Gender, Status and Power in Discourse Behavior of Men and Women
Gender, Status and Power in Discourse Behavior of Men and Women
Peter Kunsmann (Freie Universität Berlin)
In her book Women, Men and Language Coates (19932:3) asks the question whether women and men talk differently at all. She then proceeds to collect evidence for this claim trying to establish in what way these two groups differ, and finally providing explanations for these differences. In this article I will ask the additional question whether the differences are related specifically to gender or, alternatively, to status and power.
The first part will be devoted to an overview of scholarship in the field of male-female speech. It will be followed by a description of some of the linguistic variables that have been studied in the relevant literature, such as question-tags and interruptions. The second part will discuss the two major hypotheses which have been advanced as explanations for the differential use of language by male and female speakers: the dominance and the difference approaches. Finally, the question will be discussed whether gender-specificity or power related factors account for the differences observed.
Fed by the advances of civil rights in the United States during the 1960s and 1970s, the feminist movement also gained momentum. At the same time, sociolinguistics provided mechanisms for the scientific investigation of language variation on the basis of both socio-economic and gender factors. With respect to a number of sociolinguistic factors including gender these studies investigated linguistic features such as phonological variability of male and female differences. The goal, on the one hand, was to determine the stratification of these variables and, on the other hand, to find support for a mechanism of synchronic change. The differential use of these variables was interpreted as constituting a gender pattern. Women were found to be closer to a prestige norm (i.e. RP: received pronunciation) than men. In particular, studies by Martin (1954), the Norwich studies by Trudgill (1972, 1978, 1998) and Portz (1982), and Fasold (1990) provided support for this position. As Martin (1954) put it:
Women, it seems, are considerably more disposed than men to upgrade themselves into the middle-class and less likely to allocate themselves to the working-class – a finding which confirms the common observation that status consciousness is more pronounced among women. (Martin 1954:58)
Additional confirmation comes from self-evaluation studies. The percentages of over-reporting by women and those of under-reporting by men on the (er)-variable in East Anglia, for instance, show a significant difference.
(based on Trudgill 1998:26)
These quantitative studies on the relationship between gender and variable language use are contrasted with qualitative (interpretive) approaches. In her seminal publications Lakoff (1972, 1973, 1975) claimed that the differential use of language needed to be explained in large part on the basis of women’s subordinate social status and the resulting social insecurity. Lakoff observed that women’s use of color terms (mauve, ecru, lavender), of adjectives (divine, adorable), their frequent use of tag-questions (John is here, isn’t he?) and weak expletives (Oh fudge I’ve put the peanut butter in the fridge again!) differed radically from male use. Taking her cue from Bernstein’s (1972) theory of language codes she claimed that women’s linguistic behavior is deficient when contrasted with male speech behavior. As one explanation for this deficiency she pointed to the differences in the socialization of men and women.
At the same time another qualitative approach to male-female speech variation developed (Thorne/Henley 1975, Maltz/Borker 1982). Cultural rather than factors of socialization were seen as being responsible for speech differentiation. Women and men are seen as constituting subgroups of the speech community.
Finally, social constructivist approaches emphasize the notion of “doing gender” which “involve complex socially guided perceptual, interactional, and micropolitical activities” (West/Zimmermann 1991: 13-14). This in turn was the basis for the suggestion to use the category gender as a point of departure for the analysis of variation (Kramarae 1986, Torres 1992).
The quantitative and qualitative approaches were judged differently by various sociolinguistic scholars (cf. the following table):
|Basic question:||What are the social phenomena of language in a speech community?||What are the linguistic phenomena of language in a speech community?|
|Labov||language and society||secular linguistics|
|Trudgill||language in use||linguistics proper|
|Coates/Cameron||(interpretative) paradigm||quantitative paradigm|
One of the major criticisms of quantitative studies regarding the gender variable is related to methodological considerations. In particular, Labov and Trudgill would assign socio-economic status to women on the basis of masculine norms, thus, skewing the interpretation of the data in the direction of the male variable. Coates/Cameron (1989:17) remark that “men’s linguistic behavior is seen as normal; when women’s differs, it has to be explained”. Claims made by the qualitative or interpretative paradigm, on the other hand, do not only include the linguistic and social contexts of the speech situation, but also the cultural and psychological ones.
