Annotated version of the first 10 pages of: 

BEING ONE OF THE GROUP
Patrick Boylan and Nadia Mari


0. Abstract/Introduction

This paper analyzes how second-language learners and native speakers interact in small groups. It focuses on how apparently marginal behavior, such as eye and head movement, may contribute to determining each participant's status as an ‘insider’ or ‘outsider’. Data is taken from spontaneous small-group conversations among American and Italian college students, surreptitiously filmed and analyzed by a research team composed of teachers and students of English as a Second Language (henceforth ESL) in Rome.

To what degree, then, are eye/head movements culturally distinctive? What messages do they convey? How can we be sure we have understood them? Can adapting our eye/head movements improve communication? The paper argues that by studying conversations as ‘texts’, such questions cannot fully be answered. It then proposes the notion of ‘enactment of intent’ as the basic ‘unit’ making up conversations, and outlines an innovative research procedure based on experiential knowledge (phronesis), better suited to ascertaining the local meaning of somatic as well as verbal and prosodic messages. This, it claims, is the kind of knowledge that learners of a second language (henceforth L2) need and that conversation analysis (henceforth CA) should consider investigating.
 
 

1. Aims

The aims of the present research project were:

a. to observe and record spontaneous conversational interaction in informal settings between self-selected members of culturally heterogeneous small groups, specifically, between mixed-gender American and Italian university students and teachers surreptitiously filmed as they chatted together informally in English after a lecture/debate;
b. to take one of the features of the video-recorded interaction — here, gaze and head movement — and, using classical research methods (Kendon 1967, Birdwhistell 1970, Allen & Guy 1974, Argyle & Cook 1976), to verify, together with the participants:
¨ if each cultural group has a ‘typical somatic behavior’ in conversing and, if so,
¨ whether a participant can be more successful in communicating with members of the other culture by adopting their conversational behavior;
c. to determine — this time with only the Italian participants, all of whom were ESL students — whether the kind of knowledge classical research methods furnish gives them a better insight into the communicative competence they wish to acquire, i.e., native-like conversational skill in multi-party multi-cultural encounters;
d. if the response to (c.) is negative, to determine what kind of research questions are not currently being asked in the field of conversation analysis (but that the ESL learners deem of interest) and to propose a methodology for answering those questions.
 
 

2. Experimental task and setup

Fifteen American third-year university students studying art history in Italy were invited by fifteen Italian ESL students to a lecture/debate in English on cultural differences between their two countries, held at the University of Rome. After the debate, the students were free to mill and converse in the lecture hall while a lawn picnic was being set up outside. A video camera was present in the room and was constantly pointed at the table up front where the lecturer and the discussion moderator sat. The camera was equipped with an Eibl-Eibesfeldt lens, i.e.,a hidden mirror arrangement which permitted filming anywhere in the room unnoticed; audio came from a tiny radio microphone concealed on one of the students. By rotating the hidden mirror, the cameraman was able to keep the camera constantly focused on the ‘wired’ student as she wandered from one group to another engaging in conversation. To make sure the student acted as naturally as possible, she was told, when the lecture began, that the microphone batteries had gone dead and thus there would be no audio recording; since the microphone was sewn inside her dress, she had to leave it there. One wall of the lecture room had a panel mirror so that the faces of any group standing in front of it could be filmed either directly or as a reflection. Unfortunately, no group spontaneously formed in front of the mirror; thus in some cases, gaze must be reconstructed through the interactional dynamics. All students were informed afterwards of the video recording and consent was obtained for its use.

The reader is now invited to turn to Appendix 1 and inspect the drawings (called a ‘storyboard’ in cinema jargon) of a fragment of the filmed conversation. Appendix 2 gives a verbal description of the movements;Appendix 3, the transcription conventions.
 
 

3. Previous studies; new directions

3.1 Typical research into conversational interaction centers on culturally homogenous dyads (see Allen & Guy 1974:54 for a justification of the preference for dyads). Less studied are triads (Kerbrat-Orecchioni 1990-4), large groups (Edelsky 1981) and, least of all, small groups (Berrier 1997:326). While studies of intercultural interaction are now fashionable in our multi-cultural societies, they also concern dyads or triads (Orletti 1992, Shea 1994, Jensen et al. 1995) and only rarely small groups (Berrier 1995).

