Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, ANTHROPOLOGY ( (c) Oxford University Press USA, 2020. All Rights Reserved. Personal use only; commercial use is strictly prohibited (for details see Privacy Policy and Legal Notice).

date: 04 June 2020

Language as Social Action

Summary and Keywords

Linguistic anthropology is the study of language as social action. Linguistic anthropologists study how people use language, and how, in using language, people are also defining and displaying who they are, enacting their membership in particular groups, and bringing various types of truths into being. Language, then, is a set of practices that people engage in every day in numerous forms, which helps to define their positions in their families, communities, workplaces, schools, and even nation-states. How one speaks is not only who one is—it is what one does. This is possible because language is multifunctional, that is, it works in many different ways to connect people, convey meanings and feelings, move people to action, and define who they are. The major functions of language are the referential function, the emotive function, the conative function, the poetic function, the phatic function, the metalinguistic function, and the indexical function, which often overlap when people use language and are shaped by language ideologies, that is, the beliefs and attitudes that shape speakers’ relationships to their own and others’ languages, mediating between the social practice of language and the socioeconomic, historical, and political structures within which it occurs. Language use is part of what makes humans human, and as anthropologists, focused on how humans live and make sense of each other and the world, language should always be part of what anthropologists attend to and investigate.

Keywords: language, linguistic anthropology, multifunctionality, reference, emotion, performative language, poetics, phatic language, metalinguistic practice, indexicality, language ideology


Linguistic anthropology is the study of language as social action. To study language as a linguistic anthropologist is to study how people use language and how, in using language, people are also doing a number of other things, such as defining and displaying who they are, enacting their membership in particular groups, and bringing various types of truths into being. Language, then, is not just an abstract object of study that lives in textbooks and that one can have or not have; it is a set of practices that humans engage in every day in numerous forms (as you are doing now as you read these words), which helps to define their positions in their families, communities, workplaces, schools, and even nation-states. How you speak is not only who you are—it is what you do. This makes the study of language in use essential for any understanding of people and their lives.

Many types of scholars study language, and they tend to define it slightly differently. I often have to interrupt colleagues or students when I hear them describing what I do as “linguistics.” I stop them and ask them to leave the “s” off of the end of the word and add “anthropology” after it. “LinguisticS is an academic discipline in its own right; it focuses largely on language as an abstract, formal system. It has its own academic departments, conferences, graduate programs, and central concerns. Linguists study the rules of language as a system, looking at the particular sound systems of a language or languages, the norms for how words are strung together to make sentences, and the ways in which languages portray past or future tenses (if at all). But I am a “linguistic anthropologist,” and we study how people use language. Whereas linguists most often approach language as an object in its own right, linguistic anthropologists approach language through speakers. We begin from an understanding that language is a human universal: all humans have language, and all humans use it as part of their everyday lives. But we move toward understanding language use as incredibly specific in practice, differing from place to place, group to group, and even person to person.

A central tenet of studying speakers and how they use language is that linguistic variation is almost always socially meaningful. Pronouncing a sound in one way rather than another (saying “New Yawk,” for instance, rather than “New York”), adding jargon to a sentence or not, speaking briefly or waxing on, even choosing which language to speak when you know more than one—all of these may be meaningful. I say “may” because it depends on the context in which such variation occurs, as well as speakers’ own background, community membership, and linguistic practices and abilities. For some speakers of English, calling a meal “dinner” rather than “supper” may be meaningful; for others, it is not. For Italian speakers, a speaker’s accent may work to immediately anchor them to a particular place and the history of that place (“Aha! They’re from Rome!”); less fluent speakers may not hear a difference. When lawyers or doctors or engineers speak to one another about work, chances are good that they will use so many specialized words that an outsider would be hard pressed to follow what they are saying. To others in their field, this will simply be how “we” speak.

The social meaning of linguistic variation is also closely pinned to culture, or, we may say, communities of speakers. The meaning and use of language vary tremendously across cultural settings in ways that shape speakers, their relationships to each other, and their natural environment, structures of power and solidarity, and how speakers’ see the world. This includes such fundamental practices as how and why parents speak to their babies and children, how women and men and other potential genders may speak as themselves and to each other, how public speaking may—or may not—be undertaken, how conflict is enacted and resolved, even how people undertake romance and love. Linguistic anthropologists investigate such cultural variation through ethnographic fieldwork, that is, through extended time (usually at least a year) spent both observing people as they go about their lives, but also participating in many of the activities that make up those lives. They also record as much as they are able and are allowed, so as to be able to transcribe these recordings and see the linguistic details of how people are using language in those transcripts. This linguistic variation is meaningful in speech, but also in other forms of communication, such as signing, texting, and other mediated forms. Linguistic anthropologists have focused not just on speech, then, but also on various forms of writing, from letters to YouTube comments to court documents, as well as sign languages across a range of communities.

