Abstract

Family dinnertime conversations provide us with information on many aspects of language use and gender identification. In a previous study by Ochs and Taylor, data collected on family dinnertime narratives revealed patterns in which the father maintained a position of power in the family, both socially and linguistically. We argue that in a comparison between two families, gender-linked language use patterns will arise in the areas of total talking time, turn taking delegation and violations, and supportive interjections. Furthermore, we propose that some of these differences can be accounted for as a result of different family arrangements. One of the families we studied consists of two high-profile professional parents, while the other family is comprised of the husband and wife in traditional breadwinner/homemaker roles. Indeed, after examining the data sample of taped dinner conversations, significant differences were found in our areas of concentration, with the dual-career family exhibiting less instances of gender-linked language behavior. While we cannot deduce causal factors for the families’ language practices, our findings point to factors such as education level and adherence to the traditional family hierarchy as potentially influencing language use and gender socialization.

Introduction

Ochs and Taylor claim that "family exchanges do not simply exemplify gender relations otherwise shaped by forces outside the family but, rather, are the primordial means for negotiating, maintaining, transforming, and socializing gender identities" (1995:98). They maintain this viewpoint in their study on family dinnertime narratives. There is no doubt that a significant portion of early gender identification and socialization occurs within the realm of the home. Family dinner conversations also provide us with a wealth of information on language practices and behaviors among the genders. Ochs’ and Taylor’s study inspired us to examine the linguistic behaviors of two families in their dinnertime conversations. Specifically, we were interested in how a high-profile, professional couple and their child would compare in conversation behavior with a more "traditional" family, in which the husband is the sole financial provider and the mother is a self-described homemaker. We projected the professional family would exhibit less gender-linked linguistic behavior, while the "traditional" family would adhere to said behavior more rigidly. The behaviors we chose to measure include total talking time, turn taking delegation, and supportive interjections.

Language behavior differences among men and women have been the subject of many linguistic studies. In reviewing West’s and Zimmerman’s research on cross-sex conversations (1975, 1983), Johnson concludes, "Women fare poorly in comparison with men in terms of turn taking, interruptions, and holding the floor" (1997:9). Likewise, Livia discusses the results of many gender-linked linguistic studies in the introduction to her essay analyzing gender construction of fictional butch lesbians, noting "the findings are weighted toward men’s talking more" (1995:248). The notion that men engage in greater talking time and conversation dominance comes as no surprise to modern linguists familiar with the research of previous decades.

In a comprehensive survey reviewing sixty-three studies, James and Drakich (1993) further research gender-linked differences and the significance of formal versus informal circumstances in which conversation takes place. A logical, but incorrect, assumption would be that men carry more conversational influence in the public sphere while women dominate private-sphere conversation. However, as Livia correctly points out, the results of James’ and Drakich’s study show: "Despite the popular image of the nagging wife and her taciturn husband, men tend to talk more, in almost all discourse environments, than women" (1995:249). This would include informal, domestic situations such as the family dinner.

Thus, in addition to accruing the most total talking time, the men and boys in our study would be expected to take more turns, interrupt more frequently, and generally dominate the direction of conversation, based on the findings of previous research. Previous research also supports the expectation for the women and girls in our study to use more supportive statements (Haas, 1979, cited in Hyde, 1996). In our comparison of the two families, however, we anticipated less gender-linked conversation behavior within our dual-career family than in the "traditional" family. We would like to make it clear that our intent was to make a comparison between two families and not to draw any conclusions about the population at large.

Methods

The two families volunteered to take part in this study by signing a consent form and receiving a blank audio cassette with which to tape three of their dinner conversations. We were not present during their dinners, as we did not want to inhibit their discourse. The shortest dinner conversation was twenty minutes long, while the longest was an hour and forty-five minutes. In order to have a uniform data sample, we extracted twenty-minute samples from each tape. Along with listening to and analyzing the data sample, we made transcriptions of the conversation segments.

The first family can be described as a "traditional" family -- the mother, Susan, is a self-described homemaker, and her husband, Larry, maintains a nine-to-five job in a corporate setting. They have been married for eighteen years and have three children: two girls, Erin (age 16) and Stacy (13), and one boy, Ray (10). The second family consists of two high-profile professionals, Lurleen and Clyde Fowler. Their daughter, Leanne, is ten years old. Lurleen and Clyde have been married for eight years; both partners had previously been married. Lurleen has completed the highest level of education in her field, as has Clyde. The subjects were chosen because of their contrasts from one another regarding occupation status, levels of education, and type of marital relationship. All subjects, however, are white and middle class. Although this allows for control of ethnicity and class as independent variables, the results of our study are ultimately limited by this small and homogeneous sample.

