INTRODUCTION
Studies of the effects of gender and social dominance on
conversational interaction in English have suggested that the
domination of male over female is manifested in various ways in this
context as a micro-institution, just as it is manifested in
macro-institutions (e.g., DeFrancisco 1998,
West & Zimmerman 1983, West & Garcia 1988, Woods
1988, Zimmerman & West 1975).
Studies of gender dominance have identified interactional features that
are indicative of conversational dominance, such as overlap,
interruption, topic change, turn length, and distribution of number of
turns, and researchers have conducted quantitative analyses of the
distribution of such features. For example, Zimmerman & West 1975, 1983 found that,
although the distributions of overlaps and interruptions were more or
less equal in single-sex conversations, in mixed sex-conversations all
the overlaps and 98% of the interruptions were made by male speakers
(for critiques of Zimmerman & West, see Murray
1998, Murray & Covelli 1988). West & Garcia 1988 analyzed topic change in
conversational data from mixed-sex dyads of university students and found
that only male speakers initiated unilateral topic changes. DeFrancisco's
study (1998) of mixed-gender conversations from
seven married couples shows that the men interrupted and withdrew responses
more frequently than women did (see also Fishman
1983).
The ways in which research into gender dominance in conversation has
been conducted are problematic for the following reasons. First, in
these studies inequality in power and social status between men and
women is seen to be reflected in the distributional analyses of
interactional features of conversational dominance. Interactional
features of a particular type are treated as a sign of dominance, and
they are aggregated without considering the specific contexts in which
they occur. However, interruptions, for example, can be of different
types with varying degrees of control. An interruption can be resisted
by the current speaker who holds onto his or her turn-at-talk,
resulting in the interrupter's yielding the floor. This kind of
interruption is different from one that successfully causes the current
speaker to yield the floor. Similarly, although questions can be
considered as controlling in the sense that they constrain what counts
as an appropriate subsequent contribution, the degree to which they
control the interlocutor's following contribution is different
between questions that obtain a response and those that fail to do so
(for a discussion of signs of dominance, see Itakura 2001a, 2001b; Linell et al. 1988). Identification of the interactional
features of dominance, therefore, needs to be revised so as to be more
context-sensitive.
Second, interactional features identified as signs of dominance can
in fact be motivated by different conversational goals. For example, an
interruption can usurp the current speaker's turn-at-talk and
therefore can indicate dominance, but it can also reflect a highly
participatory style in which the interrupter's conversational goal
is to be a supportive conversationalist (see Coates
1997, James & Clarke 1993, Tannen
1986, 1994). Murray
& Covelli's study (1988) on interruptions
and gender in different contexts also suggests that interruption may be
related more strongly to conversational styles (“behavioural
repertoire”) rather than to dominance.
The intersection of conversational style and gender difference has
been extensively studied by Tannen 1986,
1990, 1994 (see
also Coates 1996, 1997). Tannen claims that women's
conversational style is characterized by their tendency to speak and
hear a language in terms of “connection and intimacy,”
while men tend to speak and hear a language in terms of “status
and independence” (Tannen 1990:42).
Hence, females try to establish rapport by emphasizing similarities and
matching experiences, whereas males tend to assert their independence
and status by “exhibiting knowledge and skill, and by holding
center stage through verbal performance such as story-telling, joking,
or imparting information” (Tannen
1990:77). Moreover, males tend to be more oriented to themselves
and to downplay the concerns expressed by their interlocutors. They
tend to lead rather than support the conversation. Style differences
seem to be related to differing use of the same interactional features.
For example, Coates 1996, 1997 writes that women tend to use
“other-oriented” questions (referred to as
“addressee-oriented” questions by Holmes 1984, 1995), which
include questions used to invite others into talk, to extend the topic
under discussion, and to avoid playing the expert. In contrast, men
tend to ask speaker-oriented questions that are used to seek
information, to demonstrate power and expertise, and to introduce new
topics (on conversational style and orientation, see also Hirschman
1973, 1994; Tannen
1994).
More recent studies on gender and language suggest that gender
dominance is a social construction (Baron &
Kotthoff 2002, Kiesling 1998, Wodak 1997, West &
Fenstermaker 2002). They reveal how it is achieved, maintained,
and reproduced by speakers by taking full account of its context.
