'[The ethnographer] looks upon the task of getting names for things not as an exercise in linguistic recording, but as a way of finding out what are the '"things" in the environment of the people being studied.
C. Frake(1980:2)


We have reached the point where we can summarize our more discursive discussions of the various rhetorical modes. I now formalize the specific characteristics of the modes in their linguistic realization and in their semantic connotations.

We saw that the school's rhetoric deals with the social nature of man through four symbolically signified positions. These four are:

  1. The singular person by himself;

  2. The person in group (the person with other persons);

  3. The human activitiy;

  4. The collectivity of performers of the activity.

Each aspect is signified by a particular way of organizing a local utterance. This form then feeds back into the local situation to suggest a definition to the interlocutors who remain "free" to accept or reject the suggested identification. Each form constitutes what I have labelled a "rhetorical mode." I refer to each mode in the following manner:

  1. The personal mode;

  2. The group mode;

  3. The generic mode;

  4. The collective mode.

We may think of this paradigm as being part of a kind of rhetorical morphology. The paradigm defines the forms which rhetorical expression can take and their semantic value.

Each mode is strongly differentiated from all other modes by a set of distinctive features derived from the possibility of using the presence of internal variation in syntactic structure for new purposes. This is the mechanism which allows phonemic signification to be built upon phonetic variability[ftn 1] These distinctive features are the following:

  1. Personal mode: First and second person marked for "exclusivity" (Lyons, 1968: 277) or "minimal membership" (Hymes, 1972), first names, third person singular or plural forms with immediate deictic reference to previously named persons. This mode can be used for address and reference;

  2. Group mode: First person plural forms marked for inclusivity in address and inclusivity or exclusivity in reference;

  3. Generic mode: Lexical words labelling human activities used in the singular, third person singular referring to a human activity but not to nameable persons, last name of persons;

  4. Collective mode: Lexical words used in the plural labelling human activities and referring to a collectivity (whether this collectivity can be labelled in terms of specifiable generic activity or not).

Each mode is associated with a specific manner of predicating attributes to the various signified units:

  1. the personal mode: Any predicate is specifically identified with the subject who is not marked for anything except his position as subject;

  2. the group mode: Any predicate is assumed to identify every person encompassed within the suggested group who is not marked for anything except its positions as subject;

  3. the generic mode: Any predicate is assumed to be associated with the activity; the predicated form is explicitly marked for a content which summarizes the predication (i.e., the total statement is a tautology);

  4. the collective mode: Any predicate is assumed to belong to the whole class; the predicated form is explicitly marked for a content which summarizes the predication.

We saw repeatedly that any mode can be realized in extremely miscellaneous manners. Any first name will do to signify the personal mode, any label of activity to signify the generic mode. Many forms cannot, by themselves, perform the signifying task. Outside of a context, "they...," can be a realization of either personal, generic or collective modes. Only its immediate textual environment (which, in oral discourse could very well be non-verbal with any specifying referent remaining implicit) can help to differentiate a personal from a generic from a collective "they..." Thus it is not any one morphological or syntactic feature which distinguishes a mode, it is a bundle of features. This should remain constantly in the back of our minds as we move on, particularly when we read the shorthand form I am adopting to symbolize each mode in my writing:

  1. Personal mode        /I/

  2. Group mode        /we/

  3. Generic mode        /it/

  4. Collective mode        /they/

As I did earlier, I an using the convention /x/ to symbolize a significant -emic unit. I have chosen the symbols I, we, it, they to stand for the "x" mainly for their evocative value. I am not intending thereby to suggest that each symbol is somehow 'typical' of the mode or its most common realization.

The principle of contextual signification must be generalized. What makes an instance of [they] a realization of either /1/, /it/ or /they/ is the accompanying text, including the type of predicate and the, perhaps implicit, possibility of identifying specific persons with the predicate. However, what makes /I/ and /we/, for example, different from each other are not qualities which they would possess. For example, there is no social need to distinguish collective from personal attributes. And yet, the school, in its speech and disputes, insists strongly that collectivity and personality be distinguished. A discourse about personality is not a discourse about a collectivity and vice versa. But it is precisely because collective or generic discourses are possible that a personal discourse can stand out as a specializedform. Thus, within single utterances, participants can manipulate, directly or indirectly, implicitly or not, all modes without confusing the signifying features.


