The Empirical Investigation of Conversation:
The Closing Problem
By: Dr. Leon James
Part 1 of 3
This section will deal with the specific problem of how conversations are
brought to an end. To begin with, it is necessary to understand the way in
which the closing problem is a problem. That is, what sorts of considerations
are involved in the first place to suggest that closing a conversation is a
problem as opposed to the view that the end of a conversation occurs when it
occurs, i.e. when co-
conversationalists
stop talking to each other. The nature of the issue we are raising here will be
better appreciated if it can be seen that the closing problem raises issues
similar to those when considering other types of problems of the same sort, as
for instance the problem of opening a conversation. In other words,
conversations are to be looked at as joint co-
ordinated
work accomplished by participants rather than as an event that merely happens
to occur when persons are in physical proximity of each other. This becomes
clear when you reflect upon the fact that persons may be physically present to
each other without being engaged in a conversational exchange, so that, when
conversation does take place, it is not a result of merely the fact of being
physically proximate, but implies that some joint cooperative work is being
done to put persons into a state of conversation, and we are concerned in
characterizing the nature of this work. Without developing the details
involved, it can nevertheless be seen that a person who wishes to talk to some
other party must initiate a conversation in some particular way, and in doing
so he cannot walk up to someone and just say anything and still hope to open
up a conversation. He cannot for instance accost
a stranger on the street and talk right off, in his first utterance, about
his wife's psychosomatic. illness, though he may ask a question about how to
get somewhere and expect some useful talk to ensue. Similarly, though there are
certain instances in which a person may open a conversation with an
acquaintance without a prior greeting exchange, there are other situations in
which, should he attempt to do this, this unusual action will be noticed by the
other party and will affect what happens next in the exchange. These elementary
considerations are sufficient to indicate that conversations are organized
sequences of activities requiring joint coordinated activity, and whatever
happens in conversations does not merely happen, but happens because
participants made them happen through specific sorts of cooperative
activity.
We wish to consider, then, in greater detail how conversations end or, more
precisely, what joint work do participants do in order to bring an on-
going
conversation to a close. Our first step will be to attempt to delineate the
boundary limits of closings i.e. to establish the notion of a "closing
section". It is clear what the end of the closing section is, i.e. the end of a
state of talk, but it is necessary to distinguish between a silence that occurs
within a conversation and the end of the conversation itself. Note that a
conversationalist cannot walk away at any point of silence and end the state of
talk that way. Should someone attempt to do that he will not have succeeded in
closing the conversation, but will be seen instead as having interrupted it,
the problem of closing it still will have remained unsolved. On the other hand,
a silence that follows, say, an exchange of good-
byes,
will be seen as the end of a state of talk and participants could then
walk away and be seen as having succeeded in bringing the conversation to a
close. So, an exchange of good-
byes can be seen as a device conversationalists can use to determine the end
point of a particular exchange. But this is not the only device available to
participants for closing a conversation, as the following illustrations show:
(1)
A: All right? B: Yeah, O.K. A: O.K.?B: Yeah, thanks.
(2)
A: Well, that's all, I guess. B: Alrighty. A: Fine B: Right-O.
If we look at closing sections of recorded conversations, i.e., at the last
exchange of utterances before the end of the record for an encounter, it will
be seen to contain a terminating exchange of an adjacency pair of utterances,
spoken in alternating sequence by the parties involved, such that no other type
of utterance but the terminating one will occur as the second (and final) pair
of the terminal one. Thus, in the case of leave-
taking,
one party will initiate the first pair of the terminating exchange (e.g. "Good-
bye",
"See you later"), which is immediately followed by the second pair spoken by
the other party, and this second pair will constitute the very last utterance
of the conversation, marking it's terminal boundary. If anything else happens,
viz. if the second party makes an utterance that is not the second pair of the
leave-
taking
exchange (e.g. "O.K. Good-
bye",
"Yeah, see you later"), then the conversation does not reach an end and
participants continue in a state of talk until the proper leave-
taking
exchange is recycled at some later point and successfully accomplished. In the
case of terminating exchanges other than leave-
taking,
as was the case in the two illustrations just cited, the same considerations
hold. That is, a party initiates the first pair of a terminating exchange (e.g.