Yet conversational competence in small-group situations is what learners of any L2 seem to find most difficult. In paired exchanges with members of the other culture (service encounters or one-to-one socializing), they automatically get individualized attention from their interlocutor, however begrudging it may be. In large groups, L2 learners usually have the option of creating a dyadic relationship to get individualized attention and to isolate themselves from the crowd. Not so in multi-cultural small group conversations: here they are mostly on their own and, in order to cope, need "a superior know-how" (Kerbrat-Orecchioni 1997:10). They either ‘fit in’ or get left out.

Unfortunately, current language teaching methodology gives L2 learners neither the know-how Kerbrat-Orecchioni speaks of, nor the intellectual tools to acquire such know-how empirically in real-life situations (Boylan 1998). Thus it is common for even moderately fluent L2 learners to report that, in small group conversations with native speakers, they often feel stifled, ignored or, worse yet, condescendingly listened to during the occasional lulls in the ‘real’ conversation. In other words, these L2 learners do not feel accepted as ‘one of the group’. But, of course, how could they be? They have no idea of what it takes to gain a "footing" (Goffman 1981) in the new culture.

3.2 What can we do to prepare L2 students to understand and assimilate the dynamics of the conversational interactions in which they find themselves?

A first step would be to determine what indeed is required to gain a footing in a given foreign culture. Frake (1964) suggests, rather convincingly, that this involves learning to feel as ‘real’ what is ‘real’ for one's interlocutors. There is no doubt, in fact, that in multi-cultural group conversations L2 learners will tend to be ignored if the topics they raise — ‘real’ to people back home — appear pointless to the group or out or place (see Tenny 1989 for rules of topic appropriateness). The same goes if these L2 learners, by lack of appropriate reaction, treat as pointless or irrelevant the topics that the rest of the group finds ‘really’ significant, controversial, humorous, scandalous or whatever.

Indeed, to use a metaphor from transplant surgery, can we really blame the group for rejecting a body felt as foreign? After all, why should the group take seriously people who don't take seriously things that matter? It may be instructive but it is certainly not pleasurable to let an outsider, by her apparent disregard, call into question what one always thought mattered; and it must be remembered that conversation (fun) is different from discussion (work) precisely because it operates on the pleasure principle. If, on top of it all, the outsider makes conversation management a chore — by sending confusing signals through ‘strange’ eye and head movements or by maintaining a speaking style diametrically opposite from her interlocutors on such culturally indicative scales as physical-contact ~ distance, one-speaker-at-a-time ~ multiple-flooring, explicitness ~ allusiveness, (Hall 1966, Jensen et al. 1995) — she can hardly complain at being left out of the conversation. She has in fact opted out by not having opted in.

In a perfectly just world, of course, she would not have to ‘opt in.’ She and her conversational partners would want to learn from each other in order to create a terrain of authentically shared values and creative diversity (see the special issue of Pragmatics 4/3, September 1994). This is not, however, the world that many L2 learners live in.

What L2 learners need, then, is the capacity to meet their interlocutors a little more than half-way culturally. This does not mean memorizing lists of conversational topics and Do's & Don'ts. Actors get quickly unmasked. It means something much simpler but much more radical: sharing, at least in part, the existential value system at the bottom of the other culture. To be truly ‘one of the group’, L2 learners must truly feel they are.

3.3 Thus, the basic task facing L2 teachers is to get learners to internalize their interlocutors' culture as a value system or Weltanschauung (Boylan 1998). Doing so will allow them to forget the lists of rules and play by ear. Far from being cultural imperialism, this technique permits the L2 learner to communicate her values as a member of a different culture deserving respect much more effectively and convincingly than if she spoke as an ‘outsider’. In addition, it gives the L2 learner a negotiating edge, since she controls the flow of information. (These points are developed in Boylan, forthcoming/a.).

