Identification and recognition

At a very basic level, names and name-like signs (“appearance, tone of voice”) serve as an infrastructure of person identification. Recognizable “distinguishing marks” are an affordance for the continuity of persons across events of communication; the establishment of the continuity of persons in social life (not to be confused with their coherence) is accomplished through identification. Phenomenologically, identification is prototypically realized as recognition, as an experience of the identity of this sensible “entity” (in the here-and-now event) with that one (in some prior there-and-then event). Recognition finds its apotheosis in the phatic experience of presencing the person themself—in recognizing the person you are waiting for to get off a train, say. Here recognition draws upon “event-based memory” (as opposed to “semantic memory”) in achieving identification. But the experience of recognition does not depend upon co-presence with the person recognized. We recognize a name dropped in conversation, a face in a photo. Or rather, putting the matter somewhat differently, recognition, whether or not it involves presencing the person in their actuality, is always (also) a recognition of signs of the person—Goffman’s “distinguishing marks” and “person-differentiating device[s]”.

Here I would emphasize the double meaning of the word “recognition”, which is heuristic for thinking a sociolinguistics of personhood:

First, “to recognize another” means to know who another is by virtue of perceptible indexes of them. One may recognize someone’s handwriting, or recognize a name; one may even recognize a plot, plan, or product as the handiwork of someone in particular. But the prototypical sign of recognition is the face, and recognition is most emblematic in face-to-face “contact”. Consider here the classic scene of a witness recognizing the accused in a courtroom or identifying them in a line-up. Recognition involves an intertextual harkening back to an earlier event of perception—the scene of the crime. It is a presencing that depends upon an earlier, now absent, presence.

Second, “to recognize another” is not only to recognize the other but to show a sign of that recognition to the person so-recognized and of their positive value for ego. The mutual demonstration of signs of recognition often constitutes an interactional opening or greeting. We can hear this second sense of “recognize” in speech act formulae that ritually regulate the order of talk in the proceedings of official bodies, like Congress: “I hereby recognize so-and-so …”. In many social worlds, recognition of the first kind requires recognition of the second kind. That is, one has an obligation to recognize (in the sense of “to call upon”) an other upon recognizing them (in the sense of “to realize who someone is”). The right or obligation to recognize the other often reflects social roles and statuses, as in the rural Wolof greetings described by Judith Irvine (Reference Irvine, Bauman and Sherzer1989 [1974]), where lower-status participants are obliged to initiate a greeting sequence. But greetings often additionally presuppose a prior intersubjective history, being only obligatory when we know the other and the other knows us—when we are knowns to one another. Take, for instance, the following description of greeting customs among Tingrinya and Amhara speakers in Ethiopia:

There are thus two levels to interpersonal recognition—the recognition itself and the sign of recognition. It is the second—the sign of recognition—that becomes a particularly fertile site of sociolinguistic and sociosemiotic elaboration for greetings and other “etiquette” of interpersonal encounter. As we will see shortly in a case study from Victorian England, whatever else they achieve, greetings function as rites of phatic recognition.Footnote ² But before turning to that case study, let us think a little more about how person identification is achieved beyond the face-to-face encounter, and in particular, how person identification relies upon rigid signs and presupposes social networks.

Names

Philosophers and conversation analysts speak of the problem of successfully referring to someone through the use of a proper name or a definite description, a demonstrative gesture or a pronoun of address, as a problem of “securing reference”. Analogously, ascertaining the identity of a person (place, thing, etc.) in social life depends upon securing identification.

Securing identification—knowing who someone is—need not depend upon any act of reference to the person in question. Almost anything can serve as a sign that achieves or participates in securing identification. Nevertheless, there is a class of indexical signs whose default function upon use is to referentially identify the same entity across all events (tokens) of the sign type’s occurrence. We call these signs “names”. Although names are not necessary for recognizing persons in all (even most) instances, they are a necessary condition on (or affordance for) human sociocultural life and its historicity. As a first pass at the problem, we can think of entities, like human persons, as bound variables and cultural symbols, categories, and concepts as functions applied to those bound variables. In sociocultural life, these bound variables are not “given” in experience but are themselves constituted, and this in part through acts of naming. The mediating role of names and naming is crucial to the way in which persons, and other unique self-same “objects” of experience, become invested with cultural qualities and are interleaved into social orders (Lévi-Strauss Reference Lévi-Strauss2021; Pina-Cabral Reference Pina-Cabral2015). Indeed, as cultural logics of naming underscore, the person-specific and the social-categorial are always dialectically entwined; they stand in an “affirmative relation” (Rupert Stasch, personal communication) rather than a mutually exclusive relation to one another (see the Conclusion for further development of this theme). Nevertheless, at the basic level of the semiotic organization of noun phrase types, unlike role designators, status terms, pronouns, or other nouns that refer to humans, names are unique among linguistic signs in presupposing the continuity of personhood. In this sense, and in addition to their indexical reference, names are iconic of personhood as an enduring essence, substance, or process—the identity of the sign-vehicle (“Bob”) across instances of its use is iconic of the identity of its semiotic-object (Bob, the person). This is reflected in the importance of name changes at important moments of identity transformation—when coming out as transgender, upon marriage, after the death of a relative (for Western Penan), or upon assuming the papacy.

