Chapter 9: Dr. James Reports
9
The Psychology
of Ordinary Language Use
The question of when a specialized area of study is, is not, or is
not yet a 'scientific' discipline appears to us an interesting one.
Whatever else science is, one of its important characteristics is that of
being a social enterprise and, as such, it becomes a matter of
definition, of social agreement or sanction, as to whether a particular
activity or body of knowledge is to be considered a science, and its
self-appointed experts or "mandarins" (Chomsky, 1969) as
'scientists' .
Parapsychology has associated with it all the trappings of a social
or behavioral scienceÑstatistics, control conditions, the Null
Hypothesis, Journals, and Chairmen of Departments, but the
mandarins of the other sciences have so far successfully neutralized
their claim to science-hood. Psychologists and sociologists refuse to
acknowledge the science-hood claims of 'political science' (although,
one must admit that the choice of the title was a nice exercise in one
upmanship), but they in turn are looked upon with kind-hearted
paternalism and tongue-in-cheek non-recognition by chemists and
engineers.
The Claim to Science-Hood
The base for a claim to science-hood has always been that of being
engaged in the study of 'natural' (real) phenomena by means of
systematic discovery procedures, in conjunction with the claim that
the body of knowledge thus accumulated will afford a considerable
amount of predictability and control, these being socially useful
skills, and it is here that the problem assumes its interesting
characteristic. Although we believe it is possible to present a general
argument in this connection, we shall restrict the following
discussion to the 'language sciences,' a phrase which has come to
include a number of currently flourishing scientific disciplines which
include psycholinguistics, sociolinguistics, ethnolinguistics, as well as
psychology, linguistics, anthropology, and sociology proper.
The interesting problem, as we see it, is this: Why and how does it
come about that successful claims can be made in the sociopolitical
sphere for the naturalness or real-ness of phenomena, and their
predictability and control, when these alleged phenomena are neither
natural nor real, and furthermore, when the claimed predictability
and control are not only never achieved but could not possibly be
achieved. In point of fact, the analysis of this interesting problem
will show that it is precisely because the alleged phenomena in the
language sciences are not natural and not predictable that it is
possible to lay successful claims as to their naturalness and
predictability. We were lead to this kind of a shockingly perverse
conclusion through our readings in 'ethnomethodology'Ñthe latest of
disciplines to claim science-hood (see Garfinkel, 1968; Sacks, 1971).
Let us proceed while we ask your indulgence and patience for a little
while at least.
Lewis Carroll's Alice in Wonderland and Milne's Winnie the Pooh
aren't usually on the required reading list in psycholinguistic courses
and probably never will be, and we, at least, consider this most
unenlightened, for here are two masterpieces in psychological insight
and, incidentally, excellent treatises on ethnomethodology. The trick
is to realize that the wonderland Alice has the good misfortune to
fall into is in no ways different from the wonderland she fell from.
"A rose by any other name would smell as sweet" is simultaneously
true and false. It is true because a "rose" and a "meigwei" smell and
look equally sweet, this being a consequence of a cross-cultural
coincidence whereby speakers of the English and Chinese tongues
assign a similar conceptual and affective status to what botanists have
come to refer to as "Genus Rosa." It is false wherever such a happy
coincidence does not occur, as is well known by butchers and
managers of supermarkets where "hocks" is considerably more
marketable than "pigs' feet," and by furriers who sell a coat made of
Zorina rather than South American skunk, and it's no longer good
practice to publicly refer to sanitary engineers as garbage collectors.
The rose dictum is flatly contradicted by the doctrine that has
established the magical power of euphemisms throughout man's
social history, whereby if something is unacceptable by one name, it
becomes acceptable by another name, and vice versa.
Ordinarily, for euphemisms to be euphemisms, it is necessary to
deny that they are euphemisms. Thus, it is the very denial of the
thing that establishes the realness of it. Here we have another one of
those perverse conclusions that we started out with, and thereby,
another instance of the interesting aspect of the problem that
interests us: why and how does it come about that the very denial of
something is precisely what establishes that thing as real?
Speech Acts as Institutional Facts
Ordinary language philosophers have examined this problem in
ways that we find instructive (see Austin, 1965; Searle, 1969;
Vendler, 1967). They have drawn attention to acts within ordinary
language useÑspeech acts that are 'illocutionary' in nature and these
differ from non-illocutionary acts in specific ways whose analysis
throws some light on the paradoxical conclusions of the type we are
concerned with here. Take for instance 'promising' as such a speech
act. Although there are socially appropriate ways of making
promises, such as the prefacing of an utterance by "I promise that
. . ., " the form of the expression of promising is not by itself what
constitutes a promise. For one, one can make a promise without such
a prefacing statement as for instance in "I will be there at six, sharp";
for another, one can preface a statement in that way without thereby
establishing the utterance as a promise, as for instance in "I promise
that it will rain tomorrow" or as in the threat, "I promise that I will
kill you if you hurt him." Thus for an utterance to be a successful
rather than a defective promise, it must have certain criterial features
which, as Searle (1969, p. 57 ff) points out include, among others,
the following: that the topic of the promising act be an act whose
perpetrator must be the individual doing the promising act (cf. "I
promise that he will do it" which is defective, vs. "I promise to see to
it that he will do it") that the topical act in question be in the future,
not the past or the present; that the listener to whom the promising
act is addressed be knowledgeable about acts of promising; that the
speaker believe that the listener would prefer his doing the topical
act rather than his not doing it (this requirement distinguishes
promises from other statements that are prefaced by "I promise" to
indicate obligation or emphatic assertion, as in "I didn't do it, I
promise you I didn't"); that the topical act of the promise not be an
act that the speaker would do in the normal course of eventsÑas
Searle (p. 59) points out, "a happily married man who promises his
wife he will not desert her in the next week is likely to provide more
anxiety than comfort"; that the speaker claim that he intends to do
it; and, that he claim that he intends that the promising act will place
him under the obligation to do it. Thus, the successful performance
of the illocutionary act of promising necessitates certain specificable
conditions related to the speaker, the speaker's claimed intentions,
the hearer, the hearer's alleged interests, and the knowledge of the
rules of promising on the part of the speaker and the listener.
