1
Occasioned Semantics: A Systematic Approach
to Meaning in Talk
Jack Bilmes
University of Hawaii
(This paper was published in Human Studies, 34:129-153, 2011. The final publication is
available at link.springer.com.)
Abstract: This paper puts forward an argument for a systematic, technical
approach to formulation in verbal interaction. I see this as a kind of expansion of
Sacks’ membership categorization analysis, and as something that is not offered (at
least not in a fully developed form) by sequential analysis, the currently dominant
form of conversation analysis. In particular, I suggest a technique for the study of
‘‘occasioned semantics,’’ that is, the study of structures of meaningful expressions in
actual occasions of conversation. I propose that meaning and rhetoric be approached
through consideration of various dimensions or operations or properties, including,
but not limited to, contrast and co-categorization, generalization and specification,
scaling, and marking. As illustration, I consider a variety of cases, focused on
generalization and specification. The paper can be seen as a return to some classical
concerns with meaning, as illuminated by more recent insights into indexicality,
social action, and interaction in recorded talk.
Sequential Analysis and Membership Categorization Analysis
The work of Harvey Sacks resulted in two organized traditions of research—
sequential analysis (SA) and membership categorization analysis (MCA). After
Sacks’ death, SA emerged as by far the dominant approach. This is due to a number
of factors. SA has the more elegant and, many would argue, the more rigorous
methodology. Moreover, it has a broader and more flexible approach than MCA.
And we should not underestimate the influence of Emanuel Schegloff, the most
prominent practitioner of conversation analysis (CA) since Sacks’ death in 1975.
Schegloff has been a firm advocate of SA. However, in recent years, MCA has
begun, in a modest way, to flourish again. A major turning point was the publication
of the edited volume Culture in Action (Hester and Eglin 1997 ). Perhaps the most
telling indication of the increased interest in MCA is the recent publication of two
papers on the topic by Schegloff (2007a , b ). One gets the impression (I am going out
on a limb here) that these papers were written less from Schegloff’s own interest in
the topic than as a response to others’ interest.
A central question with regard to the re-emergence of MCA is why it is
happening at all. What analytic need has SA left unfulfilled? To begin with, SA
tends to avoid (explicit) reliance on, or analysis of, cultural resources (but see, e.g.,
Turner 1971). The notion is that the analysis should be demonstrable from the data
at hand. This approach has obvious virtues. But cultural resources (language to
begin with) are brought into play in SA in a generally unexplicated way. Our
cultures undoubtedly are a major factor in the way we talk, and the refusal to deal
with culture explicitly can be argued to be a weakness of SA. The central tenet of
2
SA, to my mind, is the determination to attend to meaning in terms of the
understandings of the participants in the conversation, since it is, after all, they who
are constructing the interaction. This, of course, requires recordings of actual
situations of talk. The way that this sort of analysis is achieved is through attending
to participant responses. The response to a prior utterance demonstrates the
interlocutor’s understanding of that prior utterance, and that proposed understanding
is, in its turn, subject to acceptance or contradiction by the original speaker
(Moerman and Sacks 1988 ). In addition, the CA approach to evidence has been
more or less the opposite to that proposed by Karl Popper (1959 ). A claim in CA is
not falsified by the fact that, in some particular instance, a thing does not occur
where it ‘‘should’’. CA is not after laws but looks instead for resources and
sensitivities. If a phenomenon sometimes, intelligibly, occurs, that is proof of its
availability as a resource for speakers.1 Falsification would occur in CA only in
response to a claim that something always or never (intelligibly) happens.
These were brilliant innovations, and they have resulted in new approaches to
such fundamental matters as context (Schegloff 1991 ), description (Schegloff
1988a ), rules and analytical categories (Bilmes 1988 ), and social structure (Drew
and Heritage 1992 ), as well as methodological innovations; but, I would like to
argue, they constitute only a partial approach to meaning in conversation. It is a
result-oriented approach. It deals with what utterances end up meaning for
conversational purposes. There is no comparable systematic methodology in SA for
dealing with speakers’ choices, that is, with the rhetorical aspects of talk, or with
how certain understandings are promoted by those choices. This inevitably involves
a consideration of culture and particularly of linguistic resources and of possible but nonoccurring alternatives. Thus, such an analysis, while taking the actual talk as its
object, must, to some extent, depart from the data at hand. In addition, it involves
dealing with such matters as categorization, ‘‘scripting’’ (Edwards 1994 , 1995 ), and
other features of presentation, that is to say, with semantics and rhetoric. Not, to be
sure, with semantics and rhetoric as general, contextless topics, but with the
semantics and rhetoric ‘‘of the moment.’’
In addition, SA’s handling of conversational structure is somewhat limited.
It deals with beginnings and endings, sequencing of actions, insertion sequences,
preliminaries, prefaces, and such Schegloff 2007c ), but it falls short in analyzing
the structures of meaning that are built up over the course of a conversation.
A conversation is an ongoing construction of interrelated agreements and understandings,
of claims and presuppositions. We require a methodology that is capable of
1 CA does look for normative generalizations, but, again, the fact that a norm is sometimes violated, and
that sometimes such violations are not sanctioned, is not taken as falsification. Rather, the occasional
occurrence of sanctions in response to violations is taken as proof that participants are, or may be,
oriented to such norms and, therefore, that the norms exist. It is true that conversation analysts sometimes
make quasi-statistical empirical generalizations (Schegloff 1993), but the lack of adequate sampling
procedures makes these claims less than scientific.
3
dealing in a systematic way with the construction and transformation of meaning over
the course of interaction, and that is capable of assessing the final product as a
produced structure of meaning.
I am not saying that practitioners of SA have not dealt with any of these matters.
One thinks, for example, of Heritage and Raymond’s (2005 ) and Raymond and
Heritage’s (2006 ) work on epistemics or Stivers on marking (2007 ) or Charles
Goodwin on a variety of subjects, such as highlighting (1994 ). Ideas such as scripting
and epistemic authority are contributions to the wider field of formulation, but not to
occasioned semantics. They are themselves made up of meanings and meaning
relationships.
What is lacking is a broad framework, a methodology, for dealing with
the organization of meaning in conversation, and it is this, I think, that accounts for
lasting appeal of MCA. MCA, while rather limited in its scope, deals with some of the
shortcomings of SA as an approach to meaning in conversation, in that it attends to
choices of categories in actual situations of talk and the effects of those choices in
promoting certain understandings. In addition, it asks how categories are invoked and
related to one another. Sacks had a lively interest in categories in general, but he
developed his analytical ‘‘machinery’’ in terms of categories of persons, and his
successors, by and large, have stayed within those limits.
The best known of Sacks’ work on MCA, and the major charter for the movement,
is his paper analyzing ‘‘The baby cried. The mommy picked it up’’ (1972 ). Although
the analysis is, in my view, seriously flawed, the ideas he presents are seminal.2
I cannot see any cogent reason why these ideas cannot be applied to the analysis of all
types of categories. Furthermore, limiting the analysis to categories seems unnecessarily
restrictive. We can aspire to deal with all signifying expressions (I will return
to this), including those which are arguably not categories at all. A clear example of
the is the name of an individual, such as, ‘‘President Barack Obama’’. ‘‘New York
City,’’ although it consists of a variety of more specific places, such as Times Square,
does not seem to qualify as a category. In many cases, it is not clear what is to be
counted as a category. Housley and Fitzgerald (2009 ) treat expressions such as
‘‘untrustworthiness’’ and ‘‘incompetence’’ as categories. Perhaps more questionably,
Billig (1996 ) writes that ‘‘To say that ‘this cup is red’… is to make a categorization.
… This cup has been categorized as belonging to the wider category of red objects’’
(151).
Although this transformation from description to set may be logically
unassailable, I’m not sure that the set is being relevantly invoked. It seems simpler to
just say that ‘‘red’’ is a signifying expression. What is needed, I am proposing, is a
methodology for analyzing all signifying expressions, including categories, references,
and descriptions.
I will, shortly, propose an approach, which I am calling ‘‘occasioned semantics,’’
2 See the appendix to this article for a critique of Sacks’ paper.
4
that deals with the limitations, mentioned above, of MCA and offers a greatly
expanded methodology. It can be viewed as an attempt (a partial one, to be sure) to
deal with conversational meaning and rhetoric in a systematic way.
