Professional Documents
Culture Documents
Context Counts
ii
iii
Context Counts
Papers on Language, Gender, and Power
Robin Tolmach Lakoff
1
iv
1
Oxford University Press is a department of the University of Oxford. It furthers
the University’s objective of excellence in research, scholarship, and education
by publishing worldwide. Oxford is a registered trade mark of Oxford University
Press in the UK and certain other countries.
9 8 7 6 5 4 3 2 1
Paperback Printed by WebCom, Inc., Canada
Hardback printed by Bridgeport National Bindery, Inc., United States of America
v
CONTENTS
Acknowledgments ix
Contributors xi
Introduction 1
By Laurel A. Sutton
1. Language in context (1972) 7
Introduction by Sally McConnell-Ginet
2. The logic of politeness; or, Minding your P’s and Q’s (1973) 37
Introduction by Sachiko Ide
3. Excerpts from two 1974 papers: Pluralism in linguistics;
Linguistic theory and the real world 57
Introduction by Birch Moonwomon
4. You say what you are: Acceptability and gender-related
language (1977) 85
Introduction by Mary Bucholtz
5. Stylistic strategies within a grammar of style (1979) 101
Introduction by Deborah Tannen
6. When talk is not cheap: Psychotherapy as conversation (1979) 137
Introduction by Joan Swann
7. Some of my favorite writers are literate: The mingling of oral
and literate strategies in written communication (1982) 151
Introduction by Jenny Cook-Gumperz
8. Persuasive discourse and ordinary conversation, with examples
from advertising (1982) 183
Introduction by Janet S. Shibamoto-Smith
9. Doubletalk: Sexism in tech talk (1983) 209
Introduction by Susan M. Ervin-Tripp
10. My life in court (1986) 225
Introduction by Susan Blackwell
viii
11. The way we were; or, The real actual truth about generative
semantics: A memoir (1989) 241
Introduction by Georgia Green
12. Review essay: Women and disability (1989) 299
Introduction by Suzette Haden Elgin
13. Pragmatics and the law: Speech act theory confronts
the First Amendment (1992) 315
Introduction by Susan C. Herring
14. The rhetoric of reproduction (1992) 335
Introduction by Laurel A. Sutton
15. True confessions? Pragmatic competence and criminal
confession (1996) 353
Introduction by Linda Coleman
16. Afterword 371
by Robin Lakoff
Index 385
[ viii ] Contents
ix
ACKNOWLED GM EN TS
This book would not have been possible without the expert assistance of
Julia Bernd, who provided eagle-eyed editing, insightful comments, and
tireless research, all of which helped shape this book into a volume worthy
of its author.
Thanks also to Jocelyn Ahlers for creating Chapter 3 through her wise
choices, and for giving me the final push to finish this volume.
This book’s contributors have shown more patience than I thought pos-
sible; I offer thanks and apologies in equal measure. I will always mourn the
fact that the amazing Suzette Haden Elgin passed away before she could
see this book in print.
And finally, thanks to Robin Lakoff for being Robin: she was, and is, the
feminist linguist we need—and sometimes even deserve.
x
xi
CONTRIB U TORS
[ xii ] Contributors
1
Introduction
BY L AUREL A . SUT TON
her work is Lakoff’s understanding of the many ways in which power and
social relations are expressed in everyday utterances; and it is this under-
standing, always expressed so clearly, that keeps these essays fresh and rel-
evant. Naturally, the examples and “current” linguistic theory discussed in
the oldest papers included here (1970s) are dated, but even these provide
an insider’s view into the field and the feeling of the times.
All of the contributors—Lakoff’s peers, noted linguists in their own
right—gave generously of their time and energy to provide introductions
to each chapter, framing Lakoff’s work in a historical and personal context.
The introductions span a broad range of perspectives, from mini-research
papers to deeply personal anecdotes. They are truly the icing on the cake.
Those who know Lakoff only as a feminist may be pleasantly surprised
by the diversity of subjects covered in this volume; those who know her
only as the author of Language and woman’s place will now have an account
of her linguistic research and writing from that time until the late 1990s.
And for those young scholars just beginning to think about language,
I hope this book provides a rich resource of intelligent commentary and
analysis to which they will return again and again.
REFERENCES
[ 2 ] Context Counts
3
“generative semantics” camp (with which Lakoff was identified) held that
semantic phenomena “drive” the grammar. That is, semantic representa-
tions are in some sense basic (and, for most of these and other linguists,
universal), and it is language-particular grammars that yield actual surface
forms “from” semantic input. “Interpretive semantics,” on the other hand,
took syntactic representations of some kind as input to the semantics, with
semantics feeding on syntax rather than vice versa. But perhaps the critical
difference was that the generative semanticists took semantic representa-
tions to be fundamentally syntactic, governed by exactly the same prin-
ciples that operate generally in syntax; interpretive semantics left open
the possibility that semantic representations might be quite different from
syntactic.
Now the question of how best to describe and analyze the dependence
of interpretation on context—a major issue that “Language in context”
raises—was not really being systematically addressed by either semantics
camp. Even less attention was being paid to the social ramifications of lan-
guage use. Although a few linguists had begun to explore proposals from
the philosophy of language about the heavily contextual character of lin-
guistic communication (especially those inspired by Grice 1968), the field of
linguistic pragmatics did not yet exist. Sociolinguistics was also in its very
early stages, and virtually all the work available (e.g., Labov 1972) empha-
sized socially conditioned variation of the sort that distinguishes dialects
and plays a role in language change. Ideas about what is now sometimes
called “communicative competence” were still in their infancy; “interac-
tional sociolinguistics” and “discourse analysis” were not yet part of lin-
guists’ vocabularies. Lakoff’s article, published in the prestigious and
widely read journal Language, was an important spur to subsequent work
on questions of the importance of context to the understanding of natural
language utterances and their social effects.
In rereading “Language in context,” I was surprised to find that it did
not actually address in much detail the question of how investigations of
contextual matters were to be integrated with the rest of linguistic inquiry.
As I remembered, there is the claim that “traditional transformational
grammar” (which was accompanied by no semantics, much less any prag-
matics) fails when confronted with language in context. The idea is that
the “applicability” of grammatical rules must be conditioned by contextual
phenomena because a rule is considered “inapplicable” in contexts where
its application would produce any kind of oddness at all. Contextual mat-
ters as well as other semantic phenomena are seen as essential inputs to
the grammar, the task of which Lakoff assumes (without argument) is to
predict the acceptability or appropriateness (and perhaps even the social
[ 4 ] Context Counts
5
L A N G UAG E I N C O N T E X T [ 5 ]
6
REFERENCES
Grice, H. P. 1968. The William James Lectures. Published with other material
in H.P. Grice (1989), Studies in the way of words. Cambridge, MA: Harvard
University Press.
Labov, William. 1972. Sociolinguistic patterns. Philadelphia: University of
Pennsylvania Press.
Lakoff, Robin. 1973. Language and woman’s place. Language in Society 2.45–79.
[ 6 ] Context Counts
7
CHAPTER 1
Language in context (1972)
This paper originally appeared in Language 48:907–27. Reprinted here with permis-
sion of the publisher.
I should like to thank the following people, who have served as informants or made
valuable suggestions regarding the Japanese data: Chisato Kitagawa, Tazuko Uyeno,
and Kazuhiko Yoshida. I should also like to thank George Lakoff for much helpful dis-
cussion. All errors and misinterpretations are, of course, my own responsibility.
8
The purpose of this paper is to explore these questions. I will not really
attempt to answer the question, “How do they do it?”—we don’t know how
people do even the simplest and most obvious linguistic operations. But
what I will show is that these phenomena also occur in English. It is often
not superficially obvious that we are dealing, in English, with phenomena
analogous to politeness or hesitance markers in other languages; there
are often no special separate readily identifiable morphological devices.
Rather, these distinctions are expressed by forms used elsewhere for other
purposes. Therefore it is easy to imagine that they are not present at all.
But I hope to show that the reverse is true; and further, that if the pres-
ence and uses of these forms are recognized, several of the most difficult
[ 8 ] Context Counts
9
we know. So it is normally true in all languages and all situations that one
must somehow make clear the type of speech act involved: are you asking
a question, making a statement, or giving an order? Ambiguities in this
regard are generally not tolerated.1 Some languages require that you know
more about the speech situation than this. English sometimes requires
overt notice as to whether the speaker believes a past-time event is rele-
vant to the present, by the use of the perfect tense rather than the preterit.
Other languages require that there be overt expression of the identity of
speaker and/or addressee: What are their respective social positions? And,
related to this, of course, what are their respective ages and sex? English
only sometimes requires that these be recognized overtly; other languages,
such as Japanese, require it much more often. But it is hard to think of a
language that requires one special overt marker if the speaker has blue eyes,
and a different one if the speaker has brown eyes. This is contextual infor-
mation, as real and available to the speakers of a language for the purpose
of making distinctions as are differentiation of age and sex; yet only the
latter two often occur as linguistically significant contextual information.
In any case, I trust that, by the end of this discussion, it will be per-
fectly clear that there are areas of linguistic competence that cannot be
described in any theory that does not allow an integration of information
about the context in which the discourse takes place—sometimes erro-
neously referred to as “realworld” as opposed to “linguistically relevant”
situation—and the purely linguistically relevant information the sentence
seeks to convey: superficial syntax, choice of lexical items, and semantics
aside from contextually relevant meaning elements.
I shall try to substantiate some of the claims I have been making by
looking at examples.
We all know, or at least know of, languages that employ honorifics
as essential elements in sentences. Sometimes they occur with personal
names, and in these cases it is fairly easy to see what is going on: one usu-
ally assumes that the speaker either actually is lower in status than the
addressee, or is speaking as if he were. In the latter case, which is perhaps
the more usual in conversational situations, it is assumed that this linguis-
tic abasement occurs for reasons of politeness. But an important question
1. Gordon & G. Lakoff (1971) discuss a number of interesting cases where, if one looks
only at superficial syntactic configurations, apparent ambiguities of this type do in fact
exist: e.g., It’s stuffy in here, most normally a declarative statement, may, under specific,
contextually determined conditions, be interpretable as an imperative, equivalent to
Please open the window. As they show, this does not indicate that such sentences really
are ambiguous between the two interpretations: it indicates rather that context must
play a role in the interpretation of sentences.
[ 10 ] Context Counts
11
is usually glossed over: Why is it polite for the speaker to suggest that the
addressee surpasses him in status? In some languages we find honorifics
related to non-human items, to show that the speaker considers them of
importance in one way or another. How is this related to any notion of
politeness, which is a concept involving behavior between human beings?
Another problem is that many languages apparently have two kinds of
honorifics. One is the kind I have just mentioned. But going hand in hand
with this is the use of forms that humble or debase the speaker himself,
or things connected with him. Translated into English, this often has ludi-
crous results, e.g., “Honorable Mr. Snarf have some of my humble apple
pie?” This sort of translation is ludicrous for several reasons, but perhaps
principally because, by translating the honorific and dis-honorific, if I may
use that term, with overt adjectives, the sense of the sentence has been
palpably altered. In the original language, the sense of superiority or infe-
riority conveyed by the honorifics is presupposed, or implicit. The use of
adjectives like honorable and humble makes these concepts explicit. So
what had been a tacit suggestion, in effect, is now made overt. The English
translations do not, I think, allow the monolingual speaker of English to
get any sense of how a speaker of Japanese feels when he is addressed
with -san. But I believe there are locutions in English whose force comes
close to that of the true honorific, because the differentiation in status
they establish is implicit rather than overt. These forms are also used for
the sake of politeness (as adjectives like humble and honorable never are).
I said earlier that these contextually linked forms had not been recog-
nized in English partly because the forms utilized for this purpose had
other, more obvious uses. English modals are a case in point. Certain uses
of the modal must are parallel to the use in other languages of special
honorific forms:
Let us assume, for the purpose of analyzing these sentences, a special social
situation: a party, at which the hostess is offering the guests a cake that
she baked herself or at least selected herself, and which she therefore takes
responsibility for. In such a social context, (1) is the most polite of these
forms, approaching in its range of appropriateness that of a true honorific
in languages that have such forms. Further, although in theory (2) should
be more “polite” than (1), in actual use it is not: in the situation established
above, the use of (2) would be rude, while (1) would be polite. And (3),
which might at first seem the most polite form, actually is the least. Why
is this?
Finding the answer lies partly in determining what constitutes polite-
ness, and of course, its opposite, rudeness. If we can define these notions,
then the uses of these modals will be seen to be governed by the same
assumptions of politeness as govern the use of honorifics; once the prin-
ciple is understood, it can be transferred from language to language. What
we are dealing with here is something extralinguistic—the way in which
individuals relate to one another—that directly affects the use of lan-
guage. We must understand something about non-linguistic social inter-
action before we can see the generalization that is in effect regulating the
use of sentences like (1)–(3), along with the use of affixes like -san and o- in
Japanese.
It is obvious, of course, that what passes for politeness in one culture
will appear to a member of another culture as slavishness or boorishness.
We are all familiar with examples of this. Then how can we talk about
universal conditions governing the use of honorifics and other politeness
markers? I think we can assume that there is a universal definition of
what constitutes linguistic politeness: part of this involves the speaker’s
acting as though his status were lower than that of the addressee. What
may differ from language to language, or culture to culture—or from
subculture to subculture within a language—is the question of when it
is polite to be polite, to what extent, and how it is shown in terms of
superficial linguistic behavior.2 Although a speaker may know the univer-
sal definition of politeness, he may apply it at the wrong time or in the
wrong way if he attempts to transfer the uses of his own language directly
into another; hence the ludicrousness that results from taking a polite
concept implicit in one language and making it explicit. If, in a given lan-
guage, one’s own possessions are customarily followed by a marker of
humility (a situation which perhaps can be symbolized by (4) below), it
does not follow that (5), in which what is implicit in the marker in (4)
is made explicit, is a reasonable English translation of (4). In fact, as has
2. So, for example, if an officer in the Army (a subculture with special status-related
rules) gives a command to a private, he will not normally preface his command with
please. Although in most English-speaking groups the use of please prefaced to an
imperative is a mark of politeness, to use please in this situation will be interpretable as
sarcastic. Again, in some cultures it is considered polite to refuse an invitation several
times before one is conventionally “prevailed upon” to accept: if a speaker from such a
culture finds himself in one where it is considered polite to accept invitations at once
with thanks, confusion and worse will inevitably ensue, with each party impressing the
other as unbelievably boorish or stupid.
[ 12 ] Context Counts
13
(a) The speaker is higher in rank than the superficial subject of must, in
sent. (1) identical with the addressee. As such the former can impose
an obligation on the latter.
(b) The thing the addressee is told to do is distasteful to him: he must be
compelled to do it against his will.
(c) Something untoward will happen to the addressee if he does not carry
out the instruction.3
3. It seems reasonable to believe that, of the three assumptions comprising the
meaning of must, (a) and (c) are first-order presuppositions, and (b) second-order. The
reason for making this claim is that (a) and (c) can be questioned, as is typical of first-
order presuppositions, while (b) cannot, as the following examples show. In reply to,
e.g., You must take out the garbage!, the respondent might retort, under the appropriate
circumstances, with You can’t make me! or Who’s gonna make me? (which contradict (a),
and are equivalent to “You don’t have the authority”), or with So what if I don’t? (which
contradicts (c), and is equivalent to “If I don’t do it, I won’t suffer”). But he cannot reply
with *I want to anyhow! (which would be a contradiction of (b) and equivalent to “I am
not being made to do it against my will.”).
[ 14 ] Context Counts
15
(6) You must clean the latrine, Private Zotz: this is the Army, and I’m
your sergeant.
(7) You must take this medicine, Mr. President, or you will never get
over making those awkward gestures.
But if another guest is offering the cake, both (2) and (8) are perfectly
appropriate and usual, since the guest is not praising his own property. This
shows that implicit and explicit assumptions—in this case, of the value of
Finally, it is now easy to see why may in (3) is not a polite form: in fact,
its use makes two assumptions, both of which are counter to the conven-
tions of politeness: (a) that the person who is able to grant permission (by
the use of may) is superior to the person seeking it; (b) that the person
seeking permission not only is not averse to doing the act indicated, but
wishes to do it. Then the further assumption is that, as far as the person
receiving permission by sentence (3) is concerned, having the cake is a good
thing. As with should, this is counter to the usage of politeness.
These examples show several things. First, there are uses of the modals
that reflect politeness, in terms of relative status of speaker and hearer,
and implicit desirability of the act in question. In this respect these modal
uses are parallel to the use of honorifics in other languages. Second, in
order to tell how a modal is being used, and whether certain responses to
it are (linguistically) appropriate, one must be aware of many extralinguis-
tic, social factors. Just as, in speaking other languages, one must be aware
of the social status of the other participants in a conversation in order
to carry on the conversation acceptably, so one must at least some of the
time in English, a language usually said not to require overt distinctions of
this sort.
There are many other examples of politeness conventions explicitly
realized in English. One is the use of imperatives, a task fraught with
4. According to Tazuko Uyeno, although not every Japanese noun may receive the
o-honorific prefix, those that can behave as suggested in the text. E.g., the word taku
“house” will take the prefix o- when it refers to the home of someone other than the
speaker and will occur without o- when the speaker’s own house is being referred
to. The same informant points out an interesting difference in polite usage between
Japanese and English, also relevant at this point: I have noted above that in English the
modal must, ordinarily not a polite form, may be interpreted as polite in specific social
contexts where one is able to “ignore” certain aspects of the meaning of must. But in
Japanese, this is not the case: I cannot use the word-for-word equivalent of “You must
have some cake” as a polite utterance equivalent to its English translation. It would, in
fact, be interpreted as rude under the circumstances. One must rather say something
like, “Please have some cake as a favor to me.” Thus it is not necessarily true that one
can “ignore” the same aspects of meaning in two languages.
[ 16 ] Context Counts
17
perils for one who does not understand the application of levels of polite-
ness in English. For example, consider the following ways of giving an
order. When can each be used appropriately? What happens if the wrong
one is used?
It would seem clear that these sentences are ranked in an order of descend-
ing politeness. To use (17), your status must be higher than that of the
addressee; moreover, you must be in such a situation that you don’t even
care to maintain the conventional pretense that you are addressing him
as an equal. That is, (17) deliberately asserts the superiority of the speaker
over the addressee, and as such is rude in a situation in which it is not nor-
mal to make this assertion.5 By contrast, (15) merely implies this assump-
tion of superiority: it assumes compliance, and hence suggests that the
speaker has the right to expect this compliance, and that the speaker
therefore outranks the addressee; but it does these things much more
covertly than (17). But, though not normally a rude form, it is still not
really a polite one. Again, however, we must make an exception for one
case, analogous to the one made in the first set of cases with modals: if the
5. This claim ignores the “jovial” use of sentences like (17) as used between close
friends, almost invariably male. Other examples are: Get your ass in here, Harry! The
party’s started! and What makes you think you can go by my house without coming in, you
asshole? It seems that, between close male friends in some American subcultures at any
rate, the purpose of such otherwise unpardonably rude exclamations is to say, “We’re
on such good terms that we don’t have to go by the rules.” This linguistic impropri-
ety occurs in relationships of the same degree of closeness as those which allow their
members, for example, to invite themselves over to each other’s houses—otherwise a
non-linguistic breach of propriety of similar magnitude. This illustrates again the par-
allelism of linguistic and non-linguistic concepts of politeness. These examples show,
incidentally, that English, like Japanese, makes sex distinctions in the types of sen-
tences possible. While a woman in most American subcultures would never use the
above sentences, she might use the following to much the same effect, but lacking the
obscenities: Go ahead, have some more cake, Ethel—you’re so fat, who’ll notice if you get
fatter? Between very close friends, such a remark might be taken as an acceptable joke,
but under any other conditions it is an unpardonable insult. There are other expres-
sions confined to the feminine vocabulary: in particular epithets like “gracious!” or
“dear, dear.” So English is again not so very unlike Japanese, except that the speaker
of English can refrain from these usages altogether, but the speaker of Japanese must
make his (or her) sex explicit in most conversations.
addressee is at the speaker’s door and is a friend, (15) is much more nor-
mal than (13)–(14) as an invitation to enter. The first two, in fact, do not
seem polite in this context: they give the impression of forced hospitality.
Here again we seem to be depending on a more complex notion of polite-
ness: both (13) and (14), like (2), imply that the addressee has the choice of
complying or not—that his status is sufficiently high with respect to the
speaker that he can obey or not as he sees fit—while (15), like (1), seems at
first to suggest that the addressee has no choice, that his status is so low
that he is obliged to obey. Yet both are relatively polite in this sort of social
context. The reason in the case of (16) is parallel to that in (1): the speaker
is implying here (by convention: he doesn’t really make this assumption, of
course; it would be bizarre if he did) that the addressee doesn’t really want
to come in, that he will enter only under duress. Since (13)–(14) do not
allow this assumption, they are less polite. So again the two definitions of
politeness—status vs. desirability of the speaker’s offering—are at odds,
and again the latter seems stronger. When the speaker is not really offer-
ing something of his own, the status assumption becomes paramount,
and (13)–(14) become more polite than (15). This is the ease in a doctor’s
office, for example, where the receptionist is more likely to use (13)–(14).
Again, (13)–(14) are likely to be used for “forced” politeness—e.g., when
inviting an encyclopedia salesman in, under duress. I am not sure why
this is so. But it is also true that a superior may address an obvious infe-
rior (for example, in the Army) by (15), with no sense of sarcasm, i.e., no
sense that he is being inappropriately polite. But if an officer addresses a
private with (13), he is necessarily being sarcastic. There is no possibility
of sarcasm, however, in the use of humbling forms of politeness, such as
are found in (1) and (15). This is reminiscent of a fact that has been known
for some time about presupposition in general: a first-order presuppo-
sition may be negated or questioned, under some conditions; a second-
order presupposition cannot be. This suggests that the type of politeness
involved in a usage like (15) or (1) is more complex in derivation than is the
simple status-equalizing case in (13) or (2). In fact, it is probably true that
the humbling type allows the status type to be deduced from it (if what
I have is no good, one can deduce that I don’t outrank you, in this respect
anyway), so that the humbling type of politeness is one level deeper than
the status type.
There are other assumptions, made in normal conversation, that are not
tied to concepts of politeness. These, too, show up in non-obvious ways in
the superficial structure. Some types which have been discussed by Grice
([1967] 1975), as well as by Gordon & G. Lakoff, are rules of conversation.
[ 18 ] Context Counts
19
All these assume, in addition, that the status of speaker and hearer is
appropriate with respect to each other. (Of course, there are special situa-
tions in which all these are violated: lies, “small talk,” tall stories, riddles of
certain types, and requests as opposed to commands. But in general these
conditions define an appropriate conversational situation.)
But sometimes, even in ordinary conversational situations, some
of these rules are violated. This is analogous to violating a rule of gram-
mar: normally we should expect anomaly, lack of communication, etc. When
this is done baldly, e.g., by small children or by the insane, we do in fact
notice that “something is missing”; the conversation does not seem right.
But in ordinary discourse among normal individuals we can often discern
violations of these rules and others, and yet the total effect is not aberrant.
One way in which apparent contradictions are reconciled is by the use of
particles like well, why, golly, and really. Although these are often defined
in pedagogical grammars as “meaningless” elements, it seems evident that
6. These implicit rules show up overtly in certain locutions. Cf. the following:
{ } believe
John is a Communist, and if you don’t *obey
*answer
me, ask Fred.
{ }
*believe
Get out of here, and if you don’t obey me, I’ll sock you.
*answer
{ } *believe
I ask you whether John left, and if you don’t *obey
answer
me, I’ll be furious.
These examples show that, with each type of speech act—declaring, ordering, and
asking—an “appropriate” type of response is associated, and that this association
shows up linguistically in superficial structures.
they have real, specific meanings, and therefore can be inappropriately used.
It is therefore within the sphere of linguistics to define their appropriate
usage. Moreover, this appropriateness of usage seems at least sometimes
to involve the notion of “violation of a normal rule of conversation.” These
particles serve as warnings to participants in the discourse that one or more
of these rules is about to be, or has been, violated. When this warning is
given, it is apparently legal to violate the rule—that is, of course, only the
specific rule for which the warning was given. Otherwise confusion results.
I have shown elsewhere (R. Lakoff 1973) that the English particles well and
why function in this way. Well serves notice that something is left out of the
utterance that the hearer would need in order to understand the sentence—
something, normally, that he can supply, or that the speaker promises to sup-
ply himself shortly. That is, well marks a violation of the second part of Rule II.
Why indicates that the speaker is surprised at what the addressee has said: it
suggests that perhaps the prior speaker has violated Rule I, in the case of a
statement, or II, in the case of a question. Other analogous cases in English
involve special syntactic configurations rather than particles. Consider sen-
tences like the following:
In the even- numbered examples above, we have tag- forms, one for a
command—as discussed earlier—and one for a statement. It is worth ask-
ing whether these superficially similar structures have any deeper similar-
ity: whether the reasons for applying tag-formation to imperatives are related
to the reasons for applying this rule to declarative statements. It is more or
less traditional in transformational literature to suggest that the two types
of tags have little in common aside from superficial similarities of forma-
tion. But there are reasons for supposing that, in fact, there are real semantic
reasons for this apparent superficial coincidence. This is a rather satisfying
hypothesis, if it can be substantiated: it would suggest that these two bizarre
and highly English-specific formations have a common function, so that two
mysteries may be reduced to one.
With reference to Rules III and V above, one way in which a tag question
like (20) is distinguished from an ordinary statement like (21) is that the
speaker really is asking less of the hearer. A speaker can demand belief from
someone else only on condition that he himself fully believes the claim he
is making. But the function of the tag is to suggest that the speaker, rather
[ 20 ] Context Counts
21
7. As is well known, English has at least two intonation patterns associated with
tag-statements (or, as they are more commonly called, tag-questions). One, rising, is
closer to a question, as is predictable from the intonation pattern; this expresses less
certainty on the speaker’s part, and less hope of acquiescence by the addressee. The
other, falling, is nearer to a statement, expressing near-certainty, with just the mer-
est possibility left open that the addressee will fail to agree. The second type is often
found as a kind of gesture of conventional politeness, meaning something like, “I have
enough information to know I’m right, but I’m just letting you have your say, in order
to be polite.” It is interesting that some verbs of thinking, in the 1sg. present, have the
same ambiguity resolved by the same difference in intonation pattern, and both types
of locutions are used for similar purposes (cf. fn. 8 below). There is a third type of tag-
question, used when the speaker definitely knows something is true, based on personal
observation, and merely wishes to elicit a response from the addressee. This has the
particle sure inserted, as in It sure is cold in Ann Arbor, isn’t it?, vs. It’s cold in Ann Arbor,
isn’t it? The latter sentence might be used if the speaker had merely read reports that
the average temperature in Ann Arbor was 19˚. He could not, under these conditions
and if he had never been in Ann Arbor, use the former sentence. In Japanese, accord-
ing to Uyeno (1972), the particle ne expresses both the senses of the second sentence,
while its longer form nee corresponds to the first sentence.
8. In fn. 7 I alluded briefly to the uses of verbs of thinking. My point is that verbs such
as guess, suppose, believe, and sometimes think, when used in the 1sg. present, do not
describe acts of cogitation: rather, they are means of softening a declarative statement.
Consider the following sentences:
If sentences like (a) and (b) (cf. Ross 1970) express certainty on the part of the speaker
through the (overt or covert) performative verb of declaration, then (c) expresses the
speaker’s feeling that the event described in the complement of the verb of thinking
is a probability rather than a certainty. As (b) corresponds to (a), it is my contention
(made for other reasons in R. Lakoff 1969) that (d) corresponds to (c): in fact, they are
closely synonymous in many of their uses, just like (a) and (b). As pointed out in fn. 7,
the same disambiguation by intonation exists for both.
9. All examples here are from Uyeno, who discusses these cases, with many more
examples as well as other extremely interesting particle uses in Japanese, in her dis-
sertation. Kazuhiko Yoshida and Chisato Kitagawa provided enlightening discussion
and further examples of these constructions.
[ 22 ] Context Counts
23
(22) John is here ne. “John is here, isn’t he?” (a declarative, but
without the normal declarative demand for the hearer’s belief)
(23) Come here ne. “Come here, won’t you?” (an order, but without the
normal imperative demand for the addressee’s obedience)
(24) Is John here ne? “I wonder if John is here …?” (a question, but
without the normal interrogative demand for the addressee’s
response)
(i) The status of the addressee should be somewhat higher than that of
the speaker, since offering a choice is an act of deference. (This may
A similar situation can be shown to pertain in the case of yo, the other
particle mentioned above, which may be appended to declaratives, impera-
tives, and interrogatives:
(25) John is here yo. “I tell you John is here, (and you’d better believe
it).” (a declarative in which the speaker explicitly demands the
addressee’s observance of Rule III)
(26) Come here yo. “I’m telling you to come here, (and you’d better
obey).”10 (an order in which the speaker explicitly demands the
addressee’s observance of Rule V)
(27) Is John here yo? “What do you mean, is John here?” “Are you
asking me, ‘Is John here?’ ”
10. Kazuhiko Yoshida points out that, though this sentence may be used by both men
and women, the effect is different. The translation given here is the sense it would have
when spoken by a man. If spoken by a woman, it would mean something like, “I really
hope you will come here. Please don’t forget.” A strong command has been replaced
by an earnest request. The effect of yo here in women’s speech seems to be something
like an attempt to express the idea that the speaker wishes she had the status to insist
on the observance of Rule V. The use of yo by a speaker of much lower status than
the addressee (which, in conventional Japanese society, presumably automatically
includes all women) is in a sense contradictory for reasons to be discussed below. The
contradiction is resolved by using yo to indicate a strong request, rather than a strong
injunction that cannot be disobeyed. This is still another example of how non-linguis-
tic context (such as the sex of participants in a discourse) affects the interpretation of
sentences, and therefore must be considered part of the linguistic information avail-
able to a speaker.
[ 24 ] Context Counts
25
[ 26 ] Context Counts
27
if one is sufficiently superior, he has no reason at all to fear that his injunc-
tions, explicit or implicit, will be disobeyed. Therefore, yo is most apt to be
used where the speaker is somewhat superior to the hearer, so that he has
the right to make demands, but not so much higher that he has no need to
make them. Obviously, yo cannot be used by someone of lower social status
than the addressee.11
Having given evidence that English speakers are capable of making dis-
tinctions of the first two types alluded to at the beginning of this paper, let
us now examine the third. What about the use of “dubitatives” and their
opposites, as endings on verbs or particles, to express uncertainty or cer-
tainty on the speaker’s part? I have already given examples of “dubitatives”
in English: the use of I guess or of tags, as has been shown, is essentially
dubitative in function; it is a sign that the speaker is not altogether pre-
pared to stand by his assertion, in the sense that he does not have complete
confidence in what he is asserting, since he does not—he cannot, as we
11. An apparent problem for this analysis (or, so far as I can see, any analysis) is the
fact that yo and ne may occur in sentences like this:
This is equivalent to something like, “I guess this is your book—I certainly hope you’ll
agree.” Such a sentence might be used in circumstances like this: suppose that the
speaker of (c) has previously borrowed the book in question from the addressee. The
addressee has throughout the transaction, behaved as though the book were his to
lend. But now a third person accosts the speaker of (c), demanding the book back,
as if it had always been his. The speaker of (c)—partly because he knows or likes the
addressee better than the third person and therefore trusts him more, partly because it
is to his advantage for the book to belong to the addressee—still feels fairly confident
that the addressee really owns the book. But he is not as sure as formerly, and needs
confirmation. His use of (c) is equivalent to saying, first, “I say this is your book, and
I hope you believe it” (i.e., “This is your book ne.”) Then he adds, “I really hope you
can go along with this hypothesis; you’d better (for my sake) agree to this”—where yo
modifies and strengthens the hope of the speaker that the addressee will be able to
acquiesce: i.e., “I’m giving you a chance to relax Rule III, but I hope you don’t take it.”
I am not sure that this is precisely correct; but in any event the effect is not contra-
dictory, and it does seem as though yo modifies ne rather than the utterance itself,
particularly as an utterance like (a) is more apt to be used by women than is a normal
yo-sentence.
and it later turns out that (30) is not the case, my addressee may later say,
“You were wrong about (30).” If he does, I have no recourse but to agree,
providing his evidence is incontestable. But if instead I say
under the same conditions, and later the addressee says, “You were wrong,”
then I have the option of replying, “No, I only said I thought (30) might be
the case.” That is, I can claim I was not really making that assertion. Thus
verbs such as guess in the 1st person singular, like tags, function as subjunc-
tives do in languages like Latin:
}
(32) Marcus Publium interfecit quod
uxorem suam corrupisset. “Marcus killed Publius
(33) Marcus Publium interfecit quod because he seduced his wife.”
uxorem suam corrupit.
Here the presence of the subjunctive in (32) indicates that the speaker
is not prepared to take responsibility for the claim that the alleged rea-
son is in fact the real reason for an action. With the indicative, as in (33),
the speaker implicitly takes responsibility. We have no natural means of
expressing this difference in English: we must resort to paraphrase. But we
do have analogous devices, illustrated by (31), usable under other gram-
matical conditions. If we were teaching English to a speaker of Latin, we
might want to exemplify this use of guess (which I do not believe is found in
Latin in this sense) by suggesting parallels with sentences like (32), rather
than by resorting to elaborate circumlocutions, which, as we have seen,
don’t really give the same idea.
English has other devices to express the speaker’s acceptance or denial
of responsibility for something in an utterance. Like the honorifics dis-
cussed earlier, these are not generally recognized as dubitatives or “cer-
taintives” (if I may coin that term), because they are not characteristically
obligatory morphemes, and because they function in only a limited subset
of sentence-types. It is not surprising, in view of our earlier findings, to
see that modals perform these functions along with many others. With
verbs of perception, the modal can displays certain semantic properties not
[ 28 ] Context Counts
29
derivable from any normal definition of can. As first noted by Boyd and
Thorne (1969), under certain conditions sentences containing can appear
to be synonymous to sentences without it:
But in some contexts where this should be true, particularly when the verb
is non-1st-person present, we find that although the denotative content of
the sentence pairs remains the same, one member often contains implica-
tions that are lacking in the other; e.g.,
(36) That acid-head John hears voices telling him he is Spiro Agnew,
so don’t play golf with him.
(37) That acid-head John can hear voices telling him he is Spiro
Agnew, so don’t play golf with him.
(38) Mrs. Snickfritz has eyes like a hawk: she can spot dust on your
carpet even if you just vacuumed.
(39) Mrs. Snickfritz has eyes like a hawk: she spots dust on your
carpet even if you just vacuumed.
By using (42), the speaker takes responsibility for the obligation. But (43)
is neutral; he may merely be reporting an obligation he does not necessarily
approve of. Compare:
[ 30 ] Context Counts
31
In this case, the speaker is not taking or refusing responsibility for the fac-
tual content of the sentence, as he was in the other cases. Here the truth of
the modal notion itself is at issue—whether there really is a true “obliga-
tion” involved. I am not sure whether dubitatives in other languages can
affect or cast doubt on modality, as these can.
There are examples parallel to Larkin’s with other modals; this is not an
isolated fact about must/have to, as the previous examples were isolated
cases with can or will. This fact suggests a pervasive property throughout
the modal system. The existence in English of periphrastic modal forms
may not be due wholly to the fact that modals are syntactically defective;
there is a real need for the periphrastic forms at a semantic level as well.
Compare the following:
In the first set, will is the root sense will of command: “I order that …”
Thus (46) is a direct order, for which the speaker is assuming responsibility.
In (47), he is still transmitting an order, but it may have originated with
someone else; it may not be an order he goes along with. For most speakers
of American English, (48) expresses the direct giving of permission by the
speaker. But (49) may be used to report someone else’s giving of permis-
sion. Thus, will/be to and may/be allowed to are parallel to must/have to.
I have, then, given examples of phenomena in English and other lan-
guages that bear out certain contentions:
REFERENCES
Boyd, Julian, and James Thorne. 1969. The deep grammar of modal verbs. Journal of
Linguistics 5.57–74.
Gordon, David, and George Lakoff. 1971. Conversational postulates. In Papers
from the seventh regional meeting of the Chicago Linguistic Society (CLS 7),
Chicago, 63–84.
Grice, H. Paul. (1967) 1975. Logic and conversation. Syntax and semantics 3: Speech
acts, ed. by Peter Cole and Jerry L. Morgan, 41–58. New York: Academic Press.
Lakoff, Robin. 1969. Syntactic arguments for negative transportation. In Papers from
the fifth regional meeting of the Chicago Linguistics Society (CLS 5), Chicago,
140–47.
Lakoff, Robin. 1970. Tense and its relation to participants. Language 46.838–49.
Lakoff, Robin. 1973. Questionable answers and answerable questions. In Issues in
linguistics: Papers in honor of Henry and Renee Kahane, ed. by Braj B. Kachru
et al., 453–67. Urbana: University of Illinois Press.
Larkin, Don. 1969. Some notes on English modals. University of Michigan Phonetics
Lab Notes 4.314.
Ross, John Robert. 1970. On declarative sentences. Readings in English
transformational grammar, ed. by Roderick A. Jacobs and Peter S. Rosenbaum,
222–72. Waltham, MA: Blaisdell.
Uyeno, Tazuko. 1972. A study of Japanese modality: a performative analysis of
sentence particles. Doctoral dissertation, University of Michigan.
12. As should be apparent to anyone familiar with other than purely transforma-
tional linguistic tradition, the notion that contextual factors, social and otherwise,
must be taken into account in determining the acceptability and interpretation of sen-
tences is scarcely new. It has been anticipated by a veritable Who’s who of linguistics
and anthropology: Jespersen, Sapir, Malinowski, Firth, Nida, Pike, Hymes, Friedrich,
Tyler, and many others. But the idea has not merely been forgotten by transforma-
tional grammar; rather, it has been explicitly rejected. Therefore, to bring up facts such
as these within the framework of recent linguistic discussion is to do more than merely
restate an old platitude. I hope that by discussing new facts, and expatiating on their
theoretical implications, I have shown that contextual factors cannot be avoided by the
linguist of any theoretical view, if he is to deal honestly and accurately with the facts
of language.
[ 32 ] Context Counts
33
D eborah Tannen once said to me, “Robin was the one who inspired the
whole field” of the study of women’s language. Indeed, many women
were awakened by her seminal article “Language and woman’s place” (1973)
and were inspired both to pursue the topic of women’s language and, just
as importantly, to pursue careers as professionals in a field that had been
traditionally dominated by men and male-oriented scholarship. Needless
to say, I am one of the many women who took inspiration and strength
from Robin’s work to find a place of my own in the field.
Until I encountered Robin Lakoff’s writing, I had thought that to be a
scholar was to view the world from the ivory tower. It was when I was read-
ing a passage in “Language and woman’s place” that I realized that I had
been struggling hard to gear my mind to a male way of thinking. In that
article, she argued that to be both a woman and a scholar meant that she
had to be culturally bilingual. I could not agree more! In short, she taught
me that a woman can and should be a scholar in her own way without hav-
ing to become bilingual. She showed it herself in her thinking and writing.
Second, Robin has contributed greatly to the study of linguistic polite-
ness. Among the theories on linguistic politeness that were developed
during the 1970s and early 1980s, the theory outlined in this paper still
stands strong today as an illuminating paradigm for research. Her theory
34
[ 34 ] Context Counts
35
lives, not arguments in the ivory tower. This attitude helps me understand
what I can do as a scholar of a humane discipline.
REFERENCES
T H E L O G I C OF P OL I T E N E S S [ 35 ]
36
37
CHAPTER 2
This paper originally appeared in Papers from the Ninth Regional Meeting of the Chicago
Linguistic Society (CLS 9), edited by Claudia Corum, T. Cedric Smith-Stark, and Ann
Weiser, 292–305. Chicago: Chicago Linguistic Society, 1973. Reprinted here with per-
mission of the publisher.
38
And perhaps other cases are related, where although truth-values are not
explicitly involved in the test, the assumption we are talking about (the
pragmatic presupposition) will be held by virtually anyone. So we might say
that sentence (2a) is acceptable only in case (2b) is pragmatically presup-
posed by the normal addressee:
(That is, assumption (2b) permits us to use the perfect tense in (2a).)
But as we ventured out into this new way of looking at sentences, we
found more and more that a sentence which was perfectly acceptable
under one specific set of conditions might be bad under another, both
quite conceivable in the real world. I have (1969) given examples such as
those in (3): each of these sentences is good, within its particular set of
contexts:
(The difference being, as has been pointed out, that in (a) the speaker
either has no idea whether or not beans are wanted or else assumes they
are not not wanted; but in (b) he must assume that there will be a posi-
tive response; he may, of course, prove to have been wrong, but this is the
assumption he is working under as he utters (b).)
