein and UrbanUrban 1996). To be a sign, then, implies this independence from any particular context. “This breaking force is not an accidental predicate but the very structure” of the sign, Derrida writes. It constitutes its “essential drift” (Reference Derrida1988a, 9).

There is a doubled excess to any sign: as a token relative to a type there is an indexical excess of context specificity—any sign can be “grafted” onto a new context to generate new, singular meanings; as a type relative to its token, there is always an excess of intelligibility which maintains itself across contexts. Iterability means, then, that every sign is context dependent and context independent. That is, the fit between a sign and its present/determinate meaning in context is loose, just as the fit of the sign in context to its symbolic “code” is loose. Derrida’s point is that the sign must have both kinds of looseness of fit inherent to it. This indexical (under)specification splits the sign so that it can be what it is (its transcontextual identity) and different from that (as materialized in particular contexts) at the very same time.Footnote ⁶ The life of any sign, then, is caught between its double life as a type and as a token. The sign is both, always. Even as it stands apart from its context and its “code,” it conserves and maintains both. It is repeated as the same even as it is constantly made different from itself.Footnote ⁷ It is this tension that Derrida sees as central to the ways that all signs (do) work. As we saw, this tension is reflexively diagrammed in the citational act’s semiotic form. The canonical citation re-presents some semiotic act, but always marked with a difference and a disavowal. The citation’s relatively transparent and localizable play on sameness and difference is why Derrida’s focuses on it as the metonym of the larger class of iterability, as the materialization of what he calls a “generalized citationality.”

If citationality undergirds all signs, it undergirds performatives as well. To exclude the citation as a “parasite” or “etiolation” of “ordinary language,” as Austin does, then, is to exclude from analysis that which makes a performative able to be performative. Derrida (Reference Derrida1988a, 9) writes: “For, ultimately, isn’t it true that what Austin excludes as anomaly, exception, ‘nonserious,’ citation (on stage, in a poem, or a soliloquy) is the determined modification of a general citationality—or rather, a general iterability—without which there would not even be a ‘successful’ performative? So that—a paradoxical but unavoidable conclusion—a successful performative is necessarily an ‘impure’ performative, to adopt the word advanced later on by Austin when he acknowledges that there is no ‘pure’ performative.” Derrida’s counter to Austin, then, is that at the heart of all performatives is citationality. It is not an externality to performativity, but is internal to it.

Derrida is right to emphasize that the citation underwrites the performative—but perhaps for the wrong reasons. As we saw above, the explicit performative is constituted by its citational semiotics. In this sense, citationality is the semiotic heart of explicit performativity. The explicit performative reanimates the grammatical structure of the citational act, interdiscursively bringing into play multiple events that it sutures together through the very event of its utterance. In doing so it balances the tension of iterability—the duality of the sign as token and type—that Derrida identifies (also see Reference DerridaDerrida 2002). But more than balancing this tension, the performative effaces, or conflates, it such that this token event of utterance here and now is the type which it invokes. It evaporates that duality and tension, bringing the events which it calibrates into alignment, if only in that act and for that moment.

It is that conflation, that closing of the gap between token and type, between citing and cited event that the (non-explicit-performative) citation keeps open. Indeed, as we noted above (constructions 11a–11b), shift any of the indexical signs of an explicit performative and the effect vanishes. It vanishes because the superposition between the utterance’s semantic sketch and its moment of utterance are disarticulated. Following Austin (Reference Austin1962, 22), when a stage actor says, as part of a play, “I hereby promise to pay you tomorrow,” the actor isn’t promising—he is playing at promising. As Austin puts it, such “parasitic” uses are “hollow” and “void.” By being framed as part of a play, the performative force of the utterance is denuded. But why? As we indicated in our initial discussion of Frege, citations shift the indexical ground of that which they cite, they untether and relativize the cited event to its citing frame, be it a matrix clause or the proscenium of the stage. The explicit performative, when cited, is no longer reflexively calibrated to the moment of its utterance, but instead is reportively calibrated to a different time-space. Hence its default performative entailment evaporates. As Derrida (Reference Derrida1988a, 89) points out, however, the issue isn’t that the promise on stage isn’t a promise. It is that it isn’t a promise for the actor. It is, however, a promise for the character in the play. The whole theatrical frame of a play shifts the indexical ground to the imagined world of the play’s diegesis.

