How to distinguish a wink from a twitch

This work is licensed under the Creative Commons | © Jack Sidnell, N. J. Enfield. ISSN 2049-1115 (Online). DOI: http://dx.doi.org/10.14318/hau7.2.038


How to distinguish a wink from a twitch

Jack SIDNELL, University of Toronto

N. J. ENFIELD, The University of Sydney

Comment on Duranti, Alessandro. 2015. The anthropology of intentions: Language in a world of others. Cambridge: Cambridge University Press.

Recall the passage from Clifford Geertz’s (1973) Thick description in which, drawing on an example from Gilbert Ryle, he asks what distinguishes a wink from a twitch. Geertz imagines that the “two movements are, as movements, identical” but notes that the difference between them, “however unphotographable, . . . . is vast.” What accounts for the difference? The winker, Geertz tells us, is communicating, “in a quite precise and special way.” Specifically, “to wink is to try to signal to someone in particular, without the cognisance of others, a definite message according to an already understood code,” as Ryle (2009 [1968]: 494) put it in the original essay. And of course,

A mere twitch, on the other hand, is neither a failure nor a success; it has no intended recipient; it is not meant to be unwitnessed by anybody; it carries no message. It may be a symptom but it is not a signal. The winker could not not know that he was winking; but the victim of the twitch might be quite unaware of his twitch. The winker can tell what he was trying to do; the twitcher will deny that he was trying to do anything.

Whether we call it intention or something else, Geertz and Ryle conclude that there’s something about a wink that’s missing in a twitch. That something may not be completely controlled, it may not be totally independent of other occurrences, it may not be purely mental (as opposed to social), and it may not be traceable to some original psychological event and yet is nevertheless, according to them, [458]essential to human action. It distinguishes, as Ryle says, those things which one may be properly described as having done (or at least as having tried to do), from everything else. And, according to Geertz, that something is the key to good ethnography. It’s what makes our descriptions thick rather than thin. It is, in the end, the source of motive, of intrigue, of strategy and, ultimately, of human action itself.

But, as Geertz goes on to write, like “so many of the little stories Oxford philosophers like to make up for themselves,” all this may seem a bit artificial. Anthropologists confront not imagined, winking schoolboys, but complex situations which are spread across time, space, and persons. And anthropologists are much more concerned with what the participants in these interactions say about them—how they themselves describe them, both in situ and in retrospect. Understanding this relation between what someone did and what they can be described as having done represents a central challenge for work in this area.

In The anthropology of intentions, Alessandro Duranti addresses these issues by returning to an influential set of his writings, revising and re-evaluating them, and in some ways tempering them in light of more recent thinking. It is part intellectual autobiography, part empirical analysis, part theoretical exploration. He draws together in illuminating ways three modes of analysis: conceptual, comparative semantic, and interactional.

The conceptual analysis of intention takes off from what Duranti refers to as “the standard theory . . . in linguistics, the philosophy of language and the cognitive sciences in the twentieth century” (11). He locates the beginnings of this theory in the writings of Grice, Austin and Searle, with a focus on the first and the last. According to Duranti’s analysis, the standard theory assumes that “there is a particular and common type of ‘meaning’ in human actions” (12). This is the non-natural meaning or meaningNN associated with someone yelling “Fire” but not with smoke pluming from a fire. As Grice (1957) defined it, “A meantNN something by x” iff “A uttered x with the intention of inducing a belief by means of the recognition of this intention.” Intention is criterial.

Duranti then turns to consider the speech act theory of Austin and Searle in more detail, arguing that “Searle combines Austin’s informal account of the act of promising with Grice’s definition of non-natural meaning” (13). The focus here is on promising (as it was in Searle’s classic 1965 article “What is a speech act?”) and specifically on the conditions of satisfaction for the act. These conditions require that the speaker “intends to do” the thing promised, “intends that the utterance . . . will place him under an obligation” to do it, and “intends that the utterance” will induce the hearer to believe that these conditions obtain “by means of the recognition of the intention to produce that belief” (13). Again, intention is criterial.

Duranti then argues that, for Searle at least, “there is human action if and only if what occurred matches the conditions of satisfaction as defined by the intentions of the person who performed the action” (19). So, on this view, “Oedipus married his mother” does not describe the action that Oedipus performed. Oedipus could not have intended to marry his mother in marrying Jocasta because he did not know she was his mother. On the standard theory, in other words, there are no “unintended actions.”