Now, if we examine transactions that socialized human beings
execute, we find non-illocutionary ones such as staring,
kicking, crying, making love, blushing, eating and so on, and these
differ in a crucial respect from the illocutionary ones, as follows:
that the criterial features that establish non-illocutionary acts are
objectively and operationally definable in terms of a system of
description that is independent of the very act whose status or reality
we are trying to establish. The denial by the author of an
illocutionary act that he has performed the act, is tantamount to
establishing its non-existence, whereas this is not the case for
nonillocutionary acts. Thus, if a person denies that he has made or
broken a promise, it will do you no good to present evidence such as
a witness or a tape recording, since then he can claim that he did not
intend to make a promise, that circumstances forced him to act as if
he was promising, but that in fact he did not promise. You may then
call him a liar or a hypocrite, but you cannot reasonably claim that
he broke his promise. The breaking of a promise is thus in itself
another illocutionary act, and such an act can come into existence
only with the cooperation of the originator of the alleged promise, as
when someone admits to having broken a promise. There are of
course social and legal sanctions one can impose on people who lie,
but there is no way of establishing the making or the breaking of a
promise without the consent of the alleged perpetrator. This is not
true of non-illocutionary acts, and indeed if a person is presented
with a film in which he is seen kicking someone, there is no
reasonable way in which he can deny the fact that he has kicked the
person in question, even though he may present alleviating
rationalizations for having kicked someone, such as ("It was an
accident" or "I didn't mean it" or "It was his fault" or whatever.)
Now we wish to show that the distinction between illocutionary
and non-illocutionary acts just considered is an instance of a
fundamental and general distinction that is just the distinction that
we need to resolve paradoxes of the kind we mentioned above,
namely, first, that the establishment of the language sciences rests on
assertions about the realness, predictability, and control of
phenomena that can't possibly have these characteristics and, second,
that the establishment of euphemisms rests on assertions about their
denial. Let us refer to a distinction Searle (1969, p. 33 ff) makes
between systems based on "constitutive rules (like the illocutionary
acts or the game of football) versus systems based on "regulative"
rules (like the prohibition against murder or the laws of planetary
motion). We shall eventually try to show that while many basic
phenomena involved in language use represent institutional facts that
belong to both regulative and constitutive systems, the traditional as
well as current scientific concepts in the language sciences treat them
as brute facts belonging to regulative systems only, and that
therefore no adequate, productive, non-trivial accounts of language
use phenomena are possible with these concepts.
Assuming for the moment that we can show what we claim we
shall show, it is pertinent to ask what is it then that the language
sciences are actually about? It seems to us that the language sciences
actually deal with a hypothetical type of reification of what human
language use phenomena would be like if they were of the type
posited by these sciences (i.e., brute facts). In other words, if there
were such a thing as a human language that is based on a system of
natural regulative (as opposed to institutional) rules, then the
language sciences now in vogue might possibly represent an adequate
descriptive and discovery procedure for their study. Thus, in answer
to the question of what it is that the language sciences are about, we
would say that they are about non-existent and hypothetical
phenomena that are proper to a non-existent hypothetical human
language, that is, an artificial language rather than a natural one. The
substitution of an artificial language as an object of scientific study
in lieu of the natural one while simultaneously denying that that is
precisely what is being done, is a process that becomes under-
standable when we view the enterprise of the language sciences
as an instance of social activities that humans engage in within a
constitutive system of rules, just as we have seen is the case with
illocutionary acts like promising. In other words, the body of
knowledge accumulated under the language sciences constitutes a set
of statements that are illocutionary in character. Therefore it does us
no good for advancing our argument to point out to the mandarins
of the language sciences that what they are doing does not constitute
a description of natural language phenomena, since they can claim
that they never intended to study anything but what it is that they
were studying, just as the person confronted with evidence of having
made a promise can claim that he never intended to make a promise,
and therefore no promise was made.
There are numerous examples of this sort in the history of
psychology and we want to mention two that are well known. One is
the counter claim to criticisms of the notion of 'intelligence'
whereby psychometricians, in the face of evidence that what they
had claimed intelligence was, did not appear to be that, then claimed
that what they meant by intelligence was that which their
intelligence tests measured, nothing more, nothing less. The other
instance is the counter claim by verbal learning theorists to criticisms
that their theories do not account for classroom learning whereby
they then asserted that their laboratory experiments on rote
memorization were not intended to be anything but experiments on
rote memorization.
In the relatively short history of psycholinguistics, hundreds of
experiments have been conducted on language phenomena such as
the grammatical transformation of sentences, the speed and accuracy
with which sentences were decoded and encoded, the efficiency with
which specific types of information was being transmitted and
received, the ease with which sentences of a particular type were
recalled after given amounts and conditions of exposure, and the
like. In the past few years an increasing number of critics have
consistently pointed out the irrelevance of these phenomena to such
everyday commonplace events as conversations. These criticisms,
whatever their merit may be, can have no force since the "new
scientists of language" as Miller ( 1964) has called these
psycholinguist mandarins, can always claim that psycholinguistics
consists of the body of knowledge adduced by these experiments and
therefore one has no reasonable claims against the Pope for doing
that which the Pope is sanctioned to do.
