I use the formulas "etc." and "and so on" as much as anyone else. Linguists have in the past been interested in how I could possibly assume that others knew what I was talking about without having to "spell it out". Sometimes communication seems more effective when (severely) incomplete. But for Garfinkel, "etc." is only one instance of the speaker's incomplete communication; and as a well accepted, "conscious" technique, it also serves as proof of the soundness of one's argumentation(1). If I can say "etc." and be understood, the logic of my argumentation is demonstrated.

But in fact, everything we say is in the form of "etc", i.e. it assumes something like background knowledge in the listener. That is what ethnomethodologists call "indexicality": the fact that most parts -- if not all -- of any discourse have a specific resonance ("unstated understandings") for specific listeners: and to be understood, speakers and listeners must agree on the rational soundness of every separate unit of the discourse. That is possible only for members of "groups" who share common "methods" for relating or "telling" the story of reality.

This would seem obvious at a conference, for example, where neophyte listeners might not understand some scholarly language. However, Garfinkel is in fact talking about something quite different. The point is not that given the right dictionary any listener could figure out what has been said. In fact this reference to a lexicon of some sort is infinite: the problem is not the meaning of the words in question, but how this meaning is constructed by the speaker and the listener in a particular situation at a particular time. "Looking up" the words is of no help if you do not understand the method by which they are used. This was Garfinkel's students' problem when they attempted to "decipher" an everyday conversation: simply explaining the words led to an infinite regression. He says that "the appropriate image of common understanding is therefore an operation rather than a common intersection of overlapping sets"(2) (emphasis added). Understanding a dialogue implies understanding how words are used in a particular situation, not simply what they mean; in that way, understanding is participating in the construction of meaning. When people say they are "on the same wavelength" they do not mean simply that they are using the same lexicon, but rather that they are producing communicable accounts in the same way, with the same tools; furthermore, they are also assuming that this will remain the case at least in the foreseeable future, if not forever.

The concepts of "operation", "method", etc. are interesting. Ethnomethodology posits the relation of the individual member of the "group" to reality as "work" to be done. Reality is not a given and it has to be "made up", as was already clear with our extension of Becker; but here, the act of "labelling" is much more subtle. First, there is a reason I have been using the word "group" in quotation marks -- it is not merely to make the text less legible: this is a different concept than that of Becker, or Miller, for instance [compare]. It is not a political or normative group in the usual sense; it is also not a cultural group in the sense that it is identifiable by norms or even symbols and meanings: it is a group of people equipped with common tools for building these symbols, meanings and other communicable metaphors(3). Group membership could be described as "ethnomethodological competence". The distinction may at first appear to be without consequence, but in fact is crucial, especially on three fronts: (1) the scientific approach will be concerned with lifting the veil that metaphors tend to throw upon their own process of creation, not to stop at the description of the results. (2) The assumption that groups are immediately identifiable units disappears. (3) Interaction between groups is unpredictable, because at a certain level members do not know who belongs to which group, and because even if they did they would be incapable of readily identifying each other's ethnomethods.

Second, the work that needs to be done by members of the group is not simply to apply labels, but to create them in the first place. At this point, the image of "label" is probably too coarse to really fit this theory: it includes assumptions that make it too simple and too general. It assumes that there are observable objects to label; it assumes that language is uniform and readily applicable to those objects, and finally when it is said that a label can "stick" or not, it becomes obvious that the concept is static and cumbersome in terms of the microscopic work needed to create everyday reality.

Another modified concept is that of "rationality". According to Garfinkel, people think of the world and of themselves in terms of stories, or "accounts", which main characteristic is apparent logic. An account will not ring true to a listener (or to the speaker/thinker in the first place) if it lacks internal logic, if it is not rational from her point of view. This is "common sense" among members of groups, or recognizable sense, since it is particular to the group. Now, the work of social science so far has been to try to decode common sense and indexical expressions alike, to make them objective, to shape them into predictable, universally rule-governed facts. That is what "lay" scientists, people, also do: from the start common sense is the transformation of raw perception into organized accounts. We do this essentially in order to decide how to act and react. At this level, traditional science itself becomes a target of ethnomethodological criticism. If to live is to construct, good science is to stop taking for granted that (1) rationality can be adopted as immediate explanation for social phenomena but even less, at the other end of the spectrum (2) that common sense (as well as "scientific" sense) is a given, and not a "practical accomplishment" itself worthy of inquiry or (3) that all the reality-producing activities of members, lay and professional, are interchangeable across any group. The reasons for this will -- hopefully -- become clearer in a moment. In a nutshell, the world does appear to be Durkheimian in nature (rule based, objective, etc.) but this is a facade that hides the true mechanisms of the social.