One essential thing that linguistic anthropologists have learned through these techniques is that language use is learned; that is, no one just speaks: they have learned and been taught how to speak in the right ways, at the right times, and in the right places for the types of people that they are. This process is as culturally varied as any other (e.g., some groups talk a lot to their babies while others wait for children to be able to talk themselves before they are engaged in conversation), but it happens everywhere in some way. This is foundational to why language use is a form of social action because we learn how to use language through practicing it in culturally appropriate activities.1

Language in Interaction

It is common to think of language as something “out there,” a resource that individuals draw on to express what they mean. If we as speakers—by which I mean those using verbal language, but also users of other channels of communication, such as signing or writing—can just select the right set of words and say them in the right way, we will get our point across. However, meaning is created among people in interaction and does not depend simply on what the speaker intends to convey but, just as consequentially, on what the addressee hears or sees (for signing) or reads. By addressee, I mean here the person who is being spoken to (“addressed”) and—usually—is expected to respond, to become a speaker as well. There are others, however, who may be involved in an interaction also, such as audience members, who will listen but are not expected to speak, and overhearers, who may be acknowledged or not.

Consider a group text chat where a number of speaking and hearing roles may be in action: when one person is writing, they are the speaker and the others are the addressees. These addressees will probably also become speakers (or writers) at other moments. At times, speakers may address one group member in particular, rendering that person the addressee and the others the audience. Members may take their phones and show the chat to outsiders who are not on the group text chat, such that these outsiders become overhearers of this interaction. They may report what happened in the chat to others as well. The group member may in turn share this with others in the group (as in “I just showed this conversation to my mom and she said. . .”), which may be problematic or unremarkable, or they may not report it, which can be similarly fine, or not.

Alternately, imagine having a conversation with a friend on a bus. The roles of speaker and addressee switch back and forth as one person tells a story and the other listens and then responds, maybe taking up their own story. Others on the bus may also be listening, intentionally or not. They may be other friends who chime in, acquaintances who listen with interest, or strangers who have no stake in the stories but wish for a little quiet on their bus ride home.

The point here is both to show some of the many ways in which people interact through language and the roles that they take up in interaction, but also to highlight that each of these participants may make a different type of sense of what is said. A speaker may intend to say something funny; their addressee or an overhearer may laugh, or take offense, or not get it. A story told to a close friend on the bus about an encounter with a teacher that day may be full of meaning to the addressee, but confusing or totally nonsensical to others who may overhear it. A group text chat may be full of hilarious good fun for those who participate in it, but someone who is not in the group but reads the chat may feel left out or alarmed or bored by it. In sum, speakers do not control the meaning of what they say, whatever their desires or intentions, which may be displayed in what and how they say something (“What I meant to say was. . .”), but not determinatively. The meaning of language, rather, depends on many other things, such as who hears it, the context in which it is heard, the relationship among speakers (and overhearers and those who are spoken about), as well as who else has uttered words or phrases in the past that are similar to what is said now, among other things.

This open-ended nature of language and its meanings and effects is because language is multifunctional, that is, it works in many different ways to connect us, convey meanings and feelings, move us to action, and define who we are. It has, in other words, many different functions. Important scholars have contributed to the development of this view, including Roman Jakobson, Edward Sapir, Benjamin Whorf, Mikhail Bakhtin, Michelle Rosaldo, Charles Peirce, and Zora Neale Hurston.2 They are mentioned here because readers may be interested in finding some of this foundational scholarship, but also because it is important to recognize where ideas originate. Many scholars have contributed to developing just how multifunctional language is, and in what follows, I will be synthesizing and bringing together ideas from them to lay out the major functions of language. While I will introduce and discuss each function largely individually, they are not mutually exclusive, but often co-occur in any particular instance or speech, or utterance.

Referential Function of Language

The referential function of language is often the most intuitively simple to grasp: it is the ability of language to mean in the most usual sense, or at least for those who predominantly speak English or other Western European languages. Individual words have meaning: we can look them up in the dictionary, or on Wikipedia, or ask a parent or a friend to tell us what a word means. “Cat,” whether spoken or written, in this sense “means” those small, furry animals with long tails that like to nap and may chase string. In the utterance, “I like cats,” each of these individual words has a meaning that can be looked up and must be decoded to reveal the referential meaning of the utterance, that is, that the speaker has a positive orientation toward this set of beings. Whether we are speaking, signing, or reading, the referential function is an essential part of what language does: it represents the world around us, as well as our thoughts, our dreams, imagined worlds, and past events.3

For the referential function to work, speakers must share a language to some degree. If you don’t know the three sounds that the letters c, a, and t stand for and what these symbols add up to when assembled in the right order to make “cat,” then you won’t know what that combination of sounds or letters represents. But this point is never absolute, as speakers may point to a cat as they say the word such that the referential meaning of the word becomes clear. Even when speakers share a language, they may not share the same inventory of words (you may know “cat,” for instance, but not “feline”), so reference may require working together to define a word or words.