Although names are exclusively and universally found in human languages, they nevertheless have clear precursors in other systems of animal communication. Numerous species have “signature calls” or other signals that point toward specific individuals. In studies conducted in the wild, these rigid signs have been shown to be produced either by, or in making communicative contact with, the index individual (Berg and Delgado Reference Berg and Delgado2011 for parrots; King and Janik Reference King and Janik2013 for bottlenose dolphins; and Pardo et al. Reference Pardo2024 for African elephants). Personal names share this same invariant or rigid structure. The difference is that the rigid relation between sign and object is, for humans, a referential one. That is, the rigid sign is subordinated to and enveloped within linguistic symbolism. Correspondingly, it is not only the signal producer (speaker) or recipient (addressee) that is indexed by the name, but the discourse referent. Unlike the “signature calls” of elephants or dolphins or parakeets, the occurrence of the name does not presume that the individual rigidly indexed by the sign is present. With names, the “signature call” is brought into the holism of language. Like other nouns, names pick out the arguments of propositions and serve as the subjects and objects of verbal and other predication. Correspondingly, there is a symbolic classification of human persons effected by the integration into language of rigid signs of them.

On Saul Kripke’s (Reference Kripke1980) influential “causal theory of reference”, personal names are referential indexes that rigidly designate their referents in all possible worlds, a designation established through (ritual) baptismal events that link the name to the name-bearer. This rigid indexicality distinguishes names from other referential indexes like pronouns and anaphors. The reference of pronouns like English “I” and “you” is reflexively mapped onto the very event of signaling in which tokens of the forms occur—they refer to the persons who occupy the participant roles of speaker and addressee, respectively, at the time of their occurrence. The reference of names, contrastingly, is “nomically calibrated” to its referent across all possible worlds (terminology after Silverstein Reference Silverstein and Lucy1993).

Networks

For Kripke, names are not, as Bertrand Russell had argued, “definite descriptions” that semantically classify their referents with such great specificity that one and only one thing in the world corresponds to the description. Rather, they are “rigid designators” whose reference, once established through a more or less ritually elaborate baptismal event, is indexically tethered to its named-particular henceforth and forevermore. The “rigid designation” of personal names foregrounds their historicity. Names are anchored in historical events of baptism and the transmission of the name that flows out from that initial baptismal event.

Names are transmitted through speech chains which, taken as a set, compose networks. The successful use of personal names presupposes that interlocutors are part of a shared interpersonal network (the discussion here follows that of Agha Reference Agha2007, 64–71). Employing the terminology proposed by Asif Agha, we can call the set of people capable of recognizing the person-reference of a name its social domain. All words and expressions (not just names) have a social domain; all vocabulary has been transmitted via networked speech chains. Nevertheless, the global patterning of the social domain of names differs radically from that of other linguistic units. The social domain of a noun like table is co-terminus, or very nearly so, with that of the entire English language community. The social domain of the name Richard Feynman (he was an important quantum physicist) is not. Even less so, the social domains of the names of our less-famous friends and family members. While the social domains of high-frequency words of the English language (e.g., foot, house, walk, early) are largely isomorphic with one another, the social domains of personal names (e.g., Harry Tegnaeus, Guy Lanoue, Courtney Handman) are highly variable and—at least in the massive national-language communities of which most readers of this essay are a part—almost always a small subset of the speech community.Footnote ³ In this way, the social domain of a personal name conforms more to a rhizomatic network than to a dialect map.Footnote ⁴

To briefly review, names afford the socio-historical continuity of persons. But they only do this because of the way in which persons similarly socialized to the rigid relation between name-and-referent together constitute a speech chain network, one that (for almost all names) is a subset of the broader speech community. The social domain of person-specific indexes maps onto networks rather than communities because of the way in which securing person-identification hinges on having an event-based and existential (node-to-node-like) relation to the particular person in question (whether directly or as deferred through a speech-chain). Again, we can contrast this to the acquisition of symbolic oppositions shared across a cultural community, where the opposition does not need to be grounded in the same specific particulars in the world in order to be acquired by different people. Putting aside for the moment the treatment of names (or other rigid signs), I want to argue that this complementarity between the social-network-of-a-name and the broader speech community is an important one to keep in view in thinking through a sociolinguistics of personhood more generally. The sociolinguistics of interpersonal networks require more of our collective attention and theorizing as linguistic anthropologists. Recent scholarship points toward the largely untapped richness of phatic communication as a domain of study (Nozawa Reference Nozawa2015; Slotta Reference Slotta2015; Zuckerman Reference Zuckerman2016). Interpersonal networks are higher-order formations emergent out of events of interpersonal contact, the ways they are chained together, and the phatic ritual orders that mediate them. As an example of the ways in which interpersonal networks are mediated by phatic infrastructures, I turn to the “Good Society” of Victorian England.

Victorian rites of recognition and the “Good Society” as a social network

To gesture toward the richness of this empirical field, consider the semiotics of interpersonal contact among elites in Victorian England (technically, 1837–1901), an elite network that called themselves the “Good Society” or simply “Society”, during this period. In the first decades of the 19th century, an elaboration of rituals of interpersonal recognition emerged within the context of the socioeconomic, political, and demographic transformations of British social worlds. In The Best Circles: Society Etiquette and the Season, Leonore Davidoff summarizes the shifts in historical conditions that led to the hypertrophy of rites of recognition thusly:

In “on the record” interaction, signs of recognition were not extended between elite aspirants without an explicit or formal introduction having been obtained first. Here, good form required that the intermediary ask permission of the parties involved, especially the higher-status individual, before making the introduction. Introductions were not binding: there was a “right of recognition”—women (as opposed to men) and those of higher status (as opposed to the lower-classed party) had the option to re-initiate contact on the next encounter—“[i]f they did not exercise this option, then the gentleman or the inferior was expected to keep his distance” (Curtin Reference Curtin1987, 131). (In the USA, at this time, introductions were even required at evening parties among guests, though this was not the case in England where the fact of having both been invited to the same “private” party was considered sufficient for legitimate social intercourse—hence the expression, “the roof is the introduction”.)