"O.K.?"; "Fine."), and the second party then supplies the second part ("Yeah,
thanks."; "Right-
O")
without inserting anything else in between, otherwise the attempt at
terminating is recycled and must be attempted again at some later point.
Thus far, we have succeeded in showing part of the solution to the problem
of reaching the end of a conversation, namely, by showing that the lower
boundary limit of the closing section, the terminus, is reached at the end of a
terminating exchange properly performed, i.e. when the second party utters the
second pair of a terminating exchange initiated by the first party when
uttering the first pair of the terminating exchange, and the second party doing
precisely that, i.e. not inserting anything else but the second part of the
adjacency pair. Before this solution can be considered complete, however, it
will be necessary to consider when in an on-
going
conversation may a first party properly initiate a terminating exchange
sequence. Thus, it will be seen that should a party attempt to initiate a
.erminating exchange sequence "in the middle of the conversation", i.e. before
it is appropriate, the second party will not cooperate by automatically
supplying the second pair, but instead will do something else, as illustrated
in the following case:
A: . . . and he wouldn't let me make the announcement. He kept
interrupting
and raising all sorts of irrelevant issues and I was getting
more and more
frustrated . . . sitting there.
B: See you later.
A: Wait a minute . . . I haven't finished.
It is then a matter of crucial
importance to consider the problem of how a party who wishes to end a
conversation may do so by setting up a terminating exchange in such a way as to
enlist the cooperation of the second party such that the first pair of the
terminating exchange he initiates might properly be seen by the other as a
first pair of an adjacency pair, and respond accordingly by supplying the
second pair that will terminate the conversation.
If the terminating exchange sequence is seen as the end part of the closing
section, then we can ask what constitutes the beginning part of the closing
section, and in looking at that, we might find an answer to the problem of how
terminating exchange sequences are introduced. To do this, we have to look at
what sorts of things do participants do just before the exchange of the
terminal adjacency pair. In the two illustrations given at the beginning of
this section, the events that immediately precede the terminating adjacency pair consist of the following:
(1)A: All right?
B: Yeah, O.K.
(2)
A: Well, that's all I guess.
B: Alrighty.
We wish to focus on the sort of thing that A, the initiating
party, accomplishes by making that particular utterance. There is a class of
transactional idioms ( ) that, when used by a first party, have
the function of either indicating to the second party that he has no further
things to mention in this conversation (e.g. "Well, that's all I guess")
and/or, specifically inquiring of the second party whether
he has anything else to bring up (e.g. "O.K.?"). Let us refer to this kind
of an intervention as a proposal to foreclose the conversation. How, even
though it may turn out upon further investigation that a foreclosing proposal
may not be placed just anywhere in the conversation (e.g. right after the
initial greeting-
exchange,
it would be inappropriate, doing something else), it nevertheless turns out
that there are many more opportunities in a conversation to insert a
foreclosing proposal than there are places for inserting a terminal exchange
sequence. It will then be appreciated that the availability to
conversationalists of the device of foreclosing proposals may be one solution
to the problem of properly initiating a terminal exchange. That is, a
foreclosing proposal may be initiated by a first party, just in case it is
accepted by the second party, in which case the first party can then
initiate a terminating exchange with some reasonable chance of success. Thus,
since the initiation of a terminating exchange cannot be done appropriately at
just any point in the conversation, one point where it can occur appropriately
is immediately following the acceptance by the second party of a foreclosing
proposal. We need, now, to look at the problem of the proper placement of a
foreclosing proposal.