A lot of people, from journalists to cultural anthropologists and historians, consider what people say to be the sum of what they mean; that is, they read people’s utterances (or writing) for their referential content, taking it as evidence of what is meant, generally prioritizing what is said over how it is said. Although the how may be occasionally reported as well (as in, for instance, “they said angrily”), the what is generally valued over the how. But the how often matters just as much as, if not more than, the what in interaction, and linguistic anthropologists take this insight as a starting point for their own research on language in use. The referential function (the what) is often only one part of what is happening when people use language.

Emotive Function of Language

Language is essential to how we express emotion, and the emotive function is how this occurs. We use words like “sad,” “distraught,” “elated,” “blue,” or “down in the dumps” to describe how we feel, utilizing the referential function of language to do so. But there are other ways in which we express emotion, such as through laughter, sobbing, glaring at someone, gasping, or whistling. These are all paralinguistic cues or clues, which may occur alone or with language.

Certain elements of language, like exclamations of all sorts, are highly emotive, indicating a speaker’s stance toward something that is happening or has happened or may happen; this is when the emotive function of language is activated or comes to the fore. The effect of words like “wow” or “dang” or “my lord!” is achieved not by adding referential information to an utterance (which is often non-existent, minimal, or beside the point with exclamations), but by expressing a feeling, reaction, or state of mind. In an utterance like “Wow! A cat!,” “wow” doesn’t add new (referential) information about the cat in question, as “In the window!” or “Stinky!” would. Rather, “wow” is part of how the speaker is expressing his or her feelings about that cat, its appearance before them, or perhaps their feelings about cats in general. It matters how they say, “wow,” of course—in wonder, in anger, in delight. Again, none of this is conveyed referentially but rather in how the word is said, which word is chosen, and what that word means in that context. If a curse word were uttered instead of “wow,” it still may express wonder, anger, or delight, but would also convey something different about the speaker and the context, such as that the speaker is the right type of person to be saying such words (not a young child, for instance) and is doing so in a context in which cursing is allowed (among peers, say, and not in a classroom), or perhaps even that they are doing so in violation of those and perhaps other norms. Curse words may act to intensify the emotion being expressed, as they are often considered “taboo language” or language that may only be spoken in certain situations by certain people. Violating those norms or pushing against them may also enhance their effect, highlighting the intensity of feeling being expressed.

Curse words remind us that the different functions of language may and do co-occur in particular instances of language use. Their emotive effect depends on being taboo, which in turn often rests on what they mean, as they often refer to bodily parts or functions, deities, family members, or sexual acts that should not be discussed openly.4 Other exclamations, which do not have referential meaning—or have very little of it, like “dang” or “wow”—also work very effectively in expressing emotions.

The emotive function may occur via individual words such as these, but also in entire utterances. How utterances are spoken—or written or signed—expresses emotions in multiple ways. Whispered, shouted, ALL CAPPED, spoken in a rush or in a slow measured fashion, riddled with curse words or hypercorrected to sound like a newscaster—these are all ways in which the emotive function manifests in language use. Consider how the phrase, “I’m fine,” may be very differently spoken and thus express different emotions. Said enthusiastically—“I’m fine!”—it can show a speaker’s positive feelings toward the day, her own mental state, or her reaction to what is happening. “I’m—sniff!—fine,” however, may indicate that the speaker is the opposite of fine, that he is distressed, in need of assistance, and upset. Said in a monotone—“I’m fine”—this utterance may pass unnoticed, as not really saying anything in particular about a speaker’s state, or may tell the addressee that the speaker’s state is not up for discussion.

In the case of any of these utterances of “I’m fine,” it is up to the addressee to choose their response. I don’t mean choose here in a necessarily conscious way. One rarely pauses to think to oneself, “Oh no! My friend is obviously upset and I need to ask them what is wrong and comfort them!” or, conversely, “Oh no! My friend seems to be upset, but I have to get home and I don’t have time to ask and get caught up in a long conversation, so I’ll just act like I don’t notice.” Usually, we respond according to the relationships we have with the speaker, the context in which such utterances are made, our senses of self and responsibility, as well as the cultural norms that shape how one should, or should not, respond.

