transactional resources (Goffman's ritual idiom)that are to be used by conversationalists in the service of particular face claims, such as the reliance of one upon another that the latter would not demand an accounting of certain face claims he makes, illustrates an important aspect of the nature of the interlock between transactional work and conversational displays. To begin with, we note that certain particular conversational displays have acquired, by convention, a prominent role, a visibility status whose use is thusly marked, and hence, could be called idiomatic. Examples of such transactional idioms include both non-verbal and phonological performances: (a) gestures and utterances that serve ceremonial functions (e.g. greetings, apologies, invitations, forms of denial, promises, acceptances, orders, polite requests, and so on); (b) gestures and utterances that serve signaling functions (e.g. hand-holding, tongue-in-cheek, intent listening posture; address forms, pronouns, qualifiers of relationships-my brother, my friend, and so on); (c) gestures and utterances that serve legitimizing functions (e.g. nodding, embracing, agreeing, voting, recognizing, believing, and so on): and, to mention just one more in a list of presently unknown length, (d) gestures and utterances that serve stylistic functions (e.g. ties, recognizable gate, body shape; the overuse of certain expressions, the bonding of others to a restricted context-either linguistic, topical, or physical setting).

To pick up the main thread of this section: we have been discussing the nature of the relationship between transactional work and conversational displays. We have given a personal illustration of how one of us uses a particular gestural display-eye scanning with a preoccupied demeanor, as a strategy to block a countermove. We have speculated on why the strategy succeeds and proposed that it is an instance of a larger category of resources that are available to conversationalists in the service of ritual work. We have noted that such transactional resources may become conventionalized in form as transactional idioms, which are conversational displays that are thusly marked, i.e. they are seen by participants as having particular functions. We mentioned four kinds: ceremonial, signaling, legitimizing, and stylistic functions.

Because transactional idioms are overt displays that are both fixed in form and marked for their function, they represent a category of conversational events that can be usefully exploited in the task of determining the nature of the interlock between transactional work and conversational displays. The reason we, as analysts, can exploit transactional idioms in this way for our purposes is not unrelated to the fact that conversationalists themselves exploit their use in the service of their face work, for it is the study of the latter's strategies in using them that makes them serviceable for our purposes. The symmetry between our theoretical concepts and methods of empirical analysis, on the one hand, and on the other, the participant based oriented to features is therefore nicely maintained.

IV. REGISTER MODALITY

One definite thing we can say at this stage about the relationship between displaying gestures and utterances and transactional work is, then, the fact that conversationalists will exploit the use of transactional idioms in the service of their transactional intentions. This use is routinely permitted by the transactional code and requires the organized cooperative interplay of others to succeed.

The specific use of transactional idioms in the service of individual momentary transactional intentions is a notion that we shall refer to as a register modality, and it will be seen presently, that register modality is precisely the interlocking mechanism we've been looking for between conversational displays and transactional work. We hope to show that our elaboration of the notion of register modality as this kind of an interlocking mechanism solves rather nicely certain difficult problems about "register" as this concept has been traditionally discussed in the literature on speech varieties.

In the illustration given at the beginning of this chapter, see Section II, there was a bracketed section of the piece of conversation we analyzed that we discussed as "playtalk" (alternating turns (3) to (12) ). The opening boundary of the section consists of an utterance that has the following structural properties: from the language-as-communication point of view, it is a reply to a request for information ( (2) B: What's up?; (3) A: I am tape recording you; (4) B: You're kidding me?). This is missing the point of what's going on. Plainly, what is going on here is that B is startled at the unexpected sight of the tape recorder and displays it in his utterance and gesture. In what sense his utterance is a request for information (what information?), and in what sense hers is a reply, is not clear to us. Instead, we'll try a transactional point of view. When A opens the door to B and they suddenly find each other in a standing conversation, there is an instant in that confrontation, a brief holding moment that symbolizes nicely its fateful significance. For in that moment of transition from being alone to being with, the formidable transactional machinery is being switched on and it is upon its smooth operation that both are replacing their faith to escape injury and avoid disaster. With the transactional machinery now clearly in operation (they both can be seen to be smiling), he notices, among other things, that she is wearing a dress, that she has a yellow egg smear on the lower left corner of her mouth, and that she is holding a tape recorder in one hand. All those are public facts, seen as such by all concerned. Among these public facts is the fact that the dress on her and the egg smear on her face are not as unusual events as the tape recorder she is plainly holding. The transactional machinery alters the status of this noticeable feature and assigns it to the category of "potential first topic".