Specifically, we are considering a transition problem of how to get from
the end of talk about a topic to the beginning of the closing section marked by
a foreclosing proposal. One sort of general problem that is involved here is
when is talk about a topic considered by participants as complete so
that the talk can move off it, on to a new topic, and possibly, to a
foreclosing proposal. One very general aspect of the organizational structure
of conversation is that a first party's utterance that appears
to be complete need not be treated as such by another party, who may
choose to treat it as incomplete by adding something to it and attempting to
complete it, which in turn may be treated as incomplete by the first party by
adding something to it and attempting to complete it, and so on. We shall not
discuss this particular problem here (but see Sacks, 1971) but we are
mentioning it in connection with the problem of the completion of topic talk as
a reminder that our present concern relates to the more basic issue of
completion in general, and seeing that relation will help us see what sorts of
specific issues are involved in solving this particular problem of topic
completion and topic transition. That is, it will never be the case that a
party can singly, solely by his own intervention, succeed in closing a topic,
and hence be in a position of attempting to imitiate a foreclosing proposal to
be followed by a terminating exchange, since it will always be the case that
the second party may choose to consider the topic as incomplete by continuing
to talk about it. On the other hand, a first party may always have the option,
at particular appropriate places, to act as if he considers the
topic talk completed, just in case the second party might wish to accept that
claim. And one way that is available to a party to claim that he considers
topic talk to be completed is by simply making a foreclosing proposal, if
appropriate, or, if not, by announcing that he is going to make a foreclosing
proposal. Consider what's involved in the following illustration:
A: . . . and, anyway, no vote was in fact taken because by the time the
voting came up several people had left and the chairman noted that there was no
longer a quorum present.
B: We-
el,
so be it. Maybe next time round.
A: Yeah, we'll see what develops in the meantime.
B: O.K. I appreciate the
information.
A: Sure thing.
B: See you tonight.
A: Yep.
B: So long.
A: Right.
Bye.
B: Bye. Bye.
Note the nature of B's first utterance in this section of the
conversational exchange ("We-
ell,
so be it . . ."). As was pointed out by Shegloff and Sacks (1969) in their
analysis of opening up closing section in conversation, there is a class of
transactional idioms like "We-
ell"
(with rising intonation contour) and "O.K." (with a finality accentuation) that
serve the function of announcing a foreclosing proposal. That is, this class of
utterances, when used by a first party, can be seen by the second party as his
attempt to bring topic talk to a close, with the possibility that he will
follow it up with a foreclosing proposal. In the present illustration, A agrees
to treat the topic talk as complete by responding with an utterance ("Yeah,
we'll see what develops . . .") that can be seen to fall in the same category
as B's prior utterance which served to indicate that he considers the topic
talk complete. In other words, what goes on in the transition period between
end of topic talk and beginnir.g of closing section is this: the first party
initiates an utterance that serves to indicate that he considers topic talk
complete; let us call this an attempted topic completion utterance; this serves
as an announcement that he either wishes to move off that topic to a new one,
or that he wishes to make a foreclosing proposal; the second party, then
responds with an utterance of the same class, i.e. another attempted topic
completion utterance; this serves to indicate his acceptance of the first
party's topic completion attempt; whereupon, the first party then proceeds with
a foreclosing proposal ("O.K. I appreciate the information."); the second party
then responds in kind, viz. his next utterance confirms the foreclosing
proposal ("Sure thing."); then, finally, the first party follows it up by
initiating the possible first pair of a terminating exchange ("See you
tonight"), followed by the second party supplying the second part of the
adjacency pair ("Yep"); at this point, the conversation can, but need not be,
brought to an abrupt close, and in the illustration given, the parties continue
a second round of terminal exchanges, after which the conversation reaches its
terminal point.