3 Linguistic Variables
Interactional patterns in same-sex and cross-sex studies provide evidence for the fundamental difference between men’s and women’s linguistic behavior. In this section I will give a short account of question tags and interruptions. Other studies on language use in communication with respect to amount of speech and control of topic show similar patterns for the distribution of variation between men and women.
3.1 Question Tags
Lakoff (1975) observed that, in certain contexts, women use question tags more frequently than men do. She defines the tag-question as
a declarative statement without the assumption that the statement is to be believed by the addressee: one has an out, as with questions. [The] tag gives the addressee leeway, not forcing him [sic] to go along with the views of the speaker. (Lakoff 1975:16)
Furthermore, she claims that downtoning a statement shows lack of confidence. Support for this position comes from those situations in which either verification of the statement can be made by mere inspection: John is here, isn’t he? or where it reflects the opinion of the speaker: The way prices are rising these days is horrendous, isn’t it? Clearly, these sentences need not be questioned and, thus, demonstrate the speaker’s insecurity.
As early as 1975, a study by Dubois/Crouch showed that these observations would not hold up. In fact, in their study men produced more tag-questions than women. Recent evaluative reaction studies also support this position. Bock (1996) conducted a poll of 122 American college students on the question: “Women use more tag-questions than men”. Less than 41% agreed with this statement while 17.2% disagreed and 41% were undecided.
Further analyses reveal additional properties of tag-questions. In addition to expressing uncertainty, insecurity and the wish to be accepted (cf (1)), tag-questions also function as expressions of politeness, as hedging and boosting devices. Moreover, they facilitate communication. In (2) the speaker’s haven’t you gives the addressee, Andrew, a chance to pick up on the topic suggested by the speaker and get into a conversation with Frank.
(1) Showing insecurity: I graduated last year, didn’t I?
(2) Facilitating conversation: Andrew this is our new neighbor, Frank. Andrew has just changed jobs, haven’t you? (Holmes (1992: 318))
For the different functions of the tag-question, Holmes (1992: 319) reported the following results:
|Function of tag||Women||Men|
As can be seen, men use question tags more often to express uncertainty while women use them largely to facilitate communication.
A different division of the function of tag-questions, originally proposed by Holmes (1984), is the basis for a study by Coates/Cameron (1989). On the one hand, Coates/Cameron define an affective function for tags (cf (3)) which are directed toward the addressee and signal solidarity. On the other hand, tag-questions also serve a modal function (cf (4)). As such they are speaker oriented and indicate a request for information or a confirmation of the information.
|(3)||Showing solidarity:||His portraits are quite static by comparison, aren’t they?|
|(4)||Indicating uncertainty:||You were missing last week, weren’t you?|
(Coates/Cameron 1989: 82)
Coates/Cameron’s study of a 45,000-word corpus of the “Survey of English” at University College, London, showed similar results to Holmes’ study. Women use more affective-facilitative tags while men use more modal ones. In contrast to Holmes no softening tags were found. With respect to the function of uncertainty and insecurity a further differention should be considered. Coates/Cameron assume an uncertainty on the part of the speaker for the modal function. However, as the contrast with (1) demonstrates, the uncertainty in (4) concerns the information and not the speaker.
|Function of tag||Women||Men|
|Total Number of Tags||36||60|
(based on Coates/Cameron 1989:85)
3.2 Interruptions, Overlaps and Minimal Responses
Interruptions are generally considered to be violations of the rules of conversation. According to Sacks/Schegloff/Jefferson’s (1974) model of the structure of conversation, turns of speech are assigned such that the current speaker has the largest options. Particularly in cultures such as the United States, furthermore, it is important that the gap between turns be kept short. This may lead to overlaps at the end of the first speaker’s turn and the beginning of the next speaker’s turn. By observing the no gap-rule overlaps in conversations, therefore, are generally considered as facilitating. Finally, another facilitating strategy is the use of minimal responses. During the turn of the first speaker the addressee will provide agreement or encouragement through short interjections.