Formulation
Signifying expressions are formulations in themselves, and units of larger,
complex formulations. I stipulate at the outset that I am not talking about formulation in
the sense proposed originally by Garfinkel and Sacks (1970 ) and adopted by various
other conversation analysts (e.g., Antaki 2008 ; Arminen 2005 ; Barnes 2007 ; Drew
2003 ; Gafaranga and Britten 2004 ; Heritage and Watson 1979 ; Walker 1995 ); that
is, as a statement of the gist, meaning, or upshot of previous conversation. In fact,
Sacks himself, in his lectures, uses the term ‘‘formulation’’ frequently, and not in
this sense. What Garfinkel and Sacks are talking about is more properly called
‘‘reformulation,’’ because the previous conversation already consists of formulations.
So, then, what is a formulation? We can start by examining some examples of
Sacks’ (1992 ) usage of the term.
formulating one’s present state (V. 1: 69)
he formulates his remarks in terms of ‘you’ (V. 1: 166)
there are alternative ways that he and those he is dealing with… may be
categorically formulated (V. 1: 205)
formulating the session as a ‘group therapy session’ (V. 1: 515)
assign name-formulations to the actions (V. 1: 515)
In these passages, and many more, Sacks usage of the term is consistent with
the dictionary definition—‘‘to put into words’’ (wordnetweb.princeton.edu). To be
sure, Sacks sometimes uses the term in somewhat divergent and idiosyncratic
ways (‘‘we could formulate the omnirelevance of patient-therapist’’ [V. 1: 515]),
but, in general, he uses formulation in its common meaning. Now, ‘‘to put into
words,’’ implies a ‘‘something’’ that is being put into words, an object, concept,
attribute, situation, action, etc. That is, a formulation has a referent, a signification
in Saussure’s sense. I want to stress that I am using ‘‘referent’’ in a broad sense;
whatever can be referred to, from an object to a concept to a state of affairs, is a
referent. ‘‘Hello,’’ as opposed to ‘‘hi’’ or ‘‘how do you do,’’ is a rhetorical choice
but not a formulation, because it has no referent.3 A formulation may do pure reference,
as with a proper name.4 Or, it may have a descriptive, reality constructive
aspect. The notion of formulation bridges Schegloff’s (2007b )
distinction between category and reference. I am not questioning the utility of
3 It is to be noted that an expression may signify by virtue of its placement. So, ‘‘yes’’ or ‘‘he did’’ has a
particular signification when preceded by ‘‘Did he do his homework?’’
4 I am not really sure that there is such a thing as ‘‘pure reference’’. Even the use of a proper name is a
choice from among other referential possibilities and may therefore have descriptive implications or
overtones.
5
this distinction, but, given that we may speak of someone as ‘‘John Smith’’ or
‘‘the guy with the hat’’ or ‘‘the mailman,’’ we will need a term that covers the
possibility of these alternative ways of speaking.
In it’s minimal form, a formulation is a single meaningful item, such as a word.
However, it may also be a much longer unit. A narrative, by my definition, may be
considered a formulation insofar as it may be said to have a referent (e.g., what
happened at the ballgame). A formulation is always usable as an utterance but is
somewhat different than an utterance. An utterance must actually be uttered.5
A formulation may occur as writing as well as speech. When a formulation is in fact
uttered, we include as elements not only words but also features of performance,
such as gesture,6 stress, ‘‘tone of voice,’’ etc. (In fact, such elements may be
included even in unuttered formulations, insofar as they are indicated. For example,
a written formulation may include underlining to indicate stress.) The way an
utterance is formulated is obviously (partially) determinative both of its meaning
and its character as social action. More nebulously, but perhaps most importantly,
when we use ‘‘utterance’’ in analytical talk, we tend to be focusing on its role in
interaction, whereas ‘‘formulation’’ stresses ways of identifying, categorizing,
describing, and persuading, that is, the rhetorical and reality-constructive aspects of
linguistic action. (This is not to deny that formulations occur in some context and
derive part of their situated meaning from the context.)
Looking further into this notion of formulation, here is another passage from
Sacks:
Members can’t do pure formulating. That is to say, you can’t be engaged in
‘merely’—non-consequentially, non-methodically, non-alternatively—saying
‘This is, after all, a group therapy session’. To do that—even though you’re
merely invoking one thing that’s true about this—is to do other things as well,
e.g., put somebody down for something they said, propose special relevancies,
propose that some topic ought to be discussed or not be discussed, invoke a status
hierarchy, etc. At any rate, in each case that a formulation of a setting, or an
identity, is done, that’s something that has some line of consequences, and some
analyzable bases, for participants, which can be one differentiated from another
possible formulation, and also from not doing it at all (1992 , Vol. 1: 516).
Sacks is making two major points about formulations. One, of course, is that they
are inevitably a way of doing, not merely saying, something, a notion familiar to us
from speech act theory, although Sacks’ notion of ‘‘doing’’ is broader. But he makes
5 I have run across definitions that seem to include written language, but this, I think, does not represent
the common understanding.
6Some gestures are significant, in the sense that they have referents and may substitute for words.
They are, like signifying words, to be considered, in themselves, formulations. Thus, the notion of a
formulation as ‘‘putting into words’’ is too narrow.
6
another crucial point with ‘‘non-alternatively’’ and ‘‘differentiated from another
possible formulation,’’ namely, that a formulation is a choice from among a number
of alternative ways of identifying or describing the referent or producing the
conversational action. So, a formulation is a consequential choice, what the
discursive psychologists call ‘‘rhetoric’’ (Billig 1996 ; Edwards 1997 ; Potter 1996 ).
Presumably, the particular choice made forwards some line of argument, defends
some position, or otherwise serves the purposes of the speaker, but it is sufficient,
and preferable, to simply say that it has consequences. It follows that a formulation
is always local—it cannot be fully accounted for in cultural or semantic terms.
Central to the subject of formulation and occasioned semantics is the concept of
choice. When describing a person, we may say ‘‘He is tall’’ or ‘‘He is smart’’. Both
may be true, but they are very different. That is, we are ‘‘putting into words’’ two
different ‘‘somethings’’. But in saying ‘‘He is smart,’’ we are also choosing to say
that in preference to ‘‘He is clever,’’ which may express essentially the same
concept, although with perhaps differing connotations. So, in describing him as
smart, we are making, simultaneously, two types of choices. Sometimes these
choices are made obvious in the talk, as when an item is recategorized. Sometimes it
is implicit—it is up to the analyst to imagine the possible alternatives.
What, then, is a ‘‘possible alternative’’? I would suggest that what makes an
alternative ‘‘possible’’ within a particular conversation is similar reference, plausibility,
relevance, and non-contradiction. If two formulations do not refer to the same thing,
they are not alternatives. The plausibility criterion provides that a formulation that is
clearly false cannot be offered as an alternative to one that is possibly true, given the
recipient’s knowledge. So, for example, ‘‘I saw a short man’’ cannot (normally) be
alternatively formulated by ‘‘I saw a man who was about three inches tall,’’ because the
first is possibly true, that is, plausible, and the second is transparently false. The third
criterion is relevance. Consider a math teacher talking to John’s parents at a teacherparent conference. ‘‘John is tall, but he doesn’t try’’ is not a relevant alternative to ‘‘John
is clever, but he doesn’t try’’. Perhaps we could say that the relevance criterion is simply
a way of claiming that the context constrains the possible referent—so, in this case, the
referent is not John in toto but John’s intellectual ability. In any event, a fact may be
free-floating truth in theory, but in practice it is bound to a context. The final criterion is
non-contradiction: ‘‘John is short’’ and ‘‘John is tall’’ are not alternative formulations.
They both refer to John’s height, they may both be plausible and relevant, but they are
contradictory—they cannot both be true.
Formulations are not, in the first instance, ways of doing things (e.g., eliciting
offers), although they may serve such purposes. Saying is doing, to be sure, but
saying is also saying. (So, for instance, asserting something to be true, an
illocutionary act, requires the communication of that which is being asserted.)
I would argue that even performatives, such as ‘‘hello,’’ say before they do. We
must understand them as language before we can see what act they are performing.7
7 I am indebted to Arnulf Deppermann for reminding me that this is not necessarily true in the case of
7
This may seem a regression to a pre-Austinian, pre-Wittgensteinian mentality, but
we will retain the conversation analytic focus on recordings and transcripts of actual
occasions of talk, on indexical meaning, and on participant interpretations, and we
will retain a sensitivity to interaction as an analytical resource. And we are not
losing sight of the fact that formulations do something; we are merely insisting that
they also say something. In addition, as with CA in general, we will avoid
theoretical preoccupations and mentalistic explanations. We will look for what the
data has to offer.