The general solution to the problem of bringing a conversation to a close
can thus be seen to involve a type of joint coordinated sequential activity
whereby a party chooses at any point in the conversation to treat topic talk as
complete, announces a foreclosing proposal, makes the foreclosing proposal, and
initiates a possible terminal exchange. In order for the steps in this sequence
to succeed and allow a continuation to the very end, it is necessary for the
second party to agree at every point, and the way in which he indicates his
agreement is to refrain from continuing topic talk at those points and instead,
to supply an utterance that is of the same transactional type as the first
party's utterance. This general solution-
can
now be applied to solve certain more specific particular problems involved in
closing conversations. The following brief discussion is not intended to be
exhaustive of the many particular issues that are involved, but only to indicate how the general
solution is relevant in helping solve the specific problems that may be
involved.
One specific issue of interest relates to the problem of how contextual
setting features of a conversation restrict the possible placement of a
foreclosing proposal. For example, telephone conversations are different from
face-
to-
face
interactions (this being a contextual setting difference) in that in a
telephone conversation, but not in a face-
to-
face
interaction, the parties involved can be seen to have a differential status in
terms of caller and called. This difference is a noticeable conversational
feature to which participants are oriented to as evidenced by such facts as:
the called supplies the first utterance in the conversational exchange (e.g.
"Hello") and can disrupt the ordinary sequence by not supplying the expected
first utterance; the caller is in first position to introduce a first topic
which is often a reason-
for-
call
topic or a disclaimer attending a reason-
for-
call
statement (e,g, "I'm just calling to chat a bit, if you're not busy"); and so
on. Now, in considering the problem of the placement of a
foreclosing proposal, we might suspect that the solution involved in the case
of telephone conversations are potentially different in specific terms from the
solution involved in face-
to-
face
conversations. Specifically, while in face-
to-
face
conversations either party may be in a position at any point in the
conversation to attempt a foreclosing proposal, in telephone conversations the
asymmetrical status relationship between caller and called may restrict this
possibility in specific ways. For instance, if the caller does not specifically
make a disclaimer about this being a reason-
for-
call
call, and in the meantime talk proceeds without getting to a recognizable first
topic, called may not attempt a routine foreclosing proposal without risking being brusque,
unfriendly, angry, odd, etc. By routine foreclosing proposal, we are referring
to the generalized solution proposed earlier for face-
to-
face
conversations. In that case, there are non-
routine,
extraordinary devices that called has to use to engineer a foreclosing
proposal, should he wish to do so at any point, such as, for instance, an
appeal occasioned by an extra-
conversational
event (e.g. "Listen, there's somebody at the door. I'll call you back later.")
This solution is a type of foreclosing proposal that is immediate and forced,
unlike ordinary foreclosing proposals that need the agreement of the other
party. Forced foreclosing proposals are also available in face-
to-
face
conversations
but they will be of a different sort since the other party is a joint witness
to the occurrence of the unexpected extra-
conversational
event and therefore other sorts of issues are plainly involved (e.g. the second
party may not readily accept the unexpected event as one that warrants
immediate foreclosing).
We have just been considering a specific problem with-
regard
to the first party's placement of a foreclosing proposal as affected by certain
contextual setting features. Another sort of problem we wish to mention is that
involving the second party's response to a foreclosing proposal. Now, the
general solution we have proposed for opening closing sections entailed that
the second party agree to treat topic talk as completed. But there are
instances in which the second party may wish to agree to foreclose, but not
immediately, because he wishes to mention one more thing before agreeing to
foreclose. Now there is a difference to be noted between a second party's
action whereby he declines a foreclosing proposal, on the one hand, and on the
other, accepts it, except for the fact that he wishes to insert "one more thing" and only one thing. The
following three exchanges provide an illustration of this difference:
(1)
A: .
. . I dunno. Seems to me it'd be better to leave at six-
ten
rather than eight-
forty
so we can avoid the morning traffic rush.
B: We-
ell,
I see what you mean. O.K. In that case I'll pick you up at five-
fifteen
sharp.
A: O.K. That's fine.
