Journal of Pragmatics 14 (1990) 397-412 397

North-Holland

CONVERSATION MANAGEMENT IN CONTRAST:

Listener Response in Japanese and American English

Senko K. MAYNARD*

This paper proposes a methodological framework, C(ontrastive) C(onversation) A(nalysis) within

which some of the listener back-channel expressions in Japanese conversation are analyzed in
contrast with those observed in _American English. i, ack-channel expressions examined are limited
to uh-huh's and the like, brief comments, punctuated head movements and laughter.
After ana'3zing data consisting of 40 dyadic casual conversations videotaped in Japan and the
United States, it is concluded that in Japanese casual conversation, listener's response such as
brief comments and head movements occur far more frequently than in comparable American
situations. Relevant contexts for listener back-channels in each speech community are found to
differ significantly. In Japanese, grammatical cc.mpletion, sentence-final particles and speaker's
vertical head movement provide the relevar~t context while in English grammatical completion
provides the single most significant context.
The results of the contrastive study are evaluated and assessed in light of the potential problems of
CCA, with the issue of "equivalence' being the most serious and problematic. Additionally, as a step
beyond CCA, four intercultural conversations by American and Japanese speakers are examined.

* Correspoedence address: S.K. Maynard, 231 S. Adelaide Ave., Highland Park, NJ 08904. USA.
This is a revised versioa of my paper with the same title presented at the 1987 International
Pragrnatics Conference held in Antwerp, Belgium. A detailed report on the contrast of listener
back channels based on 12 pairs in Japar~ese and American English has appeared in Linguistics
(1986), 24(6) 1079-1108. The method, Contrastive Conversation Analysis, is also proposed in the
author's book which has appeared since, Japanese Conversation: Self-Contextualization Through
Structure and Interactionai Management, 1989, Norwood, NJ: Ablex.
I would like to express my gratitude to the Toyota Foundation, which funded the collection of
data analyzed herein. I also wish to express my gratitude to the National Endowment for the
Humanities for supporting the 1985 Summer Institute, 'Humanistic Approaches to Linguistic
Analysis' which I was fortunate enough to parti(2ate in. My special thanks go to Deborah
Tannen who directed the Institute. A portion of this research was also assisted by a grant from the
Joint Committee on Japanese Studies of the American Council of Learned Societies and the Social
Science Research Council with funds provided by the National Endowment for the Humanities,
the Ford Foundation, and the Japan-United States Friendship Commission. The collection of
videotaped English conversations, performed by native American English speakers and Japanese
students studying in the United States, was funded by the 1986-1987 Rutgers University Grant
Program in International Studies. I thank the audience at the presentation of the present paper a,
the 1987 International Pragmatics Conference, especially Charles Goodwin, who offered critical
commentary. Naturally, I am solely responsible for any shortcomings and inadequacies contained
in this paper.

0378-2166/90/$03.50 © 1990 - - Elsevier Science Publishers B.V. (North-Holland)

398 S.K. Maynard / Conversation management in contrast

1. introduction

This paper introduces a method in contrastive study which analyzes conversa-

tional discourse. Given this purpose: a brief review of contrastive analysis is in
ordci. The field of Contrastive Analysis recognizes two different trends: (1)
Contrastive Analysis in applied linguistics, primarily subscribed to and later
criticized and, in some cases, abandoned by scholars in the United States and
(2) Contrastive Linguistics or Contrastive Study, primarily conducted by
European linguists. Since the appearance of Lado's Linguistics Across Cul-
tures in 1957, a great many publications have appeared in America advocating
the use of Contrastive Analysis in language instruction. In the late 1960s and
throughout the 1970s, however, Contrastive Analysis came under severe
attack. The attack came primarily from applied linguists who have held unfair
expectations of Contrastive Analysis to predict language learners' errors. The
general atmosphere of doubt and disappointment in Contrastive Analysis is
expressed by Wardhaugh (1970: 123) when in 1970 he wrote that "a period of
quiescence is probable for contrastive anai),~is".
The American predicament, however, dGes not represent a universal
approach to how contrasting different languages should be conducted. In fact,
as Fisiak (1980) states, the situation has been entirely different in Europe. The
research area called Contrastive Linguistics has attracted many European
linguists from a number of countries who have made theoretical contributions
to contrastive studies since the 1950s. To these linguists, Contrastive Linguis-
tics may be defined as "a subdiscipline of linguistics which is concerned with
the comparison of two or more languages (or subsystems of languages) in
order to determine both the differences and similarities that hold between
them" (Fisiak (1980: 1)).
Among studies which contrast various aspects of Japanese with those of
English - for example, Kleinjans (1958), Kunihiro (1974), Kokusai Kooryuu
Kikin (1976), Monane and Rogers (!977), Kenkyuusha (1978), Azuma (1981),
and Higa (1982) - some of the e~r!ier works have included stronger pedagogical
implications. Recent studies contrasting Japanese to English have had a
weaker pedagogical orientation in general, with less emphasis given to the
theme of error prediction. For example, contrast-oriented studies, such as
Ikegami (1981a,b), Kageyama (1981), and Kuno (1982), are conducted
without pedagogical applications in mind.
Earlier contrastive analysts concentrated on phonology, morphology, and
lexicography, with moderate interest in syntax and semantics. In more recent
years, a host of studies has emerged in which languages on a level beyond the
inidividual sentence are contrasted. For example, Gleason (1968) contrasts the
narrative structure of the Kate text to its semiliteral English translation. More
recently, Hartmann (1978, 1980) introduces the term 'contrastive textology',
through which he aspires to develop a discourse-based contrastive study.
 S.K. Maynard / Conversation management in contrast 399