Occasioned Semantics
Occasioned semantics deals with a subset of formulation. It deals with words or
phrases (or signifying gestures), what I will call ‘‘expressions,’’ as opposed to longer
units, such as sentences, narratives, or accounts.
My approach has been to look at various properties or operations or dimensions
associated with signifying expressions. An initial, nonexhaustive list includes.
1. Co -categorization and contrast . I place these together because they are
opposite sides of the same coin. Co-categorization emphasizes the similarity of
constituent formulations. Contrast also implies that the items belong to the same
overarching category, but emphasizes their differences.
2. Generalization and specification . Frequently, a formulation in conversation is
later reformulated at a more general or specific level. Even when this is not so, many
formulations may be said to be given at some particular level of generality. Thus,
when we say ‘‘tree,’’ we have said something more general than ‘‘pine’’ and more
specific than ‘‘plant’’. The choice of any particular level of generality is a rhetorical
choice, with certain consequences (see Deppermann 2011 , this volume; Hauser
2011 , this volume).
3. Scaling . This involves the arrangement of a set of items from less to more
extreme or intense. My interest is primarily in ‘‘implicative scales’’ in actual talk
(see Bilmes 2010 for an initial treatment). Two general characteristics of implicative
scales are that at least some of the items are not mutually exclusive and that the use
of a less extreme item implicates the absence of the more extreme items. So, for
example, offering a weak excuse implicates, but does not logically imply, that one
does not have a stronger excuse available (see Bilmes 1993 , 1995 , on response
priority). CA already incorporates ideas of upgrading and downgrading, but these
notions have been used in an ad hoc manner and have not, to my knowledge, been
topicalized or given technical specification. My approach is clearly related to
Gricean analysis and to so-called Horn scales (Grice 1975 ; Horn 1972 ) but there are
a number of differences, including attention to actual data, on-the-spot creation of
scales, and sensitivity to cultural, indexical, and sequential—not just logical or
semantic, in-the-language—factors.
language acquisition. Perhaps in the case of formulaic performatives, to understand what act they perform
is to understand them as language.
8
4. Marking . In linguistics, marking consists of several loosely related criteria.
In my usage, though, it refers to an opposition between a formulation that would
be normally expected in a particular situation and one that has the same reference
but is not, in some sense, normal. Thus, when a wife says to her husband, ‘‘Do you
know what your son did today,’’ referring to their child as ‘‘your son’’
rather than by his name, this marked usage will give rise to inferences. In this case,
we can be fairly certain that their son did something wrong (see Geoghegan 1973 ;
Stivers 2007 ).
The approach is basically structural—it deals with expressions in relation to other
expressions used in the conversation and in relation to other expressions that might
possibly have been used. I have, thus far, relied not only on Sacks’ work on
categories, but also on adaptations of ethnosemantic taxonomic analysis (see
Conklin 1969 ; Frake 1969a , b for representative examples) and implicature theory
(Levinson 1983 , Chap. 3). The result is an analysis which is systematic and
technical rather than ad hoc and intuitive. I view occasioned semantics as a step
toward a more inclusive methodology of formulation analysis, midway, as it were,
between MCA and the analysis of broader units such as accounts and narratives.
Most of my work so far has dealt with generalization and specification, and that is
what I will focus on in the remainder of this paper. Although my general approach
to talk and data is primarily rooted in CA and, in particular, in Sacks’ work on
categories, it was anthropological ethnosemantics and, especially, a fugitive
paragraph in an article by Charles Frake (1961 ) that started me in the particular
direction that I took. At a time when anthropologists were interested in taxonomies
primarily as abstract, decontextualized, static cognitive or cultural ‘‘maps,’’ Frake
recognized that taxonomic hierarchies gave individual speakers choices in regard to
how they might categorize particular phenomena, and that these choices were
strategic resources that could be used in conversational environments. Noting that
skin disease terms may enter into situations such as bride-price negotiation,
competitive joking, and maligning, Frake observes that ‘‘in many of these situations
it is imperative to speak at just the level of generality that specifies the pertinent
information but leaves other, possibly embarrassing, information ambiguous’’
(122).8
I chose to use the conventions of taxonomic analysis rather than a Sacksian
presentation for heuristic reasons. Taxonomic analysis deals with multilevel
hierarchies. Although, as Watson (1978 ) pointed out, we can conceive of ‘‘devices
within devices’’ (in Sacks’ terms), MCA has not capitalized on that possibility.
MCA-oriented work has dealt with a category or set of related categories and a more
general term that includes the subordinate categories. It is not very well adapted for
8 Categorizing, in this quotation from Frake, involves withholding of information. Typically, within the
conversation analytic tradition, the emphasis is on meaning or relevancies that are added in the process of
categorization. That is, the choice of this category rather than that is informative. By categorizing, we
reveal the nature of the thing as we would have it known on this occasion.
9
dealing with deeper taxonomic relations because it does not use diagrams. It is not
visual. In dealing with multilevel taxonomies, one would have to narrate one’s way
through a thicket of relations of generality and contrast. When dealing with
questions concerning level of generalization, a taxonomic presentation seems
clearly superior. In addition, the diagrammatic visualization forces one to consider
items that were not mentioned but might be implicated by the talk.
Case 1: Generalization and Implication
I will illustrate with material from a previously published article (Bilmes 2009a).
Although many expressions can be seen to be produced at a particular level of
generalization, in preference to being more or less specific, I have looked
particularly at what Garfinkel has called ‘‘perspicuous settings,’’ (2002: 182) that is,
settings that are especially amenable to analysis. I have examined exchanges where
a signification appears and then reappears at a more general or specific level. Such
settings made visible the act of generalization (or specification) and its conversational
functions. The following data, first presented by Sacks (1992, Vol. 1:
113–118; also 184–185 and 411–412) and elaborated by Edwards (1997: 96–98),
presents such a setting. The police have been called to intervene in a domestic
situation. They instruct one of the parties (‘B’ below) to confer with a social worker.
Sacks provides us with an excerpt from B’s discussion with the social worker (‘A’).
Here is the excerpt, in Sacks original format:
(Sacks 1992, vol. 1: 113)
(1) A: Yeah, then what happened?
(2) B: Okay, in the meantime she [wife of B] says, ‘‘Don’t ask the child nothing.’’
Well, she stepped between me and the child, and I got up to walk out the
door. When she stepped between me and the child, I went to move her out
of the way. And then about that time her sister had called the police. I don’t
know how she…what she…
(3) A: Didn’t you smack her one?
(4) B: No
(5) A: You’re not telling me the story, Mr B
(6) B: Well, you see when you say smack you mean hit
(7) A: Yeah, you shoved her. Is that it?
(8) B: Yeah, I shoved her
Sacks’ major point is that A, not being a party to the events in question, is
nevertheless able to see that B is not telling ‘‘the story,’’ and B can recognize A’s
grounds for doing so, because B has not provided an adequate warrant for the police
being called. I was particularly interested in line 6. It is phrased as an explanation,
but what does it explain? Of course, when A says ‘‘smack,’’ he means ‘‘hit’’. What
else would he mean? What is accomplished by saying this? We begin by noting that
‘‘hit’’ is more general than ‘‘smack’’. The relationship is illustrated in Fig. 1.
10
Fig. 1 ‘‘Hit’’ (The lines indicate ‘‘kind of’’ relationships; e.g., smack is a kind of hit.
Actually occurring items are in bold)
By ruling out ‘‘hit,’’ B is instructing A that (for example) ‘‘punch’’ would not be
an appropriate second guess. ‘‘Punch’’ is never mentioned but it is nonetheless
implicated. ‘‘Smack’’ is proposed by A as part of ‘‘the story’’ and, since ‘‘hit’’ is
suggested ascategorizing ‘‘smack,’’ it too is a possible version of the story. This gives us
Fig. 2 .
Fig. 2 ‘‘The story’’ (The arrow indicates a ‘‘part of’’ relation)
This is what I have called a ‘‘hybrid taxonomy,’’ in that it combines two types of
relations. “Hit” is part of ‘‘the story,’’ whereas “smack” is a kind of hit. The crucial
point for my purposes is that each ascending level is more general than the last.