All family members were aware they were being recorded. Subjects reported feeling comfortable with the audiotapes and did not feel overly inhibited by the intrusion. In fact, during one conversation, Susan and her family forgot about the tape entirely and did not remember to stop recording until well after dinnertime. Lurleen reported being somewhat embarrassed by her family’s lack of inhibition as she handed us her audio cassette, which included a heated argument involving summer vacation plans.

In analyzing the dinner conversations for instances of sex-stereotyped language use, we observed the frequency in which participants engaged in certain conversation behaviors. The specific behaviors we collected data on include total talking time, turn taking delegation and violation, and supportive interjections. We chose to observe numerical frequencies of conversation behavior rather than evaluating content in order to reduce subjective analysis. However, we have included excerpts from the conversations in this paper to provide examples of our measurement factors and to allow the reader to understand the consistency of these conversations.

Total Talking Time

We measured the amount of time each family member spoke within the twenty-minute time frames of each conversation. Subjects were timed whenever they spoke, including short utterances, humming/singing (behavior exhibited by some of the children), and instances of overlap, which Nofsinger defines as a generic term for any simultaneous talk (1991:101).

Turn-Taking Delegation and Violation

The turn-taking system we applied in studying our families is derived from previous linguistic studies and detailed in Nofsinger (1991). Nofsinger and preceding linguists assume conversation participants can project when another’s turn has ended. This spot,"where a transition from one speaker to another becomes relevant," is referred to as the transition relevance place (Sacks et al., 1978, cited in Nofsinger 1991:81). In conversational turn-taking, the practices that participants use to switch speakers at a given moment operate at the TRP.

For our data, we recorded the frequency of turn-taking allocations. According to Sacks et al., as cited in Nofsinger (1991), there are three possible techniques by which conversation participants determine the next speaker. The first of these techniques is referred to as current speaker selects next (1991:82). This occurs when the current speaker directs a statement or question to a specific participant, as Lurleen does in the following example:

(1) Clyde: He constantly carries with him, uh, uh, an oxygen bottle.

Lurleen: You know what, we saw somebody walking around town yesterday with a big oxygen thing, remember that Leanne?

Leanne: It’s not a tank, mom.

As Lurleen selects the next speaker within her turn -- that is, before the TRP -- Leanne has exclusive rights to become the next speaker.

A second, and more common, technique is the self-select. If the current speaker does not choose the next speaker, any member of the conversation may choose to self-select at the TRP. Nofsinger notes "This rule puts a premium on starting quickly, since several participants may self-select and failure to be first may mean either that one drops out or that one continues with the risk of being judged as ‘butting in’ on someone else’s turn" (1991:83). Lurleen’s response to Clyde in the previous excerpt is an example of self-select.

The last tur- taking allocation technique occurs when the current speaker has neither chosen the next speaker, nor do the other participants opt to self-select. In these instances, when the current speaker continues to speak, extending her/his turn, s/he is engaging in speaker continuation (Nofsinger 1991:85). Lurleen’s final comment in this next segment is an example of speaker continuation:

(2) Clyde: That means spend the money, close the practice. . . whatever it takes.

Lurleen: Have three months of no income?!

Clyde: Yeah!

Lurleen: Where is that money going to come from? What sacrifice are you going to make to get that money?

(0.5 second pause)

You gonna sell the house?!

Nofsinger correctly points out, "The existence of the speaker-continues option is another reason for listeners to self-select quickly, for any delay may result in the current speaker resuming his or her turn" (1991:85).

The notion that participants will want to self-select quickly at the TRP in order to ensure legitimate status as next speaker assumes the basic rule of conversation etiquette that only one speaker should be talking at a time. While there are often instances of overlap, particularly when several speakers opt to self-select at a TRP, participants usually drop out of the conversation to leave only one speaker. If a speaker continues to talk despite another participant’s status as legitimate speaker, then he or she is said to be in violation. Turn taking violation also occurs when a participant self-selects before the TRP, such as Larry’s interruption during Susan’s turn in the following excerpt:

(3) Susan: So, we’ll just fly Alaska up to Seattle, take the cab, take a cab over to the--

[

Larry: Hey, can you guys shut the door when you let the dogs out? You’ve gotta shut it, close it all the way. . .and the little latch ‘cause they can just open it.

In addition to recording the frequency of turn-taking allocation activity among the family members, we also tallied each participants’ turn-taking violations. We did not include the brief overlap which is common at the TRP as turn taking violation. Likewise, we made a distinction between interrupting behavior such as Larry’s in the last example and supportive interjections, which we studied as a separate category.