Introductions created links between nodes in the social network, very febrile links, but links nonetheless. An introduction served to “make people acquaintances” and thus entailed a future course of action—for instance, those introduced should bow or curtsy upon meeting each other. The metapragmatic term “a bowing acquaintanceship”—used in etiquette manuals of the era (Wouters Reference Wouters2007, 65)—described a loose tie, and its associated “tie-sign”, of this kind.Footnote ⁵ The formation of an acquaintanceship was a necessary precondition for making a social call at someone’s residence. This could not be done right away, however. Rather, one would first leave a printed calling card bearing one’s name and address with the butler or housekeeper. (If a corner of the calling card was folded down, this indicated that the indexed individual, rather than contracting someone to do the task, had left the card themself [Humphry Reference Humphry1904, 197].) Once again, the higher-status person had to initiate this step in the relationship, and calling cards had first to be reciprocally exchanged before a visit could be attempted. (The calling cards—text-artifactualized person-indices—were often collected within households. Etiquette manuals of the period “strongly recommended the maintenance of a register of cards and visits paid and received, including dates” [Curtin Reference Curtin1987, 141].) Many houses had set visiting hours in the afternoon during which callers would be received in the drawing room. Visitors should not over-impose themselves, however—15 minutes was considered the maximum length of an appropriate visit, and as a sign of its fleeting character good taste required “that a gentleman must keep with him his hat and stick and the lady her bonnet and shawl rather than depositing them with a servant” (Curtin Reference Curtin1987, 139). Introductions, followed by cards and house calls, were the prerequisites for invitations—signs of a more fulsome incorporation into social circles. There were different classes of events—“concerts, receptions, balls, afternoon parties”—but it was “[i]nvitations to dinner [which] were most honorific” (Curtin Reference Curtin1987, 143). These various “access rituals” were sutured together in such a way as to provide the higher-status person a means of egress from a possibly socially polluting relation with those not in the “Best Circles”. A rich metapragmatic vocabulary to describe the opening or closing of phatic channels was elaborated; for instance, to “snub” someone or to “cut” (cf. “cut off”) a relation was to withhold signs of recognition. To ignore completely one with whom an acquaintanceship had already been established was called the “cut-direct” (Curtin Reference Curtin1987, 144).

The class-inflected characterological attribute of “English reserve” was undergirded by this heightened sensitivity to the phatic function and constant monitoring of the signs of recognition one gave off. If reserve at one level “meant the suppression of emotional expression”, at another it entailed “the careful maintenance of public and private boundaries or, more simply, privacy” (Curtin Reference Curtin1987, 128). Elite Victorian English culture involved an elaborate social semiotics of personhood that relied upon a speech-chain logic of authorization whereby well-established in-group members mediated the introduction (and induction) of new entrants. Incorporation through rites of recognition into social networks that were hypostatically referred to by its members as, simply, “Society”, implied a transformative access to “social capital”, understood in the canonical and person-centric sense that Pierre Bourdieu gave to the expression.

Positively, by virtue of residing in the network, and negatively, because of the tight ritual restrictions on entry into it, “one of the most important privileges of being in Society is having access to a vast information network: information about jobs, investment possibilities, secret political decisions” (Davidoff Reference Davidoff1973, 37). “At the same time”, incorporation into society conferred honor and status in and of itself: “participation in the group is a reward and a badge of arrival into these positions, a public seal of acceptance into elite status” (Davidoff Reference Davidoff1973, 37). This second dimension was particularly palpable in the Durkheimian effervescence of elite events—balls at Almacks or promenades along Rotten Row. In and through the collective events of what was called “the Season” (the period from April to July during which the Queen had residency near London), rites of phatic recognition not only mediated a diffuse network of the “power elite” but ritually materialized it as the face-to-face community of the state. Here the role of introductions to Queen Victoria at Court was symbolically vital, an iconic-index of social incorporation through legitimizing contact with those higher in the social order.

In summary, the interpersonal networks of elite Victorian England were policed by elaborate rituals of incorporation. Ceremonial was necessary for establishing phatic rites of recognition and to mark degrees of phatic access between persons: from acquaintanceships signed with brief bows, to 15-minute afternoon home visits negotiated through the reciprocal leaving of visiting cards, to ornate and elaborate dinner parties (where the gender-paired procession to the table, “precedence” in seating order and placement constituted the central rite, at once figurative of status differentiation and of social incorporation into elite “circles” via commensality).

Rigid signs

If the speech-chain-dependent circulation of personal names is crucial to the particular semiotic organization of person-indexicality in language, the strategic deployment and withholding of tie-signs among members of the “Good Society” of Victorian England shows that person-indexicality is also socially ordering. Importantly, in both cases, persons have their social existence—are recognized and reside—within networks rather than communities. Having made this point about the intimate connection between person recognition and social networks, let me now return to the semiotic organization of person-indexicality by generalizing from the special case of personal names. Beyond names, there is a range of other rigid signs—not necessarily referential—that have the same invariant indexical relation to persons and thus can similarly secure identification. The visible face has perhaps the claim to the greatest universality here. Multiple lines of evidence demonstrate that a dedicated part of the cerebral cortex, located specifically in the fusiform gyrus, is specialized for facial recognition (Kanwisher, McDermott, and Chun Reference Kanwisher, McDermott and Chun1997). Facial recognition has been subject to evolutionary selection “for us and other primates, because [faces] provide the principal basis for recognizing individuals, as well as a high-bandwidth channel for social communication” (Connor Reference Connor2010, 764). The face is thus intimately linked to both social and communicative recognition.