What this shows, however, is that in focusing exclusively on the duality of token and type in the locutionary act, and the ways in which the citation complexly plays with it, Derrida has missed what is crucial about both explicit performatives and citations, what makes them similar and yet crucially different from each other: the ways in which they reflexively diagram iterability. There is a critical difference between the citation of the performative and the performative’s citationality. Both are citational, and yet they reflexively diagram that citationality very differently. In contrast to the canonical citation, the explicit performative turns on obscuring the gap between token and type, on conflating them, on treating what is otherwise a difference as an identity without residue. In short, Derrida’s argument about the citationality of performativity is at odds with the very denial of citationality that explicit performative acts presume upon as performatives. While the citation draws out and exposes that quality of the explicit performative which is immanent in it—its reflexive interdiscursivity—doing so can act to denude the performative of its capacity to function as a performative.Footnote ⁸ This is Derrida’s point, of course. But in making this point, Derrida problematically shifts focus away from the crucial fact about explicit performativity: the explicit performative is a citation which typifies itself as not a citation (even as it draws on, or cites, we might say, its grammatical form); it eschews its citationality and reflexively presents itself not as a description of an act, but as that act, not as a relation of re-presentation, but of identity.

While all semiotic acts are beholden to what Derrida calls their citationality, not all signs foreground and diagram the inherent tension between token and type in the same way, if at all. Some, like explicit performatives and universal philosophical statements, require that this relationship be downplayed. Other acts, like mimicry and acting, require that token and type be held is suspended animation, not resolved, but made to be the very hinge upon which semiotic activity unfolds. And this matters. It matters because the token–type relationality that underwrites all semiotic acts—their iterability—is distinct from the reflexive materialization of that potentiality. What is crucial about citational acts is not, as it turns out, their citationality (a feature, perhaps, of all semiotic acts, as Derrida emphasizes). It is that they diagram that very fact (something which only some acts seem to do). While all acts are iterable or citable, not all are reflexive about that fact, or even reflexive about it in similar ways.Footnote ⁹

Derrida’s (Reference Derrida1988a) arguments in “Signature Event Context” and “Limited Inc. a b c” take the following form: a possibility of infelicity (say, by citing a performative), because it is a possibility, is also a necessity (it must be possible). As a “necessary possibility” (Reference Derrida1988a, 15), that necessity imprints itself on the sign as an exteriority residing in its heart, anticipating its future actualization (Reference DerridaDerrida 1988a, 64). But in emphasizing this necessary possibility Derrida brackets the question of actuality (Reference LoxleyLoxley 2007, 83–5). The problem is that not every possibility is equally actualized, or recognized, as a possibility in social life, and this, not as a contingent fact about this or that sign, but as a regular fact about social semiosis. This indicates that citationality is perhaps best seen not simply as an inherent fact about all semiosis, but as an empirical fact about particular (types of) semiotic acts.Footnote ¹⁰ To frame citationality in this way is to shift focus to the ways in which particular (types of) signs in context reflexively typify themselves (and other signs) in particular events of sign use.

This also implies that just as performatives are achievements, so are citations. Citational acts misfire and fail all the time: plagiarism when the citation isn’t explicit enough (or is dissimulated as the citing agent’s own words), falsity when the cited event is different from what is reported, misunderstanding when no one “gets it,” or, as I discuss below, as taboo when one ought not to cite. There are no guarantees that a citational act will be seen as a citation and this necessary possibility, as we might say channeling Derrida, is inherently inscribed in the notion of citationality: for citationality to be possible, its impossibility must also be in play, its seal must be split and doubled. That is, citationality must itself (necessarily possibly) fail in particular contexts. Its ability to surveil the scene of semiosis must be limited.

In citing the concept of citation—that is, turning it into something new by repeating it in new contexts (and talking about that very repeating)—Derrida’s denotational text glosses over an important quality of the citational act: its reflexivity.Footnote ¹¹ Citations do their work by pointing to the ways in which they are not what they reanimate. Without this metasemiotic layer of intelligibility, citational acts are not citational at all. It is not the mark of difference that matters; it is the mark that points to difference that matters.

Beyond Performativity and Citationality

As I have argued, citing brackets what it cites, repurposing it by making it relative to the event which cites it. Oratio obliqua and other such constructions bracket reference, thereby entailing the realm of Fregean sense. Explicit performatives bracket Fregean sense and reference, thereby bringing into being social facts of various kinds. Further, citing performatives bracket, or shift, their (un)happiness and, in so doing, open up social horizons of possibility beyond the illocutionary forces (and felicity conditions) which they cite. Deconstruction is one such example, a method of critical reading that opens “the crevice through which the yet unnameable glimmer beyond the closure can be glimpsed” (Reference DerridaDerrida 1976, 14), that is, that through citational practice can decenter Western metaphysics and open up new ways of doing philosophy (the glimpse of the unnameable glimmer beyond the closure).