James (1980) explores the possibility of Contrastive Analysis on the macro-

linguistic level, including both written and spoken discourse. Special attention
is paid to the conversation analysis in Ventola (1980) in which Finnish and
English conversational interaction is analyzed.

2. Toward Contrastive Conversation Analysis

With increasing interest in discourse and conversation analysis in several

speech communities, I find it necessary to turn our attention to a methodolog-
ical framework within which the results of conversation analysis may be
more accurately contrasted and evaluated. In this paper I explore the method
'Contrastive Conversation Analysis', or, CCA. As a case study of this
methodology, a contrastive analysis of listener response behavior in Japanese
versus American English is discussed. As a step beyond the proposed CCA
proper, toward the end of the paper the contrast of listener response behavior
between Japanese and Americans as observed in English casual conversation
by native (American English) and nonnative (Japanese) speaker pairs is
discussed.
In contrasting conversational strategies across languages, data often
consists of a few conversational interactions collected in non-controlled
sociolinguistic environments. For example, in Ventola (1980) recordings of
casual conversations conducted by two Finnish and two English speakers are
used. Unfortunately, the English data analyzed is produced by a nonnative
speaker of English (herself) and a native Australian English speaker. In terms
of adequate representation of English casual conversation, her data is less
than ideal. The CCA framework is proposed as a possible improvement to
completely ad hoc data collection, and as a framework in which the con-
trasted results may be appropriately assessed and re-evaluated.
Among various aspects of conversation that may be contrasted, this study
focuses on interactional management, specifically, some of the listener back-
channel responses observed in Japanese and American English casual conver-
sation. The research areas available for potential CCA include a contrast of
macrostructures such as (I) thematic structure, (2) narrative structure, (3)
'preference organization' (Pomerantz (1984)), and (4) conversational goals
and functions as well as a study of the interrelations among macrostructures.
Among primarily local phenomena, the speaker's iUocutionary acts or 'moves'
(as suggested by Edmondson (1981) and Stubbs (1983)) and 'adjacency pairs'
~Schegloff and Sacks (1973)) are useful for contrastive purposes. Conversa-
tional maxims and implicatures as proposed by Grice (1975), as well as a
variety of strategies for performing speech acts constitute interactional aspects
worthy of investigation. A contrast may be made in terms of the interactional
management of the conversation, such as strategies for turn-taking, usage of
400 S.K. Maynard / Conversation management in contrast

back-channel expressions, on the one hand, and what Tannen (1984) calls
'conversational style' on the other. Although this area has not been explored
until recently, the variability in conversation management provides an inter-
esting area for Contrastive Conversation Analysis.
In C C A , the first step requires data collection in two contrasting speech
communities. First, the data to be contrasted must come from the same genre.
In this study the genre under discussion is what is normally called casual
conversation. When deciding which conversational data to analyze, it is
important to choose a type such that collection procedures may be replicated
in both communities. Since conversation occurs in a specific context, the
social situation must be predetermined - for example, sex, age, social status,
relationship between subjects, number of participants, and the setting in which
actual conversation takes place. The type of conversational data selected in
one speech community must be comparable to the data chosen from the
second speech community. Moreover, the method for data collection must be
applicable to both speech communities.
The second step is data analysis. Even whe n comparable data is co,,e,.,ed
. . . . II ,-.*

from two contrasting speech communities, it is logically impossible to contrast

two different facts without postulating a common framework in which the
actual contrast is performed. An appropriate research design is such that two
languages are analyzed by an identical procedure. In the third step the results
of data analysis are brought into focus.
The fourth step involves actual contrast and comparison between the
analyzed results. Following the initial preliminary contrast, it is important to
assess and interpret the results in light of linguistic and sociocultural idiosyn-
crasies of the speech communities under investigation (the fifth step of CCA).
Even when the same methodology is applied, the subject matter being
analyzed may bear different significance in different societies. By concentrat-
ing on pre-selected linguistic and interactional strategies, for example, the
researcher may overlook seemingly unrelated devices which in fact serve
similar or nearly equivalent functions in the contrasted speech communities.
Too narrow an assumption of what the research will yield can result in a
wealth of useful information going completely unexamined. This point should
be noted and taken into consideration when assessing the results at this final
stage of CCA. This final step also requires re-evaluation of the quality of data
analyzed, the accuracy of analysis made as well as of the adequacy of
conclusions drawn.