The preceding outline can be related to some of the things which we discussed much earlier in relation to the nature of teaching (and acting in general) both in its quality and quantity. In the process, we begin to characterize the consequences of the possibility of establishing various syntagmatic relationships between the different units which constitute the rhetorical morphology. Even though I have not stressed the matter much yet, it should be clear that no total utterance (including both a statement by a particular person and the various possible responses) is ever made that is totally set in only one mode. All actual utterances contain parts set in various styles. This is how they gain their overall rhetorical power (including the power to convince, to confuse and to raise scepticism) over a particular situation. This indicates the need to look at what happens when various modes are used as immediate contexts to each other. This analysis will help us understand the cultural source of certain of the most powerful difficulties that all the participants encountered. In this section we do this in terms of the formal symbolization just introduced. In the next two parts of the book I expand on the analysis in a less formal manner, in the terms used by the participants.

At the close of our earlier analysis of the talk that surrounded the definition of teaching, of its evaluation and administration (Chapter V), we abstracted a set of topics from the participants' conversations, things which they insisted had to be distinguished and handled in someway. We could not then formalize these units any further. These topics of discourse were the following:

It can be seen immediately that these topics can be divided into definitions of unit (/role/ and /per-son/) involved in interactions, and definitions of interaction (all other topics). The definitions of interaction essentially consist in statements of organization of action. The definitions frame a certain sequence of action in its semantic power.

It is easy to identify the topics of /role/ and /person/ with the emic units /it/ and /1/. To signal that one is talking about a role, one would use the generic mode. Conversely, use of the mode signals that the utterance is about a role. It is easy to verify that this association between role as a topic and the generic mode is an extremely powerful one in American culture, from halting verbal utterances in everyday speech, to administrative rule writing, to the specialized discourse of the social sciences. We might want to see in /role/ the signified of which the generic mode is the signifier. My temptation in this work is to reverse this equation and to anchor the ability of Americans to talk in terms of roles in the availability of a rhetorical mode. The same thing can be said about the topic of /person! and the personal mode. Persons exist to the extent that /I/ speech is available.

The other topics cannot be dealt with so easily. It must first be noticed that none of the topics correspond directly to the two other units we identi-fied. This is an artifact of our starting with conflict over evaluation. In the preceding chapter, we began to see that the group mode corresponds to the topic of /friendship/ and that the collective mode corresponds to the topic of /cliques/. The other topics correspond to actions performed by the various units as they relate to each other. The units, we saw, do not stand alone. All persons, for example, have roles. All persons must deal with other persons from the point of view of their reality as persons. Wherever we start the paradigm formed by the rhetorical mode, each unit has to deal with the three other units and this relationship is symbolically marked. In more abstract terms, the syntagmatic relationships that can be established between the paradigmatic units specify the exact rhetorical power of utterances and, thereby, contribute by suggesting a pattern of interpretation for any utterance.


  1. The relating of /I/ to /it/ is done through attention to the manner an individual performs acts directly relating to his role definition. Any attention to this implies that the individual is treated in relation to his role (/actor-in-role/);

  2. The relating of /I/ to /we/ is done through attention to the manner the individual participates in a group, how he is intuited by members of the group (/the person's friends/) ;

  3. The relating of /we/ to /it/ is done through attention to the way a group can be identified with a role which many of its members may hold (/persons in role/);

  4. The relation of /it/ to /they/ is done through attention to the fact that any role is held by many individuals and that these can be dealt with a collectivity (/actor-in-role collective/);

  5. The relation of /we/ to /they/ is done through attention to the manner two groups interact (segrelation/);

  6. The relation of /I/ to /they/ is done through attention to the manner a person relates to people in groups in which the person does not participate (/prejudice/);[ftn 2]

These six sets or relationships can be displayed visually through four triangles:

1)        /I/                
/the        /actor-in        



/segregation/ /the person's friends/

/they/ /prejudice/ /I/

3) /it/

/persons in role/

/ator in role/

/we/ /segregation/ /they/


/actor in role collective/
/it/        /I/ /actor-in-role/


It may seem redundant to list each signifying relation twice. This is, in fact, central to the analysis. It can help us understand formally how people can be both persuaded and confused to the point of anger by rhetorically appropriate speech. For example, /segregation/, the relating of /we/ to /they/ can be realized either in the context of /I/ or in the context of /it/. In the first case, the emphasis is thereby put on the personal aspect (friendship) of the relation-ships within the two groups ("the one to which I belong vs. the one to which you belong"). Also enphasized is the fact that one cannot quite specify what separates the two groups. It directly relates the talk to general discourses about the "freedom of individual choice" and the ensuing legitimate segregation that is the product of the individual's "freedom of assembly." In the second case, the relating of /we/ to /they/ in terms of /it/, places the emphasis on the role content of the various associations, and the fact that it is possible to specify what separates the groups. The role content can be an administrative category (teacher, administrator) or it can be a label referring to an ensemble of life style choices (e.g., "jock," or "freak"—see Chapters X and XI). Emphasized is the absence of freedom of the participants, and the presence of conflicts that transcend personalities (or that should do so). This kind of talk is directly related to general discourses, like those that make it illegitimate for supervisory personnel to belong to the same union as the people whom they supervise. It is the same discourse which makes it legitimate for the State to infringe upon misapplied freedoms (the grounds on which desegregation suits are argued, the grounds that make cliquishness shameful). As we see, the possibility that talk about groups can proceed either in terms of the personality of the persons composing them, or in terms of a general label, has the effect of making an intermittent phenomenon highly dependent upon the symbolic form of the expression out of something as apparently "solid" as cliques.

All this brings us back to the kinds of matters that have been with us from the beginning, matters of confusion, anger, uncertainty and doubt. Until now I have mentioned them only to emphasize that all conflicts are fought in an atmosphere rigidly structured by the redundant suggestion that these few "things" are the things to fight about. Now that we have formalized what these topics of discourse are, what their formal shape is, and how they function, it is time to focus more deliberately on the conflicts themselves. We begin with a brief re-analysis of a text in the formal terms just adopted.

An Analysis of Text

The preceding is a systematic formal summary of the analyses we conducted earlier in a more discursive fashion. Let us now look again at one of the texts we already discussed (T24, see p 136), the "teacher surplus" memo. This will illustrate the usefulness of the formal analysis:

1) "TO: FAC"

With no further specification, this could be an /it/ or a /they/. The specification of activity (faculty) means that this could not be an /I/ (which could have been realized as *[each of you]). By itself, however, [FAC] does not allow us to distinguish whether it is intended as a diacritic marker to help the secretary decide to whom to give the memo (making it a realization of /it/), or whether it is intended as a collective. Given the rarity of /they/ in address, this is the less likely alternative.


The subject is impersonal. Nobody in particular is intrigued though it can be assumed that it is human beings who can be intrigued. No person or collectivity are specifically referred to in "teacher surplus" (the memo does not say *"there are too many of you," nor does it say *"there are too many teachers"). This is a prototypical example of the generic /it/mode.


[Now] and [our] both anchor the sentence in a time frame, a place and a social group neither of which, however, are exactly specified lexically in any way. The reference is all inclusive. The /we/ thus frames the /it/ of the beginning of the sentence. It opposes an unspecified personal group (of friends?) to a specific issue that has the same impact on all the members of the group. At this point, there is still no specific marking of the way /it/ relates to /we/ (beyond the fact that they are placed side by side). The distinction of /it/ from /we/ could be framed either by a /they/ or an /I/. In the first case the /we/ would be made to refer to people who are not teachers (e.g. administrators, parents, community members). In the latter it would be made to refer to all individuals who are intrigued.


This is an expansion of 2) and 3). Same balance exists between the /we/ of [us] and the /it/ of [education]. No further specification of the broader context.


The memo continues to assume that the unspecified personal group is still the overall subject. The earlier phrase about [teacher surplus] is expanded into [some teachers]. By specifying [some], the memo marks it as not referring, at this point, to the collectivity of teachers. It refers to a few people who happen to hold the position of teachers. To the extent that these teachers are potentially nameable after an investigation, this could be an /I/, thereby, suggesting that the relating of /we/ to /it/ is in fact to be done in relation to /1/. But some ambiguity remains.


Further expansion of the opening sentences. To the extent that [individualized instruction] is an /it/ closely related to [teacher] and, in the memo, directly associated to /we/, the implication necessarily is that, if the /we/ is to be specified by an /it/ it is /teacher/ that is intended thereby identifying the author of the memo with the teachers.