Conative, Directive, or Performative Function of Language

Language also acts in the world, either of its own accord or in moving others to action. Roman Jakobson, who was one of the first to write extensively about the multifunctionality of language, wrote about the “conative” function of language as the ability of language to engage the addressee directly in some way. It is useful to divide this broader function into two more specific but highly interrelated functions: the “directive” function and the “performative” function. Both are ways in which language is action, not describing or expressing but doing.

The directive function is how language gets other people to do things, often through what can be called vocatives, or namings, and imperatives. Saying someone’s name may be a way to get them to do something and, perhaps especially, if that name is paired with a command of some sort. So, “Nathan! Stop it!” or simply, “Nathan R. Peréz!” may result in the addressee ceasing an activity. Here, versions of Nathan’s names are the vocatives and “Stop it!” is a command or imperative. This is a direct command, where what should be done is clear through the referential meaning of the utterance. The emotive function is additionally layered in here, as the speaker’s feelings about the importance of Nathan stopping whatever he was doing will also be expressed in the utterance.

Imperatives can be indirect. “It is too noisy in here!” may be stated as an observation, but may also result in the noise stopping. “I’m cold” may be both a statement about the speaker’s well-being and about getting someone to turn off the air conditioning. Requests are imperatives as well and may function on a spectrum of directness and indirectness, explicitness and implicitness. “Turn it off!” “Could you turn off the light?” “Would you mind getting the lights, please?” “Lights?” “Could we have it dark?” or simply, “Let’s get started” (if the activity is usually undertaken in the dark, as with a presentation of a film, for instance) may all have the same directive effect of getting someone to turn the lights off.

The relative directness and indirectness of directives are connected to norms of politeness and the relative status of speakers. I sometimes do an exercise in class where I see if my students can come up with a way to get me to turn off the lights for them. I often get them to turn off the lights in the classroom for presentations or films by using the phrases listed in the previous paragraph, and indeed, may have just done so to show what directives are when I ask them how and if they could get me to do it without violating the often unspoken norms of a college classroom. Any requests they can come up with I can usually put off onto another student (e.g., “That’s right, we do need the lights off. Amira, would you mind getting them?”), for classroom norms mean that the professor usually gets to direct others, and not the other way around. The only possibility that they can come up with that may actually work is if I were to be already standing next to the light switch and no one else were closer, and we all needed the lights to go off. And even in that case, it would have to be cloaked in extensive politeness, with pleases and thank yous and non-command-like phrasing, such as “could you?” or “would you mind?”

The use and efficacy of the directive function of language, then, is a form of social action that is deeply connected to the contexts in which it occurs. Certainly phrases such as “Do it!” or “Stop!” or “Quiet!” work to direct the actions of other people, but their deployment, as with the use of all directive utterances, will be shaped by the conditions under which they are uttered. In other words, their use must be calibrated to meet the social rules and expectations of who can get whom to do what, and under what circumstances. This is always culturally specific, as different groups will have rules about who can move others to action and how, as linguistic anthropologists like Michelle Rosaldo have shown. Language may have the capacity to direct, but it is people who must do the directing, and in ways that will be more or less effective.

Likewise, the performative function of language is highly dependent on context. The formulation of the performative function of language comes from the work of philosopher John L. Austin, who wrote about how, while some language describes the world (the referential function), other language acts on the world, performs actions. He pointed to utterances such as “I promise,” “I bet,” and “I now pronounce you husband and wife” that did not describe actions that lay outside of language but were actually the actions themselves. To say “I promise” is to enact the promise; “I bet” (often with a handshake) seals the deal; being pronounced husband and wife changes the status of those two individuals. All of these utterances must be said under the right circumstances to work performatively, or “felicitously” as Austin put it. I can pronounce anything I want, but if I don’t say the right words during a marriage ceremony to two people who intend to get married and have the authority myself to marry them, then it does nothing. Similarly, a promise with fingers crossed behind one’s back may nullify its force, and failing to shake hands may put the commitment underlying the bet in question.