The Practice of Ordinary Science
Kuhn's revolutionary thesis in his The Structure of Scientific
Revolutions was based on a survey of paradigm switches in the
"natural sciences," mostly-physics and chemistry, but his greatest
influence may yet be not in those sciences, but rather in the social
sciences.
According to Kuhn "Normal science, the activity in which most
scientist inevitably spend almost all their time, is predicated on the
assumption that the scientific community knows what the world is
like. Much of the success of the enterprise derives from the
community's willingness to defend that assumption, if necessary at
considerable cost" ( 1962, p. 5 ). When anomalies occur, an
extraordinary set of investigations begin which lead to a new set of
commitments, "a new basis for the practice of science," a "scientific
revolution." "They are the tradition-shattering complements to the
tradition-bound activity of a normal science" (p. 6). A shift occurs in
"the standards by which the profession determine(s) what should
count as an admissible problem or as a legitimate problem-solution
(p. 6). "For these men (the established mandarins) the new theory
implies a change in the rules governing the prior practice of normal
science. Inevitably, therefore, it reflects upon much scientific work
they have already successfully completed" (p. 7).
We are particularly interested in Kuhn's use of the concept of
"normal science" in its ethnomethodological sense, even though
Kuhn himself does not explicitly discuss it from this perspective. The
meaning of "normal science" that we see being of great interest is
this: the behavior of scientists when they are being ordinary. We
don't mean "ordinary" in the usual statistical sense of the mode or
the most frequently occurring pattern of behavior. We mean it in the
sense of the expected norm or the appropriate pattem of behavior
under prescribed conditions. For instance, the way an academic
psychologist at all of the well known universities can be ordinary,
rather than nonordinary or special, is to conduct research and to
publish in journals that have a reputation for high scientific standards
in editorial policy. This is one of the things an academic psychologist
is expected to do, and in doing so, he is being ordinary qua an
academic psychologist. Knowing how to be ordinary in a particular
social endeavor requires the acquisition of certain specialized
competencies that are defined with reference to the behaviors that
are expected and appropriate in those particular circumstances. Thus,
we can view the function of graduate training in academic
psychology as being that of facilitating the acquisition of a set of
skills that will enable the graduate to later be ordinary as an
academic psychologist.
One of the required rules for behaving ordinarily or being ordinary
in one's practice of academic psychology has to do with following
certain prescribed methodological procedures in research and theory
building. In the field of psychology, the nature of acceptable
methodological procedures have changed over historical epochs, so
that what it was to be ordinary in Wundt's or in Titchener's time
does not constitute the same pattem of behavior (in fact, they are
contradictory) as the pattem of behavior that is associated with
being ordinary in Watson's or our own time. When differences in
what is being ordinary in two epochs involve the nature of the
methodological procedures that generate data, we have reached an
instance of what Kuhn refers to as a "paradigm switch." It is a fact in
the sociology of science that the innovators of a new paradigm
believe it to be superior to the old paradigm. For instance, American
Behaviorism believed itself to be superior to Introspectionism.
Similarly, generative transformational linguistics is believed to be by
their proponents a significant advance over American structuralist
linguistics. It is important to realize that the alleged superiority
attributed to the new paradigm appears to be a self-evident truth to
the innovators. In other words, the adoption of a new paradigm does
not represent an "objective" decision on the part of impartial judges.
The character of the arguments involved in this kind of a decision is
that of an oriented to operative, by which we mean, along with the
ethnomethodologists (see below), that type of situation in which the
innovators reject the features which the old paradigmers are oriented
to by prior agreement and training, and the new features they pick to
be oriented to are just those features which the old paradigmers are
not oriented to. For instance, in the recent revolution in linguistics,
Chomsky and the new paradigmers chose as a central feature of
language to be oriented to the underlying deep structure of sentences
and to deliberately reject the fundamental significance as a source of
input to theory of the overtly manifested surface features of
language, precisely those that formed the foundation of structuralist
linguistics. It was not the case that structuralist linguistics could
either be the foundation upon which transformational linguistics
could build upon, or that.its findings and theories could be built
upon cumulatively with the refinement of newer and stronger
methodologies. If the facts of fifty years of structuralist linguistics
were suddenly wiped out, current generative transformational
linguistics would not be the worse off in terms of its development.
This is the argument of non-cumulative increments of scientific
revolutions that Kuhn has tried to develop for physics and chemistry,
and it appears to be true in linguistics as well, and we would offer the
generalization that it is probably true in all other scientific
disciplines.
By proposing a paradigm switch in psycholinguistics, we are
engaging in persuasive argumentation rather than in what would be
considered in the old paradigm as "objective" ratiocination. This is
necessarily so inasmuch as we are "advocating" that we become
oriented to a new set of features, these features being in some crucial
instances contradictory to the existing set of features. Therefore, we
cannot present "scientific" arguments within the present paradigm in
favor of the new paradigm. Kuhn tries to show that the direction
new paradigms take is influenced not solely by rational or logical
considerations but by social psychological, religious and philosophi-
cal ones as well. Part of being ordinary as a scientist is to resist
changes of a certain sort, the sort that question fundamental
assumptions, this resistance going under the vigilante activity of
"maintaining high scientific standards," We would like to show how
this resistance process in being ordinary as a scientist operates in
psycholinguistics.