The obvious questions are how do members know what makes sense? How do they know how to produce a rational account? Why does this account have to be rational in the first place? In everyday conversation, one does not stop to think about what they will say or do, and this is why, in a sense, ordinary activities are so interesting to ethnomethodologists: they seem utterly automatic, immediate and "natural"; in short, they seem to prove that ethnomethods do not exist after all. But in fact the appearance of "impossible" immediacy comes from the intuitive mistake of confusing internal and external logic. When the wife speaks to her husband, she talks until she is satisfied that her discourse has fulfilled its logical necessity for her; beyond that, she can only assume that his mind works the same way, i.e. that what she said will be equally sensible from his point of view. She does not need to ask herself how he thinks, because she works on the premise that she already knows. And if this is true of trivial exchanges, it must be true of more meaningful discourse, otherwise complex thought would be impossible to communicate (and to produce). This is not to say that people understand each other all the time; in even the most intimate relationships people fail to communicate. In fact, it is entirely possible that we are always mistaken about what the other is trying to say; Garfinkel would explain that it is still acceptable to simply conclude that people share a way to communicate, regardless of its efficiency, to demonstrate the need for studying it. He would also conclude that trivial exchanges, as the most skeletal form of communication, are the closest to the "raw" ethnomethods.

Hobbes said that the polity was simply built by its citizens, in their own image as reasonable creatures. Durkheim assumed, on the contrary, that people's rationality was shaped by the nature of the social. Garfinkel has a radically different explanation. Members indeed use "law-like" standards of accounting for perceived reality, be it complex "social facts" or everyday activities like taking an umbrella when leaving for work. However the rules only exist as they are invoked and used.

To use a Star Trek illustration, life without social meanings and metaphors is like an empty "holodeck"(4), where no program is active, and there is nothing to do. The holodeck gets progressively furnished and populated when objects, people, situations, etc. get defined by the player/programmer and appear as material: she can only interact with the elements she has attached meaning to. In fact, like in the holodeck, even purely physical objects are not socially visible(5) until they are defined. Like a piece of artsy, unrecognizable furniture, you cannot decide to sit on an object before deciding that it is a chair (and if it turns out to be a coffee table you will be the laughing stock of the party). Now when you start interacting with people, negotiation on meanings begins; you will have to adjust your behaviour and your discourse in function of the reactions they produce, as perceived by your own senses -- and subsequently (re)defined. The trick is, if you do not want to live in an artificial holodeck, you have to forget ("gloss over") the fact that you are responsible for everything that exists. The objects you create have to possess the varnish of reality, of continuity, of "solidity", all the while allowing for a certain dose of unpredictability (the "loose" logic of the human as perpetual investigator) otherwise the illusion is not complete, not sufficiently convincing and you cannot play. In the end, metaphors and accounts are always rational simply because their existence includes a constructed version of what rationality and logic are (they come equipped with a veil that hides their birth).

One last example, this time closer to the nitty-gritty reality: most undergraduate students operate on the very common misconception that if something is said to be "socially constructed" it means that it is somehow less serious or concrete than a "real" thing. Some would be genuinely shocked to hear that rape, for instance, is a ("mere") social construction. It would seem to diminish the trauma of the experience, to dismiss it in some way. From what has just been said, however, one could easily argue that what is socially constructed in rape is exactly what makes it as terrible as it is. It is in fact obvious that (1) there is no socially useful "reality" "underneath" any construction; (2) constructions -- including construction of the self -- are what allow people to live in the world; (3) extreme social deconstruction of the self amounts to "social death" -- and that is the trauma of rape; (4) when dismissing (3), the most "real" version of rape is a reduction to physical bruising (if that).

One of the most annoying things about ethnomethodology is that few objections are possible, and yet it seems intuitively incorrect or at least incomplete. First, it fails completely to account for differential power relations and material conditions in the social world. However, if the objection sounds obvious, in a way it is inconsequential. Garfinkel underlines from the start that conventional social sciences can still work on those problems. Beyond that, it could simply be said that the scientific reality of social classes, power stratification, differential relations based on "race", gender, age, etc. themselves (e.g. not the relations, that is obvious, but the very concept of relation) are constructions (professional and lay) members use to make sense of the world. Anything else that can be said about these constructions is ethical/normative and not scientific.

A more difficult objection to reject has to do with the move away from subjectivism that ethnomethodology seems to imply. Up to this point, the theories we have looked at were slowly evolving into a model that left less and less room for free agency and the subjectivity of individuals. While at first sight Garfinkel seems to contradict this by regarding as important even trivial conversations and everyday actions, and concluding that the world is constantly made by "rational" actors, a more attentive reading tends to uncover an extreme proximity to the opposite pole. The study of minute, everyday life actions, for Garfinkel, uncovers the fact that there is simply no such thing as a subject. If I take the concepts one by one: (1) the world is constructed, but not by voluntary choice; it is rather a necessity of each member if she "wants" to function in the social reality, and a standardized necessity if she "wants" to function in her group (in fact, there is no other place for her to go, and if there was, the same problem would apply there too, i.e. we can only want things we can conceptualize). She has to learn ethnomethods, and she can only do this by participating in group life. I call this the "perpetual investigator" model of human life(6). The french saying "c'est en forgeant qu'on devient forgeron" is a good illustration. Indeed, one can best learn how to be a blacksmith by "doing" it; but that does not provide or even allow the choice of becoming a baker.