B: Right. See you in the morning.
A: Mm hum. Good-
bye.
B: Good-
bye.
(2)
A: . . . I dunno. Seems to me it'd be better to leave at six-
ten
rather than eight-
forty
so we can avoid the morning traffic rush.
B: We-ell, I see what you mean. O.K. In that case I'll pick you up at five-
fifteen
sharp.
A: Well, don't you agree that'd be better? Last time I had to drive to the
airport in the morning it took me over an hour. It's really a scandal. What's
happening with the planned airport thruway, anyway? You know, I hope O'Connor
loses the race this time. He's just no good. So inefficient. Did you hear about
the cancellation of the Moikiki housing project? Etc., etc.
(3)
A: . . . I dunno. Seems to me it'd be better to leave at six-
ten
rather than eight-
forty
so we can avoid the morning traffic rush.
B: We-
ell,
I see what you mean. O.K. In that case I'll pick you up at five-
fifteen
sharp.
A: All right. Oh, yes. Now, listen. Make sure you remember to take Ward
Avenue rather than Vinyard. They're still doing work on the Chester interchange
and it's really bad even that early in the morning. O.K?
B: Right-
O.
Don't worry. See you in the morning.
A: Mm humm. Good-
bye.
B: Good-
bye.
These three variant versions of a conversation illustrate, respectively, the
following events. In (l), B announces a foreclosing proposal, which
A accepts, whereupon B initiates the terminatlng exchange. In (2), B
inserts a foreclosing proposal at the same place, but A declines it
outright and the conversation moves on to other related and unrelated business.
In (3), A accepts B's foreclosing proposal but chooses to insert one more thing
before accepting it. The difference between (2) and (3) can be seen to matter
in that in (3) A succeeds in mentioning another topic without requiring B to
attempt another foreclosing proposal at some later time, as will be necessary
in (2). This consequence may be seen to be a virtue for a number of reasons,
one being that the problem of bringing the conversation to a close is thereby
facilitated, another being that, with this solution, A maintains the
status quo whereby it is B that is seen to have made the foreclosing proposal.
Since this sort of fact, i.e. which party initiates the closing section may be
a matter of relevance in a relationship (cf. "You're always running off
somewhere when I try to talk to you"), it can be appreciated that the solution
B uses in (3) of inserting one more thing before agreeing to foreclose may have
a special virtue to it that (2) does not share.
We have discussed, thus far, two specialized applications of the general
solution
offered for the problem of closing a conversation: the way in which the
placement of a foreclosing proposal is affected by certain contextual setting
features, and the way in which a second party might deal with a foreclosing
proposal other than accepting it or turning it down. Now we wish to mention a
third, and final, specific issue, which is the special status that some topics
have as doing foreclosing work. That is, there is a class of topics that can
automatically serve as an announcement of a foreclosing proposal
merely by raising them in the conversation, their virtue being that no other
foreclosing attempt need be made as a way of bringing a conversation to an end.
For instance, in certain formal conversational settings such as a university
departmental committee meeting, a participant can be seen to attempt to
foreclose merely by raising the topic of the time of a subsequent meeting (e.g.
"Mr. Chairman, when are we scheduled to meet again?"). The reason that such an
intervention constitutes a foreclosing attempt lies in the fact that the topic
of the time of a next meeting is seen by participants as a "last topic", a last
order of business that is to be handled after all other topics have been
exhausted, and thus the mere fact of someone raising that topic indicates to
others that he is proposing that all other business has been exhausted and
that, therefore, it is time to foreclose the meeting. In cases where the
committee meetings are not seen as recurrent, and hence, there is no
anticipated issue of setting a next meeting date, another topic may serve as
the last topic, with equivalent consequences for foreclosing, by reference to
an agenda, where an individual raising the last mentioned topic on the agenda
can be seen by others to therefore be raising a last topic.