Duranti complicates this picture first at a conceptual level by drawing on the arguments of practice theorists such as Bourdieu and Giddens and some of their [459]phenomenological antecedents in Husserl and Heidegger. The discussion begins with Giddens’ distinction between a “reflexive monitoring of conduct,” which is a “routine feature” and “definite goals consciously held in mind in the course of . . . activities.” It branches out from there to consider Husserl’s conception of a “natural attitude” (as opposed to a theoretical one), Heidegger’s notion of Dasein, and Bourdieu’s theory of a generative habitus. It is a familiar idea: The fact that social actors can “stop and think” about their actions does not accord them special and total control of either their “intentions” or the “meaning of their actions” (21). Part of the reason for this, of course, is that any stepping back from action is also inevitably a stepping into some other action. As Atkinson and Heritage (1984: 6) put it, following Garfinkel, there is “no escape or time out” in interaction. Everything will be interpreted in some way, whether we like it or not. So reflexivity introduces important complications into an account of action but it does not confer on the actor any kind of divine control.

Having successfully complicated any simple notion of intention on a conceptual level, Duranti turns to a comparative semantic analysis of certain English, Latin, and Samoan terms bearing on the concept of intention. This is a welcome move, given that our concepts are denoted by words in languages, with all their historical and cultural baggage. The Latin terms, which provide the etymological root of the English ones, “have a range of meanings and associations that are more amenable to an understanding . . . rooted in embodied action and interaction” (31). The Samoan words (manatu, loto) commonly used in, for instance, translating the Bible, are semantically more extensive than English “intention.” They seem to combine meanings of purpose or reason with a sense of hope or feeling and are, often at least, used in collocations that evaluate, as good or bad, the dispositions to which they refer (35). Duranti’s Chapter 4 extends this analysis further to consider translations of the English word promise in Samoan translations of the Bible.

Duranti’s third and final move is to turn to recordings of spontaneous interaction. This is both where his investigations began (in the original article that is the basis for Chapter 3) and where he has journeyed to since (with his work on jazz musicians, the political campaign of Walter Capps, and in his rethinking of Samoan interaction in Chapter 8).

In a key example from Duranti’s Samoan interactional materials, a village man named Loa is formally held to account for a public transgression. Loa had announced that a newly elected district MP “was going to come and present some goods to the village matai (chiefs and orators).” Villagers assembled and waited, but the MP did not show up. A fellow villager Iuli suggests that Loa should be fined for this. Iuli says that Loa was “sitting around” and not worrying rather than going to get some food for the assembly which was left to wait and, by this, humiliated (55). In his analysis of the fono in which this matter is deliberated, Duranti points out that what Loa had intended—what he meant to do, his motives—is never discussed. Rather, Duranti suggests, it is the consequences of what he did that constitute the focal concern. Duranti does not mean that Samoans hold a theory of strict or absolute liability. Rather the focus is a matter of preference, contextually variable and shaped by social relations of the people involved.

Later chapters consider other cases. In Chapter 6, Duranti examines three versions of the same speech by political candidate Walter Capps. When Capps asks, in [460]the first of these speeches, “How do I know we’re going to win?”, this elicits uninvited laughter. Subsequent versions of the speech are modified to elide this question. In his considerations of jazz musicians, who present a classic Rylean case of “knowing how” (as opposed to “knowing that”), Duranti insightfully analyzes their own teaching practices as evidence of an “understanding of the creative process,” which “locates agency outside the individual mind and sees it as independent of individual intentionality or will.” Such remarks can be applied mutatis mutandis to talk (along lines of reasoning articulated by markedly different thinkers from Wittgenstein to Garfinkel, Ryle to Nietzsche).