3. On the 'equivalence' for contrast

The single most critical notion in CCA is the issue of 'equivalence'. In earlier
studies, ' equwalence
," ,~. . . . , . . . . ~ L , • +~, . . . . . . t;,~ ,. . . . ;,,olo,,.o, of , l , 0 r d ~
' .....
 S.K. Maynard/Conversation management in contrast 401

Contrast was often made on the basis of translation, whose semantic equiva-
lence was either assumed or specifically displayed through an identical deep
structure. When Gleason (1968) introduced the notion of discourse-level
contrast, rendering useless the concept of the common deep structure, he
turned to a text of one language (Kate) and its semi-literal English translation.
ZydatiB (1982), on the other hand, analyzes a German text and its published
English version. James (1980:117) suggests that in a bilingual society, 'paired
text' may be available. Although such text should ideally be equated texts,
that is~ independently produced texts of two languages, James suggests that
normally there is evidence of translation. In Maynard (1983) an analysis of
relative clauses in written text in Japanese and English was made on the basis
of data consisting of the published Japanese and English translations of two
short stories, one originally in Italian and the other in Spanish. These
published texts were translated by speakers of Japanese and English, the
rationale for their choice being to avoid the translation of original Japanese or
English works into English or Japanese. This is expected to avoid or at least
to minimize the unnecessary influence on or prejudice of one of the languages
for or against the other. Chafe's (1980) project analyzes oral narratives
produced by subjects after they viewed a common film. The assumption here
is that a common visual stimulus creates almost identical situations for
subjects of different languages.
In the present study, equivalence is sought in the sociolinguistic context in
which the conversational data is produced. That is, in both Japanese and
American English, the social context has been defined and structured so that
the participant behaves in a specific manner, thereby rendering the data
obtained not only genre equivalent (casual conversation) but also type equiva-
lent (dyadic casual conversation between same gender speakers, and so on).
However, I must mention that there are several unresolved issues regarding
the equivalence in the present work. For example, variables such as choice of
discourse theme, and ethnic and regional differences among subjects were not
controlled. Depending on the type of themes discussed, discourse is more or
less prone to produce argumentative or supportive discourse interaction,
which is likely to influence listener back-channel behavior. Likewise, as
suggested by Tannen (1984), ethnic differences within a speech community are
likely to bring forth different conversational styles including listener back-
channel strategies. However, on the level of national speech community, all
our American subjects identified themselves as native speakers of American
English and they are fair representatives of American speakers with expected
diversities.
1 In selecting American subjects, only those whose parents were native speakers of American
English were considered. All subjects had no (or only minimum) exposure to foreign cultures and
identified themselves as native speakers of American English. Regarding their ethnic background,
the o..k~,,o =~s:-.,ere--! in ,ho.:r ~ ..... ~,,~, .... questionnaires as follows. The question asked was:
402 S.K. Maynard / Conversation management in contrast

4. Data

The data used for this study consists of 40 videotaped dyadic conversations
taped in Tokyo and New Jersey, in May and February, 1985, respectively. 2
The pairs selected were of the same sex (10 female and 10 male pairs in each
country) and were all college students. Subjects identified each other as friends
and they had had so-called casual conversation on numerous occasions. At
both locations, an unattended video camera was used and the subjects were
left alone in a room after being instructed to talk as naturally as possible
about anything they liked. So as to minimize the degree of subjects' awareness
of being filmed from being reflected in our data, the initial two-minute
segments were categorically excluded and the following three-minute segments
were selected as data relevant to this study)

5. Defining gstener back-channel expressions

Within the proposed framework of CCA, we examine listener responses - or

what Yngve (1970) refers to as back-channel expressions - in Japanese and
American English. Before discussing the specifics of back-channel expressions,
however, the concept of back channel itself must be clarified. In the present
study back channel is defined strictly in the context of turns within the turn-
taking system and refers to occurrences of listener behavior where an interlo-
cutor, who assumes primarily a listener's role, sends brief messages and signs
during the other interlocutor's speaking turn. 4 Beyond verbalized uh-huh's
and brief comments, various aspects of nonverbal behavior may be called
back-channel expression. In the present study only clearly visible punctuated

'Do you consider yourself to belong to a specific ethnic group in the United States? If your answer
is Yes, please comment.' Out of 4() subjects, 25 responded with 'no'; 9 indicated they were Jewish;
4, African American; and 2, Irish American.
' For further discussion regarding the characteristics of data used for this study, see Maynard
(1989).
3 The assumption that speakers grow accustomed to being recorded and that unnatural speech
decreases with time is shared among conversation analysts in general; cf., for example, Duncan
and Fiske (1977: 37).
¢ Yngve 0970: 568) states that back channel is observed "when the person who has the turn
receives short messages such as 'yes' and 'uh-huh' without relinquishing the turn". In the present
study, in order to identify when the state of 'in-the-turn" starts and ends, the following definition
of turn is adapted. Without this definition it is difficult, if not impossible, to identify some of the
back-channel-like utterances observed in the conversation-in-progress.
Turn is a fundamentally solo-speaking unit recognizable by participants as carrying some
referential and/or functional meaning. Moreover, for an utterance to be considered a turn,
both talking and non-talking participants must recognize that the taker of the turn says
something and thus his/her activity is recognized as such while the non-talking partner assumes
a complimentary listener's role.
 S.K. Maynard / Conversation management in contrast 403