To do this, we would like initially to establish the concept of
"askable questions" and its converse of "non-askable questions."
Consider the following question: Does St. Peter admit to Heaven
clergymen who have supported the principle of killing in religious
wars? Cabell answers this question in the negative in the
mythological fiction of Jurgen, but there is no way in which this
question, or its equivalent, can be asked in the language of
psychology. As soon as you think about it you realize that this is not
a special or abnormal case, and that there must be an infinity of
questions of this sort; in short, there is a set of questions in a
scientific discipline that are "not-askable" and a set that are
"askable."
Consider, next, a definition we propose for a new field to be called
''ethno-methodology.''l Ethnomethodology is the study of the
discovery procedures used by scientists (the latter term is to be given
a special definition in what ensues). By "discovery procedures" we
mean the particular methodological steps (data collection and status
of concepts) to be followed in being ordinary as a researcher in a
particular field or sub-field of specialization. Thus, "types of
methods" is to "ethnomethodology" what "types of cultures" is to
"ethnolinguistics~' (or "anthropology"). Furthermore, ethnomethod-
ology restricts its perspective to "operatives," which is to say it will
define all problems as "interactional processes," so that "scientific
methods" becomes "the procedures carried out by various scientists
in their activities qua scientists . " These differences in what
constitutes ordinary behavior in particular socially defined
circumstances (e.g., scientific research) are ascribable to the rules of
the ethnic subculture represented by a particular paradigm in a field
of specialization. Ethnomethodology thus becomes the study of the
subcultural rules of discovery of a particular group of paradigmers.
Thus far, we expect no special opposition from the psycholinguists.
They would no doubt say "Well, if that's what you want to do, go
ahead." and might add: "And, anyway, isn't that what the history or
philosophy of science is supposed to be doing already." (The latter
question we must answer in the negative, but we shall not elaborate
at this point.)
Now, as we go on to elaborate our definition of ethno-
methodology, in particular the term "scientist," we shall be stating
what to us seems readily obvious, even self-evident, yet it will be
non-evident and inacceptable to psycholinguists in the old paradigm.
Ethnomethodology studies the practices that constitute being
ordinary as a scientist. Its interest in the scientist does not arise out
of a special interest in the scientist, as traditionally identified, but
in the fact that scientists, when behaving ordinarily, claim to subscribe
to a set of explicitly stated discovery procedures. Both because his
interest is not exclusive in any one social group, and because a
comparative study of discovery procedures of other social groups
sheds light on the scientists as well, the ethnomethodologist is
committed to the general study of discovery procedures of other
groups of individuals (including, of course, their own). But the
concept of "social group" represents an outcome of and derives its
meaning from the methodological procedures of existing paradigms
in sociology and psychology. Therefore, the concept of "social
group" remains an object of ethnomethodological study and is not a
concept that can be directly incorporated as a working concept in it.
So, in fact, ethnomethodology does not study the practices of any
particular group of individuals, be they scientists or whatever. It
studies the discovery practices of individuals in their social
interactive setting. The only meaning "scientist" can have in the
definition of ethnomethodology is that of an individual engaging in
systematic discovery practices. The task of ethnomethodology is to
show that the discovery practices of any individual when being
ordinary are systematic and rule governed. Now, to get back to the
psycholinguist in the current paradigm.
We have already pointed out that many questions, such as St.
Peter's admission practices at the gates of Heaven, fall in the category
of non-askable questions in the old paradigm. This leads to a proper
and interesting question in ethnomethodology, an askable question
as follows: How does it come about that in the language and
conception of current psychology, the St. Peter question falls in the
non-askable category? To study this question the ethnomethod-
ologist will compare the practices of two "groups" of individuals:
those for whom the St. Peter question falls in the non-askable
category (this will include academic psychologists) and those for
whom it falls in the askable category (this will include theologians,
classical scholars, mythologists, many ordinary people, and the like).
Ethnomethodology is the study of the discovery practices of
individuals when they are being ordinary. What is here the proper
sense of "discovery practices?" Among all the practices that
individuals engage in which one contains the sub-set of "discovery
practices?" Once again we are not permitted to give a traditional
definition since that is precisely what forms the object of study, viz.
how does it come about that individuals classify different items of
practice into the category of "discovery practices?" The answer to
this question, and other questions of this sort, must lie in an
elaboration of what constitutes "being ordinary" in various
interactional circumstances. Our original definition has now been
transformed to read as follows: Ethnomethodology is the study of
the transactional practices of individuals when they are being
ordinary. Our subsequent discussion will clarify the following
additional points that can be made about this definition: that
"transactional" refers to joint cooccurrent interaction that stems
from coordinated work of individuals who are oriented to the same
features of the interaction; that "practices" refers to operatives, viz.
that it is the doing of the activity that constitutes the activity; and
that "being ordinary" refers to the set of operatives whose oriented
to features are jointly defined by interactants as ordinary, not
special.
To summarize these points we shall attempt one final definition as
follows: Ethnomethodology is the study of those joint cooccurrent
operatives which participants in a transaction are oriented to in
circumstances they describe as ordinary. Now there remain three
points for us to argue in order to show the inadequacy of the old
paradigm in psycholinguistics. These are as follows: (I) that the set
of operatives thus defined is not made up of unusual or special cases,
but that they constitute a large and significant body of the social
practices of individuals in conversational interaction; (2) that the
study of this important set of verbal interactions falls in the
nonaskable category in the old paradigm; and (3) that it falls in the
category of askable questions in the new paradigm. To accomplish
this task we shall have to consider first what it is that
ethnomethodologists mean when they talk about "ordinary
commonplace conversational interaction." Because the exposition of
this concept is fairly long, we would like to reassure the reader that
this is not an irrelevant digression and we promise to return to these
three points in due course.