While the first part of the memo established that its author was part of a group that is its audience, this introduction to the second part subtracts the unspecified author. The author expresses a view point redundantly marked as his own. While the comments made at the beginning apply equally to the group, the following only strictly applies to the speaker/author. This is an affirmation that the contextualization of the /we/ /it/ balance in the first part of the memo is done in terms of /I/. /tie/, up to now is 'everybody,' including the speaker.


[Teachers] is a collective. No attempt is made to subtract a subset ([some]) to whom the comments would specifically apply. Same for [students]. There is no marking that [teachers] refers to the audience of the memo. However, the ambiguity that has been resolved in 7) is reintroduced though it has been displaced. To the extent that the initial /we/ and /I/ were not specified in terms of an activity (even though there was a'. some point--6)--a suggestion that, if there was to be specification, it would lead to [teachers]), the implication is now, through the present /they/ specified for [teachers], that /we/ did not include [teachers]. If sentence 7) is treated as a new beginning With /I/ opposed to /they/, a way is open to contextualize this opposition in terms of an /it/ that would specify the speaker as a [principal].

The rest of the memo continues with an alternation between the generic and collective modes as the memo alternatively refers to external collectivities and to administrative rules for the evaluation of single teachers.

No reader of the memo will ever read it solely in itself, as we just did, without bringing to the interpretation his full knowledge of the total situation. The /I/ refers most directly to [John], a person that is not marked for any administrative position within the school. But all the people involved knew that this [John] is also [Foster, John, Principal] and that his opinions carried a weight which their own did not have. Indeed an audience to such a memo is never compelled to react with texts that adopt the identification of the various protagonists and their relation-ships which it suggests. We, as analysts, are not entitled to assume that the memo directly expresses what its author had in mind. Individual authors did not have control of the rhetorical power of their discourse and the rhetoric can be shown not to have universal value of direct expression. A memo like T24 should be seen as the struggle of an author with the means at his disposal. It should not be treated as the product of someone who has vanquished a rhetoric which is now fully serving his purposes. As we know, the memo did not accomplish what Foster hoped it would. It aroused some teachers and established clique identifications within the teacher body which, two years later were fully alive.

Certainly, what made the performance of the memo something which the principal had to do to establish his authority and what transformed it into a definite (speech) act to which teachers had to respond, wasthe fact that its social referent was the structure of evaluation. But this is not all there is to say about the memo. We cannot ignore the form which this reference took. However transparent may be the attempt at co-optation which the initial adoption of the group mode suggests, the fact that there was such an attempt is signalled precisely by the use of the mode. The same act of eventual reference to a socio-economic, material reality could have been performed by many other texts written in many different ways but they would not have been equivalent. For example, a memo such as the following (rewritten using alternate rhetorical modes) could have been referentially synonymous to the one that was actually written:

"To: Each of you
    Your principal finds it intriguing that there are so many of you around. It will permit him to be very selective and to fire some of you whom my predecessor should not have hired. Since you realize that I have a definite power over you, I expect each of you enthusiastically to adopt my favorite innovation.
    Don't forget the following:
    You are accountable for your students' failure. If your students fail, you are in imminent danger of being fired whatever else you do..."

In fact, would such a memo have been truer to the situation? Probably not. Not only would it have been much more scandalous than the original one, it would also have been the signal of a deeper rift between principal and teacher than actually existed between Foster and the teachers of Sheffield High. They expected him to talk about certain things in certain ways and they probably evaluated him in terms of his ability to co-opt them more successfully. Most teachers probably accepted that they were in an unspecified personal collectivity, a group of friends, a "big family," with the principal, that they were, indeed, /we/ with him. And their complaint, more than any-thing else, was that they could not achieve this weness with him, because he insisted too often on being /they/ to them. That he could not be /we/--for social structural reasons--was the true, unspeakable, scandal.



1) The possibility of manipulating syntax in such a manner, and to such an effect, means that the traditional manner of interpreting syntax as directly reflective of situational impulses must be reappraised. See Varenne (forthcoming) on the interpretation of pronominal morphology in light of the rhetorical analysis conducted here. back to text

2) There was little occasion to talk about these topics in Sheffield. I cannot do more than suggest the direction that an analysis could take of the dual implication of the ability to form groups ac-cording to personal criteria: Such groups cannot be indefinitely large. Thus, some people will be left out. These may form their own--separate--group (/segregation/). They may also wish to be part of a group that will not receive them, and resent the rejection (/prejudice/).back to text