Supportive Interjections

Our distinction between supportive interjections and turn taking violations which occur before a TRP is based on the intent of the participant engaging in these behaviors. We categorized interrupting comments intended to disrupt the original speaker, change the direction of the conversation, or present opposing views as turn taking violations. Interrupting comments which were not intended to discount the original speaker and, in fact, offered supportive feedback were categorized as supportive interjections. Clyde’s overlapping "vegetable comment" in the following conversation sample is an example of supportive interjection:

(4) Lurleen: Oh, I love that delicatessen! <<Clyde laughs>> It’s so pretty in there with all those roosters and ceramic roosters everywhere, and. . . big bottles of olive oil. And pretty colored vegetables and fruits. It’s just a gorgeous market.

[

Clyde: vegetables, yeah.

In addition, we marked backchannel feedback as supportive interjections, as well.

Results

The following table presents our data. All numbers are averages taken from the three twenty-minute dinner conversation segments. All numbers were rounded to the nearest tenth. The first five entries represent the Zimmerman’s data; the following three are taken from the Fowler’s data.

Discussion

The data we collected overwhelmingly supports our expected outcome that the Fowler family would exhibit less gender-linked linguistic behavior. Lurleen’s average total talking time, ten minutes and fourteen seconds, was more than double the talking time of her husband. Furthermore, Lurleen consistently accumulated the highest numbers overall in terms of self-selecting, choosing the next speaker, continuing after a pause, and violating turn etiquette. In addition to holding the floor for longer periods of time, Lurleen was interrupted less than fellow family members. In other words, when Lurleen took on the storyteller role, Clyde and Leanne listened. However, when Clyde began a story or topic to discuss, Lurleen often cut in to ask questions or offer her comments. This behavior can be seen in the following excerpt in which Clyde begins a story about a man who raised acrobatic pigeons:

(5) Clyde: And then he got into, and then he went off to war, he was in the Vietnam War . . . and then he came back and he, and he got into more sophisticated pigeons that don’t fly when you breed them. And over here, they, they tumble over backwards -- they somersault backwards. And they have contests to decide, to see who can finish the projects fastest before they start walking --

[

Lurleen: Huh? What is the point? What’s the point of a pigeon that doesn’t fly?

Clyde: That’s a good question, I was just telling about that. But he very carefully grubbed the best tumblers. The best boy tumbler, the best girl tumble --

[

Lurleen: Is it a sport?

Clyde: Yeah. He would breed them. And he now holds the world record. He kept breeding them very carefully and now he holds the world record. And one of the sheiks in Qatar -- you know where Qatar is, Leanne? It’s down there by Saudi Arabia -- he wanted to buy one. He was invited on the Da --, the David Letterman show --

[

Lurleen: Oh, the Stupid Pet Tricks, yeah! <<Lurleen laughs>>

Clyde: Yeah! <<Clyde laughs>> And you know what? So he went back . . . and they, they didn’t put him on ‘cause, you know, they went back and a wing kicked up -- a feather came off. And so Letterman was afraid that one of the rights people would say something about it.

[

Lurleen: Oh no! So he didn’t make it?

Clyde: He didn’t make it on.

In this small portion of the Fowler dinner conversation, Clyde’s pigeon story runs for one minute and seven seconds. Within this small time frame, Lurleen makes four turn-taking violations. In contrast, Lurleen holds the floor for a minute and twenty-nine seconds in a later segment of the conversation without a single interruption. In some instances, Lurleen’s violations can be considered supportive of Clyde’s story -- indeed, the general mood of the conversation at this point was lighthearted and Clyde did not seem to take offense to Lurleen’s interruptions. It should be noted, however, that the questions and comments Lurleen poses before the TRP’s dramatically shift the direction of the story. Lurleen’s interruptions give her partial control of how Clyde’s story unfolds.

Lurleen’s high number of total talking time and turn-taking violations is inconsistent with gender-linked expectations based on prior research of women’s language behavior. Larry’s comparable frequencies within the Zimmerman family conversations, however, match gender expectations. In the turn taking violation category, Larry averaged 17.7 violations per twenty-minute dinner conversation sample. This accounted for 41.7% of the family’s accumulated turn taking violations. We noted several instances in which Larry’s interruptions served the function of disciplining/lecturing the children. We did not include this as a separate category in our numerical data as this practice was unique to Larry. The following excerpt provides an example of Larry’s disciplinary interruptions:

(6) Susan: Oh great, I didn’t know she was going to be home from the interview.

Erin: She’s home now.

[

Larry: Stacy! We do not lie on the table and drink! We pick our arms up off  the table!

Stacy: Yes, Sir.

Erin: I got salsa on my sweatshirt.

Susan: That’s why god made washing machines, honey.

In this example, Larry’s authoritative turn-taking violation does not re-route the direction of the conversation as Lurleen’s comments to Clyde had in the example above. Rather, Mr. Zimmerman’s interruption completely stops current discourse. Although Erin and Susan resume their position as legitimate holders of the floor upon completion of Larry’s lecture, the topic they had been discussing is dropped.