Quantitative analyses alone, therefore, cannot reveal the complex
relationship between conversational dominance and style and associated
features. Kiesling 1998, a study of a meeting
of an American college fraternity, is interesting in this respect
because it demonstrates how quantitative and qualitative analyses of a
social variable can be usefully combined in order to understand how
men's identities are constructed.1
Kiesling's quantitative analyses captured different
patterns in the pronunciation of -ing among the male
graduates seeking to join an American fraternity club: While some
members showed a higher rate of use of vernacular variant
-in' in formal weekly meetings than in informal social
gatherings, others showed the opposite tendency. Kiesling's qualitative
analyses revealed that although all of them wanted to impress the seniors by
the way they talked, those from a working-class background tended to use the
vernacular -in' more in formal meetings to demonstrate their
values associated with hard work, physical power, and camaraderie, whereas
those from a middle-class background tended to use the standard
-ing more in formal meetings to display their identity based on
their socio-economic and structural power.
A fuller picture of gender dominance in conversation, therefore,
needs refinement of the analytical framework for its quantitative
analysis. Findings of quantitative analysis need to be triangulated and
interpreted in light of qualitative analysis of conversational styles
and orientations. The questions that this article addresses are, first,
in what ways can the operationalization of conversational dominance be
refined to take into account the specific contexts in which the
interactional features occur? Second, how far can quantitative analyses
of these features capture conversational dominance, and what insights
can qualitative analyses of the same interactional features provide?
To address these questions, the study reported here uses as data for
investigation conversations between Japanese men and women, because
gender plays a central role in determining relationships in Japanese
culture (Ide 1992. Okamoto
1995, Uchida 1997). The data set
consists of conversations produced by eight mixed-gender dyads of
Japanese university students. We will focus on two dimensions of
conversational dominance: The first relates to the nature of the turns
taken and is referred to as “sequential dominance”; and the
second relates to participation and is referred to as
“participatory dominance.” (For a detailed explanation of
these two dimensions, see the following sections.) We propose a
framework of analysis that takes into account the specific context in
which the interactional features occur in these two dimensions. A
quantitative analysis of sequential dominance showed no significant
male dominance over female, but significant female dominance over male
in one of the eight dyads. As for participatory dominance, male
dominance over female was observed in two of the dyads, and female
dominance in one of the dyads. A qualitative analysis of two dyads in
each dimension revealed that the male participants in these dyads
consistently showed a strong self-orientation in conversation, and the
female participants a strong other-orientation (see also Uchida 1992). It also showed that interaction
between the male participants' self-orientation and the female
participants' other-orientation contributed to male dominance (see
also DeFrancisco 1998, Kendall & Tannen 1997).
The findings suggest that conversational dominance is constructed
through interaction between the different conversational orientations
and styles of male and female, leading to its complex and multifaceted
nature, and that its interactional features are situated in highly
specific contexts. Quantitative analyses of the relative distribution
of interactional features should be interpreted in the conversational
context of which the interactional features are a part. They should
also be complemented by qualitative analyses of these features that
consider the larger social and cultural context that shapes and is
shaped by conversation.
FRAMEWORK OF ANALYSIS
The term “dominance” in this article refers to manifest
(observable) dimensions or structures in dialogue and discourse (see
Linell 1990, Linell et al.
1988). The framework proposed in this study draws on Linell et al. 1988, who observe that dominance is
multidimensional because there are many ways in which a party can be
considered to “dominate” in the shared construction of
conversation, in the sense of possessing or controlling the
“territory” to be shared by the communicating parties (1988: 158). Conversational dominance in the present
study is therefore seen from the point of view of the interactional effect
of controlling action during conversation, rather than that of speaker
intent (Tannen 1994, Gunnarsson
1997). Unlike Linell and his collaborators, we distinguish between
attempts to perform controlling actions that are successful and those that
are not. Only the former would be considered as indicative of conversational
dominance.
This study focuses on interactional features that are commonly
identified as indicative of conversational dominance, such as
initiations, overlaps, and interruptions. We will subsume them under
two dimensions, sequential dominance and participatory dominance,
explained below (see also Itakura 2001b).
Sequential dominance
The term “sequential dominance” refers to one
speaker's tendency to control the other speaker with respect to
the direction of the interaction. This is evidenced by how the speakers
share initiating and responding moves, and by how often a speaker makes
initiating moves, because a speaker who makes an initiation determines
what constitutes an appropriate response (see Schegloff & Sacks 1973, Coulthard & Brazil 1981, Rommetveit 1987). As Linell argues, “Asking
questions is a matter of trying to condition the other's
contributions more or less strongly, whereas simply answering questions
may amount to little more than just complying with the other's
conditions” (1990:153). Moreover, as Linell
further points out, although initiatives and responses are local events,
the aggregated patterns of initiatives and responses emerging over sequences
of utterances will indicate the dominant party in the conversation. The
significance of initiations for topic determination and conversational
dominance has also been noted by Sacks as follows: “The person
who is asking the questions seems to have first rights to perform an
operation on the set of answers. You can call it ‘draw a
conclusion’ … As long as one is in the position of doing
the questions, then in part they have control of the
conversation” (1992a:54).