vertical and horizontal head movement is examined, s Laughter is another

utterable frequently observed in our data. As suggested by Schenkein (1972)
laughter plays an important rote in communication and we identify the
listener's laughter during the other interlocutor's turn to be a case of back
channel
The back-channel behavior among Japanese - commonly known as aizuchi -
has been noted by more than a few scholars. (See for example, Mizutani
(1983, 1984), Clancy (1982), and Miller (1987).) Of the studies that analyze
Japanese aizuchi in contrast with American listener response, Mizutani's
(1983, 1984) works are most significant to the present one. Mizutani (1983,
1984) analyzes aizuchi (brief utterances only) in one radio and two television
programs, totalling approximately 34 minutes, with one interaction lasting
over 20 minutes. Four speakers participate in one television program while
two speakers participate in two other interactions. Although interesting, I find
Mizutani's work lacking in some aspects. First, Mizutani does not raise
potential problems of analyzing both dyadic and multiple-participant conver-
sations as homogeneous data without making distinctions. (How did she
analyze (almost) simultaneous aizuchi (which I suspect actually occurred) sent
by multiple listeners to one prominent speaker?) Second, there is no analysis
of non-verbal aizuchL although head movement is dearly a prominent part of
the aizuchi strategy. Third, although she points out the difficulties of non-
native (English) speakers of Japanese learning how to send aizuchi - due to
the differences of listener responses between Japanese and English speakers -
the listener response among English speakers is merely anecdotally comment-
ed. The present study attempts to resolve some of these points within the
proposed CCA framework.

s Vertical head movement is defined as a clearly visible vertical head movement which
accompanies at least one occurrence of lowering the head immediately followed by a movement of
raising the head approximately back to the starting position. Horizontal head movement is
defined as movement which involves turning the head either left or right immediately followed by
a movement turning back to the starting position. Excluded from this analysis are head raising,
head lowering and other subtle movements. This exclusion was made due to the difficulties of
accurately identifying some of the more subtle non-verbal signs on the videotape. Obviously this
does not mean that the non-verbal behavior excluded in this study is not functionally significant;
in fact the reverse may be true. As suggested by an anonymous reviewer (to whom I am thankful),
some of the more subtle signs may be more significant than the ones I examined in this study.
Such analysis, however, must await future stuoy.
Head movement was counted and recounted before we positively identified it as such. First, the
researcher identified cases of head movement. Approximately six months later, the researcher
viewed the tapes again, this time with an assistant (a non-linguist native speaker of Japanese who
was coached to point out clearly identifiable head movement as defined in this study). When faced
with disagreement as to whether or not head movement occurred, the researcher and the assistant
discussed the issue and reached agreement.
404 S.K. Maynard / Conversation management in contrast

6. Analyzing listener back-channel response in Japanese

Having identified what is investigated in the present study, let us turn to the
examination of data set (1), a ! 7-second segment. 6

(1)
I.A: Dakedo/
but
'But'
. are atsuryoku ga tsuyoi n da yo ne hora/
that pressure S strong NOM BE FP FP you-see
'there's great pressure,'
. hoogakubu jan./
law-school TAG
'cause (l'm graduating from) Law School.'
H H H
(B:I Aa soo ka boo ka uun)
oh so Q so Q u h h u h
"Oh, I see, I see.'
. dakara/
SO
"so,'

H
. mottainai to ka iwarete sa./
wasteful QT Q say-PASS-and FP
"I'm told ~hat it's not good enough for me.'
H H
6.B: A mawari kara he./
oh surrounding from FP
'You mean (you hear that) from people around you.'
(A:I H)
The following abbreviations and transcription symbols are used.
recognizable pause
utterance final contour
H indicates vertical head movement
BE copulative verb, be
FP final particle
LK linker (linking nominals)
NEG negative
NOM nominalizer
PASS passive morpheme
Q question marker
QT quotative marker
T theme marker
TAG tag questions including jan. ja-nai, etc.
 S.K. Maynard/Conversation management in contrast 405

. Oya kara sureba kodomo ga sureba i y a / L A U G H H

parent from do-if child S do-if no
'From your parents' view, if the child does ...'
H H H
(A:2 Soo soo soo soo)
yeah yeah yeah yeah
'Yeah, yeah, yeah, yeah."
8.A: Demo oya oya wa ne moo saikin s o o m o /
but parent parent T FP any more recently so even
'But nowadays parents don't'
(B: 2 Soo)
so
bl see.'
. iwanaku-natta kedo/H
say-NEG-become but
'say those things."
(B:3 H H)
!0. tomodachi toka wa sa mottainai yoo toka sa/
friend or T FPwasteful FP or FP
'The way my friends look at it, they say things like, "It's not good enough for you."'
H
(B: 4 Uun)
uh huh
"Uh huh.'

