Indexicality’s Ambivalent Ground

Abstract

Through its theorization and elaboration in the path-breaking work of Michael Silverstein, indexicality has served as a foundational analytic category for linguistic anthropology, both in its ethnographic analyses as well as in its theoretical interventions into key issues in the philosophy of language, linguistics, and sociocultural anthropology. The working out of indexicality’s implications for the study of semiosis—and more particularly, language in culture—has and continues to produce a vital and dynamic theoretical field of scholarly activity. This vitality, I argue, emerges from a foundational ambivalence within the category of indexicality: between, on the one hand, immediacy and presence and, on the other hand, mediation and representation. Productively unresolved, this ambivalence is less a problem than an opportunity and invitation for further ethnographic and analytic refinement of our accounts of (meta)semiosis and social life. A reflexive and deconstructive turn to indexicality’s ambivalent ground, then, is implied and necessitated by the category itself, though this in no way obviates its utility for semiotic and ethnographic theory and analysis; indeed, as I argue, such a turn is critical to indexicality’s ongoing utility to both.

Michael Silverstein’s essay “Shifters, Linguistic Categories, and Cultural Description” (1976b) provides linguistic anthropology’s most important touchstone on indexicality, a concept that has since been central to the field’s interventions into the philosophy of language, linguistics, and sociocultural anthropology. Synthesizing long-standing genealogies of anthropological, linguistic, and philosophical thought (Boas and his students, the Prague circle, and the American pragmatists, among others; cf. Silverstein 2004, 650), the thrust of these interventions has been to show how language and culture are inherently context- and event-mediated and mediating phenomena; that is, they are indexical phenomena. The focus on indexicality, and the various and far-reaching implications entailed therein, is the fundamental basis of linguistic anthropological analysis and theory today (Nakassis 2016b).

The term indexicality was coined by the American philosopher Charles Sanders Peirce (1839–1914) to denote a particular kind of semiotic “ground.” One among his well-known trichotomy of icon, index, and symbol, Peirce characterized indexicality as based on a “real connection” (CP 2.287) or “existential relation” (CP 2.243) between a material, token-representamen and its object, such as causality, co-presence, or contiguity (CP 2.306). Typical examples include smoke and fire, a man’s rolling gait and his (probably) being a sailor, the personal pronoun I and the utterer of the token-sign “I,” and, perhaps most canonically, an index finger and what it points to. An indexical ground is not a function of a conventional rule or habit (what Peirce called a Thirdness, as a symbol is) or of shared qualities (a Firstness, as an icon is), but by the actuality (or Secondness) of the relation between some existents in time and space (or at least, the presumption thereof).¹ A common way of saying this is that indexical signs are intrinsically related to the context of their occurrence (however capacious that context might be; Silverstein 1992).

While Peirce’s semiotic, and the place of indexicality within it, has been in the continual state of being discovered (and forgotten) by scholars since Peirce’s lifetime, it is from Roman Jakobson, who introduced the term indexicality in the latter phase of his career at Harvard in his classic discussion of the grammatical structure of the Russian verb ([1957] 1984; see Silverstein 1975), that the term entered into linguistic anthropological discourse via the work of his student Michael Silverstein. Silverstein brilliantly generalized Jakobson’s insights, showing not just the full range of indexical relations in linguistic structure but also its articulation to the poetic pragmatics of speech in interactional and cultural contexts of usage (Lee 1997).

The importance of Silverstein’s expansion of indexicality has not just been that it brought to light new objects or aspects of the study of language and related communicative modalities; nor that it brought together a number of long-standing objects of study into a unified analytic category, though it has done both things, as I discuss below. Rather, it is that the contemplation of indexicality’s implications for the study of semiosis—and more particularly, language in culture—has generated a vital and dynamically growing host of interrelated analytic concepts (Nakassis 2016b). This vitality, I suggest, emerges from a foundational ambivalence within the category of indexicality, and mutatis mutandis within the field of linguistic anthropology. Productively unresolved, this ambivalence—theorized as a dialectics of indexicality—has aided in the progressive program of linguistic anthropology as a science, though this is not without its ironies.

In order to unpack this ambivalence, I first underscore the critical intervention this analytic has been used to make by linguistic anthropology against idealist and structuralist conceptions of language and culture. From there I show that the ground of indexicality conserves something of what it is used to critique, namely, a representationalist ideology and metaphysics of presence.² I then detail how this internal tension has worked itself out through a proliferation of analytic distinctions and supplemental analytics that both explain and exemplify the dialectical semiotic processes that indexicality ambivalently points up. This is, I argue, as it should be, though a heightened reflexivity to this ambivalence is critical, as it opens new directions for the conceptualization of semiosis (and indexicality) in social and cultural life.