Facial recognition has long been augmented by technological prostheses of various sorts—prostheses that sometimes call into question the transparency of the ways in which the visible face is linked to personhood. The way in which the technology of photography allows for the visible face to outlive the embodied person may be experienced as a boon—as when death photography was practiced on the American frontier to preserve personhood by capturing its photographic likeness. Or it may be experienced as uncanny and be avoided—as in those Aboriginal Australian communities where photographic images of deceased persons are tabooed, much as are their names. The topic of person indexing has risen to the fore in the contemporary moment in no small part because of the ways in which the social currency of some of the most taken-for-granted rigid signs of persons has been transformed by the Internet and by Artificial Intelligence (AI) technologies themselves trained upon data transmitted through the Internet—“machines built out of other people’s words” (Gershon and Handman Reference Gershon and Handman2024).

First, AI has undermined the fidelity of rigid signs of personhood. Before AI, the audible voice and the visible face were hard-to-fake “honest” signals of persons (terminology after Zahavi and Zahavi Reference Zahavi and Zahavi1997). And yet, under current socio-technological conditions, we cannot be sure that the voice that we hear in a recording or the face that we see in a video was “authored” by the person it purports to point toward. Presented with real persons in virtual (possible) worlds, we cannot be sure of the facticity of the represented (narrated) events whose mechanical reproductions nevertheless enable us to secure person identification. Because of the ways in which they afford the appropriation and counterfeiting of identity, AI technologies reveal the genres of identification that were until recently highly naturalized aspects of a background infrastructure for knowing who’s who.

Second, AI has undermined the intimacy of the networks in which rigid signs of persons circulate. Users of the Internet have long complained about a lost sense of “privacy” or voiced anxieties about the uncertain and never-ending circulations of comments, videos, or other online effluvia linked to their personhood. Newer technologies allow for securing identification in contexts where, historically, anonymity would have been expected. I recently (September 2025) traveled from Canada to the USA by plane. As an American passport holder, I was sent to a line specifically for American citizens. The border agent greeted me by saying: “Meet my assistant, Tony”. Tony was a camera, placed at about my head level on top of a long vertical pole. I was instructed to come a little closer. Tony took a photo. And then the border guard said: “Have a good day, Luke”. Importantly, he never once looked at my passport. There was a double edge to the encounter. On the one hand, the event figured American citizenship as a face-to-face community of acquaintanceship where citizens are on a first-name basis. On the other hand, it socialized me to the powers of the surveillance state to secure identification of me with or without my state-issued IDs.

Of course, science and technology have been historically entangled with identification in complex ways that I can only gesture toward here. Hilary Putnam’s externalism (an allied theory to Kripke’s causal theory of reference) frames scientifically discovered properties—the chemical composition of water, the genome of an elm tree—as providing the final backing for securing the identification of natural kinds. And, of course, forensic genres of identification do have authority and primacy in a division of identificatory labor, with dental records, DNA samples, fingerprints, and facial recognition technology providing privileged rigid signs of persons in legal and academic regimes of knowledge. Nevertheless, the focus on materialized corporeal traces in securing identification suggests that these technologies are revelatory only within a particular cultural ontology and cosmology of personhood—one in which personhood is reductively equated with “the unique organic continuity imputed to each individual” (Goffman Reference Goffman1971, 189). Comparative ethnography belies the limits of this model of personhood.

As a passing example, consider Greenlandic Inuit name-sharing relationships, as these occur within a cosmological logic of reincarnation. Greenlandic Inuit often name newborns after those who are recently deceased. Mark Nuttall’s account of Inuit naming practices includes a description of a specific kind of atsiaq or name-sharing relationship called angerlartoqut (or “one who has returned home”). This occurs where a child is named after a deceased sibling of the same sex. As in other systems of homophonous kinship, relatives transpose the referent-of-kinterm-use from the newborn onto his or her namesake—the deceased sibling in this case. Siblings will employ the kin terms that they would have employed for the deceased sibling—for instance, relative-age sibling terms may be switched, with a sibling addressing his newborn sibling as “elder” brother or sister. Further, the parents of an angerlartoqut will avoid his or her name.

Note that in this example, it is diffuse aspects of character and personality, rather than material traces like fingerprints and retinal scans, that serve as the rigid signs of personhood. In European Modernity, a materialist cosmology of personhood is hypostatically located in the corporeally bounded individual. But different genres of identification are possible where cultural ontologies of personhood are not bound by corporeality or reduced to the biological organism.

Rigid repertoires and person-indexing registers

As a number of essays in this Special Issue attest, rigid signs of persons may be bundled together and become elaborated into person-indexing registers—“registers whose non-referential indexical value is an individuated ‘entity’, prototypically, but not necessarily, a person” (Nakassis, this issue; also see Nakassis Reference Nakassis2025). The most salient examples involve widely known public figures (celebrities like film stars or popular singers, political heroes or villains). In these cases, the social domain of persons socialized to recognize the repertoire of signs as rigid indices of the person may be quite large. Celebrities are knowns to large swaths of people. (The deference which celebrities receive in “public” reflects the asymmetric and unidirectional social capital that flows toward them as a function of their recognizability and its social facticity [Collins Reference Collins2000, 29].) But person-indexing registers do not only point toward “public” figures—to persons we all know. Person-registers may be more intimate, involving relations of intense mutuality between members of relatively small social networks.