Still more complex cases are those in which the sentence reflects the
speaker’s attitude toward his social context: more specifically, his assump-
tions about (1) the people he is communicating with: their feelings about
him, their rank relative to his; (2) the real-world situation in which he is
[ 38 ] Context Counts
39
And so we might want to say that a sentence like (6a) is a good sentence
on pragmatic (and, of course, other) grounds: it violates no assumptions
about real-world interaction; but (6b) is a very odd sentence, and purely for
pragmatic reasons.
Very briefly, and glossing over exactly what is going on here, we can say
that the question form (as in (b)) is polite and the declarative form of
this sentence (as in (a)) impolite, but the communicative content of
this utterance is unalterably impolite. So if we match a polite construc-
tion with an impolite meaning, either irony or out-and-out aberrancy
will occur, on pragmatic grounds alone. A sentence like (6b) is perfectly
constructed syntactically and violates no imaginable semantic con-
straints (as would a sentence like (7), which is bad in a very different way
from (6b)):
(7) *That rock can take its methodology and shove it.
Another test of syntactic rules that has proved very useful is that of
ambiguity. Syntactic ambiguity is enthroned in the literature, proving the
need for underlying structures and transformational rules to disambiguate
utterances like (8) on purely syntactic grounds:
(9) The police came into the room and everyone swallowed his cigaret.
[ 40 ] Context Counts
41
That is, (10) might be a truly subservient utterance in case its speaker
was really subordinate to the addressee, or equal to him and not a close
acquaintance. In this case please might mean “I’m asking you to do this
as a favor to me, since I can’t constrain you to do it.” But suppose Sp is
superior to A. In this case, the use of please is conventional. The speaker
really means something like “I’m asking you to do this, but I really have the
power to force you, I’m just acting like a nice guy.” The difference is clear
to the addressee: in the first case, he can refuse; in the second, he’d better
not, without good reason. So a response like (11) is valid as a refusal in case
(1) but not in case (2):
RULES OF CONVERSATION
1. Quantity: Be as informative as required
Be no more informative than required
[ 42 ] Context Counts
43
RULES OF POLITENESS
1. Don’t impose
2. Give options
3. Make A feel good—be friendly
Now sometimes two or more of these rules may be in effect together, rein-
forcing each other; just as often, we must make a choice—are we in an R1
or an R3 situation?—and one will cancel the other out. One, that is, may
supersede the other. But how do we tell which of the rules we are enforcing
at a given time? Let us give some examples of how these rules operate:
1. Rule 1: Don’t impose. This can also be taken as meaning, remain aloof,
don’t intrude into “other people’s business.” If something, linguistically or
otherwise, is nonfree goods, in Goffman’s sense, this rule cautions us to
steer clear of it, or in any event to ask permission before indulging in it. So
we request permission to examine someone else’s possessions; and simi-
larly, if we are about to ask a question that is personal, we must normally
ask permission before we do it:
(12) May I ask how much you paid for that vase, Mr. Hoving?
[ 44 ] Context Counts
45
it’s the thought that counts. You appear to give the addressee an out (see
Rule 2), even though actually you don’t.)
Also governed by Rule 1 are certain linguistic devices: passives and imper-
sonal expressions, for instance, tend to create a sense of distance between
speaker and utterance, or speaker and addressee. Hence sentences contain-
ing these forms tend to be interpreted as polite, like the other forms of
Rule 1-governed behavior. The proper butler says (14a), not (14b).
And academic authors, as is well known and often bemoaned, tend toward
passive and impersonal sentences in their work, as well as the authorial
we, which, like the polite vous in French (and analogous cases in many
other languages), creates a distance between himself (in the vous case,
the addressee) and the other people involved in the communication. (The
authorial we is thus parallel to the vous of egalitarian nonsolidarity, as dis-
cussed by Brown and Gilman (1960); but the imperial we is parallel to the
vous of superior status.)
Related to the prohibition on nonfree goods under R1 conditions, we
find the use of technical terms to avoid mentioning unmentionables, like
sex, elimination, or economic difficulties (in our culture; other cultures
may have, of course, different unmentionables). This is the practice of
bureaucratese, medical and legal terminology, and general stuffiness. It is
to be distinguished from euphemism, which is an R2-related use. Technical
terminology seeks to divorce the subject from its emotional impact: “We’re
talking about IT all right, but it doesn’t have its usual connotations because
we’re divorcing it from all emotional content.” So you say copulation, or def-
ecation, or disadvantaged, if you have to say any of them.
Rule 2 operates sometimes along with R1, sometimes in cases where R1
would be inappropriate. R2 says, “Let A make his own decisions—leave
his options open for him.” This may seem the same as remaining aloof,
but actually it only sometimes is. So certain particles may be used to give
the addressee an option about how he is to react: some of these particles
(hedges) have been discussed by George Lakoff (1972). Some hedges also
have the effect of suggesting that the speaker feels only a weak emotional
commitment toward what he’s discussing: that is, they reflect the speaker’s
feelings about the sentence. So the use of such hedges violates R1: they are
not in place in truly formal discourse.
A sentence like (15) could not be used, say, in a New York Times editorial, a
format which invokes R1 almost exclusively. But it might be used in the inter-
ests of politeness: under some circumstances a speaker whose true opinion
is representable as (16) may utter (15) to avoid social unpleasantness.
That is, quite apart from their basic functions, such sentences may also
function as politeness devices in accordance with R2. Obviously, both (17a)
and (17b) can be used when the speaker genuinely is uncertain of what he is
asserting. But often, too, they are used where the speaker speaks with full
confidence: he knows what he’s talking about, but does not wish to assert
himself at the risk of offending the addressee. Such sentences, under these
conditions, mean something roughly equivalent to “I say this to you, but
you’re under no compunction to believe it: I’m not trying to buffalo you.”
Whether for the sake of politeness or because the speaker really doesn’t
know the answer himself, such sentences leave the final decision as to the
truth of the sentence up to the addressee.
Just as technical terms for unmentionables are R1 devices, euphemisms
are in the realm of R2: they retain the presumption that the topic under
discussion is forbidden, but they seek to dispel the unpleasant effect by sug-
gesting that A need not interpret what is being said as THAT. (Of course,
he obviously does, so this is again conventional; but he is at least appar-
ently being given the chance of opting out, pretending that the unmention-
able topic has not been broached, and this is what makes euphemism an R2
device.) So we talk, under R2 conditions, about making it, or doing number
two, or being hard up (in the economic sense). People holding professional
discussions, where R1 holds sway generally, will resort to technical terms
rather than euphemisms, while at polite cocktail parties, if one must talk
about the thing, one uses euphemisms, as one also uses other R2 devices in
these situations; contrast the sentences of (18), the R1 cases with those (19),
and imagine, first, an anthropologist at an anthropological meeting using
each of (18) and then a society matron at an elegant party saying (19).
[ 46 ] Context Counts
47
{
copulate
*do it }
they…
(b)
{
Defecation
*Making number 2 }
is generally expedited by the use of
large banana leaves, or old copies of the New York Daily News.
(c) Many of the residents of the ghetto are
{
underprivileged.
*hard up. }
(19) (a) I hear that the butler found Freddy and Marion
{
making it
?copulating }
in the pantry.
{
hard up.
*unprivileged. }
The point is that we respond differently emotionally to R1 and R2
devices, and hence the careful speaker will tailor his device to his purpose.
Rules 1 and 2 may be applicable together, as we have seen: avoidance of
nonfree goods may be interpreted both as a means of not imposing and
as a way of letting the addressee have his freedom. But Rule 1 and Rule 3
seem to be mutually contradictory: if they coexist in the same conversa-
tion, we must assume that, for any of various extra-linguistic reasons, the
participants are, really or conventionally, shifting their relationships with
each other. Rule 3 is the rule of politeness that seems the least “hypocriti-
cal,” although it, too, is very often used conventionally when there is no real
friendship felt. This is the rule producing a sense of camaraderie between
speaker and addressee. The ultimate effect is to make the addressee feel
good: that is, it produces a sense of equality between Sp and A, and (provid-
ing Sp is actually equal to or better than A) this makes A feel good. (But of
course, if Sp really is of lower rank than A, his invoking R3 will be seen as
“taking liberties,” and will result in the termination of the conversation on
an unsatisfactory basis.) Now it is true that R1 and R2 also are designed to
“make A feel good.” In fact, one might try to generalize and say that this
was the purpose of all the rules of politeness. But they all do it in differ-
ent ways, and R3 does it by making A feel wanted, feel like a friend. Rule 3
is the rule that produces tu in appropriate situations, in those languages that
use tu, when it is used to express solidarity. Here too, rather than the last
[ 48 ] Context Counts
49
Now these are incontrovertibly rude, in that they impose on the addressee
and destroy his options (he must recognize the speech act for what it is,
and must respond appropriately). Also, since they insist on the address-
ee’s appropriate response, they treat him not as a friend, but as an express
[ 50 ] Context Counts
51
unequal. So they violate Rules 2 and 3. But they are unambiguously clear,
too: they make it very plain what the speaker intends. So in this case, the
R1–R3, or clarity-politeness conflict, is resolved in favor of R1, where ordi-
narily it is resolved in favor of R3. Is this an exception? We must first ask
when these sentences may be used. Generally they are used in desperation—
when an ordinary speech act has previously been ignored. Now in this situ-
ation, politeness may be waived—the addressee, by ignoring the speaker
in the first place, has forfeited his rights. And further, the speaker has to
get his message across, and by now has good reason to believe only the
most forceful speech act will accomplish this. So these sentences are not
true exceptions, but rather show that our rules are in general correct: they
enable us to cope with apparent exceptions, and to show why they exist.
In conclusion, then, I hope to have shown the following:
REFERENCES
Brown, Roger, and Marguerite Ford. 1964. Address in American English. Language in
culture and society: A reader in linguistics and anthropology, ed. by Dell Hymes,
234–44. New York: Harper and Row.
Brown, Roger, and Albert Gilman. 1960. The pronouns of power and solidarity.
Style in language, ed. by Thomas A. Sebeok, 253–76. Cambridge, MA &
New York: MIT Press & John Wiley and Sons.
Chomsky, Noam. 1965. Aspects of the theory of syntax. Cambridge, MA: MIT Press.
Gordon, David, and George Lakoff. 1971. Conversational postulates. Papers from the
seventh regional meeting of the Chicago Linguistic Society (CLS 7). Chicago, 63–84.
Grice, H. Paul. (1967) 1975. Logic and conversation. Syntax and semantics 3: Speech
acts, ed. by Peter Cole and Jerry L. Morgan, 41–58. New York: Academic Press.
Lakoff, George. 1972. Hedges, fuzzy logic, and multiple meaning criteria. Papers from
the eighth regional meeting of the Chicago Linguistic Society (CLS 8). Chicago,
183–228.
Lakoff, Robin. 1969. Some reasons why there can’t be any some-any rule. Language
55.608–19.
Introduction to excerpts
from two 1974 papers
Pluralism in linguistics; Linguistic
theory and the real world
BY BIRCH MOONWOMON
I n the early 1970s Lakoff was writing critically about the newly established,
but already solid, generative syntax from the center of its academic cul-
ture. Long before there was an acknowledged need for an “extension” to the
standard theory, when the nonstandard challenge to the Aspects (Chomsky
1965) model was barely being developed, Lakoff was articulating criticism
of what did not even have to be called “standard,” since there was little
with which to contrast it. She was writing as a member of the generative
semanticist group. Something of particular interest in understanding the
development of her scholarship and the independence of her stance is that
she was interested in a socio-syntax, and quite willing to show this, among
a group of colleagues who would talk about supposed semantic universals,
but not social particulars.
Two shortcomings of the Aspects model especially get Lakoff’s atten-
tion. Criticism of these shows up in each of the articles reprinted
here: the privileging of formalism that (still) plagues generative gram-
mar; and the failure of the theoretical construct of autonomous syntax.
Lakoff acknowledges the importance of the profound transformation in
linguistic analysis that generative syntacticians’ work has meant, but
54
[ 54 ] Context Counts
55
that the theorist needs to know what deviations from an ideal com-
petence model are found in actual language use. Sociolinguistic infor-
mation affects linguistic form. Performance facts are not random, and
selection among variants performs specific functions. Lakoff proposes
that there be pragmatic rules with accompanying syntax and semantics
rules, all fuzzy.
The promotion of fuzzy rules, evoking Zadeh’s (1965) fuzzy categories
concept, is one of several things that dates these articles, more than do the
concern with overemphasis on formalism and dependence on autonomous
levels in grammatical analysis. One is Lakoff’s faith in presumably cog-
nitively based universal information categories for the encoding of prag-
matically relevant information. It seems now strangely naive to say that it
would not be a great burden to represent all the information that a speaker
has, that this would reduce to a few things like whether information is new
or old, the uniqueness of a subject, the goodness or badness of the topic.
She is depending on a universal grammar model in terms of which it is
proper to ask questions such as how many categories from the inventory of
universal ones are pertinent in a given language’s encoding. With pragmat-
ics, Lakoff hopes that a universal underlying structure can be described in
a non-ad hoc and non-English-dominated way.
Another now-debunked idea is that women’s language is the language
of the powerless. This concept was given voice by Lakoff more than by any
other 1970s student of women’s language use, and it had her in trouble
from the start. Her remarks about women’s language anticipate her posi-
tion in Language and woman’s place (1975). Here and there it is not clear
whether she has in mind simply a stereotype of women’s use of language
and—not the same thing—an accompanying assessment of that stereo-
type as representing powerless speech or language as actually used by
women and an accompanying devaluing of that. If the latter, one would
have to know by what kinds of women and in what kinds of speech situ-
ations such language is used. Whether Lakoff has in mind what (some)
women actually do, or what they are thought to do, it is the negative eval-
uation of such behavior that is a problem. Lakoff clearly does understand
that women’s ways of doing and being are liable to be undervalued. She
notes that the people who have been turned off to formalism in syntax
have been largely women who, because only formalism is valued, have
been told their talents are not needed. A quarter of a century later it turns
out not that women do not do syntax just like men—increasingly, women
do—but that work in sociolinguistics, discourse analysis, pragmatics, and
applied linguistics is dominated by women. Lakoff as a pragmaticist has
a history of passing through the deep woods of early generative theory
I N T R O D U C T I O N T O E X C E R P T S F R O M T W O 1 9 74 PA P E R S [ 55 ]
56
REFERENCES
[ 56 ] Context Counts
57
CHAPTER 3
PLURALISM IN LINGUISTICS
In the days when the white engineers were disputing the attributes of the feeder
system that was to be, one of them came to Enzian of Bleicheröde and said, “We
cannot agree on the chamber pressure. Our calculations show that a working
pressure of 40 atü would be the most desirable. But all the data we know of are
grouped around a value of only some 10 atü.”
“Then clearly,” replied the Nguarorerue, “you must listen to the data.”
“But that would not be the most perfect or efficient value,” protested the
German.
“Proud man,” said the Nguarorerue, “what are these data, if not direct revela-
tion? Where have they come from if not from the Rocket which is to be? How
do you presume to compare a number you have only derived on paper with a
[Because these papers contain extensive overlap, material original to each has been
collated here. Deletions are indicated with ellipses, and the footnotes renumbered. The
references for each paper have been combined and edited, and appear at the end of the
chapter.]
This paper first appeared in Berkeley Studies in Syntax and Semantics (1974), vol. 1, XIV
1–29. Berkeley: Department of Linguistics, University of California. Research underly-
ing this paper was partially supported by the NSF under grant Number NSF GS-38476.
58
number that is the Rocket’s own? Avoid pride, and design to some compromise
value.”
Thomas Pynchon, Gravity’s Rainbow ([1973] 1987) p. 314–315
[ 58 ] Context Counts
59
important; solve them and you will understand language, you will have
a full-fledged theory. The idea that a theory might stand or fall on its
formal assumptions alone is what I consider the most dangerous heresy
within the field.1
Ill-advised, maybe, you say; boring, often; but dangerous? Yes, danger-
ous in the sense that it has led transformational grammar down the gar-
den path many times, to bad analyses and untenable generalizations, and
into theoretical positions that are decidedly unhealthy for anyone who
wants to understand how language really works—as opposed to some-
one who wants to study mathematically pretty notation. I assume that
linguistics is to be done by those in the first group, and mathematics by
the second; and I therefore assume that an emphasis on formalism for
its own sake is deleterious to the development of an adequate theory of
language use.
There has, thus, over the years, been a great deal of too-glib generaliza-
tion based on the doctrine of the neatest formalism. The idea is that if you
can formalize an analysis, it’s right, and if you can’t either it’s wrong or it
isn’t within our provenance as linguistic theorists… .
[A]pparently beautiful formalization[s] turn out to be inadequate in
many ways. It is true that no perfect solution of the passive problem has
been forthcoming from any quarter, but at least it seems more honest to
admit that your tentative solution is interesting but unworkable (as has
been done by Postal and others within generative semantics who have
made proposals concerning passivization), rather than putting a superfi-
cially attractive solution up as a panacea without regard for the data that
contradict it.
I said too that overdependence on formalization might have the effect
of discouraging research from being done that, though incapable of per-
fect formalization at present, nevertheless could be fruitful, insightful, and
good for linguistics through attracting new recruits.
I refer to two kinds of cases: first, work within the domain, as it is gen-
erally envisioned at present, of “linguistics proper”; then, work that goes
beyond the boundaries of our field. In both cases, of course, what fre-
quently happens is that first the informal work is done, pointing out the
1. I am not saying that interest in formalism per se is destructive; much less that we
should abandon theory and formal rigor in favor of purely informal description and
nontheoretical data collection. Certainly anyone would agree that a theoretical view-
point is necessary, at the very least, to show us what data are interesting and how to
categorize the facts: it is unarguable that formalism is very useful, properly used, in
determining the accuracy of our observations. I am inveighing here only against blind
dependence on formalism for its own sake.
nature of the problem and why it is of interest; formal work follows upon
this, some time later. But if we insist on formalization at the outset, much
preliminary investigation will be squelched, and we will be cut off from
doing some of the most interesting potential research.
Consider work by linguists such as Georgia Green, Deborah James, Ann
Borkin, and John Lawler.2 These scholars were among the first to point out
that classical transformational grammar was formally incapable of dealing
with many areas of language use: types of speech acts, interjections, polar-
ity assignment, and genericity, to name a few. It is sometimes held by more
orthodox transformationalists that if a solution can’t be formalized, the
problem is of no interest to linguistic theory; but this sort of work shows that
nothing is farther from the truth. The very fact that relationships existed
that were incapable of formalization was sufficient to show that the formal-
isms thought by orthodox transformational theory to be adequate for the
complete description of language were not so by any means. But the facts had
to be presented first; had we looked only at what our formal apparatus could
handle, we would never have discovered the inadequacy of that apparatus.
Now what these and other investigators have found, the more doggedly
they tried to account for the distributions they discovered, was, first, that
syntactic criteria alone could not be used to account for the data; then, that
semantic criteria were not sufficient either; and most recently that, if we are
going to talk about the occurrence of sentence-types, we must extend our
vision beyond even what is normally considered “linguistics” proper, and
erase some of the boundaries that have been imposed on our domain: we
must become engaged in research that we might be tempted to call “psy-
chology” or “psychiatry,” or anthropology,” or “sociology” or “literary crit-
icism”—anything in fact but linguistics. Yet if we are going to be serious
linguists, we must agree that our field is part of theirs (and vice-versa).
However, if formalization within the accepted bounds of linguistics is
unlikely at present, complete formalization within these broader defini-
tions of the field is, presently, impossible. This is not to say that it will never
be possible, but only that its realization demands much more knowledge
than we have available at present. Yet we will never be able to integrate our
knowledge about linguistics with these various other fields and arrive at a
formal system incorporating concepts of the various branches of human
knowledge, unless we first discover what sorts of facts we will have to deal
with: precisely where, and how, linguistic data interact with other kinds
of information. So once again, close examination of the facts comes first,
2. For instance, Borkin 1971, Green (1972) 1975, James 1972, and Lawler 1972.
[ 60 ] Context Counts
61
and only after we have dealt informally with the problems for some time
will a means of formalization of what we have found out be likely to arise.
And even here, we ought not to expect a formal system for the integration
of all human knowledge to arise one day full-fledged and complete, from
someone’s mind. The process of devising formalisms will be long and often
frustrating, and we will have to be content with only partial and rather
unsatisfactory representations for some time to come. It will be a process
of trying a formalism, seeing how it deals with the facts, modifying or
discarding—and starting anew. But always we must match formalism to
the facts, which come first—not the other way round.
In these days of declining enrollments and student interest turning away
from pure research to real-world relevance, it behooves us to consider the rel-
evance of linguistics to the real world. And the interaction of pure language
phenomena with these other fields is precisely where linguistics can be rel-
evant, interesting, and fruitful. The majority of students are not apt to be
turned on by Boolean conditions on analyzability; rather, they find it interest-
ing to ask how one’s mental state and status in society are reflected in one’s
use of language. This is where we ought to be exploring; but this is also where
we have no rigorous methodology, and hence, where the traditional transfor-
mational linguist tells us to stay away from, since it isn’t “linguistics.” This has,
in my view, always been poor advice, but now it is becoming utterly dangerous.
Let me talk a little about the kinds of questions I think it is time we
started addressing ourselves to, as linguists: questions about the intent
and total effect of communication—both purely linguistic, and, where nec-
essary, paralinguistic as well: gestures, intonation patterns, posture, and so
on—rather than merely its superficial purely linguistic form:
1. The use of language by minorities, or, more generally, groups who don’t
set the dominant style. Among such groups, we might identify:
a. blacks
b. Chicanos
c. lower and higher (than middle-class) economic groups
d. women
e. academics
f. children
g. hippies
h. militants of various persuasions
This is just a partial list, but I think everyone will agree that each of these
groups possesses a language pattern that stamps it as not within the English
[ 62 ] Context Counts
63
no language and no dialect is more logical than any other, as long as both
retain the ability to express what the speakers of the dialect in question wish
to express. It is perfectly possible that speakers of Black English want to
express different ideas than do speakers of the standard language, and that
this is reflected in their language. But obviously it is impossible and stupid
to make value judgments on this basis. The same may be said of women’s
language, and the language of other non-influential groups. The speakers
of these sublanguages should know that they may be misjudged according
to the way they speak; they should know the stereotype that exists in the
mind of the public of speakers who speak this dialect, and decide if they
wish to fit this image or not. (Gloria Steinem obviously does not want to
fit the image of someone who speaks women’s language; Pat Nixon, on the
other hand, might, faced with all the knowledge we could give her, make
quite the opposite choice.) At the same time, we must educate everyone to
realize that there exists the possibility of alternative modes for expressing
similar ideas, and that none is by fiat “better” than any others. This is not
a simple task, but it will be greatly facilitated by an understanding of the
data, hopefully along with both formal and informal explanations.
understand what we mean when we talk about their interaction with the
already formalizable rules of syntax.
3. Pathological and aberrant language. Here I am referring to the special lin-
guistic usages of people in one or another abnormal mental state. I am thinking
most specifically of the language of schizophrenics, but we might also include
aphasia and perhaps also the language characteristics of mysticism and trance-
states. I don’t think there’s much virtue in trying to construct a schizophrenic
grammar; rather, the task before us here is to categorize the ways in which
schizophrenics use language in ways that psychiatrists can use in making
their diagnoses, and then see how they deviate from the “normal” grammar.
What I suggest we will find is that the schizophrenic’s grammar (up till a very
advanced state of decay, perhaps more due to the confines of institutionaliza-
tion than to any inherent mental deficiency) is perfectly all right, but that his
world-view is quite different from the norm. (Laing3 would have us question
the reasonableness of the norm.) And since, as we have seen, normal grammar
is strongly dependent on normal world-view, we should expect to find devi-
ations in the schizophrenic’s use of language that mirror the abnormal way
in which he perceives his world. We might ultimately be able to use language
behavior as a fairly precise diagnostic: if trait x is there, the schizophrenic is
thus-and-so far advanced toward normality, and so on. But we need to view
the aberrancy from norms as a point on a continuum: it isn’t that everybody
who is accounted “sane” speaks one way, and everyone who is “crazy” speaks
an entirely different way. Rather there is an infinity of points along the line,
and some of us fit in at one point, others at others. And certainly we would put
Thomas Pynchon, say, at a different point in our line than we would put Henry
James, but this does not, again, mean that we are making value judgments or
casting aspersions on anyone’s sanity or creativity. It is, however, the business
of the linguistic theorist to try to figure out what the possible sorts of world-
view variations are, and how they are connected to different forms of linguistic
expression. Just as Chomsky showed in Syntactic structures (1957) that it was
impossible to define a grammar without reference to ungrammatical struc-
tures, so it is impossible for the linguist to talk about his real-world-relevant
grammar without referring to alternative views of the real world, normal and
aberrant, possible and impossible. Formalizable or not at present, this infor-
mation must be considered a part of the baggage of linguistic theory.
3. Cf. Laing 1967.
[ 64 ] Context Counts
65
make contributions—if we train them right and don’t turn them off. I feel
that one of the kinds of damage that has been done in the last dozen years
by the overemphasis on formal description of language for its own sake is
the very business of discouraging people who weren’t interested in formal-
ism, weren’t interested in pure syntax, were interested in “relevance.” For
whatever reason, we tended to sneer at such people, on the grounds that
whatever they were doing, it wasn’t linguistics. And no one will stay in a
field after you’ve told him that he isn’t capable of working on those ques-
tions that the field in its wisdom considers germane, and that the questions
that interest him are irrelevant or trivial.
I am also aware that many of those who have in the past been turned off
by undue obeisance to formalism have been women. It is a well-known and
ill-explained fact that among undergraduate linguistic majors, more than half
are women; among graduate students in their first few years, about half in
most schools; and then, as we get closer to the Ph.D., the number of women
inexplicably diminishes, so that many fewer women than men actually get
that degree, and the number of women teaching in what you might consider
respectable places is a tiny proportion of the total. Something funny has hap-
pened. You can get a lot of different answers, but I feel that it is the emphasis on
formal descriptions of the superficial aspects of language that many of us find
discouraging: hence many women, in an attempt to escape into relevance, go
into psycholinguistics and sociolinguistics and TEFL, but are lost to the main-
stream and often end up dissatisfied anyhow. I don’t know, nor does anyone,
whether there is an inherent indisposition toward formalism among women,
or whether it is a learned trait that may eventually be overcome; I know merely
that it is the case now and is apt to remain so for some time to come. I think it
is criminal to attract people into a field and then waste their abilities and insult
their intelligence by telling them there is no place for the talents they have.
Don’t tell me this isn’t done—I’ve been there. And when jobs are given out,
we all know that no one these days is such a male chauvinist pig as to refuse to
think about hiring women—it’s just that the prestigious kind of work, the sort
of specialization the department wants to acquire, is that typically possessed
by men, not women. The only way to equalize things in the long run, rather
than making token appointments that do no lasting good, is to broaden the
field’s view of itself, to agree that many things can be done equally respectably
within a linguistics department. But we must educate people, first, to be able
to do this kind of work, and second, to appreciate it. …
[ 66 ] Context Counts
67
This paper first appeared in Berkeley Studies in Syntax and Semantics (1974), vol. 1,
XVIII 1–53. Berkeley: Department of Linguistics, University of California.
1. A very interesting discussion of some of these problems is to be found in Larkin &
O’Malley 1973.
2. This paper was prepared for presentation at the 1974 Annual Conference of the
B.C. Association of Teachers of English as an Additional Language and for the 1974
Convention of the Teachers of English to Speaker of Other Languages. Research for
this paper was partially supported by the National Science Foundation, under Grant
NSF GS-38476.
L I N G U I S T I C T H E OR Y A N D T H E R E A L W OR L D [ 67 ]
68
[ 68 ] Context Counts
69
L I N G U I S T I C T H E OR Y A N D T H E R E A L W OR L D [ 69 ]
70
[ 70 ] Context Counts
71
L I N G U I S T I C T H E OR Y A N D T H E R E A L W OR L D [ 71 ]
72
[ 72 ] Context Counts
73
But the notion that enlightenment was a two-way street—that each could
learn from the other—seldom occurred to anyone, least of all to the theo-
rists, whose theories as a result tend to get more arcane and less relevant
to the real-world tasks at hand with each passing revision.
We can look at an illustration of what I mean: a theory developed pretty
much in isolation, by theorists trained not in language learning as their
principal interest, but in philosophy, information science, mathematics—
disciplines remote from real-world intervention. And they devised a very
elegant system of language description, most attractive, very compelling—as
long as you didn’t look too hard at the facts you were intuitively aware of.
Intuition was scorned: if you couldn’t formalize a relationship, if you couldn’t
write rules, make it submit to the grammatical theory that supposedly was
a tool for the understanding of language, then you’d do best to forget the
whole thing. And this is just why the system was, superficially, so very ele-
gant: the nasty complicated fuzzy parts, the parts that did not fit neatly into
the scheme, were assumed not to exist, or at least, not to be of interest to the
linguist. Rugs in that part of the world got very bulgy.
I an not inventing a parable or a fairy tale. I am talking about transforma-
tional grammar, as should be no surprise to you. I think the sins committed,
the confusion perpetrated in its name have been legion, over the past dozen
years, and I think it’s precisely because all the input came from the theorists,
most of whom had never taught languages, or were ashamed to admit it if
they had, particularly if comparison among the languages they knew revealed
data that could not be reduced to the neat form prescribed by orthodox trans-
formational theory.
Actually, I am being perhaps nastier to transformational theory than
I ought. There is a time in the life of every theory4 when oversimplifica-
tion is not only inevitable, but necessary for the development of greater
understanding. In order to be able to get an allover picture of the sort that
a theory of language must be, one must first oversimplify, pretend that the
facts are more amenable to organization than, actually, they are. Ideally, it
is better even at this stage not to fool oneself; to admit privately anyway
that the treatment gains its elegance through the overlooking of data that
will have to be dealt with some day. The ostrich-treatment, in which one
declares in effect, “My nice elegant theory can’t handle this phenomenon,
so I hereby declare it nonlinguistic and thus keep my theory correct and
neat,” is less desirable, as should be obvious, because it will prevent any
developments or improvements in the theory, ultimately rendering it too
L I N G U I S T I C T H E OR Y A N D T H E R E A L W OR L D [ 73 ]
74
rigid to be a useful heuristic tool—which in the end is the very least as well
as the most we ask of a theory. …
It should be clear that … a theory [which incorporates the full com-
plexity of language use] is far more complex, far harder both to formulate
and to understand, far less immediately elegant than was classical trans-
formational grammar. But I hope I have already made it clear that those
losses in beauty are necessary, if the truth is to be approached at all. So
as I detail a few of the assumptions of the beginnings of a theory that
will, if it is ever fully formulated, embody all those desirable character-
istics, I would hope the reader will bear in mind that its awkwardnesses
are necessary—some because of its relative youth, others because of the
scope of its ambitions.
Suppose first of all we question the most sacrosanct assumption of clas-
sical (Aspects-vintage) transformational grammar. Is there a demonstrated
need for a syntactically relevant level of deep structure, and therefore a
level of autonomous syntax, with a semantic component not directly relat-
ing to the syntax, but merely interpreting its output? For obviously if it
cannot be demonstrated that deep structure is necessary, it must be dis-
carded. Its necessity could be proved by showing that there existed at least
one generalization in one language that could be expressed by rules assum-
ing a deep structure, and not without. To date, despite strenuous efforts by
its adherents, no such case has been found. Indeed, cases have been discov-
ered that seem to suggest the very opposite: if we assume a level of deep
structure, there are explanations for phenomena we become incapable of
making.5…
This is the basis of the theory that has been called generative seman-
tics. Very simply, there is no separation of levels: a single, highly abstract,
underlying structure underlies the semantics, the syntax, and the phonol-
ogy, and further, syntactic information may be used in the statement of
phonological or semantic rules, and conversely. (Examples showing the
necessity of this are given in G. Lakoff 1970.)
Such a theory allows, in fact forces, us to take real-world contextual
information into account in determining whether a sentence is well-
formed. Real-world-linked information that a speaker uses to determine
whether a sentence may be grammatically or appropriately uttered in a
given context is, in such a theory, available to the speaker as part of his
grammatical information.
[ 74 ] Context Counts
75
L I N G U I S T I C T H E OR Y A N D T H E R E A L W OR L D [ 75 ]
76
[ 76 ] Context Counts
77
L I N G U I S T I C T H E OR Y A N D T H E R E A L W OR L D [ 77 ]
78
So we see that the theoretical linguist must deal with problems of the
intellect and morality, with reality and sanity. In order to do linguistic the-
ory at all adequately or interestingly, he must come to grips with the effect
of language in the world in which it is used, with the intent and effect of
communication. The applied linguist must concern himself with decisions
among possible theories, universals of grammar, relations among gram-
matical systems. The differences between the two types of linguists are fast
becoming less interesting than their similarities.
REFERENCES
[ 78 ] Context Counts
79
L I N G U I S T I C T H E OR Y A N D T H E R E A L W OR L D [ 79 ]
80
81
[ 82 ] Context Counts
83
REFERENCES
YO U S AY W H AT YO U A R E [ 83 ]
84
85
CHAPTER 4
[ 86 ] Context Counts
87
(1) I wrote my grandmother a letter yesterday and six men can fit in
the back seat of a Ford.
(Gleitman 1965:262, ex. 11)
That is one obvious case of a sentence that is fully grammatical, that is,
linguistically unexceptionable; and yet, if we are judging acceptability—that
is, the probability of such a sentence being uttered, or the number of con-
ceivable real-world circumstances or the normality of the real-world cir-
cumstances in which this sentence is apt to be used—we find a different
judgment pertains, and we must rank this sentence low on an acceptability
hierarchy.
On the other hand there are sentences which, taken out of context, are
bizarre; some even violate selectional restrictions, such as Morgan’s cel-
ebrated example:
A sentence like (2) would have been asterisked in any classical transfor-
mational discussion, as it violates certain selectional restrictions. Yet sen-
tences such as this are frequent in ordinary speech and hence would rank
high on any rational acceptability hierarchy. So there are sentences that are
ungrammatical but of relatively high acceptability in context. Thus, given
the correct set of social, situational, and linguistic contexts, both (1) and
(2) might be considered good sentences of English. The difference between
the two in this regard is that, while one can imagine a reasonably high
number of contexts—and “plausible” contexts, too—in which (2) might be
uttered, the same cannot be done for (1). So, defining acceptability in this
way, the “ungrammatical” (2) is a “better” or more acceptable sentence than
the “grammatical” (1).
Another way to view the grammaticality/acceptability distinction is
to say that grammaticality is a special case of acceptability. A sentence
is grammatical if it is acceptable according to purely linguistic criteria.
Grammaticality is acceptability shorn of social and psychological differen-
tiations. Then it seems fairly apparent that grammaticality is a very highly
specialized and not terribly useful concept, outside the realm of strictly
autonomous syntax. As soon as we concur that autonomous syntax is not a
viable level of analysis (as various works written in the last ten years have,
I feel, conclusively proved), we see that a separate notion of grammaticality
is neither necessary nor possible within a coherent linguistic theory. But if
[ 88 ] Context Counts
89
[ 90 ] Context Counts
91
Then we may say that three basic trends characterize women’s language
as a deviation from the standard:
1. Nondirectness: e.g., 5, 6, 7
2. Emotional expression: e.g., 2, 3, 6, and 7
3. Conservatism: e.g., 4 and 6
Point (1) does not figure in this summary, and indeed it is rather mislead-
ing to categorize “special vocabulary” as idiosyncratic to women’s language.
The deeper point to be made here is that every subculture has its own
vocabulary and that vocabulary will involve terms that are of specific use
to the particular culture in question. Whorf, of course, was the first to raise
this issue when he pointed out that Eskimo has six words for “snow” where
English has just one—presumably because snow is much more important
in the Eskimos’ lives than it is in ours, and minute differences in the quality
and quantity of snow are for them of crucial importance, as they are not for
us. So they need precise words to make these crucial distinctions, as we do
not. As I have said elsewhere, this is undoubtedly the reason for the more
precise color-discrimination vocabulary among women: because it has
traditionally been considered important for a woman to possess this sort
of expertise, for fashion and interior decoration have both been women’s
work. Of course, men have their own special highly developed vocabularies,
[ 92 ] Context Counts
93
interpreted it as (5):
(5) None of the boys like certain girls: namely, Mary, Alice, Nancy… .
Whereas, faced with the same sentence, all English-speaking women inter-
preted it as (6):
But, as I say, I know of no such cases. Rather, even where syntactic rules
such as question formation are involved, the difference between the dia-
lects lies in the fact that in one a sentence is usable in more social or psy-
chological situations than in the other. So we cannot define the two dialects
in terms of purely linguistic autonomous-syntactic distinctions. Additional
reason for this belief is the fact that even if a woman will not or cannot
use certain forms that are not parts of traditional women’s language, she
[ 94 ] Context Counts
95
from that norm was censurable. I recall that Paul Goodman—no doubt in
good company—made this mistake some years back in an article in the
New York Review of Books (1964), in which he criticized Chomsky’s lin-
guistic work on the grounds that—by distinguishing sentences that were
grammatical from those which were not—he was squashing linguistic
creativity and innovation. But the distinction between description and
prescription is still blurrier, even among linguists, than we might like, par-
ticularly in areas where we have been brought up to make value judgments
before we learned to be disinterested academic observers. The question of
gender-related roles is one of these highly charged areas, and it therefore
behooves us in discussing acceptability as a factor in understanding wom-
en’s language to bear in mind that there is a danger that we will confuse
linguistic norms with social values.
There are other confusions to be avoided. The notion of acceptability,
I have said above, implies a standard against which a speech act may be
judged. It has also been pointed out that, in talking about acceptability as
opposed to grammaticality, that standard is grounded in social and psy-
chological context: an act of speech or behavior is judged acceptable in a
specific context. Now, it would seem at first glance that men and women,
being these days participants in the same activities, similarly educated, at
least superficially raised alike, would share this set of contexts. A male and
a female, participating in a specific kind of behavior, linguistic or other-
wise, would perceive the context in which the behavior was to take place
similarly. But if acceptability implies appropriateness within a particular
social-psychological setting, and as we have suggested men’s and women’s
languages differ somewhat in terms of what is acceptable, then we are faced
with a paradox.
We must, rather, assume, I think, that a given context is interpreted one
way by a male speaker, another by a female. Actually, it will be recalled we
are dealing with a complex hierarchy of acceptability, and it is in principle
not at all unlikely that different speakers will arrange their worlds in quite
different ways. It is not as though every male speaker of English defines
Contexts A-L , let us say, as “situations requiring directness” and M-Z as
“requiring nondirectness” while all women interpret situations A-P in the
former way, Q-Z in the latter. We would prefer to say—continuing for the
sake of clarity to look at the division of social contexts in this extremely
simplistic way for the moment—that women would tend to interpret
situations as requiring nondirectness until further down in the alphabet
than men typically would. But all sorts of variations are conceivable. That
is, what I mean by saying that traditionally women’s language has tended
toward nondirectness is that women will interpret a greater number of
[ 96 ] Context Counts
97
REFERENCES
[ 98 ] Context Counts
99
[ 100 ] Context Counts
101
CHAPTER 5
This paper originally appeared in Language, Sex and Gender: Does La Différence Make
a Difference?, edited by Judith Orasanu, Mariam K. Slater, and Leonore Loeb Adler,
53–78. Annals of the New York Academy of Sciences 327. New York: New York Academy
of Sciences, 1979. Reprinted here with permission of the publisher.
Acknowledgments: The preparation of this paper has profited from my discussion of
its contents with many people, as well as numerous opportunities to present its ideas
in various stages, to many groups. In particular I would like to thank Linda Coleman
for her very helpful suggestions leading to the final form of the stylistic continuum as
it appears here; and Deborah Tannen for innumerable valuable improvements in both
style and content.
102
[ 102 ] Context Counts
103
where the use of the future tense conflicts with the use of yesterday, refer-
ring to past time, or:
[ 104 ] Context Counts
105
(A) Someone who dresses conservatively and very carefully can be expected
to use language carefully and equally precisely.
(B) Someone who divides humanity into black and white also, cog-
nitively, divides ideas into right and wrong, with no gray areas in
either case. Again, there are people who violate these expectations,
but they are experienced as aberrant. If the aberration is clearly
purposeful, for effect, it is comic: much of Woody Allen’s humor,
for example, depends on incompatibilities of stylistic modalities.
But if it seems unintentional, we consider it indicative of serious
psychopathology—schizophrenia again.
1. The work of Paul Ekman is particularly interesting in this regard, indicating that
subjects, when forced to decide between conflicting interpretations of a contribution
based on facial expression and verbal statement, will opt for the truth of the former.
But to say this is to assume that not only do these sophisticated interpret-
ers of discourse (i.e., you and I and everyone else) employ their implicit
rules in the interpretation of linguistic structures (as we have been per-
fectly well aware for quite some time) but that we use the same system
for what is sometimes called extralinguistic or paralinguistic functioning,
and we detect mismatches by our application of these rules. Woe betide
anyone caught in a mismatch (Richard Nixon is a notorious example).