Given the interactional situation the relationship between interruptions, overlaps and minimal responses is a gradual one. Therefore, it will be necessary to distinguish them on a scalar dimension. The three categories may be delineated along such a scale representing five different aspects of turn-taking:
1. Outright interruptions
2. Overlaps in which the second speaker takes the floor by default (i.e. based on an ensuing silence of the first speaker)
3. Overlaps that allow for a soft transition between the first and second speaker
4. Overlaps at the end of the first speaker’s turn that are supportive and may encourage the speaker to continue
5. Minimal responses during a turn
Early studies on interruptions and related phenomena seem to indicate a larger tendency on the part of men to interrupt in cross-sex conversations while in same-sex conversations no significant differences were found. Zimmerman/West (1975) reported the following results:
|Same-sex conversations||1st Speaker||2nd Speaker||Total|
(based on Zimmerman/West 1975: 115-116)
The studies on question tags and interruptions, on amount of speech and control of topic in conversations amount to a considerable body of research regarding the differential use of language by male and female speakers. Where the studies differed in the observation and interpretation of the data, significant changes of method resulted. Question tags were seen as constituting different functions depending on context, and interruptions had to be defined more carefully. Overall, the studies support Lakoff’s initial observation that status is a major component in the explanation of male-female differences.
4 Dominance and Difference
As we have seen in the studies presented in Section 3, women and men behave differently in a speech situation. This difference manifests itself, for instance, in linguistic behavior by the differential use of question tags, and in communicative behavior by the use of interruptions. How can this variability be explained? Scholars differ in their interpretation. Two hypotheses, however, have been suggested in the literature: the dominance and the difference approach.
The participants in a conversation use a number of strategies to achieve their conversational goals. One of these goals may be to dominate other participants of the speech situation. The question whether gender or status and power is the motivating force for conversational behavior has been resolved in favor of status and power in the literature. Most studies find that in mixed talks men tend to be more dominating than women.
One of the obvious strategies for achieving this goal, as we have seen, is the use of interruptions. Their use is generally explained by the relative power of the participants which derives from their social status. The higher incidence of interruptions, thus, is seen in the relatively high social and economic status of men. Women, on the other hand, are powerless regarding their social position. This is reflected in fewer interruptions in cross-sex conversations. Similarly, as Lakoff (1975), Trudgill (1978) and others have pointed out, low social status is often characterized by passivity and low vitality. This in turn results in the wish to be accepted by the dominating group. The verbal expressions of this accommodating behavior are, among others, tag-question. Nevertheless, personality differences will have to be considered as well. Individual subjects will react differently in certain situations. In addition, maleness and femaleness are not discrete categories. While it is true that statistical means show specific features for men and women (cf. Maltz/Borker 1998: 418-419) standard deviations can be fairly large, resulting in overlap. Indeed, psychological tests of college students in the United States found such an overlap between women and men on a scale of masculinity.
The difference approach attempts to explain the differential communicative behavior of men and women by assuming two subcultures in the speech community: men and women. In these different subcultures separate linguistic strategies for interactional behavior are acquired.
Starting with the analysis of issues that are relevant for miscommunication, Maltz/Borker (1998) point to the different rules which govern the behavior of the two subcultures. Misunderstandings can arise, for example, from the differential use of minimal responses.
Imagine a male speaker who is receiving repeated nods or “mm hmm”s from the woman he is speaking to. She is merely indicating that she is listening, but he thinks she is agreeing with everything he says. Now imagine a female speaker who is receiving only occasional nods and “mm hmm”s from the man she is speaking to. He is indicating that he doesn’t always agree; she thinks he isn’t always listening. (Maltz/Borker 1998: 422)
From this observation, Maltz/Borker conclude that the differences between men’s and women’s speech can be explained using an anthropological approach in the study of “culture and social organization.”