Indexicality and SAE Philosophy of Language

One long-standing and consistent strain in Western thinking about language—what we might dub SAE (Standard Average European) philosophy, to adopt a fashion of speaking from Benjamin Lee Whorf ([1939] 1956, 138)³—has framed language as a unique and exemplary semiotic, as manifesting a horizon of aspiration that, ironically—as SAE philosophers have consistently lamented—“natural” language, alas, fails to arrive at: a truly autonomous denotational-referential medium that captures its object in its totality, be that Platonic Form in its beautiful ideality and heavenly reason or reality in all its earthy Aristotelian actuality. Language, in this view, stands apart from what it denotes (be this heaven or earth), and its genius (or sin) is that in that distance it (never quite) realizes the universal and true by representing it. The semiotic property of language that allows this is its propositionality, the symbolic ground of language that functions independently of context, free from the messy Matter weighing down sublime (Logical) Form.

Yet if indexicality is that aspect of semiosis that grounds language back in its context of occurrence, one can see why it has long been treated as a residual problem for SAE philosophy of (SAE) language(s). As Bar-Hillel (1954) and Garfinkel and Sacks (1970, 347–48) both pointed out, the context-bound nature of semiosis has been an insistent thorn in the side of philosophy since the Greeks, one that logicians and others have continually attempted to eliminate or explain away, if only to always find it returned. (Garfinkel and Sacks compare it, in a psychoanalytically redolent and strange midcentury American metaphor, to a housewife compulsively attempting to clean one piece of floor over and over again, without success. We might suggest the equally Sisyphean, yet less gendered, allegory of whack-a-mole.)

Closer to home, we can point to two avatars of this SAE desire: Gottlob Frege (1848–1925), a contemporary of Peirce who mathematized logic (introducing functional analysis to describe propositional structure) and “discovered” the referent-autonomous, logically modelable realm of linguistic sense (Frege [1892] 1980); and Ferdinand de Saussure (1857–1913), who, along with Frege, set modern linguistics on its course with his distinction of langue and parole (speech), the latter of which, Saussure ([1916] 1986) famously argued, had to be excised from langage (language) before a proper linguistic science (of langue) could be possible. Langue was not langage, even if it was its ideal core, that essential kernel that was unique to it: namely, the synchronic, autonomous “system” of differential entities (signifiers) whose nonrandom and language-specific distribution and paradigmatic and syntagmatic relations formed the formal basis for (Fregean) sense (Lee 1997; Nakassis 2013a). Critical to see is how langue and Fregean sense stand apart from the world they denotationally encode, their significance and meaning not dependent on that world in the first instance. In this view, language in its essence is (or should be) free of indexicality, and if indexicality remains, this is as an inessential residue that can, or should, be hived off.

The intervention of linguistic anthropology, and the work of Michael Silverstein in particular, has been to put indexicality at the center of the study of language, and thus reveal the inherent limits of the idealist, semanticizing conceit that has animated both modern formal linguistics and SAE philosophies of language. On the one hand, this has taken the form of arguments that what has been taken as langue has, at its base, indexical relations of various kinds.⁴ The thrust of this critique is that at the heart of langue is parole,⁵ not as an inessential exteriority but as an animating principle of what Silverstein (1985b), following Mauss and Durkheim, has called the “total linguistic fact.”⁶ On the other hand, the focus on indexicality has clarified aspects of linguistic form/meaning and function/practice that are problematically theorized (or simply ignored) within a more narrowly focused formal linguistics, such as deixis, the poetics of interaction, performativity, and language variation, among other topics.⁷ As these various interventions have shown, the indexical processes that characterize language are not unique to it but are more general features of “culture” (Silverstein 2005b; Parmentier 2016, 15); and moreover, what is unique to language—namely, langue—thus serves as an untenable basis for a model of culture—namely, structuralism (Silverstein 1975, 1976b, 12).⁸

In short, linguistic anthropology has refused the notion of language as an autonomous “system” (Nakassis 2016b), as purely symbolic, purely defined by modalized reference and predication, and thus as essentially divorced from events of use, social relations, or culture. In this, linguistic anthropology has disavowed the representationalist ideologies carried forth by SAE philosophy in its tendencies to semanticize language and culture. Important to highlight is that this semanticization turns on distinguishing and keeping apart word and world, representation and represented. This cut is as much a semiotic ideology (Keane 2003) as it an ontology (in the sense of a theory [-logy] of Being [onto-]), one that, as Whorf and Peirce both noted, is characteristic of both SAE philosophy and folk theories of language. This semiotic ideology, we might further suggest, is animated by a particular metaphysics and desire, which we can characterize through the work of Jacques Derrida (1976): namely, a metaphysics of presence and desire for immediation, be it of thought or thing, word or world.