But getting learners to internalize a second language does not simply mean getting them to playact. It presupposes a thorough contrastive study of the cultures and communication modalities involved. In fact, to return to the topic at hand, one way to help L2 learners ‘enter into’ the new language and culture is to get them to formulate (and then try to imitate) how people manage conversations in the target culture — including how they use their eyes and heads — as the expression of a different, felt Weltanschauung. Kendon (1967) reported long ago that eye and head movements are our principal signaling devices in conversation. Allen & Guy (1974:142) confirm his assertion: "Eye contact is an important supplement to the conversational encounter. It is a good indicator of the degree to which the channel is open or closed, [...] the degree of bonding and the level of attention in the listening mode." Participation, bond of solidarity, interest: these are precisely the qualities we just mentioned as essential for ‘being one of the group.

3.4 Several classification schemes have been devised over the years — see Section 1 b; for the most recent, see Pezzato & Poggi 1998 — which give a research group interested in replicating a typical ‘candid camera’ ethnographic experimentation a framework for ordering observations of eye and head movement. What remains to be seen, however, is the usefulness of these schemes to L2 learners or, for that matter, to anyone interested in communication. Classical research methods measure, for example, phenomena like the length of time a typical subject spends gazing at an interlocutor when speaking or when listening. But these observations are seldom related to what the recorded phenomena mean. Average "length of gaze" is an interesting concept; but when does a glance — in a foreign culture, or even in one’s own culture — become ‘fleeting’ or ‘overly insistent’? What non-average "length of gaze" enhances participation, creates bonds of solidarity, manifests interest or, on the contrary, appears brazen, standoffish or bored? These are the findings that people who use language need to have. More importantly, they need to know what procedures will permit them to learn these things in the culture in which they may one day have to live and work, if no studies have been made yet.

The fundamental problem, then, is how to assign meaning to behavior. And given that communication is multi-modal and holistic (Poggi 1997), not only must we take note of eye and head movements as potentially meaningful behavior, but we must also record behavior in every other modality (verbal included) and weave all the data together into a single fabric. Note that the fabric thus obtained is not the ‘meaning’ of the communicative event, but simply its material substratum, no more than the documentation of an automobile accident (photos, declarations, measurements) is the ‘event’ determining a claimant's legal rights. In the case of an accident, legal rights derive from an interpretation of available documentation by an insurance expert or a judge. They alone are authorized to decide what really happened, by using applicable law and jurisprudence to link intentionality, behavior, concurrent circumstances and social responsibility into a legally-binding whole (the ‘event’). In the same way, the fabric of a conversation that we weave when we analyze a video recording, is not the ‘event’ but rather the material to which we assign a sense in the very act of weaving it, thereby creating the ‘event’ (Gadamer 1960). Like a judge, rightly or wrongly we determine what the conversational ‘event’ was through the hermeneutics, not of law and jurisprudence, but of psychology and cultural practices. It is not clear, however, what authorizes us to pass judgments.

Let us therefore look for a moment at how we go about ‘making sense’ of talk-in-interaction. This will allow us to answer some of the crucial methodological questions that Grossen & Orvig (1998:153-154) raise concerning the validity of clinical interviews, and which apply to any exchange that involves ‘judging’ what an interlocutor means.
 
 

4. General considerations: making sense of a conversational event

4.1 A key step in making sense of any event is defining the units that go to make it up. This of course depends on the traits we have selected to foreground in viewing the event: in any empirical science, what we discover depends on what we have chosen to notice — or, more exactly, not to notice. Indeed, as Gadamer (1960) argues, researchers (such as ourselves in this very paper) may be said to take de facto a social and ethical stand in finding ‘interesting’ certain phenomena which they then elaborate as knowledge.

4.2 For Berrier (1997), revisiting Sacks et al. (1974), the basic unit of conversational interaction is the utterance, defined as "any verbal attempt at turn-taking" (p. 326). However Ford et al. (1996) question whether the floor is in fact held or ceded on purely verbal grounds. Real-life conversational turns, they assert (p. 449), are "constellations of convergence and divergence" of verbal/prosodic/somatic signals, the weightings of which co-determine possible turn junctures dynamically. In this perspective, then, conversations should be seen as composed not of utterances but rather of moves (Goffman 1981) to gain or hold the floor by whatever means available, verbal, prosodic or somatic.