Holmes (1998) extends this approach to formulate a set of sociolinguistic universals. Among these are:
1. Women and men develop different patterns of language use. (1998: 462)
2. Women tend to focus on the affective functions of an interaction more often than men do. (1998: 463)
3. Women tend to use linguistic devices that stress solidarity more often than men do. (1998: 468)
4. Women tend to interact in ways which will maintain and increase solidarity. (1998: 472)
5. Women are stylistically more flexible than men. (1998: 475)
Whereas Maltz/Borker and Holmes see the difference approach from a cultural point of view, Chambers (1992) gives a biological explanation. Claiming an innate, albeit small, neurological advantage for women, Chambers assumes that this advantage is realized in the use of verbal skills and transferred to other behavioral skills. Using data from studies in Detroit and Belfast, from Japan and the Middle East, Chambers argues for a sex-based analysis of variability. Although pointing to the tentative nature of this explanation he claims that
female precocity in verbal skills beginning in infancy predisposes them to apply their verbal skills to all kinds of situations as they grow up. (Chambers 1992: 201)
Furthermore, Chambers cites Sherman (1978) in support of his position:
[T]he early female advantage bends the twig toward female preference for verbal approaches to problem solution. This bent is then increased by the verbal emphasis of the educational system and by aspects of sex roles that do not encourage girls’ develeopment of visual-spatial skills. (Sherman 1978: 40)
Indeed, whether due to biological or factors of socialization, results of college entrance examinations (i.e. the verbal part of the scholastic aptitute test) or other verbal measures in the United States have continuously shown higher verbal scores for women of this age group.
4.3 A Note on Ethnolinguistic Vitality
Women’s language was described as weak, unassertive, tentative, and women were presented as losers, as victims. (Coates 1998:413)
In recent years a line of inquiry has developed that analyzes the relationship between language and social identity. Giles and others in the 1970s, Gumperz (1982) and Tajfel (1974) formulated a theoretical base for this analysis: the ethnolinguistic identity theory. Given the notion of a male and female subculture we can regard women as the minority group and men as the majority group.
A major factor in the theory is the vitality of a speech community. Accordingly, a group’s vitality rests on three main features: its distinctiveness, its social activity and its solidarity when confronted by another group. When all three features are strongly represented the group’s vitality will be high. Low vitality leads to the characteristics described by Coates above.
Status, demography and institutional support are the three causal factors for measuring vitality. Demographic considerations are not applicable to male-female analyses. Status and insitutional support, however, are. While social, economic, and sociohistorical status seem to favor male vitality, language status can be viewed from two sides. On the one hand, we have seen that RP and other standard varieties have a high prestige value. Since women are closer to this norm it is possible to derive higher vitality from this factor. On the other hand, in the literature men are seen as using a dominant style. For instance, in reviewing Tannen’s (1990) publication Trömel-Plötz (1998) argues against Tannen’s failure to discuss “concepts like dominance, control, power, politics of gender, sexism, discrimination” and points to her own work in which she found the “gender hierarchy stronger than the hierarchy created by social status.”
Men, the speakers of the dominant style, have more rights and privileges. They exhibit their privileges and produce them in every conversational situation. (Trömel-Plötz 1998: 447).
Nevertheless, style in itself is a function of non-linguistic status factors. Thus, linguistic factors remain functional for the potetial increase of women’s status. In addition, institutional support, though still largely the domain of male activity, can be the source of increasing the vitality of women. In particular informal, affective support may be seen as a source.
5 Language in Relation to Gender, Status and Power
Gender related variability seems to be associated with the difference approach, while the dominance approach is supported mainly by variability on the basis of power. Power, on the other hand, is derived from social, economic and sociohistorical status. The question whether variability must be explained in terms of gender or status, however, cannot be answered conclusively on the basis of the literature discussed so far. Most studies cited here seem to favor the dominance approach. Nevertheless a number of studies cast doubt on this conclusion.
5.1 The Use of Questions as a Controlling Strategy
Zimmerman/West (1975) and others state that
just as male dominance is exhibited through male control of macro-institutions in society, it is also exhibited through male control of at least a part of micro-institutions. (Zimmerman/West 1975: 125)
As we have seen, one of the controlling mechanisms in micro-institutions is related to the strategy of interrupting. As men are interrupting more often than women, male dominance can be established in conversations. Thus, turns are claimed, topics are initiated and maintained by men or abandoned by women.