I take it, then, as proved that nonlinguistic style is as much rule-governed
behavior as is linguistic style, and I would further venture—it is not
yet proved by any means—that the same kinds of rules, the same basic
devices, are operative in all human behavioral systems. Style is a unity,
not only in its superficial manifestations in the normal individual but
also in its basic mechanisms.
It is therefore incumbent on the linguist determined to give evidence
for this last contention to show that those traits ascribed to the linguis-
tic grammar can be justified and indeed are essential as parts of a wider
stylistic grammar. If this should be provable, we can be content in that we
have given yet another instance of the working of Occam’s Razor: After all,
doesn’t it make sense that the human mind should function with as little
baggage as it can get away with? If it can be shown to need only one general
rule system, from which all forms of psychic behavior can be derived, isn’t
this preferable to a theory that requires separate theories, and thus sepa-
rate rules, for each subtype of behavior? But even if the former is ideal, it
naturally remains to be proved. I do not intend to make a watertight case
for psychic economy here, but merely to give some preliminary arguments
showing that a unified system is tenable. I use linguistic theory as my tem-
plate, and more specifically a kind of generalized theory derived from vari-
ous realizations of generative grammar. The reason I do so is partly that
I am trained as a linguist, so it comes naturally to me, but also because it is
the only attempt at formalization of psychic structure that I know, the only
attempt, I should say, at a predictive rather than a merely taxonomic or
classificatory framework, such as are found in many psychological theories
of character.
To return to the point: linguistic theory assumes a bipartite syntax
(here I am oversimplifying considerably, but the complexities are not
important for our purposes). In such a syntactic model, two levels of syn-
tactic structure are relevant: a superficial level, which is the utterance as it
appears in spoken or written form, and the underlying or deep structure,
which is related to the meaning of the sentence, and contains in a form
accessible to inspection all the elements needed to account for the mean-
ing of the sentence. Since every sentence in a language that is found in
[ 106 ] Context Counts
107
3. I am oversimplifying greatly here, and probably distorting the facts more than
I ought. The basic premise is true, but the discussion of the intricacies involved in the
differences of meaning and usage in the members of each pair, however valid, would
take us far afield.
[ 108 ] Context Counts
109
also indicate what does not exist, in normal persons on the one hand, and
in general, on the other.
With this definition in mind, let us turn to the examination of stylistic
behavior. I have said that, to be a style, behavior must be able to be seen
as both coherent across modalities and consistent across time, and that
one form of stylistic pathology may lie in the failure to be either coher-
ent or consistent, or both. Further, a style assumes a match between dual
levels of structure—the one that is superficially accessible to participants
in an interaction, and the other, the level of intention, which itself may
be multileveled and accessible not at all or in varying degrees to partici-
pants in their interactions. Also, participants generally assume a one-to-
one relationship between intention and execution: if I perform Behavior
A, my interlocutor is apt to assume that I have intended to communicate
B. There is some possibility for ambiguity, even as there is in language,
but as with language, in actual practice the context, for normal individu-
als, tends to disambiguate most potential ambiguities. So, for instance, if
a female participant does something, others will interpret it one way; if a
male, another—since “female” and “male” are different contexts. As also
with language, some users are much better at realizing that two readings
might exist, and determining which is the one that is correct in that con-
text. And the fact that many people apparently are not capable of doing this
efficiently causes great difficulty in interpersonal relationships—difficulty
that is particularly troublesome because participants are seldom aware of
its cause, or even of its precise form.
While Shapiro described one contributor to style— the nature and
degree of attentiveness—style has many other potential points of differ-
entiation. We could think of it as a grid, with Shapiro’s distinctions being
one column of many. This makes sense, in that we do not typically con-
sider a person completely classified when we have labeled him or her, as,
say, hysterical or obsessive-compulsive. We also define such an individual
as demure or assertive; as reticent or loquacious; and so forth ad infini-
tum. Another point of differentiation is the mode of rapport: the relative
importance for the individual of making the content of the communication
clear versus establishing a personal relationship with others, where this
choice must be made; and, where the latter is determined to be important,
how rapport is to be effected, what mode of presentation of self the indi-
vidual chooses to adopt. There are obvious correlations between Shapiro’s
attentiveness-oriented and my rapport-oriented categorizations, so that
the grammar should allow us to predict one category of behavior from
another: e.g., someone who tends toward diffuse attentiveness is apt to
display rapport-related behavior. But on the other hand, a person whose
[ 110 ] Context Counts
111
characteristics; (12), to, let us say, a hysteric—we are not quite sure what
to make of the behavior, it is not what we expect, but we can learn to make
allowances; (13), to a schizophrenic, whose behavior may be totally uninter-
pretable, and frightening to us as the hysteric’s is not, for just that reason.
Further, just as with linguistic rules, we have an implicit notion of sty-
listic rules and recognize when they are violated, even though we usually
are at a loss to say what it is that we recognize. Perhaps the most strik-
ing instance of this unconscious recognition of character co-occurrence
is found in literature: in the novel, and perhaps most crucially in drama.
What distinguishes a good novelist or playwright, as well as a good actor or
director, perhaps above all else, is the creation of credible characters. And
what we mean by “credible” has largely to do with the co-occurrence of sty-
listic traits, traits drawn from the repertoire of the personality as a whole.
A writer, a director, and an actor probably perceive this intuitive coherence
in different ways, or at least make different use of their intuitions: but what
distinguishes the great from the second-rate is an almost uncanny sense of
which combinations are plausible, which not.
These observations that I have made on style can be described in terms
of a theory of communicative competence such as I have discussed in
somewhat different form elsewhere (1977). What is relevant to the pres-
ent discussion is the part of the system that describes how participants
determine the appropriate mode of presentation of self in discourse with
others, on the basis of their own personal habits, the relationship between
participants, and the subject matter under discussion. On the basis of
this intuitive judgment, the speaker selects— fluent speakers, gener-
ally unconsciously—a matching strategy, dictating the point on the scale
shown in Figure 5.1 that is deemed appropriate by the speaker’s subcul-
ture for the type of interaction in which the speaker assesses himself or
herself to be engaged. It is the Rules of Communicative Competence that
enable speakers to make judgments as to the use and interpretation of sen-
tences. And, like the latter, the implicit Rules I am describing here func-
tion as models both for production and interpretation of the contributions
of others, so that—a significant point, as we shall see later—one under-
stands the contributions of others only in terms of one’s own internalized
strategies. But it is important to bear in mind that the named points on
the continuum represent strategies, or modalities, of interaction, rather
than Rules: the rules themselves, then, can be seen as metastrategies. The
modalities themselves, and the relationships among them, are universal;
but the statement of the Rules about which point is appropriate in a par-
ticular context—as well as the decision as to how the context itself is to be
interpreted—differs from culture to culture as well as from individual to
[ 112 ] Context Counts
113
Figure 5.1
[ 114 ] Context Counts
115
offer autonomy. Then the addressee is often faced with the impossible task of
determining which of the levels of communication to recognize—the overt
offer, or the covert injunction? Either way is risky. Where Distance avoids
uncomfortable topics by the pretense of intellectual uninvolvement, that
is, by denying the emotive force of the contribution, Deference denies the
cognitive content, conventionally of course (or else understanding could not
take place), saying in effect, “It’s up to you to translate my message: if you
decide we’re talking about THAT, we are.” Hence euphemism is characteristic
of this modality.
Camaraderie, necessitating as it does direct confrontation, is the
modality least in accord with what we usually think of as “politeness.”
For Camaraderie explicitly acknowledges that a relationship exists and is
important, whether one of friendliness or of hostility. Camaraderie is the
level of direct expression of orders and desires, colloquialism and slang,
first names and nicknames—much that is considered good and typical con-
temporary American behavior.
The strategies have been described as points on a continuum partly
because they can be used in isolation or in combination: a contribution can
be pure Distance, or mostly Distance, with a little Deference, or half of each.
Some people are very formal, distant, and aloof; others less so; still others,
hardly at all. Figure 5.1, however, makes some combinational possibilities
hard to visualize: a situation such as that exemplified in Figure 5.2 is no
problem—a style falling between Distance and Deference, which are contig-
uous; but this diagram would make it impossible to represent, say, a combi-
nation of Distance and Camaraderie, which is just as theoretically possible.
One way of avoiding these difficulties is shown in Figure 5.3, with
the left–right axis maintained intact, but some means of indicating that
combinations between any of the modalities are possible. We still can-
not graphically represent combinations of three and four modalities, but
Figure 5.3 is at least a step toward more accurate representation. A specific
linguistic entity—whether a lexical item, an intonation pattern, or a syn-
tactic formula—is typically categorizable as assigned to one of the named
points on our diagram, while the overall stylistic behavior on the part of
individuals can fall at nodal points or between them. So Figure 5.1 can still
be thought of as an appropriate model for the subcomponents of which
style is composed, but Figure 5.3 seems a better representation of a style
as a whole.
Figure 5.2
[ 116 ] Context Counts
117
DISTANCE
CLARITY CAMARADERIE
DEFERENCE
Figure 5.3
she is talking about (nor does she), but that she is very concerned with
whomever she is talking to, concerned with whether he is interested in
her, and whether his needs are being met. Here we have the reverse of
Gable/Cronkite style: Clarity is entirely absent, and there is no evidence
of Distancing behavior. In contrast to Gable’s characteristic poker face, we
have Monroe either smiling or looking sensuous, but certainly wearing an
identifiable facial expression. She uses interjections and hedges freely and
her dialogue is sprinkled with I guess and kinda in distinction to Gable’s
unembellished yups. Her sentences seem not to end, but rather to be ellipti-
cal, as if in invitation to the addressee to finish them for her—classic femi-
nine deference.
These diverse styles are also classifiable on Shapiro’s scale—Gable/
Cronkite’s as representative of obsessive-compulsive style, Monroe’s of
what Shapiro terms hysterical. More accurately, we would want to place
truly neurotic examples of stylistic behavior farther along our lines than
any of these more or less normal examples, but certainly it is arguable that
this culture’s paradigmatic masculine style shares a great many traits with
a “neurotic” obsessive-compulsive style; and our prototypical example of
femininity is exemplary in many of its aspects of neurotic hysterical style.
Shapiro shows too how the communicative behaviors I have just discussed
correlate with cognitive, perceptual, and motoric forms of functioning to
produce the coherency we call personal style.
Now we are faced with some interesting observations. I have remarked
that those linguistic traits that are characteristic of men in our society cor-
relate with a form of functioning that verges on the obsessive-compulsive;
and, feminine traits correlate with hysterical behavior. It is a truism among
psychological theorists that hysterical traits of character are predomi-
nantly found in women, so that it has sometimes been thought they were
exclusively the province of women. And similarly, though less strikingly, a
correlation has been claimed between obsessive-compulsive style and mas-
culine behavior.
Students of style in general, and linguistic behavior more specifically,
have tended to talk in terms of women’s behavior, women’s style—as
opposed to a more general notion, people’s style. Women have always been
classified as the other, the not-quite-human, whether by medieval theo-
logians who claimed women had no souls, or by more modern psycholo-
gists (as demonstrated in the Broverman et al. 1970 study) who would
claim that a “healthy woman” and a “healthy man” are characterologi-
cally very different things: a “healthy man” is identified with a “healthy
human being” or “healthy person,” while a “healthy woman” is not. This
extremely dangerous prejudice infects much of our scholarly behavior
[ 118 ] Context Counts
119
4. It is true that obsessive-compulsive style is more prevalent among women than
hysterical style is, or has ever been, among men. As far as I know, no reason for this
is advanced in the literature dealing with neurosis. One possible line of explanation is
suggested here: if neurosis is another term for the inflexible adherence to a cultural tar-
get, it makes more sense to cling to a target that is given some respectability than one
which is ridiculed. There is a sort of hierarchy, perhaps, in the selection of a neurotic
style: first, choose the one that is the exaggeration of one’s own group’s preferred strat-
egy; failing that, for whatever reason, choose what is culturally valued. So a woman
would most naturally select hysteria, second, obsessional neurosis. A man presumably
has only the latter option.
[ 120 ] Context Counts
121
to a less pronounced degree in all our behavior. We need not sneer or pro-
test our total sincerity: were it not for conventional politeness and our will-
ingness to use it and interpret it kindly, civilization would likely dissolve in
no time at all.5
But conventional application of a strategy occurs only in case that
strategy is seen as the target for an individual in a particular cultural con-
text. For the Japanese, male and female, conventional Deference is com-
mon, where for the German it is not. For the traditional American male,
Distance/Clarity has been the target, and so we find in this group’s behav-
ior much conventional Distance and Clarity. But Deference is not targeted,
and is not conventionally expected, of a male in this culture. And even when
overt Deference occurs, we still do not find the extralinguistic indicators
of Deference in the behavior of a normal American man—e.g., downcast
eyes, giggle, an attitude of general helplessness—although we do find such
behavior among Japanese males, as we might expect. We might interpret
this discrepancy as showing that the extralinguistic devices are the most
apt to be used, and interpreted, as conventional. So these are good tests of
the targeted politeness for a group—what indirect, and therefore perhaps
not entirely conscious or intentional, indicators of a strategy are employed.
Now if a strategy is the target for a particular group, a member of that
group will know that behavior in accordance with that strategy may be
interpretable either as real or as conventional. But for nontarget strategies,
only real behavior can be inferred. And although one might think that peo-
ple might learn that other groups make different strategies their targets
and so apply them conventionally, people generally seem to be unable to
transcend their own systems: participants can interpret others’ strategies
conventionally only in case they themselves could use them conventionally
(because their group makes them targets). People who retain perspective
and flexibility in interpretation better than others are felt to have special
understanding of groups and cultures other than their own, but they are
admired as the exception rather than the rule.
American women’s traditional target is Deference, as men’s is Distance/
Clarity. So we find certain kinds of conventional behavior based on these
strategies in women’s behavior, which men do not normally adopt. This is
the reason why men are in general unable to interpret feminine behavior as
conventional, rather than real, and hence regularly misinterpret women’s
intentions. And then, when women respond to their misapprehensions by
5. The same point is made about the social utility of disclaimers of action (e.g., the
same sorts of hedges and indirections discussed here) by Roy Schafer (1976).
[ 122 ] Context Counts
123
[ 124 ] Context Counts
125
these Rapport rules have changed over time: this is not the first time that
we have switched our targets. This is an important point, since one might
otherwise think that the shift from Clarity/Distance to Camaraderie was
an unprecedented step in human behavior, induced by the media, greater
mobility, and so forth. It is undoubtedly true that it is harder to live com-
fortably in an era where the rules are in flux, as is the case in this society
at present, than in a relatively stable order—say, Victorian Europe. But
our present situation is by no means unique; indeed, just like the linguis-
tic grammar, stylistic grammar must shift and always is partially in flux,
although the changes are seldom evident to a contemporary member of
the culture.
There is an interesting illustration of an earlier period of rule change
in Philippe Ariès’s book (1962). Although Ariès does not address the
issue directly, much of what he talks about can be interpreted in terms of
Rapport rules and change in target. During the thirteenth century, vari-
ous stylistic modalities were in flux in Europe. People were beginning to
develop a notion of privacy: they began building houses with separate
rooms, wearing more constraining and concealing clothing. About this
time too, I believe, last names were being devised and used. All of these
work toward a Distancing form of rapport. So we can infer from this sort of
evidence that European society during this era was shifting from a target
of Camaraderie to one of Distance. We would expect in such a period of flux
to find more discomfort than usual among people as they realized their old,
internalized rule systems were inappropriate but did not know yet what
was expected of them in the new. And so it is not surprising to find in Ariès
a discussion of the vogue of etiquette books, in the form of so-called “cour-
tesy manuals,” in this historical period. Courtesy manuals played a crucial
part in the medieval educational system, and these etiquette books, judg-
ing from Ariès’s excerpts, are designed to teach Distancing behavior: they
contain rules like “Don’t touch yourself in public, nor other people”; “Don’t
be too familiar in address”; “Dress modestly”; “Don’t tell your dreams to
people.” The aim is to appear as unobtrusive and unintrusive as possible—
clearly, the establishment of conventions of Distance.
It is true that similar Distancing etiquette manuals have remained
around up till the present, but seldom have they been accorded the impor-
tance that they were during this period. There will always be some residual
insecurity about the correct formulas for achieving the idealized and con-
ventional mode of rapport, but never so much as when the targets them-
selves are shifting.
Then if it is true that we are now in just such a state of flux, we should
expect to find a similar desperate emphasis on etiquette manuals. If we
[ 126 ] Context Counts
127
[ 128 ] Context Counts
129
REFERENCES
[ 130 ] Context Counts
131
[ 132 ] Context Counts
133
W H E N TA L K I S N O T C H E A P [ 133 ]
134
[ 134 ] Context Counts
135
REFERENCES
W H E N TA L K I S N O T C H E A P [ 135 ]
136
137
CHAPTER 6
This paper originally appeared in The State of the Language, edited by Leonard Michaels
and Christopher Ricks, 440–48, Berkeley: University of California Press, 1980. It is
reprinted with permission of the publisher. © 1980 The Regents of the University of
California.
1. To enumerate and describe, however briefly, a reasonable subset of relevant recent
work in conversational analysis would occupy an inordinate amount of space. To sum-
marize briefly, contributions in this area have been made from the vantage points of
several fields: formal and informal linguistic pragmatics (of which this paper is an
example); ordinary-language philosophy, of which the work of Grice (1975) is of espe-
cial importance here; and ethnomethodology, of which the work of Sacks, Schegloff,
and Jefferson (1974) is exemplary.
138
2. See Bateson 1972.
[ 138 ] Context Counts
139
[ 140 ] Context Counts
141
[ 142 ] Context Counts
143
6. The survey of Luborsky, Singer & Luborsky (1975) is evidence for this claim.
What therapy teaches by linguistic precept and example is that the same
phenomenon can be perceived in more than one way at once. This is what
therapeutic theorists like Gregory Bateson and Jay Haley refer to as learn-
ing to metacommunicate, to communicate about communication, to see
an utterance at once as both itself and a statement about itself. The allu-
sive nature that therapeutic theory attributes to all discourse helps make
understandable another peculiarity of therapeutic discourse: its fondness
for metaphor and parable.
A great deal of psychotherapeutic theoretical writing of all schools is
expressed in metaphorical terms. This is partly because no one has ever
seen a human mind working and we can only visualize this process in terms
of what we have seen. But more pertinent is the fact that stating something
as a metaphor or an allusion is a form of reframing or recontextualizing.
Metaphor often consists in placing a familiar idea in a new context. So
we find metaphor on all levels of therapeutic discourse: Freud talks about
the ego as a rider trying to control a horse that is the id; of psychological
processes as transfers of mechanical or electrical energy. More recently, as
substitutes for these physicalizations or concretizations of highly abstract
processes, we find mental processes described in terms derived from infor-
mation theory, decision theory, or computational theory. And replacing
Freud’s metapsychological metaphors—the horse and rider, or armies
advancing and retreating—we find images of game playing, of parent-child
relationships, of concrete spatial relations, prevalent in the speech of grad-
uates of the human-potential movement (represented, albeit parodically,
in Cyra McFadden’s The Serial): people are “uptight,” “up front,” “laid back,”
or “with it”; they can or can’t get “behind things”; people “feel” or “experi-
ence” rather than think things. A different class of metaphor, certainly: but
what is unchanging is the representation of abstract psychological states
and processes as concrete, physical realizations.
Interestingly, recent work by Roy Schafer (1975), within a quite ortho-
dox psychoanalytic model, has criticized this fondness for metaphor and
urged the use of literal description on the grounds that the use of meta-
phor prevents the patient from acquiring a sense of responsibility for his
or her actions.7 The suggestion is, then, that the allusive conventions,
implicitly a part of traditional therapeutic discourse, are infantilizing; they
keep patients from achieving maturity. And yet—infantilizing or no—
recontextualization is a necessary part of therapy through the agency of
transference, a necessarily regressive relation developed between therapist
7. See Schafer 1975.
[ 144 ] Context Counts
145
REFERENCES
in either italic or gothic script), we can now expand our design capabili-
ties in a moment by choosing a font style as easily as choosing the color of
the paper.
When rereading Robin’s article, I was struck by the relevance of her argu-
ment for some of the changes that computer literacy has brought into our
lives. I had not stopped to consider how much these typographical choices,
while aesthetically communicative, also affect the rhetorical import of the
written communication. The pragmatic consequences of these styles go far
beyond the surface design of the text. The adoption of computer literacy
has in fact brought about a commingling of the styles of the oral, imme-
diate, and telegraphic with the more considered compositions of written
text. Not only is the process of creation of textual products changed, but
the way we undertake written tasks is also different. Writing and creat-
ing text has become more like tinkering at a workbench than like grand
construction; whether writing a book, a journal account, a personal letter,
or an office memo, we undertake the task in the same contextual environ-
ment and with a similar compositional strategy. Text can be written and
rewritten, spliced, edited, and designed until the producer is satisfied. Or
it can be regarded as a finished written text at the very moment it emerges
from thought, through the fast-moving and often inaccurate finger pres-
sure on keys, and in the same way, it can be immediately transmitted to
others. Thus computer email has made written conversationalists of us
all. And in such composing activities as these, the text is seen to mirror
fleeting thoughts with momentary decisions. Composing activities such as
these has led to a shift in the signaling load between written and spoken
communication that has begun to influence many other forms of textual
composition in new ways (Alvarez-Caccamo & Knoblauch 1994).
When I first read Robin’s article sixteen years ago, I was immediately
interested in its implications for understanding the processes of children’s
early literacy acquisition, with which I was involved at that time. I found
most revealing the possibility of seeing a developmental growth of literate
strategies from those that were more oral and depended on spoken lan-
guage to those that were more governed by written textual conventions.
However, later I discovered this was not completely the case. For many chil-
dren, both middle-class and lower-class, the typographical or illustrated
page has a salience from very early in life that shapes the very idea of what
a text is and what it can convey (Cook-Gumperz & Gumperz 1981). The
linguist Georgia Green made a similar discovery when, while watching her
child choosing books for a bedtime story, she noticed that her daughter was
rejecting some stories as having been read before. When she questioned
her, her daughter said that she recognized the story, although her mother
[ 148 ] Context Counts
149
knew she could read neither story nor title. Apparently, the child identified
the style of the illustrations as being the same as those in books that had
been read to her before. When Green went on to devise further informal
experiments with other preliterate children, she found that these children
similarly paid close attention to stylistic characteristics of books and even
made predictions about text and content based on styles of presentation
(Green 1982).
Robin’s examples of the textual intermingling of written and spoken
strategies draws attention to the fact that the pragmatic function of text
goes far beyond the salience of its surface effects. These assumptions have
a direct parallel in children’s lives, where the understanding of the rela-
tionship of speaking to writing begins with an early awareness of textual-
ity which then accompanies their move into full literacy. While children’s
primary literacy experience is with the printed text, as books of stories for
the most part written for children, when these are read to them orally they
become not written speech but spoken writing. The notion of a text that
exists independently of any person, that once created can stand alone as its
own textual self, is born at the very moment that a child says, “You missed
a page, read it again.” The Bahktinian notion of textuality is rediscovered
over and over again at many different bedtimes.
Such a notion of textuality, once realized, is a guiding force in early lit-
eracy development. Children very early in their written literate develop-
ment learn to make use of such displays as size of letters, placement of
text relative to pictures, and, later, quotation marks and other punctuation
devices to do the work of building textuality that they will later be able
to accomplish by relying on the accepted lexical and grammatical means.
In children’s early attempts at writing, the typographical conventions of a
written page often take over the communicative power of words that can-
not yet be fully realized because of difficulties in spelling or problems of
word order. Robin’s article makes one realize that the choice of linguistic
conventions and orthographic displays can offer important cues to shifts
in pragmatic function. Successful early literate development can be seen as
dependent on two different areas of language experience that contribute to
children’s pragmatic understanding of textuality: (1) the iconicity of a text
as a visual entity and its relation to the mapping of different verbal forms
into different patterns of meaning, which ultimately should lead to a sense
of genre; and (2) storytelling as it constitutes the narrative creation of a
shared, linguistically mediated reality that orders the temporality of the
universe and can exist as a stand-alone text. Once told and written down,
the story can stand by itself and continue to provide possible solutions to
all manner of new problems (Cook-Gumperz 1995).
S O M E OF M Y FAVOR I T E W R I T E R S A R E L I T E R AT E [ 149 ]
150
However, Robin’s warnings about the myths that guide the interpreta-
tion of what constitutes a literate consciousness and an intelligent under-
standing are still needed. Even though a new iconicity can be said to pervade
the world of the computer-literate, the idea that clear thinking is demon-
strated in an orderly presentation of written words has not yet been sur-
passed. After children’s early encounters with written words and textuality,
their experience of the power of written words as active textual entities
becomes a part of their past. However, the schoolroom literacy experience
continues. In this environment, the injunction is often heard that children
should say “one thing and say it clearly” (Michaels 1986). The blooming
confusion of the multivocality of emergent texts as pictures, words, quota-
tions, and typographical devices, and all manner of meanings being packed
into a page, like a Richard Scary children’s book, becomes, under the ruling
hand of school literacy, the linear presentation of a single idea.
Robin, we still need to heed your warnings if we are to understand the
shifting literacy of the computer age.
REFERENCES
[ 150 ] Context Counts
151
CHAPTER 7
INTRODUCTION
This paper originally appeared in Spoken and Written Language: Exploring Orality and
Literacy, edited by Deborah Tannen, 239–60. Norwood, NJ: Ablex, 1982. Reprinted
here with permission of the publisher.
152
[ 152 ] Context Counts
153
paradox. But we must recognize the paradox if we are to emerge from our
confusion with honor.
We can see signs of change and confusion in a number of places, if we
correctly interpret what we see. In this paper I want to bring together a
group of seemingly unrelated or random facts that I have been noticing
over the last couple of decades, with the intention of arguing that these
illustrate the shift in our society from a literacy-based model of ideal
human communication to one based on the oral mode of discourse.
This is not, of course, the only such changeover in our culture. We can
be confident that several millennia ago contemporaneous civilization
underwent an equally agonizing shift, only in the opposite direction, as
writing became widespread and overtook oral modes of literature as a
means of recording present and past events. Of course, we have no record
of the wrenching effect of the changeover: the wrench would have princi-
pally affected the Old Guard—in this case, the nonliterate—and they, of
course, being nonliterate, have left us no record of any pain they felt. More
recently, within the past several centuries, a more minor change seems
to have taken place. Within a culture prizing literacy for the recording of
memorable events, there was a shift from a state where most people were
in fact nonliterate or barely functionally literate, and depended on another,
or scribal, class, to achieve their memorability for them, to one in which the
majority of people were, in the worst case, at least expected to be literate,
to make use of the written medium for information and amusement with
relatively little effort. It is not clear whether, in fact, this ideal ever actually
existed, but it certainly has been present as an ideal.
[ 154 ] Context Counts
155
lacks many of the devices oral, present discourse utilizes as carriers of emo-
tional tone: intonation, pitch, gesture, eyes, and so on. On the other hand,
nonspontaneous media, by their ability to capture, through planning, the
appropriate mood, the description, may help the reader or hearer form an
ultimately more lasting and more vivid memory in the mind, and allows a
reader as well as a writer to rethink, re-experience, and revise impressions
which, in traditional forms of oral discourse, are lost forever once uttered.
There are additional “meanings” that we attribute to the choice of
medium. Written communication is memorable, the stuff of history and
reliability. Hence—and also because it is acquired and utilized with more
difficulty—it is more formal, the bearer of respectability. Oral discourse
can be colloquial or dialectal; the representation of nonstandard dialect in
writing—as a reader of, say, Mark Twain or George Ade will attest—tends
to give a reader a tired throat after a short period of reading: we cannot
help subvocalizing as we read “dialect”; it exists only in oral form.
Written discourse, then, is respectable; spoken, more heartfelt. A cul-
ture at any point in time has to decide whether the preferred mode of pre-
sentation of self is as a respectable or as a feeling creature. There may, at
some times in some situations, be available the chance to be both at once,
so that no such decision must be made. But in the matter of form of com-
munication, a society must decide whether the ideal is that of writing or
that of talking—reliability or warmth, respectability or ability to convey
emotion. For the last several centuries, we have, where possible, opted for
the first, assuming that the written channel was in some sense primary
or preferable. For various reasons, some social, some technological, we are
at present in the process of shifting, so that we prefer and respond most
appropriately to communications in any mode couched in an oral frame-
work. This switch, like any profound cultural revolution, is creating severe
confusion and dislocation, especially in those who perceive themselves as
holdovers of the old order. But rather than take moral or aesthetic posi-
tions in favor of one or the other, we will do better to examine the evi-
dence of the claim that this change is in progress, and then consider why, if
indeed it is, and what if anything we are to do in response.
If the written medium is primary, we can expect to find that even when
people speak or are assumed to speak “spontaneously,” their contributions
are represented via the conventions of the written mode. Thus quotations
in biography or fiction can be expected generally to sound like “written”
discourse, to utilize its conventions. We might find other, special uses of
an assumption of the primacy of writing: for instance, serious people,
in serious situations, in a work of art, might be represented as adopting
stylized “written” modes of discourse even when they are supposed to be
[ 156 ] Context Counts
157
QUOTATION MARKS
[ 158 ] Context Counts
159
They are used to enclose, and exonerate, anything that might be consid-
ered nonliterate form, however mildly colloquial:
Or, indeed, they may be used to indicate anything that is, until entrusted
to paper, information that only the writer, and not the reader, possesses.
This use may seem to contradict the claim that these quotation marks
indicate denial of responsibility; but in fact, information not certainly
shared by the reader is apt to make a fledgling writer especially nervous.
One recent use of quotation marks outside of written communication
is particularly noteworthy. Traditionally, of course, quotation marks are
1. A striking example of this use of quotation marks was seen in the television reports
of the freeing of the US hostages by Iran on January 20, 1981. The hostages were wel-
comed to the U.S. military installation in Wiesbaden, Germany, by crowds, some of
whose members were carrying signs. The most prominent of these was one which
said: “Welcome Home!” (with the quotation marks).
ITALICS
Other devices, on the other hand, are more direct in their communicative
effect, dictating a specific emotional response just as stress or pitch might,
for instance, in oral discourse. This is a principal function of italics. Italics,
like quotation marks, serve two distinct but related purposes. One, their
[ 160 ] Context Counts
161
No, this is Mrs. Jones, and that is Mr. Smith, with the carnation in
his lapel.
As illnesses go, hypoglycemia has a good deal of flair, but another debility is
almost as status-laden these days: low back pain. No one is sure why an ach-
ing back should have such panache. Perhaps the appeal of this disorder is that
it has traditionally been associated with vigorous, earthy, physical types… .
(Cosmopolitan, July 1980)
The Cosmo italic is really a sort of amalgam of the two conventions I have
identified. The reader is undoubtedly supposed to imagine the writer’s
voice rising to a squeal; but at the same time, the words italicized, as in
the quotation above, would not normally receive unusual stress or pitch in
speech. The italics then signal “N.B.”
Italics, however, have a meaning not shared by quotation marks: they are
popularly characterized as a written manifestation of “women’s language.” It is
popularly thought that women, when they write, overuse italics (for example,
in letter writing). This stereotype no doubt helps to account for the prevalence
of italics in Cosmopolitan. Presumably this accords with the wider inflectional
range of women’s speech. As far as I know, no serious research has been done
on the prevalence of italics in women’s written prose. Any correlation with
the wider range of oral intonational possibilities would be rather unexpected
and inexplicable, unless we assume that in writing we literally encode a spo-
ken “voice” in our minds, which seems quite implausible. In any case, whether
because of their association with “feminine” style or because they too heavy-
handedly attempt to manipulate the reader’s emotional response, italics are
much riskier in writing than their counterparts in speech.
CAPITALIZATION
[ 162 ] Context Counts
163
Here capitalization does indeed serve to mark the words of greatest sig-
nificance in the narrative, rather like the Cosmopolitan italics. But it does so
without recourse to spoken convention.2
NONFLUENCIES
There are still other devices of spoken language that are carried into writ-
ten language: for example, ellipses, repetitions, and vocalized pauses.
These, though, fall into a different category. Where the earlier types are
intentionally used to create emotional involvement, these tend to be more
or less involuntary. Indeed, encountering oneself on tape using them to
excess frequently causes chagrin: “Do I talk like that?” They do not, typi-
cally, mark emotional directness, but rather simple unpreparedness—the
negative side of spontaneity. For while in writing we can marshal our
thoughts and present them in coherent and rhetorically effective order, in
speech we seldom have the opportunity, and at least currently, if we do,
we try to look as if we are not making use of it. It is conventional—and,
probably, realistic—to assume that the more emotionally involved one
is, the more one’s thoughts are confused, or at least presented linguisti-
cally in incoherent and rambling form. Therefore, the use of these devices,
primarily resulting from the normal inability in spontaneous discourse to
know in advance what to say, secondarily can signify emotional turmoil in
oral communication, at least if carried to excess, or used in unusual places
in the conversation.
Perhaps, in fact, it is this that gives metered dialogue, in older drama-
tists, its sense of seriousness and majesty: if we are working within the
constraints of iambic pentameter (for instance), the form cannot tolerate
any additions or deletions, no vocalized pauses, hesitations, or repetitions.
Distress must, then, be represented lexically and explicitly. In more mod-
ern forms of constructed dialogue, however, we find these devices used for
2. The reverse of A. A. Milne is, perhaps, e. e. cummings, who eschewed capitalization
entirely. If Milne (and others) use capitals as directives to readers as to what is to be
considered important, then we can take cummings’s avoidance of the device as a kind
of anarchy: readers are left to their own devices, with no authorial guidance.
very specific effects. Soap opera is one genre that strives for the appearance
of spontaneity, both to manipulate the audience’s emotions and to increase
the verisimilitude of the dialogue. Besides, soap opera can afford it while
other genres can’t: a script aiming for artistic effect must compress and
telescope its exposition, leaving explicit only what is necessary to further
the development of the plot and the characters. Therefore, those aspects
of spontaneous dialogue that are truly random and accidental—the result
of universal human difficulties in organization and memory, rather than
problems specific to the character and his or her situation—are eliminated.
But soap opera runs by different rules. Soap opera tries to provide a literal
slice of true experience for its audience, so that life, and dialogue, in that
genre move just as slowly as they do in real life. Hence there is a time for
ellipsis, correction, hesitation, and so on.
One random episode in a randomly chosen half-hour daily soap opera
(All My Children) illustrates this aspect of soap-opera grammar interest-
ingly. A male character has, in the way common to soaps, unbeknownst
to himself fathered a daughter years ago by a woman who has since (as a
result, we may surmise) become a nun. This girl and the man’s legitimate
son (unaware of course of their true relationship) are having an affair. They
are somehow involved in a murder, and are in jail. The son, it is implied,
has fallen into these bad ways as a result of his father’s neglect, of which
his father is keenly aware. In the randomly chosen segment, the father first
encounters the nun, who indirectly informs him of the identity of the girl,
and then encounters his son and the girl, separately, in jail. Thus, we have
one character who, in the course of the episode, has increasingly more to
feel uncomfortable about, and whose every confrontation with another
character is wracking to him; and several other characters who, while they
have their various guilts, have no guilt with respect to this particular charac-
ter. This character, incidentally, is a lawyer and politician, so articulateness
is expected of him. As the episode progresses, while the speech patterns
of the other characters remain more or less stable, only occasionally laps-
ing into real-speech conventions, the father progressively gets closer and
closer to transcript form.
Another interesting use of “spontaneous” form in nonspontaneous oral
media is seen principally in commercials. The commercial is, in one sense,
the diametric opposite of the soap opera: time goes by in a flash; every
micro-millisecond must count. At the same time, the actors must seem
real, must be people the audience can identify or at least sympathize with.
Because of the time constraint, we see very little use of time-wasting oral
devices. Rather, we get a different set of quasi-oral conventions. In true
spontaneous discourse, sentence structure such as is expected in written
[ 164 ] Context Counts
165
prose is not strictly adhered to. We find run-ons and fragments quite typi-
cally. So one way to approximate spontaneous speech is via telegraphic
fragments:
Aside from the question of whether normal adults ever really find them-
selves conversing at length about breakfast cereal, this passage is unnatural
in a couple of respects. If telegraphic fragments do in fact occur in ordi-
nary conversation, they tend to occur in the body of the discourse—not
as introductions, where the topics of discourse tend to be made explicit.
It also seems to me that the third contribution is unlikely: repetitions of
this type normally signal serious disagreement, not merely the need for
a bit of amplification (for which “Oh, really?” normally suffices). But in
the commercial, every syllable must add to the informational content.
Finally, notice the word crispy. This is certainly not a word characteristic
of formal written register—in fact, crispy is not oral colloquial American
English: I don’t think I have ever heard it uttered spontaneously. It is one
of a set of words used only, or almost exclusively, in writing to suggest oral
register, but in fact seldom or never in true spontaneous oral discourse
(tyke is another).
I have contrasted these special narrative conventions with transcripts,
and indeed if we continue this comparison we can begin to understand
why, accustomed to these narrative conventions, we find accurate tran-
scripts of spontaneous conversation especially hard to interpret. We have
been trained to believe that, when we encounter these devices in written
communication, we must translate them as signals of emotional intensity.
But in ordinary transcripts, we find them in every sentence, in every con-
text. Either we are dealing with a feverish emotional pitch—belied by other
clues in the recorded conversation—or we are truly contending with a for-
eign language, or perhaps a pidgin—as indeed a transcript is, half written
and half spoken grammar.
In general, then, the borrowing of a device from one medium into
another is always overdetermined: it carries with it the communicative
effect, or “feel,” of one medium into another (the metacommunicative
effect) and at the same time attempts to utilize the language of one mode
to communicate ideas in another (the communicative effect). It is no won-
der that this sort of translation can create confusion in readers (or hear-
ers), and can also create in them very strong feelings—typically negative.
THE COMIC STRIP
The negative impact of such borrowings, particularly from the oral to the
written mode, is exacerbated by the fact that the least “desirable” forms
of communication are the first to show the traces—or perhaps this is the
effect of the derogation of oral style, rather than the cause of its initial
appearance in unrespectable places. One of the earliest forms of commu-
nication to attempt to convey essentially oral concepts in print was the
comic strip. Interestingly, in the newer and more sophisticated strips (e.g.,
Doonesbury by Garry Trudeau), the conventions of the old strips are miss-
ing, and we have, as it were, reverted to the “respectability” of print con-
vention. But the strips I remember reading as a child, though they utilized
print, actually approximated in many ways a blend of oral and literate cul-
ture. It is easy to recall examples: the fact that comic-book sentences never
ended with a period but, if not questions, with an exclamation point, seems
to have been an attempt analogous to the use of italics or initial capitals to
impart importance and emotional immediacy to the text. (Alas, with over-
kill it lost its potential force.) There were frequent attempts to reproduce
oral self-correction devices, for emotional effect, e.g., pauses and repeti-
tion. Additionally, there was the attempt to reproduce nonstandard dialect
and colloquialism. Very commonly, this was done by “dropping g’s” from
the suffix -ing, but rather more interesting were other attempts to create
the same effect.
I remember as a child trying to imagine pronouncing some of the comic
book writers’ attempts at “natural” speech. There were basically two types
of problems. One occurred when the form could not in fact be pronounced
as written by anyone with the normal articulatory apparatus; the other
involved spelling representations that, in fact, represented the only way a
form could be pronounced at all, outside of profound affectation, so that it
was hard to see what was gained by the special “colloquial” spelling. In fact,
it sometimes led to more confusion than necessary. Examples of the first
type include the dropping of vowels to form impossible consonant clus-
ters: T’ th’ store; of the second, spellings like yuh or ya for unstressed you
(always pronounced [jə]); ta for to; and the mysterious (to me for years)
locution Omigosh, which I perceived then as an exotic exclamation [ˈɑmigɑʃ]
(years later, I realized that it was intended to represent the much less inter-
esting Oh, my gosh). But running it together as a single word (which led to
my confusion) served the purpose of making it seem more quickly articu-
lated, hence more exciting as well as more “colloquial.” We sometimes find
phonetic spelling used to represent slang, although the word itself would
not be pronounced differently in standard and nonstandard dialects: wuz
[ 166 ] Context Counts
167
for was, for example. In all these cases, then, special spellings are used not
simply as a guide to pronunciation but as a way of indicating, “Since this
representation is different from the ‘formal’ forms of written language, it is
to be taken as ‘oral,’ i.e., immediate, emotional, colloquial.”3
Abuse of the trust thus engendered is always a temptation. I remem-
ber a striking case from the 1964 presidential primary campaign. George
Wallace was running in this campaign, a fact the New York Times, for rea-
sons good and bad, viewed with dismay. Wallace was perceived as a red-
neck, an illiterate—a far cry from the hallowed JFK, and not even up to
the style of the folksy but nonetheless minimally literate LBJ. At one point
the Times attributed to Wallace, in an interview, a locution the Times repre-
sented as “could of.” What is remarkable is that the newspaper was thereby
representing an utterance that could not in fact have been pronounced,
in nonstilted speech, by a native speaker of English in any way other than
[kʊ́dəv] or, even more colloquially, [kʊ́də]. It corresponded, of course, to
the written form “could have,” but would never normally have actually been
pronounced that way. So the Times was engaging essentially in comic-strip
tactics, representing a form via nonstandard spelling not in order to accu-
rately distinguish a special pronunciation from the standard, but for some
other reason. I would argue that the reason was to make Wallace look like
an illiterate redneck; that if someone of unimpeachable intellectual creden-
tials, say Adlai Stevenson, uttered the same phonetic segments, they would
have been represented as “could have”; had they been uttered by the infor-
mal but respectable Lyndon Johnson, perhaps as “could’ve”; but “could of”
marked Wallace as unmistakably of the booboisie, an implicit editorializa-
tion in the paper of record’s news columns.