In this study, sequential dominance is defined as an overall pattern
in which one speaker tends to control the direction of the
conversation. This is measured by the way in which the speakers share
initiating and responding moves. Following Sinclair
& Coulthard 1975, we define each turn as one or more than
one move. Hence, a speaker who responds to a preceding question and
then asks a question in the same turn will be performing two moves.
In investigating sequential dominance, we propose the concept of
“controlling action” (see Linell et al.
1988, Linell 1990, Linell
& Luckmann 1991), which is defined as an action that has the
effect of restricting the other speaker's contribution to the
development of the conversation (referred to as “sequential
controlling action” or SCA). If an initiation is followed by a
complying action – that is, the next speaker produces a
“prospected response” (see Sinclair
& Coulthard 1975, Tsui 1994)
– then the first speaker has successfully performed a controlling
action over the next speaker in terms of sequential dominance. We shall
refer to such a sequence as “sequential successful controlling
action” (SSCA). If an initiation is not followed by a complying
action – that is, the next speaker makes another initiation
instead of producing the prospected second pair-part, or challenges the
presupposition in the initiation – then it is treated as an
instance of “sequential attempted controlling action”
(SACA). Instances of SACA are not treated as contributing to dominance
in conversation because they fail to control the other speaker's
conversational actions.
The description of initiations in this paper is based on Tsui 1994, who classifies initiations as
“requestives,” “directives,”
“elicitations,” and “informatives.” Requestives
and directives prospect a nonverbal action from the next speaker; a
requestive offers the option of complying or not complying, whereas a
directive does not. Elicitations subsume all utterances that elicit a
verbal response. Elicitations can be in the form of interrogatives,
tag-questions, or declaratives. Informatives include all utterances
that prospect an acknowledgment or a comment. They subsume reports
(Drew 1984) and assessments (Pomerantz 1978, 1984), which can be
positive or negative evaluations of the speaker, the other speaker, or
a third party (see Tsui 1991, 1994).
Example (1) is an excerpt from the data that illustrates instances of
SSCA and SACA:
In turn 17, M3 produces an informative in which he makes a negative
evaluation of his private student. This is acknowledged by F3 in turn
18 by echoing the evaluation. Turn 17 is therefore an SSCA. Turn 19 is
also an SSCA because in turn 20, F3's tag question is spoken with
falling intonation and functions as an acknowledgment of M3's
negative evaluation by upgrading not understanding “they”
to not understanding “anything.” This is a common type of
response to evaluative statements (see Pomerantz
1984). That F3's response was prospected can be seen from
M3's total agreement with her upgrade. Turn 22 is a SACA because
F3's assessment of the student as knowing only how to say ‘My
name is …’ in English is challenged by M3, who points out
that he does not even know that. Hence, she fails to elicit a prospected
response that agrees with her statement.
Dominance on the sequential dimension is measured by comparing the
proportions of SSCAs achieved out of the total number of SCAs performed
by the male and the female speakers. If the male speaker achieved a
significantly higher proportion of SSCAs than the female speaker, he
would be considered to have dominated her.
Participatory dominance
The term “participatory dominance” refers to one
speaker's tendency to control the other speaker's
participation in interaction by restricting his or her speaking rights,
in particular through interruptions, overlaps, and completion offers
(Grimshaw 1987; these are referred to as
“appendors” by Sacks 1992a, 1992b).
Following Zimmerman & West 1975 and
West & Zimmerman 1983, we define
“interruption” here as simultaneous speech that results
when a second speaker begins to speak in the middle of a current
speaker's utterance, and not at the first possible completion
point (or transition-relevant place). “Overlap” denotes
simultaneous speech that results when a speaker begins to speak at or
very close to a transition-relevant place in a current speaker's
utterance. Although interruptions are often viewed as violations of the
turn-taking rule whereas overlaps occur at a legitimate place for
speaker alternation, they are similar in that they may affect the other
speaker's participatory rights.
Again, taking into consideration how interruptions, overlaps, and
completion offers are related to their immediate discourse context, we
consider them as indicators of dominance only if the speaker is
successful in maintaining the floor and getting the other speaker to
relinquish it. Using the notion of controlling action, a speaker is
considered to have successfully performed the controlling action in
participation (participatory successful controlling action, or PSCA) if
he or she succeeds in completing a turn-at-talk and in getting the
other speaker to withdraw. If, instead, the speaker who interrupts,
overlaps, or offers a completion of the current speaker's turn has
to relinquish the floor, then this is considered a participatory
attempted controlling action (PACA) and does not contribute to
participatory dominance.