4.2.1 The two definitions just given offer useful insights. However both follow a consolidated but questionable tradition: that of treating conversations as ‘texts’, i.e., bounded, finalized semantic representations, transferable onto paper for easy inspection and dividable into verbal (or verbal/prosodic/somatic) units with the stroke of a pen. This approach, we hold, while applicable to text productions proper (a sonnet, a joke, ritual insulting) is misleading when applied to conversation. Conversations are not simply ‘texts’ — neither verbal texts, nor multi-modal texts (i.e., concurrent verbal/prosodic/somatic realizations pictured as a musical score: Poggi 1997), nor semiotic texts (e.g., in the way some semiologists attempt to reduce culture to a text). This is because a ‘text’, by definition, is a framed representation to be viewed from without, while conversations (like cultures) are boundless events, the sense of which can only be experienced from within. Conversations (and cultures) produce ‘texts’ as their residue. While much can be learned by studying that residue, to fully grasp a conversation (or a culture), a procedure other than textual reconstruction is needed. Let us briefly clarify these assertions.

4.2.2 When one delineates (‘frames’) an object in order to observe it as a ‘text’, one endows it with meaning potential. Pop Art did that with soup cans. It is also what CA does in framing conversation fragments as ‘adjacency pairs’ or ‘side sequences’.

If a stretch of speech has been conventionally bounded and finalized by the speakers (even unconsciously) — for example, a ‘side sequence’ — and if the subsequent framing operation by the linguist respects those boundaries and puts that finalization into perspective, then textual analysis of the framed object is both practicable and instructive. If on the other hand, as in (Pop) Art, the framing does not correspond to the way whoever created the object sees it, but rather to how the framer sees it — i.e., to the personal mental world the framer projects upon the object — then the analysis will end up telling us more about the framer than about the object. This, too, can, be a perfectly legitimate and instructive operation, if we are interested in the framer’s ‘perceptions’ or metadiscourse. It must not, however, be confused with scientific inquiry, which imposes on us to accept the dictates of the object as other in order to come to grips with it (7.1.8.) .

The first question to resolve, then, is whether conversations, as such, are bounded and finalized and thus frameable as ‘texts’. If they are, we can then move on to defining procedures that guarantee that our framing respects the object. If they are not, we have a more difficult task — discovering a non-textual approach for coming to grips with them.

4.2.3 Conversations, as Garfinkel (1967) argues convincingly, are on-going, situated, collegial, intentional attempts at existential meaning-making (or meaning-confirmation); they are the very stuff out of which our conscious lives are made. That is why Aristotle, long before ethnomethodology, considered societies to be communities of discourse more than communities of people (Lo Piparo 1996:40). In this perspective, therefore, the stretches of talk ordinarily called ‘conversations’ — e.g., the talk that occurs between the lifting and the lowering of a phone receiver are simply fragments of a single, uninterrupted conversation punctuating the lifetime of individuals and their community (and still unachieved when both pass away). Chats on the phone will of course have, like all conversational fragments, recognizable textual features (phone calls are partly ritual, as CA has shown) and even goals. But insofar as the call is ‘conversation’, it will not be a ‘text’. Thus, a specific kind of non-textual competence is needed for analyzing a conversation as a researcher, as well as for co-creating one as a participant.

Travelers, for example, know that, while mastering a community's conversational rituals is sufficient for service exchanges (e.g., knowing "Excuse me; how do I get to..." is sufficient for asking directions), it is rarely sufficient for entering into communion, through conversation, with the members of that community — no more than mastering the rituals of prayer in a given religion puts one in communion with the god being prayed to. Conversation is different from ‘information exchanges’ precisely because it is an ‘entering into communion’. As such it requires making a community's existential value system one's own, in order to spontaneously recognize and react to the units of intent that make up ‘free talk’ in that community. Recognizing existential units of intent as an outsider is not easy. This is why it is harder, in a foreign language, to learn how to converse as ‘one of the group’ than to create sophisticated forms of bounded, explicitly finalized discourse (i.e., ‘texts’: academic papers, sales talks, literary reviews...), the structures of which can be thoroughly described by teachers or learned from handbooks.