While this may make it sound like there are distinctive boundaries between performative and non-performative language, in fact, the lines are extremely permeable. Judith Butler, writing about the performative nature of gender, argued that statements like “It’s a girl!” by a doctor as she holds a newborn are crucial to making that newborn into a girl. They also are part of what produces femininity more broadly, as something on its own as well as distinct from other achievements, like masculinity. Gender itself is performative in this sense, produced by utterances and other actions, like dressing a child in blue or pink, admonishing a little boy not to cry, or discussions about how and if to use pronouns like him or her or they. Consider, for instance, the growing tendency for some writers of emails to include in their signatures the pronouns they prefer, with phrases such as “pronouns in use: she, her, hers” or simply “pronouns: she/her/hers.” These are descriptive, referring to things that happen (such as others referring to these email writers using “her” or “she”), but also performative in that they help to produce the gender of the email writers as normatively feminine. Similarly, newspaper articles that use “they” to refer to nonbinary-identifying individuals (as The New York Times began to do in 2018) don’t just describe these individuals—they help to make them into members of this relatively new category of people while simultaneously contributing to the formation of this category.

Poetic Function of Language

In some contexts, how language sounds or looks matters to what it means and does as much as any other aspect. This is the “poetic” function, when attention is drawn to the form of the language at hand. Song lyrics, rhyming couplets in poems, even the delivery of the punchline of a joke are all instances in which the poetic function of language comes to the fore. Song lyrics are composed so that their form is just as important as their content. How a song flows, its words rhyme, or how it turns a phrase so that the song will get stuck in your head, for instance, are as important as what the words say (the referential function), if not more. Poems adhere to particular formats, such as sonnets, limericks, epics, or even free verse; otherwise they don’t count as poems, as there are rules surrounding length, rhymes, syllables per line, topic, and even the way the speaker portrays themselves, depending on the poetic form or tradition being adhered to. Similarly, there are rules about how jokes can be structured: a knock-knock joke only counts as such if it follows the familiar pattern of call and response, followed by laughter (or groans, as the case may be). All these rules are the poetic function at work, for they have to do with what we can call the structure of an utterance or set of utterances (as a song or joke would be), and the relationship between how these are built and the effects they have on those who encounter them, such as a laugh or singing along. There can, of course, be good and bad songs, better or worse poems, hilarious and awful jokes. These judgments depend on individual evaluations, but also on how well such examples successfully follow culturally appropriate poetic norms and expectations.

The poetic function of language is often highlighted and important when language is performed or is part of performances, when the sound of language matters. From play performances to hip-hop concerts and opera recitals, those uttering the words (or singing them) are supposed to get them just right and adhere to an already established structure for the language they use. With certain poems, how they are laid out on the page may also contribute to their effect and how readers encounter them. In signing, the repetition or alteration of particular handshapes may impact the emotional effect of an utterance or a poem. We may be moved to action by a well-constructed phrase, such as President John Kennedy’s well-known plea to “ask not what your country can do for you, ask what you can do for your country,” which is still so widely recalled in large part because of how it is built, with repetition and parallel structuring across its two parts. But the poetic function may be present even in everyday language, which we may recognize when we appreciate someone’s turn of phrase or observe that someone “speaks well,” even when we may be hard-pressed to pinpoint exactly what we mean. This would be the poetic function of language, resonating through how an utterance sounds to shape its effect.

Phatic Function of Language

Sometimes language is about connecting speakers above all else. This is when the “phatic” function of language becomes important. The phatic function of language is the way in which we use language to attend to the channel among speakers. Think about a mic check (“one, two, three—is this thing on?”), and all the ways we use language to make sure that the metaphorical mic is on and people are engaging in an interaction. Greeting routines are extremely phatic in that there is very little or no reference happening in phrases like “Hey!” or “What’s up?” or even “Good morning.” Instead, speakers are opening a channel of communication—“Hi there!”—and then recognizing that the channel has been opened through responding in kind—“Hi yourself!” Back channel cues (those words or phrases that are uttered to show that an addressee is listening, such as “mmm-hmmm” and “Right” and “I see”) are also phatic, demonstrating that the channel is open and functioning. Asking “Are you even listening?” or “Can you hear me?” are phatic as well, as questions like these draw attention to the channel and its perhaps precarious or questionable state. Leave-taking routines like “Bye!” or “Catch you later” are phatic in that they safely close the channel. Exchanging farewells is usually the socially acceptable way to exit an interaction, in effect saying, “We’re done here, right?” without the social risks of breaking off an interaction too suddenly. As with other language functions, how the phatic function is undertaken and manifested in greetings and leaving-takings is culturally shaped and socially consequential. In some communities, there must be several rounds of “goodbyes” for an interaction to end satisfactorily; in others, one round of “Bye!” said by each speaker may suffice. But phatic language is often a necessary part of starting, maintaining, and ending interactions, and forgoing it can cause problems (think about how problematic hanging up on someone without saying goodbye is!).