In some cultures, on the other hand, questions may also be used as controlling mechanisms. Similar to the pressure by the no-gap-rule mentioned above exerted on the participants in a speech situatuation in these cultures, questions require answers in many conversational situations. When questions in form of facilitative rather than polite or modal tag-questions, therefore, are combined with a specific statement they can be used to maintain or to control the direction of the conversation. As women use this type of question more often than men, female dominance can be established.
5.2 Interruptions Revisited
Eakins/Eakins (1978) report that status seems to be a factor in the pattern of interruptions. While males initiated more interruptions than females in their study of faculty meetings there was a clear ranking along status lines. The chair of the department, for instance, suffered the least number of interruptions. Nevertheless, the most interrupted person was a woman.
Holmes (1992), on the other hand, found that in doctor-patient conversations female doctors were interrupted more often than male physicians. In addition, in business organisations, men but not women tended to dominate the interactions.
West (1998) came to similar conclusions in her study of interaction between doctors and patients. Her findings contrast with those of the earlier study (Zimmerman/West 1975). Here, it is clear that status alone can not account for the results. Female physicians were interrupted more often by patients of all social status groups than male physicians.
|Gender of Physicians||Number of Interruptions||Physicians’ Interruptions||Patients’ Interruptions||Number of Patients|
(based on West 1998:400)
A study by Beattie (1981) examining the relationship between status, gender and number of interruptions came to quite different conclusions. Analyzing the extent to which interruptions account for speaker switches in ten tutorials at Sheffield University he found that more than one third of changes in turns resulted from interruptions. In addititon, status seemed to affect interruptions inversely while gender had no significant effect. While both results may be accounted for by the specific conditions of the study they give further support to the supposition that gender as well as status need to be considered.
We have seen that both the dominance and the difference approaches can be employed to explain variation in speech situations. In addition, the personality of the individual and the vitality of the group are also involved in the explanation of variability in language use. There is evidence that the vitality factor in the female subculture is increasing resulting in growing assertiveness. Given such a process, the significance of structures such as tag-questions and of behavioral patterns such as interruptions will diminish for a determination of the differentiation of men and women with respect to language use. Gender and status rather than gender or status will be the determinant categories.
Beattie, Geoffrey (1981): “Interruption in Conversational Interaction and its Relation to the Sex and Status of the Interactants”. Linguistics 19: 15-35.
Bernstein, Basil (1972): Studien zur sprachlichen Sozialisation. Düsseldorf.
Bock, Ute (1996): Frauensprache – Männersprache: Fakt oder Artefakt. Berlin.
Chambers, J.C. (1992): “Linguistic Correlates of Gender and Sex”. English World Wide 13: 173-218.
Coates, Jennifer (19932): Women, Men and Language. London.
Coates, Jennifer (ed) (1998): Language and Gender: A Reader. Oxford.
Coates, Jennifer/Cameron, Deborah. (eds.) (1989): Women in Their Speech Communities. London.
Dubois, Betty/Crouch, Isabel (1975): “The Question of Tag Questions in Women’s Speech: They Really Don’t Use More of Them”. Language in Society 4: 289-294.
Dubois, Betty/Crouch, Isabel (eds.) (1976): The Sociology of the Languages of American Women. San Antonio, TX.
Eakins, Barbara/Eakins, Gene: “Verbal Turn-Taking and Exchanges in Faculty Dialogue”. In: Dubois, Betty/Crouch, Isabel (eds.) (1976): The Sociology of the Languages of American Women. San Antonio, TX: 53-61.
Fasold, Ralph (1990): The Sociolinguistics of Language. Cambridge, MA.
Giles, Howard (ed.) (1977): Language, Ethnicity and Intergroup Relations. London.
Gumperz, John (ed.) (1982): Language and Social Identity. Oxford.
Holmes, Janet (1984): “Women’s Language: A Functional Approach”. General Linguistics 24/3: 149-178.
Holmes, Janet (1992): An Introduction to Sociolinguistics. London.