Once we are thinking in multilevel terms, we are led to ask why B’s claim was not
generalized further. He might have said ‘‘I didn’t do anything to her’’ or ‘‘That is the
whole story’’. If there were no conceivable warrant for his sister-in-law’s call to the
police, one might expect a broader denial than the one he actually produces. We
evaluate what people say by reference to what they might, relevantly, have said. When
someone generalizes, we may ask, why did he generalize just so far and no further? The
very act of generalization invites such speculation. Thus, B’s turn 6 can be heard as an
invitation to A to look for some act other than hitting that would warrant the call to the
police. He doesn’t deny that there is a warrant for calling the police, just that smacking
or hitting is that warrant. A responds by guessing at ‘‘shove,’’ an action that has not
been ruled out and that is a subcategory of B’s initial description (‘‘move’’).
This analysis suggests that the degree of generalization may generate certain
implicatures that have consequences for the further course of the conversation. The
full analysis of the Sacks’ segment (not presented here) produced a complex
taxonomy. Some of the relations in that taxonomy were wholly indexical rather than
being based in a pre-existing linguistic semantics. (The relation of ‘‘smack’’ to ‘‘hit’’
11
is explainable on linguistic grounds, but ‘‘smack’’ as, potentially, part of ‘‘the story’’
is a product of the particular circumstances.) The taxonomy is a field of meaning,
co-constructed in and for the moment, with sequential implications.
Case 2: Generalization and Discursive Environment
In another article (Bilmes 2008 ), I dealt with constraints on the level of
generalization. The setting was a discussion of a passage in a draft memo in a
US regulatory agency. The passage offered a very specific reason for not enforcing a
certain regulatory provision. The provision required that loan companies inform
female applicants that they need not mention income from alimony or child support.
In recommending that the agency not enforce this provision, the memo authors
suggested that giving applicants the required notice might have the effect of
discouraging them from providing full information on their income, which might, in
turn, prevent them from getting the requested loan. Ben, one of the discussants,
reformulated this reason as ‘‘We’re not enforcing this portion of Reg B because we
think it’s bad public policy’’. This was, of course, a very generalized version of the
original reason. It would have been possible to generalize less radically; for
example, instead of ‘‘bad public policy,’’ he might have said ‘‘harmful to female
loan applicants’’. However, as I tried to show, a less radical generalization would
not have made sense in terms of the then-current discursive environment of the
agency or in terms of the argument that the speaker was constructing.
The politically appointed administrative officers of the agency were free market
advocates and were hostile to what they considered overregulation. Thus, their idea
of what was bad public policy was rather different from the generally more activist
legal staff. The staff countered the new agency policies by insisting that their job
was to enforce the law. So, by saying ‘‘We’re not enforcing this portion of Reg B
because we think it’s bad public policy,’’ Ben was pointing to an inconsistency in
the staff rhetoric. He was also suggesting that the argument legitimized administration
obstructionism in regard to enforcement of the law. This become apparent in
his follow-up statement: ‘‘If Ted [a member of the administration] didn’t enforce
everything he thought was bad public policy where would you be’’. Given these
considerations, a less extreme generalization would not have served Ben’s purposes,
that is, it would not have constituted a coherent argument against keeping the
targeted passage in the memo. The level of generalization was fitted to the
administration’s economic/legal arguments and to the staff’s counterargument.
Sometimes we need to consider not only possible alternatives, but also those
which are related but impossible. This is what I did with ‘bad public policy’. In
particular, it may be profitable to look at upgradings and downgradings along
various scales (intensity, generality, etc.)
Case 3: Generalization and Implied Categorization
A focus on generalization and specification, as it turns out, yields a variety of
insights into the structuring of meaning in conversation. The following extract is
from a mediation session at mediation agency known as the Neighborhood Justice
12
Center. L, the landlord, and T, the former tenant, have come for mediation of a
dispute. After T moved out of the house which he and a roommate had been renting
from L, L refused to return any of their deposit. The session began with L, T, and M,
a trained mediator, and myself as nonparticipating observer, sitting around a table.
L and T described the situation and their positions to M. Then M had L leave the
room, so that she could speak privately to T, after which she spoke privately to
L. They then reassembled as a group. M stated that they seemed to be at an impasse,
and L and T reiterated their positions. Then L asked T to have his former housemate
call her ‘‘so that I can explain’’ and T replied ‘‘Yah (whatever) I can explain it
(too)’’. M seizes on this as an agreement and suggested that they ‘‘write it up’’.
(NJC)
1. T: Aoh:: that's s'ch a um: a poor: um: y'know uh- 'at's
2.
not even worth the time t' write it up y'know I c'n
3.
gi[ve ya John's uh: (.5) John's (work) phone number is:=
[( )
4. ?:
5. T: =if you'd like [( )
[No- b't- b't Lloyd (how I feel)
6. M:
7.
about it is this=
8. T: =I do hav I do have 'n appointmenttuh: y'know I I
9.
feel th't uh .hh John Seiden's number at work is
10.
((telephone number)) an:d ah: (.) don' need any
11.
signatures on anything.
12. M: Wull (2.5) okay [is is
[Yeah (.) ('n) I have another meeting
13. T:
14.
too so I- I can' see ( [) (even spending) around uh=
[Okay
15. M:
16. T: =spending the time to write anything down (there).
17.
(1.3)
18. M: Okay so: that you'll give Linda the (ud number)
19.
c'd- could we- (.) have that number again: jus' so th'
20.
I c'n (.) communicate that to (
)?
21. T: ((repeats the telephone number)) (1)
22. M: four two nine six.
23. T: °umhm
24. M: .hh Well (.) (bLloyd) I'd like y' ta- (.8) would
25.
you: be willing t' take a minute (.) just to: (.) let
26.
us put that down in writing?
27. T: No I can't see any reason f'r that.=It's (j's a)
28.
phone number it's available y'know: uh: (.4) he works
29.
at Pearl Harbor 'nd uh (.7) so on 'n' so forth (.5)
30.
.hh .hh °y'know° I din' come in here 'n ho- hour 'nd
31.
a half 'r whatever j'st ta j'st t' give a phone
32.
number 'n': 'n'put something down 'n writing (.)
13
33.
34. M:
35. T:
36.
37. M:
38. M:
39.
40. T:
like that [(y'know)=
[Okay (.7) wull:
=(.) I come in here f'r: substantial uh (1.2)
[reasons
[Okay
'n' I c'n see how- (.) y'know y' could be
disappointed=
=resolution yeah
My interest, for present purposes, is in lines 28–41. T is rather specific in saying
what he did not come here for. He mentions two items (give a phone number and put
something like that in writing)without offering a categorization of those items. One
might expect the contrast to be at a similar level of specificity (I came to get my money
back). However, the contrast is actually made at a more general level (substantial
reasons). The direct contrast to substantial reasons is insubstantial (i.e., trivial) reasons.
He has thus, by implication, categorized his original two items. This way of doing it
makes a certain amount of sense, given that he begins his turn with ‘‘No I can’t see any
reason for that’’. But then he mentions giving a phone number (which M has brought up
earlier as an agreement and, therefore, something which should be recorded) and putting
something down in writing (which M later mentions as worthwhile in itself, since it
acknowledges them both for attending the mediation). Having first claimed that he could
see no reason to accede to M’s request, and then mentioned two possible reasons, T uses
the procedure I have described to more or less negate the two reasons by classifying them
as trivial. This analytic outcome (i.e., the implied categorization of the two items as
trivial) is, I think, fairly obvious. What is significant here is the mechanism by which the
outcome is produced. When item(s) A is contrasted with a more general item B, this
suggests a categorization of item A. The situation is illustrated in the following diagram
(Fig. 3 ).
trivial reasons (implied contrast) substantial reasons
(contrast)
give a phone number
put something
in writing
get my money back
Fig. 3 ‘‘Trivial reasons’’ (Actually occurring items are in bold)
Case 4: Tactical Generalization and Specification
I turn now to a short example of the rhetorical possibilities of generalization and
specification. Sara Taylor, a high ranking member of President George W. Bush’s
Justice Department, has been summoned to appear before a congressional committee.
14
Taylor testimony (originally transcribed by Matthew Prior)
1. S: Did you recei:ve (.) telephone calls or other forms
2.
of communication (1.6) in regards (.) to (.) the (.)
3.
US attorneys (.) that were fired.