In d a t a set (1) 4 t o k e n s o f B's b a c k c h a n n e l are f o u n d d u r i n g A's turn, a n d

2 instances o f A's back c h a n n e l d u r i n g B's turn. 7 S o m e b a c k - c h a n n e l devices
are strictly verbal as in the case o f B's b a c k c h a n n e l (B: 2), soo q see'; s o m e
are strictly n o n v e r b a l as in A's b a c k c h a n n e l (A: i). 8 S o m e listener back-
c h a n n e l b e h a v i o r c o m b i n e s verbal a n d n o n v e r b a l signs as in B's short
u t t e r a n c e (B: I), A a soo k a hoo k a uun ' O h I see, I see' a c c o m p a n i e d by three
repetitious h e a d m o v e m e n t s . In a n a l y z i n g b a c k - c h a n n e l behavior, we focus on
its two aspects, n a m e l y , frequency a n d discourse context.

7 The reader may argue that A's back channel (A: 2) does not occur during B's turn and should
therefore not be included in our analysis. I included, however, back channels that occur
immediately after the current speaker stops talking and are followed by a pause before the next
turn starts. This decision is reached because such back-channel response gives the impression that
it is the listener's response to the current speaker during his/her turn and it occurs before a turn
transitional period starts. Back-channel-like utterables during prolonged inter-turn pauses and
turn-internal lapses, however, are excluded, since their primary function seems to be that of fillers.
See Maynard (1987) and (1989) for a detailed analysis of head movement in Japanese
dyadic conversation, which discusses some head movement functioning as back-channel
response.
406 S.K. Maynard / Conversation management in contrast

The overall frequency of back-channel expressions (as defined in this study)

across 40 Japanese speakers totalled 871. Out of the 871 cases, 703 were
identified as back channel occurring in the immediate neighborhood of
recognizable pauses or breaks in :cmpo made by the current speaker. The
most frequently occurring types among all back-channel expressions found in
our data were brief utterables such as un 'uh-huh', honto 'really', and soo 'I
see', which totaled 70.49%. Head movement accompanied these brief utter-
ables 62.87% of the time. In the second most frequent category, vertical head
movement (which does not accompany verbal expressions), accounted for
18.83%. Head movement occurred either independently or with verbal back-
channel in 63.15% of all back-channel expressions.
Let us now focus on the context in which back-channel behavior occurs.
First, in the majority of occurrences, listener back-channel behavior is near or
at speaker pauses. Second, these pauses are marked by frequently occurring
linguistic devices. One such device is the final particle. For example, in data
set (1) the particles ne and sa each occur twice while y o occurs once at or near
the pause. Ne, which Uyeno (1971) calls a 'particle of rapport', solicits listener
response - either as a back channel or as a request to change the turn. The
particles sa and y o which appear in data set (!) also serve to solicit listener
response but do so less overtly than does the tag-question-like ne. Sa and y o
function as emphatic markers in part and mark phrasal or clausal boundaries
that are likely to be followed by a short pause.
Particle endings in general signal, together with the pause itself, the moment
where some manner of feedback may be relevantly performed. In fact, particle
endings marked 40.84% of all instances where back-channel expressions
occurred near or at identifiable pauses. Auxiliary verb endings also function
similarly to final particles. Auxiliary endings glossed as TAG, such as deshoo
'isn't it right?' and ja-nai 'isn't it?', marked 54 locations. The discourse
context in which listener response was overtly solicited by final particles or
auxiliary verb forms characterized 48.69% of back-channel expressions.
Thirdly, even when no particles are attached, pauses occur at the major
clausal and sentential junctures - such as at the juncture of subordinate
clauses as in (1.9.). Another utterance ending often observed is that of the
gerundive -te form of the verb. The -te forms which appear without stress and
with falling intonation also function as grammatically complete points. Out of
all locations where back channels occurred, 51.02% occurred at major gram-
matical junctures, some accompanied by sentence final particles and/or head
movements.
Data set (I) also reveals an interesting nonverbal behavior on the part of
the speaker at the point where back channel is observed, namely, head
movement that co-occurs at or near the final syllable of the speaker's
utterance. For example in (1.6) a case is observed where back-channel devices
follow the speaker's head movement. Similar cases of head movement are
 S.K. Maynard / Conversation management in contrast 407

observed quite frequently across all Japanese data examined. Out of all the
back channels, 38.08% occur in the context of the speaker's head movement.
One point should be noted here. The present study does not address the
predictability of back-channel occurrences, given the discourse contexts iden-
tified above as providing 'cues'. We have only observed actual occurrences o f
back-channel expressions and the characteristics of their discourse contexts.

7. Analyzing Hstener back-channel expressions in American English

Based on identical methods, back-channel behavior (specifically, u h - h u h ' s and

the like, brief comments, head movements and laughter) among Americans
was analyzed. In data set (2) given below, a 47-second segment taken from
our data, four cases of back channel are found, two utterances each by
speaker A and B.