The following excerpt from the data illustrates instances of PSCA and
PACA:
In turn 69, F3 interrupts M3 soon after the first transition-relevant
place. However, instead of withdrawing, M3 continues to complete his
turn (turn 68). Turn 69, therefore, is an instance of PACA. In turn 70,
M3 interrupts F3, also before a possible transition-relevant place, and
F3 relinquishes the floor soon after M3's interruption and M3
completes his turn. Turn 70 is therefore an instance of PSCA. The same
principle applies to overlaps. One speaker may insist on completing the
turn and the other speaker may then withdraw, in which case the former
will have successfully performed a controlling action over the latter.
Alternatively, both of them may hold onto their turns, in which case
neither will have established control over the other.
Participatory dominance is measured by comparing the proportions of
PSCAs achieved out of the total number of PCAs performed by the male
and the female speakers. The speaker who achieves a significantly
higher proportion of PCAs is considered to have dominated the other
party.
DATA COLLECTION
The participants in this study consisted of eight males and eight
females in the same age group, all first-year undergraduate students at
Kanda University of Foreign Languages in Chiba, near Tokyo. The
participants did not know each other well when the recording took place
in June, which was approximately two months after they entered
university. Therefore, there is minimal difference within and across
the dyads in two variables, age and social distance, that may affect
interactional dynamics.
To investigate the effect of gender on conversational interaction,
the participants were grouped into eight mixed-gender dyads. A list of
topics familiar to both participants, such as university life, food,
and part-time jobs, were given to them. The aim was to minimize the
possibility of one speaker's dominating another because he or she
had greater knowledge of the topic (Leet-Pellegrini
1980). They were allowed to choose any topic that they wanted to
talk about in Japanese for approximately 10 minutes. Each dyad was free
to discuss the topic in whatever way they wished. The participants were
not told the purpose of the recording. Each dyadic conversation was
recorded and transcribed verbatim.
Although the conversations were recorded in a quasi-experimental
setting and hence cannot be considered naturally occurring in a strict
sense, the transcripts of the data show that the recording setting did
not seem to cause any unnaturalness in the discourse. In order to
ensure that the data captured their behavior when they were fully
involved in the conversation, approximately the last 100 turns were
selected from each of the 8 dyadic conversations for analysis (the
number ranged from 98 to 102 turns, to make sure that each data set
started and ended with a complete exchange).
RESULTS
Sequential dominance
The results of the analysis of sequential dominance are presented in
Table 1.
Sequential dominance: number of SSCAs and SACAs produced by male and
female speakers in mixed gender dyads.
A total of 211 SCAs (sequential controlling actions) were achieved by
the male participants, among which 154 were SSCAs and 57 were SACAs. A
total of 185 SCAs were performed by the female participants, among
which 131 were SSCAs and 54 were SACAs. In order to see if there was a
significant difference between the proportions of SSCAs and SACAs
achieved by the male and the female participants in each dyad,
Fisher's Exact Tests were performed on the eight dyads. Only Dyad
3 showed a significant difference (p < 0.05) between male
and female, with female achieving 20 SSCAs out of 23 SCAs (87%),
whereas male achieved only 14 SSCAs out of 24 SCAs (58%).
Participatory dominance
The results of the analysis of participatory dominance are presented
in Table 2.
Participatory dominance: number of PSCAs and PACAs produced by male
and female speakers in mixed gender dyads.
A total of 88 PCAs (participatory controlling actions) were performed
by the male participants, among which 36 were PSCAs and 52 were PACAs.
A total of 86 PCAs were performed by the female participants, among
which 30 were PSCAs and 56 were PACAs. Among the eight dyads, Dyads 1,
6, and 8 showed significant differences between males and females in
the proportions of PSCAs and PACAs achieved. In Dyad 1, the male
participant achieved 7 PSCAs out of 11 PCAs (64%), whereas the female
participant did not achieve any PSCA out of 9 PCAs (0%). In Dyad 6, the
male participant achieved 8 PSCAs out of 12 PCAs (67%), whereas the
female participant achieved only one PSCA out of 12 PCAs (8%). Both
Dyads 1 and 6 showed significant participatory domination of male over
female (p < 0.00 for the former and p < 0.01
for the latter). Dyad 8, however, showed significant female dominance
over male (p < 0.05). In Dyad 8, the male participant
achieved 2 PSCAs out of 11 PCAs (18%), whereas the female participant
achieved 8 PSCAs out of 11 PCAs (73%).