4. T: I don’t recall. (.) I don’t recall getting
5.
communications about them. (3.0)
6. S: hhhh (2.0) hh (.) You don’t reca:ll if (.) someone
7.
called you to complain about a US attorney? (.07)
8. T: Uh (.) Senator (.) uh- (.) I’m sure you can
9.
appreciate (.) uh (.) that somebody (.) who (.) uh
10.
(0.9) i- u was in my position (.) who got (.) and I’m
11.
estimating here (.) who got roughly twenty phone calls
12.
a day (.) roughly three hundred emails (.) a day. Each
13.
and every day (.) uh (.) about a myriad of topics (.)
14.
uh (.) any and everything you could probably (.) uh
15.
(.) uh (0.9) would not recall (.) conversations (.) uh
16.
or uh (.) phone calls that came to her (.) I-I (.)
17.
senator I-I can’t remember what I had for breakfast
18.
last week=I just don’t recall any of those
19.
conver[sations]
20. S:
[I] I assume what you had for breakfast
21.
last week has not been the subject of considerable
22.
national attention
In lines 1–2, the senator asks about ‘‘telephone calls or other forms of
communication in regards to the US attorneys that were fired’’. When Taylor says that
she does not recall such communication, the senator becomes more specific (lines 6–7).
He narrows his question to calls and specifies the subject matter (complaints about state
attorneys). This seems to be based on an intuitive notion of what one would be likely to
remember. Taylor responds by dealing with the matter in more general terms. She gets a
large number of communications each day and so cannot reasonably be expected to
remember their content. Instead of mentioning complaints about state attorneys
specifically, she speaks only of a nondescript ‘‘myriad of topics’’. She thus generalizes
telephone calls to communications and omits the feature that might make certain calls
memorable. In structure and function, this is rather similar to Frake’s skin disease
example. The level of generalization is manipulated to fit the argument that they are
making and, more particularly, to bolster the credibility of the participants’ claims.
Case 5: Specification and Meaning
Another aspect of the relationship between generality and specificity is revealed in
15
the following extract, from Lawrence 1996 . The manager of a brothel in a midWestern town is being interviewed on the radio.
(Lawrence 1996)
1. IR: And you were providing a service
2. IE: .h Yes services were provided .h Mister Mellish I
3.
can tell you (.) I can remember a time when fathers .h
4.
brought their sons there .h to keep them from uh getting
5.
young girls pregnant (.) .hhh uh picking up u:h (.)
6.
something off of the street (0.6) picking up a venereal
7.
disease (0.8) .hhh u:h I can remember a motherh that had
8.
a somewhat retarded so:n (1.3) that waited across the
9.
tra:ck (.) wh- while he came down to my house .h my
10.
particular house .hhhhhhhhhh u:h I can remember picking
11.
up the social section of the newspaper and seeing young
12.
men that I knew as customers .hh that married nice
13.
gir:ls .hh uh but while they were courting .h um while
14.
they were (0.3) into u:h .hh the relationship .h they
15.
didn't want they didn't want to go out on these girls
16.
(1.1) they used the house and its facilities
17.
(1.3)
18. IR: .hh Rather than: (0.3) what is the saying violate
19.
the one that they love
20. IE: Yes or as opposed to going to a bar and maybe
21.
getting mixed up with a bar gir:l .h o:r u::h (0.8) .h
22.
(0.2) u:h there is a difference (.) between a house
23.
girl and a street girl an' anything .h we- u- let me
24.
define what street means that means any .h girl that
25.
workd on the street .h or in a tavern (0.2) or- (.)
26.
massage parlor .h where uh: (0.8) she doesn't have a
27.
place to live right along with it (.) those are street
28.
girls
My attention was drawn to the word ‘‘service’’. One might think that the claim
that a brothel provides a service requires no discussion, since prostitution is, more or
less by definition, a service. It turns out, though, that she is talking about service to
the community and not simply the sexual services that are provided to her
customers. One such service was to help prevent unwanted pregnancies and
venereal disease. Note that it is the father who brings his son—she is providing a
service to the family. The second story is similar in that it is the mother who brings
her retarded son for a life experience that he might otherwise have no access to. She
provides a service to courting couples, in the third example, by providing the male
an alternative to what she calls street girls. She goes on to point out the way in
which house girls are superior to street girls. Thus, each of these narratives
represents a kind of service (Fig. 4 ).
16
Fig. 4 Service
Note that this taxonomy combines a single word (‘‘service’’) with more complex
discursive units. That is, the taxonomy includes both an expression and several
narratives, thus crossing the line between occasioned semantics and formulation in
the wider sense. The narratives help to define what IE means as ‘‘service,’’ which
might otherwise be understood to refer to a purely physical service—giving the
customer what he wants. The work of defining is frequently done, as it is here,
through specification. At the same time, IE reflexively classifies these narratives as
service, thus instructing us on how they are to be understood. Note, for instance, that
the first two narratives could be offered, in other circumstances, as examples of
child abuse, or at least poor parenting.
Case 6: Specification and Sequence
This example is taken from a phone conversation between Linda Tripp and
Monica Lewinsky. They are discussing the fact that Lewinsky intends to lie in
court about her relationship to the President. Tripp says that, if she is called to
testify, she will feel compelled to tell the truth, which will contradict Lewinsky’s
account.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Tripp 6-1.1, 4+ minutes into the recording. T=Tripp, L=Lewinsky
T: (*) Look Monica: .hh We already know: that you're
gonna lie under oath. .hh We also know that I want out
of this big time. (.8) I mean (.) awh: how else is this
gonna go.=If I have to testify, (.6) if I am forced to
answer questions and I answer truthfully, it's gonna be
the opposite of what you say so therefore it's a
conflict right there. (.8)
L((plaintively)): But it doesn't have t'be a conflict.=
T: =Whataya mea:n. (1) Ho:w. Tell me ho:w. (2) Tell me
how. .hh I'm supposed to say=if they sa:y has Monica
Lewinsky ever said to you (.6) tha::t she:: (.3) is in
love with the president or is having a physical
relationship with the president=if I say no that is (*)
perjury .hh (.) that's the bottom line.
L: (****)
T: I will do everything I can: (.) not to be in that
position. (.6) That's what I'm trying to do. But I
don't know how else to do this. .hh I'm coming up with
foot surgery in California.=I will do anything=
17
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
L: =I ^kno:wˇ (.5)
T: But no you really don't know cause you don't believe
me. (.5) I think you really belie:ve (.) [that this is=
L:
[(*)
T: =very ea:sy –an' I should just say ((one or two
syllables deleted)) if they can't prove it
L: I believe you but obviously: -I don't have the same
feelings about the situation that you do so I [(*)
T:
[What do
you mean.
L: .hh Because i- if I had the same feelings that it was
so .hh (.5) wron:g (.5) t'do: (.) t t t deny- to deny
something (.8) then: then I would not be doing it. Ya
see what I mean?
T: .hh I think down deep ya don't like t having to lie.
(.5)
L: ^Of course I don't think anybody ^likes to:.^ (.5) I
don't think anybody likes to –but if they [(**)=
T:
[but the
scary thing to me is
L: =this is how family is (.) This is- I would- lie on the
stand for my family. That is how I was raised as
family. (.5)
T: I don't think I- ya know what Monica you're gonna die
here an' I would do almost anything for my kids but I
don't think I would lie on the stand for them.
L: Wull- I mean I would just that's just- for me: ya know
that's just how: (1.5)
T: I mean: [(.) I have (this)
L:
[(an') I'm sure you were raised too in a very:
(.) sort of honorable: (.5) [(**)
T:
[I would have b- I would
have probably been: (.5[) tarred and feathered (.) if=
L:
[(*)
T: =I even considered saying something like that in my
house.=
L: =Right (.) [(*) an' I'm- an' I was brought up with lies.=
T:
[(look)
L: =(.) all (.) the time. (.5) So tha- that was how: (.)
that was how: you got along -in life -was by lying.
T: I don't believe that. Is that true?
L: ^Yes it's true. Y I wanted something from my dad, well
first my parents were divorced.=If I wanted money from
my dad, I had to make up a story. .hh When my parents
were married .hh my mom would –always (.) lying to my
dad (.) for everything. (.5) Everything. My mom helped
me sneak out of the house (1) I mean that's just how:
(.) that's just how I was raised. (2)
18
68
69
70
71
72
73
74
75
76
77
T: Well (.) [In in the Catholic religion there are white=
L:
[(*)
T: =lie:s and there are black lies.=Those are white lies.