(2)
i.A: I ordered some escargots/
2. and got me a coke./
3. I was like/
4.B: i have never been to K. Miller./
5.A: I don't know just like/
6. strikes me as being very pseudointellectual./
7. Don and I were walking past (?) going to that little shop
8. past it's open only three days or something./
(B:I Um hum)
. you know the one I bought my uh
10. dice bag.
il.B: Yeah I think I know what you mean./
(A:! Yeah)
12.A: And we were going there and this guy came out of K. Miller because he notices
us looking at the menu and he goes/
13. Hey, Babe, want ~ drink? Come on inside rll pay for you./
(B: 2 LAUGH)
14. And we were like "Oh go away"./
15.B: Weird/
(A: 2 Yeah)
16. No I heard the food's actually good though./
17.A: All I know is Polly offered me a slimy little escargot and I said thank you but
no. LAUGH/
18.B: Oh I like escargot./
19.A: I don't./
20. I I just keep on thinking slime/
21. sledge/
22. sea bottoms, you know./
408 S.K. Maynard / Conversation management in contrast

After examining three-minute segments of 20 American pairs, we observe a

total of 428 cases of back channels, 373 of which occur near or at an
identifiable pause. In terms of frequency, the American pairs produced far
fewer back channels. The lowest frequency of back-channel near or at the
pause was observed in an American male pair (6 times), while the highest
frequency was found in a Japanese ?emale pair (55 times). Although there are
instances among the American pairs in which more back-channels are ob-
served than the least frequent case among the Japanese pairs (18 times), the
difference between the means of the two groups is highly significant (the t-test
for two sample means, with p value less than 0.001, two-tailed test), and,
therefore, there is reasonable evidence to conclude that Japanese send more
back channels (those investigated in this study) than do Americans in casual
conversation.
The most frequently occurring back-channel strategies among the American
pairs were brief utterables such as uh-huh, yeah, and right, which resulted in
50.23%. Head movement accompanied these brief utterables 50.70% of the
time, somewhat less freouently, than in the_ Japanese_
_ data ~62.,,~7°/-~,,,,. As for
head movement (which does not accompany verbal cues), American pairs
engaged in this form of feedback 150 times (35.05%), while the Japanese pairs
did so 164 times (18.83%). The laugh category was observed 63 times
(14.72%) among American pairs as compared with 93 times (10.68%) among
Jat~anese pairs. In terms of contexts for back-channel strategy in American
English the devices simi'~r to those examined in Japanese were focused on,
naauely, grammatical comp!ction, phrase ending markers such as you know, or
what Bernstein ~.,962" 723~ calls "sympathetic circularity sequences", plus tag
questions and head movement. 82.84% of back channel in the American data
occurred a' the point of grammatical completion. Sympathetic circularity
sequences provided context in only 6.97%, while head movement occurred in
only 7.77%. In English, then, the grammatical completion point is the single
most powerful contex~ for back channel; other criteria appear to mark only
marginally the points of relevance for back channel.

8. Assessing and re-evaluating ~he results of analysis

Up to this point, the first four steps of CCA have been performed. Now the
fifth and final step. In assessing the observed differences in listener back--
channel behavior in Japanese and American English, we must note the
following. First, the observed difference is partly a function of the language
itself in that certain devices are available only in one of the languages
contrasted. It is sometimes suggested that Japanese final particles function
similarly to English tag questions. Although in Japanese such particles can be
placed within the sentential boundary, in English the tag question is used at
 S.K. Maynard / Conversation management in contrast 409

the end of the full sentence only - its usage being much more restricted. While
in the Japanese data final particles are used a total of 863 times, in the
American data there is a total of only 67 cases of sympathetic circularity
sequence. 9 This means that the availability of linguistically marked environ-
ment where potentially the listener may send responses differs in these two
languages.
Second, although the continuous flow of brief utterables and head move-
ments suggest that Japanese interactants possess a strong inclination for
mutual monitoring and cooperation, this does not mean that American
English lacks in listener back-channel responses per se. It is possible in
English that other speaker behavior and listener back-channel strategies that
we have not investigated, such as eyegaze and subtle shifts of head, for
example, are used for similar purposes. We can only conclude that the
frequent back channel monitoring through abundant brief utterables and head
movements is characteristically Japanese (in contrast with American English).
Third, although care was taken in selecting genre- and type-equivalent data
in the two contrasted speech communities, other variables were not ade-
quately controlled. In future studies, variables such as choice of discourse theme,
type of verbal text, and ethnic and regional differences among subjects must
be carefully addressed before we draw any final conclusions. Additionally, in
the future, inquiry should be made as to which functions of back-channel
expressions are characteristically associated with the discourse context in
which they occur. Naturally, whether the results of the present study based on
dyadic casual conversational discourse are applicable to other genres awaits
further study.l°