Holmes, Janet: “Women’s Talk: The Question of Sociolinguistci Universals”. In: Coates, Jennifer (ed.) (1998): Language and Gender: A Reader. Oxford: 461-483.
Kramarae, Cheris (1986): “A Feminist Critique of Sociolinguistics”. Journal of the Atlantic Provinces Linguistic Association 8: 1-22.
Kramarae, Cheris/Spender, Dale (eds.) (1992): The Knowledge Explosion. New York.
Lakoff, Robin (1972): “Language in Context”. Language 48/4: 907-927.
Lakoff, Robin: “The Social Context of Language Use”. (1973): Lecture delivered at the Linguistic Summer Institute. Ann Arbor.
Lakoff, Robin (1975): Language and Women’s Place. New York.
Martin, F.: “Some Subjective Aspects of Social Stratification”. In Glass, D. (ed.) (1954): Social Mobility in Great Britain. London.
Maltz, Daniel/Borker, Ruth: “A Cultural Approach to Male-Female Miscommunication”. In: Gumperz, John (ed.) (1982): Language and Social Identity. Oxford: 281-312. Also in: Coates, Jennifer (ed.) (1998): Language and Gender: A Reader. Oxford: 415-434.
Portz, Renate (1982): Sprachliche Variation und Spracheinstellungen bei Schulkindern und
Sacks, Harvey/Schegloff, Emanuel/Jefferson, Gail (1974): “A simplest Systematics for the Organization of Turn-Taking for Conversation”. Language 50: 696-735.
Sherman, Julia (1978): Sex-related Cognitive Differences: An Essay on Theory and Evidence. Springfield, IL.
Tajfel, Henri (1974): “Social Identity and Intergroup Behavior”. In: Social Science Information 13: 65-93.
Tannen, Deborah (1990): You Just Don’tUnderstand. New York.
Thorne, Barrie/Henley, Nancy (eds.) (1975): Language and Sex: Difference and Dominance. Rowley, MA.
Thorne, Barry/Kramarae, Cheris/ Henley, Nancy (eds.) (1983): Language, Gender and Society. Rowley, MA.
Torres, Lourdes: “Women and Language. From Sex Differences to Power Dynamics”. In: Kramarae, Cheris/Spender, Dale (eds.) (1992): The Knowledge Explosion. New York: 281-290.
Trömel-Plötz, Senta: “Selling the Apolitical”. In: Coates, Jennifer (ed.) (1998): Language and Gender: A Reader. Oxford: 446-458.
Trudgill, Peter (1972):”Sex, Covert Prestige and Linguistic Change in the Urban British English of East Anglia”. Language in Society 1: 179-195.
Trudgill, Peter (1978): Sociolinguistic Patterns in British English. London.
Trudgill, Peter: “Sex and Covert Prestige”. In: Coates, Jennifer (ed) (1998): Language and Gender: A Reader. Oxford: 21-28.
West, Candace/Zimmerman, Don: “Small Insults: A Study of Interruptions in Cross-Sex Conversations between Unaquainted Persons”. In: Thorne, Barry/Kramarae, Cheris/ Henley, Nancy (eds.) (1983): Language, Gender and Society. Rowley, MA: 127-150.
West, Candace: “When the Doctor is a ‘Lady’: Power, Status and Gender in Physician-Patient Encounters”. In: Coates, Jennifer (ed) (1998): Language and Gender: A Reader. Oxford: 396-412.
West, Candace/Zimmerman, Don: “Doing Gender”. In: Lorber, Judith/Farrell, Susan (eds.) (1991): The Social Construction of Gender. London: 13-37.
Zimmerman, Don/West, Candice: “Sex Roles, Interruptions and Silences in Conversations”. In: Thorne, Barrie/Henley, Nancy (eds.) (1975): Language and Sex: Difference and Dominance. Rowley; MA: 105-129. Sex Roles, Interruptions and Silences in Conversations”. In: Thorne, Barrie/Henley, Nancy (eds.) (1975): Language and Sex: Difference and Dominance. Rowley; MA: 105-129.