There are some very interesting sorts of questions that relate to the possible format and content of the transactional code. In this instance, we may wonder whether such categorical structural components as "potential first topic" are content filled as well as process marked, or only the latter, and in either case, which way and how. Thus, are there inquiries about the health of the other and the other's relatives, that the code specifies in advance as possible first topics, and thus excludes topics that are not thusly marked, or, does the code specify process features only, that, when run through in a particular setting, select out possible first topics for that conversational occasion? More discussion on this will be found in the next chapter where specific problems in topicolization in conversation are examined.

To get back to our illustration. It is plain that much more is involved in the exchange "What's up?", "I'm tape recording you" than a question-reply exchange, for having said the latter, even if it were true, not much has been said. Reference to the transactional machinery allows us to make more substantial statements about what's going on. Here, we are discussing in particular, register modality which refers to the syntactic mechanisms in operation at the interface between two modality systems in interpersonal exchanges, the underlying transactional system and the surface display of conversational events.

Students of contemporary linguistics will no doubt recognize the theoretical affinity between the concept of register modality and "transformations" as the interface mechanism between the underlying deep structure syntactic system and the surface structure system of phonological strings of sentences. In the experimental literature on psycholinguistics over the past decade, a great number of studies are to be found where the impression is left that it is assumed by their authors that the visible surface structure of sentences is an index of the invisible underlying syntactic structure, and in consequence, they are led to carry out experiments on the rate and efficiency of surface sentence processing as an index of underlying processes. Without arguing the point here (but see Jakobovits, 1969) we shall flatly state that we do not see great merit in such an approach. Similarly, and this time we are arguing the point, conversational displays are not to be usefully seen as representing an index of the underlying transactional work. We agree with Goffman on this point, who states that gestural and verbalized surface displays in conversation are, instead, visible evidence of the transactional machinery in operation. A gesture or an utterance is a display of a public alignment taken up by a participant vis-a-vis a transactional problem, a visible stand; though an utterance is certainly an expression of something, it is foremost a performative, not an index of something else, just as an illocutionary act (Searle, 1969) is a performative. "I hereby promise to do it" is a statement that may or may not be an index of an underlying intention to do something (who knows for sure?), but it certainly, and most fundamentally, is a speech act, a performance, a promising, hence, an alignment vis-a-vis the issue of whether or not one has promised something to someone.

Register modality can thus be seen as a transformation mechanism that enables a participant to engineer visible displays that count as or are seen by participants as taking a stand vis-a-vis some transactional issue that comes up in the interchange. A detailed examination of this transformation mechanism is, thus, another central task in the analysis of the structure of conversation.

To get back to the illustration at hand, we have already discussed, much earlier, (see Section II), what's involved in talking turns (3) to (12), namely that it is a bracketed section of playtalk that is to be distinguished from adjacent sections of serious talk. We might suspect that playtalk and serious talk form two distinct register modalities, though we must now make the case for this supposition. By saying that they are two different register modalities, we are, in essence, claiming that they are two different register modalities, we are, in essence, claiming that the manner of taking up an alignment in playtalk is distinguishable from the manner in which an alignment is taken up in serious talk. One such global difference has already been pointed out in the earlier discussion: given a first transactional move by one participant, a move that raises a specific transactional issue vis-a-vis which the other participant is expected to take up an immediate alignment (in this case, "What's up?"), a return move by the other participant can now be cast in either the serious or the playtalk register. If the latter is chosen, the form of the second, return move must be such as to be appropriately responsive to the transactional intentions of the first move, as displayed by his alignment taken up in his first utterance. In this case, B, in "What's up?" is plainly to be seen as taking up an alignment vis-a-vis the problematic status of the setting, namely, the unusual presence of the tape recorder, assigning that issue the status of a first topic. An appropriately responsive display in the serious talk register would now require A to take up an alignment vis-a-vis B's earlier alignment, viz to provide a suitable remedy as given by the transactional code: if a participant, B, finds himself in jeopardy as a result of the doings of another participant, A, the latter is expected to come to his rescue by providing some remedy that will extirpate B this predicament. The transactional machinery may allow a choice between a number of equivalent remedies to a problem. One kind of remedy to be had within the serious talk register, in the present case, is an accounted apology and a display that constitutes such an alignment might be an utterance of the following form: "I'm sorry to thrust this thing upon you, like this, unexpectedly, but I need some data for a course on conversational materials. I hope you won't mind. Come in. I'm almost ready. Etc." An utterance display of this form can be seen to be an alignment A takes up whereby she shows awareness and concern for B's predicament, offers an apology to repair any potential offense, and requests cooperation in the endeavour.

A second method of providing a remedy is here available in the playtalk register: given that A and B are friends ("pairs"), they are allowed to display alignments which ordinarily would be considered an indignity to another's face, but in this case, it is not, given the special dispensations that pairs owe each other by virtue of their benign intimacy. In such circumstances, it is then possible to perform an otherwise indignity, such that the offensive act will be seen by all as the type of indignity that pairs are permitted to perform upon each other. Because such acts are now marked for their special status, i.e. they are idiomatic, they may be displayed without running the usual risks, and such actions will then be seen to belong to a playtalk register.