We share Duranti’s commitment to studying people in real life, to see “how their language reveals their thoughts, stances, beliefs, feelings, and aspirations” (3). He is clearly correct that the “salience of intentions” is something that “varies across cultural contexts,” but as Duranti acknowledges, this context-specific variation applies most clearly to explicit discourse about intention rather than to the tacit appeal to actors’ intentions or goals that drive fundamental interpretive processes such as low-level Gricean inference and that underwrite the architecture of intersubjectivity in interaction (Enfield 2013; Grice 1975; Sidnell 2014). In a welcome development, Duranti reflects this by introducing a “universalistic” element to his account, leading him to confirm that we cannot completely reject the notion of intentions. Rather, in the book’s conclusion, Duranti (239) introduces the notion of an intentional continuum which encompasses:

a range of graded ways of being disposed or mentally (or sensorially) connected with some entity in the world,1 from a basic relationship between our consciousness and some entity that attracts our attention (through sound, vision, smell, taste, or any combination of them) all the way to our partly explicit anticipation of planning of complex activities, resulting from several hierarchically organized dispositions, like telling a story, having a dinner party, or preparing for a business meeting and then running it.

For us, the way to get at the universal nature of intention while acknowledging both (1) the diversity of cultural ideologies about others’ mental states, and (2) the impossibility of knowing what is actually in others’ heads, is to reframe the concept of intention, by acknowledging that mental states are actually not what we are working with when oriented to intention. We need to define intention in terms of social accountability, and in turn see that social accountability is impossible without language.

Elizabeth Anscombe (1976 [1957]) proposed that an action could be properly understood as intentional under some (accurate, truthful) linguistic descriptions but not under others. For example, the action of a man moving his arm up and down while holding a handle may be intentional under the description “pumping water” but not under other descriptions such as “contracting these muscles,” [461]“displacing the air,” “poisoning the household” (see Anscombe 1976 [1957]: 13).2 Ryle makes a similar point in his discussion of winking, noting that such an action can be described thinly as “contracting his right eye lid” or thickly, to paraphrase Ryle, as trying to get ready to try to amuse one’s cronies by grimacing like Tommy trying to signal covertly to his accomplice by trying to contract his eyelids (2009 [1968]: 496). And Ryle goes on to exemplify the same point with reference to language use (2009 [1968]: 498):

You hear someone come out with “Today is the 3rd of February”. What was he doing? Obviously the thinnest possible description of what he was doing is, what would fit a gramophone equally well, that he was launching this sequence of syllables into the air. A tape-recording would reproduce just what he was doing, in this thinnest sense of “doing”. But we naturally and probably correctly give a thicker description than this. We say that he was telling someone else the date. He was trying to impart a piece of wanted calendar-information, so that his attempt was unsuccessful (1) if his companion did not hear or misheard the noises, or (2) did not understand or misunderstood what he had heard, or (3) did not believe or already knew what he was told, or (4) if the speaker had himself got the date wrong.

Anscombe’s (and Ryle’s) point was that the concept of intention depends crucially on the possibility of linguistic description. So, contents of mind can be real but they may be irrelevant. What’s relevant is the accountability that is thematized by linguistic descriptions (by means of their norm-entwined inferential articulation; Brandom 1979; Enfield and Sidnell in press). In part this has to do with the fact that a person can be held accountable only for what they can be described as having done. As Anscombe noted, an important set of those descriptions of what someone has done are “intentional.” Boasting, lying, stealing are intentional descriptions. Killing may or may not be while murdering and assassinating, by definition, are. There are of course cases in which the description held to be relevant in terms of accountability does not, necessarily, imply intention. I can insult you without meaning to just as I can “hurt your feelings” or “ignore you” without realizing it.3

So the central question of intention is whether I can be rightfully held to account for having done something. While these considerations obviously complicate the [462]notion of intention, we must remember that the very act of speaking, exceptional circumstances notwithstanding, presupposes intention. Without some notion of intention there would be no way to distinguish speaking a language from making exactly the same sounds (in the case of a spoken language, manual gestures in the case of a signed one) inadvertently (and automatically or unconsciously). Exceptional circumstances might include talking in one’s sleep or in trance or under the possession of a spirit. However, even in those cases there will likely be traces of an intentional structure to be recovered by, e.g., psychoanalysis, or, perhaps, ethnography.

If cultures differ in their orientation to intention, then by definition they differ in their descriptions of intention. This has to be true in terms of the semantic structures that their languages provide for such descriptions (a form of relativity). But because all cultures use language to describe actions, then it cannot be true that a culture does not deal with intention. (We also know this because of the universal indispensability of enriching inferences à la Grice for going from literal readings of utterances to their often obvious contextual import.)