So anything deviating from the written medium and its ways of expres-
sion has tended to be viewed as suspect, a bad second best, certainly not to
find its way into respectable print. Indeed, a good part of the fashionable
3. On the Phil Donahue show of January 23, 1981, Donahue read a letter from a
viewer whose essence was this: She had always found the show one of the most “liter-
ate” on television. Recently she had received her first transcript of the show, and was
dismayed to discover that Donahue is cited as saying ya and +-in’—e.g.:
“I have to tell ya… .”
“It’s comin’ to the time… .”
There are two relevant points here. First, it is perhaps curious that transcripts of
the Donahue show provide this sort of close phonetic transcription (and even more
curious, their fidelity appears somewhat selective: if ya, why not hafta or haveta?); sec-
ond, it is striking that—as my foregoing discussion would predict—the viewer was
perfectly comfortable with [jə] and [kʌmɪn] in oral discourse, but reinterpreted these
forms when they were encountered in print.
“But,” began Oedipa, then saw how they were suddenly out of wine.
“Aha,” said Metzger, from an inside pocket producing a bottle of tequila.
“No lemon?” she asked, with movie-gaiety. “No salt?”
“A tourist thing. Did Inverarity use lemons when you were there?”
(p. 19)
[ 168 ] Context Counts
169
The thing was, he said, the Mercury system was completely automated.
Once they put you in the capsule, that was the last you got to say about
the subject.
Whuh!—
“Well,” said Yeager, “a monkey’s gonna make the first flight.”
A monkey?—
The reporters were shocked … Was this national heresy? What the hell
was it? …
But f’r chrissake …
(p. 100)
or:
And in this new branch of the military, no one outranked you. (p. 108)
Sympathy … because our rockets all blow up. (p. 109)
… he had literally written the book on the handling characteristics of
aircraft … (p. 108–109)
In all the last three examples, italics are used to indicate, almost paradox-
ically, a sort of ironic detachment: “They keep saying this, this is a cliche.”
Sometimes, too, they mirror directly the vocal inflections of incredulity:
The italics here can represent nothing but vocal intonation, since in
writing italics are never used to underline part of a word, except to indicate,
for instance, misspelling—deviations from written convention.
Examples could be multiplied at will to illustrate the really shocking
change that Wolfe has visited on English narrative prose convention. What
is surprising is that it works. Purists may complain, but Wolfe has had a
[ 170 ] Context Counts
171
changes nothing before its time. Thus, if we look at what literacy has
achieved for humanity in the past—as well, by the way, as what the advent
of literacy has cost us, for there have indeed been costs—we may well find
that there are now other means at our disposal to achieve these same ben-
efits, perhaps with fewer unfortunate side effects; and that, in fact, the
new mode that is gaining strength at the expense of literacy will enable us
to communicate more beautifully and forcefully with one another than can
be envisioned now. Since the change is barely in its infancy, is in fact still
in gestation, we cannot see the final product yet. But we can at least take a
different tack in viewing the oral/literate dichotomy. Rather than wringing
our hands about the loss of literacy, let us ask what is replacing literacy, and
why this change is occurring, and finally, what the gains as well as losses are
apt, in the long run, to be.
While most of what we encounter contrasting literate with oral culture,
or discussing the apparent decline of literacy among ourselves, explicitly
or implicitly assumes that literacy is the desirable state, and to forswear
it is to invite the return of the Dark Ages, there are other possible points
of view. Marshall McLuhan, for example, has argued in his writings (e.g.,
1964) that a literate culture loses something: immediacy, warmth—the
qualities I have been suggesting we are trying to put back into our liter-
ate productions, albeit not really successfully. McLuhan suggests that the
coming of literacy caused great, and undesirable, changes in human char-
acter and social behavior: we retreated, physically and emotionally, from
one another. But now that nonliterate media are gaining influence at the
expense of literacy, we are beginning to see the return of the old virtues.
McLuhan, then, sees literacy and nonliteracy as heavily influencing charac-
ter and behavior, as well as less clearly social psychological attributes.
On the other hand, it is sometimes argued that scientific progress, in
terms of linear logical thought, can only be found where there is literacy;
that the written word, in fact, and its linear syntactic arrangement into
sentences going from one side of the page to the other (or, for that matter,
up and down, but in any case, in a visually perceptible pattern implying
sequencing) has formed our thought and enabled us to formulate concepts
like logical causality and temporal sequentiality. Without literacy, the argu-
ment continues, we would lose all ability to think logically, and revert to a
state of animalistic savagery.
To my knowledge, no evidence for either position, in any empirical
sense, has surfaced. A fairer representation may be found in the writing of
scholars like Olson (1977), who sets forth advantages and disadvantages
in cognitive style for both states. In particular, they note that members
of nonliterate societies have far better short-term memory capacity—as
[ 172 ] Context Counts
173
CONCLUSION: A NEW NONLITERACY
[ 174 ] Context Counts
175
REFERENCES
Coleman, Lee. 1980. “In” diseases: A look at the current crop of chic complaints.
Cosmopolitan, July 1980.
Donahue. 1981. Episode first broadcast 23 January 1981 by WGN. Directed by Ron
Weiner.
MacGraw, Ali. 1980. Interview. San Francisco Chronicle, 20 October 1980.
McLuhan, Marshall. 1964. Understanding media: The extensions of man.
New York: McGraw-Hill.
Milne, A. A. 1926. Winnie-the-Pooh. London: Dutton.
Olson, David R. 1977. From utterance to text: The bias of language in speech and
writing. Harvard Educational Review 47:257–81.
Pynchon, Thomas. 1966. The crying of Lot 49. New York: Bantam.
Wolfe, Tom. 1979. The right stuff. New York: Picador USA.
Introduction to “Persuasive
discourse and ordinary conversation,
with examples from advertising”
BY JANET S. SHIBAMOTO-SMI TH
[ 178 ] Context Counts
179
P E R S UA S I V E DI S C O U R S E A N D OR DI N A R Y C O N V E R S AT I O N [ 179 ]
180
[ 180 ] Context Counts
181
REFERENCES
P E R S UA S I V E DI S C O U R S E A N D OR DI N A R Y C O N V E R S AT I O N [ 181 ]
182
183
CHAPTER 8
The very work that engaged him [in an advertising agency] … wafted him into a sphere
of dim platonic archetypes, bearing a scarcely recognizable relationship to anything
in the living world. Here those strange entities, the Thrifty Housewife, the Man of
Discrimination, the Keen Buyer and the Good Judge, for ever young, for ever hand-
some, for ever virtuous, economical and inquisitive, moved to and fro upon their com-
plicated orbits, comparing prices and values, making tests of purity, asking indiscreet
questions about each other’s ailments, household expenses, bedsprings, shaving cream,
diet, laundry, work and boots, perpetually spending to save and saving to spend, cutting
out coupons and collecting cartons, surprising husbands with margarine and wives with
patent washers and vacuum cleaners, occupied from morning to night in washing, cook-
ing, dusting, filling, saving their children from germs, their complexions from wind and
weather, their teeth from decay and their stomachs from indigestion, and yet adding so
many hours to the day by labour-saving appliances that they had always leisure for visit-
ing the talkies, sprawling on the beach to picnic upon Potted Meats and Tinned Fruit,
and (when adorned by So-and-so’s Silks, Blank’s Gloves, Dash’s Footwear, Whatnot’s
Weatherproof Complexion Cream and Thingummy’s Beautifying Shampoos), even
This paper originally appeared in Analyzing discourse: Text and talk, edited by Deborah
Tannen, 25–42. Georgetown University Round Table on Languages and Linguistics
1981. Washington, DC: Georgetown University Press, 1982. Reprinted here with per-
mission of the publisher.
Acknowledgments: The ideas in this paper owe much to discussion with many oth-
ers. In particular, Linda Coleman’s (1983) work on advertising, as well as other forms
of persuasive discourse, has been stimulating. Students in Linguistics 153, Pragmatics
(spring 1981), have been invaluable in refining my thoughts, and their forbearance in
serving as sounding boards is gratefully acknowledged. Deborah Tannen’s suggestions
and inspiration are similarly much appreciated.
184
attending Ranelagh, Cowes, and Grand Stand at Ascot, Monte Carlo and the Queen’s
Drawing-Rooms.
Dorothy L. Sayers, Murder Must Advertise ([1933] 1986)
I n the field of discourse analysis, much attention has been focused on cer-
tain forms of discourse, much less on others. In particular, scholars have
been concerned with ordinary conversation, on the one hand, and written
expository text, on the other. While the treatment of these types as sepa-
rate entities has certainly taught us a great deal about the characteristics
of each of them individually, and something about the nature of discourse
in general, it leaves a great many questions unexplored. For one thing, how
do they relate to other types of discourse—which may superficially, or even
more deeply, resemble one or the other? What are the universal charac-
teristics of all types of discourse, and what characteristics are specific to
just one or two? Can we devise a taxonomy of discourse types, a means of
unambiguously differentiating among them?
Classifying the basic forms of discourse in terms of their differing and
similar characteristics seems rather less glamorous than writing a grammar,
a system of rules, for one or all discourse types. But in fact, it is impossible
to write a grammar without knowing the basic units involved: grammar
consists of instructions for the combinations of these basic elements.
It would also be profitable to look at discourse not from its surface form
(as conversation, say, or literary text) but more deeply, with interest in its
deeper purpose. When we look at a range of discourse types, we notice
that they appear quite different from one another. Clearly, this disparity
is due to differences in what each is intended to accomplish, so that a suc-
cessful performance of a type of discourse is one that accomplishes what
the speaker, or the participants as a whole, had set out to do, rather than
merely one that conforms to some particular surface configuration.
This paper is somewhat experimental in nature, as I want to address
some of these questions, and see how far we can get toward at least prelimi-
nary answers. As a start, I want to consider one possible distinction among
discourse types in terms of function or purpose: ordinary conversation, on
the one hand, and something we can call “persuasive discourse” (PD), on
the other. I will come to the problem of the definition of the latter shortly.
I am not suggesting that this is the only distinction one can make among
discourse types, nor that it is necessarily the major one. Certainly, we can
[ 184 ] Context Counts
185
the other’s personal appearance), the other has the same privilege in turn,
and if one can refuse to answer, so can the other. Violations of this principle
do, of course, occur in OC, but when they do, participants feel a rule has
been violated, that the conversation is making them uncomfortable, while
nonreciprocity in a lecture is expected and reasonably comfortable.
An example of an intermediate, hence problematic, case is psycho-
therapeutic discourse. In many of its forms, there is the appearance of an
egalitarian, reciprocal conversation, but in terms of deeper intention, the
reciprocity turns out to be only superficial. The therapist can ask questions
which the client soon learns not to ask; and if the latter should attempt to
ask such a question, the therapist, rather than give an answer, will usually
treat the question as a tacit invitation to ask another question, or make an
interpretation: “I notice you’re curious about my personal life.” Further,
many of the marks of power that belong to the lecturer in the classroom
also belong to the therapist: the decision when to begin and end, and—
while the client ostensibly picks the topics of discourse—the determina-
tion of what the client’s contributions actually mean. The client, however,
typically holds the floor for the major part of the discourse—an anomaly
in terms of the relation between floor holding and power holding in typical
conversational settings, which makes it especially difficult to acquire profi-
ciency in therapeutic dialogue.
With these assumptions in mind, one can attempt a definition of persua-
sive discourse as a type of discourse that nonreciprocally attempts to effect
persuasion. Discourse, then, is to be considered persuasive only in case it
is nonreciprocal, and the intent to persuade is recognized explicitly as such
by at least one party to the discourse. By persuasion I mean the attempt or
intention of one participant to change the behavior, feelings, intentions,
or viewpoint of another by communicative means. The last is important.
Communicative means may be linguistic or nonlinguistic (say, gestures),
but they are abstract and symbolic. A gun held to the head may indeed
induce a change in someone’s behavior, but it is not communicative in this
sense. Hence I do not consider a direct physical threat a type of persuasive
discourse. Types such as advertising, propaganda, political rhetoric, and
religious sermons clearly do fall into this category. While lectures, psycho-
therapy, and literature might belong here under some interpretations, they
are problematic and are dealt with later: they seem intermediate between
PD and OC. On the other side, while direct physical intervention is clearly
outside of our definition of PD, brainwashing is more difficult to assign to
either category. It is true that very often there is no direct physical force
involved. But in brainwashing, as the term is ordinarily used, there almost
necessarily are physical interventions— whether isolation from other
[ 186 ] Context Counts
187
and should be considered value-free, with the added assumption that some
discourse types must, to be effective, be nonreciprocal and power imbal-
anced, the only issue in this kind of word being whether it is effective, not
the value of the method by which that effect is achieved. We must try not to
heap obloquy on the commercial, and praise on the therapeutic discourse,
because of their purposes, at least not while we are trying to discover their
properties and their positions within a taxonomy, and eventually a gram-
mar, of discourse.
With this problem out of the way, let us turn to the question of defini-
tion. Some of the factors have already been alluded to at greater or lesser
length in this paper, but I summarize here all the relevant points as they
appear to me, in classifying discourse types, determining how they differ
and what aspects are universal, and differentiating between persuasive and
nonpersuasive discourse.
First, and perhaps most important, is reciprocity, about which much
has been said earlier in this paper. Connected with reciprocity is bilateral-
ity. A discourse may be reciprocal and bilateral, like OC; nonreciprocal and
unilateral, in that true participation occurs on only one side, like a class-
room lecture (though I make amendments even to this statement shortly);
or, most complex, nonreciprocal but bilateral, like psychotherapeutic dis-
course, where both parties most typically can make true contributions to
the conversations, but the contributions may be of different surface forms,
and certainly are open to different interpretations. Turn taking is a natu-
ral concomitant of reciprocity, though (as in psychotherapeutic discourse,
which involves turn taking) the two can be separated; but a non-turn tak-
ing, or unilateral, discourse can never be reciprocal.1
Discourse may be spontaneous or not, or rather, can be conventionally
spontaneous or not. Thus, OC is at least conventionally spontaneous: we
distrust apparent OC if we have reason to suspect any of the participants is
working from a script, or has planned significantly in advance. But a work
of art, or a lecture, is not supposed to be spontaneous, and the lecturer
feels no embarrassment about referring to notes. Hence, we find differ-
ences in the use of hesitation devices, pragmatic particles, cohesion, and
so forth. A spate of y’knows distresses us far less in OC than in a lecture—
and an ordinary conversation style without hesitations and other devices
reflective of spontaneity would make most of us very uncomfortable.
1. An odd apparent counterexample occurred at the end of the Phil Donahue show
(Donahue), June 16, 1981. At this point Donahue turned to the camera (not the studio
audience) and said: “We’re glad you were with us. You were terrific!” It is hard to imag-
ine how the TV audience could have demonstrated its terrificness.
[ 188 ] Context Counts
189
inform the other about necessary facts, are absent, and other things are
present which would ordinarily not be found in OC because all participants
are already aware of them and the contribution is therefore redundant. But
the audience is not aware, and needs to be apprised of the information,
and so it is. An audience has a different role than does an eavesdropper in
true OC.
In addition to the major distinction of audience/addressee, there are
distinct types of audience, based on the role the latter is to play in the dis-
course. The role of audience ranges from totally nonparticipatory to a con-
scious and active involvement. Indeed, an audience can exist even in case
the speaker is not aware of its presence—an eavesdropper, that is, whose
role is precisely to remain undiscovered and therefore to give no clues
whatsoever as to its presence. In ceremonial functions—at a wedding, for
instance—the audience is known to be present, and indeed its presence as
witness to the ritual is necessary if the ritual is to be transacted success-
fully, but there its role ends. It does not participate, it is not expected to
understand, or signal its understanding. Typically, the audience at a wed-
ding does not indicate by any means, verbal or otherwise, its agreement or
complicity. In fact, it does not matter to the members of such an audience
if the ceremony is conducted in a language they do not understand (say,
Hebrew at an Orthodox Jewish wedding). Their role is simply to be present,
not to derive anything themselves from what they hear. On the other hand,
an audience at a classroom lecture has as part of its function to understand
and, perhaps, give approbation—in the form of evaluations later, yes, but
more importantly, immediately, in the form of nodding and other non-
verbal backchannels. A good lecturer, however large the audience, devises
ways of discovering and utilizing this nonverbal response, and without it—
say, speaking to a television audience—someone accustomed to lecturing
in the presence of an audience is lost. Hence, we often find television dis-
course being taped in the presence of a live audience, to provide speakers
with this all-important, though nonverbal and unconscious, feedback.
A third level of audience response is seen in certain forms of religious
and political gatherings: the audience is expected to participate with audi-
ble and explicit backchannels: amen/right on! or applause and cheers. The
audience here is very important to the speaker: it signals by its response
whether persuasion is occurring in these maximally persuasive forms of
discourse.
Related to the foregoing, but perceptible at a more abstract level of
analysis, is the use of the Gricean conversational maxims. The maxims
themselves are problematic for conversational analysis in that they were
formulated by Grice (1975) for quite different purposes than providing an
[ 190 ] Context Counts
191
[ 192 ] Context Counts
193
children, a dog, a white picket fence. The irony is that, in fact, the second
approach is far more effective than the first, since it circumvents intellec-
tual judgment (“How can using a detergent make Harry more responsive
to me?”) and goes directly to the realm of the unconscious, capable only of
desires, fears, and needs.
If much of this discussion has the ring of a psychology textbook, many
modern advertising techniques trace their genealogy directly to the far-
famed couch in Vienna. Indeed, much of modern motivational psychology,
the basis of advertising, derives from Freud, directly or indirectly. For it
was Freud ([1900] 1953) who pointed out the basic distinction between
the processes of the conscious and the unconscious mind. The unconscious
works by the laws of the primary process; the conscious, by the secondary
process. Primary-process thought is preverbal—symbolic, nonsequential,
and visual—while secondary-process thought, more directly “rational,” is
auditory and verbal. Hence, if one wants to persuade by circumventing the
processes of rational thought, it makes good sense to emphasize abstract
images—music and pictures, for instance, rather than words in logical
sequence. And if the FTC is concerned with deception in advertising, it
would do better to pay attention to the nonverbal means of persuasion,
which can be much more deceptive. But the problem is, of course, that it is
difficult for the investigator to prove what the picture of the happy family,
or the vital adolescents guzzling Coke, is communicating, since we think of
communication in terms of logical symbols—words.
Edward L. Bernays (perhaps not coincidentally Freud’s nephew), the
inventor of modern public relations and hence of many of the modern
forms of persuasion, spoke of “the engineering of consent” (1952). It is this
that upsets us as consumers and intrigues us as investigators. If we are so
suspicious of persuasion and its techniques, how are we “engineered” to
give consent? Bernays referred not only to the relatively harmless influence
of advertising, but also to the more baneful effects of political rhetoric,
including propaganda. The latter term itself can be given some scrutiny.
Originating in a religious context, with positive connotations (the
“propagating” of the faith), the term propaganda eventually came to mean,
viewed from the perspective of another religion, a form of improper influ-
ence or pressure. Hence propaganda today is exclusively a pejorative term
(though Bernays makes a plea for its rehabilitation).3 One question is
3. Curiously, the Compact edition of the Oxford English dictionary ([1933] 1971) has a
definition of propaganda that is wholly without negative connotations—quite differ-
ent, I think, from its normal use: “Any association, systematic scheme, or concerned
movement for the propagation of a particular doctrine or practice.”
whether the term really has true denotative meaning, as a special kind of
persuasion identifiable even when the argument is one with which we are
in sympathy; or whether propaganda simply means emotional persuasion,
when the argument is not one we approve of. We would agree that pro-
paganda is mainly applicable to forms of persuasiveness that utilize emo-
tion, usually of a high intensity, often invoking fear and irrational desires.
But advertising typically does this, yet is not considered propaganda. So
we might want to add the proviso that propaganda concerns changes in
beliefs, rather than concrete buying habits alone. Yet we might argue that
much advertising is propagandistic—not only the “public service” advertis-
ing and institutional advertising by power companies, but even advertise-
ments for soft drinks that suggest indirectly that youth and a svelte body
are minimal conditions for being allowed to exist, or that create and rein-
force all sorts of traditional sexual and other stereotypes. This then sug-
gests another facet of propaganda: it is normally indirect. It is not present
in the explicit message, but somewhere in the presuppositions. When we
are serenaded to the effect that “Coke adds life,” the propaganda is not in
the admonition to buy Coke, but in the inference that youth is desirable
(and young people drink Coke).
This brings us back a bit roundaboutly to the relation between the
Gricean maxims and discourse types. I have argued that the Cooperative
Principle (CP) is more directly applicable to the classroom lecture than to
any other type of discourse—certainly more than to ordinary conversation.
But even OC makes reference to the maxims, if only via implicature. And
in OC, we understand that flouting of the maxims is due to our desire to
adhere to more socially (as opposed to intellectually) relevant rules, rules of
Rapport: when we have a choice between being offensive and being unclear,
we invariably choose the latter, and a majority of cases of OC implicature
can be seen to stem from this assumption.
But in persuasive discourse, the situation is rather different. Whereas
in OC and the lecture, our aim is to inform—at least, it is so, other things
being equal—in PD our aim is, of course, to persuade. The politician does
not especially want a knowledgeable electorate: he wants votes. The adver-
tiser is not interested in educating people about hygiene: he wants to
sell deodorant. It is not that the maxims are violated only in case a more
peremptory need intervenes, as in OC—they are regularly infringed with-
out explanation, cue, or apology.
Indeed, it might be argued that the PD need for novelty alone is respon-
sible for many instances of violations of the maxims, especially in advertis-
ing. For in OC, we have seen, novelty is not especially valued. Part of the
[ 194 ] Context Counts
195
reason for this is that familiarity aids intelligibility—that is, aids in the
keeping of the Cooperative Principle. What is new and requires interpreta-
tion is in violation of the maxim of Manner (be clear). But in PD it is this
very violation that is striking, memorable—efficacious.
If we understand this preference for novelty as an intrinsic aim of PD,
we can perhaps understand better why advertisers cling to certain formu-
las despite—or rather, because of—the contempt to which they are sub-
jected by critics. For example, many of us remember a slogan from the
fifties: “Winston tastes good like a cigarette should.” The criticism heaped
on the hapless preposition was staggering, yet the commercial continued
to appear in that form. (Lip service was later paid to linguistic chastity by
the addition, after the infamous slogan, of the rejoinder, “What do you
want, good grammar or good taste?,” which was cold comfort for traditional
grammarians.) We can understand the company’s clinging to the solecism
if we understand it as a Manner violation: focusing the hearer’s (or reader’s)
attention on style tended to obscure content, and thus to flout Manner, if a
bit indirectly. Indeed, anything neologistic will have much the same effect,
and will serve as good persuasion for two reasons: first, because it vio-
lates Manner as just explained, and thus attracts the audience’s attention;
and second, in the case of obvious neologism—as opposed to the simple
ungrammaticality of like for as—it forces the audience to interpret—as any
violation of the Cooperative Principle does.
It is axiomatic among proponents of all forms of persuasive or semi-
persuasive discourse (e.g., psychotherapeutic discourse and the classroom
lecture) that if the audience can be made to participate at some level—
that is, to function as addressee, not wholly as audience—learning, or per-
suasion, will be much more successful. Hence, if the audience is forced to
interpret neologisms, or relate “ungrammatical” forms to their textbook
version, they will probably remember better. And memory is the name of
the game in PD.
We find violation of the Cooperative Principle not only in Manner, but
in other maxims as well. In terms of neologisms and other sorts of linguis-
tic novelty—most easily considered as Manner violations—we find:
4. An added benefit of this form of utterance is that it mimics a casual register and
thus suggests folksy informality. Comparison of these segments with true OC casual
register makes it clear, however, that the truncations found in commercials do not
occur in OC. One more advantage, of course: a micro-millisecond saved is a tidy sum of
money earned, given current rates on prime-time television.
[ 196 ] Context Counts
197
5. Pragmatic novelty:
(This includes anything aberrant about the discourse form itself.
In particular, since many commercials are framed as minidramas,
we see many unusual bits of dialogue within these thirty-second
segments. One such type much in vogue lately is the following.)
—You still use Good Seasons Italian?
—Not any more!
—No?
—I use new, improved Good Seasons Italian!
—Fill it to the rim!
—With Brim! [laughter]
5. The use of your here is typical of many kinds of commercials (preceding the name
of the sponsor’s product, or its generic category, as here) and is characteristic of the
genre and not OC. Normal here is “another cup of coffee,” perhaps “that coffee.” Your
in this environment (a quantity violation in that it provides unnecessary information)
perhaps is intended to suggest the addressee’s responsibility for the adequacy of the
product.
with no pause between the clauses. In fact, the stress of this utterance is
what we would expect if we had a syntactic dependent clause at the left
(e.g., “Since I like it so much …”) The intonation here is steady rather than
falling, as would be normal for two syntactically independent sentences.
I think the only way we can make sense of this is as an attention-getting
novelty.
It seems that in the past, most advertisers focused their attention on
lexical and syntactic novelty, while more recently, pragmatic novelty seems
to be favored. What we may be seeing here is metanovelty: the audience
is jaded with the older forms of newness, and advertisers must press ever
onward into the more mysterious reaches of language to get a response
from us at all.
Some of the novelties I have been talking about can be justified on addi-
tional grounds, beyond their usefulness as floutings of some part of the
Cooperative Principle. For example, many of these work for brevity, a real
desideratum when the point must be made in thirty seconds. But I think
that brevity alone seldom justifies any special usage. Rather, all the special
features of PD are overdetermined.
I have illustrated the flouting of the Cooperative Principle with regard to
advertising, in which it is particularly glaring and has, indeed, been exalted
to an aesthetic feature of the medium. But it can be found just as easily in
other forms of PD. Its function in psychotherapeutic discourse is special,
and is not dealt with here. But certainly we are accustomed to it in political
rhetoric—especially in the form of hyperbolic violations of Quantity, and
metaphorical floutings of Manner. It is noteworthy that, of all the forms
of violations of the CP that are found and tolerated in PD, Quality alone
is missing. Quality violations are what bring the FTC down on the spon-
sor, impeachment threats down on the politician. The others seem beyond
reproach. We may want to examine again the assumption that Quality is on
[ 198 ] Context Counts
199
the same level as the other maxims, when in fact we feel that abrogation of
this maxim has very different psychological and social implications. Indeed,
violation of the other maxims can always be justified for either aesthetic
or social reasons, without further ado; but we have only a subcategory of
Quality violations that are tolerable—and only dubiously so: the “fib” or
“white lie.” In fact, Quality violations are the only ones ordinary language
has a separate word for—and a word with bad connotations at that.6
I have referred to the importance of Rapport rules and strategies in OC,
and their superseding of the Cooperative Principle when they conflict with
it. Since PD is nonreciprocal and has an audience, rather than participating
addressees, it is not surprising that Rapport as such plays no role. We do
find, in some types of commercials, attempts to establish a one-to-one rela-
tionship with the audience on television: eye contact and twinkling, casual
register, and the like. But in fact, most of the conventions of OC that exist
for Rapport purposes are absent from commercials, and from most forms
of PD as well—openings and closings, for example, are either totally miss-
ing or heavily truncated (as in psychotherapeutic discourse).
All this discussion raises a very troubling issue in discourse analysis of
the kind I am attempting here: What is the role of the CP in persuasive
discourse, and especially in advertising? I have spoken as if the CP were
involved in our understanding of commercials as it is in OC or the lecture—
and yet we saw it was utilized very differently in even these two types of
discourse. Certainly the CP gives us a much-needed and illuminating han-
dle on the workings of commercials and our understanding of them. But at
the same time, invoking the CP as the basis of our understanding of a dis-
course type makes an implicit claim: since the CP enjoins us to be as infor-
mative as possible, to say that our discourse is understood in relation to it
is to imply that the true underlying purpose of that discourse is to inform;
and that, if information is not exchanged in some utterance in an opti-
mal way, there is a special reason for it—politeness, for instance, in OC.
This assumption works splendidly for discourse that is explicitly intended
above all to inform—that is, of course, the lecture. But even with OC, it
gets us into trouble. For the main purpose of OC is not to inform, or even
to exchange information. This can occur, and conversations frequently are
loaded with useful information, but most often the information serves
only as a sort of carrier, enabling the real business of the interaction to get
[ 200 ] Context Counts
201
discourse until our classification includes, and compares, form and func-
tion together, and we can hope to have a satisfactory grammar of discourse
only when we have arrived at a valid taxonomy. This brief discussion of
one parameter—persuasive and nonpersuasive discourse—is presented as
a beginning.
REFERENCES
Introduction to “Techtalk
as group talk”
BY SUSAN M. ERVIN-T RIPP
Members’ knowledge of how Members behave is so organized that items of that knowl-
edge may, discriminatively, be taken as expectably descriptive if no more than a single
referentially adequate category has been asserted to hold for some Member in question.
It is that Members have such categorically localized knowledge that provides for what it
is they employ to make the assessments just referred to.
Harvey Sacks, “An initial investigation of the usability of conversational
data for doing sociology” (1972)
The SPARCclassic system with its super aggressive price points is targeted at the client
portion of the downsizing/rightsizing phenomenon that is occurring in the mission criti-
cal business application markets… . This model advertises the most aggressive street
price possible to gain mindshare in the market… .
From a SUN workstation ad in 1992
I n this brief paper, Robin Lakoff raises the question of how technical
writing becomes incomprehensible and brings to bear two other per-
spectives—the social dynamics of in-group talk and the place of gender in
the social construction of talk.
We all know what technical writing is and teach it around the world
as language for special purposes. Students are to use arcane vocabulary
and the syntax of impersonality—the passive and the generalization—to
sound scientific. We intend this kind of language to obscure the voice and
204
[ 204 ] Context Counts
205
features according to age and gender. Labov’s famous paper (1963) on why
young people on Martha’s Vineyard who wanted to stay on-island had dif-
ferent vowels than those aiming not to stay, even to the point of hyper-
correction, is a good example of these factors, and more recently Eckert’s
(1989) work on high-school cliques analyzes phonological effects of social
structure. You could talk often to someone in an emotionally neutral, busi-
ness relationship, or be close to a lover or grandparent with whom you
don’t identify, or identify with someone you don’t know personally.
It is a speculation that speech similarity and writing similarity have
the same social dynamics, but presumably modeling at a distance could be
more possible in writing. That is, without meeting a writer, one could read
enough to identify with him or her and emulate stylistic features. Could
the reverse happen; can copying lead to identification? We know students
can copy styles; this is a standard assignment in some writing classes. How
writing is related to identification remains to be studied.
Specialized language develops—and here I mean especially vocabulary—
whenever a group speaks more to one another than to nonmembers, or
whenever there are new activities or contexts. This happens in both closed
networks of close friends and in open networks in cities. We see these sim-
ilarities whether with Valley Girl accents, group slang, or morphological
or syntactic preferences. The spread of such features is primarily through
contact and emulation. Rampton (1991) even notes borrowing of Pakistani
words and phrases for particular functions by English teenagers as a sign
of group solidarity.
As Lakoff points out, the semantic utility of new vocabulary is only a
part of its function. Nobody is surprised by terms like phonology or mor-
pheme, DNA or neutrino. These terms were necessary for progress in talking
about new concepts and means of analysis. Technical terms when needed
are of high value, and hijacking of them by outsiders creates problems. An
example is the theft of neural, which at one time referred to the nervous
system, but now appears in electrical-engineering and machine-tool jour-
nals. Now biologists use the phrase neuronal network instead.1
What Lakoff points out, however, is that the new vocabulary can go well
beyond what we need to identify new concepts, into renaming old concepts
by new names. What does this accomplish? Most obviously, it marks the
boundary between insiders and outsiders.
1. Even word choices have a social component. Quark (from Finnegan’s Wake) entered
physics because of the literary tastes of Murray Gell-Mann. His preference defeated its
competitor ace, which was offered by a physicist of less prestige.
D O U B L E TA L K : S E X I S M I N T E C H TA L K [ 205 ]
206
[ 206 ] Context Counts
207
REFERENCES
D O U B L E TA L K : S E X I S M I N T E C H TA L K [ 207 ]
208
[ 208 ] Context Counts
209
CHAPTER 9
Doubletalk
Sexism in tech talk (1983)
LANGUAGE CODES
[ 210 ] Context Counts
211
TECHNOLOGIZING THE HUMANITIES
not unlike the church, the army, medicine, law, or government. We would
expect the chosen code of the profession to reflect this bias, and indeed it
does. In Western culture, the stereotypical conventionalization of behav-
ior toward which men are generally socialized is distancing: a power-based
strategy involving impersonality and nonemotionality, the denial of the
interaction and its effects on the participants.
Women, on the contrary, have traditionally been encouraged toward a
deference strategy, oriented not to power, but to acceptance or love, and
the fear of its loss. Hence, women’s language allows for openness to others’
needs and a reluctance to acknowledge one’s own needs. But as a strategy
concerned with acceptance and rejection, emotional expression is valued
over intellectual clarity in case of conflict.
I am not suggesting here that either of these strategies is in any sense
better, or that the users of either are better people; and less still that it is
male and female biology or physiology that motivates us in this direction.
But I do suggest that these have been, and for most people still are, the
prevalent patterns. And even those of us who have struggled hardest to
rid ourselves of stereotypical behavior often find ourselves confronted by a
society that still holds conventional models up to our children.
What is important in all this is that technical codes, as they were
described earlier, can be understood as exaggerations of masculine distanc-
ing, or power-oriented style. Hence bureaucratese and its relatives make
outsiders angry precisely because they implicitly involve a power state-
ment: “I (the user) am more powerful and valuable than you (the nonuser).”
And since it is implicit in the style, not explicitly present, it cannot be nego-
tiated. So, as outsiders faced with this kind of language, we feel impotent
and resentful.
Academese is no different. As the language of the male power elite, it
continues to be used in a way that excludes women—and probably those
male members of minorities, where the white-male stereotype is differ-
ent from their own. And it does so with the hidden implication that any-
one who cannot use the language is unworthy of membership in the club.
What is illegitimate in this argument is that the selected language is itself
exclusionary on two levels. First, it excludes, in the way any code does,
those who are not yet full members and, second, since it is based upon an
exaggeration of a male style of interaction, it is relatively easier for men
to assimilate. This is, after all, what graduate and professional schools are
largely set up to accomplish, but the burden on women is greater. This isn’t
deliberate racism or sexism, but it works to exclude women and minorities
from positions of influence and authority on what are presented as valid
intellectual grounds. Affirmative action doesn’t touch this problem, which
[ 212 ] Context Counts
213
[ 214 ] Context Counts
215
their doubts about whether their work is masculine enough, since it isn’t
bringing home the NSF bacon, and about whether they can compete with
women once the barriers are down. Therefore the style of discourse in the
humanities has become scientized without any concomitant scientization
of content. Hence, humanists feel still more uncomfortable: What if they
are found out in their little game? So, men in the humanities are creating
a secret language all their own, to shore up their own sense of being real
and legitimate in the face of doubt. They draw the inner circle ever tighter,
enforcing the shibboleths ruthlessly and continually raising their require-
ments, linguistic and otherwise, for belonging to the club.
Interestingly, the situation we are speaking of as current in the univer-
sity has analogies in older cultures. In the contemporary case, the language
used by both men and women is the same, English. But during the medi-
eval period in Japan, a remarkably similar pattern seems to have existed.
There was an indigenous language, Japanese, which had not long before
been given written form in the katakana syllabary. At the same time,
Chinese was extremely influential in Japan, both in terms of cultural and
linguistic influence, and was considered by the Japanese to be the lan-
guage appropriate to literate expression. But women were not educated
in Chinese during this period; this kind of literacy and familiarity with
another society was considered men’s prerogative. Therefore, there was
no literature in Japanese written by men during this period. What men
wrote were imitative and stilted exercises copying Chinese sources in infe-
rior Chinese. But women, relegated to the vernacular, developed a lively
literature in Japanese, which endures today, recognized as a much stronger
form of writing than the men’s, largely because it was intended as a means
of communication, rather than show.
Returning to the contemporary situation, we find that this divisive and
ominous trend toward the technologization of language is not limited to
academic, bureaucratic, or professional parlance; sadly, it is becoming char-
acteristic of our literature as well.
the police make arrests, detain suspects, interrogate them, and—by means
which vary in their legal and ethical legitimacy—obtain statements, it is
with the objective of producing evidence that can be presented in court.
When people are sentenced to periods of imprisonment, it is as a result of
the evidence presented on both sides; the availability and quality of legal
advice, representation, and interpretation available to the accused; and
the expectations and beliefs of the magistrates, judges, or jury, which are
products of their culture and life experiences. If all African Americans had
access to O. J. Simpson’s legal team, would their ethnic group still make up
over 40% of those on Death Row?
Since lawyers and judges are still overwhelmingly male, white, and
middle-or even upper-class, while the defendants are not, there are likely
to be acute differences in language and discourse style between the parties.
This is important, because everything that takes place in the courtroom is
mediated through language.
Of course, not all evidence is verbal in nature, but nonverbal evidence
must be introduced to the court by verbal means. When Sergeant Stacey
Koon and Officers Lawrence Power, Theodore Briseno, and Timothy Wind
were charged in 1992 with beating Rodney Glen King, the prosecutor pre-
sented the amateur video of the incident with the words “What more could
you ask for?” However, police counsel provided the jury not only with plas-
tic overlays outlining King’s body in white on the frames but also with a
set of verbal categories to analyze the sequence of events (Goodwin and
Goodwin 1997). Repeatedly freezing the frame and focusing the court’s
attention on King rather than the officers, the defense invited the jury to
view the tape through the discourse of the police, labeling their weapons as
“tools” and categorizing the beatings into ten distinct “uses of force,” each
with an escalation and a de-escalation triggered by King’s movements on
the ground. The repeated evidence of an expert witness that these move-
ments were “starting to be” aggressive apparently convinced the jury that
the amount of violence used was reasonable: at any rate, on April 29, 1992,
they acquitted all four officers, a verdict that led to the Los Angeles riots
of that year.2 The riots in turn led to further abuses: seventeen thousand
people were imprisoned as a result and hundreds deported to Mexico and
2. A year later there was a retrial. Although the defense again tried to argue that King
acted in a threatening manner and was therefore “in charge of the beating,” one of
their own witnesses contradicted this interpretation of events. On April 17, 1993, after
seven full days of deliberation, the jury found Officer Lawrence Power and Sergeant
Stacey Koon guilty of violating King’s civil rights, while acquitting officers Briseno and
Wind. (United States v. Koon, 1993 U.S. Dist. LEXIS 17926 (1993).)
[ 218 ] Context Counts
219
1. To take a tape recorder into court and then make one’s own transcript
from the recordings. This is the most desirable but also the most imprac-
ticable option: tape-recording of proceedings is unlikely to be permitted,
at least in a British courtroom.
2. To make one’s own notes at the time. However, even if the linguist
is competent in shorthand transcription, she is unlikely to be able to
transcribe everything. For psycholinguistic as well as practical reasons,
such items as speech errors, hesitations, discourse markers, repetitions,
and pauses are likely to be omitted, even by a linguist (Blackwell 1996).
Without the aid of a tape recorder, it is impossible to be sure that one’s
transcription is accurate, or to get it checked by a colleague.
3. To rely on the transcript made by the official Court Reporter, which may
or may not be based on a tape recording. While accurate enough for legal
purposes, the same kinds of omissions are likely to occur as in (2) above
(Walker 1990). In any case, solicitors often have to pay substantial
Lakoff does not say how she obtained her courtroom data, but it seems
likely that she relied upon her own notes. With the benefit of what we now
know about the deficiencies of transcription techniques, we might well ask,
for instance, how she could be so sure that she had not missed any simul-
taneous speech:
Indeed, the hedges “virtually” and “perhaps” suggest that Lakoff herself
was not so sure. Nonetheless, one gains the impression that she instinc-
tively knew what sort of discourse features to look for. Now that accurate
transcripts of courtroom interaction are gradually becoming available, lin-
guists who have access to them should rise to the challenge and test Lakoff’s
observations on this one case against a larger and more reliable corpus.