Indexicality as a non-representationalist (i.e., nonsymbolic) form of mediation has been invoked to refuse this metaphysics and its attendant desires. Yet, as I touched on above and detail below, an ambivalence lingers in the category of indexicality and thus in the interventions made in its name. Indexicality retains something of the larger ideological enclosure of representation and metaphysics of presence that it displaces. Perhaps this is because of indexicality’s very definition. Recall Peirce’s formulation of indexicality as a semiotic relation based on a “real connection” or “existential relation.” Might we detect a tension held taut in this definition and diagrammed by it: a tension between questions of mediation (existential relations, real connections) and immediacy (existential relations, real connections), presence and representation, being and semiosis? What kind of tension is this, and by playing on it what notes might we sound?

The Primal Scene of Indexicality

Consider the primal scene of indexicality, a scene that appears in most every foundational treatment of the topic of indexicality and that constitutes, I would argue, our default conceptualization of indexicality: an arm and index finger extended, pointing at some physically co-present object along with the gaze of the eye, attracting and then directing another’s eye to that object. In most accounts, this primal gesture is accompanied by a linguistic sign: a deictically determined noun phrase like “this hat” (Bühler [1934] 1990, 103–4), a sentential proposition such as “There is a balloon” or “That chair is yellow” (CP 2.293, 7.635), or simply a word-sentence like “slab” (Wittgenstein 1953, pt. 1, §6, §8) or “Gavagai” (Quine 1960, 29ff.; 1969, 28ff.; cf. Jakobson 1953, 21).⁹

This pointing index finger is a familiar ur-gesture,¹⁰ often figured as the basic communicative form that denotational propositionality either aspires to or completes (as in Aristotelian-inspired treatments of the subject), or as that to which propositionality eternally returns so as to actually refer (as Peirce [CP 2.357] and Bühler [1934] 1990, 43 both suggest).¹¹ In the telling of this primal scene, indexicality in the form of a pointing finger stands at the outset as ambivalent to language (and, by extension, thought), as external to it and yet as somehow central to what it is or was or should be. This extimacy is often emplotted as an origin myth of one sort or another, be it as an ontogenetic narrative of childhood development (Vygotsky 1978, 56; cf. Silverstein 1985a, 231–34), a phylogenetic narrative about the evolution of human communication (Tomasello 2008; cf. Levinson 2004, 98), or even as an anthropological linguistic myth of first contact (Quine 1960, 1969; also see Wittgenstein 1953, pt. 1, §§1–8).¹²

Embedded in this primal scene, I suggest, is a nagging worry and desire: the worry of mediation and the desire for more immediate encounter (if only at some origin), an ambivalence that pointing is both a self-evident, unproblematic operation (and thus that indexicality a self-evident, unproblematic semiotic ground) and yet also fundamentally underdetermined, ambiguous, shifty, never quite, and thus deeply mediated. Hence, the problem that Willard Quine (1960, 32; 1969, 6; also see Jakobson 1953, 21) famously drew attention to in his thought experiment of the anthropological “jungle” linguist trying (and failing) to determine the referent of the “heathen” “native’s” utterance of “Gavagai” while pointing to a white rabbit scurrying by: even with such a simple, basic act of ostension, how are we to know that the referent or denotation is (the co-present) “rabbit” and not “undetached rabbit parts” or some other ontological and epistemological peculiarity?

This primal scene of pointing dramatizes the basic ideological form of representation that Martin Heidegger ([1938] 2002) discusses in his essay “The Age of the World Picture.” As Heidegger suggests, the classic notion of representation entails a making present of an object for a subject (say, a rabbit for a linguist), an object figured as exterior to that subject so that it can be seen, and thus known and named, computed and controlled by him (Derrida [1980] 2002). Here representation operates within a historically and culturally particular “ocularcentric” regime, as some historians of philosophy have put it (Levin 1993), one that construes semiosis and knowledge as a species of vision and imaging (Rorty 1979; Judovitz 1993).¹³ Consider, for example, some of the visual language in Peirce’s remarks on the index:

The index asserts nothing; it only says “There!” It takes hold of our eyes, as it were, and forcibly directs them to a particular object, and there it stops.

(CP 3.361)

… which like a Demonstrative pronoun, or a pointing finger, brutely direct the mental eyeballs of the interpreter to the object in question, which in this case cannot be given by independent reasoning.