Some Features of Ordinary Talk
We begin with the commonplace observation that ordinary talk is
partly made up of talk about talk. For instance, you can say "I)id
you see George, yesterday?" and some utterances later you can say
"Why aren't you answering my question?" The latter utterance is no
less a part of ordinary talk, yet in relation to the first utterance, it
is an utterance about an utterance, talk about talk or metatalk. That
metatalk is part of ordinary talk appears to be a fundamental aspect
of human conversational interaction that allows its present
constitutive character. In other words, we are claiming that this is a
nontrivial feature of ordinary talk whose absence would funda-
mentally alter the character of ordinary conversations and would
become immediately noticeable to participants leading to their
qualification of talk without the possibility of metatalk as
nonordinary or abnormal.
All illocutionary acts belong to the metatalk set of the full set of
ordinary talk. In general, acts that belong to the class of metatalk are
just those acts that are mentioned when in ordinary talk someone
asks "What is he doing?" by which they intend to mean "What is he
doing in his talk and through it?", such as: promising,
complimenting, lambasting, flattering, joking, being smart-Alecky,
polite, obtuse, fastidious, imaginitive, sweet, supportive, cold,
immature, opinionated, incomprehensible, thankful, electrifying,
vulgar, and so on. This is, as you no doubt realize, an open set, and it
may be an interesting problem to look into as to why and how it
comes about that it is an open set.
To inquire into this problem it might be useful to recall certain
distinctions in types of speech acts which Austin (1965) and Searle
(1969) among others have explicated, in particular the difference
between illocutionary, and perlocutionary acts. An'illocutionary'
act, as we have discussed, involves such things as promising,
criticizing, requesting, approving, apologizing, objecting, comment-
ing, questioning, and so on. A 'perlocutionary' act involves having an
effect on the intended listener such as persuading, inspiring,
surprising, misleading, confusing, alarming and so on. It seems to us
that there exists another set of locutionary acts that are different
from propositional, illocutionary, or perlocutionary acts, a set that is
exemplified by such qualifiers as being polite, pompous, immature,
opinionated, supportive, warm, cold, sweet, understanding, creative
and so on. In ordinary talk we sometimes use these terms to refer to
'personality traits' although we do not necessarily insist that they be
permanent or even characteristic. For instance, we can be polite in
one conversation and impolite in another, pompous and opinionated
in one context, and supportive and understanding in another. What
the members of this set seem to have in common is the attribution of
a descriptive trait to something the speaker is doing in a specific
conversational circumstance or episode, but unlike so-called
personality traits, the description refers to the conversationalist's
work qua conversationalist in a particular conversational interaction.
Note that it is not the traits of the speaker that are being qualified,
but rather the speaker's conversational acts in the particular
interaction. This, we think, is precisely the difference that exists
between "He is so-and-so" and "He is being so-and-so," as for
instance, "George is opinionated" (i.e., generally and characteris-
tically), versus "George is being opinionated" (i.e., now, as he is
acting in this part of the conversation). We propose to call these
locutionary acts, translocutionary acts (for locutionary plus
transactional), to distinguish them from propositional, illocutionary,
and perlocutionary acts.
Translocutionary as well as perlocutionary acts differ from
illocutionary acts in the criteria that must hold for their successful
performance. As we have seen, illocutionary acts cannot reasonably
be defined independently of the speaker's definition of what it is
that he is doing. A speaker cannot properly promise without claiming
that he intends to promise, but he can be opinionated or offensive
without claiming that he intends to be so. The reason for this
difference is that participants treat translocutionary and perlocution-
ary acts as acts that are to be evaluated by a regulative system of
rules. These rules are stated in terms that are independent of the
speaker's claimed intentions.
Searle ( 1969) claims that all institutional facts belong to a
constitutive system, and that all facts belonging to regulative systems
are brute facts. We think this is an oversimplification that eliminates
certain distinctions that are useful to make. For instance, we would
say that both illocutionary and translocutionary acts are institutional
facts, but the latter belong to a regulative system while the former
belong to a constitutive system. In other words, we wish to
distinguish between two sorts of facts that yet belong to regulative
systems: brute facts such as kicking or caressing someone and
institutional facts such as being opinionated or impolite. This
distinction is necessary for specifying the nature of the observation
procedures that are capable of establishing the existence of
institutional (regulative) -facts and brute (regulative) facts. An
independent observer armed with a camera can establish the brute
fact that one native aborigine kicked another member of the tribe,
but never could he establish the institutional fact that a native was
being opinionated or offensive with nothing but a camera or tape
recorder. Facts of this sort have no meaning or existence
independently of the system of (regulative) rules that governs
interpersonal interaction among the tribesmen. A psychologist or
anthropologist armed with a universal descriptive system of behavior
could never discover a single institutional fact, whether regulative or
constitutive. The reason for this is that institutional facts are all
"operatives," which is to say that it is the doing of the activity that
establishes them as instances of that activity, and this establishment
procedure may be done in two ways: by appeal to the joint
cooccurrent oriented to features of the ongoing activity (constitutive
systems), or by appeal to a standard of reference that is usually
practiced in situations of the sort that the present activity is claimed
as being an instance of (regulative systems). These relationships are
illustrated in Fig. 1.