Those ar- (.) those are like white lies.=Those are kind
of like lies wL: You look great.
T: When you really look (.) t.hhh That's .hh that's not
what I'm talking about here.
L: I mean I have outright (.) I lie: to my dad. I have
lied my entire life
19
When T asserts that she would not lie on the stand even for her own children
(lines 44–45), L begins to make a direct contrast between herself and T, but aborts this in
favor of a comment on T’s upbringing. This sets up a contrast with her own upbringing.
Her argument, as it develops, is not that she is lying for her family, but rather that she
does not mind lying because of her family. (It appears that her ‘‘family,’’ in this case, is
her mother. Her father is the one who was lied to.) T agrees that she was raised in an
‘‘honorable’’ way, thus providing the first item in the contrast.9 L continues:
L: =Right (.) [(*) an' I'm- an' I was brought up with lies.=
T:
[(look)
L: =(.) all (.) the time. (.5) So tha- that was how: (.)
that was how: you got along -in life -was by lying.
When T expresses doubt (line 61), L offers specifics:
L: ^Yes it's true. Y I wanted something from my dad, well
first my parents were divorced.=If I wanted money from
my dad, I had to make up a story. .hh When my parents
were married .hh my mom would –always (.) lying to my
dad (.) for everything. (.5) Everything. My mom helped
me sneak out of the house (1) I mean that's just how:
(.) that's just how I was raised. (2)
This instance of specification is somewhat different from that offered in case 5. In
the radio interview the categorization and the events categorized are discrete items.
That is, the narratives exist independently of the way they are categorized. On the
other hand, in case (6), the category ‘‘lie’’ more or less constitutes the examples
rather than telling us how we are to interpret them. The events, as formulated, are
not only categorized by the concept of lying; also they consist of lying.
We have seen in the radio interview example one function of offering specifics. It
helps to define what one is talking about. Also, it seems that somehow specifics
stand as a form of evidence, even when they are no more proven than the original
assertion. Therefore, expressions of doubt regularly elicit specifics. Perhaps the
specifics show how the general statement could be true, thus making it more
credible. At any rate, it is T who seizes on the defining function of L’s specifics. She
no longer resists L’s assertion that she was brought up with lies, but she interprets
the statement in a way that suits her.
The conversation to this point might be (partially) conceptualized in the
following diagram (Fig. 5 ).
9 This is a common way to do certain kinds of tactical contrast—get the other to provide or accede to the
first term. It is also a way to introduce an item into a conversation. In this case, show a proper (and, at
least in this case, sequentially motivated) interest in the other’s upbringing, then reciprocate with
information about one’s own upbringing.
20
Fig. 5 Lies
T uses L’s specifications to introduce a new categorical distinction: white lies vs.
black lies. She associates this with the Catholic religion, perhaps to explain why she
is making a distinction which seems to have escaped L. L may have grown up lying
but her lies were white lies and therefore insignificant. That is not, as T says, what is
being discussed here (i.e., lying under oath). Therefore, L’s upbringing, although it
may be as she describes it, has no application to the present situation. The diagram,
incorporating L’s proposal now looks like this (Fig. 6 ).
her father
Fig. 6 White and black
In line 73, L offers an understanding of ‘‘white lie,’’ which T validates in line 74.
But in lines 76–77, L asserts that she has ‘‘lied my entire life’’. She appears to be
ignoring the distinction that T is making, or at least rejecting the significance that T
assigns to it. We have here a clear example of how categorical distinctions are
sequentially motivated. L’s offering of details follows T’s expression of doubt. And T’s
introduction of the distinction between white and black lies utilizes materials
supplied by L’s attempt at specification.
Conclusions
In the above examples, I have illustrated various manifestations, implications, and
functions of a general feature of expressions in conversation—generalization/
specification. Some findings:
1. The level of generalization may limit what can be said next and may supply a
21
basis for advancing the conversation in particular directions by providing for
certain implicatures or inferences.
2. The direction and level of generalization may be, at least in part, determined by
such matters as the kind of argument being constructed and the discursive
environment.
3. The use of a generalized contrast may provide for the categorization of some
original item or set of items.
4. Generalization and specification may be used strategically to achieve rhetorical
effects; to persuade, justify, excuse, and so forth.
5. The relation between a general item and its specifications may reflexively
reveal the nature of the items concerned. That is, the meaning of the general
expression may be defined by its specifications at the same time that the
meaning of the specific items are defined by the general expression.
6. Categorizations and specifications are sequentially motivated.
7. Specification provides a kind of quasi-evidence.10
If nothing else, I hope that I have demonstrated the value of studying this
conversational phenomenon. I have chosen diverse examples from a variety of
settings in an attempt to demonstrate the richness of the subject. However, I have
presented my analyses as exemplifications of one possible aspect of an occasioned
semantics and, ultimately, an approach to formulation. Other features of conversational
meaning, such as contrast (Deppermann 2005 ) and scaling, presumably
have a similar diversity of function.
Whether a fully formed, systematic approach to formulation will take the form I
have envisioned and partially exemplified remains to be seen. But I am convinced
that the current analytical boundaries of conversation analysis need to be broadened
to include matters which traditionally concerned linguistic pragmatists, semanticists,
linguistic philosophers, and other students of meaning and rhetoric, even
including scholars practicing so-called critical discourse analysis and so-called
postmodern approaches. In recent decades, there has been a turn, led by Wittgenstein
(1958 ), Austin (1962 ), Bakhtin (1986 ), Garfinkel 1967 ), Sacks(1992 ), and others,
from language as saying something to language as doing something and from
semantics to indexicality. It is time to unite the old concerns with the new, as Sacks
began to do in his discussions of categories. We cannot go back to abstract
discussions of meaning in language—when we speak or write, we are doing
something, and we are doing it in a context that partially determines what we are
taken to mean. But we cannot lose sight of the fact that the doing is accomplished by
means of saying, or, to put it in Austinian terms, illocution is accomplished through
10 I have argued in another publication (Bilmes 2009b) that at least some metaphors may also be
considered generalizations of a sort, but I have chosen not to consider metaphor in this article, since it
raises issues that are not especially relevant to the current discussion.
22
locution. We need to concern ourselves, once again, with the ‘‘what’’ as well as the
‘‘how,’’ with content as well as operations.
We have, on the one hand, the technical, methodologically-informed studies of
conversational interaction; on the other, we have the concern with the topics and
meanings expressed through language, approached largely through ‘‘interpretation,’’
akin to literary criticism. In the former case, the focus has been primarily on
conversation and related forms of verbal interaction; the latter tends to focus
primarily on texts and monologs (although Bakhtin has shown us that these too are
interactive). In terms of this distinction, linguistic pragmatics may be said to occupy
a middle ground—it deals with the meanings of utterances and textual passages in a
technical way. In fact, some pragmatists (e.g., Labov and Fanshel 1977 ) even work
with extended stretches of actual conversation. But their analyses and results—
which frequently draw on unexamined ‘‘mentalistic’’ resources, such as intention
and motivation—tend to be dominated by their theoretical orientations, which limits
their approach to meaning, interaction, and conversational (or textual) structure and
constrains their sensitivity to indexical matters in favor of general rules (Streeck
1980 ; also Bilmes 1988 ; Schegloff 1988b ). I am suggesting, therefore, that we find a
new way of bridging the gap between action and meaning, between doing and
saying, an approach that maintains something of the CA sensibility exemplified in
Sacks’ lectures. I propose that this approach consist of a methodologically-informed
study of formulation, rhetoric, and meaning structure. I view this as an attempt to
enrich CA, to broaden its reach.
SA is, and always has been about more than the ‘‘clacking of turns’’ [in
Moerman’s (1988 : xi) memorable phrase]. Nevertheless, SA has been primarily
concerned with the organization of conversational interaction, with the organization
of turns and actions. I am proposing a parallel concern with the organization of
meaning in conversation.