Susan often maintained legitimate speaker status, which parallels her high frequency of self-selects and total talking time. This data is inconsistent with our hypothesis in which we projected her adherence to feminine language behavior as described by linguistic analysts. It is important to note, however, that despite Susan’s greater amount of total talking time compared to other members of her family, she does not engage much in the more conversation-dominating behavior, such as current speaker selects next, speaker continuation, or turn taking violations. Alternatively, she scores high in the supportive interjections category -- making 55% of the family’s total supportive interjections. This finding concurs with our expectations.

Susan’s high frequency of supportive interjections coupled with her low score of speaker continuation and turn-taking violations does not necessarily mean her language behavior places her in a subordinate position relative to her husband or other family members. The assumption that supportive interjections indicate passivity would follow a dominance approach in understanding gender differences in linguistic practice -- "which invariably places women in a position of deficit" (Johnson 1997:10). Albrecht and Adelman point out the potential influence of a supportive interjection: "When one person communicates a supportive message to another, that behavior can affect both persons’ feelings and cognitions, the state of the relationship and future message exchange sequences" (1987:20).

Although Susan demonstrated some instances of non-gender-linked language behavior, Lurleen clearly exhibited said behavior to a much greater extent. Clyde’s linguistic data also illustrates a high degree of non-conformity to gender expectations. Note the instances of supportive feedback in the following excerpt:

(7) Lurleen: I mean, every spring the kids come up with a plan to outsmart the grownups. . . last year it was on the track bus, you know when they bought a lot of liquor and they were all drinking.

[

Clyde: Right.

Lurleen: And it’s always the leaders! Last year it was the principal’s son. ‘Cause they think, you know, we’re leaders, we can get away with it.

Clyde: Well they think they’re leaders because they, they want to lead.

[

Lurleen (finishes sentence with Clyde): want to lead

Clyde: One of the, one of the, uh, one of the characteristics of a leader is they figure out what everybody wants and then meet it.

[

Lurleen (finishes sentence with Clyde): and then do it.

(0.45 second pause)

Lurleen: The superintendent does not want her back in the high school because it’s important to the other kids--

[ Clyde: It is!

Lurleen: If you let the, quote, good kids get away with murder, then all the other kids are sooo resentful. So, she’s a symbol. She has to be punished. Probably because of her high visibility.

Clyde: Yeah.

Clyde’s average number of supportive interjections per twenty minute sample (23.7) made up 60.2% of the total support statements of the Fowler family. This statistic is especially noteworthy considering his accumulated talking time was less than half of Lurleen’s. Mr. Fowler’s high ratings in this category may be due to many factors, including his occupational position outside the home which requires astute listening skills on his part. Likewise, his decidedly more egalitarian style of discourse compared to that of Larry’s could potentially be attributed to his higher education level and less-traditional marital relationship.

The data from all four children in our study matches our projections about gender-linked language behavior. In all cases, the children seemed to model their language behavior after their same-sex parent. Thus, Leanne displayed non-gender-linked behavior like her mother, while Ray’s behavior shadowed Larry’s, and Stacy’s and Erin’s behaviors were strikingly similar to Susan’s. An interesting phenomenon which occurred in Susan’s family -- automatic gender segregation in simultaneous conversations -- points to the presence of subtle encouragement for the children to adopt gender-specific behavior. Larry and Ray carried on exclusive conversations with each other throughout the dinner, as did Susan with Erin and Stacy. The only time Stacy spoke during our first Zimmerman conversation sample (other than to say "Yes, Sir," to her father) was in one of these all-female discussions.

While the children of both families demonstrated turn-taking behavior which was consistent with our hypothesis, we believe Leanne and Ray’s conversational behavior may be additionally influenced by their age. Both children are ten, and may not have mastered the subtleties of conversation etiquette yet, as evidenced by Ray’s incessant bids for his mother’s attention -- "Mommy, mommy, mommy" -- while she is engaged in discourse with other family members. Similarly, Leanne violates Clyde and Lurleen’s discussion of their impending summer vacation several times with the question "How do you spell Europe?" Though we are unable to legitimately deduce causal factors for the language behavior of our participants, Leanne’s and Ray’s repetitious questions point toward their unsophisticated comprehension of conversation etiquette.

There is no doubt that a multitude of factors influence discourse. Our study points to some potential factors which may influence gender-linked language behavior in the family setting, such as similarly ranked occupation status among spouses and higher education levels. However, we do not wish to make the reductionistic claim that these are the only variables influencing the participants’ language practices. With economic necessity and more opportunities for women in the workplace, the stay-at-home mom is becoming increasingly rare. Non-gender-linked language may evolve to some degree due to women’s increased visibility in the public domain. However, we acknowledge the slow-changing nature of discourse practices, and we do not foresee a progressive, linear change, as such change is rare in both language and social movements.