In the following section, we will conduct a qualitative analysis of
the interactional features in three of these four dyads to explore the
nature of gender dominance, or what underlies dominance identified
through quantitative analysis. For sequential dominance, we will
examine Dyad 3 because the statistical tests found that in this dyad, F
was the dominant speaker. Although it would have been interesting to
compare this dyad with a dyad that showed the opposite direction of
gender dominance, we were unable to do so because there was no dyad
that showed significant male dominance. Dyad 3 will therefore be
compared with Dyad 6 as an example of dyad that did not show any clear
male dominance. For participatory dominance, we will conduct a
qualitative analysis on two dyads, 6 and 8, in order to compare male
dominance in the former with female dominance in the latter.
QUALITATIVE ANALYSIS
Sequential dominance: Dyad 3
In this dyad, M3 produced 14 SACAs and F3 produced 20 SACAs. For M3,
out of the 14 initiations, only one is an elicitation asking for
agreement; the rest are informatives. For F3, out of the 20
initiations, 9 are informatives, 10 are elicitations asking for
information, agreement, and confirmation, and one provides a
suggestion. All instances of M3's informatives are about a
secondary student whom he was giving private tuition and whom he
describes as ‘super weak’ and unable to handle mathematical
problems. These informatives receive supportive responses from F3. The
following excerpt from turns 3 to 36 is typical of the interaction:
We can see from (3) that F3 plays a supportive role by echoing
M3's informatives (turn 18), upgrading M3's negative
evaluation of his student as ‘super weak’ and not motivated
to ‘hopeless’ (mazui yoo) (turn 10) and
‘incredible’ (Shinjirannai!) (turn 24), and
empathizing, ‘You've got a real problem student, haven't
you?’ (Sugoi seito motchatta nee) (turn 28), and ‘Oh
my goodness!’ (‘Kuutt!’) (turn 36).
The 20 initiations made by F3 are almost equally divided into
informatives and elicitations. An analysis of the informatives produced
by F3 shows that they are very different from those made by M3. All of
the informatives made by M3 are self-oriented in the sense that they
are on the topic that he wants to talk about, his private student and
how weak he is. In contrast, out of the nine informatives produced by
F3 that elicit acknowledgment or comments from M3, six express her lack
of confidence in herself in giving private tuition; in other words, six
of her informatives are self-denigrations. For example, in (2) above,
turn 67, F3 expresses her lack of self-confidence to be a private
tutor. Instead of responding with a positive evaluation of F3, which is
a preferred response to a self-denigration (Pomerantz 1984, Tsui
1994), M3 affirms that F3's negative evaluation of herself
is well founded. F3's diffidence stands in strong contrast to
M3's self-confidence expressed in turn 70, to which F3 responds
with an oh-receipt acknowledgment (Drew
1984) and an agreement. The only instance where F3 makes a
positive evaluation of herself is in turn 46 in (4):
In turn 46, F3 is pleased with herself for being able to get the
right answer to a mathematical problem. However, this was not responded
to with a positive evaluation from M3, but simply with a reaffirmation
of the answer she gave, followed by a description of how his student
handled the problem.
The ten elicitations made by F3 consist mostly of elicitations that
ask for more information about M3's job as a private tutor –
such as where he found the job, the average rate paid for private
tuition, whether the job was difficult for him – and those that
ask for confirmation of her interpretation of the information provided
by M3, or seek agreement from M3 with her opinion. Excerpt (4) contains
four elicitations that are typical of the interaction. In turn 38, on
the basis of the information provided by M3 about his private student,
F3 makes an evaluation of the student and invites M3 to agree with it.
In turns 40, 44, and 50, she asks for confirmation from M3. In all four
instances, M3 is the source of information and hence authority.
F3's lack of self-confidence can be seen best in her deference to
M3 for a verdict in turn 44.
To summarize, qualitative analysis of sequential dominance in Dyad 3
suggests that, although quantitative analysis of controlling actions on
this dimension shows female dominance, the female speaker's
initiations tend to be informatives that are used to echo the male
speaker's views and empathize with them, or to express lack of
self-confidence, and the elicitations are addressee-oriented questions
or clarification questions used to encourage the male speaker to
elaborate on his story. As for the male speaker, although quantitative
analysis found him to be dominated, most of his informatives are used
to pursue his own storytelling, and his initiations support rather than
contradict F3's self-denigration. The analysis therefore suggests
that M3's initiations reflect self-oriented conversational style,
and F3's initiations other-oriented conversational style.