(CP 8.350; also see CP 8.41).¹⁴

For Heidegger, it is through such an ocularcentric formulation that Being comes to be articulated in modernity as representation (and denuded as simple beings); that is, that which “is” becomes that which can be made present to a subject as an image-object (Rorty 1979; Staten 1984). Underwriting the writings of SAE thinkers from Plato and Aristotle to Descartes, Husserl, Saussure, and beyond, this metaphysics—or semiotic ideology—has treated Being as the immediated beginning and end of thought and reason, normatively privileging the self-present, self-identical, and the origin over absence, difference, and mediation (Derrida 1976). This ontology contrasts with what Heidegger describes as an alternate metaphysics of presence: that of the pre-Platonic Greeks, for whom being (“what-is”) is less that which is looked upon by man as representation or as perceptually co-present object than that which looks upon man himself, which “rises up and opens itself” and “comes upon man” (Heidegger [1938] 2002, 68), and which in its presencing grabs and attaches itself to and in him (Derrida [1980] 2002, 109; cf. Nakassis 2017b).

The primal pointing finger, in this classic representationalist metaphysics, then, doesn’t just render the object present as an object; it serves it up to us as an image of itself, as something to be seen so that it can be more immediately and truly known, to the eyes if not the transcendental mind (cf. Peirce’s “mental eyeballs”). And indeed, at the anxious limit of this account of the origins of conceptual thought—where seeing a present token-being enables the Object’s essence/type to be “seen” by the mind—the Form of the Object, its identity, and its Being are secured and purified, if always open to the haunting risk of mediation that this purification keeps at bay.¹⁵

And herein is the irony with which we began, for it is against this enclosure and its variably cloaked, if still transparent yearning for immediation that indexicality rebels, be it in the hands of Peirce in his critiques of Descartes and Kant, or as grasped by linguistic anthropologists in critiques of structuralist approaches to language and culture, subject–object dualisms, and their related ideologies of representation.¹⁶ The indexical sign—as with the prototypic pointing finger—depends on, and is defined by, Peirce’s Secondness: hence, its value can’t be fully specified solely by appeal to transcontextual rule, law, or essence, but only ever relative to its particular contexts of use, that is, to other arrays of sign-tokens that, in their indexical co-relatedness, reflexively frame and determine the value/reference of such an indexical sign (if only for then, there, and them). As such, indexicality is always residually enmeshed in that (Aristotelian) Matter that eludes Form, in that materiality that eludes “meaning,” even as it, in matter of fact, has form and is meaningful.

But if indexicality gives itself over in this way to a distributed, non-self-presentist account of mediation and if it does so by an appeal to and desire for co-presence, real existence, and the like, then perhaps we can see the paradoxical ambivalence at work in the foundational critiques made through it. Indexicality forsakes a metaphysics of presence and idealist representationalism while, at the same time, promising to get us closer to phenomena in their actuality, to make them more present and epistemologically available to our social scientific gaze in ways that idealist representationalist approaches, alas, fail to deliver. This is indexicality’s ambivalent ground.

From the Fingertip to the Wrist: Ethnographies of Presence

We can detect in indexicality, and thus in linguistic anthropology, an implicit metaphysics of presence, making both of a piece with that which they have been devoted to critiquing. What are we to make of this? While I discuss one important upshot of this ambivalence in the next section, let me note here that to the extent that this ambivalence is problematic, it is because it remains tacit and untheorized, and thus unworked upon and undeveloped. This allows analysis to proceed as if indexicality stands on solid ground, as if its ground was self-evident when it is anything but.

My claim, however, is not that such a naïve appeal to and desire for presence is avowed by any linguistic anthropologist. Indeed, as I’ve suggested, the field is founded on disavowing any such appeal or desire. Rather, my point is that such disavowals rest on, and stand in tension with, tacit assumptions smuggled in by the notion of indexicality itself, assumptions regarding questions of presence, existence, and being that thereby risk reappearing in our analyses, if not in our theory. This structure of ambivalence repudiates that which thereby may return, continually holding the immediate and the self-present at a distance, even as we, in turn, are held by them.¹⁷

Pointing out this ambivalence forces such seemingly self-evident categories to become reflexive objects of worry; and, more to the point, ethnographic objects of worry that prompt us to reexamine our analytic concepts, in this case, indexicality. Indexicality has, of course, been central to anthropological analyses of how presence is achieved and mediated in variegated cultural contexts.¹⁸ Silverstein’s (2004, 626–27) masterful analysis of the Eucharist, for example, shows in semiotic detail how the presence of Christ is indexically manifest—as a function of an emergent, iconic diagrammaticity standing under a Biblical metapragmatics—through those comestibles (wafer, wine) that by their ingestion incorporate congregants, in a chiastic reversal, into the body of Christ. A particular kind of presence, here, is the achieved effect of the ritual’s iconic-indexical “dynamic figuration” in its historical and cultural context of real-time and cosmic unfolding.¹⁹

But if one finding of these studies has been the multiplicity of metaphysics of presence and thus the multiplicity of ontologies, and if indexicality, as it stands, itself presumes some such metaphysics, then on what other grounds might we find and found indexicality? That is, how might we make a critical and reflexive ethnography of presence the basis of our analytic, conceptual work (namely, in theorizing indexicality)? Rather than assuming that we know what indexicality is so as to show how presence is realized, how can an openness to the question of presence—in all its ethnographic complexity—reformulate our conceptualization, and thus analytic use, of indexicality?