To summarize what we have said thus far, ordinary talk is made up
of locutionary acts of various types that are identified by
participants in answer to the question "What is he doing in his talk
and through it?" Now, a second question that can ordinarily be
asked about talk is "What is he talking about?" and the answer to
this question ordinarily refers to what participants call the 'topic' of
the conversation or utterance. Topic refers to the set of things that
are talkable about and this set is defined in reference to the
conversational circumstance. Not any topic is appropriate within a
particular conversational circumstance. Thus, when we ask in any
particular conversational circumstance "What did he talk about?" we
specifically mean to exclude from the answer that set of topics that
are not talkable about in that circumstance. Should someone want to
talk about a topic that isn't talkable about, he must preface the
doing so by some statement that indicates to the participants that he
is not engaging in an ordinary conversation. For instance, he might
say "I know I am not supposed to talk about that but . . . " or "Hey,
Dad, I am doing an experiment for my Psych 100 course. What is
your name, please." In the absence of such a prefatory statement he
will not succeed in talking about what he wants to talk about; in that
case, the other participants would not say that he is talking about
something; instead, he would be sounding off, being smart,
complaining, insulting, going mad, talking for the sake of talking,
imitating, fooling around, being obnoxious, talking nonsense, or
whatever, but not talking about something.
Talking ordinarily restricts not only what you can talk about in a
particular conversational circumstance, but also the sequence of the
things you can talk about. For instance, you cannot transact an
ordinary conversation by performing the illocutionary acts of
greeting and leave-taking in the middle of the conversational episode
as opposed to the beginning and end, respectively. Should someone
attempt that, the conversation will either come to an abrupt end or
the act will be treated according to the conversational rules for
handling interruptions, in this case probably by a topic switch that
makes the interruption the next topic of conversation as in "What's
the matter with you, anyway" or "What are you trying to do. Have
you been feeling all right lately?" or whatever. Goffman (1961) has
made the observation that the reason inmates of an asylum strike us
as very odd is not so much due to the fact that they say or do very
odd things, but rather because they break the most ordinary rules of
talking and behaving publicly, and it is the doing of that that strikes
us as very odd.
While it is possible to talk about a number of different things in
any particular conversational circumstance, it is not possible to do so
without performing the illocutionary act of announcing a topic
switch every time you wish to change topics. Such acts may take the
form of "Oh, by the way, . . . " or "Incidentally . . . " or "That
reminds me of . . . " and so on in which the illocutionary act of
announcing a topic switch is accomplished by an utterance, or
altemately, by the vehicle of a lull in conversation. In the latter case,
it cannot be just any silence; it must be a silence that participants
recognize and identify as the type that constitutes an occasion for
topic switch and not some other kind of silence such as pondering an
answer, or refusing to participate in the conversation, or seeing the
effect of a perlocutionary act, or whatever.
It may be useful at this point to introduce some additional terms
and distinctions that are involved in the successful performance of an
ordinary conversation. A conversational circumstance is an
interactional episode that is made up of a structured sequence of
conversational events being transacted within the framework of an
identifiable background context. By 'interactional' episode we mean
to indicate not merely the trivial feature that more than one person
must be involved, but the nontrivial fact that it is cooccurrent
oriented to operative work (see Sacks, 1971) in which the listener
contributes to the joint interaction just as crucially, definitionally or
constitutively as the speaker in ways that shall become clear later. By
'identifiable background context' we mean to indicate the fact that it
matters critically who the participants are, where they are talking
and for what purpose. The nature of conversational events that make
up an episode needs to be examined. Basically, we are dealing with
silences and utterances in a structured sequence. Silences can be
distinguished by length and locus of occurrence (within and between
utterances). Utterances themselves vary in syntactic type, locution-
ary type, communicative function, and integrity level. By syntactic
type we mean to refer to their syntactic value within the structured
sequence that constitutes the episode. For instance, an interruption
is an utterance that prevents the completion of an ongoing utterance;
an opener marks the beginning of a conversational sequence, either at
the beginning of a conversational episode or during the introduction
of a topical switch; a repartee is an utterance that is appropriate
during a talker switch and serves to continue or maintain a
conversational episode; and so on. Locutionary type refers to the
type of locutionary act it is, namely propositional, illocutionary;
perlocutionary, or translocutionary, as discussed earlier. The
communicative function of an utterance relates to the purpose or
motive of the conversational interaction as defined by the set of
socially legitimate uses of talk: reporting an event, instructing,
problem solving, being sociable, describing something, expressing an
opinion, making a judgment, establishing one's status vis-a-vis
participants, laying claims for some special competence, and so on.
The integrity level of an utterance refers to its status vis-a-vis it
being
what it appears to be on the surface. There appear to be at least five
levels of talk in addition to "direct representational " talk, namely
"indirect representational" talk (as in making allusions or talking
around the subject when participants know what it is but agree
tacitly not to mention it directly); "misrepresentational" talk as in
propaganda, dissimulation and other manipulative situations);
"tactful" talk (in which, like in indirect representational talk, the
topic is not mentioned directly, but in this case, one or more of the
participants can claim that they are unaware of the topic);
"non-representational" talk (in which the expressed topic is
incidental to another underlying topic with which the participants
actually deal, knowingly or unknowingly, such as in beating around
the bush or, in instances where a participant wishes to change its
non-representational character would say "Come on, what is it really
that you want to tell me" or some such thing); and finally, "other
directed" talk (in which the effective audience is other than the
addressee, as in "loud-talking" (see Mitchell-Kerman, 1969, or in
statements to reporters made at press conferences).