The second significant contribution that this paper makes to the study
of language and the law is its attempt to articulate exactly what it is that
makes legal discourse different from ordinary verbal interaction, e.g.:
and
“It is extraordinary that the one participant in the discourse who must be silent
[i.e., the jury] is the one who, in the end, plays the truly significant and predomi-
nantly active role… .” (p. 228)
[ 220 ] Context Counts
221
of such a framework using models that have, by and large, stood the test of
time: those of Grice, Gumperz, Schegloff, and Searle.
The third outstanding feature of this paper is its attempt to illuminate
how inequalities of power in the courtroom, such as those with which
I opened this piece, are manifested linguistically. It is interesting that Lakoff
compares the language of the courtroom to that of the classroom with regard
to the nonreciprocity of speaker/hearer roles in both (p. 227). Other linguists
have noted further, more specific features that the two genres have in com-
mon: for instance, Eades (1996) demonstrates that courtroom interrogation
is typically not dyadic but triadic, with the lawyer repeating part of the wit-
ness’s answer by way of evaluative feedback in a way that is reminiscent of the
Initiation-Response-Feedback structure of a classroom-teaching exchange
(Sinclair & Coulthard 1975). Interestingly, the third lawyer “feedback” move
is one of the items frequently omitted from the official transcript, probably
because it contributes nothing in terms of content from a legal point of view.
Another feature noticed here by Lakoff that is shared with “teacher talk” is
the asking of questions to which one already knows the answer.
Lakoff’s now notorious claim (R. Lakoff 1975) that women use more tag
questions than men, thereby indicating their need for (male?) approval,
has been much criticized by other feminist linguists (e.g., Dubois & Crouch
1976; Cameron, McAlinden, & O’Leary 1989). In “My life in court,” how-
ever, Lakoff makes it clear that while she does indeed view tag questions as
“disempowering devices” in ordinary conversation, their role in courtroom
discourse can be the very opposite of this:
An utterance type normally used to hand power over to the addressee has been
utilized here by the prosecutor to underline and reinforce his own already explicit
power; what is ordinarily an invitation becomes an act of coercion. (p. 230)
Lakoff should be given some credit, then, for acknowledging that tag
questions are a complex phenomenon that can function in a variety of ways
depending on social and linguistic context. That they are even more complex
than she realized is beyond doubt. Michael Cooke, relating his experiences
of interpreting at an inquest in the Northern Territory of Australia, states:
An example:
Counsel for the Police: But the old man didn’t go in the boat,
did he?
Witness: Yes.
Counsel: I beg your pardon?
Witness: Yes.
Interpreter: Yes, he’s affirming [that] he didn’t go in the boat.
Coroner: The old man didn’t go in the boat.
2nd Counsel: He’s answering you exactly on point.
Coroner: You ask these questions that way and that’s what you get.
(Cooke 1995 : 81)
Imagine the prosecutor hurling tag after tag at the reluctant respondent, and
think of the latter’s tricky position. If he replies, he damages himself. But if he
doesn’t … he risks labeling himself conversationally inept: he shows he doesn’t
know how to respond to a tag question, which demands a response. (p. 229)3
O’Barr and Atkins had argued in 1980, from their studies of courtroom
testimony, that so-called “women’s language” was actually a marker of sta-
tus rather than gender, being used by low-status men and avoided by high-
status women. It now appears that tag questions are best omitted from any
catalogue of either “women’s language” or “powerless language,” but credit
for leading off the whole debate must go to Lakoff.
This article, of course, has its limitations. The exchanges presented here
for discussion are mere snippets of courtroom discourse, probably not tran-
scribed verbatim (whatever that means: see Eades 1996), and little linguis-
tic or nonlinguistic context is provided with them. Although Lakoff was a
juror in this case, we are not given any glimpse of the jury deliberations,
probably for legal reasons. And of course, as Lakoff herself acknowledges
3. Lakoff does use generic he in her writing; but since the defendant in this case was
male I think the pronouns refer more to him than to a hypothetical generic respondent.
[ 222 ] Context Counts
223
at the outset, it is only one case and is not presented as “what necessarily
always happens in court” (p. 226).
One can also, with hindsight, criticize some of Lakoff’s linguistic analysis
and sociolinguistic comment. The position with which I would personally
take most issue is her claim that “we should think long and hard before we
dispense with [the pomp and circumstance of the courtroom and its commu-
nications] in the name of ‘accessibility’ ” (p. 232). An ever-increasing num-
ber of people processed by the courts—the pieces of meat in the “sausage
machine,” to use Goldflam’s (1995) metaphor—do not have English as their
first language, and may not have standard British/US/Australian English as
any of their languages. This does not, unfortunately, mean that they will
necessarily be offered the services of an interpreter, as Goldflam (1995) and
Cooke (1995) explain at some length. For such people, who may well be at a
disadvantage already because of their ethnicity, the strangeness of English
legal discourse may reduce their chances of getting justice still further.
To say that the court system is “a basically conservative structure” is an
understatement. When Lakoff refers to the courts’ “symbolic function as
the upholder of traditional mores, the enforcer of law and order” (p. 231),
I am moved to respond: whose traditional mores, whose law, and whose
order? For the powerless groups I have described above, the archaisms of
courtroom discourse do indeed express “the strength of the justice system
as rooted in tradition,” and for that very reason they intimidate, alienate,
and oppress.
Yet again, however, it is Lakoff who has set the ball rolling by attempting
to characterize exactly what it is that is strange about courtroom discourse.
The insight and analysis that she brought to bear on her “life in court” have
proved extremely suggestive for further research, some of which I have
attempted to describe above. Lakoff is one of a number of scholars who,
over the past decade or two, have sought to bring together the fields of
law and linguistics. The results of that union have been fruitful indeed,
and some of the work now emerging from it should provide the basis for
empowerment of those who, because of their gender, race, social class, or
linguistic background, currently experience the legal system as a “sausage
machine” that mutilates their human rights.
REFERENCES
[ 224 ] Context Counts
225
CHAPTER 10
My life in court (1986)
1. INTRODUCTION
[ 226 ] Context Counts
227
one member of the defense team (who actually got a good deal better as he
went along), the attorneys on both sides were articulate and organized, and
insulted the intelligence of the jury rather less than I would have antici-
pated. The jurors, drawn from the proverbial cross-section of the commu-
nity, were interested, dedicated, and responsible from the start. During
deliberations, I was impressed with their intelligence in being able to grasp
and use complicated instructions (given only orally); their articulateness in
arguing for their positions; and their compassion. As the presiding official,
the judge in a very real way set the style for everyone else. His lucidity in
explaining things to the jury, coupled with his very real sympathy for what
they were going through—demonstrated in many ways throughout—made
the experience memorable.
As a linguistic event, a trial is unique in many ways. I will organize my
remarks about courtroom discourse under six subheads: (1) the trial as a
kind of discourse situation; (2) special functions of speech acts; (3) con-
versational rules; (4) conversational logic; (5) ritual and spontaneity; and
(6) the creation of sympathy—persuasion.
[ 228 ] Context Counts
229
Similarly, speech acts may play unusual roles in this situation, related to the
special powers and functions of the participants. Tag questions have been
discussed at length, largely from the perspective of their role in ordinary
conversation, where they may function as disempowering devices: their
use invites the addressee to control the conversation, both in terms of
defining content and holding the floor (for more discussion, see R. Lakoff
1985). But if we superimpose the special power dynamic of the courtroom,
we see that even a form such as the tag, with its clearly specified pragmatic
functions, can have quite different implications. Consider that a tag, in its
ordinary-conversation function, is an elicitor par excellence of response
from an addressee. In dyadic discourse, in which participants really or con-
ventionally are equal in power, this is most often felt as an opening up of
the floor. But now imagine a prosecutor cross-examining a defendant who
has already been found guilty of murder with special circumstances. The
attorney clearly has the power; he has no need to invite a response, since
the witness is compelled by law to reply. And the questioner, as we have
seen, doesn’t even need the information himself, so the addressee can’t
do anything for him. Imagine the prosecutor hurling tag after tag at the
reluctant respondent, and think of the latter’s tricky position. If he replies,
he damages himself. But if he doesn’t, not only does he risk a citation for
contempt but (probably worse) he risks labeling himself conversationally
inept: he shows he doesn’t know how to respond to a tag question, which
demands a response.
Thus, the prosecutor, by placing him in a double bind, drew from the defen-
dant “crazy” behavior, losing him any possible sympathy from the jury and
earning him a warning from the judge. An utterance type normally used to
hand power over to the addressee has been utilized here by the prosecutor
4. CONVERSATION ANALYSIS
Of all types of discourse, a trial is one in which “the truth, the whole truth,
and nothing but the truth” is explicitly and exclusively enjoined as permis-
sible. That is, the Cooperative Principle is explicitly and formulaically stated
to be in effect. Here more than anywhere information is what is sought, and
anything clearly noninformational (in some sense of the term) is imper-
missible. Further, more than is ordinarily the case, the participants in the
business normally have no prior relationship with one another, and there-
fore the CP violations and implicatures that intimacy allows do not occur.
[ 230 ] Context Counts
231
The automobile she had for sale—did she place an ad in the paper for the pur-
pose of selling it?
With regard to the sale of the truck, was there anything unusual about the sale
of the truck?
(We might note here that ordinarily, when the judge sustained an explicit
objection, “Thank you” would have been inappropriate: it was part of
his job.)
Finally, we can look at some of the more obvious ways in which lawyers do
their job: winning sympathy from the jury for or against the defendant by
[ 232 ] Context Counts
233
8. CONCLUSIONS
We see, then, from these few examples, that in the courtroom language
is used differently in many significant ways than it is elsewhere, and all
participants must learn new rules and new interpretations of old ones.
But at the same time, we necessarily bring to the courtroom our familiar
grammars; these may be used or abused by the various participants. It is
REFERENCES
Bernstein, Basil. 1973. Class, codes and control. St. Albans, UK: Paladin.
Garfinkel, Harold. 1967. Some rules of correct decisions that jurors respect.
Studies in ethnomethodology, by Harold Garfinkel, 104–15. Englewood Cliffs,
NJ: Prentice-Hall.
Grice, H. Paul. 1975. Logic and conversation. Syntax and semantics 3: Speech acts, ed.
by Peter Cole and Jerry L. Morgan, 41–58. New York: Academic Press.
Gumperz, John J. 1982. Discourse strategies. Cambridge, UK: Cambridge
University Press.
Lakoff, Robin Tolmach. 1985. The politics of language. CATESOL Occasional Papers
11:1–15.
Schegloff, Emanuel A. 1972. Sequencing in conversational openings. Directions
in sociolinguistics, ed. by Dell Hymes and John J. Gumperz, 346–80.
New York: Holt, Rinehart, and Winston.
Searle, John R. 1969. Speech acts: An essay in the philosophy of language. Cambridge,
UK: Cambridge University Press.
[ 234 ] Context Counts
235
“
T he way we were” is a diversely textured social and intellectual history
of fifteen exciting years (1962–1977) in the development of modern
syntactic theory. We get an insider’s view, unabashedly subjective, in a deli-
cious stew of metaphors: theoretical debate seen alternately as combat,
seduction, and religion. It should be required reading in every history of
linguistics course. Of course, I don’t buy every word of it; it’s as hyperbolic
and permeated with procedural metaphors (e.g., replace, change, delete,
choose, operate) as generative semantics ever was, and while this re-creates
the flavor of the times Lakoff1 chronicles, it’s still as likely to lead to misun-
derstanding as it was in the old days.
My working title for this note was “Highly abstract snake oil: An appre-
ciation,” in honor of one of those delicious metaphors in Lakoff’s essay.
But much as the thought of using that title brought back the flavor of the
1. In the absence of instructions to the contrary, I use standard academic style for
personal references here, noting that until I made an issue of it with reference to
the publication of my review of R. Lakoff 1968 (Green 1970), women linguists were
referred to in the pages of Language as “Mrs. Husband’s-last-name.”
236
VALEDICTION
What did finally become of generative semantics? Did it fold its tents and
steal quietly into the night, as Lakoff implies at the end of section 3? Did it
implode under pressure of self-induced marginalization from topical exam-
ple sentences and cute titles, as Newmeyer (1986:137) has implied? (In
general, it is not true, as Lakoff asserts, that generative semantics evolved
into functionalism, or that the central figures adopted functionalist agen-
das. Some (e.g., Schmerling, Abbott, Dowty, Green, Postal) eschewed the
demonizing of formalism and, finding it empowering instead, went on to
work in formal semantics (incorporating as much pragmatics as they could
into it), or formal syntax constrained by an explicit compositional seman-
tics.) Was it that the effort of trying to represent syntactically significant
pragmatic properties of sentences within the constraints of the available
formalism for syntactic structure caused generative semantics to collapse
of its own weight? The details can be found in articles by Morgan and
Sadock with the sort of typical generative-semantics titles (“How can you
be in two places at once when you’re not anywhere at all?” Morgan (1973);
“The soft interpretive underbelly of generative semantics,” Sadock (1975))
that the curmudgeon-in-chief found so embarrassing. (Wouldn’t I have
liked to be a fly on the wall to see how Fritz Newmeyer reacted to being
called a curmudgeon by a female Presence a couple of years his senior!)
Lakoff characterizes these papers as despairing, but this only goes to show
[ 236 ] Context Counts
237
how personal a ground-level view of events must necessarily be.2 Some lin-
guists took inspiration from them3 and embarked on a program to tackle
questions of pragmatics head-on, with an eye to articulating a language-
independent account of pragmatics that would be compatible with a theory
of linguistic structure that was independent of use.4
Science: Values
Perhaps the aspect of this article that I like best is the support it provides
(e.g., at the end of sections 1 and 2.4) for some observations Thomas Kuhn
was criticized for making,5 to the effect that theoretical disputes may be
unresolvable to the extent that values which form the foundations for prem-
ises in arguments are not shared, since what is valued is beyond reasoning
(i.e., not generally subject to alteration as a result of hypothetico-deductive
argument—this is what Kuhn means by saying that “the superiority of
one theory to another is something that cannot be proved in the debate”
(1970:198)).6 Often, of course, it is not acknowledged, or even recognized,
that incompatible points of view, proposals, and conclusions devolve from
unshared values. The combatants in the “linguistics wars” can’t be blamed
for failing to make their goals, hopes, aims, and beliefs explicit, because
as individuals they took it for granted that their particular values were all
universally shared.
Science: Methods
2. Cf. O’Meara 1996.
3. E.g., Morgan, Green, Horn, and Lakoff herself.
4. See Morgan 1975 for discussion of some of the difficulties they faced.
5. Apparently by scholars who lost track of the premise that science is not a math-
ematical function but an activity of human beings.
6. Cf. also these remarks:
There must also be a basis, though it need be neither rational nor ultimately cor-
rect, for faith in the particular candidate chosen. Something must make at least
a few scientists feel that the new proposal is on the right track, and sometimes
it is only personal and inarticulate aesthetic considerations that can do that.
(1970:158)
T H E WAY W E W E R E [ 237 ]
238
for the details of the cause and effect. In any case, I have to take issue with
the assumption that pragmatics cannot be studied scientifically. As I have
discussed elsewhere (Green 1995), the questions we want to answer dic-
tate the means we will use to eliminate wrong answers, but nothing in the
nature of pragmatics precludes approaching pragmatics questions as scien-
tifically as any others. As human beings, we cannot help but form hypoth-
eses, and once we have a hypothesis, we make a beeline for data that will
test it, and either corroborate it or disconfirm it. (We do this not because
we are following some prescribed scientific method, but just because we
have operational minds and act rationally.) The hard part of research, as
always, is figuring out which questions to ask, and being able to roll with
the punches, and adjust both the question and the means for testing it, as
preliminary results tell us more about the domain of inquiry.
PERSPECTIVE
The sections of Lakoff’s essay all begin with epigrams (in three languages).
My favorite is the quote from Santayana, which has been used elsewhere
(Green & Morgan 1996) for the same purpose. While generative seman-
tics may be today no more than a colorful backwater in the history of gen-
erative grammar, at the time of its ascendancy it saw itself as a natural
development of the standard (“classical”) theory, despite getting a cool
reception from Chomsky. Nonetheless, many of the leading ideas of gen-
erative semantics have been increasingly incorporated in work developing
Chomsky’s more recent elaborations—see Pullum (1989) for a detailed
account of where generative-semantics ideas show up in the more recent
history of current incarnations of transformational grammar. Hold on to
your hat when you read “The way we were”; you’re in for a ride!
REFERENCES
[ 238 ] Context Counts
239
T H E WAY W E W E R E [ 239 ]
240
241
CHAPTER 11
2. Readers should note that the intended audience for this paper is diverse. It was
originally written for the first-year graduate students in a proseminar I was teaching,
and some of the background information reflects this origin. On the other hand, I sus-
pect that in some cases I have presupposed knowledge on the reader’s part of particular
[ 242 ] Context Counts
243
points of linguistic theory that are not part of the linguist’s current armamentarium.
I would hope that readers would try to piece things together, skimming over what they
already know, or not worrying too much about what is obscure, in order to get the gist
of what I’m trying to say.
A significant part of the attraction of CTG, and its meteoric rise, lay in
its presentation of a formal linguistic theory, the promise of a complete and
rigorous model. It was covertly assumed that all the properties of language
could be described by a system that utilized dichotomous choices, as a for-
mal system must. The “justification” of the idea that a formal system of this
kind was adequate for linguistic description mainly consisted of the sorts
of “fragments” of grammars of particular languages (most often English,
or anglocentric versions of other languages) found in Aspects. While this
showed that, for carefully selected cases, a particular formalism could work,
it didn’t show that it could work for the grammar as a whole, much less that
it really captured what was going on. The assumption was that the frag-
ments we had in 1965 would, in a couple of years, increase and meet one
another to form a complete grammar of English. Students would then take
this grammar and apply it to many “exotic” languages, shortly producing
complete grammars for each and all. The distillation of this effort would be,
in the not-too-far-distant future, an inductively derived universal gram-
mar which would (undoubtedly) jibe smoothly with the deductive version
then being discussed, based on the way language “had to be” due to the
theoretical assumptions of CTG. It was a heady promise and we all believed
it wholeheartedly. But at least some of us wanted to be more precise about
the details of this grammar than Aspects or its author allowed.
Important at this point and later, overtly and implicitly, was the “para-
digm argument” of CTG, the way in which DSs could be justified: both the
idea of implicit underlying structure, and particular realizations thereof.
To avoid phenomenological criticism about postulating forms that were in
some sense interpretively derived, theorists had to offer a rigorous means
of deriving the abstract from the concrete, as well as reasons why this
potentially dangerous step was justified in terms of the increase in gener-
alizing power of the grammar. These arguments were made elegantly in a
1964 article by Paul Postal. He took an example familiar and comfortable
not only to transformational grammarians and structural linguists, but to
grammar teachers everywhere: the idea that, intuitively, English impera-
tive sentences were not, at heart, “subjectless” as they appeared; but, in
order to explain what could and could not occur as imperative sentences,
one had to “understand” a second-person subject not ordinarily accessible
to superficial observation. Postal’s example was inspired, as the presence of
a “you” in our understanding of imperatives was uncontroversial; his leap
was to build that intuition into the syntactic grammar, to justify the for-
mal, systematic existence at a specified level of structure of entities whose
presence could only be discerned indirectly. He did it by arguing that such
an assumption simplified the grammar, allowed its rules to be more general
[ 244 ] Context Counts
245
[ 246 ] Context Counts
247
men had perfectly good vision, but the elephant was behind a screen, per-
haps less elephant than chimera.
But the parable applies to the situation: the question of the nature of
language and therefore the linguistic theory needed to capture it. As I said,
two kinds of people entered the field, in those first days.3 One group were
basically mathematicians and logicians, by temperament if not by trade.
Their fascination with language was in seeing it as a quasi-mathematical
system, in isolation, like the systems of topology or algebra. They were con-
cerned with predictable regularities, patterns that recurred, and the for-
malisms necessary to capture those generalizations. To show that language
functioned in this way would be to make a deep claim about the logical
capacities of the human mind, to give deeper and more rigorous meaning
to the Cartesian claim that man was a reasoning animal and that those for-
mal rational capacities were intrinsic to and universal in—that is, provided
a definition of—humankind.
If that was one’s aim in doing linguistics, certain assumptions were
natural. You would tend to search for generalizations, stop with the sim-
pler cases on the grounds that they represented the deeper reality; more
complex examples did not necessarily show the system to be wrong, or
entail more abstract or more complex versions, but merely were static,
interference with the deep patterns based on mathematical logic by sur-
face annoyances—other psychological capacities and incapacities, social
involvements: interesting to other kinds of social scientists, maybe, but
off-limits to linguists, irrelevant and uninteresting. And just as, for the
logician or mathematician, the universe could be fragmented up without
distortion into subsystems, smaller worlds, within which generalizations
were more readily accessible, so too language could be seen as a network
of autonomous systems: phonology, syntax, semantics, interdependent
but not formally interconnected. The rules of syntax did not—could not—
mention semantic criteria, and vice versa. Hence, the existence of abstract
3. This dichotomy, like others, is deceptively sharp. It would be erroneous for example
to suggest that the ESTists had no interest in the cognitive aspects of language. This
of course was one of Chomsky’s motivating concerns, and one of particular interest to
Ray Jackendoff among others. Similarly, it would be absurd to suggest that GSists were
uninterested in formalism: George Lakoff and James McCawley in particular were at
pains to develop formal devices and systems and considered these central to GS. I mean
rather that each group tended toward their own emphasis, the end toward which they
strove. EST was—and its descendants remain—chiefly concerned with the description
of language as an autonomous system (which might, significantly, shed light on other,
autonomous, human psychological processes); GS moved more and more in the direction
of seeing language as the reflex of—and inextricably interconnected with—other human
processes, social and cognitive. It is a matter of preferred emphasis more than anything.
[ 248 ] Context Counts
249
interested in, not because they determined that their preexisting interests
were better served by the other side’s model.)
2. THE THEORY
I have touched briefly upon the basic tenets of CTG, and the ways in which
GS differed, and some reasons why the proponents of each felt as they did
and worked as they did. But before we leave the arena of scientific dispu-
tation, we should examine more closely the claims of GS. As I see it now,
although it seemed then that there were myriad aspects comprising a richly
textured conceptual system (as indeed there were), they pattern together
into four major claims, each diametrically opposed to some tenet of CTG,
and all fitting together into a cohesive whole, all interdependent.
We have seen that, in some of his writings, Chomsky held forth the promise
of a syntactically rooted deep structure that, nevertheless, was based upon
the universal rational capacities of the human mind. We have also seen
that, as he exemplified the model DS in other works, it could not have func-
tioned in this way. GS theorists, troubled by the inconsistency, resolved to
reconcile it (as, indeed, Chomsky did later, with Extended Standard Theory,
though in the opposite direction). They took the less-formal Chomsky at
his word: the basis of syntax was logical and universal. But then, all traces
of English-specific features in the DS, or Base Component as a whole (the
latter term unlike the former including the lexicon) had to be eliminated in
favor of forms that could fit in agreeably with the underlying structure of
any natural language. Problems with the CTG model included:
(a) The lexicon. It was often noted that words in one language did not
generally correspond, in meaning or syntactic constraints, to their “syn-
onyms” in any other. The structures in which a causative verb like kill could
be inserted in English were not identical to their Japanese counterparts.
And, clearly, the more figurative uses of words—kill time, a killer exam—
would not necessarily transfer from one language to another. One solution
might be to assume that, in the basic lexicon, “complete” surface-structure
type words did not exist. Rather, what were found were atomic elements,
semantic primes, basic concepts common to all languages which (more or
less language-specifically) could be combined together by transformational-
type processes to form the surface words of each language. Kill in this the-
ory is decomposable into several primes: cause, become, not, alive.
A bigger problem than the composition of individual lexical items was
the nature and number of lexical categories at the basic level. In CTG, to
the extent that the issue was dealt with, the categories in the Base were
essentially those found at the surface. Later, this covert assumption was
explicitly codified by the descendants of CTG: no categories were permitted
in the Base that did not exist in the language at the superficial level; and
no lexical item could change categories by transformational rules (or any
other way).
The first postulate meant that “abstract” entities of various types, with
one or two exceptions, could not exist. Both in Aspects and earlier, Chomsky
permitted two types of abstract markers. One was the “dummy” symbol, Δ,
used to indicate (as for the underlying subject of an agentless passive) an
item that was necessarily deletable because it was semantically not fully
specified, and was not needed to function syntactically at a superficial level.
For instance, the grammar recognized that the underlying subject of the
passive was an NP, and that it was semantically not incompatible with the
verb selected from the lexicon. Since the passive transformation removed it
from subject position, it was no longer syntactically required, and could be
deleted: all relevant semantic information about it could be discerned from
the choice of main verb. The other abstract category was more problematic.
It included a set of items dictating that specific transformations were to
operate on the trees which contained them. Each referred to a category
of sentence types: Imp, Q. Thus, if Q was selected in the DS, in English,
the transformational rule of subject-auxiliary switch was triggered. A late
transformational rule deleted these markers; their only surface trace (mak-
ing them recoverable) was in the transformations they had triggered. It
was never really clarified in Aspects just what sorts of phenomena could be
handled by these mechanisms. By permitting into the DS abstract catego-
ries (both the “dummy” and these categories were chosen from the lexicon,
like nouns or verbs), CTG seemed to promise that other abstract categories
could be justified, as long as they were recoverable from the sentence’s later
transformational history and superficial co-occurrence patterns. These,
then, functioned as an Open Sesame to GS, particularly combined with
Postal’s paradigmatic illustration.
GS argued that abstract lexical items parallel to these types existed, as
long as they could be justified as suggested above. Postal’s discussion of
[ 250 ] Context Counts
251
(b) The nonexistence of VP. CTG followed traditional rules of sentence
parsing in dividing a sentence basically into two major components: NP
(née “subject”) and VP (or “predicate”). This division worked well for an
English-specific DS in which main verb and direct object were more or less
inseparable. But it was obviously less viable for a language in which the
verb was followed by subject, then DO (VSO), or any other such possibil-
ity. As long as the opening rule of the phrase structure (PS) separated sen-
tences into NP and VP, it had to be nonuniversal.
Further, GS began to find arguments for a single underlying word order
(VSO) for all languages quite different from the prevailing SVO order of
surface-structure English. Such an order was essential for a theory involving
[ 252 ] Context Counts
253
(c) Word order. In CTG, the (again covert) assumption was that the word
order postulated for a language in the DS should be as close as possible to
the normal or prevalent word order in surface structure. (In general, CTG
preferred a rather spare transformational component, with as few rules
as possible, working on as few sentence types as possible.) So in the DS,
English was represented in Aspects as SVO, since that is the normal order
of the declarative, assumed to be the “basic” sentence type on statistical
grounds. (Covert again.) But in other languages (as Greenberg (1966) had
demonstrated), there was reason to believe that at a superficial level, many
other either possible or mandatory word orders occurred, and CTG would
treat each of these as having a different DS word order (that is, if the pre-
vailing order in surface declaratives was VSO, that would be taken as the
base order, and so on). One problem CTG couldn’t cope with at all was so-
called “free word-order” languages, where, as we would say now, word-order
was governed pragmatically rather than semantically. (Latin is an excellent
case.) In CTG, such a language could be dealt with only very artificially and
ad hocly: the most “prevalent” surface order was decreed to be the “basic”
order, and generated in the Base—e.g., in Latin, SOV. But the “basicness”
of SOV in Latin was of a very different status from the “basicness” of SVO
in English, since word-order changes in the former did not affect semantic
reference, but rather pragmatic function: vividness, topic, cohesion, etc.
More troubling than the false analysis itself is the fact that CTG theory did
not provide a way to preclude it. A favorite validation for formalism is that
it provides a “garbage detector” for incorrect analyses. This one didn’t.
The postulation of a universal basic VSO order had more than aesthetic
justification. McCawley argued that it was not coincidental that VSO was
analogous to the order of items in the propositions of symbolic logic, where
V = predicate and N (that is, S and O) = arguments. GS suggested that the
[ 254 ] Context Counts
255
NP VP
Floyd Aux V NP
T D N
V NP NP NP
say I to you S
V NP NP
cause Floyd S
V NP
happen S
V NP
not S
V NP
be whole glass
Figure 11.2 One version of the GS Underlying Structure for Floyd broke the glass.
grammar ought to do, which in turn went back to a covert dispute about
the nature of what the grammar was supposed to describe—language.
From the beginning of the schism, one of the favorite insults the tradition-
alists could fling at the rebels was to accuse their grammar of being “too
powerful.” The proper riposte was that the former’s grammar was insuf-
ficiently powerful; or that the GS version was, in fact, no more powerful
than its competitor, if fully understood. Assuming the validity of formal-
ism itself, the criticism was serious if justified: under the assumption of
Occam’s Razor, the best theory is the one which can accomplish all that is
necessary with the least amount of machinery. A too-powerful theory is
one whose mechanisms allow it to do more than is actually needed, and
is therefore uneconomical. Hence, it was important that transformations
be limited both in number and in type: the fewest possible rules, and the
fewest possible kinds of rules, made for the simplest system.
The problem with this assumption, said GS, is that it is difficult or impos-
sible to gauge simplicity over the grammar as a whole. Maybe you can count
rule applications, or features, in fragments of the grammar, but until you
have a complete grammar of a language (by this time, recognized as not
an imminent possibility), you cannot talk sensibly of economy or power.
What might look like an unnecessary efflorescence in the transformational
component when the latter is considered by itself might arguably effect
significant savings overall, by radically simplifying the Base Component.
Indeed, this is just what was argued.
It is true that in GS, the transformational component did a good deal
more work. It did so most obviously because the DS trees needed more
processing by the same rules, cyclically iterated level by level, than did
their equivalents in CTG. (But it was by no means clear that merely hav-
ing rules apply more often made for nonsimplicity of an interesting kind.)
Then too, GS introduced new rules into the grammar: there were prelexi-
cal transformations, for example, which combined the atomic predicates of
the lexicon into the nouns and verbs of the surface structure. There were
rules deleting abstract elements such as the higher performative. There
were global rules. Additionally GS developed a rich stratum of constraints
on the applicability of rules, of which the profusion of cyclical types is but
one instance. There were Ross’s islands (1967); notions of precedence and
command (Langacker 1969); governed rules and major and minor rules
(G. Lakoff 1970); transderivational constraints; and much, much more.
[ 256 ] Context Counts
257
That leads directly to the next point of difference. I have already noted that
CTG and EST were predicated on a notion of autonomous language, as well
as autonomous levels of grammar. In a way there is an irony here. It is well
known that CTG took pains to distinguish itself from structuralism (the
“Bad Guys”) on the phonological level by rejecting the latter’s maintenance
of a level of autonomous phonemics.4 (No such effort was made for syn-
tax, largely because structuralists generally left syntax alone.) In general,
said CTG, levels are nonautonomous, systems are interconnected. But in
fact, CTG (to the extent that it troubled itself with semantics at all) saw
4. It is curious that each side saw the other as relapsing into Bloomfieldian heresy.
While the GSists saw the ESTists as neo-crypto-Bloomfieldians because of their non-
universalist Deep Structure, their belief in the autonomy of levels, and their rejection
of true mentalism, EST saw GS as similarly benighted because of their “empiricism”—a
dirty word in the Cartesian circles of CTG/EST, albeit a curious charge against a group
who were at least as introspective and intuitive in their methods as they were them-
selves. (Cf. Katz & Bever 1976.)
[ 258 ] Context Counts
259
Yes. But “order” is not synonymous with “discreteness,” particularly when imposing
the latter creates distortions. In that case, it is the job of the responsible investigator
to divest his or her mind of outmoded beliefs in what “science” or “theory” must be
(according to whom) and tailor explanation to observations, not vice versa. Granted
that fuzziness is unsettling, even frightening: that is no reason to deny its reality. To
abjure nondiscrete theories because they are unsettling, or because they conflict with
the kinds of formalisms we currently feel comfortable with, is antiscientific in the most
dangerous way: analogous to the Church’s determination that Galileo’s claims were
heretical because they were antithetical to current established wisdom.
[ 260 ] Context Counts
261
change meaning, and when did it not? And if it was doctrine that trans-
formations did not change meaning, was that to be taken to imply that, if
two surface structures were equivalent in meaning, they must (the inverse
of K-P) be taken to have a common underlying source? This was the stron-
gest interpretation of Katz-Postal, and GS essentially adopted it, without
too much explicit consideration. The argument was that it was always in
the interest of economy to derive paraphrases from the same underlying
source: thus, selectional restrictions and other constraints need be stated
only once in the Base. Then, if two sentences were arguably paraphrases
(which was defined as, if neither could be true in a context where the other
was false), Occam’s Razor required and K-P allowed that they share a com-
mon source. But what was a true and complete paraphrase? Was (3) accu-
rately paraphrased by (4)?
These issues turned out to be unresolvable, with GS saying “yes” and CTG/
EST “no,” with no agreed-upon way to decide.6
Moreover, the old active/passive bugbear was understood very differ-
ently by GS and EST. The latter (as discussed by Partee (1971) among oth-
ers) kept to their guns: it demonstrated a deep problem with K-P; and the
introduction in EST of interpretive semantics, essentially the reintroduc-
tion of pre-A spects projection rules operating on surface structures, made
K-P unnecessary or untenable in many cases. GS, on the other hand, saw
(i) The mani who deserves itj will get the prizej hei wants.
If (as GS argued) SS pronouns represented underlying full (and concrete) NPs, there
was no way this example could be completely accounted for. But EST ran into the same
problem, though it might transfer the problem of the representation of the full NPs to
the level of surface structure, via a Semantic Interpretation Rule (SIR).
the problematic examples as special cases that did not cast doubt on K-P,
but only on the CTG statement of passive and its lexical theory, as well as
its difficulties in dealing with multiple meaning and nondichotomy. The
problem with the active-passive pair given above is not with active-passive
per se; the transformation is involved in the problem only in that it moves
subject NPs and direct object NPs over each other. When quantifiers cross
this way (as argued by Postal (1971)), changes of meaning may occur, unre-
lated to the transformational operation itself. Only where quantifiers play
the roles of subject and object do passives display this meaning change, so
it is not characteristic of passivization itself, and therefore, K-P does not
fail with passivization. Moreover, it might also be argued, CTG suggested
that the two sentences in question had distinct and different meanings,
which would cause trouble for K-P. Rather, both sentences have the pos-
sibility for both meanings; but each tends to favor one meaning over the
other. Thus, (1) has (in isolation—always a tricky criterion) the primary
meaning “Each person speaks two languages, but they could be any two”;
(2) most likely means “Two languages (the same two) are spoken by all the
people.” Within autonomous syntax, the meaning preference must remain
mysterious. If, however, we introduce pragmatic, textual, or functional
considerations, things get clearer. There is a tendency in English, other
things being equal, to use the subject position for topics, for focal points.
So (1), Everyone in the room… ., suggests, “I’m talking about these people,
my emphasis is on them,” and there is no reason to think that the lan-
guages themselves are being stressed and pointed out. But in the passive
case, the only reason (in isolation again) why the sentence is passivized (a
marked construction) is that we are being asked to focus our attention on
the languages themselves. It is a great deal easier to imagine doing this
in the case of specific languages that are under discussion, rather than
the vague reference point of languages in general. So functional consid-
erations and discourse expectations motivate one interpretation over the
other, and even so, there is no 100% correlation, as would be expected if
the distinction were based on syntax—that is, on the operation of pas-
sivization. So passivization doesn’t pose a problem for K-P, except within
a theory of autonomous syntax.
This doesn’t resolve the issue about the strength of K-P; but again, this
cannot be resolved theory-externally. If you believe in autonomous syntax
and semantic interpretation of surface structures, then a weak or nonex-
istent version of K-P works best for you; if you believe in a semantics-or
pragmatics-driven syntax, then paraphrase relations are deep and impor-
tant, and K-P must be maintained.
[ 262 ] Context Counts
263
[ 264 ] Context Counts
265
Figure 11.3 Continua.
7. The immediately preceding statement should, of course, be taken cautiously (cf. fn.
3 above). In fact, there were many intermediate positions, between the far left of GS
and the far right of EST. We might think in this connection of Paul and Carol Kiparsky,
Susumo Kuno, David Perlmutter, and perhaps Barbara Partee and Charles Fillmore as
centrists of diverse kinds.
3. BEHIND THE RIFT
[ 266 ] Context Counts
267
and then, I will talk about the personal styles of the two groups, and the
consequences for their theories and for linguistics, with some thoughts on
how to do “science” of this kind, or perhaps better, how not to.
One reason for the bitterness of the fight that eventually erupted lay
in the origin of CTG as a union of empathic souls, like-minded thinkers
fiercely arrayed against a common enemy. The sense of camaraderie was
there from the start, in TG’s “us against the world” format. It should be
noted as well that the earliest adherents tended to be people who had tried
other fields and found them, or been found by them to be, wanting. They
were outcasts looking for a group to belong to and be accepted by, some-
thing larger than themselves to which they could make a meaningful con-
tribution. In other ages they might have chosen the Church or the cloister,
only to be expelled for heresy and found a nonconformist group of their
own. We find, then, an unusual group: unusually close-knit, since until the
late 1960s virtually all had spent significant time at MIT officially or oth-
erwise, and all felt they owed an allegiance deeper than professional con-
nection to Chomsky—it verged on worship; unusually committed—they
were unwilling to stay in fields that didn’t promise to make a difference,
to allow them to do something important. They were, then, ambitious,
as well, but at the same time willing to run risks, as they were willing to
break away from careers in established fields. When they joined the move-
ment, therefore, they tended to be older than beginning graduate students
ordinarily would be, with predeveloped ideas and personalities. And this
suggests too that they were unusually strong-minded, even abrasive: they
could and did tell their superiors in established branches of linguistics and
other fields that the new work they were leaving to do was the “real thing”;
that the old stuff was not worth staying around for or committing your life
to. They didn’t all say it directly in so many words, but by leaving they said
it. By breaking those ties, professional and personal, with their own past
and the history of their field they effectively isolated themselves, making
all the more crucial the relationships that were to be forged within the new
group: it was to be family, world, church.
So the rise of dissension around 1965 was unwelcome and frightening,
not unlike the parents’ divorce to a child. It was different, of course, in that
they themselves played a pivotal role in the disagreement—and so each
side tended to blame the other as the starter of the fight. Then not only
did each side have its sense of its own intellectual rightness and the other’s
wrongness to drive them apart, but the additional rancor based on the feel-
ing that they had broken up an idyllic family.
There was added to the undercurrent one more ingredient that made
the eventual fight nastier than it perhaps needed to have been. Any
academic field, over its history, develops or borrows the means to defend
itself: evolves, if you will, teeth and claws. In the humanities, which at that
time would have been taken by many if not most linguists to encompass
their field, the notion of argumentation was scholarly and gentlemanly.
Unlike the hard sciences, the humanities did not have (as they still, for
the most part, do not have) the concept of work being done by competing
paradigms, one of which wins out over the others by dint of demonstrated
superiority. Rather, one achieved repute by demonstrating a fine aesthetic
sense, good judgment, the ability to see many facets of a work, to digest all
that had previously been learned about it and add to that—not overturn
it. So it didn’t make sense to engage in brawls about who was the best: it
was in bad taste, and there was no way to determine the “right” idea in any
case: it didn’t even make sense.
In the “hard” or physical sciences, on the other hand, there was such a cri-
terion: the “best” theory was the one that most economically accounted for
the observed data. Both what was “economical” and what was “data” were
givens: everyone in the field could be counted on to agree, at least in prin-
ciple. One theory could and in time probably would destroy another: pat-
ricide and fratricide were daily events, necessary parts of progress. But at
least according to the mythos of the sciences, these debates were solved
by the data, external to the investigators and equally accessible to all: it
was impersonal and objective, it came out of microscopes and telescopes,
not one’s own mind. (Of course, occasional leaks about the way it really is
in science, such as the autobiographical writings of James Watson, should
encourage us to view this myth with a little skepticism; but there was a
truth to it: it was a feasible position because the data were objectively veri-
fiable and competing positions could be tested in replicable ways.)