Beyond these particular categories of utterances, certain phases or types of interaction may be largely phatic, as when strangers or acquaintances talk about the weather, or “small talk” in general. What is said matters little in such interactions; it is really only that things are said at all, that the channel of communication is open, that is important. Topics in phatic interactions are often very narrowly defined—nothing very personal, no subjects that may be controversial—so that speakers can participate without what they say really mattering. Phatic interactions may also be a way to open up other types of conversations such that talk about the weather may end when one speaker says, “So what I really wanted to talk to you about was. . .” or something similar, to cue that a different type of interaction—with different linguistic functions at the fore—is being shifted into.

In many ways, phatic utterances conjure the interaction and transform people into participants in a particular interaction. To not respond to “Hello!” is to refuse the implicit invitation to interact and to not open up the channel between you and the other person. To keep quiet and not nod and murmur things like “Right . . . I get it . . . and then?” while someone else tells a story may make one appear to be a bad listener, or not interested, or a snob. People may be held accountable for not participating in phatic communication (“Why didn’t you say hi to me yesterday?”), as this may be read as a refusal to engage and meaningful as either willful or accidental. Such accountability works differently for different types of speakers: small children in mainstream US society are instructed to greet people and say goodbye, but expectations of their behavior are different than that afforded to adults, for instance.

Phatic behaviors may also be highly visible as markers between different types and groups of speakers, as, for instance, particular types of greeting routines may be the norm among certain people but not others. During a class project, several of my male students realized that while they had particular stylized handshakes paired with specific verbal greetings with which they greeted male friends and acquaintances, they never did these handshakes or used these greetings with their female friends; instead, they hugged them and said “Hey!” In class discussion, we observed that they did neither with me (indeed, they laughed at the thought). A simple wave and minimal verbal greeting (“What’s up, Professor?”) was how they greeted me and all their professors, but this would have been weird or inappropriate with their friends, both male and female. Phatic activity, then, was one way that they enacted different types of social relationships and entered into different types of interactions.

Metalinguistic and Metapragmatic Functions of Language

Language also has the ability to refer to itself. This is the “metalinguistic” function of language, when language reflects upon language. Critiquing an accent, correcting someone’s spelling, gushing over how well someone speaks, as well as discussions of particular languages as being hard, sexy, or sounding guttural are all instances of the metalinguistic function of language at work, as are labels used for linguistic activities, such as gossip, dirty jokes, compliments, or interrogations. It is also metalinguistic to instruct someone that “we” don’t talk like that, or that they are speaking too loudly (or too softly), or when signs are posted that inform people that languages other than English are not welcome. We can think of these instances of monitoring or encouraging or curtailing someone’s speech as forms of social action because they have effects on what people do, but also because they are often expression of ideas about right and wrong ways of speaking—and being or acting. To say that someone’s language is not welcome is essentially to tell someone that they are not welcome.

It is possible to expand the metalinguistic function of language to encompass not only reflections on language (“Navajo is a beautiful language”), but also on what people are doing with language (“The use of Navajo is on the rise” or “We only speak Navajo here”). This is the “metapragmatic” function of language, pragmatics here being the ways in which language in use functions. To study pragmatics is in a way to study the traffic rules of how language is used, such as how to thank or be thanked, what types of address terms can be used in a particular situation (“Mom,” “Ms. Shu,” “Sweetie,” and “Uncle John” are all address terms), or how to use emojis when texting. People reflect on these rules for using language, which is metapragmatic. Metapragmatic utterances include telling some not to interrupt or that they should say please or thank you. They may also include reflections on how a particular person should be addressed in an email or letter (by Dr. or Professor, rather than first name, for instance), admonitions to speak only when spoken to, or that bilingual speakers should avoid code-switching (moving between their languages) in particular situations.

The metalinguistic and metapragmatic functions of language may be particularly involved in helping novices learn how to use language, but also in delineating boundaries between groups, including and excluding particular speakers. To insist that “we don’t talk like that” may be both an instruction for someone who is relatively new to this type of talking (like a child) as well as a way to indicate that there are people who are not like us, who do speak in that way. In doing so, the meaning and effect of metalinguistic and metapragmatic language (which, of course, will also involve other functions, such as the referential and the emotive, for instance) is heavily tied to the contexts in which it is deployed.

Indexical Function of Language

The meaning and effect of language practices are indeed always closely bound to the contexts in which they are used, and the “indexical” function of language, which emerges from language’s connections to context, is an essential part of how language in use works in contextually specific ways. The concept largely comes from the philosopher Charles S. Peirce, who saw that how language works—or more broadly, how meaning-making works—is variable, interactional, and inherently processual. Indexical meaning is produced through contiguity (being in proximity and/or connected to something) or causality (being caused by something), that is, utterances are contextually situated and take meaning from the specifics of that context.