Consider here the normative proscriptions, attested in diverse languages-in-culture, on the use of certain people’s names. These restrictions are often at the root of more elaborate name registers where particular social others are not only respected through the avoidance of their names but also by the avoidance of homophones or near-homophones—words phonetically iconic—of those names (Fleming Reference Fleming2024, 83–134). Although actors may more studiously follow this linguistic etiquette in the co-presence of the target-of-avoidance or of his or her close kin, in normative framing the negative performativity of saying the name or of uttering any of a range of tabooed words situated in its sonic surround is understood to be constant and context-independent; the social indexical connection between sign and person, like the link between the name and its referent, is rigid.

These restrictions are often attested in kinship avoidance relationships. To give an example, Tobelo (Papuan; Halmahera, Indonesia) in-laws (moyoka) are supposed to avoid each other’s names and “words containing the names (or certain parts of the names)” (Taylor Reference Taylor and Masinambow1984, 132). Because of a marked absence of elaborate marriage rituals, these practices of name and word taboo constitute “the most ‘public’ guide to Tobelo marital status” (Taylor Reference Taylor and Masinambow1984, 132). By way of illustration of the Tobelo materials, Paul Michael Taylor offers the following hypothetical example:

The use of substitute vocabulary items to replace homophones (called hohono-words in Tobelo) signals deference toward the bearer of the name, and “true” parents-, children- and siblings-in-law can be fined for employing the offending lexemes. Alternatively, “[w]hen a person has truly been wronged … he may say a sibling-in-law’s or child-in-law’s name to his face, in anger” (Taylor Reference Taylor and Masinambow1984, 133). Semiotically, name avoidance operates according to a logic of iconic-indexicality: respect is enacted by avoiding phonological strings iconically resonant with the name that is indexically attached to the person of an in-law. Mirroring the sensorial interpersonal avoidances of touch and gaze so often attested in in-law avoidance, speakers attend to and circumvent those elements of everyday language that “participate” in the name of the in-law. (If persons are partible, so are the semiotic media—like language—through which relational-personhood-as-participation is achieved.) Underscoring this reflexive engagement with their own speech, “certain sound sequences of an in-law name may simply be hummed or skipped when pronouncing other words in which they occur” (Taylor Reference Taylor and Masinambow1984, 134).

It is not just names but a whole range of signs rigidly indexical of particular persons that may be tabooed in registers of kinship avoidance. Just as they are enjoined to avoid speaking one another’s names, avoiding kin may actively work to avoid beholding the countenance of the other. The face, which, as we have noted, is the visual analog of the name, is subject to avoidance. The Daribi say: “To look at one’s mother-in-law is like looking at the sun” (Wagner Reference Wagner1967, 173). Jicarilla Apache call avoidance kin—most particularly a man’s mother-in-law or a woman’s son-in-law—“one whom I cannot see”, and they contend that if one deliberately stares at one’s mother-in-law for too long, one will become blind (Opler Reference Opler1947).

To summarize, person-indexing registers mediate kinship avoidance. The mutuality of kinship avoidance is achieved through the avoidance of (signs of) the other person—their name, their face, their body. Rigid signs of the other serve as anchoring points for a more general phenomenological separation, the sacrifice of shared experience (Fleming Reference Fleming2024, 269). If Goffmanian sociology imagines a “private” sphere of “anchored relations” floating on a sea of stranger-sociality, closely related avoidance partners in smaller-scale societies actively produce their mutual estrangement through “effacement work” (Nozawa Reference Nozawa2020). As Rupert Stasch (Reference Stasch2009, 40) writes concerning the life-world of Korowai, for whom in-law avoidance is a ubiquitous practice, “social connection and estrangement [are taken] to be correlates rather than opposites”. Crucially, these modes of intersubjective relating are managed by means of semiotic registers of person.

It is not surprising that mourning rituals—enacted as they are by a close network of relations and focused upon a particular person—are also a site for the elaboration of person-indexing registers. Name-linked repertoires of verbal taboo and substitution are commonly attested after death (see, for example, Nash and Simpson Reference Nash, Simpson, Masek, Hendrick and Miller1981). But mourning typically entails consideration of a much wider semiotic range of rigid signs than just the name. For exemplification, consider the Western Penan communities of central Borneo as described in J. Peter Brosius’s brilliant ethnographic materials (Brosius Reference Brosius1992, Reference Brosius1995, Reference Brosius1995/Reference Brosius1996; see also Guilbeault Reference Guilbeault2024, 158–228).

Western Penan enact mourning through the avoidance of a repertoire of signs perspectivally and experientially anchored to the deceased person. They avoid places, words, foods, among other media, that were important sensibilia of the deceased’s life-world—phenomenological relics of the deceased. These sensorial experiences participate in (and, more ominously, potentially recruit) recently deceased persons. Mourning is a reciprocal relation between the bereaved and the deceased; “the pain of death is felt to be mutual: both the bereaved and the deceased long to be with each other again” (Brosius Reference Brosius1992, 445). The dead wish to return, and the living wish to die. Close kin who experience intense feelings of sorrow, misery, listlessness, and “extreme despondency” (muəng) must be closely monitored, and knives must be kept out of their reach, as it is thought that they will stab themselves if given the chance.

The key cultural concept informing mourning practices and the period of mourning (lumu) is that of the uban, the presence of a living being’s absence which attaches, in particular, to places.

These uban serve as points-of-participation or portals that mediate between bereaved kin and their newly deceased relatives. Two kinds of speech registers are anchored to uban.