In that respect we can note that, in translation at least, Loa is described as having “lied” and having “boasted.” It may be that the Samoans in the discourse described are relatively uninterested in why he lied or why he boasted. They avoid attribution of motives. But they do use action descriptions—metapragmatic descriptors à la Silverstein (1993)—that presuppose a complex underlying intentional structure, a model of behavior in which persons are taken as intending (trying, attempting, endeavoring, etc.) to carry out particular courses of conduct.

This suggests an alternative way in which to understand Duranti’s Samoan data, toward spelling out those “preferences” he alludes to. We could seek to describe the contextual parameters, the particular social relations, which make speculation about another’s motives (rather than intentions per se) more or less appropriate. If parents don’t attribute intention to children, is this because children are not seen as possible intentional agents? If matai don’t conjecture why another has announced the imminent visit of an MP, is this a matter of delicacy and etiquette? Would doing so violate norms of politesse or politics? Would the one conjecturing be culturally unintelligible or seen as vulgar? To ask these questions is to reframe the analysis with accountability at the center.

Key to moving forward in research on intention is to overcome the persistent unclarity of the concept itself. Duranti’s work is a welcome attempt to undertake this work, and Throop’s contribution to an email thread reproduced as Chapter 7 constitutes an important step in this direction. We think more can be done in this respect.

One locus for conceptual unclarity around intention is the very ambiguity of the word, giving us the oft-confused ideas of (1) being deliberate or purposeful, (2) the directedness of mental states such that they are “about” things, (3) the general content of others’ mental states, such as their beliefs, desires, and goals, and (4) the specific motives or reasons which are assumed or asserted to have provoked some particular action. Keeping these meanings distinct should in principle be a straightforward matter.

For us, a deeper and more significant move toward clarity around the concept of intention is to recognize that when we talk about intention—in any of the senses just mentioned—we are talking about something that is, first and foremost, created and maintained by public, dialogic processes of social accountability. Of course there [463]are mental states such as beliefs, desires, and goals. People of all cultures must effectively assume this—independent from their ideology regarding overt talk about such states—if they are going to coherently and efficiently engage in conversations and other forms of social interaction. As Duranti repeatedly notes, a group may prefer not to entertain overt discussion about others’ mental states, but this is not incompatible with the idea that the same people do effectively orient to intentions when they make inferences as to what others are meaning to say, and meaning to do, based on their words. After all, our words are only a partial index of our meanings and actions.

But the key move that we think is required is a necessary consequence of the obvious fact that, regardless of our ideology, we really do not have access to others’ mental states. This is the “no telepathy” premise (Hutchins and Hazlehurst 1995). It entails that, as analysts and as participants in social life, we have two things to work with. The first is the claims that others make, as coded in their words. The second is the immediately observable value of those claims, as revealed in their effects in the enchronic frame, in which every move is immediately treated to a certain interpretation by the move that is touched off next. In this sense, actors are “not so much constrained by rules or sanctions, as caught in a web of inferences” (Levinson 1983: 321), or as we put it, a tyranny of accountability.

To summarize, whether or not we approve of talk about mental states, we (1) have no direct access to any such states, yet (2) act in ways that look as if we do entertain those mental states (e.g. when we do what conversation analysts describe as “third position repair,” or when we otherwise infer more than what is said). These points are seemingly contradictory, and they position both Samoans and WMC people at a step of remove (in opposite directions) from reality. Samoans embrace (1), yet by demonstrating (2), they appear to contradict that stance, while WMCs appear to act as if (1) were not true yet it must be.

The key questions now are: If we so clearly cannot read minds, how is it that cultures can vary in the degree to which they acknowledge this, and how do we pull off the many behaviors that so closely resemble mind-reading? The answer, we argue, is to be found in dialogic practices of accountability in the enchronic frame of real-time social interaction. Actual mental states, or intentional states, are irrelevant. What matters is the public, social accountability that a person can rightfully be subject to.

We see two related ways forward for research here.

  1. We need to study actual instances of accountability (including expressions of surprise, sanction, wherever subprehension is contravened and surfaces) in order to map the normative coordinates that pertain in a community; importantly, looking at the right data allows us to see both (1) the explicit coordinates that we refer to as ideology, as well as (2) the implicit but equally real, though on a different level, coordinates that we refer to as practice.
  2. We need to study the practices of describing actions, thus on Anscombe’s account defining intentions in so far as these descriptions thematize individuals’ accountability (and through that, following Garfinkel, Sacks, et al., structure the entire culture and society by making explicit its norms); perhaps most important is to study practices of description in the course of expressing surprise, sanction, and otherwise holding others accountable for what they say and do.[464]


Anscombe, G. E. M. 1976 [1957]. Intention. Oxford: Blackwell.