Indexes point to something else, which is their “meaning.” Smoke is a physical index, pointing to its source (or cause): fire. An accent is a linguistic index, pointing to the origin (source) of a particular style of speaking, often related to where the speaker is from. Words like “tomorrow,” “him,” “this,” or “that” are indexical, as they point to something that can only be understandable from the context. Which day is being picked out among all possible days as “tomorrow” depends on which day that utterance of “tomorrow” occurs. For instance, I can write, “Tomorrow will be Saturday” and that will be true as I write this (today is Friday). But it may or may not be true for you, the reader, as you read it—that will depend on which day is “today” for you. If you say “tomorrow,” that word will index, or point to, a totally different day (Wednesday, perhaps). Similarly, when someone says something about “him,” the person being picked out in the world will only be decipherable for those who know which “him” is being pointed to. It will be clear to those in the interaction because they will know, for instance, that someone named Wilson is being discussed and that “him” is Wilson.

“Him” and “tomorrow” are indexes in that one must be part of the interaction in which they are uttered for them to “mean” the same thing, that is, they pick out or point to the same thing in the world. Thus these words—whose meaning also depends on their reference, as you need to know the definition of “tomorrow” and “him” to understand them when they are deployed—are particularly indexical, but all language in use is potentially indexical, and frequently in ways that help to define who we are, what we are doing, and why.

Let’s recall cursing, which may be an important part in how emotions are expressed. Cursing is also indexical of certain types of people, activities, and ways of approaching the world. The appropriateness of cursing is contextually defined; what may be fine on the ball field or during an argument may be highly objectionable in a classroom or at the dinner table. This is a first clue that cursing is indexical: its use in different contexts potentially changes its meaning and the effects it may have. Another clue is that when people curse, it makes them sound like particular types of people (such as “potty mouths,” “nasty women,” or just “teenagers”). These types are highly variable across communities of speakers as well as within them.

For people who speak more than one language, choosing one language over another is highly indexical. For instance, if a community of speakers uses one language at home with their family and another at work, then the use of those respective languages indexes—points to—those particular contexts. When those speakers use those languages in those contexts, they sound and feel right, appropriate, and unremarkable. When they use those languages in other contexts, they may take these associations with them such that using the work language outside of work may make one sound authoritative, or business-like, or make the context in which it gets used seem like a place in which to get work done. To switch and speak the work language at home, and vice versa, could potentially result in dissonances, as the home language may sound and feel too intimate for work, while the work language may sound and feel too formal and authoritative for home. This happens because those languages continue to index the contexts in which they are usually used.

The indexical function, which can also be called “indexicality,” permeates language use because as speakers, we are almost always orienting ourselves to each other, to the contexts in which we are interacting, to the roles that we take on in these contexts (and are available to us), and to any number of relevant past events. But the indexical function is shaped by other things as well, including the historical, socioeconomic, political, and cultural conditions within which language use occurs. For instance, to understand what is indexed by a particular accent in English—one associated with Alabama, or Liverpool, or New Delhi—in any specific context of use, we need to understand the history of the people who speak in that way as well as their places within contemporary political, economic, and cultural conditions, and the relationships among speakers (are they from the same group, a different group, another country?). For this type of perspective, we turn to language ideology.

Language Ideology

The functions just laid out are how language works within interactions. But there is another essential aspect of how language works as social action that linguistic anthropologists and others call “language ideology.” Language ideologies are the beliefs and attitudes that shape speakers’ relationships to their own and others’ languages, mediating between the social practice of language and the socioeconomic, historical, and political structures within which it occurs. They are not just ideas about language (although they are certainly these), but are also the combination of these ideas and the ways in which these ideas shape how people use language, on the one hand, and how these ideas contribute to and are an essential part of upholding structures of power, on the other.

An example here will help. In my research in a community in northern Italy, where speakers use both standard Italian and the local language, Bergamasco, people have very strong ideas about language, as many people do: what sounds good and bad, sexy and dowdy, prestigious and rude, homey and formal. These ideas show up in lots of places, such as in everyday conversation when people may comment on their own or someone else’s accent or turn of phrase; in public events, when a local politician may, in Italian, praise Bergamasco but also apologize for his own heavy Bergamasco accent; in print, in newspapers that praise local play performances in Bergamasco or Italian; or on websites, where speakers may discuss someone’s language in the comments section. Often, these comments align the two languages with particular constellations of values. Italian is seen as the language of the Italian nation-state, education, and modernity, making it a good language to speak in public contexts such as school, work, or public events. Bergamasco is seen as the language of the place and its past, and speakers often find it homey and intimate, but also crude and sometimes embarrassing.