In ordinary face-
to-
face
conversations, it is sometimes possible for a participant to announce a
foreclosing proposal by directly raising a topic that can be seen to be a
potential last topic, in which case it automatically serves as a
foreclosing attempt. The problem, in this case, is more conplicated than in the
case of formal meetings both on account of the fact that in
ordinary conversations there is no agenda to refer to in the determination of a
possible last topic, as well as on account of the fact that there is no readily
available procedural matters such that any one of them can be seen as a
habitual last order of business. Instead, considerations of another sort will
be necessary to successfully claim that a particular topic is to be
seen as a last topic. These considerations will have to relate to the interaal
organization of the single conversational episode involved. For instance, a
previously mentioned topic may be reintroduced by a party or, an on-
going
topic may be commented on in such a way as to be seen to be presenting an end-
of-
topic
statement. He may, for instance, use a transactional idiom that indicates that
he is summing-
up
or making a final conclusion: "Well, in any case, . . ."; "To cut a long story
short . . ."; "Well, that's life"'; "I guess, you can't win them all"'; "So be
it." etc. These sorts of comments are seen by other participants as end-
of-
topic
comments, i.e. the party producing the utterance is seen as wanting to complete
topic talk. Now we have seen that announcements of foreclosing proposals may
properly occur at the transition point that is marked by end-
of-
topic
talk and thus an utterance that serves as an end-
of topic
utterance can be seen as a possible immediate place for inserting
foreclosing attempt. The question then is, when does an end-
of-
topic
utterance serve the double function of possibly ending topic talk and
possibly announcing a foreclosing proposal, as opposed to merely serving
as end-
of-
topic
X and a possible beginning of topic Y? The overall solution to this problem is
quite complex, and we shall only present here some general schematic
considerations as to the sorts of issues that would be involved in a fully
adequate description.
There is a class of conversational events that are interactionally marked
for behavioral consequences of specific sorts that are distinctly different
from other conversational events not marked thusly. For instance, extending an
invitation to a party has very definite behavioral implications that are absent
when discussing a party as an event, or when discussing the events that are
reported to have occurred at a party. Thus, parties to a conversation will be
seen to sometimes re-
introduce
the topic of an invitation by merely re-
extending
the same invitation as proposed earlier, or reintroducing a confirmation of an
acceptance, without adding anything new to it, as transacted earlier.
This does not occur with topics that are not marked for behavioral
implicativeness, in which case re-
introducing
a previous topic is not done unless it is accompanied by some after thought
comment which is new. Thus, when a party re-
introduces
a previous topic without adding anything new to it, he can be seen as
attempting to raise a last topic, since he would not do such a thing if he were
interested in continuing topic talk. Once the intervention is seen as raising
a last topic, it can then serve the additional function of announcing
a foreclosing proposal that is to come next.
We have been discussing the way in which merely raising a particular sort
of topic can serve, by virtue of the nature of the topic, as being a last
topic, as constituting an announcement at foreclosing a conversation. We have
seen that in certain formal settings, some particular topics are
automatically marked as a last topic by reference to an habitual
procedural sequence or an agcnda. In ordinary face-
to-
face
conversations, re-
introducing
a topic without adding anything new to it, such as those having behavioral
implicativeness, are marked thereby as a last topic, and therefore,
can serve as announcing a foreclosing attempt.
Re-
introducing
an invitation to a party, and other sorts of business having to do with
finalizing arrangements, is but one possible device fGr marking a topic as a
possible last topic announcing foreclosing. A topic can be directly marked as a
possible last topic, even if it is not reintroduced but is raised for the first
time: e.g. "Listen. Before I go, I just want to check with you about . . .".
Although we have not fully investgated this matter, it is our hypothesis that when topics are thusly marked in a
direct way, the justifying marker will inv41ve an appeal to some sort of
behavioral implicativeness (e.g.-
"Well, I gotta go now. What about ..."), and thus these two alternative
solutions can be seen to be: related.