This is a question that I have recently broached in a study of cinematic images of so-called mass heroes in the south Indian film industry of Tamil Nadu, such as the “Superstar” and “King of Style” Rajinikanth (Nakassis 2017b). Filmic images of such auratic, charismatic personages are, for their Tamil filmic audiences, not simply representations of fictive characters or even of the absent star-actor. Rather, they are often citationally figured by fans as akin to religious idols and other figures of sovereignty, which in a Hindu idiom of darśan transubstantiate with devotees through seeing and being seen. There is an ontological identity between screen image and its object (Rajinikanth, the Superstar). Every avatar-character “played” by Rajinikanth is Rajinikanth, performatively presenced in the moment of the image’s apperception. This presence is the basis for a “cine-political” potency (Prasad 2014); indeed, such stars are expected to and on occasion do translate their populist screen/star images into political “message” (Lempert and Silverstein 2012), segueing from their filmic worlds to elected political office.

The textuality of Rajinikanth’s films cultivate this politicized presence, with copious shots of Rajinikanth looking and pointing at, and speaking to, the audience, often in highly charged moments of thinly veiled allegorical political oration (Nakassis 2017a), with the audience yelling back, clapping, whistling, and touching the screen in response. Such scenes do not simply put Rajinikanth into the co-presence of his audience. Dynamically figurating them as metonyms of the “masses,” and him as an exemplary co-member who is also their representative, in such scenes Rajinikanth performatively encompasses, in ways not unlike the Eucharist (though in inversion of it), the audience/polity into his body politic. When the vector of Rajinikanth’s index finger fixes on the camera/fourth wall, he is not simply pointing at the audience. He is touching and grabbing them, transporting and transmuting himself/them across the diaphanous membrane of the screen so as to englobe them in his being. Reconstructing the indexical ground of such gestures and their presencing effects, I argue, requires ethnographically teasing out a metaphysics and metapragmatics of presence within and across film texts (since such presencing turns on an insistent citationality between films; see Nakassis 2016a, 188–223) in relation to theatrical moments of screening and uptake. It is this metaphysics and metapragmatics (qua Rajinikanth’s being, his “mass”) that constitutively enables indexicality in such scenes. Here, the ground of indexicality is not a self-evident relation of causality, contiguity, or co-presence, but a highly mediated relation of encompassment, transposition, and incorporation. The point, however, is not that the category of indexicality is unnecessary for the analysis of Rajini’s finger (or for ethnographic inquiry more generally); it is absolutely critical. Rather, it is that to account for the indexicality of Rajini’s finger prompts our critical reflection on the category itself, enabling its further development for ethnographic analysis and theory.

The Dialectics of Indexicality and the Lexical Register of Linguistic Anthropology

One might take the argument presented so far as a deconstructive critique of indexicality. In certain measure, it is. But it is more a provocation to continue to think with the category, to keep unpacking it, to not presume that its faults and fault lines can simply be resolved or dismissed (Derrida 1976, 13–14, 24; Nakassis 2013c). The ambivalence of indexicality (and whatever problems this entails) is an opportunity and invitation, a site for analytic and theoretical refinement and productivity. This is implicitly realized, I contend in this section, in the relentless splitting and supplementing that the category has been subjected to by those who have theorized its implications.

Consider how indexicality is often bifurcated into two terms. Peirce, for example, distinguished “genuine” (or “perfect”) and “degenerate” indexes (CP 5.75) for relations of relatively immediated dynamical (i.e., causal) connection and relations of mere spatiotemporal association (e.g., a punch in the face versus a pointing finger; Parmentier 2016, 54, 70–71). Quine (1969, 39–40), for his part, distinguished “direct” and “deferred” ostension for acts individuating concrete referents at the end of the pointing vector (e.g., a rabbit) and those denoting some indirectly associated referent or quality (e.g., whiteness). Through a similar trope, Elinor Ochs’s (1992) “direct” and “indirect” indexes differentiate social indexes (e.g., of gender) based on degrees of exclusivity and specificity (i.e., in/directness) vis-à-vis their object (cf. Lefebvre 2007, 231).²⁰ In each of these cases, the former of each pair is relatively more basic and immediated—more present, as it were—in relation to the latter, which is more mediated by or mediating of the context in which it is embedded (or embeds itself).