The social sciences have always held a problematic place in human
knowledge. As ways of understanding reality, they came last, after the
humanities and the physical sciences. Their very name seems an attempt
to filch some of the glory that the latter fields had achieved. Why “social
sciences” rather than “social humanities”? (I mean, of course, not just the
name but the associated method.) Their proponents will claim, naturally,
that the “scientific” aspects of the fields are extant and valid: quantifica-
tion, hypothesis formation, falsification, replication, lots of impressive-
sounding -ations. The question is what they tell us, what we are enabled
to know with certainty as a result of these methods. What we note in all,
over the last three-quarters of a century or so, is steady factionalization,
into smaller or different, competing, often acrimonious fields. Psychology
begets sociology; anthropology begets linguistics; linguistics splits into
infinite subfields, hyphenated and otherwise. Each generation has its own
[ 268 ] Context Counts
269
paradigm, or many, its own revolution, or many; in each case, at least judg-
ing from the perspective of my own field, the new one appears to conquer
because it elegantly handles data that the old one could not, or did not
address itself to. But always overlooked is the fact that the old one elegantly
did things that the new one cannot, or disdains to do. There seems to be no
theory remotely capable of rigorously making order of all the richness of
the data, and one chooses one’s theory partly according to where one goes
to school and what generation one belongs to, and partly, as I suggested
earlier, based on one’s personal preference about what issues are interest-
ing, central, and crucial.
In other words, my feeling at this juncture is that there is no hard evi-
dence that the social sciences really are “sciences” rather than other modes
of organized knowledge acquisition masquerading in the garments of sci-
ence, but no more science than I would be Einstein if I put on an Einstein
mask and talked with a German accent. One can imitate science—as astrol-
ogy does in a somewhat different way—by insisting on its surface fea-
tures: avoidance of “mentalism”; quantification; formalism; discreteness of
categories. But the results do not resemble those of science, nor does the
behavior of social scientists resemble the ways in which the participants
in those fields behave. It may be that we are taking the image of “science”
based, after all, ultimately upon astronomy as the first modern science,
and basing our behavior as linguists upon that. But this might be a fal-
lacy. Astronomy worked according to its rules—as originally postulated by
the likes of Ptolemy, Galileo, Copernicus—and worked well as a predictive
model because it was focused on objects totally remote from ourselves, not
in any direct way connected to us. That disconnection allowed the kinds
of observations that in turn permitted the quantification and discreteness
that underlie the physical sciences and allow them to produce their pro-
found results. But when observation of necessity turns inward, when the
investigator in one way or another must be the instrument that evaluates
the data, or at least some human mind must gauge the meaning of prod-
ucts of some human mind, then those methods become less reliable. When
the data themselves do not come in finite, discrete, and unambiguous for-
mat, but everything is relative to its context, and context itself is highly
subjective, and these are basic realities about language use, crucial parts of
the structure, not annoying encumbrances—then the methods that work
so well for other kinds of data cannot be automatically appropriated for
these new ones. It may be that they could yet be shown to be equally valid,
though probably with profound modifications. But the social sciences have
adopted the methods as unthinkingly as they did the name; and the fact
that they get “results” thereby doesn’t mean they are correct or meaningful
results. Of all the social sciences I fear linguistics is the most culpable here,
since it depends the most crucially on the kind of data I am talking about—
the artifacts of the mind.8
Linguistics, like the other social sciences, assumes that it can and should
operate via the “scientific method” developed for the physical sciences since
the fifteenth century; assumes that it is both possible and desirable to treat
the artifacts of language, which are discoverable only through the use of the
investigator’s mind as an instrument (that is, partially or wholly through the
use of intuition), as though they were molecules or stars. But some residual
doubt remains in the social sciences about whether this assumption is rea-
sonable; the problem is that the doubt remains repressed, covert. It seems
to me the question should be asked and investigated, or we should admit
we cannot do so, leave the question open, and call our various disciplines by
some other blanket term than “social science.” Otherwise, as psychoanaly-
sis (another field subject to the same sorts of uncertainties, for many of
the same reasons) would argue, the repressed fear is apt to lead to distress-
ing superficial behavior. For instance, philosophers of science, taking the
unquestioned physical sciences as their model, have proposed definitions of
“science” meant to generalize about what links these fields conceptually. For
practitioners of those fields, if the question is of interest at all, it is of pass-
ing interest, and has not (to my knowledge) sparked much soul-searching.
But the work of these scholars (especially Popper and Kuhn) has been taken
up with passion within a number of the social sciences, where it has been
treated as a litmus test for legitimacy or entry in the club. Therefore, the
claims assume huge symbolic importance: if you can prove that the findings
8. It may be useful here to recognize a distinction between “doing science” and “play-
ing science”—as children play Doctor, or House. Doing science entails utilizing the
scientific method because it has proved, over time, to be useful in facilitating lasting
discoveries and deep understanding of natural phenomena. On the other hand, when
children play at adult occupations, they grasp at superficial behaviors without under-
standing their deeper purpose. We must be very sure that that is not the case when we,
as linguists, call ourselves “social scientists.” There is a danger of a valuing the superfi-
cial manifestations of “scientific” behavior as validating for their own sake: quantify-
ing, formalizing, replicating, and so on make us feel like real scientists, grownup and
responsible—but do they produce lastingly valid results in our field? A corollary of
“playing science” is the overvaluation of theory at the expense of observed data, as
represented perhaps most obviously in the many Chomskyan gibes at “empiricism”
as stupid, culminating in utterances like Newmeyer’s cited in footnote 5, or Gazdar
and Klein’s (1978:666) statement to the same effect. It may also be germane to note
that the paradigm “science” social scientists like linguists are prone to take as a model
is Newtonian physics, with its dichotomies, objectivity, and certainty. But quantum
physics has cast doubt on all these vaunted desiderata, and we might ponder the dubi-
ous advantages of modeling our own theory and method on those of an obsolescent
field. Phlogiston, anyone?
[ 270 ] Context Counts
271
of your field are falsifiable, that your field has paradigms, etc., you are
respectable. The implicit assumptions, of course, boggle the mind.
The regrettable surface behavior that arises from these self-doubts is
not long in coming. I recall an article written by Ray Dougherty (1974). In
it he argued that Extended Standard Theory was a science because it had
scientific revolutions; and indeed a better science than any of its competi-
tors because it had more revolutions. I cringed at the time to see this sort
of argument appear in a refereed journal (its very appearance might be said
to disprove its claim); but my chagrin was somewhat assuaged a few years
later (misery loves company) when I encountered, in an equally august psy-
choanalytic journal, an article by Kurt Eissler, one of the giants of the field,
arguing that psychoanalysis was too a science, pace Popper, because it had
lotsnlots of paradigms. Case closed.
Suppose what I have proposed has some truth to it: suppose the social
“sciences” in general, and linguistics in particular, have yet to demonstrate
the appropriateness of the scientific method to their subject matter, the
working of the human mind. Suppose a large part of the work of this field
is, despite our recent disclaimers, still humanistic at heart: dedicated to
figuring out what is individual, how a person creates himself or herself,
and therefore beyond the reach of statistics, of replicable experimentation.
Then what happens if we graft the argumentative techniques that work so
well for science upon a humanistic study? What will we get? Modern lin-
guistics, I suggest—to its misfortune.
Science, with its distinct and contrasting paradigms, works by an adver-
sarial method: only one of us is right, and I intend to show that it has to be
me. Because there is agreement on the basic issues—method and data—if
one interpretation can be shown to be superior in its explanatory capacity,
everyone will sooner or later agree on it, and that will become the prevail-
ing model. While at any moment, in any active science, there are always
several areas of passionate disagreement, there is normally expectation
that sooner or later evidence will transpire that will be persuasive for one
approach. And while scientists do of course take the disagreements some-
what personally and get excited and even angry with one another when
their views are challenged, it does not seem that these passions normally
result in formal rifts. (One can, of course, think of occasional exceptions,
such as the Leakey/Johanson fracas in physical anthropology*. It may not
* In the late 1970s, the discovery of hominid bones (including Lucy) in the Olduvai
Gorge by a team led by Donald C. Johanson and the discovery of fossilized hominid
footprints at Laetoli by a team led by Mary Leakey were each interpreted rather dif-
ferently by the other. Johanson believed both to have been left by the newly named
be coincidental that this field is on the border between science and social
science, nor that the necessary data are inadequate, perhaps forever.)
So adversarial argumentation is not a serious problem in fields in which
there is confidence that external and objective grounds are bound to emerge
to prove one side right. And in those fields in which this is not part of the
world-view, this kind of conflictual discussion is not encountered, for it would
clearly accomplish nothing. But the social sciences, and most especially lin-
guistics, are in the middle, with the focus of the humanities but attempting
the methods of the sciences. And Chomsky brought to bear one further tool
of the latter, the better to make linguistics rigorous and respectable: the tra-
dition of contentious and acrimonious adversarial argumentation.
It was first applied, of course, to foes within linguistics or allied fields
(Lees (1960) against Bolinger; Chomsky (1959) versus Skinner). This
served, more than anything, to create the climate in the early years of CTG
of “us against the world,” TG surrounded by vicious enemies. I remember
well the times that nontransformationalists would speak at MIT, in those
early years when the field still saw itself as fighting for survival in a hostile
world. Rather than attempting to charm, conciliate, find points of connec-
tion, the circle at MIT regularly went for blood. Points were made by obvi-
ous public demolition; the question or counterexample that brought the
offender to his knees was repeated for weeks or months afterward with
relish. TG did not win, then, by gradually persuading its opposition, but
mostly by waiting until they retired or died. Since the field had been quite
small, this didn’t take very long. Those who were not won over or gotten rid
of were rendered ineffectual. There was no place for pluralism.
This habit of victorious battle felt very good to these young people,
ardent and ambitious. But by 1964, certainly, the battle was won. No more
opponents came riding into Cambridge eager to joust with the champion.
Could they let their lances rust, slide into a gentle middle age? Impossible!
So when the time came that dissension arose within their own ranks, they
were primed for blood. Everything led inevitably to conflict and implacable
hostility: (i) the fact that people had entered TG with two very different
agendas, and didn’t know it; (ii) the fact that linguistics was caught between
the methods and data of science and humanities, and wasn’t cognizant of
[ 272 ] Context Counts
273
this; (iii) the fact that early transformationalists had emerged, rather late
in life, from other fields that had disappointed them, and they were thus
feisty, ambitious, and in a hurry to prove themselves; (iv) the fact that they
acquired a taste for blood early on, then ran out of prey, and had no one to
turn their weapons against but one another. Thus the bitterness, the divi-
siveness, the insolubility of the struggle. The seeds were sown long before
dissension appeared; once any source of serious disagreement manifested
itself, the outcome would be inevitable.
The extramural hostility had one other unfortunate result, at least for
one side in the eventual struggle. It was seen as very important not to let
the Bad Guys see your weaknesses, if any: rather than be honest, acknowl-
edge that TG had its flaws, things it couldn’t do, the requirement of adver-
sarial discourse was that one present a pose of perfect poise and complete
certainty. Within, at least sometimes, one could groan about the failings
of one’s field, the immensity of the labor (although there was normally an
atmosphere of buoyancy, the sense that all apparent problems were capable
of solution, and very soon); but to the world, one presented one’s best face.
Not only did this serve to further alienate the outsiders, the brighter of
whom could clearly see the weaknesses inherent in the model, but were
hooted at by the faithful if they attempted any critique; but it meant that
generations of students were educated to believe that success involved PR
as much as insight or hard work. If you could overlook your own inadequa-
cies, maybe they would go away, and at least no one else would tease you
about them. The latter might be true, but they didn’t go away, and the atti-
tude created closed minds and an unwillingness to question the established
doctrine, to explore, to combine their theory with others. Along with the
adversarial argumentation, it guaranteed that CTG would remain closed to
outside influences—ensuring a short life, or at least a tedious one.
Because they had closed themselves off to acceptance by the outside
world, it was critically important for transformationalists that they feel
loved and accepted by each other. It was of equal importance that they be
able to see their colleagues as worthy of love and respect. Both of these atti-
tudes were especially true with regard to Chomsky, and became more true
with regard to him as time passed and his reputation soared in the outside
world. In the early days, now barely remembered folklore, before his fame,
it is related that he was readily available—physically and even psychologi-
cally. You could wander into his office, and he’d take time to talk to you.
He might even be persuaded by something you’d said (Postal managed to
persuade him of the need for deep structure, quite a change in theory!),
and maybe if you were favored you’d get a footnote or bibliographical ref-
erence in his next article. As time went on, both kinds of availability were
lessened. One would think that as someone’s repute grew, they would feel
more secure, would be able to be more open to reasoned discussion, new
ideas… . One would think that, but in some cases one would be wrong.
It might work the other way: such a person might feel insecure despite
the laurels, and feel that only by maintaining a firm hold on the doctrine
could he maintain his influence. Or someone might feel that his glory only
proved that he alone had possession of the truth, and therefore to listen to
anyone else was injurious to the field. In either case, the result was closure
of the mind.
Not really paradoxically, as Chomsky himself became less available in
mind and body, his status among the students as cult figure rose. He had
always been seen as a figure of towering intellect and integrity; these per-
ceptions increased. Hence, to have Chomsky’s approval meant even more
than it had before, even more than a prominent professor’s blessing means
to graduate students generally. Yes, to be in Chomsky’s good graces meant
mentions in his writings, getting your work published, getting a job; but it
also meant that you were worthy of him, you partook in some small way in
the godhead. For that to be withdrawn was equivalent to banishment from
the Kingdom of Heaven. I realize I am again straying into hyperbole and
religious imagery, and I can only assure the reader that it is employed to
capture a mood. To lose the goodwill of one’s fellow-transformationalists
was less serious than losing Chomsky’s but was still painful in this close-
knit and indrawn community. For all these reasons, then, when the split
came, it had to be highly unpleasant.
The worst of it was that, when Chomsky finally did address himself offi-
cially to the arguments of GS (as early as 1967 in lectures, later (1970) in
print), the level of hostility and close-mindedness was truly disillusioning.
It wasn’t even the sense of having fallen from favor that stung the most,
though that certainly smarted; worse was the sense that an idol had fallen,
leaving nothing else to believe in. The late 1960s was an era of idealism
and hope for a better world, and to see for the first time that the person
one was following to create the New Jerusalem had serious human foibles
probably hit worse, at this moment in history, than it might have earlier or
later, in more cynical times. So people unconsciously blamed Chomsky not
only for being unreasonable as a scholar (which was legitimate), but also
for destroying their illusions (which scholars are not supposed to carry into
their professional lives).
The result can be seen as instantiating a process first described by
Elisabeth Kubler-Ross in those facing death, and then extended to people
in situations in which their sense of self and their relation to their real-
ity is shaken: divorce, serious illness, job loss. It happened here too. First
[ 274 ] Context Counts
275
[ 276 ] Context Counts
277
of GS had very different notions about practically all the tenets that
Newmeyer identifies as Generative Semantic Orthodoxy. Indeed, judging
from where we each stand now, years later, it seems fair to say that GS
represented at best a loose coalition of interests: we joined together only at
the points where our interests happened to coincide.
It makes sense, I suppose, that people with a high tolerance— or
preference—for chaos would tolerate and delight in a theory which pre-
sumed and necessitated chaos, one which encompassed as its domain all
human endeavor. And those who preferred a tighter organizational struc-
ture would also prefer a more rigid theory that drew strict limits on what
was a part of its realm, what could be part of the theory and what could not.
One might also argue that these basic personality structures colored other
aspects of the theories of each side. If your preference is an organization in
which one prestigious person controls what is believed, then there will be a
certain tendency to authoritarian argumentation; a desire not to make one’s
case too simply, in such a way that it might be intelligible to hoi polloi, the
nonelect. In such a system profundity or brilliance might be directly equated
with turgidity; preferred style would bristle with arcane references, innu-
merable vaguely relevant footnotes, untranslated quotations in fifteenth-
century French or nineteenth-century German: anything to intimidate and
stun the reader (something no representative of the GS mentality would
ever do, as the epigraphs scattered herein make clear). Examples were
shunned, to be used sparingly only when utterly unavoidable: otherwise
the opposition might see what facts you meant to allude to, and test your
theory out on them. In short, orderly minds too often display a fondness for
authoritarian systems. Mussolini made the trains run on time.
CTG and EST, then, as well as their current descendants, are hierarchical
in organization. In this respect as well as in their dependency on formal-
ism, they are masculine.9 I had mentioned in earlier work that formalism
was a male perspective, and Newmeyer (1980:169) has seen fit to make fun
9. Once again the reader is adjured to beware of easy dichotomies, including M/F. But
although individuals display a range of behaviors and orientations along a continuum
between those points, societies tend to see the sexes as polarized: every behavior,
every physical and psychological trait, is identified as either Male or Female, Masculine
or Feminine. It is in this sense that we might say that GS reflects certain stereotypi-
cally feminine properties, EST more (stereotypically) masculine ones. The fact that
the majority of both groups were males (as was to be expected in academia of that
period—and this) is not especially relevant in this perspective. We would say rather
that GS allowed both its male and its female members to have access to aspects of their
psychologies that this society has identified as typically feminine; and the opposite for
EST. I do find it surprising though that more than one of my colleagues of the GS per-
suasion, and of both genders, have responded to the suggestion that GS was in some
sense “feminine” as though it were an insult. Rest assured, it isn’t.
[ 278 ] Context Counts
279
of the statement, apparently without asking what lay behind it. I meant
more than the obvious fact that men tend to be overrepresented in the for-
mal end of the field—the more formal, the more masculine—and women
at the informal, data-oriented end. Of course there are counterexamples—
it would be stupid, especially for a generative semanticist, to claim that
humans are irrevocably dichotomized by gender. I am speaking, as we must
in talking of human possibilities, of general tendencies. But I meant more
than this. There is now an impressive array of evidence that, from earliest
infancy, males respond to stimuli differently from females. Male neonates
are more responsive to their external physical surroundings: light, warmth.
As they get older, they are more active—interactive with their physical
environs, more eager to play with things, objects. This involvement with
inanimate externals is evident again in the games they play, having to do
with external goals, manipulation of objects. Little girls, on the other hand,
are immediately more responsive to people, being held, voices, eyes. They
smile sooner, recognize others sooner. Later, their games are interactive,
concerned with reaching out to others: House, Doctor, and so on.
We can see formalism as maximally noninteractive; and autonomous-
language theories as treating language as an external, impersonal object.
On the other hand, a theory that is concerned with language as an interac-
tive strategy, linking people with one another more or less successfully, is
closer to the way women tend to approach the world. And a hierarchical
theory is masculine as well in that it tends to recapitulate the structure
of male institutions: government, the military, the university, and so on
are and have always been organized in a hierarchical fashion, with a single
authority at the top—like CTG. On the other hand, female institutions
or groups, in those cultures in which they exist, tend to be more collec-
tive, cooperative, or collaborative: there is no formal leadership structure.
This tends to make such groups more fluid and impermanent, and some-
times more disorderly, but they usually get things done. I am—it should be
clear—not saying that EST was the field for manly men and only those, and
GS only for womanly women. But I am equating the theoretical preferences
I wonder, though, whether the “feminine” nature of GS and the “masculine” one
of EST is what has led to the perception of the former within academic linguistics as
less successful than the latter. After all, a “feminine” perspective is antithetical to the
hierarchical, dichotomizing tendencies that are characteristic of academic discourse
and masculine world-view. (And the academic perspective is inextricably masculine,
as it has been for the last couple of millennia in the West.) Here, as often, to under-
stand what the reality is, we may have to wrest ourselves from the prevailing ideology.
Several books have recently appeared within a feminist framework offering reassess-
ments of some of the culture’s unexamined verities: “objectivity,” “science,” “reason,”
and so on (Belenky et al. (1986); Gilligan (1982); Keller (1985)).
10. There is more to the GS romance with examples, though I don’t think we saw it
at the time. Examples are egalitarian: they allow each reader to form his or her own
conclusions, based on direct access to the same evidence used by the writer. The her-
meneutic Master, who controls the disciples’ exegesis, in disciplines as diverse as CTG
and deconstructionism, makes decisions on an ipse dixit basis: he (and it is by no acci-
dent virtually always a he) decides what his pronouncements meant, mean, and will
mean; he controls the development of the theory because he is the source of under-
standing. By contrast, a tradition relying on prodigality of examples is saying to the
reader, insider or outsider: Here are the facts. Make what you will of them. The reader is
thereby empowered, relative to the author(ity). One can see why this style was natural
to a group that came to maturity in the late 1960s, and why it might be threatening
to those, younger or older, who preferred a more authoritarian relationship between
Master and disciples.
[ 280 ] Context Counts
281
will be more apt to make sense of an argument if its examples are fun,
they’ll be less apt to get drowsy, and less apt to forget: if they at least
remember the examples, they may be able later to reconstruct from them
the arguments they supported. I don’t want to suggest that we invented
those examples from such laudable motives, but they can be defended in
this way. Certainly I would take issue with those curmudgeons (I think of
Stockwell and Newmeyer)11 who object to GS writing on the grounds that
examples were funny, names for rules or principles were frivolous (“Pied
Piping,” “Flip,” “WH-iz deletion”), or they wrote using slang or colloquial-
isms, or heaven forfend, puns. Their argument seems to be that it is inde-
corous, unscholarly—scholars don’t want to Have Fun, and so anyone who
is observably Having Fun is not a scholar.
If GS had had fun at the expense of accurate description, that would be
grounds for criticism, of course. But this is not anyone’s claim. Just as the
establishment made assumptions about the scientific status of linguistics,
it made the same unexamined assumptions about what constituted respon-
sible scholarship. One can certainly ask which stylistic strategy is the more
preferable: murky unintelligibility or quirky frivolity. I don’t know how to
decide, but I do know which I’d prefer to read.
Their collective high spirits point up another salient trait of GS as a
field: the tendency for its writings to point inward, for arguments and
claims often to be implicit. I know I just said that GS writing is especially
accessible. But I ought perhaps to qualify that statement, and shed more
11. Stockwell 1977:131, fn. 2:
We might note in rebuttal that many CTG/EST names are no more mnemonic: Root
Transformations, Strict Subcategorization, the A-over-A principle being just the first
that spring to mind. But the real point is that there is no reason for names to be mne-
monic: one associates the names with processes, just as one associates human names
with faces. Stockwell doesn’t especially look like a Robert, but that doesn’t make it
harder to recognize him.
In the same vein, Newmeyer (1980:171f.) grouses:
Such stylistic traits [of GS: he is referring to its practitioners’ “whimsical style of
presentation”] only served to give extra credibility to the charge of lack of seri-
ousness… . Indeed, it is tempting to speculate that generative semantic style is
but a classic example of content both shaping form and dominating it.
A comment which reflects a lack of understanding of both the style and the
content of GS.
[ 282 ] Context Counts
283
fields are peopled by scholars who have spent time in the field, have seen
with their own eyes, heard with their own ears, the diversity among human
languages, and see this as the interesting fact. One of the structuralists,
Martin Joos, once made a statement to the effect that, as far as he knew,
languages could differ from one another in innumerable and unpredictable
ways12—an extreme form of the relativist position, to be sure, and one the
transformationalists seized upon with glee as attesting to the imbecility of
the Bad Guys. On the other side, the anthropological types found amus-
ing, or horrifying, the TG tendency to abstract “universals” from thin air: if
a phenomenon was true of English and German, it was said, TG would
declare it a universal—an overstatement, but not by much. The dispute is
far from dead. Recently Dell Hymes (a representative of the anthropologi-
cal linguists) has written at length excoriating the universalist tendencies
of current linguistics (1986), pointing out the impossibility of proving most
such contentions in terms of any current theories, and the misstatements
and wrong turns they have led to. His examples are in many cases cogent
and sobering, but I think his arguments on the whole miss the point.
Language is neither pure surface diversity, nor pure deep identity. Both
exist as vital parts of linguistic activity, and it is the connection between
the two that composes grammar and indeed necessitates it. There must
(empirically) be universals, or second-language learning would be impos-
sible or tremendously difficult; there must be typological differences or
else it would be trivial, or nonexistent. But even if we grant these proposi-
tions, questions still remain that divide theorists: which of the two is more
important (if either); and if there are agreed to be universals, what are they
and how are they discovered?
Anthropologically minded linguists would say that the only way to prove
the universality of a phenomenon is to demonstrate it empirically, for
instance as Greenberg (1966) did in his work on universals of word order.
To make his case that there were statistical correlations between basic sen-
tence word order and other intrasentence orderings, he and his investiga-
tors examined superficial sentences in a large number of diverse languages.
So their universality claims had a firm empirical base.
Transformationalists, on the other hand, tended to reason deduc-
tively: a principle would be identified as necessary for the description of
[ 284 ] Context Counts
285
#1 #2
Sources anthropology philosophy
Patron saint Bloomfield Descartes
Level of observation surface/concrete logical/abstract
Philosophical stance empirical rationalist
World-view relativistic absolutist
Professional identification social science “hard” science
Scientific method quantification formalism
Discovery procedure empirical intuitive
Perspective diversity universality
Interaction theory sociolinguistics pragmatics
[ 286 ] Context Counts
287
In the end, we must ask: Why did GS fail—for all its hope, its optimism,
and its energy? If we accepted the question as valid, we would have to
answer: because it tried to do too much, dared too much. It tried to encom-
pass what could not be formally or rigorously controlled, and this was intol-
erable to those who see linguistic investigation, or scholarly work generally,
as taming the universe, getting things under our control. It’s a ruggedly
masculine image: knowledge, like nature, must be broken, tamed, bent to
our will. We cannot tolerate slippage or disorder. GS reveled in disorder,
and in its ability to tolerate it lay its contribution—and, if we accept the
myth, its downfall.
I have offered disclaimers in lamenting the downfall of GS. For I think
there is an excellent argument that GS never died. The conventional wis-
dom is that the Chomskyan branch of the theory continues in a straight
line, with only minor modifications separating the theories of Syntactic
structures, Aspects, early lexicalism, EST, and Government and Binding—
and on into the future. For Chomsky’s writings imply that position: had
we, the readers, but read Aspects intelligently, we would have understood
how fully it presaged EST… . Nothing really changes, it’s just made more
explicit by the Master, pushed by the stupid misunderstandings of the
opposition. So we have a sense of growing, flourishing, success. But the GS
people, as we have seen, were of a different kind. Their distaste for author-
ity brought them together, and it also made them quick to acknowledge
when some premise or formula fell apart. As Chomsky can be said to be
too slow to acknowledge paradigm shifts, GS may have been too quick to
despair—and to make its despair known to all, as loudly as possible (cf.
Morgan (1973) and Sadock (1975)). So the legend has it that GS finally
exploded, of its own impossible convolution, about 1975. But I think a bet-
ter understanding is that, at that time, we began to move away from the
idea that social and psychological context could be represented as the basis
of syntax in a syntax-central grammar such as GS was. We understood that,
to deal with the phenomena we had uncovered, the relationships between
form and function that our work had made manifest and unavoidable, we
needed to shift the emphasis of the grammar from syntax to semantics
and pragmatics. No longer did we use pragmatics to “explain” the central
point, syntax. More and more we started from function, and saw syntactic
devices as the servants of that function—so function still explained form,
but form depended on function, not vice versa. Yet, there is no sharp break.
I would say the difference between early-1970s GS and what its descen-
dants practice today is less strong than the difference between Syntactic
structures and Aspects, the latter and EST, or the latter and Government
and Binding. It’s just the salesmanship that makes it seem otherwise. One
(a) GS died, and serves it right. Seen from the CTG/EST side, GS failed
because it was chaotic, and that chaos was the direct outcome of a theory
that strayed too far from autonomous syntax, tried to incorporate the
whole world, and choked on it. I would argue rather that GS did die: but
“GS” in its literal acceptance, in the form its practitioners specified up
until the mid-1970s. GS in this sense died because it was too conserva-
tive, its practitioners too much enslaved by their early training as classical
transformationalists.
The problem was not the attempt to connect syntax, semantics, and
pragmatics in pragmantax; the problem was that whatever we called it,
syntax was still central and syntactic modes of argumentation and proof—
distribution and co- occurrence— were still the only options. But this
doesn’t really make sense: Why should syntactic form be what drives the
grammar, if meaning and function were basic to form? We gave lip ser-
vice to the concept of semantic and pragmatic conditioning for syntactic
rules, but never really got away from the idea that everything had to be
syntactically justified; that is, that the other two levels existed to serve
syntactic form.
Then, too, although late in the history of GS we began to realize the
importance of continua, we never followed through; the grammatical
devices and categories GS developed remained as dichotomous and discrete
as any within CTG. And as the meanings and functions we sought to cap-
ture were of course continuous, we went crazy finding more and more and
finer and finer subcategorizations in an attempt to account for the com-
plexity and richness of the data we wanted to incorporate into “linguistics”
and “language.” Because we clung too desperately to the mindset of our
youth, we made GS untenable.
There is another way we can account for the perception that GS failed
and its competitor survived, and that is by considering the role the devices
of public relations and salesmanship play in the ivory tower. GS is pre-
sumed to have failed because its practitioners said it did: how often, pub-
licly and privately, we bemoaned the fact that our theories didn’t fit the
data! That everything was getting too complex for our understanding! That
the phenomena of language went beyond the realm of science and into the
realm of aesthetics and even the supernatural! All this was interpreted as
[ 288 ] Context Counts
289
a statement that we gave up, it was no good. We didn’t really intend it that
way (Newmeyer cites such statements as evidence of our despair, but as so
often he misunderstands), but that was what was understood, and eventu-
ally we began to wonder if they maybe weren’t right. In any case, we never
made a vigorous defense.
Newmeyer suggests that salesmanship was the style of GS, with CTG/
EST peopled by earnest Young Doctor Pasteurs, concerned only for the
truth, while GSists ran around the world making conversions and selling
highly abstract snake oil. But the premise is contradictory; at least, if we
were salesmen, we were bad ones. We had no journal; we controlled few
departments, if any; we couldn’t even keep our own gang in order.
CTG, on the other hand, had in its progenitor a superb persuader,
whose buoyant optimism about his theory swept others irresistibly
along. Chomsky’s style (and that of his disciples) always was to down-
play sharp revisions in the theory. One reads Aspects without encoun-
tering any suggestion that deep structure represents a revolutionary
departure from the TG of kernel sentences; one reads “Remarks on
nominalization” without a clue that the description there of the lexicon
and of deep structure existed nowhere before and was developed only
because GS pushed Chomsky to redefine his position—quite radically.
The fissures separating early TG, Classical TG, EST, and GB are at least
as deep as those dividing GS and what its developers are doing today.
But the Chomskyans present themselves as coherent, united, and essen-
tially unchanging—a reassuring position for the neophyte looking for
a wing to shelter beneath; GS shows itself as disorganized, chaotic, and
unfocused—not a position to attract new adherents, or even hold ones
who are afraid of uncertainty.
(b) GS never died, and thank heaven for it. What GS strove to do, its ideal
version of itself, is still healthy, and if anything can save linguistics and
make it once again a rational field, I think it is this perspective. It is a highly
ambitious program, as well as (currently) a rather nebulous one, not fully
articulated by any of its practitioners and practiced by people with a wide
diversity of agendas who do not necessarily see eye to eye (any more than
the old-time GSists did). But the contributions toward which GS was reach-
ing when it fell into crisis are still part of all of our active agendas, though
we approach it in diverse ways. Basically, I think the lasting contributions
GS made, and its successor(s) will continue to make, to our understanding
of language are these:
First, the realization that the phenomena of language are continuous,
and that this continuity extends across all the data as well as the theoretical
7. CONCLUSION
What are we finally to make of the upheavals of those years and the com-
plexities of feeling and belief that still remain? For surely there is no reason
to dwell at length on a moment’s misguided squabbles and the bad feelings
they engendered, unless there is something to learn from the experience.
It seems to me now that the “wars” were an unfortunate outgrowth of a
failure on the part of all of us to make explicit—to ourselves and others—
our covert hopes, aims, and beliefs concerning language and its analysis.
Not surprisingly, the unaskable and unanswerable questions festering a
decade ago remain largely unmentionable. Only by bringing them out into
the light, as I have tried to begin to do here, can we hope to achieve some
resolution and perhaps ultimately reunify the field, or at least make it pos-
sible for us to hold rational discourse with one another.
I see this as the major, unaskable, festering question: Is linguistics a sci-
ence?; and if not, What happens? Before you clench those muscles in your
jaws, dear reader, consider awhile: Do we want our field to be a “science”
because that name makes our enterprise feel prestigious and worthwhile;
or are there really justifications?
Let us drop back a bit and ask: What does it mean to be a science? Does
linguistics work that way? Does language fit the model of the proper object
of scientific inquiry?
[ 290 ] Context Counts
291
We have tried for most of this century to force language into the
Procrustean bed of “science,” and the chaos and dissension that we have
experienced in the field are the result. If we are a science, we must assume
that only one paradigm has access to the truth, and it had better be our
own. But the impossibility of getting everyone in the field to accept a sin-
gle paradigm, to settle down to Kuhnian “normal science,” demonstrates
that we have been seeing things incorrectly. Just as every serious literary
critic who has had something to say about Hamlet has added to our under-
standing of that work, although each sees it in a very different way, so each
linguist, or each theoretical perspective, captures a different vision of the
linguistic reality, and all, though incompatible as scientific theories, have
something to add to our knowledge. But we can no longer require that per-
spectives be combinable into one single theory: we must settle for differ-
ent, but equally valid, viewpoints.
We must see that our work is not, on the whole, “objective,” since we
can only view language through the filter of our own individual minds,
themselves working through a lifetime of diverse and unique experiences.13
Language use is subjective, and much of our treatment of it, if it is to repre-
sent it accurately, must mirror that subjectivity: talk about how the inves-
tigator reacts to the data as a human being rather than speaking as though
that unique experience were somehow that of all of us.
The introduction of pragmatics into linguistics in the early 1970s brought
(without our realizing it) this subjective and mentalistic aspect of language
use into focus, where syntax-based approaches allowed the problem to be
glossed over. Therefore, it was at the time of the introduction of pragmatics
into GS that the dispute between the two sides got most heated and bitter.
Not because one side wanted to see linguistics as “scientific” and objective,
and the other didn’t; but because both did, and neither could reconcile that
desire satisfactorily with the data that were now turning up. Both felt frus-
trated with the inability of their own theories to deal with these facts fully;
one turned away completely, and the other attempted to incorporate the
problems within the domain of linguistic science. The internal frustrations
created irritation—for which each side blamed the other.
If we now take seriously the implications of the existence of a pragmatic
aspect of language, we must recognize that the nature of language is not as
13. It should be noted throughout this discussion that when I talk about the nec-
essary objectivity of linguistic investigation, I really mean to confine my comments
to the domain of syntax, semantics, and pragmatics. Phonetics and phonology, being
grounded in instrumentally examinable, instrumentally verifiable actualities, might
certainly be scientizable.
[ 292 ] Context Counts
293
we had been assuming, and the way in which it has to be studied is much
more complex and sophisticated than any method available to scholarship
at present. We have two choices, as I see it: to continue to pretend to objec-
tivity, discreteness, the artifacts of Newtonian physics; and as an inevitable
concomitant, brace ourselves for another generation of noncooperation,
frustration, and bad feelings. Or, we must put aside all of our previous
ways of looking at the linguistic universe, and carve out a new perspec-
tive: toward language, toward knowledge, toward one another.
REFERENCES
[ 294 ] Context Counts
295
[ 296 ] Context Counts
297
And we know that’s true. Our doctors are no exception; they make it very
clear that they don’t want to listen to us either. It’s very very hard—when
you have worn a gag all your life, when you have been endlessly shamed
for trying to go gagless—to remove it and use language that is elegant and
polished. I know how rough this introduction I am writing is; several times
I’ve almost given up and told the editor I can’t do it. The fact that I’ve pub-
lished dozens of books and written a dissertation helps very little. The gag
I wear, like my twisted spine and my distorted neural wiring, is invisible to
most eyes, but I can’t get away from either one, not even for a few minutes.
Like Lakoff, I have no solution to offer for the linguistic problems. Some
of the other problems are closer to solution today, I think, than they have
been in the past. I could no longer be forced into retirement for lack of a
handicapped parking space or an office in the building where I taught my
classes, for example—I would have to need much more exotic concessions
than those today, and that’s an improvement. But the linguistic problems
remain the same, and no one is interested in solving them. The interest
ends, predictably, at finding a “name” with which those who must fit us
into the society of the Whole-and-Able could feel comfortable. There is still
no interest in learning how we cope, or what might be done to make it eas-
ier for us to do so. There is still no interest in finding a way to increase soci-
ety’s tolerance for the way disabled women often look. (I am quite certain
that if Stephen Hawking were a woman, that woman would never appear
on television.)
My personal vote is for the old-fashioned word cripple. Not “I was
crippled by spinal polio,” the passive participle that Lakoff rightly identi-
fies as demeaning, but the straightforward “I am a cripple.” She says that
she would prefer “Are you Jewish?” to “Are you a Jew?” because the lat-
ter question suggests that Jewishness is “the totality of her … identity”
(p. 301). But there is no “Are you cripplish?”; in many important ways, the
crippledness (no word for that either, you perceive) is your whole identity.
None of the euphemisms—disabled, handicapped, differently abled, physi-
cally challenged—expresses the simple truth of my situation in the way that
cripple does. If I’d insisted on that word all these years instead of being
politically correct and “nice,” I think some of the most painful parts of my
life could have been avoided. The reaction of men to seeing my naked back,
for example. If I’d said to them beforehand, “Be warned—I’m a cripple,”
perhaps they’d have been less startled and better able to disguise the fact
that they found me repulsive; perhaps those who knew they wouldn’t be
able to disguise it would have been wise enough to decline with thanks.
I am grateful to Lakoff for raising these issues, as she has raised so many
others, and for making it explicitly clear that there are linguistic problems
W O M E N A N D DI S A B I L I T Y [ 297 ]
298
[ 298 ] Context Counts
299
CHAPTER 12
Review essay
Women and disability (1989)
This paper was originally published in Feminist Studies 15(2), 1989, 365–75. It is
reprinted here by permission of the publisher: Feminist Studies, 0103 Taliaferro Hall,
4280 Chapel Lane, University of Maryland, College Park, MD 20742.
300
existence a guilty secret. The strength, poignancy, and beauty of these writ-
ings communicate the error of such attitudes, which have caused a great
loss to all of us over the millennia. And how fortunate it is that the barriers
are beginning to be torn down, although barely beginning, as many of the
testimonials make painfully clear.
“The disabled,” I said at the outset; and most of the books incorporate
that term into their titles. As clear as it is, the term is in some respects
problematic. The basic dilemma is familiar to most nondominant groups.
By what name are its members to be called, and do they have the right
to name themselves? What is the consequence of any particular choice?
(Are they to be Negroes, African-Americans, or blacks? Ladies or women?
Old people, the elderly, senior citizens, or golden agers? No choice is without
difficulties.)
A first reaction may be dismissive: “What’s in a name? Let’s get to the
realities!” Our culture is ambivalent about the power of language, swinging
proverbially between “Sticks and stones can break my bones, but words
can never harm me” and “The pen is mightier than the sword.” But words
are actions; like actions, words alter reality. What you call someone is what
that person becomes—most often for both of you. Naming has too often
been the privilege of the powerful, further reinforcing their power. To
name oneself is to take back one’s power, to assert the right to determine
one’s own identity.
The problem for the group under discussion is complicated by the
fact that membership is seen, almost universally, as unfortunate, if not
accursed. Therefore, the term disabled tends to elicit reactions of pity and
fear and patronizing attitudes and requires, all the more, a new terminol-
ogy to suggest a new point of view. But because of the strength of the nega-
tive image, it is especially hard to find a name that carries the appropriate
denotation and connotations and none of the undesirable ones. The oldest
term in modern English is cripple, unsatisfactory in many ways, because of
the Tiny Tim overtones, the implications of ineffectuality, and the associa-
tion with pity.
Around the turn of the present century, an attempt was made to find
a more neutral (and “objective”) term; and handicapped was substituted.
More recently, though, this term has been criticized as implying helpless-
ness. But it has been harder to find a term that both achieves descriptive
accuracy and avoids negative implications. Among those I have seen in
print, probably the one closest to universal acceptance (as all of these writ-
ings suggest), is disabled. (I have also encountered differently abled, which
I think we could do without, and challenged, which I like, suggesting as it
does the eventual transcendence of the difficulty.)
[ 300 ] Context Counts
301
1. For discussion of the concept and uses of euphemisms, see R. Lakoff (1975).
point merely to show that these books, both by their very existence and by
the points they raise, force us to reflect, to rethink, to recognize that ambi-
guity is very much a part of our world.
As women and feminists, readers of these books may feel particularly
at a loss for solutions. All the books specifically address the situation of
the disabled woman. As several authors point out, she is in significant
ways worse off than a correspondingly disabled man, although not in every
way. In terms of policy, the disabled woman, like other women, tends to
become invisible or to get the raw end of the deal; she receives less sup-
port, economic and other; more paternalism; more interference with basic
human needs, especially reproductive rights. Just as women have fought
for visibility, for public existence, so have the disabled, and disabled women
have had to fight twice as hard. But there is a sense, as Michelle Fine and
Adrienne Asch note in Women with disabilities: Essays in psychology, culture,
and politics (p. 23), in which a disabled man is in a worse position than his
female counterpart. Disabled man is a self-contradiction, because men are
stereotypically supposed to be “able,” strong, and powerful. Hence, society
is less concerned with helping disabled women to achieve competence and
autonomy—as women, they aren’t supposed to want them, much less get
them. Disabled woman, on the other hand, is redundant. She is a super-
woman in an ironic sense: superchildlike, dependent, incompetent, even
beyond the stereotypical ideal female.