4.2.4 If conversation proper is not a ‘text’, then what is it concretely? In this paper, it will be treated as an ‘intentional event’ — specifically, as an on-going, situated enactment of a collegial will to make sense of ordinary experience, accompanied by verbal/prosodic/somatic improvisations. Thus, while any verbal interaction has textual features which must certainly be accounted for, we shall consider the basic units of conversation to be neither physical objects (‘utterances’) nor discourse functions (‘turns’) nor logical constructs (‘felicitous acts’) but rather intuitively-grasped enactments of will, or ‘stances’. A stance is the perceived intent behind a basic behavioral pattern.

How can such a subjective criterion be scientific? For now we simply appeal to the ordinary experience of the reader. In general, when we encounter manifestations of intent — a strange movement made by a passer-by in a dark alley or human-like cooing coming from the apartment next door — we recognize them as willful acts on the strength of intuition alone, i.e., without explicitly calculating the degree of their deviation from randomly generated acts of similar nature. In other words, an intent ‘makes itself felt’ to us clearly and distinctly, even if we are not sure of what it means. Of course at times we are mistaken: we project intentionality on phenomena we later find have none; thus we discover the pitfalls of intuition and learn to practice preventive strategies, like inner as well as outer listening. Whatever level of expertise we reach, however, sensing intentionality still remains the soul of our conversational ability. It is what conversation analysts rely on all the time when they confidently assert the ‘meaning’ of other people’s talk.

4.2.5 Treating conversations as interplays of ‘intentional events’ — and not exclusively as verbal or verbal/prosodic/somatic ‘texts’ — is not only closer to our experience of them but also more useful. First, it permits us to use a wider range of data in our textual analyses of the frameable parts of a conversation (to be specified in 4.5). This allows us to explain interactions like a TV talk show on politics in which the skirting of a certain issue by A provokes a protest from A’s rival B (reticence is, in fact, an ‘intentional event’); or one in which A respectfully lets B talk at length without interruption (deference also constitutes an ‘intentional event’); or one in which A sadistically sits in silence letting his expensive clothes show as B, stammering, constantly tugs at his overly-short jacket cuffs (this normal exhibition of self by A is just as much an ‘intentional event’ as was the act of putting those clothes on at home); or, finally, one in which A chooses to remain in silence after B, having made a terrible admission, relinquishes his turn (if A’s silence lasts a long time, the turn goes back to B; if B has nothing to add, his silence, with A’s, becomes a simultaneous silent turn). Occurrences such as these constitute genuine communicative acts by A, through doing nothing (see Hall 1966 and Watzlawick et al. 1967 on covert communication; novelists, too, regularly note occurrences like these in describing conversations). And yet they are not "verbal attempts at turn-taking" nor even, strictly speaking, multi-modal attempts. If, however, we consider conversations to be made up of units of intent, then we can treat them structurally, in order, as: alternating turns in a side sequence, alternating turns with immediate turn relinquishment, backchanneling, and an alternating turn followed by a simultaneous turn.

More importantly, by considering experiential investigation through cultural assimilation as a valid research tool for grasping units of intent (the procedure we shall suggest in Section 7), we can gain access to forms of communication that remain obscure if analyzed as ‘text’. An obvious example are certain Native American cultural practices:

When used in a special way by Blackfeet, the term 'listening' refers to a form of communication that is unique to them; when enacted in its special way, 'listening' connects participants intimately to a specific physical place. [...] 'Listening' this way can involve the listener in an intense, efficacious, and complex set of communicative acts in which one is not speaking, discussing, or disclosing, but sitting quietly, watching, and feeling-the-place [communally], through all the senses. — Carbaugh (1998)

Indeed, the complementary heuristic we shall be proposing at the end of this paper can give investigators better access to any intercultural interaction (native/non-native, boss/worker, woman/man...). Only by learning how to sense, represent and introject the Weltanschauung of their conversational interlocutors can investigators grasp the ‘enactments of intent’ that direct the flow of the conversation in which they seek to participate.