In this light, we might similarly consider Michael Silverstein’s important distinction of “presupposing” and “creative” (or “performative” or “entailing”) indexes, introduced in his 1976 “Shifters, Linguistic Categories, and Cultural Description” essay. Of the pair, the felicity of the first—exemplified in the essay by referential indexes—is a function of the representamen presupposing the independent existence of some aspect of the speech situation in which it occurs (namely, its object); by contrast, the latter—exemplified by nonreferential indexes—is a function of the index entailing the existence of some aspect of that context. This is not a hard split but a gradient, or better, two sides of a nonsynthesizing dialectic.

Of interest here is how “existence” is epistemologically and graphically bracketed in Silverstein’s text with consistency. In discussing relatively presupposing indexes (e.g., deictic phrases like this table) Silverstein notes that the referent of the token must be “identifiable, must “exist” cognitively, for the deictic itself to be interpretable” (Silverstein 1976b, 33; my emphasis).²¹ By contrast, relatively entailing indexes (e.g., deference indexes such as second-person pronouns in so-called T/V systems) bring “into sharp cognitive relief part of the context of speech” for participants and thus often “seem[] to be the very medium through which the relevant aspect of the context is made to ‘exist’” (ibid., 34; my emphasis; cf. Silverstein 2004, 626). Here indexicality is re-grounded in cognition (cf. Peirce’s “mental eyeballs”) and the question of existence and co-presence bracketed out from the frame of analysis, deferred to their presence to thought, or to the context-oriented norm, or “rule of use,” which governs the meaningfulness of the sign.

This passage suggests a further distinction in addition to presupposing/entailing, referential/nonreferential indexes: between the localizability or globality of the indexical act (also see Silverstein 1976b, 49). The felicity of some indexical acts is (seemingly) relatively more localizable “in” a particular segmentable form in relation to its context (e.g., a relatively presupposing referential act of deixis). By contrast, other indexical acts involve relatively more distributed, global arrays of signs. The latter is the case with nonreferential entailing indexicalities (e.g., signs of deference), whose entailments are less the effect of the particular “salient” or “overt” sign in question (e.g., a pronoun) than the total effect of a textual configuration of indexical signs (e.g., the pronoun, previous/subsequent address practices, bodily hexis, etc.). Speakers, however, tend to focus on such a salient, overt sign rather than the total text-level indexicality (Silverstein 1976b, 47–50, 1981; Agha 2007, 24, 286).

In short, certain forms of indexicality (presupposing, referential, localizable) are more likely to come before our “mental eyeballs” than others (entailing, nonreferential, distributed), and when directing our mind’s eyes (“brutely” or otherwise) the latter tend to present themselves as if they functioned on the model of the former. Hence the passage quoted above (Silverstein 1976b, 34) hedges on the mediating power of entailing indexicalities by saying that overt and salient signs “seem[] to be the very medium …” (my emphasis) through which “existence” is manifest, for the reality of such entailed existence, as analysis reveals, is a function of much more than what appears to (explicit) consciousness.²²

In the few cases in Silverstein’s text where “existence” is not put in scare quotes, it is deferred to the fact of successful (i.e., “proper”) uses, in which case the analyst, like the sign user/perceiver, can assume there to have been an “actual” (ontological) dynamical connection of token-sign and object (Silverstein 1976b, 33, 43–44; cf. later definitions of indexical presupposition as “appropriateness” to context [Silverstein 2003a, 193]).

In all these examples the ontological question of existence and co-presence is epistemologically relativized to cognition (and to the legisign, or “rule”/norm, of the index) and rhetorically distanced through scare quotes, even as it is tacitly put into play (if with some epistemological wiggle room) and assumed to be, in matter of fact, presumable upon. This assumption, of course, follows from the conceit of introducing indexicality in the first place: namely, to “anchor” language back into cultural praxis, into its actually existing contexts and events of use (Silverstein 1976b, 24–25, 1985a, 224–25; see n. 22). Through this conceit and its deferral, existence and presence are slipped in through the backdoor (through which the analyst subsequently follows close behind).

Notice the delicate, perhaps even knowing, dance of ambivalence. We find a hedging and deferral of the (ethnographic) question of “existence”—and its mediation by cognition and norms of use—and a positive commitment to its (analytic) promise. This deferral and avowal is diagrammed by Silverstein’s splitting of indexicality into token and type, presupposition and entailment, referential and nonreferential, and localizable and configurational, wherein, as above, the former of each pair presumes upon a more immediated relation of existence relative to the latter, which is both more mediating and mediated.