Sequential dominance: Dyad 6
The quantitative analysis of this dyad shows no significant gender
domination. Their conversation is mostly on the topic of living abroad
and learning English. Prior to the 100 turns that were part of the
corpus analyzed in this study, M6 tells F6 that he lived abroad for
more than two and a half years. F6 then tells M6 about her one-month
visit to Canada. However, in F6's account, she portrays herself in
a negative way, saying that she was not able to ‘speak much
English’ and that she got through an examination and an interview
by sheer luck.
M6 achieves 26 instances of SSCA out of 34 SCAs (76.5%). The SSCAs
consist of 17 informatives, 8 elicitations and one “challenging
move” (Burton 1981) that challenges the
presupposition of the preceding move. An analysis of the SSCA sequences
and their relation to the topic of talk will provide a better
understanding of the nature of the interaction. Let us start with
M6's informatives.
The informatives produced by M6 are mainly on the topic of studying
abroad and learning English, both overseas and in Japan. Through these
informatives, M6 tries to reinforce the claim to expertise that he has
established earlier in the conversation because he has lived abroad
much longer than F6. As the conversation develops, M6's claimed
expertise appears to have become a shared assumption that allows him to
speak with authority on the subject of learning English.
Excerpt (5) is a typical example of the informatives produced by M6
to reinforce his claim to expertise. He establishes his expertise not
only based on his length of stay overseas and his knowledge of English,
but also on his knowledge of universities in Australia and how
difficult it is to graduate from these universities:
Earlier in the conversation, F6 has indicated that she wants to visit
Australia. M6, in turns 84, 87, and 89, reports on university life
there. However, the upshot of these reports (Drew
1984) seems to suggest that it may be too difficult for F6 to
keep up with her study at an Australian university. This can be seen
from the fact that when F6 indicates that she would like to go and
study at an Australia university, M6 ignores her. Instead, he just
carries on with his previous turn by reiterating that it is hard to
graduate from these universities, and he supports his claim by citing
examples of his Japanese friends. In the subsequent development of the
conversation, M6's implicit suggestion is made explicit when F6
says that, given such tough requirements, there would be
‘absolutely no hope’ for her, and M6 agrees.2
Because of space limitations, this part of
the transcript is not presented (for the whole transcript, see Itakura 2001a).
Let us now turn to M6's elicitations. There are eight instances,
and they are used in chains of initiations together with informatives,
which facilitates his claim to expertise in learning English.
In turn 38, M6 criticizes English teaching in Japan by eliciting
F6’s agreement that common expressions such as ‘see
you’ are not taught. This elicitation is linked to his
informative in turn 40 and to his elicitation in turn 42, thereby
forming a chain of initiations in which he claims expertise on English
teaching in Japan. Similar to his use of informatives, M6 uses
elicitations to demonstrate his expertise through his criticisms of how
poorly English is taught in Japan. The vocabulary taught is criticized
as ‘useless' and ‘never used’.
F6 achieves 11 instances of SSCAs out of 16 SCAs (69%), of which four
are informatives and seven are elicitations. Among the seven
elicitations that she produces to which she successfully obtains
responses, three invite M6 to provide further information on a topic
that M6 has introduced. Turn 15 in (7) is an example of this:
Of the remaining elicitations, two ask for confirmation from M6,
which acknowledges his authority with regard to information about
universities overseas and learning English.
As for F6's four informatives that elicit acknowledgment or
comments from M6, one serves to elaborate on the topic that M6 has
introduced, and one is a self-denigration in which she describes
herself as ‘such a hopeless or inadequate, or incapable person
like myself’ (watashi nanka). (In Japanese,
nanka is often used after watashi ‘me’ to
provide the additional meaning of ‘hopeless, inadequate,
incapable’.) When F6's informatives change the topic, M6
shows a lack of interest, and subsequently F6 returns to the topic
previously introduced by M6. For example:
After having listened to M6's detailed description of his class,
F6 attempts to describe her own class in turns 24 and 27. However, her
attempt to shift the topic from M6's class to F6's class is
unsuccessful because M6 expresses a lack of interest in listening to
F6's report in turn 25 and produces a minimal acknowledgment with
no elaboration in turn 28. As in turn 88 in (5), F6 is unsuccessful in
developing the topic concerning herself and her experience. However, as
soon as F6 changes the topic back to the one in which M6 claims
expertise, he responds more enthusiastically and continues to develop
the topic (turns 34–37).