(2) Garfinkel's "rational" actors are of a very specific breed. Here, rationality is but a gloss, another ethnomethod (standards of rationality do differ). I have mentioned earlier a difference between internal and external logic. Without going much further, I will only briefly clarify what I meant. When the wife talks to her husband, she tries to make sense to herself, assuming that her husband thinks the same way; there may or may not be clues given that he is following, but they need interpretation. There is an internal logic to her discourse even though, as stated before, her husband might turn out not to understand at all. In effect, internal logic here is strictly limited to individual conformity to an ethnomethod: it is therefore not a product of reason in the usual sense of the word. External logic, i.e. statements that seem(7) to conform with the canons of classical rationality (laws, necessity and universality, which in my view consistently fails to convincingly prove its relevance to the social sciences) on the other hand does not apply at all. Obviously, the wife cannot analyse her words in advance in function of this criterion and this is why I said that the immediacy of her discourse can seem impossible (it cannot be both immediate and analysed). Conformity to external rationality is only an educated guess, a matter of subjective interpretation and a continuous local accomplishment.

(3) The "World" (and I kept this one for last because I am going to go out on a limb here) is a story that members tell themselves, and this is why Garfinkel calls accounts "reflexive" and "incarnate". It is this last term that for me evokes the most problematic question. Obviously, Garfinkel's inspiration in the matter is Husserl's phenomenological approach, the idea that truth, or true Reality is not available to our senses because it in fact does not exist (there are no kantian noumena), while only "phenomena", as productions of the observer, are. Therefore conscience observes and constitutes reality simultaneously (i.e. it fills in inevitable sensory gaps), and only the subject has reality. The point is this: for Husserl, there is a reason for me to be writing this paper: there are other subjects who might be interested in reading it. But for Garfinkel, there is absolutely no reason, because there is no such thing as a subject, only a self-perpetuating story: the subject is disincarnated. This paper is in fact a total waste of time (but then again what is time? Will it still exist if I stop writing, furnishing it with ethnomethod-based actions?).

So in this perspective individual experience loses most of its probative usefulness in social sciences: experience is again explained (as opposed to understood) by external and independent structural factors that seem more determinant than ever. The reason I said ethnomethodology was annoying in its completeness, is that while it squarely negates any conventional concept of free will, it can account for it seamlessly. Humanist remnants aside, nothing can stand in its way, as opposed to the gaping holes left in functionalist theories, for instance. Saying "I have the power to change things", here, simply means "my ethnomethodological competence allows for a different conceptualization of the world". That said, the next question is, if we are entirely colonized by a structural grammar of thought, how is it possible for us to think about it? Nietzsche circumvented the problem by reversing the analytical train in the direction of the "genealogy of language", to avoid being bogged down in speculation about its relation to hypothetical reality: he says, "I can tell you that what you say is not true, not because I know what truth is, but because I know that the way you got there is independent of it." But in this model, the capacity to think about ethnomethods seems to lead to either of two conclusions, first, we are absolute geniuses, second, the theory does not hold water.

Finally, I still have a problem with the notion of "groups" in the cultural sense (I had the same objection with Miller). Exactly like the indexical character of language, it seems that trying to define "group" leads to an infinite regression. What does it mean to be a "member", when intimate couples fail to communicate (or am I just being too cynical about that)? In short form, the question is: what level of consistency in ethnomethods is required to form a group?



1. Another example is irony; when one is saying something by verbally expressing its opposite, she is using a technique that she might not identify as such but has to be fully aware of. What is said here is that while the rest of her discourse is also in the form of irony (i.e. it is the product of a technique that is assumed to be shared by the listener), she wants it to be "real", and not a technique, and so she cannot let herself be aware of (she has to ignore) the technique. As we will see, the elements of her discourse automatically contain tools that make this possible for her.

2. (1967) Studies in Ethnomethodology, Cambridge, Polity Press, 1984, p. 30.

3. The word "metaphor" may remind the reader of a piece by Nietzsche titled Truth and Lies in the Extra-Moral Sense, in which he demonstrates that the necessary steps towards the creation of language infailably lead to a construction of reality (a lie) and that truth is simply adherence to conventional language (a consistent lie). Words are metaphores used to generalize the ungeneralizable complexity of perception.

4. For non-trekkers, the "holodeck" is a tridimensional environment where holograms of objects as well as people take physical consistency and can interact with players according to a pre-written program.

5. They may in fact be visible from a purely sensory point of view, and they may be analyzable by exact science techniques, microscope, mass spectrograph, etc. but any object still will not socially exist in terms of how the individual can relate to it until she has "named" it. It is of little help to know that an object is made of glass particles when you are trying to decide whether to drink from it or put flowers in it. And if this is true of physical objects, it is easy to imagine when conversations, social situations, etc. are in question. One can remember Mr. Kane uttering "rosebud" before passing on.

6. I would however like to stress that it is not entirely compatible with Garfinkel in the sense that once ethnomethods are acquired, little change is allowed for.

7. I will not discuss it, but conformity of natural science to pure classical rationality has nowadays been transformed to a falsification model, where the rationality is not in the definitive answer but in the absence of a definitive rebuttal. See Carl Popper on that.