Or perhaps not so delicate and very knowing. Indeed, in its use, such splitting of indexicality has been leveraged to unambiguously critique the semanticizing, representationalist tendencies that have haunted studies of the social aspects of language—where creative, nonreferential, and nonsegmentable and configurational (token-)indexicalities are reduced to presupposing, referentialist, segmentable and localizable (type-)indexicalities—as evinced for example, in Silverstein’s (see, e.g., 1976b, 1979, 1985b, 2003a) withering reanalyses of classic studies of phonetic variation (Labov 1972), explicit primary performativity (Austin 1962), and honorification in pronominal systems (Brown and Gilman 1960) as various forms of folk ethnometapragmatics. In such cases, what mediates and constitutively enables indexicality (its cognitive reality and existence) misconstrues its actuality in the contexts of its occurrence, as present to (the “mental eyeballs” of) the semiotically sensitive analyst at least.

But if one half of this split seems to ideologically skew the phenomenon of indexicality, even as it constitutes it, the lesson is not to discard it for the other. Rather, the point is that there is an intrinsic dialectic here, one that can never rest on the immediacy of presence or existence (namely, “context”) and yet must presume upon it all the same (indeed, since the signs and their users do); which presumes upon the mediation of existence by semiosis and yet cannot take it at face value. And it is this tension, as I have argued, that is the Janus-faced ambivalence of indexicality itself, a tension mediating between immediacy and mediation.

This dialectic tension, however, is itself an indexical process, subject to the same ambivalence. This has also been tacitly registered in linguistic anthropology, worked through the copious analytic supplements to indexicality that the field has generated.²³ Such supplementing recognizes the inherently underdetermined, and hence mediated nature of indexicality as constitutive of indexicality’s actual effectivity in particular events of semiosis.

From the various terms that could similarly be worried, I consider Silverstein’s discussion of indexical order (1996, 2003a) and, relatedly, metapragmatics (Silverstein 1976b, 1985a, 1993a) and language ideology (Silverstein 1979, 1985b, 1998b).

In a now classic article, first published twenty years after his “Shifters, Linguistic Categories, and Cultural Description” essay, Silverstein (1996, 2003a) splits indexicality into “n-th” and “n+1-th” order indexes. Their dialectical relation comprises the earlier, related splitting of “presupposing” and “entailing” indexes and the supplements “metapragmatics” and “ideology.” They are related in the following way: any n-th order index (e.g., a second-person plural pronoun) indexically presupposes something of its context as a condition of its “appropriate” use (e.g., the existence and co-presence of a potential addressee), and in doing so may, under particular (co-)textual conditions, indexically entail something in the context of its use (e.g., honorification to singular addressee). As Silverstein shows, such entailments, however, are also inevitably a function of some metapragmatic, “ideological” regimentation, a “cultural construal of the n-th order usage” (2003a, 193–94) that may itself, through an interdiscursive process of enregisterment (or conventionalization), come to stand in a presupposing relation to the context (and n-th order indexicality) it itself reflexively mediates in the instance. Hence, to continue our example, the use of a second-person plural pronoun to honorificate a singular addressee can also function to index qualities of the speaker—as gentile, polite, old-fashioned, and so on. Such an n+1-th order indexicality, thus, stands as a “virtual” contextualization of the (n-th order) token index. This is in contrast to a (by implication) more “real,” “lower” order of contextualization; in our simplified example of the pronoun, to its “micro,” interactional text-in-context of prototypic face-to-face use.

Supplementing analytics like indexical order, metapragmatics, ideology show that the construability and efficacy of indexical acts are always mediated by some metafunction (as denoted by these analytics). Because the pragmatics of indexical signs are indeterminate out of context, they depend on some reflexive framework, or metapragmatics (or ideology, indexical order, etc.) to “fix” their otherwise shifty values (Silverstein 1998b). Such metafunctions stand apart from and yet constitute indexical relations and processes. At the same time, these metafunctions are also emergent from, that is, are mediated by (or presuppose) the indexical processes they in turn dialectically mediate (or entail).²⁴ Ideologies and cultural construals of semiosis (n+1-th orders) respond to, rationalize, and unfold alongside and thus are motivated out of the indexicalities they thereby constitute in dialectical and historical turn. The pragmatics of n-th and the metapragmatics of n+1-th indexicalities presuppose and entail each other, each serving as the other’s conditions of possibility, actuality, and generality.

Here we see the splitting and supplementing tendencies arising from the ambivalence of indexicality brought into clear focus, where indexicality’s splitting implies its supplementation by some metaindexical (or metapragmatic) framing that is itself indexical, and thus similarly split. The splitting and supplementing of indexicality by linguistic anthropologists theorizes this dialectic—itself the unresolved ambivalence of indexicality, as I have been suggesting—and in doing so conserves something of that ambivalence. Telling in Silverstein’s 1996/2003a article is how the hedging around “existence” and the deferral to cognition in the 1976 essay more or less disappears. Instead, the scare quotes and references to cognition are substituted for by the analytic supplements metapragmatics, ideology, and indexical order, in effect, internalizing that ambivalence to these dialectics of indexicality (even as it thereby externalizes cognition back into the social reality of such metafunctionalities). Consider, for example, how the notion of an indexical order imagines a stratified, even stratigraphic, layering of indexicalities with the more basic, originary indexicalities (and contextualizations) at a “lower,” more “micro” (face-to-face) and more present order entailing (the possibility of) “higher” and more “macro” “virtual” orders that, in dialectical turn, presuppose (and even replace or transform) those “lower” orders over interactional and historical time (cf. Silverstein 2004, 650, 2013, 2014a, 147, 2016; and see discussion below). While this framework troubles the representationalist ideologies and presentist metaphysics of the study of language, both persist, permanently problematized, in that framework’s very chronotopic form.