———. 1958. “On brute facts.” Analysis 18(3): 69–72.

Brandom, Robert. 1979. “Freedom and constraint by norms.” American Philosophical Quarterly 16(3): 187–96.

Enfield, N. J. 2013. Relationship thinking: Agency, enchrony, and human sociality. New York: Oxford University Press.

———, and Jack Sidnell. in press. “The normative nature of language.” In The normative animal? On the anthropological significance of social, moral and linguistic norms, edited by Kurt Bayertz and Neil Roughley. Oxford: Oxford University Press.

Geertz, Clifford. 1973. “Thick description: Toward an interpretive theory of culture.” In The interpretation of cultures, 3–30. New York: Basic Books.

Grice, H. P. 1957. “Meaning.” The Philosophical Review 66 (3): 377–88.

———. 1975. “Logic and conversation.” In Syntax and semantics, volume 3: Speech acts, edited by Peter Cole and Jerry L. Morgan, 41–58. New York: Academic Press.

Heritage, John, and Maxwell Atkinson. 1984. “Introduction.” In Structures of social action: Studies in conversation analysis, edited by Maxwell Atkinson and John Heritage, 1–15. Cambridge: Cambridge University Press.

Hutchins, Edwin, and Brian Hazlehurst. 1995. “How to invent a shared lexicon: The emergence of shared form-meaning mappings in interaction.” In Social intelligence and interaction: Expressions and implications of the social bias in human intelligence, edited by Esther Goody, 53–67. Cambridge: Cambridge University Press.

Levinson, Stephen C. 1983. Pragmatics. Cambridge: Cambridge University Press.

Pierce, Chester, Jean V. Carew, Diane Pierce-Gonzalez, and Deborah Wills. 1977. “An experiment in racism: TV commercials.” Education and Urban Society 10 (1): 61–87.

Ryle, Gilbert. 2009 [1968]. “The thinking of thoughts: What is ‘Le Penseur’ doing?” In Collected Essays 1929–1968: Collected Papers, Volume 2, 494–510. London: Routledge.

Searle, John. 1965. “What is a speech act?” In Philosophy in America, edited by Max Black, 221–39. London: Allen and Unwin.

Sidnell, Jack. 2014. “The architecture of intersubjectivity revisited.” In The Cambridge Handbook of Linguistic Anthropology, edited by N. J. Enfield, Paul Kockelman, and Jack Sidnell. Cambridge: Cambridge University Press.

Silverstein, Michael. 1993. “Metapragmatic discourse and metapragmatic function.” In Reflexive language: Reported speech and metapragmatics, edited by John A. Lucy, 33–58. Cambridge: Cambridge University Press.

Sue, Derald Wing, et al. 2007. “Racial microaggressions in everyday life: Implications for clinical practice.” American Psychologist 62 (4), 271–86.[465]


Jack Sidnell
Department of Anthropology
University of Toronto
19 Russell St.
Toronto, Ontario
M5S 2S2

N. J. Enfield
Department of Linguistics
The University of Sydney
John Woolley Building A20
Sydney, NSW 2006


1. And presumably also entities that are not, in any straightforward sense, “in the world,” e.g., justice and jurisprudence, kindness and kindedness, as well as such remarkable “things” as, for instance, intentions and intentionality!

2. Anscombe’s own understanding of intention was subtle and nuanced. She notes, for instance, that, “the fact that something is done in a society with certain institutions, in the context of which it ordinarily amounts to such-and-such a transaction, is not absolute proof that such-and-such a transaction has taken place. Is it intention that makes the difference? Not if we think of intention as purely interior” (1958: 70).

3. An interesting case of an “action” which involves no intention and is assessed strictly in terms of effects is the so-called “micro aggression.” The classic example of this is a white person remarking about or even complimenting a black person’s hair and asking if they can touch it. The notion was introduced in Pierce et al. (1977) where it was described as “subtle, stunning, often automatic, and nonverbal exchanges which are ‘put downs’ of blacks by offenders” (61). As Sue et al. (2007) write, “Perpetrators of microaggressions are often unaware that they engage in such communications when they interact with racial/ethnic minorities” (271).