First, there is a genre of laments produced by the bereaved and called pia pətosən ‘co-knowingness speech’ (pia means ‘speech’ in Penan). These are laments spoken with “rising and falling tone broken by brief episodes of sobbing”. The content of the laments is made up “of recitations of events that the bereaved and the deceased shared together, places they were both at, acts of kindness by the deceased, and their good qualities in general” (Brosius Reference Brosius1992, 425). These laments are triggered by sensorial experiences that evoke the deceased—that is, by uban of the deceased. As an example, Brosius recounts hiking with a man whose wife had recently died:

Second, there is an extensive set of enregistered avoidances of uban of the deceased. Particular spaces and places are avoided. Traditionally, the deceased was buried under the hearth of his or her lamin “household”, and that structure would itself be torn down, and the entire settlement would be abandoned. Even at the time of Brosius’s fieldwork, when Penan were living in more permanent settlements, he reports that they found “the lamin of someone who has died an unbearable place to be: the sight of the place where the deceased used to sleep or sit” (Brosius Reference Brosius1992, 432). Avoidances are not only enacted in physical space; forms of spoken language that formed part of the intimate life-world of the deceased should also be avoided. The names of the deceased and occasionally similar-sounding homophones, but also the names of their close kin, are tabooed. The changed names of close kin and other bereaved are called ngaran nasu “moved names”. Even the kin terms that would have been used by the deceased in referring to social others are avoided in referring to those persons. So, for instance, a man whose father has died should not be called “son” but rather by a birth-order term that presupposes that the origo-of-kin-reference is dead. These substitute kin terms, called ngaram lumu “mourning terms”, are employed in the place of the unmarked kin terms.

As this brief synopsis illustrates, Western Penan mourning registers are organized around signs indexically anchored to the life-world of the deceased. The cultural pragmatics of register use allows the bereaved to participate in the personhood of the deceased in and through the signs of their absence. For the bereaved and the deceased alike, it is upsetting to hear the names and kin appellations, to see the places, to taste the foods—in short, to experience what the deceased experienced during life. Respect for the deceased and for the bereaved is accomplished by a shared sacrifice of the salient signs of a life-world which the departed can no longer enjoy. But even if mourning repertoires are composed of rigid signs, their cultural pragmatics and poetics are not purely person-specific. Rather, these rigid signs are infused with social-categorial resonances of kinship and grief that color experiences of the traces of persons past. As Rupert Stasch (personal communication) emphasizes, the “Penan avoidance actions turn on personal indexes of the deceased, but they are also indexically creating the ruptured temporal present of simultaneous memory-based presence of the deceased and physical absence of the deceased, which is also indexing the mourner’s social and emotional condition and an overall temporalized social field of orientation to the condition of loss.” Much as with Korowai “mourners [who] understand media of social relating as channels of separation as well as contact” (Stasch Reference Stasch2009, 253), Penan avoidances involve the paradox of avoidance as both cutting and maintaining, but also indexically construing, a social relation with the deceased.

The known/unknown distinction in T–V systems

Even beyond person-indexing registers as such, a focus on “categoric”-type meanings in the study of language variation has occluded from view the important roles that “individual”-specific signs play in a wide variety of sociolinguistic systems and speech registers. To see this, consider that old chestnut of sociolinguistic analysis, honorific distinctions in personal pronouns or T–V systems.Footnote ⁶ T–V systems present an intriguing “formal-functional analogy” (Silverstein Reference Silverstein, Basso and Selby1976, 41): cross-linguistically the same linguistic forms (e.g., grammatically plural or third-person pronouns) accomplish the same social functions (i.e., the enactment of respect and the demarcating of social “distance”). Thus, in languages as disparate as Usan (Papuan), Djaru (Pama-Nyungan), Temiar (Aslian), Tamil (Dravidian), Russian (Slavic), and Mijikenda (Bantu), grammatically plural forms are employed to enact respect toward a unique discourse addressee or referent. And yet, although T–V systems compose tightly structured, analogous systems of social indexicality, they vary widely in how they incorporate person-relative and category-relative inferences and indexicality.

In South Asian language communities, the social-categorial character of T–V alternations, at least as articulated in widely voiced indexical stereotypes, is often patently clear. I provide a few examples in Table 1. And yet, a presentation of the sociolinguistic data in this manner glosses over the important role that person recognition has in undergirding such categorical social indexicality.

Table 1.

Stereotypes of pronominal honorification linked to rank hierarchy, gender, and kinship categories in some South Asian languages

Table 1 Long description

Three language columns list category honored and form employed. First column describes Telugu usages from the 1970s taking a Brahmin speaker as the origo. The second column describes Konkani usages from the 1970s, taking a woman as the speaker. The third column describes Santali patterns of pronominal use described as early as the 1890s. In Telugu, the high honorific (a plural form) is used in reference to a Panchayat member; the mid honorific, aayana, is used for a Brahmin; the non honorific form, atanu, is used for a middle-ranked caste member; and the anti-honorific form, waaDu, is used for a member of the so-called Schedule Castes. In Konkani, the honorific plural us used in reference to male in-laws; the non-honorific singular is used for a woman of the same age as the speaker; and the anti-honorific neuter gender is used in reference to a younger woman. In Santali, co-parents in law use the honorific plural in addressing one another; parent-in-law and child-in-law exchange the honorific dual in addressing one another and everywhere else the non-honorific singular is employed. Telugu includes four category rows, while Konkani and Santali each include three.

Sources: Telugu (Murty Reference Murty1979; Miller Reference Miller2013); Konkani (Katre Reference Katre1970; Miranda Reference Miranda2014); and Santali (Ghosh Reference Ghosh and Anderson2008; Choksi Reference Choksi2021).