To summarize, the analysis of sequential dominance in Dyad 6 shows
that, by skillfully using his informatives to create a gap between the
two speakers in the amount of experience relevant to the topic under
discussion, the male speaker assumed the role of an expert. His
assumption of this role is supported by the female speaker in different
ways. First, she accepts the male speaker's claim of superior
knowledge and experience with her response moves. Second, she
facilitates further development of the male speaker's topic by
inviting him to provide further information on his topic of talk.
Third, she lowers her own position by self-denigration. The female
speaker's attempt to take control over the topic is inhibited by
the male speaker's lack of interest. Thus, similar to the findings
from Dyad 3, the male speaker's initiations seem to indicate his
self-oriented conversational style, and the female speaker's
initiations indicate other-oriented conversational style, and these
styles seem to interact with each other to accomplish male
dominance.
Participatory Dominance: Dyad 6
In Dyad 6, the male speaker (M6) achieves 8 instances of PSCAs
(participatory successful controlling actions) out of 12 PCAs
(participatory controlling actions) (67%), whereas the female speaker
(F6) achieves only one PSCA out of 12 PCAs (8%). The eight instances of
PSCA by M6 consist mainly of M6's continuing his control over the
topic, either when F6 is about to contradict him or when F6 is
developing the topic. For example, in (5) above, F6 tries to come in at
the first possible completion point of turn 84 to develop her
elicitation for agreement in turn 83. However, M6 holds onto his turn,
and F6 relinquishes the floor. In (7), turn 10, just as F6 is going to
put forth something contrary, her turn is interrupted. In turns 12 and
13, there is an overlap and F6 relinquishes her turn, whereas M6
reiterates the topic in turn 14. F6 goes along with M6's topic
control and produces an acknowledgment followed by an elicitation to
enable M6 to develop the topic further.
When F6's turns are interrupted by M6 or when there is an
overlap, she readily relinquishes the floor and allows M6 to take
control. However, when she tries to interrupt M6 and introduce contrary
views, M6 seldom relinquishes his turn. Consequently, the conversation
develops very much around topics that are oriented to M6 rather than to
F6, allowing him to claim superior knowledge and expertise. It also
puts F6 in a position where the only role she is allowed to play is
that of a subservient supporter to M6.
Participatory dominance: Dyad 8
In Dyad 8, the male speaker (M8) achieves 2 PSCAs out of 11 PCAs
(18%), and the female speaker (F8) achieves 8 PSCAs out of 11 PCAs
(73%), showing female dominance (see Table
2).
The two instances of PSCAs achieved by M8 discourage F8 from
developing her topic either by withholding interest in her topic or by
displaying his superior knowledge or experience. Below is an example of
this:
In turn 3, M8 interrupts F8 to express a lack of interest in
continuing her topic, some of the courses that she is taking (‘I
don't know anything about Anthropology…’).
As for F8's eight PSCAs, five of them are used to show her
interest in her male interlocutor's storytelling through offering
her comments, clarification questions, or addressee-oriented questions
that facilitate expansion of his talk; two are used to express her
self-denigration or inferior experience, and the remaining one attempts
to continue her topic of talk but fails. The collaborative nature of
her interruptions is illustrated in (10):
In turn 88, F8 interrupts M8 and stops M8 from completing his turn,
but her interrupting utterance in fact conveys her interest in his
storytelling. Similarly, her interruption in turn 92, where she asks a
clarification question, helps M8 to develop his story further.
F8's interruption (turn 92), his response (turn 93), and her
follow-up (turn 94) comprise an inserted sequence that holds up
M8's storytelling temporarily but helps him to elaborate on its
details.
In addition to two PSCAs (participatory successful controlling
actions), M8 produces nine PACAs (participatory attempted controlling
actions), where he does not complete his utterance following F8's
interruption. However, in the majority of these cases (6 of 9),
F8's interruptions in fact elicit further information or
clarification of M8's story and therefore play a facilitating
role.
DISCUSSION
If the findings of the quantitative analysis were to be considered
separately from those of the qualitative analysis, the
participants' interactional behavior in most of the dyads might be
described as displaying no gender dominance. However, if we see them in
the context of the entire conversation and the way the features
identified have shaped the development of the conversation, we can
understand the situation differently.