4.3 This paper, then, will view conversations as boundless meaning-making events, moved forward by bounded sequences of stances, i.e.,perceived minimal enactments of intentionality. (See ahead for the bounding criteria.) It is possible to sense a stance without fully grasping the intent behind it, as we have seen, simply by noting a suggestive pattern of behavior. But to understand a stance fully, we must grasp the intent giving it shape within the context of both the whole conversation and the particular sequence in which it appears. (See Gumperz 1992, 1995, on contextualization.) As a unit of effect — what we are ‘summoned’ to feel — a stance is the equivalent, in the realm of intentionality, to a ‘speech act’ in the realm of discursive thought (Austin 1962: see ahead). As a unit of perceived expression, a stance is a bundle of signals, i.e., triggering occurrences. Examples of signals: stammering, keeping silent, letting one's expensive clothes show... Example of a stance: while B stammers, A seems to keep purposefully silent, letting his expensive clothes show... (which is seen as a sadistic stance only if the dynamic of the whole conversation is grasped). The basic unit of conversations is thus not the signal but the bundle of signals (triggering occurrences) within which each occurrence may acquire its presumed intentional value — i.e.,the stance.

4.4 The Oxford Ordinary Language philosophers, and many pragmaticists with them, have adopted a different, quasi-legal perspective in defining the basic units of communication. To converse is to ‘felicitously’ do something with words and/or gestures (Austin 1962). This is clearly an oversimplification, as Goffman (1981) has pointed out: people do much more with language than accomplish speech acts; they indulge in phatic communion, for instance, or talk to keep from thinking. Thus, while Austin's ‘speech acts’ may constitute the building blocks of the world of discursive talk, the world in which some philosophers choose to live, they do not qualify as the building blocks of all talk, much less conversation. Moreover, a definition of the basic units of communication should establish at least some correlation between units of expression and units of effect. But speech act theory can do so only with propositions, not with the co-constructions and multi-purpose utterance-turns typical of conversation (see Goffman 1981 for examples). Nonetheless, Austin's basic notion of speech as ‘doing’ (1.creating a value; 2.modifying/advancing a state of affairs’) seems worth keeping. In fact, applied to enactments of intent, the first sense defines ‘stance’; the second, ‘sequence of stances’.

But what exactly constitutes a ‘sequence’? Taking inspiration from Kerbrat-Orecchioni (1997), Colas & Vion (1998) define the minimal unit of conversation as a multi-phase contribution. (Also see the notion of ‘mutual adjustment’ in Clark & Wilkes-Gibbs 1986, ‘contribution’ in Clark & Schaefer 1989, and ‘exchange’ in Kerbrat-Orecchioni 1994.) Each ‘contribution’ has a three-part structure: A makes a verbal offering;B acknowledges the offering; A acknowledges the acknowledgment (explicitly or tacitly).

This schema elegantly explains how a whole series of utterances can constitute a single functional unit. Nonetheless, as worded, it clarifies conversational exchanges uniquely as exchanges of information — more precisely, as epistemic or deontic transactions (a tribute to the Austinian tradition?). Our objective, on the other hand, is to clarify how conversation works as an alogical world-building practice. Thus, we shall borrow the term, conserve its primitive sense ("finalized transaction made up of a series of turns"), but define it teleologically as a "perceived claim to value" (Stuart Hall, lecture). Contributions make a conversation advance, as political struggles make a nation advance. They modify a state of affairs. The entire storyboard in Appendix 1, for example, represents a single contribution: PHIL’s defense of an American idiom (a claim to value), with acknowledgment by the GROUP and PHIL’s acknowledgment of the acknowledgment.

We shall not consider here units superior to the ‘contribution’. Our thesis is that intentionality drives conversations. Larger units (cognitive ‘enhancements’) are derivative.

4.5 Drawing on Goffman's notions of ‘frames’ and ‘moves’ (1974, 1981), Schank's notion of ‘scripts’ (revisited in Schank & Leake 1989) and the existentialist notion of ‘will affirming itself’ (Sartre 1943), we have defined stances as the basic units of intent which go to make up contributions and thus conversations. But what are they in practice?

For the moment the reader can get some idea by glancing once again at the conversational fragment given in Appendix 1 in the form of a cinema storyboard. Clearly, in order to choose a video sequence for the storyboard, the authors of this paper must have had some idea of what had struck the participants as a ‘claim to value’ (or ‘contribution’) during the conversation. Debriefing revealed that PHIL’s irruption into the conversation to comment on the expression "Give me a break!" had struck everyone. Next, the authors had to define the beginning and end of that contribution and divide it into stances. Those choices were not automatic. A video tape is a continuous flow of visual and audio phenomena: some principle must have therefore guided the authors not only in subdividing that flow, before and after PHIL’s guffaws, into a coherent unit with a start and a finish (the ‘contribution’), but also in subdividing that contribution into single units to be pictured as ‘frames’, i.e.,the six boxes making up the storyboard. Whatever that principle was — and we shall try to describe it further on — Appendix 1 illustrates it.