In recent research in rural Kongu Nadu in South India, Constantine Nakassis and I have found important differences between the ways in which respectful language operates in village and town contexts. Distinctions in levels of respect are reflected both in Tamil linguistic media—like the distinction between the singular second-person pronoun, nī, and the honorific plural, nīṅka—and in non-linguistic signs of (dis)respect—who can enter whose house, eat with whom, or leave their banana leaf-plate behind for someone else to clean up after eating. In villages, these signs of respect are more closely tied to caste-group membership than in the town. Crucially, this reflects the fact that the indexicability of caste-as-social-category is contingent on interpersonal knowledge—on knowing who others are. When persons can be placed within social networks and located in terms of kinship ties of filiation or marriage—a condition readily met in village life but not in towns—even unknown social others can be ascribed caste-identities. Contrastingly, the social-categorial sensitivity of the Tamil T–V system is muted in town—and even more so in metropolises, like Coimbatore and Chennai. This reflects the relative opacity of caste in urban contexts. Caste group membership is something that is not asked about directly. (As one interviewee put it, “You cannot and you should not ask what someone’s caste is.”) While in the village (gramam) everyone can place everyone else within a genealogical network of relations, the relative anonymity of towns, but especially of cities, derives from not being person-indexicable in this way. Social actors are, of course, well aware of—and affectively experience—these differences between village and town. In village life, acts of respect and disrespect can be at once highly personal and simultaneously charged with a caste-relative coding. One young man who was in the village to visit his parents after taking work in Chennai described his mixed feelings about coming back to the village. A member of one of the castes whose members are systematically disrespected by the demographically dominant land-holding caste, he appreciated the intimacy and familiarity of home while he bristled at being put “in his place” by the same families that had always denied him respect based upon his caste identity. The point that I want to underscore here is that social-categorial indexicality often (not always, but often) depends upon person-indexicality.Footnote ⁷

In European language communities, T–V alternations are even more directly linked to person-indexicality. Modern European languages have proven stubbornly resistant to treatment in terms of categorical stereotypes.Footnote ⁸ So, for instance, in her article, “On sociolinguistic rules”, Susan Ervin-Tripp (Reference Ervin-Tripp, Gumperz and Hymes1986) tried to exhaustively model T–V usage as a flow-chart plotted with social-categorial distinctions like kinship relation, relative-age, and social rank. Nevertheless, she was forced to include parameters like “familiarity” to generate the correct outputs. Of course, “familiarity” is not a social-categorial distinction but rather one based precisely upon an interpersonal history of contacts between interlocutors—it is a pattern of usage that reflects a cultural elaboration of the known/unknown distinction.

In the early modern period, French vous, English you, and German Sie did reflect social-categorial distinctions linked to the feudal order—land-holding elites, for instance, could expect to receive V from serfs who would be addressed by T in return. In the modern period, those who diffused liberal, equalitarian ideologies linked to various political and religious movements actively sought to undermine the category-based, hierarchical, and non-reciprocal exchange of T and V. Quakers in the 17th century engaged in a universal use of thou that emphasized the equality of all humans before God (Bauman Reference Bauman1983). Jacobins during the French Revolution tried to legislate the universal use of tu on grounds of human égalité and shared citoyenneté (Frey and Frey Reference Frey and Frey2018). In the 20th century, soldiers in the Red Army successfully demanded that they be addressed by vy by their higher-ups, replacing the non-reciprocal rank-based pattern employed within the Imperial Russian Army (Comrie, Stone, and Polinsky Reference Comrie, Stone and Polinsky1978). These pronominal reforms—whether institutionalized or emergent—effected a steady erosion of the pattern of usage that sketched a (diagrammatic) icon of asymmetrical status: non-reciprocal address. In most contemporary European language communities, non-reciprocal address is virtually only attested in strongly age-gapped relations—in addressing very elderly persons with V but receiving T in return, or in addressing children with T and receiving V in return. Between most dyads, either reciprocal V or reciprocal T is employed. For example, “[m]ost Italians are likely to expect to receive the same address form that they give; the only clearly functional non-reciprocal relationships involve differences in age rather than status” (Bates and Benigni Reference Bates and Benigni1975, 271).

As social category-linked T–V usage has been undermined over the past two hundred and fifty years, European T–V systems have become increasingly organized around the interpersonal “closeness” or “distance” of the pair of interlocutors. Of course, social proximity is not divorced from roles and statuses. The shift toward reciprocal usage has not entailed the elimination of status-hierarchies in language, though it has masked them. Status-asymmetric role relations like professor/student, boss/employee, server/client still more often than not involve the use of V, only now couched within a reciprocal pattern of “distanced” address that does not make explicit (the nevertheless contextually evident fact of) whose status is higher than whose (Morford Reference Morford1997). Nevertheless, the T–V distinction is increasingly understood as a demarcating mechanism between an “inner” and “outer” sphere of social relations, between knowns and unknowns. This is seen in how “anonymous” versus “anchored” relationships (to return to Goffman’s language) are reflected in T–V usage. During the 20th century, V increasingly became the default form employed in addressee-reference between strangers in Western European speech communities (Clyne, Norrby, and Warren Reference Clyne, Norrby and Warren2009, 53, 55; Comrie, Stone, and Polinsky Reference Comrie, Stone and Polinsky1978, 255; Halmøy Reference Halmøy, Peeters and Ramière2009). This default usage underscores how the T–V system has been increasingly reoriented away from indexing social category and toward indexing interpersonal relationship. Intersubjective recognition as a cultural form—not simply the facticity of person identification—has become the central axis around which pronominal address has become organized.