This creative potential of citation is, as I take it, Derrida’s most important point. Citation conjures something new into the world by deconstructing the intelligibility and legibility of the social forms that it reanimates, introducing an alterity through repetition (Reference DerridaDerrida 1988a, 40). Citations focalize new qualities, eliciting and entailing them out of what is cited. This capacity, as I would argue with Derrida’s method/poetics but against his own theoretical treatment of citationality, is a function of the reflexivity of the act of citation. Indeed, is not deconstruction’s defining feature its reflexive, poetic style? Is it not a citational strategy that attempts to open a discursive space beyond that which it cites by compulsively repeating it, and simultaneously calling attention to that repetition as an appropriation, until it gives itself over to its self-difference? More than a work of “theory,” texts like Limited, Inc. are poetic and comic acts of mimicry, chock full of double meanings, puns, inside jokes, biting satire, quotations (within quotations), allusions, and self-reflexive meditations; in a word, they are acts of citation.

But is the moment of citation as agentive and all-powerful as it might seem? By embodying and re-presenting that which it simultaneously disavows as not what it (re-)presences, does not the citing event necessarily come to share qualities which that which it suspends? Is it necessarily tethered and entangled by that which it cites? Is it, too, engaged in a risky liminality, always potentially too close to the awesome powers of the cited? Is not every citation a prophylactic? Or put otherwise, is there an excess to citation’s own excess, a remainder which defies the bracketing which the citation attempts? Does the cited resist its citation, resist being shifted and suspended? Are there some signs that cannot, or perhaps ought not, be cited?

Speak of the _ _ _ _ _, and He Shall Appear

Citations repeat but with a difference, a difference that is marked at a different logical type than that which is cited. The difference that the citation introduces turns on partially negating or bracketing that which is repeated, displacing its status as itself, marking it as not what it otherwise would seem to be. In doing so the citation suspends something of what it re-presents, be it its quality, its semanticity or denotation, or even its performativity. And yet, the capacity to bracket is always partial, for just as the citation depends on its metamarks of difference, it also depends on an irreducible sameness, an identity in abeyance, a quality of similarity between the act of citation and the cited.Footnote ¹² This implies that there is a recalcitrance of that which is cited, a resistance to the bracketing force which the citation attempts, always partially, to instate. Even as the citation brings something new into the world, under the metasign of the citation, something of the cited always remains.Footnote ¹³

This makes the citation not simply an act which brackets or appropriates the power of that which is cited, but one which does so with risk. All prophylaxis risks inverting into that which it protects against (Reference DeanDean 2013), all negations risk the affirmation of what they deny (Reference Freud and GayFreud [1925] 1989), and all expropriations risk being themselves purloined (Reference Lacan and FinkLacan 2007). The citation as a (meta)semiotic form is defined by these risks, for while it attempts to bracket and shift, it always deals with material events/signs which, by giving themselves over to citation also resist being bracketed, which resist being shifted, being cited.

It is this resistance, of course, which makes citation possible, for only that which can withstand the force of de- and re-contextualization can be cited. But harnessing this resistance is also what makes the citation powerful, it is what the citing act draws upon so as to be a citation.Footnote ¹⁴ And yet, in this liminal space between de- and re-contextualization, in the gap between the citing and the cited act, some acts would seem to be too risky. Some signs are too recalcitrant. They require careful care. Sometimes only a trained expert—such as a shaman (Reference TaussigTaussig 1993) or a philosopher (Reference ButlerButler 1997)—can handle such signs without getting burned. The risk is not simply that of failure or defeasibility, but of being overwhelmed by that which is cited. This risk points to a different kind of excess: not the excess generated by the “necessary possibility” of citation (i.e., the excess generated by the iterated sign’s “graft”) but of the excess immanent within every sign, its stubborn remainder, its reminder of what Peirce (Reference Peirce, Hartshorne and WeissCP 1.322, 1.325) called its “Secondness.”

Here I draw on Luke Fleming’s (Reference Fleming2011) work on taboo speech and what he calls rigid performativity, signs whose performativity seems indefeasible. Fleming is concerned to explicate the semiotics of signs which are framed in particular communities as unmentionable, curse words (e.g., the “N-word,” the “F-word,” and the like in American media discourse) and taboo words (e.g., the proper names of certain “core” affines or of the dead, as among the Banggi of East Malaysia or the Twana of the Pacific Northwest, respectively). With such signs, as Fleming (Reference Fleming2011, 153) puts it, “attempts to neuter them of their performative effects by recontextualization always run the risk of replicating their taboo effects.” Among some Tamils, for example, the utterance of the word for ‘snake’ (pāmbu) cannot be said outside at night without risking the appearance of a snake. Such a word cannot be used or cited without entailing this risk and eventuality.