A nice piece of evidence for this interpretation is to be found in situations where third persons are present when a pair interacts, where these third persons cannot be assumed to be knowledgeable about the status of the with as a pair. In that case, performing the usual allowable indignities by the second of the pair becomes a potential source of embarrassment to the first, in that he can be seen by others as someone who has been treated in an undignified fashion. Similarly, there is a source of embarrassment to the second person of the pair, in that she could be see as a type of person who inflicts indignities upon another. In the face of such dangers, the pair can be seen to engage in additional remedial displays, this time for the benefit of the audience, to reassure them that what they are forced to witness is not really an exchange of malevolent indignities. They may, for instance, be seen to sustain a fixed jovial facial expression throughout the interchange as a way of indicating that they are only joking. Or, they may engage in side-talk or sub-grouping (see Kochman, (1969) and Mitchell-Kernan (1969) on loud-talking in the Black Ghetto), performing various other displays to indicate the same position.

In the traditional literature, types of registers are often discussed in terms of setting features: the face-to-face register, the telephone register, the academic register, the polite register, the informal register, the lecture register, the debating register, the office register, the formal invitation register, and so on. Though it may be true that setting features such as physical location and role of participants may be used to identify patterns of exchange displays, we see this possibility as arising from a correlational status, and, in any event, the problem of specifying synctactic differences between registers remains unsolved. Furthermore, changes in the register of displays can be seen to routinely occur within a conversational episode, while setting features remain visibly constant, thus casting further doubt on the traditional approach. The discussion that follows will elaborate upon the solution we favour.

Analogic reasoning has its inherent problems, yet we risk the following analogy, not to profit from analogic extention but as an aid in pointing up some of the features of the mechanism of register modality. Consider the functional properties of a car engine's transmission. The gear system functions as a transformation device, putting two different functional component sytems into effective interaction: on the one hand, the rotation power of the engine, which is controlled by the accelerator pedal, and on the other hand, the speed rolling capacity of the front wheels. A coupling system is necessary inasmuch as the structure of a car engine is such that there is an inverse relation between its ability to overcome the inertia of the car's weight and its ability to turn at high speeds for faster forward locomotion. Thus, when a car is at rest, it has maximum negative inertia, and consequently, if the engine were turning at very high speeds, it would not be able to get it moving (as we know so well when the engine stalls on cool mornings). On the other hand, if the engine were turning at a very slow speed, it could get the car moving but it would then take an inconvenient length of time to get it to move up to a much faster speed. The transmission system is a nice solution to the problem: when the car is at rest, it couples the wheel to a slow turning, but heavy pulling engine; when the car is moving slowly, but moving, it uncouples the two systems sufficiently long to allow the engine to speed up quickly to a higher mark, without the encumbrances of the car's inertia; when the engine reaches its second hgih mark, the transmission recouples the wheels to the engine long enough to get the two moving at the same higher rate; at that point, the coupling operation is recycled (up to "three gears" in ordinary cars, many more in racing cars).

Now consider the functional properties of register modality, as we have thus far proposed. This structural mechanism functions as a transformation device, putting two different functional component systems into effective interaction: on the one hand, the dynamic properties of transactional intentions and face claims, whose legitimacy is both dictated and upheld by the prescriptions of the transactional code, and on the other hand, the time-bound, sequential performance of conversational displays. A coupling system is necessary inasmuch as different and independent time requirements govern the operation of the two system. There is no one-to-one relationships between underlying transactional motivations and conversational displays. Thus, while it is ordinarily the case that a participant has a large number of transactional requirements at any moment in an encounter, only a limited proportion of these can find expression in overt displays in any particular exchange, so some kind of device is necessary to order transactional requirement sin terms of immediate priority and moment by moment selection for expression. Register modality accomplishes this selectional task by effective moment-by-moment re-ordering or priorities for the various simultaneously relevant transactional requirements. We now need to look a the details of this selection mechanism. Specifically, we may ask what are the bases of ordering transactional requirements that have a momentary priority of expression? One entry to this problem might be the intuitively plausible hypothesis that there are at least two kinds of transactional requirements, short term and long term. The operational rule that a sort term requirement ordinarily (i.e. if not specifically marked otherwise) takes precedence over a long term requirement is a first sort of cut whereby the register modality machinery may begin its selectional activity. Short vs. Long term requirements are defined by the transactional code. At one end of the scale, we find the highest priority of expression attached to displays that are structurally second-pair members of adjacency-pairs whose first member was displayed

(continue here)