The characteristics of utterances that we have elaborated thus far,
namely syntactic and locutionary type, communicative function, and
integrity level belong to considerations about the syntax of
conversations (for additional observations on the syntax of
conversations see the recent and stimulating, work of Harvey Sacks,
pg71). Now we wish to make certain comments on the semantics of
conversations.
The Meaning of an Utterance
When in ordinary conversations we wish to inquire about the
meaning of an utterance, we ~normally ask four sorts of questions:
"What did he say?", "Why did he say that?", "What did he mean?"
and, "What was he trying to do in saying that?" That these represent
different questions, rather than being paraphrases of one another is
evidenced by the fact that different kinds of answers are appropriate
to each of them, and an answer that is satisfactory to one question is
not satisfactory to another, as shown by the following illustration:
A1: What did he say?
B1: He said he is not satisfied with Helen.
A2: Why did he say that?
B2: He doesn't think she is suited for that kind of work.
A3: What did he mean?
B3: She doesn't show sufficient initiative.
A4: What was he trying to do?
B4: He wanted me to try to get her fired.
Note that B's answers are different for each of the four questions and
furthermore, they are not interchangeable: Bl is an indirect quote or
a paraphrase; B2 is a justification; B3 is an elaboration and B4 is an
explanation. Thus, the 'meaning' of an utterance in ordinary
conversation is to be given in terms of a paraphrase or quote, a
justification, an elaboration, or an explanation (see McCawley,
1971). Note that B's answers can be tied into one utterance as
follows:
"He said he is not satisfied with Helen because
she is not suited for this kind of work inasmuch
as she doesn't show sufficient initiative and therefore
he wants me to fire her."
Expressions like "He said that . . ., " "because," "inasmuch as," and
"therefore" are discourse hinges which underlie the structure of
utterances whether or not they are actualized in surface structure
(see Jakobovits, 1968). The semantics of utterances in ordinary
conversation must take into account the contextual structure in
which they occur no less than the meaning of a word in an utterance
must be derived from the underlying structure of the sentence. What
constitutes the contextual structure of an utterance in ordinary
conversation? Let us attempt a preliminary formulation to this basic
problem in the theory of communicative competence.
Consider once again the four questions that seem to be involved in
what people mean ordinarily when they inquire about the meaning
of a locutionary act:
1. What did he say?
2. Why did he say it?
3. What did he mean?
4. What was he trying to do (in his talking)?
Except in the case of a direct quote the answer to the first
question involves paraphrasing. Paraphrasing is a selectional and
creative process and, with the possible exception of the paraphrasing
of a single sentence, it involves inferences about what kind of
information the questioner may have an interest in at the moment.
The ability to make appropriate selections in answer to the question
"What did he say?" is part of the competence participants need to
have to engage in ordinary conversations. There are rules of
conversational interaction that tag topical information of conversa-
tions in some such terms as: reportable information, non-reportable
information, reportable to so-and-so and not reportable to so-and-so
or in such-and-such a place, and so on. Thus, if a participant is asked
"What did he say?" (e.g., "What did the President say at his press
conference yesterday?"), his answer will vary in significant and
nontrivial ways depending on the conversational circumstance in
which the request occurs: who the questioner is (child or adult,
friend or foe, peer or subordinate, etc.), what the social setting of the
request is (e.g., are there other listeners present, etc.), what the
interpersonal import is of the participants' reaction to his choice of
information to select, etc.). It is the non-trivial specification of these
various "etceteras" that needs to be done if we are interested in
describing the semantics of ordinary conversations.
The Inadequacies of the Old Paradigm
Earlier, we claimed that we shall argue the following three points:
(1) that joint cooccurrent oriented to operatives constitute a
significant core set of phenomena that cannot reasonably be
excluded from a theory of communicative competence; (2) that this
significant set is yet excluded from study in the old paradigm; and
(3) that it forms the central focus of the new paradigm. We believe
we have adequately dealt so far with the first and third questions.
What the ethnomethodologists call ordinary commonplace conversational
interaction cannot reasonably be excluded from study and
still claim an interest in communicative competence. Our discussion
on the nature of these phenomena, as studied from the perspective of
the new ethnomethodological paradigm, constitutes a partial
argument to point number two, as well, but it might be best to
elaborate further at the risk of being tedious.
The central issue has already been stated: the facts that constitute
the data for the study of locutionary phenomena, whether regulative
or constitutive, are institutional in character (i.e., oriented to
operatives). The facts that constitute the data for the study of
language use phenomena in the old paradigm are (regulative) brute
facts (i.e., independently observable). Facts that are oriented to
operatives and facts that are independently observable are worlds
apart and separated by an insurmountable chasm. The discovery of
oriented to features cannot be accomplished through objective
observations by an independent investigator. When we think of what
is needed in psycholinguistics today, we have in mind a reversal of
the choice made by "scientific" psychologists, one that is in favor of
substance and against methodology.2 We are using 'against' advisedly
because there is strong professional pressure to maintain the present
choice in favor of current conceptions of methodology. The pressure
takes the form of the maintenance of so-called 'high' scientific
standards sanctioned by the established mandarins: the journal
editors, the book reviewers, the research evaluators of granting
agencies, the senior personnel in academic departments, the officers
of professional organizations, the graduate curriculum in psychology.