The women’s movement and the disability-rights movement have many
common aims. Both are dedicated to extending opportunities and increas-
ing options for everyone. What women have struggled to achieve over
the last century—equality of education, access to employment, economic
equity, opening of the professions, reproductive autonomy—the disabled
have also historically been denied, often with the same excuses. There are
also strong parallels to the civil rights movement. Thus, a century ago,
women were denied access to higher education because they weren’t capable
of it, because they wouldn’t use it productively, and because their presence
in a coeducational classroom would be distracting to the dominant group
(whose education matters). Similar things were said about black school
integration half a century later. And the same arguments flourish today in
opposition to the mainstreaming of the disabled in public education. There
is little apparent commonality among blacks, nondisabled women, and the
disabled. But society’s insistence on treating them similarly and subjecting
them to analogous stereotypical judgments suggests that all nonpowerful
groups serve similar functions for the dominant. Their existence gives the
dominant group a comfortable sense of its own superiority, as long as it is
not forced to share its privileges with the inferior group.
[ 302 ] Context Counts
303
2. Cf. Freud (1916) 1955. Women, Freud says, feel shame because of their lack of the “nor-
mal” male sexual organ; this defect gives them the status of “exceptions,” including the dis-
abled. He assigned to this category Richard III and women in general, on similar grounds.
[ 304 ] Context Counts
305
[ 306 ] Context Counts
307
and their partners; race, class, gender, and disability; disabled women who
have achieved satisfying work (how and what?); disabled women as moth-
ers; and the disabled in institutions (most of the articles in these volumes
are by or about women living at home).
Even more than women and blacks, the disabled have lived in shadows,
if not invisible, certainly unmentionable; very little is written about them,
still less from their own perspective. Much that has been written, most
often from a vocational-rehabilitative perspective, is (to judge from the
excerpts quoted in the articles) insensitive when not downright paternal-
istic. And, of course, society considers it neither possible nor quite proper
for the disabled to speak for themselves. I have heard many descriptions by
disabled women and men of their treatment in service interviews—even
though they are not visually or aurally impaired, waiters and store clerks
address them through their nondisabled companions: “And what does she
want?” These books begin to give the disabled a voice, with which, ulti-
mately, will come visibility and full participation.
Collections like these serve two functions seldom combined elsewhere
(as academia tends, improperly, to find such combinations improper),
analysis and advocacy. Some of the writings stress one over the other or
develop one explicitly, the other as subtext. But it is impossible to read even
the coolest of the descriptive articles without asking, “What is to be done?
Why did this happen?” and the reader comes away from even the most stir-
ringly persuasive contribution with deeper intellectual understanding.
One great benefit of reading all these books in all their variety might
be to move the reader past one undiscussed, but highly toxic, disability—
dichotomization. Humanity as a whole takes comfort from polarizations—
he male, she female; this right, that wrong; I able, you disabled—forgetting
that all these properties function as continua, not dichotomies. Academics
are, if possible, more susceptible to this disease than most people. We
like our neat disciplinary categories. Most universities make it difficult if
not impossible to teach interdisciplinary courses or do interdisciplinary
research. Social scientists like writing by social scientists, in social-scientific
vocabulary, with nicely contorted social-scientific syntax; humanists pre-
fer a more graceful, if less general or definitive, approach. It is considered
unseemly—a form of miscegenation—for the two to meet. In fact, aca-
demics agree on only one thing: that nonacademics should be kept out of
the discourse, as irresponsible and unscholarly.
Therefore, these collections serve a variety of purposes. For the disabled
reader, they provide ideas, role models, possibilities. The nondisabled are
afforded a view of another way of life, more demanding, certainly different.
Both groups achieve an understanding of the connectedness between the
disabled and the rest of society, of the need for demolishing the barriers
that create and enforce invisibility.
And finally, the reader may learn something else as well. The content of
the books, which stresses the commonality of the human experience, and
their format, combining disparate methods, approaches, and perspectives,
make, finally, one profound statement. We come away with the knowledge
that full understanding of anything is possible only if we achieve these
combinations. Just as the juxtaposition of all these collections, and the
varieties of writing within each, add up to a richer whole, so by including
everyone within the sphere of “us-ness,” including, finally, the disabled, we
will achieve together the complexity, dignity, and wholeness that define the
property of being human.
REFERENCES
[ 308 ] Context Counts
309
Introduction to “Pragmatics
and the law: Speech act theory
confronts the First Amendment”
BY SUSAN C. HERRING
of crime and punishment. Her next court appearance was in the role of an
expert witness, in a 1989 case involving allegations of “false advertising,”
and I also went to hear her speak about that experience. It seems she had
given the court an introductory lesson in presupposition and implicature,
complete with illustrations on a chalkboard brought in specially for the
purpose. I inwardly delighted at the thought of a linguist teaching judges
and lawyers about pragmatics—maybe linguistics could contribute some-
thing practical to the real world after all! In this way, through public pre-
sentations (and, later, in published articles as well), Robin transformed her
courtroom experiences into linguistic insights.
The paper reprinted here was first delivered in April 1991 as a plenary
speech at the annual meeting of the Chicago Linguistic Society. In it,
Lakoff draws on various legal interpretations of the First Amendment—
the constitutionally guaranteed right to “freedom of speech”—in order to
critique speech-act theory. Particularly targeted are two legacies of clas-
sical speech-act theorists: a tendency to focus mainly on the speaker’s
intent, or illocutionary force, to the exclusion of an utterance’s perlocution-
ary effect, and a tendency to treat speech acts as either performative or
not, and if performative, to treat them all as equally performative, that
is, capable of acting on the world with equal force. As Lakoff points out
in the article, these tendencies are problematic enough within linguistics
itself. But they become even more problematic when confronted with the
real-world complexities involved in legislating cases involving the First
Amendment.
At the root of the problem is the fact that not all forms of speech are
protected. The U.S. government has always reserved the right to limit indi-
vidual expression when it goes against the common good, as, for example,
anti-draft rhetoric during wartime, or when it inflicts harm on others, as
in the case of defamation, or “fighting words” whose sole purpose is to pro-
voke conflict. That is, the legal system recognizes that some language use is
merely expression, while other language use constitutes action—and lan-
guage as action is not protected by the First Amendment. The problem is
and has always been: How can we tell the difference between the two?
As if this were not difficult enough, there is a second problem: what
has harmful effects on others may not have been intended as action by
the speaker, but rather as simple expression. (Such, for example, is often
claimed to be the case in defenses of “racist” and “sexist” language use.)
Whose perspective should prevail in determining whether the speaker
is guilty of a “language crime,” or whether he was merely exercising his
First Amendment right to free speech—that of the speaker or that of the
offended party? Lakoff argues insightfully that these are false dichotomies,
[ 310 ] Context Counts
311
both in legal theory and speech-act theory. Rather than classifying every
speech act in terms of either action or expression, intent or effect, she bases
her analysis on the notion that language use is gradient and complex. That
is, words constitute action (they are performative) to varying degrees, and
the effects (perlocutionary force) words have on the hearer are part of the
socially situated meaning of an utterance, along with the intentions (illocu-
tionary force) of the speaker.
Lakoff is not the first to critique speech-act theory on the grounds that
it falls short of capturing the complexity of actual language use (see, e.g.,
Stubbs 1983). She is, however, the first to do so with evidence from court
cases, and the first, to my knowledge, to address linguistic issues sur-
rounding the First Amendment. In this, she has targeted a key issue of our
times—the First Amendment is increasingly evoked in contemporary pub-
lic discourse in defense of controversial behaviors ranging from flag burn-
ing to posting offensive speech on the Internet.
Here I find myself once again inspired by Lakoff’s work, this time not
as a graduate student seeking reassurance that linguistic research can be
socially meaningful, but as a professional linguist interested in how social
inequality is perpetuated through language use on the Internet. In the
course of my investigations into gender-based asymmetry on the Internet,
I have observed a disturbing correlation between the use of misogynistic,
harassing language by some individuals (all of whom appear to be male) and
explicit appeals to the First Amendment. Lakoff’s article provides a useful
perspective from which to understand and critique this phenomenon.
I should state first off that defense of misogyny is but one manifesta-
tion of a more general civil libertarian-influenced ideology on the Internet
in which the concept of “free speech” plays a crucial role. In this view, “free
speech” is conceptualized in the strongest possible terms, as absolute
and unrestricted. Such an extreme stance is justified, in part, in terms of
the noncorporeal nature of computer-mediated interaction, which alleg-
edly ensures that all that takes place is “just text,” even when the words
imitate actions. It cannot physically harm you (the argument goes), and
if you don’t like it, you aren’t forced to read it—you can delete the mes-
sages on your computer screen, or avoid participating in electronic forums
containing the kinds of messages you find offensive. Reasoning of this
sort was evoked—successfully—in 1995 to defeat the first introduction
of the Communications Decency Act in Congress which aimed to restrict
the transmission of pornography via computer networks (Herring 1996).
(The Act was later amended and passed in 1996.) According to this view, all
communication on the Internet is, by technical definition, expression and
not action, and therefore is protected by the First Amendment.
P R AG M AT I C S A N D T H E L AW [ 311 ]
312
This view is at odds with three important points made in Lakoff’s arti-
cle: (1) that words can constitute actions; (2) that the effects of the words
on others are relevant in assessing their harmfulness; and (3) that “free
speech” is not protected when it harms others. Of course, Lakoff did
not write her article with the Internet in mind—perhaps the electroni-
cally mediated nature of computer communication takes us beyond such
considerations, rendering them obsolete. Some postmodern theorists of
cybercommunication would no doubt wish to claim this. But does com-
puter mediation really mean that people don’t “do things with words” on
the Internet, and that they aren’t accountable for their words if others are
adversely affected by them?
Revealing counterevidence appears as soon as we set theorizing aside
and look at actual computer-mediated interaction. Consider, for example,
a much-publicized event that took place on the Internet several years ago
involving a “virtual rape” (Dibbell 1993). A male character in a recreational
MOO (an abbreviation for “multi-user dungeon, object-oriented,” a type of
real-time electronic forum) publicly posted graphic descriptions of a vio-
lent rape having as its primary object a female character in the forum. The
descriptions were clearly performative in structure, making use of immedi-
ate present tense verbs in sentences of the type “X does Y to Z.” However,
the man responsible for the actions of the male character claimed after-
ward that his words were not intended to be taken seriously; rather, he was
experimenting with the freedom to construct different textual personas
in the new medium. That is, he appealed to the prevailing libertarian view
that “anything goes” on the Internet, because it is “just text.” In contrast,
other members of the group noted the psychological distress of the woman
whose character was raped, and the disruptive effect the event had on the
group as a whole, and maintained that the behavior constituted a harm-
ful action. After much discussion, the group voted to deny the man future
access to the MOO.
In this case, a community of users decided that computer-mediated
words were more than “just text,” that their meaning was determined in
part from their effects on others, and that because the effects were harm-
ful, the speech was sanctionable. In other words, the virtual community,
like courts of law in the physical world, concluded that in order to preserve
the common good, some restrictions on absolute free speech were neces-
sary. This outcome is fully consistent with Lakoff’s claims, but is inconsis-
tent with the arguments of male cyberspace libertarians, which in light of
this and similar incidents, seem not so much “postmodern” as self-serving
and pragmatically naive.
[ 312 ] Context Counts
313
Although the group legislative process that transpired in the MOO case
was an isolated event, it opens the floodgates, it seems to me, to the same
kinds of messy deliberations faced by courts of law in the “real world”
regarding the First Amendment. That is, the Internet, too, will ultimately
have to confront in a systematic way the fundamental problems identified
in Lakoff’s article—expression or action? intent or effect?—and when it
does, it would do well to avoid false dichotomies.
REFERENCES
P R AG M AT I C S A N D T H E L AW [ 313 ]
314
315
CHAPTER 13
I n this paper,1 I want to discuss a serious problem for two fields seriously
concerned with language use: pragmatics, and therefore linguistics in
general, and the law. It seems to me that speech-act theory, as framed by
J. L. Austin (1962), an essential underpinning (in some form) of pragmat-
ics (as well as other areas of linguistics), is in direct conflict with current
interpretations of the First Amendment of the Bill of Rights of the United
States Constitution (cf. Tribe 1988). While it is possible to imagine ways
out of the morass, they will probably be anathema to legal and linguistic
theory.
Linguistics and (in some respects) the law are social sciences; both
deal with the meaning and intention of language—the understanding
of human motives as represented in linguistic form, language as a link
between the mind and reality. Additionally, linguistics and legal schol-
arship are both interpretive disciplines. Both have access only to super-
ficial representations, and both achieve their results by making sense of
those surfaces: by finding recurrent patterns, deeper rationales, and gen-
eralizations. Linguistics, it is true, places its emphasis on description and
This paper originally appeared in Papers from the Twenty-Seventh Regional Meeting of
the Chicago Linguistic Society (CLS 27). Vol. 1, The Main Session, edited by Lise M. Dobrin,
Lynn Nichols, and Rosa M. Rodriguez, 306–23. Chicago: Chicago Linguistic Society,
1992. Reprinted here with permission of the publisher.
1. I would like to thank Elizabeth Bader, Barbara Bryant, and Jesse Judnick for their
insightful discussion and other help with this paper.
316
prediction, the law on prescription, but both seek to relate form and func-
tion. Both linguistics and the law have theoretical and applied branches.
But in linguistics, theory has always been preeminent, with concern for
real data secondary and practical application still further downgraded; the
law, on the other hand, begins with the reality of statute and precedent,
and derives theory therefrom, with the sense that the theory should clarify
and simplify the task of those who are engaged in real-world application
(trial and appellate attorneys and judges).
So if a crucial tenet of linguistic theory is directly at odds with a basic
aspect of American constitutional law, and if we agree that theories of lan-
guage use, meaning, and function should be consistent across disciplinary
boundaries, we are faced with a paradox. In this paper, I will offer one pos-
sibility for a rapprochement, more as a first attempt, to see if it’s feasible,
than as a complete and cohesive model. I assume that, in the long run, a
unified understanding of the functioning of language among the social sci-
ences (including, as here, linguistics and the law) is not only desirable but
essential.2
Central to the problem is speech-act theory, as originally developed by
Austin and developed by many others in linguistics and the philosophy
of language, perhaps most notably John Searle (1969, 1979). While there
has been a great deal of rethinking and revision of speech-act theory since
Austin’s proposal, its basic claims retain their validity, and it is those essen-
tial claims that are troublesome in this perspective.
Austin’s aim was a critique of theories of language that equated truth,
or truth- functionality, with meaning: a sentence whose truth- value
could not be computed (anything other than a declarative), was, for
the purposes of logical positivism, technically meaningless. To Austin
as an ordinary-language philosopher, this conclusion was dissatisfying,
2. I am aware that the suggestions I am making here and below could not readily
be amalgamated into legal theory and much less into legal practice. That is why I am
making these proposals in a linguistic, rather than a legal forum. It is not that linguists
are necessarily more openminded than lawyers, or more willing to abandon treasured
theoretical beliefs. But changes in linguistic theory do not require changes in reality
(for instance, in the way we speak). But if legal scholars were to re-evaluate their inter-
pretations of the First Amendment (and other linguistically based legal formulations)
and practitioners were obliged to implement their proposals, the application of the law
would be radically altered, with resultant chaos.
As linguists, we have the luxury of being able to concern ourselves with matching the-
ory to reality, rather than vice versa. It is our task to ask how well a theory accommo-
dates itself to real language observations, simplifies real-world complexities (without
oversimplifying), and relates phenomena that have something in common. We may
wish the realities of linguistic form molded themselves more readily to the theoretical
devices we have at our disposal—but we cannot require that they do so.
[ 316 ] Context Counts
317
removing as it did large parts of language from the realm of the analyz-
able, the interesting, or even the meaningful. Yet it had its charms. By
reducing complexity to dichotomy, it allowed “meaning” to be formalized.
Thus it offered an opening to the scientization of language—to many,
then as now, a consummation devoutly to be wished and eventually to be
achieved.
Austin offered an alternative, less formalizable, but permitting clari-
fication and taxonomization— themselves prerequisites to scientific
treatment. In his framework, truth was reduced to a condition on the
appropriate usage of one class of verbs, rather than the major criterion of
meaning. Austin also shifted the central focus of investigation from mean-
ing, that is, semantics, to function, or pragmatics. (Where meaning is com-
puted in terms of truth, function is judged by appeal to appropriateness in
context.) Austin saw language as functioning, that is, doing things, altering
reality rather than merely describing it.
Austin begins his argument rather blandly. Some utterances, which he
called constatives, are evaluated in terms of their truth-value; others, a
smaller category, performatives, cannot be assessed for truthfulness, but
only for appropriateness, or as he called it, felicity. They are judged by their
success or failure in bringing about real-world changes. Such utterances
are speech acts: they perform actions (alter reality) through the medium of
words.3
It ought to be the case that utterances with similar meanings, func-
tions, etc. are analyzable as members of the same category (performative
or constative). But there exist pairs very closely related in meaning, but one
member constative and the other performative: I promise to return this book
on Friday, a performative; and I will return this book on Friday, a constative.
A performative/constative split creates a paradox. Austin’s remedy was to
make all utterances performative: some, like the first example, explicitly
so; others, like the second, implicit. Then truth becomes one of the felicity
conditions on the usage of a subset (verbs of asserting) of expositives, one
class of performatives. So, for Austin, all language is action: all utterance is
performative, and all equally performative.
But many of us had nagging doubts about Austin’s resolution of the
problem, even as we found it a source of significant insight, though in our
4. Ross’s discussion of squishes (e.g., 1972, 1973), somewhat after the period of
strongest interest in performatives, offers us a way to begin to deal with these prob-
lems more lucidly.
5. Another way to frame this apparent contradiction is to say that, at a general and
theoretical level, the illocutionary act is superficially apparent, much easier to deter-
mine and define than its complement. But on a case-by-case basis, looking at actual
utterances produced by real people and heard by others, the perlocutionary act seems
much more reliably determined.
6. I told you not to mention pragmatics!
[ 318 ] Context Counts
319
For about the first 130 years of its existence, the First Amendment led a
quiet existence (cf. Kairys 1982). The occasional cases that invoked it before
the Supreme Court tended to be settled in a way that restricted its domain,
rather than extending it.7 Its scope was first directly tested in the years
immediately following the first World War, in a series of cases pitting the
First Amendment against the government’s right to restrict speech in situ-
ations of “clear and present danger” such as occur most often in wartime.
Most of these cases involve persons who, during wartime, urged the United
States to withdraw from the conflict or counseled young men to resist the
draft. The first such case, Schenck v. United States,8 was decided against the
appellant, Schenck, by a unanimous court, in which Justice Oliver Wendell
Holmes wrote the opinion, including as part of its justification the famous
analogy: “The most stringent protection of free speech would not protect
a man in falsely shouting fire in a theatre and causing a panic.”9 But in
a rather similar case decided later that same year, Abrams et al. v. United
States,10 Holmes, along with Justice Louis Brandeis, dissented from the
majority affirmation of the verdict because the wording of the pamphlets
involved does not allow the inference that the defendants acted “with
intent by such curtailment to cripple or hinder the United States in the
prosecution of the war,” as the statute under which they were prosecuted
requires. Holmes and Brandeis required proof that the speakers had an
intent to do what their language appeared to suggest. In other words, the
majority of the Supreme Court, in accordance with normal judicial behav-
ior of the times, was requiring only a demonstration of perlocutionary force
that the word intent as vaguely used in ordinary legal discourse means no more
than knowledge at the time of the act that the consequences said to be intended
will ensue… . But, when words are used exactly, a deed is not done with intent
to produce a consequence unless that consequence is the aim of the deed.11
This two-man dissent gradually developed, over the next dozen years or so,
with later changes in Supreme Court membership, into a majority which
helped to give the First Amendment more or less its current understanding.
The issue or responsibility for the interpretation of an utterance was
not, however, fully resolved by Holmes’s discussion: later courts continued
to argue over whether responsibility for meaning resided with speaker or
recipient. Another problem, for both pragmaticists and legal scholars, lies
in the interpretation of even simpler language: What does the “free speech”
clause of the First Amendment actually refer to?
The clause says that Congress shall make no law abridging the free-
dom of speech. The words italicized are simple, but their range of meaning
broad: they are vague, or ambiguous. And, as is frequently the job of appel-
late courts, this vague language must be reduced to dichotomous precision
in order to conform to the needs of legal reality, to set precedents and allow
appropriate remedies to be taken.
The first of these terms, freedom, has been seen by the Supreme Court
throughout its history as relative rather than absolute, guaranteed as long
as it does not conflict with another constitutionally recognized right or
obligation. In the case of “clear and present danger,” as most typically in
wartime (or its moral equivalent, for example the Red Scare of the 1950s or
today’s “war on drugs”), the duty of the government to protect its citizens
from enemies foreign and domestic is frequently deemed by the Supreme
Court to supersede the First Amendment right of citizens to free speech.
Likewise, in its obligation to ensure domestic peace and tranquillity, the
government has been adjudicated to have the right to restrict speech in
order to prevent breaches of the peace, so that “what men of common
intelligence would understand [as] … words likely to cause an average
addressee to fight …,’’12 that is, “fighting words,” do not receive constitu-
tional protection. But the line between potentially troublesome utterances
11. Id. at 627–28.
12. Chaplinsky v. New Hampshire, 315 U.S. 568 (1942) at 573.
[ 320 ] Context Counts
321
that are unprotected “fighting words,” and those that are not, and there-
fore protected by the First Amendment, has been a tricky one for appellate
courts and constitutional scholars to draw. And one traditional distinction
on which both “fighting words” and “clear and present danger” interpreta-
tion has been based conflicts directly with Austinian speech-act theory.
Jurists and scholars have, since at least the 1920s, wrestled with this
problem: how to draw lines between the kind of speech they want to pro-
tect, and other kinds of social behavior that should be subject to criminal
or civil prosecution. They have frequently drawn those lines on the basis
of hypotheses about the Founders’ original intention in putting the First
Amendment into the Constitution. One frequently cited justification is
Alexander Meiklejohn’s (1948) doctrine of the marketplace of ideas, based
on Holmes’s dissent in Abrams: the First Amendment exists because, in a
democracy, citizens need to have access to a range of information in order
to vote intelligently. Thus, any restriction on the flow of relevant informa-
tion is destructive to the functioning of a democratic government.
This rationale allows a distinction to be made between conduct and
expression. Language does express ideas, the articles of trade in Holmes’s
marketplace. But it also can be used simply to let off steam, or to abuse
others, not unlike the ringing of bells or a punch in the nose. Such language
does not materially aid the business of informing the citizenry; indeed, it
may hinder it. That sort of behavior, the reasoning goes, is tantamount
to action, indeed it is action. And the law has never recognized protection
for actions per se. There are uncontested laws on our books dealing with
disturbance of the peace, physical assault, and other crimes of aggressive
interpersonal conduct. When language is used as action, it, too, can expect
no constitutional protection.
In the abstract, as is the case with most jurisprudential reasoning, the
argument makes a great deal of intuitive sense. But while the law (like too
much linguistic theory) is predicated on dichotomies such as this, the real
world does not return the favor: it organizes its phenomena as continua. So
trial courts instruct juries hoping to reduce the blooming, buzzing confu-
sion of the conduct they are called upon to judge into the either/or catego-
ries mandated by law: guilty or innocent, intentional or unintended, sane
or insane … expression or conduct. What trial juries frequently do, in this
impasse, is craft a compromise: find the defendant guilty of a lesser offense
than the prosecution demands and the evidence permits, as a way of say-
ing, “Guilty, but… .”
Appellate courts adopt analogous stratagems, albeit of a more explicit
and sophisticated form. Often uncertainties about the plausibility of the
legal dichotomy take the surface form of sharply divided courts, with
The conviction quite clearly rests upon the asserted offensiveness of the words
Cohen used to convey his message to the public. The only “conduct” which the
State sought to punish is the fact of communication. Thus, we deal here with a
conviction resting solely upon “speech” … not upon any separately identifiable
conduct… . At least so long as there is no showing of an intent to incite disobe-
dience to or disruption of the draft, Cohen could not, consistently with the First
and Fourteenth Amendments, be punished for asserting the evident position on
the inutility or immorality of the draft his jacket reflected.15
So the majority in Cohen based its reversal both on the lack of intention,
and on the fact that Cohen’s jacket was an example of expression, rather
[ 322 ] Context Counts
323
harmless to another. The law is too generous in its assumption that a nor-
mal speaker can make an accurate assessment of a hearer’s sensibilities; to
do so accurately would require knowledge of, and the ability to compute,
the hearer’s frame of reference, the status of hearer and target relative to
speaker, the physical and psychological setting in which the speech act took
place, and much more—far more than it is reasonable to suppose is avail-
able to the average speaker, especially in oral discourse.
On the other hand, an entirely speaker-based theory of defamation (or
language crime generally)—that is, one based purely on the computation
of illocutionary force—is equally untenable. Especially in an adversarial
situation where the utterer has much to lose, the latter’s assessment of
his/her intention is not wholly trustworthy. The same applies in cases of
harassment, “fighting words,” and linguistic crimes or torts in general. It
is all too easy to plead “only joking,” or “never realized,” or myriad other
bluntings of the responsibility for the force of the speech act. And it is fre-
quently impossible to prove otherwise.
A case of sexual harassment that recently made its way through the fed-
eral appellate courts illustrates the problem. In this case, Ellison v. Brady,18
a woman is suing a coworker for harassment, charging that he pursued her
with bizarre notes, although she asked him several times to desist, and
denied interest in him.
The trial court granted the employer summary judgment (dismissed the
plaintiff’s claim) on the basis that the conduct of the coworker was not
“sufficiently severe or pervasive” to constitute sexual harassment. Based
on the standard of what would be threatening and offensive to the “reason-
able man,” the trial court found the harassment isolated and trivial.
The Ninth Circuit reversed, on the grounds that the “reasonable man”
standard was inappropriate in this case; that women were in a more vulner-
able position, physically and psychologically, than men, reasonable or oth-
erwise, and that to invoke a hypothetical “reasonable man” was to adopt
a standard inappropriate to the plaintiff in this case. In other words, the
Circuit Court held that to some degree and for some cases at any rate, the
particular world-view of the addressee had to be taken into account, by
requiring that the jury consider a standard based on a “reasonable woman.”
The US Supreme Court came to a different conclusion in a criminal
case, Rhode Island v. Innis.19 A suspect was arrested on suspicion of murder.
18. Ellison v. Brady, 91 Cal. Daily Op. Service 605 (9th Cir. 1991), as discussed in
Bryant 1991.
19. Rhode Island v. Innis, 446 U.S. 291 (1980). My summary is taken from Tiersma
(1987:315 n. 52).
[ 324 ] Context Counts
325
Read his Miranda rights, he asked for a lawyer (that is, refused to waive his
rights). The law is firm here: at this point, all questioning should stop; any
statement obtained without counsel present is legally inadmissible if it is
elicited via a question. The arresting officers were, though, understandably
eager to discover the location of the murder weapon. As they were driving
to the police station, one officer said to the other that there was a school
for the handicapped in the vicinity of the crime scene, and “God forbid”
that one of the children should get hold of the weapon. The suspect then
revealed where he had hidden the gun. The confession was deemed admis-
sible by the trial court, and the defendant was convicted. The Supreme
Court was asked to determine whether the officer’s communication had
the force of a question, which would have required the verdict to be over-
turned. The Court decided that it did not, defining interrogation as “words
or actions on the part of police officers that they should have known were
reasonably likely to elicit an incriminating response.”20 The dissent, writ-
ten by Justice Stevens, argued on the other hand that “a statement that
would be understood by the average listener as calling for a response is the
functional equivalent of a direct question, whether or not it is punctuated
by a question mark.”21 The opinion of the court in this case, then, is to the
effect that the determination of meaning or intent rests with the speaker
of the utterance.
Cases like Ellison suggest that the likely interpretation of language by a
hearer in a particular situation, not only what a hypothetical “reasonable
man” might deduce, and certainly not only what the speaker might claim to
have meant, must be part of any decision in a linguistic crime or tort action.
On the other hand, Innis comes to the opposite conclusion (and is the judg-
ment of a higher court, too). So, as they say, the jury is still out on this one.
Another case currently before the Supreme Court will also raise the
issue of responsibility for interpretation. The Court is being asked by
Jeffrey Moussaieff Masson to decide whether there are grounds for a
libel suit against reporter Janet Malcolm and the New Yorker magazine.
Masson alleges that, in two articles published in 1983, Malcolm libeled him
by grossly distorting the content and implications of his remarks during
interviews with her. At issue is the definition of paraphrase.
It is part of a journalist’s normal job to “fix” the spontaneous oral
remarks of subjects, putting them into the form readers will consider
articulate and intelligible in writing.22 People who have had the misfortune
20. Id. at 322.
21. Id. at 309.
22. Cf. in this regard Malcolm (1990) herself.
* The Supreme Court decided in June 1991 that Masson’s lawsuit against Malcolm
should be allowed to go forward, stating that the issue of whether Masson’s statements
had been materially altered in meaning was sufficiently in question to be a matter for
jury trial (Masson v. New Yorker Magazine, 1991 U.S. LEXIS 3630 (1991). After a series
of jury trials and appeals, the suit was eventually decided in favor of Malcolm (Masson
v. New Yorker Magazine, 1996 U.S. App. LEXIS 13326 [9th Cir. 1996]).—Ed.
23. Oral communication with J. Masson, March 18, 1991, and column by James
J. Kilpatrick in the San Francisco Chronicle, October 12, 1990.
24. Milkovich v. Lorain Journal Company et al., 110 S. Ct. 2695 (1990).
[ 326 ] Context Counts
327
[ 328 ] Context Counts
329
constative performative
1 2 3 4 5 6
expression conduct
1 2 3 4 5 6
REFERENCES
[ 330 ] Context Counts
331
1. Guttmacher Institute (2015); report published in partnership with the World
Health Organization.
332
2. “Abortion should not only be safe and legal, it should be rare.” Source: Speech by
President Bill Clinton at the Democratic National Convention, August 29, 1996.
[ 332 ] Context Counts
333
with the word “abortion,” the side supporting the choice to abort went
with, well, “pro-choice.” The Center for Reproductive Rights, a global orga-
nization focused on securing women’s rights to abortion, studiously avoids
the word “abortion” in their mission statement, instead using “reproduc-
tive freedom,” and listing abortion as only one of many aspects of women’s
health care to which they are dedicated. Planned Parenthood, a nonprofit
organization that provides reproductive health services, similarly avoids the
word “abortion” in its profile, positioning itself as a “trusted health care pro-
vider.” And the National Abortion Rights Action League, usually shortened
to NARAL, has officially changed its name to NARAL Pro-Choice America;
their domain name is simply “prochoiceamerica.org.” The White House, in
its annual statement marking the anniversary of Roe v. Wade, hasn’t used
the word “abortion” since 2011, preferring the “choice” language instead.
News outlets have become uncomfortable with the soi-disant “pro-life/
pro-choice” dichotomy. In a 2010 survey of the terminology used by major
news organizations (NBC, CBS, CNN, the Associated Press, the New York
Times, the Washington Post, and the Philadelphia Inquirer), the National
Public Radio ombudsman found that not one of them used the terms
“pro-life” or “pro-choice,” most preferring “anti-abortion” and “abortion
rights,” respectively, as they are seen as more neutral.3 Yet most impor-
tantly, regardless of what either side chooses to be called, the majority of
Americans (55%) continue to support the legality of abortion, according to
a Pew Research Poll in late 2014.4
Of all of Lakoff’s suggestions for abortion-rights slogans, “forced moth-
erhood” has actually gained some traction and spawned another compelling
phrase, “forced pregnancy.” Both are used to refute the notion of women
as vessels, and to evoke a system of near-slavery of women akin to that
in Margaret Atwood’s dystopian novel The handmaid’s tale. This rhetoric
appeals strongly to American anti-authority sentiments while at the same
time acknowledging the consequences for women of being denied abor-
tion—something which previous abortion-rights language had skirted, as
it focused on abortion as a medical procedure quite divorced from the life
and experiences of the woman who chose it.
In the 21st century, the continued attacks on abortion rights, such
as attempted federal defunding of Planned Parenthood clinics and the
3. NPR itself uses “pro-choice” and “pro-life,” because it’s “more conversational and
not as cumbersome” (Shepard 2011).
4. This information was gathered as part of a Pew Research Project on the influence
of religion in the United States (Pew Research 2014).
T H E R H E T OR I C OF R E P R O D U C T I O N [ 333 ]
334
introduction of state legislation that would assign full human rights to fer-
tilized eggs5 (Associated Press 2012), have led to a new type of discourse
that could not have been foreseen from Lakoff’s vantage point in 1992—
that of social media, and the online social networking service Twitter, in
particular. Here, women felt free to discuss their own abortions openly,
rejecting the stigma of abortion itself, as they shared their stories with
each other publicly. In contrast to the type of organized, marketable rheto-
ric analyzed and suggested by Lakoff, the Twitter hashtag (searchable key-
word) #ShoutYourAbortion arose organically as a response to the efforts
to cut Planned Parenthood funding by the U.S. House of Representatives.
The American feminist and activist Lindy West explained her motivation
for catalyzing the moment: “I set up #ShoutYourAbortion because I am not
sorry, and I will not whisper” (West 2015). Tens of thousands of women
responded with the experiences of their own abortions, reclaiming the
word—not with a quiet voice, but with a shout.
The war on women’s rights is far from over, but it is clear that the new
channels of social media have provided an important alternative to the
type of rhetorical discourse Lakoff examines here. The same questions
remain: What is life? What is a person? How shall we speak of such things?
Perhaps the new access to public discourse will allow women to confront
these questions, as Lakoff emphasizes, “directly, intelligently, and fear-
lessly”—and in their own words.
REFERENCES
Associated Press. 2012. “Va. House GOP muscles through abortion curbs.” USA Today,
February 14, 2012, http://usatoday30.usatoday.com/news/nation/story/2012-
02-14/virginia-abortion-legislation/53097654/1
Guttmacher Institute. 2015. Facts on induced abortion worldwide.
New York: Guttmacher Institute, http://www.guttmacher.org/pubs/f b_IAW.
html
Pew Research Center. 2014. “Public sees religion’s influence waning.” September 22,
2104. Washington, DC: Pew Research Center. 37.
Shepard, Alicia C. 2011. “In the abortion debate, words matter.” NPR Ombudsman,
March 18, 2010, http://www.npr.org/sections/ombudsman/2010/03/in_the_
abortion_debate_words_m_1.html
West, Lindy. 2015. “I set up #ShoutYourAbortion because I am not
sorry, and I will not whisper.” Guardian [UK], September 22, 2015,
http://www.theguardian.com/commentisfree/2015/sep/22/
i-set-up-shoutyourabortion-because-i-am-not-sorry-and-i-will-not-whisper
5. Although the Virginia State Senate shelved this bill in 2012, it has been reintro-
duced every year since.
[ 334 ] Context Counts
335
CHAPTER 14
In the propaganda war over reproductive rights, the forces arrayed against
choice generally represent themselves as the moral side, their way as the
high road. They portray the pro-choice position as a slippery slope of
expediency. Yet, while flashy, deceptive rhetoric seldom is considered the
ethical or moral approach, opponents of choice make regular use of such
illegitimate rhetorical strategies. That they claim to act out of religious
motives only exacerbates the dilemma such strategies must pose for them.
Confronted with this paradox, they have argued that because rhetorical
razzle-dazzle works, and the cause is worthy, the end justifies the means—
an argument rarely associated with the high moral ground.
The advocates of reproductive rights, often accused of expediency and
slippery ethical stands, are in the opposite position. Their rhetoric is above-
board and eminently reasoned; but it has been, considering the significance
of its message, singularly unsuccessful in persuasion. When this problem
is discussed, suggestions for more enticing rhetoric often are dismissed, on
the grounds that persuasive techniques that circumvent the intellect are
dishonest and dishonorable—an argument not typically associated with
expediency.
It is time we become concerned with the ways in which anti-choice rhe-
torical strategies are deceptive and hypocritical, and time we find means to
This paper first appeared in Conscience 13(2):4–12, 1992. It is reprinted here with
permission of the publisher. Conscience is the newsjournal of Catholics for Choice,
Washington, DC.
336
overcome them. This task will require an understanding of the real reasons
for many people’s enthusiasm for that side. Further, an effective opposing
strategy must discover why the persuasive attempts of reproductive-rights
advocates have been less than successful, and it must devise ways to change
the balance of the debate.
Rhetorical skill never has been more important in this struggle. As the
U.S. Supreme Court dismantles Roe v. Wade,1 abortion is thrown more and
more into the political arena, to become the business of legislatures in the
fifty states as well as the Congress. It is not so much that effective rhetori-
cal strategies are needed to influence legislators; rhetoric seldom influences
legislators—votes do. But voters can be affected by rhetoric, as indeed they
have been by pro-life rhetoric. Now it must be our rhetoric.
Following the Supreme Court’s Roe opinion in 1973, the anti-choice
side wasted no effort in developing and putting into practice hard-hitting,
gut-wrenching forms of persuasion. As many of its arguments have been
disproved or have become ineffectual through overuse, even more compel-
ling tactics have taken their place. Among these are sloganeering, semantic
obfuscation, and slick advertising (verbal and pictorial). While these are
not independent of one another, it is useful to discuss each separately to
see how they work independently as well as together.
[ 336 ] Context Counts
337
“Coke adds life.” So what is more natural (to borrow a technical term) than
to unite the two?
Other slogans are close relatives, for example, the refrain of the slick
new Arthur S. DeMoss Foundation television spots, “Life. What a beautiful
choice.” Notable in right-to-life campaigns is the deliberate cooptation of
the language of the opposition—“choice”—which would, in the world of
Mammon, be actionable as trademark infringement.
But that is the least of the ethical quagmires into which such slogans,
and the Madison Avenue expertise behind them, lead their sponsors. In
the first place, it has been noted often that advertising dollars are spent
most freely to promote those products that respond to no real need.
Advertising slogans suggest reasons to choose among many products that
are either identical to each other or irrelevant to real needs. Advertising
offers apparent, but unreal, choices to consumers. The illusion of choice
hides the fact that we have few real choices to make as consumers. In
this sense, advertising itself constitutes a lie or at least a false statement
(quite aside from specific truth-in-advertising problems like marbles in
the soup)—behavior that is immoral on its face. For religion to be mean-
ingful, it must ally itself with the forces of morality. For religion to form
alliances with immoral or at best ambiguous forces is to make a mockery
of its reason for being.
The irony is compounded and deepened in the debate over reproductive
rights. Here not only does the advertising obfuscate the reasons for which
people choose abortion; so far, this is merely the business of advertising.
But advertising in the reproductive arena also goes further. A fake choice—
“choosing life” for lack of any other option—is offered in exchange for the
ability to make a real and meaningful choice. Beguiled by the slogans, the
pictures of dewy infants, one renounces the human right to control one’s
personal destiny, to make informed decisions about issues affecting one’s
very life and body.
Juxtaposing “life,” “natural,” and “choice” is a masterstroke. Every adver-
tiser knows that those are among the ten words that provoke the strongest
positive responses. But what is “life” in this context, and who has it? When
two “lives” are at odds, whose is the more “natural”?
Worse still, the sloganeering obscures a delicate psychological problem
and theological problem. That is, do you in fact make a “choice” for life,
natural or not, if you have been driven to that decision by fear or emotional
blackmail? If you “choose life” under these circumstances, can your deci-
sion be considered an act of virtue or piety? And if abortion is made a crime
and is already a sin, can “life” still meaningfully be called a “choice”?
SEMANTIC OBFUSCATION
Debaters who believe their positions are right will choose language that
does not intentionally create ambiguities or otherwise mislead listeners.
Such debaters will be confident that (a) their position is sufficiently strong
as to prevail on its own merits, and (b) listeners are intelligent enough to
be trusted to select the winner on rational grounds. The anti-abortionists
show, by their preference for slippery semantics, that they disbelieve one
or both propositions.
Partial equation masquerading as total equivalency is one favorite tech-
nique, especially when buttressed by another propagandist’s trick, presup-
posing the equivalence rather than stating it overtly. An example of both
tactics is the use of words like “child” or “baby” to refer to a first-trimester
fetus or even an embryo (the accurate but rarely used term during the first
eight weeks of development). In anti-choice statements, the equivalence of
“fetus” and “child” is taken for granted, so that something else is asserted,
depending on it: “Baby killing is murder.”