Appendix
Three of Sacks’ major points on ‘‘The baby cried. The mommy picked it up’’.
are: (a) Sacks asks how we know that the mommy is the mommy of the baby. ‘‘Baby,’’
he claims, belongs to two ‘‘devices’’—a device being a collection (i.e., a set of
categories that go together, that is, a set of categories which are themselves
members of a more general category, as ‘‘male’’ and ‘‘female’’ are members of
the category ‘‘sex’’) and its rules of application—a stage-of-life device and a
family device. ‘‘Baby’’ is a member of the category ‘‘age group’’ and also a
member of the category ‘‘family member’’. The ambiguity is resolved by the
use of ‘‘mommy’’. Sacks proposes three rules to explain this. According to the
‘‘economy rule,’’ ‘‘a single category from a membership categorization device
can be referentially adequate’’. So, for instance, the category ‘‘baby’’ is
sufficient reference for the purposes of this story. The consistency rule states
‘‘If some population of persons is being categorized, and if a category from
some device’s collection has been used to categorize a first member of the
23
population, then that category or other categories of the same collection may
be used to categorize further members of the population’’. So, if ‘‘oak’’ is
mentioned, there is a distinct possibility that ‘‘pine’’ will be mentioned also.
There is also a ‘‘corollary’’ of the consistency rule which holds that ‘‘if two or
more categories are used to categorize two or more members of some
population, and these categories can be heard as categories from the same
collection, then: Hear them that way’’. So, given that both ‘‘mommy’’ and
‘‘baby’’ can be heard as members of the family category, that is how we will
hear them. Moreover, since families, like teams, are ‘‘duplicatively
organized,’’ we will hear them as belonging to the same family.
I think that Sacks’ argument on this point is flawed. To begin with, ‘‘baby,’’
unlike ‘‘child,’’ is not ambiguous as between age and family membership.
‘‘Baby,’’ in its literal usage, refers to an age-group. For this very reason, one
cannot literally refer to an adult as a baby, although an adult can, literally, be
someone’s child. (The online Miriam-Webster dictionary supports me on this.)
Many languages have two words that would be translated as ‘‘child,’’ one for
child as offspring and one that refers to an individual as a member of an age
group. Furthermore, Sacks ignores a crucial aspect of the word’s context—its
situation within a nominal phrase. Since, I have argued, ‘‘baby’’ is not
ambiguous to begin with, let’s use the semantically ambiguous ‘‘child’’ as our
example. The phrase ‘‘the child’’ signals, in most contexts, that we are
referring to age, whereas ‘‘my child’’ refers to family membership. So, even if
‘‘baby’’ were ambiguous, the phrase ‘‘the baby’’ would make it clear that the
reference is to age. (Of course, the phrase ‘‘my baby’’ does invoke family membership,
but so does ‘‘my teen-ager’’ or ‘‘my little fireman’’. It is the possessive pronoun that
recruits teen-agers and firemen—and babies—into the family.) Why, then, do we hear
‘‘the mommy’’ as the parent of ‘‘the baby’’ in this story? Given that the adult in question
could be formulated as, e.g., woman or lady, we are led to ask why this more marked
formulation. ‘‘Mommy’’ is unambiguously a family member. We assume, therefore, that
she is formulated in this way because she is the mother of the baby who cried. (Consider:
‘‘The teen-ager was rude. The mother scolded him.’’) But, again, the nominal phrase
is crucial. As Arnulf Deppermann has pointed out to me, if it were ‘‘a mommy’’ who
picked the baby up, we would probably understand that it was not the baby’s mother. We
require a linguistic, as well as categorial, analysis.
One further thought: If the story was ‘‘The baby cried. The nanny (or the
babysitter) picked it up,’’ would we not be likely to hear ‘‘the nanny’’ as the
nanny of the baby? How does Sacks’ analysis apply to this case? I think, as in
the case of ‘‘mommy,’’ that the interpretation is guided by the categorization of
the adult. Why formulate the woman as a nanny, unless she is the nanny of the
baby?
Despite these flaws in Sacks’ analysis, his economy and consistency rules are
powerful. I am interested in particular in the corollary to the consistency rule.
Take, for instance, the statement ‘‘He hates rats and squirrels.’’ These animals
24
can be heard as belonging to the same collection—rodent or small mammal or
some such—but I don’t think that we necessarily hear them in that way. It may
be that he doesn’t like rats and he doesn’t like squirrels, period. In contrast,
‘‘He hates rats and squirrels and so forth’’ calls for the inference of a
superordinate category, with ‘‘rodents’’ as one possibility. Compare ‘‘He hates
rats and cockroaches and so forth,’’ which might lead us to infer ‘‘vermin’’ as
the collection.
(b) The second major point of Sacks’ analysis is that membership categories are
associated with particular activities. He calls these ‘‘category-bound’’ activities.
[Watson (1978 ) has proposed the phrase ‘‘category-bound predicate,’’ to
accommodate the fact that categories may have associated attributes other than
activities.] Thus, the category ‘‘baby’’ is associated with crying. (Note that this
lends further weight to my contention that ‘‘baby’’ in this story refers
unambiguously to the age-group category.) When a category is invoked, it
brings with it a set of possibly relevant predicates. And mention of an activity
or attribute, in turn, may invoke the associated category. The concept of
category-bound activity is not exactly original—one thinks, for example, of
role theory—but Sacks’ use of it to analyze conversation was an innovation,
and subsequent studies in MCA rely heavily on this concept.
(c) Sacks points out that the baby’s crying and the mommy’s picking it up are
ordered acts. First the baby cried and then the mother picked it up. He proposes
that this perception of sequence is produced not so much by the ordering of the
sentences (cf., Labov and Waletzky 1967 ) as by our notion that a mother
picking up a baby is commonly occasioned by the baby crying. (See,
especially, Sacks 1992 : 244–245.) That is, we see not merely a sequence of behaviors,
but events linked by causation. We suppose that the mother’s action
was consequent on the baby’s. I think Sacks is not entirely correct on this
point. How would we hear ‘‘The mommy picked the baby up. It cried.’’?
It would seem to me that the ordering of the sentences is, in this case,
determinative. I think this story would normally be heard as suggesting that the
mother, by picking the baby up, caused it to cry. Perhaps the more crucial point
that Sacks is making, though, is that, when we hear of some sequentially
ordered sequence of events, we infer, when necessary, some mechanism—
causation, purpose, etc.—that connects the events and accounts for their
ordering.
One can go to other sources (e.g., Lepper 2000 ; Schegloff 2007a ; Silverman
1998 ) for more complete (and less critical) expositions of MCA. Some, most
notably Jayusi (1984 ), have tried to elaborate Sacks’ notions regarding categories.
Others (e.g., Watson 1997 ) have argued for a more complete integration of MCA
with SA.
References
Antaki, C. (2008). Formulations in psychotherapy. In A. Perakyla, C. Antaki, S.
25
Vehvilainen, & I. Leudar (Eds.), Conversation analysis and psychotherapy (pp. 26–42).
Cambridge, England: Cambridge University Press.
Arminen, I. (2005). Institutional interaction: Studies of talk at work. Aldershot, England:
Ashgate Publishing.
Austin, J. L. (1962). How to do things with words. Oxford: Clarendon.
Bakhtin, M. (1986). Speech genres and other late essays. Austin: University of Texas
Press.
Barnes, R. (2007). Formulations and the facilitation of common agreement in meetings
talk. Text & Talk, 27(3), 273–296.
Billig, M. (1996). Arguing and thinking: A rhetorical approach to social psychology (2nd
ed.). Cambridge, England: Cambridge University Press.
Bilmes, J. (1988). Category and rule in conversation analysis. Papers in Pragmatics, 2(1),
25–59.
Bilmes, J. (1993). Ethnomethodology, culture, and implicature: Toward an empirical
pragmatics. Pragmatics, 3(4), 387–409.
Bilmes, J. (1995). Negotiation and compromise: A microanalysis of a discussion in the
Federal Trade Commission. In A. Firth (Ed.), The discourse of negotiation: Studies of
language in the workplace (pp. 61–81). Oxford: Pergamon Press.
Bilmes, J. (2008). Generally speaking: Formulating an argument in the US federal trade
commission. Text & Talk, 28(2), 193–217.
Bilmes, J. (2009a). Taxonomies are for talking: A reanalysis of a Sacks classic. Journal of
Pragmatics, 41(8), 1600–1610.
Bilmes, J. (2009b). Kinship categories in a Northern Thai narrative. In H. Nguyen & G.
Kasper (Eds.), Talk-in-interaction: Multilingual perspectives (pp. 29–56). Honolulu:
National Foreign Language Resource Center.
Bilmes, J. (2010). Scaling as an aspect of formulation in verbal interaction. In Language
learning and socialization through conversations (pp. 3–9). Conference proceedings.
Center for Human Activity Theory, Kansai University, Osaka, Japan.