9. An examination of intercultural discourse

As a step beyond CCA proper, I made a preliminary examination of English

conversations produced by four pairs (2 female and 2 male) of Anierican/
Japanese speakers. The results discussed above suggest that - since Japanese
tend to send brief utterables and head movements more frequently than do
Americans - perhaps this difference will be made manifest in intercultural
communication. The data analyzed consists of four dyadic casual conversa-
tions videotaped in New Jersey in March, 1987 with the same method as used
for the collection of American conversations described at the beginning of this
paper. The subjects were all college students of ages ranging between 19 and

o While 49.48% of Japanese final particles occurred clause-internally, no case of sympathetic

circularity sequence in the American data appeared within the clause boundary.
1o For example, in an argumentative discourse the primary function of back channel may be an
expression of disagreement. Different discourse genres may foreground specific functions of back
channel and background others.
410 S.K. Maynard / Conversation management in contrast

28. The Japanese students had studied in America between 7 months to 3

years, and all ranked approximately at the lower to middle intermediate levels
of English language proficiency.
From each videotaped conversation, the first two-minute segment was
excluded to preclude any initial awkwardness from being reflected in the
analysis. Starting after the two-minute segment, the first two speaker turns
lasting longer than 15 seconds were selected for each speaker and the listener
response behavior (as defined earlier in this paper) during these long turns was
examined. The total occurrence of back-channel response sent by Japanese
toward American speakers was 68 times in 193 seconds, while back-channel
response sent by Americans toward Japanese speakers was 31 times in 184
seconds. In all cases, Japanese listeners sent a back-channel response within a
shorter span of time than did the American listeners. On average, Japanese
listeners sent one listener back-channel response per 3.00 seconds, in contrast
with American listeners who sent one back-channel response per 6.16 seconds.
The types of back-channel responses sent by Japanese and American
listeners were similar, both Japanese and American speakers used brief
utterances and head movements. The major difference was in the frequency
and the discourse contexts in which back channels occurred. Among Japanese
listeners there were more repetitious back-channel responses (as punctuated
vertical head movements repeated up to four times consecutively) and these
back-channel strategies frequently overlapped with the American speaker's
utterance. Back channels sent by Americans were almost exclusively limited to
occurrence during intra-turn pauses.
We observe here that listener response in actual Japanese (nonnative
speakers of English)/American (native speakers of English) intercultural
conversations is conducted in such a way as to reflect the result of CCA based
on native/native interaction. Although it is beyond the scope of this paper, it
should be pointed out that many further questions remain to be addressed in
analyzing native/nonnative interaction; for example, the nonnative compe-
tence level, types of conversation topic, which language is chosen during
interaction, and so on.

10. Concluding remarks

To conclude, this paper has explored CCA - the concept of contrasting

linguistic and interactionai strategies in conversation across speech communi-
ties, based on the analysis of genre and type equivalent data. We also observed
in the preliminary examination that actual intercultural communication be-
tween native/normative speakers is created in such a way as to reflect the
result of CCA based on native/native interaction.
However, as we have seen, the CCA method is not without its problems,
 S.K. Maynard ] Conversation management in contrast 41 !

the most serious and difficult being the issue of 'equivalence'. Although
unresolved issues remain concerning CCA, this methodology provides a
framework within which more careful examination of conversation manage-
ment strategies can be conducted and within which a meaningful contrast can
be made by scrutinizing the actual verbal text produced by speakers in
comparable social situations. As exemplified by the case study reported in this
paper, the CCA approach makes it possible to avoid unnecessarily careless ad
hoc and anecdotal accounts of conversation management in contrast.