A smooth new use of this unstated equivalence is the latest DeMoss
slogan, “Everyone deserves a chance to be born.” Everyone presumes the
personhood of its referent. The verb it agrees with, deserves, strengthens
that presumption; personhood is normally a precondition for “deserving”
anything. So the slogan both presupposes and asserts the purported sen-
tience of the fetus—a form of argumentation that is both logically invalid
and ethically dishonest.
2. George H. W. Bush added passion but little light to the discussion in his 1989
inaugural address, in which he spoke of “young women … who are about to become
mothers of children they can’t care for and might not love … though we bless them
for choosing life.”
[ 338 ] Context Counts
339
“The argument is over when life begins,” [Reagan] said. “Well, look, if that’s the
argument: If there’s a bag in the gutter and you don’t know if what’s in it is alive,
you don’t kick it, do you?” (Noonan 1990:159–60)
In other words, you forbear to hurt a woman not because it might injure
her, as a fellow human being or fellow creature, but because she might be
carrying a life you wouldn’t want to hurt.
A third problematic equivalence, also typically implicit, is the equation
of abortion with slavery. Everyone agrees that slavery is evil. Most also
would agree that extreme measures (including civil war) are justified in
order to extirpate it. The rhetoric of the anti-slavery movement was full of
passion, passion justified by the cause it supported. So the anti-abortion/
[ 340 ] Context Counts
341
A serious moral cause should use its terms literally. Suppose we take “pro-
life” literally, that is, assume that its proponents are unequivocal support-
ers of “life” in all its manifestations. (Let us grant them a little latitude,
interpreting “life” as “human life” in their extended sense, so as to exclude
animal rights.) Then it would stand to reason that any person or organiza-
tion claiming a “pro-life” position would have to take a public, explicit, and
unambiguous stand that was:
1. Anti-war because war kills. One hundred thousand Iraqi children cannot
be less valuable than one unborn American embryo.*
2. Opposed to capital punishment. It is argued that execution is the taking of
“non-innocent” human life. But who is to be the judge of that? Moreover,
the simultaneous belief in capital punishment and intolerance of abor-
tion leads to a serious moral paradox. Murder is the unlawful killing of
a human being. Capital murder, in those states that employ that statu-
tory term, is premeditated murder committed under circumstances so
heinous as to justify execution. One such circumstance under these laws
is “killing for financial gain,” which subjects both parties in the hiring
agreement to capital punishment. If a fetus is defined as a human being,
then abortion becomes premeditated murder and, further, capital mur-
der, as the doctor usually is paid by the woman. Then, once it is legally
established that an abortion has taken place, two human deaths beyond
that of the fetus are mandated. Is this “pro-life”? Prosecutors might
agree tacitly that the death penalty never would be sought in such cases,
but that would be piling hypocrisy on hypocrisy.
3. Absolutist. All abortions would be forbidden. While perhaps abortion
to save the mother’s life would be debatable (but by no means a guar-
anteed right), the termination of pregnancies incurred through rape or
incest would be precluded; those fetuses are as human and “innocent” as
*
This article was written in the wake of the 1991 Persian Gulf War, in which a U.S.-led
coalition invaded the nation of Iraq.—Ed.
any others. But the majority of pro-lifers would permit abortions under
these circumstances, leading to the suspicion that they see pregnancy
as punishing a woman for having “fun”; as long as the sex demonstrably
wasn’t fun, abortion is countenanced. (The anti-abortionists also realize,
to be sure, that the absolutist position would turn many in the middle
against their position. But if this motivates their position, they are trad-
ing moral consistency for a few votes—again, not the higher choice.)
4. In favor of contraception and sex education. If the aim of the movement is
to prevent the destruction of human life, then high on its agenda should
be avoiding the conception of unwanted human life. Many anti-abortion
groups claim to be “neutral” on these issues. But others—notably the
Roman Catholic bishops and Randall Terry (founder of the anti-abortion
group Operation Rescue)—actively oppose contraception and compre-
hensive sex education. Moreover, even neutrality is inappropriate and
hypocritical here.
. Opposed to violence, including clinic-blockade tactics—such as obstruc-
5
tion, jeering, and shoving—that incite violence, as violence tends to
lead to injury and death.
Further, a position that was strongly pro-child and based on the pas-
sionate love for children that anti-abortionists claim as their motivation
would have to take certain explicit positions, to avoid a charge of hypocrisy
or at least confusion:
[ 342 ] Context Counts
343
unconscious, to one that is false but can be acknowledged publicly and pri-
vately.5 I assume that the majority of “pro-lifers” are sincere and intelligent
people who are, nevertheless, driven by forces they cannot control or admit
to consciousness. Then it becomes clear why their dubious rhetoric has the
power to sway many people and why, when pro-choice rhetoric rationally
addresses the ostensible arguments of the anti-abortionists, it has little
persuasive force.
The underlying emotion of “pro-life” people is not love but fear, which is
at least as strong a motivator but is less comfortably admitted. It is a fear
of change in a rapidly changing world, of loss of control, of no longer under-
standing what is expected, of no longer being able to count on actions hav-
ing predictable consequences. The fear arouses a desire for “business as
usual” or, even better, a return to a golden age, when men were men and
women were women, with sharply polarized roles: men held power and
women didn’t have it or desire it. There was certainty: the authorities of
church and state set forth unambiguous standards for behavior, with seri-
ous punishments for falls from grace. Children respected their parents; vir-
tue was honored; and individuals had relatively few choices, making life (if
less pleasant, in some respects, than we are used to) quite simple.
Many innovations over the last century or so (especially in the United
States) have rendered this myth inoperative. Perhaps the most significant
is the gradual acquisition by women of control over their bodies, beginning
with laws against wife beating. Women’s control over their bodies came
much closer to reality with the mid-20th-century development of reliable
contraceptives, culminating in the Pill. Much of the social ferment of the
1960s and the ensuing social changes stem from the availability of effective
contraception, which freed women not only sexually but also profession-
ally and politically. The Supreme Court’s decision in Griswold v. Connecticut6
ensured that contraception was a legal right: the genie was out of the bot-
tle. But it took many years for these changes to have their full effects on
American life.
In fact, the ferment peaked just in time to coincide with Roe v. Wade,
which thus became the scapegoat for all the distress the earlier changes
evoked in those unready for change. The Pill was a reality, and Griswold was
firmly enshrined in law and enjoying too high a degree of public approval to
be criticized. But Roe was different; unlike Griswold, it could be construed
5. Displacement is defined as “the process by which the individual shifts interest
from one object or activity to another in such a way that the latter becomes an equiva-
lent or substitute for the other” (Rycroft 1973:35).
6. Griswold v. Connecticut, 1965 U.S. LEXIS 2282 (1965).
[ 344 ] Context Counts
345
A MODEST PROPOSAL
• Women are not slaves (with the Battle Hymn of the Republic in the
background)
• Remember the Middle Ages? (with a picture of a woman being burned at
the stake)
• Forced motherhood—the “choice” of the past
A REAL CHOICE
Those possibilities are entertaining, and I would not dismiss them out-
right. But before we rush to hire our own propaganda thugs, let us try once
more to do the right thing. We must appeal to Americans, male and female,
as rational and caring people, people who will take the right public-policy
position if the false attractions of the wrong position are made manifest.
Our rhetoric should compel recognition of the fact that life is full of gray
areas, which make personal decision making (rather than government
dictates) all the more important. Choices are hard precisely because there
are pros and cons, ifs, ands, and buts, and modern life brings with it ever-
harder choices. That is why flashy rhetoric is dangerous. It makes it impos-
sible to think rationally through the options. We must expose the cynicism
of the other side, always emphasizing that women are smart enough and
brave enough and moral enough to choose.
We must acknowledge that there is a vast gray territory: What is life,
what is human life, and when does it start? What is a person, a human being,
and what are a person’s rights, duties, and entitlements? The mindless
intoning of slogans temporarily makes those difficult questions seem easy,
but in the long run, the questions must be confronted directly, intelligently,
[ 346 ] Context Counts
347
REFERENCES
W hen I first read this piece, I was transported back to the classroom,
listening to one of Robin Tolmach Lakoff’s extraordinary lectures.
It’s all here: the clear prose style, the memorable turn of phrase, the
straightforward presentation of stance and theory, the careful laying out of
analytical tools, the precisely mapped-out line of reasoning, supported by
just the right examples, and, throughout, the perfectly timed wit. Also here
is the deceptive simplicity of presentation. Like many other gifted lectur-
ers, Lakoff moves with such ease from point to point that listening to her
go over some complex analysis can leave you feeling that, really, the whole
thing is perfectly obvious, and perhaps even wondering, with a touch of
hubris, why you hadn’t thought of it yourself.
Such clarity has its pitfalls for the unwary. When I started taking courses
from Professor Lakoff, I was a sophomore, and it was in that first year or
so of the 1970s that still counted as part of “the sixties.” We returned to
the earth and boasted of our peasant roots, whether we actually had them
or not. Handicrafts, along with granny glasses and prairie skirts, became
popular. In fact, knitting or crocheting in class became something of a fad,
and in many classes you could find a couple of students, usually but not
always women, working on a sweater or a cap or pieces of an afghan. I don’t
recall that anyone noted at the time that it was, at the very least, extremely
350
rude to sit in class and look otherwise occupied. The knitter wouldn’t admit
to being distracted and probably didn’t think that others might be.
Rudeness aside, an accomplished knitter could probably knit easily
enough while absorbing the lecture content. I, however, was not an accom-
plished knitter. Nonetheless, in the second or third class I took with Lakoff,
I started work on something or other and, recalling how very clear her lec-
tures always were, foolishly imagined that I would be able to knit and keep
up with the analysis. It was an act of rudeness for which, Robin, I belatedly
but no less shamefacedly apologize.
It was also a serious mistake. I learned quickly—not quickly enough—
that I had to pay attention, and that it wasn’t going to work if I was going
to be counting stitches at crucial points in a carefully constructed argu-
ment. After a few weeks of frantically increased study time to make up for
my lack of concentration in class, I concluded that the knitting would be
best left at home. Besides, I realized, it was silly to distract myself from the
class I enjoyed most.
[ 350 ] Context Counts
351
There is, of course, much that could be said, if space permitted, about
the peculiarities of a “community” in which the competent members are
law-enforcement personnel and experienced criminals, while at least a
significant number of those who become involved with this community
are not competent in its interactional patterns. This is one of the issues
explored in Tom Wolfe’s Bonfire of the vanities (1987), which follows the
story of Sherman McCoy, a successful investment banker who gets in trou-
ble with the law, unwillingly develops expertise in the roles and behaviors
of the world of the accused suspect, and, somewhat to his own surprise,
finds himself taking pride in his new competency and, indeed, in the new
identity it gives him.
This development of identities, how we speak of ourselves, how we cre-
ate and sustain our public personae, is not a simple issue. It is delicate
enough when we are interacting with our peers on equal grounds. It is
more complex when our peers are judging us. And it becomes overwhelm-
ingly problematic when we are presenting ourselves to people who perceive
themselves entirely as judges rather than peers—that is, when our disposal
is in the hands of those who believe themselves to share very little common
ground with us. Such is Virgil Reilly’s case.
For purposes of comparison, we might look at the case of Karla Faye
Tucker, a confessed ax murderer who became a born-again Christian in
prison (Pressley 1998). Tucker was ultimately executed on February 4,
1998, but her case attracted widespread attention, in part because many
proponents of the death penalty, including many evangelical Christians,
supported reduction of her sentence to life imprisonment. Tucker’s appeal
hinged on issues of identity: when she committed the crime, she was, in
the words of one news account, “drug-addled”; now, it was argued, no lon-
ger using drugs and having experienced a religious conversion, she was
in essence a different person. More than the simple facts of the case, the
appeal was based strongly on the construction of the audience as peo-
ple who believed in the power of religious conversion. This may explain
why claims of personal change that moved members of the evangelical
Christian community may have been less moving to the Texas State Board
of Pardons and Paroles, which has doubtless heard more than a few claims
that a religious conversion has given an inmate what amounts to a new life
and indeed a new identity.
Nonetheless, Tucker did gain a hearing from some normally reluctant
members of the broader society. And, while some of that may be attributed
to her status as an attractive, articulate, and young white woman in jeop-
ardy, still, that achievement was to a significant degree the result of her abil-
ity to divorce herself rhetorically from her earlier “murderer” identity by
demonstrating that she identified with the values and beliefs of the larger
society and stood with them in condemnation of the “old” Karla Faye Tucker.
The problem is, of course, that it is difficult to find any sort of clear divid-
ing line between Reilly’s confused and possibly pathological disconnection
from his own actions and Tucker’s carefully constructed (and religiously
sanctioned) rhetorical self-distancing. Nor is there a clear boundary between
those and the kind of casual and commonplace self-distancing of normal
interaction. Imagine a situation in which you and I meet at a conference (let
us say) and enter into a conversation, part of which involves trading stories
about our favorite professors, about stupid things we did when young, or
about how behavior changes with time—the “true confession” with which
I began this introduction could fit into any of those categories. If I told you
that story in that kind of casual context, you would be unsurprised, in all
likelihood, by my self-distancing “What could I have been thinking?”
But that is another element of Lakoff’s work, for her longstanding inter-
est in special types of language use (language associated with the law, with
advertising, with politics) works in both directions. That is, while showing
how linguistics can inform our understanding of legal (or advertising or
political) issues, she demonstrates that the special demands and circum-
stances of legal (or advertising or political) contexts can serve as a labora-
tory in which to test our assumptions about day-to-day communication.
As a result, her work has located itself on the boundaries linguistics
shares with psychology, in its concern with the representation of identity;
with anthropology, in its work on the construction of community; and with
rhetoric, in its efforts to account for decisions made by individual speakers
selecting from the available choices in order to create a specific effect in the
audience. In short, Lakoff’s work was interdisciplinary before interdiscipli-
narity became quite as popular as it is now, and the students she trained
became, as a result, well equipped to handle interdisciplinary issues. It is in
fact in these interdisciplinary connections that we are most likely to find
answers to the kinds of difficult real-world questions that Lakoff’s work
consistently forces us to examine.
REFERENCES
Pressley, Sue Anne. 1998. For first time since Civil War, Texas executes a woman.
Washington Post, 4 February 1998, A1, A10.
Wolfe, Tom. 1987. Bonfire of the vanities. New York: Farrar, Straus and Giroux.
[ 352 ] Context Counts
353
CHAPTER 15
1. INTRODUCTION
There are two especially difficult problems for pragmatics and sociolin-
guistics. One is the connection between purely linguistic form (phonology,
syntax, morphology) and discourse function. Eventually we must, however
reluctantly, consider to what degree and in what ways real-world situations
and communicative needs of speakers govern syntactic form; and the sense
in which syntax, semantics, and pragmatics are truly, bidirectionally, inter-
dependent. The second is the definition of “pragmatic” or “communicative”
competence: What does the normally competent speaker know?
The case I will discuss here raises another issue of interest to linguistics
and other social sciences. Over the last fifteen years or so, there has been
much discussion in several fields concerning the nature and/or reality of
the “self.” Some of the data we will be examining provide more evidence of
the tenuousness and fuzziness of that concept.
The examples I will use are drawn from a pair of criminal confessions in
a death-penalty case currently (June 1993) under appellate review. I was
This paper originally appeared in Social Interaction, Social Context, and Language: Essays
in Honor of Susan Ervin-Tripp, edited by Dan Isaac Slobin, Julie Gerhardt, Amy Kyratzis,
and Jiansheng Guo, 481– 93. Mahwah, NJ: Lawrence Erlbaum Associates, 1996.
Reprinted here with permission of the publisher.
354
2. FACTS OF THE CASE
The facts of the case are briefly these. When the crimes were committed,
in November 1984, Virgil Reilly was a black male of twenty-five, who had
never been in any previous legal trouble.
VR’s IQ falls between the mid-sixties and mid-seventies—borderline
or “dull normal.” Psychological testing reveals an array of impairments—
particularly relevant to our concerns are verbal deficits, including “severe
impairment in attention, memory for verbally presented material, verbal
fluency, cognitive flexibility, and adaptability to novel cognitive tasks.”1
He is a native speaker of English, and has attended high school. He
reports, prior to receiving a Miranda warning during one of his confes-
sions, that while he “know[s]how to read,” he doesn’t read “that good you
know” (SB 2:5,7).2
VR’s life prior to the crimes to which he is confessing was typically
dreary for a defendant convicted of a capital crime. His mother used
alcohol rather heavily during her pregnancy. She was twenty when VR
was born, three months prematurely, after a difficult pregnancy, the
first of eight children born over a twelve-year period. VR was regularly
abused by his father and other relatives (physically and sexually) and
neglected by his mother. He also displayed, from early childhood, a vari-
ety of physical and psychological problems, created or exacerbated by
the abuse.
In October of 1984 he had begun to use large quantities of over-the-
counter and prescription analgesic drugs (especially Tylenol with codeine)
to self-medicate a variety of complaints, including but not restricted to per-
sistent headaches and pain from dental procedures. He was also drinking
large quantities of wine.
[ 354 ] Context Counts
355
3. Miranda v. Arizona, 384 U.S. 436 (1966). At 479, Miranda states that the suspect
“must be warned prior to any questioning that he has the right to remain silent, that
anything he says can be used against him in a court of law, that he has the right to the
presence of an attorney, and that if he cannot afford an attorney one will be appointed
for him prior to any questioning if he so desires.” Receiving these warnings will enable
him to “knowingly and intelligently waive these rights.”
4. For the definitions and functions of these terms, cf. Austin 1962 and Searle 1969
and 1979; for conversational logic, Grice 1975, and for conversational structure, Sacks,
Schegloff & Jefferson 1974.
[ 356 ] Context Counts
357
5. For discussion of this term, as well as the notion of truth-seeking versus other
discourse genres, cf. Lakoff 1990.
[ 358 ] Context Counts
359
And shortly later:
[ 360 ] Context Counts
361
For a confession to have been made with informed consent, the suspect
must realize that the confessional frame itself places the confessor in jeop-
ardy, and that that risk is the interrogator’s desired outcome: it is a win-
lose situation. Since as a criminal suspect VR enters the discourse under
a disadvantage, full understanding would make clear to him the neces-
sity of keeping up the boundaries, maintaining distance and formality at
all points, giving no more information than is essential, and not expect-
ing the interlocutors to do anything to his advantage. This is the reason
why Miranda requires interrogators to offer suspects explicitly the right
to have an attorney present, and also why interrogators do their best to
discourage suspects from invoking those rights6: the presence of an attor-
ney would substantially equalize the participants. In VR’s case, the waiver
was almost too easily accomplished. In the SB confession, for example, the
Miranda statement is read rapidly and without inflection, in long phrases.
So while VR might have understood all or most of the words used, in iso-
lation, it is very possible that the warning as uttered, presented to him
in a state of agitation and exhaustion, might not have been fully compre-
hended. He answers, “Uh huh,” but it’s not clear whether this response
signifies understanding or merely compliance: “Just do what you like, fine
with me.” In the Riverside confession, after the detective has read VR his
rights, he asks: “Having these rights in mind, [Virgil], do you wish to talk
to us now?” (R 2:10). There follows a four-second pause on the tape, an
indication that VR fails to understand fully the requirement that, at this
explicit transition-relevance place, he take a turn. As the silence length-
ens, the second interrogator takes over. There is a marked stylistic shift
between Cornejo, who speaks first, and Ferguson: the former is formal
and distant, the latter spontaneous-sounding and colloquial. VR responds
At this point, Cornejo takes over again and makes VR’s role in the confession
into a kind of voluntary game: “Well, I’ll ask you a question. If you want to
answer it, answer it” (R 2:13). Although VR has never explicitly in so many
words indicated his willingness to give a confession, he is now locked into
a discourse frame in which he would have to aggressively perform an active
(rather than passive) violation of conversational preference rules: refuse
to carry out a directive. For someone who has been repeatedly diagnosed
as having a compliant personality (and someone with few communicative
skills), this would be particularly difficult. From VR’s subsequent behavior,
it seems clear that, rather than recognizing Cornejo’s Miranda warning as a
statement of a boundary, as the beginning of an adversarial and high-risk
interrogation frame, VR is encouraged by Ferguson’s quasi-offer of help to
see the ensuing discourse as collaborative, win-win: a safe place wherein he
can ask for advice, or openly digress about things that trouble him, in the
hope of getting help or clarification, as in examples (1)–(3) above, and his
repeated questions such as “What should I do?” “What should I say?”
[ 362 ] Context Counts
363
speaker making the confession is the same person as the one who per-
formed the action being acknowledged.7 That is, an identity relation is pre-
supposed between the two: they are, in some sense, “the same” person.
So stated, the relationship seems simple and unambiguous: either A1 = A2,
and (other things being equal) the confession is appropriate; or A1 ≠ A2, in
which case it is aberrant—both semantically meaningless and pragmati-
cally nonfunctional. The clearest cases of the latter are literally false confes-
sions, in which speakers “confess” to deeds they have not done. Such cases
are recognized as inappropriate speech acts, and have no further legal nor
moral status as confessions.
That is the simplest and most obvious problem case. It suggests that
identity relations are analyzable dichotomously: A1 = A2, or the reverse.
But in fact, both in human life and in syntactic construction, there exist
numerous intermediate possibilities: possibilities abhorred alike by lin-
guists, psychologists, and legal scholars, but nonetheless very real. Because
both our minds and (therefore) our language seem to demand strictly
delimited categories rather than continua, we either refuse to recognize
such cases, or relegate them to the margins of our analytic systems as “spe-
cial cases,” willfully refusing to recognize them as central and highly signifi-
cant. Yet these cases define us and are crucial to our language, and must be
acknowledged, and eventually accounted for, in a complete theory linking
language, mind, and the real world in which and on which they operate.
At this time, in Western culture, it is commonplace to feel that “normal”
human beings possess a unified and cohesive “self”—a persona that is felt
to be constant over time as a cohesive and rational agent, a person more or
less consistent at all times, under all conditions. These definitions seem to
us basic, self-evident, and eternal, as aspects of the human condition.
Over the last dozen years or so, postmodernist scholars in a diverse
array of fields have questioned and largely discarded the certainty and
universality of the concept of selfhood.8 This research demonstrates per-
suasively that the “self” is only a convenient organizing fiction of the
7. This definition is somewhat too narrow as stated, in that under some conditions it
is possible to “confess” appropriately to an act performed by someone with a close con-
nection to the confessor, and for whom the confessor has some degree of responsibility
or control—rather like the case of apology. For instance, a mother might confess that
her small child had damaged a neighbor’s property. But even in this case there is still a
presumptive identity relation between the confessor and the doer of the action, albeit
at one remove.
8. For example: in history, work by Davis and Starn (e.g., 1989); in anthropology,
perhaps the earliest and most influential discussion is that of Geertz (1983, ch. 3);
in discourse analysis, cf. Linde 1993; in the field of ethics, cf. the collection edited by
Johnson (1993); and in psychology, cf. Sass 1992.
It might be argued that even in these “simple” cases, there are two distinct
notions of “identity” operating: in (5i) it is psychological, but in (5ii) physi-
cal. The problem only worsens with
in which the obvious identity of subject and object implied in (5i) and (5ii)
gives way to a split between the interests and perspectives of subject and
object, yet not a complete schism. The complexity increases in (5iv),
in which there is still a notion of identity between the subjects of the two
clauses, but not the complete merger assumed in (5v):
In each of the sets in (6), both examples represent some sense of iden-
tity between the subject and direct object of the clause. If identity, or
[ 364 ] Context Counts
365
full recognition –- “neurotic” –- multiple personality–- retrograde –- schizophrenia
of identity noncohesion disorder amnesia10
9. For extensive and interesting discussion of several problems of this kind, see
A. Lakoff & Becker 1991.
10. Some clarification of the points on this scale may be helpful. The second point,
“ ‘neurotic” noncohesion,” refers to the argument in much recent psychoanalytic
writing that what brings people into therapy is a disorganized or incoherent life-
story narrative. Therapy, then, consists of restoring the narrative to coherency. In
this situation, the patient senses gaps and discrepancies among points in the narra-
tive, but recognizes him-or herself as its subject throughout. “Multiple personality
disorder” (MPD), is a condition in which, as a result of early and persistent abuse,
patients experience themselves as fragmented into multiple discrete personalities
with no continuity of memory or personality among them. Each one typically main-
tains its own continuity across time; but synchronically the patient feels divided
into multiple selves, rather than selfless. Retrograde amnesia is that rare state
beloved by writers of soap opera in which, as a result of physical or mental trauma,
memory of all past history is lost. Hence the individual has no “self ” at all, no mem-
ory providing an identity and a means of predicting and stabilizing future behavior.
Finally, in schizophrenia, especially as interpreted by Sass (1992), the individual’s
total sense of self, synchronic and diachronic, is shattered, and boundaries between
the self (physical and psychological) and other individuals and the outside world
cease to exist.
(7) Cornejo: You don’t sound like [you’re upset]. I’ve accused you of
killing a little boy and you haven’t yelled at me and said
“You’re a goddamned liar,” or nothing. You just sit there as
calm as can be. I don’t understand your attitude.
VR: Well, I’ve never, you know, there’s something that I did,
I know I didn’t do it, you know.
(R 23:1-2)
(9) VR: But, but if I killed somebody I am deeply, deeply sorry that I’ve
done it. By being knowing myself, I probably did kill somebody.
(R 27:8)
And he speaks similarly later on, even though by that point a more ade-
quate “confession” had already been obtained:
[ 366 ] Context Counts
367
(11) Cornejo: This is the little boy right here. Handsome little boy, isn’t
he? Eleven years old. A handsome eleven-year-old little boy.
VR: I killed him?
Cornejo: You killed him [Virgil]?
VR: I killed him?
Cornejo: Uh-hum.
VR: I don’t know why I killed him.
(R 30:17-22)
[ 368 ] Context Counts
369
REFERENCES
CHAPTER 16
Afterword
BY ROBIN L AKOFF
1.
There has been some commentary in the popular media about a current
tendency to premature, and postmodern, autobiography: people of ten-
der age and therefore moderate accomplishments publishing full-length
autobiographical statements whose factuality is open to dispute, if they
are not frankly fictional or “fictionalized.” In this case, the first criticism
is easily disposed of, since I am no longer of tender age. The second is
more troublesome. It is worth asking how truthful any autobiographical
statement can be, and how truthful this one actually is. All I can say is
that I think the way I describe my career as represented in these papers
is the way it happened; what I say that I meant at the time to accomplish
was indeed my intention; my current interpretations of what I was saying
are at least not strongly antithetical to the ones I would have made, had
I been asked, contemporaneously. But of course we know memory is slip-
pery. Caveat lector.
The academic autobiography is a genre with relatively few prototypes
available for imitation. Should I look to one of the heroes of my youth,
Julius Caesar, and write in the third person—suggesting disinterestedness
and thus reliability? Or represent my life as travelogue—first I went here,
then I moved there; the natives were friendly, the scenery enchanting… .
The reader may be getting impatient here: Why not just get down to work,
move through the material, leave yourself out of it like a good academic?
But these papers are my self. They were a part of me when I wrote them,
372
and they still are close to me, like children. And no one can reasonably
claim to be objective about her children.
At the same time, this is an academic retrospective. Therefore I may, or
must, omit all sorts of juicy matters, and include some rather dry ones.
Arguably I am writing not about my self, but about my work, if these can
be differentiated.
A possible compromise is to use the texts as the bases of my discussion,
focal points; but allow myself to reflect in passing not only on what the
papers in and of themselves were about, but also why I wrote them, why
I wrote them as I did, and when I did; why I said some things explicitly,
some more or less roundaboutly, and others not at all—often things that,
seen from the vantage point of the present, deserve lengthy and explicit
explication. In that sense, scholarly explication du texte meets personal
memoir. Let’s see what happens.
2.
“Once,” says the Mock Turtle in Alice’s Adventures in Wonderland, “I was a real
Turtle.” And once, reader, I was a real Linguist. In fact, before that I was even
realer: I was a classicist, more specifically, a Latinist. Like his schoolmate the
Gryphon, I studied Laughing and Grief. I have retained a soft spot in my heart
for Latin, and go back to it from time to time, on any excuse I can concoct.
But I was young back then, in the early 1960s, and I was in Cambridge,
Massachusetts, and I was hungry for answers, real truths, such as the study
of a language and its literature seemed unable to provide. At that time,
in that place, there was a compelling alternative: down the block from my
own institution, the field of linguistics was being re-created, language itself
being redefined. I could spend the morning in the tenth century bce (an
amusing enough place, but not au courant), get on the bus, and get off at
RIGHT NOW, in fact, TOMORROW. It was irresistible. At MIT, linguistics
was a combination of cutting-edge science and religious cult—a delectable
if corrupting combination for the youth. So I became a linguist, first an
Indo-Europeanist and then a transformational grammarian.
For a few years we thought, my cohorts and I, that we were following
the approved transformational program. But a problem arose because (as
is not uncommon in cults) the program itself was not completely clear,
and subject to sudden reinterpretations ex cathedra. Many of us had joined
up on the promise that the goal of transformational generative gram-
mar was to use language as a “window into the mind,” a way of rigorously
[ 372 ] Context Counts
373
A fterword [ 373 ]
374
3.
Though these origins can be seen in their clearest form in the first three
chapters (NB: chapter 3 includes excerpts from two papers written during
this period), written under the aegis of the GS project or during its imme-
diate aftermath, much of what I have done since then is also better under-
stood if their GS roots are made explicit. For this reason I am spending
what may seem a profligate amount of space on the development of GS and
its influence on my work. But I think it is generally true that the beliefs of
one’s formative years continue to exert a strong, if sometimes subliminal,
influence on the work of one’s maturity. Besides, the motives and findings
of GS have been misunderstood and maligned by writers who have little or
no personal knowledge of the movement, and I think it’s important to try
to set the record straight. Finally, I am spending a disproportionate amount
of time on the early papers because, as the theoretical (and theological)
premises under which they were written fade into the mists of time, they
become almost impossible to understand without a gloss.
In his “Retrospective Epilogue” to his posthumously published collec-
tion of writings, Studies in the way of words (1991), the philosopher Paul
Grice sorts the nineteen papers included in the volume into eight “strands.”
In the fashion of analytic philosophy, he doesn’t discuss his choice of ter-
minology. But I think it a profound choice, one well worth imitating; and
perhaps, in the forthright manner of a linguist, I can explain why I find the
notion of “strands” of topics useful enough to appropriate.
Strands are long and thin, like necklaces or pasta. A strand of a necklace
can stand alone, but it is in combination with others that it achieves sub-
stantiality and beauty. The strands are sometimes left separate, one on top
of the other, but often intertwined, so that the observer cannot tell one
from another. Likewise with pasta—while you can eat spaghetti strand by
strand, it is not really satisfying to do so. Tangled together, the spaghetti
picks up the sauce and other goodies, making a complex and interwoven
experience for the eater. Attractive, substantial, and tasty—that’s how
I hope you will see these papers, so I offer them as strands of a complex and
composite totality.
Something else that comes in strands is DNA. DNA differs from spa-
ghetti or necklaces in that its strands may reorganize themselves. An item
originally on one strand may link up with another, with interesting and
[ 374 ] Context Counts
375
4.
The first three chapters here represent an attempt to find a way to incorpo-
rate extralinguistic information into the syntactic component of a trans-
formational generative grammar. In many ways they are characteristic of
the whole GS project, especially in its earlier stages, both in terms of con-
tent and style. They are an attempt to move from a binary, determinate,
linguistic-autonomous understanding of grammar, and especially syntax,
to something fuzzier and more complex.
Indeterminacy was a major bête noire for the Extended Standard
Theorists. As it appeared to me and still does, they were more concerned
with making a pure science of linguistics—removing it from its tawdry
humanistic origins and social-science companions—than with actually
looking at and accounting for the way real people actually use language,
often imprecisely and ambiguously. Some of my colleagues were obsessed
with fitting linguistics into a Kuhnian paradigm or achieving Popperian
falsifiability, since only thus (they argued) would linguistics be a True
Science. What they missed is that science has achieved its results and thus
its prestige by observing and generalizing carefully from natural data.
Transformational grammar, on the other hand, derived significantly from
philosophy of language, a field in which the data were produced in the mind
of the investigator, then subjected to his (just imaginably her) own analy-
sis. So it was more important to produce a formal rule, process, or theory
than to capture what happens when people speak. That focus made it easy
A fterword [ 375 ]
376
[ 376 ] Context Counts
377
A fterword [ 377 ]
378
5.
There is no sharp distinction between the first two strands, but merely
a shift in focus. If you join a movement advocating the incorporation of
psychosocial information into syntactic description, then you must believe
that the correct and complete identification of those psychosocial forces,
as they are represented in linguistic form, is of the utmost importance. So
even as the GS-ST (Standard Theory) wars abated and each side retreated
to its own corner, the former combatants on the GS side took their marbles
and went home: each of us returned from the battle to concentrate on the
particular aspect of the GS project that had drawn us into linguistics, and
onto the GS side, in the first place. My initial interest was psychological,
and so that occupies a large place both in the papers of strand one, and in
the first post-GS writings here. But I realized at the same time that no man
is an intrapsychic island: we create our internal selves, our belief systems,
our sense of who we are, in part on the basis of what we get from outside,
how others communicate with us and what our communicative possibili-
ties are. At the same time, these social aspects of constructed identity arise
partially out of our internal, psychic representations of the world, others,
and ourselves. And many of these considerations have explicit linguistic
representation. By recognizing these undercurrents we can carry out one
of the original missions of transformational grammar: to use language as
a window to the mind, where the “mind” is both the internal and external
selves.
Some aspects of this linguistically made selfhood are more readily con-
fined to the intrapsychic mind, others more extrinsically determined. In the
first set I would locate presupposition (both “semantic” and “pragmatic,”
if that distinction is useful or possible). In the second, I would place the
analysis of directness and politeness. But as these papers show, that neat
division fails almost at once: presupposition shapes our choice of direct or
indirect form; the decision about whether and how to speak politely to an
interlocutor is based on our assessment (“presupposition”) of the relation-
ship between us and how our culture expects us to treat each other. So
strands two and three also tend to coalesce.
This difficulty I am having in keeping these papers distinct reflects my
belief that the world and its categories are fuzzy rather than sharply deter-
minate. As I noted, later GS took explicit cognizance of indeterminacy,
diverging from the TG belief in distinct, often binary categories of many
[ 378 ] Context Counts
379
A fterword [ 379 ]
380
[ 380 ] Context Counts
381
disabled women, a group with two strikes against their achieving social
equity. This paper links strands two and three, with reference to how
extrinsic traits (gender, abledness) affect one’s communicative options and
the meanings one makes; and strand four as well, in its concern with the
making of public meaning by individuals and institutions.
6.
Language creates and is created by the self and at the same time creates
and is created by community. Even as humans are cognitive creatures (a
focal point of strand two), we are at the same time and equally importantly
social creatures, and each of these identities permits and potentiates the
other. So while strand 3 represents a change of emphasis from strand two,
it moves in many of the same directions. That connection makes sense
since often, the same linguistic devices work either as intrapsychic/cogni-
tive or as interpersonal/social, or even both at once.
While c hapters 1 and 2 were listed in the first strand, the arguments
used in the cause of theoretical persuasion were based in part on the social
functions of language, for instance the use of forms like honorifics (which
symbolize the relationships between participants) and politeness as a rea-
son for indirectness. Chapter 4 considers the gender of participants as a
significant factor in the choice of forms and their understanding. Chapter 5
also considers politeness as a function of grammar, in this case both syn-
chronically and diachronically, as well as male versus female strategies
and cross-cultural differences. In this one can see changes from earlier,
GS-oriented work like chapter 2: politeness is seen here as much as a lin-
guistically mediated social phenomenon as a social relationship with syn-
tactic consequences. Chapter 6 considers a particular form of interpersonal
communication, psychotherapy, looking at the therapeutic interchange as
a special sort of discourse in which participants engage for explicit pur-
poses; the peculiar forms of therapeutic discourse arise out of and facilitate
the functions the discourse is meant to serve. Chapter 7, like chapters 5
and 6 discussed above, is concerned in part with the social and interactive
function of literacy, versus orality. The choice of the literate channel makes
statements about the nature of the communication and the relationship
among the participants, and the currently shifting valuation of literacy will
bring changes in society.
A fterword [ 381 ]
382
7.
Institutions are created by individuals and small groups who unite into
large groups and so gain authority, and which are held together internally
by conventional rules and externally by the larger society’s belief in the
institution’s value. Institutional language is particularly interesting: it
must work for a large, often culturally diverse group; it must serve at once
as an internal “secret handshake,” and an external “keep off the grass” sign;
it must also work as shorthand, to express the cognitive concerns of the
institution in an economical way. It is often defensive, but as often vainglo-
rious. In short, it is an embodiment of all that makes us human, and plays
on all of the strands in these papers.
[ 382 ] Context Counts
383
Hence the papers in this group have almost all been discussed under
other subheads, and need not be mentioned specifically here. One excep-
tion, chapter 8, “Doubletalk: Sexism in tech talk,” on technology and its
language, is an attempt to confront a contemporary issue fraught with
problems. When a particular form of language is connected with an espe-
cially prestigious institution, and when, further, that form of language and
that institution are, for whatever reasons, open only to certain groups in a
society, how does that affect social equity? In the case of technology, what
is the relation between the impenetrability of its discourse style and the
tendency for women to feel unwelcome? Should (can) inaccessible language
be democratized by fiat, bringing with it real-world equity? Or does lan-
guage follow reality, so that once enough women (or other disenfranchised
minorities) enter a field, the language will change with them? Or will they
merely be corrupted by internalizing the language, and with it the mores,
of the institution, turning into honorary white males?
Chapter 14, “The rhetoric of reproduction,” is an investigation of the
rhetoric of the public sphere, concerning policy on reproductive rights. The
major question here is: What is appropriate public rhetoric? In a hotly con-
tested struggle, what are the rhetorical responsibilities of each side? Is it
possible to distinguish legitimate from illegitimate persuasion?
8.
And…
I have, over the quarter-century represented by these papers, moved
from being a relatively conventional ivory- tower linguist, concerned
mainly to establish my own theory of language and linguistics, to a more
public stance, a concern with language as a vector of politics and power,
which can be used for good or ill. Many of my colleagues view this stance
with dismay, believing that a scholar must be disengaged and apolitical.
That may be an option, or even a desideratum, in some fields. But a
thinking member of a democratic society has to have feelings about the
right way for her society to go. I don’t think those feelings can or should
be compartmentalized into someone’s “political component,” brought out
only the first week in November. We are—besides being social and cogni-
tive creatures—political animals, instinctively interested in power and its
distribution, and unless we are candidates for sainthood, eager to get our
fair share of what’s going around. It would be peculiar if our work could be
insulated from those concerns, and if it even could, it would become unat-
tractively passionless.
A fterword [ 383 ]
384
It’s much better, if we’re doing work that might have any foreseeable
political consequences, to state our own biases at the outset. It is the rare
linguist these days who can rightly believe his or her work resides exclu-
sively in the ivory tower. Current political and social debates over English-
Only legislation and “Ebonics” support that argument. There are two ways
to reconcile one’s academic and political selves. One, exemplified by the dual
oeuvres of at least one eminent linguist, is the path of multiple personal-
ity: one persona publishing inaccessible linguistic theory, the other writing
impassioned political diatribes with very little direct analysis of language
per se. But if my training enables me to understand language in depth,
I should apply my skills to the analysis of language and its sociopolitical
functions. I don’t find it easy or intuitive to split myself into scholarly and
political selves, or intellectual and emotional beings. I also think—despite
a lot of conservative criticism of this position—that emotional energy does
not damage intellectual analysis, when the two are applied together to the
understanding of linguistic problems; rather, both work together to reveal
how language works, since language itself is both cognitive and emotional.
You can’t use pure logic to understand passion, and passion underlies many
of our most crucial linguistic debates and choices.
As these papers suggest, linguistics is my field because it is a universal
donor. Most of what we do as humans is done via the medium of language.
The understanding of language therefore illuminates other aspects of being
human, perhaps more than any other field. Only linguists have a way of
making sense of the human proclivity for using language to construct our-
selves, our cultures, and our societies—a task as absorbing as it is unnerv-
ing, as endless as it is immediate.
REFERENCES
[ 384 ] Context Counts
╇ 385
INDEX
[ 386 ] Index
387
Index [ 387 ]
388
[ 388 ] Index
389
Index [ 389 ]
390
[ 390 ] Index
391
Index [ 391 ]
392
[ 392 ] Index
393
Index [ 393 ]
394
[ 394 ] Index
395
Index [ 395 ]
396
[ 396 ] Index
397
Index [ 397 ]
398
[ 398 ] Index
399
Index [ 399 ]
400
[ 400 ] Index
401
Index [ 401 ]
402
[ 402 ] Index
403
Index [ 403 ]
404
[ 404 ] Index
405
Index [ 405 ]
406
[ 406 ] Index
407
Index [ 407 ]
408
[ 408 ] Index
409
410
411
412