Conklin, H. (1969). Lexicographical treatment of folk taxonomies. In S. Tyler (Ed.),
Cognitive anthropology (pp. 41–59). New York: Holt, Rinehart and Winston.
26
Deppermann, A. (2005). Conversational interpretation of lexical items and conversational
contrasting. In A. Hakulinen & M. Selting (Eds.), Syntax and lexis in conversation:
Studies on the use of linguisticresources in talk-in-interaction (pp. 289–306). Amsterdam:
John Benjamins.
Deppermann, A. (2011). Notionalization: The transformation of descriptions into
categorizations. Human Studies, 34(2). doi:10.1007/s10746-011-9186-9.
Drew, P. (2003). Comparative analysis of talk-in-interaction in different institutional
settings: A sketch. In P. Glenn, C. LeBaron, & J. Mandelbaum (Eds.), Studies in
language and social interaction: In honor of Robert Hopper (pp. 293–308). Mahwah, NJ:
Lawrence Erlbaum Associates.
Drew, P., & Heritage, J. (1992). Analyzing talk at work: An introduction. In P. Drew & J.
Heritage (Eds.), Talk at work: Interaction in institutional settings (pp. 3–65). Cambridge,
England: Cambridge University Press.
Edwards, D. (1994). Script formulations: An analysis of event descriptions in
conversation. Journal of Language and Social Psychology, 13(3), 211–247.
Edwards, D. (1995). Two to tango: Script formulations, dispositions, and rhetorical
symmetry in relationship troubles talk. Research on Language and Social Interaction,
28(4), 319–350.
Edwards, D. (1997). Discourse and cognition. London: Sage.
Frake, C. (1961). The diagnosis of disease among the Subanun of Mindanao. American
Anthropologist, 63(1), 113–132.
Frake, C. (1969a). The ethnographic study of cognitive systems. In S. Tyler (Ed.),
Cognitive anthropology (pp. 28–41). New York: Holt, Rinehart and Winston.
Frake, C. (1969b). A structural description of Subanun ‘‘religious behavior’’. In S. Tyler
(Ed.), Cognitive anthropology (pp. 470–487). New York: Holt, Rinehart and Winston.
Gafaranga, J., & Britten, N. (2004). Formulation in general practice consultations. Text,
24(2), 147–170.
Garfinkel, H. (1967). Studies in ethnomethodology. Englewood Cliffs, New Jersey:
Prentice Hall.
Garfinkel, H. (2002). Ethnomethodology’s program: Working out Durkheim’s aphorism.
(edited and introduced by A. Rawls). Lanham, Maryland: Rowman & Littlefield.
27
Garfinkel, H., & Sacks, H. (1970). On formal structures of practical actions. In J.
McKinney & E. Tiryakian (Eds.), Theoretical sociology: Perspectives and developments
(pp. 337–356). New York: Appleton-Century-Croft.
Geoghegan, W. (1973). Natural information processing rules: Formal theory and
applications to ethnography. Monographs of the Language-Behavior Research
Laboratory, University of California, Berkeley.
Goodwin, C. (1994). Professional vision. American Anthropologist, 96(3), 606–633.
Grice, H. P. (1975). Logic and conversation. In P. Cole & J. Morgan (Eds.), Syntax and
semantics 3: Speech acts (pp. 41–58). New York: Academic Press.
Hauser, E. (2011) Generalization: A practice of situated categorization in talk. Human
Studies, 34(2).
Heritage, J., & Raymond, G. (2005). The terms of agreement: Indexing epistemic
authority and subordination in talk-in-interaction. Social Psychology Quarterly, 68(1),
15–38.
Heritage, J., & Watson, D. R. (1979). Formulations as conversational objects. In G.
Psathas (Ed.), Everyday language: Studies in ethnomethodology (pp. 123–162). New
York: Irvington.
Hester, S., & Eglin, P. (1997). Culture in action: Studies in membership categorization
analysis. Washington, DC: University Press of America.
Horn, L. (1972). On the semantic properties of logical operators in English. PhD thesis,
University of California, Los Angeles.
Housley, W., & Fitzgerald, R. (2009). Membership categorization, culture and norms in
action. Discourse and Society, 20(3), 345–362.
Jayusi, L. (1984). Categorization and the moral order. London: Routledge and Kegan
Paul.
Labov, W., & Fanshel, D. (1977). Therapeutic discourse: Psychotherapy as conversation.
New York: Academic Press.
Labov, W., & Waletzky, J. (1967). Narrative analysis. In J. Helm (Ed.), Essays on the
verbal and visualarts (pp. 12–44). Seattle: University of Washington Press.
Lawrence, S. (1996). Normalizing stigmatized practices: Achieving co-membership by
28
‘‘doing being ordinary’’. Research on Language and Social Interaction, 29(3), 181–218.
Lepper, G. (2000). Categories in text and talk: A practical introduction to categorization
analysis. London: Sage.
Levinson, S. (1983). Pragmatics. Cambridge: Cambridge University Press.
Moerman, M. (1988). Talking culture: Ethnography and conversation analysis.
Philadelphia: University of Pennsylvania Press.
Moerman, M., & Sacks, H. (1988). On understanding in the analysis of natural
conversation. In M. Moerman (Ed.), Talking culture: Ethnography and conversation
analysis (pp. 180–186). Philadelphia: University of Pennsylvania Press.
Popper, K. (1959). The logic of scientific discovery. London: Hutchinson. (First
published in German in 1935).
Potter, J. (1996). Representing reality: Discourse. rhetoric, and social construction.
London: Sage.
Raymond, G., & Heritage, J. (2006). The epistemics of social relations: Owning
grandchildren. Language in Society, 35(5), 677–705.
Sacks, H. (1972). On the analyzability of stories by children. In J. Gumperz & D. Hymes
(Eds.), Directions in sociolinguistics: The ethnography of communication (pp. 329–345).
New York: Holt, Rinehart and Winston.
Sacks, H. (1992). Lectures on conversation. Edited by G. Jefferson. Oxford: Blackwell.
Schegloff, E. (1988a). Description in the social sciences I: Talk-in-interaction. Papers in
Pragmatics, 2(1/2), 1–24.
Schegloff, E. (1988b). Presequences and indirection—applying speech act theory to
ordinary conversation. Journal of Pragmatics, 12(1), 55–62.
Schegloff, E. (1991). Reflections on talk and social structure. In D. Boden & D.
Zimmerman (Eds.), Talk and social structure (pp. 44–70). Cambridge, England: Polity
Press.
Schegloff, E. (1993). Reflections on quantification in the study of conversation. Research
on Language and Social Interaction, 26(1), 99–128.
Schegloff, E. (2007a). A tutorial on membership categorization. Journal of Pragmatics,
29
39(3), 462–482.
Schegloff, E. (2007b). Categories in action: Person-reference and membership
categorization. Discourse
Studies, 9(4), 433–461.
Schegloff, E. (2007c). Sequence organization in interaction: A primer in conversation
analysis 1.
Cambridge, England: Cambridge University Press.
Silverman, D. (1998). Harvey sacks: Social science and conversation analysis. New
York: Oxford
University Press.
Stivers, T. (2007). Alternative recognitionals in person reference. In N. Enfield & T.
Stivers (Eds.),
Person reference in interaction: Linguistic, cultural and social perspectives (pp. 73–96).
Cambridge, England: Cambridge University Press.
Streeck, J. (1980). Speech acts in interaction: A critique of Searle. Discourse Processes,
3(2), 133–154.
Turner, R. (1971). Words, utterances, and activities. In J. Douglas (Ed.), Understanding
everyday life:
Toward the reconstruction of sociological knowledge (pp. 169–187). Chicago: Aldine.
Walker, E. (1995). Making a bid for change: formulations in union/management
negotiations. In A. Firth
(Ed.), The discourse of negotiation: Studies of language in the workplace (pp. 101–140).
Oxford:
Pergamon Press.
Watson, D. R. (1978). Categorization, authorization and blame: Negotiation in
conversation. Sociology,
12(1), 105–113.
Watson, R. (1997). Some general reflections on ‘categorization’ and ‘sequence’ in the
analysis of
conversation. In S. Hester & P. Eglin (Eds.), Culture in action: Studies in membership
categorization analysis (pp. 49–75). Washington, DC: University Press of America.
Wittgenstein, L. (1958). Philosophical investigations (2nd ed.). Oxford: Blackwell.