These judgments, which speakers make both explicitly (“Please excuse my rough Bergamasco accent!”) and implicitly (such as consistently replying to an interlocutor’s Bergamasco with Italian), shape how and when speakers use the two languages. These ideas about language emerge out of the construction of the modern Italian nation-state in which Italian was and is held up as the only true, fitting language for its citizens. Thus the views that portray Italian as the best language in various ways (prestigious, modern, desirable) are part of a collective order that constructs membership in the nation-state—and the related privileges thereof—as at least in part dependent on speaking Italian. Local languages like Bergamasco (and there are many of them, most descended from Latin and whatever was spoken in particular places before the Romans arrived) are seen as linked to distracting local allegiances, where being from a place (and sounding like it) is akin to favoring that place over all others. There’s a long history of this type of friction among local places as well as between the local and the national in Italy; the language ideology that prizes Italian over Bergamasco emerges from but also upholds this history as relevant now. And since the 1950s, it has been a key part of why fewer and fewer people choose to teach their children Bergamasco as they are growing up. Therefore these are not just ideas about language; they are ideas that shape what people do with language that are also part of how the nation-state of Italy makes the people who live within its borders into citizens, or at least some of them. Notice that the many newly arriving immigrants who speak other languages are totally erased from this picture, as are their languages. These speakers and their languages are basically made invisible in this language ideology, as they are in many political debates.

Language ideologies are enacted and expressed through the various functions of language. Bergamascos often metalinguistically reflect on the referential differences between Bergamasco and Italian, pointing out words that exist in one or the other language, but not both. A Bergamasco accent indexes that a speaker is from this place, as does the common exclamation “Póta!” (a phatic utterance) that speakers may say to express surprise or dismay. Poetry written in Bergamasco is regarded as simple but emotionally affecting, and writing and reciting it are ways in which speakers may performatively enact their Bergamasco-ness.

Multifunctionality and Social Action

Language is a form of social action because it is multifunctional in the ways presented here. As a social practice that communicates referential content, but also expresses our feelings, connects us to contexts and to each other, enables us to performatively enact who we are, and situates us within sociohistorical and political economic situations, language use is part of what makes us human. It is also an essential part of who we are as individuals and as members of our communities. As anthropologists, focused on how humans live and make sense of each other and the world, language should always be part of what we attend to and investigate.

Further Reading

Austin, John L. 1962. How To Do Things With Words. Cambridge, MA: Harvard University Press.Find this resource:

Bakhtin, Mikhail. 1981. The Dialogic Imagination: Four Essays. Austin: The University of Texas Press.Find this resource:

Butler, Judith. 1990. Gender Trouble: Feminism and the Subversion of Identity. New York: Routledge.Find this resource:

Cavanaugh, Jillian R. 2009. Living Memory: The Social Aesthetics of Language in a Northern Italian Town. Malden, MA: Wiley-Blackwell.Find this resource:

Hurston, Zora Neale. 2018. Barracoon: The Story of the Last “Black Cargo.” New York: Amistad.Find this resource:

Jakobson, Roman. 1960. “Closing Statement: Linguistics and Poetics.” In Style in Language. Edited by T. Sebeok, 350–359. Cambridge, MA: MIT Press.Find this resource:

Peirce, Charles S. 1955. Philosophical Writings of Peirce. Edited by Justin Buchler. New York: Dover.Find this resource:

Rosaldo, Michelle Z. 1982. “The Things We Do with Words: Ilongot Speech Acts and Speech Act Theory in Philosophy.” Language in Society 11 (2): 203–237.Find this resource:

Sapir, Edward. 1985. “Selected Writings in Language, Culture, and Personality. Epilogue by Dell H. Hymes.” Edited by David G. Mandelbaum. Berkeley, CA: University of California Press.Find this resource:

Whorf, Benjamin L. 1995. “The Relation of Habitual Thought and Behavior to Language. In Language,Culture, and Society. Edited by B. Blount, 2nd ed., 62–84. Long Grove, IL: Waveland Press.Find this resource:


(1.) This process may occur differently across individuals who are differently abled, a subject that has been richly studied elsewhere and is beyond the scope of this article.

(2.) For additional reading, see the “Further Reading” citations in that section.

(3.) Philosophers, such as Gottlieb Frege, have long engaged with the referential function of language as its most important, or even its only function.

(4.) This is a highly culturally specific notion, which varies greatly in terms of local community rules about who should say what, and to whom. What may be accepted and appropriate in one cultural setting may be objectionable and beyond the pale in another, and vice versa.