The people who maintain these standards are people who helped
define and establish them, who are 'achievers' in them, and believe in
them. We are faced with the necessity of a 'paradigm switch' and in
such a situation, as Kuhn (1962) so convincingly documented for the
physical sciences, there is left no contact point between the 'old' and
the 'new' paradigms.
An analogous necessity has faced the field of linguistics at the
beginning of the Chomskian 'revolution' and it exists now in the field
of education as well. Theoretical and experimental psychology has
disqualified itself from the task of explicating the instructional
process, both on account of their practical failure in enabling us to
deal effectively with the current problems in education, as well as on
account of the disclaimers of some of their foremost spokesmen
(e.g., Postman, 1961). On the other hand, we do not believe that the
problems encountered in education, psychotherapy, human
communication, are beyond the reach of systematic study, but we
need to make them the focus of our concern, the departing point of
our study rather than the elusive and ideal goal of it.
It seems that we are in search of an explanatory theory; where do
we begin? It seems to us a reasonable strategy to begin at precisely
the point the 'old' paradigm tried to get away from, namely common
sense. We have in mind the 'commonness' feature of this strategic
goal rather than its 'common sensical-ness.' Humans strike us as
marvelously complex and powerful organisms when we focus on all
the things they can do 'ordinarily.' Think of the intricacies involved
in the successful completion of a conversational episode, and yet this
is a very ordinary, common competence most, if not all humans,
have. At the same time, the scientific competence needed to
successfully complete a trip to the moon is most certainly not
ordinary or common. Wherein lies the difference? Man is the creator
of constitutive systems and the institutional facts in them; he is a
master at it. But he is not the creator (in the same sense as above) of
the universe and the brute facts in it; the so-called physical world is a
regulative system, not constitutive, and man feels himself being
regulated by it. Here he is a child, not a master. And so, the flight to
the stars is one he has to take on the wings of those amongst us, the
Newtons and the Einsteins, who have very special, non-common,
competencies.
But here we are concerned with very common competencies, in
fact, the more ordinary they are the more they point up the superb
and admirable mastery which we possess. Should we then not trust
our common synthetic powers for creating constitutive systems just
because they are of little use in the analysis of regulative systems? We
believe one of the important reasons psychologists have come to
disdain popular wisdom is that the language of folk theories and
popular writers has remained inaccessible to the scientific language.
In the field of philosophy, for instance, many traditional
epistemological problems that have been debated by scholars for
centuries, have been resolved as soon as a few philosophers got the
unusual idea of looking for their solution in ordinary language, and
today, "ordinary language philosophy" is the most influential and
powerful development in the field of philosophy. One of the most
crucial innovations of transformational generative linguistics was to
begin using in systematic and non-trivial ways the linguistic
competence of the ordinary native speaker as opposed to relying
exclusively on the non-ordinary special competencies of the linguist.
It is to the competence of the ordinary conversationalist that we
entrust the beginning of our search for a theory of communicative
competence.
Our previous discussion on the analysis of the meaning of an
utterance has taken such a frame of reference. We have asked how
the participant in an ordinary conversation construes the problem of
utterance meaning and using our own ordinary conversational
competence we have adduced four queries that focus on that which
we might mean when we ask about what an utterance might mean.
These were: What did the speaker say? Why did he say it? What did
he mean? What was he trying to do through his talking?
We do not necessarily expect, nor do we think it important, that
should any one else consider this problem he would come up with
the same set of questions that we have adduced. The relevant
consideration here would be, not that the two sets are different, but
rather how they are different. Suppose, for instance, that someone,
perhaps an old paradigmer, would list the following queries when
considering the problem of the meaning of an utterance:
1. Was he wearing a hat when he said what he said?
2. Were the listener's hands in his pocket at the time the utterance was
made?
3. Did the utterance in question contain a word that was of low frequency
according to the Thorndike-Lorge count?
4. Was the syntactic form of the utterance in question left-branching,
right-branching, or something else?
It is conceivable that the meaning of an utterance may be affected by
conditions of the sort that this new set of questions refer to, but to
claim that they tap in a general way that which we mean when we
inquire about the meaning of an utterance sounds totally
implausable. Our rejection of this second set would thus not
constitute merely an arbitrary and idiosyncratic decision. We have
publicly statable and verifiable grounds for rejecting these questions
as a general paradigm for how people determine the meaning of an
utterance.
Suppose, as a third instance, that someone proposes the following
questions as a general solution:
1. When did he make the utterance in question?
2. Was the speaker in a normal state or did he seem disturbed?
3. Who were the other people present when he said it, if any?
4. Did he seem to be aware of the effect his utterance had on the
other participants?
Now this third set is also different from the first set we suggested but
this difference can be seen upon examination to be of another sort
than the difference between our first set and the second one. If we
went on to elaborate the implications of the third set, we would be
led to considerations about the meaning of an utterance which would
be similar to those we discussed earlier, namely a specification of the
characteristics that relate to joint, cooccurrent oriented to interactive
operatives that constitute conversational work and the specific
background context of the conversational circumstance in which the
utterance in question was embedded.
We are considering this methodological problem in some detail in~
anticipation of methodological criticisms that a traditional
psycholinguist might raise in connection with an approach to the
theory of communicative competence such as the one we are
proposing here. The specific form that common explanations or
accounts take is not of critical importance. The elaboration of these
implications, however, does constitute a critical problem. Those
elaborations that seem to run against what is common knowledge
and practice as to the meaning of an utterance, become ipso facto
suspect.
Go to Part Two