4.6 We may therefore define a ‘frame’ as the pictorial representation of a stance. It is a bundle of individual/collective signals (e.g., changes in gaze and head position) seen as enacting a personal/group intent. This is represented pictorially by keeping all visual elements intact from one frame to the other, except those in which change is seen as potentially significant and indicative of intent (see Note 6). Since all other changes are perceived by the participants as either background or as noise, they need not be pictured. As a rule of thumb, frames — like stances themselves — represent the shortest sequences (or ‘flashes’) that a Video Editor might cut from the tape of the interaction to make a brief TV spot to publicize the conversation, like spots made for films. Minimal units give enough (not necessarily all) of an utterance or gesture to make it seem like the concrete, intelligible expression of some will directed to moving events forward.

The storyboard makes it clear that, in any case, the minimal units of a conversation are neither utterances/utterance-turns, nor speech acts, nor transactions. The frames were not chosen on linguistic grounds (one contains no utterance and one contains two that work as a single enactment) or to represent discourse functions (although all, including the silence in frame 6, perform one or more ‘illocutionary acts’ and frame 2 contains a topic/comment exchange). The criterion used was existential: every frame had to capture one minimal aspect of the overall attempt (on the part of the participants in the conversation both as individuals and as a group) to ‘get a hold on things,’ and thus to create or reinforce a world-view. To borrow an expression from an Italian theoretician of literature inspired by Althusser, every frame shows, in miniature, an "attempt to implement a project of acting on the world" (Liborio 1979:9, our transl.).

4.7 Now that we have ‘units’ suitable for analyzing conversations as existential meaning-making events, we may proceed to examine a segment of our video recording. We propose to: 1. identify all phenomenological regularities using traditional nomenclature (‘posture’...) and procedure (distributional analysis...) and, simultaneously, 2. assign meaning to these regularities using our new nomenclature (‘stance’...) and a yet-to-be-defined method for grasping experientially the overall sense of the conversation.

The need for such a method is evident. We have asserted, in fact, that signals can be perceived as intentional only within a stance; a stance can be perceived as such only within a contribution; and a contribution, only within the development of the conversation-up-to-that-point. Moreover, we have asserted that a conversation is not frameable and can be grasped only by experiencing it. These assertions would seem to imply that to be able to recognize a ‘stance’ in the verbal/prosodic/somatic ‘posture’ of a person, we must actively participate in the conversation in which the posture is struck; what is more, we must be a bona fide member of the interaction, able to ‘get the feel’ of ‘what’s going on’ and, insofar as possible, to work out meanings dynamically with the group.

If instead we try to make sense of a conversation from the outside, we may fail to notice many of the stances and contributions as perceived by the participants: but these are what determines (subjectively) the development of the conversation! The risk is especially great with covert stances (4.2.5). In addition, we may misinterpret the various movements we notice by projecting onto them our own world of values (‘projective interpretation’), imagining ‘stances’ and ‘contributions’ where there are none (for the participants). In other words, we may manage to make sense of the conversation for us, but fail to capture what is going on "from the natives’ point of view" (Malinowski 1944).

To conclude, it would seem that there is no way we can reliably establish the idiosyncratic meanings of the interactions we recorded. Current methodology can legitimately uncover only their prototypical meanings by using psychological/cultural universals to frame the interactions (e.g., Grice’s conversation ‘maxims’, Goffman’s conversation ‘system requirements’, Sacks’ ‘turns’, Brown & Levinson’s ‘face-work’...).

This dilemma, called the hermeneutic circle, has various solutions (Gadamer 1960:312 seq.); we shall propose ours in Section 7. In carrying out the present experiment, however, we chose to take a shortcut in order to expedite analysis. We substituted actual participation in the conversation with ‘virtual participation’ obtained by debriefing the participants and then attempting to view the video with their eyes (as though we were living it as they had lived it). How well this procedure works is what we shall now see.