A dialectic view of indexicality implies that, in principle, it is indexicality “all the way down” (Silverstein 1998b, 128). Yet if we could muster the courage to peer “all the way down” into the hoary recesses of this bottomless origo, we might surprisingly quickly find the primal scene of indexicality (rock bottom, as it were).²⁵ That trepidation, and perhaps vertiginous excitement, is, I have suggested, lexically (and textually) manifest in our continually progressive Eleatic splitting and supplementing of indexicality. As noted above, this doesn’t mean that indexicality is a naively physicalist or presentist concept unnecessary for serious ethnographic or semiotic inquiry. Exactly the opposite. It is precisely because of its ambivalences that indexicality has been, and continues to be, one of the most fertile and necessary tools to theorize semiosis and social life.

Semiotic Realism and the Ambivalent Dialectics of Immediacy/Mediation

In the lexical proliferation that has grown our subfield’s technical register—that is, these linked movements of splitting and supplementing—indexicality’s ambivalence vis-à-vis immediacy/mediation is both belied and betrayed, but also productively worried to theorize communicative action. Such splitting and supplementing take up indexicality while framing it as always mediated in ways that defer and problematize any simple notion of an immediated presence or existential connection; and yet, at the same time, both such movements conserve this very notion, they hold on to it, if only perhaps by letting it go underground and out of sight.

But as I have suggested, this ambivalence should be kept front and center, made visible and put under ethnographic scrutiny. Rather than something to be resolved or done away with, this compulsion to split, defer, and supplement can and should be pushed continually further. It is through this compulsion that our field has made progress and grown its conceptual riches. And this is why, I have suggested, the very ambivalent ground of indexicality is itself an important if perhaps underexplored site for empirical inquiry. Problematizing this ground, it seems to me, is implied by indexicality itself and by the ways in which our own pragmatically deconstructive engagements with it have mined and massaged its intrinsic ambivalence as the basis of its various dialectics (namely, presupposing/entailing, n-th/n+1-th, and each of these dialectics in dialectical relation to metapragmatics, ideology, indexical order, and so on).

In this, I think, is a larger lesson about the epistemology and ontology demanded by our semiotic realism (EP 1:28–55; Silverstein 2004, 651)—namely, our stance that semiosis is the basis of whatever we call reality (and our knowledge of it), and vice versa. From this semiotic realism, of which indexicality presents a metonymic object lesson, we see that mediation and immediacy are themselves phases or aspects of an irreducible dialectic, each presupposing the other, each entailing the other. This mutual relation makes each “side” of this dialectic both necessary and impossible. And it is this that gives rise to and drives semiosis itself, as a movement of immediacy toward ever-approaching mediation and from mediation to an ever-receding immediacy (Nakassis 2013b). If indexicality as a concept is ambivalent, then, this is as it should be, for semiosis itself is ambivalent.

The grooves of this movement from Firstness to Thirdness, from Thirdness to Firstness is, as Peirce saw, Secondness. That is, actuality and existence are this movement, just as it is actuality that mediates the quality and possibility and generality and necessity of the worlds we inhabit and study. Perhaps it is not coincidental, then, that it is indexicality that has captivated linguistic anthropology’s “mental eyeballs” (though really, honed its disciplinary habits of semiosis: research, writing, reading, and teaching), for the ambivalence of indexicality is itself a perfect indexical (!) icon of the potencies and generalities, and thus reality, of semiosis as such.

One of the signature contributions of linguistic anthropology, exemplified in and developed by the work of Michael Silverstein, is not just to offer an analytical language to theorize this ambivalence and dialectic motility, but in doing so making both the driving engine of our own intellectual practice. As a semiotic practice spurred on by ethnographic engagements in worlds of Secondness (i.e., as a social science), linguistic anthropology has turned its own thought around this ambivalence, generating with each turn (or better, spiral) more and more developed interpretants of this particular ground. Looking back on this semiotic chain of interpretants as a way to look forward to their futures, Michael Silverstein has been a foundation and guiding light, providing pointed examples and real riches of analysis and critique. No doubt he will continue to provide us with all of the above in the years to come.