Although the quantitative analysis of sequential dominance conducted
on the eight dyads did not show any obvious pattern of male dominance,
the qualitative analysis of the same dimension revealed self-oriented
conversational style for the male speakers and other-oriented
conversational style for the female speakers. These two different
conversational styles contributed to the asymmetrical relationship
between male and female, with the male speaker playing the dominant
role. The male speakers' self-oriented conversational style on the
dimension of sequential dominance was manifested in their pursuit of
center stage in storytelling, claiming expertise, and gaining the
status of an authority figure. To these ends, they used various kinds
of discourse strategies: using initiating moves (informatives) to
create a gap between their knowledge and experience relevant to the
topic and those of the female speakers; claiming superior experience
relevant to the topic and aligning with a powerful persona; and
agreeing with a female speaker's self-denigration or undermining
her confidence. In addition, the male speakers used informatives to
defend their topic control and preempt attempts to take control by
female speakers. The male speakers' informatives were therefore
used to create a gap between the two speakers in level of knowledge and
amount of experience relevant to topic development, in order to raise
the males' status or lower that of the females. In contrast,
female speakers appeared to use initiating moves mostly to support the
male speakers' storytelling in various ways – for example,
by showing interest, asking for more details, endorsing male
speakers' claim of superior knowledge and expertise, deferring to
their authority, creating opportunities for them to demonstrate their
superior knowledge and expertise, and expressing self-denigration.
Findings from the qualitative analysis of participatory dominance
resemble those from that of sequential dominance: Regardless of the
different directions of gender dominance found in the quantitative
analysis, the male speakers seem to have used interruptions along with
self-oriented conversational styles, and the female speakers along with
other-oriented conversational styles. The male speakers tended to
interrupt the female speakers when the female speakers started to
present a contrary view, and they resumed their storytelling
immediately following the interruption. In contrast, the female
speakers' interruptions mostly supported the male speakers'
storytelling by way of clarification questions or addressee-oriented
questions that encouraged the male speakers to elaborate on their
story, or through comments on it. When the male speakers resumed their
topic after being interrupted, the female speakers readily supported
their resumption of topic. Thus, although the female speakers'
interruptions constrained the male speakers' participatory right
temporarily, they normally served the function of an inserted sequence
that provided the male speakers opportunities to elaborate on their
talk. Female speakers' interruptions also expressed
self-denigration, in which they displayed their supposedly inferior
knowledge and experience relevant to the topic.
In sum, the qualitative analyses on these dimensions of
conversational dominance showed that male speakers' self-oriented
conversational style and female speakers' other-oriented
conversational style are complementary and mutually reinforcing rather
than competing. In other words, male dominance is not something
predetermined and imposed upon female speakers. It is instead mutually
constructed by the two parties.
The study also confirms the importance of integrating qualitative
analysis and quantitative analysis. Only by doing this are we able to
situate the production of interactional features (see Freed & Greenwood 1996, Uchida 1992) and to uncover intricate relationships
among power, conversational style, and interactional strategies (see
Kiesling 1998). The integrated approach taken
in this study, therefore, opens up the possibility for investigating
different ways in which “doing gender” is accomplished
within specific contexts. This would allow us to see “doing
gender” as a composite of conversational styles and goals, and
the discourse strategies that underlie them, leaving opportunities for
individual speakers to take an active role in using different
components to suit their own gender identity.
CONCLUSION
In this article, we have investigated the phenomenon of gender and
conversational dominance and how it can be described adequately and
interpreted on the basis of conversational data from dyads of mixed
gender. We proposed a framework of analysis based on Linell et
al.'s (1988) notion of controlling action,
and the premise that interactional features are highly situation-specific,
and that they should not be taken as indicative of conversational dominance
without considering their relation to the preceding and subsequent discourse.
We focused on two important dimensions of conversational dominance:
sequential dominance and participatory dominance. The quantitative
analyses of each dimension did not show any obvious tendency of gender
dominance if they were taken separately and not in the context of the
conversational goals of the speakers. However, when they were taken
together and as situated in the specific contexts of the conversations,
there was a clear tendency of male dominance over female. This suggests
that conversational dominance is multifaceted, and that it would be
misleading to just take a single interactional feature and draw
conclusions on the basis of its frequency of occurrence.
The qualitative analysis of three dyads that showed different
patterns of gender dominance presented remarkable similarities. There
was a strong tendency of male dominance over females in all three:
Males played a central role in shaping the development of the
conversation; all the conversations evolved around the interests of the
male speakers; and the female speakers were forced to play a
subservient role and their contributions to the conversation were
trivialized. Such differences can hardly be explained away as mere
differences in conversational style and in the orientation of people
being brought up in different subcultures. The findings we have
reported suggest that it is very important to examine instances of
dominance in the light of how they are related to the rest of the
conversation, how dominance affects the development of the topic of
talk, how it shapes the dynamics of conversational interaction, and
what impact it has on the achievement of conversational goals.
Moreover, they suggest that gender dominance in conversation is
dialectical, jointly constituted by the behaviors of the male and the
female speakers.