While Derrida may be correct in saying that all signs are citable, what their citation does (to them) is not always the same. Not all signs are citable to the same pragmatic effect. As Fleming argues, the capacity of various semiotic forms to reanimate and bracket other signs depends on larger institutional and semiotic organizations of (non-)use (status hierarchies, kinship systems, cosmological beliefs, and the like) within which certain signs may become more and more difficult to deploy.Footnote ¹⁵ The point for all such cases, however, is not that such signs cannot be cited, but that they ought not to be cited and, more to the point, that if cited certain entailments (necessarily) follow.

What Fleming shows is that signs differ with regards to the degree that they give themselves over to bracketing through citation. For the citation of a performative on stage, the construal of the citational frame subordinates the reanimated event, whereby the cited performative is nearly completely suspended and shifted. For the citation of a taboo word, the attempt to indexically differentiate the reanimation of the word as a mere re-presentation of the taboo type is less successful. A remainder of performative effect characterizing its taboo status always seems to linger, the citational frame enveloping the taboo word failing to fully decenter its performativity. By contrast to both, as we saw, the explicit performative balances and aligns the citing and the cited event through its reflexive and nomic calibration, conflating them as part of one and the same iconic indexical act. In this way the explicit performative stands between the rigidly performative unmentionable act—the act whose performative effects hold regardless of whether it appears within citational marks or not—and the felicitous citational act—the act which in citing another act marks itself as a citation of it and thereby successfully brackets something of it (e.g., its performativity).

Note, of course, that “rigid performatives” like taboo words are not explicitly citational. In fact, quite the opposite. As Fleming’s discussion suggests, the reflexive metapragmatics of their performativity is nonlocalizable to their form (it being, for example, part of a larger cosmology or set of kinship practices). And yet, when seen relative to such implicit, virtual framings (which they must be so as to be intelligible as taboo and as performative), the taboo word, like the explicit performative, is both reflexively calibrated to its token event of utterance and nomically calibrated to some event type (taboo, disrespect, cursing, etc.), the latter often grounded in some transcendental, social imaginary of perlocution (myth, kinship, law, etc.). If the citational frame of the explicit performative serves as the metapragmatics that aligns these different semiotic levels within the utterance act, in taboo speech—as with indirect speech acts, nonexplicit performative acts, and the like—the relevant metapragmatics is exogenous to the act of utterance, which is to say, following Agha (Reference Agha2007), that its performative effect follows from its enregisterment. Indeed, to construe such acts as performative requires being socialized to certain presuppositions (e.g., about the power of the tabooed word, its place in a larger cosmology, etc.) through which the felicity of the act is felt to be inherent, necessary, and even materially “in” the words themselves.Footnote ¹⁶ Just as the citational, metapragmatic framing of allusion is “elsewhere,” so too is the metapragmatics of performative signs like taboo words. For our discussion, the point is that for such performatives, their metasemiotic framing as taboo resists being shifted by other, more localizable, framings, which is to say that their performativity cannot be easily defused by citation.

Such rigid performatives, seemingly immune to citation and devoid of citational semiotics, are quite distinct from the citational acts that I have explored in this article. And yet, they interestingly act on the signs that fall within their ambit similarly. Both metasemiotic frames decenter, and even decompose, the semiotic form of those signs that they frame. The avoidance of taboo speech leads to forms of phonetic leakage, where the semantics, and even morphology, of taboo lexical types come to be subordinated to the qualia of the forms in question. In such cases of extreme taboo, even words of different lexical types that sound similar to the tabooed word come to be avoided (Reference FlemingFleming 2011, 153–59). Similarly, as we’ve seen, citations act to bracket something of the cited event (e.g., its reference, propositionality, performativity) so as to elicit some other quality, or newness, out of it. Taboo frames, however, act to proliferate iconic identities out of distinct (token-)types, while citations tend to proliferate indexical differences of out iconic replicas. Second, both framings decenter as a function of how they reflexively play on token–type relations. Homophone avoidance of tabooed (qualia-token-)types, as Fleming (Reference Fleming2011, 155–60) describes it, is driven by a reflexive anxiety concerning the ways in which tokens of different lexical types may come to stand as tokens of the same taboo type. Relatedly, though in inverted fashion, citationality is a reflexive play on the iterability of the sign, the capacity of every (token) sign to conserve and break with its context and its “code” (type).