Because the norm between unknowns is the use of reciprocal V, the switch to reciprocal T—elaborated in rites of “dispensation” (Paulston Reference Paulston1976)—comes to figurate the deepening of an interpersonal connection between once-autonomous individuals. Dispensation rites mediate between a “before” of reciprocal V and an “after” of reciprocal T. Their symbolism underscores how the central axis of these T–V systems hinges on the signaling of degrees of interpersonal mutuality. The German ritual of Bruderschaft trinken “drinking brotherhood”, variants of which are found throughout eastern and central Europe, involves partners interlinking arms and simultaneously taking a sip or a shot of an alcoholic beverage (see, for instance, Thornton Reference Thornton and Douglas1987, 108 on the use of schnapps in Austrian Bruderschaft toasting). Observe here how both the commensality realized in the shared sensorial experience of the alcohol and the figuration of the intertwining of bodies indexically iconize—much as in rites of, for instance, “blood brotherhood” (Tegnaeus Reference Tegnaeus1952)—the identification of the members of the dyad with each other.

To summarize, T and V in contemporary European language communities are not so much associated with particular social-categorial kinds (human types) as they are associated with particular dyads of individuals (human tokens). Indeed, these systems may make quite onerous demands upon “episodic” or “event-based” memory, as opposed to the “semantic” memory that psychologists see as undergirding social categorization. As an example, a Dean at a Swiss German University complained to me about how he has to remember, for more than 200 members of the faculty, whether he employs du or Sie with them, as slipping up—in either direction—can be insulting. Pronominal tokens are “tie-signs” that presuppose interpersonal histories of discursive interaction, dyad-specific speech chains.

Conclusion: indexical personhood

Although I have found Goffman’s distinction between the “categoric” and the “individual” incredibly helpful to pursue an exploration of person-indexicality, in many ways this framing is trapped within the binary of the individual versus society. Sociocultural anthropologists have been keen to relativize the concept of the “individual”, pointing toward the ways that persons only ever have their being in relation (Strathern Reference Strathern1988 is the touchstone text here). As the case studies in this paper illustrate, it is not just persons, but the rigid signs that point to them, that are “partible”. Take, for instance, name transmission traditions among Gê-speaking groups of lowland Amazonia (Maybury-Lewis Reference Maybury-Lewis1979). The possession of name sets entails kin-like relations between those who possess them. But it also brings along with it rights over particular dance moves, design motifs, and obligations to perform particular ceremonial roles, all of which, because they are sutured to the name set, become rigid signs of the name-holder. Names fill social roles, make kinship relations, and produce persons.

Indeed, names seem at once to disclose persons but also to encompass or “dividuate” them; names afford human tokens while at the same time making their existence contingent upon their “participation” within broader social types, series, and sets (Maybury-Lewis Reference Maybury-Lewis1979; Strathern Reference Strathern1988; Pina-Cabral Reference Pina-Cabral2015). In being named, people become identified … identified with gender (USA), with clan (Osage), with generational moieties (Mijikenda), etc. Institutions of name transmission (and rigid signs more generally) interleave the “categoric” and the “individual”.

The paradox of identification—at once individuation and essentialization, specification and speciation—recalls the tension between the philosopher’s and the ethnologist’s understandings of personal names. On the one side, we have the American philosopher, Saul Kripke, for whom names do not have semantic sense or descriptive backing. They are rigid designators whose reference, once established through a metapragmatic event of baptism, propagates out through rhizomatic speech chains that indexically tether them to their named-particulars. As we have seen, not only for names and naming but also for phatic rituals of greeting and address, it is these speech chains, conceptualized as social networks of person recognition, and not speech communities, that are arguably the fundamental sociocultural “units” for a sociolinguistics of personhood. Further, this structure of rigid designation corresponds to a semiotic-functional organization attested not only in referential indexes like names but also non-referential ones like faces, fingerprints, and even favorite foods. And yet we must resist the desire to equate a structure of reference—names as labels for “unique individuals” in the world—with an ontology of personhood. To this point, contrasting with the philosophical reduction of the name, ethnological research on name sets and naming conventions reveals a consistent tendency toward the sublation of the individual by classification in acts of naming—or better, the co-dependent origination or “triangulation” of personhood through naming institutions (Pina-Cabral Reference Pina-Cabral2010). In La pensée sauvage (newly translated as Wild Thought by Jeffrey Mehlman and John Leavitt), one of the central thrusts of Lévi-Strauss’s argument against the reification of “totemism” depends upon his observation that structures of classification do not stop at the group (the moiety, the clan, the lineage) but percolate all the way “down” to single humans. The French ethnologist’s book insistently asserts the lack of a distinction, for structuralist analysis, between human types and human tokens, or what he calls the “species” and the “individual”. After an exhaustive analysis of the social organization of name sets, he writes that “there exists a homology between the system of individuals within the class and the system of classes within higher ranking categories” and he asserts that “proper names do not form a simple practical modality of classificatory systems, which it is enough to mention after the other modalities” (Lévi-Strauss Reference Lévi-Strauss2021, 193). Rather,

(Recall that Kripke [Reference Kripke1980, 26–28] begins Naming and Necessity with a recuperative gesture toward John Stuart Mill’s proposition “that names have denotation but not connotation”.)

Our objective should not be to resolve the seeming conflict between Lévi-Strauss’s and Kripke’s accounts of names. Rather, it should be to work with this productive tension inherent in the Janus-faced character of names and indeed all rigid signs—names as social classifiers and names as rigid designators, names as culture and names as networks. We need to attend to the ways in which rigid signs interleave the personal and the social-categorial, to the ways in which we are unavoidably identified with culture through names and the other rigid signs of our indexical personhood.