References

Azuma, Nobuyuki, 1981. "Gogi no hikaku'. In: Tetsuya Kunihiro, ed., Nichieigo hikaku kooza,
Vol. 3, Imi to goi. Tokyo: Taishuukan. pp. 101-163.
Bernstein, Basil, 1962. Social class, linguistic codes and grammatical elements. Semiotica 23(I/
2): 29-52.
Chafe, Wallace, ed., 1980. The pear stories. Norwood, NJ: Ablex.
Clancy, Patricia, 1982. 'Written and spoken style in Japanese narratives'. In: Deborah Tannen,
ed., Spoken and written language. Norwood, N J: Ablex. pp. 55-76.
Duncan, Starkey and Donald Fiske, 1977. Face-to-face interaction: Research, methods, and
theory. Hillsdale, NJ: Lawrence Erlbaum.
Edmondson, Willis, 1981. Spoken discourse: A model for analysis. London and New York:
Longman.
Fisiak, Jacek, ed. 1980. Theoretical is~aes in contrastive linguistics. Amsterdam: Benjamins.
Gleason, H.A. Jr., 1968. 'Contrastive analysis in discourse structure'. In: James E. Alatis, ed.,
Contrastive linguistics and its pedagogical implications (Georgetown University Round Table).
Washington DC: Georgetown University Press. pp. 39-63.
Grice, Paul, 1975. 'Logic and conversation'. In: Peter Cole and J. Morgan, eds., Syntax and
semantics, Vol. 3. ?~e v York: Academic Press. pp. 41-58.
Hartmann, Reinhard. '978. Contrastive textology in descriptive and applied linguistics. Sophia
Linguistica 4: l-l?.
Hartmann, Reinha:d, 1980. Contrastive textology - Comparative discourse analysis in applied
linguistics. Heide berg: Julius Groos.
Higa, Masanori, 19~2. 'Kaiwa koozoo no hikaku'. In: Te~suya Kunihiro, ed., Nichieigo kikaku
kooza, Vol. 5, Bmka to shakai. Tokyo: Taishuukan. pp. 83-106.
Ikegami, Yoshihiko, 1981a. Gengo no kata to bunka no kata. Gengo 10(12): 36-44.
Ikegami, Yoshihiko, 1981b. Suru to naru no gengogaku. Tokyo: Taishuukan.
James, Carl, 1980. Contrastive analysis. London: Longman.
Kageyama, Taroo, 1981. Nichieigo no kyoozoo kankei. Gengo 10(12): 54-61.
Kenkyuusha, Ed, 1978. Nichieigo no hikaku. Gendai no eigo kyooiku 8. Tokyo: Kenkyuusha.
Kleinjans, Everett, 1958. A descriptive-comparative study predicting interference for .!apanese in
learning English noun-head modification patterns. Tokyo: Taishuukan.
Kokusai Kooryuu Kikin, 1976. Nihonjin no hassoo kara eigo no hyoogen e. Tokyo: Kenkyuusha.
Kunihiro, Tetsuya, 1974. Nichieigo hyoogen taikei no hikaku. Gengo Seikatsu 3: 46-52.
Kuno, Susumu, 1982. 'Danwa no koozoo: Nichi eigo'. In: Kooza nihongogaku, Gaikokugo to no
taishoo, Vol 12. Tokyo: Meiji Shoin. pp. 120-154.
Lado, Robert, 1957. Linguistics across cultures. Ann Arbor, MI: University of Michigan Press.
Maynard, Senko K., 1983. 'Flow of discourse and linguistic manipulation: Functions and
constraints of the Japanese and English relative clause in discourse'. In: Shiro Hattori and
Kazuko Inoue, eds., Proceedings of the 13th International Congress of Linguists. pp. 1028-1031.
412 S.K. Maynard/Conversation management in contrast

Maynard, Senko K., 1986. On back-channel behavior in Japanese and English casual conversa-
tion. Linguistics 24(6): !079-1108.
Maynard, Senko K., 1987. Interactional functions of a nonverbal sign: Head movement in
Japanese dyadic conversation. Journal of Pragmatics 1 ! : 589-606.
Maynard, Senko K., 1989. Japanese conversation: Self-contextualization through structure and
interactional management. Norwood, N J: Ablex.
Miller, Laura, 1987. Listening behavior in conversations bctwcen Japanese and Americans. Paper
presented at the International Pragmatics Conference held in Antwerp, Belgium.
Mizutani, Nobuko, 1983~ "Aizuchi to ootoo'. In: Osamu Mizutani, ed., Hanashi kotoba no
hyoogen. Tokyo: Chikuma Shoboo. pp. 37-44.
Mizutani, Nobuko, 1984. 'Nihongo kyooiku to hanashi kotoba no jittai: Aizuchi no bunseki'. In:
Kindaichi Haruhiko hakase koki kinen ronbunshuu. Tokyo: Kuroshio Shuppan. pp. 261-279.
Monane, Tazuko A. and Lawrence W. Rogers, 1977. Cognitive features of Japanese language and
culture and their implications for language teaching. Proceedings of the 2nd HATJ-UH
Conference on Japanese Language and Linguistics. pp. 129-137.
Pomerantz, Anita, 1984. 'Agreeing and disagreeing with assessments: Some features of preferred/
dispreferred turn shapes'. In: J. Maxwell Atkinson and John Heritage, eds., Structures of social
action. Cambridge: Cambridge University Press. pp. 57-101.
Schegloff, Emanuel and Harvey Sacks, 1973. Opening up closings. Semiotica 8: 289-327.
Schenkein, Jim, 1972. Towards the analysis of natural conversation and the sense of heheh.
Semiotica 6: 3--~4--377.
Stubbs, Michael, 1983. Discourse analysis. Chicago, IL: University of Chicago Press.
Tannen~ Deborah, 1984. Conversational style: Analyzing talk among friends. Norwood, NJ:
Ablex.
Uyeno, Tazuko, 1971. A study of Japanese modality - A performative analysis of sentence
particles. Unpublished dissertation. University of Michigan.
Ventola, Eija, 1980. 'Conversation management from a contrastive point of view'. In: Kari
Sajavaara and Jaakko Lehtonen, eds., Papers in discourse and contrastive discourse analysis.
Jyv/iskyl/i Cross-Cultural Studies 6: 109-144.
Wardhaugh, Ronald, 1970. The contrastive analysis hypothesis. TESOL Quarterly 4: i 23-130.
Yngve, Victor H., 1970. On getting a word in edgewise. Chicago Linguistics Society 6: 567-578.
ZudatiB, Woifgang, 1982. Text type oriented contrastive linguistics and its implications for
translation pedagogy at university level. International Review of Applied Linguistics
20(3): 175-191.